{"Affiliation":[{"label":"Affiliation","value":"Arts, Faculty of","attrs":{"lang":"en","ns":"http:\/\/vivoweb.org\/ontology\/core#departmentOrSchool","classmap":"vivo:EducationalProcess","property":"vivo:departmentOrSchool"},"iri":"http:\/\/vivoweb.org\/ontology\/core#departmentOrSchool","explain":"VIVO-ISF Ontology V1.6 Property; The department or school name within institution; Not intended to be an institution name."},{"label":"Affiliation","value":"Linguistics, Department of","attrs":{"lang":"en","ns":"http:\/\/vivoweb.org\/ontology\/core#departmentOrSchool","classmap":"vivo:EducationalProcess","property":"vivo:departmentOrSchool"},"iri":"http:\/\/vivoweb.org\/ontology\/core#departmentOrSchool","explain":"VIVO-ISF Ontology V1.6 Property; The department or school name within institution; Not intended to be an institution name."}],"AggregatedSourceRepository":[{"label":"Aggregated Source Repository","value":"DSpace","attrs":{"lang":"en","ns":"http:\/\/www.europeana.eu\/schemas\/edm\/dataProvider","classmap":"ore:Aggregation","property":"edm:dataProvider"},"iri":"http:\/\/www.europeana.eu\/schemas\/edm\/dataProvider","explain":"A Europeana Data Model Property; The name or identifier of the organization who contributes data indirectly to an aggregation service (e.g. Europeana)"}],"Campus":[{"label":"Campus","value":"UBCV","attrs":{"lang":"en","ns":"https:\/\/open.library.ubc.ca\/terms#degreeCampus","classmap":"oc:ThesisDescription","property":"oc:degreeCampus"},"iri":"https:\/\/open.library.ubc.ca\/terms#degreeCampus","explain":"UBC Open Collections Metadata Components; Local Field; Identifies the name of the campus from which the graduate completed their degree."}],"Creator":[{"label":"Creator","value":"M\u00fchlbauer, Jeffrey Thomas","attrs":{"lang":"en","ns":"http:\/\/purl.org\/dc\/terms\/creator","classmap":"dpla:SourceResource","property":"dcterms:creator"},"iri":"http:\/\/purl.org\/dc\/terms\/creator","explain":"A Dublin Core Terms Property; An entity primarily responsible for making the resource.; Examples of a Contributor include a person, an organization, or a service."}],"DateAvailable":[{"label":"Date Available","value":"2008-06-24T23:52:43Z","attrs":{"lang":"en","ns":"http:\/\/purl.org\/dc\/terms\/issued","classmap":"edm:WebResource","property":"dcterms:issued"},"iri":"http:\/\/purl.org\/dc\/terms\/issued","explain":"A Dublin Core Terms Property; Date of formal issuance (e.g., publication) of the resource."}],"DateIssued":[{"label":"Date Issued","value":"2008","attrs":{"lang":"en","ns":"http:\/\/purl.org\/dc\/terms\/issued","classmap":"oc:SourceResource","property":"dcterms:issued"},"iri":"http:\/\/purl.org\/dc\/terms\/issued","explain":"A Dublin Core Terms Property; Date of formal issuance (e.g., publication) of the resource."}],"Degree":[{"label":"Degree (Theses)","value":"Doctor of Philosophy - PhD","attrs":{"lang":"en","ns":"http:\/\/vivoweb.org\/ontology\/core#relatedDegree","classmap":"vivo:ThesisDegree","property":"vivo:relatedDegree"},"iri":"http:\/\/vivoweb.org\/ontology\/core#relatedDegree","explain":"VIVO-ISF Ontology V1.6 Property; The thesis degree; Extended Property specified by UBC, as per https:\/\/wiki.duraspace.org\/display\/VIVO\/Ontology+Editor%27s+Guide"}],"DegreeGrantor":[{"label":"Degree Grantor","value":"University of British Columbia","attrs":{"lang":"en","ns":"https:\/\/open.library.ubc.ca\/terms#degreeGrantor","classmap":"oc:ThesisDescription","property":"oc:degreeGrantor"},"iri":"https:\/\/open.library.ubc.ca\/terms#degreeGrantor","explain":"UBC Open Collections Metadata Components; Local Field; Indicates the institution where thesis was granted."}],"Description":[{"label":"Description","value":"This thesis considers the reference system of Plains Cree, an Algonquian language spoken in Canada. I argue that the referential system of this language can be understood as coding distinctions in extentionality; it distinguishes between referents that possess perspectives (\u2018intentional\u2019) and referents that do not (\u2018extentional\u2019). With respect to perspectival possession, Plains Cree distinguishes four referential classes: (i) inherently extentional \u201cInanimate\u201d referents, (ii) contextually extentional \u201cObviative\u201d referents, (iii) contextually \nintentional \u201cProximate\u201d referents, and (iv) unspecified \u201cAnimate\u201d referents. I then show \nthat the referential class \u201cObviative\u201d is decompositional; it is constructed out of \ncomponents that code referential dependency, which is the confluence of structural \nordering and perspectival embedding. Finally, I consider the methodological issues raised by the study of referential types, showing how different data-collection methods interact with the semantics of perspectival possession.","attrs":{"lang":"en","ns":"http:\/\/purl.org\/dc\/terms\/description","classmap":"dpla:SourceResource","property":"dcterms:description"},"iri":"http:\/\/purl.org\/dc\/terms\/description","explain":"A Dublin Core Terms Property; An account of the resource.; Description may include but is not limited to: an abstract, a table of contents, a graphical representation, or a free-text account of the resource."}],"DigitalResourceOriginalRecord":[{"label":"Digital Resource Original Record","value":"https:\/\/circle.library.ubc.ca\/rest\/handle\/2429\/940?expand=metadata","attrs":{"lang":"en","ns":"http:\/\/www.europeana.eu\/schemas\/edm\/aggregatedCHO","classmap":"ore:Aggregation","property":"edm:aggregatedCHO"},"iri":"http:\/\/www.europeana.eu\/schemas\/edm\/aggregatedCHO","explain":"A Europeana Data Model Property; The identifier of the source object, e.g. the Mona Lisa itself. This could be a full linked open date URI or an internal identifier"}],"Extent":[{"label":"Extent","value":"5755216 bytes","attrs":{"lang":"en","ns":"http:\/\/purl.org\/dc\/terms\/extent","classmap":"dpla:SourceResource","property":"dcterms:extent"},"iri":"http:\/\/purl.org\/dc\/terms\/extent","explain":"A Dublin Core Terms Property; The size or duration of the resource."}],"FileFormat":[{"label":"File Format","value":"application\/pdf","attrs":{"lang":"en","ns":"http:\/\/purl.org\/dc\/elements\/1.1\/format","classmap":"edm:WebResource","property":"dc:format"},"iri":"http:\/\/purl.org\/dc\/elements\/1.1\/format","explain":"A Dublin Core Elements Property; The file format, physical medium, or dimensions of the resource.; Examples of dimensions include size and duration. Recommended best practice is to use a controlled vocabulary such as the list of Internet Media Types [MIME]."}],"FullText":[{"label":"Full Text","value":" k\u00e2-y\u00f4sk\u00e2tahk \u00f4ma n\u00eahiyaw\u00eawin: THE REPRESENTATION OF INTENTIONALITY IN PLAINS CREE by JEFFREY THOMAS M\u00dcHLBAUER B.A., The University of Wisconsin. 2001 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Linguistics) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) June 2008 \u00a9 Jeffrey Thomas M\u00fchlbauer, 2008 ii Abstract This thesis considers the reference system of Plains Cree, an Algonquian language spoken in Canada. I argue that the referential system of this language can be understood as coding distinctions in extentionality; it distinguishes between referents that possess perspectives (\u2018intentional\u2019) and referents that do not (\u2018extentional\u2019). With respect to perspectival possession, Plains Cree distinguishes four referential classes: (i) inherently extentional \u201cInanimate\u201d referents, (ii) contextually extentional \u201cObviative\u201d referents, (iii) contextually intentional \u201cProximate\u201d referents, and (iv) unspecified \u201cAnimate\u201d referents. I then show that the referential class \u201cObviative\u201d is decompositional; it is constructed out of components that code referential dependency, which is the confluence of structural ordering and perspectival embedding. Finally, I consider the methodological issues raised by the study of referential types, showing how different data-collection methods interact with the semantics of perspectival possession. iii Table of Contents Abstract ...........................................................................................................................................ii Table of Contents ...........................................................................................................................iii List of Tables ..................................................................................................................................x List of Symbols and Abbreviations ..............................................................................................xii Acknowledgements .....................................................................................................................xiii Citation ..........................................................................................................................................xv 1. Introduction: Making truth happen ...............................................................................1 1.1. Introduction: k\u00e2-y\u00f4sk\u00e2tahk \u00f4ma n\u00eahiyaw\u00eawin..............................................................1 1.2. How does \u201ctruth happen\u201d in Plains Cree? ....................................................................2 1.2.1 The \u201cCalvin and Hobbes\u201d problem in Plains Cree ........................................3 1.2.2 The \u201c6-Degrees of Separation\u201d problem in Plains Cree ............................\u20265 1.3. Proposal: Individuals related to perspectives ......................................................\u2026....6 1.3.1. Perspective possession .............................................................................\u2026.6 1.3.2. Discourse Representation Theory with perspectives ................................\u20268 1.3.3. Proposal: Four kinds of referents in Plains Cree ....................................... 11 1.3.4. Proposal: Constructing obviation via referential dependency ................\u202616 1.4. Relation to previous work .....................................................................................\u2026.17 1.4.1. Models of subjective meaning ....................................................................17 1.4.2. Perspectives vs. possible worlds .................................................................19 1.5. The Plains Cree language ...........................................................................................21 1.5.1. Sources ..................................................................................................\u202621 1.5.2. Orthography ................................................................................................23 1.6. Outline of the thesis ...................................................................................................24 2. The Form, Content, and Context of Animacy...............................................................26 2.1. Proposal \u2026................................................................................................................26 2.2. Previous accounts .....................................................................................................30 2.2.1. Syntactic accounts of Plains Cree ...............................................................30 2.2.2. \u201cInanimate\u201d as semantically specified ........................................................33 2.2.3. \u201cInanimate\u201d as extentional ..........................................................................34 iv 2.2.4. Animacy is syntactically determined ..........................................................37 2.2.5. The data set........................................................................................\u2026......38 2.3 The form of animacy: Distributed throughout the syntax .........................................38 2.3.1. Animacy is not specified in the nominal......................................................42 2.3.2. Animacy contrasts are not suppletive..........................................................44 2.3.3. The necessity of homophony ......................................................................46 2.3.4. Animacy coding does not determine the distribution of bare nouns...........49 2.3.5. Animacy coding is not asymmetric..............................................................52 2.3.51 Animacy without nominals.............................................................53 2.3.52 Animacy without verbs...................................................................56 2.3.6. Conclusion: Animacy is distributed.............................................................58 2.4. The content of animacy: \u201cInanimates\u201d are inherently extentional .............................59 2.4.1. Asymmetries in the interpretation of the two classes .................................62 2.4.2. Abstract nominals are always \u201cInanimate\u201d..................................................64 2.4.3. Extentionality in the verb system: \u201cInanimates\u201d can\u2019t think, speak, or feel......................66 2.4.31. Expletive arguments must be \u201cinanimate\u201d ...................................67 2.4.32. Psych verbs prohibit \u201cinanimate\u201d arguments................................69 2.4.33. Reflexive verbs prohibit \u201cinanimate\u201d arguments .........................72 2.4.4. Animates are unspecified ..........................................................................73 2.4.5. Only the \u201cinanimate\u201d form has specification ............................................75 2.4.6. \u201cInanimate\u201d nominals are inherently extentional .....................................79 2.5. The context of animacy: Relativization to individual perspectives............................81 2.5.1. \u201cInanimates\u201d as preconditions on embeddings ...........................................82 2.5.2. Relativizing to the Speaker: Changes in belief ...........................................84 2.5.3. Relativization to the Hearer: Effects of common ground ...........................86 2.5.4. Relativizing to third persons: Perspective shifts .........................................90 2.6. Conclusion .................................................................................................................97 3. The Form, Content, and Context of Obviation.............................................................98 3.1. Proposal: Obviation as contextual extentionality ......................................................98 3.2. Previous accounts .....................................................................................................102 3.2.1. Obviation as discourse, not argument structure.........................................102 v 3.2.2. Obviation as a construct, not a primitive...................................................105 3.2.3. \u201cObviative\u201d is specified, \u201cproximate\u201d is contextually determined............105 3.2.4. The meaning of obviation.........................................................................107 3.2.5. The data set ...............................................................................................108 3.3. The form of obviation: no dedicated \u201cobviative\u201d marking .....................................108 3.3.1. The nominal suffix \u2013a................................................................................110 3.3.11 The traditional view: \u2013a marks \u201cobviative\u201d.................................111 3.3.12 The current analysis: \u2013a marks extentional referents...................111 3.3.13 Modelling the \u201cobviative\u201d effects of \u2013a.......................................114 3.3.2. The demonstrative suffix \u2013ihi....................................................................115 3.3.21 The traditional view: \u2013ihi marks \u201cobviative\u201d...............................117 3.3.22 The current analysis: \u2013ihi marks extentional referents.................118 3.3.23 Modelling the \u201cobviative\u201d effects of \u2013ihi.....................................120 3.3.3. The theme sign \u2013\u00ea\u2013....................................................................................120 3.3.31 The traditional view: \u2013\u00ea\u2013 marks \u201cthird person interactions\u201d........121 3.3.32 The current analysis: \u2013\u00ea\u2013 marks extentional objects....................122 3.3.33 Modelling the \u201cobviative\u201d effects of \u2013\u00ea\u2013......................................126 3.3.4. The predicate suffix \u2013im\u2013..........................................................................127 3.3.41 The traditional view: \u2013im\u2013 marks \u201cobviative\u201d objects.................128 3.3.42 The current analysis: \u2013im\u2013 marks argument disjunction..............128 3.3.43 Modelling the \u201cobviative\u201d effects of \u2013im\u2013....................................133 3.3.5. The predicate suffix \u2013yi\u2013...........................................................................137 3.3.51. The traditional view: \u2013yi\u2013 marks \u201cobviative\u201d arguments ...........139 3.3.52. The current analysis: \u2013yi\u2013 marks disjoint subjects .....................139 3.3.53. Modelling the \u201cobviative\u201d effects of \u2013yi\u2013....................................144 3.3.6. Summary: The derivative nature of obviation.........................................148 3.4. The content of obviation: contextual extentionality ................................................148 3.4.1. Intentional verbs restricted \u201cobviative\u201d reference ....................................154 3.4.11. \u201cObviative\u201d referents are restricted with psych verbs ................155 3.4.12. \u201cObviative\u201d referents are restricted with speaking verbs ...........158 3.4.2. \u201cObviative\u201d referents lack awareness .......................................................161 3.4.21. If contextual absence, referent is \u201cobviative\u201d .............................162 3.4.22. An \u201cobviative\u201d referent that is present is unaware .....................166 vi 3.4.3. The emergence of \u201cproximate\u201d via contrast with \u201cobviative\u201d ..................168 3.4.31. \u201cObviative\u201d as context for \u201cproximate\u201d.......................................169 3.4.32. Paradigmatic contrast of \u201canimate\u201d with \u201cobviative\u201d .................172 3.4.4. Summary: \u201cObviative\u201d as contextual extentionality ................................173 3.5. The context of obviation: Speaker knowledge of intentions ...................................174 3.5.1. \u201cObviative\u201d perspectives force Speaker dissociation ...............................178 3.5.2. \u201cObviative\u201d perspectives force invented \u201cproximate\u201d ..............................179 3.5.3. Obviation patterns with indirect evidentiality ...........................................182 3.5.31. \u201cObviative\u201d psych arguments and indirect knowledge ..............186 3.5.32. \u201cObviative\u201d speakers and indirect knowledge ............................188 3.5.33. \u201cObviative\u201d referents and the indirect evidential \u00easa .................196 3.6. Conclusion ...............................................................................................................198 4. Referential dependency in Plains Cree........................................................................200 4.1. A structural and a semantic condition on referential dependency ...........................200 4.2. Previous work ..........................................................................................................204 4.2.1. Structural conditions on anaphora.............................................................204 4.2.2. Clausal dependency and referential dependency.......................................206 4.2.3. Possession and grammatical obviation .....................................................210 4.2.4. The data set................................................................................................211 4.3. Cross-predicate dependencies and obviation ...........................................................212 4.3.1. Nominal ordering and referential dependency ..........................................213 4.3.11. Nominal orderings in texts ..........................................................215 4.3.12. Elicitation and nominal-nominal ordering ..................................217 4.3.2. The suffix -yi- constructs referential dependency .....................................223 4.3.21. Structural conditions on -yi- ........................................................225 4.3.211. \u2013yi\u2013 is c-commanded by its antecedent ........................226 4.3.212. \u2013yi\u2013 is linearly preceded by its antecedent....................227 4.3.22. Semantic conditions on \u2013yi\u2013 .......................................................231 4.3.221. The antecedent of \u2013yi\u2013 must be in previous discourse..232 4.3.222. The antecedent of \u2013yi\u2013 must be animate .......................233 4.3.223. The antecedent of \u2013yi\u2013 must have a perspective ..........233 4.3.224. Propositional attitude verbs force \u2013yi\u2013..........................236 vii 4.3.23. Conclusion: -yi- constructs referential dependency ....................238 4.4. Predicate-internal dependencies and obviation ........................................................239 4.4.1 Possession as referential dependency ........................................................241 4.4.11. Structural subordination of possessums to possessors ................242 4.4.12. Perspectival embedding with possession ....................................245 4.4.121. \u201cInanimate\u201d referents cannot be possessors ..................246 4.4.122. Possessors are always existential ..................................247 4.4.123. Possession blocks attitudes towards possessum .......... 249 4.4.124. Possession and speech act participants .........................251 4.4.125. Changing possession changes the Speaker....................254 4.4.13. Conclusion: Possession is referential dependence ......................258 4.4.2. Theme signs and referential dependency...................................................259 4.4.21. The direct theme sign \u2013\u00ea\u2013 constructs referential dependency.....262 4.4.211. Structural conditions on \u2013\u00ea\u2013 .........................................263 4.4.212. Semantic conditions on \u2013\u00ea\u2013 ..........................................265 4.4.2121. \u2013\u00ea\u2013 occurs with perspectivally-embedded objects .....................................................266 4.4.2122. \u2013\u00ea\u2013 cannot occur when the subject is not a perspective holder ................................272 4.4.213. Conclusion: \u2013\u00ea\u2013 constructs referential dependency .....276 4.4.22. The inverse theme \u2013ikw constructs referential dependency ........277 4.4.221. Structural conditions on \u2013ikw .......................................279 4.4.2211. Evidence for IP position of theme sign \u2013ikw...279 4.4.2212. Evidence for structural relations of \u2013ikw ........282 4.4.222. Semantic conditions on \u2013ikw ........................................291 4.4.2221. Semantic restrictions on topic op ....................292 4.4.2222. Perspective-less properties of \u2013ikw .................296 4.4.223. Conclusion: \u2013ikw as referential dependency .................301 4.5. Conclusion ...............................................................................................................301 5. Eliciting obviation and animacy .................................................................................303 5.1. The place of methodology in linguistic analysis.......................................................303 5.2. Fieldwork in linguistics.............................................................................................303 viii 5.3. Obviation and animacy data fluctuates in elicitation ...............................................304 5.4. A taxonomy of elicitation tasks and their effects on obviation and animacy ..........307 5.4.1. Translation tasks .......................................................................................307 5.4.11. Confounds in translation: Non-correlation and priming .............308 5.4.12. Obviation in translation tasks .....................................................309 5.4.13. Animacy in translation tasks .......................................................310 5.4.2. Judgment tasks ..........................................................................................312 5.4.21. Confounds in judgment tasks: Metalinguistics ...........................314 5.4.22. Judgment tasks and obviation .....................................................315 5.4.23. Animacy in judgment tasks .........................................................316 5.4.3. Supporting judgment tasks: Utterance-in-context tasks ...........................319 5.4.31. Obviation in utterance-in-context tasks ......................................321 5.4.32. Animacy in utterance-in-context tasks ........................................323 5.4.4. Correction tasks .......................................................................................324 5.4.41. Confounds in correction tasks: Un-noticed corrections ..............325 5.4.42. Obviation in correction tasks ......................................................328 5.4.43. Animacy in correction tasks ........................................................329 5.4.5. Analysis tasks ...........................................................................................331 5.4.51. Confounds: Decontextualizing forms, pseudo-linguistics ..........331 5.4.52. Obviation in analysis tasks .........................................................332 5.4.53. Animacy in analysis tasks ...........................................................333 5.5. Elicitation and impoverished context ......................................................................334 5.5.1. The proliferation of simple animates in elicitation ...................................335 5.5.2. Obtaining \u201cobviative\u201d forms in elicitation.................................................338 5.5.3. Contextless obviation = \u201cInanimate\u201d ........................................................339 5.6. The consultant as linguist .........................................................................................341 5.6.1. The \u201cexperimentalist\u201d position: Consultant as data bank .........................341 5.6.2. The \u201ccooperative\u201d position ......................................................................342 5.6.3. The two types of fieldwork and obviation data ........................................343 5.7. Conclusion: Elicitation as performance, not competence ........................................346 6. Conclusion and Further Research................................................................................347 6.1. Conclusion: Constructing intentionality in Plains Cree ...........................................347 ix 6.2. Approaches to \u2013a via the verbal suffix \u2013i : Plurality and obviation ........................348 6.2.1. The \u201cplural\u201d reading .................................................................................350 6.2.2. The habitual reading .................................................................................350 6.2.3. The irrealis reading ...................................................................................351 6.2.4. Conclusion ................................................................................................352 6.3. \u2013yi\u2013 across Cree........................................................................................................353 6.3.1. All Cree languages have a reflex of \u2013yi\u2013...................................................354 6.3.2. \u2013yi\u2013 occurs with \u201cobviative\u201d possessors....................................................354 6.3.3. \u2013yi\u2013 occurs with intransitive subjects........................................................355 6.3.4. \u2013yi\u2013 occurs with transitive subjects...........................................................355 6.3.5. \u2013yi\u2013 is insensitive to animacy....................................................................357 6.4. A comparison of the Plains Cree system to Athabaskan..........................................359 6.4.1. Athabaskan yi-\/bi- are pronominals, Plains Cree\u2019s are not........................362 6.4.2. Plains Cree \u2013yi\u2013 and Athabaskan yi- code disjoint reference\u2026\u2026............364 6.4.3. Plains Cree \u2013yi\u2013 and Athabaskan yi- differ in transitivity.........................366 6.4.4. A and A\u2019 dependencies. ............................................................................367 Glossary ....................................................................................................................................370 References .................................................................................................................................377 x List of Tables Table 1.1. Two conditions on referential dependency ..................................................................17 Table 1.2. The forms of obviation and their dependencies ...........................................................17 Table 2.1. Summary of animacy coding for verb, demonstrative, and nominal .......................... 39 Table 2.2. Summary of animacy shifting for the Severed Head story ......................................... 43 Table 2.3. Forms commonly found in both animacy classes ....................................................... 46 Table 2.4. Summary of predictions for three hypotheses ............................................................ 59 Table 2.5. Semantic gaps among verbs ........................................................................................ 70 Table 2.6. A summary of \u201cinanimate\u201d\/\u201canimate\u201d shifting for the severed head story ................ 91 Table 3.1. No dedicated obviation forms ..................................................................................... 99 Table 3.2. \u201cInanimate\u201d plural and \u201cobviative\u201d across Algonquian ........................................... 113 Table 3.3. Canonical demonstrative organization in Plains Cree ............................................. 115 Table 3.4. Demonstrative organization of S2 ............................................................................ 116 Table 3.5. \u201cInanimate\u201d plural and \u201cobviative\u201d demonstratives across Algonquian .................. 119 Table 3.6. Independent order distribution of -\u00ea- ........................................................................ 125 Table 3.7. No -\u00ea- in conjunct order ............................................................................................ 125 Table 3.8. Distribution of -im- with \u201cobviative\u201d arguments ...................................................... 127 Table 3.9. Distribution of \u2013yi\u2013 with \u201cobviative\u201d arguments ...................................................... 139 Table 3.10. Distribution of \u2013yi\u2013 with \u201cinanimate\u201d arguments ................................................... 140 Table 3.11. \u201cObviative\u201d forms and their patterning .................................................................. 148 Table 3.12. Summary for psych verbs built with \u2013\u00eayim\/\u2013\u00eayiht\/\u2013\u00eayimo ..................................... 157 Table 3.13. Summary of other psych verb forms ........................................................................157 Table 3.14. Summary of preverb forms ......................................................................................157 Table 3.15. A Summary of psych verbs in 3 texts ......................................................................158 Table 3.16. Speaking verbs and obviation ..................................................................................159 Table 3.17. Intransitive itw\u00ea- and obviation .............................................................................. 160 Table 3.18. \u201cObviative\u201d referents and speaking in a text ...........................................................160 Table 3.19. \u201cObviative\u201d reference shifts when topic event shifts ...............................................163 Table 3.20. Obviated animal in a story .......................................................................................170 Table 3.21. Chart of obviation for modified re-telling .............................................................. 171 Table 3.22. Categorization of \u201cobviative\u201d with psych verbs ......................................................186 Table 3.23. Transitive it- and obviation ......................................................................................190 xi Table 3.24. Transitive w\u00eehtamaw- and obviation ........................................................................191 Table 3.25. Transitive speaking and obviation in a text ............................................................ 192 Table 3.26. \u201cObviative\u201d referents speaking in a text ..................................................................194 Table 3.27. Obviatiation in an \u00easa evidential span .....................................................................197 Table 4.1. \u201cObviative\u201d constructions meet the conditions of referential dependency ................204 Table 4.2. Nominal-nominal ordering meets the conditions of referential dependency .............214 Table 4.3. \u201cProximate\u201d referring nouns precede \u201cobviatives\u201d in texts .......................................216 Table 4.4. Nominal-nominal ordering patterns in elicitation ......................................................218 Table 4.5. The suffix \u2013yi\u2013 meets the conditions of referential dependency ...............................225 Table 4.6. Possession meets the conditions of referential dependency ......................................242 Table 4.7. The inverse meets the conditions of referential dependency .....................................278 Table 4.8. Restrictions on the identity of topic op in Plains Cree ..............................................292 Table 4.9. Restrictions on the identity of the argument introduced by \u2013ikw ..............................296 Table 4.10. Topic operator and Inverse pro are opposites in their properties ............................301 Table 6.1. A summary of the occurrences of \u2013i ..........................................................................349 Table 6.2. A summary of contexts for plurality ..........................................................................353 Table 6.3. Navajo, Dogrib and D\u00ebne S\u0173\u0142in\u00e9 pronominal affixes ................................................359 Table 6.4. Comparing Plains Cree \u2013ikw and \u2013yi\u2013 to Athabascan yi- ..........................................361 Table G1. Summary of verbal morphology ................................................................................374 xii List of Symbols and Abbreviations x \uf0e0 y = if x, then y x \u2194 y = iff (if and only if) x, then y x \u2227 y = x and y \u2248 = approximately \u2260 = does not equal \u2200 = universal quantifier \u2203 = existential quantifier \u03bb = lambda operator \u03c8 = perspective 0 = inanimate agreement 1 = first person 1>3 = first person acting on third person 2 = second person 3 = third person 3>1 = third person acting on first person AI = animate intransitive AIT = animate intransitive-transitive ACT = actor AN = animate APPLIC = applicative C1 = changed conjunct 1 C2 = changed conjunct 2 CP = clause phrase DEMP = demonstrative phrase DIM = diminutive DIR = direct DP = determiner phrase DRS = discourse representation structure DRT = discourse representation theory DS = disjoint subject DSJ = disjoint marker DST = distal EP = epenthetic EXCL = exclamation EXT = extentional EXTP = extentional phrase FUT = future GEN = generic HABIT = habitual HES = hesitation II = inanimate intransitive IN = inanimate INCH = inchoative INTERJ = interjection INT = intentional INV = inverse IP = inflectional phrase LOC = locative LP = local person NEG = negation NOM = nominalizer NP = noun phrase nP = little noun phrase NUMP = number phrase OBV = obviative PERSP = person phrase PL = plural PREV = previous PROX = proximate PRX = proximal R = relation RECIP = reciprocal RED = reduplication RESUM = resumptive RFLX = reflexive RR = relative root SG = singular STAT = stative TA = transitive animate TI = transitive inanimate UP = unknown possessor VP = verb phrase vP = little verb phrase x = variable XP = unidentified phrase XT = extentional y = variable xiii Acknowledgements First and foremost, I thank the Plains Cree and M\u00e9tis people who have contributed their linguistic knowledge to this thesis. This includes Wally Aw\u00e2sis, who showed up to our first session in an Cleveland Indians jacket (\u201cI\u2019m an Indian \u2013 they made the jacket just for me\u201d), Rita Daniels, whose kindness won\u2019t be forgotten, Joseph Deschamps, who has repeatedly provided crucial insights with his knowledge of Cree culture and language, Josephine Small, whose linguistic flexibility and patience has provided many crucial data points, and Dorothy Thunder, who gave me many good questions to ask about the language. Toni Cardinal\u2019s exceptional linguistic abilities were crucial to the construction of the generalizations that are the backbone of this work. Repeatedly, and over many years of work, she turned her intuitive understanding of the Plains Cree language into clear, concise descriptions that far exceeded any analytic work done on these aspects of her language. In addition, I would also like to thank speakers and community members, including Beverly Crier, Judy Louis, Roy Louis, Barbara MacLeod, Tom McCallum, Jean Okim\u00e2sis, Maryann Palmer, Solomon Ratt, Henry Rain, Wayne Roan, Blanche Steinhauer, and Simon Threefingers. I know it is hard to listen to ignorance speaking with unearned authority, and I appreciate your forbearance and compassion. kin\u00e2naskomitin\u00e2w\u00e2w. My supervisor, Rose-Marie D\u00e9chaine, is the reason I survived UBC academics, the expert hand behind all my successful grants and any papers you liked, and the main reason my thesis is at all coherently written (You readers think this is bad \u2013 you should have seen Draft #2!). From the admissions process until I filed this thesis, I cannot think of any stage that she was not involved in. I am greatly indebted to her for her time, her energy, and her ideas. The other two members of my committee, Guy Carden and Hotze Rullmann have also been of central importance to this work. Guy is one of the great linguists, and his sharp questions and encouragement have been much enjoyed. I am grateful that he took time for me. Hotze has suffered my semantic excesses (and their ensuing chaos) more cruelly than anyone, and any clarity in the final semantic discussion is to his credit. He has given me significant amounts of his time, carefully reading many drafts and patiently trying to pull me from the mud. I would like to thank the four additional members of my examining committee: Andrew Irvine, Katarzyna Jaszczolt, Leslie Saxon, and Geoff Williams for their voluntary and insightful contributions to this work. I would especially like to thank Kasia and Leslie for going through my thesis in such fine detail, and the final outcome has been much improved by their ideas. In addition, Tamina Stephenson was a silent member of my examining committee, and her open mind and flexibility provided me with much help when it was most needed. H.C. Wolfart has suffered through both my theoretical and editorial contortions with much patience (Chris: Please don\u2019t look at the references to this thesis). His scholarship, fueled by a sincere interest in the language, has provided an excellent role-model for what linguistic work on a language ought to be and how it ought to proceed, and his academic manner represents the best standard of what I had hoped Academia would be like. I owe Monica Macaulay a great thank you for introducing me to Algonquian languages. Out of her own good will and entirely on her own time, she spent many hours working through texts, discussing Bloomfield\u2019s grammar, and going through teaching materials, as well as literally driving me up to the reservation. She also gave me much advice about graduate school, and introduced me to fieldwork methodology via Menominee and isiZulu. Thanks to those who shared their language knowledge with me, including Ol\u00e1di\u00edp\u00f2 Aj\u00edb\u00f3y\u00e8 (Yor\u00f9b\u00e1), Beatrice and Melinda Bullshields (Blackfoot), Adriano Barbosa (Brazilian Portuguese), Seok Koon Chin (Mandarin, Cantonese, Hokkien), Kerim Demirci (Turkish), Peter Jacobs (Squamish), Bekisizwe Ndimande (isiZulu), and Lena Russell (Blackfoot). xiv This work has been supported by the Jacobs Fund, Phillips Fund, UBC, and SSHRC. Without Edna Dharmaratne, nothing would have gotten done, I would have never gotten any money, and I would have starved in the rain. All that, and she\u2019s a nice person. Thanks to my teachers at UBC, including Gunnar Hansson, Lisa Matthewson, Doug Pulleyblank, and Martina Wiltschko. Martina\u2019s teaching helped clarify a great deal of syntactic theory to me, and Lisa\u2019s semantic instruction and advising has been of great assistance. Thanks also to the Algonquian scholarly community, including Eleanor Blain, Phil Branigan, Amy Dahlstrom, Inge Genee, Ives Goddard, Marguerite MacKenzie, Jean Okim\u00e2sis, Arden Ogg, David Pentland, Marie Odile-Junker, Nicole Rosen, and Arok Wolvengrey. They have sat through my talks, shared their work, offered me help, and have still somehow not given up hope. I am fortunate to have worked with so many talented people. I would like to thank the scholars of ancient languages at the University of Wisconsin, Jack McKeown, Victoria Pagan, Barry Powell, Richard Ringler, and Andrew Wolpert, who gave me a clear picture of the first and best model of linguistic scholarship, and much laughter. Thanks also to my linguistics teachers at the University of Wisconsin, including Marie- Helene C\u00f4te, M\u00fcrvet Enc, Yafei Li, Matt Pearson, and Manindra Verma. In particular, I have vivid memories of Matt\u2019s brutal tests and M\u00fcrvet\u2019s fantastic rants (and fashion sense!). Thanks to my teachers at the University of Wisconsin extension in Marathon County, Harlan Grinde, Theo Koupelis, Keith Montgomery, and Peter Okray, who spent a great deal of time and energy teaching me, and introduced me to ideas that still inform my perspective. Thanks to the graduate student group for putting up with me: Ol\u00e1di\u00edp\u00f2 Aj\u00edb\u00f3y\u00e8, Solveiga Armoskaite, Leszek Barczak, Fiona Campbell, Mario Chavez-Peon, Yunhee Chung, Ramona McDowell, Atsushi Fujimori, Diana Gibraiel, Carrie Gillon, Fusheini Hudu, Yoko Ikegami, Peter Jacobs, Masaru Kiyota, Karsten Koch, John Lyon, Calisto Mudzingwa, Nahal Namdaran, Jeremy Perkins, Sugunya Ruangjaroon, Scott Shank, Tanya Slavin, Dominique Quis, Ian Wilson, Rachel Wojdak, and Florence Woo. I share your pain; I\u2019m either glad you\u2019re done or hope it will be over soon. Ryan Waldie told me about fairies in linguistics and botanical varieties in Vancouver, while also fixing my busted semantics. Better in a gutter than on a pedestal. Chin Seok Koon \u79e6\u6dd1\u541b has helped the teacher of thieves (\u8d3c\u592b) out of many of his worst scrapes, and has gracefully endured swashbuckling and barbarity from the Western Pirates. Many people showed me hospitality and kindness through these years, including Solveiga Armoskaite and family, Joe and Nella Cook, Rita Daniels, Denis D\u00e9chaine and family, Lucienne D\u00e9chaine, Rose-Marie D\u00e9chaine, Joseph Deschamps, the Demirci family, Tom Nook, Poncho, Paola Quintanar, Totakeke, and Juliana Wolfart. For still being willing to admit my acquaintance after 6 years of insanity, I would like to thank Kerim Demirci and David Tang. Kerim provided me with the only real break I had during this work, and David has somehow managed to eat most of my ice cream. I owe my interests and abilities to my parents, Gerald and Suzanne Muehlbauer, who have steadily tried to support my intellectual interests since I was in three-cornered pants. Topics have shifted from dinosaurs to languages, but I\u2019m not sure much else has changed. Thanks also to the rest of my family, including Alma Johnson, my grandparents Robert and Elizabeth Muehlbauer, Kathy Ritenour, who has lavished attention on me all my life, Jane \u201cPufferfish\u201d Ritenour, Ned and Debbie Wicker (and feline and canine housemates), who provided much encouragement over many years, and the Muehlbauer and Klann clans. Thanks for not being academics, and for trying to understand why on earth someone would do such a thing. (Answer: I still don\u2019t know!) Finally, there\u2019s Clare. You never can tell with porcupines, can you? xv Citation \u2018Most men will not swim before they are able to.\u2019 Is not that witty? Naturally, they won\u2019t swim! They are born for the solid earth, not for the water. And naturally they won\u2019t think. They are made for life, not for thought. Yes, and he who thinks, what\u2019s more, he who makes thought his business, he may go far in it, but he has bartered the solid earth for the water all the same, and one day he will drown.\u201d \u2013 Hermann Hesse Steppenwolf 1 Chapter 1 Introduction: Making truth happen 1.1. Introduction: k\u00e2-y\u00f4sk\u00e2tahk \u00f4ma n\u00eahiyaw\u00eawin When I asked a Plains Cree (Algonquian, Northern Plains) speaker about the linguistic structures considered in this thesis, I was told \u2018k\u00e2-y\u00f4sk\u00e2tahk \u00f4ma n\u00eahiyaw\u00eawin,\u2019 literally \u2018The Plains Cree language is soft.\u20191 (1) \u1473 \u152b\u1422\u1473\u1456\u1541 \u1406\u14aa \u14c0\u1426\u1403\u152d\u140d\u140f\u1423. k\u00e2-y\u00f4sk\u00e2tahk \u00f4ma n\u00eahiyaw\u00eawin. k\u00e2-y\u00f4sk=\u00e2t =an-k aw =ima n\u00eahiyaw=\u00ea =win C2-soft =by.air=II -0 PRX=IN.SG cree =AI=NOM \u2018The Plains Cree language is soft.\u20192 Here, the use of the concept of \u2018soft\u2019 (y\u00f4sk-) is meant to convey a particular philosophy about the purpose of communication and the way it ought to proceed. As Lightning (1996:62) explains it, the Speaker makes themselves \u201cvulnerable and open\u201d to their audience, not trying to impress their assertions on the audience by force of personality or argument. Rather than speaking to convince (cf. Aristotle\u2019s \u2018Rhetoric\u2019), a Plains Cree speaker\u2019s communicative intent is understood as coming from a need to express what they believe is the truth, individualized to the particular audience they are speaking to (Lightning 1996:63). The mechanisms of discourse, then, are organized to set up the proper events for \u201ctruth to happen\u201d (Lightning 1996:63). This means that accuracy in representation, both of the Speaker\u2019s beliefs and the beliefs they convey from others, is absolutely crucial to the Speaker\u2019s goals for the discourse; the Hearer(s) must trust the Speaker, and the Speaker must be trustworthy. The Plains Cree language is \u2018soft,\u2019 then, because its users prize the \u201cmutual-thinking\u201d (Lightning 1996) that develops between Speaker and Hearer through careful representation, and they value this over the force of logical or charismatic persuasion. This means that, to a speaker of Plains Cree, the grammatical material covered in this thesis relates to how the Speaker \u201cmakes truth happen.\u201d 1 Thanks to Joseph Deschamps, \u1405\u146d\u14ab\u1424 \u1473\u1439\u140f\u1424, for this form and discussion of its meaning. \u146d\u14c7\u14c8\u1422\u146f\u14a5\u144e\u1423. 2 Thanks to H.C. Wolfart (p.c.) for help with the composition of this form. 2 1.2. How does \u201ctruth happen\u201d in Plains Cree? This thesis considers the construction and maintenance of \u201ctruth happenings\u201d in Plain Cree. By \u201ctruth happenings,\u201d I mean the way that a speaker of Plains Cree ties together some proposition (e.g. a sentence) with a means to evaluate it. As is demonstrated in this thesis, Plains Cree possesses a rich set of grammatical mechanisms that the Speaker can manipulate to construct these representations of truth. I claim that there are two basic ingredients to the \u201ctruth happening\u201d process in Plains Cree: (i) A proposition (ii) A thinker. That is, truth \u201chappens\u201d when a proposition is held by someone; without this \u201csomeone\u201d there can be no truth. Making truth happen amounts to connecting propositions to thinkers. Truth, then, is the result of relativizing a proposition to a person; it is inherently relational or individualized in the sense of Lightning (1996). If \u201ctruth happening\u201d is the process of connecting a proposition to a thinker, we expect that Plains Cree grammar will be concerned with the identification of thinkers, and the linking of propositions to these thinkers. In particular, we expect two kinds of operations to be at play: (i) Identifying thinkers (ii) Coding the relation of these thinkers to propositions. The first process (identifying thinkers) is the primary of the two; it does not matter what the relation of a thinker is to a proposition if there is not yet a thinker. In studying the process of \u201ctruth happening\u201d in Plains Cree, then, we should begin by looking at mechanisms that are used to pick out thinkers. This is the core topic of this thesis. To illustrate these two operations in Plains Cree, let us consider two limited cases of \u201ctruth happening.\u201d In the first case, the \u201cCalvin and Hobbes\u201d3 problem (\u00a71.2.1), we see two things: (i) There are grammatical forms that distinguish referents that can \u201cthink\u201d from referents that can\u2019t. (ii) These grammatical forms can only be interpreted once a decision has been made about who thinks that the referent can think. 3 The characters and comic strip \u201cCalvin and Hobbes\u201d are copyright Bill Watterson. Its use here is for scholarly and illustrative purposes only, and is thus allowed under \u201cFair Use\u201d copyright laws. 3 Without this relativization, the grammatical terms still have only the status of propositional elements. In the second case, the \u201c6 degrees of separation\u201d problem (\u00a71.2.2), we see two things: (i) There are grammatical forms that distinguish which referents are thinking in the given context. (ii) These grammatical forms are always relativized to what the Speaker knows. Plains Cree grammar, then, possesses a rich set of devices for discriminating those who can make \u201ctruth happen\u201d from those who can\u2019t. The rest of the thesis is concerned with the proper description and modeling of these phenomena. 1.2.1 The \u201cCalvin and Hobbes\u201d problem in Plains Cree Anyone that has been to a book store or read a newspaper in the last 20 years will be familiar with the comic strip \u201cCalvin and Hobbes\u201d by Bill Watterson. Calvin is a young boy who has a best friend that is a tiger named Hobbes. Being anthropomorphic, Hobbes can walk and talk just like a human, and he goes on many adventures with Calvin. However, at other times in the comic strip, Hobbes is simply an inert stuffed animal. This means that there are two versions of Hobbes presented in the comic strip: (i) Hobbes is an anthropomorphic tiger. (ii) Hobbes is a stuffed toy. This I term the \u201cCalvin and Hobbes\u201d problem \u2013 how are we to understand the changing status of Hobbes? In the Plains Cree language, there are two grammatical classes for nominals, let us call them \u201cA\u201d and \u201cB\u201d for now, and these two classes map directly onto the Calvin and Hobbes problem. Consider a context in which Calvin has lost Hobbes out in the woods, and his father went out to look for him. After finding Hobbes, the stuffed animal, out behind some bushes, the father puts it in bed with Calvin. Upon waking, Calvin runs to inform his mother that Hobbes came back from his evening wanderings. In responding, the mother must either mark Hobbes with grammatical category \u201cA\u201d (2a) or grammatical category \u201cB\u201d (2b).4 4 This context is an actual comic strip by Bill Watterson, and was used to elicit the Plains Cree forms. 4 (2) a. CATEGORY A \u146f\u1426\u1456\u140f\u1429 \u146e \u14a5\u1422\u1472\u1466 Hobbes. koht\u00e2wiy k\u00ee-miskam Hobbes. k-oht\u00e2wiy k\u00ee- m =isk -am Hobbes 2-father PREV-find=by.body.TI-TI Hobbes \u2018Your fatherAN found HobbesIN.\u2019 (Presented S2) b. CATEGORY B \u146f\u1426\u1456\u140f\u1429 \u146e \u14a5\u1422\u1472\u140d\u1424 Hobbes\u140a. k\u00f4ht\u00e2wiy miskaw\u00eaw Hobbesa. k-oht\u00e2wiy m =iskaw -\u00ea -w Hobbes-a 2-father find=by.body.TA-DIR-3 Hobbes-XT5 \u2018Your fatherPROX found HobbesOBV.\u2019 (Presented S2) If the mother marks Hobbes with grammatical category \u201cA,\u201d which is shown by the \u2013isk\u2013am set of suffixes on the verb in (2a), she is saying that Hobbes is a stuffed animal. If she instead chooses grammatical category \u201cB,\u201d which is shown by the combination of the \u2013iskaw\u2013\u00e2\u2013w suffixes on the verb and the \u2013a on the noun (2b), she is saying that Hobbes is a walking, talking tiger. For example, if the \u201cA\u201d form is used, subsequent discourse cannot allow Hobbes to speak, walk, or believe anything (e.g. being happy about being found) (3a), while the \u201cB\u201d form can allow Hobbes to talk, walk, and believe things (e.g. being happy about being found) (3b). (3) a. CATEGORY A # \u1401\u1426\u140b, \u146f\u1426\u1456\u140f\u1429 \u146e \u14a5\u1422\u1472\u1466 Hobbes; \u1401 \u14a5\u1429\u140d\u1528\u1426\u1455\u14a5\u1541. # \u00eah\u00e2, k\u00f4ht\u00e2wiy k\u00ee-miskam Hobbes; \u00ea-miyw\u00eayihtahk. \u00eah\u00e2 k-\u00f4ht\u00e2wiy k\u00ee- mi =sk -am Hobbes \u00ea- miyw=\u00eayiht -am-k yes 2-father PREV-find=by.body.TI-TI Hobbes C1-good=by.mind.TI-TI -3 Intended: \u2018Yes, your fatherAN found HobbesIN; [Hobbes] was happy.\u2019 (Presented S2) b. CATEGORY B \u1401\u1426\u140b, \u146f\u1426\u1456\u140f\u1429 \u146e \u14a5\u1422\u1472\u140d\u1424 Hobbes\u140a; \u1401 \u14a5\u1429\u140d\u1528\u1426\u1455\u14a5\u1528\u141f. \u00eah\u00e2, k\u00f4ht\u00e2wiy k\u00ee-miskaw\u00eaw Hobbesa; \u00ea-miyw\u00eayihtamiyit. \u00eah\u00e2 k-\u00f4ht\u00e2wiy k\u00ee- mi =skaw -\u00ea -w Hobbes-a \u00ea- miyw=\u00eayiht -am-yi-t yes 2-father PREV-find=by.body.TA-DIR-3 Hobbes-XT c1-good=by.mind.TI-TI-DS-3 \u2018Yes, your fatherPROX found HobbesOBV; heOBV was happy.\u2019 (Presented S2) In this situation, then, the mother must choose between representing her own belief about Hobbes and that of her child; does she use the \u201cA\u201d form, which corresponds to what she herself sees, or does she use the \u201cB\u201d form, which corresponds to what her son thinks of his tiger? 5 For an explanation of \u2018extentional\u2019 (XT), see \u00a71.3.1. below. 5 The Plains Cree grammatical categories \u201cA\u201d and \u201cB,\u201d then, do two important things relative to the current discussion: (i) They discriminate between non-thinkers (\u201cA\u201d) and potential thinkers (\u201cB\u201d). (ii) They require that someone think they are non-thinkers or thinkers. As such, these two grammatical categories are intimately involved in making \u201ctruth happen\u201d in Plains Cree. These two categories are the focus of Chapter 2. 1.2.2 The \u201c6-Degrees of Separation\u201d problem in Plains Cree In considering our relation to famous or important people, we sometimes count in terms of our relation to people who have been related to famous people. For example, my uncle once wrestled Hulk Hogan. Counting from The Hulk to me, I can say that I am separated from him by one degree of separation (i.e. my intermediate Uncle Ned). Of course, I can\u2019t say that I \u201cknow\u201d The Hulk \u2013 only that I know someone who knew him. This game of counting intermediate people has sometimes been done as a logic puzzle \u2013 as in the hobby of connecting actors to other actors until one of them has starred in a movie with Kevin Bacon (i.e. the \u201cSix Degrees of Separation from Kevin Bacon\u201d game). Suppose that I am speaking Plains Cree, and I want to tell you something that Hulk Hogan said to my Uncle. In this language, I have a choice of three ways to present this (4). (4) a. \u1401\u146f\u14ef \u1403\u1458\u1424 \u14c8\u142f\u1424. \u00eakosi itw\u00eaw n\u00e2p\u00eaw \u00eakosi it =w\u00ea-w n\u00e2p\u00eaw so thus=AI -3 man \u2018That\u2019s what the manAN said\u2019 (Volunteered S4, Presented S2,S3) b. \u1401\u146f\u14ef \u1401 \u1403\u1458\u141f \u14c8\u142f\u1424. \u00eakosi \u00ea-itw\u00eat n\u00e2p\u00eaw \u00eakosi \u00ea- it =w\u00ea-t n\u00e2p\u00eaw so C1-thus=AI -3 man \u2018That\u2019s what the manAN said\u2019 (Volunteered S4, Presented S2,S3) c. \u1401\u146f\u14ef \u1401 \u1403\u1458\u1528\u141f (\u1401\u14f4) \u14c8\u142f\u1418. \u00eakosi \u00ea-itw\u00eayit (\u00easa) n\u00e2p\u00eawa \u00eakosi \u00ea- it =w\u00ea-yi -t \u00easa n\u00e2p\u00eaw-a so C1-thus=AI -DS-3 EVID man -XT \u2018That\u2019s what the manOBV said\u2019 (Volunteered S4, Presented S2,S3) 6 While the form in (4a) indicates that I heard the Hulk myself, and the form in (4b) is non- committal on how I heard about this, the final form, in (4c), is what I would say if I wanted to make it clear that I wasn\u2019t there to hear Hulk Hogan tell this \u2013 it was something he said to someone else, that I\u2019m passing on. By marking the noun n\u00e2p\u00eaw \u2018man\u2019 with the suffix \u2013a, and putting the suffix \u2013yi\u2013 in the verb, I am telling you that I don\u2019t know this Hulk Hogan fellow personally, and thus I can\u2019t vouch for this directly. You\u2019d have to go check with my Uncle to make sure that this was really what The Hulk said. Thus, Plains Cree has a way to do this \u201cdegree-counting\u201d right in its grammar, by specially-marking the noun and verbs that relate to it. 1.3. Proposal: Individuals related to perspectives In this thesis, I attempt to model \u201ctruth happening\u201d in Plains Cree. In doing this, it is important to remember the distinction between a phenomenon (Greek: \u03c6\u03b1\u03b9\u03bd\u00f3\u00b5\u03b5\u03bd\u03bf\u03bd \u2018that which appears\u2019) and the model of the phenomenon. I am not claiming that the model I am constructing is, in any sense, \u2018real.\u2019 Rather, it is a constructed representation that covers the observable facts and gives us a systematic way to look for new facts to model. In other words, the model is a map, not a territory (cf. Korzybski 1958, Bateson 1971). 1.3.1. Perspective possession In this thesis, I take \u201ctruth happening\u201d to be the relation of individuals to propositions. This is an inherently relativized process; the truth of a proposition is relative to some individual. In particular, the truth of a proposition is relative to the individual\u2019s perspective (in the sense of K\u00f6lbel 2002) \u2013 the way that individual sees the world. The process of \u201ctruth-happening,\u201d then, is the process of mapping a proposition into some individual\u2019s perspective. Perspectives can be thought of as analogous to other possessed entities (K\u00f6lbel 2002); each individual owns one, and uses it to evaluate the truth of propositions. In a sense, it is like a body part that the individual takes with them wherever they go. Parallel to possession, then, we can define a perspective and an individual as being related by possessor relation \u2018R\u2019 (Higginbotham 1983). (5) a. [[John\u2019s dog]] = R(John,dog) = There is a relation between John and the dog. b. R(x,\u03c8) = There is a relation between an individual x and a perspective \u03c8. 7 In most cases, the relation between the possessor and the perspective is unspecified, analogous to normal possession. However, there are ways to further restrict this relation, by introducing predicates that explicitly restrict this unspecified \u2018R\u2019 relation, just as with kinship terms (Burton 1995). With perspectives, these predicates are typically called \u201cpropositional attitude\u201d predicates (Russell 1918), and comprise predicates like \u201cthink,\u201d \u201cfeel,\u201d or \u201csay.\u201d (6) a. [[John\u2019s mother]] = mother-of(x,John) = There is a mother relation between John and x. (Burton 1995) b. [[John thinks]] = think(John, \u03c8) = There is a think relation between John and \u03c8. A perspective, then, is a domain that an individual possesses with which to evaluate the truth of a proposition. If the \u2018R\u2019 relation of perspective possession is further specified by the addition of predicates, we could expect a language to grammatically separate perspective possession from the relation to the perspective. In fact, in Plains Cree, many of these propositional attitude predicates are built off of one element \u2013 the suffix \u2013\u00eayiht \u2018by mind.\u2019 Thus, to \u2018believe\u2019 is t\u00e2pw\u00eawak\u00eayihtam \u2018to hold it true in the mind,\u2019 while to \u2018think\u2019 is it\u00eayihtam \u2018to do thus with it in the mind.\u2019 (7) a. \u1456\u143b\u1418\u146b\u1528\u1426\u1455\u1466. t\u00e2pw\u00eawak\u00eayihtam. t\u00e2pw\u00eawak=\u00eayiht -am true =by.mind.TI-TI \u2018S\/heAN believes in itIN, holds it to be true.\u2019 b. \u1403\u144c\u1528\u1426\u1455\u1466. it\u00eayihtam. it =\u00eayiht -am thus=by.mind.TI-TI \u2018S\/heAN thinks thus of itIN.\u2019 Plains Cree, then, could be thought of as coding the having of a perspective (-\u00eayiht) and the relation to the perspective (e.g. t\u00e2pw\u00ea- \u2018true\u2019) as separate linguistic elements, which crucially exist in an structurally-conditioned, asymmetric relation to the holder of perspective.6 The concept of perspective possession provides a way to model this pattern. 6 This asymmetry presents challenges for accounts that treat the perspective holder, the perspective, and the relation to the perspective as an undifferentiated triple (e.g. Kamp 1990). While the current account does not provide a full 8 The definition of perspective possession adopted here gives rise to two kinds of individuals: (i) Those that hold a perspective, (ii) Those that don\u2019t. For perspective-holders, I use the term \u201cintentional,\u201d because they have \u201cintentions,\u201d which means that they have mental processes directed at something (Brentano 1874). For individuals that do not possess a perspective, I adopt the term \u201cextentional,\u201d since they do not have mental processes directed at anything.7 \u201cIntentionality,\u201d then, is the property of possessing a perspective. 1.3.2. Discourse Representation Theory with perspectives To model the contexts for perspective possession in Plains Cree, I employ a modified form of Discourse Representation Theory (DRT: Kamp 1981) that is built off of the work of Farkas (1992) and Smith (2004). In this model, each Discourse Representation Structure (DRS: symbolized by a \u2018box\u2019 in the representation) represents a perspective (the domain in which the truth of a proposition is evaluated). This perspective is necessarily \u2018anchored\u2019 to an individual (Fillmore 1971, Ruwet 1982, K\u00f6lbel 2002, Lasersohn 2005, Stephenson 2007), creating a perspective-individual pair (e.g. R(x,\u03c8)). This perspective-individual pair is represented in the DRS through the following conventions: (i) The perspective is represented by the \u2018box\u2019 of the DRS itself. (ii) The individual that possesses this perspective is marked above the upper-left corner of the box. (iii) The relation between the individual and the perspective is represented by the predicate in angle brackets to the right of the individual. x (8) = R(x,\u03c8) model of this asymmetry, it is amenable to future work that considers the compositional nature of propositional attitudes. 7 Thanks to Hotze Rullmann (p.c.) for suggesting this term. 9 Elements embedded within a perspective are thus placed inside the corresponding \u2018box\u2019 of the DRS, and variables related to this perspective are introduced in the frame at the top of each \u2018box.\u2019 (9) Jeff: The sun feels good.8 Jeff x = say(Jeff,\u03c8) sun(x) feel.good(x) The relation between the individual and proposition is neutrally unspecified If the Speaker wishes to represent the perspective of some other individual, a second layer of embedding can be added.9 (10) Clare: Jeff thinks the sun feels good. Clare x Jeff = say(Clare,\u03c8) sun(x) think(Jeff) Jeff = think(Jeff,\u03c8') x sun(x) feel.good(x) Of course, the DRS in (10) is not a complete representation of the utterance in (10); I have abstracted away from places and times. For the purposes of understanding the reference-typing 8 Regarding the argument structure of propositional attitude verbs, Cook (2008) provides evidence that propositions (syntactically CPs) are never the objects of verbs. This is the reason for the treatment of predicates like say in the formalism. 9 Although the implementation here is similar in method to that considered by Kamp (1990), which would model \u201cintentional\u201d as an \u201cexternal anchor\u201d and \u201cextentional\u201d as an \u201cinternal anchor,\u201d the two accounts differ with respect to their ability to model multiple layers (>2) of embedded perspectives, which is crucial for the Plains Cree data considered in Chapter 3. 10 phenomena in Plains Cree, there does not appear to be a need to distinguish between times and places, although a full model of Plains Cree perspectival meanings will necessarily require their inclusion. In the study of obviation phenomena (Chapter 3), I employ the context variable C of Partee (1989), which can be conceived of as a collapsing of place and time together. This is outlined in section 1.3.3 below. The use of a DRS to represent a perspective implicates several alterations to the standard DRT developed by Kamp (1981). In particular, the typical DRT convention of carrying referents into embedded boxes cannot here be maintained; to do so would indicate that any referent known to the Speaker would also be known to the possessor of embedded perspectives. Altering this convention allows for the modeling of differences in perspectives regarding referents. For example, the \u201csun\u201d in the perspective of Clare in (10) is repeated in the embedded perspective of Jeff. This represents that both Clare and Jeff have the same referent in mind. If the variable of the embedded perspective were switched (e.g. to y), then Clare and Jeff would have different versions of the \u201csun\u201d in mind, as schematized in (11). (11) Clare: Jeff thinks the sun feels good. Clare x Jeff = say(Clare,\u03c8) sun(x) think(Jeff) Jeff = think(Jeff,\u03c8') y sun(y) feel.good(x) The ban on referents carrying into embedded DRS structures also has implications for existential quantification. In typical DRT frameworks, the referent is taken to exist within the DRS it is located in. In the current framework, this existence only holds of a perspective. For example, consider a context in which Clare and Jeff were walking in the woods together and they saw a 11 large creature run by. Jeff claims this creature was a hodag,10 while Clare thinks it was merely an exceptionally large dog. We can model this as in (12), where the predicate \u2018large\u2019 is located in both perspectives, while the predicate \u2018hodag\u2019 is located only in the perspective possessed by Jeff. (12) Clare: Jeff thinks he saw a large hodag. Clare x Jeff = say(Clare,\u03c8) think(Jeff) large(x) Jeff = think(Jeff,\u03c8') x see(Jeff,x) hodag(x) large(x) This represents that Clare is only committing to the referent\u2019s bigness, while the idea that it was a hodag is entirely Jeff\u2019s. Within this model, \u201cTruth-happening\u201d is understood as the process of connecting a proposition to a perspective-individual pairing. This connection provides the means for an individual to evaluate the proposition. 1.3.3. Proposal: Four kinds of referents in Plains Cree I propose that Plains Cree can be understood to refer to four kinds of individuals, defined in terms of perspectival possession (i.e. intentionality): (i) Individuals that can never possess a perspective. (ii) Individuals that cannot currently possess a perspective. (iii) Individuals that currently possess a perspective. (iv) Individuals unspecified for perspective possession. 10 A \u2018hodag\u2019 is a legendary beast that is thought to live in Northern Wisconsin woodland areas, particularly around Rhinelander, Eagle River, Minocqua, and Hayward. 12 The first kind of individuals, those that can never possess a perspective, correspond to the grammatical category used to mark Hobbes as a stuffed animal in example (2). This category is termed \u201cinanimate\u201d in the literature (cf. Wolfart 1973). (13) CATEGORY (I) = INANIMATE \u146f\u1426\u1456\u140f\u1429 \u146e \u14a5\u1422\u1472\u1466 Hobbes. koht\u00e2wiy k\u00ee-miskam Hobbes. k-oht\u00e2wiy k\u00ee- m =isk -am Hobbes 2-father PREV-find=by.body.TI-TI Hobbes \u2018Your fatherAN found HobbesIN.\u2019 (Presented S2) Here, the verb is coded with two forms that pick out individuals that never possess perspectives: (i) The suffix \u2013isk \u2018done by the body to an inanimate thing\u2019 (ii) The suffix \u2013am \u2018inanimate object\u2019 I propose that this category marks individuals that are never able to possess a perspective; they are inherently extentional. This is represented in the formalism as in (14). (14) EXT(x) \u2194 \u2200\u03c8\u2200y(R(y,\u03c8) \uf0e0 x \u2260 y) x is Extentional if and only if for all Perspectives \u03c8 and all individuals y, if there is a relation R between individual y with perspective \u03c8, then x is not y. As such, referents marked with this category will be unable to speak, think, or feel, since those are predicates that introduce a relation between a perspective and an individual (\u00a71.3.1). Modeling the difference in who believes the referent is extentional thus reduces to a differential embedding of the extentional-marking predicate. (15) \u146f\u1426\u1456\u140f\u1429 \u146e \u14a5\u1422\u1472\u1466 Hobbes. koht\u00e2wiy k\u00ee-miskam Hobbes. k-oht\u00e2wiy k\u00ee- m =isk -am Hobbes 2-father PREV-find=by.body.TI-TI Hobbes \u2018Your fatherAN found HobbesIN.\u2019 (Presented S2) 13 a. CONTEXT 1: EXTENTIONAL W.R.T. SPEAKER11 Speaker x Hobbes father(x,Hearer) EXT(Hobbes) find(x,Hobbes) b. CONTEXT 2: EXTENTIONAL W.R.T. HEARER Speaker Here, the location of the EXT predicate in the perspective of one or the other perspective represents who conceives of Hobbes as extentional. This category is considered in detail in Chapter 2. The second kind of individuals, those that are currently unable to possess a perspective, are signified by the structure used to mark n\u00e2p\u00eaw \u2018the man\u2019 in example (4c), termed \u201cobviative\u201d in the literature (cf. Wolfart 1973). (16) CATEGORY (II) = OBVIATIVE \u1401\u146f\u14ef \u1401 \u1403\u1458\u1528\u141f (\u1401\u14f4) \u14c8\u142f\u1418. \u00eakosi \u00ea-itw\u00eayit (\u00easa) n\u00e2p\u00eawa \u00eakosi \u00ea-itw\u00ea -yi -t \u00easa n\u00e2p\u00eaw-a so C1-say-DS-3 EVID man -XT \u2018That\u2019s what the manOBV said\u2019 (Volunteered S4, Presented S2,S3) 11 I am here treating proper names like other variables. This departs from treatments of proper names as \u2018rigid designators,\u2019 in the sense of Kripke (1980), but allows for the modeling of different perspectives on a referent identified by a proper name. x Hobbes father(x,Hearer) find(x,Hobbes) Hearer Hobbes EXT(Hobbes) 14 Here, the nominal bears a suffix \u2013a, and the verb carries a special suffix \u2013yi\u2013. I propose that this construction codes that the individual is unable to possess a perspective relative to the specified context; the individual is contextually extentional. (17) EXT(x,C) \u2194 \u2200\u03c8\u2200y(R(y,\u03c8,C) \uf0e0 x \u2260 y) x is extentional at context C if and only if for all perspectives \u03c8 and all individuals y, if there is a relation R of y with \u03c8 at context C, then x is not y. This contextual extentionality is always relative to the Speaker\u2019s perspective, and means that \u201cobviative\u201d referents are unable to be assigned a perspective by the Speaker. (18) a. \u1401\u146f\u14ef \u1401 \u1403\u1458\u1528\u141f (\u1401\u14f4) \u14c8\u142f\u1418. \u00eakosi \u00ea-itw\u00eayit (\u00easa) n\u00e2p\u00eawa \u00eakosi \u00ea-itw\u00ea -yi -t \u00easa n\u00e2p\u00eaw-a so C1-say-DS-3 EVID man -XT \u2018That\u2019s what the manOBV said\u2019 (Volunteered S4, Presented S2,S3) b. WELL-FORMED DRS FOR 18A Speaker x C say(x) man(x) EXT(x,C) c. ILL-FORMED DRS FOR 18A Speaker x say(x) man(x) EXT(x,C) x say(x) 15 Here, the well-formed DRS is one in which there is no perspective assigned to the man, who is marked as obviative (18b). Introducing an embedded perspective for this obviative referent is ill- formed (18c). Modeling \u201cobviation,\u201d then, reduces to modeling a contextual ban on perspective possession. This category is considered in detail in Chapter 3. The third kind of individuals is those that currently possess a perspective, termed \u201cProximate\u201d in the literature. This kind of referent is created by contrast between the \u201cObviative\u201d form and a normal nominal, rather than by anything special about the grammatical forms involved. For example, in (19), the perspective of n\u00e2p\u00eaw \u2018man\u2019 is represented, as shown by the explanation offered by the consultant. (19) CATEGORY (III) = \u201cPROXIMATE\u201d \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u00ea-w\u00e2pam\u00e2t n\u00e2p\u00eaw iskw\u00eawa \u00ea-w\u00e2p=am -\u00e2 -t n\u00e2p\u00eaw iskw\u00eaw-a c1-see=by.eye.TA-DIR-3 man woman-XT \u2018The manPROX sees the womanOBV.\u2019 COMMENT (S2): \u201cIn this example, you\u2019re hearing what the man has to say about it.\u201d Speaker x y man(x) EXT(y) x y woman(y) see(x,y) The \u201cProximate,\u201d then possess a perspective for the evaluation of the truth proposition. Being a function of contrast with the \u201cObviative,\u201d this category is also considered in Chapter 3. The last kind of individuals is the elsewhere case, termed \u201cAnimate\u201d in the literature. This kind of individual has neither intentional or extentional properties, as shown by the pair of examples in (20). 16 (20) CATEGORY (IV)= \u201cANIMATE\u201d a. \u14c2\u141a\u1438\u14ab\u1424 \u140a\u14c7 \u140a\u141a\u14ef\u1422 niw\u00e2pam\u00e2w ana aw\u00e2sis ni-w\u00e2pam -\u00e2 -w an=a aw\u00e2sis 1- see=by.eye.TA-DIR-3 DST=AN.SG child \u2018I seeAN thatAN child \/ ribbon.\u2019 Comment (S2): \u201cI don\u2019t know if the child knows about this or not.\u201d b. \u14c2\u141a\u1438\u14ab\u1424 \u140a\u14c7 \u14ed\u14c7\u1439\u1423 niw\u00e2pam\u00e2w ana s\u00eanap\u00e2n ni-w\u00e2pam -\u00e2 -w an=a s\u00eanap\u00e2n 1- see=by.eye.TA-DIR-3 DST=AN.SG ribbon \u2018I seeAN thatAN child \/ ribbon.\u2019 Comment (S2): \u201cThe ribbon couldn\u2019t know about this.\u201d I consider this category in Chapter 2. 1.3.4. Proposal: Constructing obviation via referential dependency Turning from the kinds of individuals coded in Plains Cree to the manner in which these individuals are coded, I propose that the three categories of individuals are not grammatically equal. In particular, I argue that \u201cinanimate\u201d and \u201canimate\u201d are coded by dedicated forms (Chapter 2), while obviation, by contrast is constructed out of other resources available to the grammar (Chapter 3). Animacy, then is a primitive property of Plains Cree grammar, while obviation is not. In constructing obviation, I propose that Plains Cree recruits elements that code referential dependency (i.e. the process of making one referent dependent on another for its interpretation). This referential dependency can be understood as the satisfaction of two conditions: (i) a structural condition (c-command\/linear precedence, shown on the left side of the table) and (ii) a semantic condition (perspectival embedding, shown on the right side of the table). 17 STRUCTURAL CONDITION SEMANTIC CONDITION 1 V y V V x V 2 PRED1 \uf0e0 PRED2 5 5 y x Table 1.1.: Two conditions on referential dependency Depending on the configuration that an obviative referent is used in, a different piece of this referential dependency system will be recruited. In Table 1.2, we see that there are five different configurations for obviation, and they correspond with five different dependency types. FORM DEPENDENCY TYPE Noun-Noun Nominal to previous nominal -yi- Subject to previous referent Possession Possessum to Possessor -\u00ea- Object to subject -ikw Subject to Topic Table 1.2.: The forms of obviation and their dependencies Obviation, then, is just the application of referential dependency operations to animate referents. The more general property of Plains Cree is referential dependency. This is explored in detail in Chapter 4. 1.4. Relation to previous work In this section, I consider the ways in which the current model relates to previous work on these topics. In addition, each chapter has its second section devoted to previous work specific to those topics (i.e. \u00a72.2, \u00a73.2, \u00a74.2, \u00a75.2). 1.4.1. Models of subjective meaning Models of meaning in natural language have developed from a philosophical tradition that sought to develop an explicit language for expressing philosophical concepts. Since Aristotle, this philosophical tradition was crucially concerned with constructing arguments that entail 18 contradictions (Hume 1748, Tarski 1944), and numerous proposals were made to pare down linguistic forms to their most abstract, objective forms, in order to reduce ambiguity. Imported into work on non-philosophical language (i.e. natural language), this desire for objective, non- ambiguity was maintained (Cresswell 1985:5). This had the result that semantic analyses of natural language tended to focus on contradiction as a means to define the boundaries of meaning in the formal semantic sense (i.e. \u201cIt is raining\u201d entails that \u201cit is not raining\u201d cannot also be true). The judge for these contradictions was taken to be someone external to the speech- act \u2013 the knower of \u201chow it actually is\u201d out there in the world (Tarski 1944). While this convenient way of talking about truth has proven useful in the description of meaning in natural language, it has long been noticed that numerous components of the natural languages considered (e.g. English, French, etc.) do not inherently entail contradictions, even when supplied with a specific time and place (e.g. Austin 1962). This includes, minimally, linguistic forms like questions (\u201cMay we come in?\u201d), commands (\u201cOpen the door!\u201d), and conditionals (\u201cForget to close the door and you\u2019ll be sorry!\u201d)(Fillmore 1975, Karttunnen 1977). Further, it was noticed that some of these elements in natural language are dependent on a specific individual involved in the speech act (cf. Searle 1965, Cresswell 1985, etc.). For example, an element like the English word \u201clocal\u201d could be true when applied to the University of British Columbia for a speaker in Vancouver, but simultaneously untrue for a speaker in Boston (cf. Fillmore 1975, Mitchell 1987). Thus, the move has been away from the calculation of linguistic meaning solely in terms of contradiction-laden, individual-independent \u201cobjective\u201d meaning and towards relativized, individual-dependent \u201csubjective\u201d meaning. One attempt at formalizing these \u201csubjective\u201d contexts was Farkas (1992), which focussed on modeling the distribution and meaning of Romance subjunctives. To model these, Farkas (1992) employed a DRT model in which an individual was paired with a world. More recently, Smith (2003) has employed DRT to model point of view effects in English discourse. The current model takes much of its mechanisms from these works, though exchanges worlds for perspectives. Recent work on the formalization of subjective meaning has focussed on predicates that trigger so-called \u201cfaultless disagreement\u201d (K\u00f6lbel 2002, Lasersohn 2005, Stephenson 2007). These are contexts in which one person can utter a proposition, and the other person can disagree without saying the first person was wrong.12 12 I here cite the judgements of Stephenson (2007). I myself cannot utter \u201cNo, it isn\u2019t!\u201d in this context. 19 (21) Speaker A: This cake is tasty. Speaker B: No, it isn\u2019t! (Stephenson 2007) Lasersohn (2005) and Stephenson (2007) analyze these forms by appealing to the notion of a \u201cjudge.\u201d This judge is an individual who judges the proposition to be true relative to a time and world. Being thus relativized, these propositions are inherently \u201csubjective.\u201d By employing a parallel relativization mechanism, the current work can be considered an extension of that done by Lasersohn (2005) and Stephenson (2007), but is much closer in its implementation to that of Farkas (1992). In all three accounts, truth evaluation is relativized to a particular individual, although some of the other particulars differ (e.g. perspectives vs. sets of worlds). Setting aside these more minor notational and philosophical differences, the important difference between the current model and these other works is that, whereas these other accounts relativize only certain predicates with respect to an individual, the current account relativizes all propositions to a perspective-holder (minimally, the Speaker). 1.4.2. Perspectives vs. possible worlds In this thesis, I employ the notion of a \u201cperspective\u201d for the evaluation domain of a proposition\u2019s truth. While this follows K\u00f6lbel (2002), it departs significantly from the treatments typically used for the formal semantics of natural language. This means that a comparison of the current theory to other work is in order. It should be noted at the outset that none of the reasons for employing \u201cperspectives\u201d are absolutely vital to the current work, representing instead ontological and philosophical disagreement, and thus a semanticist that uses a possible worlds framework can readily translate the current formalisms into those more comfortable to them (i.e. exchange \u201cperspective of x\u201d for \u201cset of worlds epistemically accessible to x\u201d or \u201cdoxastic alternatives of x\u201d and other modifications as necessary). In most theories of meaning, the domain of evaluation for propositions is a \u201cworld\u201d (Hintikka 1962, Kripke 1963). What exactly is in a \u201cworld\u201d appears to be a matter of some debate, but the canonical view is expressed by Hintikka (1962): a world contains a truth value for every proposition. Sets of worlds, then, are sets of sets of truth values for propositions. Relevant to our current discussion, this kind of system was used by Hintikka (1962) to model belief. For example, were we playing poker, and I wanted to guess the set of cards in your hand, I could write down every possible set of cards you could have, writing down one possible 20 hand on each slip of paper. As the game progresses, I could continually narrow down this set of papers based on new information, until I restrict the set papers to only one or two slips of paper. These would be the \u201cworlds\u201d that still depict possible states of the cards in your hand (i.e. epistemic alternatives; Hintikka 1962). Belief, then, is modeled as the continual narrowing of the set of worlds, based on additional knowledge. To believe something is to know the set of worlds that are accessible given your current belief. This is sometimes called an individual\u2019s \u201cdoxastic alternatives.\u201d The core difference between a possible worlds model like this and the perspective model I am here using lies in the notion of exhaustivity; worlds are taken to be exhaustively defined but perspectives are not. That is, a world has a truth value for every proposition in it, whereas a perspective only has a truth value for those propositions that are given to it. This has the result that individuals can only make reference to sets of worlds (since they do not know the truth value of every proposition in existence), but can make reference to a single perspective (since all they have to know is the set of propositions in it). Worlds, then, are always being restricted into a more narrow set (as the beliefs of the referent increases), whereas perspectives are always being specified with more belief content. In this sense, perspectives are something like \u201cpartial worlds\u201d or \u201csituations,\u201d depending on the model employed (Barwise 1981). The advantage of the perspective model is this orientation towards specification. Rather than defining the acquisition of new information as the reduction in number of the set of worlds in the individual\u2019s set of doxastic alternatives, we can define \u201cbelief\u201d as the addition of some proposition to some individual\u2019s perspective. More precisely, we can say that an individual possesses different perspectives for different kinds of propositional attitudes; there are perspectives filled with the things they \u201cthink\u201d versus perspectives filled with the things they \u201cbelieve\u201d or \u201cdoubt.\u201d Thus, there is the perspective, and then there is the relation to this perspective. This allows us to think carefully about the ways that propositional attitudes are constructed in natural languages like Plains Cree, where these predicates often have two elements (e.g. \u2013\u00eayiht \u2018by mind\u2019 and t\u00e2pw\u00ea- \u2018true\u2019; \u00a71.3.1). 21 1.5. The Plains Cree language Plains Cree is an Algonquian language originally spoken in the northern Plains of North America by approximately 30,000 speakers.13 Speakers are typically in their 50\u2019s or older, with some reports of children acquiring the language in the less-populated northern areas (Northern Alberta and Northern Saskatchewan). Since the language has failed to transfer to between two and three generations of speakers in most communities, it is safe to conclude that the language is now in a moribund state. Plains Cree is part of a larger language group of Central Algonquian languages sometimes called the \u201cCree Dialect Continuum,\u201d which stretches from Labrador to Alberta. The precise division between dialects and languages is difficult, but the speakers I work with completely lose intelligibility around James Bay (Moose Cree). Speakers are unable to recognize the more eastern varieties (e.g. Montagnais, Innu-Aimun) as a related language. The languages share basic lexical forms, and much morphology, but the phonology, morphophonology, and syntax are significantly different. 1.5.1. Sources Plains Cree has a strong descriptive tradition, stretching back to the 1840\u2019s (Howse 1844). I here offer a review of available materials that the reader can refer to for particular forms and discussion. I have also included a glossary section at the end of the thesis, which identifies and defines every form used in this thesis. Using this glossary in concert with other grammatical materials should provide the necessary tools to understand the Plains Cree data presented here. There are three major grammatical descriptions of Plains Cree; Howse (1865), Lacombe (1874), and Wolfart (1973). Of these, Wolfart (1973) is the most accessible, and provides a discussion and analysis of the earlier two works. It represents a clear statement of the principles of grammatical description that Hockett (1966) and Bloomfield (1962) developed, and is thus of interest for theoretical as well as practical reasons. 13 Note that these estimates are quite old (SIL: 1982). 22 Beyond these more formal grammatical descriptions are those aimed at a less linguistically-trained audience, written with the intent of introducing beginners to the language, training teachers, and instructing students of the language. In particular, the grammatical descriptions of Wolfart and Carroll (1981), Ahenakew (1987), and Okim\u00e2sis and Ratt (1999) are useful sources. The first book is a succinct introduction to the most notable features of Plains Cree grammar. In addition to presenting user-friendly grammatical discussion, the latter two books offer the insights of native Plains Cree speakers, which are sometimes found nowhere else (for example Ahenakew\u2019s discussion of \u2018factive\u2019 \u00f4ma). In terms of dictionaries, there are three major ones to choose from: LeClaire and Cardinal (1998), Wolfart and Ahenakew (1998), and (Wolvengrey 2001). In work with modern Plains Cree speakers, the 2-volume dictionary by Wolvengrey (2001) is the most useful. It serves as a fully-functional dictionary, comparable to an modern dictionary for Spanish or German. Wolfart and Ahenakew (1998), which is constructed from corpus work, is an excellent companion to published texts, but can only offer words that have been recorded there (e.g. if the texts have no instance of a word for \u201chigh-heeled shoes,\u201d the dictionary will not have an entry for this word.) Plains Cree has one of the best text collections of any language indigenous to North America. Starting with Bloomfield\u2019s work (1930, 1933), high-quality textual work has been consistently produced. Beginning in the late 1980s, the partnership of Ahenakew and Wolfart produced a set of Plains Cree-language books (1993, 1997, 1999, 2000) that provide a significant body of high-quality data, complete with translation, commentary, and glossary. More recently, Wolvengrey (2007) has begun an initiative to publish sets of texts. Formal linguistic work on Plains Cree has focussed largely on the complex morphosyntax of the language. Relevant work that deals with morphosyntax includes Dahlstrom (1986), Dryer (1996), Hirose (2000), and D\u00e9chaine (1999, 2003, 2008). The complex particle system of Plains Cree has been described in detail by Ogg (1991). Clausal structure has been considered by Blain (1997), Long (1999), and Cook (2008). Obviation phenomena has been considered by Wolfart (1978), Russell (1991, 1996), and 23 M\u00fchlbauer (herein). Other grammatical features that have been considered are pronouns (Blain 1995), weather verbs (Blain 1987), incorporation (Wolfart 1971), possession (M\u00fchlbauer 2004, 2007), word order (M\u00fchlbauer 2003, Wolvengrey 2007, D\u00e9chaine 2007), demonstrative ordering (Wolvengrey 2003), preverbs (Cook 2006, Wolvengrey 2006), relative roots (Cook 2003), evidentials (Blain et al. 2006, Blain & D\u00e9chaine 2002, 2007), and reduplication (Ahenakew & Wolfart 1983). Phonetic and phonological discussions include treatments of prominence (M\u00fchlbauer 2006), ablaut (Wolfart 1973), epenthesis and deletion (Wolfart 1973), syllable structure (Cook 2003), diminutive palatalization (Hirose 1999), prosody (Cook 2006), and phrasing (Cook & M\u00fchlbauer 2005). 1.5.2. Orthography There are two orthographies for Plains Cree: (i) the standard roman orthography (SRO), and (ii) the syllabic orthography. I here offer a short guide to understanding these systems, as they relate to the current thesis. The standard roman orthography\u2019s main notable features are the circumflexes over vowels, which represent vowel quality contrasts (e.g. a [\u04d9] vs. \u00e2 [a], i [\u026a] vs. \u00ee [i], o [\u028a] vs. \u00f4 [u]), the use of \u2018c\u2019 for the palatal [ts], and the use of pre-consonantal \u2018h\u2019 to code pre-aspiration. It should be noted that this orthography aims for a highly phonemicized spelling system, and does not closely approximate the surface forms of the language as it is currently spoken. As such, it does not record the wide dialect variation found within Plains Cree. For example, one consultant (S2) turns all [ks] clusters into geminate [ss], often turns CVC sequences into [\u0294]C, and metathesizes all [wa] sequences to [\u0254]. Using the standardized orthography, none of this is written down \u2013 just as none of the dialect variation of English is captured in the orthography (e.g. you can\u2019t tell by reading this that I often drift into a heavy Northern Wisconsin accent). The data in this thesis, then, is not suitable for phonological analysis. The syllabic writing system is favored by many speakers of the language. It represents the same basic vowel and consonant contrasts as the standard roman orthography, but does so syllabically. The shape of the symbol codes the consonantal value, while its orientation codes the vowel. Thus, \u142f\u1431\u1433\u1438 all represent different combinations of [p] with a vowel. This orthography has been used in this thesis because of its value for speakers of the language. Detailed explanations of this system can be found in the dictionary of Wolvengrey (2001), and elsewhere. 24 1.6. Outline of the thesis There are five chapters in this thesis, as well as a glossary. They are organized in the following way. Chapter 2 considers the form, content, and context of the two animacy classes in Plains Cree. I argue there that the formal organization of these two classes can be best understood as dedicated coding that is distributed across all syntactic positions that code reference (e.g. nominal and argument structure). I then argue that the inanimate class is inherently extentional in meaning, while the inherently animate class has no inherent content at all. Contextually, these two classes of nominals are manipulated to code different individual\u2019s perspectives about the referent. Chapter 3 considers the form, content, and context of obviation. Rather than being a basic grammatical category of Plains Cree, I argue that obviation is best understood as the result of using several independent kinds of morphosyntax to construct the referential category \u201cobviative.\u201d In terms of its specific referential properties, I analyze the obviative as denoting referents that are contextual extentionality. This referential information is then situated within the Speaker\u2019s perspective as a filter on potential perspective embeddings for this referent. In Chapter 4, I turn to the forms used to code obviation. I propose that Plains Cree constructs obviation out of forms that code the more general property of referential dependency. This referential dependency obeys a structural and a semantic condition that is operative across Plains Cree grammar. Obviation, then, is a by-product of more basic (morpho)syntactic operations of Plains Cree. In Chapter 5, the focus shifts from analysis of the data to how the data was obtained. Here, I consider the ways that different kinds of data collecting methods affect the outcome, and catalogue how each kind of method interacts with obviation and animacy. From this consideration, I argue that all data is good, so long as we think carefully about the context of its collection. In Chapter 6, I conclude the discussion and consider its implications. In particular, I consider approaches to the plural\/obviative parallels, the behaviour of the suffix \u2013yi\u2013 across the Cree languages, and the similarities and differences of the Plains Cree system described here to that shown in Athabaskan languages. Because Algonquian linguistics has its own terminology, I have included a glossary at the end of the work. This glossary contains every term and every gloss used in this thesis, and what 25 they mean. It is intended to not only make the thesis more accessible to a non-Algonquianist audience, but to also help make the entire field of Algonquian linguistics accessible to non- specialists. 26 Chapter 2 The Form, Content, and Context of Animacy 2.1. Proposal In this chapter, I consider the form, content, and context of the two basic referential classes of Plains Cree, traditionally termed \u201cAnimate\u201d and \u201cInanimate.\u201d1 (1) a. \u201cANIMATE\u201d \u14c2\u141a\u1438\u14ab\u1418\u1420 \u140a\u14c2\u146d \u14aa\u1422\u146d\u14ef\u14c7\u1420 niw\u00e2pam\u00e2wak aniki maskisinak ni-w\u00e2p=am -\u00e2 -w-ak an =iki maskisin-ak 1-bright=by.eye.TA-DIR-3-PL DST=AN.PL shoe -PL \u2018I seeAN thoseAN shoesAN.\u2019 (Presented S2) b. \u201cINANIMATE\u201d \u14c2\u141a\u1438\u1426\u144c\u1423 \u140a\u14c2\u1426\u1403 \u14aa\u1422\u146d\u14ef\u14c7 niw\u00e2paht\u00ean anihi maskisina ni-w\u00e2p=aht -\u00ea -n an =ihi maskisin-a 1-bright=by.eye.TI-TI-LP DST=XT shoe -XT \u2018I seeIN thoseIN shoesIN.\u2019 (Presented S2) The form in (1a) shows 4 distinct morphemes that code \u201cAnimate\u201d referents (\u2013am\u2013, \u2013\u00e2\u2013, \u2013ak, and \u2013iki), and these are mirrored by 4 distinct morphemes in (1b) that code \u201cInanimate\u201d referents (\u2013aht\u2013, \u2013\u00ea\u2013, \u2013a, and \u2013ihi). After introducing the proposal, I review the previous literature on animacy in Plains Cree, paying special attention to how the current account relates to these others (\u00a72.2). I then argue that animacy does not have a dedicated locus in the syntax of Plains Cree (\u00a72.3). Rather, animacy is coded in all places that reference is coded. In the verb system, these locations minimally include verbal argument positions (e.g. \u2013am\u2013, \u2013\u00e2\u2013, \u2013\u00ea\u2013). Following the syntactic work of Hirose (2000), D\u00e9chaine (2003), and D\u00e9chaine & Reinholtz (2008) of the verb system of Plains Cree, I schematize this as in (2). 1 While my definition of these two classes is new, I have here adopted the traditional terms for them. I used scare quotes in this thesis to show that the terms themselves do not carry any ontological significance. I have maintained this terminology because these are the names used throughout the literature; if I were to change terminology, comparing different approaches would be made much more difficult. 27 (2) niw\u00e2paht\u00ean ni-w\u00e2p=aht -\u00ea-n 1- see =by.eye.TI-TI-LP \u2018I see itIN\u2019 CP 3 C IP 3 I vP 3 w\u00e2p- vP 3 pro 3 v VP -aht- 3 [EXT] pro 6 V -\u00ea- [EXT] In the nominal system, referential positions minimally include demonstrative positions (e.g. \u2013iki, \u2013ihi), and determiner-related positions (e.g. \u2013a, \u2013ak).2 (3) anihi maskisina an=ihi maskisin-a dst=XT shoe -XT \u2018Those shoesIN\u2019 DEMP 3 DEM EXTP an- 3 pro 3 EXT NUMP -ihi 3 [EXT] NUM NP -a 6 [EXT] maskisin Animacy, then, is distributed across the syntax of Plains Cree. 2 For a consideration of the syntax of demonstratives, please turn to \u00a73.3.2. 28 In section four, I argue that the concept of \u201canimacy\u201d is more precisely defined in terms of extentionality. I show that the two sets of forms (\u201cInanimate\u201d and \u201cAnimate\u201d) map on to the semantic property of extentionality in the following way: (i) The grammatical class traditionally called \u201cInanimate\u201d specifies that the referent never possesses a perspective (i.e. are inherently extentional; [EXT]). (ii) The grammatical class traditionally called \u201cAnimate\u201d is unspecified in its content; it does not specify anything about a referent\u2019s extentionality or intentionality. Thus, Plains Cree has a privative opposition between a form that is specified for extentional content (the \u201cInanimate\u201d) and a form that is unspecified (the \u201cAnimate\u201d). (4) a. \u201cINANIMATE\u201d FORM : [EXT] b. \u201cANIMATE\u201d FORM : [\u00d8] I define an extentional referent as one that cannot be paired with a perspective (cf. \u00a71.3.1). (5) EXT(x) \u2194 \u2200\u03c8\u2200y(R(y,\u03c8) \uf0e0 x \u2260 y) x is Extentional if and only if for all Perspectives \u03c8 and all individuals y, if there is a relation R between individual y with perspective \u03c8, then x is not y. Applying this to the two referential classes of Plains Cree, I claim that the \u201cInanimate\u201d forms will have the content of (6), while the \u201cAnimate\u201d forms will have the content of (7). (6) \u201cINANIMATE\u201d FORM = \u03bbx \u00b7 [EXT(x) \u22c0 PRED'(x)] x, such that x is extentional and x is a member of the set of referents denoted by the predicate. (7) \u201cANIMATE\u201d FORM = \u03bbx \u00b7 [PRED'(x)] x, such that x is a member of the set of referents denoted by the predicate. Finally, in section five I argue that these form-content pairs bear contextual information: a referent\u2019s extentionality is always evaluated within the perspective of some referent. This may be any of three possible perspective possessors: (i) The Speaker (ii) The Hearer (iii) Some prominent third-person (Kuno\u2019s 1972 \u201cDiscourse Perspective\u201d) To adequately model this, I utilize a modified version of the Discourse Representation Theory (DRT: Kamp 1981) proposed by Farkas (1992) and Smith (2003). In this formulation, Discourse 29 Representation Structures (DRS) are taken to represent perspectives (K\u00f6lbel 2002). Consider the Calvin & Hobbes example (\u00a71.3.1). (8) HOBBES IS \u201cINANIMATE\u201d \u146f\u1426\u1456\u1411 \u14a5\u1422\u1472\u1466 \u1426\u140b\u1431\u1422 koht\u00e2wiy miskam Hobbes k-oht\u00e2wiy m=isk -am Hobbes 2-father find=by.body.TI-TI Hobbes \u2018Your father foundINAN Hobbes.\u2019 (Translated S2) Using the proposed model, we can represent the different individual\u2019s perspectives in the following way: (i) When Hobbes is \u201cInanimate\u201d to Speaker, [EXT(Hobbes)] is embedded in the Speaker\u2019s perspective. Speaker x Hobbes father(x,Hearer) EXT(Hobbes) find(x,Hobbes) (ii) When Hobbes is \u201cInanimate\u201d to the Hearer, [EXT(Hobbes)] is embedded in the Speaker\u2019s representation of the Hearer\u2019s perspective. Speaker x Hobbes father(x,Hearer) find(x,Hobbes) Hearer Hobbes EXT(Hobbes) 30 (iii) When Hobbes is \u201cInanimate\u201d to some third person (y), [EXT(Hobbes)] is embedded in the Speaker\u2019s representation of that referent\u2019s perspective. Speaker x y Hobbes father(x,Hearer) find(x,Hobbes) y Hobbes EXT(Hobbes) The modeling of the contextual component of animacy reduces to differential embedding within perspectives. 2.2. Previous accounts In order to understand the discussion of animacy that follows, it is necessary to consider its relation to previous work. In particular, it is necessary to understand previous work on the syntax of Plains Cree, which I review in section 2.2.1. While the current proposal is built from the significant body of previous work on this topic, it generally departs from previous accounts in four crucial ways: (i) The treatment of \u201cInanimate\u201d as the marked member of the opposition (\u00a72.2.2) (ii) The equating of the \u201cInanimate\u201d class with extentionality (\u00a72.2.3) (iii) The treatment of animacy as a syntactic element (\u00a72.2.4) (iv) The data set considered (\u00a72.2.5) 2.2.1. Syntactic accounts of Plains Cree Algonquian languages have a great deal of morphology, which has attracted much interest from many of the most influential scholars of the last century, including Leonard Bloomfield, Mary Haas, Ken Hale, Charles Hockett, C.C. Uhlenbeck, and Morris Swadesh. This work can be (very 31 roughly) broken into two types: (i) American Structuralist accounts and (ii) Generative Grammar accounts, with work in both frameworks continuing to present. In his account of the grammar of Menominee (Central Algonquian, Wisconsin), Leonard Bloomfield (1962) outlines a positional account of the morphosyntax of the verb and nominal which has been influential for all subsequent work on Algonquian. In this account, the system is conceived of as having slots, for which morphemes were specified. When two morphemes were shown to be in complementary distribution, they were assigned the same slot. Following this method, Bloomfield constructs a description of Menominee that includes 10 suffixal positions and two prefixal positions. Subsequent work in Plains Cree (e.g. Wolfart 1973, Dahlstrom 1986) has posited a similar system. One of the crucial generalizations that has come out of this work is the decompositional nature of the Algonquian stem. In particular, verbs are composed of multiple morphemes, conveniently labeled \u201cInitials,\u201d \u201cMedials,\u201d and \u201cFinals\u201d based on their order of occurrence. A sample for Plains Cree is shown in (9). (9) \u1411\u14f4\u146d\u148b\u1426\u148c\u14c0\u1424. w\u00eesakicihc\u00een\u00eaw. w\u00eesak -icihciy- -in- -\u00ea- -w pain -hand- -by.hand- -dir- -3 INITIAL MEDIAL FINAL AFFIX AFFIX \u2018s\/he hurts his\/her hand with his\/her hand.\u2019 (Volunteered S3) The affix that immediately follows the final in transitive verbs (\u2013\u00ea\u2013 above) is labeled a \u201ctheme sign.\u201d Building on these generalizations about the decompositionality of verbal predicates, Hirose (2000) observes that verbs in Plains Cree canonically have one morpheme for each of their arguments; intransitive verbs have one affix beyond the root (\u201cInitial\u201d), transitive verbs have two affixes, and di-transitive verbs sometimes have three.3 Modeling this in a generative syntactic framework (Chomsky 1982, etc.), Hirose proposed that the affixes that code transitivity be located in the two heads of a split VP, each of which introduces an argument (taken to be pronominal in nature; pro). The root is introduced in the complement position of the lowest verb phrase. This is all schematized in (10). 3 This is, of course, a gross characterization. There are numerous cases of single-morpheme verbs in Plains Cree \u2013 a trend which appears to be generalized in Blackfoot. 32 (10) niw\u00e2paht\u00ean ni-w\u00e2p=aht -\u00ea-n 1- see =by.eye.TI-TI-LP \u2018I see itIN\u2019 vP 3 pro 3 v VP -\u00ea- 3 pro 3 V ROOT -aht- w\u00e2p- To account for the linear order of these affixes, Hirose (2000) posited a cyclic raising mechanism; the ROOT raises to the left of the first verbal head V, and then these two raise to the left of the second verbal head v, giving the linear order ROOT > V > v. Considering Algonquian verb syntax in the context of typological work, D\u00e9chaine (2003) proposes modifications to Hirose\u2019s (2000) treatment in order to model the relation between Algonquian verbal structure and that of other languages (e.g. English, Semitic, Salish). In particular, she reorders the morphemes and locates the root as an adjunct to the outer verb phrase (vP in transitive structures). (11) niw\u00e2paht\u00ean ni-w\u00e2p=aht -\u00ea-n 1- see =by.eye.TI-TI-LP \u2018I see itIN\u2019 vP 3 ROOT vP w\u00e2p- 3 pro 3 v VP -aht- 3 pro 6 V -\u00ea- This structure linearizes via phrasal encliticization; each phrase is spelled out as SPEC > HEAD > COMPLIMENT, yielding a linear order ROOT > v > V for transitive verbs. This is the version of Plains Cree argument structure that is adopted in the current thesis. 33 In contrast to the verbal system, the nominal system of Algonquian has received relatively little attention since the first detailed descriptions of Bloomfield (1962), Hockett (1966), and Wolfart (1973). For Plains Cree, the only generative account of nominal morphosyntax that I am aware of is D\u00e9chaine (1999). Working from the generalizations made by Hockett (1966), D\u00e9chaine argues that affixation in the nominal domain can be represented as in (12). (12) nimaskisinin\u00e2na ni-maskisin-n\u00e2n-a 1-shoe -1pl -XT \u2018Our shoes\u2019 (Presented S2) DP 2 D PERSP ni\u2013 2 2 PERS NumP \u2013n\u00e2n 2 2 NUM NP \u2013a 2 2 N maskisin Arguing on the basis of selectional restriction, D\u00e9chaine locates the suffixes for possessor number (e.g. \u2013n\u00e2n) in the head of PersP, introduced in the compliment of D, where the possessor prefix is located (e.g. ni\u2013). Elements that code plurality (e.g \u2013a) are introduced in NumP related to the NP itself (e.g. maskisin). To linearize this, D\u00e9chaine employs two mechanisms: (i) cyclic movement of the NP to the spec of PersP, and (ii) phrasal encliticization. Together, this yields the linear order D > NP > PERS > NUM. The basics of this model are adopted for this thesis, although some modifications will be necessary, which are outlined as introduced. 2.2.2. \u201cInanimate\u201d as semantically specified Most analyses of the \u201cAnimate\/Inanimate\u201d distinction in Algonquian languages take \u201cAnimate\u201d to be the specified, contentful member of the contrast. This approach is exemplified by 34 Dahlstrom (1995:64-65), who builds a picture of \u201cAnimate\u201d in terms of semantic prototypes. In some form or other, most formal linguists have adopted this view, particularly those linguists who have appealed to a \u201cperson hierarchy,\u201d wherein referents are ranked based on their semantic properties, with [\u00b1ANIMATE] playing a crucial role (e.g. Blain (1997), Junker (2005), Bliss (2005), Ritter & Wiltschko (2007), etc.). By contrast, the current account treats \u201cAnimate\u201d as having no semantic properties, while \u201cInanimate\u201d is specified. However, Hockett (1966:62) argues that the systematic shifting of \u201cInanimate\u201d referents to \u201cAnimate\u201d (but not the reverse) points to \u201cInanimate\u201d as the contentful member of the pair, a position that is also echoed by some other linguists (e.g. Wolfart 1973).4 The specified\/unspecified contrast, then, is based on a combination of discoursal and grammatical properties. This account is further developed in the current proposal. A third position, proposed most clearly by Goddard (2004) takes the specification relation to be different along different dimensions. On this account, the specification of an animacy form is a function of the context it is used in. For Goddard (2004), \u201cAnimate\u201d is unspecified semantically, being \u201ca function of contrast with inanimate,\u201d (Goddard 2004:224), but \u201cInanimate\u201d is the general member in many morphosyntactic constructions.5 Put in the terms of this thesis, \u201cInanimate\u201d is unspecified in form, but specified in semantic content. Focusing on semantic specification, the current account is only directly at odds with accounts that adopt \u201cAnimate\u201d as the semantically-specified class (e.g. Dahlstrom 1995, Ritter & Rose 2005). By contrast, the current proposal is consistent with Hockett\u2019s (1966) generalizations; given the semantically-specified status of \u201cInanimate,\u201d it should be easier for nominals to slip into the unspecified class (\u201cAnimate\u201d) than the other way around. 2.2.3. \u201cInanimate\u201d as extentional One can identify three types of semantic proposals that have been put forward to account for animacy patterns in Algonquian. As will be seen, the current proposal, while not completely converging with any of them, assembles the core observations of each line of thought into one unified analysis. One approach to animacy which I call the \u201cconcrete\u201d proposal analyzes Algonquian animacy in terms of the semantic properties that are directly accessible to speakers and 4 See Valentine (2000:118-119) for a detailed discussion of animacy shift in Ojibwa (Central Algonquian). 5 Note, however, that most of Goddard\u2019s (2004) morphosyntactic data is from Fox (Central Algonquian). The facts gathered from that study do not all have direct correlates in Plains Cree, although many do. 35 observers. Proposals of this kind fall into two camps: (i) animacy codes a \u201cliving\/non-living\u201d distinction, and (ii) animacy codes a culturally-defined concept of \u201cpower.\u201d The first proposal is the earlier of the two, with its first form coming in 1634 by LeJeune (Thwaites 1896-1901:7.22- 23), and being subsequently adopted by Eliot (1666). For example, in his grammar of Massachusset (Eastern Algonquian), Eliot (1666), states \u201cThe Animate form or declension is when the thing signified is a living Creature \u2026The Inanimate form or declension of Nouns, is when the thing signified is not a living Creature\u201d (Eliot 1666:10). Of course, even in these early treatments, linguists were well aware that there were significant groups of nouns that were grammatically animate but could not be said to be alive. For example, articles of clothing like socks (13a) and non-living things like dead snakes (13b) are coded as \u201cAnimate\u201d in Plains Cree. (13) a. ASIKAN \u2018SOCK\u2019 = \u201cANIMATE\u201d \u14c2\u141a\u1438\u14ab\u1424 \u140a\u14c7 \u140a\u14ef\u1472\u1423. niw\u00e2pam\u00e2w ana asikan. ni-wap=am -\u00e2 -w an=a asikan 1- see =by.eye.TA-DIR-3 DST=AN.SG sock \u2018I see that sockAN.\u2019 (Presented S2) b. DEAD ANIMALS = \u201cANIMATE\u201d \u14c2\u141a\u1438\u14ab\u1424 \u140a\u14c7 \u146d\u14c0\u1431\u146f \u1401 \u14c2\u1431\u141f. niw\u00e2pam\u00e2w ana kin\u00eapik \u00ea-nipit. ni-wap=am -\u00e2 -w an=a kin\u00eapikw \u00ea- nip=i-t 1- see =by.eye.TA-DIR-3 DST=AN.SG snake C1-die=AI-3 \u2018I see that dead snakeAN.\u2019 (Presented S2) The recognition of this descriptive inadequacy caused a disjunction among theorists, such that one group pursued a more detailed concrete analysis, while others moved away from concrete treatments entirely (see below). Among the remaining concretists, perhaps the most well-known is the work of Darnell and Varnek (1976), which claims that animacy \u201cdeals with power to maintain and balance the universe and to interact with persons and other interactive beings.\u201d Because accounts of this kind have invariably settled on animate as the marked member, assigning it semantics of the sort just described, these accounts have been widely criticized for being forced to make numerous ad hoc claims. For example, what is \u201cpowerful\u201d to a Cree speaker about socks (14a), but not hats (14b)? 36 (14) a. \u2018SOCK\u2019 = \u201cANIMATE\u201d \u1406\u146d \u140a\u14ef\u1472\u14c7\u1420 \u00f4ki asikanak aw=iki asikan-ak PRX=AN.PL sock -PL \u2018these socksAN\u2019 (Presented S2) b. \u2018HAT\u2019 = \u201cINANIMATE\u201d \u00f4hi astotina aw=ihi astotin-a PRX=XT hat -XT \u2018these hatsIN\u2019 A second approach, conveniently termed the \u201cabstract\u201d approach, analyzes Algonquian animacy as coding a kind of semantics that is an abstraction from the observable facts. An account of this kind is suggested by early French-speaking linguists (e.g. Nicolas 1672, in Daviault 1994), who divided nominals between a class called \u201cnoble\u201d and \u201cignoble.\u201d Most often, grammarians defined these terms similarly to \u201canimate\u201d and \u201cInanimate\u201d (i.e. living\/non-living), but sometimes more abstract discussion was offered, referring to concepts like \u201cesteem\u201d or \u201cobjects of consideration.\u201d Goddard (2004) takes this initial step towards a more abstract analysis further, arguing that a useful way to understand Algonquian animacy is to strip the \u201cnoble\/ignoble\u201d dichotomy of its notions of \u201cesteem\u201d and \u201cliving,\u201d leaving only the more abstract notion of \u201chigh\u201d versus \u201clow.\u201d Many of the linguists that ascribe to a \u201cperson hierarchy\u201d (e.g. Blain 1997, D\u00e9chaine and Reinholtz 1999, Junker 2005) must adopt, in some form, this kind of an \u201cabstract\u201d approach to the semantics of animacy. Where the previous accounts attempted to ascribe the semantic value to the two classes, analyses of the third kind, which I label \u201cgrammar\u201d approaches, treat Algonquian animacy as a purely grammatical distinction, with no meaning ascribed to the classes at all. Proponents of this approach point to the apparent inconsistencies in the semantic approach, and draw parallels to the gender systems of Indo-European, which are assumed to be strictly grammatical devices. Frantz (1995), for example, explains that \u201cgrammatical gender is a classification of the noun stems themselves, not of the entitites to which they refer\u201d (Frantz 1991:8). Most linguists that work within a formal framework have adopted this kind of approach (e.g. Blain 1997, Hirose 2000, M\u00fchlbauer 2007). The current account agrees with parts of each of these three basic positions, but is not classifiable as any one of these approaches. Like the \u201cconcrete\u201d position, I take the animacy 37 contrast to have a clear, defineable set of semantics that bear some resemblence to the description offered by Darnell and Varnek (1976). However, the specific definition I have offered is more abstract; animacy situates references within the discourse constructed by the Speaker, placing restrictions on their ability to host embedded propositions (i.e. \u201cInanimate\u201d referents are extentional). This means that the \u201cInanimate\u201d class of referents cannot have a mental state directed at something (i.e. are extentional). This more abstract characterization subsumes previous attempts to characterize \u201cInanimate\u201d as \u201clacking power\u201d or being \u201cignoble.\u201d 2.2.4. Animacy is syntactically determined It is often assumed that the division between \u201cAnimate\u201d and \u201cInanimate\u201d is a lexical division. That is, animacy is an inherent property of stems; stems are either inherently \u201cAnimate\u201d or inherently \u201cInanimate\u201d (cf. Frantz 1995). This can most easily be seen by checking glossing (e.g. Bloomfield 1962, Wolfart 1973, Valentine 2001) and dictionaries (e.g. Wolvengrey 2001), which universally code nominals as either inherently \u201cAnimate\u201d or \u201cInanimate.\u201d This is done so systematically that a nominal form known to regularly appear with both \u201cAnimate\u201d and \u201cInanimate\u201d marking is given two lexical entries in the dictionary. Consider as an example the entry for mistikw \u2018tree\/stick\u2019 in Wolfart and Ahenakew\u2019s (1998) dictionary of Plains Cree. (15) a. mistikw- NA tree, post [sic] [sic:NA] b. mistikw- NI stick, pole, post, log, wooden rail [sic:NI] (Wolfart & Ahenakew 1998:55) Here, the lexical item mistikw has been entered twice, once under \u201cAnimate\u201d (NA) and once under \u201cInanimate\u201d (NI), with notes about speakers shifting from one form to the other (e.g. there is apparently a text example of mistikw occurring in \u201cAnimate\u201d contexts but denoting a \u2018post\u2019). This set of lexicographic conventions has been standard for several hundred years. When formal treatments are developed that involve animacy (e.g. Hirose 2000, Bruening 2001, Branigan & MacKenzie 2002), it is assumed that the nominal is inherently coded for animacy, and the other morphology agrees with it. Syntactic analyses of this kind include both Minimalist (e.g. Bruening 2001) and Relational Grammar treatments (e.g. Frantz 1991). In his discussion of animacy in Menominee (Central Algonquian, Wisconsin), Bloomfield (1962) takes a different view. Bloomfield describes the nominal maeqtek as \u201cAnimate\u201d in its meaning of \u2018tree.\u2019 However, he notes that \u201cthe same word\u201d can be used in 38 \u201cInanimate gender\u201d to describe a \u2018stick\u2019 (Bloomfield 1962:29). By treating the same nominal as occurring in two different \u201cgender\u201d contexts, Bloomfield is treating the nominal as independent of its animacy specification. In treating animacy as a part of the syntax of Plains Cree rather than the lexicon, then, the current account follows this kind of approach. 2.2.5. The data set The current analysis brings to light six new generalizations about animacy: (i) The lack of psych-predicates inflected for \u201cInanimate\u201d referents (\u00a72.4.3.1) (ii) The restriction of reflexive-marked verbs to \u201cAnimate\u201d referents (\u00a72.4.33) (iii) Speaker judgments about the non-specificity of \u201cAnimate\u201d coded referents (\u00a72.4.4) (iv) The correlation between Speaker belief and animacy coding (\u00a72.4.2) (v) The ability of animacy coding to be manipulated for common-ground effects (\u00a72.4.3) (vi) The ability of animacy coding to be manipulated for point-of-view effects in narrative contexts (\u00a72.4.4) Any subsequent account of animacy will have to expand to cover this new data set. 2.3. The form of animacy: Distributed throughout the syntax A consideration of Plains Cree\u2019s grammatical forms shows a basic distinction between \u201cAnimate\u201d forms and \u201cInanimate\u201d forms. The \u201cAnimate\u201d structure in (16a) shows \u201cAnimate\u201d marking within the verb stem (-am-), in its argument-structure (-\u00e2-), and in its pronominal agreement (-w-ak). The demonstrative shows \u201cAnimate\u201d marking in its suffixes (-iki). Finally, the nominal shows animate marking in the suffix attached for plurality (-ak). A comparison of these animate forms in (16a) with their \u201cInanimate\u201d counterparts in (16b) shows the contrast clearly (e.g. -am- vs. -aht- within the verb stem). (16) a. \u201cANIMATE\u201d MORPHOLOGY \u14c2\u141a\u1438\u14ab\u1418\u1420 \u140a\u14c2\u146d \u14aa\u1422\u146d\u14ef\u14c7\u1420 niw\u00e2pam\u00e2wak aniki maskisinak ni-w\u00e2p=am -\u00e2 -w-ak an =iki maskisin-ak 1-bright=by.eye.TA-DIR-3-PL DST=AN.PL shoe -PL \u2018I seeAN thoseAN shoesAN.\u2019 (Presented S2) 39 b. \u201cINANIMATE\u201d MORPHOLOGY \u14c2\u141a\u1438\u1426\u144c\u1423 \u140a\u14c2\u1426\u1403 \u14aa\u1422\u146d\u14ef\u14c7 niw\u00e2paht\u00ean anihi maskisina ni-w\u00e2p=aht -\u00ea -n an =ihi maskisin-a 1-bright=by.eye.TI-TI-LP DST=XT shoe -XT \u2018I seeIN thoseIN shoesIN.\u2019 (Presented S2) The four domains of animacy coding are summarized in Table 2.1 below. VERB DEMONSTRATIVE NOMINAL ROOT FINAL THEME ROOT AFFIX ROOT AFFIX ANIMATE w\u00e2p am \u00e2 an iki maskisin ak INANIMATE w\u00e2p aht am an ihi maskisin a Table 2.1. Summary of animacy coding for verb, demonstrative, and nominal Within models of syntax built on the generative program of Chomsky (1982), there are at least three possible ways to implement a model of animacy forms in Plains Cree, differing in where they locate animacy in the grammatical structure: (i) HYPOTHESIS ONE: A \u201clexical\u201d model. The locus of animacy is the nominal root. The feature [\u00b1EXTENTIONAL] is a specified on nominal roots; animacy does not have any structural locus. (17) a. n\u00e2p\u00eaw : [-EXT] \u2018man\u2019 b. maskisin : [+EXT] \u2018shoe\u2019 These roots are then merged into the syntax bearing this [\u00b1EXTENTIONAL] feature, where their inherent properties trigger agreement with other elements in the syntax (cf. Hirose 2000, Bruening 2001). (18) a. DEMP qi DEM N awa n\u00e2p\u00eaw [-EXT] [-EXT] 40 b. DEMP qi DEM N \u00f4ma maskisin [+EXT] [+EXT] (ii) HYPOTHESIS TWO: A \u201clocal\u201d model. Animacy is confined to a single, dedicated head in the syntax. If this head is in the nominal syntax, it is a classifier feature introduced external to the nominal root (18) (M\u00fchlbauer 2007). (19) a. CLASSP qi CLASS N \u00d8 n\u00e2p\u00eaw [-EXT] [-EXT] b. CLASSP qi CLASS N \u00d8 maskisin [+EXT] [+EXT] If this head is instead located in the verbal syntax, it restricts the argument structure (e.g. Theta-roles) of the verb. (20) a. VP qi ARG V \u00d8 -am- [-EXT] [-EXT] b. VP qi ARG V \u00d8 -aht- [+EXT] [+EXT] On either view, animacy is reduceable to a choice in the content of a single, dedicated position. (iii) HYPOTHESIS THREE: A \u201cdistributed\u201d model. Animacy is coded in all positions that referential distinctions are coded, with no preference for one position over another. 41 (21) \u14c2\u141a\u1438\u1426\u144c\u1423 \u1406\u1426\u1403 \u14aa\u1422\u146d\u14ef\u14c7. niw\u00e2paht\u00ean \u00f4hi maskisina. ni-w\u00e2p=aht -\u00ea-n aw=ihi maskisin-a 1- see =by.eye.TI-TI-LP PRX=XT shoe -XT \u2018I see these shoesIN.\u2019 (Presented S2) qp CP DEMP 3 3 C IP aw- EXTP 3 3 I vP pro 3 3 EXT NUMP ROOT vP ihi 3 w\u00e2p- 3 [EXT] NUM NP pro 3 -a 6 v VP [EXT] maskisin -aht- 3 [EXT] pro 6 V -\u00ea- [EXT] Animacy syntax, then, is a distributed set of operations aimed at coding a referential property. In deciding between these different models, it is important to keep in mind that any set of data can be accounted for within any syntactic theory, if we allow enough additional mechanisms. This means that a discrimination between the three possible models involves the reader\u2019s presuppositions about the kinds of structures available to the grammar. A \u201ccostly\u201d stipulation to one analysis may be taken for granted in another, meaning that we must have a notion of \u201ccost\u201d before we begin to choose among models. In the following discussion, I consider a stipulation to be more \u201ccostly\u201d when it invokes a mechanism that has no surface exponent in the grammar. Thus, the more surface-true a mechanism is, the more desireable it is. With a notion of \u201ccost\u201d now in hand, let us consider what kind of data each of the three analyses covers, and what kind of data it cannot cover. The lexical account requires three things, which are different from the other two accounts: (i) Animacy cannot be actively manipulated (\u00a72.3.1). (ii) Animacy manipulations can only be accomplished by suppletion (\u00a72.3.2). (iii) Plains Cree has a high degree of homophony (\u00a72.3.3). 42 As discussed in \u00a72.3.1 through \u00a72.3.3, the lexical model is undesireable on all three counts. Setting this model aside, we can then consider the requirements of the other two kinds of models. These models differ in several respects: (i) whether or not bare nominals must occur in the scope (i.e. c-commanded by or linear preceded by) of an element that codes animacy (\u00a72.3.4) (ii) whether or not verbal coding can function in the absence of nominal coding (\u00a72.3.5) (iii) whether or not nominal coding can function in the absence of verbal coding (\u00a72.3.6) When the different models are compared to the attested language patterns, along with a notion of cost, it is clear that the distributed hypothesis covers the data with the least amount of extra stipulations. 2.3.1. Animacy is not specified in the nominal If nouns are specified for animacy, then a shift in animacy value should not be possible. This is because an element that lacks structural expression cannot be manipulated via structural operations. Put another way, we expect that the animacy value of a noun - [+EXTENTIONAL] or [-EXTENTIONAL] \u2013 should be a stable property.6 However, it turns out that animacy values may be contextually manipulated. As we have already seen several times, speakers are capable of manipulating animacy coding. For example, although the typical animacy value for \u2013sk\u00e2t \u2018leg\u2019 is \u201cInanimate,\u201d in contexts where the Speaker views legs as intentional, the nominal can be marked as \u201cAnimate.\u201d This is illustrated in (22), with (22a) having the typical \u201cInanimate\u201d value (nisk\u00e2ta \u2018my Inanimate legs\u2019), and (22b) having the animate value (nisk\u00e2tak \u2018my animate legs\u2019). (22) a. \u201cINANIMATE\u201d LEGS \u14c2\u1422\u1473\u1455 \u1406\u1426\u1403 nisk\u00e2ta \u00f4hi ni-sk\u00e2t-a aw =ihi7 1-leg-XT PRX=XT \u2018TheseIN are my legsIN\u2019 (Judgment by S4) CONTEXT (S4): Pointing to legs to identify the proper word for them. 6 This assumes that all nominals are lexically specified to the same degree (i.e. all nouns are either specified for animacy, or all nouns are underspecified). Because the evidence shows that any noun in the language can have its animacy value manipulated, there is no evidence for class cleavage among nominals in terms of their animacy. This makes the assumption of uniform specificity warranted. 7 This parsing is based on stem-internal phonology; aw + i \uf0e0 \u00f4. There are places where this alternation is still unstable, and thus provides crucial evidence: \u00e2sawinamaw- (Minde 1998) vs. \u00e2s\u00f4namaw- (canonical). 43 b. \u201cANIMATE\u201d LEGS \u14c2\u1422\u1473\u1455\u1420 \u1406\u146d nisk\u00e2tak \u00f4ki ni-sk\u00e2t-ak aw =iki 1- leg -PL PRX=PL \u2018TheseAN are my legsAN\u2019 (Judgment by S4) CONTEXT (S4): The speaker\u2019s legs are out of his control, acting on their own. They are kicking him or other people. An example of the manipulation of these animacy contrasts in a discourse is shown by Louis Moosomin\u2019s telling of the \u2018Rolling Head\u2019 story (in Bloomfield 1930:\u00a71).8 In the version of this story told by Moosomin, a woman\u2019s head has been cut off by her husband, but continues to be inhabited by the woman\u2019s mind. The head rolls along the ground, pursuing its children, who flee from it (see \u00a72.5.3 for a more detailed discussion). Throughout the story, the animacy value of the head shifts back and forth from \u201cInanimate\u201d to \u201cAnimate.\u201d A summary of the locations of \u201cInanimate\u201d and \u201cAnimate\u201d shifting for the severed head is given in the table below. Note that in normal Plains Cree speech, the nominal -stikw\u00e2n \u2018head\u2019 is usually framed with \u201cInanimate\u201d morphology (i.e. its \u2018cultural default\u2019 is \u201cInanimate\u201d as in Jaszczolt 2004). Line Event SEVERED HEAD ANIMACY i Severed Head opens eyes INANIMATE ii Severed Head speaks MIX \u2026 iv Severed Head asks Utensils ANIMATE \u2026 x Boy sees Severed Head talking INANIMATE \u2026 xiii Severed Head stopped ANIMATE \u2026 xv Only Severed Head going INANIMATE \u2026 xvii Severed Head passes fire ANIMATE \u2026 xxxii Boy cries that Severed Head kill Boy INANIMATE \u2026 xxxiv Severed Head sees Great Serpent ANIMATE Table 2.2. Summary of animacy shifting for the severed head story 8 This is a story common to Algonquian peoples across North America (Cowan 1980). 44 As can be seen, the animacy classification of the severed head changes seven times in the span of approximately 40 clauses. These shifts are considered in detail in section 2.5.3 below, where I argue that they code shifts in the perspective represented in the discourse. From this evidence, we should conclude that there is active manipulation of the \u201cInanimate\/Animate\u201d contrast to code referential distinctions. The fact that animacy values can shift according to the perspective represented indicates that nounrs are not inherently specified for animacy. This runs counter to the expectations of a lexical treatment of animacy, but is expected by both of the other syntactic accounts. 2.3.2. Animacy contrasts are not suppletive If animacy is a property of nominal stems rather than a property of the syntax, as the lexical theory posits, speakers should not be able to access it to alter its content. Not having a position in the grammatical structure of the language, speakers would have to perform substitution operations, changing one nominal form for another, in order to change animacy values. This would follow patterns of suppletion seen elsewhere in the language, for example the stem miy- \u2018give\u2019 encompasses both the root and final of the verbal complex. (23) a. TRANSITIVE VERB W\u00c2PAM- = 2 MORPHEMES (W\u00c2P- + -AM-) \u141a\u1438\u14a3\u1424. w\u00e2pam\u00eaw. w\u00e2p=am -\u00ea -w see =by.eye.TA-DIR-3 \u2018S\/hePROX sees him\/herOBV.\u2019 (Presented S2) b. TRANSITIVE VERB MIY- = 1 MORPHEME (MIY-) \u14a5\u1526\u1424. miy\u00eaw. miy -\u00ea -w give.TA-DIR-3 \u2018S\/hePROX gave it to him\/herOBV.\u2019 (Presented S2) When the transitivity of this verb form is altered, Plains Cree has to employ completely different verb stem, as shown in (24), where the transitive stem is miy- (24a) and the intransitive (AIT) stem is m\u00eaki- (24b). 45 (24) a. TRANSITIVE VERB MIY- \u2018GIVE\u2019 \u14a5\u1526\u1424. miy\u00eaw. miy -\u00ea -w give.TA-DIR-3 \u2018S\/hePROX gave it to him\/herOBV.\u2019 (Presented S2) b. INTRANSITIVE SUPPLETIVE FORM M\u00caKI- \u2018GIVE\u2019 \u14a3\u146d\u1424. m\u00eakiw. m\u00eaki -w give.AI-3 \u2018S\/heAN gives things.\u2019 (Presented S2) Thus, whenever transitivity is a property of stems instead of affixes, the stem has to be changed when transitivity changes. Supposing that animacy, like transitivity, were a property of stems, predicts that, whenever there is a shift in animacy, there ought to be suppletion of one stem for another. For example, kin\u00eapikw \u2018snake\u2019 (\u201cAnimate\u201d) ought to change to (e.g.) s\u00e2p\u00e2wisk \u2018snake\u2019 (\u201cInanimate\u201d). We can find no evidence for suppletion in Plains Cree. When a speaker alters the animacy of a referent, they employ the same nominal stem but with different morphosyntax. (25) a. \u201cANIMATE\u201d \u2018SHOE\u2019 = maskisin + [-si-]v + [-a]DEM \u140a\u1418 \u14aa\u1422\u146d\u14ef\u14c7 \u14a6\u1426\u146f\u14f2 awa maskisin m\u00eehkosiw aw =a maskisin m\u00eehkw=si-w PRX=IN.SG shoe red =AI-3 \u2018ThisAN shoe is redAN.\u2019 (Presented S2) CONTEXT: Magic shoe that talks to speaker. b. \u201cINANIMATE\u201d \u2018SHOE\u2019 = maskisin + [-\u00e2-]v + [-ima]DEM \u1406\u14aa \u14aa\u1422\u146d\u14ef\u14c7 \u14a6\u1426\u1473\u141d \u00f4ma maskisin m\u00eehkw\u00e2w aw =ima maskisin m\u00eehkw=\u00e2-w PRX=IN.SG shoe red =II-3 \u2018ThisAN shoe is redAN.\u2019 (Presented S2) CONTEXT: Normal shoe laying on the floor. This runs counter to the lexical analysis, which expects that one form or the other of this pair ought to employ a different nominal stem, in parallel to the suppletion pattern seen in verbs. 46 2.3.3. The necessity of homophony Plains Cree appears to be a system that has numerous nominal forms that are readily able to occur with both \u201cAnimate\u201d and \u201cInanimate\u201d marking. FORM \u201cINANIMATE\u201d \u201cANIMATE\u201d asiniy bullet rock \u00e2tay\u00f4hkan sacred story spirit being cik\u00e2hkw\u00e2n lance gambling toy kay\u00e2s\u00eey\u00e2kan old dish Old Dish (person) kistik\u00e2n garden seed maskipit\u00f4n twisted mouth Twisted Mouth mistikw stick tree mistikw\u00e2skisin wooden shoe Dutch person mit\u00e2s leggings pants piw\u00e2piskwastotin steel helmet German w\u00e2pistikw\u00e2n white head person with white hair w\u00e2pimin white berry white bead w\u00e2posway\u00e2n rabbit hide The Rabbitskin People Table 2.3. Forms commonly found in both animacy classes 9 In lexicographic work on Plains Cree, these forms are listed in the dictionary twice; once under \u201cAnimate,\u201d and once under \u201cInanimate\u201d (cf. Wolfart & Ahenakew 1998, Wolvengrey 2001, and all other lexical work done on the language). (26) a. \u00e2tay\u00f4hkan \u140b\u1455\u152b\u1426\u1472\u1423 NA spirit being, spirit power, spirit guardian, spirit animal b. \u00e2tay\u00f4hkan \u140b\u1455\u152b\u1426\u1472\u1423 NI sacred story; legend [cf. \u00e2tay\u00f4hk\u00eawin] (Wolvengrey 2001) The question that arises is whether this lexicographical homophony should be carried over into the analysis of animacy. A number of nominal forms in Cree are found to alter the kind of referent they denote when framed with each of the two sets of nominal morphosyntax. For example, kistik\u00e2n is equivalent to English \u2018farm\u2019 when the \u201cInanimate\u201d set of morphology is applied (27a), but \u2018grain\u2019 when the \u201cAnimate\u201d set is used (27b). 9 Note that the pattern of animate forms correlating with proper names is only accidental; inanimate forms can easily be used as proper names as well. For example, the Plains Cree name for Blaine lake, Saskatchewan is s\u00eehw\u00eeht\u00e2kani- s\u00e2kahikan (lit: salt-lake), a form that is invariably inanimate. 47 (27) a. KISTIK\u00c2N AS \u201cINANIMATE\u201d = \u2018FARM\u2019 \u1401\u146e\u14ab\u1426\u14a5\u14f5\u146d \u1401\u146f\u14c2 \u146d\u1422\u144e\u1473\u14c7 \u00ea-k\u00ee-m\u00e2h-mis\u00e2ki \u00eakoni kistik\u00e2na \u2026 \u00ea- k\u00ee- m\u00e2h-mis=\u00e2-k-i \u00eakoni kistik\u00e2n-a C1-PREV-RED-big =II-0-PL RESM farm -XT \u2018thoseIN farms were bigIN\u2026\u2019 (Minde 1997:\u00a772) b. KISTIK\u00c2N AS \u201cANIMATE\u201d = \u2018GRAIN\u2019 \u14c2\u1472\u14c2\u1450\u1419\u1455\u1419\u1473\u1423 \u140a\u1418 \u146d\u1422\u144e\u1473\u1423 nika-nitaw-\u00e2t\u00e2w\u00e2k\u00e2n awa kistik\u00e2n \u2026 ni-ka- nitaw-at\u00e2waw=ik\u00ea -n aw =a kistik\u00e2n 1- FUT-go- sell =GEN-LP PRX=AN.SG grain \u2018I will go and sell thisAN grain\u2026\u2019 (Ahenakew 2000:\u00a72.3) A similar phenomenon occurs when a nominal is used as the proper name of a person. Thus k\u00eeskihk\u00f4m\u00e2n refers to a kind of \u2018cut off knife\u2019 when marked \u201cInanimate\u201d (28a), but refers to a particular person named \u2018Cut Knife\u2019 (a Sarcee chief) when marked with the \u201cAnimate\u201d set (28b). (28) a. K\u00ceSKIHK\u00d4M\u00c2N \u2018CUT KNIFE\u2019 AS \u201cINANIMATE\u201d \u140a\u14c2\u14aa \u146e\u1422\u1470\u1426\u1470\u14ab\u1423 anima k\u00eeskihk\u00f4m\u00e2n an =ima k\u00eeski-mohk\u00f4m\u00e2n DST=IN.SG cut -knife \u2018thatIN cut knife\u2019 (Wolvengry 2000, S2) b. K\u00ceSKIHK\u00d4M\u00c2N \u2018CUT KNIFE\u2019 AS \u201cANIMATE\u201d \u140a\u14c7 \u146e\u1422\u1470\u1426\u1470\u14ab\u1423 ana k\u00eeskihk\u00f4m\u00e2n an =a k\u00eeski-mohk\u00f4m\u00e2n DST=AN.SG cut -knife \u2018thatAN Cut Knife\u2019 (Wolvengrey 2000, S2) Lest this be assumed to be restricted to a few forms, almost any nominal can be coerced in this way with enough context, as the examples in (29) show. 48 (29) a. NOUN KIN\u00caPIKW \u2018SNAKE\u2019 AS \u201cANIMATE\u201d \u14c2\u146e\u141a\u1438\u14ab\u1418\u1420 \u140a\u1418 \u146d\u14c0\u1431 nik\u00ee-w\u00e2pam\u00e2w awa kin\u00eapik ni-k\u00ee-w\u00e2p =am -\u00e2-w aw=a kin\u00eapikw 1-PREV-see=by.eye.TA-DIR-3 PRX=AN.SG snake \u2018I sawAN thisAN snake\u2019 (Presented S2) CONTEXT (S2): Speaker sees a rubber toy snake, thinks it is a real snake, is frightened. b. NOUN KIN\u00caPIKW \u2018SNAKE\u2019 AS \u201cINANIMATE\u201d \u14c2\u146e\u141a\u1438\u1426\u144c\u1423 \u1406\u14aa \u146d\u14c0\u1431 nik\u00ee-w\u00e2paht\u00ean \u00f4ma kin\u00eapik ni-k\u00ee-w\u00e2p =aht -\u00ea-n aw=ima kin\u00eapikw 1-PREV-see=by.eye.TI-TI-LP PRX=IN.SG snake \u2018I sawIN thisIN snake\u2019 CONTEXT (S2): Speaker sees a snake, then realizes it is just a rubber toy. Here, the use of \u201cInanimate\u201d morphosyntax makes the nominal stem kin\u00eapikw \u2018snake\u2019 refer to a plastic representation of a snake (29b), while framing it with the \u201cAnimate\u201d structure allows it to be construed as an animal (29a). If we transfer the lexicographic decisions to the formal analysis of animacy, Plains Cree would be analyzed as having more homophony in its nominal domain than any language that has ever existed. For example, a count of a small, 203 page dictionary (Wolfart & Ahenakew 1998) yields roughly 933 nominals.10 If the lexical approach is correct, this same set of nominals would have to be doubled to 1,866. Taking into account that every nominal form can be treated as \u201cAnimate\u201d or \u201cInanimate\u201d (27-29), these homophonous nominals would increase exponentially. Another option available is to employ statistical sampling. For example, kin\u00eapikw most often means \u2018a real snake\u2019 in texts, being marked with the \u201cAnimate\u201d set of morphology. Therefore, this nominal is canonically viewed as \u201cAnimate\u201d and listed as such in the lexicon, with the proviso that animacy distinctions may be contextually-shifted. Importantly, this shifting would have to be done in the (morpho)syntax of the language, since this is where all of the mechanisms that overtly code animacy exist (27-29). This means that an analysis of this kind has to posit two layers of modeling in order to account for the data: a lexical component (carrying the inherent specification of the nominal\u2019s animacy), and a syntactic component (carrying the mechanisms for contextually shifting this property). 10 The counting of distinct nominal forms is methodologically problematic; nominals are freely derived from other nominals (mistikw tree to mistikomin \u2018acorn\u2019 to mistikomin\u00e2htikw \u2018oak tree\u2019), and they are also freely derived from verbs (kiskinoham\u00e2k\u00ea- \u2018teach\u2019 to kiskinoham\u00e2k\u00eawin \u2018leaning, education\u2019). Likewise, the demarcation between classes is also doubtful. These numbers, then, are meant only as metrics of comparison. 49 While both of these analytic decisions are, in principal, possible, they come at a cost. Either the system has tens of thousands of homophonous nominals, or it posits two grammatical systems and a set of rules to relate them. The three other analyses, which locate animacy in the syntax, do not have to posit either homophony or lexical-syntactic rules. 2.3.4. Animacy coding does not determine the distribution of bare nouns One possible analysis is that \u201cInanimate\u201d nominals have a classifier-like functional head in their syntax that codes the animacy of the nominal, and by extension, the referent denoted by the nominal (M\u00fchlbauer 2007). However, there is no evidence in Plains Cree for the existence of an overt classifier. Most11 nominals freely occur in a bare, unaffixed form (M\u00fchlbauer 2006). (30) NOUN WITHOUT AFFIXATION \u14aa\u1422\u146e\u1426\u146e \u1401\u146f\u1455 \u1451\u14f0\u1426\u1456\u152e\u1423 \u14c3\u152d\u1424 \u1405\u1426\u148b maskihkiy \u00eakota t-\u00f4s\u00eeht\u00e2yan niyaw ohci, maskihkiy \u00eakota ta- os\u00eeht\u00e2-yan ni-yaw ohci medicine there FUT-make-2s 1 -body from \u2018\u2026for you to make medicine there from my body,\u2019 (AA:4.2) This being the case, the purported classifier is generally null. (31) a. CLASSP qi CLASS N \u00d8 n\u00e2p\u00eaw [-EXT] [-EXT] 11 The only exception to this generalization are nominals that do not meet the minimal prosodic constraints of the language. Nouns of less than two syllables, for some speakers, have an obligatory vowel added (cf. Wolfart 1973). For example, the stem w\u00e2w- \u2018egg\u2019 has an additional vowel added when it occurs in non-suffixed forms; w\u00e2w \uf0e0 w\u00e2wi. This vowel is said to change with the animacy of the nominal; thus maskw- \u2018bear\u2019 becomes maskwa. While this could be interpreted as evidence for the classifier head, two difficulties present themselves: (i) this suffix could just as easily be coding number (singular), as it is taken to do by Wolfart (1973) and in parallel to Blackfoot (Frantz 1970), and (ii) these augment vowels are subject to extreme speaker variation, such that no two speakers I work with have the same organization, and none have the organization described by Wolfart (e.g. S1: w\u00e2ws and maskwa, while S2 has w\u00e2w and maskwa. For S1, the \u2013a deletes in maskwa under suffixation, while for S2 it does not.). Worse, some speakers appear to have multiple strategies (e.g. S2: w\u00e2w but m\u00eehti) 50 b. CLASSP qi CLASS N \u00d8 maskisin [+EXT] [+EXT] If empty categories need to be licensed (Rizzi 1987), this means that this classifier head needs to be within the scope (e.g. c-command) of some other element, in order to receive content, in parallel to other empty categories in natural language (e.g. traces, pro, etc; Chomsky 1982, Longobardi 2004). (32) XP qi X CLASSP [-EXT] qi CLASS N \u00d8 n\u00e2p\u00eaw [-EXT] [-EXT] This would lead us to expect that the distribution of bare nominals (with null classifier heads) should be restricted in some way. In particular, bare nominals should show a different distribution pattern than forms that include overt animacy coding. Taking stock of the candidates for giving this classifier head content, we see that Plains Cree has several sets of elements that code animacy distinctions (cf \u00a72.4.1 above). In particular, one set of these elements affixes to the nominal itself (33). (33) a. \u201cINANIMATE\u201d PLURAL = -a \u14a5\u1429\u141a\u14ef\u14cc \u14aa\u1422\u146d\u14ef\u14c7. miyw\u00e2sinwa maskisina. miyw=\u00e2si-n-w-a maskisin-a good=II -0-3-XT shoe -XT \u2018The shoesIN are nice.\u2019 (Presented S2) b. \u201cANIMATE\u201d PLURAL = -ak \u14a5\u152a\u14ef\u1418\u1420 \u14c8\u142f\u1418\u1420. miyosiwak n\u00e2p\u00eawak. miyw=is-w-ak n\u00e2p\u00eaw-ak good=AI-3-PL man -PL \u2018The menAN are good-looking.\u2019 (Presented S2) 51 In these cases, the nominal bears overt animacy coding, and thus the putative null classifier would be given content local to the nominal syntax.12 (34) a. NUMP qi NUM CLASSP -a qi [+EXT] CLASS N \u00d8 maskisin [+EXT] [+EXT] b. NUMP qi NUM CLASSP -ak qi [-EXT] CLASS N \u00d8 n\u00e2p\u00eaw [-EXT] [-EXT] Since these plural forms carry overt animacy coding, they should have a different distribution than bare forms. There is no evidence for such a difference in distribution; bare nominal forms can occur either before or after animacy coded elements (cf. Reinholtz 1997, Wolvengrey 2003),13 dependent on information structure (cf. Wolvengrey 2003, M\u00fchlbauer 2003, D\u00e9chaine 2007). (35) a. BARE NOUN FOLLOWS VERB: VERBAN > DEMAN > N \u146e\u1439\u1426\u1431\u1424 \u140a\u1418 \u140a\u141a\u14ef\u1422. k\u00ee-p\u00e2hpiw awa aw\u00e2sis. k\u00ee- p\u00e2hpi -w aw =a aw\u00e2sis PREV-laugh.AI-3 PRX=AN.SG child \u2018This childAN laughed.\u2019 (Wolvengrey 2003) b. BARE NOUN PRECEDES VERB: N > DEMAN > VERBAN \u140a\u141a\u14ef\u1422 \u140a\u1418 \u146e\u1439\u1426\u1431\u1424. aw\u00e2sis awa k\u00ee-p\u00e2hpiw. aw\u00e2sis aw =a k\u00ee- p\u00e2hpi -w child PRX=AN.SG PREV-laugh.AI-3 \u2018A\/the childAN here laughed.\u2019 (Wolvengrey 2003) 12 Locating \u2013a in NumP follows D\u00e9chaine (1997). It could, in theory, be located in a classifier phrase, but the difference is not here crucial. For a consideration of the complications of number and extentionality, turn to \u00a76.2. 13 According to Wolvengry (2003), certain orders of noun-demonstrative-verb triples are ruled out (e.g. N V DEM), but the reasons for this are not relevant to the current point about scope. If demonstratives are quantificational (cf. King 2001), these other effects reduce to the scope-taking properties of quantifiers. 52 In this respect, bare nominals behave exactly like affixed forms; both are insensitive to their ordering with respect to other animacy coded elements. (36) a. BARE NOUN FOLLOWS VERB: VERBAN > DEMAN > N-ak \u146e\u1439\u1426\u1431\u1418\u1420 \u1406\u146d \u140a\u141a\u14ef\u14f4\u1420. k\u00ee-p\u00e2hpiwak \u00f4ki aw\u00e2sisak. k\u00ee- p\u00e2hpi -w-ak aw =iki aw\u00e2sis-ak PREV-laugh.AI-3-PL PRX=AN.PL child -PL \u2018This childAN laughed.\u2019 (Presented S2) b. NOUN NOT IN SCOPE OF ANIMACY: N > DEMAN > VERBAN \u140a\u141a\u14ef\u14f4\u1420 \u1406\u146d \u146e\u1439\u1426\u1431\u1418\u1420. aw\u00e2sisak \u00f4ki k\u00ee-p\u00e2hpiwak aw\u00e2sis-ak aw =iki k\u00ee- p\u00e2hpi -w-ak child -PL PRX=AN.SG PREV-laugh.AI-3-PL \u2018A\/the childAN here laughed.\u2019 (Presented S2) This generalization is also supported by a great deal of textual evidence (cf. Wolfart 1996, M\u00fchlbauer 2003); nominals are ordered based on their information structure properties rather than the presence\/absence of animacy-coding affixation. In conclusion, then, it is the universal consensus of linguists who have considered Plains Cree (e.g. Wolfart 1973, Dahlstrom 1991, Wolfart 1996, D\u00e9chaine 1997, M\u00fchlbauer 2003, D\u00e9chaine 2007) that nominals are not restricted in their ordering with respect to elements that carry animacy coding (i.e. demonstratives, verbs). This means that we have no concrete evidence for a null classifier head in the nominal syntax. 2.3.5. Animacy coding is not asymmetric Within current theories of syntax (e.g. Kayne 1994, Chomsky 1995), syntactic structure is inherently asymmetric. For example, merging two syntactic elements together results in a phrase projected from one or the other, but not both. (37) a. \u03b1 V = \u03b2 is complement to \u03b1 \u03b1 \u03b2 b. \u03b2 V = \u03b1 is complement to \u03b2 \u03b1 \u03b2 53 c. \u273d \u03b1\u03b2 V = Neither \u03b1 or \u03b2 is a complement to the other \u03b1 \u03b2 When we consider cases where \u03b1 and \u03b2 share some feature (e.g. plurality), the inherent asymmetry of syntax means that the feature must originate with only one of \u03b1 or \u03b2; one carries the feature and the other matches (\u201cagrees\u201d) with it. The feature-matching element, then, is dependent for its content on the feature-carrying element. Under an agreement model of this kind, an analysis that posits a dedicated head for animacy in either the nominal or verbal syntax predicts a specific kind of asymmetry between nominal and verbal coding of animacy. If animacy is a property of nominal syntax, the verbal exponents of animacy are agreement (i.e. verbs agree with nouns). If, on the other hand, animacy is a property of verbal syntax, the nominal exponents of animacy are agreement (i.e. nouns agree with verbs). The evidence from Plains Cree shows that neither the verb\u2019s animacy coding nor the nominal\u2019s is required for the identification of animacy. The system appears to be evenly weighted; neither nominals nor verbs have an asymmetric dominance in the coding of animacy. The only account that can explain this kind of behaviour is one that divorces animacy from localized coding; the distributed hypothesis is the only one that can model the full breadth of Plains Cree animacy data. 2.3.51. Animacy without nominals If verbs agree with nouns, then nominal syntax should be necessary for the licensing of verbal syntax. This predicts that verbal forms should not be able to have animacy coding without a nominal present. These predictions are not borne out in the data. When there is an overt nominal with overt animacy coding, this asymmetry is not relevant; the nominal gets its animacy value from its local affixation, and the verb reflects this assignment, as shown in (38), where the nominal itself (maskisin \u2018shoe\u2019) carries plural marking that codes animacy (either \u2013a in 38a, or \u2013ak in 38b). 54 (38) a. \u201cINANIMATE\u201d PLURALITY ON NOUN AND VERB \u14aa\u1422\u146d\u14ef\u14c7 \u1401\u14a5\u14f5\u146d maskisina \u00ea-mis\u00e2ki maskisin-a \u00ea- mis=\u00e2-k-i shoe -XT C1-big=II-0-PL \u2018\u2026 (as) the shoesINAN are bigINAN.\u2019 (Presented S2) b. \u201cANIMATE\u201d PLURALITY ON NOUN AND VERB \u14aa\u1422\u146d\u14ef\u14c7 \u1401\u14a5\u14ef\u146d\u144e\u148b\u1420 maskisinak \u00ea-misikiticik maskisin-ak \u00ea-misikiti-t-ik shoe -PL C1-be.big.AI-3-PL \u2018\u2026 (as) the shoesAN are bigAN.\u2019 (Presented S2) As we have seen above (\u00a72.3.4), the case becomes more complex when there is no affixation on the nominal. (39) a. BARE NOMINAL WITH \u201cINANIMATE\u201d VERB MORPHOLOGY \u14aa\u1422\u146d\u14ef\u1423 \u1401\u14a5\u14f5\u1420 maskisin \u00ea-mis\u00e2k maskisin \u00ea-mis=\u00e2-k shoe C1-big=II-0 \u2018\u2026 (as) the shoe is bigINAN.\u2019 (Presented S2) b. BARE NOMINAL WITH \u201cANIMATE\u201d VERB MORPHOLOGY \u14aa\u1422\u146d\u14ef\u1423 \u1401\u14a5\u14ef\u146d\u144e\u141f maskisin \u00ea-misikitit maskisin \u00ea-misikiti-t shoe C1-be.big.AI-3 \u2018\u2026 (as) the shoe is big.\u2019 (Presented S2) The verb\u2019s agreement will now have to provide the content for the null functional head in the nominal syntax. In the cases where there is no overt nominal form present, the case becomes still more complex. 55 (40) a. \u201cINANIMATE\u201d VERB WITHOUT NOMINAL \u1401\u14a5\u14f5\u1420 \u00ea-mis\u00e2k \u00ea-mis=\u00e2 -k C1-big=II-0 \u2018\u2026 (as) it is bigINAN.\u2019 (Presented S2) b. \u201cANIMATE\u201d VERB WITHOUT NOMINAL \u1401\u14a5\u14ef\u146d\u144e\u141f \u00ea-misikitit \u00ea- misikiti -t c1-be.big.AI-3 \u2018\u2026 (as) it is bigAN.\u2019 (Presented S2) Now, we will have to even more heavily rely on the verbal agreement; not only will it have to identify the null head, it will now have to create the head itself out of thin air, similar to the syntax posited for \u2018pro-drop\u2019 systems like Italian (cf. Rizzi 1986, etc.), which use verbal morphology to identify null subjects. While it is possible to construct a model that allows for the verbal agreement to reconstruct the identity of a nominal head present in the discourse but not present locally, the case becomes more difficult when we turn our attention to verbs that never take an overt nominal (Pentland 1996). (41) a. \u201cINANIMATE\u201d VERB WITH EXPLETIVE ARGUMENT \u1401\u146d\u14a5\u1418\u157d \u00ea-kimiwahk \u00ea-kimiwan-k C1-rain -0 \u2018\u2026 (as) it is rainingIN.\u2019 (Presented S2) b. \u201cINANIMATE\u201d VERB WITH EXPLETIVE ARGUMENT CANNOT HAVE OVERT NOMINAL \u273d \u1401\u146d\u14a5\u1418\u157d \u14c2\u1431\u1429 \u273d \u00ea-kimiwahk nipiy \u00ea- kimiwan-k nipiy C1-rain -0 water Intended: \u2018Water is rainingIN.\u2019 (Presented S2) In these cases, there is never any nominal syntax present, but the verb is still inflected for animacy; the \u201cInanimate\u201d form is required (42a), while the \u201cAnimate\u201d form is disallowed (42b). 56 (42) a. EXPLETIVE ARGUMENT MUST BE \u201cINANIMATE\u201d \u1401\u146e\u14ef\u1473\u1420 \u00ea-k\u00eesik\u00e2k \u00ea- k\u00eesik\u00e2 -k C1-be.day-0 \u2018\u2026 (as) it is dayINAN.\u2019 (Presented S2) b. EXPLETIVE ARGUMENT CANNOT BE \u201cANIMATE\u201d \u273d \u1401\u146e\u14ef\u1473\u141f \u273d \u00ea-k\u00eesik\u00e2t \u00ea- k\u00eesik\u00e2 -t c1-be.day-3 Intended: \u2018\u2026 (as) it is dayAN.\u2019 (Presented S2) In these cases, there is no nominal head to reconstruct, leaving the nominal-local hypothesis with a significant problem. If animacy is located in nominal syntax, why is there still animacy when there is no nominal syntax? It is evidence of this kind that is taken to support an analysis in which the verb codes animacy and the nominal agrees with it. Following the argumentation of Jelinek (1984) and Baker (1996) among others, the absence of a nominal is taken to mean that the nominal\u2019s syntax is peripheral to the central verbal syntax. 2.3.52. Animacy without verbs As we have just seen, an analysis that locates animacy in a verbal head (i.e. nouns agree with verbs), makes correct predictions for the behaviour of many noun-verb pairings; nominals are optional or may be non-existent while animacy coding still persists. However, there are also cases in which verbs are optional or non-existent but animacy coding still persists. Thus, the argumentation that best supports the verb-local hypothesis is also the argumentation that defeats it. In connected speech, Plains Cree makes extensive use of verbal ellipsis. In these structures, some previous constituent is gapped in a subsequent, parallel structure. For example, when Alice Ahenakew relates her work with a white family while the mother was away giving birth, she first describes what she did (i.e. she watched the other two girls), then says what the husband did (i.e. he worked outside on the farm), and then repeats, in ellided form, her own tasks. 57 (43) ELISION OF VERB IN DISCOURSE i. \u1401\u146f\u144c \u1401\u147f \u14c2\u152d \u14c2\u1455\u1429 \u140a\u152e\u1423 \u1406\u146d \u140a\u141a\u14ef\u14f4\u1420, \u2026 \u00eakot\u00ea \u00eakwa niya nitay-ay\u00e2n \u00f4k \u00e2w\u00e2sisak kotakak, \u2026 \u00eakot\u00ea \u00eakwa ni-ya nit-ay- ay\u00e2 -n aw -iki \u00e2w\u00e2sis-ak kotak-ak there then 1-body 1- RED-be.at-LP PRX-AN.PL child -PL other-PL \u2018I stayed there \u2026 \u2019 LIT: \u2018I stayed there, these other children, \u2026\u2019 14 ii. \u14c3\u14f1 \u1403\u1422\u1475\u14f1\u1420, \u1401 \u1472\u14c7\u140d\u1528\u14aa\u146d\u1420. n\u00eeso iskw\u00easisak, \u00ea-kanaw\u00eayimakik. n\u00eesw iskw\u00eaw=sis -ak, \u00ea-kanaw=\u00eayim -ak -ik two woman=DIM-PL C1-care=by.mind.TA-1>3-PL \u2018\u2026 and kept the other children, two little girls.\u2019 LIT: \u2018\u2026two girls, taking care of them.\u2019 iii. \u1401\u147f \u140f\u152d \u140a\u1418 \u14c8\u142f\u1424 \u14ab\u14c7 \u1401 \u140a\u1450\u1422\u146b\u141f, \u1401 \u146d\u1422\u144e\u146b\u141f \u140f\u152d\u1411\u144e\u14a5\u1541, \u00eakwa wiy \u00e2wa n\u00e2p\u00eaw m\u00e2n \u00ea-atosk\u00eat, \u00ea-kistik\u00eat wiyaw\u00eetimihk, \u00eakwa w-iya aw =a n\u00e2p\u00eaw m\u00e2na \u00ea-atosk\u00ea-t, \u00ea-kistik\u00ea-t wayaw\u00eetim-ihk and 3-body PRX=AN.SG man usual c1-work-3 c1-farm -3 outside -LOC \u2018The man worked outside, on the farm, \u2026\u2019 LIT: \u2018\u2026and this man worked, farming outside,\u2019 iv. \u1401\u147f \u140a\u152d, \u1401\u146f\u14c2\u1420 \u1401\u147f \u1406\u146d, \u00eakwa aya, \u00eakonik \u00eakwa \u00f4ki, \u00eakwa aya, \u00eakoni -k \u00eakwa aw =iki and HES RESUM-PL and PRX=AN.PL \u2018and I looked after themAN,\u2019 (AA \u00a75:5) LIT: \u2018and uh, theseAN were then the onesAN,\u2019 Here, the form in (iv) is lacking a verb, but it is not lacking animacy coding; the demonstrative forms present bear marking for \u201cAnimate\u201d plural (-iki). This kind of ellipsis can be used with overt nominals (44a) as well as the demonstrative case in (iv). (44) a. ELLIPSIS WITH OVERT NOMINAL Clare \u146e \u14c3\u14a6\u1426\u1403\u1450\u1424, \u1401\u147f \u14a6\u14c7 \u140a\u141a\u14ef\u14f4\u1420. Clare k\u00ee-n\u00eem\u00eehitow, \u00eakwa m\u00eena aw\u00e2sisak. Clare k\u00ee- n\u00eem\u00eehito-w, \u00eakwa m\u00eena aw\u00e2sis-ak Clare PREV-dance -3 and also child-PL \u2018Clare danced, and so did the childrenAN.\u2019 (Presented S2) 14 In order to give the reader a more direct representation of Cree sentence structure, I am here offering more literal translations than are available in the published edition. This is particularly important for the ellipsis example in (iv). The translations done by Ahenakew & Wolfart (1997) are, of course, better for apprehending the broad meaning and intent of the Plains Cree, representing true translation work rather than my rough pony. 58 b. NON-ELLIPSIS FORM Clare \u146e \u14c3\u14a6\u1426\u1403\u1450\u1424, \u1401\u147f \u14a6\u14c7 \u140a\u141a\u14ef\u14f4\u1420 \u146e \u14c3\u14a6\u1426\u1403\u1450\u1418\u1420. Clare k\u00ee-n\u00eem\u00eehtow, \u00eakwa m\u00eena aw\u00e2sisak k\u00ee-n\u00eem\u00eehtowak. Clare k\u00ee- n\u00eem\u00eehito-w, \u00eakwa m\u00eena aw\u00e2sis-ak Clare PREV-dance -3 and also child-PL \u2018Clare danced, and the childrenAN also danced.\u2019 (Presented S2) Here, the verb can be forced to appear in the second clause (44b), but it is considered awkward. Thus, the verb begins to look as \u201coptional\u201d to animacy coding as the noun is. While ellipsis shows verbs to be optional, there are also constructions in which verbs are forbidden. This is most clearly shown by equational sentences, which make use of a postposed demonstrative. (45) EQUATIONAL CONSTRUCTIONS LACK A VERB \u14a5\u14c5\u14f4\u1420 \u1406\u146d min\u00f4sak \u00f4ki min\u00f4s-ak aw =iki cat -PL PRX=AN.PL \u2018TheseAN are catsAN.\u2019 (Translated S1, Presented S2) It is impossible to give a negative example containing a verb, because there is no verb in Plains Cree directly equivalent to English \u2018is\u2019 in these constructions; all verbs available entail some notion of locational presence (e.g. \u2018these cat are here\u2019). These constructions, then, show an instance where animacy coding persists without the possibility of even a null verb form. If animacy is located in the verbal syntax, why is there still animacy when there is no verb? 2.3.6. Conclusion: Animacy is distributed Summarizing, we see that none of the accounts that posit a dedicated locus for animacy can model the data without invoking additional mechanisms. 59 DEDICATED LOCUS FOR ANIMACY \u00a7 DOMAIN ATTESTED LEXICAL LOCAL N LOCAL V NO LOCUS 2.2.1 Active manipulation \u2714 \u273d \u2714 \u2714 \u2714 2.2.2 Suppletion \u273d \u2714 \u273d \u273d \u273d 2.2.3 Homophony - \u2714 \u273d \u273d \u273d 2.2.4 Scope \u273d \u273d \u2714 \u273d \u273d 2.2.51 Nounless animacy \u2714 \u273d \u273d \u2714 \u2714 2.2.52 Verbless animacy \u2714 \u2714 \u2714 \u273d \u2714 Table 2.4. Summary of predictions for three hypotheses A model that posits animacy coding to be distributed across the referential positions of the syntax (VP, DP, DemP, etc.) is most consistent with the data generalizations. 2.4. The content of animacy: \u201cInanimates\u201d are inherently extentionality In the following section, I construct a model in which the \u201cInanimate\u201d class of nominals in Plains Cree bears the semantic feature [EXTENTIONAL], while the \u201cAnimate\u201d class of nominals bears no semantic features at all. Informally, a referent is considered extentional if they do not have a mental process directed at some object (Brentano 1874); an extentional referent cannot offer a point of view for a proposition to be considered from. This means that extentional referents are incapable of having beliefs or attitudes. Philosophers sometimes call these kinds of referents \u201cobjects\u201d rather than \u201csubjects\u201d (Lyons 1982), a use that has entered common language in the notion of \u201cobjectifying\u201d someone.15 These \u201cobjects\u201d cannot perceive anything, believe anything, or be said to think, feel, or speak. I am claiming, then, that the referential class \u201cInanimate\u201d in Plains Cree carries the meaning that the referent is never able to think, feel, or speak \u2013 the referent is just an \u201cobject\u201d out there in the world. More formally, a referent is extentional if (and only if) they cannot be paired with a perspectival domain that provides truth-conditions for propositions. This is represented by the formula in (46) (cf. \u00a71.3.1). 15 Statements of the kind \u201cPornography objectifies human beings\u201d are good examples. Unfortunately, the cross-over of terminology with the more common grammatical notions of \u201csubject\u201d and \u201cobject\u201d make this common usage unappealing for use in linguistic analysis. 60 (46) EXT(x) \u2194 \u2200\u03c8\u2200y(R(y,\u03c8) \uf0e0 x \u2260 y) x is Extentional if and only if for all Perspectives \u03c8 and all individuals y, if there is a relation R between individual y with perspective \u03c8, then x is not y. Extentionality, then, can be thought of as a checking function; it checks all perspective- individual pairs to ensure that none of them contain the referent given to it. As a checking function, extentionality is restrictive; it restricts possible discourse constructions to those that do not have the extentional referent paired with a perspective. Extentionality, then, is a precondition on the construction of subsequent discourse structures \u2013 a property that becomes central when we consider the context of animacy below in section 2.5. Applying the formalism in (46) to \u201cInanimate\u201d nominals, we can say that Plains Cree\u2019s \u201cInanimate\u201d class of nominals are inherently extentional; they can never have a point of view). This yields the formula in (47). (47) a. \u201cINANIMATE\u201d FORM = \u03bbx \u00b7 [EXT(x) \u22c0 PRED'(x)] x, such that x is extentional and x is a member of the set of referents denoted by the predicate. b. \u201cINANIMATE\u201d CONTEXT \u14c2\u14a5\u1422\u146b\u1423 Hobbes. nimisk\u00ean Hobbes. ni-m =isk -\u00ea-n Hobbes 1-find=by.body.TI-TI-lp Hobbes \u2018I foundINAN Hobbes (the stuffed animal).\u2019 (Presented S2) By contrast, Animate nominals have no semantics beyond their basic nominal meanings. Nominals in this class may relate to referents that are interpreted as extentional, or they may be intentional (i.e. offer a point of view); there is no restriction on them in the grammar.16 (48) a. \u201cANIMATE\u201d = \u03bbx \u00b7 [PRED'(x)] x, such that x is a member of the set of referents denoted by the predicate. b. \u201cANIMATE\u201d CONTEXT \u14c2\u14a5\u1422\u1472\u141a\u1424 Hobbes. nimiskaw\u00e2w Hobbes. ni-m=iskaw -\u00e2 -w Hobbes 1-find=by.body.TA-DIR-3 Hobbes \u2018I foundAN Hobbes (the walking, talking tiger).\u2019 (Presented S2) 16 That is, there is no inherent restriction on them. As we will see in Chapter 3, the grammar does place contextual restrictions on them. 61 The difference between \u201cAnimate\u201d and \u201cInanimate,\u201d then, is a difference between a referent who may have intentions (\u201cAnimate\u201d), in some context, and a referent who can never have intentions in any context (\u201cInanimate\u201d). If we reconsider the possible organizations of the two classes, we now see that the organization must be fundamentally asymmetric. The \u201cInanimate\u201d class is contentful, specifying that the referent is extentional, while the \u201cAnimate\u201d class is contentless, specifying nothing at all. (49) a. \u201cINANIMATE\u201d : [Ext] b. \u201cANIMATE\u201d : [\u00d8] The \u201cAnimate\u201d class, then, can be thought of as the general class, employed whenever the inherent extentionality of a referent is desired to be left vague.17 In contrast, the \u201cInanimate\u201d can be thought of as the specific class, employed whenever the inherent extentionality of a referent is needs explicit coding. The use of a general form (\u201cAnimate\u201d) where a specific form (\u201cInanimate\u201d) is expected implies that the general form is carrying the opposite feature; thus, if an \u201cInanimate\u201d form is expected, the use of an \u201cAnimate\u201d form can imply that the referent is inherenty intentional (cf. Wolfart 1978 for obviation; \u00a73.5). Thus, modeling the relation between the two animacy classes as asymmetric results in a logic that balances unrestricted meaning with contrastive meaning. The model I have outlined here makes specific claims about the distribution of animacy in Cree: (i) The analysis requires that \u201cInanimate\u201d nominals do not think, feel, or speak (inherently intentional events; Banfield 1982), because they have been coded as strictly extentional (\u00a72.4.1-2.4.3). (ii) The analysis of \u201cInanimate\u201d as specified and animate as unspecified means that \u201cAnimate\u201d nominals will freely shift their interpretation; \u201cAnimates are vague\u201d (\u00a72.4.4). (iii) The asymmetric relation between a form that carries a specified feature (here, extentional \u201cInanimates\u201d) and one that does not (here, unspecified animates) means that in certain contexts a blocking relationship is established, wherein the unspecified form acquires the opposite meaning of the specified form; for example, via a Gricean implicature as in Grice (1989) (\u00a72.4.5). 17 See Green (1989), among others, for a discussion of purposeful vaguess in English discourse. 62 (iv) The current analysis treats the extentional property of \u201cInanimate\u201d nominals as inherent to the referent. \u201cInanimate\u201d nominals, then, should not be able to identify referents that have ever been intentional, an expectation that the data from Plains Cree corroborates (\u00a72.4.6). 2.4.1. Asymmetries in the interpretation of the two classes In the present analysis, \u201cInanimate\u201d forms carry inherently extentional content, which means that they refer to inherently extentional referents. By contrast, the animate form is unspecified, which means that it should be unrestricted in its interpretation. Recall from section 2.4.3. that there are a number of nominal forms that are commonly found with both animate and \u201cInanimate\u201d marking. A small list of these is repeated below. FORM INANIMATE ANIMATE asiniy bullet rock atay\u00f4hkan sacred story spirit being cik\u00e2hkw\u00e2n lance gambling toy kay\u00e2s\u00eey\u00e2kan old dish Old Dish (person) kistik\u00e2n garden seed maskipit\u00f4n twisted mouth Twisted Mouth mistikw stick tree mistikw\u00e2skisin wooden shoe Dutch person mit\u00e2s leggings pants piw\u00e2piskwastotin steel helmet German w\u00e2pistikw\u00e2n white head person with white hair w\u00e2pimin white berry white bead w\u00e2posway\u00e2n rabbit hide The Rabbitskin People (Repeated) Table 2.3. Forms commonly found in both animacy classes Taking stock of the interpretations18 offered in this list, we see that the \u201cInanimate\u201d coded forms denote entities that are unable to think, speak, or feel. This is corroborated by discussion with native speakers of the language; rabbit hides, sticks, and bullets are considered mere objects, not intelligent entities. Likewise, while some of the animate forms do denote referents capable of thinking, speaking, or feeling (e.g. piw\u00e2piskwastotin \u2018a German,\u2019 \u00e2tay\u00f4hkan \u2018a spirit being,\u2019 etc.), others are not (e.g. kistik\u00e2n \u2018seed,\u2019 mit\u00e2s \u2018pants,\u2019 or w\u00e2pimin \u2018white bead). Being unspecified, native speakers often express confusion about the reasons for referents being 18 Note that by \u201cinterpretation,\u201d I mean more than merely translation. Rather than using interpretation to mean \u201ca set of corresponding English words,\u201d I mean the kind of \u201cthing\u201d out there in the world that the form signifies. Thus, a Plains Cree speaker can say \u201c\u00f4ma asiniy\u201d (\u201cthisIN rock\u201d) and point to a bullet lying on the table. The act of pointing, and the understanding on the part of the speaker-hearer pair is the \u201cinterpretation\u201d of the form, not an English translation of it. 63 marked animate (e.g. S2: \u201cI don\u2019t know why rocks are alive. They aren\u2019t, but I\u2019m talking about them like they are.\u201d). Thus, the \u201cInanimate\u201d class is consistently extentional, while the \u201cAnimate\u201d class is inconsistently intentional. Any nominal can be coded with either animacy class.19 For example, the nominal stem kin\u00eapikw \u2018snake\u2019 is accepted in both forms, as in (50). (50) a. ANIMATE CONTEXT FOR KIN\u00caPIKW \u2018SNAKE\u2019 \u14c2\u146e \u141a\u1438\u14ab\u1424 \u140a\u1418 \u146d\u14c0\u1431\u146f. nik\u00ee-w\u00e2pam\u00e2w awa kin\u00eapik. ni-k\u00ee-w\u00e2p =am -\u00e2-w aw =a kin\u00eapikw 1-PREV-see=by.eye.TA-DIR-3 PRX=AN.SG snake \u2018I saw this snakeAN.\u2019 (Presented S2) b. \u201cINANIMATE\u201d CONTEXT FOR KIN\u00caPIKW \u2018SNAKE\u2019 \u14c2\u146e \u141a\u1438\u1426\u144c\u1423 \u1406\u14aa \u146d\u14c0\u1431\u146f. nik\u00ee-w\u00e2paht\u00ean \u00f4ma kin\u00eapik. ni-k\u00ee-w\u00e2p =aht -\u00ea-n aw=ima kin\u00eapikw 1-PREV-see=by.eye.TI-TI-LP PRX=IN.SG snake \u2018I saw this snakeIN.\u2019 (Presented S2) The same is possible with proper names, as shown in (51). (51) a. ANIMATE CONTEXT FOR HOBBES \u14c2\u146e \u141a\u1438\u14ab\u1424 Hobbes. nik\u00ee-w\u00e2pam\u00e2w Hobbes. ni-k\u00ee- w\u00e2p=am -\u00e2 -w Hobbes 1-PREV-see=by.eye.TA-DIR-3 Hobbes \u2018I saw HobbesAN.\u2019 (Presented S2) b. \u201cINANIMATE\u201d CONTEXT FOR HOBBES \u14c2\u146e \u141a\u1438\u1426\u144c\u1423 Hobbes. nik\u00ee-w\u00e2paht\u00ean Hobbes. ni-k\u00ee- w\u00e2p=aht -\u00ea -n Hobbes 1-PREV-see=by.eye.TI-TI-LP Hobbes \u2018I saw HobbesIN.\u2019 (Presented S2) In these cases, the \u201cInanimate\u201d form correlates with an interpretation of the referent as an extentional \u201cthing\u201d in the world; either a plastic snake, or a stuffed animal. By contrast, the animate form may have intentions (e.g. if the Speaker wants to make the snake or Hobbes think, speak, or feel), but is not required to. 19 See Valentine (2000) and King (1997) for discussion. 64 2.4.2. Abstract nominals are always \u201cinanimate\u201d An extentional referent does not have a perspective. If \u201cInanimate\u201d nominals are extentional, then we expect that nominals that denote abstract concepts should always be \u201cInanimate\u201d: the concepts \u201clove\u201d or \u201cJohn went to the store yesterday\u201d or the event of stubbing my toe on a drawer should never be able to speak, think, or feel. A productive way to form abstract nominals in Plains Cree is to add the suffix \u2013win20 to an intransitive verb stem. This creates derivationally-related sets as in (52). (52) FORMS BUILT OFF OF ROOT s\u00e2k- \u201cLOVE\u201d a. \u14f5\u146d\u1426\u1401\u1424. s\u00e2kih\u00eaw. s\u00e2k =ih -\u00ea -w attach=by.neut-DIR-3 \u2018s\/hePROX loves him\/herOBV.\u2019 (Presented S2) b. \u14f5\u146d\u1426\u1403\u1450\u1418\u1420. s\u00e2kihitowak. s\u00e2k =ih =ito -w-ak attach=by.neut=RECIP-3 -PL \u2018TheyAN love eachother.\u2019 (Presented S2) c. \u14f5\u146d\u1426\u1403\u1450\u140f\u1423. s\u00e2kihitowin s\u00e2k =ih =ito -win attach=by.neut=RECIP-NOM \u2018loveIN\u2019 (Presented S2) These nominalized forms are used to denote the concept associated with the activity, and are frequently used in philosophical discourse. In (53), we see three such forms: kis\u00eaw\u00e2tisiwin \u2018grace,\u2019 w\u00eec\u00eahtowin \u2018living in harmony,\u2019 kitim\u00e2k\u00eayihtowin \u2018compassion.\u2019 (53) \u2026 \u1470\u1426\u1456\u1411\u14c7\u1424 \u1405\u146d\u14ed\u141a\u144e\u14ef\u140f\u1423 \u14a5\u152a \u1411\u1489\u1426\u1450\u140f\u1423 \u146d\u144e\u14ab\u146b\u1528\u1426\u1450\u140f\u1423, \u2026 \u2026 k\u00f4ht\u00e2w\u00eenaw okis\u00eaw\u00e2tisiwin miyo-w\u00eec\u00eahtowin kitim\u00e2k\u00eayihtowin, \u2026 k-\u00f4ht\u00e2wiy-naw o-kis\u00eaw\u00e2t=isi -win miyo-w\u00eec\u00ea=h=ito -win 2-father -21PL 3-merciful=STAT-NOM good-live=by.neut=RECIP-NOM kitim\u00e2k\u00eayihto-win have.pity -NOM \u2018\u2026Our Father\u2019s grace, living in harmony, compassion for one another, \u2026\u2019 (JKN:1.7) 20 It is possible, that this suffix is actualy two morphemes \u2013 the third person suffix -w and a derivational suffix -(i)n. I currently do not have enough evidence to warrant this break, and thus leave it as one unit. 65 Since these nominalized forms denote concepts, rather than entities in the world, they should never be capable of thinking, speaking, or feeling. Being unable to engage in these activities, they cannot offer a perspective for the evaluation of a proposition; they are always extentional. This means that they should always be coded as \u201cInanimate,\u201d which is exactly what Plains Cree does. Consider the examples from elicitation in (54), where a verb that takes an abstract nominal s\u00e2kihitowin \u2018love\u2019 as its argument must be inflected for \u201cInanimate\u201d (-k) rather than \u201cAnimate\u201d (-t) agreement. (54) ABSTRACT NOMINALS MUST BE \u201cINANIMATE\u201d a. \u14f5\u146d\u1426\u1403\u1450\u140f\u1423 \u1401 \u14a5\u1429\u141a\u14ef\u1420. s\u00e2kihitowin \u00ea-miyw\u00e2sik. s\u00e2k =ih =ito -win \u00ea- miyw=\u00e2si-k attach=by.neut=RECIP-NOM C1-good=II-0 \u2018LoveIN is good.\u2019 (Presented S2) b. \u273d \u14f5\u146d\u1426\u1403\u1450\u140f\u1423 \u1401 \u14a5\u152a\u14ef\u141f \u273d s\u00e2kihitowin \u00ea-miyosit s\u00e2k =ih =ito -win \u00ea-miyo =si-t attach=by.neut=RECIP-NOM c1-good=AI-3 Intended: \u2018LoveAN is good.\u2019 (Presented S2) Both in texts and elicitation, there are no attested cases of derived, abstract nominals being treated as animate (cf. Wolfart 1973:23). Speakers can make reference to things larger than events or referents, and in all cases, this reference is coded as \u201cInanimate\u201d. For example, a Speaker can make reference to a proposition, like \u00ea-ahkosit \u2018...that she is sick\u2019 as in (55). (55) PROPOSITIONS ARE \u201cINANIMATE\u201d \u14c2\u144c\u1528\u1426\u144c\u1423 \u1401 \u140a\u1426\u146f\u14ef\u141f. nit\u00eayiht\u00ean \u00ea-ahkosit. ni-it =\u00eayiht -\u00ea -n \u00ea-ahkosi -t 1-thus=by.mind.TI-TI-LP C1-be.sick-3 \u2018I thinkINAN that s\/he\u2019s sick.\u2019 (Translation S2) Lit: \u2018I think thus of itINAN, that s\/he is sick.\u2019 A speaker can also make reference to non-linguistic actions or events in the speech context, such as the dropping of a plate on the floor, as in (56). 66 (56) SPEECH CONTEXTS ARE \u201cINANIMATE\u201d abe, \u140a\u14c2\u14aa \u14a7\u152d \u14a5\u1429\u141a\u14ef\u1423. aben, anima moy miyw\u00e2sin. aben an =ima moya miyw=\u00e2si-n-w EXCL DST=IN.SG NEG good=II -0-3 \u2018Darn, thatINAN is badINAN.\u2019 (Translation S2) Context: Speaker has accidentally dropped a plate and broken it. A speaker can also make reference to entire discourses, such as the purpose for a counselling speech, as in (57), which is elaborated on throughout the following dialogue. (57) DISCOURSES ARE \u201cINANIMATE\u201d \u1401\u1418\u146f \u1406\u14aa \u146e\u147f\u1429, \u1473 \u14c2\u1455\u140d\u1528\u1426\u1455\u1426\u146d\u1420 \u1406\u146d \u1472 \u1411\u1426\u1455\u14aa\u1418\u146d\u1420, ... \u00eawak \u00f4ma k\u00eekway, k\u00e2-nitaw\u00eayihtahkik \u00f4ki ka-w\u00eehtamawakik, \u2026 \u00eawakw aw=ima k\u00eekway, k\u00e2-nitaw =\u00eayiht -am-k-ik aw=iki resum PRX=AN.SG thing c2-towards=by.mind.TI-TI -0-PL PRX=PL ka-w\u00eeht=amaw -ak -ik FUT-tell=APPLIC-1>3-PL \u2018There is somethingIN which they wantIN me to tellIN them \u2026\u2019 (JKN: 1:1) In all of these cases, the referent must be coded as \u201cInanimate.\u201d This is expected under an analysis that treats inanimate morphology as coding extentionality; propositions, speech contexts, and discourses can never, under any circumstances, think, feel, or speak.21 2.4.3. Extentionality in the verb system: \u201cinanimates\u201d can\u2019t think, speak, or feel Verbs systematically code the animacy for their arguments; this coding occurs in two places: (i) In the final element of the stem (the \u201cfinal\u201d), including the \u201cAnimate\u201d final -isi- in (58a) and the inanimate final -\u00e2- in (58b). (58) a. \u201cANIMATE\u201d FINAL: -isi- \u14a5\u1422\u144e\u146f \u1401 \u146d\u14c4\u14ef\u141f. mistik \u00ea-kinosit. mistikw \u00ea- kinw=si-t wood c1-long=AI-3 \u2018The treeAN is tall.\u2019 (Presented S2) 21 In the West, there is a tradition of personifying abstract concepts (e.g. Philosophy visiting Boethius in the form of a woman). Note that this is no longer the concept of philosophy, but rather the concept made into a human being. As an abstract concept, Philosophy cannot talk. 67 b. INANIMATE FINAL: -\u00e2- \u14a5\u1422\u144e\u146f \u1401 \u146d\u14ce\u1420. mistik \u00ea-kinw\u00e2k. mistikw \u00ea- kinw=\u00e2-k wood c1-long=II-0 \u2018The stickIN is longIN.\u2019 (Presented S2) (ii) In the elements that affix to the stem, including pronominal marking like the \u201cAnimate\u201d \u2013t in (59a) and \u2013k (59b). (59) a. \u201cANIMATE\u201d PRONOMINAL \u2013t \u14a5\u1422\u144e\u146f \u1401 \u146d\u14c4\u14ef\u141f. mistik \u00ea-kinosit. mistikw \u00ea- kinw=si-t wood c1-long=AI-3 \u2018The treeAN is tallAN.\u2019 (Presented S2) b. \u201cINANIMATE\u201d PRONOMINAL \u2013k \u14a5\u1422\u144e\u146f \u1401 \u146d\u14ce\u1420. mistik \u00ea-kinw\u00e2k. mistikw \u00ea- kinw=\u00e2-k wood c1-long=II-0 \u2018The stickINAN is longINAN.\u2019 (Presented S2) When we compare the kinds of content verb stems can have with the kind of animacy marking they allow, we find that the gaps in the verbal system\u2019s inventory of stems surface exactly where an extentional account of the nominal semantics would expect them. Whenever the event described by the verb is not specifying anything about intentionality, both inanimate and \u201cAnimate\u201d verb forms exist. Whenever the verb describes a referentless event, the \u201cInanimate\u201d form is the only attested case (\u00a72.4.31). Whenever the event described by the verb is intentional for one of its arguments, verb forms that mark that argument as \u201cAnimate\u201d are the only ones attested (\u00a72.4.32). Finally, reflexive verb forms always specify \u201cAnimate,\u201d rather than \u201cInanimate,\u201d referents (\u00a72.4.33). From this data, I conclude that verbs marked with \u201cInanimate\u201d agreement cannot be ascribers of intentions. 2.4.31. Expletive arguments must be \u201cinanimate\u201d Extentionality is the referential property of not having a perspective to evaluate the truth of a proposition in; extentional referents do not have intentions. Among other things, claiming that the \u201cInanimate\u201d class has inherent extentional content means that whenever there is a need for a 68 strictly syntactically-conditioned argument, the verb will invariably use \u201cInanimate\u201d morphology. Since these expletive arguments lack reference, they can never be associated with an intentional referent. The data from Plains Cree supports this conclusion, both for expletive subjects and expletive objects. Verbs whose subject cannot be construed with a referent are always inflected as agreeing with an \u201cInanimate\u201d argument (e.g. the pronominal suffix \u2013k). This includes: (i) Weather verbs, such as kimiwan \u2018it rains.\u2019 (60) \u1401 \u146d\u14a5\u1418\u1541 \u140a\u14c4\u1426\u1428. \u00ea-kimiwahk anohc. \u00ea- kimiwan-k anohc C1-rain.II -0 now \u2018ItIN is raining today.\u2019 (Presented S2) (ii) Verbs that denote temporal events, such as \u2018it is Tuesday.\u2019 (61) \u1401 \u14c3\u14f1 \u146e\u14ef\u1473\u1420 \u140a\u14c4\u1426\u1428. \u00ea-n\u00eeso-k\u00eesik\u00e2k anohc. \u00ea- n\u00eesw-k\u00eesik\u00e2 -k anohc c1-two- be.day-0 now \u2018ItIN is Tuesday today.\u2019 (Presented S2) (iii) Verbs that give the characteristic properties of something, for example kin\u00eapikosk\u00e2w \u2018there are many snakes there.\u2019 (62) \u1401\u146f\u1455 \u1401 \u146d\u14c0\u1431\u146f\u1422\u1473\u1420. \u00eakota \u00ea-kin\u00eapikosk\u00e2k. \u00eakota \u00ea- kin\u00eapikw=sk =\u00e2-k there C1-snake =HABIT=II-0 \u2018ThereIN are many snakes there.\u2019 (Presented S2) Thus, any verb that denotes a time or a weather condition is systematically coded as \u201cInanimate.\u201d There are also known cases of morphologically transitive verbs that do not appear to ever take an object. Termed \u2018formal objects\u2019 by Bloomfield (1962) and \u2018pseudo-intransitives\u2019 by Goddard (1967), these are forms that inflect for an \u201cInanimate\u201d object, but cannot occur with an overt argument. While these forms are well-attested in Menominee (Bloomfield 1962), in Plains Cree, there are very few attested forms of this kind (cf. Wolfart 1973:39), with the best-known example being shown in (63). 69 (63) FORMAL TI VERB \u14ab\u1426\u140a\u1466. m\u00e2ham. m\u00e2h -am paddle.downriver-TI \u2018s\/he paddles downriver.\u2019 (Wolfart 1973, Wolvengrey 2001) Here, the verb is inflected for an \u201cInanimate\u201d object. According to the consensus among Algonquianists (cf. Bloomfield 1958, 1962, Goddard 1967, Wolfart 1973, Valentine 2001), this phenomenon can only occur with transitive verbs coded for an \u201cInanimate\u201d object. While more restricted than expletive subjects, these expletive objects show the same pattern of \u201cInanimate\u201d only morphology. 2.4.32. Psych verbs prohibit \u201cinanimate\u201d arguments Verbs that denote events that involve thinking, speaking, or feeling can be thought of as introducing a perspective to evaluate the truth of a proposition in. Verbs of these semantic classes are typically grouped under the heading of \u201cpsych verbs\u201d in the literature (Ruwet 1972, 1982, Banfield 1982, Bouchard 1995, etc.), and are thought to be semantically special because they subordinate the proposition with respect to some referent other than the Speaker (Lewis 1979, Cresswell 1985, Mitchell 1987, Farkas 1992). For example, in the sentence \u201cBeth thinks that Norway was a place in Central America,\u201d the embedded proposition is not evaluated with respect to the Speaker\u2019s beliefs, but rather the subject of the psych verb (\u201cBeth\u201d). Thus, these verbs introduce a referent and a referent\u2019s perspective as the relevant domain for the evaluation of the truth of the proposition (cf. K\u00f6lbel 2002). This means that, in the terms of the current analysis, verbs of this kind introduce intentional referents (cf. Bouchard 1995). Plains Cree has numerous verbs of this kind, with an example of each group shown in (64). (64) a. THINKING VERB \u1456\u1445\u146b\u1528\u1426\u1455\u1466. t\u00e2pwak\u00eayihtam. t\u00e2pw\u00eawak=\u00eayiht -am truth =by.mind.TI-TI \u2018s\/he believes itINAN.\u2019 (Volunteered S4) 70 b. SPEAKING VERB \u1403\u1458\u1424 itw\u00eaw it =w\u00ea-w thus=AI -3 \u2018s\/he says\u2019 (Volunteered S2) c. FEELING VERB \u1472\u1422\u146b\u1528\u1426\u1455\u1466. kask\u00eayihtam. kask =\u00eayiht -am desolate=by.mind.TI-TI \u2018s\/he feels lonely\u2019 (Presented S2) Being inherently intentional, these verbs provide an excellent place to test the semantic content of animacy forms. When we cross-classify verbs based both on their stem morphology and their meanings, we observe that no psych verbs introduce an extentional (\u201cInanimate\u201d) argument. In other words, psych verbs, being inherently intentional, prohibit extentional arguments. A preliminary classification is given in the table below. Class INANIMATE ANIMATE Sensory (e.g. \u2018appear\u2019) isin\u00e2kwan- isin\u00e2kosi- Undergo (e.g. \u2018be smoked\u2019) kask\u00e2paht\u00ea- kask\u00e2paso- Emotion (e.g. \u2018x feels well\u2019) \u00d8 miyomaciho- Intellect (e.g. \u2018x remembers\u2019) \u00d8 kiskisi- Speech (e.g. \u2018x speaks\u2019) \u00d8 p\u00eekiskw\u00ea- Table 2.5.. Semantic gaps among verbs For verb forms denoting characteristics of a referent that are observable to an outsider, termed \u2018Sensory verbs\u2019 by Hunter & Karpinski (1994), both \u201cInanimate\u201d and \u201cAnimate\u201d forms are attested. Consider the root mihko- \u2018red.\u2019 This root can be formed into a verb that codes either \u201cAnimate\u201d reference (65a), or \u201cInanimate\u201d reference (65b). 71 (65) a. EXTERNAL CHARACTERISTIC VERB WITH \u201cANIMATE\u201d STEM \u2026\u1401 \u14a5\u1426\u146f\u14ef\u1424. \u00ea-mihkosit \u00ea-mihko=si-t c1-red =AI-3 \u2018\u2026(as) s\/he is redAN.\u2019 (Presented to S2) b. EXTERNAL CHARACTERISTIC VERB WITH \u201cINANIMATE\u201d STEM \u2026\u1401 \u14a5\u1426\u141a\u1420. \u00ea-mihkw\u00e2k \u00ea-mihkw=\u00e2-k C1-red =II-0 \u2018\u2026(as) it is redIN.\u2019 (Presented to S2) However, when we consider psych verbs miyomaciho- \u2018feel well,\u2019 we see that there is no attested form that bears \u201cInanimate\u201d agreement; no dictionary lists one, and I have never seen one in any text. This non-occurrence of intentional verbs with \u201cInanimate\u201d (extentional) referents could, in theory, be accidental. However, due to Plains Cree\u2019s pronominal system, we can construct a test to ensure that it is not. This is because, with verbs that carry specific kinds of clause-typing (conjunct order verbs), the two classes of referents (\u201cAnimate\u201d and \u201cInanimate\u201d) have dedicated pronominal marking. Thus, the form in (66a) has -t for an \u201cAnimate\u201d referent, while the form in (66b) has -k for an \u201cInanimate\u201d referent. This means that the textual pattern can be tested by constucting the form using the available morphology, as in (66b), and presenting it to a speaker.22 (66) a. PSYCH VERB WITH \u201cANIMATE\u201d AGREEMENT \u2026\u1401 \u14a5\u152a\u14ab\u148b\u1426\u1405\u141f. \u00ea-miyom\u00e2cihot \u00ea-miyw=m\u00e2ciho-t c1-good=feel -3 \u2018S\/heAN feels wellAN.\u2019 (Presented to S2) 22 Note that the \u201cinanimate actor\u201d suffix \u2013makan (Wolfart 1973) could also be used to build these constructions. However, this suffix appears to be archaic, since three of the speakers I have worked with have not been able to recognize it, either synthetically or in texts. One speaker (S4) uses it in natural speech and is able to understand it in texts, but I have currently been unable to get him to recognized it in elicitation-type settings. 72 b. PSYCH VERB WITH \u201cINANIMATE\u201d AGREEMENT \u273d \u2026\u1401 \u14a5\u152a\u14ab\u148b\u1426\u1405\u1420. \u273d \u2026 \u00ea-miyom\u00e2cihok \u00ea-miyom\u00e2ciho-k C1-feel.well -0 Intended: \u2018ItIN feels well.\u2019 (Presented to S2) Here, the \u201cInanimate\u201d-marked verb is rejected by the Speaker, in line with the attested corpus patterns. 2.4.34. Reflexive verbs prohibit \u201cinanimate\u201d arguments Another place that we could expect in\/extentionality to surface is the case of reflexives. In many languages (e.g. English, Icelandic, etc.), the reflexive form of a predicate is licensed when its antecedent is interpreted as intentional (i.e. Logophoric; Hag\u00e8ge 1974, Sells 1987, Stirling 1993, etc.). For example, in English, a reflexive form (e.g. itself) cannot be felicitously used as the object of an eventive verb that has an extentional subject (e.g. \u2018a dish\u2019). 23 (67) a. When Shujun tried to get a dish out of the cupboard, she dropped it. b. # When Shujun tried to get a dish out of the cupboard, it dropped itself. c. When Shujun tried to get a dish out of the cupboard, it dropped. Here, the application of -self to the pronominal it in (67b) infelicitously conveys that its antecedent \u2018the dish\u2019 would have to be purposefully breaking itself \u2013 that is, the dish had a desire, the capacity for action necessary to carry out this desire, and then set about a course of action that fulfilled it (a sort of dish suicide). This means that the dish would be construed as intentional. Of course, as many mothers have told their guilty children, \u201cdishes don\u2019t break themselves.\u201d24 In Plains Cree, reflexivization is done on the verbal predicate via the affixation of -iso- to the stem, providing another testing ground for the correlation between intentionality and animacy. 23 Judgments come from multiple speakers of Wisconsin English. Since judgments of speakers of this area are consistently at odds with \u201ccanonical\u201d judgments for English pronominals, this is noted here in case the forms are troubling to some native speakers of other dialects. 24 In a situation in which a vase lies broken on the floor, and the only \u201cadult-licensed\u201d referents are a child and a vase, children sometimes resort to the creation of an additional intermediary referent. For example, one friend, as a child, named her hand \u201cDayday,\u201d and blamed Dayday for the act. That is, the child has imbued some typically \u201cinanimate\u201d referent with intentionality, suitable for ascribing the act to. 73 (68) a. TRANSITIVE VERB \u2026\u1401\u1411\u148b\u1426\u140b\u141f Shujun Martha\u1418. \u2026 \u00ea-w\u00eecih\u00e2t Sujun Marthawa \u00ea-w\u00eec =ih -\u00e2 -t Sujun Martha-a c1-along=by.neut-DIR-3 Shujun horse -XT \u2018ShujunPROX helpedAN MarthaOBV.\u2019 (Presented S2) b. REFLEXIVIZED VERB \u2026\u1401 \u1411\u148b\u1426\u1403\u14f1\u141f Shujun. \u2026 \u00ea-w\u00eecihisot Sujun \u00ea-w\u00eec =ih -iso -t Shujun C1-along=by.neut-RFLX-3 Shujun \u2018ShujunAN helped herself.\u2019 (Presented S2) Reflexive verbs cannot be marked with \u201cInanimate\u201d pronominal marking, and cannot be connected to \u201cInanimate\u201d-marked nominals. (69) a. TRANSITIVE VERB WITH \u201cINANIMATE\u201d OBJECT \u2026\u1401 \u1432\u146f\u14c7\u1541 Shujun \u1406\u14aa \u1411\u152e\u1472\u1423. \u00ea-p\u00eekonahk Sujun \u00f4ma w\u00eey\u00e2kan. \u00ea-p\u00eekw =n -am-k Sujun aw =ima w\u00eey\u00e2kan C1-break=by.hand-TI -0 Shujun PRX=IN.SG dish \u2018Shujun brokeIN thisIN dish.\u2019 (Presented S2) b. \u201cINANIMATE\u201d SUBJECT OF REFLEXIVE IS UNGRAMMATICAL \u273d \u2026\u1401 \u1432\u146f\u14c2\u14f1\u1420 \u1406\u14aa \u1411\u152e\u1472\u1423. \u273d \u00ea-p\u00eekonisok \u00f4ma w\u00eey\u00e2kan \u00ea-pikw =n -iso -k aw =ima w\u00eey\u00e2kan c1-break=by.hand-RFLX-0 PRX=IN.SG dish Intended: \u2018ThisIN dishIN brokeIN itself.\u2019 (Presented S2) Thus, \u201cInanimate\u201d forms are banned in a structure that (may) give rise to an intentional interpretation. 2.4.4. Animates are unspecified The current analysis treats \u201cInanimate\u201d referents as inherently extentional, and animates as unspecified. If this is correct, \u201cAnimate\u201d forms should behave in particular ways: (i) They should be capable of being construed as extentional. (ii) They should be capable of being construed as intentional. 74 (iii) They should be capable of being construed as neither intentional nor extentional. \u201cAnimate\u201d forms, then, should be entirely unspecified for the semantics associated with \u201cInanimate.\u201d \u201cAnimate\u201d referents can be construed as extentional. For example, an \u201cAnimate\u201d referent can be both an animal and dead, as the example in (70) shows. Here, the nominal kin\u00eapikw \u2018snake\u2019 is modified both by an \u201cAnimate\u201d coded demonstrative awa (coded with the \u201cAnimate\u201d singular suffix \u2013a), is the argument of an \u201cAnimate\u201d coded verb nipit \u2018s\/he is dead\u2019 (coded with the \u201cAnimate\u201d pronominal -t), and is the object of a transitive verb w\u00e2pam- \u2018see\u2019 that is coded for an \u201cAnimate\u201d object (the three morphemes -am-, -\u00e2-, and -w). (70) \u201cANIMATE\u201d NOMINAL IN EXTENTIONAL CONTEXT \u14c2\u146e \u141a\u1438\u14ab\u1424 \u140a\u1418 \u146d\u14c0\u1431\u146f. \u1401 \u14c2\u1431\u141f. nik\u00ee-w\u00e2pam\u00e2w awa k\u00een\u00eapik \u00ea-nipit. ni-k\u00ee- w\u00e2p=am -\u00e2 -w aw=a kin\u00eapikw \u00ea-nipi -t 1-PREV-see=by.eye.TA-DIR-3 PRX=AN.SG snake C1-be.dead-3 \u2018I sawAN thisAN deadAN snake.\u2019 (Translation by S2) In this context, it is not possible for the snake to be assigned any kind of intentions. For example, it cannot be made the subject of a psych verb like miyom\u00e2ciho- \u2018be happy.\u2019 (71) \u201cANIMATE\u201d NOMINAL IN EXTENTIONAL CONTEXT # \u14c2\u146e \u141a\u1438\u14ab\u1424 \u140a\u1418 \u146d\u14c0\u1431\u146f. \u1401 \u14c2\u1431\u141f. # nik\u00ee-w\u00e2pam\u00e2w awa k\u00een\u00eapik \u00ea-nipit. k\u00ee-miyom\u00e2cihow. ni-k\u00ee- w\u00e2p=am -\u00e2 -w aw=a kin\u00eapikw \u00ea-nipi -t k\u00ee- miyom\u00e2ciho-w 1-PREV-see=by.eye.TA-DIR-3 PRX=AN.SG snake C1-be.dead-3 PREV-be.happy -3 \u2018I sawAN thisAN deadAN snake. ItAN was happy.\u2019 (Translation by S2) Thus, \u201cAnimate\u201d referents can occur in contexts where they are forbidden to have intentions; they are interpreted as strictly extentional. \u201cAnimate\u201d-marked nominals may be construed as intentional. In fact, as demonstrated with every example in this entire thesis, intentional contexts always require an \u201cAnimate\u201d nominal form. While the requirement of \u201cAnimate\u201d forms for intentional contexts could suggest that \u201cAnimate\u201d forms carry intentional semantics, the numerous \u201cAnimate\u201d forms that cannot be intentional, mean that the relationship between \u201cAnimate\u201d forms and intentionality is more complex. For example, attempting to connect the \u201cAnimate\u201d-marked nominal s\u00eanap\u00e2n \u2018ribbon\u2019 to a verb requiring an intentional argument results in infelicity (72a), even though the morphology is well-formed (72b). 75 (72) a. \u201cANIMATE\u201d AWA S\u00caNAP\u00c2N \u2018THIS RIBBON\u2019 CANNOT BE SUBJECT OF PSYCH VERB # \u140a\u1418 \u14ed\u14c7\u1439\u1423 \u148c\u1426\u146b\u1528\u1426\u1455\u1466. # awa s\u00eanap\u00e2n c\u00eehk\u00eayihtam aw=a s\u00eanap\u00e2n c\u00eehk=\u00eayiht -am-w PRX=AN.SG ribbon joy =by.mind.TI-TI-3 Intended: \u201cThis ribbonAN is happy.\u2019 (Presented S2) b. \u201cANIMATE\u201d AWA AW\u00c2SIS \u2018THIS CHILD\u2019 CAN BE SUBJECT OF PSYCH VERB \u140a\u1418 \u140a\u141a\u14ef\u1422 \u148c\u1426\u146b\u1528\u1426\u1455\u1466. awa aw\u00e2sis c\u00eehk\u00eayihtam aw=a aw\u00e2sis c\u00eehk=\u00eayiht -am-w PRX=AN.SG child joy=by.mind.TI-TI -3 \u201cThis child AN is happy.\u2019 (Presented S2) An \u201cAnimate\u201d-marked nominal can be thought of as a condition on the construction of intentionality, but animacy cannot itself be coding it. Most often, \u201cAnimate\u201d nominals are construed as neither intentional nor extentional. Consider the example in (73). (73) UNSPECIFIED \u201cANIMATE\u201d W\u00c2POS \u2018RABBIT\u2019 \u14c2\u141a\u1438\u14ab\u1424 \u140a\u14c7 \u141a\u1433\u14f1 niw\u00e2pam\u00e2w ana w\u00e2pos ni-w\u00e2pam -\u00e2 -w an =a w\u00e2posw 1- see=by.eye.TA-DIR-3 DST=AN.SG rabbit \u2018I seeAN thatAN rabbit.\u2019 (Presented S2) Here, the Speaker has expressed that they see a distant rabbit. Asking consultants to comment on the mental state or perspective of the rabbit results in confusion; the rabbit\u2019s mental state is irrelevant. That is, stating that I have seen this rabbit does not commit me to any comment at all on their mental state. They may have also seen me, or they may be entirely unaware of the event. As one consultant (S2) put it, \u201cWho knows what rabbits think about?\u201d 2.4.5. Only the \u201cinanimate\u201d form has specification In modeling animacy, I have employed a predicate that restricts a nominal to extentional meanings (EXT(x)), rather than intentional ones (INT(x)). There are two logical possibilities to this categorical opposition; either everything that is not extentional is intentional, or everything that is not extentional is simply unmarked semantically. This is exactly the markedness puzzle 76 found so often in phonological analyses; do we posit a single privative feature of [high], or is it a contrast between two features (e.g. [high] vs. [low]), or a single bivalent feature (e.g. [\u00b1high]) (cf. Steriade 1995)? (74) a. EXTENTIONAL AS A BINARY FEATURE \u201cInanimate\u201d : [+EXT] \u201cAnimate\u201d : [-EXT] b. EXTENTIONAL AS A PRIVATIVE FEATURE \u201cInanimate\u201d : [EXT] \u201cAnimate\u201d : [\u00d8] I here consider what the predictions of these different analyses are, concluding that treating extentionality as a privative feature is more consonant with Plains Cree language data. The predictions of the potential treatments are clear. If the division is binary ([\u00b1EXTENTIONAL]), then the two classes should exhibit symmetric behaviour. This is because each class will have distinct content associated with it. If, however, the division is between one class with a privative feature ([EXTENTIONAL]) and an unmarked form ([\u00d8]), then the two classes will exhibit asymmetric behaviour. This is because one class will have a marked semantic meaning, and the other class will have nothing associated with it. There are ways to test this difference in Plains Cree. Marking a referent with \u201cInanimate\u201d morphosyntax means that it has the property of extentionality; it cannot be said to think, feel, or speak. Thus, in the pair of sentences involving the nominal form kin\u00eapikw \u2018snake,\u2019 the \u201cInanimate\u201d-marked form denotes a plastic snake. (75) \u201cINANIMATE\u201d-MARKED REFERENT \u14c2\u146e \u141a\u1438\u1426\u144c\u1423 \u1406\u146d\u14c0\u1431\u146f. nik\u00ee-w\u00e2paht\u00ean \u00f4ma kin\u00eapik. ni-k\u00ee-w\u00e2p=aht -\u00ea-n aw =ima kin\u00eapikw 1-PREV-see=by.eye.TI-TI-LP PRX=IN.SG snake \u2018I sawIN thisIN snake.\u2019 CONTEXT (S2): Speaker sees a snake, then realizes it is just a rubber toy. This plastic snake has no intentions; it does not think, it does not feel, it does not speak. It is just an object in the world, incapable of having a perspective for propositions. What, then, does the \u201cAnimate\u201d marked form in (76) signify? 77 (76) \u201cANIMATE\u201d-MARKED REFERENT \u14c2\u146e \u141a\u1438\u14ab\u1424 \u140a\u1418 \u146d\u14c0\u1431\u146f. nik\u00ee-w\u00e2pam\u00e2w awa kin\u00eapik. ni-k\u00ee- w\u00e2p=am -\u00e2 -w aw=a kin\u00eapikw 1-PREV-see=by.eye.TA-DIR-3 PRX=AN.SG snake \u2018I sawAN thisAN snake.\u2019 CONTEXT (S2): Speaker sees a rubber toy snake, thinks it is a real snake, is frightened. If it were a full, semantically-marked opposite to \u201cInanimate\u201d, it should signify a referent that has an intention. Thus, the sentence in (76) should mean that the snake is speaking, feeling, or thinking. In fact, this is not the case; the snake in (76) can be entirely intention-less \u2013 even dead (77). (77) \u201cANIMATE\u201d MARKED REFERENT IS DEAD \u14c2\u146e \u141a\u1438\u14ab\u1424 \u140a\u1418 \u146d\u14c0\u1431\u146f. \u1401 \u14c2\u1431\u141f. nik\u00ee-w\u00e2pam\u00e2w awa k\u00een\u00eapik \u00ea-nipit. ni-k\u00ee- w\u00e2p=am -\u00e2 -w aw=a kin\u00eapikw \u00ea-nipi -t 1-PREV-see=by.eye.TA-DIR-3 PRX=AN.SG snake C1-be.dead-3 \u2018I sawAN thisAN deadAN snake.\u2019 (Translation by S2) In such a context, the snake has no intentional state assigned to it whatsoever. It is simply a referent in the world. This interpretive asymmetry between \u201cAnimate\u201d and \u201cInanimate\u201d is robust in Plains Cree; the \u201cAnimate\u201d form is semantically unrestricted in its distribution, whereas the \u201cInanimate\u201d form is semantically restricted. If \u201cAnimate\u201d referents are not inherently intentional, why is it that they often appear to be? For example, the legs of one speaker, normally an \u201cInanimate\u201d referent marked with the plural \u2013a and the demonstrative \u00f4hi in (78a), when marked with \u201cAnimate\u201d morphology (the plural \u2013ak and the demonstrative \u00f4ki), take on a mind of their own (78b). (78) a. \u201cINANIMATE\u201d \u2013SK\u00c2T \u2018FOOT\u2019 : FEET ARE EXTENTIONAL \u14c2\u1422\u1473\u1455 \u1406\u1426\u1403. nisk\u00e2ta \u00f4hi. ni-sk\u00e2t-a aw =ihi 1-leg -XT PRX=XT \u2018TheseIN are my legsIN.\u2019 (Volunteered S4) CONTEXT (S4): Pointing to legs to identify the proper word for them in a \u2018normal\u2019 context. 78 b. \u201cANIMATE\u201d \u2013SK\u00c2T \u2018FOOT\u2019 : FEET ARE INTENTIONAL \u14c2\u1422\u1473\u1455\u1420 \u1406\u146d. nisk\u00e2tak \u00f4ki. ni-sk\u00e2t-ak aw =iki 1-leg -PL PRX=AN.PL \u2018TheseAN are my legsAN.\u2019 (Volunteered S4) CONTEXT (S4): The speaker\u2019s legs are out of his control, acting on their own. They are kicking him or other people. Thus, the \u201cInanimate\u201d morphology in (78a) is contrasted with the \u201cAnimate\u201d set in (78b), with the result that the \u201cAnimate\u201d form in (78b) now denotes a referent who is thinking on their own. Contexts such as those in (78) might lead to the conclusion that \u201cAnimate\u201d referents really are inherently intentional.25 But such accounts fail to account for the contexts where \u201cAnimate\u201d referents are not intentional (e.g. 77). Typically, analysts who view animates as inherently intentional appeal to some notion of \u201cfuzzy\u201d categories (cf. Lakoff 1987, Dahlstrom 1995, Bliss 2005); there are \u201cprototypical\u201d members of the \u201cAnimate\u201d class, and \u201cless- prototypical\u201d members. On this view, animacy is a gradient, scalar property; \u201cAnimate\u201d rocks are \u201cless-prototypical\u201d members of the \u201cAnimate\u201d set, while \u201cAnimate\u201d children are \u201cmore prototypical.\u201d As Goddard (2004:224) points out, the apparent intentionality of animates is a by-product of their contrast with \u201cInanimates.\u201d \u201cAnimate\u201d referents appear to carry intentional meaning only in contexts where they are contrasted with \u201cInanimate.\u201d Returning to the examples in (78), the referent \u2013sk\u00e2ta \u2018legs\u2019 is typically \u201cInanimate.\u201d This means that changing the morphology to \u201cAnimate\u201d is contrastive. Thus, \u201cAnimate\u201d referents only have an intentional meaning when they are set up in contrast with an extentional (\u201cInanimate\u201d) form.26 2.4.6 \u201cInanimate\u201d nominals are inherently extentional In addition to \u201cInanimate\u201d marked nominals being interpreted as extentional, I also claim that they are inherently extentional. An \u201cInanimate\u201d marked nominal has the property of 25 See Bliss (2005) for related discussion of the notion of \u201csentience\u201d in Blackfoot. 26 Hockett (1966) arrived at the same conclusions about the two classes, saying that the \u201cAnimate\u201d class is \u2018absorptive,\u2019 but his reasons for concluding this are, ironically, entirely at odds with the argumentation used in this analysis. Like other Algonquianists, Hockett viewed the \u201cInanimate\/Animate\u201d contrast as a lexical property, specified for each noun. He then set about considering when this lexical property could be changed, and determined that it almost universally went from \u201cInanimate\u201d to \u201cAnimate.\u201d Thus, a linguist arguing on entirely different foundational assumptions arrived at the same asymmetry. Ironically, his conclusion of an inherent asymmetry was what caused the present analysis to be developed. 79 extentionality for all time and in all contexts. In the formalism used here, the claim that extentionality is a stable property of \u201cInanimates\u201d is modeled by the lack of contextual dependency. (79) EXT(x) \u2194 \u2200\u03c8\u2200y(R(y,\u03c8) \uf0e0 x \u2260 y) x is Extentional if and only if for all Perspectives \u03c8 and all individuals y, if there is a relation R between individual y with perspective \u03c8, then x is not y. Without contextual coding, the predicate holds over all contexts. By contrast, a contextual form of extentionality would be formulated as in (80). (80) EXT(x,C) \u2194 \u2200\u03c8\u2200y(R(y,\u03c8,C) \uf0e0 x \u2260 y) x is extentional at context C if and only if for all perspectives \u03c8 and all individuals y, if there is a relation R of y with \u03c8 at C, then x is not y. Here, context-dependency is coded by a context indexical \u2018C\u2019 (cf. Partee 1989). While the form in (80) is possible (and attested: see Chapter 3 on obviation), I take (79) to be the correct formulation for \u201cInanimate\u201d nominals. This means that \u201cInanimate\u201d nominals should lack context sensitivity; once a Speaker codes a referent as \u201cInanimate,\u201d it should not change. Examing Plains Cree data, this inherent extentionality of \u201cInanimates\u201d is exactly what we find; if we consider only one perspective in a discourse, animacy cannot be altered once an value is selected. Animacy forms must stay consistent throughout a discourse. There are no instances in texts or in elicitation of a referent slipping unsystematically between \u201cAnimate\u201d and \u201cInanimate\u201d forms. Attempts to make this happen are always rejected. For example, in (81), I have tried to introduce a referent (kin\u00eapik \u2018snake\u2019) as \u201cAnimate\u201d (coded by the verbal agreement \u2013am\u2013\u00e2\u2013), and then, in a subsequent clause, tried to conjoin an \u201cInanimate\u201d coded verb (\u00ea-m\u00eehkw\u00e2k \u2018it is red\u2019) with an \u201cAnimate\u201d coded one (\u00ea-misikitit \u2018s\/he is big\u2019) to refer to this same referent (kin\u00eapik \u2018snake\u2019). The result is rejected by consultants. (81) SWITCHING ANIMACY IS NOT POSSIBLE # \u14c2\u146e \u141a\u1438\u14ab\u1424 \u146d\u14c0\u1431\u146f. \u1401 \u14a5\u1426\u1481\u1420 \u1401\u147f \u1401 \u14a5\u14ef\u146d\u144e\u141f. # niw\u00e2pam\u00e2w kin\u00eapikw. \u00ea-mihkw\u00e2k \u00eakwa \u00ea-misikitit. ni-w\u00e2p=am -\u00e2 -w kin\u00eapikw. \u00ea-mihkw=\u00e2-k \u00eakwa \u00ea-misikiti-t 1-see=by.eye.TA-DIR-3 snake c1-red =II-0 and C1-be.large-3 Intended: \u2018I sawAN a snake. It was redIN and largeAN.\u2019 (Presented S2) 80 Instead, once a referent is set up as \u201cAnimate\u201d or \u201cInanimate,\u201d all subsequent reference to it must maintain this assignment.27 This is shown by the corrected form offered in (82). (82) ANIMACY MUST BE MAINTAINED IN SUBSEQUENT CLAUSES \u14c2\u146e \u141a\u1438\u14ab\u1424 \u146d\u14c0\u1431\u146f. \u1401 \u14a5\u1426\u146f\u14ef\u141f \u1401\u147f \u1401 \u14a5\u14ef\u146d\u144e\u141f. niw\u00e2pam\u00e2w kin\u00eapikw. \u00ea-mihkosit \u00eakwa \u00ea-misikitit. ni-w\u00e2p=am -\u00e2 -w kin\u00eapikw. \u00ea-mihkw=\u00e2-k \u00eakwa \u00ea-misikiti-t 1-see =by.eye.TA-DIR-3 snake C1-red =II -0 and C1-be.large-3 \u2018I sawAN a snake. It was redAN and largeAN.\u2019 (Presented S2) Animacy is not sensitive to changes in the state of the referent. This can be demonstrated by making reference to a dead person\u2019s body. A dead human being was undeniably intentional in the past (humans think, talk, and feel), but their ability to think, talk, or feel has now ceased.28 In Plains Cree, if I saw the body of a person we both knew, who had now died and was laid in state at a funeral, I must use \u201cAnimate\u201d forms (83a). \u201cInanimate\u201d forms are not allowed (83b). (83) a. DEAD PERSON CAN \u201cANIMATE\u201d \u14c2\u141a\u1438\u14ab\u1424 Heather. nik\u00ee-w\u00e2pam\u00e2w Heather. ni-k\u00ee- w\u00e2p=am -\u00e2 -w Heather 1- PREV-see=by.eye.TA-DIR-3 Heather \u2018I sawAN Heather.\u2019 (Judgment by S2) CONTEXT: Speaker was at Heather\u2019s funeral. b. DEAD PERSON CANNOT BE \u201cINANIMATE\u201d # \u14c2\u141a\u1438\u1426\u144c\u1423 Heather. # nik\u00ee-w\u00e2paht\u00ean Heather. ni-k\u00ee-w\u00e2p=aht -\u00ea -n Heather 1-PREV-see=by.eye.TI-TI-LP Heather \u2018I seeIN Heather.\u2019 (Judgment by S2) CONTEXT: Speaker was at Heather\u2019s funeral. Here, the verb form bearing \u201cAnimate\u201d agreement (w\u00e2pam-\u00e2-) must be used, even though the referent is dead (and thus technically incapable of having an intention). The verb form bearing \u201cInanimate\u201d agreement (w\u00e2paht-am) is not allowed, even if the Speaker is specifically indicating 27 Controlling, of course, for perspectival shifts of the kind discussed in \u00a72.5 below. 28 The question of whether or not a dead person can still have (non-corporeal) intentions is avoided here, by making explicit reference to the corpse of the person, not to the person themselves. In discussions with Plains Cree speakers, it is clear that many people would be willing to consider the referent (as a spirit) capable of intentions even when dead; hence the explicit pointing at the body in the examples. 81 the dead body. In the present analysis, the shifting from \u201cAnimate\u201d to \u201cInanimate\u201d is disallowed because \u201cInanimate\u201d codes an inherent property of the referent (inherently extentional), not a contextual one. This refusal to shift the animacy class of dead people is known from textual sources as well (see Goddard 2004). While animacy is not sensitive to either changes in subsequent discourse or changes in the status of the referent, it is sensitive to shifts in perspectives. For example, if the Speaker\u2019s perspective establishes that a referent is inherently extentional (\u201cInanimate\u201d), but some other person in the discourse believes that the referent is not inherently extentional (\u201cAnimate\u201d), a change in perspective from the Speaker\u2019s to the other person will correlate with a change in animacy. This will be taken up in detail in section 2.4 below. So long as we hold our consideration to only one perspective, the context insensitivity is clear. 2.5. The context of animacy: Relativization to individual perspectives The content of animacy is linked with its context of use (cf. Fillmore 1975, Cook & M\u00fchlbauer 2007): (i) In section 2.4.1 we relied on an alteration in felicitous contexts to discern the content of \u201cInanimate\u201d forms. (ii) In 2.4.4, we used the context insensitivity of \u201cAnimate\u201d forms to argue for its lack of content. (iii) In section 2.4.5, we relied on contexts of use to determine the contentful status of \u201cInanimate\u201d and the contentless status of \u201cAnimate.\u201d (iv) In section 2.4.6, we had to carefully control context in order to consider the inherent properties of animacy. In this section, I consider the contextual properties of animacy in finer detail, arguing that the content of animacy can never be calculated without making reference to someone\u2019s perspective; animacy always codes the beliefs of someone. The shifts in context for animacy, then, are necessarily shifts in perspectives. Modeling the context of animacy thus becomes a task of modelling these perspective shifts. To model these contextual properties, I employ a modified form of Discourse Representation Theory (DRT: Kamp 1981) that is built off of the observations of Farkas (1992) and Smith (2004). Please turn to Chapter 1, section 1.3.2 for a discussion of how this model works. 82 This mechanism models the two core properties of animacy: (i) \u201cInanimate\u201d has extentional content; \u201cInanimates\u201d cannot possess a perspective. (ii) Animacy can be used to represent different individual\u2019s perspectives. In section 2.5.1, I show how the model accounts for the inherent extentional content of \u201cInanimate\u201d predicates and the contextually-defined properties of \u201cAnimate\u201d predicates. I then show how the analysis accounts for the perspectival properties of animacy. First, in section 2.5.2., I show that all nominals are invariably interpreted within a perspective, which is minimally the Speaker\u2019s perspective. In section 2.5.3., I show that nominals can be embedded in the perspective of the hearer. In section 2.5.4., I show that nominals can be embedded within the perspective of third persons (Kuno\u2019s \u201cDiscourse Perspective\u201d \u2013 Kuno 1972, 1987, and similar ideas in Banfield 1982). 2.5.1. \u201cInanimates\u201d as preconditions on embeddings The \u201cInanimate\u201d form, being inherently extentional, carries an inherent disjunction from all individual-perspective pairs (cf. \u00a71.3.1). (84) a. EXT(x) \u2194 \u2200\u03c8\u2200y(R(y,\u03c8) \uf0e0 x \u2260 y) x is Extentional if and only if for all Perspectives \u03c8 and all individuals y, if there is a relation R between individual y with perspective \u03c8, then x is not y. b. \u201cINANIMATE\u201d FORM = \u03bbx \u00b7 [EXT(x) \u22c0 PRED'(x)] x, such that x is extentional and x is a member of the set of referents denoted by the predicate. Adequate modelling of this extentional property is straightforward within the DRT framework I have constructed. Within the current framework, a perspective is represented as a Discourse Representation Structure (DRS), with dependent perspectives represented as embedded inside of the outermost perspective. In selecting the \u201cInanimate\u201d form, then, a Speaker is placing a restriction on possible perspectival embeddings, making the DRS in (85) ill-formed because it says both that the referent kin\u00eapik \u201csnake\u201d is inherently unable to possess a perspective (extentional) and possesses a perspective. 83 (85) ASSIGNING A PERSPECTIVE TO AN \u201cINANIMATE\u201d IS NOT ALLOWED \u273d \u140a\u14c2\u14aa \u146d\u14c0\u1431\u146f \u148c\u1426\u146b\u1528\u1426\u1455\u1466. \u273d anima kin\u00eapik c\u00eehk\u00eayihtam. an =ima kin\u00eapik c\u00eehk=\u00eayiht -am PRX=IN.SG snake joy =by.mind.TI-TI Intended: \u2018That (plastic) snakeIN feels happy.\u2019 (Presented S2) Speaker x EXT(x) snake(x) x happy(x) By contrast, the \u201cAnimate\u201d form carries no restrictions, allowing it to be freely identified with perspectives, should the Speaker desire it. (86) \u140a\u1418 \u146d\u14c0\u1431\u146f \u148c\u1426\u146b\u1528\u1426\u1455\u1466. awa kin\u00eapik c\u00eehk\u00eayihtam. aw =a kin\u00eapik c\u00eehk=\u00eayiht -am PRX=AN.SG snake joy =by.mind.TI-TI Intended: \u2018ThatAN snake feels happy.\u2019 (Presented S2) Speaker x snake(x) x happy(x) 84 Inanimacy, then, can be seen as carrying a pre-condition on the construction of discourses; it limits the possible embeddings to \u201cAnimate\u201d (non-extentional) forms. 2.5.2. Relativizing to the Speaker: Changes in belief In the present analysis, animacy values are perspectival in nature. If a Speaker views an entity as being inherently extentional (i.e. as inherently lacking a perspective), then the referent will be classified as \u201cInanimate.\u201d If a Speaker instead views this same referent as either possessing a perspective (i.e. intentional) or unspecified for perspectival properties, then the referent will be classified as \u201cAnimate.\u201d Consider a minimal pair of sentences involving a plastic snake. In this context, I had a brightly colored rubber snake sitting on a shelf in my house. The speaker caught sight out of it out of the corner of her eye and thought it was a real snake for a moment. She gasped, but, upon closer inspection realized that the snake was, in fact, a piece of rubber. I asked her how she would say this in Plains Cree. When the speaker first thinks the snake is \u2018real,\u2019 the \u201cAnimate\u201d system of morphology is used (87). (87) SPEAKER THINKS SNAKE IS REAL = \u201cANIMATE\u201d \u14c2\u146e \u141a\u1438\u14ab\u1424 \u140a\u1418 \u146d\u14c0\u1431\u146f. nik\u00ee-w\u00e2pam\u00e2w awa kin\u00eapik ni-k\u00ee- w\u00e2p=am -\u00e2 -w aw =a kin\u00eapikw 1-PREV-see=by.eye.TA-DIR-3 PROX=AN.SG snake \u2018I sawAN thisAN snake\u2019 (Judgment by S2) Later, when she realized that it was not a real snake, but rather a rubber representation, she shifts the nominal to the \u201cInanimate\u201d set of morphology (88).29 29 In this context, an alternative form was also offered, which employs the suffix \u2013hk\u00e2n \u2018pretend or created form of x\u2019 as shown in (i): (i) \u14c2\u141a\u1438\u1426\u144c\u1423 \u1406\u14aa \u146d\u14c0\u1431\u146f\u1426\u1473\u1423. niw\u00e2paht\u00ean \u00f4ma kin\u00eapikohk\u00e2n. ni-w\u00e2p=aht -\u00ea -n aw=ima kin\u00eapikw-hk\u00e2n 1-see=by.eye.TI-TI-LP PRX=IN.SG snake -constructed \u2018I sawIN thisIN fake snake.\u2019 (Volunteered S2) Both forms with and without \u2013hk\u00e2n are considered acceptable, but (i) is considered more specific (and thus preferred). 85 (88) SPEAKER REALIZES SNAKE IS PLASTIC = \u201cINANIMATE\u201d \u14c2\u146e \u141a\u1438\u1426\u144c\u1423 \u1406\u14aa \u146d\u14c0\u1431\u146f. nik\u00ee-w\u00e2paht\u00ean \u00f4ma kin\u00eapik. ni-k\u00ee- w\u00e2p=aht -\u00ea-n aw =ima kin\u00eapikw 1-PREV-see=by.eye.TI-TI-LP PROX=IN.SG snake \u2018I sawIN thisIN snake.\u2019 (Judgment by S2) Here, the Speaker begins by thinking the snake is a real snake, and frames it with \u201cAnimate\u201d morphology, which conveys nothing about its mental state (\u201cAnimate\u201d is unspecified). Later, she realizes her mistake \u2013 the snake is really just made out of plastic and incapable of trying to bite her \u2013 and she thus changes the coding of the referent to \u201cInanimate.\u201d Crucially, this shift from \u201cAnimate\u201d to \u201cInanimate\u201d happened because the Speaker altered their understanding of the referent. The shift from \u201cAnimate\u201d to \u201cInanimate\u201d or \u201cInanimate\u201d to \u201cAnimate,\u201d then, is necessarily a shift in the speaker\u2019s conceptions of the inherent properties of the referent. This alteration in the Speaker\u2019s perception of the referent\u2019s properties can be modeled in a straightforward way. In the \u201cAnimate\u201d form in (87), the referent is simply a referent; there are no special properties ascribed to the snake by the Speaker. Thus, in the DRS model, the referent kin\u00eapik \u2018snake\u2019 is embedded within the Speaker\u2019s perspective; it is true in the Speaker\u2019s perspective that they saw a snake. This matches the interpretation given by the native speaker (S2). (89) Speaker x snake(x) see(S,x) In the \u201cInanimate\u201d form in (88), the snake is coded as inherently extentional. In the DRS model, then, the extentional content of the \u201cInanimate\u201d form (EXT(x)) is embedded within the Speaker\u2019s perspective. This says that the Speaker is claiming that it is true in their perspective that they saw a snake and the snake was an extentional entity (e.g. plastic). This also matches the interpretation given by the native speaker (S2). 86 (90) Speaker x snake(x) EXT(x) see(S,x) The current model, then, successfully covers the contextual properties of animacy when relativized to the Speaker. 2.5.3. Relativizing to hearers: effects of common ground While it is generally assumed that speech act participants will seek to reach a \u2018common ground\u2019 where the knowledge of both participants are brought into harmony (Stalnaker 1974), there are times where this simply cannot happen. For example, if one participant has had an experience that the other participant believes to be impossible, the two will not be able to come to a \u2018common ground\u2019 for the event. This kind of asymmetry in belief is a good place to test relativization to referents other than the Speaker. In Plains Cree, events experienced only by the Speaker require that the Speaker\u2019s beliefs be coded. Consider a context in which I am alone in my room and suddenly a red shoe on the floor starts speaking to me. Later, I want to tell a friend about this strange experience. In Plains Cree, the shoe must be marked with \u201cAnimate\u201d forms in such a context, as the example in (91) shows (maskisin \u2018shoe\u2019 is modified by the animate demonstrative awa \u2018this one\u2019 and is the argument of a TA verb p\u00eekiskw\u00e2t- \u2018talk\u2019). (91) A SPEAKING SHOE MUST BE \u201cANIMATE\u201d \u140a\u1418 \u14a5\u1426\u146f \u14aa\u1422\u146d\u14ef\u1423 \u1401 \u1432\u146d\u1422\u1481\u14ef\u141f. awa mihko-masksin \u00ea-p\u00eekiskw\u00e2sit aw =a mihkw-maskisin \u00ea- p\u00eekiskw\u00e2s-it PRX=AN.SG red- shoe c1-speak.TA -3>1 \u2018This red shoeAN talked to me.\u2019 (Judgment S2) Here, \u201cAnimate\u201d forms are used, because the Speaker perceives the referent to be capable of speech. The hearer, not being present at this event, is simply going to have to accommodate the Speaker\u2019s beliefs about this event, regardless of what they may personally think about it. 87 Suppose my friend walks in immediately after (or interrupting) the event of this shoe speaking to me, and sees me looking stunned, with a red shoe on the floor. While I believe the shoe has spoken to me, my friend sees only a regular shoe laying there. In this context, I have to make a decision about whose perspective to represent \u2013 my own, in which the shoe was an \u201cAnimate\u201d referent that was capable of talking, or my friend\u2019s, in which the shoe was just a shoe \u2013 an inherently extentional referent (\u201cInanimate\u201d), incapable of speaking. If I want to accommodate my friend\u2019s perspective on this shoe in Plains Cree, the shoe can now be marked \u201cInanimate\u201d via an inanimate demonstrative \u00f4ma, and a derived intransitive verb stem can be used to talk about it (p\u00eekiskw\u00eaw\u00e2t-). (92) INANIMATE SHOE TO ACCOMMODATE HEARER\u2019S PERSPECTIVE \u1406\u14aa \u14a5\u1426\u146f \u14aa\u1422\u146d\u14ef\u1423 \u1401 \u1432\u146d\u1422\u1475\u141a\u144e\u141f. \u00f4ma mihko-masksin \u00ea-p\u00eekiskw\u00eaw\u00e2tit aw =ima mihkw-maskisin \u00ea- p\u00eekiskw\u00ea-w\u00e2-t -it PRX=IN.SG red- shoe c1-speak.AI -IN-EP-3>1 \u2018ThisIN red shoe talkedIN to me.\u2019 (Judgment S2) Here, \u201cInanimate\u201d forms are used, because the hearer conceives of the shoe as extentional (not having heard it speak, and assuming things about shoes that people generally assume about them). The Speaker, who heard the shoe speak, suppresses their own version of the experience to accommodate the hearer\u2019s version.30 Animacy can also be relativized to the hearer when the referent and the event are both known equally well to the Speaker and Hearer. Consider the Calvin and Hobbes examples discussed in the first chapter. Calvin is a child who has a tiger for a best friend; this tiger is taken to be a living, talking tiger by Calvin, but is taken to be a stuffed animal by everybody else. In a context where Hobbes is lost in the woods, and found by the parent, Calvin and the rest of the world have clashing beliefs about the event. This presents the parent with a complex task in Plain Cree; when informing Calvin of the event of finding Hobbes, should they employ \u201cInanimate\u201d (93a) or \u201cAnimate\u201d (93b) forms for Hobbes? 30 It is also possible that the Speaker has here aligned their own beliefs to the Hearer\u2019s (e.g. upon considering what the Hearer would say, the Speaker became convinced that they had experienced some kind of delusion). In that case, the accomodation would be complete; the hearer\u2019s perspective would take over the Speaker\u2019s entirely. 88 (93) a. SPEAKER PRESENTS SPEAKER\u2019S PERSPECTIVE \u146f\u1426\u1456\u140f\u1429 \u146e \u14a5\u1422\u1472\u1466 Hobbes. koht\u00e2wiy k\u00ee-miskam Hobbes. k-oht\u00e2wiy k\u00ee- m =isk -am Hobbes 2-father PREV-find=by.body.TI-TI Hobbes \u2018Your fatherAN found HobbesIN.\u2019 (Presented S2) Context: Mother speaking to Calvin. b. SPEAKER PRESENTS HEARER\u2019S PERSPECTIVE \u146f\u1426\u1456\u140f\u1429 \u146e \u14a5\u1422\u1472\u140d\u1424 Hobbes\u140a. k\u00f4ht\u00e2wiy miskaw\u00eaw Hobbesa. k-oht\u00e2wiy m =iskaw -\u00ea -w Hobbes-a 2-father find=by.body.TA-DIR-3 Hobbes-XT \u2018Your fatherPROX foundAN HobbesOBV.\u2019 (Presented S2) Context: Mother speaking to Calvin. Here, the choice of animacy forms depends on whose beliefs the Speaker desires to represent. According to the consultant (S2), if Calvin\u2019s mother uttered (93a) to Calvin, she would be over- riding Calvin\u2019s belief that Hobbes was real with her own, as if \u201cshe were telling him there were no Santa Claus.\u201d Thus, (93a) represents the belief of the Speaker (Calvin\u2019s mother). By contrast, if Calvin\u2019s mother uttered (93b), she would be accomodating Calvin\u2019s belief that Hobbes was real, \u201cbecause he is a child and you want to respect what he believes.\u201d Thus, (93b) represents the beliefs of Calvin, rather than his mother. These kinds of relativizations can be represented straightforwardly in the current model of animacy context. The accommodation of the Speaker\u2019s account of the talking shoe simply relativizes the event to the Speaker\u2019s perspective. This is represented by embedding the predicates associated with the shoe (e.g. shoe(x)) in the Speaker\u2019s perspective. The model says that the Speaker is saying that there was a shoe, it was red, and it spoke to them. This matches the judgments of the consultants (S2). (94) Speaker x shoe(x) red(x) speak(x,S) 89 Likewise, the Speaker\u2019s accommodation of the hearer\u2019s beliefs can be represented by an embedding of what the Speaker takes the hearer\u2019s perspective to be inside their own perspective. Here, the model says that the Speaker says that there was a shoe, it was red, it spoke to them, and the hearer conceives of this shoe as inherently extentional (i.e. \u201cInanimate\u201d). (95) Speaker x shoe(x) red(x) Speak(x,S) Hearer x EXT(x) The same model can also cover the Calvin and Hobbes cases in (93). To represent the Mother\u2019s beliefs about the Hobbes, we embed the extentional coding of Hobbes within the Speaker\u2019s perspective, as in (96). This says that it is true in the Speaker\u2019s perspective that Hobbes was found by the hearer\u2019s father and Hobbes is an inherently extentional entity (\u201cInanimate\u201d). (96) Speaker x Hobbes father(x,Hearer) Ext(Hobbes) find(x,Hobbes) In the example in (93b), the Mother is accommodating Calvin\u2019s conception of the event; Calvin\u2019s father bumped into Hobbes in the woods while Hobbes was out looking for a tuna fish sandwich and they walked home together, chatting as they went. We can model this in a DRS by embedding the event inside Calvin (the Hearer\u2019s) perspective, as in (97). This says that it is true in the Speaker\u2019s perspective that the hearer\u2019s father found Hobbes, and it is also true in the hearer\u2019s perspective that hearer\u2019s father found Hobbes. 90 (97) Speaker x y Hobbes father(x,Hearer) find(x,Hobbes) Hearer x Hobbes find(x,Hobbes) The current model, then provides a logic for modelling the complex shifting of Speaker\/Hearer perspectives possible with animacy. 2.5.4. Relativizing to third persons: perspective shifts The current model treats the extentionality of a referent as relativized to a referent. In the cases considered so far, this has been either the Speaker (section 2.5.2) or the Hearer (section 2.5.3). Based on Kuno\u2019s (1972, 1987) work on discourse-prominent third persons in English and Japanese, we could expect that Plains Cree would be able to invoke a \u201cDiscourse Perspective\u201d some perspective other than the Speaker or Hearer\u2019s. Animacy can be relativized to a referent who is not present for the speech act, but whose perspective is considered relevant. Consider again the situation of Calvin losing his tiger Hobbes. If his father finds the tiger in the woods and reports it to Calvin\u2019s mother, while Calvin is not around, he can choose to either represent his own beliefs about Hobbes or Calvin\u2019s. In (99a), he uses \u201cInanimate\u201d marking on the verbal predicate (-isk-\u00ea-), which codes that Hobbes is inherently extentional (Father\u2019s Perspective). In (98b), he uses \u201cAnimate\u201d marking on the verbal predicate (-iskaw-\u00e2-), which codes that Hobbes is capable of possessing a perspective (Calvin\u2019s Perspective). 91 (98) a. FATHER PRESENTS OWN PERSPECTIVE, HOBBES = \u201cINANIMATE\u201d \u14c2\u146e \u14a5\u1422\u146b\u1423 Hobbes. nik\u00ee-misk\u00ean Hobbes. ni-k\u00ee- m=isk-\u00ea-n Hobbes 1-PREV-find=by.body.TI-TI-LP Hobbes \u2018I foundIN Hobbes.\u2019 (Judgment S2) b. FATHER PRESENTS CALVIN\u2019S PERSPECTIVE, HOBBES = \u201cANIMATE\u201d \u14c2\u146e \u14a5\u1422\u1472\u141a\u1424 Hobbes. nik\u00ee-miskaw\u00e2w Hobbes. ni-k\u00ee- m=iskaw -\u00e2 -w Hobbes 1-PREV-find=by.body.TA-DIR-3 Hobbes \u2018I foundAN Hobbes.\u2019 (Judgment S2) One consultant explained the Father\u2019s ability to use \u201cAnimate\u201d marking in this context by saying that \u201cparents can get used to talking about these things like their kids think about them.\u201d Thus, the (98b) example has relativized the animacy of Hobbes to someone who is not present for the speech act (i.e. a third person). In complex narratives told by gifted storytellers in earlier times, we find evidence for relativizing a third person\u2019s perspective. As we have already seen in section 2.3.1 above, the \u201cRolling Head\u201d story told by Louis Moosomin (Bloomfield 1930:\u00a71) makes use of animacy shifting. The table summarizing these shifts is repeated here. Line Event SEVERED HEAD STATUS i Severed head opens eyes INANIMATE ii Severed head speaks MIX \u2026 iv Severed head asks Utensils ANIMATE \u2026 x Boy sees Severed head talking INANIMATE \u2026 xiii Severed head stopped ANIMATE \u2026 xv Only Severed head going INANIMATE \u2026 xvii Severed head passes fire ANIMATE \u2026 xxxii Boy cries that Severed head kill Boy INANIMATE \u2026 xxxiv Severed head sees great serpent ANIMATE Table 2.6. A summary of \u201cinanimate\u201d\/\u201canimate\u201d shifting for the severed head story 92 When we consider the places where animacy shifts in detail, it is clear that the contrast between \u201cAnimate\u201d and \u201cInanimate\u201d is heavily exploited by the speaker to code the different knowledge states of the two main points of view; the narrator\u2019s and the boy fleeing from the severe head. Let us consider these shifts in finer detail. When the head is first severed, it is marked as an \u201cInanimate\u201d referent as its eyes mechanically open; severed heads, after all, do not have the possibility of intentions. (99) i. SEVERED HEAD VIEWED AS INHERENTLY EXTENTIONAL (\u201cINANIMATE\u201d) \u146e\u1455\u1426\u1455\u140d \u1451\u1426\u1473\u1432\u14aa\u1472\u1423 \u1406\u14aa \u1431\u14ef\u14ef\u1420 \u14a5\u1422\u144e\u1481\u14c2\u1422. k\u00eetahtaw\u00ea t\u00f4hk\u00e2p\u00eemakan \u00f4ma pisisik mistikw\u00e2nis. k\u00eatahtaw\u00ea t\u00f4hk\u00e2p\u00ee -makan aw =ima pisisik mi-stikw\u00e2n-is one.time open.eyes-INANACT PRX=IN.SG routine UP-head -DIM \u2018thatIN severed head presently openedIN its eyes.\u2019 Here, the independent-mode verb is marked with the suffix \u2013makan, which is used when something is done in a purely extentional way (see, e.g. all uses in k\u00e2-pimw\u00eaw\u00eahahk 1998), and is frequently used with mechanical apparatus (e.g. a book or tape-recorder; see Wolvengrey 2001:44).31 The nominal is marked with the prefix mi- which is used when a body-part\u2019s owner is not known to the speaker (see M\u00fchlbauer 2004, 2007). It is then framed with the \u201cInanimate\u201d demonstrative \u00f4ma, and the modifier pisisik (not translated by Bloomfield) is crucially inserted between the demonstrative and the nominal. When we consider the patterning of this modifier pisisik in other discourse contexts, it appears that it means something like \u2018routinely\u2019 or \u2018every time.\u2019 (100) THE SEVERED HEAD IS \u201cROUTINE\u201d \u1431\u14ef\u14ef\u1420 \u140a\u1528\u1422 \u14c2\u146e \u1438\u1439\u14a5\u1431\u148b\u14c8\u1423 \u14ab\u14c7; pisisik ayis nik\u00ee-pap\u00e2mipicin\u00e2n m\u00e2na; pisisik ayis ni-k\u00ee- pap\u00e2m=ipici -n\u00e2n m\u00e2na routine for 1-PREV-about=by.pulling.AI-1PL usual \u2018for we always used to move our camp about;\u2019 (Ahenakew 2000:\u00a71.8) Here the \u201cInanimate\u201d properties of the nominal are being presented as they are usually conceived of. So far, the story is going according to routine; the referent is in its proper extentional class. The next moment, something strange happens; the head begins to speak. 31 I cite only text examples of -makan because none of the speakers I have done elicitation with recognize it. 93 (101) SEVERED HEAD IS TRANSITIONAL BETWEEN \u201cANIMATE\u201d AND \u201cINANIMATE\u201d ii. \u1401\u147f \u146e\u1455\u1426\u1455\u140d \u1473 \u1432\u146d\u1422\u1475\u141f \u1406\u14aa \u1406\u1422\u144e\u1481\u1423. \u00eakwa k\u00eetahtaw\u00ea k\u00e2-p\u00eekiskw\u00eat om \u00f4stikw\u00e2n. \u00eakwa k\u00eatahtaw\u00ea k\u00e2-p\u00eekiskw\u00ea-t aw =ima o-stikw\u00e2n then one.time C2-speak -3 PRX=IN.SG 3-head \u2018then presently [suddenly] thatIN head spokeAN.\u2019 iii. \u201c\u140b, \u14c2\u1451\u152e\u1472\u1423, \u1456\u14c2\u140d\u1426\u1473\u1541?\u201d \u201c\u00e2, nit\u00f4y\u00e2kan, t\u00e2niw\u00eahk\u00e2k?\u201d \u00e2 nit-wiy\u00e2kan t\u00e2niw\u00eahk\u00e2k INTERJ 1- dish where.are.they \u201cCome, my dishes, where are they? [the children]\u201d Here, the first clause has been marked with the k\u00e2-form of the changed conjunct (see Wolfart 1973, Cook 2007, etc.). Combining this kind of clause with the temporal adverb k\u00eatahtaw\u00ea yields a meaning of suddenness (Wolfart p.c., Cook, p.c., Cook 2007), as seen in the following example from Alice Ahenakew. (102) K\u00caTAHTAW\u00ca + K\u00c2- = \u201cSUDDENLY\u201d \u1401\u147f \u146b\u1455\u1426\u1455\u140d \u1473 \u1455\u146f\u1438\u1528\u141f \u140a\u140f\u152d\u1420 \u2026 \u00eakwa k\u00eetahtaw\u00ea k\u00e2-takopayit awiyak\u2026 \u00eakwa k\u00eetahtaw\u00ea k\u00e2-tako =payi -t awiyak and one.time C2-arrive=INCH-3 someone \u201csuddenly someone drove up\u2026\u2019 (Ahenakew 2000:\u00a72.1) Thus, the severed head has suddenly begun to speak. This sudden act signals a clash in the knowledge states about this severed head; is it Extentional or Intentional? The verb is marked with \u2013t, which is only used with the unspecified \u201cAnimate\u201d class. This is expected, since the act of speaking is inherently intentional (e.g. Banfield 1982). However the nominal is once again marked with the \u201cInanimate\u201d form of the demonstrative, \u00f4ma and not awa. Thus, the verbal agreement is coding an \u201cAnimate\u201d referent, while the nominal agreement is coding an \u201cInanimate\u201d referent. As Goddard (2004) points out, this is one of the only recorded cases of a mismatch between the nominal and verb agreement. In the context, it makes sense, however, because the referent\u2019s extentional state is now under question: is it \u201cInanimate\u201d or isn\u2019t it? The severed head then questions its utensils, finds where they went, and rolls off in pursuit of them. During this span, there is no overt nominal or demonstrative, but there are other reasons to suppose that the severed head is now grammatically an \u201cAnimate\u201d referent. It is 94 connected to a set of verbs that code \u201cAnimate\u201d actors and introduce strong intentional contexts (iv-vi) (103). (103) iv. SEVERED HEAD VIEWED AS INTENTIONAL (\u201cANIMATE\u201d) \u14f2\u1422\u1481\u1428 \u1472\u1426\u146d\u152d\u1424 \u1472\u1475\u148b\u14a3\u1424 \u1405\u1456\u1438\u148b\u1426\u148b\u1472\u14c7. s\u00f4skw\u00e2c kahkiyaw kakw\u00eacim\u00eaw ot\u00e2pacihcikana. s\u00f4skw\u00e2c kahkiyaw kakw\u00eac=im -\u00ea -w ot-\u00e2pacihcikan-a straight.out all try =by.mouth-DIR-3 3-utensil -XT \u2018Without delay, she askedAN all her utensils.\u2019 Further, the speaking severed head forces the obviation (see Chapter 3) of other referents (103), shown in (104) by the suffix -yi- on the verb kot\u00e2win\u00e2- \u2018sink underground.\u2019 She also becomes the possessor of the nominal n\u00e2p\u00eam- \u2018husband,\u2019 which is only allowed for \u201cAnimate\u201d referents (see \u00a74.3.1). (104) vi. SEVERED HEAD FORCES OBVIATION, BECOMES POSSESSOR \u1434\u144e \u140a\u14ef\u14c2\u152d \u1473 \u1411\u1426\u1455\u14ab\u146f\u141f \u140a\u1422\u146e\u1541 \u1401 \u146e \u146f\u1456\u140f\u14c8\u1528\u141f \u1405\u14c8\u142f\u14aa p\u00f4ti asiniya k\u00e2-w\u00eehtam\u00e2kot ask\u00eehk \u00ea-k\u00ee-kot\u00e2win\u00e2yit on\u00e2p\u00eama p\u00f4ti asiniy-a k\u00e2-w\u00eeht=amaw-iko-t askiy-ihk \u00ea-k\u00ee-kot\u00e2win -\u00e2 -yi-t at.last stone -XT C2-tell=APPLIC -INV-3 earth-LOC C1-PREV-sink-DIR-DS-3 o-n\u00e2p\u00eam -a 3-husband-XT \u2018at last a stoneOBV told herAN that her husbandOBV had sunk themOBV into the earth\u2019 Thus, we can conclude that the severed head has here become fully an \u201cAnimate\u201d referent; it is now conceived of by the Speaker as capable of having intentions (here, expressing a mental state by speaking). It is no surprise, then, when the nominal finally resurfaces in (vii) bearing the \u201cAnimate\u201d demonstrative awa (105). (105) vii. SEVERED HEAD MODIFIED BY \u201cANIMATE\u201d AWA \u2018THIS\u2019 \u1401\u147f \u144c\u143b\u1424 \u140a\u1418 \u1405\u1422\u144e\u1481\u14c2\u1422 \u140a\u1418. \u00eakwa t\u00eapw\u00eaw awa ostikw\u00e2nis awa. \u00eakwa t\u00eapw\u00ea-w aw =a o-stikw\u00e2n-is aw-a then call -3 PRX=AN.SG 3-head -DIM PROX=AN.SG \u2018Then thatAN head began to call.\u2019 Now the severed head\u2019s transformation from an extentional referent into an intentional one has been completed. The narrator, and his audience, have experienced a significant perspectival shift, and the grammar reflects it. 95 While the audience and the narrator have experienced this perspectival shift, a referent in the story is more recalcitrant. The boy pursued by the head is not so ready to acknowledge that the head is still inhabited by his dead mother. The narrator shifts to him as the central figure by employing a significant discourse shift; he is introduced via a left-dislocation (M\u00fchlbauer 2003). (106) ix. NOMINAL N\u00c2P\u00caSIS \u2018BOY\u2019 IN LEFT DISLOCATION \u1401\u147f \u140a\u1418 \u14c8\u142f\u14ef\u1422 \u1403\u1456\u14aa\u1422\u1472\u14a5\u1420 \u1473 \u1403\u14ef \u1455\u1439\u14ef\u141f, \u00eakwa awa n\u00e2p\u00easis it\u00e2maskamik k\u00e2-isi-tapas\u00eet, \u00eakwa aw =a n\u00e2p\u00ea=sis it\u00e2maskamik k\u00e2-isi-tapas\u00ee-t and PRX=AN.SG man=DIM underground c2-RR-flee -3 \u2018And that littleAN boy that was fleeingAN underground,\u2026\u2019 When the head calls to the boy, he turns to look at it. Suddenly, the head is completely \u201cInanimate\u201d once again (107). This is shown by the use of a verb coding \u201cInanimate\u201d objects (-aht-am), an \u201cInanimate\u201d demonstrative \u00f4ma, and a subsequent intransitive verb inflected for an \u201cInanimate\u201d actor using \u2013makan. This last predicate is also coded with \u2013yi\u2013, which codes that the subject is embedded in the perspective of some preceding third person (cf. \u00a74.3.2). (107) x. \u141a\u1426\u152d\u1424 \u1405\u1426\u148b \u141a\u1438\u1426\u1455\u1466 \u1406\u14aa \u1431\u14ef\u14ef\u1420 \u14a5\u1422\u144e\u1481\u1423 \u1401 \u1432\u146d\u1422\u1475\u14aa\u1472\u14c2\u1528\u1420. w\u00e2hyaw ohci w\u00e2pahtam \u00f4ma pisisik mistikw\u00e2n \u00ea-p\u00eekiskw\u00eamakaniyik. w\u00e2hyaw ohci w\u00e2p=aht -am aw =ima pisisik mi-stikw\u00e2n far from see=by.eye.TI-TI PRX=IN.SG routine UP-head \u00ea-p\u00eekiskw\u00ea-makan -yi -k c1-speak -INACT -DS-0 \u2018From afar he [the boy] sawIN thatIN severed speakingIN head\u2019 Here, the location of the event has been shifted; we now see the head wahyaw ohci \u2018from far away.\u2019 The child looks and sees the head, whose speaking is now mechanical and intentionless, marked overtly with the suffix \u2013makan (see above). Lest we not understand that this boy considers the head to be a purely extentional entity, the narrator introduces a direct quote from the boy to this effect (108). (108) xii. BOY CLAIMS HEAD IS \u201cINANIMATE\u201d \u201c\u2026\u14c7\u14aa \u146e\u147f\u1429 \u146d\u147f\u140f\u1528\u14c7\u1424. \u1405\u1422\u144e\u1481\u1423 \u1431\u146f \u1473 \u1432\u146d\u1422\u1475\u14aa\u1472\u1541 \u2026\u201d \u201c\u2026 nama k\u00eakway kik\u00e2wiyinaw. ostikw\u00e2n piko k\u00e2-p\u00eekiskw\u00eamakahk\u2026\u201d nama k\u00eekway ki-k\u00e2wiy-inaw o-stikw\u00e2n piko k\u00e2-p\u00eekiskw\u00ea-makan -k NEG thing 2- mother-21Pl 3-head only C2-speak- INACT -0 \u201c\u2026our motherAN is not there. It is only a talkingIN head\u2026\u201d 96 Here, the boy overtly denies that the head is inhabited with the mind of his mother. The head is modified with piko, which means something like \u2018just\u2019 or \u2018only\u2019 when postposed to a noun, and the verb of speaking is once again marked with \u2013makan. The narrator could not be any more explicit about what the boy believes. Thus, the inherently extentional, \u201cInanimate\u201d form of \u2013stikw\u00e2n is now been relativized to the boy\u2019s perspective. Throughout the rest of the passage, this contrast between the boy\u2019s perspective and the narrator\u2019s is repeatedly brought to light. Thus, the story presents two opposing perspectives on the extentionality of the head. The narrator and his audience have inside information on the head that indicates it is capable of intentions; they saw it question the utensils about its children, and heard the rock confirm what her husband had done with them (i-vi). The two boys, by contrast, were not there for those events, and are furthermore naturally resistant to identify the monstrous severed head that pursues them as their mother\u2019s spirit calling them. Thus, the Narrator and his audience are able to treat the severed head as an intentional referent, and when their knowledge is referenced, the narrator uses numerous grammatical devices to convey the intentionality of the head. All of these devices depend on the \u201cAnimate\u201d status of the head: it is marked \u201cAnimate,\u201d it speaks with animate-marked verbs, and it obviates other referents. By contrast, when the boy\u2019s perspective is referenced, the narrator uses every grammatical device available to code the inherently extentional, \u201cInanimate\u201d status of the head: it is marked \u201cInanimate,\u201d all verbs are marked with \u201cInanimate\u201d morphology, and overt quantifiers are used. Thus, the grammar of Plains Cree has here been exploited to maximal rhetorical effect by a gifted speaker, who is playing on the difference between what his audience expects a severed head to be capable of and what it here becomes. Modeling this perspectival shift to third persons is straightforward. In all cases, it simply involves an embedding of the third person\u2019s perspective inside of the Speaker\u2019s. For example, when the boy sees the severed head chasing him, the perspective on the severed head is the boy\u2019s (109). This is modeled by embedding the extentional predicate (EXT(y)) inside of the boy\u2019s perspective. (109) x. \u141a\u1426\u152d\u1424 \u1405\u1426\u148b \u141a\u1438\u1426\u1455\u1466 \u1406\u14aa \u1431\u14ef\u14ef\u1420 \u14a5\u1422\u144e\u1481\u1423 \u1401 \u1432\u146d\u1422\u1475\u14aa\u1472\u14c2\u1528\u1420. w\u00e2hyaw ohci w\u00e2pahtam \u00f4ma pisisik mistikw\u00e2n \u00ea-p\u00eekiskw\u00eamakaniyik. w\u00e2hyaw ohci w\u00e2p=aht -am aw =ima pisisik mi-stikw\u00e2n far from see=by.eye.TI-TI PRX=IN.SG routine UP-head \u00ea-p\u00eekiskw\u00ea-makan -yi -k c1-speak -INANACT-DS-0 \u2018From afar he [the boy] sawIN thatIN severed speakingIN head\u2019 97 Speaker x y \uf0df HEAD IS NOT EXTENTIONAL TO SPEAKER boy(x) head(y) x \uf0df HEAD IS EXTENTIONAL TO BOY y EXT(y) head(y) see(x,y) speak(y) 2.6. Conclusion Context informs content (Bateson 1972, Fillmore 1975), and content and form are inseparably linked (McCawley 1988). The two animacy classes of Plains Cree show a convergence of form, content, and context. In terms of form, the two animacy classes are used to code referential contrasts across multiple positions in the clause, including V, v, NumP, and DemP. In terms of content, the \u201cInanimate\u201d codes extentionality while the \u201cAnimate\u201d is unspecified. Form and content are manipulated by speakers to present the perspectives of different individuals. Considering forms allows us to control content and context, while context provides cues to content, and content provides cues to context. 98 Chapter 3 The Form, Content, and Context of Obviation 3.1. Proposal: Obviation as contextual extentionality In this chapter, I consider the form, content, and context of a morphological subclass of \u201cAnimate\u201d referents in Plains Cree, traditionally termed \u201cObviative.\u201d1 A minimal pair is given in (1); the form in (1a) has only \u201cAnimate\u201d coding on the verb (\u2013t), while the form in (1b) has a suffix \u2013a attached to the nominal, and an additional suffix \u2013yi\u2013 attached to the \u201cAnimate\u201d coded verb (w\u00e2piskisi-). (1) a. \u201cANIMATE\u201d \u141a\u1433\u14f1 \u1401 \u141a\u1431\u1422\u146d\u14ef\u141f w\u00e2pos \u00ea-w\u00e2piskisit w\u00e2posw \u00ea- w\u00e2piski=si-t rabbit C1-white =AI-3 \u2018A\/the rabbitAN is white.\u2019 (Presented S2) b. \u201cOBVIATIVE\u201d \u141a\u1433\u1501 \u1401 \u141a\u1431\u1422\u146d\u14ef\u1528\u141f \u2026w\u00e2poswa \u00ea-w\u00e2piskisiyit w\u00e2posw-a \u00ea- w\u00e2piski=si -yi -t rabbit -XT C1-white =AI -DS-3 \u2018\u2026 (as) a\/the rabbitOBV was white.\u2019 (Presented S2) After reviewing previous work on obviation (\u00a73.2), I then argue that obviation is a construct (\u00a73.3); Plains Cree constructs obviation by combining forms that code \u201cAnimate\u201d reference with forms that code dependency. \u201cObviative,\u201d then, does not have any dedicated forms in the syntax of Plains Cree. A summary of the forms associated with obviation and the present analysis of them is given in Table 3.1. 1 While my definition of this category is new, I have here adopted the traditional term for it. 99 FORM CONTENT EXAMPLE N-a V-a EXTP x is Extentional n\u00e2p\u00eawa n\u00e2p\u00eaw-a man -XT \u2018manOBV\/IN\u2019 DEM=ihi DEMP x is Extentional \u00f4hi aw =ihi PRX=XT \u2018This oneOBV\/IN\u2019 N-yi- V-yi- IP xSUBJ \u2260 y nikamoyiwa nikamo-yi-w-a sing -DS-3-XT \u2018s\/heOBV sings\u2019 N-im V-im nP VP xOBJ \u2260 y niw\u00e2pamim\u00e2wa ni-w\u00e2p=am -im -\u00e2 -w-a 1- see =by.eye.TA-DSJ-DIR-3-XT \u2018I see him\/herOBV\u2019 V-\u00ea- VP xOBJ is Extentional w\u00e2pam\u00eaw w\u00e2p=am -\u00ea -w see =by.eye.TA-DIR-3 \u2018s\/hePROX sees him\/herOBV\u2019 Table 3.1. No dedicated obviation forms I then argue that the referent associated with an \u201cObviative\u201d construct is extentional within a specified context (\u00a73.4). I define a contextually extentional referent as one that cannot be paired with a perspective in a given context (signified by \u2018C\u2019 in the formalism). (2) EXT(x,C) \u2194 \u2200\u03c8\u2200y(R(y,\u03c8,C) \uf0e0 x \u2260 y) x is extentional at context C if and only if for all perspectives \u03c8 and all individuals y, if there is a relation R of y with \u03c8 at context C, then x is not y. Applying this to the \u201cObviative\u201d form in Plains Cree, I claim that a referent classified as \u201cObviative\u201d will have the content of (4). (3) \u201cOBVIATIVE\u201d CONSTRUCT = [\u03bbx \u00b7 EXT(x,C)] x is extentional at context C if and only if for all perspectives \u03c8 and all individuals y, if there is a relation R of y with \u03c8 at context C, then x is not y. This specified \u201cObviative\u201d referent can be made to contrast with the unspecified \u201cAnimate\u201d.\u201d In this contextual contrast, the unspecified \u201cAnimate\u201d referent takes on the property of contextual intentionality (sometimes called \u201cProximate\u201d). 100 (4) \u201cANIMATE\u201d ISKW\u00caW \u2018WOMAN\u2019 BECOMES \u201cPROXIMATE\u201d IN CONTRAST TO \u201cOBVIATIVE\u201d \u00ea-w\u00e2pam\u00e2t iskw\u00eaw n\u00e2p\u00eawa \u00ea-w\u00e2p=am -\u00e2 -t iskw\u00eaw n\u00e2p\u00eaw-a c1-see=by.eye.TA-DIR-3 woman man -XT \u2018The womanPROX saw the manOBV.\u2019 (Presented S2) = \u201cProximate\u201d Woman has perspective on event. \u2260 \u201cObviative\u201d Man has perspective on event. This gives us a typology of referential contrasts that depend on two variables: (i) extentionality, and (ii) inherent vs. contextual properties. (5) a. \u201c\u201cInanimate\u201d : [EXT] b. \u201cAnimate\u201d : [\u00d8] c. \u201cObviative\u201d : [EXT,C] d. \u201cProximate\u201d : [-EXT, C] I then argue that the \u201cObviative\u201d form-content pair also bears contextual information (\u00a73.5): a referent\u2019s contextual extentionality is evaluated within the perspective of the Speaker. Building on the model constructed for animacy in Chapter 2, I analyze the obviative as placing restriction on perspectival embeddings. Within the context of the Speaker\u2019s perspective, \u201cObviative\u201d referents cannot possess a perspective; they cannot both be extentional and perspective- possessing. (6) ILL-FORMED DRS: \u201cOBVIATIVES\u201d CANNOT POSSESS PERSPECTIVES EMBEDDED IN SPEAKER\u2019S Speaker x C EXT(x,C) \uf0df x DOES NOT HAVE A PERSPECTIVE x \uf0df x HAS A PERSPECTIVE 101 In the context of this extentional form, the unmarked \u201cAnimate\u201d form gains a perspective within the Speaker\u2019s perspective, and can thus serve as a source of information to the Speaker about the proposition. (7) Speaker x y C EXT(y,C) \uf0df y DOES NOT HAVE A PERSPECTIVE x \uf0df x HAS A PERSPECTIVE y y \uf0df y HAS A PERSPECTIVE From this organization of perspectives, it follows that something that is true in the \u201cObviative\u2019s\u201d perspective is not directly true in the Speaker\u2019s; there is an intermediate perspective that always must be considered. This opacity of the \u201cObviative\u2019s\u201d perspective gives rise to an accessibility condition (cf. Kratzer 1977, 1991), that defines the relation between perspectives (i.e. truth- evaluation domains), as given in (8). (8) ACCESSIBILITY CONDITION: A perspective \u03c82 is accessible to another perspective \u03c81 iff \u03c82 is embedded inside \u03c81 and there are no intervening embeddings between \u03c82 and \u03c81. Obviation, then, can be used to create chains of perspectives, each of which is accessible only to the perspective immediately outside of it. 102 (9) Speaker Proximat Accessible Obviative Accessible This gives rise to the evidential effects of obviation; the \u201cProximate\u2019s\u201d perspective is a domain that is accessible to the Speaker (i.e. open to direct evidentiality), while the \u201cObviative\u2019s\u201d perspective is inaccessible to the Speaker (i.e. open only to indirect evidentiality). 3.2. Previous accounts Obviation is one of the most-studied properties of Algonquian languages, with descriptions going back to the 17th Century (Eliot 1666). The claims made in these different analyses are divergent, making it both important and difficult to situate the current analysis at the outset of the discussion. While the current proposal is built from previous work on this topic, it departs from previous accounts in five crucial ways: (i) The treatment of \u201cObviative\u201d as a property of both discourse and argument structure, rather than primarily one or the other (\u00a73.2.1) (ii) The treatment of \u201cObviative\u201d as a construct, rather than a primitive (\u00a73.2.2) (iii) The treatment of \u201cObviative\u201d as specified and \u201cProximate\u201d as contextually-conditioned, rather than \u201cProximate\u201d as specified (\u00a73.2.3) (iv) The analysis of the \u201cObviative\u201d referential class as denoting contextual extentionality rather than empathy, sentience, topic, or focus (\u00a73.2.4) (v) The range of the data set that the analysis accounts for (\u00a73.2.5) 3.2.1. Obviation as discourse, not argument structure There are two positions that have been taken on the relevant linguistic domain for obviation in Algonquian: (i) Obviation is argument structure, with discourse effects being derived. 103 (ii) Obviation is discourse structure, with argument structure effects being derived. These approaches differ over the direction of investigation; approach (i) begins with local relations and intends to derive discourse effects, while approach (ii) begins with discourse effects and attempts to derive local relations. The position that obviation is argument structure, which I term \u201cGrammatical Obviation,\u201d was suggested by the earliest grammarians. For example, Eliot (1666) devotes a great deal of space to verbal paradigms, which has the result of centering the grammatical discussion on argument structure. In linguistic work on obviation in the 20th century, this focus on argument structure has developed in two different directions. One approach, favored by Relational Grammarians (Perlmutter 1983), analyzes obviation as a by-product of the interaction between verbal argument structure and ranking effects. The most extensive implementation of this approach is Rhodes (1976) for Ojibwa, but similar approaches have been implemented in Optimality-theoretic syntax (Aissen 1997). The other argument structure approach has been to analyze obviation as derived from general principles of syntactic structure. In approaches of this kind, obviation can be characterized in terms of binding relations, either simply via c-command or some version of the conditions of Binding Theory (Chomsky 1982). In these analyses, the disjunction properties of obviation are given primary importance (see Chapter 4). Implementations in this approach include Grafstein (1984), Bruening (2001), D\u00e9chaine and Wiltschko (2002), Branigan and MacKenzie (2004), Piriyawiboon (2007), and Lochbiler (2007), among many others. In all of these approaches, the focus is on relations within the clause \u2013 there is a division between the grammatical component of obviation and its discourse component. Transitive verbs, possession, and dependent clauses are the primary locus of \u201cgrammatical\u201d obviation, and \u201cdiscourse\u201d obviation is either derived (e.g. Aissen 1997, Piriyawiboon 2007) or set aside for future research (e.g. Rhodes 1976). A consideration of data sets from Plains Cree shows that obviation cannot be a by- product of verbal argument structure, because obviation holds in many domains where there is no verbal coding to support it, including nominal predication (10a) and AIT (ANIMATE INTRANSITIVE-TRANSITIVE) forms (10b) (see D\u00e9chaine 1997, M\u00fchlbauer 2002, etc.). 104 (10) a. NOMINAL PREDICATION \u140a\u141a\u14ef\u14f4 \u1406\u1426\u1403 aw\u00e2sisa \u00f4hi aw\u00e2sis-a aw =ihi child -XT PRX=XT \u2018These are the childrenOBV.\u2019 (Presented S2) b. OBVIATIVE ARGUMENTS CAN OCCUR WITH VERBS THAT LACK TRANSITIVITY CODING \u1401 \u141a\u1438\u1426\u148b\u146b\u141f \u140a\u14c2\u1426\u1403 \u141a\u1434\u1501 \u00ea-w\u00e2pahcik\u00eat anihi w\u00e2p\u00f4swa \u00ea-w\u00e2p=aht =ik\u00ea-t an =ihi w\u00e2p\u00f4sw-a c1-see=by.eye.TI=GEN-3 DST=XT rabbit-XT \u2018HePROX watches those rabbitsOBV.\u2019 (Presented S2) Without a strict correlation between obviation and the verb system, analyses that treat obviation as a kind of argument-structure relation are not tenable. In the model proposed here, however, this kind of data is expected; \u201cObviative\u201d is a referential category rather than a grammatical one, meaning that it is only opportunistically coded in the morphosyntax of Plains Cree. An alternative view, which I term \u201cDiscourse Obviation,\u201d has been suggested by Bloomfield (1962) and subsequent work (e.g. Wolfart 1973, Goddard 1991, Russell 1996). Describing Menominee (Central Algonquian, Wisconsin), Bloomfield considered the referential properties of obviation to be primary, though they fell largely outside of the scope of his linguistic investigation. He described obviation as coding distinguishing between a referent that is the topic of discourse, and those that are not within a given context (Bloomfield 1962:38). Goddard (1991) expands on this discussion to consider the ways that obviation can be manipulated to code changes in discourse structure, and Russell (1996) considers obviation phenomena in terms of point of view. In these accounts, then, the primary focus is on the discoursal properties of obviation, rather than its local argument structure properties. The current analysis treats the \u201cObviative\u201d as a referential class, and obviation as a referential distinction. In terms of obviation\u2019s syntactic properties, I argue that they opportunistically employ underspecified forms to build \u201cObviative\u201d reference. Obviation, then, is a property of discourse, but this discoursal property is constructed through the manipulation of dependency-building syntax (cf. Chapter 4). This combines the observations of both the \u201cGrammatical\u201d obviation and the \u201cDiscourse\u201d obviation approaches. 105 3.2.2. Obviation as a construct, not a primitive In many analyses, obviation is taken to be located in particular places in the Plains Cree grammatical system. For example, Bruening (2001:121) posits \u201cObviative\u201d to be a syntactic feature [OBV] in Passamaquoddy (Eastern Algonquian, Maine), which is spelled out by the \u201cObviative\u201d suffix \u2013ol, following claims made for Potawatomi by Halle and Marantz (1993). He then posits a phrase \u2018HP\u2019 (\u201cHead Phrase\u201d) that forces these \u201cObviative\u201d features to move there to be checked, which is intended to model direct-inverse verbal behaviour. This means that \u201cObviative\u201d must be a primitive feature of the grammar. Obviation is subsumed within the characterization of local relations between the verb and its arguments. Other analyses of obviation have followed a similar logic. Aissen (1997) implements an Optimality-Theoretic syntax treats \u201cObviative\u201d as a primitive, as do many other analysts (Bliss 2005, Branigan & MacKenzie 2004, Grafstein 1984, Piriyawiboon 2007, among others). However, there are several significant problems with analyses of this kind, which warrant a new direction of research. Obviation has no dedicated forms in the syntax. None of the forms that are typically treated as coding obviation are in fact dedicated \u201cObviative\u201d morphemes. The absence of dedicated forms challenges accounts that treat obviation as a syntactic primitive. Instead, an adequate account will have to model a system that uses underspecification and restrictions to construct \u201cObviative\u201d contexts. This underspecification clarifies the case for the formal identity between \u201cInanimate\u201d and \u201cObviative\u201d forms (see Hockett 1966, Wolfart 1973), which I pursue in section 3.4. To my knowledge, the only formal account, besides the current analysis, that has attempted to move in that direction is the one proposed by Piriyawiboon (2007), which uses a minimalist framework (Chomsky 1995) and a person-hierarchy analysis to attempt to derive the parallelism. 3.2.3. \u201cObviative\u201d is specified, \u201cproximate\u201d is contextually-determined Analyses of obviation phenomena pick one or the other member of the \u201cProximate\/Obviative\u201d contrast to focus on, with the choice being a function of the domain of inquiry. If the analyst is modelling the syntax of obviation, the focus will be on the \u201cObviative\u201d class (Grafstein 1984, Bruening 2001, D\u00e9chaine & Wiltschko 2002), since this class is coded with numerous pieces of morphology, as discussed in section 3.4. Sometimes, these accounts 106 refer to \u201cProximate\u201d agreement (e.g. D\u00e9chaine & Wiltschko 2002:433), but this approach is untenable for two resons: (i) There is no morphological distinction between a \u201cProximate\u201d nominal form and the normal \u201cAnimate\u201d form (11). (11) a. \u201cANIMATE\u201d AGREEMENT: \u2013ak, \u2013iki \u14a5\u152a\u14ef\u1418\u1420 \u1406\u146d \u14c8\u142f\u1418\u1420. miyosiwak \u00f4ki n\u00e2p\u00eawak. miyw=si-w-ak aw =iki n\u00e2p\u00eaw-ak good =AI-3-PL PRX=AN.PL man -PL \u2018These menAN are good.\u2019 (Presented S2) b. \u201cPROXIMATE\u201d AGREEMENT: \u2013ak, \u2013iki \u141a\u1438\u14a3\u1418\u1420 \u1406\u146d \u14c8\u142f\u1418\u1420 . w\u00e2pam\u00eawak \u00f4ki n\u00e2p\u00eawak aw\u00e2sisa. w\u00e2p=am -\u00ea -w aw =iki n\u00e2p\u00eaw-ak aw\u00e2sis-a see =by.eye.TA-DIR-3PRX=AN.PL man -PL child -XT \u2018These menPROX see the childOBV.\u2019 (Presented S2) (ii) There is no distinction between \u201cProximate\u201d agreement and \u201cObviative\u201d agreement on verbs (12); the \u201cObviative\u201d has all the agreement of the \u201cProximate\u201d and then more (cf. Wolfart 1973, 1978). (12) a. \u201cPROXIMATE\u201d VERBAL AGREEMENT: -\u00ea-w \u141a\u1438\u14a3\u1424. w\u00e2pam\u00eaw. w\u00e2p=am -\u00ea -w see =by.eye.TA-DIR-3 \u2018S\/hePROX sees him\/herOBV.\u2019 (Presented S2) b. \u201cOBVIATIVE\u201d VERBAL AGREEMENT: -\u00ea-yi-w-a \u141a\u1438\u14a3\u1528\u1418. w\u00e2pam\u00eayiwa. w\u00e2p=am -\u00ea -yi -w-a see =by.eye.TA-DIR-DS-3-XT \u2018S\/heOBV sees him\/herOBV.\u2019 (Wolfart & Carroll 1973) Lacking any morphosyntactic correlate, the category \u201cProximate\u201d must necessarily be derived from context, while the \u201cObviative,\u201d which has morphosyntactic correlates, can be constructed in the syntax. 107 In analyses of the meaning of obviation, the focus of investigation is the \u201cProximate\u201d referent. Accounts consider the properties of this referential class; it is the \u201cTopic\u201d of the discourse (Bloomfield 1962), the one \u201cin focus\u201d (Wolfart 1973), or the \u201chero\u201d of the narrative (Goddard 1991). Its \u201cpoint of view\u201d properties are considered (Russell 1996), and accounts look at its \u201cempathetic\u201d properties (Oshima 2007). Positioning this \u201cProximate\u201d class within a discourse, analysts ask how long a \u201cProximate\u201d can hold its status before being replaced with a different referent (i.e. \u201cProximate spans\u201d). In discourse studies (e.g. Hasler 2002), the \u201cProximate\u201d referents are counted and their contexts of use analyzed. Because the current analysis seeks to account for both the syntactic and semantic properties of obviation phenomena, it must combine the requirements of the syntactic analysis with the observations of the semantic work. By analyzing \u201cObviative\u201d as a construct and \u201cProximate\u201d as contextually-conditioned, the morphosyntactic generalizations (i.e. that \u201cObviative\u201d is coded with additional morphology, and the \u201cProximate\u201d is indistinguishable from \u201cAnimate\u201d) are accounted for. By bringing these morphosyntactic generalizations to bear on the semantic problem, the current account departs from previous semantic work and treats the \u201cObviative\u201d as the specified class; the \u201cObviative\u201d is specified for contextual extentionality, while the \u201cProximate\u201d is only a function of contrast with the \u201cObviative.\u201d This analysis, then, constitutes a different direction for work on obviation. 3.2.4. The meaning of obviation A detailed treatment of the meaning of obviation has not been developed outside of the current thesis. Analyses of obviation have either identified its discourse properties in general terms (e.g. \u201cTopic\u201d or \u201cProminent\u201d or \u201cFocus\u201d as in Bloomfield 1962, Wolfart 1973, Russell 1996), employed functionalist person hierarchies (Frantz 1976, Giv\u00f3n 1979, Aissen 1997, Ritter & Rosen 2005, Oshima 2007), or have not addressed the meaning of obviation at all (e.g. Bruening 2001). Wolfart (1973:14) suggests a future line of research by noting that the formal parallels between \u201cInanimate\u201d and \u201cObviative\u201d referential classes likely point to a common semantic property. By analyzing the \u201cObviative\u201d class as denoting a contextually-extentional referent (a contextual form of the inherent extentionality of \u201cInanimate\u201d), this analysis develops the suggestion of Wolfart (1973), and offers the first detailed treatment of the semantics of obviation. 108 3.2.5. The data set Syntactic accounts of obviation focus on verbal argument structure (e.g. Lochbihler 2007) and, in some cases, clause-level properties (e.g. Bruening 2001, Branigan & MacKenzie 2007). This yields a data set restricted to pairs of single sentences that are not controlled for context or, in some cases, single words that are not complete utterances (e.g. Lochbihler 2007). In functionalist accounts, the data is usually restricted to labels given to forms by previous linguists; a categorical decision by an earlier linguist will be taken as a kind of primary data, and then the different possible semantic models suggested by this label are considered (e.g. Ritter and Rosen 2005, Oshima 2007). Goddard (1984), who views obviation as discourse-based, provides a wide range of data, introducing concepts like \u201cProximate and Obviative spans\u201d to describe the large-scale, discourse manipulations of obviation that he demonstrates. However, the data sets are limited to specific kinds of formal narratives (e.g. at\u00e2y\u00f4hkana \u2018sacred stories\u2019), so that the effects of different kinds of speech situations are not controlled for. In comparison to previous work, then, the most significant contribution of the present analysis is actually the data set it is derived from: (i) The effects of physical absence on the \u201cObviative\u201d status of a referent (\u00a73.4.21) (ii) The interpretation of \u201cObviative\u201d referents as unaware of the event (\u00a73.4.22) (iii) The interpretations of \u201cProximate\u201d referents as speakers, even in contexts where they do not speak (\u00a73.4.3) (iv) The correlation between \u201cObviative\u201d and indirect evidentials (\u00a73.5.4). Regardless of the success or failure of the model constructed here, the enrichment of the data set for obviation provides future linguistic work with crucial generalizations. 3.3. The form of obviation: No dedicated \u201cobviative\u201d marking The \u201cObviative\u201d construction is formed by combining any of a set of morphemes with a verbal structure that codes an \u201cAnimate\u201d referent. For example, the form in (13a) shows n\u00e2p\u00eaw \u2018man\u2019 as the argument of a verb bearing \u201cAnimate\u201d marking (-si-t), while the form in (13b) shows the same noun affixed with \u2013a and connected to a verb bearing both \u201cAnimate\u201d marking (-si-t) and an additional suffix \u2013yi\u2013. 109 (13) a. \u201cANIMATE\u201d \u1401 \u14a5\u152a\u14ef\u141f \u140a\u14c2\u140a \u14c8\u142f\u1424 \u00ea-miyosit ana n\u00e2p\u00eaw \u00ea-miyw=si -t an =a n\u00e2p\u00eaw c1-good=AI-3 DST=AN.SG man \u2018\u2026(as) thisAN man is goodAN.\u2019 (Presented S2) b. \u201cOBVIATIVE\u201d \u1401 \u14a5\u152a\u14ef\u1528\u141f \u140a\u14c2\u1426\u1403 \u14c8\u142f\u1418 \u00ea-miyosiyit anihi n\u00e2p\u00eawa \u00ea-miyw=si -yi -t an =ihi n\u00e2p\u00eaw-a c1-good=AI-DS-3 DST=XT man -XT \u2018\u2026(as) thisOBV manOBV is goodAN.\u2019 \u2018\u2026(as) thoseOBV menOBV are goodAN.\u2019 (Presented S2) A diverse set of forms can be shown to occur with the \u201cObviative\u201d class. This includes the three affixes seen above (\u2013yi\u2013, \u2013ihi, and \u2013a), as well as the suffix \u2013im\u2013 in (14a) and the theme sign \u2013\u00ea\u2013 in (14b). (14) a. SUFFIX \u2013im\u2013 OCCURS WITH \u201cOBVIATIVE\u201d \u14c2\u141a\u1438\u14a5\u14ab\u1418 \u141a\u1438\u1422\u144e\u1466 \u1405\u14a5\u14c5\u14ef\u14aa niw\u00e2pamim\u00e2wa W\u00e2pastim omin\u00f4sima ni-w\u00e2p=am -im -\u00e2 -w-a w\u00e2pastim o-min\u00f4s-im -a 1-see=see.by.eye.TA-DSJ-DIR-3-XT W\u00e2pastimw 3-cat -DSJ-XT \u2018I saw W\u00e2pastim\u2019sPROX catOBV \u2026\u2019 (Volunteered S1) b. THEME SIGN \u2013\u00ea\u2013 OCCURS WITH \u201cOBVIATIVE\u201d \u141a\u1438\u14a3\u1424 w\u00e2pam\u00eaw w\u00e2p=am -\u00ea -w see =by.eye.TA-DIR-3 \u2018s\/heprox sees him\/herobv\u2019 (Presented S2) There are two ways to think about the form of obviation phenomena in Plain Cree: (i) HYPOTHESIS 1: \u201cObviative\u201d is a primitive feature, like \u201cAnimate\u201d and \u201cInanimate.\u201d (ii) HYPOTHESIS 2: \u201cObviative\u201d is a construct. Depending on the choice made in this issue, a number of different expectations about the organization of the system arise. In the subsequent sections, I argue that treating \u201cObviative\u201d as a construct is a better account. 110 A grammatical \u201cprimitive\u201d is an atomic building block of the grammar; it is used to construct, but is not itself constructed (cf. Link 1984 for atomicity). If we consider \u201cObviative\u201d to be a primitive (Aissen 1999, Bruening 2001, Ritter & Rosen 2005, Bliss 2005), we expect that there should be a dedicated form or a set of dedicated forms coding \u201cObviative.\u201d If we find such forms, we can then ask the same questions about its exponence in the grammar that we did with animacy: Is the coding localized to a particular syntactic position or is it the case that it has no particular locus? This would lead us to diagnostics and discussion similar to the issues considered for the syntax of animacy in Chapter 2. A grammatical \u201cconstruct\u201d is a non-atomic, concatenation of primitives or other constructs (cf. Link 1984 for atomicity). If we consider the \u201cObviative\u201d to be a construct, we expect that there will not be any form, or set of forms, that are dedicated to coding it. We can ask how the grammar constructs obviation: what kinds of forms are recruited for the construction of obviation, and in what places in the grammar? If the Plains Cree shows dedicated \u201cObviative\u201d forms, obviation can be thought of as an primitive of Plains Cree, akin to animacy. If, however, the system shows no forms dedicated to obviation, but instead shows obviation to be built additively off of more basic forms, then obviation can be thought of as a construct of Plains Cree. This can be tested by considering all the forms in Plains Cree that potentially code obviation, and asking in every case if it codes obviation exclusively. Let us now consider each of these forms in detail. 3.3.1. The nominal suffix \u2013a The suffix \u2013a occurs on nominals when they refer to an \u201cObviative\u201d referent. (15) \u201cOBVIATIVE\u201d NOMINAL AFFIXED WITH \u2013a \u1401 \u14a5\u152a\u14ef\u1528\u141f \u140a\u14c2\u1426\u1403 \u14c8\u142f\u1418 \u00ea-miyosiyit anihi n\u00e2p\u00eawa \u00ea-miyw=si -yi -t an =ihi n\u00e2p\u00eaw-a c1-good=AI-DS-3 DST=XT man-XT \u2018\u2026(as) thisOBV manOBV is goodAN.\u2019 \u2018\u2026(as) thoseOBV menOBV are goodAN.\u2019 (Presented S2) The suffix \u2013a also occurs on Independent Order verbs when an \u201cObviative\u201d interacts with a local person on a transitive verb (16a), or when an \u201cObviative\u201d referent is the subject of an intransitive (16b) (Wolfart 1973, etc.). 111 (16) a. \u201cOBVIATIVE\u201d OBJECT OF VERB + 1ST\/2ND PERSON SUBJECT = -a \u14c2\u141a\u1438\u14a5\u14ab\u1418 \u141a\u1438\u1422\u144e\u1466 \u1405\u14a5\u14c5\u14ef\u14aa niw\u00e2pamim\u00e2wa W\u00e2pastim omin\u00f4sima ni-w\u00e2p=am -im -\u00e2 -w-a W\u00e2pastimw o-min\u00f4s-im -a 1-see =by.eye.TA-DISJ-DIR-3-XT W\u00e2pastim 3-cat -DISJ-XT \u2018I see W\u00e2pastim\u2019sPROX catOBV.\u2019 (Translation S1) b. \u201cOBVIATIVE\u201d SUBJECT OF INTRANSITIVE VERB = -a Clare \u1405\u14ab\u14ab\u1418 \u1472 \u14c2\u1472\u14a7\u1528\u1418 Clare om\u00e2m\u00e2wa ka-nikamoyiwa Clare o-m\u00e2m\u00e2 -a ka- nikamo-yi -w-a Clare 3-mother-XT FUT-sing -DS-3-XT \u2018Clare\u2019sPROX motherOBV will sing.\u2019 (Presented S2) This form, then, is systematically associated with \u201cObviative\u201d nominals, as well as some \u201cObviative\u201d arguments of verbs. This makes it a likely candidate for dedicated \u201cObviative\u201d marking. 3.3.11. The traditional view: \u2013a marks \u201cobviative\u201d Analyses of the suffix \u2013a in Plains Cree and related forms in other Algonquian languages treat it as dedicated to coding \u201cObviative.\u201d Bloomfield (1933), Dahlstrom (1991), and Goddard (1991) have all treated this form as dedicated to coding the \u201cObviative\u201d class of referents. Following this approach, many presentations of Plains Cree data adopt an \u201cObviative\u201d gloss for this suffix (e.g. M\u00fchlbauer 2007, D\u00e9chaine & Wiltschko 2002). Wolfart (1973) adopts the same usage, but notes reservations, based on the suffix\u2019s formal parallels to the \u201cInanimate\u201d suffix \u2013a. If these analyses are correct, this suffix \u2013a is dedicated to coding \u201cObviative,\u201d and thus constitutes evidence for \u201cObviation\u201d as a primitive of the grammar. 3.3.12. The current analysis: \u2013a marks extentional referents There are reasons to believe that \u2013a is not dedicated to coding \u201cObviative\u201d in Plains Cree. There are contexts where \u2013a occurs, but does not code \u201cObviation.\u201d However, in all these contexts, \u2013a codes that the referent is extentional \u2013 either \u201cInanimate\u201d or \u201cObviative.\u201d A suffix that is formally identical to the \u201cObviative\u201d suffix \u2013a in (17a) occurs with nominals that refer to non-singular \u201cInanimate\u201d referents (17b). 112 (17) a. \u201cOBVIATIVE\u201d NOMINAL WITH \u2013a \u1401 \u14a5\u152a\u14ef\u1528\u141f \u140a\u14c2\u1426\u1403 \u14c8\u142f\u1418 \u00ea-miyosiyit anihi n\u00e2p\u00eawa \u00ea-miyw=si -yi -t an =ihi n\u00e2p\u00eaw-a C1-good=AI-DS-3 DST=XT man -XT \u2018\u2026(as) thisOBV manOBV is goodAN.\u2019 \u2018\u2026(as) thoseOBV menOBV are goodAN.\u2019 (Presented S2) b. \u201cINANIMATE\u201d NOMINAL WITH \u2013a \u1401 \u14a5\u1429\u141a\u14ef\u146d \u140a\u14c2\u1426\u1403 \u14aa\u1422\u146d\u14ef\u14c7 \u00ea-miyw\u00e2siki anihi maskisina \u00ea- miyw=\u00e2si-k -i an =ihi maskisin-a C1-good=II -0-PL DST=XT shoe -XT \u2018\u2026(as) thoseIN shoes are niceIN.\u2019 (Presented S2) This putative \u201cInanimate\u201d suffix also occurs on intransitive Independent Order verbs, as the example in (18) shows. (18) \u201cINANIMATE\u201d SUBJECT OF INTRANSITIVE VERB WITH \u2013a \u14a5\u1429\u141a\u14ef\u14cc \u140a\u14c2\u1426\u1403 \u14aa\u1422\u146d\u14ef\u14c7 miyw\u00e2sinwa anihi maskisina miyw=\u00e2si-n-w-a an =ihi maskisin-a good=II -0-3-XT DST=XT shoe -XT \u2018ThoseIN shoes are niceIN.\u2019 (Presented S2) Both the phonological shape of the suffix (\/\u04d9\/) and its positioning are identical in all these cases. This total formal identity means that it is impossible to tell whether a nominal affixed with \u2013a denotes an \u201cInanimate\u201d or \u201cObviative\u201d referent, unless there is supporting context that codes the animacy of the referent (cf. Wolfart 1973:14,23). This apparent homophony is not specific to Plains Cree. Across Algonquian, the \u201cInanimate\u201d \u2013a and \u201cObviative\u201d \u2013a suffixes have shared an identical shape for a very long time, and across many languages. Both the \u201cInanimate\u201d suffix \u2013a and the \u201cObviative\u201d suffix \u2013a in Plains Cree have been reconstructed to have the Proto-Algonquian shape of \u273d-ali (Wolfart 1973:14). This means that, in the hypothesized proto-language that Plains Cree is descended from, \u201cInanimate\u201d and \u201cObviative\u201d nominals have a suffix whose phonological shape is also identical. This would mean that the shape of the \u201cInanimate\u201d and \u201cObviative\u201d suffixes has remained stable for somewhere around 3,000 years (cf. Proulx 1984). A survey of other 113 Algonquian languages shows the same stability, with some interesting reorganization in Blackfoot (\u201cInanimate\u201d singular is homophonous with \u201cObviative\u201d rather than \u201cInanimate\u201d plural), and a singular\/non-singular distinction in Fox. Notably, even where there has been a reorganization of the forms, the \u201cObviative\u201d form still parallels the \u201cInanimate\u201d one, as shown in table 3.2.2 \u201cINANIMATE\u201d \u201cOBVIATIVE\u201d LANGUAGE Singular Plural Singular Plural Plains Cree (-i) -a -a -a Swampy Cree -a -a -a East Cree -h -h -h Cree Innu -a -a -a Fox -ani -ahi -ani -ahi Potawatomi \u00d8 -n -n -n Menominee -an -an -an Shawnee \u00d8 -li -li -li Central Algonquian Ojibwa -i -an -an -an Eastern Algonquian Passamaquoddy -ol -ol -ol Algonquian Blackfoot -(y)i -istsi -(y)i -(y)i Table 3.2. \u201cInanimate\u201d plural and \u201cobviative\u201d across Algonquian3 This means that any account of \u201cObviative\u201d and \u201cInanimate\u201d \u2013a that treats them as different suffixes must explain 3,000 years of formal identity, holding across languages that bear significant linguistic, spatial, and temporal distance. As I show in \u00a73.4, \u201cObviative\u201d shares content with \u201cInanimate;\u201d both \u201cInanimate\u201d and \u201cObviative\u201d code an extentional referent, differing only in how long this extentionality is taken to hold. For the current problem, this means that both the \u201cInanimate\u201d usage of \u2013a and the \u201cObviative\u201d usage of \u2013a code that the referent is extentional. Thus, the content contribution that the structural addition of \u2013a makes to a nominal is, minimally, a coding of extentionality. Since both the \u201cInanimate\u201d use of \u2013a and the \u201cObviative\u201d use of \u2013a occur in exactly the same position in nominal and verbal affixation, we could put them in the same position in the syntactic structure. Based on the logic of affixation argued for by D\u00e9chaine (1999) (see \u00a72.2.1), we expect that the linear order of affixes correlates with their hierarchy in the syntax: the closer 2 For a discussion of approaches to the plural-obviative parallel, see \u00a76.2. 3 Sources for this table are Wolfart (1973) (Plains Cree), Ellis (1983) (Swampy Cree), MacKenzie (1980) (East Cree and Innu), Goddard (2004) (Fox), Hockett (1966) (Potawatomi), Bloomfield (1962) (Menominee), Costa (2001) (Shawnee), Valentine (2001) (Ojibwa), Bruening (2001) (Passamaquoddy), and Frantz (1976) (Blackfoot). Where I have no evidence for a form, I leave the table blank. 114 the suffix is to the root, the higher its position in the tree. Since this \u2013a suffix is linearly the most peripheral affix in the nominal structure (cf. Wolfart 1973, D\u00e9chaine 1999), it belongs in the lowest position in the nominal syntax4. (19) DP 3 D PERSONP 3 PERS EXTP 3 -a ROOT (Adapted from D\u00e9chaine 1999:44) 3.3.13. Modelling the \u201cobviative\u201d effects of \u2013a The resolution of a referent as \u201cObviative\u201d or \u201cInanimate\u201d plural, is accomplished by conjoining the structure in (19) with a verb that codes either \u201cInanimate\u201d or \u201cAnimate\u201d for the same referent.5 Combining a nominal marked with \u2013a with an \u201cInanimate\u201d marked verb (e.g. w\u00e2pahtam \u2018see itIN\u2019) yields an \u201cInanimate\u201d plural interpretation (20a), while combining it with an \u201cAnimate\u201d marked verb (e.g. w\u00e2pam\u00eaw \u2018S\/hePROX sees him\/herOBV\u2019) yields an \u201cObviative\u201d interpretation (20b). (20) a. VIN + N-a = \u201cINANIMATE\u201d PLURAL \u141a\u1438\u1426\u1455\u1466 \u146d\u14c0\u1431\u147f w\u00e2pahtam kin\u00eapikwa w\u00e2p=aht -am kin\u00eapikw-a see =by.eye.TI-TI snake -XT \u2018S\/he sawIN (the) snakes.\u2019 (Presented S2) b. VAN + N-a = \u201cOBVIATIVE\u201d \u141a\u1438\u14a3\u1424 \u146d\u14c0\u1431\u147f w\u00e2pam\u00eaw kin\u00eapikwa w\u00e2p=am -\u00ea -w kin\u00eapikw-a see =by.eye.TA-DIR-3 snake -XT \u2018S\/he sawAN a\/the snake(s)OBV.\u2019 (Presented S2) 4 Spec-to-spec movement then derives the final ordering of the morphemes; see D\u00e9chaine (1999:44) for details. 5 For a consideration of the plural\/\u201dObviative\u201d ambiguity, see Appendix A below. 115 This is why, in the verb system, the affixation of \u2013a never appears underspecified; it is always affixed to a predicate that already codes animacy for the referent. Affixing \u2013a to a verb coded for \u201cInanimate\u201d arguments (e.g. miyw\u00e2sin \u2018ItIN is good\u2019) yields an \u201cInanimate\u201d plural interpretation (21a), while affixing it to a verb coded for \u201cAnimate\u201d arguments (e.g. niw\u00e2pamim\u00e2wa \u2018I see him\/herOBV\u2019) yields an \u201cObviative\u201d interpretation (21b). (21) a. VIN + -a = \u201cINANIMATE\u201d PLURAL \u14a5\u1540\u141a\u14ef\u14cc miyw\u00e2sinwa miyw=\u00e2si-n-w-a good =II -0-3-XT \u2018They are niceIN.\u2019 (Presented S2) b. VAN + -a = \u201cOBVIATIVE\u201d \u14c2\u141a\u1438\u14a5\u14ab\u1418 niw\u00e2pamim\u00e2wa ni-w\u00e2p=am -im -\u00e2 -w-a 1-see =by.eye.TA-DISJ-DIR-3-XT \u2018I seeAN him\/her\/it\/themobv.\u2019 (Translation S1) The suffix \u2013a, then, introduces a restrictive operation on the coding of animacy, further narrowing the set of potential referents denoted. 3.3.2. The demonstrative suffix \u2013ihi Plains Cree has a demonstrative system that canonically inflects for two spatial distinctions (proximal \u2018near Speaker\u2019 and distal \u2018far from Speaker\u2019), animacy (\u201cInanimate\u201d and \u201cAnimate\u201d) and plurality (singular and plural). \u201cInanimate\u201d \u201cAnimate\u201d Singular Plural Singular Plural \u201cObviative\u201d Proximal \u00f4ma \u00f4hi awa \u00f4ki \u00f4hi Distal anima anihi ana aniki anihi Table 3.3. Canonical demonstrative organization in Plains Cree A sample form is given in (22); here, the \u201cAnimate\u201d marked demonstrative \u00f4ki modifies the nominal ayisiyiniwak \u2018people.\u2019 116 (22) \u201cANIMATE\u201d DEMONSTRATIVE \u00f4ki \u1401\u146f\u14c2 \u140a\u14c2\u146d \u14c3\u14f1 \u140a\u1528\u14ef\u1528\u14c2\u1418\u1420 \u1401 \u146e \u14c3\u1438\u1424\u140f\u1422\u1455\u14ab\u146f\u152e\u1426\u146d\u1420\u2026 \u00eakonik aniki n\u00eesw \u00e2yisiyiniwak \u00ea-k\u00ee-n\u00eepawistam\u00e2koy\u00e2hkik\u2026 \u00eakonik an =iki n\u00eesw \u00e2yisiyiniw-ak \u00ea- k\u00ee- n\u00eepaw=stamaw-iko -y\u00e2hk-ik resum DST=AN.PL two person -PL C1-PREV-stand =APPLIC -INV-1PL -PL \u2018these two [people] stood up for us\u2026\u2019 (Minde 1997:\u00a742) One consultant I have worked with has a different organization; instead of a division between \u201cAnimate\u201d and \u201cInanimate,\u201d this speaker has a division between forms for \u201cIndependent\u201d referents (i.e. referents that are not embedded in a 3rd person\u2019s perspective; \u00a74), and forms for \u201cDependent\u201d referents (i.e. referents that are embedded in a 3rd person\u2019s perspective; \u00a74). Independent Singular \u201cInanimate\u201d \u201cAnimate\u201d Plural Dependent Proximal \u00f4ma awa \u00f4ki \u00f4hi Distal anima ana aniki anihi Table 3.4. Demonstrative organization of S2 A sample set is given in (23); here, the dependent form \u00f4hi cannot be used in new discourse contexts involving no other third persons (23a). Instead, the preferred demonstrative in this context is \u00f4ki, as in (23b), which is used canonically for \u201cAnimate\u201d plurals. However, when the same nominal maskisin \u2018shoe\u2019 is related to a third person, \u00f4hi is now felicitous (23c). (23) a. \u00f4hi CANNOT MODIFY AN INDEPENDENT \u201cINANIMATE\u201d PLURAL \u273d \u1406\u1426\u1403 \u14aa\u1422\u146d\u14ef\u14c7 \u14a5\u1429\u141a\u14ef\u14cc. \u273d \u00f4hi maskisina miyw\u00e2sinwa. aw =ihi maskisin-a miyw=\u00e2si-n-w-a PROX=XT shoe -XT good=II -0-0-XT Intended: \u201cThese shoesIN are nice.\u201d (Presented S2) b. \u00f4ki CAN MODIFY AN INDEPENDENT \u201cINANIMATE\u201d PLURAL \u1406\u146d \u14aa\u1422\u146d\u14ef\u14c7 \u14a5\u1429\u141a\u14ef\u14cc. \u00f4ki maskisina miyawsina aw =ihi maskisin-a miyw=\u00e2si-n-w-a PROX=XT shoe -XT good =ii -0-0-XT Intended: \u201cThese shoesIN are nice.\u201d (Volunteered S2) 117 c. DEPENDENT \u201cINANIMATE\u201d MODIFIED BY \u00f4hi \u140a\u1411\u14c7 \u1406\u1426\u1403 \u14aa\u1422\u146d\u14ef\u14c7? aw\u00eena \u00f4hi maskisina? aw\u00eena aw=ihi maskisin-a who PRX=XT shoe -XT \u2018Whose shoesIN are these here?\u2019 (Volunteered S2) For both systems, the \u201cObviative\u201d is coded with the suffix \u2013ihi. (24) DEMONSTRAIVE SUFFIX =IHI MODIFIES \u201cOBVIATIVE\u201d REFERENT \u14c7\u14b9\u1428, \u140b\u1426\u148b \u1431\u146f \u1401 \u1434\u14ef\u140d\u1433\u1426\u141a\u141f \u1406\u1426\u1403 \u146d\u1422\u144e\u1472\u14c7 namw\u00e2c, \u00e2hci piko \u00ea-p\u00f4siw\u00eapahw\u00e2t \u00f4hi kistik\u00e2na\u2026 namw\u00e2c, \u00e2hci piko \u00ea-p\u00f4si =w\u00eap =ahw -\u00e2 -t aw =ihi kistik\u00e2n-a NEG still only C1-ride=throw=by.tool.TA-DIR-3 PRX=XT grain -XT \u2018It was in vain, instead hePROX shovelled the grainOBV onto the wagon\u2026\u2019 (AA 2:3) Since this suffix occurs whenever the demonstrative codes an \u201cObviative\u201d referent, it is possible that \u2013ihi is dedicated to coding obviation. 3.3.21. The traditional view: \u2013ihi marks \u201cobviative\u201d As with the analysis of the suffix \u2013a, the demonstrative suffix \u2013ihi has been treated as dedicated to coding \u201cObviative\u201d in most analyses of Plains Cree. Bloomfield (1933), Dahlstrom (1991), and Goddard (1991) have all treated this form as dedicated to coding the \u201cObviative\u201d class of referents. Following this approach, many presentations of Plains Cree data adopt an \u201cObviative\u201d gloss for this suffix (e.g. M\u00fchlbauer 2007, D\u00e9chaine & Wiltschko 2002). Wolfart (1973) adopts the same usage, but, as with \u2013a, notes reservations based on the suffix\u2019s formal parallels to the \u201cInanimate\u201d suffix \u2013ihi. If these analyses are correct, this suffix \u2013ihi is dedicated to coding \u201cObviative,\u201d and thus constitutes evidence for \u201cObviation\u201d as a primitive of the grammar. 118 3.3.22. The current analysis: \u2013ihi marks extentional referents There are reasons to believe that \u2013ihi is not dedicated to coding \u201cObviative\u201d in Plains Cree. There are contexts where \u2013ihi occurs, but does not code \u201cObviation.\u201d However, in all these contexts, \u2013ihi codes that the referent is extentional \u2013 either \u201cInanimate\u201d or \u201cObviative.\u201d As we saw with the suffix \u2013a, the suffix \u2013ihi occurs with \u201cInanimate\u201d plural referents in the canonical organization of Plains Cree demonstratives. (25) DEMONSTRATIVE SUFFIX \u2013IHI MODIFIES \u201cINANIMATE\u201d PLURAL REFERENT \u1401\u147f \u1403\u1528\u146f\u1541 \u1401 \u1401 \u14ab\u1426 \u14a5\u14f5\u146d \u1406\u1426\u1403 cathedrals \u1401 \u1432\u1426\u144e\u1475\u152e\u1541, \u2026 \u00eakwa \u00eayikohk \u00ea-~ \u00ea-m\u00e2h-mis\u00e2ki \u00f4hi cathedrals \u00ea-p\u00eehtikw\u00eay\u00e2hk, \u2026 \u00eakwa iyikohk \u00ea-~ \u00ea- m\u00e2h-mis=\u00e2-k-i aw =ihi cathedral-s \u00ea- p\u00eehtikw\u00ea-y\u00e2hk and so.much C1-~ C1-red- big=II-0-PL PROX-XT cathedral-PL C1-enter -1pl \u2018And the cathedrals we went in were so bigIN, \u2026\u2019 (AA 3:2) As with the suffix \u2013a, this formal identity means that an account that treats \u2013ihi as dedicated to \u201cObviative\u201d coding must consider this a case of accidental homophony.6 The diachronic work on \u2013ihi has not yet been done definitively (cf. Proulx 1988). It is thought that the suffixes used with demonstrative roots were historically particles, and were incorporated into the demonstratives at a later date (Proulx 1988). Since different languages incorporated particles differently, there is no straightforward mapping to be found. Further, there are thought to be at least two sets of demonstrative roots (Proulx 1988),which makes pinpointing reflexes in daughter languages quite complex. This lack of reconstructive evidence means that the historical argument against accidental homophony that was available for the \u201cObviative\/Inanimate\u201d suffix\u2013a is not currently available here. However, if we consider the following sampling of demonstrative forms across Algonquian (given in table 3.5), we observe that the \u201cInanimate\u201d form corresponds to the \u201cObviative\u201d form. For example, the \u201cInanimate\u201d plural proximal demonstrative is noti \u2018these onesIN\u2019 in Potawatomi (Central Algonquian, Wisconsin), and this corresponds to the \u201cObviative\u201d form noti \u2018these\/this oneOBV.\u2019 Further, in all but the Blackfoot case, the parallel is between \u201cInanimate\u201d plural and \u201cObviative.\u201d 6 Note that in some dialects, the form of \u2013ihi is actually \u2013iha (Wolfart 1973). This may be a generalization of \u2013a to the demonstrative system, which normally uses a distinct form for a similar function. 119 PROXIMAL DISTAL \u201cINANIMATE\u201d \u201cINANIMATE\u201d LANGUAGE Sg Pl \u201cOBVIATIVE\u201d Sg Pl \u201cOBVIATIVE \u201d Plains Cree -ima \u2013ihi \u2013ihi -ima \u2013ihi \u2013ihi Cree Atikamekw -he: -hi -hi -he: -hi -hi Fox -i -ini -ini Potawatomi \u0294oti noti noti \u0294i \u0294eni \u0294eni Menominee y\u014dw anoh anoh eneh enoh enoh Central Algonquian Shawnee yoma yohoma yohoma hini nihi nihi Eastern Algonquian Delaware -\u00fa -\u00f3:l -\u00f3:l -\u00ed -\u00e9:l -\u00e9:l Algonquian Blackfoot -\u00ed\u00ed -\u00ed\u00edstsi -\u00ed\u00ed -\u00ed\u00ed -\u00ed\u00edstsi -\u00ed\u00ed Table 3.5. \u201cInanimate\u201d plural and \u201cobviative\u201d demonstratives across Algonquian This means that languages separated by great temporal (3,000 years) and spatial distance (more than 4,000 miles) maintain a parallel between \u201cInanimate\u201d plural and \u201cObviative.\u201d In fact, these languages maintain this parallel even when the forms used are different (compare PC \u2013ihi to PQ \u2013\u00f3:l). Were the system in Plains Cree a case of accidental homophony, these parallels across Algonquian would be difficult to explain. Extending the argumentation from the modelling of \u2013a (\u00a73.3.12), we can say that there is only one suffix \u2013ihi, and two contexts of occurrence for it. As with the nominal suffix \u2013a, the contribution of \u2013ihi to the structure is extentionality; the referent associated with the demonstrative is extentional. The two morphological components of a demonstrative like \u00f4hi (aw=ihi) can be represented as being in two different positions. The first element, which codes the spatial deictic component (e.g. aw\u2013 \u2018near Speaker\u2019), heads the demonstrive phrase. The second element, which codes the features of the referent (e.g. \u2013ihi \u2018extentionality\u2019 or \u2013iki \u2018plural\u2019), introduces an argument in the phrase that is complement of the demonstrative. For example, in (26) =ihi heads an extentional phrase and introduces an extentional argument. (26) DEMP 3 DEM EXTP aw- 3 pro 3 EXT DP \u2013ihi 6 120 3.3.23. Modelling the \u201cobviative\u201d effects of \u2013ihi Just as with \u2013a, the coding of extentionality is all that the suffix \u2013ihi needs to do. Further determination of an \u201cObviative\u201d (contextually extentional) or \u201cInanimate\u201d (inherently extentional) referent is accomplished by combining this demonstrative with other animacy-coded forms. (27) a. VIN + DEM=ihi = \u201cINANIMATE\u201d PLURAL \u141a\u1438\u1426\u1455\u1466 \u1406\u1426\u1403 w\u00e2pahtam \u00f4hi w\u00e2p=aht -am aw =ihi see =by.eye.TI-TI PROX-XT \u2018S\/he sawIN these.\u2019 (Presented S2) b. VAN + DEM=ihi = \u201cOBVIATIVE\u201d \u141a\u1438\u14a3\u1424 \u1406\u1426\u1403 w\u00e2pam\u00eaw \u00f4hi w\u00e2p=am -\u00ea -w aw =ihi see =by.eye.TA-DIR-3 PROX-XT \u2018S\/he sawAN this oneOBV \/ theseOBV.\u2019 (Presented S2) The demonstrative suffix \u2013ihi introduces a restriction on the coding of animacy, just as the suffix \u2013a does. 3.3.3. The theme sign \u2013\u00ea\u2013 The theme sign \u2013\u00ea\u2013 is one of a set of morphemes known as theme signs; they code part of the argument structure of transitive verbs (see \u00a74.4.2 for detailed discussion). Within the independent order of clause-typing, this particular theme sign covers two argument structure configurations (Wolfart 1973). First, when the subject of the verb is \u201cProximate,\u201d and the object is \u201cObviative,\u201d we get \u2013 \u00ea\u2013, as in (28). 121 (28) 3 > 3\u2019 \u141a\u1438\u14a3\u1424. w\u00e2pam\u00eaw. w\u00e2p=am -\u00ea -w see =by.eye.TA-DIR-3 \u2018S\/heprox sees him\/herobv\u2019 (Presented S2) Second, when the subject of the verb is \u201cObviative\u201d and the object is also \u201cObviative\u201d we get \u2013 \u00ea\u2013, as in (29). (29) 3\u2019 > 3\u2019 \u141a\u1438\u14a3\u1528\u1418 w\u00e2pam\u00eayiwa w\u00e2p=am -\u00ea -yi-w-a see=by.eye.TA-DIR-DS-3-XT \u2018S\/heobv sees him\/herobv\u2019 (Wolfart & Carroll 1973) The theme sign \u2013\u00ea\u2013 appears to be insensitive to the \u201cObviative\u201d status of the verb\u2019s subject, but always occurs with an \u201cObviative\u201d object. This would seem to indicate that \u2013\u00ea\u2013 codes obviation for objects, meaning that it is dedicated to coding \u201cObviative.\u201d 3.3.31. The traditional view: \u2013\u00ea\u2013 marks \u201cthird person interactions\u201d Analyses of \u2013\u00ea\u2013 are widely divergent, but very little has been said specifically about this morpheme in the literature. In particular, two main views may be distinguished: (i) \u2013\u00ea\u2013 codes \u201cDirection\u201d on a \u201cPerson Hierarchy\u201d (Blain 1997, Ritter & Rosen 2005). Under such a view, \u2013\u00ea\u2013 codes that the third person subject of the verb is ranked higher than the object of the verb along a hierarchy of persons. (30) PERSON HIERARCHY ANALYSIS OF \u2013\u00ea\u2013 \u2013\u00ea\u2013 \uf0e0 2 1 3 3\u2019 0 (ii) \u2013\u00ea\u2013 codes that the subject-object interaction is between third persons, with the rest of the specification being dependent on further affixation (Wolfart 1973, Dahlstrom 1991). On this view, only strings of forms can be characterized for person interactions. For example, Wolfart (1973) discusses forms like \u2013\u00eawak as \u20183rd plural acting on obviative\u2019 (31a) or \u2013 122 \u00eayiwa as \u2018Obviative acting on obviative\u2019 (31b), but does not analyze the theme sign \u2013\u00ea\u2013 in much detail. (31) a. \u2013\u00eawak = \u201cPROXIMATE\u201d PLURAL ACTING ON \u201cOBVIATIVE\u201d \u141a\u1438\u14a3\u1418\u1420. w\u00e2pam\u00eawak. w\u00e2p=am -\u00ea -w-ak see =by.eye.TA-DIR-3-PL \u2018S\/hePROX sees him\/herOBV.\u2019 (Presented S2) b. \u2013\u00eayiwa = \u201cOBVIATIVE\u201d ACTING ON \u201cOBVIATIVE\u201d \u141a\u1438\u14a3\u1528\u1418. w\u00e2pam\u00eayiwa. w\u00e2p=am -\u00ea -yi-w-a see =by.eye.TA-DIR-DS-3-XT \u2018S\/heOBV sees him\/herOBV.\u2019 (Dahlstrom 1986) On both views, \u2013\u00ea\u2013 is restricted to third person \u201cAnimate\u201d interactions, and has either a \u201cPerson Hierarchy\u201d function, or an abstract argument structure. 3.3.32. The current analysis: \u2013\u00ea\u2013 marks extentional objects There are reasons to believe that \u2013\u00ea\u2013 is not dedicated to coding \u201cObviative\u201d or even \u201cAnimate\u201d interactions in Plains Cree. There are contexts where \u2013\u00ea\u2013 occurs, but does not code \u201cObviation.\u201d However, in all these contexts, \u2013\u00ea\u2013 codes extentional objects \u2013 either \u201cInanimate\u201d or \u201cObviative.\u201d In elicitation, the object of an independent verb marked with -\u00ea- need not have overt nominal \u201cObviative\u201d marking (\u2013a) (cf. Cook & M\u00fchlbauer 2007b). (32) NOMINALS LACK \u201cOBVIATIVE\u201d CODING Clare \u1405\u144e\u14c0\u1424 \u140a\u14a5\u1422\u146f \u140a\u1528\u1422 \u1401 \u1411 \u14a7\u141a\u141f Clare otin\u00eaw amisk ayis \u00ea-w\u00ee-mow\u00e2t Clare ot=in -\u00ea -w amiskw ayis \u00ea- w\u00ee- mow-\u00e2 -t Clare take=by.hand-DIR-3 beaver for C1-intend-eat -DIR-3 \u2018Clare? took a beaver? to eat it.\u2019 (Translation S2) Here, the object of the verb otin\u00eaw \u2018take\u2019 is amisk \u2018beaver,\u2019 which is not marked for extentionality. This state of affairs is extremely common in elicitation when there is other means to disambiguate referent. Contexts that allows the non-occurrence of the suffix \u2013a are: 123 (i) The subject occuring linearly before the object. (ii) The alternative reading is pragmatically difficult (e.g. a beaver taking Clare to eat her). Thus, in examples like (33), the reference is easy to resolve because the subject (Solveiga) precedes the object (Clare). (33) SUBJECT PRECEDES OBJECT Solveiga \u1434\u14c2 \u14a5\u1526\u1528\u14a3\u1424 Clare Solveiga p\u00f4n-miy\u00eayim\u00eaw Clare Solveiga p\u00f4n- miyw=\u00eayim -\u00ea -w Clare Solveiga stop-good=by.mind.TA-DIR-3 Clare \u2018Solveiga(?) stopped liking Clare(?).\u2019 (Translation S2) While the nominals in (32) and (33) were bare, they can also be overtly coded for \u201cAnimate\u201d (non-obviative), as the example in (34) shows. Here, the nominal atimw \u2018dog\u2019 has the \u201cAnimate\u201d plural suffix \u2013ak applied (34a), rather than the expected extentional suffix \u2013a (34b). (34) \u201cANIMATE\u201d PLURAL \u2013ak UNEXPECTEDLY OCCURS IN \u201cOBVIATIVE\u201d CONTEXTS a. David \u14a5\u1426\u1489\u141f \u140a\u144e\u14b7\u1420 \u140a\u152e\u140d\u1424. David mihcet atimwak ay\u00e2w\u00eaw. David mihc\u00eat atimw-ak ay\u00e2w -\u00ea -w Davice many dog -PL have.TA-DIR-3 \u2018DavidAN has many dogsAN.\u2019 (Translation S2) b. David \u14a5\u1426\u1489\u141f \u140a\u144e\u14b7 \u140a\u152e\u140d\u1424 David mihcet atimwa ay\u00e2w\u00eaw David mihc\u00eat atimw-a ay\u00e2w -\u00ea -w Davice many dog -XT have.TA-DIR-3 \u2018DavidAN has many dogsOBV.\u2019 (Translation S2) In elicitation, then, we see that \u201cnon-Obviative\u201d referents may act on each other with the morpheme \u2013\u00ea\u2013. Even if we discount this elicitation data, which does not match any known text data, there are further difficulties for an obviation-specificity account of \u2013\u00ea\u2013. There are extant textual cases where the object of \u2013\u00ea\u2013 marked verbs appears to be \u201cInanimate,\u201d not \u201cObviative.\u201d 124 (35) \u2013\u00ea\u2013 OCCURS WITH \u201cINANIMATE\u201d OBJECT \u14f2\u1422\u1481\u1428 \u1472\u1426\u146d\u152d\u1424 \u1472\u1475\u148b\u14a3\u1424 \u1405\u1456\u1438\u148b\u1426\u148b\u1472\u14c7 s\u00f4skw\u00e2c kahkiyaw kakw\u00eacim\u00eaw ot\u00e2pacihcikana s\u00f4skw\u00e2c kahkiyaw kakw\u00eac=im -\u00ea -w ot-\u00e2pac=iht =ikan-a straight.out all try =by.mouth-DIR-3 3- use =by.tool.TI=NOM-XT \u2018Without delay, shePROX asked all her utensils\u2019 (Bloomfield 1933, Cited in Wolfart 1973) In cases like this, we could claim that the noun is de facto \u201cObviative\u201d because the verb has the morpheme -\u00ea-. However, Wolfart (1973:14) uses this form to exemplify the ambiguity of nominal marking (is it \u201cObviative\u201d \u2013a or \u201cInanimate\u201d plural \u2013a?) and concludes that the form is entirely ambiguous. Thus, there appear to be possible cases where \u2013\u00ea\u2013 codes an \u201cInanimate\u201d, not an \u201cObviative,\u201d object. The use of \u2013\u00ea\u2013 with a potentially \u201cInanimate\u201d object in this last form, kakw\u00eacim\u00eaw, suggests that the object of a verb marked with \u2013\u00ea\u2013 may, under certain conditions, be \u201cInanimate.\u201d If we turn from \u201cAnimate\u201d stems to \u201cInanimate\u201d stems, we again find a theme sign \u2013\u00ea\u2013. When the stem codes an inanimate object, and a speech-act-participant is acting on this \u201cInanimate\u201d object, \u2013\u00ea\u2013 is used (36b). (36) a. 3 > \u201cINANIMATE\u201d \u141a\u1438\u1426\u1455\u1466. w\u00e2pahtam. w\u00e2p=aht -am see =by.eye.TI-TI \u2018s\/he seesIN it\u2019 (Presented S2) b. SAP > \u201cINANIMATE\u201d \u14c2\u141a\u1438\u1426\u144c\u1423. niw\u00e2paht\u00ean. ni-w\u00e2p=aht -\u00ea -n 1- see =by.eye.TI-TI-LP \u2018I seeIN it\u2019 (Presented S2) In these constructions, the theme sign \u2013\u00ea\u2013 combines with stem-level \u201cInanimate\u201d agreement and pronominal marking to code an \u201cInanimate\u201d object. This \u2013\u00ea\u2013 is in complementary distribution with the -\u00ea- we have seen so far. The first \u2013\u00ea\u2013 occurs only with third-person interactions in the independent order TA paradigm, while the second \u2013\u00ea\u2013 occurs only with speech-act participants acting on \u201cInanimate\u201d referents the independent order TI paradigm. 125 PATIENT THEMES (IND) 1 2 3 OBV INAN 1 \u273d -iti- -\u00e2- -\u00e2- -\u00ea- 2 -i- \u273d -\u00e2- -\u00e2- -\u00ea- 3 -ikw- -ikw- \u273d [-\u00ea-]7 -\u00ea- -am- OBV -ikw- -ikw- -ikw- -\u00ea- -am- AGENT INAN -ikw- -ikw- -ikw- -ikw- (-am-) Table 3.6.: Independent order distribution of -\u00ea- Neither of these theme signs are used in the Conjunct order. Instead, the theme sign \u2013\u00e2\u2013 is used for all \u201cAnimate\u201d direct forms, and the theme sign \u2013am is used for all \u201cInanimate\u201d direct forms.8 PATIENT THEMES (CON) 1 2 3 OBV INAN 1 \u273d -iti- -\u00e2- -\u00e2- -am- 2 -i- \u273d -\u00e2- -\u00e2- -am- 3 -ikw- -ikw- \u273d [-\u00e2-] -\u00e2- -am- OBV -ikw- -ikw- -ikw- -\u00e2- -am- AGENT INAN -ikw- -ikw- -ikw- -ikw- -am- Table 3.7. No -\u00ea- in conjunct order When two elements are in this kind of perfect complementary distribution, there are two possible solutions: (i) The two \u2013\u00ea\u2013 theme signs are entirely different; there is accidental homophony. (ii) The two \u2013\u00ea\u2013 theme signs are exactly the same. Considering both the phonological identify (both are [e]) and the structural identity (both are theme signs), it is likely that the two putative morphemes are really only one form. The evidence suggests that \u2013\u00ea\u2013 occurs with extentional objects. It occurs in two places: (i) \u201cObviative\u201d objects (ii) \u201cInanimate\u201d objects. On the current view, these two referential classes share the property of extentionality; the \u201cInanimate\u201d class codes inherent extentionality, and the \u201cObviative\u201d class codes contextual extentionality. The use of one morpheme for both classes is expected. As with the nominal suffix \u2013a, and the demonstrative suffix \u2013ihi, the contribution of the theme sign \u2013\u00ea\u2013 is extentionality; the referent is coded as extentional. Following D\u00e9chaine (2003), Hirose (2000), and D\u00e9chaine & Reinholz (2008), the morpheme \u2013\u00ea\u2013 can be represented as 7 The brackets indicate the form gotten in elicitation under these conditions. 8 Note that this is not a cross-Algonquian pattern. While Menominee shows the same pattern (Bloomfield 1962), balancing \u2013a\u00b7\u2013 with \u2013ae\u2013, other languages like Nishnabemwin appear to not have the TA version of \u2013\u00ea\u2013 (Valentine 2001). 126 occupying the head of VP (see \u00a72.2.1 for discussion). It introduces an extentional argument in its specifier. (37) vP 3 ROOT vP 3 pro 3 v VP 3 AA pro 3 -\u00ea- [EXT] 3.3.33. Modelling the \u201cobviative\u201d effects of \u2013\u00ea\u2013 I have claimed that \u2013\u00ea\u2013 is the head of VP, and introduces an extentional argument, which does not distinguish between \u201cInanimate\u201d and \u201cObviative.\u201d Distinguishing between \u201cInanimate\u201d and \u201cObviative\u201d is done by merging vP, which codes the inherent extentional properties of this argument. If the final in the head of vP is \u201cInanimate\u201d (inherently extentional) the argument introduced by \u2013\u00ea\u2013 in the head of VP is \u201cInanimate\u201d (38). (38) \u201cINANIMATE\u201d FINAL IN VP = \u201cINANIMATE\u201d ARGUMENT FOR VP vP 3 ROOT 3 pro 3 v VP [EXT] 3 AA pro 3 -\u00ea- [EXT] If the final in the head of vP is \u201cAnimate\u201d (unspecified for inherent extentionality), then the argument introduced by \u2013\u00ea\u2013 in the head of VP is interpreted as \u201cObviative\u201d (39). 127 (39) \u201cANIMATE\u201d FINAL IN VP = \u201cOBVIATIVE\u201d ARGUMENT FOR VP vP 3 ROOT 3 pro 3 v VP [\u2205] 3 AA pro 3 -\u00ea- [EXT] 3.3.4. The predicate suffix \u2013im\u2013 Another verbal morpheme that appears to correlate with \u201cObviative\u201d referents is the suffix -im-. This morpheme affixes directly to the stem, interior to theme signs. (40) \u14c2\u141a\u1438\u14a5\u14ab\u1418 \u141a\u1438\u1422\u144e\u1466 \u1405\u14a5\u14c5\u14ef\u14aa niw\u00e2pamim\u00e2wa W\u00e2pastim omin\u00f4sima ni-w\u00e2p=am -im -\u00e2 -w-a w\u00e2pastim o-min\u00f4s-im -a 1-see=see.by.eye.TA-DISJ-DIR-3-XT W\u00e2pastimw 3-cat -DISJ-XT \u2018I saw W\u00e2pastim\u2019sPROX catOBV \u2026\u2019 (Volunteered S1) This morpheme occurs when the \u201cObviative\u201d referent is the object of a verb that has a speech act participant as its subject. In table 3.8, we see that \u2013im\u2013 is possible when an \u201cObviative\u201d acts on an \u201cObviative,\u201d when a \u201cProximate\u201d acts on some \u201cObviatives,\u201d and when a speech act participant acts on an \u201cObviative.\u201d INTRANSITIVE TRANSITIVE X > OBV OBV OBV > X OBV > OBV PROX > OBV 1\/2 > OBV -im- \u273d \u273d \u2714 \u273d\/\u2714 \u2714 Table 3.8. Distribution of -im- with \u201cobviative\u201d arguments This distribution suggests that \u2013im\u2013 could be dedicated to coding obviation with objects. 128 3.3.41. The traditional view: \u2013im\u2013 marks \u201cobviative\u201d objects The descriptions of this morpheme that are most influential have treated it as an \u201cObviative\u201d agreement marker. Wolfart (1973), for example, calls it a \u201cthematic obviative sign\u201d (Wolfart 1973:47), while Dahlstrom (1986) refers to it simply as an \u201cObviative\u201d suffix,\u201d saying that it \u201cmarks obviative objects\u201d (Dahlstrom 1986:60), and \u201cspecifies that the object is the \u2018further obviative\u2019\u201d (Dahlstrom 1986:56). This is based on evidence such as the example in (41); under some contexts, -im- shows up when the object is \u201cObviative.\u201d As we will see, however, a close consideration of these contexts casts doubt on the status of -im- as an \u2018\u201cObviative\u201d\u2019 morpheme in the verbal complex. 3.3.42. The current analysis: \u2013im\u2013 marks argument disjunction There are reasons to believe that \u2013im\u2013 is not dedicated to coding \u201cObviative\u201d objects. There are contexts where \u2013im\u2013 occurs, but does not code \u201cObviation.\u201d However, in all these contexts, \u2013im\u2013 codes that the object is disjoint from another potential argument in the clause (A disjunction: Saxon 1986 for Dogrib). Troubling for an account of -im- that treats it as simply \u201cObviative\u201d agreement, most \u201cObviative\u201d objects do not have it. For example, when \u201cProximate\u201d interacts with an \u201dObviative\u201d, there is no \u2013im\u2013. (41) \u201cPROXIMATE\u201d ACTS ON \u201cOBVIATIVE\u201d = no \u2013im\u2013 \u1401\u146f\u1455 \u1401\u147f \u1403\u1450\u1426\u1455\u1426\u1401\u1424 \u140a\u14c7 \u1411\u1418 \u00eakot\u00ea \u00eakwa itohtah\u00eaw ana w\u00eewa \u00eakot\u00ea \u00eakwa it =oht =ah -\u00ea -w an -a w-\u00eew -a there then thus=walk=CAUS-DIR-3 DST-AN.SG 3-wife-XT \u2018then thatPROX [man] took hisPROX wifeOBV there;\u2019 (Ahenakew 2000:5.5) Nor does a possessed \u201cObviative\u201d require it. (42) POSSESSED \u201cOBVIATIVE\u201d OBJECT DOES NOT REQUIRE \u2013im\u2013 \u14c0\u1413 \u146b\u147f\u1429 \u146e \u14a5\u1526\u1424 \u1405\u1455\u141a\u14ef\u14a5\u14f4 \u140a\u1418 \u14c8\u142f\u1424, n\u00eawo k\u00eakway k\u00ee-miy\u00eaw otaw\u00e2simisa awa n\u00e2p\u00eaw, n\u00eawo k\u00eekway k\u00ee- miy -\u00ea -w ot-aw\u00e2s-im -is -a aw-a n\u00e2p\u00eaw four thing PREV-give-DIR-3 3- child-DISJ-DIM-XT PROX-AN.SG man \u2018Four things that man had given his children,\u2019 (Moosoomin in Bloomfield 1930) Neither does -im- occur regularly when an \u201cObviative\u201d is acted on by a local person, although this interaction is so rare in texts (i.e. almost non-existent) that elicitation data had to be relied on 129 instead. Only one of the consultants I have worked with (S1) appears to use this suffix at all, and in examples like (43), where a first person acts on an obviative referent, \u2013im\u2013 is not necessary. (43) \u2013im\u2013 NOT NECESSARY WITH \u201cOBVIATIVE\u201d OBJECT \u14c2\u141a\u1438\u14ab\u1424 \u1405\u14a5\u14c5\u14ef\u14aa niw\u00e2pam\u00e2w omin\u00f4sima ni-w\u00e2p=am -\u00e2 -w o-min\u00f4s-im -a 1-see =by.eye.TA-DIR-3 3-cat -DSJ-XT \u2018I saw [W\u00e2pastimw\u2019sPROX ] catOBV\u2019 (Volunteered S1) In all, textual occurrences of -im- are exceedingly rare; only 10 cases were found in four book- length texts (Ahenakew 2000, Minde 1997, Whitecalf 1993, K\u00e2-pimw\u00eaw\u00eahahk 1998). This would be puzzling were -im- a kind of \u201cObviative\u201d object marking. If it were, it should be mechanically applied to every verb that bears an \u201cObviative\u201d object. However, as the data clearly demonstrates, this is not the case; the actual conditions of use on -im- with \u201cObviative\u201d objects are much more specific. Usually, when the suffix -im- is used in an independent order verb, it co-occurs with the extentional suffix \u2013a. (44) \u201cOBVIATIVE\u201d OBJECT WITH \u2013im\u2013 CO-OCCURS WITH EXTENTIONAL \u2013a \u14ab\u1472 \u140b\u14f4\u1429 \u14a5\u1456\u1455\u1426\u141f \u1405\u144c\u14a5\u141a\u1418 \u1406\u1455 \u14c2\u142f\u14f0\u14ab\u1418 m\u00e2ka \u00e2say mit\u00e2taht ot\u00eamiw\u00e2wa \u00f4tah nip\u00eas\u00eem\u00e2wa! m\u00e2ka \u00e2say mit\u00e2taht o-t\u00eam-iw\u00e2w-a \u00f4ta ni-p\u00easi -im -\u00e2 -w-a but already ten 3-horse-3pl -XT here 1- bring-DISJ-DIR-3-XT \u201cBut I have already brought ten of theirPROX horsesOBV here!\u201d (Dahlstrom 1986:117) If \u2013im\u2013 could only occur with \u201cObviative\u201d objects, these two morphemes should always pattern together in this context. However, there are known exceptions to this correlation. The example in (45a) shows a verb form nik\u00ee-w\u00e2pamim\u00e2wak \u2018I saw them\u2019 that bears both \u201cAnimate\u201d plural agreement (-ak), and the suffix \u2013im\u2013. 130 (45) \u2013im\u2013 CO-OCCURS WITH \u2013ak \u146b\u1455\u1426\u1455\u140d \u14ab\u14c7 \u14c2\u146e \u141a\u1438\u14a5\u14ab\u1418\u1420 \u14a6\u14c7 \u14ab\u14c7 \u1401 \u146e \u1472\u14c7\u140d\u1528\u14ab\u148b\u1420 \u1406\u1426\u1403 \u140b\u152d k\u00eatahtaw\u00ea m\u00e2na nik\u00ee-w\u00e2pamim\u00e2wak m\u00eena m\u00e2n \u00ea-k\u00ee-kanaw\u00eayim\u00e2cik \u00f4h \u00e2ya k\u00eatahtaw\u00ea m\u00e2na ni-k\u00ee- w\u00e2p=am -im -\u00e2 -w-ak m\u00eena m\u00e2na that.time usual 1- PREV-see=by.eye.TA-DSJ-DIR-3-PL also usual \u00ea- k\u00ee- kanaw=\u00eayim -\u00e2 -t-ik aw =ihi aya C1-PREV-keep =by.mind.TA-DIR-3-pl PRX=XT CONN \u2018Later I also used to see [these little shells], theyPROX used to keep themOBV\u2026\u2019 (AA 10.4) If \u2013im\u2013 codes obviative, the form in (45) should be as shown in (46), where the suffix \u2013a occurs instead of \u2013ak. (46) \u2013im\u2013 CO-OCCURS WITH \u2013a \u14c2\u141a\u1438\u14a5\u14ab\u1418 \u141a\u1438\u1422\u144e\u1466 \u1405\u14a5\u14c5\u14ef\u14aa niw\u00e2pamim\u00e2wa W\u00e2pastim omin\u00f4sima ni-w\u00e2p=am -im -\u00e2 -w-a w\u00e2pastim o-min\u00f4s-im -a 1-see=see.by.eye.TA-DISJ-DIR-3-XT W\u00e2pastimw 3-cat -DSJ-XT \u2018I saw W\u00e2pastim\u2019sPROX catOBV \u2026\u2019 (Volunteered S1) This is a case where the suffix \u2013a and \u2013im\u2013 are not appearing together, which contradicts the expectations of an analysis that treats \u2013im\u2013 as coding obviation. In the verbal domain, there are many appearances of the suffix \u2013im\u2013 or a homophonous counterpart: First, it occurs when a verb\u2019s agent is suppressed and the patient is \u201cObviative\u201d (47). (47) IMPERSONAL SUBJECT WITH \u201cOBVIATIVE\u201d OBJECT \u14a5\u1450\u14c2 \u140b\u1426 \u140a\u1473\u141a\u144e\u14ab\u1418, mitoni \u00e2h-ak\u00e2w\u00e2tim\u00e2wa, mitoni \u00e2h- ak\u00e2w\u00e2t-im -\u00e2 -w-a very RED-lust -DSJ-DIR-3-XT \u2018There was great desire for herOBV on all sides,\u2019 (AA 12.3) Second, extending this usage into the relational system, suffix -im- is added when kinship terms are not possessed by any discourse referent (Pentland; p.c. cited in Junker 2003), as exemplified in (48b). 131 (48) a. POSSESSED RELATIONAL NOMINAL \u1405\u146f\u14ef\u14f4 okosisa o-kosis-a 3-son -XT \u2018his\/her son\u2019 (Translation S2) b. NON-POSSSESSED RELATIONAL NOMINAL TAKES \u2013im\u2013 \u1405\u146f\u14ef\u14ef\u14ab\u1424 okosisim\u00e2w o-kosis-im-\u00e2-w 3-son -DISJ-DIR-3 \u2018a son\u2019 Lit: \u2018he is had as a son\u2019 (Presented S4) Third, a morpheme with the identical form \u2013im\u2013 is used to make relationalized constructions out of intransitive verbs, usually in concert with the prefix w\u00eet-. This morpheme also appears to occur in exactly the same place in the verb system as the \u201cObviative\u201d \u2013im\u2013, immediately after the stem. (49) \u2013im\u2013 OCCURS IN RELATIONAL CONSTRUCTIONS a. \u1473 \u142f \u1411\u1455\u1431\u14aa\u146d\u1420 \u1406\u146d \u2026 k\u00e2-p\u00ea-w\u00eetapimakik \u00f4ki. k\u00e2-p\u00ea- w\u00eet- api-im -ak -ik aw-iki C2-come-with-sit -DISJ-3>1-pl PROX-AN.PL \u2018\u2026 when I come and sit here with them.\u2019 (JKN \u00a71.2) b. \u1401 \u1411\u1455\u1431\u14ab\u141f \u00ea-w\u00eetapim\u00e2t \u00ea- w\u00eet =api-m -\u00e2 -t c1-with=sit -DISJ-DIR-3 \u2018s\/hePROX sits with him\/herOBV.\u2019 (Presented S2) The application of this -im- before theme signs is common enough that it can cause language learners to overgenerate forms; learners can accidently use full TI verbs (which end in the theme -am) as though they were relational constructions. 132 (50) a. LEARNER INCORRECTLY OVER-GENERALIZES \u2013im\u2013 \u273d \u1456\u14c2\u14ef \u1401 \u1403\u14ef \u1403\u144c\u1528\u1426\u144e\u14ab\u148b\u1420 \u14c0\u1426\u1403\u152d\u1418\u1420 \u14a5\u1422\u1455\u144e\u14b7? \u273d t\u00e2nisi \u00ea-si-it\u00eayihtim\u00e2cik n\u00eahiyawak mistatimwa? t\u00e2n=isi \u00ea- isi- it =\u00eayiht -im -\u00e2 -t-ik n\u00eahiyaw-ak mistatimw-a Q=thus C1-thus-thus=by.mind.TI-DISJ-DIR-3-PL cree -PL horse -XT Intended: \u201cWhat do the CreePROX think about horsesOBV?\u201d (Produced by me when interviewing a Speaker) b. CORRECTED FORM \u1456\u14c2\u14ef \u1401 \u1403\u14ef \u1403\u144c\u1528\u1426\u144e\u14ab\u148b\u1420 \u14c0\u1426\u1403\u152d\u1418\u1420 \u14a5\u1422\u1455\u144e\u14b7? t\u00e2nsi \u00ea-si-it\u00eayim\u00e2cik n\u00eahiyawak mistatimwa? t\u00e2n=isi \u00ea- isi- it =\u00eayim -\u00e2 -t-ik n\u00eahiyaw-ak mistatimw-a Q=thus C1-thus-thus=by.mind.TA-DIR-3-PL cree -PL horse -XT \u201cWhat do the CreePROX think about horsesOBV?\u201d (Corrected form) Finally, there is an \u2013im\u2013 that also occurs immediately after the stem of some nominal possessor constructions, as in (51b). (51) a. POSSESSED NOMINAL WITHOUT \u2013im\u2013 \u14c2\u14aa\u1422\u146d\u14ef\u1423 nimaskisin ni-maskisin 1-shoe \u2018my shoe\u2019 (Volunteered S2) b. POSSESSED NOMINAL WITH \u2013im\u2013 \u14c2\u14f0\u14f0\u1431\u1466 nis\u00ees\u00eepim ni-s\u00ees\u00eep-im 1-duck-DISJ \u2018My duck\u2019 (Volunteered S1) If all these instances of \u2013im\u2013 are all really the same morpheme, then we here have numerous examples of \u2013im\u2013 occurring in non-\u201cObviative\u201d contexts. Thus, it cannot be said that the morpheme codes obviation. It occurs in many contexts where there is no \u201cObviative\u201d form present, either in the immediate string or in the discourse. Instead, the data suggests that \u2013im\u2013 is a morpheme that affects relational argument structure, and has been recruited to code obviation in some circumstances. 133 3.3.43. Modelling the \u201cobviative\u201d effects of \u2013im\u2013 The suffix -im- can be understood as a restrictor on arguments, coding that they are disjunct from some other argument. When -im- is introduced in the higher head (vP or nP), this disjunction ranges between its argument and the argument of the lower phrase (VP, NP). When it is introduced in the lower head, it ranges across possible objects in the clause. Following D\u00e9chaine (2003), I take objects to be associated with the lower of the two v heads of the verbal syntax, as exemplified in (52). (52) vP 3 ROOT 3 Subject 3 v VP 3 AA Object 3 V The suffix \u2013im\u2013, then, is inserted into the argument position of the lower VP (Spec,VP), restricting its possible reference, in much the same way that an incorporated form does (cf. Hirose 2000:128, Chung & Ladusaw 2003). For example, (54) shows the structure of a verb niw\u00e2pamim\u00e2wa \u2018I see him\/herOBV,\u2019 which has an \u201cObviative\u201d object. The suffix \u2013im\u2013 is inserted in the specifier of the VP, and codes that the object is disjoint from some other clausal nominal (here, W\u00e2pastim). (53) \u2013im\u2013 = DISJOINT OBJECT \u14c2\u141a\u1438\u14a5\u14ab\u1418 \u141a\u1438\u1422\u144e\u1466 \u1405\u14a5\u14c5\u14ef\u14aa niw\u00e2pamim\u00e2wa W\u00e2pastim omin\u00f4sima ni-w\u00e2p=am -im -\u00e2 -w-a w\u00e2pastim o-min\u00f4s-im -a 1-see=see.by.eye.TA-DISJ-DIR-3-XT W\u00e2pastimw 3-cat -DISJ-XT \u2018I saw W\u00e2pastim\u2019sPROX catOBV \u2026\u2019 (Volunteered S1) 134 CP 3 C IP 3 I VP 3 ROOT 3 w\u00e2p- pro 3 V VP -am 3 -im- 3 V -\u00e2- This parallels the structure for an incorporated nominal. (54) shows a verb form kisip\u00eakin- \u2018clean by hand,\u2019 which has an incorporated nomina \u2013iy\u00e2kan \u2018dish.\u2019 This incorporated nominal is inserted in the specifier of the VP, to restrict the interpretation of this argument to only dishes (cf. Chung & Ladusaw 2003). (54) INCORPORATED NOMINAL SYNTAX kis\u00eep\u00eak=in -iy\u00e2kan-\u00ea- clean =by.hand-dish -AI CP 3 C IP 3 I VP 3 ROOT 3 kis\u00eep\u00eak- pro 3 V VP -in 3 -iy\u00e2kan- 3 V -\u00ea- (Adapted from D\u00e9chaine 2003, Hirose 2000) 135 When -im- is used to add an argument to an intransitive verb (the relational use described in (50) and exemplified in (55) below), it is inserted in the specifier of the higher, transitive head (vP). In this role, it implicates the existence of a verbal head that is not spelled-out overtly.9 (55) RELATIONAL VERB SYNTAX \u1401 \u1411\u1455\u1431\u14ab\u141f \u00ea-w\u00eetapim\u00e2t \u00ea- w\u00eet =api-im -\u00e2 -t c1-with=sit -DSJ-DIR-3 \u2018s\/hePROX sits with him\/herOBV.\u2019 (Presented S2) CP 3 C IP 3 I VP 3 ROOT 3 api- -im- 3 V VP \u2205 3 pro 3 V -\u00e2- This same structure is employed with the nominal uses of \u2013im\u2013. The suffix is inserted in the specifier higher, transitive nominal phrase (nP), as in (57). The presence of this suffix in the specifier position implicates the existence of the transitive head of its phrase, which is not spelled out (Koopman 2000). 9 Deciding on the right structural representation of intransitive verbs is complicated by the lack of finals (i.e. overt V heads) in many forms. In earlier analyses (e.g. Bloomfield 1962), these intransitive forms were posited to have a null final. Hirose (2000), on the other hand, modelled these forms by placing the root in the VP itself, which made it function as a final itself. For the sake of a systematic representation, I have here stayed with the more traditional framework suggested by Bloomfield (1962) and others; the intransitive verb is a root without a final. 136 (56) SYNTAX OF \u2013im\u2013 ON NOMINALS \u1405\u14a5\u14c5\u14ef\u14ef\u141a\u1418 omin\u00f4simisisiw\u00e2wa o-min\u00f4s-im -sis -w\u00e2w-a 3-cat -DISJ-DIM-3PL -XT \u2018TheirPROX cat(s)OBV\u2019 DP 3 D PERSONP o- 3 PERS EXTP -w\u00e2w 3 EXT NP -a 3 -im- 3 N NP 3 pro 3 N -sis- min\u00f4s In these structures, what is the content added by the use of the suffix -im-? When considering the distribution of the nominal form, Mailhot (in Clarke 1982) points out that \u2013im\u2013 occurs only when the possessum can be conceived of as disjoint from the possessee.10 The suffix cannot occur when an inalienably-possessed body part (e.g. \u2013stikw\u00e2n \u2018head\u2019) is possessed by its original owner (57). (57) INALIENABLE POSSESSION = NO \u2013im\u2013 a. \u14c2\u1422\u144e\u1481\u1423 nistikw\u00e2n ni-stikw\u00e2n 1-head \u2018My head\u2019 (Volunteered S1) 10 Note that this argument, which appears to be supported by data, suggests that the referential types of nominals is more finely subdivided than commonly considered. Shoes would not be disjoint in a manner that ducks are, for example. Considering that shoes are articles of clothing, this distinction should not be surprising, but it deserves further research. 137 b. \u273d \u14c2\u1422\u144e\u1481\u14c2\u1466 \u273d nistikw\u00e2nim ni-stikw\u00e2n-im 1-head -DISJ --- (Presented S1) However, the suffix \u2013im\u2013 is obligatory when an inalienably-possessed body part is owned by someone besides its original owner (e.g. when a human owns an animal\u2019s body part), as shown in (58). (58) SECONDARILY-POSSESSED INALIENABLE = \u2013im\u2013 a. \u14c2\u1450\u1422\u144e\u1481\u14c2\u1466 nitostikw\u00e2nim nit-o-stikw\u00e2n-im 1- 3-head -DISJ \u2018The head that I have as a possession\u2019 (Presented S1) b. ?\/\u273d \u14c2\u1450\u1422\u144e\u1481\u1423 ?\/\u273d nitostikw\u00e2n nit-o-stikw\u00e2n 1-3-head --- (Presented S1) Based on evidence of this kind, Junker (2003) concludes that the nominal form of -im- marks a kind of disjoint reference (x \u2260 y). More specifically, we can say that \u2013im\u2013 marks that the argument is disjoint from some other argument (\u2018A disjunction\u2019 as in Saxon 1986 for Dogrib). The suffix \u2013im\u2013, then, is more general than simply coding \u201cObviative.\u201d 3.3.5. The predicate suffix \u2013yi\u2013 In the verbal morphosyntax, sometimes the only morpheme that distinguishes between the \u201cObviative\u201d and simple \u201cAnimate\u201d form of the verb is the suffix -yi-. For example, the verb form in (59a) \u00ea-miyosit \u2018s\/heAN is good\u2019 and the verb form in (59b) \u00ea-miyosiyit \u2018s\/heOBV is good\u2019 are only distinguished by the presence of \u2013yi\u2013 in the \u201cObviative\u201d case in (59b). 138 (59) a. \u201cANIMATE\u201d SUBJECT \u1401\u14a5\u152a\u14ef\u141f \u140a\u14c7 \u14c8\u142f\u1424 \u00ea-miyosit ana n\u00e2p\u00eaw \u00ea-miyw=si -t an -a n\u00e2p\u00eaw c1-good=AI-3 DST-AN.SG man \u2018ThatAN man is goodAN.\u2019 (Presented S2) b. \u201cOBVIATIVE\u201d SUBJECT = \u201cANIMATE\u201d SUBJECT + \u2013yi\u2013 \u1401\u14a5\u152a\u14ef\u1528\u141f \u140a\u14c2\u1426\u1403 \u14c8\u142f\u1418 \u00ea-miyosiyit anihi n\u00e2p\u00eawa \u00ea-miyw=si -yi -t an =ihi n\u00e2p\u00eaw-a c1-good=AI-DS-3 DST=XT man-XT \u2018ThisOBV manOBV is goodAN.\u2019 \u2018ThoseOBV menOBV are goodAN.\u2019 (Presented S2) On an intransitive verb, the suffix occurs immediately after the verb stem (60) and before any person marking. (60) \u201cOBVIATIVE\u201d SUBJECT OF INTRANSITIVE VERB RECEIVES \u2013yi\u2013 \u1401 \u140a\u1450\u1422\u146b\u1528\u141f \u00ea-atosk\u00eayit \u00ea- atosk\u00ea-yi -t C1-work -DS-3 \u2018\u2026s\/heOBV works\u2019 (Presented S2) When -yi- occurs on a transitive verb, it is positioned between theme sign (e.g. \u2013\u00e2\u2013 \u2018direct\u2019) and the person marking (e.g. \u2013t \u2018Animate\u2019), as shown in (61). (61) \u201cOBVIATIVE\u201d ACTING ON \u201cOBVIATIVE\u201d RECEIVES \u2013yi\u2013 \u1401 \u141a\u1438\u14ab\u1528\u141f \u00ea-w\u00e2pam\u00e2yit \u00ea- w\u00e2p=am -\u00e2 -yi-t c1-see =by.eye.TA-DIR-DS-3 \u2018\u2026s\/heOBV sees him\/herOBV\u2019 (Presented S2) This suffix also occurs as a post-stem suffix on possessed nouns, as shown in (62), where the possessed nominal ot\u00eam- \u2018his\/her horse\u2019 receives the suffix \u2013yi\u2013 when the possessor is \u201cObviative.\u201d 139 (62) \u201cOBVIATIVE\u201d POSSESSOR RECEIVES \u2013yi\u2013 \u1405\u144c\u14a5\u1528\u1418 ot\u00eamiyiwa o-t\u00eam -yi-w-a11 3-horse-DS-3-XT \u2018his\/herOBV horseOBV\u2019 (Presented S2) Summarizing, this morpheme occurs when the \u201cObviative\u201d referent is the subject of a verb that has either no object (Intransitive) or an object that is not \u201cProximate.\u201d In table 3.9, we see that \u2013 yi\u2013 is possible when an \u201cObviative\u201d is the subject of an intransitive verb, when an \u201cObviative\u201d acts on an \u201cObviative,\u201d and when an \u201cObviative\u201d acts on a speech act participant. INTRANSITIVE TRANSITIVE OBV > X OBV OBV > OBV OBV > PROX OBV > 1\/2 X > OBV \u2013yi\u2013 \u2714 \u2714 \u273d \u2714 \u273d Table 3.9. Distribution of \u2013yi\u2013 with \u201cobviative\u201d arguments This co-occurrence of -yi- and \u201cObviative\u201d referents makes it a candidate for dedicated \u201cObviative\u201d coding. 3.3.51. The traditional view: \u2013yi\u2013 marks \u201cobviative\u201d arguments Analysts have always treated \u2013yi\u2013 as coding \u201cObviative.\u201d Wolfart (1973) calls it an \u201cobviative theme.\u201d Dahlstrom (1986, 1991) continues this treatment, noting that it is restricted to \u201cObviative\u201d subjects, and all subsequent work on Plains Cree (e.g. Blain 1997, Hirose 2000, M\u00fchlbauer 2007) agrees that the suffix codes \u201cObviative.\u201d 3.3.52. The current analysis: \u2013yi\u2013 marks disjoint subjects There are reasons to believe that \u2013yi\u2013 is not dedicated to coding \u201cObviative\u201d or even \u201cAnimate\u201d interactions in Plains Cree. There are contexts where \u2013yi\u2013 occurs, but does not code \u201cObviation.\u201d However, in all these contexts, \u2013yi\u2013 codes that the subject of the predicate is disjoint (x \u2260 y) from 11 Wolfart (1973), among others, breaks this suffix up into -iyi-wa, where -wa is taken to be a variant of the \u201cObviative\u201d suffix -a. I have separated the -wa into two morphemes, based on the occurrence of forms like omaskisiniyiw \u2018his\/herOBV shoe,\u2019 where the suffix -w occurs without a following -a. The difference is not here crucial, but the reader should be aware that there are differences in treatment of this form. 140 a preceding third person (i.e. A' disjoint reference; Horseherder 1998 for Navajo) \u2013 either \u201cInanimate\u201d or \u201cObviative.\u201d While \u2013yi\u2013 occurs with \u201cObviative\u201d subjects in certain configurations, this is not the only place that this suffix is found. It also occurs with \u201cInanimate\u201d subjects in specific contexts. In particular, if the \u201cInanimate\u201d subject of a verb is dependent on some previous third person in the discourse (\u00a74.3.2), \u2013yi\u2013 is affixed to the verb. This is seen in (63b), where the verb \u00ea-kinw\u00e2yik \u2018itIN is long\u2019 is the complement of a propositional attitude verb kisk\u00eayihtam \u2018s\/heAN know itIN.\u2019 In this context, the inanimate subject of the complement verb receives the suffix \u2013yi. (63) a. \u201cINANIMATE\u201d SUBJECT WITHOUT \u2013YI\u2013 \u1401\u146d\u14ce\u1420 \u14aa\u1422\u146d\u14ef\u1423 \u00ea-kinw\u00e2k maskisin \u00ea- kinw=\u00e2-k maskisin c1-long=II -0 shoe \u2018\u2026the\/a shoe is longIN\u2019 (Presented S2) b. \u201cINANIMATE\u201d SUBJECT OF VERB IN PROPOSITIONAL ATTITUDE CONTEXT = \u2013yi\u2013 \u14c8\u142f\u1424 \u146d\u1422\u146b\u1528\u1426\u1455\u1466 \u1401\u146d\u14ce\u1528\u1420 \u14aa\u1422\u146d\u14ef\u1423. n\u00e2p\u00eaw kisk\u00eayihtam \u00ea-kinw\u00e2yik maskisin. \u00ea- kinw=\u00e2-yi -k maskisin c1-long=II -DS-0 shoe \u2018The manAN knows the shoe is longIN.\u2019 (Presented S2) In table 3.10, we see that \u2013yi\u2013 is possible when an \u201cInanimate\u201d is the dependent subject of an intransitive verb, when an \u201cInanimate\u201d acts on an \u201cObviative,\u201d and when an \u201cInanimate\u201d acts on a speech act participant. INTRANSITIVE TRANSITIVE IN > X DEP IN > X IN DEP IN IN > OBV IN > PROX IN > 1\/2 IN > OBV IN > PROX IN > 1\/2 X > IN \u2013yi\u2013 \u273d \u2714 \u273d \u273d \u273d \u2714 \u273d \u2714 \u273d Table 3.10: Distribution of \u2013yi\u2013 with \u201cinanimate\u201d arguments Comparing these generalizations to those for the \u201cObviative,\u201d there are two differences: (i) \u2013yi\u2013sometimes does not occur with \u201cInanimate\u201d subjects, but it always occurs with \u201cObviative\u201d subjects. 141 (ii) \u201cInanimate\u201d referents can never be possessors in Plains Cree (\u00a74.4.1).12 This eliminates possession constructions as places for \u2013yi\u2013 to occur with \u201cInanimate\u201d referents. Turning to the shape of the \u201cInanimate\u201d nominal in these constructions, we see a contrast with the \u201cObviative\u201d forms. The \u201cInanimate\u201d nominal connected to a verb without \u2013yi\u2013 (64a) has exactly the same form as the \u201cInanimate nominal connected to a verb with \u2013yi\u2013 (64b). (64) a. \u201cINANIMATE\u201d NOMINAL \u1401\u146d\u14ce\u1420 \u14aa\u1422\u146d\u14ef\u1423 \u00ea-kinw\u00e2k maskisin \u00ea- kinw=\u00e2-k maskisin c1-long=II-0 shoe \u2018\u2026the\/a shoe is longIN\u2019 (Presented S2) b. \u201cINANIMATE\u201d NOMINAL AS SUBJECT OF VERB MARKED WITH \u2013yi\u2013 \u1401\u146d\u14ce\u1528\u1420 \u14aa\u1422\u146d\u14ef\u1423 \u00ea-kinw\u00e2yik maskisin \u00ea- kinw=\u00e2-yi -k maskisin c1-long=II-DS-0 shoe \u2018\u2026the\/a shoe is longIN\u2019 (Presented S2) With \u201cInanimate\u201d referents, the singular\/plural contrast is maintained \u2013 a contrast not available for \u201cObviative\u201d forms in Plains Cree. In (65a), the \u201cInanimate\u201d plural coding on the verb (-i) co- occurs with the nominal suffix used for plural \u201cInanimate\u201d (\u2013a). By contrast, (65b) shows that an \u201cObviative\u201d nominal cannot co-occur with \u201cAnimate\u201d plural coding on the verb. (65) a. PLURAL \u201cINANIMATE\u201d CODING CO-OCCURS WITH \u2013yi\u2013 \u2026\u1401 \u146d\u14ce\u1528\u146d \u14aa\u1422\u146d\u14ef\u14c7. \u2026 \u00ea-kinw\u00e2yiki maskisina. \u00ea- kinw=\u00e2-yi -k-i maskisin-a c1-long=II -DS-0-PL shoe -XT \u2018\u2026the shoesIN are long.\u2019 (Presented S2) b. \u201cOBVIATIVE\u201d NOMINALS CANNOT HAVE PLURAL VERB CODING WITH \u2013yi\u2013 \u273d \u2026\u1401 \u146d\u14c4\u14ef\u1528\u148b\u1420 \u14c8\u142f\u1418. \u273d \u2026 \u00ea-kinosiyicik n\u00e2p\u00eawa. \u00ea- kinw=si-yi -t-ik n\u00e2p\u00eaw-a c1-long=AI -DS-3-PL man -XT \u2018\u2026the manOBV is tall.\u2019 (Presented S2) 12 To my knowledge, this fact has never been pointed out. In Plains Cree, possession constructions of the kind \u201cshoe\u2019s laces\u201d or \u201cmen\u2019s socks\u201d are instead compounds (masksin\u00eay\u00e2piy lit: \u201cshoe-string\u201d and n\u00e2p\u00eawasikana lit:\u201cman-socks\u201d, respectively). This is relevant to analyses of English prenominal genitives that treat these as deriving from compound-like underlying structures (e.g. McCawley 1988:385). 142 The distinction in plurality associated with (65) provides evidence that the \u201cInanimate\u201d referent is not equivalent to the \u201cObviative\u201d referent. Thus in (65a), \u2013yi\u2013 must be marking something else. In other members of the Cree language family (e.g. Moose Cree), the \u201cInanimate\u201d nominal does have additional forms added to it. Whereas \u2013yi\u2013 is only marked on the verb in Plains Cree (66a), its Moose Cree counterpart (\u2013li\u2013) is marked on both the verb and the \u201cInanimate\u201d nominal. (66) a. PLAINS CREE: -yi- OCCURS ONLY ON VERBS \u1401\u1473 \u1401 \u1405\u1426\u148b \u140b\u1455\u140d\u1528\u1455\u1541 \u1401 \u140b\u1528\u14aa\u14c2\u1528\u1420 \u140a\u1450\u1422\u146b\u140f\u1423. \u00eak\u00e2 \u00ea-ohc-\u00e2taw\u00eayihtahk \u00ea-\u00e2yimaniyik atosk\u00eawin. \u00eak\u00e2 \u00ea- ohc-\u00e2taw =\u00eayiht -am-k \u00ea-\u00e2yiman -yi -k atosk\u00eawin neg C1-rr- discount=by.mind.TI-TI -3 C1-difficult-DS-0 work \u2018[my husband] did not think anything of hard work.\u2019 [EM 28] (67) b. MOOSE CREE : -li- OCCURS ON BOTH VERBS AND NOUNS \u149e\u14d0 \u1405\u1410\u146d\u14ab\u1472\u14c7 \u146e\u14ef\u14f4\u14a5\u14d5\u1417 \u14a8\u14f1\u1410\u152e\u14ef\u14d5\u1424 cw\u00e2n ow\u00eekim\u00e2kana k\u00eesisamiliwa m\u00f4so-w\u00eey\u00e2siliw cw\u00e2n o-w\u00eekim\u00e2kan-a k\u00ees =is -am-li -w-a m\u00f4so- w\u00eey\u00e2s-li-w John 3-spouse -XT cook=by.heat.TI-TI -DS-3-XT moose-meat -DS-3 \u2018John\u2019s wife is cooking some moose meat.\u2019 (Ellis 2000:107) Crucially, these other Cree languages do not use the nominal suffix added for \u201cObviative\u201d (-a). This is what would be expected if \u2013yi\u2013 marked \u201cObviation.\u201d There are two sets of facts that challenge the possibility that an \u201cInanimate\u201d argument associated with \u2013yi\u2013 is an \u201cObviative\u201d referent: (i) When \u2013yi\u2013 occurs on a verb, the referential contrasts for \u201cInanimate\u201d arguments are different than for \u201cObviative.\u201d \u201cInanimate\u201d is be specified as either plural or singular, whereas the \u201cObviative\u201d cannot be specified for plurality. (ii) When \u2013yi\u2013 occurs on a verb, other Cree languages (e.g. Moose Cree) mark nominal arguments associated with \u2013yi\u2013 instead of \u201cObviative\u201d coding. I now consider what an account of \u2013yi\u2013 would look like that does not treat it as a dedicated \u201cObviative\u201d morpheme. First, we notice that \u2013yi\u2013 always and only occurs with subjects (Dahlstrom 1991, M\u00fchlbauer 2007), which implies that \u2013yi\u2013 is located in the IP domain, since that is thought to be 143 the locus of subjecthood in clauses (cf. Chomsky 1986). Putting this within the analysis developed by Hirose (2000) and D\u00e9chaine (2003) for Plains Cree\u2019s verb system, we get a structure as in (68). (68) CP 3 C IP 3 -yi- VP 3 ROOT 3 V VP 3 V (Adapted from D\u00e9chaine 2003) Here, the presence of the suffix \u2013yi\u2013 signals that the subject of the verb is not some previous third person referent. This means that \u2013yi\u2013 is a component of Plains Cree\u2019s switch reference system. Its content, then, is disjoint reference (x \u2260 y). This is discussed in Chapter 4 (\u00a74.3.2). The position of \u2013yi\u2013 accounts for the difference in pronominal forms in the TI conjunct paradigm. Consider the examples in (69). The form with \u2013yi\u2013 shows \u201cAnimate\u201d marking (69a), while the form without \u2013yi\u2013 shows \u201cInanimate\u201d pronominal marking (69b). (69) a. TI CONJUNCT VERB + \u2013yi\u2013 HAS \u2013t \u1401 \u14b9\u1438\u1426\u1455\u14a5\u1528\u141f \u00ea-w\u00e2pahtamiyit \u00ea-w\u00e2p=aht -am-yi-t c1-see=by.eye.TI-TI -DS-3 \u2018\u2026 (as) s\/heOBV seesIN it.\u2019 (Presented S2) b. TI CONJUNCT VERB HAS \u2013k \u1401 \u14b9\u1438\u1426\u1455\u1541 \u00ea-w\u00e2pahtahk \u00ea-w\u00e2p=aht -am-k c1-see=by.eye.TI-TI -0 \u2018\u2026 (as) s\/he seesIN it.\u2019 (Presented S2) 144 If -yi- codes IP-level subject properties, the suffix following it could be introduced as IP-level affix (i.e. marking subjects; 63a), whereas the form without -yi-, lacking IP suffixation, could have the suffix introduced in the lower, object position (63b).13 (70) a. CP 3 C IP 3 -t 3 -yi- VP 3 ROOT 3 V VP 3 V b. CP 3 C IP 3 I VP 3 ROOT 3 V VP 3 V -k 3.3.53. Modelling the \u201cobviative\u201d effects of \u2013yi\u2013 If \u2013yi\u2013 codes that subject of the predicate is disjoint from some previous third person (i.e. A' disjoint reference: Horseherder 1998 for Navajo, \u00a74.3.2 for Plains Cree), we expect a particular distribution for this morpheme. First, the suffix \u2013yi\u2013 should not occur in out-of-the-blue contexts. This is confirmed; with just a single clause in the discourse, \u2013yi\u2013 cannot occur. This is true for both \u201cAnimate\u201d (71a) and \u201cInanimate\u201d (71b) arguments. 13 I am here avoiding the issue of linearization. See D\u00e9chaine & Reinholtz (2008) for a consideration of these issues, where they argue that VPs are constructed via encliticization, which then undergoes snowball movement (cf. Aboh 2004) to the CP level. 145 (71) a. \u201cANIMATE\u201d VERB WITH \u2013yi\u2013 INFELICITOUS OUT-OF-THE-BLUE # \u1401 \u14c2\u1472\u14a7\u1528\u141f \u140a\u14c2\u1426\u1403 \u1403\u1422\u1475\u1418 # \u00ea-kinosiyit anihi iskw\u00eawa. \u00ea- kinw=si -yi-t an =ihi iskw\u00eaw-a C1-long=AI-DS-3 DST=XT woman-XT \u2018That womanOBV is tall.\u2019 (Presented S2) b. \u201cINANIMATE\u201d VERB WITH \u2013yi\u2013 INFELICITOUS OUT-OF-THE-BLUE # \u1401 \u146d\u14ce\u1528\u1420 \u140a\u14c2\u14aa \u14aa\u1422\u146d\u14ef\u1423 # \u00ea-kinw\u00e2yik anima maskisin. \u00ea-kinw =\u00e2 -yi -k an =ima maskisin c1-long=II-DS-0 DST=IN.SG shoe \u2018That shoe is long.\u2019 (Presented S2) If the morpheme \u2013yi\u2013 codes A' disjoint reference, this behaviour makes sense; coding disjointness when there is nothing to be disjoint from would be a strange thing to do. In fact, the repair strategies used by consultants support this; the -yi- forms can only be made felicitous if the consultant imagines some other referent in the discourse. Second, when two clauses are chained together, \u2013yi\u2013 should occur on the second clause if and only if its subject is different from some previously-established third person referent. This is again confirmed by the data. Consider the \u201cAnimate\u201d form in (72a) and the \u201cInanimate\u201d form in (72b). In (72a), the initial verb\u2019s subject is n\u00e2p\u00eaw \u2018man,\u2019 who is the \u201cProximate\u201d referent of a TA verb that uses a direct theme sign (-\u00e2-). The second verb \u00ea-nikamoyit \u2018s\/he sings\u2019 carries the suffix \u2013yi\u2013. In (72b), the intitial verb\u2019s subject is again a \u201cProximate\u201d n\u00e2p\u00eaw \u2018man.\u2019 The second verb, \u00ea-kimiwaniyik \u2018it is raining,\u2019 carries the suffix \u2013yi\u2013. (72) a. SECOND \u201cANIMATE\u201d VERB CARRIES \u2013yi\u2013 \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u1401 \u14c2\u1472\u14a7\u1528\u141f n\u00e2p\u00eaw \u00ea-w\u00e2pam\u00e2t iskw\u00eawa \u00ea-nikamoyit. n\u00e2p\u00eaw \u00ea- w\u00e2p=am -\u00e2 -t iskw\u00eaw-a \u00ea-nikamo-yi-t man C1-see=by.eye.TA-DIR-3 woman-XT C1-sing -DS-3 \u2018The manPROX saw the womanOBV when sheOBV was singing.\u2019 (Presented S2) b. SECOND \u201cINANIMATE\u201d VERB CARRIES \u2013yi\u2013 \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u1401 \u146d\u14a5\u1418\u14c2\u1528\u141f n\u00e2p\u00eaw \u00ea-w\u00e2pam\u00e2t iskw\u00eawa \u00ea-kimiwaniyik. n\u00e2p\u00eaw \u00ea- w\u00e2p=am -\u00e2 -t iskw\u00eaw-a \u00ea-kimiwan-yi-k man C1-see=by.eye.TA-DIR-3 woman-XT C1-rain -DS-0 \u2018The manPROX saw the womanOBV when it was raining.\u2019 (Presented S2) 146 This suffix \u2013yi\u2013 occurs, then, when the second subject is different from the first. Third, we expect that the suffix -yi- should be used for reference disambiguation when both referents are \u201cAnimate.\u201d This is also confirmed. Consider the pair in (73). In (73a), the presence of \u2013yi\u2013 on the second verb signals obligatory disjoint reference (73a) from the first verb\u2019s subject, while the absence of \u2013yi\u2013 in (73b) signals obligatory co-reference. (73) a. \u2013yi\u2013 ON SECOND VERB = DISJOINT REFERENCE \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u1473 \u14c2\u1472\u14a7\u1528\u141f. n\u00e2p\u00eaw \u00ea-w\u00e2pam\u00e2t iskw\u00eawa k\u00e2-nikamoyit. n\u00e2p\u00eaw \u00ea- w\u00e2p=am -\u00e2 -t iskw\u00eaw-a k\u00e2-nikamo-yi-t man C1-see=by.eye.TA-DIR-3 woman-XT C2-sing -DS-3 \u2018The manPROX saw the womanOBV when sheOBV was singing.\u2019 (Presented S2) \u2260 Man is singing = Woman is singing b. NO \u2013yi\u2013 ON SECOND VERB = CO-REFERENCE \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u1473 \u146d\u14a5\u1418\u14c2\u141f. n\u00e2p\u00eaw \u00ea-w\u00e2pam\u00e2t iskw\u00eawa k\u00e2-nikamot. n\u00e2p\u00eaw \u00ea- w\u00e2p=am -\u00e2 -t iskw\u00eaw-a k\u00e2-nikamo-t man C1-see=by.eye.TA-DIR-3 woman-XT C2-sing -3 \u2018The manPROX saw the womanOBV when hePROX was singing.\u2019 (Presented S2) = Man is singing \u2260 Woman is singing Here, the only difference in the two examples is the presence\/absence of the morpheme \u2013yi\u2013, and this difference correlates with a difference in the interpretation of the subject of the second clause. This suffix \u2013yi\u2013 is also used to disambiguate possessors of nominals, as shown in (74). In (74a), the presence of the suffix \u2013yi\u2013 signals the the possessor is different from the subject of the verb (n\u00e2p\u00eaw \u2018man\u2019). (74) a. \u2013yi\u2013 ON POSSESSED NOMINAL = DISJOINT REFERENCE \u14c8\u142f\u1424 \u141a\u1438\u14ab\u1424 \u1405\u144c\u14a5\u1528\u1418 n\u00e2p\u00eaw w\u00e2pam\u00eaw ot\u00eamiyiwa n\u00e2p\u00eaw w\u00e2p=am -\u00ea -w o-t\u00eam -yi-w-a man see =by.eye.TA-DIR-3 3-horse-DS-3-XT \u2018The manPROX saw his\/herOBV horseOBV.\u2019 (Presented S2) \u2260 Man\u2019s horse = Some other third person\u2019s horse 147 b. NO \u2013yi\u2013 ON POSSESSED NOMINAL = CO-REFERENCE \u14c8\u142f\u1424 \u141a\u1438\u14ab\u1424 \u1405\u144c\u14aa n\u00e2p\u00eaw w\u00e2pam\u00eaw ot\u00eama n\u00e2p\u00eaw w\u00e2p=am -\u00ea -w o-t\u00eam -a man see =by.eye.TA-DIR-3 3-horse-XT \u2018The manPROX saw his\/herOBV horseOBV.\u2019 (Presented S2) = Man\u2019s horse = Some other third person\u2019s horse There is a complication with the possessed nominals. Consider the examples in (75), which show that the forms where the nominal lacks \u2013yi\u2013 are not interpreted as obligatorily co-referent, but are instead ambiguous. (75) a. NO \u2013yi\u2013 = EITHER CO-REFERENCE OR DISJOINT REFERENCE \u14c8\u142f\u1424 \u141a\u1438\u14ab\u1424 \u1405\u144c\u14aa n\u00e2p\u00eaw w\u00e2pam\u00eaw ot\u00eama n\u00e2p\u00eaw w\u00e2p=am -\u00ea -w o-t\u00eam -a man see =by.eye.TA-DIR-3 3-horse-XT \u2018The manPROX saw his\/her? horseOBV.\u2019 (Presented S2) = Man\u2019s horse = Some other third person\u2019s horse b. NO \u2013yi\u2013 = EITHER CO-REFERENCE OR DISJOINT REFERENCE \u14c8\u142f\u1424 \u141a\u1438\u1426\u1455\u1466 \u1405\u14aa\u1422\u146d\u14ef\u1423 n\u00e2p\u00eaw w\u00e2pahtam omaskisin n\u00e2p\u00eaw w\u00e2p=aht -am o-maskisin man see =by.eye.TI-TI 3-shoe \u2018The manPROX saw his\/her? shoeIN.\u2019 (Presented S2) = Man\u2019s shoe = Some other third person\u2019s shoe The pattern of -yi- can best be understood if we posit it to code disjoint reference. As a disjoint reference marker, it is infelicitous with single-clause discourses, as we saw in (71), but it is used when its argument is not some previous argument (here, restricted to subjects), as we saw in (72). In this usage as a disjoint reference marker, it is used in the disambiguation of subsequent subjects, as we saw in (73). From this, I conclude that the content of -yi- is best characterized as carrying disjoint reference across subjects, either verbal or nominal (possessors). 148 3.3.6. Summary: The derivative nature of obviation The results of this study are shown in the table below. Occurrence Forms \u201cObviative\u201d Other Atomic Construction \u00a73.2.1. [PREDICATE]-a YES YES \u2013 INAN \u273d \u2714 \u00a73.2.2. [DEM]=ihi YES YES \u2013 INAN \u273d \u2714 \u00a73.2.3. [Verb]-\u00ea YES YES \u2013 INAN \u273d \u2714 \u00a73.2.4. [PREDICATE]-im YES YES \u2013 ANIM \u273d \u2714 \u00a73.2.5. [PREDICATE]-yi YES YES \u2013 INAN \u273d \u2714 Table 3.11. \u201cObviative\u201d forms and their patterning From this table, it is clear that Plains Cree systematically lacks any dedicated forms that code obviation. Instead, all of the forms that code \u201cObviative\u201d also can be used in other constructions that do not code \u201cObviative\u201d reference. \u201cObviative,\u201d then, is a derived class, and is constructed from several different components of the Plains Cree grammar. The logic of this construction is considered in detail in Chapter 4. I now turn from the form of \u201cObviation\u201d to its semantic content. 3.4. The content of obviation: contextual extentionality We have just seen that the \u201cObviative\u201d forms are systematically identical to \u201cInanimate\u201d forms, but range over \u201cAnimate\u201d contexts. For example, the suffix \u2013a is used to code obviation on nominals in \u201cAnimate\u201d contexts (76a), but also appears in \u201cInanimate\u201d contexts to code number (76b). (76) a. \u201cOBVIATIVE\u201d NOMINAL WITH \u2013a \u1401 \u14a5\u152a\u14ef\u1528\u141f \u140a\u14c2\u1426\u1403 \u14c8\u142f\u1418 \u00ea-miyosiyit anihi n\u00e2p\u00eawa \u00ea-miyw=si -yi -t an =ihi n\u00e2p\u00eaw-a C1-good=AI-DS-3 DST=XT man -XT \u2018\u2026(as) thisOBV manOBV is goodAN.\u2019 \u2018\u2026(as) thoseOBV menOBV are goodAN.\u2019 (Presented S2) 149 b. \u201cINANIMATE\u201d NOMINAL WITH \u2013a \u1401 \u14a5\u1429\u141a\u14ef\u146d \u140a\u14c2\u1426\u1403 \u14aa\u1422\u146d\u14ef\u14c7 \u00ea-miyw\u00e2siki anihi maskisina \u00ea- miyw=\u00e2si-k -i an =ihi maskisin-a C1-good=II -0-PL DST=XT shoe -XT \u2018\u2026(as) thoseIN shoes are niceIN.\u2019 (Presented S2) The demonstrative sets used for both \u201cInanimate\u201d and \u201cObviative\u201d are the same, as shown in (77), where the demonstrative anihi \u2018that one\u2019 can either modify an \u201cObviative\u201d (77a), or an \u201cInanimate\u201d (77b). (77) a. \u201cOBVIATIVE\u201d DEMONSTRATIVE WITH \u2013ihi \u1401 \u14a5\u152a\u14ef\u1528\u141f \u140a\u14c2\u1426\u1403 \u14c8\u142f\u1418 \u00ea-miyosiyit anihi n\u00e2p\u00eawa \u00ea-miyw=si -yi -t an =ihi n\u00e2p\u00eaw-a C1-good=AI-DS-3 DST=XT man -XT \u2018\u2026(as) thisOBV manOBV is goodAN.\u2019 \u2018\u2026(as) thoseOBV menOBV are goodAN.\u2019 (Presented S2) b. \u201cINANIMATE\u201d DEMONSTRATIVE WITH \u2013ihi \u1401 \u14a5\u1429\u141a\u14ef\u146d \u140a\u14c2\u1426\u1403 \u14aa\u1422\u146d\u14ef\u14c7 \u00ea-miyw\u00e2siki anihi maskisina \u00ea- miyw=\u00e2si-k -i an =ihi maskisin-a C1-good=II -0-PL DST=XT shoe -XT \u2018\u2026(as) thoseIN shoes are niceIN.\u2019 (Presented S2) Both the \u201cObviative\u201d and \u201cInanimate\u201d occur with the suffix \u2013yi\u2013. For example, in (78a), the \u201cObviative\u201d referring nominal iskw\u00eawa \u2018womanOBV\u2019 is the argument of \u00ea-kinosiyit \u2018sheOBV is tall,\u2019 which bears \u2013yi\u2013, while, in (78b), the \u201cInanimate\u201d verb \u00ea-kinw\u00e2yik \u2018itIN is long\u2019 also bears the suffix \u2013yi\u2013. (78) a. \u201cOBVIATIVE\u201d CO-OCCURS WITH \u2013yi\u2013 \u140a\u141a\u14ef\u1422 \u1403\u144c\u1528\u1426\u1455\u1466 \u1401 \u14c2\u1472\u14a7\u1528\u141f \u140a\u14c2\u1426\u1403 \u1403\u1422\u1475\u1418 aw\u00e2sis it\u00eayihtam \u00ea-kinosiyit anihi iskw\u00eawa. aw\u00e2sis it=\u00eayiht -am \u00ea- kinw=si -yi-t an =ihi iskw\u00eaw-a child RR=by.mind.ti-TI C1-long=AI-DS-3 DST=XT woman-XT \u2018The childPROX thinks that that womanOBV is tall.\u2019 (Presented S2) 150 b. \u201cINANIMATE\u201d CO-OCCURS WITH \u2013yi\u2013 \u140a\u141a\u14ef\u1422 \u1403\u144c\u1528\u1426\u1455\u1466 \u1401 \u146d\u14ce\u1528\u1420 \u140a\u14c2\u14aa \u14aa\u1422\u146d\u14ef\u1423 aw\u00e2sis it\u00eayihtam \u00ea-kinw\u00e2yik anima maskisin. aw\u00e2sis it=\u00eayiht -am \u00ea-kinw =\u00e2 -yi -k an=ima maskisin child RR=by.mind.ti-TI c1-long=II -DS-0 DST=IN.SG shoe \u2018The childAN thinks that that shoeIN is long. (Presented S2) When the \u201cObviative\u201d or \u201cInanimate\u201d class of referents is the object of a verb, the theme sign \u2013\u00ea\u2013 is used. In (79a), we see the verb w\u00e2pam\u00eaw \u2018s\/hePROX sees him\/herOBV,\u2019 which has an \u201cObviative\u201d object, and uses the theme sign \u2013\u00ea\u2013. In (79b), the verb niw\u00e2paht\u00ean \u2018I see itIN\u2019 has an \u201cInanimate\u201d object, and uses the same theme sign \u2013\u00ea\u2013. (79) a. \u201cPROXIMATE\u201d ACTS ON \u201cOBVIATIVE\u201d = \u2013\u00ea\u2013 \u14c2\u141a\u1438\u14a3\u1424. w\u00e2pam\u00eaw. w\u00e2p=am -\u00ea -w see =by.eye.TA-DIR-LP \u2018S\/hePROX sees him\/herOBV.\u2019 (Presented S2) b. SAP ACTS ON \u201cINANIMATE\u201d= \u2013\u00ea\u2013 \u14c2\u141a\u1438\u1426\u144c\u1423. niw\u00e2paht\u00ean. ni-w\u00e2p=aht -\u00ea -n 1- see =by.eye.TI-TI-LP \u2018I see itIN.\u2019 (Presented S2) Both the \u201cObviative\u201d and the \u201cInanimate\u201d trigger the inverse theme sign \u2013ikw\u2013 when they act on other referents. For example, the form in (80a) shows an \u201cObviative\u201d referent aw\u00e2sisa \u2018childIN\u2019 acting on a \u201cProximate\u201d referent nim\u00e2ma \u2018my mother, which requires the use of the theme sign \u2013 ikw\u2013 on the verb \u00ea-s\u00eakihikot \u2018s\/heOBV scared him\/herPROX.\u2019 In (80b), an identical verb form, \u00ea- s\u00eakihikot, is used when an \u201cInanimate\u201d referent (mohkom\u00e2n \u2018knife\u2019) acts on an \u201cAnimate\u201d referent (nim\u00e2ma \u2018my mother\u2019). 151 (80) a. \u201cOBVIATIVE\u201d ACTS ON \u201cPROXIMATE\u201d = \u2013ikw\u2013 \u14c2\u14ab\u14ab \u1401 \u14ed\u146d\u1426\u1403\u146f\u141f \u140a\u141a\u14ef\u14f4. nim\u00e2ma \u00ea-s\u00eakihikot aw\u00e2sisa. ni-m\u00e2ma \u00ea- s\u00eak =ih -ikw-t aw\u00e2sis-a 1-mother C1-scare=by.neut-INV-3 child -XT \u2018The childOBV scared my motherPROX.\u2019 (Presented S2) a. \u201cINANIMATE\u201d ACTS ON \u201cANIMATE\u201d = \u2013ikw\u2013 \u14c2\u14ab\u14ab \u1401 \u14ed\u146d\u1426\u1403\u146f\u141f \u140a\u14c2\u14aa \u14a7\u1426\u146f\u14ab\u1423. nim\u00e2ma \u00ea-s\u00eakihikot anima mohkom\u00e2n. ni-m\u00e2ma \u00ea- s\u00eak =ih -ikw-t an=ima mohkom\u00e2n 1-mother C1-scare=by.neut-INV-3 DST=IN.SG knife \u2018The knifeIN scared my motherAN.\u2019 (Presented S2) This formal parallelism between \u201cInanimate\u201d and \u201cObviative\u201d led Wolfart (1973) to suggest that there may be some underlying content that the two referential categories share: If we rule out accident as the cause of the identity of the animate obviative and the inanimate plural, we have to look for the semantic feature of Cree which these categories have in common. (Wolfart 1973:14) Formal identity, then, points to content identity, but, until now, the thread of meaning that unites these forms has not been found. A hint at what the shared content of \u201cInanimate\u201d and \u201cObviative\u201d might be is suggested by data like (81). Here, sentence describes an event in which a man is seeing an \u201cObviative\u201d woman. The nominal iskw\u00eaw \u2018woman\u2019 bears the suffix \u2013a, and the verb bears TA morphology. When asked about what this sentence could mean, one speaker (S3) responded, \u201cIn this example, the woman is not aware that she\u2019s being seen.\u201d (81) \u201cOBVIATIVE\u201d REFERENT IS NOT AWARE OF EVENT \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u00ea-w\u00e2pam\u00e2t n\u00e2p\u00eaw iskw\u00eawa \u00ea-w\u00e2p=am -\u00e2 -t n\u00e2p\u00eaw iskw\u00eaw-a c1-see=by.eye.TA-DIR-3 man woman-XT \u2018The manPROX sees the womanOBV.\u2019 (Presented S3) COMMENT (S3): \u201cIn this example, the woman is not aware that she\u2019s being seen.\u201d By contrast, a simple \u201cAnimate\u201d form evokes no psychological discussion at all, as shown in (82). Here, the nominal aw\u00e2sis \u2018child\u2019 is not marked with any suffixation, and the verb carries 152 TA marking; the referent is \u201cAnimate\u201d. When asked about the awareness of the child for this event, speakers are non-committal; the child could be aware of the event, or he could not. (82) \u201cANIMATE\u201d REFERENT IS NOT SPECIFIED FOR AWARENESS \u14c2\u141a\u1438\u14ab\u1424 \u140a\u14c7 \u140a\u141a\u14ef\u1422 niw\u00e2pam\u00e2w ana aw\u00e2sis ni-w\u00e2pam -\u00e2 -w an -a aw\u00e2sis 1- see=by.eye.TA-DIR-3 DST-AN.SG child \u2018I seeAN thatAN child.\u2019 (Presented S2) Comment (S2): \u201cI don\u2019t know if the child knows about this or not.\u201d Interpretations like those for (81) bear a striking resemblance to the readings for \u201cInanimate\u201d nominals that we considered in Chapter 2. There, we saw that the \u201cInanimate\u201d forms were associated with a referent that was never aware of any event. (83) \u201cINANIMATE\u201d REFERENT IS NEVER AWARE \u14c2\u146e\u141a\u1438\u1426\u144c\u1423 \u1406\u146d\u14c0\u1431\u146f nik\u00ee-w\u00e2paht\u00ean \u00f4ma kin\u00eapik ni-k\u00ee-w\u00e2p=aht -\u00ea-n aw-ima kin\u00eapikw 1-PREV-see=by.eye.TI-TI-LP PROX-IN.SG snake \u2018I sawIN thisIN (plastic) snake\u2019 (Presented S2) Comment (S2): \u201cIt\u2019s just a plastic snake. It doesn\u2019t think anything.\u201d Here, the speaker is considering a rubber representation of a snake, rather than a real one. The nominal is associated with an overt demonstrative \u00f4ma \u2018this,\u2019 which codes \u201cInanimate\u201d referents, and the verb is in the TI form. As far as the utterer of this sentence is concerned, this referent cannot have awareness in any context. This, of course, could not be said of the child in example (81); she has, no doubt, had many intentions in her life, but has none associated with this particular context. What we see is that an \u201cInanimate\u201d referent can never have a intention (i.e. is never \u2018aware\u2019), while an \u201cObviative\u201d referent doesn\u2019t have an intention in the context of the event described. The difference between \u201cInanimate\u201d and \u201cObviative\u201d, then, appears to be a difference between inherent lack of awareness (\u201cInanimate\u201d), and contextual lack of awareness (\u201cObviative\u201d). \u201cObviative\u201d is a contextual version of \u201cInanimate,\u201d which means that \u201cObviative\u201d is contextual extentionality. Extending the model constructed for animacy (Chapter 2), I define a contextually extentional referent as one that cannot be paired with a perspective in a given context. Following 153 Partee (1988), I represent the discourse context of an utterance using the variable C. Contextual extentionality, then, is the property of being extentional at context C (84a). Inherent extentionality, by contrast, is the context-independent property of being extentional (84b). (84) a. CONTEXTUAL EXTENTIONALITY EXT(x,C) \u2194 \u2200\u03c8\u2200y(R(y,\u03c8,C) \uf0e0 x \u2260 y) x is extentional at context C if and only if for all perspectives \u03c8 and all individuals y, if there is a relation R of y with \u03c8 at context C, then x is not y. b. INHERENT EXTENTIONALITY EXT(x) \u2194 \u2200\u03c8\u2200y(R(y,\u03c8) \uf0e0 x \u2260 y) x is Extentional if and only if for all Perspectives \u03c8 and all individuals y, if there is a relation R between individual y with perspective \u03c8, then x is not y. Applying this to the \u201cObviative\u201d form in Plains Cree, I claim that a referent classified as \u201cObviative\u201d will have the content of (85). (85) \u201cOBVIATIVE\u201d FORM = [\u03bbx \u00b7 EXT(x,C) \u22c0 PRED'(x)] = Give some x, such that x is extentional at C and x is a member of the set of referents denoted by the predicate. This differs from the \u201cInanimate\u201d form, which is analyzed as inherently extentional, as in (86). (86) \u201cINANIMATE\u201d FORM = \u03bbx \u00b7 [EXT(x) \u22c0 PRED'(x)] Give some x, such that x is extentional and x is a member of the set of referents denoted by the predicate. When the \u201cObviative\u201d is set up in contrast to an unspecified \u201cAnimate\u201d referent, the \u201cAnimate\u201d receives the opposite extentional value; it is contextually non-extentional (i.e. intentional). (87) a. \u201cAnimate\u201d : [\u00d8] b. \u201cObviative\u201d : [EXT] at C c. \u201cProximate Animate\u201d : [-EXT] at C The formal identity between \u201cObviative\u201d and \u201cInanimate,\u201d then, paralleles the two classes\u2019 referential identity. This model of obviation in Plains Cree makes specific claims about what the distribution of \u201cObviative\u201d constructions must be: 154 (i) The model requires that extentional \u201cObviative\u201d referents should not be able to occur with intentional verbs (psych verbs; Bouchard 2005, etc.). An examination of the distribution of \u201cObviative\u201d referents with these verbs supports this (\u00a73.4.1). (ii) The model requires that an \u201cObviative\u201d referent should either have spatial-temporal or intentional deficits in the context, making them \u201cunaware\u201d of that context. An examination of the distribution of \u201cObviative\u201d referents when there are spatio-temporal or intentional deficits supports this (\u00a73.4.2). (iii) Because Plains Cree has coding only for extentional referents (\u201cInanimate\u201d and \u201cObviative\u201d), intentional referents will be a by-product of contrast with an extentional referent. This is confirmed by a consideration of the \u201cProximate\u201d referential class (\u00a73.4.3). All evidence supports the conclusion that the \u201cObviative\u201d is used when a referent has no contextual intention. This is the contextual parallel to the inherent extentional specification of \u201cInanimate\u201d referents; \u201cInanimate\u201d forms are used when a referent has never had a mind (see Chapter 2), while \u201cObviative\u201d constructions are used when the referent has no contextual intention. 3.4.1. Intentional verbs restrict \u201cobviative\u201d reference Verbs that are inherently intentional (i.e. introduce a perspective) can occur with an \u201cAnimate\u201d argument, but not an \u201cInanimate\u201d one (88). This is because \u201cInanimate\u201d referents are inherently extentional, and, as such, can never be intentional. (88) a. INTENTIONAL VERB WITH \u201cANIMATE\u201d AGREEMENT \u1401 \u14a5\u152a\u14ab\u148b\u1426\u1405\u141f \u140a\u1418 \u1403\u1422\u1475\u1424 \u00ea-miyom\u00e2cihot awa iskw\u00eaw \u00ea- miyom\u00e2ciho-t aw =a iskw\u00eaw C1-feel.well -3 PRX=AN.SG woman \u2018This woman feels wellANIM.\u2019 (Presented to S2) b. INTENTIONAL VERB CANNOT HAVE \u201cINANIMATE\u201d AGREEMENT \u1401 \u14a5\u152a\u14ab\u148b\u1426\u1405\u1420 \u1406\u14aa \u14aa\u1422\u146d\u14ef\u1422 \u273d \u00ea-miyom\u00e2cihok \u00f4ma maskisin \u00ea-miyom\u00e2ciho-k aw =ima maskisin c1-feel.well -0 PRX=IN.SG shoe Intended: \u2018This shoe feels wellINAN.\u2019 (Presented to S2) 155 If \u201cObviatives\u201d are the contextual equivalent to \u201cInanimate\u201d nominals, they should show a similar kind of restriction. However, unlike \u201cInanimate\u201d nominals, this restriction should be conditioned by the intentional properties of the referent within the context, not conditioned by any intentional properties inherent in the referent. Consider verbs that introduce intentional content: psych verbs (\u00a73.4.11) and speaking verbs (\u00a73.4.12). Denoting intentional events (cf. Banfield 1982, Bouchard 2005), verbs of this kind introduce a context that is not consonant with the meaning of obviation\u2019s eventive extentionality: the verb\u2019s experiencer is intentional (i.e. they have a perspective and there is a proposition evaluated as true\/false within it), but the referent, lacking a perspective (being extentional), can not be intentional. In fact, this is exactly what we find; \u201cObviative\u201d constructions cannot freely occur with intentional predicates. 3.4.11. \u201cObviative\u201d referents are restricted with psych verbs As we saw in Chapter 2, a Speaker is able to use a psych verb to ascribe an intention to some referent. In Plains Cree, some psych verbs are lexically-specified, employing a root that holds inherent intentional content (89). Here, the root kis\u00ee- carries the semantics of \u2018angriness\u2019 in (89a), and the stem p\u00f4m\u00ea- carries the meaning of disappointedness or discouragement in (89b). (89) a. ROOT kis\u00ee- = \u2018angriness\u2019 \u2026, \u14c7\u14a8\u152d \u1411\u1426\u1473\u1428 \u140a\u152d \u14c5\u1426 \u146d\u14f0\u1422\u1450\u141a\u1424 \u14c2\u1473\u140f\u1429, \u2026 \u2026, nam\u00f4y w\u00eehk\u00e2c aya n\u00f4h-kis\u00eestaw\u00e2w nik\u00e2wiy, \u2026 nam\u00f4ya w\u00eehk\u00e2c aya ni-\u00f4h-kis\u00ee=staw -\u00e2 -w ni-k\u00e2wiy NEG ever CONN 1- RR-angry=APPLIC-DIR-3 1-mother \u2018\u2026, I never stayed angry at my mother \u2026\u2019 (EM 1997:\u00a719) b. ROOT p\u00f4m\u00ea- = \u2018disappointedness\u2019 \u1401 \u146e \u14ab\u14c7 \u14a5\u1450\u14c2 \u1434\u14a3\u148b\u1420, \u2026 \u00ea-k\u00ee-m\u00e2na-mitoni-p\u00f4m\u00eacik, \u2026 \u00ea- k\u00ee- m\u00e2na-mitoni-p\u00f4m\u00ea -t-ik C1-PREV-usual- very- be.disappointed-3-PL \u2018They used to be very disappointed \u2026\u2019 (EM 1997: \u00a769) Other psych verbs are built using one of Plains Cree\u2019s \u2018finals\u2019 or \u2018manner suffixes\u2019 (cf. Wolfart 1973, Hirose 2000), in particular -\u00eayim\/-\u00eayiht\/-\u00eayimo \u2018by mind.\u201914 In (90), the final \u2013\u00eayiht in 14 Wolfart (1973), among others, treats this as two morphemes, -\u00eayi- plus -m\/-ht. The differences are not here relevant. However, the division into two morphemes may help explain \u2013\u00eayimo-, which could then be taken to have the common AI final \u2013mo\u2013 in it. 156 (90a) and \u2013\u00eayimo in (90b) contribute a psychological meaning to the roots, building a verb that conveys a mental property of its subject. (90) a. FINAL \u2013\u00eayiht = \u2018MENTAL STATE\u2019 \u1401 \u1418\u14c0\u1528\u1426\u1455\u1426\u146d\u1420 \u140a\u14c2\u14aa \u140a\u152d, \u2026. \u00ea-wan\u00eayihtahkik anim \u00e2ya, \u2026 \u00ea- wan=\u00eayiht -am-k-ik an- ima aya c1-lose=by.mind.TI-TI -0-PL DST-IN.SG CONN \u2018their minds are blurred \u2026\u2019 (EM 1997: \u00a736) b. FINAL \u2013\u00eayimo = \u2018MENTAL STATE\u2019 \u1470\u1426 \u146e \u144c\u142f\u1528\u14a7\u141f \u1472 \u146d\u1426\u148b \u1411\u146d\u14a5\u141f; \u2026 \u2026, k-\u00f4h-k\u00ee-t\u00eap\u00eayimot ka-kihci-w\u00eekimit; \u2026 k\u00e2-oh-k\u00ee- t\u00eap =\u00eayimo -t ka- kihci-w\u00eeki -m -it c2-rr -PREV-enough=by.mind.AI-3 fut-great-reside-DISJ-3>1 \u2018\u2026, since he had been willing to marrying me; \u2026\u2019 (EM 1997: \u00a742) Additional elements, called \u2018preverbs,\u2019 can also be affixed to the stem to convey these kinds of meanings. Here, the preverb n\u00f4ht\u00ea- adds the meaning that the verb\u2019s agent desires the event to happen in (91a), and the preverb w\u00ee- in (91b) adds the meaning that the speaker intends for the event to happen. (91) a. PREVERB n\u00f4ht\u00ea- = \u2018WANT\u2019 \u146f\u14c2\u1455 \u1401\u147f \u1401\u146f\u1455 \u1401 \u14ab\u1450\u152e\u1423 \u1401\u147f \u1401 \u14c5\u1426\u144c \u146e\u140d\u152e\u1423, \u2026 konit \u00eakwa \u00eakota \u00ea-m\u00e2toy\u00e2n \u00eakwa \u00ea-n\u00f4ht\u00ea-k\u00eew\u00eey\u00e2n, ... konita \u00eakwa \u00eakota \u00ea- m\u00e2to-y\u00e2n \u00eakwa \u00ea- n\u00f4ht\u00ea-k\u00eew\u00ea -y\u00e2n just then there C1-cry -1 and C1-want- go.home-1 \u2018And there I was, I just cried and wanted to go home, ...\u2019 (AA 2000:\u00a72.2) b. PREVERB w\u00ee- = \u2018INTEND\u2019 \u14ab\u1472 \u14c2\u14c3\u146d\u1426\u147f\u1420 \u1401 \u1411 \u14c7\u14c7\u1426\u1403\u1426\u1455\u1418\u1420 \u1401 \u14f0\u1426\u146d\u14a5\u148b\u1420, \u2026 m\u00e2ka nin\u00eekihikwak \u00ea-w\u00ee-nanahihtawakik \u00ea-s\u00eehkimicik, \u2026 m\u00e2ka ni-n\u00eekihikw-ak \u00ea-w\u00ee- nanah=ihtaw-ak-ik but 1- parent -pl c1-intend-respect=by.ear.ta-1>3-pl \u00ea-s\u00eehk =im -it -ik c1-advise=by.mouth.ta-3>1-pl \u2018But I was going to obey my parents when they urged me, \u2026\u2019 (EM 1997: \u00a73) As this survey shows, Plains Cree has several different formal strategies in its verbal system for constructing psychological predicates. 157 When obviation is tracked with respect to these forms in two texts (Ahenakew 2000, Minde 1997),15 the following generalization emerges: the experiencer of an intentional state is almost always non-\u201cObviative.\u201d For example, this holds for psych verbs constructed from the finals \u2013\u00eayim\/\u2013\u00eayiht\/\u2013\u00eayimo, as shown in Table 3.12. \u00eayim- \u00eayiht- \u00eayimo- ANIM OBV ANIM OBV ANIM OBV Minde 17 3 25 1 4 1 Ahenakew 3 0 21 0 0 0 TOTALS 20 3 46 1 4 1 Table 3.12. Summary for psych verbs built with \u2013\u00eayim\/\u2013\u00eayiht\/\u2013\u00eayimo \u201cObviative\u201d referents are also restricted with verbs that carry intentional content in other ways (including: kost- \u2018fear,\u2019 miyaw\u00e2t- \u2018enjoy it,\u2019 kisiw\u00e2si- \u2018be angry,\u2019 kitim\u00e2kin\u00e2so- \u2018look on onself with pity,\u2019 nisitoht(aw)- \u2018understand it\/him\/her,\u2019 yaw\u00easi- \u2018be full of anger,\u2019 n\u00eap\u00eawisi- \u2018be ashamed,\u2019 kis\u00eestaw- \u2018be angry with him\/her,\u2019 ak\u00e2w\u00e2t- \u2018desire\/lust for it\/him\/her,\u2019 p\u00f4m\u00ea- \u2018be discouraged,\u2019 pakw\u00e2t- \u2018hate it\/him\/her,\u2019 atamih- \u2018make him\/her grateful,\u2019 sas\u00eehciwih- \u2018make him\/her ashamed,\u2019 k\u00eehkih- \u2018bother him\/her,\u2019 kisiw\u00e2h- \u2018make him\/her angry,\u2019 m\u00e2koh- \u2018trouble him\/her,\u2019 and paciyaw\u00eah- \u2018anger him\/her with speech\u2019), as shown in Table 3.13. Other ANIM PSYCH OBV PSYCH Minde 24 6 Ahenakew 12 2 TOTALS 36 8 Table 3.13. Summary of other psych verb forms Verbs modified by the preverbs w\u00ee- \u2018intend to\u2019 and n\u00f4ht\u00ea- \u2018want to\u2019 also are restricted in their occurrence with \u201cObviative\u201d referents, as shown in Table 3.14. n\u00f4ht\u00ea- w\u00ee- ANIM OBV ANIM OBV Minde 12 2 13 1 Ahenakew 7 0 22 1 TOTALS 19 2 35 2 Table 3.14. Summary of preverb forms Altogether, the total distribution of \u201cObviative\u201d verbs with psych verbs shows that the experiencier of an intentional state is almost never \u201cObviative.\u201d 15 The counting of \u2018words\u2019 is a problematic notion in Plains Cree (as elsewhere). The totals are meant to give the reader a rough idea of corpus size. 158 ANIM Psych OBV Psych WORD COUNT Minde 95 13 12,900 Ahenakew 65 3 10,300 Whitecalf 67 8 3,800 Totals 227 24 27,000 Table 3.15. A Summary of psych verbs in 3 texts This data confirms that there is a restriction on the co-occurrence of \u201cObviative\u201d referents with psych verbs in Plains Cree \u2013 something that is expected if \u201cObviative\u201d is a contextual form of \u201cInanimate\u201d. Notice that there are cases of \u201cObviative\u201d arguments for psych verbs \u2013 a total of 24 cases in the three texts. While these cases may appear to contradict the generalizations of this section, they in fact do not. However, understanding these cases requires a close consideration of the context of obviation, and I therefore defer a discussion of them until section 3.5. There, it will be shown that \u201cObviative\u201d referents can only be connected with psych verbs if the Speaker does not take responsibility for the assignment of this intentional property. 3.4.12. \u201cObviative\u201d referents are restricted with speaking verbs Verbs that refer to acts of speaking show properties similar to verbs of thinking and feeling. This has been documented for numerous languages, including English (Ross 1970), German (Schulz & Griesbach 1965), Mandarin (Li & Thompson 1981, Huang 1981), Romance (Hahn 1952, Farkas 1992), and Greek (Goodwin 1875), and relates to the classic Latin categorization of complementation in terms of verba sentiendi and dicendi \u2018verbs of thinking and speaking\u2019 (Hahn 1952, etc.). Banfield (1982) explains: \u201cWe can consider communication verbs as a subset of consciousness verbs \u2013 communication implies consciousness of what is being communicated.\u201d In both the case of psych verbs and speaking verbs, the agent performing the action must have an intention, else the action cannot be performed; Jane cannot think if she has no thoughts, cannot feel if she has no feelings, and neither can she be said to speak if she has no intentions.16 Within 16 It is possible for a speaker to act as a medium for someone else\u2019s message (e.g. a news anchor reading a teleprompter). If the audience understands this speech to not represent the speaker\u2019s intention (an understanding rarely, if ever, achieved), then the intention relevant is that of the writer, not the speaker. This becomes relevant when we consider examples like \u00ea-itw\u00eamakahk in Plains Cree (lit: \u2018it saidin\u2019), which inflects for an \u201cInanimate\u201d actor (-makan-k). These forms are used when a referent either (i) chatters mindlessly (as in the rolling head in Chapter 3), or (ii) is simply a vehicle for someone else\u2019s message (e.g. a pipestem or a tape recorder). In these cases, the referent performing the speech action has no intention, and is instead mechanically channeling the content of the speech from some other entity. 159 the current model, then, we can say that a Speaker is able to use a speaking verb to ascribe an intention to a referent. This makes speaking verbs a relevant area of investigation with respect to obviation, because \u201cObviative\u201d is here taken to be contextual extentionality (i.e. the contextual lack of an intention). Plains Cree has several kinds of verbs that convey speech acts. Referents may speak to every one present or no one in particular. Either of these speech contexts is coded by the stem itw\u00ea- \u2018say so\u2019 (cf. M\u00fchlbauer 2007). This is the form most often used to introduce a direct quote, and the form itself appears to be recruited as a kind of quotative morpheme (Rhodes 2002, Blain et al. 2006, Blain & D\u00e9chaine 2007). (92) QUOTATIVE USE OF itw\u00ea- a. \u201c\u2026\u201d \u1401\u146f\u14ef \u1403\u14ef \u1403\u1458\u1424 \u140a\u14c7 \u1438\u141a\u1472\u1423. \u201c\u2026\u201d \u00eakos \u00ees \u00eetw\u00eaw ana paw\u00e2kan. \u00eakosi isi it =w\u00ea-w an-a paw\u00e2kan so thus thus=AI -3 dst-AN.SG dream.spirit \u2018\u201c\u2026\u201d That is what that dream spirit said.\u2019 (Volunteered S4) b. \u201c\u142f\u152d\u1481\u1424 \u14a6\u14c7,\u201d \u1403\u140d\u1424, \u201c\u140a\u152d, \u1401 \u1434\u14ef\u1426\u1403\u1472\u140f\u152e\u1423 Debden \u2026\u201d \u201cp\u00eayakw\u00e2w m\u00een,\u201d itw\u00eaw, \u201caya, \u00ea-p\u00f4sihikawiy\u00e2n Debden \u2026\u201d p\u00eayakw\u00e2w m\u00eena itw\u00ea-w aya, \u00ea- p\u00f4s=ih -ikawi-y\u00e2n Debden once also say-3 CONN C1-ride=CAUS-IMP -1 Debden \u2018\u201cOnce also,\u201d he said, \u201cI was given a ride to Debden, \u2026\u201d\u2019 (AA 2000:\u00a78.3) When we consider verbs of speaking in Plains Cree, what we see is that \u201cObviative\u201d referents almost never speak. TEXT \u201cANIMATE\u201d \u201cOBVIATIVE\u201d WORD COUNT Minde 29 11 12,900 Ahenakew 155 9 10,300 Whitecalf 80 0 3,800 K\u00e2-pimw\u00eaw\u00eahahk 63 1 7,400 TOTALS 327 21 34,400 Table 3.16. Speaking verbs and obviation Focussing on the intransitive verb itw\u00ea- \u2018say so\u2019 shows only one instance of an \u201cObviative\u201d speaker across all texts considered (this one example, as well as the transitive cases, is discussed in section 3.5 below). 160 itw\u00ea- \u2018say\u2019 \u201cANIMATE\u201d \u201cOBVIATIVE\u201d WORD COUNT Minde 17 0 12,900 Ahenakew 148 0 10,300 Whitecalf 80 0 3,800 K\u00e2-pimw\u00eaw\u00eahahk 63 1 7,400 TOTALS 308 1 34,400 Table 3.17. Intransitive itw\u00ea- and obviation The \u201cObviative\u201d construction, then, is highly restricted in its distribution with intransitive verbs of speaking. This ban on \u201cObviative\u201ds occuring with intransitive verbs of speaking is strong enough that it will often cause a previously \u201cObviative\u201d referent to be shifted to \u201cAnimate.\u201d This can create a rapid-fire succession of shifts in the coding for a referent.17 For example, in a story by Alice Ahenakew, she and her husband Andrew go to visit her sister-in-law and her brother-in- law. In these events, the sister-in-law is \u201cAnimate\u201d and the brother-in-law is \u201cObviative\u201d when she is bad-tempered, gives dirty looks, and talks to Andrew Ahenakew about the brother-in-law (lines vi-x). When the brother-in-law begins telling his story, he is shifted from his status as \u201cObviative\u201d to become a simple \u201cAnimate\u201d, intransitive speaker. LINE EVENT ANIM OBV vi Sister-in-law is bad-tempered Sister-in-law - vii Sister-in-law gives Brother-in-law dirty looks Sister-in-law Brother-in-law viii Brother-in-law is about to tell stories - Brother-in-law ix Brother-in-law tells stories - Brother-in-law x Sister-in-law talks to Andrew Sister-in-law Andrew xi. Brother-in-law begins to tell stories Brother-in-law - xii. Brother-in-law tells story Brother-in-law - Table 3.18. \u201cObviative\u201d referents and speaking in a text (93) xii. \u201c\u142f\u152d\u1479\u1424 \u14a6\u14c7,\u201d \u1403\u1458\u1424, \u201c\u140a\u152d, \u1401 \u1434\u14ef\u1426\u1403\u1472\u140f\u152d\u1423 Debden \u2026\u201d \u201cp\u00eayakw\u00e2w m\u00een,\u201d itw\u00eaw, \u201caya, \u00ea-p\u00f4sihikawiy\u00e2n Debden \u2026\u201d p\u00eayakw\u00e2w m\u00eena itw\u00ea-w aya, \u00ea- p\u00f4s=ih -ikawi-y\u00e2n Debden once also say-3 CONN C1-ride=CAUS-IMP -1 Debden \u2018\u201cOnce also,\u201d he said, \u201cI was given a ride to Debden, \u2026\u201d\u2019 (AA 2000:\u00a78.3) Thus, a shift in referent class from \u201cObviative\u201d to \u201cAnimate\u201d has occurred simply because there has been a change in speaker. We see a similar pattern for intransitive speaking verbs in elicitation, where a consultant refused to accept any intransitive speaking verbs marked for obviation (94). 17 Thanks to Inge Genee (p.c.) for this observation. 161 (94) INTRANSITIVE SUBJECT OF SPEAKING VERB CANNOT BE \u201cOBVIATIVE\u201d \u273d\/# Clare \u1401 \u146d\u152a\u1472\u141a\u141f Martha \u1418 \u1401 \u1431\u1433\u14c2\u1528\u1420. \u201c\u14c2\u14a5\u1429\u140d\u1528\u1426\u144c\u1423 \u1406\u14aa,\u201d \u1401 \u1403\u1458\u1528\u141f. \u273d\/# Clare \u00ea-kiyokaw\u00e2t Marthawa \u00ea-piponiyik. \u201cnimiyw\u00eayihten \u00f4ma,\u201d \u00ea-itw\u00eayit. Clare \u00ea- kiyokaw-\u00e2 -t Martha-a \u00ea- pipon- yi-k. Clare C1-visit -DIR-3 Martha-XT C1-winter-DS-0 ni-miyw=\u00eayiht -\u00ea-n aw -ima \u00ea-it =w\u00ea-yi-t 1- good =by.mind.TI-TI-LP PROX-IN.SG C1-thus=AI -DS-3 \u2018ClarePROX went to visit MarthaOBV during the winter. \u201cI am happy about this\u201d, sheOBV said.\u201d\u2019 The speaker\u2019s reasons for rejecting this sentence are illuminating: \u201cIf we use \u2018\u00ea-tw\u00eayit,\u2019 it would be more like a recording, something coming off of an answering machine. It doesn\u2019t sound like it\u2019s coming from a person. Maybe it if was a big committee.\u201d This response suggests that the \u201cObviative\u201d referential class has a strong meaning of \u2018mechanical or impersonal speech\u2019 (i.e. unconscious speech); something that shows a close parallel to the content of \u201cInanimate\u201d referents considered in Chapter 2. Summarizing the patterns for intransitive verbs, we see the following generalizations: (i) \u201cObviative\u201d speakers are almost non-existent. (ii) When a referent begins to speak, they are shifted out of the \u201cObviative\u201d category. (iii) Verbs of speaking with \u201cObviative\u201d referents are interpreted as (infelicitous) mindless, mechanistic activities. These generalizations show that speaking verbs, as a means of assigning an intention to a referent, pattern with verbs of feeling. \u201cObviative\u201d referents are contextually extentional and thus cannot be assigned intentions. 3.4.2. \u201cObviative\u201d referents lack awareness The concept of awareness inherently requires a relation between a perspective and a spatio- temporal context associated with the referent that has this perspective. For example, I could be said to be aware of the sunshine at Eau Claire Dells today if I have a perspective about it and I have a spatio-temporal relation to it during that perspective possession (e.g. I am at Eau Claire Dells, I see and feel the sunshine, and I have a perspective about the sunshine). If I lack either (i) a spatio-temporal relation to this situation, or (ii) an perspective on this situation, I cannot be said to have awareness during that situation. For example, if I am in Vancouver during this sunshine, I can have all the perspectives I want about it, but cannot be said to be aware of it. Likewise, were I to go to Eau Claire Dells and drink two 24-packs of Leinenkugels, I would have a spatio- temporal relation to the situation of it being sunny there, but no associated perspective on it. This means that a spatio-temporal relation is a necessary, but not sufficient condition on awareness; if 162 I do not have a spatio-temporal relation, I cannot be aware, but if I have a spatio-temporal relation, I may still not be aware. Awareness entails a spatio-temporal relation, but a spatio- temporal relation does not entail awareness. An \u201cObviative\u201d referent is extentional in some context (EXT(x,C)). Given some context, this definition means that an \u201cObviative\u201d referent should either (i) lack a spatio-temporal relation to the context, or, if the referent has a spatio-temporal relation to it, then (ii) the \u201cObviative\u201d referent should have no perspective on it (i.e. they will not be intentional in that context). The data supports this expectation: (i) A referent that has no spatio-temporal relation to a context will be obviated (\u00a73.4.21). (ii) When an \u201cObviative\u201d referent has a spatio-temporal relation to the context, the \u201cObviative\u201d referent is interpreted as having no intention in that context (\u00a73.4.22). 3.4.21. If contextual absence, referent is \u201cobviative\u201d Based on the definition of awareness given above, one way that a referent could lack an intention in a context is by not being present in the context. Lacking a spatio-temporal relation to the context, it would be hard for a referent to have any intentions associated with it.18 When one of the referents involved in an event is not present for the event, that referent will invariably be obviated. For example, in a kak\u00eask\u00eehk\u00eamowin \u2018counselling speech\u2019 given by k\u00e2-pimw\u00eaw\u00eahahk (JKN 1998), the Speaker keeps the old man and his father unmarked throughout a discussion of the counselling they gave him. The discourse structure is schematized in Table 3.19, which shows that the old man becomes \u201cObviative\u201d when a referent that is present in the room is introduced (xii). 18 I am here avoiding the issue of chains of information. For example, if someone who is at Eau Claire Dells calls me and tells me it is sunny, I could be said to be aware of the situation there. As Cook and M\u00fchlbauer (2007) have argued, this chaining of two presences together to create awareness is crucial to Plains Cree grammar. 163 LINE EVENT NON-OBV OBV i. Father used to tell speaker things Father - ii. Speaker met old man Old Man - iii. Old man reminds speaker of father\u2019s words Old man, Father iv. F tells speaker Father - v. Old man tells speaker Old Man - vi. Old man is 88 years old Old Man - vii. Direct quote - - viii. Direct quote - - ix. Direct quote - - x. Old man speaks Old Man - xi. Speaker tells audience member about this when smoke Audience member - xii. Speaker tells audience member what old man said Audience member Old Man Table 3.19. \u201cObviative\u201d reference shifts when topic event shifts After he relates this conversation (i-x), the Speaker shifts to an event that has happened recently between him and an audience member (xi-xii). (95) xi. \u1401\u1418\u146f \u1406\u14aa \u1473 \u140f\u1426\u1455\u14aa\u1418\u1420 \u140a\u1418\u1406\u1455, \u140a\u14c4\u1426\u1428, \u00eawak \u00f4ma k\u00e2-w\u00eehtamawak aw \u00f4ta, anohc, \u00eawakw aw -ima k\u00e2-w\u00eeht-amaw-ak aw -a aw-ita, anohc, resum PROX-IN.SG C2 -tell-APPLIC-1>3 PROX-AN.SG PROX-LOC, today, \u1406\u1455 \u1401 \u1432\u1426\u1464\u152e\u1541, \u00f4ta \u00ea-p\u00eehtw\u00e2yahk, aw -ita \u00ea -p\u00eehtw\u00e2-yahk, PROX-LOC C1-smoke -21p \u2018When I told this onePROX here [1 of audience] about that today, here while we held the pipe ceremony, \u2026 xii. \u1401\u146f\u14ef \u14c2\u1455 \u140b\u148b\u14a7\u1422\u1455\u141a\u1424 \u1406\u14aa \u1401\u1418\u146f \u1406\u1432\u146d\u1422\u1475\u140f\u14c2\u1528\u1424. \u00eakosi nitay-\u00e2cimostaw\u00e2w \u00f4m \u00eawak \u00f4p\u00eekiskw\u00eawiniyiw. ekosi nit-ay -\u00e2cimo -staw-\u00e2 -w aw -ima \u00eawakw \u00f4-p\u00eekiskw\u00ea-w-in -yi-w thus 1 -RED-tell.story-BEN-DIR-3 PROX-IN.SG resum 3-speak -3-NOM-DS-3 I was repeating hisOBV words to him in this way.\u2019 (K\u00e2-pimw\u00eaw\u00eahahk 1998: 50.11-12) The change in the topic event is signalled here by the use of multiple devices in line (xi). First, the clause is begun with the cleft \u00eawak \u00f4ma \u2018it is this one\u2019 which is the Plains Cree equivalent to a paragraph marker (cf. M\u00fchlbauer 2003, Wolvengrey 2007). Then, the referent in the audience is picked out with a proximal deictic awa \u2018this one\u2019 that is in external sandhi with a locative expression \u00f4ta \u2018here\u2019 denoting the location of the Speaker. This is then combined with the particle anohc \u2018now\/today\u2019 which picks out the speech time as the reference time. Finally, in line 164 (xii), an entire clause is used to set the time and place; \u00f4ta \u00ea-p\u00eehtw\u00e2yahk \u2018here while we smoked.\u2019 Thus, the event context has shifted from a recitation of events that are not spatially or temporally connected to the present circumstances to ones that are. Now that the context has shifted to the speech situation, the suffix set -yi-w is affixed to the nominal p\u00eekiskw\u00eawin \u2018word.\u2019 In Plains Cree, this construction is only used when the possessor is an \u201cObviative\u201d referent (\u00a73.3.3). In this context, the possessor is the Speaker\u2019s teacher, who was not present for the event of the Speaker relating his speech. Thus, the teacher is obviated. This is in strong contrast to the equivalent treatment that both the teacher and the Speaker\u2019s father received previously, where both were kept \u201cAnimate.\u201d This is shown in the correlative structure in (96), where the the Speaker\u2019s father is coded as \u201cAnimate\u201d in the first clause, and the old man (awa kis\u00eayiniw) is coded as \u201cAnimate\u201d in the second clause. (96) BOTH FATHER AND OLD MAN ARE \u201cANIMATE\u201d AT SAME TIME iv. \u142f\u152d\u147f \u1473 \u146e \u1403\u14ef \u1411\u1426\u1455\u14aa\u140f\u141f, \u2026 p\u00eayakwan k\u00e2-k\u00ee-isi-w\u00eehtamawit, \u2026 p\u00eayakwan k\u00e2-k\u00ee- isi- w\u00eeht=amaw-it same C2-PREV-thus-tell=APPLIC-3>1 \u2018The same as heAN [my father] told me, \u2026 \u2019 v. \u1401\u146f\u14ef \u1401 \u1403\u14ef \u1411\u1426\u1455\u14aa\u140f\u141f \u140a\u1418\u146d\u14ed\u1528\u14c2\u1424, \u2026 \u00eakos \u00ea-isi-w\u00eehtamawit awa kis\u00eayiniw, \u00eakosi \u00ea- isi- w\u00eeht=amaw-it aw =a kis\u00ea=iyiniw thus C1-thus-tell=APPLIC-3>1 PRX=SG.AN great=man \u2018\u2026 thus the old manAN told me, \u2026\u2019 We can test this generalization in elicitation by constructing a parallel context. For example, suppose I am at a large family gathering at a house. Entering one room, I hear the local priest gossiping impolitely about my mother. After listening for a moment, I leave the room and go to another part of the house, where my mother and some of the other family is. Seeing me, someone in the family asks why I am annoyed. In explanation, I gesture at my mother and utter (97a) felicitiously. Here, ayamiyh\u00eawiyiniw \u2018the priest,\u2019 which identifies the referent who was doing the gossping, has the suffix \u2013a attached, and the TA verb is coded with the inverse -ikw-. This signifies that this referent is \u201cObviative.\u201d Crucially, I cannot, in this context, code nik\u00e2wiy \u2018my mother\u2019 as an \u201cObviative\u201d referent, as in (97b). 165 (97) CONTEXT: MOTHER IS PRESENT, PRIEST IS ABSENT a. MOTHER IS \u201cPROXIMATE\u201d AND PRIEST IS \u201cOBVIATIVE\u201d \u1401 \u14ab\u14a5\u1422\u1470\u14a5\u146f\u141f \u14c2\u1473\u140f\u1429 \u140a\u152d\u14a5\u1426\u1401\u140f\u1528\u14c2\u1418 \u00ea-m\u00e2misk\u00f4mikot nik\u00e2wiy ayamih\u00eawiyiniwa \u00ea- m\u00e2misk\u00f4=m -iko-t ni-k\u00e2wiy ayamih\u00ea-w-iyiniw-a C1-talk =by.mouth.TA-INV-3 1-mother pray -3-person-XT \u2018The priestOBV is talking about my motherPROX\u2019 (Presented S2, S3) b. PRIEST IS \u201cPROXIMATE\u201d AND MOTHER IS \u201cOBVIATIVE\u201d # \u1401 \u14ab\u14a5\u1422\u1470\u14ab\u141f \u14c2\u1473\u140f\u152d \u140a\u152d\u14a5\u1426\u1401\u140f\u1528\u14c2\u1424 # \u00ea-m\u00e2mskom\u00e2t nik\u00e2wiya ayamih\u00eaw\u00eeniw \u00ea- m\u00e2misk\u00f4=m -\u00e2-t ni-k\u00e2wiy-a ayamih\u00ea-w-iyiniw-a C1-talk =by.mouth.TA-DIR-3 1-mother-XT pray -3-person-XT \u2018The priestPROX is talking about my motherOBV\u2019 (Presented S2, S3) In explanation, the consultant explicitly noted that my mother was in the room, that I was explaining why I was angry, and that I had gestured at her before talking. Thus, I had set the context as the here and now of Speech Time; a context that the priest had no spatio-temporal relation to (i.e. he was absent for it), but my mother did. Further, I am speaking to a group that includes my mother; assuming she hears me, she must have an intention associated with speech time. Suppose that we alter the situation so that the event I am discussing is occurring while I am speaking, in the same place I am speaking in. Instead of finding him in one room and going to another room, my mother and the priest are both at the dinner table. The priest is at one end of the table, and my mother is at the other. I again hear the priest talking about my mother and become annoyed. Someone asks me why I look annoyed. In this context, the utterance in (98a) is not felicitous, because it codes n\u00eek\u00e2wiy \u2018my mother\u2019 as \u201cProximate\u201d and ayamih\u00eawiyiniw \u2018the priest\u2019 as \u201cObviative.\u201d By contrast, I can utter (98b) felicitously, which codes n\u00eek\u00e2wiy \u2018my mother\u2019 as \u201cObviative\u201d and ayamih\u00eawiyiniw \u2018the priest\u2019 as \u201cProximate.\u201d (98) CONTEXT: PRIEST IS PRESENT, MOTHER IS ABSENT a. MOTHER IS \u201cPROXIMATE\u201d AND PRIEST IS \u201cOBVIATIVE\u201d # \u1401 \u14ab\u14a5\u1422\u1470\u14a5\u146f\u141f \u14c2\u1473\u140f\u1429 \u140a\u152d\u14a5\u1426\u1401\u140f\u1528\u14c2\u1418 # \u00ea-m\u00e2misk\u00f4mikot nik\u00e2wiy ayamih\u00eawiyiniwa \u00ea- m\u00e2misk\u00f4=m -iko-t ni-k\u00e2wiy ayamih\u00ea-w-iyiniw-a C1-talk =by.mouth.TA-INV-3 1-mother pray -3-person-XT \u2018The priestOBV is talking about my motherPROX\u2019 (Presented S2, S3) 166 b. PRIEST IS \u201cPROXIMATE\u201d AND MOTHER IS \u201cOBVIATIVE\u201d \u1401 \u14ab\u14a5\u1422\u1470\u14ab\u141f \u14c2\u1473\u140f\u152d \u140a\u152d\u14a5\u1426\u1401\u140f\u1528\u14c2\u1424 \u00ea-m\u00e2mskom\u00e2t nik\u00e2wiya ayamih\u00eaw\u00eeniw \u00ea- m\u00e2misk\u00f4=m -\u00e2-t ni-k\u00e2wiy-a ayamih\u00ea-w-iyiniw-a C1-talk =by.mouth.TA-DIR-3 1-mother-XT pray -3-person-XT \u2018The priestPROX is talking about my motherOBV\u2019 (Presented S2, S3) The priest, having a spatial-temporal relationship to the context and an intention (see section 3.4.12 on speaking verbs), cannot be said to be unaware of the event, and thus must not be obviated. By contrast, my mother does not know about the gossiping (the event that sets the context), and thus has no intention associated with the context. 3.4.22. An \u201cobviative\u201d referent that is present is unaware An \u201cObviative\u201d referent has no intentional state relevant to the context. For example, in a simple transitive sentence with minimal contextual support, speakers often respond that the obviated referent is not aware that the event is happening. (99) \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u00ea-w\u00e2pam\u00e2t n\u00e2p\u00eaw iskw\u00eawa \u00ea- w\u00e2p=am -\u00e2 -t n\u00e2p\u00eaw iskw\u00eaw-a C1-see=by.eye.TA-DIR-3 man woman-XT \u2018The manPROX saw the womanOBV.\u2019 (Presented S3) Here, one speaker (S3) explained, \u201cIt\u2019s like the woman doesn\u2019t even know that she\u2019s being seen.\u201d This means that the obviation of iskw\u00eaw \u2018woman\u2019 causes the speaker to judge that referent has no relevant intentional state. If we change iskw\u00eaw \u2018the woman\u2019 to \u201cProximate\u201d and instead code n\u00e2p\u00eaw \u2018the man\u2019 as \u201cObviative,\u201d we get a concurrent alteration in judgments. Here, one speaker (S2) suggested that the sentence meant that the woman was now talking to the speaker, about this man who was staring at her. (100) \u1401 \u141a\u1438\u14a5\u146f\u141f \u14c8\u142f\u1418 \u1403\u1422\u1475\u1424 \u00ea-w\u00e2pamikot n\u00e2p\u00eawa iskw\u00eaw \u00ea- w\u00e2p=am -ikw-t n\u00e2p\u00eaw-a iskw\u00eaw C1-see=by.eye.TA-INV-3 man -XT woman \u2018The manOBV saw the womanPROX.\u2019 (Presented S2) That is, the consultant has shifted the context to one involving the Speaker and the woman, neither of whom know anything about why this man is doing this. 167 A referent cannot be coded as \u201cObviative\u201d in an event that they are both present for and must be intentional in. Recall from the discussion of intransitive speech verbs (section 3.4.12. above) that an intransitive verb of speaking does not easily accept \u201cObviative\u201d arguments. Here, the referent Martha has been coded as \u201cObviative\u201d, via the affixation of \u2013a to her name (Marthawa), and the use of a TA direct verb form (kiyokaw\u00e2-). (101) SPEAKING REFERENT CANNOT BE \u201cOBVIATIVE\u201d \u273d\/# Clare \u1401 \u146d\u152a\u1472\u141a\u141f Martha \u1418 \u1401 \u1431\u1433\u14c2\u1528\u1420. \u201c\u14c2\u14a5\u1429\u140d\u1528\u1426\u144c\u1423 \u1406\u14aa,\u201d \u1401 \u1403\u1458\u1528\u141f. \u273d\/# Clare \u00ea-kiyokaw\u00e2t Marthawa \u00ea-piponiyik. \u201cnimiyw\u00eayihten \u00f4ma,\u201d \u00ea-itw\u00eayit. Clare \u00ea- kiyokaw -\u00e2 -t Martha-a \u00ea- pipon- yi-k. Clare C1-visit -DIR-3 Martha-XT C1-winter-DS-0 ni-miyw=\u00eayiht -\u00ea-n aw =ima \u00ea-it =w\u00ea-yi-t 1- good =by.mind.TI-TI-LP PROX=IN.SG C1-thus=AI -DS-3 \u2018ClarePROX went to visit MarthaOBV during the winter. \u201cI am happy about this\u201d, sheOBV said.\u201d\u2019 According to one consultant (S2), maintaining this referent\u2019s \u201cObviative\u201d status is unacceptable with the subsequent intransitive speech verb (itw\u00ea-), because it would suggest that \u201cit would be more like a recording, something coming off of an answering machine. It doesn\u2019t sound like it\u2019s coming from a person. \u201d Instead, the speaker requires that Martha be switched to an \u201cAnimate\u201d referent, with the repetition of a kin-term used to aid reference-tracking (\u00a74.4.1). (102) SUBJECTS OF INTRANSITIVE VERBS OF SPEAKING MUST BE \u201cANIMATE\u201d i. Clare \u1401 \u146d\u152a\u1472\u141a\u141f Martha \u1418 \u1401 \u1431\u1433\u14c2\u1528\u1420. \u2026 Clare \u00ea-kiyokaw\u00e2t Marthawa \u00ea-piponiyik. \u2026 Clare \u00ea- kiyokaw-\u00e2 -t Martha-a \u00ea- pipon- yi-k. Clare C1-visit -DIR-3 Martha-XT C1-winter-DS-0 \u2018ClarePROX went to visit MarthaOBV during the winter.\u2019 ii. \u201c\u14c2\u14a5\u14ed! \u14c2\u14a5\u1429\u140d\u1528\u1426\u144c\u1423 \u1406\u14aa,\u201d \u1401 \u1403\u1458\u141f. \u201cnimis\u00ea! nimiyw\u00eayihten \u00f4ma,\u201d \u00ea-itw\u00eat. ni-mis -\u00ea ni-miyw=\u00eayiht -\u00ea-n aw -ima \u00ea-it =w\u00ea-t 1-old.sister-VOC 1- good =by.mind.TI-TI-LP PROX-IN.SG C1-thus=AI -3 \u2018 \u201cOlder sister! I am happy about this,\u201d she said.\u2019 (Volunteered S2) In light of the facts about obviation and awareness, we can now understand why this is; since the context is set as Clare\u2019s visit home, a context that Martha was present for, the use of obviation for Martha would mean that she had no intention associated with the context that she is speaking in (hence the \u2018mechanical\u2019 interpretation). 168 3.4.3. The emergence of \u201cproximate\u201d via contrast with \u201cobviative\u201d As we have seen, utterances involving only a single \u201cAnimate\u201d third person are unspecified for intentional properties. (103) \u201cANIMATE\u201d REFERENT IS UNSPECIFIED FOR INTENTIONALITY \u14c2\u141a\u1438\u14ab\u1424 \u140a\u14c7 \u140a\u141a\u14ef\u1422. niw\u00e2pam\u00e2w ana aw\u00e2sis. ni-w\u00e2pam -\u00e2 -w an =a aw\u00e2sis 1- see=by.eye.TA-DIR-3 DST=AN.SG child \u2018I see that childAN.\u2019 (Presented S2) COMMENT (S2): \u201cI don\u2019t know if this child knows anything about this or not.\u201d However, there are particular contexts in which an \u201cAnimate\u201d referent will suddenly appear to have crucial intentional information ascribed to it. In the example in (104), the referent iskw\u00eaw \u2018woman\u2019 is the object of an inverse-marked TA verb \u00ea-w\u00e2pamikot \u2018heOBV sees herPROX,\u2019 and has no nominal marking. Speakers (e.g. S2) respond that, \u201cIn [104], it seems like you\u2019re hearing what the woman has to say about it. She could be saying \u2018the man is looking at me.\u2019\u201d (104) \u201cANIMATE\u201d REFERENT GAINS INTENTIONS IN CONTEXT OF \u201cOBVIATIVE\u201d \u1401 \u141a\u1438\u14a5\u146f\u141f \u14c8\u142f\u1418 \u1403\u1422\u1475\u1424 \u00ea-w\u00e2pamikot n\u00e2p\u00eawa iskw\u00eaw \u00ea- w\u00e2p=am -ikw-t n\u00e2p\u00eaw-a iskw\u00eaw C1-see=by.eye.TA-INV-3 man -XT woman \u2018The manOBV saw the womanPROX.\u2019 (Presented S2) COMMENT (S2): \u201cThe woman is talking here.\u201d Lacking both a speaking verb and a previous speaking context, there must be something else in the sentence that conveys that the woman is speaking and possessing a perspective (i.e. is intentional). If only the \u201cObviative\u201d referent has any content, how could the \u201cAnimate\u201d form here seem to have intentional content? In this section, I argue that this reflects the contrastive relationship between the \u201cObviative\u201d and the unspecified \u201cAnimate\u201d classes. When a simple \u201cAnimate\u201d can be set up in contrast to the \u201cObviative,\u201d this gives rise to a meaning for the simple \u201cAnimate\u201d that is the opposite of \u201cObviative.\u201d The intentional \u201cProximate\u201d referent emerges froms this contrast (Wolfart 1973, 1978). The \u201cAnimate\u201d form has no inherent meaning of its own, but can acquire a meaning in the context of a contrast with the \u201cObviative.\u201d This contrast can be shown either overtly, through 169 the opposition of two referents (\u00a73.4.31.), or covertly, through the choice of one class of referent over the other (\u00a73.4.32.). 3.4.31. \u201cObviative\u201d as context for \u201cproximate\u201d As we have just seen, when an \u201cAnimate\u201d referent interacts with an \u201cObviative,\u201d the \u201cAnimate\u201d referent is judged to be intentional. One way to see this effect in full operation is to take a natural discourse involving a human and an animal and reverse the \u201cObviative\u201d\/ \u201cProximate\u201d reference. A discourse of this kind can be found in Alice Ahenakew\u2019s narrative of an elderly couple surviving starvation by shooting a goose through the smoke hole of their tipi (Ahenakew 2000:\u00a79). In its original structure, this story involves a simple-\u201dAnimate\u201d man shooting an \u201cObviative\u201d goose (105). (105) \u201cPROXIMATE\u201d MAN SHOOTS \u201cOBVIATIVE\u201d GOOSE i. \u146b\u1455\u1426\u1455\u140d \u1401\u14f4 \u1473 \u141a\u1438\u14ab\u141f \u1401\u146f\u144c \u1401 \u1431\u14a5\u1426-~, \u1403\u1422\u1431\u14a5\u1541 \u1401 \u1431\u14a5\u1426\u152e\u1528\u141f \u14c2\u1422\u1472. k\u00eatahtaw \u00easa k\u00e2-w\u00e2pam\u00e2t \u00eakot\u00ea \u00ea-pimih-~, ispimihk \u00ea-pimihy\u00e2yit niska. k\u00eatahtaw\u00ea \u00easa k\u00e2-w\u00e2p=am -\u00e2 -t \u00eakot\u00ea ispimihk \u00ea- pimihy\u00e2 -yi-t nisk -a at.time EVID C2-see=by.eye.TA-DIR-3 there above C1-fly.along-DS-3 goose-XT \u2018\u2026when suddenly hePROX saw some geeseOBV flying overhead.\u2019 ii. \u14a3\u1450\u14c2 \u1401\u14f4 \u1473 \u1401\u146d\u148b\u1473\u140f\u141f \u1401\u14f4 \u1401 \u1403\u14ef\u1456\u148b\u14a7\u141f \u1403\u1422\u1481\u144c\u14a5\u1541, m\u00eaton \u00easa k\u00e2-\u00eakicik\u00e2wit \u00eas \u00ea-isit\u00e2cimot iskw\u00e2ht\u00eamihk, mitoni \u00easa k\u00e2-y\u00eekicik\u00e2w =i -t \u00easa \u00ea- isi=t\u00e2cimot iskw\u00e2ht\u00eam-ihk, very evid C2-slow.person=AI-3 EVID C1-rr=by.crawling.AI-3 door -LOC \u2018heAN was very slow in crawling to the door,\u2019 iii. \u1401\u146f\u1455 \u1401 \u1431\u14aa\u146f\u148b\u14c2\u1528\u141f \u1401\u14f4 \u1473 \u1439\u1422\u146d\u1503\u141f \u1401\u146f\u14c2 \u140a\u14c2\u1426\u1403 \u14c2\u1422\u1472; \u00eakota \u00ea-pimakociniyit \u00easa k\u00e2-p\u00e2skisw\u00e2t \u00eakoni anihi niska; \u00eakota \u00ea-pim=akocini-yi-t \u00easa k\u00e2-p\u00e2sk=isw -\u00e2 -t \u00eakoni an=ihi nisk-a there c1-along=fly-DS-3 evid c2-shoot=by.violent.TA-DIR-3 resum DST=XT goose-XT \u2018and as the geeseOBV flew over hePROX shot at themOBV; iv. \u14c2\u1438\u1426\u1401\u1424, \u142f \u14c3\u1426\u1455\u146f\u148b\u14c2\u1528\u1418. nipah\u00eaw, p\u00ea-n\u00eehtakociniyiwa. nip=ah -\u00ea -w p\u00ea-n\u00eeht =akocini-yi-w-a die=by.tool-DIR-3 along-down=fly -DS-3-XT hePROX killed oneOBV and itOBV came falling down.\u2019 (Ahenakew 2000:\u00a79.8) Here, n\u00e2p\u00eaw \u2018the man\u2019 is treated as the \u201cAnimate\u201d referent and niska \u2018the goose\u2019 is treated as \u201cObviative.\u201d This is seen by the use of \u201cAnimate\u201d verb forms for the man (e.g. k\u00e2-y\u00eekicik\u00e2wit \u2018he was crawling\u2019), the use of TA direct verb forms when the man acts on the goose (e.g. k\u00e2- 170 w\u00e2pam\u00e2t \u2018hePROX saw itOBV\u2019), and \u201cObviative\u201d forms for the goose (e.g. p\u00ea-n\u00eehtakociniyiwa \u2018itOBV came falling down\u2019). This is summarized in Table 3.19. LINE EVENT PROXIMATE OBVIATIVE i. Man sees goose flying Man Goose ii. Man crawls to door Man - iii. Goose flies, Man shoots goose Man Goose iv. Man kills goose, it falls Man Goose Table 3.20. Obviated animal in a story If \u201cObviative\u201d codes contextual extentionality, and \u201cAnimate\u201d forms contrasted with it code contextual intentionality, switching the classes for the man and the goose in this story should result in an infelicitous interpretation; the goose would be telling the Speaker about an event of a man shooting it. To test this, I went through this story with a consultant (S2), condensing it into its essentials, and accomodating the consultant\u2019s dialect differences at the same time. I then manipulated the \u201cObviative\u201d and \u201cAnimate\u201d coding to make the man \u201cObviative\u201d and the goose \u201cAnimate.\u201d This is done by changing the nominal modifiers for n\u00e2p\u00eaw \u2018man\u2019 to \u201cObviative\u201d foms (e.g. anihi n\u00e2p\u00eawa \u2018that manOBV\u2019 rather than ana n\u00e2p\u00eaw \u2018that manAN\u2019), using \u201cObviative\u201d verb forms (e.g. \u00ea-n\u00f4ht\u00eakat\u00eayit \u2018heOBV was starving\u2019), and using inverse verb forms when the man interacts with the goose (e.g. w\u00e2pmikow \u2018heOBV sees itPROX\u2019). (106) \u201cOBVIATIVE\u201d MAN SHOOTS \u201cPROXIMATE\u201d GOOSE i. \u1473\u152e\u1422, \u140a\u14c2\u1426\u1403 \u14c8\u142f\u1418 \u1401 \u14c5\u1426\u144c\u1473\u144c\u1528\u141f. k\u00e2y\u00e2s, anihi n\u00e2p\u00eawa \u00ea-n\u00f4htek\u00e2teyit. k\u00e2y\u00e2s, an =ihi n\u00e2p\u00eaw-a \u00ea-n\u00f4hte=k\u00e2t\u00ea-yi-t. long.ago DST=XT man -XT C1-lack=eat -DS-3 \u2018In the past, this manOBV was hungry. ii. \u14a8\u152d \u146d\u14ca\u1422 \u141a\u1438\u14a5\u146f\u1424 \u1431\u1526\u14ef\u1422 \u1403\u1422\u1431\u14a5\u1541. m\u00f4y kinw\u00eas w\u00e2pamikow piy\u00easis ispimihk.19 m\u00f4ya kinw\u00eas w\u00e2p=am -ikw-w piy\u00easis ispim-ihk. NEG long.time see =by.eye.TA-INV-3 bird above-LOC \u2018In a short while, heOBV saw a birdPROX up above.\u2019 19 Two relevant notes on the data: (i) this Speaker (S2) regularly drops final pre-aspirated consonants, and (ii) the inverse form preferred here is the unexpected w\u00e2pamikow rather than the expected w\u00e2pamik (cf. Wolfart 1973). 171 iii. \u1401\u1405\u144e\u14c7\u1541 \u1405\u1438\u1422\u146d\u14ef\u1472\u1423, \u00ea-otinamiyit opaskisikan,20 \u00ea- ot =in -am-yi -t w-pask=is=ikan, C1-take=by.hand-TI -DS-3 3-shoot=by.violence.TI=NOM \u2018HeOBV took his gunIN,\u2019 iv. \u1401\u147f \u1401 \u1439\u1422\u146d\u14ef\u146f\u141f \u1431\u1526\u14ef\u1422. \u00eakwa \u00ea-p\u00e2skisokot piy\u00easis. \u00eakwa \u00ea-p\u00e2sk=isw -iko-t piy\u00easis and C1-shoot=by.violence.TA-INV-3 bird \u2018and heOBV shot the birdPROX.\u2019 (Translated from Cree to Cree by S2) In this retelling, the man has been made \u201cObviative\u201d and the bird has been made \u201cProximate.\u201d This is perfectly grammatical and was ruled well-formed by the Speaker. This is summarized in Table 3.20, which shows that the bird is now \u201cProximate\u201d and the man is now \u201cObviative.\u201d LINE EVENT \u201cPROXIMATE\u201d \u201cOBVIATIVE\u201d i. Man is hungry - Man ii. Man sees bird Bird Man iii. Man takes gun - Man iv. Man shoots bird Bird Man Table 3.21. Chart of obviation for modified re-telling Although this short text is structurally well-formed, it meant something strange. As the Speaker (S2) said, while laughing, \u201cYou\u2019re trying to say it from the point of view of the bird. You\u2019d be understood that it\u2019s from the point of view of the bird, but they [the elder Cree Speakers S2 knew in her life] would say \u2018What the hell are you talking about?\u2019 \u2018You\u2019re trying to talk about it from the bird\u2019s point of view\u2019.\u201d Thus, the juxtapositioning of an \u201cObviative\u201d form with a \u201cAnimate\u201d one creates an intentional, perspectival \u201cProximate\u201d referent out of the \u201cAnimate.\u201d There are also cases where the same referent shifts from one category to the other, creating a contrast in intentional content. For example, in Alice Ahenakew\u2019s (Ahenakew 2000) telling of a windigo21 hunting a family friend, the windigo starts out as \u201cObviative.\u201d Throughout the narrative, the windigo remains \u201cObviative\u201d, until one crucial juncture in the story (107).22 20 On some occasions, this Speaker lacks the \u201cObviative\u201d form of possession and \u201cObviative\u201d marking on TI verbs \u2013 this was one of those times. (See discussion in \u00a75) Specifically, the expected \u201cObviative\u201d forms \u00ea-otinamiyit and opaskiskaniyiw were not offered here. 21 For those unfamiliar with windigos: windigos are typically described as ex-humans who have become supernatural by consuming raw flesh, usually human flesh (cf. Brightman 1994). They typically hunt people who have spiritual powers of some kind. 22 Examples like this, in which a nominal switches from \u201cObviative\u201d to proximate across two clauses, are one of several good reasons why the theme sign system cannot be argued to produce the proximate-\u201dObviative\u201d contrast in Plains Cree. Other evidence includes AIt forms (where obviation holds in a verb that lacks a theme sign), possessor uses (where obviation holds without a verb at all), and embedded clauses such as John niw\u00eehtam\u00e2k Clara \u00ea-ahkosiyit 172 Here, the nominal w\u00eehtikow is \u201cObviative\u201d in the first clause (107i), as it presses upon the family friend. In the clause immediately following, however, w\u00eehtikow has been stripped of its \u201cObviative\u201d marking and is connected to a verb lacking the different-subject marker -yi-. (107) WINDIGO SHIFTS FROM \u201cOBVIATIVE\u201d TO \u201cANIMATE\u201d i. \u2026 \u1403\u1528\u146f\u1541 \u14ab\u14c7 \u1401 \u146e \u140a\u1528 \u14ab\u146f\u1426\u1403\u146f\u141f \u140a\u14c2\u1426\u1403 \u1411\u144e\u146f\u1418 \u1456\u143b, \u2026 iyikohk m\u00e2na \u00ea-k\u00ee-ayi-m\u00e2kohikot anihi w\u00eehtikowa t\u00e2pw\u00ea, iyikohk m\u00e2na \u00ea- k\u00ee- ayi- m\u00e2ko =h -iko-t an=ihi w\u00eehtikow-a t\u00e2pw\u00ea so.much usual C1-PREV-CONN-trouble=by.neut-INV-3 DST=XT windigo -XT true \u2018and hePROX [n\u00eaw\u00e2pisk] was truly pressed upon by that windigoOBV, \u2026\u2019 ii. \u2026\u1401 \u146e \u14f5\u1422\u1472\u1458\u14a7\u141f \u14ab\u14c7 \u140a\u14c7 \u1411\u1426\u144e\u146f\u1424. \u2026 \u00ea-k\u00ee-s\u00e2skatw\u00eamot m\u00e2n \u00e2na w\u00eehtikow. \u00ea- k\u00ee- s\u00e2skatw\u00eamo-t m\u00e2na an-a w\u00eehtikow c1-PREV-scream -3 usual DST-AN.SG windigo \u2018\u2026 so much did that windigoPROX used to wail.\u2019 (Ahenakew 2000:\u00a71.3) Going through this story with a native speaker (S2), this passage produced an interesting (and unexpected) result; upon hearing the second clause, the speaker exlaimed, \u201cSo she heard it too!! Oh my gosh!\u201d She was emphatic that this structure meant that the Speaker (Alice Ahenakew) had heard this windigo herself. Thus, switching a nominal from \u201cObviative\u201d to unmarked makes the referent in one context contrast with itself in another context. The windigo is here interpeted as \u201cProximate\u201d as a function of a contrast with its \u201cObviative\u201d form. 3.4.32. Paradigmatic contrast of \u201canimate\u201d with \u201cobviative\u201d As Cook and M\u00fchlbauer (2007) point out, there is another context in which the \u201cAnimate\u201d form has a particular meaning; paradigms volunteered for coding a chain of information. One speaker (S4) volunteered a paradigm for the appropriate way to use grammatical devices to code the utterer\u2019s spatio-temporal relationship to the referent (cf. Cook & M\u00fchlbauer 2007). In this paradigm, the simple \u201cAnimate\u201d form is used when the Speaker had direct contact with the referent (108a), while the \u201cObviative\u201d form means that the Speaker did not have direct knowledge (108b). \u2018JohnPROX told me that ClareOBV is sick.\u2019 All of these are cases of proximate\/\u201dObviative\u201d contrast without a theme sign involved. 173 (108) a. \u1401\u146f\u14ef \u1403\u1458\u1424 \u1438\u141a\u1472\u1423. \u00eakosi itw\u00eaw paw\u00e2kan. \u00eakosi it =w\u00ea-w paw\u00e2kan thus thus=AI -3 dream.spirit \u2018That\u2019s what the dream spiritAN said.\u2019 (Volunteered by S4) = Speaker heard the dream spirit say it b. \u1401\u146f\u14ef \u1401\u1403\u1458\u1528\u141f \u1438\u141a\u1472\u14c7. \u00eakosi \u00ea-itw\u00eayit paw\u00e2kana. \u00eakosi \u00ea-itw\u00ea-yi-t paw\u00e2kan -a thus c1-say-DS-3 dream.spirit-XT \u2018That\u2019s what the dream spiritobv said.\u2019 (Volunteered by S4) = Speaker heard about this from someone who did not witness it. Here, the \u201cProximate\u201d form correlates with the Speaker\u2019s direct experience, and is the form used when the Speaker knows the referent\u2019s intentional state. To understand how these examples work, it is important to remember that a native speaker always knows what has not been said, as well as what has been said. In this set of examples, the choice between forms is unconstrained; the Speaker can choose either the simple \u201cAnimate\u201d or \u201cObviative\u201d construction. By choosing the \u201cAnimate\u201d form, the Speaker is implicating that they have a reason for not doing this (i.e. is creating an implicature). Understanding this implication thus requires an evaluation of alternative forms \u2013 an operation that creates a contrastive relation between the simple \u201cAnimate\u201d and \u201cObviative\u201d forms. This gives rise to the intentional, \u201cProximate\u201d interpretation of simple \u201cAnimate\u201d. 3.4.4. Summary: \u201cObviative\u201d as contextual extentionality An \u201cObviative\u201d bears contextual, extentional semantics. This is shown by sets of evidence: (i) The distribution and interpretation of \u201cObviative\u201d with verbs that ascribe intentional properties (\u00a73.4.1.) (ii) The distribution and interpretation of \u201cObviative\u201d in contexts of awareness (presence\/absence: \u00a73.4.2.) (iii) The ability of \u201cObviative\u201d to contrastively create a contextually intentional referent, termed the \u201cProximate\u201d (\u00a73.4.3.) I now turn to the contextual properties of \u201cObviative.\u201d 174 3.5. The context of obviation: Speaker knowledge of intentions As we have seen, a consideration of the content of obviation must constantly involve its context of use. Given the content of obviation, this inherent reference to context should not be surprising; an \u201cObviative\u201d referent is extentional (i.e. lacks a perspective) in a context. Both the notion of \u201cperspective\u201d and, obviously, the notion of \u201ccontext\u201d entail that context is as crucial to understanding \u201cObviative\u201d referents as either its form or content. In this section, I argue that the \u201cObviative\u201d form is contextually extentional with respect to the Speaker\u2019s perspective. When a consultant is asked about the usage of \u201cObviative\u201d constructions, a consensus emerges; as a Speaker, consultants always know what form to say in a given context, but, as a Hearer, they are not able to identify why someone else select a form. As an example, consider the following situation; the Speaker is standing outdoors, talking on a cellphone to her friend. She is standing near a low hill and some trees. There is a female stranger standing close by, who suddenly turns to the Speaker and complains that a man over beyond the hill is staring at her. The Speaker cannot see the man, but wishes to report this event to her friend on the phone. In such a context. In this context, the man must be \u201cObviative\u201d and the woman who is telling the Speaker about it must be \u201cProximate,\u201d as shown in (109). (109) CONTEXT: UNKNOWN, OUT-OF-SIGHT MAN SEES WOMAN WHO TELLS SPEAKER a. \u201cOBVIATIVE\u201d MAN AND \u201cPROXIMATE\u201d WOMAN \u1401 \u141a\u1438\u14a5\u146f\u141f \u1403\u1422\u1475\u1424 \u14c8\u142f\u1418 \u00ea-w\u00e2pamikot iskw\u00eaw n\u00e2p\u00eawa \u00ea-w\u00e2p=am -ikw-t iskw\u00eaw n\u00e2p\u00eaw-a c1-see=by.eye.TA-INV-3 woman man -XT \u2018A\/the manOBV is seeing a\/the womanPROX.\u2019 (Presented S2) b. \u201cPROXIMATE\u201d MAN AND \u201cOBVIATIVE\u201d WOMAN # \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418. # \u00ea-w\u00e2pam\u00e2t n\u00e2p\u00eaw iskw\u00eawa. \u00ea-w\u00e2p=am -\u00e2 -t n\u00e2p\u00eaw iskw\u00eaw -a c1-see=by.eye.TA-DIR-3 man woman -XT \u2018A\/the manPROX is seeing a\/the womanOBV.\u2019 (Presented S2) In explaining this choice, consultants are confident about the judgment of felicity, and they appeal to the organization of the context \u2013 the woman is talking, the man is unknown. Suppose, now, that the speech roles are reversed, and the consultant is the hearer on the phone, not the Speaker of the sentence. She hears her friend utter the same sentence either (109a) or (109b). Suddenly, she is uncertain as to why the obviation pattern has been chosen this way, and the 175 clear felicity judgment evaporates. As the hearer, she cannot tell what information is being encoded in the \u201cObviative\u201d \u201cProximate\u201d chaining. If I ask consultants to think about it, all they can offer is why they would say it that way, and what possible context configurations would lead them to speak in that way. The \u201cObviative\u201d construction, then, appears to place constraints on speaking, but not on hearing. In this section, I argue that this asymmetry between speaking \u201cObviative\u201d forms and hearing \u201cObviative\u201d forms reflects a core property of obviation: a referent\u2019s contextual extentionality is evaluated within the perspective of the Speaker. Building on the model I constructed for animacy (Chapter 2), I once again utilize a modified version of the Discourse Representation Theory (DRT: Kamp 1981) model that was proposed by Farkas (1992). The \u201cObviative\u201d form places a restriction on embeddings; \u201cObviative\u201d referents cannot possess a perspective embedded within the Speaker\u2019s. For example, an \u201cObviative\u201d referent like iskw\u00eawa \u2018the womanOBV\u2019 in (110) cannot possess a perspective embedded inside the Speaker\u2019s perspective. (110) a. \u201cOBVIATIVE\u201d WOMAN \u1401 \u146d\u14c4\u14ef\u1528\u141f \u1403\u1422\u1475\u1418 \u2026\u00ea-kinosiyit iskw\u00eawa \u00ea-kinw =si -yi -t iskw\u00eaw-a C1-long=AI-DS-3 woman-XT \u2018\u2026(as) the womanOBV was tall.\u2019 (Presented S3) b. ILL-FORMED DRS: \u201cOBVIATIVE\u201d WOMAN HAS A PERSPECTIVE Speaker x tall(x) \uf0df WOMAN HAS NO PERSPECTIVE woman(x) EXT(x,C) x \uf0df WOMAN HAS A PERSPECTIVE tall(x) In the context of an extentional \u201cObviative\u201d referent, the unmarked \u201cAnimate\u201d form gains a perspective in which the truth of the proposition can be evaluated. For example, the sentence in 176 (111) shows a \u201cProximate\u201d iskw\u00eaw \u2018woman\u2019 set up in contrast to an \u201cObviative\u201d aw\u00e2sisa \u2018childOBV.\u2019 In this context, the woman can possess a perspective on the propositions associated with the child. (111) a. WOMAN IS \u201cPROXIMATE\u201d AND CHILD IS \u201cOBVIATIVE\u201d \u1401 \u146e \u1411\u1426\u1455\u14aa\u140f\u141f \u1403\u1422\u1475\u1424 \u1401 \u140a\u1426\u146f\u14ef\u1528\u141f \u140a\u141a\u14ef\u14f4 \u2026\u00ea-k\u00ee-w\u00eehtamawit iskw\u00eaw \u00ea-ahkosiyit aw\u00e2sisa \u00ea- k\u00ee- w\u00eeht=amaw -it iskw\u00eaw \u00ea-ahkosi-yi -t aw\u00e2sis-a C1-PREV-tell =APPLIC-3>1 woman C1-sick -DS-3 child-XT \u2018The womanPROX told me that the childOBV is sick.\u2019 (Presented S3) b. WELL-FORMED DRS: \u201cPROXIMATE\u201d WOMAN HAS A PERSPECTIVE Speaker Co x y woman(x) EXT(y,C0) tell(x,S) x \uf0df WOMAN HAS A PERSPECTIVE C1 y sick(y) child(y) When an intentional ascription is made to an \u201cObviative\u201d referent (e.g. with a verb of speaking, feeling, or thinking), this perspective is embedded within the \u201cProximate\u2019s\u201d perspective, rather than directly within the Speaker\u2019s. For example, the form in (112a) shows an \u201cObviative\u201d referent (aw\u00e2sisa \u2018the childOBV\u2019) associated with a psych verb (\u00ea-miyw\u00eayihtamiyit \u2018s\/heOBV is happy about itIN.\u2019). In this context, the \u201cObviative\u201d child is allowed to possess a perspective, because it is embedded inside of the \u201cProximate\u201d woman\u2019s perspective (112b). 177 (112) a. \u201cOBVIATIVE\u201d REFERENT HAS INTENTION IN PERSPECTIVE OF \u201cPROXIMATE\u201d \u2026\u1401 \u146e \u1411\u1426\u1455\u14aa\u140f\u141f \u1403\u1422\u1475\u1424 \u1401 \u14a5\u1429\u140d\u1528\u1426\u1455\u14a5\u1528\u141f \u140a\u141a\u14ef\u14f4 \u2026\u00ea-k\u00ee-w\u00eehtamawit iskw\u00eaw \u00ea-miyw\u00eayihtamiyit aw\u00e2sisa \u00ea- k\u00ee- w\u00eeht=amaw -it iskw\u00eaw \u00ea-miyw=\u00eayiht -am-yi -t aw\u00e2sis-a C1-PREV-tell =APPLIC-3>1 woman C1-good=by.mind.TI-TI -DS-3 child-XT \u2018\u2026the womanPROX told me that the childOBV likesIN it.\u2019 (Presented S3) b. WELL-FORMED DRS FOR (112A): \u201cOBVIATIVE\u201d HAS A PERSPECTIVE EMBEDDED IN \u201cPROXIMATE\u201d PERSPECTIVE. Speaker C0 x y z woman(x) EXT(y,C0) tell(x,S) EXT(z) x \uf0df WOMAN HAS A PERSPECTIVE y z child(y) y \uf0df CHILD HAS A PERSPECTIVE C1 z like(y,z) The \u201cObviative\u201d referent, then, is a distanced referent; one whose perspective can only be related to the Speaker\u2019s through some intermediate embedding. In order for this model to adequately cover the data, obviation must behave in a specific set of ways in a specific set of contexts: (i) Because \u201cObviative\u201d referents cannot possess a perspective embedded within the Speaker\u2019s, \u201cObviative\u201d referents should either be forbidden from speaking to the Speaker, or it should force a dissociative reading (Lakoff 1970). This is exactly the pattern of \u201cObviative\u201d Speaker interactions (\u00a73.5.1). 178 (ii) Because an \u201cObviative\u201d referent\u2019s perspective cannot be directly embedded in the Speaker\u2019s perspective, an intermediary referent must always be present; even if none has been introduced. This pattern holds in Plains Cree (\u00a73.5.2). (iii) The prohibition of directly embedding the \u201cObviative\u2019s\u201d perspective in the Speaker\u2019s means that \u201cObviative\u201d referents share many properties in common with indirect evidentials, and thus ought to pattern closely with them. The evidence from Plains Cree confirms this (\u00a73.5.3). Across all of these contexts, then, we see that \u201cObviative\u201d referents are systematically extentional with respect to the Speaker\u2019s perspective. 3.5.1. \u201cObviative\u201d perspectives force Speaker dissociation \u201cObviative\u201d referents cannot possess a perspective directly embedded inside of the Speaker\u2019s perspective. If this is so, we expect that psych predicates that involve both the Speaker and an \u201cObviative\u201d referent ought to be highly restricted. In particular, \u201cObviative\u201d referents should not be able to speak directly to the Speaker, since this would create a perspective for the \u201cobviative\u201d referent within the Speaker\u2019s perspective. In all four texts I have considered (Ahenakew 2000, k\u00e2-pimw\u00eaw\u00eahahk 1998, Minde 1997, and Whitecalf 1993), only a simple \u201cAnimate\u201d referent can speak to the Speaker, or in the Speaker\u2019s hearing. By contrast, there are no cases of \u201cObviative\u201d referents speaking to the Speaker, or being spoken to by the Speaker. Testing this in eliciation, we see that, although the grammar of Plains Cree allows for the construction of \u201cObviative\u201d Speaker forms (cf. Wolfart 1973), they are consistently rejected by consultants when used with verbs of speaking. (113) \u201cOBVIATIVE\u201d REFERENT INFELICITOUS SPEAKING TO SPEAKER a. # \u14c2\u1411\u1426\u1455\u14ab\u146f\u1528\u1418 # niw\u00eehtam\u00e2koyiwa ni-w\u00eeht=amaw -ikw-yi-w-a 1- tell =APPLIC-INV-DS-3-XT Intended: \u2018S\/heOBV told me.\u2019 (Presented S2, S4) b. # \u1401 \u1411\u1426\u1455\u14aa\u140f\u1528\u141f # \u2026 \u00ea-w\u00eehtamawiyit \u00ea-w\u00eeht=amaw -yi -it C1-tell=APPLIC-DS-3>1 Intended: \u2018s\/heOBV told me.\u2019 (Presented S2) 179 It is possible to construct a context for an \u201cObviative\u201d referent speaking to the Speaker, but it requires a dissociative context, wherein the Speaker does not remember the event, or the event happened in some other world (Lakoff 1970). Consider the example in (114), where an \u201cObviative\u201d referent (iskw\u00eawa \u2018a womanOBV\u2019) speaks to the Speaker using a conjunct order verb inflected for obviation (\u00ea-w\u00eehtamawiyit \u2018sheOBV told me\u2019). This utterance is well-formed only if indirect evidentials (e.g. \u00easa \u2018reportedly\u2019) are used and the Speaker was mentally incapacitated (e.g. unconscious) during the event and unable to remember it. (114) \u201cOBVIATIVE\u201d REFERENTS CAN SPEAK TO SPEAKER WHO DOES NOT REMEMBER IT \u2026 \u1403\u1422\u1475\u1418 \u1401 \u1411\u1426\u1455\u14aa\u140f\u1528\u141f \u1401\u14f4, \u1401 \u140a\u1426\u146f\u14ef\u1528\u141f. \u2026 iskw\u00eawa \u00ea-w\u00eehtamawiyit \u00easa, \u00ea-ahkosiyit. iskw\u00eaw-a \u00ea- w\u00eeht=amaw -yi -it \u00easa \u00ea-ahkosi -yi-t woman-XT C1-tell =APPLIC-DS-3>1 EVID C1-be.sick-DS-3 \u2018\u2026 apparently, the womanOBV told me sheOBV was sick.\u2019 (Volunteered S2) Within the current discussion, we can understand why this is. The form in (114) means that a woman that is obligatorily embedded with respect to some other referent said something to the Speaker (i.e. had Speaker-known intentions). That is, there must be an interecessory between the Speaker and the referent, but the sentence, coding an interaction between the \u201cObviative\u201d referent and the Speaker, is implying that there is no intercessory. How could such a situation happen? The only contexts in which such a thing could even be imagined to occur would be one in which the Speaker was mentally incapacitated when the woman talked to them. It follows, then, that someone has told the Speaker about what this woman said \u2013 the Speaker does not know for themselves, being mentally incapacitated at the time. A context that supports an intermediary referent thus makes this form felicitous, where it could otherwise not be imagined. 3.5.2. \u201cObviative\u201d perspectives force invented \u201cproximate\u201d A context that lacks an intermediate referent to embed the \u201cObviative\u2019s\u201d perspective in forces an estrangement reading (e.g. the Speaker was unconscious). This is because the \u201cObviative\u201d referent cannot have a perspective embedded within the Speaker\u2019s. Another logical possibility for fixing \u201cObviative\u201d contexts is to invent a \u201cProximate\u201d referent to host the \u201cObviative\u201d referent\u2019s perspective. This would solve the embedding problem in a different way; instead of the Speaker being mentally incapacitated, the \u201cObviative\u201d referent\u2019s perspective is embedded with respect to some unknown person. 180 (115) INVENTED REFERENT TO HOST THE \u201cOBVIATIVE\u2019S\u201d PERSPECTIVE S C x y EXT(y,C) \uf0df Y DOES NOT HAVE A PERSPECTIVE x \uf0df X CREATED TO HAVE PERSPECTIVE ON Y y \uf0df Y HAS A PERSPECTIVE VERB(y,S) In fact, there is evidence that Plains Cree makes use of this strategy systematically \u2013 out-of-the- blue \u201cObviative\u201d referents are regularly accomodated by inventing \u201cProximate\u201ds. There are a few occasions in elicitation settings in which the \u201cObviative\u201d form can be considered without a \u201cProximate\u201d form. For example, consider the example in (116). Without any discourse context, the referent iskw\u00eaw \u2018woman\u2019 must be coded as \u201cAnimate,\u201d as in (116a), and an \u201cObviative\u201d form as in (116b) is ruled infelicitous. (116) CONTEXT: OUT-OF-THE-BLUE a. WOMAN IS \u201cOBVIATIVE\u201d # \u1403\u1422\u1475\u1418 \u1401 \u14a5\u152a\u14ef\u1528\u141f. # iskw\u00eawa \u00ea-miyosiyit. iskw\u00eaw-a \u00ea-miyo =si -yi-t woman-XT C1-good=AI-DS-3 \u2018The womanOBV is beautiful.\u2019 (Presented S2) 181 b. WOMAN IS \u201cANIMATE\u201d \u1403\u1422\u1475\u1424 \u1401 \u14a5\u152a\u14ef\u141f. iskw\u00eaw \u00ea-miyosit. iskw\u00eaw \u00ea-miyo =si-t woman C1-good=AI-3 \u2018The womanAN is beautiful.\u2019 (Presented S2) However, if we invent enough context, the \u201cObviative\u201d form in (116b) can become acceptable. Consider a context in which I asked a consultant to picture that they had overheard (116b) at the next booth in a restaurant. The \u201cObviative\u201d form of iskw\u00eawa \u2018womanOBV\u2019 is now felicitous. (117) CONTEXT: OVERHEARD AT A RESTAURANT a. WOMAN IS \u201cOBVIATIVE\u201d \u2026\u1403\u1422\u1475\u1418 \u1401 \u14a5\u152a\u14ef\u1528\u141f. \u2026iskw\u00eawa \u00ea-miyosiyit. iskw\u00eaw-a \u00ea-miyo =si -yi-t woman-XT C1-good=AI-DS-3 \u2018\u2026some guy\u2019s womanOBV is beautiful.\u2019 (Presented S2) COMMENT (S2): \u201cThey were talking about someone\u2019s girlfriend.\u201d b. WOMAN IS \u201cANIMATE\u201d \u2026\u1403\u1422\u1475\u1424 \u1401 \u14a5\u152a\u14ef\u141f. \u2026iskw\u00eaw \u00ea-miyosit. iskw\u00eaw \u00ea-miyo =si-t woman C1-good=AI-3 \u2018\u2026the womanAN is beautiful.\u2019 (Presented S2) In such a context, the utterance was ruled felicitous, but with a crucial change in interpretation; the consultant (S2) explained that this form meant that \u201cthey were talking about someone\u2019s girlfriend, or some man\u2019s wife. It\u2019s like saying \u2018his woman is beautiful\u2019 in English.\u201d This means that the consultant has invented a \u201cProximate\u201d referent; she does not know who this referent is, but she has to posit their existence for the form to be acceptable. Thus, the form is felicitous so long as the consultant can invent some unknown \u201cProximate\u201d referent to fill the gap in the context. Another example of this kind of invention was offered by a different consultant (S4) when explaining how to construct proper chains of information regarding dreams and prophecy (Cook & M\u00fchlbauer 2007). This form employs an \u201cAnimate\u201d intransitive verb form along with - yi- and a nominal marked with \u2013a; the referent n\u00e2p\u00eaw \u2018man\u2019 is \u201cObviative\u201d (\u00a73.3). 182 (118) \u201cOBVIATIVE\u201d SPEECH IS NOT HEARD BY THE SPEAKER \u1401\u146f\u14ef \u1401 \u1403\u1458\u1528\u141f (\u1401\u14f4) \u14c8\u142f\u1418. \u00eakosi \u00ea-itw\u00eayit (\u00easa) n\u00e2p\u00eawa \u00eakosi \u00ea-itw\u00ea -yi -t \u00easa n\u00e2p\u00eaw-a so C1-say-DS-3 EVID man -XT \u2018That\u2019s what the manOBV said.\u2019 (Presented S2,S3) The consultant (S4) explained that this form could be felicitously uttered in a specific context: the Speaker had not heard about the event directly, but had instead heard about it from someone who had only heard about it second-hand.23 Within the current model, we can understand why this construction requires this specific context. The verb itw\u00ea-, being a speaking verb, is inherently intentional \u2013 it automatically assigns a perspective to its agent (\u00a73.4.1). However, the referent doing the speaker is \u201cObviative.\u201d Being \u201cObviative\u201d, this referent\u2019s perspective cannot be directly embedded under the Speaker\u2019s. Instead, there must be some intermediary perspective \u2013 the implied source of this information for the Speaker. The use of this \u201cObviative\u201d construction, then, inherently implicates an intermediate perspective between the Speaker\u2019s and the \u201cObviative\u201d referent\u2019s. 3.5.3. Obviation patterns with indirect evidentiality As we have just seen, a \u201cProximate\u201d referent can be used to facilitate the assignment of a perspective to an \u201cObviative\u201d referent. This \u201cProximate\u201d referent can possess a Speaker- embedded perspective, and thus is able to host a perspective for the \u201cObviative\u201d referent. Put another way, the \u201cProximate\u201d referent can bear the burden of assertion for whatever intentions the \u201cObviative\u201d referent has. This means that \u201cProximate\u201d referents can serve as sources of information for the Speaker. \u201cObviative\u201d referents can\u2019t possess a Speaker-embedded perspective. This means that anything that is held to be true in the \u201cObviative\u201d referent\u2019s perspective must not be directly dependent on the Speaker\u2019s perspective. This means that the \u201cObviative\u201d referent can never serve as a source of information to the Speaker. 23 I here avoid discussion of the required second-hand source; for a thorough discussion, see Cook & M\u00fchlbauer (2007). Briefly, it is the combination of obviation and clause-typing that creates this intermediary \u201cnon-co-present\u201d layer of experiential information. Since the current discussion is about obviation alone, the distinction between these two layers is not necessary. 183 At the core of this referential split is the concept of accessibility (cf. Kratzer 1981, 1991), which can be thought of as a partial ordering between the domains in which propositions are evaluated (cf. Stephenson 2007, Kratzer 1991). One domain is accessible to another domain if the second domain is ordered after it (119). (119) x > y \uf0e0 y is accessible to x A partial ordering (x > y) is one part of the logic of part-whole relations (Link 1983); part-whole relations are a particular kind of partial orderings. Since perspectival embedding is a part-whole relation (perspective x is embedded within perspective y), then, it is also a partial-ordering. This means that accessibility relations hold between perspectives. Under this logic, the \u201cProximate\u201d referent\u2019s perspective, being embedded directly within the Speaker\u2019s, is accessible to the Speaker\u2019s perspective. The \u201cObviative\u201d referent\u2019s perspective, however, is embedded inside of some other perspective, which is not directly accesible to the Speaker, but is instead mediated by a middle perspective (the \u201cProximate\u201d\u2019s). More formally, we can define this as the Accessibility Condition, as in (120). (120) ACCESSIBILITY CONDITION: A perspective \u03c82 is accessible to another perspective \u03c81 iff \u03c82 is embedded inside \u03c81 and there are no intervening embeddings between \u03c82 and \u03c81. Following this condition, propositions contained within a single embedding are accessible x. (121) (\u03a8,X) > (\u03a8,Y) : (\u03a8,Y) ACCESSIBLE TO (\u03a8,X) x y Accessible By contrast, propositions embedded more than one layer inside of x\u2019s are not accessible to the x. 184 (122) (\u03a8,X) > (\u03a8,Y) > (\u03a8,Z) : (\u03a8,Z) INACCESSIBLE TO (\u03a8,X) x z y Inaccessible If we take the outermost box to always be the Speaker, then this Accessibility Condition provides a model of Speaker knowledge. \u201cObviative\u201d perspectives must always be embedded more than one layer away from the Speaker\u2019s, while \u201cProximate\u201d referents can be directly embedded. This sets up a set of accesibility relations between referent\u2019s perspectives; the \u201cObviative\u201d\u2019s perspective is not accessible to the Speaker, while the \u201cProximate\u201d referent\u2019s is. (123) (\u03a8,SPEAKER) > (\u03a8, \u201cPROXIMATE\u201d) > (\u03a8, \u201cOBVIATIVE\u201d : (\u03a8,\u201cOBVIATIVE\u201d) IS INACCESSIBLE TO (\u03a8,SPEAKER) Speaker Accessible Proximate Obviative Inaccessible This means that all \u201cObviative\u201d perspectives are only taken by the Speaker to be contingently true; they are only true contingent on their being true in some intermediate person\u2019s perspective. Multiple embeddings, then, must be processed as chains of information (Landman 1986); accessibility to one referent\u2019s perspective is used to carry the evaluation of the proposition forward to the Speaker\u2019s perspective. 185 (124) (\u03a8,SPEAKER) > (\u03a8, \u201cPROXIMATE\u201d) > (\u03a8, \u201cOBVIATIVE\u201d : (\u03a8,\u201cOBVIATIVE\u201d) IS ACCESSIBLE TO (\u03a8, \u201cPROXIMATE\u201d) Speaker Proximate Accessible Obviative Accessible Accessibility, then, is a condition on the relations between embedded perspectives.24 Given an organization of reference into \u201cProximate\u201d referents (Speaker-accessible asserters), and \u201cObviative\u201d referents (Speaker-unknown referents), we expect a particular set of effects with respect to grammatical forms that code the Speaker\u2019s relation to the truth of a proposition, typically called evidentials (cf. Garrett 2001, Aikhenvald 2004, Blain & D\u00e9chaine 2007, etc.). \u201cProximate\u201d referents, being able to possess perspectives directly accessible to the Speaker\u2019s, ought to have direct evidential force; the Speaker ought to have direct, certain knowledge of propositions that are held to be true in this referent\u2019s perspective. By contrast, \u201cObviative\u201d referents, being unable to possess perspectives directly accessible to the Speaker\u2019s, ought to have indirect evidential force. There are three kinds of evidence that support this expectation in Plains Cree: (i) With psych verbs, which are inherently intentional, \u201cObviative\u201d referents require indirect evidentiality coding, while \u201cProximate\u201d referents do not (\u00a73.5.31). (ii) With speaking verbs, the Speaker can have direct evidence for a \u201cProximate\u201d\u2019s speech act, but can only have indirect evidence for an \u201cObviative\u201d referent\u2019s (\u00a73.5.32). (iii) \u201cObviative\u201d referents pattern closely with indirect evidential markers like \u00easa (\u00a73.5.33). Evidentials code the relation between the perspective in which the proposition is true and the Speaker\u2019s perspective, whereas obviation codes a referent\u2019s ability to possess a Speaker-known perspective. Both kinds of coding work together to connect the Speaker\u2019s perspective to that in which the proposition is true; obviation introduces the referents whose perspectives are considered, and evidentiality qualifies the Speaker\u2019s relation to those worlds. 24 As discussed in \u00a71.3.2, the current model differs from standard DRT in that it does not allow referents to be carried into embedded boxes. This is necessary both for the correlation between a DRS and a perspective, and for these accessibility relations to hold. 186 3.5.31. \u201cObviative\u201d psych arguments and indirect knowledge When we considered the correlation between obviation and psych verbs, we found that there was a strong correlation between psych verbs and obviation; an \u201cObviative\u201d referent is almost never the experiencer of a psychological state. ANIM Psych OBV Psych WORD COUNT Minde 95 13 12,900 Ahenakew 65 3 10,300 Whitecalf 67 8 3,800 Totals 227 24 27,000 Table 3.15. (Repeated): A Summary of psych verbs in 3 texts While this finding is quantitatively significant, it is not qualitatively so; there are 24 cases of \u201cObviative\u201d experiencers with psych verbs. If there was nothing else to be said about these 24 cases, we would have to conclude that there is a qualitative problem with the analysis. However, upon closer inspection, it is apparent that these 24 cases in fact support the generalization, rather than controvert it. When we consider these few cases of \u201cObviative\u201d arguments for psych verbs in more detail, definite patterns emerge: \u201cObviative\u201d referents can be used with psych verbs if either (i) an evidential is used, (ii) an irrealis construction, or (iii) the authority for the sentence has been established as someone other than the Speaker. EVIDENTIAL IRREALIS AUTHORITY Obv Psych 5 15 4 Table 3.22. Categorization of \u201cobviative\u201d with psych verbs For example, the sentence in (125) has an \u201cObviative\u201d argument for the psych verb it\u00eayihtam \u2018think thus of it,\u2019 and employs an overt evidential morpheme \u00eatokw\u00ea. (125) EVIDENTIAL WITH \u201cOBVIATIVE\u201d SUBJECT OF PSYCH VERB \u2026 \u1401 \u146e \u1401\u1450\u1475 \u140a\u152d \u1401\u144c\u1528\u1426\u1455\u14a5\u1528\u141f, \u2026 \u2026, \u00ea-k\u00ee-\u00eatokw\u00ea-aya-it\u00eayihtamiyit, \u2026 \u00ea- k\u00ee- \u00eatokw\u00ea-aya- it =\u00eayiht -am-yi-t C1-PREV-DUBIT-CONN-thus=by.mind.TI-TI-DS-3 \u2018\u2026; heobv [her brother] must have thought \u2026\u2019 (EM 1997:\u00a743) 187 In (126), we see an example where the event is coded as irrealis. In line (i), the verb miyaw\u00e2tamiyit \u2018theyOBV are happy\u2019 has an \u201cObviative\u201d subject (ayisiyiniwa \u2018peopleOBV\u2019), and is marked with the preverb ka-, which is used in contexts where the truth of the proposition is not asserted (\u2018averidical\u2019 as in Cook 2008). In line (ii), the verb kakw\u00e2tak\u00eayihtamiyit \u2018they live in torment\u2019 is negated with nam\u00f4ya \u2018not\u2019 and prefixed with the same preverb ka-. (126) \u201cOBVIATIVE\u201d SUBJECT OF PSYCH VERB AND IRREALIS MARKING i. \u201c\u146d\u14ed \u14ab\u14c2\u1450\u1424 \u1406\u14aa \u1401 \u146e \u1405\u14f0\u1426\u140b\u141f \u140a\u1528\u14ef\u14c2\u1418, \u1472 \u14a5\u152d\u141a\u1455\u14a5\u1528\u141f, \u201ckis\u00ea-manitow \u00f4m \u00ea-k\u00ee-os\u00eeh\u00e2t ayisiyiniwa, ka-miyaw\u00e2tamiyit, kis\u00ea- manitow aw -ima \u00ea- k\u00ee- os\u00ee =h -\u00e2 -t ayisiyiniw-a, gentle-god PROX-IN.SG C1-PREV-make=by.neut-DIR-3 person -XT ka- miyaw\u00e2t -am-yi -t FUT-enjoy.TI -TI -DS-3 \u2018\u201cGodPROX has created manOBV to be happy, \u2026 ii. \u14c7\u14a8\u152d \u140a\u152d \u1472 \u1472\u1481\u1455\u146b\u1528\u1426\u1455\u14a5\u1528\u141f;\u201d nam\u00f4y \u00e2ya ka-kakw\u00e2tak\u00eayihtamiyit;\u201d nam\u00f4ya aya ka- kakw\u00e2tak=\u00eayiht -am-yi-t NEG CONN FUT-suffer =by.mind.TI-TI -DS-3 \u2026not to live in torment;\u201d\u2019 (EM 1997:\u00a737) In (127), an \u201cObviative\u201d argument (ayamih\u00eawiskw\u00eawa \u2018nunsOBV\u2019) is the experiencer of a psych verb (atamih\u00e2t \u2018shePROX pleased themOBV\u2019), and this \u201cProximate\u201d third person has been established as the source of information on this state. (127) \u201cOBVIATIVE\u201d SUBJECT OF A PSYCH VERB WITH AUTHORITY OF \u201cPROXIMATE\u201d \u2026, \u140a\u152d\u14a5\u1426\u1401\u1424\u1403\u1422\u1475\u1418 \u14a6\u14c7 \u1401\u1418\u146f \u1401 \u146e \u140a\u152d \u140a\u1455\u14a5\u1426\u140b\u141f, \u2026 \u2026, ayamih\u00eawiskw\u00eawa m\u00een \u00eawako \u00ea-k\u00ee-aya-atamih\u00e2t, \u2026 ayamih\u00eaw=iskw\u00eaw-a m\u00eena \u00eawako \u00ea- k\u00ee- aya- atam=ih -\u00e2 -t pray =woman-XT also RESUM C1-PREV-conn-please=CAUS-DIR-3 \u2018\u2026, and the nunsOBV were happy with herPROX, too; \u2026\u2019 (EM 1997:\u00a764) Most often, several of these conditions are met. For example, there is evidential marking and a \u201cProximate\u201d referent provides the information. Thus, in the example in (127), Emma Minde has already explicitly framed the story as coming from her Mother-in-law, and she then carries through with this indirect knowledge by marking the clauses relating to the events described as guesses (\u00eatikw\u00ea) and reports (\u00easa) (cf. Blain et al. 2006), and keeping the experiencer of the intention as \u201cObviative\u201d (ayamih\u00eawiskw\u00eawa). This is in line with Plains Cree ideals, which hold 188 the development and maintenance of chains of information as crucial to all discourse (cf. K\u00e2- pimw\u00eaw\u00eahahk 1998, and Wolfart in K\u00e2-pimw\u00eaw\u00eahahk 1998, Chapter 6). From this closer look, we see that the cases of \u201cObviative\u201d referents being associated with psych verbs are actually not exceptions to the generalization that \u201cObviative\u201d referents are contextually extentional. In all these \u201cObviative\u201d psych cases, the Speaker does not directly know the intentional state of the \u201cObviative\u201d referent. Instead, this \u201cObviative\u201d referent\u2019s perspective is systematically distanced from the Speaker\u2019s, either through (i) indirect evidential coding, (ii) a hypothetical, unknown future context (irrealis), or (iii) an intermediate referent (authority). 3.5.32. \u201cObviative\u201d speakers and indirect knowledge When we considered obviation with verbs of speaking, we saw that \u201cObviative\u201d referents are highly restricted as speakers. In particular, intransitive verbs of speaking almost never had an \u201cObviative\u201d argument associated with them. itw\u00ea- \u2018say\u2019 \u201cANIMATE\u201d \u201cOBVIATIVE\u201d WORD COUNT Minde 17 0 12,900 Ahenakew 148 0 10,300 Whitecalf 80 0 3,800 K\u00e2-pimw\u00eaw\u00eahahk 63 1 7,400 TOTALS 308 1 34,400 Table 3.17. (repeated): Intransitive itw\u00ea- and obviation In making this generalization, there were two significant gaps: (i) a qualitative gap, since there is one intransitive \u201cObviative\u201d speaker, and (ii) a valency gap, since we did not consider transitive forms. When we turn our attention from the narrow, quantitative generalization for intransitive verbs to these two gaps, we find further evidence for the content and context properties of obviation; an \u201cObviative\u201d referent can only speak if the Speaker has not hear it. While the ban on \u201cObviative\u201ds occuring with intransitive speaking verbs is strong, it is not absolute. The only place I have ever located a textual example of this sort is in K\u00e2- pimw\u00eaw\u00eahahk\u2019s recounting of his father\u2019s account of the signing of Treaty 6 (K\u00e2-pimw\u00eaw\u00eahahk 1998:\u00a76). In this example (128), 189 (128) \u201cOBVIATIVE\u201d SUBJECT OF INTRANSITIVE SPEAKING VERB \u201c\u1401\u1426\u140a,\u201d \u1401 \u1403\u1458\u1528\u141f, \u2026 \u201c\u00eaha,\u201d \u00ea-itw\u00eayit, \u2026 \u00eaha \u00ea- it =w\u00ea-yi-t yes C1-thus=AI-DS-3 \u2018\u201cYes!\u201d heOBV [Queen\u2019s Representative] said, \u2026\u2019 (JK 1998:\u00a76.5) Here, the Speaker is reporting what his father told him about the speeches made during the signing of the treaty between the Plains Cree and the British \u2013 an event that happened before the Speaker was born. Crucially, this means that the Speaker has not witnessed this event of speaking.25 This is a second-hand narrative context, which parallels the quotative paradigm offered by one speaker (S4) to track sources of information (see section 3.5.2). (129) \u201cOBVIATIVE\u201d REFERENT IS NOT SOURCE OF INFORMATION TO SPEAKER \u1401\u146f\u14ef \u1401 \u1403\u1458\u1528\u141f (\u1401\u14f4) \u14c8\u142f\u1418. \u00eakosi \u00ea-itw\u00eayit (\u00easa) n\u00e2p\u00eawa \u00eakosi \u00ea-itw\u00ea -yi -t \u00easa n\u00e2p\u00eaw-a so C1-say-DS-3 EVID man -XT \u2018That\u2019s what the manOBV said\u2019 Here, the consultant explained that the \u201cObviative\u201d form (\u00ea-itw\u00eayit n\u00e2p\u00eawa) would be used if the Speaker did not hear the speech themselves \u2013 the same context that the intransitive, \u201cObviative\u201d speaker is appearing in the text in (128). This means that an \u201cObviative\u201d referent can be the argument of an intransitive verb of speaking, so long as the Speaker has no direct knowledge of this event. Put another way, the \u201cObviative\u201d referent can have a perspective only so long as it is not embedded immediately within the Speaker\u2019s. While intransitive speaking verbs are used when a referent speaks to whoever is present, referents may also speak to other people specifically, which calls for a transitive verb to be used. Most often, the verb used is either one composed of the relative root it- \u2018thus\u2019 as in (130), or the applicative form w\u00eehtamaw- \u2018tell about\u2019 as in (131). 25 In much of the rest of this passage, the Queen\u2019s representative is proximate and occurs with intransitive speaking verbs. This is because this other speech is embedded inside the quotation of K\u00e2-pimw\u00eaw\u00eahahk\u2019s father; e.g. \u2018\u00eakosi k\u00ee-itw\u00eaw, \u00eawakw \u00e2wa k\u00e2-k\u00ee-pimoht\u00eastamaw\u00e2t kihc-\u00f4kim\u00e2skw\u00eawa,\u201d k\u00ee-itw\u00eaw m\u00e2na k\u00e2-k\u00ee-oy\u00f4ht\u00e2wiy\u00e2n, \u2026\u2019 \u2018Thus he spoke, the one who had come representing the Queen,\u201d my late father used to say \u2026\u2019 (JK 1998:\u00a76.5) 190 (130) it- \u2018SAY TO\u2019 a. \u1403\u144c\u1424 it\u00eaw it -\u00ea -w thus-DIR-3 \u2018s\/hePROX said to him\/herOBV.\u2019 (Presented S2) b. \u2026, \u14f2\u1422\u1481\u1428 \u140a\u1528\u14ef\u1528\u14c2\u1418 \u1401 \u146e \u142f\u1426\u1455\u1418\u1420 \u14ab\u14c7 \u1401 \u146e \u1403\u1456\u141f, \u201c\u2026\u201d \u2026, s\u00f4skw\u00e2c ayisiyiniwa \u00ea-k\u00ee-p\u00eahtawak m\u00e2n \u00ea-k\u00ee-it\u00e2t, \u201c\u2026\u201d s\u00f4skw\u00e2c ayisiyiniw-a \u00ea-k\u00ee- p\u00ea =htaw -ak m\u00e2na \u00ea-k\u00ee-it-\u00e2-t simply person -XT C1-PREV-toward=by.ear.TA-1>3 usual C1-PREV-thus-DIR-3 \u2018I heard him say to anybody, \u201c\u2026\u201d\u2019 (Minde 1997:\u00a743) (131) w\u00eehtamaw- \u2018TELL ABOUT\u2019 a. \u1411\u1426\u1455\u14aa\u140d\u1424 w\u00eehtamaw\u00eaw w\u00eeht=amaw -\u00ea -w tell =APPLIC-DIR-3 \u2018s\/hePROX tells him\/herOBV about it.\u2019 (Presented S2) b. \u146e \u1411\u1426\u1455\u14aa\u140d\u1424 \u1401\u14f4 \u140a\u1528\u14ef\u1528\u14c2\u1418, \u201c\u2026\u201d k\u00ee-w\u00eehtamaw\u00eaw \u00easa ayisiyiniwa, \u201c\u2026\u201d k\u00ee- w\u00eeht=amaw-\u00ea -w \u00easa ayisiyiniw-a PREV-tell =applic -DIR-3 EVID person -XT \u2018HePROX had told the peopleOBV, \u201c\u2026\u201d\u2019 (Minde 1997:\u00a727) In comparison to the intransitive form, the transitive forms show an almost even split between \u201cAnimate\u201d and \u201cObviative\u201d speakers. The form it- \u2018say to\u2019 has a nearly even distribution of the two kinds of referents. it- \u2018say to\u2019 ANIM > OBV OBV > ANIM WORD COUNT Minde 7 6 12,900 Ahenakew 6 8 10,300 Whitecalf 1 0 3,800 K\u00e2-pimw\u00eaw\u00eahahk 4 0 7,400 TOTALS 18 14 34,400 Table 3.23. Transitive it- and obviation Mirroring this pattern, the verb w\u00eehtamaw- \u2018tell about\u2019 also shows a more even split. 191 wihtamaw- \u2018tell\u2019 ANIM > OBV OBV > ANIM WORD COUNT Minde 5 5 12,900 Ahenakew 1 1 10,300 Whitecalf 0 0 3,800 K\u00e2-pimw\u00eaw\u00eahahk 4 0 7,400 TOTALS 10 6 34,400 Table 3.24. Transitive w\u00eehtamaw- and obviation A consideration of these forms in more detail shows two related generalizations: (i) a \u201cProximate\u201d referent speaks to an \u201cObviative\u201d when the Speaker has heard the speech themselves, and (ii) an \u201cObviative\u201d referent speaks to a \u201cProximate\u201d when the Speaker only knows about it via the \u201cProximate\u201d referent\u2019s report. As the counts above showed, a \u201cProximate\u201d referent speaks to an \u201cObviative\u201d referent about half of the time in the corpora under consideration. A canonical example is seen in (132). (132) \u201cPROXIMATE\u201d SPEAKS TO \u201cOBVIATIVE\u201d IN SPEAKER\u2019S HEARING \u201c\u1401 \u142f ~ \u1401 \u142f \u1439\u1426\u1431\u152e\u1541 \u1406\u14aa \u1401\u147f \u14a6\u14c7,\u201d \u1401 \u146e \u1403\u1456\u141f \u14ab\u14c7. \u201c\u00ea-p\u00ea-~ \u00ea-p\u00ea-p\u00e2hpiy\u00e2hk \u00f4m \u00eakwa m\u00een,\u201d \u00ea-k\u00ee-it\u00e2t m\u00e2na. \u00ea- p\u00ea- p\u00e2hpi-y\u00e2hk aw -ima \u00eakwa m\u00eena \u00ea- k\u00ee- it -\u00e2 -t m\u00e2na C1-come-laugh-21PL PROX-IN.SG now also C1-PREV-thus-DIR-3 usual \u2018\u201cWe have come once again, to have a good laugh,\u201d he would say to himOBV.\u2019 (AA 2000:\u00a78.2) Here, the verb form \u00ea-k\u00ee-it\u00e2t uses the TA direct theme sign -\u00e2-, which picks up the \u201cProximate\u201d Andrew Ahenakew from previous discourse. In considering \u201cProximate\u201d speakers, a distinct pattern emerges. \u201cProximate\u201d referents speak to \u201cObviative\u201d ones when the Speaker has witnessed the speech act. Consider the data charted in here, taken from a longer span of Alice Ahenakew relating conversations. 192 LINE EVENT PROX OBV i Speaker drives over - - ii Black tells stories Black - iii Speaker calls Black a name Black - iv Black is named Black - v Andrew tells Black Andrew Black vi Sister-in-law is bad-tempered Black - vii Sister-in-law gives Black dirty looks Speaker Black viii Black is about to tell stories - Black ix Black tells stories - Black x Sister-in-law talks to Andrew Sister-in-law Andrew xi Black begins to tell stories Black - xii Black tells story Black - Table 3.25. Transitive speaking and obviation in a text Here, Alice\u2019s brother-in-law [Black] is introduced as a \u201cProximate\u201d referent in (ii). He is maintained as the sole, \u201cProximate\u201d, referent until line (v), at which time he is spoken to by Alice\u2019s husband Andrew [Andrew], which is shown in (132) above. Then, in line (vi), Alice\u2019s sister-in-law [Sister-in-law] is introduced, and heavily marked with verbs denoting intentional content, like \u00ea-m\u00e2c-\u00e2yiwit \u2018be bad tempered.\u2019 She is maintained as a \u201cProximate\u201d, while her husband [Black] is obviated throughout this short span (133). (133) \u201cBLACK\u201d IS OBVIATED WHILE SISTER-IN-LAW TALKS viii. \u2026 \u1473 \u1411 \u140b\u148b\u14a7\u1528\u141f, \u2026 \u2026 k\u00e2-w\u00ee-\u00e2cimoyit, \u2026 k\u00e2-w\u00ee- \u00e2cimo -yi-t C2-intend-tell.story-DS-3 \u2018\u2026 when heOBV was about to tell stories, \u2026\u2019 (AA 2000:\u00a78.2) When she then speaks to Andrew, the direct verb form is used, picking up her \u201cProximate\u201d status as already established (134). (134) SISTER-IN-LAW IS \u201cPROXIMATE\u201d WHEN SHE SPEAKS x. \u201c\u2026\u201d \u1401 \u146e \u1403\u1456\u141f \u14ab\u14c7, \u2026 \u201c\u2026\u201d \u00ea-k\u00ee-it\u00e2t m\u00e2na, \u2026 \u00ea-k\u00ee- it -\u00e2 -t m\u00e2na C1-PREV-thus-DIR-3 usual \u2018\u201c\u2026,\u201d she would say to himOBV, \u2026\u2019 (AA 2000:\u00a78.2) Black is then held as \u201cObviative\u201d until Alice begins reporting his story, in line (xi), as shown in (135), where the \u201cAnimate\u201d demonstrative awa is used along with \u201cAnimate\u201d verb forms. 193 (135) BLACK BECOMES \u201cANIMATE\u201d AGAIN WHEN SPEAKING xi. \u2026, \u1401\u147f \u140a\u152d, \u1401\u146f\u14ef \u1401\u147f \u14ab\u14c7 \u1401 \u146e \u14ab\u1456\u148b\u14a7\u141f \u140a\u1418, \u2026 \u2026, \u00eakwa aya, \u00eakos \u00eakwa m\u00e2n \u00ea-k\u00ee-m\u00e2t\u00e2cimot awa, \u2026 \u00eakwa aya, \u00eakosi \u00eakwa m\u00e2na \u00ea- k\u00ee- m\u00e2t= \u00e2cimo -t aw -a and CONN, so then usual C1-PREV-begin=tell.story-3 PROX-AN.SG \u2018\u2026, and then, so then, he [my late brother-in-law] would begin to tell stories.\u2019 (AA 2000:\u00a78.2) Summarizing, then, we see that, in this passage, every referent that speaks is made \u201cProximate\u201d, regardless of that referent\u2019s relation to the discourse\u2019s topic structure or to other referents. Although Alice is clearly talking mainly about her brother-in-law \u2018Black\u2019 and his outlandish stories, he is obviated whenever someone else is speaking. This results in a rapid-fire kind of \u201cProximate\u201d shifting, which would be perplexing under a topic-oriented analysis of obviation, but makes sense if \u201cProximate\u201d referents are intentional. \u201cObviative\u201d referents also frequently speak to \u201cProximate\u201ds. A canonical case is shown in (136). Here, an \u201cObviative\u201d referent (Chief Ermineskin) is speaking to a \u201cProximate\u201d referent (Dan Minde), and the verb of speaking bears the TA inverse theme sign -ikw- to signify this. (136) \u201cOBVIATIVE\u201d CHIEF ERMINESKIN SPEAKS TO \u201cPROXIMATE\u201d DAN MINDE vi. \u2026, \u1406\u14a5\u14ef \u1403\u14ef \u1401 \u146e \u1403\u144e\u146f\u141f: \u201c\u2026,\u201d \u2026 \u2026, \u00f4mis \u00ees \u00ea-k\u00ee-itikot: \u201c\u2026,\u201d \u2026 \u00f4misi isi \u00ea- k\u00ee- it- ikw-t like thus C1-PREV-thus-INV-3 \u2018\u2026 and heOBV had said thus to him: \u201c\u2026,\u201d \u2026\u2019 (EM 1997:\u00a743) When we consider the contexts in which an \u201cObviative\u201d referent speaks to a \u201cProximate\u201d, a strong generalization emerges: \u201cObviative\u201d referents speak to \u201cProximate\u201d ones only when the \u201cProximate\u201d referent being spoken to is the one that has reported this speech to the Speaker. \u201cObviative\u201d speech acts are never witnessed by the Speaker, but are instead learned about second-hand, with the \u201cProximate\u201d referent being spoken to as the source of information. An excellent example of this pattern in full operation can be found in Emma Minde\u2019s famous account of the prophecies of Chief Ermineskin. A summary looks like this: 194 LINE EVENT PROX OBV i Speaker hears Dan Minde talk Dan Minde - ii Dan Minde talks about Ermineskin Dan Minde Ermineskin iii Ermineskin counsels Dan Minde Dan Minde Ermineskin iv Ermineskin prophesies to Dan Minde Dan Minde Ermineskin v Ermineskin tells Dan Minde of future Dan Minde Ermineskin vi Ermineskin speaks to Dan Minde Dan Minde Ermineskin vii Ermineskin speaks to Dan Minde Dan Minde Ermineskin viii Speaker summarizes Ermineskin Chief People ix Speaker direct quotes Ermineskin People - x Speaker direct quotes Ermineskin People - xi Ermineskin tells things to Dan Minde Dan Minde Ermineskin x Ermineskin is named - Ermineskin Table 3.26. \u201cObviative\u201d referents speaking in a text In typical Plains Cree rhetorical style, Emma Minde first identifies the context that led to her knowledge of Ermineskin\u2019s speeches; she heard it from her Father-in-law Dan Minde (137). (137) EMMA MINDE RELATES CONTEXT OF KNOWLEDGE FOR THE SPEECH i. \u2026; \u14c2\u146e \u142f\u1426\u1455\u141a\u1424 \u14ab\u14c7 \u1401\u141a\u146f \u140a\u1418 \u14c2\u14aa\u14c8\u148b\u14ab\u1472\u1423 Dan Minde, \u2026 \u2026; nik\u00ee-p\u00eahtaw\u00e2w m\u00e2na \u00eawakw \u00e2wa niman\u00e2cim\u00e2kan Dan Minde, \u2026 ni-k\u00ee- p\u00ea=htaw -\u00e2 -w m\u00e2na \u00eawakw aw =a ni-man\u00e2cim\u00e2kan Dan Minde 1- PREV-come=by.ear.TA-DIR-3 usual resum PRX=AN.SG 1- parent.in.law Dan Minde \u2018\u2026; I used to hear my parent-in-law Dan Minde \u2026\u2019 (EM 1997:\u00a744) She then relates the connection between Dan Minde and Ermineskin, moving from the most general relation (counselling) to the most specific (an exact quote of what was said); Dan Minde had been counselled by him (iii), had received prophecies from him (iv-v), and was spoken to directly (vi). The flow of information from Ermineskin to Dan Minde to Emma is made extremely explicit. Throughout this setting process, Dan Minde is held as \u201cProximate\u201d and Ermineskin is obviated, with representative examples in (138) and (139). (138) DAN MINDE IS \u201cPROXIMATE\u201d AND ERMINESKIN IS \u201cOBVIATIVE\u201d iii. \u2026; \u140f\u1422\u1455 \u1401\u14f4 \u1401 \u146e \u1472\u146b\u1422\u146d\u14a5\u146f\u141f \u1401\u146f\u14c2 \u140a\u14c2\u1426\u1403 \u1406\u146d\u14ab\u1426\u1473\u14c7. \u2026; w\u00eest \u00eas \u00ea-k\u00ee-kak\u00easkimikot \u00eakoni anih \u00f4kim\u00e2hk\u00e2na. w-\u00eesta \u00easa \u00ea- k\u00ee- kak\u00eask=im -ikw-t \u00eakoni an =ihi okimw=ihk\u00e2n-a 3-emph EVID C1-PREV-counsel=by.mouth.TA-INV-3 resum dst=xt chief =made -XT \u2018\u2026; he, too, had been counselled by that chiefOBV.\u2019 (EM 1997:\u00a744) 195 (139) DAN MINDE IS \u201cPROXIMATE\u201d AND ERMINESKIN IS \u201cOBVIATIVE\u201d v. \u2026, \u1401 \u1401 \u146e \u1411\u1426\u1455\u146f\u141f \u1401\u14f4 \u14c2\u152e\u1420 \u140a\u152d, \u1401 \u141a\u1438\u1426\u1455\u14a5\u1528\u141f, \u2026 \u2026, \u00ea-~ \u00ea-k\u00ee-w\u00eehtam\u00e2kot \u00easa niy\u00e2k aya, \u00ea-k\u00ee-w\u00e2pahtamiyit, \u2026 \u00ea- k\u00ee- w\u00eeht=amaw-ikw-t \u00easa niy\u00e2k aya, \u00ea- k\u00ee- w\u00e2p=aht -am-yi-t C1-PREV-tell =applic-INV-3 EVID future CONN, C1-PREV-see=by.eye.TI-TI -DS-3 \u2018\u2026, he had been told by himOBV about the future, heOBV had seen it \u2026\u2019 (EM 1997:\u00a744) Throughout the reporting of Ermineskin\u2019s speech, Ermineskin is \u201cObviative\u201d and Dan Minde is \u201cProximate.\u201d At the very end, Emma re-identifies who it is that uttered these words \u2013 Chief Ermineskin. In doing this, Ermineskin remains \u201cObviative\u201d (140). (140) ERMINESKIN IS \u201cOBVIATIVE\u201d AT END OF DISCOURSE xii. \u2026, \u1472\u152e\u1422 \u140a\u14c2\u1426\u1403 \u1406\u146d\u14ab\u1426\u1473\u14c7 \u140a\u152d, \u2018\u1473 \u1406\u14ef\u1426\u146f\u14ef\u1418\u152e\u14c2\u1424\u2019 \u1473 \u146e \u1403\u14ef\u1529\u1426\u1473\u14f1\u1528\u141f. \u2026, kay\u00e2s anih \u00f4kim\u00e2hk\u00e2na aya, \u2018k-\u00f4sihkosiway\u00e2niw\u2019 k\u00e2-k\u00ee-isiy\u00eehk\u00e2soyit. kay\u00e2s an =ihi okim\u00e2hk\u00e2n -a aya k\u00e2-o=sihkosi=way\u00e2n=i-w long.ago DST=XT chief=made-XT conn C2-3=ermine=skin =AI-3 k\u00e2-k\u00ee- isi =\u00eehkaw=isw -yi-t C2-PREV-thus=make=REFLX-DS-3 \u2018\u2026 by that chiefOBV of long ago, that k-\u00f4sihkosiway\u00e2niw as heOBV was called.\u2019 (EM 1997:\u00a744) Thus, the \u201cObviative\u201d referent (Ermineskin) is talking to the \u201cProximate\u201d referent (Dan Minde) and the Speaker (Emma) has not witnessed the event. Summarizing, then, the referent that has told the Speaker about this event (Dan Minde) is \u201cProximate\u201d throughout the passage, while the referent who is doing the speaking (Ermineskin) is \u201cObviative.\u201d The referent speaking is \u201cObviative\u201d, then, because the Speaker has only indirect knowledge of their speech. We can test these generalizations about speaking verbs in elicitation settings. Let us consider the following context as an example. I am talking to you on my cell phone, and a woman is standing nearby. She turns to me and says that a man, standing out of my sight on the other side of a hill, is looking at her. I now tell you this on the phone. In such a situation, iskw\u00eaw \u2018woman\u2019 must be kept in the \u201cProximate\u201d for the sentence to be felicitous (141a,b). 196 (141) CONTEXT: WOMAN TELLS SPEAKER THAT A MAN IS LOOKING AT HER a. WOMAN IS \u201cPROXIMATE\u201d MAN IS \u201cOBVIATIVE\u201d \u1401 \u141a\u1438\u14a5\u146f\u141f \u1403\u1422\u1475\u1424 \u14c8\u142f\u1418. \u00ea-w\u00e2pamikot iskw\u00eaw n\u00e2p\u00eawa. \u00ea-w\u00e2p =am -ikw-t iskw\u00eaw n\u00e2p\u00eaw-a C1-see=by.eye.TA-INV -3 woman man -XT \u2018The womanPROX is seen by the manOBV\u2019 (Presented to S2) b. WOMAN IS \u201cOBVIATIVE\u201d MAN IS \u201cPROXIMATE\u201d # \u1401 \u141a\u1438\u14ab\u141f \u1403\u1422\u1475\u1418 \u14c8\u142f\u1424. # \u00ea-w\u00e2pam\u00e2t iskw\u00eawa n\u00e2p\u00eaw \u00ea- w\u00e2p=am -\u00e2 -t iskw\u00eaw-a n\u00e2p\u00eaw C1-see =by.eye.TA-DIR-3 woman-XT man \u2018The manPROX saw the womanOBV.\u2019 (Presented to S2) Comment (S2): \u201cYou\u2019re getting your information from the man in this one.\u201d The speaker explained the felicity of (141a) by saying, \u201cThe woman could be saying, \u2018the man is seeing me\u2019. [\u2026] The woman is telling you about it - you\u2019re hearing what she has to say about it. It\u2019s pointing at the person, not who the person is. Just pointing at the person doing the speaking.\u201d The utterance in (141b) was dispreferred because, \u201cyou\u2019re getting your information from the man in this one.\u201d Thus, in a pair of sentences contrasting only in obviation, the choice always conveys who is providing the information. In the context of the Speaker\u2019s direct experience, \u201cProximate\u201ds are speakers, but \u201cObviative\u201ds are not. When we consider all of the instances of third-person speech interactions, a qualitatively solid generalization emerges: \u201cProximate\u201d referents speak to \u201cObviative\u201d ones when the Speaker has witnessed the event, while \u201cObviative\u201d referents speak to \u201cProximate\u201d ones when the Speaker has only heard about this speech act from the \u201cProximate\u201d referent. In terms of the current model, we can understand this as a prohibition on direct embeddings for \u201cObviative\u201d perspectives; the \u201cObviative\u201d perspective cannot be embedded directly inside the Speaker\u2019s, but instead must be embedded within some intermediate perspective. 3.5.33. \u201cObviative\u201d referents and the indirect evidential \u00easa There are short spans in developed Plains Cree discourse where evidential forms like \u00easa \u2018reportedly\u2019 are marked on nearly every clause. For example, in a short story related by Alice Ahenakew about an old couple escaping starvation, there are approximately 48 clauses (not counting quotations) and 31 instances of \u00easa \u2018reportedly.\u2019 Sometimes, there are two occurrences of \u00easa marked on a single clause as in (142). 197 (142) TWO INSTANCES OF \u00easa IN ONE CLAUSE \u2026 - \u14ab\u1472 \u1439\u1422\u146d\u14ef\u1472 \u140f\u152d \u1401\u14f4 \u1401 \u146e \u14c7\u1472\u1455\u14ab\u1426\u141f \u1401\u14f4. \u2026 - m\u00e2ka p\u00e2skisikan wiy \u00easa \u00ea-k\u00ee-nakatam\u00e2ht \u00easa. m\u00e2ka p\u00e2skisikan wiya \u00easa \u00ea-k\u00ee-nakat=amaw=iht \u00easa but gun 3 EVID C1-PREV-leave.for-IMP-3 EVID \u2018\u2026 - but he had been left with a gun.\u2019 (AA 2000:\u00a79.8) Because of both its restriction to a small domain and saturation with evidentials, Blain et al. (2006) have called these \u2018evidential spans.\u2019 While evidential spans may occur for a variety of reasons, such as the entire story being a second-hand account as was the case in (142), one place that these spans consistently occur is when an \u201cObviative\u201d referent possesses a perspective. Thus, in Emma Minde\u2019s recounting of the words of Chief Ermineskin, every single clause that codes Ermineskin\u2019s speech act is marked with \u00easa \u2018reportedly,\u2019 and no other clauses are. LINE EVENT PROX OBV EVID i Speaker hears Dan Minde talk Dan Minde - Ind ii Dan Minde talks about Ermineskin Dan Minde Ermineskin Conj, iii Ermineskin counsels Dan Minde Dan Minde Ermineskin Conj, \u00easa iv Ermineskin prophesies to Dan Minde Dan Minde Ermineskin Conj, \u00easa v Ermineskin tells Dan Minde of future Dan Minde Ermineskin Conj, \u00easa vi Ermineskin speaks to Dan Minde Dan Minde Ermineskin Conj, \u00easa vii Ermineskin speaks to Dan Minde Dan Minde Ermineskin Conj, \u00easa viii Speaker summarizes Ermineskin Chief People Conj, Irreal ix Speaker direct quotes Ermineskin People - Conj, Irreal x Speaker direct quotes Ermineskin People - Conj, Irreal xi Ermineskin tells things to Dan Minde Dan Minde Ermineskin Conj, \u00easa x Ermineskin is named - Ermineskin Conj Table 3.27. Obviatiation in an \u00easa evidential span For example, in line (iv), when Chief Ermineskin prophesies to Dan Minde, the main verb is marked with \u00easa (143). (143) \u201cOBVIATIVE\u201d ERMINESKIN SPEAKS, \u00easa IS USED iv. \u1401 \u146e \u146d\u1422\u146d\u140d\u1426\u1405\u146f\u141f \u1401\u14f4, \u1401\u1473 \u1489\u1422\u147f \u146e\u147f\u1429 \u1401~ \u1401 \u1456\u140f\u14c7\u1541, \u2026 \u00ea-k\u00ee-kiskiw\u00eahokot \u00easa, \u00eak\u00e2 c\u00easkwa k\u00eekway \u00ea-~ \u00ea-t\u00e2winahk, \u2026 \u00ea-k\u00ee- kisk=w\u00ea=hw-ikw-t \u00easa \u00eak\u00e2 c\u00easkwa k\u00eekway \u00ea- t\u00e2win- am-k C1-PREV-know=AI=TA-INV-3 EVID NEG moment something C1-encounter-TI -0 \u2018He had had prophesies uttered by himOBV, about things which he had not yet encountered, \u2026\u2019 (EM 1997:\u00a744) 198 The only sentences not being marked with evidentials are ones embedded in a quotation (ix-x) or the sentences having to do with Emma\u2019s own understanding, for example lines (viii) or (x). This correlation between evidentials like \u00easa and \u201cObviative\u201d referents is freely identified by speakers, even when entirely unprompted. (144) \u201cOBVIATIVE\u201d COUNSELLOR, \u00easa IS PREFERRED \u1401 \u1472\u146b\u1422\u146d\u14a5\u146f\u141f (\u1401\u14f4) \u00ea-kak\u00easkimikot (\u00easa) \u00ea-kak\u00eask=im -ikw-t C1-counsel=by.mouth.TA-INV-3 \u2018heobv counsels himPROX\u2019 (Volunteered by S4) Comment (S4): \u201cIt would be better if you say it with \u00easa.\u201d Here, the Speaker is explaining culturally-appropriate ways to describe someone\u2019s counselling activies. In so doing, the Speaker says, \u201cYou would say \u00ea-kak\u00easkimikot if someone was getting their counselling out of a book. It would be better if you say it with \u00easa.\u201d26 This preference for the addition of \u00easa to \u201cObviative\u201d sources of information is consistent across all speakers I have worked with. Whenever information comes from an \u201cObviative\u201d referent, the Speaker marks the event as only indirectly known. Within the current analysis, this is expected, since \u201cObviative\u201d referents are those referents whose intentions (e.g. speaking) the Speaker has no direct knowledge of. Evidentiality and Obviation are then marking the same kind of Speaker knowledge, only over different domains; the one over the Speaker\u2019s relation to a proposition (Evidentiality), the other over the Speaker\u2019s relation to the (non)perspective of the referent (Obviation). 3.6. Conclusion In this chapter, I have developed an analysis of obviation phenomena in Plains Cree that models its three facets: (i) FORM: \u201cObviation\u201d is a construct (\u00a73.3). It is built by recruiting compatible pieces in compatible parts of the grammar. 26 Because of lack of overt nominals in (S4)\u2019s example to disambiguate between \u201cObviative\u201d and \u201cInanimate\u201d reference (the two are identical in the inverse; see Wolfart (1973)), and the complications surrounding books as evidential sources of information (i.e. are the books a source in themselves, or only the transmission of someone\u2019s words? (cf. Aikhenvald 2004)), this speaker could either be referring to the \u201cInanimate\u201d masinahikan \u2018book\u2019 or to the unidentified \u201cObviative\u201d writer of the book. 199 (ii) CONTENT: An \u201cObviative\u201d referent is one that is contextually extentional (\u00a73.4). It is a referent that lacks a perspective for the contextually-determined span. (iii) CONTEXT: Obviation situates a referent within a chain of information (\u00a73.5). The \u201cObviative\u201d referent\u2019s perspective is inaccessible to the Speaker. By constructing a model that covers obviation\u2019s form, content, and context altogether, we can begin, for the first time, to systematically consider all obviation phenomena together. In doing this, we can begin to ask questions about the relation of obviation to other components of Plains Cree grammar \u2013 a topic that I take up in the next chapter. 200 Chapter 4 Referential dependency in Plains Cree 4.1. A structural and a semantic condition on referential dependency Human language systematically exploits context to inform content (cf. Fillmore 1976, Bateson 1975). One of the most productive ways this is done is through the constructions of dependencies \u2013 the process of making one element the context for the interpretation of another (Safir 2004). Of the rich taxonomy of dependencies found in natural language, the current discussion has centered on only one kind: referential dependency. These are dependencies in which one referent is used as the context for the interpretation of another. This kind of dependency has been recognized by many researchers in the case of possession (e.g. Higginbotham 1983, Kuno 1988, Barker 1995, Burton 1995, Partee and Borschev 2003, and M\u00fchlbauer 2007, etc.), but the current thesis, which focusses on referents and perspective, shows referential dependence to be a broader property of Plains Cree. In this chapter, I claim that referential dependency is constructional in Plains Cree. In particular, I claim that the construal of a dependency between two referents must meet two conditions in Plains Cree: (i) A STRUCTURAL CONDITION: either linear precedence or c-command (ii) A SEMANTIC CONDITION: perspectival embedding. Together, these two components form the necessary and sufficient conditions for the construal of a referent as dependent in Plains Cree. The structural condition on referential dependency is fulfilled through one of two possible structures: (i) C-COMMAND: The dependent referent must be c-commanded by its antecedent. (1) V y V V = X DEPENDENT ON Y x V 201 (ii) LINEAR PRECEDENCE: The dependent referent must linearly follow its antecedent. (2) PRED1 PRED2 5 5 = X DEPENDENT ON Y y x In the logic of most theories of syntactic structure (e.g. Ross 1966, Langacker 1969, Lakoff 1971, Williams 1997), one structure does not exclude the other. For example, Carden (1986) argues that clause-internal relations are subject to c-command, while clause-external relations are also subject to linear precedence. The semantic condition on referential dependency is fulfilled through perspectival embedding; the dependent referent must be embedded within the perspective of the referent it is dependent on (3). y (3) x PREDICATE(x) This perspectival embedding defines the content of the relation between the two referents: the dependent referent is interpreted within the perspective of its antecedent; all access to x is mediated by the perspectival domain of y. Referential dependency in Plains Cree, then, is the convergence of structural and semantic operations for the purpose of constructing a context for a referent to be interpreted in. 202 STRUCTURAL CONDITION SEMANTIC CONDITION 1 V y V V x V 2 PRED1 \uf0e0 PRED2 5 5 y x I am therefore making a claim about the necessary and sufficient conditions for the construction of referential dependency in Plains Cree. If referential dependency is defined as the use of one referent as the context for the interpretation of another, then obviation is a kind of dependent reference. As we saw in Chapter 3, obviation inherently invokes a valuation with respect to some other referent in the discourse. This is seen in the formulation of the \u201cObviative\u2019s\u201d content (\u00a73.4): the \u201cObviative\u201d is disjoint from any referent that possesses a perspective, which entails that the \u201cObviative\u201d must be evaluated with respect to all perspective-possessing referents (4). (4) EXT(x,C) \u2194 \u2200\u03c8\u2200y(R(y,\u03c8,C) \uf0e0 x \u2260 y) x is extentional at context C if and only if for all perspectives \u03c8 and all individuals y, if there is a relation R of y with \u03c8 at context C, then x is not y. This is also seen in the formulation of the \u201cObviative\u2019s\u201d context (\u00a73.5): the \u201cObviative\u201d referent is embedded within the perspective of a \u201cProximate\u201d referent, which entails that the \u201cObviative\u201d is dependent on the truth-evaluation domain (perspective) of the \u201cProximate\u201d. \u201cProximate\u201d (5) Obv PREDICATE(Obv) PREDICATE(OBV) = TRUE W.R.T. PROX\u2019S PERSPECTIVE 203 The \u201cObviative\u201d referent, whose content is informed by the contextual availability of a \u201cProximate\u201d referent (3.4, 3.5), is thus one instance of the broader category of referential dependency in Plains Cree. Since obviation is a kind of referential dependency, the model of referential dependency developed here predicts that the structural and semantic conditions on referential dependency should be evident in all constructions involving obviation in Plains Cree. Concretely, this means that obviative constructions should show sensitivity to (i) a structural condition of either linear- precedence or c-command, and (ii) a semantic condition of perspectival embedding. Obviation constructions, then, provide an excellent testing ground for the logic of referential dependency proposed here. In section 4.3, I consider predicate-to-predicate relations and obviation. In \u00a74.3.1, I consider the ordering of \u201cObviative\u201d referring nominals with respect to \u201cProximate\u201ds, showing that the \u201cObviative\u201d must always succeed the \u201cProximate\u201d. In \u00a74.3.2, I show that the suffix -yi-, which is used to build \u201cObviative\u201d reference (\u00a73.3.5), generally codes referential dependency between the subject of the predicate and some structurally-preceding referent. Predicate- predicate relations used to construct obviation, then, demonstrate sensitivity to both the structural and semantic conditions on referential dependency. In section 4.4, I consider predicate-internal relations and obviation, which are typically subordinative (cf. Hirose 2000, D\u00e9chaine 2003). In \u00a74.4.1, I consider nominal-internal dependencies, which is typically called possession. I show that Plains Cree possession generally codes referential dependency between the possessor and the possessum. In \u00a74.4.2, I consider verb-internal dependencies, which are typically coded by theme signs in Plains Cree (cf. Wolfart 1973, etc.). I show that, for at least two of the set of Plains Cree\u2019s theme signs, both conditions on referential dependency hold independent of obviation. Predicate-internal elements that are used to construct obviation, then, also demonstrate sensitivity to both the structural and semantic conditions on referential dependency. From this discussion, we see that Plains Cree systematically exploits dependency- building operations to construct obviation. Further, we see that, in most cases (\u00a74.3.2, 4.4.1-2), the forms used to build \u201cObviative\u201d reference have the properties of referential dependency even in constructions where they do not code obviation. 204 Structural Conditions Semantic Conditions \u00a7 Construction C-command Precedence Perspectival Referential Dependency 4.3.1 N \u2013 N \u2718 \u2714 \u2714 \u2714 4.3.2 Pred \u2013 Pred-yi \u2714 \u2714 \u2714 \u2714 4.4.1 Possession \u2714 \u2718 \u2714 \u2714 4.4.21 -\u00ea- \u2714 \u2718 \u2714 \u2714 4.4.31 -ikw- \u2714 \u2714 \u2714 \u2714 Table 4.1. \u201cObviative\u201d constructions meet the conditions of referential dependency Since obviation is both a construct (\u00a73.3) and dependent (3.4-5), this kind of grammatical organization is expected. 4.2. Previous work The current proposal and the resulting data generalizations are themselves dependent on much previous work, both in Algonquian linguistics in specific and linguistics in general. For Algonquian linguistics, the previous work of Wolfart (1973) and Dahlstrom (1986) on the basic generalizations of Plains Cree morphosyntax are relevant, as well as the more recent work by Blain (1997), Cook (2007, 2008), D\u00e9chaine (1997, 2003), Hirose (2000), and Reinholtz (D\u00e9chaine & Reinholtz 1997, 2008). For general linguistic work, the current work relies most heavily on the syntactic work of Carden (1986), McCawley (1988), and Williams (1997). Some of this work is best considered when the appropriate domain of inquiry comes up (e.g. the direct\/inverse system in \u00a74.4.2), but other relevant discussion is better considered beforehand, to situate the discussion properly. 4.2.1. Structural conditions on anaphora The literature is divided on the kinds of structural relations available to natural language (cf. McCawley 1988). For our current purposes, the relevant distinction is between those who posit only one kind of structural relation (i.e. c-command), and those who posit multiple kinds of structural relations (e.g. precedence, c-command, command). The position that only one kind of structural relation is available to natural language was promoted most strongly by Reinhart (1983). Considering the resolution of anaphoric reference in 205 English sentences, Reinhart argued that all cases of anaphoric dependencies can be described in terms of one structural relation: c(onstituent)-command. This accounts, she argued, for the ungrammaticality of English sentences like those in (6) (judgments are those presented in Reinhart 1983).1 (6) a. \u273d The actressi whoi Brando kissed in heri latest film will win the Oscar. b. \u273d What actressi did Brando kiss in heri latest film? b. \u273d Whoi did the police arrest in spite of hisi alibi? c. \u273d The guyi whoi the police arrested in spite of hisi alibi has filed a complaint. Here, the referent that is supposed to be the antecedent of the pronoun linearly precedes the pronoun, but the structure is ruled ill-formed. Reinhart claims that the pronoun is in an adjunct, rather than in an argument position in these cases, and thus the antecedent does not c-command the pronoun. Based on data of this kind, Reinhart (1983) argues that the relevant structural configuration for the licensing of anaphora in English is c-command, not any kind of linear precedence. This position has been developed in much subsequent work, in particular that of Kayne (1994), who uses c-command as a fundamental component of the logic of linear-ordering relations. For linguists that adopt this position, all linear-precedence relations can be explained in terms of c-command. Any linear-order effects that are not amenable to a c-command analysis are typically judged to be part of discourse considerations, rather than part of syntax (see Van Hoek 1997 for discussion). There have been several dissenters from Reinhart\u2019s (1983) position. Numerous linguists that worked on the structure of English previous to Reinhart (1983), for example Langacker (1969) and Ross (1967) argue that the linear-precedence relation is relevant to the resolution of anaphora. They consider cases such as those in (7), arguing that pronominalization of the second instance of John in (7b) was a result of a restriction on the linear occurence of a pronoun with respect to its antecedent: a pronoun may only precede its antecedent if it is in a subordinate clause. (7) a. Realizing that hei was unpopular didn\u2019t bother Johni. b. \u273d Realizing that Johni was unpopular didn\u2019t bother himi. Beyond this, Carden (1986) and McCawley (1988) have pointed out that the c-command accounts, while yielding strong results in many areas, fail to account for some of the original data set considered for anaphora in English. For example, sentences like those in (8) are not covered 1 All of these sentences are perfectly well-formed for me, and for the other native speakers I have asked. 206 by c-command formalizations without significant ad hoc alterations (cf. McCawley 1988:345). Instead, they argued, pronominal relations are not all equal, and some must be calculated at different levels of derivation (cf. Carden 1986) (8) a. After Maryi had finished the report, shei went home. b. Near the car that Johni was repairing, hei saw a snake. c. Near himi is where Johni saw the snake. d. It was near himi that Johni saw the snake. (Carden 1986) Following this debate, Williams (1997) has argued that the English data Reinhart (1983) used to argue for c-command English must itself be reconsidered. He points out that the sentences Reinhart (1983) presents as ungrammatical are actually grammatical for many English speakers (Williams 1993:236), and argues that both linear precedence and c-command are available structural relations for natural language. This means that, in some form or other, linear precedence is still a part of modern syntactic theory. In formulating the structural conditions on referential dependency in Plains Cree, I have made reference to both structural dominance (i.e. c-command) and linear-precedence relations. As such, I have committed the analysis to a framework like that of Carden (1986) or Williams (1997), rather than that of Reinhart (1983) or Kayne (1994). In addition, this kind of division in structural relations opens the possibility that different kinds of anaphoric relations are calculated with different mechanisms, as was proposed for English by Carden (1986). As we will see, there are good reasons to suppose that this position is correct, at least for Plains Cree; the language is sensitive to both linear precedence and structural dominance, but not equally in all cases. 4.2.2. Clausal dependency and referential dependency Cook (2008) argues that the clause system of Plains Cree has at least two kinds of structural relations: (i) relations defined by structural dominance (i.e. c-command; Reinhart 1983), and (ii) relations defined by precedence (i.e. linear precedence; Ross 1967). In Plains Cree, there is a set of predicates that introduce a structurally subordinate predicate (Cook 2008). This can be shown by their special syntactic properties (see Cook 2008): (i) A quantifier that has scope over the argument of the subordinate clause can occur to the left of the matrix clause (9) but not other kinds of clausal relations. 207 (9) QUANTIFIER OF ARGUMENT OF SECOND CLAUSE APPEARS TO LEFT OF FIRST CLAUSE \u1472\u1426\u146d\u152d\u1424 \u14c2\u146d\u1422\u146b\u1528\u1426\u144c\u1423 \u1401 \u140a\u1426\u146f\u14ef\u148b\u1420 \u14c8\u142f\u1418\u1420. kahkiyaw nikisk\u00eayiht\u00ean \u00ea-ahkosicik n\u00e2p\u00eawak kahkiyaw ni-kisk=\u00eayiht -\u00ea -n \u00ea-ahkosi-t-ik n\u00e2p\u00eaw-ak all 1- know=by.mind.TI-TI-LP C1-sick -3-PL man -PL \u2018I know that all the men were sick.\u2019 (Cook 2008) (ii) A nominal relating to an argument of the subordinate clause can occur to the left of the matrix clause (10) but not other kinds of clausal relations. (10) NOMINAL ARGUMENT OF SECOND CLAUSE APPEARS TO LEFT OF FIRST CLAUSE \u14c8\u142f\u1418\u1420 \u14c2\u146d\u1422\u146b\u1528\u1426\u144c\u1423 \u1401\u14c3\u14a5\u1426\u1403\u1450\u148b\u1420. n\u00e2p\u00eawak nikisk\u00eayiht\u00ean \u00ea-n\u00eemihitocik. n\u00e2p\u00eaw-ak ni-kisk =\u00eayiht -\u00ea -n \u00ea-n\u00eemihito-t-ik man -PL 1- know=by.mind.TI-TI-LP C1-dance -3-PL \u2018I know the men were dancing.\u2019 (Cook 2008) (iii) A question word relating to an argument of the subordinate clause can occur to the left of the matrix clause (11) but not other kinds of clausal relations. (11) QUESTION WORD ARGUMENT OF SECOND CLAUSE APPEARS TO LEFT OF FIRST CLAUSE \u140a\u140f\u14c7 \u1401 \u1403\u140d\u152d\u1423 \u1401 \u1403\u144c\u1528\u1426\u1455\u14aa\u1423 John \u1401 \u1405\u1489\u14ab\u141f? aw\u00eena \u00ea-itw\u00eayan \u00ea-it\u00eayihtaman John \u00ea-oc\u00eam\u00e2t? aw\u00eena \u00ea- it =w\u00ea-yan \u00ea- it =\u00eayiht -am-an John \u00ea- oc\u00ea=m -\u00e2 -t who C1-thus=AI-2 C1-thus=by.mind.TI-TI -2 John C1-kiss=by.mouth.TA-DIR-3 \u2018Who did you say you think John kissed?\u2019 (Cook 2008) Based on data of this kind, Cook (2008) concludes that the relation between the two predicates is one of structural dominance; the main predicate has both linear precedence and c-command over the second predicate. (12) CP 5 VP 3 VP CPi 2 5 V DPi DEPENDENT (Cook 2008) 208 Cook (2008) also shows that there are other kinds of clausal relations in Plains Cree that are more complex than a simple matrix-subordinate distinction can model. Typically, a Plains Cree speaker will begin a discourse with an independent mode verb, and then introduce a series of verbs, each in one of several modes of the conjunct order, as the example in (13) shows. (13) i. \u1401\u147f \u14c0\u1413\u14f5\u144a \u14c2\u1422\u1450\u14f5\u144a \u146e\u14ef\u1473\u1424 \u14c2\u146e \u1438\u1439\u14ab\u148b\u1426\u1405\u14c8\u1423 \u1401\u146f\u144c, [Independent-Order] \u00eakwa n\u00eawos\u00e2p-~ nistos\u00e2p-k\u00eesik\u00e2w nik\u00ee-pap\u00e2m\u00e2cihon\u00e2n \u00eakot\u00ea, \u00eakwa n\u00eawo=s\u00e2p-~ nisto=s\u00e2p-k\u00eesik\u00e2 -w ni-k\u00ee- pap\u00e2m =m\u00e2ci=ho -n\u00e2n \u00eakot\u00ea and 4 =tens 3 =tens-be.day-3 1-PREV-go.about=be =MID-1pl there \u201cThen we toured about over there for fourteen -~ for thirteen days,\u201d ii. a tour \u1401 \u1405\u144e\u14c7\u14ab\u1541 \u1405\u144e, [\u00ea-conjunct] a tour \u00ea-otinam\u00e2hk oti, a tour \u00ea-ot =in -am-\u00e2hk oti a tour C1-take=by.hand-TI -1pl in.fact \u201cwe took a tour,\u201d iii. bus \u1401 \u1434\u14ef\u152e\u1541, \u140a\u152d, [\u00ea-conjunct] bus \u00ea-p\u00f4siy\u00e2hk, aya, bus \u00ea- p\u00f4si-y\u00e2hk aya bus c1-ride-1pl HES \u201cwe travelled on a bus,\u201d iv. thirteen-day tour \u1401 \u1405\u144e\u14c7\u14ab\u1541, [\u00ea-conjunct] thirteen-day tour \u00ea-otinam\u00e2hk, thirteen-day tour \u00ea- ot =in -am-\u00e2hk thirteen-day tour C1-take=by.hand-TI -1PL \u201cwe took a thirteen-day tour \u2026\u201d v. thirty-six \u1401 \u1403\u1426\u1455\u14ef\u152e\u1541, [\u00ea-conjunct] thirty-six \u00ea-ihtasiy\u00e2hk, thirty-six \u00ea- iht =asi -y\u00e2hk thirty-six c1-thus=number-1PL \u201cwith thirty-six people \u2026\u201d vi. bus \u140a\u14c7 \u1401 \u1434\u14ef\u152e\u1541. [\u00ea-conjunct] bus an \u00ea-p\u00f4siy\u00e2hk. bus an-a \u00ea- p\u00f4si-y\u00e2hk bus DST-AN.SG C1-ride-1PL \u201ctravelling on the bus.\u201d (Cook 2008) As Cook (2008) shows, these subsequent conjunct order verbs are dependent on the initial independent mode verb. They are dependent temporally on it, such that temporal coding of the initial clause carries through to the uncoded clause (14). 209 (14) INITIAL CLAUSE\u2019S TEMPORAL CODING CONTINUES FOR FOLLOWING CLAUSES \u1401 \u146e \u142f \u1403\u1451\u1426\u144c\u148b\u1420 \u14ab\u14c7 \u140a\u141a\u14ef\u14f4\u1420 \u1401 \u14c2\u1472\u14a7\u148b\u1420. \u00ea-k\u00ee-p\u00ea-it\u00f4ht\u00eacik m\u00e2na aw\u00e2sisak \u00ea-nikamocik. \u00ea- k\u00ee- p\u00ea- it =\u00f4ht\u00ea -t-ik m\u00e2na aw\u00e2sis-ak \u00ea-nikamo-t-ik C1-PREV-come-thus=by.walking.AI-3-PL usual child -PL C1-sing -3-PL \u2018The children used to come and they used to sing.\u2019 (Cook 2008) However, this relation cannot be a straightforward case of c-command, since it shares none of the properties associated with c-command seen above (9-11); here, the second clause\u2019s quantifiers (15a), nominal arguments (15b), and wh-words (15c) cannot occur in front of the initial clause. (15) a. QUANTIFIERS OF SECOND CLAUSE CANNOT OCCUR IN FRONT OF INTIAL CLAUSE \u273d \u1472\u1426\u146d\u152d\u1424 \u14c8\u142f\u14ef\u1422 \u1475\u14ef\u14a8\u148b\u146d\u1426\u1456\u1424 \u14a6\u14c2\u14f4 \u1401 \u14a6\u146d\u14f1\u141f. \u273d kahkiyaw n\u00e2p\u00easis kw\u00easim\u00f4cikiht\u00e2w m\u00eenisa \u00ea-m\u00eecisot. kahkiyaw n\u00e2p\u00ea=sis kw\u00easi =m\u00f4cik=iht\u00e2 -w m\u00eenis-a \u00ea-m\u00eeciso-t all man=DIM repeat=be.fun =make-3 berry-XT C1-eat -3 Intended: \u2018The boy had fun eating all the berries.\u2019 (Cook 2008) b. ARGUMENTS OF SECOND CLAUSE CANNOT OCCUR IN FRONT OF INTIAL CLAUSE \u273d \u14a6\u14c2\u14f4 \u14c8\u142f\u14ef\u1422 \u1475\u14ef\u14a8\u148b\u146d\u1426\u1456\u1424 \u1401 \u14a6\u146d\u14f1\u141f. \u273d m\u00eenisa n\u00e2p\u00easis kw\u00easim\u00f4cikiht\u00e2w \u00ea-m\u00eecisot m\u00eenis-a n\u00e2p\u00ea=sis kw\u00easi =m\u00f4cik=iht\u00e2 -w \u00ea-m\u00eeciso-t berry-XT man=DIM repeat=be.fun =make-3 C1-eat -3 Intended: \u2018The boy had fun eating all the berries.\u2019 (Cook 2008) c. QUESTION WORDS OF SECOND CLAUSE CANNOT OCCUR IN FRONT OF INTIAL CLAUSE \u273d \u146e\u1481\u1429 \u14c8\u142f\u14ef\u1422 \u1475\u14ef\u14a8\u148b\u146d\u1426\u1456\u1424 \u1401 \u14a6\u146d\u14f1\u141f? \u273d k\u00eekw\u00e2y n\u00e2p\u00easis kw\u00easim\u00f4cikiht\u00e2w \u00ea-m\u00eecisot? k\u00eekw\u00e2y n\u00e2p\u00ea=sis kw\u00easi =m\u00f4cik=iht\u00e2 -w \u00ea-m\u00eeciso-t what man=DIM repeat=be.fun =make-3 C1-eat -3 Intended: \u2018What did the child have fun, they ate __ ?\u2019 (Cook 2008) To model this, Cook (2008) uses Williams\u2019 (1997) model of binding effects with anaphora; the antecedent may be either linearly precedent or structurally dominant. These clause-chaining examples, then, are examples of linear precedence without structural dominance, leading Cook (2008) to build structural configurations as in (16). 210 (16) CP1 \/ i 6 CPii CPiii CPiv CPvi nik\u00ee-pap\u00e2m\u00e2cihon\u00e2n 5 5 5 5 \u00ea-otinam\u00e2hk \u00ea-p\u00f4siy\u00e2hk \u00ea-otinam\u00e2hk CPv \u00ea-p\u00f4siy\u00e2hk 5 \u00ea-ihtasiy\u00e2hk Here, the linearly initial clause introduces a set of dependent clauses which rely on it for referential information. Comparing this model of clausal dependency to the one posited here for referential dependency, the parallels become immediately apparent; in both cases, the dependency is constructed by either structural dominance or linear precedence. The current model of referential dependency, then, extends Cook (2008) from clauses to referents. 4.2.3. Possession and grammatical obviation It has long been known that obviation marking is obligatory in third-person possession contexts (cf. Bloomfield 1962, Wolfart 1973, Dahlstrom 1986, etc.). (17) THIRD-PERSON POSSESSUMS MUST BE \u201cOBVIATIVE\u201d a. \u1405\u14a5\u14c5\u14ef\u14aa omin\u00f4sima o-min\u00f4s-im-a 3-cat -DJ -XT \u2018his\/herPROX catOBV\u2019 (Presented S2) b. \u273d \u1405\u14a5\u14c5\u14ef\u1466 \u273d omin\u00f4sim o-min\u00f4s-im 3-cat-DSJ \u2018his\/her cat\u2019 (Presented S2) Here, a nominal form that lacks the suffix \u2013a (\u00a73.3.1) is rejected. It should be noted that speakers do produce forms without obviation marking in elicitation environments (cf. Cook & M\u00fchlbauer 2006), but these \u201cObviative\u201d less possessor constructions disappear as context improves. This fact has often been used to divide obviation into two cases (cf. Rhodes 1976, Dahlstorm 1986, D\u00e9chaine & Wiltschko 2002): 211 (i) \u201cGRAMMATICAL\u201d OBVIATION: Characterized by obligatory marking and local relations (i.e. possession, but also verb relations in some accounts) (ii) \u201cDISCOURSE\u201d OBVIATION: characterized by optionality and non-local relations. These two contexts of obviation have led to a development of a dichotomy in the literature, such that linguists can choose to work on \u201cgrammatical\u201d obviation, and leave \u201cdiscourse\u201d obviation for some future research (cf. Rhodes 1976, etc.). Thus, obviation is two things, not one thing, under such a view; it is a grammatical operation or a discourse operation, but the relation between the two is not clear. Notice that this account presupposes some crucial things about possession in order to make its case for the bifurcation of obviation. First, it is necessary to assume that possession relations are not optional, discourse operations \u2013 that is, the Speaker has no choice but to employ a possessive based on the clause-level properties of the grammar. Were possession shown to serve a particular set of discourse functions, and be itself an optional process, there would be no reason to divide obviation in two. The current work offers a way to unify these two domains of obviation once again. In section 4.3.1, I show that possession has discourse properties in Plains Cree, making the division between \u201cgrammatical\u201d and \u201cdiscourse\u201d possession difficult to maintain. Further, I show that these discourse properties have to do with the semantic dependency of the possessed referent, which means that the structure and semantics of possession line up with obviation; both are kinds of referential dependency. Thus, there is no division in obviation; obviation is one thing, and it works in concert with the discourse properties of possession. 4.2.4. The data set While the general problems considered in this chapter have long been considered for human language, the data from Plains Cree is less well-known, and much of it has never been documented before, even within Plains Cree linguistics. In particular, the following eleven data sets have never, to my knowledge, been documented for Plains Cree: (i) The linear sensitivity of \u201cProximate\u201d and \u201cObviative\u201d referring nominals (\u00a74.3.1) (ii) The inability of \u201cObviative\u201d marked nominals to occur in front of the matrix clause where other nominals can (\u00a74.3.12) (iii) The use of pausing to repair dispreferred \u201cObviative\u201d > \u201cProximate\u201d nominal orderings (\u00a74.3.12) 212 (iv) The structural conditions on the suffix \u2013yi\u2013 ; the antecedent of \u2013yi\u2013 must be in a clause that linearly precedes or structurally dominates \u2013yi\u2013 (\u00a74.3.21) (v) The semantic conditions on the referent that \u2013yi\u2013 is dependent on (i.e. that it must be existential and perspective-possessing (\u00a74.3.22) (vi) The ill-formedness of \u201cInanimate\u201d possessors (\u00a74.4.121) (vii) The semantic difference between compounding and possession (\u00a74.4.122) (viii) The use of possessor forms to manipulate direct vs. indirect speech (\u00a74.4.124) (ix) The interaction of possession with requirements on Speaker knowledge (\u00a74.4.125) (x) The linear dependency of independent order inverse forms that involve only third- persons (\u00a74.4.222) (xi) The relation between inverse forms and familiarity (\u00a74.4.223) These additions to the available data set on these phenomena mean that, even should another scholar consider my theoretical model to be undesireable, they will still have a much richer data set to work with in the future. 4.3. Cross-predicate dependencies and obviation \u201cObviative\u201d constructions can involve relations between two predicates. In the most simple cases, this means noun-noun relations, as exemplified by (18). Here, one nominal predicate, atim \u2018dog,\u2019 relates to another nominal predicate, min\u00f4sa \u2018catOBV.\u2019 (18) \u1456\u1426\u146f\u14a3\u1424 \u140a\u144e\u14a7 \u14a5\u14c5\u14f4. t\u00e2hkom\u00eaw atim min\u00f4sa. t\u00e2hkw=m -\u00ea -w atimw min\u00f4s-a seize=by.mouth-DIR-3 dog cat -XT \u2018the dogPROX bit the catOBV.\u2019 (Presented S2) More complex cases include the relation between the subject of one predicate and some other predicate, as exemplified by the case of \u2013yi\u2013 (19). In (19a), the second verbal predicate has \u2013yi\u2013 attached, and codes that its subject is not the previous verb\u2019s subject (Jeff). In (19b), the absence of \u2013yi\u2013 allows the subject of the second verb to be construed as co-referent with the previous subject. 213 (19) a. Jeff \u1473 \u14a5\u1422\u1472\u141a\u141f \u140a\u144e\u14b7 \u1401 \u14c5\u1426\u144c\u1426\u1472\u144c\u1528\u141f. Jeff k\u00e2-miskaw\u00e2t atimwa \u00ea-n\u00f4ht\u00eahkat\u00eayit. Jeff k\u00e2-m =iskaw -\u00e2 -t atimw-a \u00ea-n\u00f4ht\u00eahkat\u00ea-yi-t Jeff c2-find=by.body.TA-DIR-3 dog -XT C1-be.hungry-DS-3 \u2018When JeffPROX found the dogOBV, itOBV was hungry.\u2019 (Presented S2) b. Jeff \u1473 \u14a5\u1422\u1472\u141a\u141f \u140a\u144e\u14b7 \u1401 \u14c5\u1426\u144c\u1472\u144c\u1528\u141f. Jeff k\u00e2-miskaw\u00e2t atimwa \u00ea-n\u00f4ht\u00eakat\u00eat. Jeff k\u00e2-m =iskaw -\u00e2 -t atimw-a \u00ea-n\u00f4ht\u00eahkat\u00ea-t Jeff c2-find=by.body.TA-DIR-3 dog -XT C1-be.hungry-3 \u2018When JeffPROX found the dogOBV, hePROX was hungry.\u2019 (Presented S2) In each of these cases, we can ask whether or not the two conditions on referential dependency hold: (i) one referent must be perspectivally embedded inside of the other referent, and (ii) the perspectivally-embedded referent must linearly succeed or be c-commanded by its antecedent. In the following sections, I demonstrate that each of these predicate-predicate relations is sensitive to both the structural and semantic conditions of referential dependency (\u00a74). In section 4.3.1, I show these effects in the relation of nominals to eachother; \u201cProximate\u201d referring nominals must have linear precedence over obviative-referring ones in order for the \u201cObviative\u2019s\u201d\u201d dependency to be correctly constructed. In section 4.3.2, I show these effects for the predicate-suffix \u2013yi\u2013. This suffix meets the structural conditions of dependent reference; the predicate bearing \u2013yi\u2013 must either (i) be c-commanded by a predicate containing the antecedent (\u00a74.3.21), or (ii) linearly succeed it (\u00a74.3.22). This suffix also meets the semantic condition on dependent reference; the referent associated with \u2013yi- must be perspectivally embedded under this preceding referent. In all of these cases, the elements that are used to construct the \u201cObviative\u201d are used to build referential dependency; all elements have both the structural and semantic conditions of referential dependency (\u00a74). The referential dependency of the \u201cObviative\u201d, then, correlates with the referential dependency of its parts. This is expected on the view of obviation\u2019s form considered in section 3.2, which models the \u201cObviative\u201d as constructed out of non-dedicated forms. 4.3.1. Nominal ordering and referential dependency Since nominal expressions can be used to refer to \u201cObviative\u201d referents (\u00a73.4), we expect them to play a role in the construction of \u201cObviative\u2019s\u201d referential dependency. In particular, we 214 expect that nominals that refer to \u201cObviative\u201d referents should necessarily linearly-succeed nominals that refer to \u201cProximate\u201d referents. (20) NPPROX NPOBV V V YPROX V XOBV V N N We expect this behaviour because the \u201cProximate\u201d \/ \u201dObviative\u201d nominal relations already satisfy the semantic condition on obviation (embedding of the \u201cObviative\u201d in the \u201cProximate\u2019s\u201d perspective: \u00a73.5). Showing this kind of linear-precedence dependency, then, nominal-nominal ordering with obviation would satisfy both conditions of referential dependency, as modelled above (\u00a74.2). These expectations are confirmed in the data; \u201cProximate\u201d referring nominals must always precede \u201cObviative\u201d ones. First, in section 4.3.11, I show that \u201cProximate\u201d referring nominals always precede \u201cObviative\u201d ones in a discourse. Then, in section 4.3.12, I show that \u201cProximate\u201d referring nominals precede \u201cObviative\u201d ones in elicitation contexts. From this evidence, I conclude that \u201cProximate\u201d referring nominals must have linear precedence over the \u201cObviative\u201d referring nominals. From this discussion, we see that nominal-nominal ordering meets both conditions for referential dependency in Plains Cree. First, \u201cObviative\u201d referring nominals meet the structural condition on referential dependence (c-command or linear precedence:\u00a74.2); \u201cObviative\u201d referring nominals must linearly succeed \u201cProximate\u201d nominals. Second, \u201cObviative\u201d referring nominals meet the semantic condition on referentical dependence (perspectival-embedding: \u00a74.2); they pick out a referent that is perspectivally-embedded within the \u201cProximate\u2019s\u201d perspective (cf. 3.5). Structural Conditions Semantic Condition C-command Linear Precedence Perspectival Embedding N-N \u2718 \u2714 \u2714 Referential Dependency \u2714 \u2714 \u2714 Table 4.2. Nominal-nominal ordering meets the conditions of referential dependency This is what is expected under a model of referential dependency like that constructed here. 215 4.3.1.1. Nominal orderings in texts In considering the relative ordering of nominals in Plains Cree texts, it is important to note immediately that it is extremely rare for two nominals to occur in the same clause in a developed discourse (cf. Ahenakew 1987, Wolfart & Reinholtz 1993). One of the only clear cases I have found is shown in (21), which has a \u201cProximate\u201d referring nominal ayisiyiniwak \u2018people\u2019 preceding an \u201cObviative\u201d referring nominal ayisyiniwa \u2018people\/person.\u2019 (21) TWO NOMINALS IN THE SAME CLAUSE \u140a\u1528\u14ef\u1528\u14c2\u1418\u1420 \u1401 \u1529\u1472\u144c\u1422\u1455\u141a\u148b\u1420 \u1401\u146f\u14c2 \u140a\u1528\u14ef\u1528\u14c2\u1418; ayisiyiniwak \u00ea-y\u00eekat\u00eastaw\u00e2cik \u00eakoni ayisiyiniwa; ayisiyiniw-ak \u00ea- y\u00eekat\u00eastaw-\u00e2 -t-ik \u00eakoni ayisiyiniw-a person -PL C1-stay.away -DIR-3-PL resum person -XT \u2018peoplePROX stay away from that personOBV\u2019 (Minde 1997:\u00a738) Elicitation examples of double nominals are thus synthetic in the sense that they do not represent natural speech, although they are perfectly comprehensible.2 To look for nominal-nominal orderings in texts, then, we have to consider orderings in domains larger than clauses. In developed discourses, the first referent introduced is always \u201cProximate.\u201d A typical discourse is considered here (Ahenakew 2000:\u00a711.2-3), and is shown clause-by-clause in the following table. Here, the first nominal used refers to the \u201cProximate\u201d referent, and an \u201cObviative\u201d referring nominal is not introduced until clause iv. 2 The same should be said for English; I have spoken the language my entire life and have hardly ever heard someone utter a sentence like \u201cJohn saw Mary.\u201d Its use in linguistic data, then, also represents a synthetic, partial abstraction away from natural speech, which is always embedded in a context. 216 Clause Event Prox Obv Overt N i. Speaker remembers Old Woman - - Old Woman ii. S forgets Old Woman\u2019s name - - - iii. Old Woman comes driving - - - iv. Old Woman camps with Grand-daughter Old Woman Grand- daughter Grand- daughter vi. Old Woman brings Grand- daughter with Old Woman Grand- daughter Grand- daughter vii. Old Woman & Grand-daughter drive buggy Old Woman, Grand- daughter - viii. Grand-daughter beautiful Grand-daughter - Grand- daughter ix. Man desires Grand-daughter Man Grand- daughter Man, Grand- daughter x. Man flirts with Grand-daughter Man Grand- daughter - xi. Man scores Man - - Table 4.3. \u201cProximate\u201d referring nouns precede \u201cobviatives\u201d in texts When a new \u201cProximate-Obviative\u201d relation is coded (clause ix), overt nominals for both referents are used, with the new \u201cProximate\u201d oskin\u00eekiw \u2018young man\u2019 occurring as a topicalized nominal to the left of the clause. As always, the \u201cProximate\u201d referring nominal linearly precedes the \u201cObviative\u201d referring one (22). (22) \u201cPROXIMATE\u201d NOMINAL PRECEDES \u201cOBVIATIVE\u201d NOMINAL IN DISCOURSE \u140a\u147f \u140a\u1418 \u14c8\u142f\u1424, \u1418\u1426\u141a, \u142f\u152d\u1420 \u1405\u1421\u146d\u14c3\u146d\u1424, \u140a\u1473\u141a\u144c\u1424 \u1406\u1426\u1403 \u1405\u1422\u146d\u14c3\u146d\u1422\u1475\u1418, \u2026 \u00eakwa awa n\u00e2p\u00eaw, wahw\u00e2, p\u00eayak oskin\u00eekiw, ak\u00e2w\u00e2t\u00eaw \u00f4hi oskin\u00eekiskw\u00eawa, \u2026 \u00eakwa aw =a n\u00e2p\u00eaw wahw\u00e2 p\u00eayak oskin\u00eekiw ak\u00e2w\u00e2t-\u00ea -w and PRX=AN.SG man EXCL one young.one lust -DIR-3 aw =ihi oskin\u00eek=iskw\u00eaw-a PRX=XT young=woman-XT \u2018and a certain man, oh my, a certain young manPROX desired this young womanOBV, \u2026\u2019 (Ahenakew 11:2) This is a typical structure for a Cree discourse. So typical, in fact, that I have gone through many discourses looking for an instance of an \u201cObviative\u201d referring nominal being introduced before a \u201cProximate\u201d, but have found none. The fact that discourses always start with \u201cProximate\u201d forms has the practical result that \u201cProximate\u201d nominals always precede \u201cObviative\u201d forms. This parallels the few cases of double- nominal clauses that we have just seen, lending further support to the conclusion that \u201cProximate\u201ds precede \u201cObviative\u201ds in natural speech. 217 4.3.1.2. Elicitation and nominal-nominal ordering To understand what elicitation data on nominal-nominal ordering is telling us, we need to keep a crucial observation in mind: consultants will often accept orderings as comprehensible that they themselves have never produced, and cannot produce when asked. For example, when asked about the pair of sentences in (23), which differ in their ordering of the \u201cProximate\u201d referring nominal atim \u2018dog\u2019 and the \u201cObviative\u201d referring nominal min\u00f4sa \u2018catOBV,\u2019 the consultant initially responded that both forms were \u201cperfectly understandable.\u201d (23) a. \u140a\u144e\u14a7 \u1455\u1426\u146f\u14a3\u1424 \u14a5\u14c5\u14f4. [ProxN>ObvN] atim tahkom\u00eaw min\u00f4sa. atimw tahkw=m -\u00ea -w min\u00f4s-a dog seize=by.mouth-DIR-3 cat -XT \u2018The dogPROX bit the catOBV.\u2019 (Presented S2) b. \u14a5\u14c5\u14f4 \u1455\u1426\u146f\u14a3\u1424 \u140a\u144e\u14a7. [ObvN>ProxN] min\u00f4sa tahkom\u00eaw atim. min\u00f4s-a tahkw=m -\u00ea -w atimw cat -XT seize =by.mouth-DIR-3 dog \u2018The dogPROX bit the catOBV.\u2019 (Presented S2) However, there are several reasons to think that these sentences are not equally good: (i) Consultants say that they would not personally say (23b). (ii) Consultants say that they have not heard people say (23b). (iii) Consultants often have to repeat (23b) to themselves a few times before offering a grammaticality judgment on it. (iv) Consultants repeat (23b) as (23a). This kind of data means that it is necessary to make a distinction between (i) what consultants accept as comprehensible, and (ii) what consultants identify that they, themselves, would say. When we do this, we get a three-way split in data: (i) Orderings that consultants rule uninterpretable (\u273d) (ii) Orderings that are interpretable but consultants cannot themselves produce (?) (iii) Orderings that are both interpretable and able to be produced by the consultant (\u2714) 218 Cross-classifying orderings with these three variables, we see that only orderings where the \u201cProximate\u201d nominal precedes the \u201cObviative\u201d are both interpretable and produced by the consultant. This is summarized in table 4.4. PROX > OBV OBV > PROX V N N \u2714 ? N V N \u2714 ? N N V \u2714 ? N V [N V] \u2714 \u273d N V [V N] \u2714 \u273d Table 4.4. Nominal-nominal ordering patterns in elicitation All orderings are acceptable that have the \u201cProximate\u201d referring form precede the \u201cObviative\u201d. The relevant set is shown in (24). (24) a. \u1455\u1426\u146f\u14a3\u1424 \u140a\u144e\u14a7 \u14a5\u14c5\u14f4. [V > PROX > OBV ] tahkom\u00eaw atim min\u00f4sa. tahkw=m -\u00ea -w atimw min\u00f4s-a seize=by.mouth-DIR-3 dog cat -XT \u2018The dogPROX bit the catOBV.\u2019 (Presented S2) b. \u140a\u144e\u14a7 \u1455\u1426\u146f\u14a3\u1424 \u14a5\u14c5\u14f4. [PROX > V > OBV] atim tahkom\u00eaw min\u00f4sa. atimw tahkw=m -\u00ea -w min\u00f4s-a dog seize=by.mouth-DIR-3 cat -XT \u2018The dogPROX bit the catOBV.\u2019 (Presented S2) c. \u140a\u144e\u14a7 \u14a5\u14c5\u14f4 \u1455\u1426\u146f\u14a3\u1424. [PROX > OBV > V] atim min\u00f4sa tahkom\u00eaw. atimw min\u00f4s-a tahkw=m -\u00ea -w dog cat -XT seize=by.mouth-DIR-3 \u2018The dogPROX bit the catOBV.\u2019 (Presented S2) d. \u140a\u144e\u14a7 \u14c2\u146d\u1422\u146b\u1528\u1426\u144c\u1423 \u14a5\u14c5\u14f4 \u1401 \u1455\u1426\u146f\u14ab\u141f. [PROX > V [OBV > V]] atim nikisk\u00eayiht\u00ean min\u00f4sa \u00ea-tahkom\u00e2t. atimw ni-kisk =\u00eayiht -\u00ea -n min\u00f4s-a \u00ea- tahkw=m -\u00e2 -t dog 1- know=by.mind.TI-TI-LP cat -XT C1-seize =by.mouth-DIR-3 \u2018I know that the dogPROX bit the catOBV.\u2019 (Presented S2) e. \u140a\u144e\u14a7 \u14c2\u146d\u1422\u146b\u1528\u1426\u144c\u1423 \u1401 \u1455\u1426\u146f\u14ab\u141f \u14a5\u14c5\u14f4. [PROX > V [V > OBV]] atim nikisk\u00eayiht\u00ean \u00ea-tahkom\u00e2t min\u00f4sa atimw ni-kisk =\u00eayiht -\u00ea -n \u00ea- tahkw=m -\u00e2 -t min\u00f4s-a dog 1- know=by.mind.TI-TI-LP C1-seize =by.mouth-DIR-3 cat -XT \u2018I know that the dogPROX bit the catOBV.\u2019 (Presented S2) 219 Here, all word ordering permutations are well-formed, so long as the \u201cProximate\u201d referring nominal (here atimw \u2018dog\u2019) precedes the \u201cObviative\u201d one (here min\u00f4sa \u2018cat(s)OBV\u2019). In contrast to the \u201cProximate\u201d preceding \u201cObviative\u201d orderings, forms that have the \u201cObviative\u201d referring nominal preceding the \u201cProximate\u201d are dispreferred. In these cases (25a-c), consultants will respond that the forms are understandable, but cannot be uttered by the consultant. Further, consultants also cannot imagine a fluent speaker uttering them. (25) a. ? \u1455\u1426\u146f\u14a3\u1424 \u14a5\u14c5\u14f4 \u140a\u144e\u14a7. [V > OBV > PROX ] ? tahkom\u00eaw min\u00f4sa atim. tahkw=m -\u00ea -w min\u00f4s-a atimw seize=by.mouth-DIR-3 cat -XT dog \u2018The dogPROX bit the catOBV.\u2019 (Presented S2) b. ? \u14a5\u14c5\u14f4 \u1455\u1426\u146f\u14a3\u1424 \u140a\u144e\u14a7. [OBV > V > PROX] ? min\u00f4sa tahkom\u00eaw atim. min\u00f4s-a tahkw=m -\u00ea -w atimw cat -XT seize =by.mouth-DIR-3 dog \u2018The dogPROX bit the catOBV.\u2019 (Presented S2) c. ? \u140a\u144e\u14a7 \u14a5\u14c5\u14f4 \u1455\u1426\u146f\u14a3\u1424. [OBV > PROX > V] ? min\u00f4sa atim tahkom\u00eaw. min\u00f4s-a atimw tahkw=m -\u00ea -w cat -XT dog seize =by.mouth-DIR-3 \u2018The dogPROX bit the catOBV.\u2019 (Presented S2) It is ungrammatical for the \u201cObviative\u201d referring nominal argument of a subordinate clause to be positioned in front of the matrix clause. In (26), the \u201cObviative\u201d referring nominal min\u00f4sa \u2018catOBV\u2019 is an argument of the subordinate verb \u00ea-tahkom\u00e2t \u2018itPROX bit itOBV,\u2019 but is positioned in front of the matrix verb nikisk\u00eayiht\u00ean \u2018I know itIN.\u2019 This is ruled ungrammatical. (26) \u201cOBVIATIVE\u201d REFERRING NOMINALS CANNOT OCCUR IN FRONT OF MATRIX VERB a. \u273d \u14a5\u14c5\u14f4 \u14c2\u146d\u1422\u146b\u1528\u1426\u144c\u1423 \u140a\u144e\u14a7 \u1401 \u1455\u1426\u146f\u14ab\u141f. [OBV > V [PROX > V]] \u273d min\u00f4sa nikisk\u00eayiht\u00ean atim \u00ea-tahkom\u00e2t. min\u00f4s-a ni-kisk =\u00eayiht -\u00ea -n atimw \u00ea- tahkw=m -\u00e2 -t cat -XT 1- know=by.mind.TI-TI-LP dog C1-seize =by.mouth-DIR-3 Intended: \u2018I know that the dogPROX bit the catOBV.\u2019 (Presented S2) b. \u273d \u14a5\u14c5\u14f4 \u14c2\u146d\u1422\u146b\u1528\u1426\u144c\u1423 \u1401 \u1455\u1426\u146f\u14ab\u141f \u140a\u144e\u14a7. [OBV > V [V > PROX]] \u273d min\u00f4sa nikisk\u00eayiht\u00ean \u00ea-tahkom\u00e2t atim min\u00f4s-a ni-kisk =\u00eayiht -\u00ea -n \u00ea- tahkw=m -\u00e2 -t atimw cat -XT 1- know=by.mind.TI-TI-LP C1-seize =by.mouth-DIR-3 dog Intended: \u2018I know that the dogPROX bit the catOBV.\u2019 (Presented S2) 220 By contrast, \u201cProximate\u201d referring nominal arguments of subordinate verbs can occur in front of matrix verb. (27) \u201cPROXIMATE\u201d REFERRING NOMINALS CAN OCCUR IN FRONT OF MATRIX VERB a. \u140a\u144e\u14a7 \u14c2\u146d\u1422\u146b\u1528\u1426\u144c\u1423 \u14a5\u14c5\u14f4 \u1401 \u1455\u1426\u146f\u14ab\u141f. [PROX > V [OBV > V]] atim nikisk\u00eayiht\u00ean min\u00f4sa \u00ea-tahkom\u00e2t. atimw ni-kisk =\u00eayiht -\u00ea -n min\u00f4s-a \u00ea- tahkw=m -\u00e2 -t dog 1- know=by.mind.TI-TI-LP cat -XT C1-seize =by.mouth-DIR-3 \u2018I know that the dogPROX bit the catOBV.\u2019 (Presented S2) b. \u140a\u144e\u14a7 \u14c2\u146d\u1422\u146b\u1528\u1426\u144c\u1423 \u1401 \u1455\u1426\u146f\u14ab\u141f \u14a5\u14c5\u14f4. [PROX > V [V > OBV]] atim nikisk\u00eayiht\u00ean \u00ea-tahkom\u00e2t min\u00f4sa. atimw ni-kisk =\u00eayiht -\u00ea -n \u00ea-tahkw=m -\u00e2 -t min\u00f4s-a dog 1- know=by.mind.TI-TI-LP C1-seize =by.mouth-DIR-3 cat -XT \u2018I know that the dogPROX bit the catOBV.\u2019 (Presented S2) We can conclude, then, that there is a strong asymmetry, then, between orderings where the \u201cProximate\u201d referring nominal is first, and those where the \u201cObviative\u201d is first. Another variable in these orderings is the theme sign system; the use of an inverse verb form (shown by the verbal suffix \u2013ikw\u2013 on the verb) significantly degrades judgments, particularly for orderings where the \u201cObviative\u201d is in front of the verb. For example, the form in (28a), which has the verb tahkomik \u2018itPROX was bitten by itOBV\u2019 in initial position and the \u201cObviative\u201d nominal preceding the \u201cProximate,\u201d is interpretable but unnatural. Any form with the \u201cObviative\u201d nominal in initial position (28b-c) is ruled completely uninterpretable. (28) a. ? \u1455\u1426\u146f\u14a3\u1424 \u140a\u144e\u14b7 \u14a5\u14c5\u1422. [V > OBV > PROX ] ? tahkomik atimwa min\u00f4s. tahkw=m -\u00ea -w atimw-a atimw seize=by.mouth-DIR-3 dog -XT cat \u2018The dogOBV bit the catPROX.\u2019 (Presented S2) b. \u273d \u140a\u144e\u14b7 \u1455\u1426\u146f\u14a3\u1424 \u14a5\u14c5\u1422. [OBV > V > PROX] \u273d atimwa tahkomik min\u00f4s. atimw-a tahkw=m -\u00ea -w min\u00f4s dog -XT seize =by.mouth-DIR-3 cat Intended: \u2018The dogOBV bit the catPROX.\u2019 (Presented S2) 221 c. \u273d \u140a\u144e\u14b7 \u14a5\u14c5\u1422 \u1455\u1426\u146f\u14a3\u1424. [OBV > PROX > V] \u273d atimwa min\u00f4s tahkomik. atimw-a min\u00f4s tahkw=m -\u00ea -w dog -XT cat seize =by.mouth-DIR-3 Intended: \u2018The dogOBV bit the catPROX.\u2019 (Presented S2) Likely, this degradation of the ability of consultants to parse these forms has to do with the extra contextual demands that the inverse form places on the hearer (see \u00a74.4.22 below). The ungrammatical constructions that have an \u201cObviative\u201d referring nominal preceding a \u201cProximate\u201d can be repaired prosodically. If there is a pause placed between the \u201cObviative\u201d referring nominal and the rest of the clause, speakers find the examples easier to parse, and even natural-sounding. 3 For example, the form in (29a), which has the \u201cObviative\u201d referring nominal atimwa \u2018dogOBV\u2019 in initial position, is ungrammatical, but the form in (29b), where there is a pause after this same nominal, is considered natural-sounding. (29) a. \u273d \u140a\u144e\u14b7 \u1455\u1426\u146f\u14a3\u1424 \u14a5\u14c5\u1422. [OBV > V > PROX] \u273d atimwa tahkomik min\u00f4s. atimw-a tahkw=m -\u00ea -w min\u00f4s dog -XT seize =by.mouth-DIR-3 cat Intended: \u2018The dogOBV bit the catPROX.\u2019 (Presented S2) b. \u140a\u144e\u14b7, \u1455\u1426\u146f\u14a3\u1424 \u14a5\u14c5\u1422. [OBV pause V > PROX] atimwa, tahkomik min\u00f4s. atimw-a tahkw=m -\u00ea -w min\u00f4s dog -XT seize =by.mouth-DIR-3 cat \u2018The dogOBV, it bit the catPROX.\u2019 (Presented S2) Pausing can also repair the cases where the \u201cObviative\u201d nominal precedes the matrix clause, as in (30). For example, the form in (30a), which does not have a pause after the initial \u201cObviative\u201d nominal min\u00f4sa \u2018the catOBV,\u2019 is bad, but the form in (30b) which has a pause is good. (30) a. \u273d \u14a5\u14c5\u14f4 \u14c2\u146d\u1422\u146b\u1528\u1426\u144c\u1423 \u140a\u144e\u14a7 \u1401 \u1455\u1426\u146f\u14ab\u141f. [OBV > V [V > PROX]] \u273d min\u00f4sa nikisk\u00eayiht\u00ean atim \u00ea-tahkom\u00e2t. min\u00f4s-a ni-kisk =\u00eayiht -\u00ea -n atimw \u00ea- tahkw=m -\u00e2 -t cat -XT 1- know=by.mind.TI-TI-LP dog C1-seize =by.mouth-DIR-3 Intended: \u2018I know that the dogPROX bit the catOBV.\u2019 (Presented S2) 3 This pausing strategy appears to be a kind of topicalization. As such, it requires previous context to be felicitous. 222 b. \u14a5\u14c5\u14f4, \u14c2\u146d\u1422\u146b\u1528\u1426\u144c\u1423 \u140a\u144e\u14a7 \u1401 \u1455\u1426\u146f\u14ab\u141f. [OBV pause V [V > PROX]] min\u00f4sa, nikisk\u00eayiht\u00ean atim \u00ea-tahkom\u00e2t. min\u00f4s-a ni-kisk =\u00eayiht -\u00ea -n atimw \u00ea- tahkw=m -\u00e2 -t cat -XT 1- know=by.mind.TI-TI-LP dog C1-seize =by.mouth-DIR-3 \u2018The catOBV, I know that the dogPROX bit itOBV.\u2019 (Presented S2) Understanding why this pausing helps improve these orderings requires a better understanding of the role of prosodic information in the syntax of Plains Cree than is currently possible (see Cook 2006 for discussion). This linear restriction that \u201cProximate\u201d referring nominals need to precede \u201cObviatives\u201d can be exploited in context-poor sentences to determine thematic roles without any obviation marking on the nominal (i.e. the suffix \u2013a; 3.3.1 or the demonstrative suffix \u2013ihi; 3.3.2). In (31a), the first nominal is consistently interpreted as the subject of the verb, and the second nominal is consistently interpreted as the object. (31) a. \u14c8\u142f\u1424 \u14f5\u146d\u1426\u1401\u1424 \u1403\u1422\u1475\u1424. n\u00e2p\u00eaw s\u00e2kih\u00eaw iskw\u00eaw. [N V N = S V O] n\u00e2p\u00eaw s\u00e2k =ih -\u00ea -w iskw\u00eaw man love=by.neut-DIR-3 woman \u2018That man loves the woman.\u2019 (Volunteered S2) \u2260 \u2018The woman loves that man.\u2019 b. \u1403\u1422\u1475\u1424 \u14f5\u146d\u1426\u1401\u1424 \u14c8\u142f\u1424. iskw\u00eaw s\u00e2kih\u00eaw n\u00e2p\u00eaw. [N V N = S V O] iskw\u00eaw s\u00e2k =ih -\u00ea -w man woman love=by.neut-DIR-3 man \u2018The woman loves that man.\u2019 (Volunteered S2) \u2260 \u2018That man loves the woman.\u2019 Nominal-level obviation marking is only preferred when the ordering of the two arguments is reversed. For example, in (32a), the intended interpretation, that the first nominal iskw\u00eaw \u2018woman\u2019 is the \u201cObviative\u201d one, is unavailable. In (32b), this interpretation is gotten by adding the suffix \u2013a to iskw\u00eaw \u2018woman,\u2019 and a pause is again preferred (32b). 223 (32) a. \u1403\u1422\u1475\u1424 \u140a\u14c7 \u14c8\u142f\u1424 \u14f5\u146d\u1426\u1401\u1424. N N V \u2260 O S V iskw\u00eaw n\u00e2p\u00eaw s\u00e2kih\u00eaw iskw\u00eaw an -a n\u00e2p\u00eaw s\u00e2k =ih -\u00ea -w woman DST-AN.SG man love=by.neut-DIR-3 \u2260 \u2018The man loves the woman\u2019 (Presented S2) b. \u1403\u1422\u1475\u1418, \u140a\u14c7 \u14c8\u142f\u1424 \u14f5\u146d\u1426\u1401\u1424. NOBV N V = O S V iskw\u00eawa, n\u00e2p\u00eaw s\u00e2kih\u00eaw iskw\u00eaw-a n\u00e2p\u00eaw s\u00e2k =ih -\u00ea -w woman-XT man love=by.neut-DIR-3 \u2018The woman, the man loves.\u2019 (Presented S2) Thus, the unmarked ordering appears to be that \u201cProximate\u201ds linearly precede \u201cObviative\u201ds. 4.3.2. The suffix -yi- constructs referential dependency In section 3.3.5, I considered the suffix -yi-, which has traditionally been taken to code \u201cObviative\u201d arguments of predicates. For example, an \u201cAnimate\u201d intransitive subject does not need \u2013yi\u2013 (33a), while an \u201cObviative\u201d intransitive subject requires it (33b). (33) \u201cOBVIATIVE\u201d SUBJECTS REQUIRE \u2013yi\u2013 a. \u1401 \u14a5\u152a\u14ef\u141f \u140a\u14c7 \u14c8\u142f\u1424. \u2026 \u00ea-miyosit ana n\u00e2p\u00eaw. \u00ea-miyw=si -t an =a n\u00e2p\u00eaw c1-good=AI-3 DST=AN.SG man \u2018\u2026thatAN man is goodAN.\u2019 (Presented S2) b. \u1401 \u14a5\u152a\u14ef\u1528\u141f \u140a\u14c2\u1426\u1403 \u14c8\u142f\u1418. \u2026 \u00ea-miyosiyit anihi n\u00e2p\u00eawa. \u00ea-miyw=si -yi -t an =ihi n\u00e2p\u00eaw-a c1-good=AI-DS-3 DST=XT man-XT \u2018\u2026 thisOBV manOBV is goodAN.\u2019 \u2018\u2026 thoseOBV menOBV are goodAN.\u2019 (Presented S2) In that discussion (\u00a73.3.5), I showed that this suffix cannot be considered a dedicated \u201cObviative\u201d morpheme because it occurs with \u201cInanimate\u201d as well as \u201cObviative\u201d \u201cAnimate\u201d referents (34). 224 (34) THE SUFFIX \u2013yi\u2013 OCCURS WITH \u201cINANIMATE\u201d SUBJECTS i. \u14a3\u1450\u14c2 \u1401 \u14c2\u1438\u1426\u1403 \u14a5\u14f5\u1528\u1420 \u1406\u14aa mug \u1411\u1455\u1431\u1426\u1455\u1466, \u2026 m\u00eaton \u00ea-nipahi-mis\u00e2yik \u00f4ma mug w\u00eetapihtam, m\u00eatoni \u00ea- nipahi- mis=\u00e2 -yi-k aw=ima mug w\u00eet =api=ht -am very C1-to.death-big =STAT-DS-0 PRX=IN.SG mug with=sit =by.neut.TI-TI \u2018[he was] sitting before an enormously big mugIN, \u2026\u2019 ii. \u14a5\u1450\u14c2 \u1401\u146f\u1455 \u1401 \u140a\u14ef\u1418\u144c\u1528\u1420 \u140a\u14c2\u14aa yellow \u1401 \u1403\u14ef\u14c8\u147f\u1541. miton \u00eakota \u00ea-asiwat\u00eayik anima yellow \u00ea-isin\u00e2kwahk. mitoni \u00eakw=ta \u00ea-asiwa=t\u00ea-yi -k an=ima yellow \u00ea- isi =n\u00e2kw =an-k very deic=her c1-fill=II -DS-0 prx=in.sg yellow c1-thus=appear=II -0 \u2018[itIN was] filled to the brim with something that looked yellow.\u2019 (Ahenakew 2000 3:4) In \u00a73.3.5, I proposed an analysis of \u2013yi\u2013 which treated it as an IP-level coder of disjoint subjects. (35) CP 3 C IP 3 -yi- VP 3 ROOT 3 V VP 3 V (Adapted from D\u00e9chaine 2003) While this models the internal (i.e. word-level) structural properties of this form, I did not consider its external (i.e. clause-level) structural properties. To do this, it is necessary to consider both the content and context of this suffix \u2013 a task I take up here. In this section, I construct a model of -yi- that extends its coverage beyond its clause- internal properties to include its clause-external properties. Since the suffix \u2013yi- is used in \u201cObviative\u201d constructions, we expect its context of use to show both the structural and semantic conditions necessary for the construction of dependent reference. In section 4.3.21, I consider the structural context of \u2013yi\u2013. I argue that it can only occur when its predicate is in one of two structural configurations: either (i) it is structurally-dominated by another predicate (4.3.211), or (ii) linearly preceded by another predicate (4.3.212). These structural conditions match the conditions on referential dependency exactly (cf. \u00a74). I then turn to the semantic context of \u2013yi\u2013, and show that correlated to this structural dependency is a referential dependency; the subject that is coded by -yi- must be embedded within the perspective of some third person 225 (4.3.22). This semantic condition matches the semantic condition on referential dependency exactly (cf. \u00a74.1.). Structural Conditions Semantic Condition C-command Linear Precedence Perspectival Embedding -yi- \u2714 \u2714 \u2714 Dependent Reference \u2714 \u2714 \u2714 Table 4.5. The suffix \u2013yi\u2013 meets the conditions of referential dependency Taken altogether, the picture developed for -yi- bears strong parallels to the model of dependent reference proposed in this chapter (\u00a74.1.). I conclude that the this suffix thus codes dependent reference for subjects, which suggests that it falls within the domain of switch reference (cf. Stirling 1993) \u2013 a possibility explored in Appendix A. 4.3.21. Structural conditions on -yi- The suffix \u2013yi\u2013 has the exact structural requirements that referential dependency has. In the most simple case, the clause it depends on structurally dominates it (\u00a74.3.211). (36) PRED1 6 y PRED2 6 x PRED2-yi In more complex cases, the clause it depends on for reference linearly precedes, but does not structurally dominate it (\u00a74.3.212). (37) a. PRED1 PRED2 6 6 y x PRED2-yi b. PRED1 PRED2 PRED 3 6 5 6 y x PRED3-yi 226 c. PRED1 PRED2 6 6 y PRED3 6 x PRED3-yi Based on the model of referential dependency\u2019s structural characteristics that I have proposed for Plains Cree (\u00a74.1.), this is exactly the kinds of structural relations we expect for -yi-. 4.3.211. \u2013yi\u2013 is c-commanded by its antecedent As Cook (2008) has argued, one kind of clausal relation in Plains Cree is c-command (\u00a74.2.2). When we consider the occurrence of the suffix -yi- with these kind of constructions, we find that it will occur whenever its predicate is c-commanded by a previous predicate and the subject is different from a third person in this matrix predicate. In both the \u201cObviative\u201d subject in (38a) and the \u201cInanimate\u201d subject in (38b), the subordinate predicate has a subject that is not the subject of the matrix predicate. In both instances, the lower predicate receives the suffix -yi-. (38) a. \u201cOBVIATIVE\u201d SUBJECT OF VERB WITH \u2013yi\u2013 \u14c8\u142f\u1424 \u1403\u144c\u1528\u1426\u1455\u1466 \u1401 \u14c2\u1472\u14a7\u1528\u141f \u1403\u1422\u1475\u1418 n\u00e2p\u00eaw it\u00eayihtam \u00ea-nikamoyit iskw\u00eawa n\u00e2p\u00eaw it =\u00eayiht -am \u00ea- nikamo-yi -t iskw\u00eaw-a man thus=by.mind.TI-TI C1-sing -DS-3 woman-XT \u2018The manPROX thinks the womanOBV is singing.\u2019 (Presented S2) b. \u201cINANIMATE\u201d SUBJECT OF VERB WITH \u2013yi\u2013 \u14c8\u142f\u1424 \u1403\u144c\u1528\u1426\u1455\u1466 \u1401 \u14a5\u14f5\u1528\u1420 \u14aa\u1422\u146d\u14ef\u1423 n\u00e2p\u00eaw it\u00eayihtam \u00ea-mis\u00e2yik maskisin n\u00e2p\u00eaw it =\u00eayiht -am \u00ea- mis=\u00e2 -yi -k maskisin man thus=by.mind.TI-TI C1-big=STAT-DS-0 shoe \u2018The manAN thinks the shoe is bigIN.\u2019 (Presented S2) This generalization is also true of possessed predicates, although possession constructions only allow \u201cAnimate\u201d possessors (\u00a74.4.1). 227 (39) \u201cOBVIATIVE\u201d POSSESSORS RECEIVE \u2013yi\u2013 a. \u14c8\u142f\u1424 \u141a\u1438\u14ab\u1424 \u1403\u1422\u1475\u1418 \u1405\u144c\u14a5\u1528\u1418 n\u00e2p\u00eaw w\u00e2pam\u00eaw iskw\u00eawa ot\u00eamiyiwa n\u00e2p\u00eaw w\u00e2p=am -\u00ea -w iskw\u00eaw-a o-t\u00eam -yi-w-a man see =by.eye.TA-DIR-3 woman-XT 3-horse-DS-3-XT \u2018The manPROX saw the woman\u2019sOBV horseOBV.\u2019 (Presented S2) b. \u14c8\u142f\u1424 \u141a\u1438\u1426\u1455\u1466 \u1403\u1422\u1475\u1418 \u1405\u14aa\u1422\u146d\u14ef\u14c2\u1528\u1424 n\u00e2p\u00eaw w\u00e2pahtam iskw\u00eawa omaskisiniyiw n\u00e2p\u00eaw w\u00e2p=aht -am iskw\u00eaw-a o-maskisin-yi-w man see =by.eye.TI-TI woman-XT 3-shoe -DS-3 \u2018The manPROX saw the woman\u2019sOBV shoeIN.\u2019 (Presented S2) Here, the possesed predicate (-t\u00eam \u2018horse\/dog\u2019 and maskisin \u2018shoe\u2019) is possessed by someone other than the subject of the verb, and the possessed predicate is affixed with -yi-. From this data, we see that -yi- occurs on subordinate predicates. Structurally, we can represent this configuration as a c-command relation between the main clause\u2019s argument and the subject in the lower clause. (40) PRED1 6 y PRED2 6 x PRED2-yi This models one of the most basic configurations that licenses the occurrence of -yi- on predicates in Plains Cree. 4.3.212. \u2013yi\u2013 is linearly preceded by its antecedent While \u2013yi\u2013 systematically occurs on subordinate predicates, it also occurs in configurations where its predicate is part of a chain of predicates (cf. Cook 2008). (41) PRED1 PRED2 6 6 y x PRED2-yi This introduces three logical possibilities for the relation between the predicate bearing \u2013yi\u2013 and its antecedent: (i) The predicate bearing \u2013yi\u2013 may immediately follow the predicate it is dependent on. 228 (ii) The predicate bearing \u2013yi\u2013 may be separated from this predicate by other chained predicates. (iii) The predicate bearing \u2013yi\u2013 may be separated from this predicate by a predicate that it is c-commanded by. I now consider each of these three possibilities in turn. A predicate bearing \u2013yi\u2013 may occur directly after another predicate that it is dependent on, but not subordinated to. Consider the example in (42), where the last predicate (\u00ea-k\u00ee- kiskinoham\u00e2kosiyit \u2018heOBV learned\u2019) is coded with \u2013yi\u2013.4 This means that the subject of this predicate is not the previous third person \u2018him,\u2019 which is the possessor of the previous nominal ow\u00eekim\u00e2kana \u2018hisPROX wifeOBV\u2019 and the object of \u00ea-w\u00eecihikot \u2018sheOBV helped himPROX.\u2019 (42) PREDICATED BEARING \u2013yi\u2013 IMMEDIATELY FOLLOWS MAIN PREDICATE i. \u14c7\u14c8\u1450\u1541 \u1403\u14ef \u14a6\u14c7 \u1401 \u146e \u141a\u1438\u1426\u1455\u14ab\u1423 \u1401 \u1411\u148b\u1426\u1403\u146f\u141f \u140a\u152d \u1405\u1411\u146d\u14ab\u1472\u14c7, \u2026 nan\u00e2tohk isi m\u00een \u00ea-k\u00ee-w\u00e2pahtam\u00e2n \u00ea-w\u00eecihikot aya ow\u00eekim\u00e2kana, \u2026 nan\u00e2tohk isi m\u00eena \u00ea- k\u00ee- w\u00e2p=aht -am-\u00e2n \u00ea- w\u00eec=ih -iko -t aya many.way thus also C1-prev-see=by.eye.ti-TI -1 C1-with=by.neut-INV-3 HES o-w\u00eekim\u00e2kan-a 3-spouse -XT \u2018I also saw that hisPROX wifeOBV helped him in various ways, \u2026\u2019 ii. \u2026 \u1401 \u146e \u146d\u1422\u146d\u146f\u1426\u140a\u14ab\u146f\u14ef\u1528\u141f, \u2026 \u2026 \u00ea-k\u00ee-kiskinoham\u00e2kosiyit, \u00ea- k\u00ee- kisk =ino =h =amaw=ikw =si- yi-t c1-prev-know=by.hand=CAUS=APPLIC=INV=STAT-DS-3 \u2018\u2026 sheOBV had gone to school, \u2026\u2019 (EM 43) Here, the relation between this clause and previous clauses is not straightforward structural c- command; the second predicate marked with \u2013yi\u2013 (\u00ea-k\u00ee-kiskinoham\u00e2kosiyit \u2018she had gone go school\u2019) is part of a chain of clauses, rather than a subordinate clause (cf. Cook 2008), and immediately follows a the clause it is dependent on (\u2026\u00ea-w\u00eecihikot aya ow\u00eekim\u00e2kana \u2018hisPROX wifeOBV helped himPROX\u2019). (43) PRED1 PRED2 6 6 y w\u00eecihikot x kiskinoham\u00e2kosiyit 4 These examples contain \u201cObviative\u201d reference, rather than \u201cInanimate\u201d, but the \u201cInanimate\u201d forms are equally available and attested in texts. For example: \u2026, m\u00eaton \u00ea-nipahi-mis\u00e2yik \u00f4ma mug w\u00eetapihtam, miton \u00eakota \u00ea-asiwat\u00eayik anima yellow \u00ea-isin\u00e2kwahk. \u2018\u2026, [he was] sitting there before an enormously big mug filled to the brim with something that looked yellow.\u2019 (Ahenakew 2000 3:3). 229 The preceding predicate may also be a possessed nominal, as the string in (44) demonstrates. In this example, the first predicate marked with \u2013yi\u2013 (\u00ea-kiskinoham\u00e2k\u00eayit \u2018sheOBV was a teacher\u2019) linearly succeeds the initial clause that it is dependent on (\u00ea-isiyaw\u00easit \u2018he was angry\u2019). The subsequent predicates marked with \u2013yi\u2013 (\u00ea-wiyaw\u00eeyit \u2018sheOBV went out\u2019 and \u00ea-ay\u00e2yit \u2018she was there\u2019) each follow a possessed nominal predicate that makes referent to the referent that the subject of the clause is dependent on (w\u00eewa \u2018hisPROX wifeOBV\u2019 and osikosa \u2018hisPROX mother-in- lawOBV\u2019). (44) \u2013yi\u2013 FOLLOWS POSSESSED PREDICATE i. \u1456\u14c2\u14ef \u1401\u144e\u1475 \u1401 \u1403\u14ef\u152d\u140d\u14ef\u141f \u1401\u147f \u140a\u1418, \u2026 t\u00e2nis \u00eatikw\u00ea \u00ea-~ \u00ea-~ \u00ea-isiyaw\u00easit \u00eakwa awa, \u2026 t\u00e2nisi \u00eatikw\u00ea \u00ea- isi =yaw\u00easi -t \u00eakwa aw -a how EVID C1-thus=wrathful-3 and PRX-AN.SG \u2018And I do not know what heAN was so angry about, \u2026\u2019 ii. \u2026 \u1401 \u146d\u1422\u146d\u14c4\u1426\u140a\u14ab\u146b\u1528\u141f \u1411\u1418 \u1401 \u140f\u152d\u1411\u1528\u141f \u2026 \u2026 \u00ea-kiskinoham\u00e2k\u00eayit w\u00eewa \u00ea-wiyaw\u00eeyit \u2026 \u00ea-kisk =ino =h =amaw =ik\u00ea-yi -t w-\u00eew -a \u00ea-wiyaw\u00ee-yi-t c1-know=by.hand=CAUS=APPLIC=AI -DS-3 3-wife-XT C1-go.out-DS-3 \u2018\u2026 but when hisPROX wifeOBV the teacher went out \u2026\u2019 iii. \u1401\u147f \u1405\u146f\u14ef\u14f4 \u14a6\u14c7 \u1401\u146f\u1455 \u1401 \u140a\u152e\u1528\u141f, \u2026 \u2026 \u00eakwa osikosa m\u00een \u00eakota \u00ea-ay\u00e2yit, \u2026 \u00eakwa o-sikos -a m\u00eena \u00eakota \u00ea- ay\u00e2 -yi -t and 3-mother.in.law-XT also there C1-be.there-DS-3 \u2018\u2026 and hisPROX mother-in-lawOBV was there as well, \u2026\u2019 (Ahenakew 5:7) This is also a case of linear-precedence (45). (45) PRED1 PRED2 6 6 y w\u00eewa x \u00ea-wiyaw\u00eeyit In both of the case in (42) and (43), then, the predicate coded with \u2013yi\u2013 linearly succeeds the predicate it is referentially dependent on but is not structurally dominated by it. This means that the suffix \u2013yi\u2013 occurs on a predicates that are not straightforwardly c-commanded by their antecedent. While in the preceding examples pivot predicate immediately precedes the predicate bearing \u2013yi\u2013 , the two predicates may be further apart. The example in (46) shows this kind of configuration. Here, the antecedent is only associated with the first predicate (\u00ea-k\u00ee-nakiskaw\u00e2t \u2018hePROX met 230 herOBV\u2019). The next predicate (\u00ea-kiyok\u00eay\u00e2n \u2018I was visiting\u2019) makes no reference to the antecedent at all. The next clause, which is conjoined with \u00eakwa \u2018and,\u2019 has a predicate that is dependent for its reference on the initial predicate, and bears \u2013yi\u2013 (\u00ea-k\u00ee-miyohtw\u00e2yit \u2018sheOBV was a good person\u2019). (46) PREDICATE MARKED WITH \u2013yi\u2013 IS SEPARATED FROM INITIAL PREDICATE i. \u14c2\u146e \u14c2\u1455\u140f \u146d\u152a\u1473\u1423 \u14c3\u1489\u141a\u1472\u1423 \u1405\u141a\u1422\u1472\u1426\u1403\u1472\u14c2\u1541. Nik\u00ee-nitawi-kiyok\u00e2n n\u00eec\u00eaw\u00e2kan ow\u00e2skahikanihk. Ni-k\u00ee- nitawi-kiyok\u00ea-n ni-w\u00eec\u00eaw =\u00e2 =kan o-w\u00e2skahikan-ihk 1- PREV-go- visit -LP 1- accompany=DIR=NOM 3-house -LOC \u2018I went to visit my friendAN in his home.\u2019 ii. \u1401 \u146e \u14c7\u146d\u1422\u1472\u141a\u141f \u1403\u1422\u1475\u1418 \u1401 \u146d\u152a\u146b\u152e\u1423, \u00ea-k\u00ee-nakiskaw\u00e2t iskw\u00eawa \u00ea-kiyok\u00eay\u00e2n, \u00ea- k\u00ee- nak =iskaw -\u00e2 -t iskw\u00eaw-a \u00ea-kiyok\u00ea-y\u00e2n C1-PREV-meet=by.body.TA-DIR-3 woman-XT C1-visit-1 \u2018HePROX had met a womanOBV when I was visiting, \u2026\u2019 iii. \u1401\u147f \u1450\u1423\u1450\u14c0 \u1401 \u146e \u14a5\u152a\u1426\u1464\u1528\u141f. \u00eakwa tonton\u00ea \u00ea-k\u00ee-miyohtw\u00e2yit. \u00eakwa tonton\u00ea \u00ea- k\u00ee- miyw=htw\u00e2 -yi-t and intense C1-PREV-good =behave-DS-3 \u2018\u2026 and sheOBV was really kind.\u2019 (Presented S2) This is a case where another predicate intervenes in the linear-precedence relation, as schematized in (47). (47) PRED1 PRED2 PRED 3 6 5 5 y \u00ea-nakiskaw\u00e2t \u00ea-kiyok\u00eay\u00e2n x \u00ea-miyohtw\u00e2yit Thus, another predicate can intervene between the \u2013yi\u2013marked clause and its antecedent. A third possibility is that the predicate marked with \u2013yi\u2013 is linearly distant from its pivot predicate and is c-commanded by a clause that lacks reference to this antecedent. An example of this kind is shown in (48). Here, the second predicate marked with \u2013yi\u2013 (\u00ea-w\u00ee-p\u00f4n-ahkosiyit \u2018she will stop being sick\u2019) is separated from a predicate that makes reference to its pivot (nitawi- w\u00e2pam\u00eaw \u2018he went to see her\u2019 or om\u00e2m\u00e2wa \u2018his mother\u2019) by another predicate (ninitaw\u00eayiht\u00ean \u2018I want it\u2019), which it is c-commanded by (see \u00a74.3.21). 231 (48) \u2013yi\u2013 PREDICATE C-COMMANDED BY INTERVENING PREDICATE i. \u14c3\u1489\u141a\u1472\u1423 \u14c2\u1455\u140f \u141a\u1438\u14a3\u1424 \u1405\u14ab\u14ab\u1418, \u2026 n\u00eec\u00eaw\u00e2kan nitawi-w\u00e2pam\u00eaw om\u00e2m\u00e2wa, \u2026 ni-w\u00eec\u00eaw=\u00e2=kan nitawi-w\u00e2p=am -\u00ea -w o-m\u00e2m\u00e2-a 1-accompany=DIR=NOM go- see=by.eye.TA-DIR-3 3-mother-XT \u2018My friendPROX has gone to visit herPROX motherOBV, \u2026 \u2019 (Presented S2) ii. \u2026 \u1401\u147f \u14c2\u14c2\u1455\u140d\u1528\u1426\u144c\u1423 \u1472 \u1434\u1423 \u140a\u1426\u146f\u14ef\u1528\u141f. \u2026 \u00eakwa ninitaw\u00eayiht\u00ean ka-p\u00f4n-ahkosiyit. \u00eakwa ni-nitaw=\u00eayiht -\u00ea -n ka-p\u00f4n-ahkosi-yi -t and 1-toward=by.mind.TI-TI-LP C1-stop-sick -DS-3 \u2018\u2026 and I want herOBV to stop being sick.\u2019 (Presented S2) This is a case where an intervening predicate c-commands the predicate bearing \u2013yi\u2013 (49). (49) PRED1 PRED2 6 6 y w\u00e2pam\u00eaw PRED3 6 x \u00ea-w\u00ee-p\u00f4n-ahkosiyit Considered in light of Cook\u2019s (2008) analysis of clause-chaining constructions as precedence relations, we immediate see parallels in the problem of identifying the antecedent clause for \u2013yi\u2013. The clause that \u2013yi\u2013 relates to is this initial, discourse-setting clause (i.e. the \u201cpivot\u201d clause; Stirling 1993), rather than the immediately preceding, and\/or structurally dominant clause. Seen in this way, the suffix \u2013yi\u2013 can be understood as part of a more general pattern of anaphora in Plains Cree; its antecedent can either structurally dominate it or linearly precede it. 4.3.22. Semantic conditions on \u2013yi\u2013 So far, we have seen that the suffix \u2013yi\u2013 codes a disjunction from some previous referent, but we have said very little about the properties of this antecedent. When we do this, we immediately notice a cluster of related requirements for the antecedent: (i) The antecedent must be in the previous discourse (\u00a74.3.221). (ii) The antecedent must be \u201cAnimate\u201d (\u00a74.3.222). (iii) The antecedent must possess a perspective for the predicate marked with \u2013yi\u2013 to be evaluated within (\u00a74.3.223). 232 (iv) If the matrix predicate is a psych verb, the c-commanded predicate must bear \u2013yi\u2013 (\u00a74.3.224). Taken together, these four observations point to the same conclusion: the suffix \u2013yi\u2013 codes that its argument is embedded within the perspective of some previous referent. y (50) x PREDICATE(x) + -yi \uf0e0 PREDICATE(x) This is expected within the model of dependent reference constructed in this chapter (\u00a74.1.). 4.3.221. The antecedent of \u2013yi\u2013 must be in previous discourse We have already seen the requirement for the antecedent to be in the previous discourse. In section 4.3.211, we considered cases where the antecedent was in the matrix clause, while, in section 4.3.212, we considered cases where the antecedent was in a linearly precedent clause. In fact, this precedence is all that is required structurally; the antecedent can be either the subject (51a) or the object (51b) of previous clauses. (51) a. ANTECEDENT OF \u2013yi\u2013 IS SUBJECT \u14c3\u1489\u141a\u1472\u1423 \u1401 \u142f \u1411\u1455\u1450\u1422\u146b\u14a5\u141f, \u1401 \u140a\u1426\u146f\u14ef\u1528\u141f \u1411\u1418. n\u00eec\u00eaw\u00e2kan \u00ea-p\u00ea-w\u00eetatosk\u00eamit, \u00ea-ahkosiyit w\u00eewa. n-w\u00eec\u00eaw\u00e2kan \u00ea- p\u00ea- w\u00eet =atosk\u00ea=m -it \u00ea- ahkosi-yi -t w-\u00eew -a 1-friend C1-come-with=work =DJ-3>1 C1-sick -DS-3 3-wife-XT \u2018My friend came to work with me, hisPROX wifeOBV was sick.\u2019 (Presented S2) b. ANTECEDENT OF \u2013yi\u2013 IS OBJECT \u14c3\u1489\u141a\u1472\u1423 \u1401 \u14c2\u1455\u140f \u1411\u1455\u1450\u1422\u146b\u14aa\u1420, \u1401 \u140a\u1426\u146f\u14ef\u1528\u141f \u1411\u1418. n\u00eec\u00eaw\u00e2kan \u00ea-nitawi-w\u00eetatosk\u00eamak, \u00ea-ahkosiyit w\u00eewa n-w\u00eec\u00eaw\u00e2kan \u00ea- nitawi-w\u00eet =atosk\u00ea=m -ak \u00ea- ahkosi-yi -t w-\u00eew -a 1-friend C1-go- with=work =DJ-1>3 C1-sick -DS-3 3-wife-XT \u2018I went to work with my friend, hisPROX wifeOBV was sick.\u2019 (Presented S2) 233 Here, the antecedent, n\u00eec\u00eaw\u00e2kan \u2018my friend,\u2019 is the subject of (51a), as shown by the portmanteau5 verbal affix \u2013it, which codes that a 3rd person referent is acting on the Speaker. Conversely, the form in (51b) has the antecedent as the object, as shown by the portmanteau verbal affix \u2013ak, which codes that the Speaker is acting on a 3rd person referent. 4.3.222. The antecedent of \u2013yi\u2013 must be \u201canimate\u201d As Wolfart (p.c.) has pointed out, the antecedent must be \u201cAnimate\u201d in order for -yi- to occur on subsequent predicates. Consider the pairwise contrast in (52). Here, only a structure coding the referent of the matrix clause as \u201cAnimate\u201d (the TA form miskaw- \u2018find\u2019 in 52b) allows for the application of \u2013yi\u2013 to the subsequent predicate. Within the context of the current model, this ban on \u201cInanimate\u201d antecedents makes sense; \u201cInanimate\u201d referents are those that inherently lack a perspective (see Chapter 2). (52) If ANTECEDENT IS \u201cINANIMATE,\u201d NO \u2013yi\u2013 a. \u14c2\u14a5\u1422\u146b\u1423 Hobbes \u1401 \u142f \u146d\u152a\u146b\u141f Calvin. nimisk\u00ean Hobbes \u00ea-p\u00ea-kiyok\u00eat Calvin. ni-m=isk -\u00ea -n Hobbes \u00ea- p\u00ea- kiyok\u00ea-t Calvin-a 1-find=by.body.TI-TI-LP Hobbes c1-come-visit -3 Calvin-XT \u2018I found HobbesIN when CalvinAN came to visit.\u2019 (Presented S2) b. # \u14c2\u14a5\u1422\u146b\u1423 Hobbes \u1401 \u142f \u146d\u152a\u146b\u1528\u141f Calvin\u140a. # nimisk\u00ean Hobbes \u00ea-p\u00ea-kiyok\u00eayit Calvina. ni-m=isk -\u00ea -n Hobbes \u00ea- p\u00ea- kiyok\u00ea-yi-t Calvin-a 1-find=by.body.TI-TI-LP Hobbes c1-come-visit -DS-3 Calvin-XT \u2018I found HobbesIN when CalvinOBV came to visit.\u2019 (Presented S2) Without a perspective, the referent cannot force a subsequent referent to be referentially dependent on it. 4.3.223. The antecedent of \u2013yi\u2013 must have a perspective Even if the antecedent is in previous discourse (\u00a74.3.221) and \u201cAnimate\u201d (\u00a74.3.222), \u2013yi\u2013 can still fail to occur if the final condition is not met; the antecedent must possess a perspective. An instructive near-minimal pair of this is seen in Sarah Whitecalf\u2019s speech on the Sundance 5 It is possible that this form is actually -i-t, where -i- is the theme sign coding a 1st person object (see D\u00e9chaine & Reinholtz 2008), and -t codes a 3rd person \u201cAnimate\u201d subject. I am taking the less controversial approach here because the difference is not relevant to the current discussion. 234 (Whitecalf 1993:\u00a716,17). When she refers to an audience member (H.C. Wolfart) and his desire to learn about Cree ways, she uses a psych predicate (kisk\u00eayiht- \u2018know\u2019), a psych preverb (n\u00f4ht\u00ea- \u2018want\u2019), and then marks the subordinate clause \u00ea-ispayiniyik \u2018it works\u2019 with the suffix \u2013yi\u2013. (53) ANTECEDENT IS A SPECIFIC PERSON, \u2013yi\u2013 IS USED \u1401 \u14c5\u1426\u144c \u146d\u1422\u146b\u1528\u1426\u1455\u1541 \u1456\u14c2\u14ef \u1401 \u1403\u1422\u1438\u1528\u14c2\u1528\u1420 \u1406\u14aa \u146d\u14c0\u1426\u1403\u152e\u140f\u14c2\u14c7\u1424. \u00ea-n\u00f4ht\u00ea-kisk\u00eayihtahk t\u00e2nisi \u00ea-ispayiniyik \u00f4ma kin\u00eahiy\u00e2wiwininaw. \u00ea- n\u00f4ht\u00ea-kisk=\u00eayiht-am-k t\u00e2nisi \u00ea- is =payi- n-yi-k C1-want-know=by.mind-TI-3 how C1-thus=INCH-II-DS-0 aw-ima ki-n\u00eahiy\u00e2w-i-w-in -inaw. PROX-IN.SG 2-cree -AI-3-NOM-21PL \u2018[he]AN wants to understand how [our Cree culture]IN works\u2019 (Whitecalf 1993:\u00a716) Whitecalf employs a parallel structure later on in the same speech, but this time the subordinate predicate \u00ea-ispayik \u2018itIN works\u2019 lacks the suffix \u2013yi\u2013. Here, the very same verb (\u00ea-ispayik \u2018it works\u2019), in a similarly subordinative structure lacks the suffix \u2013yi\u2013. (54) ANTECEDENT IS GENERIC, NO \u2013yi\u2013 IS USED \u14a7\u1450\u14c2 \u144c\u1433 \u146d\u1422\u146d\u14c4\u141a\u1438\u1426\u1455\u1426\u146d \u1406\u14aa \u14c2\u1439\u1475\u14ef\u14a7\u140f\u1472\u14a5\u1420 \u1456\u14c2\u14ef \u1401 \u1403\u1422\u1438\u1528\u1420 mitoni t\u00eapi-kiskinow\u00e2pahtahki \u00f4ma nip\u00e2kw\u00easimowikamik t\u00e2nisi \u00ea-ispayik mitoni t\u00eapi- kisk=inw =w\u00e2p =aht -am-k-i aw-ima intense enough-know=by.hand=see =by.eye.TI-TI -0-SUBJ PROX-IN.SG nip\u00e2kw\u00easimo-w-ikamikw t\u00e2nisi \u00ea-is=payi -k sun.dance -3-place how C1-thus=INCH-0 \u2018When the WhitesAN have watched and learned enough about how the Sundance-LodgeIN works, \u2026\u2019 (Whitecalf 1993:\u00a717) When speakers are asked about these pairs of sentences, their explanations always appeal to notions of point of view. For example, the pair of sentences offered by Sarah Whitecalf have been explained by two different speakers (S2, S3) as coding a contrast in points of view; the first sentence (wherein H.C. Wolfart wishes to know how the Cree Culture works) represents H.C. Wolfart\u2019s point of view, whereas the second sentence does not represent any third person\u2019s point of view. The combination of judgments like this and the textual patterns led Wolfart (p.c.) to describe this distinction as one between a discourse-prominent third person and a generic third person. This set of interpretations by native speakers makes sense when we consider the context of Sarah Whitecalf\u2019s discussion; in the first example (53), she is referring to a specific audience 235 member, and what his intentions are (i.e. he has a desire to learn about Cree culture). In the second example (54), she is not referring to any specific referent, but rather the general behaviours of generic people (m\u00f4niy\u00e2sak \u2018the whites\u2019). Crucially, this second example employs subjunctive morphology (-i); there is no current intention for a real referent, only possible future intentions for unknown referents. The example in (53), then, represents the perspective of H.C. Wolfart, while the example in (54) represents no one in particular\u2019s. We can model this in the DRT framework developed thus far by exploiting the presence versus absence of an embedded perspective. In the case of the example in (53), which represents the perspective of H.C. Wolfart, we can assign this referent (W) an embedded perspective (55). Embedding the referent y, identified with \u201cCree culture\u201d inside of the perspective of H.C. Wolfart (W) forces y to be dependent on W. With this dependency established, the linear and structural rules for applying -yi- come into play, and the subordinate predicate ispayi- \u2018function\u2019 is coded with -yi-, since its subject is different from, and dependent on, an antecedent, which c- commands it. (55) \u1401 \u14c5\u1426\u144c \u146d\u1422\u146b\u1528\u1426\u1455\u1541 \u1456\u14c2\u14ef \u1401 \u1403\u1422\u1438\u1528\u14c2\u1528\u1420 \u1406\u14aa \u146d\u14c0\u1426\u1403\u152e\u140f\u14c2\u14c7\u1424. \u00ea-n\u00f4ht\u00ea-kisk\u00eayihtahk t\u00e2nisi \u00ea-ispayiniyik \u00f4ma kin\u00eahiy\u00e2wiwininaw. \u00ea- n\u00f4ht\u00ea-kisk=\u00eayiht-am-k t\u00e2nisi \u00ea- is =payi- n-yi-k C1-want-know=by.mind-TI-3 how C1-thus=INCH-II-DS-0 aw-ima ki-n\u00eahiy\u00e2w-i-win -inaw. PROX-IN.SG 2-cree -AI-NOM-21PL \u2018[he]AN wants to understand how [our Cree culture]IN works\u2019 (Whitecalf 1993:\u00a716) Speaker y Cree.culture(y) Wolfart y know(W,y) work(y) \uf0df \u2018WORK\u2019 EMBEDDED IN WOLFART\u2019S PERSPECTIVE 236 Turning now to the case in (54), we do not assign an embedded perspective to m\u00f4niy\u00e2sak \u2018the whites\u2019 (56). (56) \u14a7\u1450\u14c2 \u144c\u1433 \u146d\u1422\u146d\u14c4\u141a\u1438\u1426\u1455\u1426\u146d \u1406\u14aa \u14c2\u1439\u1475\u14ef\u14a7\u140f\u1472\u14a5\u1420 \u1456\u14c2\u14ef \u1401 \u1403\u1422\u1438\u1528\u1420 mitoni t\u00eapi-kiskinow\u00e2pahtahki \u00f4ma nip\u00e2kw\u00easimowikamik t\u00e2nisi \u00ea-ispayik mitoni t\u00eapi- kisk=inw =w\u00e2p =aht -am-k-i aw-ima intense enough-know=by.hand=see =by.eye.TI-TI -0-SUBJ PROX-IN.SG nip\u00e2kw\u00easimo-w-ikamikw t\u00e2nisi \u00ea-is=payi -k sun.dance -3-place how C1-thus=INCH-0 \u2018When the Whites have watched and learned enough about how the Sundance-Lodge works,\u2019 (Whitecalf 1993:\u00a717) Speaker x y whites(x) Sun.Dance(y) know(x,y) work (y) \uf0df \u2018WORK\u2019 EMBEDDED IN SPEAKER\u2019S PERSPECTIVE Here, there is no embedded perspective, and thus there is no referent for the \u201cInanimate\u201d subject of the subordinate clause to be dependent on. Lacking such a dependence, the application of -yi- is not licensed, and the predicate ispayi- \u2018function\u2019 does not receive this suffix. 4.3.224. Propositional attitude verbs force \u2013yi\u2013 If the matrix verb is a verb that expresses a propositional attitude, and its subject meets the three previous criteria (in previous discourse, \u201cAnimate,\u201d existential), subjects in the subordinate clause must be coded with \u2013yi\u2013. Consider the pairwise contrasts in (57) and (58). In (57), the matrix clause is a propositional attitude verb (it\u00eayihtam \u2018heAN think thus of itIN\u2019), and the subordinate clause \u00ea-nikamoyit \u2018she singsOBV\u2019 has iskw\u00eaw \u2018woman\u2019 as its subject. In this configuration, the lower predicate must have \u2013yi\u2013 affixed (57b). 237 (57) PROPOSITIONAL ATTITUDE VERBS FORCE \u2013yi\u2013 FOR \u201cANIMATE\u201d REFERENTS a. \u14c8\u142f\u1424 \u1403\u144c\u1528\u1426\u1455\u1466 \u1401 \u14c2\u1472\u14a7\u1528\u141f \u1403\u1422\u1475\u1418 n\u00e2p\u00eaw it\u00eayihtam \u00ea-nikamoyit iskw\u00eawa n\u00e2p\u00eaw it =\u00eayiht -am \u00ea- nikamo-yi -t iskw\u00eaw-a man thus=by.mind.TI-TI C1-sing -DS-3 woman-XT \u2018The manPROX thinks the womanOBV is singing.\u2019 (Presented S2) b. \u273d \u14c8\u142f\u1424 \u1403\u144c\u1528\u1426\u1455\u1466 \u1401 \u14c2\u1472\u14a7\u141f \u1403\u1422\u1475\u1424 \u273d n\u00e2p\u00eaw it\u00eayihtam \u00ea-nikamot iskw\u00eaw n\u00e2p\u00eaw it =\u00eayiht -am \u00ea- nikamo-t iskw\u00eaw man thus=by.mind.TI-TI C1-sing -3 woman \u2018The manAN thinks the womanAN is singing.\u2019 (Presented S2) The same holds true with \u201cInanimate\u201d subjects; the subordinate clause must be marked with \u2013yi\u2013 , as the minimal pair in (58) shows. (58) PROPOSITIONAL ATTITUDE VERBS FORCE \u2013yi\u2013 FOR \u201cINANIMATE\u201d REFERENTS a. \u14c8\u142f\u1424 \u1403\u144c\u1528\u1426\u1455\u1466 \u140a\u14c2\u14aa \u14aa\u1422\u146d\u14ef\u1423 \u1401 \u14a5\u14f5\u1528\u1420. n\u00e2p\u00eaw it\u00eayihtam anima maskisin \u00ea-mis\u00e2yik. n\u00e2p\u00eaw it =\u00eayiht -am an =ima maskisin \u00ea- mis=\u00e2 -yi -k man thus=by.mind.TI-TI DST=IN.SG shoe C1-big=STAT-DS-0 \u2018The manAN thinks the shoe is bigIN.\u2019 (Presented S2) b. \u273d \u14c8\u142f\u1424 \u1403\u144c\u1528\u1426\u1455\u1466 \u140a\u14c2\u14aa \u14aa\u1422\u146d\u14ef\u1423 \u1401 \u14a5\u14f5\u1420. \u273d n\u00e2p\u00eaw it\u00eayihtam anima maskisin \u00ea-mis\u00e2k. n\u00e2p\u00eaw it =\u00eayiht -am an =ima maskisin \u00ea- mis=\u00e2 -k man thus=by.mind.TI-TI DST=IN.SG shoe C1-big=STAT-0 \u2018The manAN thinks the shoe is bigIN.\u2019 (Presented S2) Since propositional attitude verbs always introduce a perspective for the experiencer (cf. \u00a71.3, \u00a72.4, \u00a73.4), these constructions code that the subject of the lower predicate is evaluated within the perspective of the subject of the propositional attitude verb. This is shown in (59). 238 (59) Speaker y man(y) y x shoe(x) big(x) \uf0df \u2018BIG\u2019 IS EMBEDDED INSIDE OF MAN\u2019S PERSPECTIVE 4.3.23. Conclusion: -yi- constructs referential dependency Taken altogether, we now have a model of the referential dependency coded by the suffix \u2013yi\u2013 in Plains Cree. First, we see that this dependency has structural correlates, including the necessity of one referent to structurally dominate or linearly precede the other. Second, we see that dependency has semantic properties; the dependent referent must be embedded inside the perspective of the antecedent. From this analysis, we see that the function of -yi- overlaps with the \u201cObviative\u201d referential category, but the two are not identical. This is because -yi- defines a structural position for a dependent referent (i.e. the subject of the predicate), whereas the \u201cObviative\u201d is a particular kind of referent, whose inability to convey information to the Speaker often requires the construction of chains of discourse-level dependencies (\u00a73.5). When the conditions on \u201cObviative\u201d usage coincide with those for the suffix -yi-, the \u201cObviative\u201d referent can be coded by the suffix -yi-. Thus, when an \u201cObviative\u201d referent, which is dependent on a \u201cProximate\u2019s\u201d perspective (vis-a-vis its extentional content: see chapter 3, section 3.4), is the subject of a verb, it is de facto a dependent subject, which triggers the application of -yi- to the verb. 239 (60) DEPENDENT SUBJECT \uf0e0 -yi- \u2026\u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u1473 \u146d\u14c4\u14ef\u1528\u141f. \u2026 \u00ea-w\u00e2pam\u00e2t n\u00e2p\u00eaw iskw\u00eawa k\u00e2-kinosiyit \u00ea-w\u00e2p=am -\u00e2 -t n\u00e2p\u00eaw iskw\u00eaw-a k\u00e2-kinw=si -yi -t c1-see=by.eye.TA-DIR-3 man woman-XT C2-long=STAT-DS-3 \u2018\u2026the manPROX saw the womanOBV who was tall.\u2019 (Presented S2) In other configurations, the \u201cObviative\u201d referent is still referentially dependent, but not in a configuration that licenses -yi-. Thus, when an \u201cObviative\u201d referent is the object of a verb, it is not a dependent subject, and thus there will be no -yi- on the verb. (61) NOT DEPENDENT SUBJECT \uf0e0 No -yi- \u2026\u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u1473 \u1411\u1455\u1450\u1422\u146b\u14ab\u141f. \u00ea-w\u00e2pam\u00e2t n\u00e2p\u00eaw iskw\u00eawa k\u00e2-w\u00eetatosk\u00eam\u00e2t \u00ea-w\u00e2p=am -\u00e2 -t n\u00e2p\u00eaw iskw\u00eaw-a k\u00e2-w\u00eet=atosk\u00ea=m- \u00e2 -t c1-see=by.eye.TA-DIR-3 man woman-XT C2-long=work=DJ-DIR-3 \u2018The manPROX saw the womanOBV who hePROX works with.\u2019 (Presented S2) Thus, the verbal system can be said to use a semantically-compatible morpheme to opportunistically track \u201cObviative\u201d referents. This is a significantly different conclusion than we would be led to by an analysis that posits -yi- to be \u201cObviative\u201d agreement, and leads to a significantly different picture of the grammar\u2019s organization. 4.4. Predicate-internal dependencies and obviation So far, we have considered the construction of referential dependency that involves two predicates. This included (i) nominal-nominal relations (\u00a74.4.1), and (ii) dependent subjects (\u00a74.4.2). Both of these constructions shared a structural constraint; the predicate containing the dependent referent must linearly succeed the predicate containing its antecedent precedence. While these predicate-predicate relations all have been demonstrated to code referential dependency via linear precedence, the model of referential dependency developed here also expects that referential dependencies can be coded strictly via c-command. 240 (62) V y V V = X DEPENDENT ON Y x V This could, in principle, be a relation between predicates, as shown by the c-command cases with \u2013yi\u2013 (\u00a74.3.21). (63) PRED1 V y V PRED1 PRED2 V = X DEPENDENT ON Y x V PRED2 However, the models of Hirose (2000) and D\u00e9chaine (2003) require that predicate-internal relations must necessarily use c-command, because the predicate is constructed out of successively restricting, c-commanding components (see Chapter 2, \u00a72.2.1). (64) PRED V y V V = X DEPENDENT ON Y x V Thus, when we consider predicate-internal dependencies, we necessarily consider c-command environments. In the following sections, I consider predicate-internal relations in \u201cObviative\u201d constructions. In section 4.4.1, I show that nominal-internal relations always code both structural c-command and perspectival embedding; they meet both conditions of dependent reference. The obligatory coding of obviation in these environments thus becomes transparent. In section 4.4.2, I show that verb-internal relations also code both structural c-command and perspectival embedding. In the case of the \u2018direct\u2019 forms (\u00a74.4.21), the structure is equivalent to nominals; there is both c-command and perspectival embedding. Turning to the \u2018inverse\u2019 forms (\u00a74.4.22), I show that the inverse form breaks c-command between the dependent referent and its antecedent (cf. D\u00e9chaine & Reinholtz 1997, 2008), but introduces a linear-precedence constraint. In all 241 cases of predicate-internal relations, then, \u201cObviative\u201d constructions show a sensitivity to both the structural and semantic conditions on dependent reference. 4.4.1 Possession as referential dependency Plains Cree possession constructions are strictly head-marking; the possessum receives inflection for possession, while the possessor receives no coding at all (cf. Wolfart 1973). (65) POSSESSUM RECEIVES CODING, POSSESSOR DOES NOT \u1403\u1422\u1475\u1424 \u1405\u14a5\u14c5\u14ef\u14aa iskw\u00eaw omin\u00f4sima iskw\u00eaw w-min\u00f4s=im-a woman 3-cat =DSJ-XT \u2018the woman\u2019sPROX catOBV\u2019 (Presented S2) As can be seen in (65), the third-person possessed forms obey a particular \u201cProximate\u201d to \u201cObviative\u201d pattern; the possessor is \u201cProximate\u201d and the possessum is \u201cObviative.\u201d Removing the \u201cObviative\u201d suffix \u2013a from the possessum is ungrammatical, as exemplified by (66b). (66) POSSESSOR MUST BE \u201cPROXIMATE\u201d AND POSSESSUM MUST BE \u201cOBVIATIVE\u201d a. \u1403\u1422\u1475\u1424 \u1405\u14a5\u14c5\u14ef\u14aa iskw\u00eaw omin\u00f4sima iskw\u00eaw w-min\u00f4s-im-a woman 3-cat -DSJ-XT \u2018the woman\u2019sPROX catOBV\u2019 (Presented S2) b. \u273d \u1403\u1422\u1475\u1424 \u1405\u14a5\u14c5\u14ef\u14aa \u273d iskw\u00eawa omin\u00f4sim iskw\u00eaw-a w-min\u00f4s-im woman-XT 3-cat -DSJ Intended: \u2018the woman\u2019sOBV catPROX\u2019 (Presented S2) The \u201cProximate\u201d and \u201cObviative\u201d mapping cannot be reversed in possession constructions. The \u201cProximate\u201d referent always maps onto the possessor, and the \u201cObviative\u201d referent always maps onto the possessum.6 This is a widely-known property of possessor constructions in Plains Cree 6 It is important to note that in elicitation, possessums can lose their obviation coding (cf. Cook & M\u00fchlbauer 2006). Further, even in natural speech, when possessed forms are arguments of verbs, the verb can show simple \u201cAnimate\u201d agreement, rather than the expected \u201cObviative\u201d pattern (cf. Wolfart 1973). It seems that these constructions have to do with cases where the Speaker wishes to subordinate a referent initially, to introduce the referent felicitously, and then moves that referent to \u201cProximate\u201d status (cf. English, John\u2019s brother Bill called me yesterday.). 242 (cf. Bloomfield 1962, Wolfart 1973, Rhodes 1976, Dahlstrom 1986, D\u00e9chaine & Wiltschko 2002, etc.). Given such a correlation between obviation and possession, the current model expects that possession should show the two conditions on referential dependency: (i) STRUCTURAL CONDITION: It should show c-command or linear precedence effects. (ii) SEMANTIC CONDITION: It should show perspectival embedding. In the next two sections (4.4.11-12), I demonstrate that both of these conditions hold with possession constructions; possession is a kind of referential dependence. Structural Conditions Semantic Condition C-command Linear Precedence Perspective Embedding Possession \u2714 \u2718 \u2714 Referential Dependency \u2714 \u2714 \u2714 Table 4.6. Possession meets the conditions of referential dependency The reason for the tight correlation between possession and obviation, then, becomes transparent. 4.4.11. Possessors c-command possessums Based on the pattern of head-marking and the additive affixation, D\u00e9chaine (1999) analyzed possessor constructions as being built out of a hiearchy of forms, as in (67). (67) \u14c2\u14a5\u14c5\u14ef\u14a5\u14c8\u14c7\u1420 nimin\u00f4simin\u00e2nak ni-min\u00f4s-im -n\u00e2n-ak 1- cat -DSJ-1PL-PL \u2018our cats\u2019 (Volunteered S1) 243 DP 3 D PERSP ni- 3 PERS NUMP -n\u00e2n 3 NUM NP -ak 3 N min\u00f4s (Adapted from D\u00e9chaine 1999) In Chapter 3 (\u00a73.3.4), I proposed a modification to this structure based on the properties of the suffix \u2013im. In particular, I split the NP into two heads, in the same manner that Hirose (2000) split Plains Cree\u2019s VP structure (cf. \u00a72.2.1). (68) \u1405\u14a5\u14c5\u14ef\u14ef\u141a\u1418 omin\u00f4simisisiw\u00e2wa o-min\u00f4s-im -sis -w\u00e2w-a 3-cat -DISJ-DIM-3PL -XT \u2018TheirPROX cat(s)OBV\u2019 244 DP 3 D PERSONP o- 3 PERS EXTP -w\u00e2w 3 EXT NP -a 3 ROOT 3 min\u00f4s -im- 3 N NP 3 pro 3 N -sis- Here, the higher head of nP introduces the external argument of the nominal structure (i.e. the possessor), while the lower head of NP introduces the internal argument of the nominal structure (i.e. the possessum). This corresponds directly with the model of verbal structure proposed by Hirose (2000) and modified by D\u00e9chaine (2003). (69) \u14c2\u141a\u1438\u14ab\u1424 niw\u00e2pam\u00e2w ni-w\u00e2p=am -\u00e2 -w 1-see=see.by.eye.TA-DIR-3 \u2018I see him\/her\/itAN \u2026\u2019 (Volunteered S1) CP 3 ni- 3 C IP 3 I VP 3 ROOT 3 w\u00e2p- pro 3 V VP -am 3 3 V -\u00e2- (Hirose 2000, D\u00e9chaine 2003) 245 Here, the higher phrase (vP) introduces the subject argument, while the lower phrase (VP) introduces the object argument. Thus, possessors are structurally akin to subjects, and the nominal\u2019s own referential argument (Williams\u2019 1981 \u2018R\u2019 argument) is akin to an object. Given a model of this kind, the possessor is always structurally-superior to the possessum; it is introduced in the higher head (nP), and thus c-commands the argument in the next highest head (the R argument of NP). (70) NP 3 ROOT 3 Poss\u2019r 3 N NP 3 Poss\u2019m 3 N This configuration means that possession constructions fulfull the structural condition on referential dependency (\u00a74.1); the dependent referent is structurally subordinated to its antecedent. 4.4.12. Perspectival embedding with possession In order to demonstrate that possession is a kind of referential dependency, it is not sufficient to show that it meets the structural conditions of referential dependency. In addition, it is necessary to demonstrate that these structural conditions parallel a semantic condition; the referent that is c- commanded by its antecedent must also be interpreted within the perspective of that antecedent (\u00a74.1). For possession, this means that the possessum, the lower argument of the nominal structure (\u00a74.4.11), must be embedded in the perspective of the possessor, which is the higher argument of the nominal structure (\u00a74.4.11). 246 Possessor (71) Possessum PREDICATE(Possessum) \uf0df PREDICATE = TRUE W.R.T. POSS\u2019R\u2019S PERSPECTIVE Without this, possession cannot be said to code a kind of referential dependency, and we are left wondering why it should so closely parallel obviation, which is hypothesized to code such a dependency (\u00a73.4, \u00a74.1). In fact, a systematic consideration of possessor semantics strongly suggests that the possessum is indeed embedded in the perspective of the possessor. In section 4.3.12.1, I show that \u201cInanimate\u201d referents, which are inherently unable to possess a perspective (\u00a72.4), cannot be possessors \u2013 a fact that is expected if possessors are inherently perspectival. In section 4.4.122, I show that the possessor must always be existential and specific in Plains Cree, which parallels the findings for the antecedent of \u2013yi\u2013marked subjects seen above (\u00a74.3.2). In section 4.4.123, I show that possession induces an implication that the possessor holds the proposition expressed about the possessum to be true. Following this, I show in 4.4.12.4 that possession is intimately related to indexicality, which is expected if possessors are perspective-holders. Finally, in section 4.4.125, I show that possession interacts with the knowledge states of the speech-act participants, placing knowledge restrictions that support the conclusion that possession induces an expectation that the evaluation of the proposition should happen in the perspective of the possessor. Taken altogether, I consider this evidence to strongly support the conclusion that the possessum is embedded in the perspective of the possessor. This means that possession meets the semantic condition on referential dependence. 4.4.12.1. \u201cInanimate\u201d referents cannot be possessors Unlike English, Plains Cree does not allow possessors to be \u201cInanimate\u201d. There are no forms in any text, and consultants consistently explain that they are ill-formed in elicitation (72). Here, the form that inflects the nominal \u2013sk\u00e2t \u2018leg\u2019 for possession by t\u00eahtapiwin \u2018table\u2019 is rejected. Instead, the consultant volunteers a form that compounds t\u00eahtapiwin \u2018chair\u2019 together with \u2013sk\u00e2t \u2018leg,\u2019 giving it one total pitch contour for the entire construction (cf. M\u00fchlbauer 2006 for pitch and the \u2018word\u2019). 247 (72) \u201cINANIMATE\u201d REFERENTS CANNOT BE POSSESSORS a. \u273d \u14c2\u1432\u146f\u14c0\u1423 \u144c\u1426\u1455\u1431\u140f\u1423 \u1405\u1422\u1473\u1455. \u273d nip\u00eekon\u00ean t\u00eahtapiwin osk\u00e2ta. ni-p\u00eekw=n -\u00ea -n t\u00eaht=api=win o-sk\u00e2t-a 1- see=by.hand-TI-LP upon=sit=NOM 3-leg -XT Intended: \u2018I broke the chair\u2019sIN legs.\u2019 (Presented S2) COMMENT (S2): \u2018You\u2019re saying the chair is alive.\u2019 b. \u14c2\u1432\u146f\u14c0\u1423 \u144c\u1426\u1455\u1431\u140f\u14c2\u1422\u1473\u1455. nip\u00eekon\u00ean t\u00eahtapiwinisk\u00e2ta. ni-p\u00eekw=n -\u00ea -n t\u00eaht=api=win-sk\u00e2t-a 1- see=by.hand-TI-LP upon=sit=NOM-leg -XT \u2018I broke the chair\u2019sIN legs.\u2019 (Presented S2) One consultant (S2) specifically identified animacy as the problem, saying that \u201cit seems like you\u2019re saying the chair is alive, but you\u2019re also saying the chair isn\u2019t alive.\u201d Recall from chapter 2 (\u00a72.4) that \u201cInanimate\u201d referents are unable to function as the subjects of verbs that have intentional semantics (\u00a72.4.4). Class \u201cINANIMATE\u201d \u201cANIMATE\u201d Sensory (e.g. \u2018appear\u2019) isin\u00e2kwan- isin\u00e2kosi- Undergo (e.g. \u2018be smoked\u2019) kask\u00e2paht\u00ea- kask\u00e2paso- Emotion (e.g. \u2018x feels well\u2019) \u00d8 miyomaciho- Intellect (e.g. \u2018x remembers\u2019) \u00d8 kiskisi- Speech (e.g. \u2018x speaks\u2019) \u00d8 p\u00eekiskw\u00ea- Table 2.5. (Repeated) Semantic gaps among verb classes From that data and other sources (\u00a72.4), I argued that \u201cInanimate\u201d nominals are inherently extentional; they can never possess a perspective for the truth of a proposition to be evaluated in. Following this logic through to the current discussion of possession, we can understand the restriction on \u201cInanimate\u201d nominals as providing evidence about the semantics of possessor constructions. If possession in Plains Cree is inherently perspectival (i.e. the possessum is perspectivally-embedded in the possessor), the impossibility of \u201cInanimate\u201d possessors is expected. 4.4.122. Possessors are always existential In Plains Cree, the possessor is always interpreted as existential and specific (as opposed to generic). Consider the contrast in (73). 248 (73) a. COMPOUND NOMINAL = GENERIC POSSESSOR \u14c2\u14c5\u1426\u144c \u140a\u1456\u141a\u1423 \u14c8\u142f\u1418\u1422\u146d\u14ef\u14c7. nin\u00f4ht\u00ea-at\u00e2w\u00e2n n\u00e2p\u00eawaskisina. ni-n\u00f4ht\u00ea-at\u00e2w\u00ea-n n\u00e2p\u00eaw-maskisin-a 1- want- buy -LP man -shoe -XT = \u2018I want to buy men\u2019s shoes.\u2019 (Presented S2) \u2260 \u2018I want to buy the man\u2019s shoes.\u2019 Comment (S2): \u201cYou\u2019re just going to the store to buy shoes.\u201d b. OVERT POSSESSOR = SPECIFIC, EXISTENTIAL POSSESSOR \u14c2\u14c5\u1426\u144c \u140a\u1456\u141a\u1423 \u14c8\u142f\u1424 \u1405\u14aa\u1422\u146d\u14ef\u14c7. nin\u00f4ht\u00ea-at\u00e2w\u00e2n n\u00e2p\u00eaw omaskisina ni-n\u00f4ht\u00ea-at\u00e2w\u00ea-n n\u00e2p\u00eaw o-maskisin-a 1- want- buy -LP man 3-shoe -XT \u2260 \u2018I want to buy men\u2019s shoes.\u2019 (Presented S2) = \u2018I want to buy the man\u2019s shoes.\u2019 Comment (S2): \u201cYou\u2019re talking about some particular man.\u201d Here, I have set up the context as one in which I want to go to a shoe store to purchase shoes. The consultant (S2) explains that, in this context, I can only utter (73a) felicitously. I cannot utter (73b), because \u201cyou are talking about some particular man, and how you want to buy his shoes. You\u2019re not going to a store to buy shoes anymore.\u201d Thus, the compound form n\u00e2p\u00eawaskisina \u2018men\u2019s shoes\u2019 in (73a), which has one pitch contour (cf. M\u00fchlbauer 2006) and shows a loss of the initial nasal of maskisin \u2018shoe,\u2019 correlates with the generic type of possession, while the form in (73b), which has an independent nominal and possessor inflection, correlates with an specific, existential reading. Like with the restrictions on \u201cInanimate\u201d possessors, we have also seen this kind of pattern before. In section 4.3.2, we saw that the antecedent in a construction involving \u2013yi\u2013 had to be existential. (74) EXISTENTIAL ANTECEDENT = \u2013yi\u2013 ON PREDICATE \u1401 \u14c5\u1426\u144c \u146d\u1422\u146b\u1528\u1426\u1455\u1541 \u1456\u14c2\u14ef \u1401 \u1403\u1422\u1438\u1528\u14c2\u1528\u1420 \u1406\u14aa \u146d\u14c0\u1426\u1403\u152e\u140f\u14c2\u14c7\u1424. \u00ea-n\u00f4ht\u00ea-kisk\u00eayihtahk t\u00e2nisi \u00ea-ispayiniyik \u00f4ma kin\u00eahiy\u00e2wiwininaw. \u00ea- n\u00f4ht\u00ea-kisk=\u00eayiht-am-k t\u00e2nisi \u00ea- is =payi- n-yi-k C1-want-know=by.mind-TI-3 how C1-thus=INCH-II-DS-0 aw-ima ki-n\u00eahiy\u00e2w=i=w=in -inaw. PROX-IN.SG 2-cree =AI=3=NOM-21PL \u2018[he]PROX wants to understand how [our Cree culture] works\u2019 (Whitecalf 1993:\u00a716) 249 (75) NON-EXISTENTIAL ANTECEDENT = NO \u2013yi\u2013 ON PREDICATE \u14a7\u1450\u14c2 \u144c\u1433 \u146d\u1422\u146d\u14c4\u141a\u1438\u1426\u1455\u1426\u146d \u1406\u14aa \u14c2\u1439\u1475\u14ef\u14a7\u140f\u1472\u14a5\u1420 \u1456\u14c2\u14ef \u1401 \u1403\u1422\u1438\u1528\u1420 mitoni t\u00eapi-kiskinow\u00e2pahtahki \u00f4ma nip\u00e2kw\u00easimowikamik t\u00e2nisi \u00ea-ispayik mitoni t\u00eapi- kisk=inw =w\u00e2p =aht -am-k-i aw-ima intense enough-know=by.hand=see =by.eye.TI-TI -0-SUBJ PROX-IN.SG nip\u00e2kw\u00easimo-w-ikamikw t\u00e2nisi \u00ea-is =payi -k sun.dance -3-place how C1-thus=INCH-0 \u2018When the Whites have watched and learned enough about how the Sundance-Lodge works,\u2019 (Whitecalf 1993:\u00a717) This contrast was understood in terms of perspectival possession; in order for a referent to possess a perspective, they had to be existential (i.e. they had to exist in the discourse). Following this logic in the current case, we can understand the requirement of a possessor to be existential as deriving from the requirement of this referent to possess a perspective. Possession constructions in Plains Cree, then, always assign a perspective to the possessor. 4.4.123. Possession blocks speaker attitudes towards possessum When a speaker wishes to express their opinion about some property of a referent, they cannot use a possession construction, if the possessor is someone other than themselves. Consider the following example from Alice Ahenakew\u2019s reminiscences, shown in (76). In this example, the referent on\u00e2p\u00eama \u2018her husband\u2019 is first introduced as a possessed nominal, and then shifted to an independent form ana n\u00e2p\u00eaw \u2018that man.\u2019 (76) POSSESSED NOMINAL SHIFTED TO INDEPENDENT FORM \u2026\u14ab\u1472 \u1401\u147f \u140a\u14c2\u1426\u1403 \u1405\u14c8\u142f\u14aa, \u1401, \u1401 \u146e \u14c2\u1438\u1426\u1403 \u14aa\u148b \u140a\u1528\u140f\u141f \u140a\u14c7 \u14c8\u142f\u1424, m\u00e2k \u00eakw \u00e2nih \u00f4n\u00e2p\u00eama, \u00eay, \u00ea-k\u00ee-nipahi-mac-\u00e2yiwit ana n\u00e2p\u00eaw, m\u00e2ka \u00eakwa an-ihi o-n\u00e2p\u00eaw=m -a \u00eay \u00ea- k\u00ee- nipahi- maci-ayiw =i -t but then dst-xt 3-man =DSJ-XT INTERJ C1-PREV-to.death-bad- someone=AI-3 an =a n\u00e2p\u00eaw DST=AN.SG man \u2018\u2026but herPROX husbandOBV, hey, [that manAN] had been extremely ill-tempered,\u2019 (Ahenakew \u00a75.4) Here, the Speaker (Alice Ahenakew) has been talking about the Irish woman she worked for, and the woman\u2019s husband secondarily. She has just offered an opinion on the qualities of this woman (she had been very good-natured), and now switches her topic to the man. She starts with a strongly contrastive nominal referring to the man, following by an interjection and a pause. She then re-identifies the man using the non-possessed form ana n\u00e2p\u00eaw \u2018that man.\u2019 250 I asked a consultant (S2) about this construction, and her commentary is illuminating. As the construction stands, the consultant was certain that the Speaker is expressing her own opinion about the man\u2019s qualities (i.e. that she thinks he was very ill-tempered). I then changed the utterance so that there is no independent nominal (and the concurrent switch to simple \u201cAnimate\u201d verbal morphology). (77) INFELICITOUS TO LEAVE THE REFERENT AS POSSESSED # \u2026\u14ab\u1472 \u1401\u147f \u140a\u14c2\u1426\u1403 \u1405\u14c8\u142f\u14aa, \u1401, \u1401 \u146e \u14c2\u1438\u1426\u1403 \u14aa\u148b \u140a\u1528\u140f\u141f \u140a\u14c7 \u14c8\u142f\u1424, # m\u00e2k \u00eakw \u00e2nih \u00f4n\u00e2p\u00eama, \u00eay, \u00ea-k\u00ee-nipahi-mac-\u00e2yiwiyit, m\u00e2ka \u00eakwa an-ihi o-n\u00e2p\u00eaw=m -a \u00eay \u00ea- k\u00ee- nipahi- maci-ayiw =i -yi-t but then dst-xt 3-man =DSJ-XT INTERJ C1-PREV-to.death-bad- someone=AI-DS-3 \u2018\u2026but herPROX husbandOBV, hey, heOBV had been extremely ill-tempered,\u2019 (Presented S2) When the consultant was asked to consider this change, she explained that it would be inappropriate given the context, because \u201cit could make you think that the woman thought her husband was a bad person, but it\u2019s [the Speaker] that thinks that.\u201d In the current model, we have relativized the evaluation of truth to a possessed perspective (see chapter 1). Within this framework, we can understand the different between (76) and (77) as a difference in which perspective the truth of the proposition is interpreted in. Based on the consultant\u2019s discussion, it is clear that the example in (76) is interpreted as true in the Speaker\u2019s perspective; the Speaker takes the man to be a bad person. (78) \u2026\u14ab\u1472 \u1401\u147f \u140a\u14c2\u1426\u1403 \u1405\u14c8\u142f\u14aa, \u1401, \u1401 \u146e \u14c2\u1438\u1426\u1403 \u14aa\u148b \u140a\u1528\u140f\u141f \u140a\u14c7 \u14c8\u142f\u1424, m\u00e2k \u00eakw \u00e2nih \u00f4n\u00e2p\u00eama, \u00eay, \u00ea-k\u00ee-nipahi-mac-\u00e2yiwit ana n\u00e2p\u00eaw, m\u00e2ka \u00eakwa an-ihi o-n\u00e2p\u00eaw=m -a \u00eay \u00ea- k\u00ee- nipahi- maci-ayiw =i -t but then dst-xt 3-man =DSJ-XT INTERJ C1-PREV-to.death-bad- someone=AI-3 an-a n\u00e2p\u00eaw dst-XT man \u2018\u2026but herPROX husbandOBV, hey, [that manAN] had been extremely ill-tempered,\u2019 (Ahenakew \u00a75.4) Speaker x man(x) bad.person(x) By contrast, the consultant\u2019s explanation for the example in (77) suggests that the proposition is true in the woman\u2019s perspective, rather than the Speaker\u2019s; the woman takes her husband to be a bad person. 251 (79) \u2026\u14ab\u1472 \u1401\u147f \u140a\u14c2\u1426\u1403 \u1405\u14c8\u142f\u14aa, \u1401, \u1401 \u146e \u14c2\u1438\u1426\u1403 \u14aa\u148b \u140a\u1528\u140f\u141f \u140a\u14c7 \u14c8\u142f\u1424, m\u00e2k \u00eakw \u00e2nih \u00f4n\u00e2p\u00eama, \u00eay, \u00ea-k\u00ee-nipahi-mac-\u00e2yiwiyit, m\u00e2ka \u00eakwa an-ihi o-n\u00e2p\u00eaw=m -a \u00eay \u00ea- k\u00ee- nipahi- maci-ayiw =i -yi-t but then dst-xt 3-man =DSJ-XT INTERJ C1-PREV-to.death-bad- someone=AI-DS-3 \u2018\u2026but herPROX husbandOBV, hey, heOBV had been extremely ill-tempered,\u2019 (Presented S2) Speaker x y x y bad.person(y) husband(y,x) Thus, the presence\/absence of a possessor construction changes whose perspective the truth of the proposition is evaluated within. When possession is present, the proposition associated with the possessum is neutrally interpreted within the perspective of the possessor. This suggests that possessor constructions induce a perspectival embedding of the possessum. 4.4.124. Possession and knowledge of speech act participants If the use of a possessed nominal implicates that predicates associated with the possessum are evaluated in the perspective of the possessor (\u00a74.4.123), we expect an even stronger effect when using second-person possession. For example, in English, the use of possessor structures is known to induce a presupposition violation in certain contexts. Consider the pair of forms in (80), uttered out of the blue when the speaker has just rushed into the room.7 (80) a. I\u2019m sorry I\u2019m late \u2013 my mother was sick. b. # I\u2019m sorry I\u2019m late \u2013 your mother was sick. (Irene Heim, p.c.) Here, an utterance that involves a referent related to the hearer requires a great deal of supporting context to make it felicitous. In particular, we would need some kind of contextual support to 7 Examples are those of Irene Heim, which she gave in a guest lecture in a semantics class at UBC in the spring of 2003. 252 explain how it is that the Speaker knows more about the Hearer\u2019s mother than the Hearer does (e.g. perhaps I was at her house, visiting, while you were at this meeting that I was late for). One consultant, explained that \u201cIn [the form in (80b)], it sounds like you\u2019re going to tell me I have to do something about this.\u201d A search of Google confirms this restriction for 2nd person possession: among the first 40 instances of \u201cYour mother is sick,\u201d the majority are hypothetical situations (e.g. \u201cSuppose your mother is sick\u2026\u201d). The only two realis contexts for this utterance are (i) a doctor informing the children of their mother\u2019s condition, or (ii) a father informing his children of their mother\u2019s condition. In both types of cases, the reaction on the hearer\u2019s part was one of obligation. This suggests that, at least in English, possession constructions are intimately connected to knowledge states of referents. Within the current model, this means that they are intimately involved with perspective. In this regard, Plains Cree shows exactly the same properties that English does. Possessor constructions related to the Hearer require significant contextual support to be felicitous, and cannot simply be uttered out of the blue. Here, the consultant (S2) rejects the form in (81b) out of the blue. In doing so, she asks how I would know anything about her younger sister. She then creates an imaginary context for (81b), in which I was friends with her younger sibling, and I had called her this morning to find out how she was doing. (81) 2ND PERSON POSSESSION IS INFELICITOUS OUT-OF-THE-BLUE a. \u14c2\u14f0\u14a5\u1422 \u1401 \u140a\u1426\u146f\u14ef\u141f. nis\u00eemis \u00ea-ahkosit. ni-s\u00eemis \u00ea- ahkosi-t 1- younger.sibling C1-sick -3 \u2018My younger brother is sick.\u2019 (Presented S2) b. # \u146d\u14f0\u14a5\u1422 \u1401 \u140a\u1426\u146f\u14ef\u141f. # kis\u00eemis \u00ea-ahkosit. ki-s\u00eemis \u00ea- ahkosi-t 2-younger.sibling C1-sick -3 \u2018Your younger sister is sick.\u2019 (Presented S2) Comment (S2): \u201cHow do you know anything about my younger sister?\u201d Thus, the form that relates the referent to the Hearer requires stronger contextual support than the one that relates the referent to the Speaker. Within the current model, we can understand this property of possession as a condition on perspectival embedding. Neutrally, a referent possessed by the Hearer will be evaluated in the Hearer\u2019s perspective. 253 (82) # \u146d\u14f0\u14a5\u1422 \u1401 \u140a\u1426\u146f\u14ef\u141f. # kis\u00eemis \u00ea-ahkosit. ki-s\u00eemis \u00ea- ahkosi-t 2-younger.sibling C1-sick -3 \u2018Your younger sister is sick.\u2019 (Presented S2) Speaker Hearer x sick(x) sibling(x,H) Embedded in this way, the Speaker is unintentionally suggesting that it is true in the Hearer\u2019s perspective that the referent is sick, equivalent to uttering a sentence like \u201cYou know that your younger sister is sick.\u201d If the Hearer does not know of this proposition, then the claim made about the Hearer\u2019s perspective will fail to be accepted. By contrast, if we set up a strong enough discourse context, the referent (and the proposition) can be embedded within the Speaker\u2019s perspective. (83) \u2026 \u146d\u14f0\u14a5\u1422 \u1401 \u140a\u1426\u146f\u14ef\u141f. \u2026 kis\u00eemis \u00ea-ahkosit. ki-s\u00eemis \u00ea- ahkosi-t 2-younger.sibling C1-sick -3 \u2018Your younger sister is sick.\u2019 (Presented S2) 254 Speaker x sick(x) sibling(x,H) Hearer x sibling(x,H) Possession, then, can induce expectations about whose perspective the proposition is embedded within. Neutrally, the proposition is evaluated within the perspective of the referent identified as the possessor, with strong contextual support necessary to undo this expectation. This suggests that possessors are interpreted as perspective-holders. 4.4.125. Changing possession changes the Speaker Following a long line of research (e.g. Searle 1950, Bar-Hillel 1954), Banfield (1982) considers direct and indirect quotation, concluding that forms that pick out facets of the speech situation (indexicals; Searle 1950, Kaplan 1989, etc.) can only shift their reference when there is a new speaker. For Banfield (1982), there are two ways that the Speaker can change: (i) A change in conversational turns; Some different human being is now uttering linguistic forms (ii) Some other person\u2019s speech is being represented by the same speaker (i.e. indirect quotation) In Plains Cree, possession parallels this shift in indexicality. The use of possessor terms provides crucial information about whether the Speaker is speaking as themselves or representing someone else\u2019s speech. Consider the pair of forms in (84). In a context in which I am talking to a woman who is married to my brother, the sentence in (84a) corresponds to an indirect quotation of her statement, whereas the form in (84b) corresponds to a direct quotation. 255 (84) CHANGING POSSESSOR = CHANGE FROM INDIRECT TO DIRECT QUOTATION a. \u1401 \u140a\u1426\u146f\u14ef\u141f \u14c2\u14f0\u14a5\u1422, \u146d\u144e\u1464\u1423. \u00ea-ahkosit nis\u00eemis, kititw\u00e2n. \u00ea-ahkosi-t ni-s\u00eemis kit-itw\u00ea-n C1-sick-3 1-younger.sibling 2- say -LP \u2018You said that my younger brother is sick.\u2019 (Presented S2) b. \u1401 \u140a\u1426\u146f\u14ef\u141f \u14c2\u14c8\u142f\u1466, \u146d\u144e\u1464\u1423. \u00ea-ahkosit nin\u00e2p\u00eam, kititw\u00e2n. \u00ea- ahkosi-t ni-n\u00e2p\u00ea=m kit-itw\u00ea-n C1-sick -3 1-man =DSJ 2- say -LP \u2018You said, \u201cmy husband is sick.\u201d\u2019 (Presented S2) Thus, the modification of kin terms, from one that relates only to the Speaker to one that relates only to the Hearer changes the type of speech act from indirect quotation to direct. Possession constructions are also crucial for the disambiguation of speakers. Consider the sentence in (85), which has a speaking verb \u00ea-itw\u00eat \u2018sheAN says\u2019 with a pronominal subject. Here, consultants say that the structure is well-formed, but ambiguous; we do not know whether the quotation is of Clare or Martha\u2019s speech. (85) REFERENTIALLY AMBIGUOUS VERB OF SPEAKING Clare \u1401 \u146d\u152a\u1472\u141a\u141f Martha \u1418 \u1401 \u1431\u1433\u14c2\u1528\u1420. \u201c\u14c2\u14a5\u1429\u140d\u1528\u1426\u144c\u1423 \u1406\u14aa,\u201d \u1401 \u1403\u1458\u141f. Clare \u00ea-kiyokaw\u00e2t Marthawa \u00ea-piponiyik. \u201cnimiyw\u00eayihten \u00f4ma,\u201d \u00ea-itw\u00eat. Clare \u00ea- kiyokaw-\u00e2 -t Martha-a \u00ea- pipon- yi-k. Clare C1-visit -DIR-3 Martha-XT C1-winter-DS-0 ni-miyw=\u00eayiht -\u00ea-n aw -ima \u00ea-it =w\u00ea-t 1- good =by.mind.TI-TI-LP PROX-IN.SG C1-thus=AI-3 \u2018ClarePROX went to visit MarthaOBV during the winter. \u201cI am happy about this\u201d, sheAN said.\u201d\u2019 When asked how to disambiguate this, the consultant (S2) volunteered a fix that overtly employed a kin term. (86) AMBIGUITY FIXED WITH A KINSHIP TERM i. Clare \u1401 \u146d\u152a\u1472\u141a\u141f Martha \u1418 \u1401 \u1431\u1433\u14c2\u1528\u1420. Clare \u00ea-kiyokaw\u00e2t Marthawa \u00ea-piponiyik. Clare \u00ea- kiyokaw-\u00e2 -t Martha-a \u00ea- pipon- yi-k. Clare C1-visit -DIR-3 Martha-XT C1-winter-DS-0 \u2018ClarePROX went to visit MarthaOBV during the winter.\u2019 256 ii. \u201c\u14c2\u14a5\u1429\u140d\u1528\u1426\u144c\u1423 \u1406\u14aa \u1401 \u142f \u146d\u152a\u146b\u141f \u14c2\u14a5\u1422,\u201d \u1401 \u1403\u1458\u141f. \u201cnimiyw\u00eayihten \u00f4ma \u00ea-p\u00ea-kiyok\u00eat nimis,\u201d \u00ea-itw\u00eat. ni-miyw=\u00eayiht -\u00ea-n aw -ima \u00ea- p\u00ea- kiyok\u00ea-t ni-mis \u00ea- it =w\u00ea-t 1- good =by.mind.TI-TI-LP PROX-IN.SG c1-come-visit -3 1-older.sister C1-thus=AI-3 \u201cI am happy that my older sisterAN has come to visit,\u201d sheAN said.\u2019 Here, the fixed form inserts reference to my older sister nimis in the quoted material. By doing this, we now know that Martha (who is Clare\u2019s younger sister) is the utterer of the quotation, rather than Clare. Within the model of perspective developed in this thesis, we can understand this shift in possession as a shift in perspectives. The pairwise contrast between direct and indirect speech can be understood as a contrast in the embedding of the referent denoted by the possessum. In indirect speech (84a), the referent nis\u00eemis \u2018my younger brother\u2019 is embedded within the Speaker\u2019s perspective. (87) \u1401 \u140a\u1426\u146f\u14ef\u141f \u14c2\u14f0\u14a5\u1422, \u146d\u144e\u1464\u1423. \u00ea-ahkosit nis\u00eemis, kititw\u00e2n. \u00ea-ahkosi-t ni-s\u00eemis kit-itw\u00ea-n C1-sick-3 1-younger.sibling 2- say -LP \u2018You said that my younger brother is sick.\u2019 (Presented S2) Speaker x younger.brother(S,x) Hearer x sick(x) This has the result of distributing the burden of truth conditions between the Speaker and the embedded perspective (here, identified with the Hearer); the Speaker is saying that they have a brother and that the Hearer claims this brother is sick.8 By contrast, the direct speech case in 8 It may be that the existence of the brother is presupposed rather than asserted, but this difference is not relevant to the current contrast. 257 (84b) locates the referent within the embedded perspective (again identified with the Hearer) as shown in (88). (88) \u1401 \u140a\u1426\u146f\u14ef\u141f \u14c2\u14c8\u142f\u1466, \u146d\u144e\u1464\u1423. \u00ea-ahkosit nin\u00e2p\u00eam, kititw\u00e2n. \u00ea- ahkosi-t ni-w\u00eeki=m =\u00e2 =kan kit-itw\u00ea-n C1-sick -3 1-live=DSJ=DIR=NOM 2- say -LP \u2018You said, \u201cmy husband is sick.\u201d\u2019 (Presented S2) Speaker H x husband(x,H) sick(x) Here, the burden of truth conditions is placed entirely on the embedded perspective; all the Speaker is saying is that the Hearer uttered a specific proposition. Similarly, the use of the possessor construction nimis \u2018my older sister\u2019 in the quotation in (86) tells us whose perspective to embed the proposition in. (89) ii. \u201c\u14c2\u14a5\u1429\u140d\u1528\u1426\u144c\u1423 \u1406\u14aa \u1401 \u142f \u146d\u152a\u146b\u141f \u14c2\u14a5\u1422,\u201d \u1401 \u1403\u1458\u141f. \u201cnimiyw\u00eayihten \u00f4ma \u00ea-p\u00ea-kiyok\u00eat nimis,\u201d \u00ea-itw\u00eat. ni-miyw=\u00eayiht -\u00ea-n aw -ima \u00ea- p\u00ea- kiyok\u00ea-t ni-mis \u00ea- it =w\u00ea-t 1- good =by.mind.TI-TI-LP PROX-IN.SG c1-come-visit -3 1-older.sister C1-thus=AI-3 \u201cI am happy that my older sisterAN has come to visit,\u201d sheAN said.\u2019 258 Speaker Clare Martha visit(C,M) Martha x happy(M) visit(x) older.sister(x,M) Thus, the possession construction is providing crucial information about whose perspective the proposition should be evaluated within. This suggests that possessor constructions are intimately connected to perspectival embedding. 4.4.13. Conclusion: Possession is referential dependence From this discussion, we see that possession in Plains Cree meets both of the conditions necessary for the construal of a referential dependency. First, we saw in section 4.4.11. that the possessum is always structurally c-commanded by the possessum, which means that possession meets the structural condition on referential dependence (\u00a74.1). Second, we saw in section 4.4.12. that there are good reasons to suppose that the possessum is perspectivally-embedded inside the perspective of the possessor, which means that possession meets the structural condition on referential dependence (\u00a74.1). Since possession meets both of these conditions, it should be no surprise that possession systematically correlates with obviation in constructions involving two third-person referents. Recall from chapter 3 (\u00a73.4-5) that the \u201cProximate\u201d referent, which is created via contrast with an \u201cObviative\u201d (\u00a73.4), is the referent whose perspective propositions associated with the \u201cObviative\u201d are evaluated in. This means that, in possession constructions, the possessor will invariably map onto the \u201cProximate\u201d referent (since the possessor is always the perspective holder; \u00a74.4.12) and the possessum will invariably map onto the \u201cObviative\u201d referent (since the possessum is always the referentially-dependent referent). Thus, obviation and possession code much the same kind of information, meaning that their systematic co-occurrence is expected. 259 4.4.2. Theme signs and referential dependency As is well known, Plains Cree transitive verbs are composed out of two pieces (cf. Wolfart 1973, Dahlstrom 1986, Hirose 2000). The last element of the verb stem, termed the \u2018final\u2019 in the Algonquianist literature (cf. Bloomfield 1962), codes that the verbal predicate is transitive, and introduces the verb\u2019s agent (cf. Hirose 2000, D\u00e9chaine 2003). (90) VERB FINAL : \u2013am\u2013 w\u00e2pam- w\u00e2p =am- see =by.eye.TA- ROOT=FINAL- Following the final is another suffix, termed a \u2018theme sign\u2019 (Bloomfield 1962), which codes the other half of the verb\u2019s argument structure (cf. Hirose 2000, D\u00e9chaine 2003, D\u00e9chaine & Reinholtz 2008). (91) THEME SIGN : \u2013\u00ea\u2013 w\u00e2pam\u00ea- w\u00e2p =am -\u00ea see =by.eye.TA-DIR ROOT=FINAL -THEME There are two basic types of theme signs: (i) the \u2018direct\u2019 theme signs, and (ii) the \u2018inverse\u2019 theme sign. (92) a. DIRECT THEME SIGN: -\u00ea- w\u00e2pam\u00ea- w\u00e2p =am -\u00ea see =by.eye.TA-DIR ROOT=FINAL -THEME b. INVERSE THEME SIGN: -ikw w\u00e2pamikw- w\u00e2p =am -ikw see =by.eye.TA-INV ROOT=FINAL -THEME 260 Typically, the alternation between \u201cdirect\u201d and \u201cinverse\u201d is explained in terms of a \u201cperson hierarchy\u201d (cf. Blain 1997, Ritter & Rosen 2005, etc.), but a consideration of the agreement patterns in Plains Cree shows that the predicted blocking patterns (2 > 1 > Prox > Obv > Inan) are not found, making such an explanation untenable. Instead, the theme sign system appears to be a method for solving the reference tracking ambiguities seen in languages like English, where previous argument structure sets up expectations about the reference resolution of anaphora in subsequent clauses. (93) a. Bill saw John, and then he hit him. (Carden p.c.) = Bill hit John. = (?) John hit Bill. Here, the English example has an initial transitive clause, which sets up an expectation about the identity of the two pronominals in the following clause (i.e. that the transitive relation is the sae as the previous clause). If the opposite reference for the pronouns is desired (i.e. that John hit Bill), then the structure typically requires additional context. Comparing this to Plains Cree, we see that the theme sign system explicitly disambiguates these two potential readings. (94) a. \u201cDIRECT\u201d = SAME ARGUMENT STRUCTURE AS PREVIOUS CLAUSE Bill \u141a\u1438\u14a3\u1424 John\u140a, \u1401\u147f \u1401 \u1438\u1472\u14aa\u1426\u141a\u141f. Bill w\u00e2pam\u00eaw Johna, \u00eakwa \u00ea-pakamahw\u00e2t. Bill w\u00e2p=am -\u00ea -w John-a \u00eakwa \u00ea- pakam=ahw -\u00e2 -t Bill see=by.eye.TA-DIR-3 John-XT and C1-hit =by.tool.TA-DIR-3 \u2018Bill saw John, and then he hit him.\u2019 = Bill hit John. \u2260 John hit Bill. (Presented S2) b. \u201cINVERSE\u201d = OPPOSITE ARGUMENT STRUCTURE TO PREVIOUS CLAUSE Bill \u141a\u1438\u14a3\u1424 John\u140a, \u1401\u147f \u1401 \u1438\u1472\u14aa\u1426\u1405\u146f\u141f. Bill w\u00e2pam\u00eaw Johna, \u00eakwa \u00ea-pakamahokot. Bill w\u00e2p=am -\u00ea -w John-a \u00eakwa \u00ea- pakam=ahw -ikw-t Bill see=by.eye.TA-DIR-3 John-XT and C1-hit =by.tool.TA-INV -3 \u2018Bill saw John, and then he hit him.\u2019 \u2260 Bill hit John. = John hit Bill. (Presented S2) Here, the choice of a theme sign for the subsequent clause tells us who is hitting who. Plains Cree\u2019s theme sign system, then, is overtly coding something that English has to do with context. 261 The theme sign system of Plains Cree has an intimate relationship to obviation, though none of the morphemes can be said to code obviation directly (cf. \u00a73.33). Given a verb that involves a \u201cProximate\u201d and an \u201cObviative\u201d, the direct theme sign \u2013\u00ea\u2013 maps \u201cProximate\u201d onto subject (95a), and the inverse theme sign \u2013ikw maps \u201cObviative\u201d onto subject (95b). Here, the suffix \u2013\u00e2\u2013 is used when the \u201cProximate\u201d referent aw\u00e2sis \u2018the child\u2019 is the subject, and the inverse theme sign \u2013ikw is used when the \u201cObviative\u201d referent atimwa \u2018the dog\u2019 is the subject. (95) a. \u201cPROXIMATE\u201d ACTS ON \u201cOBVIATIVE\u201d = \u2013\u00ea\u2013 \u141a\u1438\u14a3\u1424 \u140a\u141a\u14ef\u1422 \u140a\u144e\u14b7. w\u00e2pam\u00eaw aw\u00e2sis atimwa. w\u00e2p=am -\u00ea -w aw\u00e2sis atimw-a see=by.eye.TA-DIR-3 child dog -XT = \u2018The childPROX sees the dogOBV.\u2019 (Presented S2) \u2260 \u2018The the dogOBV sees childPROX.\u2019 b. \u201cOBVIATIVE\u201d ACTS ON \u201cPROXIMATE\u201d = \u2013ikw\u2013 \u140a\u141a\u14ef\u1422 \u141a\u1438\u14a5\u1420 \u140a\u144e\u14b7. aw\u00e2sis w\u00e2pamik atimwa. aw\u00e2sis w\u00e2p=am -ikw atimw-a child see=by.eye.TA-INV dog -XT \u2260 \u2018The childPROX sees the dogOBV.\u2019 = \u2018The the dogOBV sees childPROX.\u2019 (Presented S2) Given the close relationship between theme signs and obviation, we expect that theme signs should code referential dependency. This means that they should meet both of the conditions on referential dependency: (i) STRUCTURAL CONDITION: They should show either c-command or linear precedence between the dependent referent and its antecedent. (ii) SEMANTIC CONDITION: They should perspectivally-embed this same structurally- dependent referent in its antecedent. All that remains is to evaluate the theme signs\u2019 behaviour to see if these expectations are borne out. I first consider the direct theme sign -\u00ea-. In section 4.4.211, I consider the structural conditions on this theme sign, arguing that it always occurs when the argument it introduces is structurally subordinated to the subject of the verb, which is introduced by the verb\u2019s final (cf. Hirose 2000, etc.). This means that it meets the structural condition on referential dependence. Then, in section 4.4.212, I turn to the semantic conditions on the use of this theme sign. In so 262 doing, I show that the argument introduced by this theme sign is always embedded in the perspective of its antecedent - the subject of the verb. As such, the direct theme sign \u2013\u00ea\u2013 meets both the structural and semantic conditions on referential dependency. Its use in obviation constructions thus becomes transparent. I then turn to the inverse theme sign \u2013ikw. In section 4.4.221, I examine the structural properties of this theme sign. I show that, in contrast to the direct theme sign, it is not structurally-subordinated to the argument introduced by the final. Instead, it is c-commanded by the referent in the highest position of the predicate \u2013 sometimes called the \u2018topic\u2019 operator (cf. D\u00e9chaine & Reinholtz 2008). This means that the inverse theme sign meets the structural condition on referential dependence. Following this, I consider the semantic conditions on the inverse theme sign in section 4.4.222. I demonstrate that the argument introduced by the theme sign is always embedded within the perspective of its antecedent \u2013 the \u2018topic\u2019 operator. This means that the inverse theme sign \u2013ikw also meets both conditions on referential dependency. 4.4.21. The direct theme sign \u2013\u00ea\u2013 constructs referential dependency In this thesis, the theme sign \u2013\u00ea\u2013 is considered to occur in three constructions of the independent order: (i) speech-act participants acting on \u201cInanimate\u201d referents (96a), \u201cProximate\u201d referents acting on \u201cObviative\u2019s\u201d (96b), and \u201cObviative\u201d referents acting on \u201cObviative\u201d referents (96c). (96) a. SPEECH-ACT PARTICIPANT ACTS ON \u201cINANIMATE\u201d : \u2013\u00ea\u2013 \u14c2\u141a\u1438\u1426\u144c\u1423 niw\u00e2paht\u00ean ni-w\u00e2p=aht -\u00ea -n 1- see =by.eye.TI-DIR-LP \u2018I see itIN.\u2019 (Presented S2) b. \u201cPROXIMATE\u201d ACTS ON \u201cOBVIATIVE\u201d : \u2013\u00ea\u2013 \u141a\u1438\u14a3\u1424 w\u00e2pam\u00eaw w\u00e2p=am -\u00ea -w see=by.eye.TA-DIR-3 \u2018s\/hePROX sees him\/herOBV.\u2019 (Presented S2) 263 c. \u201cOBVIATIVE\u201d ACTS ON \u201cOBVIATIVE\u201d : \u2013\u00ea\u2013 \u141a\u1438\u14a3\u1528\u1418 w\u00e2pam\u00eayiwa w\u00e2p=am -\u00ea -yi-w-a see=by.eye.TA-DIR-DS-3-XT \u2018s\/heobv sees him\/herobv\u2019 (Wolfart 1996) Treating this as the distribution of \u2013\u00ea\u2013 means that, compared to the hypothesized proto-forms (cf. Goddard 1974, 1983, 2008), Plains Cree has collapsed two homophonous theme signs together; the TI theme \u273d \u2013\u00ea\u2013 and the TA theme \u273d \u2013\u00ea\u2013. As is shown in the three examples above (96), the theme sign \u2013\u00ea\u2013 is used in \u201cObviative\u201d constructions. As such, the current analysis expects that it should obey both of the conditions on referential dependency: (i) STRUCTURAL CONDITION: It should introduce either c-command or linear-precedence requirements for its referent (\u00a74.1). (ii) SEMANTIC CONDITION: It should require the referent associated with it to be embedded in the perspective of the referent it is structurally-dependent on (\u00a74.1). In the following sections, I demonstrate that this is exactly what the theme sign \u2013\u00ea\u2013 does. As such, it can be considered to code referential dependency in Plains Cree. 4.4.211. Structural conditions on \u2013\u00ea\u2013 Numerous linguists have analyzed the Plains Cree verb system as being organized into a series of hierarchically-related heads. Hirose (2000) as modified by D\u00e9chaine (2003), for example, puts the transitive verb\u2019s two argument affixes (the final and the theme) in the two heads of the decomposed VP (vP and VP, respectively). (97) vP 3 ROOT 3 pro 3 FINAL VP 3 AA pro 3 THEME 264 As such, the theme signs are expected to always enter into c-command; specifically, the argument introduced by the final should always structurally-dominate the argument introduced by the theme sign. If this is so, all that remains to be shown is the specifics regarding the theme sign \u2013\u00ea\u2013. While all the direct theme signs could be analyzed as occupying this lower head position, there are particularly good reasons to treat \u2013\u00ea\u2013 as occupying this lower head. Recall from Chapter 3 (\u00a73.3.3) that the direct theme sign \u2013\u00ea\u2013 can be analyzed as introducing an extentional object. (98) vP 3 ROOT 3 3 FINAL VP 3 AA pro 3 -\u00ea- [EXT] In such a model, the further discrimination between inherent and contextual extentionality (i.e. selection between an \u201cInanimate\u201d and \u201cObviative\u201d object) is achieved by the restrictive relation between the final and the theme sign; the choice of finals decides whether the object is inherently or contextually extentional. (99) a. niw\u00e2paht\u00ean ni-w\u00e2p=aht-\u00ea-n 1-see=by.eye.TI-DIR-LP \u2018I seeIN it.\u2019 vP 3 w\u00e2p- 3 pro 3 -aht- VP = \u201cINANIMATE\u201d OBJECT [EXT] 3 AA pro 3 -\u00ea- [EXT] 265 b. w\u00e2pam\u00eaw w\u00e2p=am -\u00ea -w see =by.eye.TA-DIR-3 \u2018S\/hePROX sees him\/herOBV.\u2019 vP 3 w\u00e2p- 3 pro 3 -am- VP = \u201cOBVIATIVE\u201d OBJECT [\u2205] 3 AA pro 3 -\u00ea- [EXT] Since the final and the theme sign are in a c-command relation, the logic of the structure entails that the arguments they introduce are in the same relation; the argument introduced by the final c-commands the argument introduced by the theme sign. (100) DIRECT THEME SIGN (\u2013\u00ea\u2013) : C-COMMAND VP 3 ROOT 3 pro 3 FINAL VP 3 pro 3 -\u00ea- This means that the theme sign \u2013\u00ea\u2013 satisfies the structural condition on referential dependency, via c-command. 4.4.212. Semantic conditions on \u2013\u00ea\u2013 Since the final informs the content of the theme sign \u2013\u00ea\u2013 via c-command, it is expected that the argument introduced by the final has a correlated effect on the argument introduced by the theme sign. In fact, when we consider the semantics of the relation between the two arguments, we see exactly this kind of correlation: the argument introduced by the theme sign is always interpreted within the perspective of the argument introduced by the final. This is demonstrated both by considering when the theme sign \u2013\u00ea\u2013 occurs, and when it cannot occur. 266 4.4.2121. \u2013\u00ea\u2013 occurs with perspectivally-embedded objects There are three contexts that allow the occurrence of \u2013\u00ea\u2013. First, it occurs when the subject of the verb is a speech-act participant, and the object is \u201cInanimate\u201d. (101) a. \u14c2\u141a\u1438\u1426\u144c\u1423 niw\u00e2paht\u00ean ni-w\u00e2p=aht -\u00ea -n 1- see =by.eye.TI-DIR-LP \u2018I see itIN.\u2019 (Presented S2) b. \u14c2\u141a\u1438\u1426\u144c\u1423 kiw\u00e2paht\u00ean ki-w\u00e2p=aht -\u00ea -n 1- see =by.eye.TI-DIR-LP \u2018You see itIN.\u2019 (Presented S2) Second, it occurs when the subject of a verb is \u201cProximate\u201d, and the object is \u201cObviative\u201d. (102) \u141a\u1438\u14a3\u1424 w\u00e2pam\u00eaw w\u00e2p=am -\u00ea -w see=by.eye.TA-DIR-3 \u2018s\/hePROX sees him\/herOBV.\u2019 (Presented S2) Finally, it occurs when both the subject and the object of the verb are \u201cObviative\u201d, although I have been unable to get a consultant to recognize or produce this form, and it does not occur in any of the four large texts I have searched (Ahenakew 2000, Minde 1997, K\u00e2-pimw\u00eaw\u00eahahk 1998, and Whitecalf 1993). This means that my only sources for it are paradigms supplied by grammatical descriptions (e.g. Wolfart 1973, Wolfart & Carroll 1973, Wolfart 1996). (103) \u141a\u1438\u14a3\u1528\u1418 w\u00e2pam\u00eayiwa w\u00e2p=am -\u00ea -yi-w-a see=by.eye.TA-DIR-DS-3-XT \u2018s\/heobv sees him\/herobv\u2019 (Wolfart 1996) Lacking data, we can only hypothesize about what the \u201cObviative\u201d acting on \u201cObviative\u201d constructions mean, and wait until data on these forms materializes. 267 First and second persons are uncontroversially capable of possessing perspectives (cf. Banfield 1982). When we check these forms with consultants, the results confirm this expectation. (104) a. \u14c2\u1432\u146f\u14c0\u1423 \u141a\u1438\u14a7\u1423. nip\u00eekon\u00ean w\u00e2pamon. ni-p\u00eekw =n -\u00ea -n w\u00e2pamon 1- break=by.hand-DIR-LP mirror \u2018I broke the mirrorIN.\u2019 (Presented S2) b. \u146d\u1432\u146f\u14c0\u1423 \u141a\u1438\u14a7\u1423. kip\u00eekon\u00ean w\u00e2pamon. ki-p\u00eekw =n -\u00ea -n w\u00e2pamon 2- break=by.hand-TI-LP mirror \u2018You broke a mirrorIN.\u2019 (Presented S2) The first-person form in (104a) is interpreted such that the Speaker knows that they broke a mirror, and cannot be felicitously uttered if the Speaker is unsure either if it was a mirror that was broken or if the mirror indeed broke. By contrast, consultants reject the second person form unless the situation is set up such that the Hearer has already informed the speaker that the proposition expressed here is true. Otherwise, the form becomes interpreted as harsh, or aggressive, and would be used if \u201cmaybe the person was denying that they broke the mirorr, and you were telling them that you know they did it\u201d (S2). Often, these 2nd person forms are only interpreted as questions that lack question coding (e.g. c\u00ee), and declarative statements are impossible to imagine. (105) 2ND PERSON AGENT FORCES QUESTION INTERPRETATION \u146d\u144e\u144c\u1528\u1426\u144c\u1423 \u1401 \u14c2\u1472\u14a7\u141f \u140a\u141a\u14ef\u1422. kitit\u00eayiht\u00ean \u00ea-nikamot aw\u00e2sis. kit-it =\u00eayiht -\u00ea -n \u00ea-nikamo-t aw\u00e2sis 2- thus=by.mind.TI-DIR-LP C1-sing -3 child = \u2018(Do) You think the children are singing?\u2019 \u2260 \u2018You think the children are singing.\u2019 (Presented S2) This elicitation data is confirmed by a survey of texts, where 2nd person subjects with -\u00ea- are typically only used with future markers in strong commands, such as the bear\u2019s semi- prophetic promise in (106a), without future-coding, such as the mind-reading contexts in (106b- c), or when the Speaker is repeating somehthing that the Hearer has already established (106d). 268 (106) a. CERTAINTY THAT HEARER WILL ACT WITH VERB BEARING \u2013\u00ea\u2013 \u201c\u2018\u14ab\u1472 \u14a5\u14f4\u141a\u1428 \u146b\u152e\u1431\u1428 \u146d\u1472 \u1403\u1451\u144c\u1423 \u1406\u14aa \u1473 \u142f \u1403\u144e\u1456\u1423,\u2019 \u14c2\u144e\u144e\u1420,\u201d \u2026 \u201c\u2018m\u00e2ka misaw\u00e2c k\u00eay\u00e2pic kika-t\u00f4t\u00ean \u00f4ma k\u00e2-p\u00ea-itit\u00e2n,\u2019 nititik,\u201d \u2026 m\u00e2ka misaw\u00e2c k\u00eay\u00e2pic ki-ka- it\u00f4t-\u00ea -n aw =ima k\u00e2-p\u00ea- it -it -\u00e2n nit-it-ikw but in.any.case still 2- FUT-do -TI-LP PRX=IN.SG C2-come-RR-1>2-1 1-RR-INV \u201cbut, in any case, you will do it nevertheless, that which I have come to tell you\u201d , \u2026 (Ahenakew 2000 4:3) b. READING HEARER\u2019S PERSON\u2019S MIND WITH VERB BEARING \u2013\u00ea\u2013 \u201c\u2018\u146d\u1456\u1423\u140d\u1426\u1455\u140f\u1423 \u1406\u14aa, \u14ab\u1472 \u146b\u152e\u1431\u1428 \u146d\u1472 \u1403\u1451\u144c\u1423 \u1406\u14aa \u1473 \u142f\u1403\u144e\u1456\u1423!\u2019 \u2026\u201d \u201c\u2018kit\u00e2nw\u00eahtawin \u00f4ma, m\u00e2ka k\u00eay\u00e2pic kika-t\u00f4t\u00ean \u00f4ma k\u00e2-itit\u00e2n!\u2019 \u2026\u201d kit-\u00e2nw\u00ea=htaw -i -n aw=ima m\u00e2ka k\u00eay\u00e2pic 2-doubt =by.ear.TA-2>1-LP PRX=IN.SG but still ki-ka-it\u00f4t-\u00ea-n aw =ima k\u00e2 -it -it -\u00e2n 2-fut-do-TI-LP PRX=IN.SG c2-RR-1>2-1 \u201c\u2018You are doubting me, but you will nevertheless do what I have told you!\u2019 \u2026\u201d (Ahenakew 2000 4:3) c. \u2018\u146d\u144e\u144c\u1528\u1426\u144c\u1423 \u1406\u14aa, \u201c\u1456\u14c2\u14ef \u1406\u14aa \u146b\u14ef \u1473\u1426\u148b\u144e\u14c2\u1456\u1423 \u1406\u14aa, \u2026 \u2018kitit\u00eayiht\u00ean \u00f4ma, \u201ct\u00e2nis \u00f4ma k-\u00easi-k\u00e2hcitinit\u00e2n \u00f4ma, \u2026 kit-it =\u00eayiht -\u00ea -n aw =ima t\u00e2nisi aw=ima 2-thus=by.mind.TI-DIR-LP PRX=IN.SG how PRX=IN.SG k\u00e2-isi- k\u00e2hciti =n -iti -\u00e2n aw=ima c2-thus-get.ahold=by.hand-1>2-1 PRX=IN.SG \u2018You think, \u201cHow will I get hold of you, \u2026\u201d (Ahenakew 2000 4:4) d. REPEATING HEARER\u2019S THOUGHTS WITH VERB BEARING \u2013\u00ea\u2013 \u201c\u14aa \u14ab\u1528\u14f0\u1426\u1456\u1418\u1420,\u201d \u146d\u144e\u144c\u1528\u1426\u144c\u1423. \u201cma-m\u00e2yis\u00eeht\u00e2wak,\u201d kitit\u00eayiht\u00ean. ma-m\u00e2yi=os =\u00eeht\u00e2-w-ak kit-it =\u00eayiht -\u00ea -n red-bad =make=do -3-PL 2-thus=by.mind.TI-DIR-LP \u2018\u201cThey make poor juice,\u201d you thought.\u2019 (Ahenakew 2000 3:4) In strong contrast, 1st person subjects with -\u00ea- are numerous in texts, and often occur with propositional attitude verbs like psych verbs (107), confirming the pattern expected by the elicitation judgments. (107) SPEAKER EXPRESSING OWN MENTAL STATE WITH VERB BEARING \u2013\u00ea\u2013 \u00eakwa m\u00eena p\u00eayak nik\u00ee-koskw\u00eayiht\u00ean m\u00een \u00ea-nan\u00e2tawih\u00e2t, \u2026 \u00eakwa m\u00eena p\u00eayak ni-k\u00ee- koskw =\u00eayiht -\u00ea -n m\u00eena \u00ea- nan\u00e2taw=ih -\u00e2 -t and also one 1- PREV-surprise=by.mind.TI-DIR-LP also C1-doctor =by.neut-DIR-3 \u2018and I was also amazed at another case of his doctoring someone, \u2026\u2019 (Ahenakew 4:15) 269 This is exactly as expected if the object is interpreted within the perspective of the subject in these cases.9 (108) a. SPEAKER AS SUBJECT : TRUE IN SPEAKER\u2019S PERSPECTIVE \u14c2\u1432\u146f\u14c0\u1423 \u141a\u1438\u14a7\u1423. nip\u00eekon\u00ean w\u00e2pamon. ni-p\u00eekw =n -\u00ea -n w\u00e2pamon 1- break=by.hand-DIR-LP mirror \u2018I broke the mirrorIN.\u2019 (Presented S2) Speaker x mirror(x) EXT(x) break(S,x) 9 I am here putting the predicate in both the Speaker\u2019s perspective and the embedded perspective because of the semantics of independent-order clauses (cf. Cook 2008); these clauses are always evaluated as true at least in the Speaker\u2019s perspective. 270 b. HEARER AS SUBJECT : TRUE IN HEARER\u2019S PERSPECTIVE \u146d\u1432\u146f\u14c0\u1423 \u141a\u1438\u14a7\u1423. kip\u00eekon\u00ean w\u00e2pamon. ki-p\u00eekw =n -\u00ea -n w\u00e2pamon 2- break=by.hand-TI-LP mirror \u2018You broke a mirrorIN.\u2019 (Presented S2) Speaker x mirror(x) EXT(x) break(H,x) Hearer x mirror(x) break(H,x) Thus, the interpretational restrictions and contextual distribution of forms coding 1st and 2nd person strongly suggest that these forms pick out perspectives for the subject. Turning to the cases of \u201cProximate\u201ds acting on \u201cObviative\u201ds, I have already argued for independent reasons that the \u201cProximate\u201d referent possesses a perspective in which the \u201cObviative\u201d referent is evaluated (\u00a73.4-5). Thus, a verb marked with -\u00ea- that involves a \u201cProximate\u201d subject acting on an \u201cObviative\u201d is going to have the object embedded inside the subject. (109) \u14c8\u142f\u1424 \u141a\u1438\u14a3\u1424 \u140a\u141a\u14ef\u14f4. n\u00e2p\u00eaw w\u00e2pam\u00eaw aw\u00e2sisa. n\u00e2p\u00eaw w\u00e2p=am -\u00ea -w aw\u00e2sis-a man see =by.eye.TA-DIR-3 child -XT \u2018The manPROX saw the childOBV.\u2019 (Presented S2) 271 Speaker x y C man(x) child(y) EXT(y,C) see(x,y,C) x y child(y) see(x,y) If we pushed this analysis, we would then predict that the \u201cObviative\u201d acting on \u201dObviative\u201d forms like (110) would mean that the subject possessed an embedded perspective in which the object is evaluated. (110) \u1403\u1422\u1475\u1418 \u141a\u1438\u14a3\u1528\u1418 \u1405\u144c\u14a5\u1528\u1418. \u2026iskw\u00eawa w\u00e2pam\u00eayiwa ot\u00eamiyiwa. iskw\u00eaw-a w\u00e2p=am -\u00ea -yi-w-a o-t\u00eam-yi-w-a woman-XT see=by.eye.TA-DIR-DS-3-XT 3-dog-DS-3-XT \u2018the womanOBV sees herOBV horseOBV\u2019 (cf. Wolfart 1996) 272 Speaker x y z C EXT(y,C) EXT(z,C) see(z,C) x y woman(y) y z horse(z,y) see(y,z) Thus, the \u201cProximate\u201d \u201dObviative\u201d forms confirm the expectation that the object introduced by \u2013 \u00ea\u2013 is embedded in the perspective of the verb\u2019s subject, while we know what to look for with \u201cObviative\u201d \u201dObviative\u201d interactions. Taken altogether, then, we see that \u2013\u00ea\u2013 consistently occurs in contexts in which the subject possesses a perspective for the evaluation of the object. This is expected if \u2013\u00ea\u2013 codes referential dependency; the object is c-commanded by the subject (\u00a74.4.211), and thus it must be the subject that this referent is perspectivally-dependent on. So far, our expectations for \u2013\u00ea\u2013 have been fulfilled. 4.4.2122. \u2013\u00ea\u2013 cannot occur when the subject is not a perspective holder The distribution of \u2013\u00ea\u2013 is only the positive half of the puzzle, and does not constitute definitive evidence for a treatment of \u2013\u00ea\u2013 as meeting the semantic constraints on referential dependency. In order to provide a definitive case, we must also consider where \u2013\u00ea\u2013 is not allowed, and ask if these restrictions follow from its posited referential dependency. Put another way, the positive data shows the sufficient conditions, but the addition of the negative data adds the necessary conditions to this. 273 In fact, there are two environments that disallow the application of \u2013\u00ea\u2013 : (i) contexts in which 1st and 2nd persons are not perspective holders, and (ii) contexts in which 3rd persons are not perspective holders. In each case, the absence of \u2013\u00ea\u2013 correlates with the absence of a perspective for the subject of the verb. This gives us both the necessary and sufficient conditions for the use of this suffix. Crucially, all of the contexts for the occurrence of the theme \u2013\u00ea\u2013 are in the Independent Order of verbal inflection. In fact, this form can never occur in any conjunct order verb. (111) ONLY INDEPENDENT ORDER \u201cPROXIMATE-OBVIATIVE\u201d INTERACTIONS USE \u2013\u00ea\u2013 a. \u141a\u1438\u14a3\u1424. w\u00e2pam\u00eaw. w\u00e2p=am -\u00ea -w see=by.eye.TA-DIR-3 \u2018s\/hePROX sees him\/herOBV.\u2019 (Presented S2) b. \u273d \u1401 \u141a\u1438\u14a3\u141f. \u273d \u00ea-w\u00e2pam\u00eat \u00ea-w\u00e2p=am -\u00ea -t c1-see=by.eye.TA-DIR-3 Intended: \u2018s\/hePROX sees him\/herOBV.\u2019 (Presented S2) (112) ONLY INDEPENDENT ORDER SPEECH-ACT-PARTICIPANT INTERACTIONS USE \u2013\u00ea\u2013 a. \u14c2\u141a\u1438\u1426\u144c\u1423. niw\u00e2paht\u00ean. ni-w\u00e2p=aht -\u00ea -n 1- see =by.eye.TI-DIR-LP \u2018I see itIN.\u2019 (Presented S2) b. \u273d \u1401 \u141a\u1438\u1426\u144c\u152e\u1423 \u273d \u00ea-w\u00e2paht\u00eay\u00e2n w\u00e2p=aht -\u00ea -y\u00e2n see =by.eye.TI-DIR-1 Intended: \u2018I see itIN.\u2019 (Presented S2) As Cook (2008) has shown, verbs in these two orders differ as to their ability to be embedded; Independent Order cannot be subordinated to another verb, while conjunct order verbs can. For example, the independent order verb kimiwan \u2018it rains\u2019 cannot be subordinated with respect to another clause (113a), but the conjunct order form \u00ea-kimiwahk \u2018it rains\u2019 can (113b). 274 (113) INDEPENDENT ORDER CANNOT BE SUBORDINATED a. \u273d \u14c2\u146d\u1422\u146b\u1528\u1426\u144c\u1423 \u146d\u14a5\u1418\u1423. \u273d nikisk\u00eayiht\u00ean kimiwan. ni-kisk=\u00eayiht -\u00ea -n kimiwan 1-know=by.mind.TI-TI-LP rain Intended: \u2018I know that it\u2019s raining\/rained.\u2019 (Cook 2008) b. \u14c2\u146d\u1422\u146b\u1528\u1426\u144c\u1423 \u1401 \u146d\u14a5\u1418\u1541. nikisk\u00eayiht\u00ean \u00ea-kimiwahk. ni-kisk =\u00eayiht -\u00ea -n \u00ea- kimiwan-k 1-know=by.mind.TI-TI-LP c1-rain -0 \u2018I know that it\u2019s raining\/rained.\u2019 (Cook 2008) This has the result of preventing these Independent Order verbs from being made the object of propositional attitude verbs, since, as in English, these always induce c-command. The conjunct order, by contrast, has no embedding restrictions (cf. Wolfart 1996, Cook 2008). It can occur in matrix environments, or embedded environments equally well, as shown by the grammaticality of the double conjunct sentence in (114). (114) CONJUNCT ORDER CAN BE EITHER MATRIX OR SUBORDINATE CLAUSE \u1401 \u146d\u1422\u146b\u1528\u1426\u1455\u14ab\u1423 \u1401 \u146d\u14a5\u1418\u1541. \u00ea-kisk\u00eayihtam\u00e2n \u00ea-kimiwahk. \u00ea-kisk\u00eayihtam\u00e2n \u00ea- kimiwan-k c1-know=by.mind.TI-TI-1 C1-rain -0 \u2018I know that it is raining\/rained.\u2019 (Presented S2) This has the result that verbs in the conjunct order can be embedded under propositional attitude verbs. The subject of this verb can be any kind of perspective-holding referent: first, (second),10 or third persons. (115) CONJUNCT ORDER CAN BE SUBORDINATED W.R.T. A PROPOSITIONAL ATTITUDE a. \u14c2\u144e\u144c\u1528\u1426\u144c\u1423 \u1401 \u14c2\u1472\u14a7\u141f \u140a\u141a\u14ef\u1422. nitit\u00eayiht\u00ean \u00ea-nikamot aw\u00e2sis. nit-it =\u00eayiht -\u00ea -n \u00ea-nikamo-t aw\u00e2sis 1- thus=by.mind.TI-DIR-LP C1-sing -3 child \u2018I think that a child is singing.\u2019 (Volunteered S2) 10 Because of the restrictions on 2nd person seen above, the 2nd person subject with intentional verbs (e.g. kitit\u00eayiht\u00ean \u2018you think thus of it\u2019) is extremely difficult to get in elicitation. Hence, it is left out of this discussion. 275 b. \u1403\u1422\u1475\u1424 \u1401\u144c\u1528\u1426\u1455\u1466 \u1401 \u14c2\u1472\u14a7\u1528\u141f \u140a\u141a\u14ef\u14f4. iskw\u00eaw it\u00eayihtam \u00ea-nikamoyit aw\u00e2sisa. iskw\u00eaw it =\u00eayiht -\u00ea -n \u00ea-nikamo-yi -t aw\u00e2sis-a woman thus=by.mind.TI-DIR-LP C1-sing -DS-3 child-XT \u2018the womanPROX thinks the childOBV is singing.\u2019 (Volunteered S2) This means that, in the conjunct mode, 1st and 2nd person subjects are not guaranteed to be perspective holders; they could be embedded under a propositional attitude connected with some other referent. (116) a. 1ST PERSON SUBJECT SUBORDINATED TO PROPOSITIONAL ATTITUDE Shujun \u1403\u144c\u1528\u1426\u1455\u1466 \u1401 \u141a\u1438\u14aa\u1420. Shujun it\u00eayihtam \u00ea-w\u00e2pamak. Shujun it=\u00eayiht -am \u00ea- w\u00e2p=am -ak Shujun thus=by.mind-TI C1-see =by.eye.TA-1>3 \u2018Shujun thinks that I saw her.\u2019 (Presented S2) b. 2ND PERSON SUBJECT SUBORDINATED TO PROPOSITIONAL ATTITUDE Shujun \u1403\u144c\u1528\u1426\u1455\u1466 \u146e\u152d \u1401 \u141a\u1438\u14aa\u141f. Shujun it\u00eayihtam k\u00eeya \u00ea-w\u00e2pamat. Shujun it =\u00eayiht -am k-\u00eeya \u00ea- w\u00e2p=am -at Shujun thus=by.mind-TI 2-body C1-see=by.eye.TA-2>3 \u2018Shujun thinks that you saw her.\u2019 (Presented S2) c. 3RD PERSON SUBJECT SUBORDINATED TO PROPOSITIONAL ATTITUDE Shujun \u1403\u144c\u1528\u1426\u1455\u1466 \u1401 \u141a\u1438\u14ab\u141f. Shujun it\u00eayihtam \u00ea-w\u00e2pam\u00e2t. Shujun it =\u00eayiht -am \u00ea- w\u00e2p=am -\u00e2 -t Shujun thus=by.mind-TI C1-see =by.eye.TA-DIR-3 \u2018ShujunPROX thinks that shePROX saw him\/herOBV.\u2019 (Presented S2) Here, the subject of the embedded verb is a speech-act participant, but it is not necessary that this proposition be evaluated in the perspective of that referent. The Speaker in (116a) or the Hearer in (116b) may have been unaware of this event, or think it did not happen (i.e. it could be true for Shujun\u2019s perspective, but false for the Speaker). This is markedly different than independent- order forms, where 1st and 2nd person necessarily introduce their perspectives on the proposition (see above). In fact, the only way for the subject of a conjunct-order verb to necessarily be a perspective-holder is if that referent is identified in the matrix clause as one. 276 (117) a. 1ST PERSON SUBJECT SUBORDINATED TO 1ST PERSON PROPOSITIONAL ATTITUDE \u14c2\u144e\u144c\u1528\u1426\u144c\u1423 \u1403\u144c\u1528\u1426\u1455\u1466 \u1401 \u141a\u1438\u14a5\u141f Shujun. nitit\u00eayiht\u00ean \u00ea-w\u00e2pamak Shujun. nit-it=\u00eayiht -\u00ea -n \u00ea- w\u00e2p=am -ak Shujun Shujun thus=by.mind-TI-LP C1-see =by.eye.TA-1>3 Shujun \u2018I think that I saw Shujun.\u2019 (Presented S2) c. 3RD PERSON SUBJECT SUBORDINATED TO 3RD PERSON PROPOSITIONAL ATTITUDE Shujun \u1403\u144c\u1528\u1426\u1455\u1466 \u1401 \u141a\u1438\u14ab\u141f. Shujun it\u00eayihtam \u00ea-w\u00e2pam\u00e2t. Shujun it =\u00eayiht -am \u00ea- w\u00e2p=am -\u00e2 -t Shujun thus=by.mind-TI C1-see =by.eye.TA-DIR-3 \u2018ShujunPROX thinks that shePROX saw him\/herOBV.\u2019 (Presented S2) This means that the perspectival properties of the subjects of conjunct-order clauses are accidental \u2013 context can always over-ride them. Summarizing, we see that the clause types that do not confer a perspective on the subject of the verb are the exact set of clause types that do not allow the affixation of the theme sign \u2013\u00ea\u2013. If \u2013\u00ea\u2013 codes that the object of the verb is embedded inside of the perspective of the subject, this behaviour is expected; a clause-type that is used where non-specification of perspective is desired (i.e. embedded environments; cf. Cook 2008) would not want to use a theme sign that codes exactly this specification. 4.4.213. Conclusion: \u2013\u00ea\u2013 constructs referential dependency When we consider of both the structural and semantic properties of the theme sign \u2013\u00ea\u2013, we find properties that look like those of referential dependency. First, this theme sign always and only occurs in in environments where it is structurally- subordinated to the subject of the verb (\u00a74.4.211). 277 (118) DIRECT THEME SIGN (\u2013\u00ea\u2013) : C-COMMAND VP 3 ROOT 3 pro 3 FINAL VP 3 pro 3 -\u00ea- According to the formulation of referential dependency in section, this means that the theme sign obeys the structural condition on referential dependency. Second, as we saw in \u00a74.4.211, this theme sign always and only occurs in constructions where its argument is embedded in the perspective of the dominating referent (i.e. the subject). This inherent subject-oriented embedding means that the theme sign also obeys the semantic condition on embedding; the dependent referent must be embedded in the perspective of its antecedent. The sensitivity to both of the conditions on referential dependency strongly suggests that this morpheme codes referential dependency, and that it does so independently of its relation to obviation constructions. This is what is expected under the model of referential dependency constructed here. 4.4.22. The inverse theme sign \u2013ikw codes referential dependency The theme sign \u2013ikw\u2013 occurs in several constructions, including \u201cObviative\u201d referents acting on \u201cProximate\u201ds (119), (119) \u201cObviative\u201d acts on \u201cProximate\u201d : \u2013\u00ea\u2013 a. \u2026\u141a\u1438\u14a5\u1420. w\u00e2pamik w\u00e2p=am -ikw see=by.eye.TA-INV \u2018s\/heOBV sees him\/herPROX.\u2019 (Presented S2) b. \u2026\u1401 \u141a\u1438\u14a5\u146f\u141f. \u00ea-w\u00e2pamikot \u00ea-w\u00e2p=am-ikw-t c1-see=by.eye.TA-INV-3 \u2018\u2026s\/heOBV sees him\/herPROX.\u2019 (Presented S2) 278 Since this morpheme is used in some contexts to construct \u201cObviative\u201d reference (119), we expect that it should code some kind of referential dependency (cf \u00a74.1). This means that we expect the theme sign \u2013ikw to show two specific sets of properties: (i) STRUCTURAL CONDITION: The suffix \u2013ikw should be sensitive to either linear precedence or c-command. (ii) SEMANTIC CONDITION: The suffix \u2013ikw should induce a perspectival embedding between the argument it is associated with and this structurally-dominating antecedent. In the following sections, I demonstrate that both of these properties are confirmed for \u2013 ikw, and they are confirmed independently of \u201cObviative\u201d contexts. In section 4.3.221, I consider the structural conditions on the use of \u2013ikw. I show that it introduces a c-command relation between an IP-level argument and a topic-level operator (cf. D\u00e9chaine & Reinholtz 2008). If the topic-level operator is absent, inverse-marked verbs require a linearly-preceding predicate or an overt operator. This means that this morpheme shows systematic sensitivity to both halves of the structural condition on referential dependency: (i) c-command and (ii) linear precedence. In section 4.4.222, I turn to the semantic properties of the inverse. I argue that this morpheme can be understood as coding that its argument is embedded within the perspective of its antecedent \u2013 the topic-level argument. This conclusion means that this theme sign codes referential dependency. Structural Conditions Semantic Condition C-command Linear Precedence Perspectival Embedding \u2013ikw \u2714 \u2714 \u2714 Referential Dependency \u2714 \u2714 \u2714 Table 4.7. The inverse meets the conditions of referential dependency Taken together, we see that the inverse theme sign \u2013ikw satisfies both halves of the conditions on referential dependency, and thus it is safe to conclude that it codes a kind of referential dependency. 279 4.4.221. Structural conditions on \u2013ikw When we consider the structural properties of the inverse theme sign \u2013ikw, we need to show two kinds of things. First, we need to identify the location of this form in the structure of the verbal complex. Second, we need to identify the kinds of structural relation this form has to other parts of the verb. When we do both of these, we find that the theme sign \u2013ikw has the following structural properties: (i) it is located in IP, and (ii) it is subordinated with respect to the topic operator in CP. This yields the structure as in (120). (120) CP 3 op 3 C IP 3 pro 3 -ikw- VP 3 ROOT 3 pro 3 FINAL Here, the inverse suffix \u2013ikw introduces an argument in the IP-layer of the clause, and is structurally c-commanded by an operator located in the specifier of CP. Existing in this kind of c-command relation, we see that the suffix \u2013ikw obeys the structural condition on referential dependency. 4.4.2211. Evidence for IP position of theme sign -ikw D\u00e9chaine and Reinholtz (1997, 2008) consider the position of the theme sign \u2013ikw relative to the rest of the verb. They conclude that, in contrast to the position of the \u2018direct\u2019 theme signs (e.g. \u2013\u00ea\u2013, \u00a74.4.21), this theme sign is located in the IP domain, introducing the subject of the verb\u2019s structure. 280 (121) a. DIRECT THEME SIGN \u2013\u00ea\u2013: VP-LEVEL ARGUMENT vP 3 ROOT 3 pro 3 FINAL VP 3 AA pro 3 -\u00ea- b. INVERSE THEME SIGN \u2013ikw: IP-LEVEL ARGUMENT CP 3 3 C IP 3 pro 3 -ikw- VP 3 ROOT 3 pro 3 FINAL For D\u00e9chaine and Reinholtz (1997, 2008), this structural difference correlates with a difference in the case-marking of the two forms: the \u2018direct\u2019 forms are Nominative-Accusative, and the \u2018inverse\u2019 forms are Ergative-Absolutive. Plains Cree, then, exhibits split-ergativity in its verb system. There are several kinds of evidence for this model. First, we notice that the inverse theme sign is in complementary distribution with the direct theme signs, but not with the verb\u2019s finals (122), which means that it likely relates to the same VP-level argument that the direct theme signs do (cf. D\u00e9chaine 2003, \u00a74.4.21). (122) a. \u2026\u1401 \u141a\u1438\u14a5\u146f\u141f. [Final + INV = \u2714] \u2026\u00ea-w\u00e2pamikot. \u2026\u00ea-w\u00e2p=am -ikw-t \u2026c1-see=by.eye.TA-INV-3 \u2018\u2026 s\/heOBV sees him\/herPROX.\u2019 (Presented S2) 281 b. \u273d ...\u1401 \u141a\u1438\u146f\u141f. [No Final = \u273d] \u273d \u2026\u00ea-w\u00e2pikot. \u2026\u00ea-w\u00e2p-ikw-t \u2026c1-see-INV-3 Intended: \u2018\u2026 s\/heOBV sees him\/herPROX.\u2019 (Presented S2) c. \u273d ...\u1401 \u141a\u1438\u14a5\u1481\u141f. [INV + DIR = \u273d] \u273d \u2026\u00ea-w\u00e2pamikw\u00e2t \u2026\u00ea-w\u00e2p=am -ikw-\u00e2 -t c1-see=by.eye.TA-INV-DIR-3 Intended: \u2018\u2026 s\/heOBV sees him\/herPROX.\u2019 (Presented S2) d. \u273d ...\u1401 \u141a\u14ab\u146f\u141f. [DIR + INV = \u273d] \u273d \u2026\u00ea-w\u00e2pam\u00e2kot \u2026\u00ea-w\u00e2p =am -\u00e2 -ikw-t c1-see=by.eye.TA-DIR-INV-3 Intended: \u2018\u2026 s\/heOBV sees him\/herPROX.\u2019 (Presented S2) Second, the inverse and direct theme signs occur on different sides of the diminutive suffix \u2013s\u2013 (123), which means that, although the two theme signs relate to the same argument, they cannot be in the exact same position (cf. D\u00e9chaine & Reinholtz 2008). (123) a. \u14c8\u1438\u1424 \u1401 \u1438\u1472\u14aa\u1426\u141a\u14ef\u141f \u140a\u144e\u14b7. [DIR > DIM] n\u00e2p\u00eaw \u00ea-pakamahw\u00e2sit atimwa n\u00e2p\u00eaw pakam=ahw -\u00ea -si -w atimw-a man hit =by.tool.TA-DIR-DIM-3 dog -XT \u2018the manPROX hit the dogOBV a little, just to scare it.\u2019 (Presented S2) b. \u14c8\u1438\u1424 \u1401 \u1438\u1472\u14aa\u1426\u1405\u14ef\u146f\u141f \u140a\u144e\u14b7. [INV > DIM] n\u00e2p\u00eaw pakamahosikot aw\u00e2sisa n\u00e2p\u00eaw pakam=ahw -si -ikw \u2013w aw\u00e2sis-a man hit =by.tool.TA-DIM-INV -3 child -XT \u2018the childOBV hit the manPROX just a little bit.\u2019 (Presented S2) Finally, the copy-to-object construction (cf. Dahlstrom 1986, 1991) picks out the verb final\u2019s argument in direct forms, but the other argument in inverse forms (124), suggesting that the inverse correlates with an alteration in the subject of the verb. (124) a. DIRECT: MATRIX VERB MARKS PROX \u14c2\u146d\u1422\u146b\u1528\u14ab\u1424 George \u1401 \u14f5\u146d\u1426\u140b\u141f \u1405\u146f\u14ef\u14f4. nikisk\u00eayim\u00e2w George \u00ea-s\u00e2kih\u00e2t okosisa. ni-kisk=\u00eayim -\u00e2 -w George \u00ea- s\u00e2k =ih -\u00e2 -t o-kosis-a 1-know=by.mind.TA-DIR-3 George C1-love=by.neut-DIR-3 3-son -XT \u201cI know GeorgePROX loves his sonsOBV.\u201d (Dahlstrom 1986: 87) 282 b. INVERSE: MATRIX VERB MARKS OBV \u14c2\u146d\u1422\u146b\u1528\u14a5\u14ab\u1418 George \u1401 \u14f5\u146d\u1426\u140b\u141f \u1405\u146f\u14ef\u14f4. nikisk\u00eayimim\u00e2wa George \u00ea-s\u00e2kihikot okosisa ni-kisk=\u00eayim -im -\u00e2 -w-a George \u00ea- s\u00e2k =ih -ikw -t o-kosis-a 1-know=by.mind.TA-DSJ-DIR-3-XT George C1-love=by.neut-INV-3 3-son -XT \u201cI know his sonsOBV love George.\u201d (Dahlstrom 1986: 88) Taken together, the model that provides the best coverage for this data set is one that locates the inverse in the IP-level (cf. D\u00e9chaine & Reinholtz 1997, 2008); this captures both its subject properties, and its alteration in argument structure from the direct forms. 4.4.2212. Evidence for structural relations of \u2013ikw If the theme sign \u2013ikw is located in IP, it is structurally higher than all the verb\u2019s other agurments. This positioning has the result that the argument introduced by -ikw cannot be in a c- command relation with any of the verb\u2019s argument positions. (125) IP 3 pro 3 -ikw- VP 3 ROOT 3 pro 3 FINAL While, at first glance, this would suggest that this argument cannot be structurally subordinated to any other argument, there are reasons to suppose that there is a higher position in the verb\u2019s structure that the argument of \u2013ikw can be subordinated to. For independent reasons, linguists working on the verbal structure of Plains Cree have argued that the system has a rich set of CP-level coding that includes reference to person features. As Cook (2008) has argued, this coding can be broken into two sets: a set that inherently picks out a speech-act participant (the prefixes ni(t)- and ki(t)- of the Independent Order), and a set that introduces a variable that requires an antecedent (e.g. the clause-typing prefixes \u00ea- and k\u00e2- of the conjunct order). D\u00e9chaine and Reinholtz (1997) and Blain (1997) consider the personal prefixes ni(t)- and ki(t)- of the Independent Order of verbal inflection (126). 283 (126) INDEPENDENT ORDER PERSON PREFIXES a. \u14c2\u141a\u1438\u1426\u144c\u1423. niw\u00e2paht\u00ean. ni-w\u00e2p=aht -\u00ea -n 1-see =by.eye.TI-DIR-LP \u2018I see itIN.\u2019 (Presented S2) b. \u14c2\u141a\u1438\u1426\u144c\u1423. kiw\u00e2paht\u00ean. ki-w\u00e2p=aht -\u00ea -n 2-see =by.eye.TI-DIR-LP \u2018You see itIN.\u2019 (Presented S2) These forms are in complementary distribution with a set of conjunct clause-typing elements (in particular \u00ea- and k\u00e2-: (127a)), and do not trigger conjunct agreement (127b) (127) a. PERSON PREFIXES CANNOT CO-OCCUR WITH CLAUSE-TYPING \u273d \u14c2\u144c \u141a\u1438\u1426\u1455\u14ab\u1423 \u273d nit\u00ea-w\u00e2pahtam\u00e2n nit-\u00ea-w\u00e2p=aht-am-\u00e2n 1-c1-see=by.eye.TI-TI-1 Intended: \u2018I see itIN.\u2019 (Presented S2) d. PERSON PREFIXES DO NOT TRIGGER CONJUNCT ORDER AGREEMENT \u273d \u14c2\u141a\u1438\u1426\u1455\u14ab\u1423 \u273d niw\u00e2pahtam\u00e2n ni-w\u00e2p=aht-am-\u00e2n 1- see=by.eye.TI-TI-1 Intended: \u2018I see itIN.\u2019 (Presented S2) Based on this evidence, Blain (1997) and D\u00e9chaine & Reinholtz (2008) conclude that ni(t)- and ki(t)- are clause-level operators located in the specifier of CP. Being located in the CP layer, the prefixes ni(t)- and ki(t)- are in a position that structurally-dominates the argument introduced by the inverse theme sign \u2013ikw. 284 (128) CP 3 nit- 3 C IP 3 pro 3 -ikw- VP 3 ROOT 3 pro 3 FINAL This means that, in a form involving a personal prefix, the argument introduced by -ikw is structurally-subordinated to that prefix. While this makes for a clean account about the structural relations when CP is filled with a person prefix in the Independent Order, the third-person set lacks a prefix in modern Plains Cree, as shown in (128). (129) NO PREFIX FOR 3RD PERSON a. \u2026\u14ed\u146d\u1426\u1403\u1420 \u2026s\u00eakihik s\u00eak =ih -ikw scare=by.neut-INV \u2018s\/hePROX was frightened by him\/herOBV.\u2019 (Presented S2) b. \u273d \u2026\u1405\u14ed\u146d\u1426\u1403\u1420 \u273d \u2026os\u00eakihik o-s\u00eak =ih -ikw 3-scare=by.neut-INV Intended: \u2018s\/hePROX was frightened by him\/herOBV.\u2019 (Presented S2) With no overt form to correspond to the other forms, a simple-minded assumption about the syntax of these forms would be that the CP lacks content. 285 (130) w\u00e2pamik w\u00e2p=am-ikw see=by.eye.TA-INV \u2018s\/heOBV sees him\/herPROX.\u2019 CP 3 \u2205 3 C IP 3 pro 3 -ikw- VP 3 w\u00e2p- 3 pro 3 -am If this is the case, and \u2013ikw has a c-command relation to the referential contents of CP, the current model of referential dependency expects a particular pattern when this CP content is lacking. Either (i) the lack of content in CP should cause the inverse to be ill-formed, or (ii) the form will depend on linear-precedence to satisfy what cannot be satisfied by structural c- command. (131) OPTION 1: LACK OF C-COMMAND = UNGRAMMATICAL CP V \u2205 V = \u273d C IP V X V -ikw- 286 (132) OPTION 2: LACK OF C-COMMAND = LINEAR-PRECEDENCE CP V PREDP1 CP \u2205 V V V C IP \uf0e0 y V \u2205 V V PRED1 C IP x V V -ikw VP x V -ikw- VP In fact, the second option is taken \u2013 the lack of a filled CP induces a linear dependency for the inverse-marked verb forms. The inverse theme sign is restricted in out-of-the-blue contexts, but the direct themes are not. In particular, the inverse cannot be used with two third persons unless there has been a previous predicate that introduced the \u201cProximate\u201d referent.11 In (133), the direct verb form tahkom\u00eaw \u2018itPROX bit itOBV\u2019 in (133a), affixed with the direct theme \u2013\u00ea\u2013, is felicitous in a context where I run into the room and announce what has happened outside, while the inverse verb form tahkomik \u2018itPROX was bitten by itOBV\u2019 in (133b), using the affix \u2013ikw\u2013, is not. Consultants (e.g. S2) explain that, although (133b) is comprehensible (i.e. the hearer knows that a dog has bitten a cat outside), it is not the way anyone would ever begin talking. (133) INVERSE \u2013ikw REQUIRES PREVIOUS PREDICATE, DIRECT \u2013\u00ea\u2013 DOES NOT a. \u140a\u144e\u14a7 \u1455\u1426\u146f\u14a3\u1424 \u14a5\u14c5\u14f4 \u1411\u152d\u1411\u1426\u144e\u14a5\u1541! atim tahkom\u00eaw min\u00f4sa w\u00eeyaw\u00eehtimihk! atim tahko=m -\u00ea -w min\u00f4s-a w\u00eeyaw\u00eehtim-ihk dog seize =by.mouth-DIR-3 cat -XT outside -LOC \u2018A dogPROX bit a catOBV outside!\u2019 (Presented S2) b. # \u14a5\u14c5\u1422 \u1455\u1426\u146f\u14a5\u1420 \u140a\u144e\u14b7 \u1411\u152d\u1411\u1426\u144e\u14a5\u1541! # min\u00f4s tahkomik atimwa w\u00eeyaw\u00eehtimihk! min\u00f4s tahko=m -\u00ea -w atimw-a w\u00eeyaw\u00eehtim-ihk cat seize =by.mouth-DIR-3 dog -XT outside -LOC \u2018A dogOBV bit a catPROX outside!\u2019 (Presented S2) This requirement of previous context for \u2013ikw is universally supported in texts as well; there are no cases of a discourse beginning with an inverse verb form involving two third persons. This is expected if the inverse requires structural-domination by something in CP. 11 Note that the word order has to be switched in these two examples because of the constraint on \u201cProximate\u201d- \u201dObviative\u201d nominal precedence, discussed in section 4.4.1 above. 287 The inverse-licensing predicate that precedes this verb can either be a verb (134a) or a possessed nominal (134b), further cementing the parallels between direct verb forms and possessor constructions (cf. \u00a74.4.1). In the pair of sentences in (134), a predicate making reference to the \u201cProximate\u201d referent linearly-precedes the inverse verb form. In (134a), this preceding predicate is a verb w\u00e2pam\u00eaw \u2018hePROX saw itOBV,\u2019 while in (134b), the preceding predicate is a possessed nominal nimin\u00f4sim \u2018my catPROX.\u2019 Both of these sentences are acceptable without previous context. (134) PRECEDING PREDICATE LICENSES \u2013ikw a. \u14c8\u142f\u1424 \u141a\u1438\u14a3\u1424 \u14a5\u14c5\u1422 \u1401 \u1455\u1426\u146f\u14a5\u146f\u141f. n\u00e2p\u00eaw w\u00e2pam\u00eaw min\u00f4s \u00ea-tahkomikot. n\u00e2p\u00eaw w\u00e2p=am-\u00ea-w min\u00f4s \u00ea- tahkw=m -ikw -t man see=by.eye.TA-DIR-3 cat C1-seize =by.mouth-INV-3 \u2018The manPROX saw the catOBV that bit himPROX.\u2019 (Presented S2) b. \u14c2\u14a5\u14c5\u14ef\u1466 \u1455\u1426\u146f\u14a5\u1420 \u140a\u144e\u14b7 \u1411\u152d\u1411\u1426\u144e\u14a5\u1541! nimin\u00f4sim tahkomik atimwa w\u00eeyaw\u00eehtimihk! ni-min\u00f4s=im tahko=m -ikw atimw-a w\u00eeyaw\u00eehtim-ihk 1-cat =DSJ seize =by.mouth-INV dog -XT outside -LOC \u2018My catPROX got bitten by a dogOBV outside!\u2019 (Presented S2) By contrast, direct verb forms need no such linear-precedence; they are acceptable out-of-the- blue as they are. The linear-precedence restriction of inverse-marked verbs shows an interesting person split. While verbs that involve only third persons require a linearly-preceding predicate, verbs that involve a speech-act participant do not. In (135), the form involving speech-act participants (134a) can be uttered in an out-of-the-blue context. For example, the sentence in (134a), which has a 1st person experiencer with a verb marked with \u2013ikw (nitahkomik \u2018I was bitten by itAN\u2019), can be uttered if the Speaker is outside, is attacked by a dog, and then comes running inside to tell people what happened. By contrast, the form in (135b), which has a 3rd person experiencer min\u00f4s \u2018cat\u2019 as the experiencer with a verb bearing \u2013ikw (tahkomik \u2018itPROX was bitten by itOBV\u2019), cannot be uttered in any parallel speech situation. 288 (135) a. 1ST PREFIX WITH \u2013ikw REQUIRES NO CONTEXT \u14c2\u1455\u1426\u146f\u14a5\u1420 \u140a\u144e\u1466 \u1411\u152d\u1411\u144e\u14a5\u1541! nitahkomik atim w\u00eeyaw\u00eetimihk! ni-tahkw=m -ikw atimw w\u00eeyaw\u00eetim-ihk 1- seize =by.mouth-INV dog outside -LOC \u2018A dog bit me outside!\u2019 (Presented S2) b. UNPREFIXED VERB BEARING \u2013ikw REQUIRES PREVIOUS CONTEXT # \u14a5\u14c5\u1422 \u1455\u1426\u146f\u14a5\u1420 \u140a\u144e\u14b7 \u1411\u152d\u1411\u144e\u14a5\u1541! # min\u00f4s tahkomik atimwa w\u00eeyaw\u00eetimihk! min\u00f4s tahko=m -ikw atimw-a w\u00eeyaw\u00eetim-ihk cat seize =by.mouth-INV-3 dog -XT outside -LOC \u2018A dogOBV bit a catPROX outside!\u2019 (Presented S2) This data makes sense if we posit that \u2013ikw needs to be structurally-dominated by the referent in CP; the first- and second-person forms have something that fills CP: ni(t)- and ki(t)- respectively, which means they do not need linear precedence to provide for the licensing of \u2013ikw (136). (136) niw\u00e2pamik ni-w\u00e2p=am-ikw 1- see=by.eye.TA-INV \u2018s\/he sees me.\u2019 CP 3 ni- 3 C IP 3 pro 3 -ikw- VP 3 w\u00e2p- 3 pro 3 -am- By contrast, the third-person forms do not have content in CP in the Independent Mode, and thus they must have linearly-preceding predicate to license -ikw. 289 (137) nimin\u00f4sim tahkomik atimwa ni-min\u00f4s=im tahkw=m -ikw atimw-a 1-cat =DIM seize =by.mouth-INV dog -XT \u2018My catPROX was bitten by a dogOBV.\u2019 (Presented S2) DP CP 3 3 ni- 3 \u2205 3 D nP C IP 6 3 y min\u00f4sim x 3 -ikw- VP 3 tahkw- 3 pro 3 -m Thus, testing the hypothesis that \u2013ikw induces a c-command relation results in a search that yields new data (the occurrence restrictions on inverse-marked clauses) and a way to understand this data. 12 Turning to the Conjunct Order, it is widely assumed that these prefixes are in C, since they code clause-typing properties (cf. Blain 1997, Cook 2008). 12 Note that this suggests something interesting about the logic of personal prefixation in Blackfoot (Algonquian, Alberta). In that language, Obv>Prox inverse forms obligatorily have prefixation for third person, and lack any kind of agreement for the \u201cObviative\u201d referent(i): i. otsik\u00e1komimmoka nohk\u00f3wa ot\u00e1niksi. ot-ik\u00e1komimm-ok-a n-ohk\u00f3-wa ot-\u00e1n -iksi 3-love-INV-AN.SG 1-son-AN.SG 3- daughter-PL \u2018Her daughtersOBV love my sonPROX.\u2019 (Frantz 1991:56) Crucially, Blackfoot lacks the Algonquian Conjunct Order, instead using Independent Order in embedded environments (cf. Frantz 1991). Assuming a parallel shift in syntactic structure, there will never be any referential material in CP sufficient to license the inverse\u2019s structural dependency. To compensate, the system inserts an overt variable (ot-) in the structure so that the inverse can be fed its antecedent via normal pronominal binding relations. 290 (138) CP 3 op 3 C IP \u00ea- 3 pro 3 -ikw- VP 3 ROOT 3 pro 3 FINAL This element is also higher than the argument introduced by \u2013ikw, and, according to Cook (2008) it introduces a variable in its specifier position, which is bound anaphorically. According to Cook (2008), the function of this variable construction is to provide a situation for the proposition to be evaluated with respect to. Since situations include referents (cf. Kratzer 2007), it follows that the clause-typing prefixes \u00ea- and k\u00e2- introduce a variable in their specifier that brings clause-external referents into the calculation. This means that they introduce a variable that can function much like the topic operators ni(t)- and ki(t)- seen above; the operator in conjunct clauses has referential content. Included in this is their ability to structurally-dominate the argument introduced by \u2013ikw. (139) CP 3 op 3 C IP \u00ea- 3 pro 3 -ikw VP 3 ROOT 3 pro 3 FINAL This means that, on a view where \u2013ikw requires c-command in order to be licensed, conjunct clauses provide the necessary structural configurations. 291 4.4.222. Semantic conditions on \u2013ikw If \u2013ikw codes referential dependency, it is not enough to show that it is c-commanded by some antecedent (\u00a74.4.221). According to the model of referential dependency developed here (\u00a74.1), it is also necessary to show that \u2013ikw codes that its referent is embedded in the perspective of this antecedent. Based on the conclusions we reached about the structural relations of \u2013ikw, this perspectival requirement means that the argument introduced by \u2013ikw must be embedded in the perspective of the referent related to the topic operator. STRUCTURAL CONDITION SEMANTIC CONDITION CP 3 y 3 C IP 3 x 3 -ikw VP 6 One way to test this expectation is to consider what kinds of referents can be associated with the topic position in an inverse construction. If only referents that hold a perspective can, then the argument introduced by \u2013ikw is always c-commanded by a perspective-possessing referent. If any kind of referent at all can be associated with this topic position, then the argument introduced by \u2013ikw is not necessarily c-commanded by a perspective-possessing referent. In fact, a consideration of the evidence shows that only perspective-possessing referents can be associated with the topic-position. Conversely, if the referent introduced by \u2013ikw is restricted in its ability to possess a perspective, then it will necessarily be the case that it is evaluated in some other referent\u2019s perspective. Taken together, the evidence supports the conclusion that the argument introduced by \u2013ikw is always evaluated within the perspective of the referent related to the topic operator in CP. This means that the theme sign \u2013ikw obeys the semantic condition on referential dependency. 292 4.4.2221. Semantic restrictions on the identity of topic op If the inverse form is used, the identity of the referent in the topic position is constrained in a particular set of ways; it can be 1st person, 2nd person, and \u201cProximate,\u201d but nothing else. 1ST \/ 2ND PROX OBV INAN Animal Unfamiliar Generic Topic op \u2714 \u2714 ?13 \u2718 \u2718 \u2718 \u2718 Table 4.8. Restrictions on the identity of topic op in Plains Cree It can be overtly filled with the 1st and 2nd person prefixes ni(t)- and ki(t)-, which code that the proposition is evaluated with respect to their perspective (cf. Cook 2008, \u00a74.4.2121), as the example in (140) shows. (140) \u14c2\u14ed\u146d\u1426\u1403\u1420. [\u2714 1OP & 3INV] nis\u00eakihik. ni-s\u00eak =ih -ikw 1-scare=by.neut-INV \u2018s\/he scared me.\u2019 (Presented S2) The topic position can also be filled with a \u201cProximate\u201d 3rd person, so long as the \u201cProximate\u201d third person precedes the verbal complex (cf. \u00a74.4.2212) and is in previous discourse (e.g. a proper name), as shown in (141). (141) Clare \u14ed\u146d\u1426\u1403\u146f\u1424 \u140a\u144e\u14b7. [\u2714 Clareop & DogINV] Clare s\u00eakihikow atimwa. Clare s\u00eak =ih -ikw-w atimw-a Clare scare=by.neut-INV-3 dog -XT \u2018ClarePROX was scared by the dogOBV.\u201914 (Volunteered S2) The topic position cannot be filled with an \u201cInanimate,\u201d as shown by the ungrammatical examples in (142a-b). In these examples, the possessed nominal nim\u00f4hkom\u00e2n \u2018my knife,\u2019 which precedes the verbal complex, cannot be construed as the topic of the verb otinikow \u2018itIN picked himAN up\u2019 that involves an \u201cAnimate\u201d (142a) or \u201cObviative\u201d (142b) referent n\u00e2p\u00eaw \u2018man\u2019. Instead, the only interpretation available is the infelicitous reading where the argument of \u2013ikw is the \u201cInanimate\u201d referent nim\u00f4hkom\u00e2n \u2018my knife,\u2019 and the other argument n\u00e2p\u00eaw occupies the topic position. 13 Since there are no examples of the inverse with an \u201cObviative\u201d antecedent in the Independent Order (i.e. -ikoyiwa) in any text that I have access to, and my consultants do not produce or recognize it, I am here agnostic on its ability to occur as the topic operator, although the lack of evidence for its existence is certainly suggestive. 14 Notice that a proper name can license the inverse form. This is likely to be a discourse effect of proper names. 293 (142) a. \u273d \u140a\u14c2\u14aa \u14c2\u14a8\u1426\u146f\u14ab\u1423 \u1405\u144e\u14c2\u146f\u1424 \u14c8\u1438\u1424. [\u273d KnifeOP & ManINV] \u273d anima nim\u00f4hkom\u00e2n otinikow n\u00e2p\u00eaw. an=ima ni-m\u00f4hkom\u00e2n ot =in -ikw-w n\u00e2p\u00eaw DST=IN.SG 1- knife pick=by.hand-INV-3 man Intended: \u2018That knifeIN of mine was picked up by the manAN.\u2019 (Presented S2) = # \u2018That knifeIN of mine picked up the manAN.\u2019 b. \u273d \u140a\u14c2\u14aa \u14c2\u14a8\u1426\u146f\u14ab\u1423 \u1405\u144e\u14c2\u146f\u1424 \u14c8\u1438\u1418. [\u273d KnifeOP & ManINV] \u273d anima nim\u00f4hkom\u00e2n otinikow n\u00e2p\u00eawa. an=ima ni-m\u00f4hkom\u00e2n ot =in -ikw-w n\u00e2p\u00eaw-a dst=IN.SG 1- knife pick=by.hand-INV-3 man -XT Intended: That knifeIN of mine was picked up by the manOBV. (Presented S2) = # \u2018That knifeIN of mine picked up the manOBV.\u2019 The topic position cannot be filled by an unfamiliar referent, as shown in (143). Here, I have set up a dialogue, in which a familiar referent Heather is introduced as a topic, identified by a proper name. In subsequent discourse, new referents must be identified with the argument introduced by \u2013ikw rather than this topic. For example, the inverse verb form \u00f4c\u00eamik \u2018s\/heOBV kissed him\/herPROX\u2019 can only mean that the new, unfamiliar referent (n\u00e2p\u00eaw \u2018man\u2019) is acting on this familiar, old referent (Heather). The other reading, in which Heather kisses the man, is unavailable in this context. (143) A: \u1401 \u1438\u1426\u1455\u14aa\u1423 \u1406\u14aa \u148c \u1473 \u1403\u1422\u1438\u1528\u1420 Heather? \u00ea-p\u00eahtaman \u00f4ma c\u00ee k\u00e2-ispayik Heather? \u00ea-p\u00eaht-am-an \u00f4ma c\u00ee k\u00e2-is-payi-k Heather CONJ-hear-TI-2 this.NI Q REL-thus-INCH-0 Heather \u2018Did you hear what happened to Heather?\u2019 (S2) B: \u14c2\u14a8\u152d. Nam\u00f4ya. nam=w=\u00eey=a Neg=3=body=XT \u2018No.\u2019 A: # \u14c8\u1438\u1424 \u1406\u1489\u14a5\u1420. [# Manop & Heatherinv] # n\u00e2p\u00eaw \u00f4c\u00eamik. n\u00e2p\u00eaw \u00f4c\u00ea=m -ikw man kiss=by.mouth-INV Intended: \u2018SheOBV kissed a manPROX.\u2019 = (?) \u2018A man? kissed her?.\u2019 294 A: \u14c8\u1438\u1418 \u1406\u1489\u14a5\u1420. [\u2714 Heatherop & Maninv] n\u00e2p\u00eawa \u00f4c\u00eamik. n\u00e2p\u00eaw-a \u00f4c\u00ea=m -ikw man -XT kiss=by.mouth-INV \u2018A manOBV kissed herPROX.\u2019 (Presented S2) It cannot be filled by a generic referent when the other referent is in the discourse. Consider the examples in (144). The verb miyw\u00eayimik \u2018s\/heOBV likes him\/herPROX\u2019 bears \u2013ikw and has a familiar referent (Shujun) as one argument and an unfamiliar, generic referent (ayisiyiniw \u2018people\u2019) as its other argument. In such a construction, Shujun must be interpreted as occupying the topic position and ayisiyiniw \u2018people\u2019 as the argument introduced by \u2013ikw (144b), rather than the reverse (144a). (144) a. # \u140a\u1528\u14ef\u1528\u14c2\u1424 \u14a5\u1429\u140d\u1528\u14a5\u1420 Shujun\u140a. [# Peopleop & Shujuninv] # ayisiyiniw miyw\u00eayimik Shujuna. ayisiyiniw miyw=\u00eayim -ikw Shujun-a person good=by.mind.TA-INV Shujun-XT Intended: \u2018ShujunOBV likes peoplePROX.\u2019 = (#) \u2018A specific personPROX likes ShujunOBV.\u2019 (Presented S2) b. Shujun \u14a5\u1429\u140d\u1528\u14a5\u1420 \u140a\u1528\u14ef\u1528\u14c2\u1418. [\u2714 Shujunop & Peopleinv] Shujun miyw\u00eayimik ayisiyiniwa. Shujun miyw=\u00eayim -ikw ayisiyiniw -a Shujun good=by.mind.TA-INV person -XT \u2018PeopleOBV like ShujunPROX.\u2019 (Presented S2) It cannot easily be filled with an animal (145a-b), and consultants will contravene all other grammatical rules of the language and interpret pragmatically-difficult meanings to avoid this configuration. Consider the examples in (145), where the verb s\u00eakihik \u2018s\/heOBV scares him\/herPROX\u2019 has a human, nip\u00e2pa \u2018my father,\u2019 and an animal, s\u00ees\u00eepa \u2018a duck\u2019 as its arguments. The only possible interpretation associates the duck with the \u2013ikw argument and my father with the topic argument \u2013 even if this means directly violating the argument coding of the nominals, as in (145b). (145) a. \u14c2\u1439\u1438 \u141a\u1438\u14a5\u1420 \u14f0\u14f0\u1438. [\u2714 FatherOP & DuckINV] nip\u00e2p\u00e2 s\u00eakihik s\u00ees\u00eepa. ni-p\u00e2pa s\u00eak =ih -ikw s\u00ees\u00eep-a 1- father scare=by.neut-INV duck-XT \u2018My fatherPROX was scared by a duckOBV.\u2019 (Presented S2) 295 b. \u273d \u14f0\u14f0\u144a \u14c2\u1438\u1426\u1403\u1420 \u14c2\u1439\u1438\u1418. [\u273d DuckOP & FatherINV] \u273d s\u00ees\u00eep nipahik nip\u00e2pawa. s\u00ees\u00eep nip =ah -ikw ni-p\u00e2pa-a duck sleep=by.tool-INV 1-father-XT \u2260 A duckPROX was killed by my fatherOBV.\u2019 = (?) \u2018My father? was killed by a duck?.\u2019 (Presented S2) The only way any consultant has been willing to allow an animal to be the topic operator is if Speaker has a special relation to the animal (e.g. a pet), and the other referent is either \u201cInanimate\u201d or a total stranger (146). Only one consultant was willing to tolerate this form. (146) \u14c2\u144c\u1466 \u1455\u1426\u146d\u1422\u1472\u140f\u1420 \u140a\u140f\u152d \u142f\u1418. [\u2714 My dogOP & ManINV] nit\u00eam tahkiskawik awiya n\u00e2p\u00eawa. ni-t\u00eam tahk=iskaw -ikw awiya n\u00e2p\u00eaw-a 1-dog seize=by.body.TA-INV some man -XT \u2018My pet dogPROX was kicked by some guyOBV.\u2019 (Presented S3) If the CP is filled with \u00ea- or k\u00e2-, the antecedent of \u2013ikw in spec CP must be one of the acceptable referential categories, and not one of the forbidden ones. This includes the following referential categories: (i) 1st and 2nd persons are possible antecedents, as in (147), where the verb k\u00e2- tahkomikoy\u00e2hk \u2018itAN bit us,\u2019 which is coded with \u2013ikw, is subordinate to a matrix clause that codes reference to the Speaker (nik\u00ee-m\u00eakwa-atosk\u00e2n\u00e2n \u2018we were working\u2019). (147) \u14c2 \u146e \u14a3\u147f \u140a\u1450\u1422\u1473\u14c8\u1423 \u1473 \u1455\u1426\u146f\u14a5\u146f. [\u2714 1OP & DogINV] nik\u00ee-m\u00eakwa-atosk\u00e2n\u00e2n k\u00e2-tahkomikoy\u00e2hk atim. ni-k\u00ee- m\u00eakwa-atosk\u00ea-n\u00e2n k\u00e2-tahko=m -ikw-y\u00e2hk atimw 1-PREV-midst- work -1pl C2-seize=by.mouth-INV-1pl dog \u2018We were working when the dogAN bit us.\u2019 (Presented S2) (ii) \u201cProximate\u201d referents are possible antecedents, as in (148), where the verb k\u00e2-tahkomikot \u2018it bit her\u2019 is coded with \u2013ikw and is subordinate to the verb k\u00ee-m\u00eakwa-atosk\u00eaw \u2018she was working,\u2019 which codes reference to a Proximate third person (Shujun). (148) Shujun \u146e \u14a3\u147f \u140a\u1455\u1422\u146b\u1424 \u1473 \u1455\u1426\u146f\u14a5\u146f\u141f \u140a\u144e\u14b7. [\u2714 ShujunOP & DogINV] Shujun k\u00ee-m\u00eakwa-atosk\u00eaw k\u00e2-tahkomikot atimwa. Shujun k\u00ee- m\u00eakwa-atosk\u00ea-w k\u00e2- tahko=m -ikw-t atimw-a Shujun PREV-amidst- work -3 c2- seize=by.mouth-INV-3 dog -XT \u2018ShujunPROX was working when the dogOBV bit herPROX.\u2019 (Presented S2) 296 (iii) \u201cInanimate\u201d referents are not allowed, as in the ungrammatical examples in (149) show. Here, the verb k\u00e2-otinikot is coded with \u2013ikw and is subordinate to the verb p\u00eekopayin \u2018itIN broke,\u2019 which codes reference to an \u201cInanimate\u201d referent. (149) a. \u273d \u140f\u152d\u1472\u1423 \u1432\u146f\u1438\u1528\u1423 \u1473 \u1405\u144e\u14c2\u146f\u141f \u140a\u141a\u14ef\u1422. [\u273d DishOP & ChildINV] \u273d \u00f4ma wiy\u00e2kan p\u00eekopayin k\u00e2-otinikot aw\u00e2sis. aw =ima wiy\u00e2kan p\u00eeko =payi-n k\u00e2- oti =n -ikw-t aw\u00e2sis PRX=IN.SG dish break=inch-LP c2-pick=by.hand-INV-3 child Intended: \u2018This dishIN broke when the childAN picked it up.\u2019 (Presented S2) b. \u273d \u140f\u152d\u1472\u1423 \u1432\u146f\u1438\u1528\u1423 \u1473 \u1405\u144e\u14c2\u146f\u141f \u140a\u141a\u14ef\u14f4. [\u273d DishOP & ChildINV] \u273d \u00f4ma wiy\u00e2kan p\u00eekopayin \u00ea-otinikot aw\u00e2sisa. aw =ima wiy\u00e2kan p\u00eeko =payi-w \u00ea- oti =n -ikw-t aw\u00e2sis-a PRX=IN.SG dish break=inch-3 c1-pick=by.hand-INV-3 child -XT Intended: \u2018This dishIN broke when the childOBV picked it up.\u2019 (Presented S2) Taking stock of these generalizations, we see that the only kinds of referents that can be connected to the topic operator position are those that possess a perspective; this includes 1st and 2nd person in the independent order and \u201cProximate\u201d third persons, and not \u201cInanimate\u201d referents, which are inherently extentional (\u00a72.4). This looks exactly like the semantic conditions on the antecedent for \u2013yi\u2013 (\u00a74.3.22), and the semantic conditions on the possessor in possession constructions (\u00a74.4.12); the antecedent of the argument introduced by \u2013ikw must be perspective- possessing. 4.4.2222. Perspective-less properties of -ikw When we consider the properties of the argument introduced by \u2013ikw in the context of the properties of the referent identified with CP, a strong generalization emerges: either the referent associated with \u2013ikw does not have a perspective, or their perspective is embedded within the perspective of their antecedent. This includes \u201cObviative\u201d referents, \u201cInanimate\u201d referents, Animals, unfamiliar referents, and generic referents, as summarized in Table 4.9. 1ST \/ 2ND PROX OBV INAN Animal Unfamiliar Generic Inverse pro \u2718 \u2718 \u2714 \u2714 \u2714 \u2714 \u2714 Table 4.9. Restrictions on the identity of the argument introduced by \u2013ikw This means the properties of the argument introduced by \u2013ikw are almost completely opposite to the properties of the referent associated with the topic op position. \u201cInanimate\u201d subjects triggers 297 the application of \u2013ikw when they are the subjects of transitive verbs, as seen in both examples in (150). (150) a. \u14c2\u146d\u14f0\u142f\u146d\u14c2\u146f\u1423 \u140a\u14c2\u14aa fountain. [\u2714 InINV] nikis\u00eep\u00eakinikon anima fountain. ni-kis\u00eep\u00eak=in -ikw-n an=ima fountain 1- wash =by.hand-INV-LP PRX=IN.SG fountain \u2018That fountainIN washed me.\u2019 (Presented S2) b. Cora \u146d\u14f0\u142f\u146d\u14c2\u146f\u1424 \u140a\u14c2\u14aa fountain. [\u2714 InINV] Cora kis\u00eep\u00eakinikow anima fountain. Cora kis\u00eep\u00eak=in -ikw-w an=ima fountain Cora wash =by.hand-INV-3 PRX=IN.SG fountain \u2018That fountainIN washed Cora.\u2019 (Presented S2) Animals can be associated with \u2013ikw when acting on human beings. Consider the example in (151), where the inverse form of the verb nis\u00eakihik \u2018itAN scared me\u2019 codes that atim \u2018the dog\u2019 is acting on the Speaker. (151) \u14c2\u14ed\u146d\u1426\u1403\u1420 \u140a\u144e\u1466 [\u2714 DOGINV] nis\u00eakihik atim ni-s\u00eak =ih -ikw atimw 1- scare=by.neut-INV dog \u2018The dog scared me.\u2019 (Presented S2) With third persons, consultants once again interpret the animal as the inverse argument regardless of the fact that the human-referring nominal is coded as associating with the \u2013ikw argument (152b). (152) a. Clare \u14ed\u146d\u1426\u1403\u1420 \u140a\u144e\u14b7. [\u2714 DOGINV] Clare s\u00eakihik atimwa Clare s\u00eak =ih -ikw atimw-a Clare scare=by.neut-INV dog -XT = \u2018The dogOBV scared ClarePROX.\u2019 \u2260 \u2018ClarePROX scared the dogOBV.\u2019 (Presented S2) c. \u273d \u14c2\u144c\u1466 \u14ed\u146d\u1426\u1403\u1420 Clare\u140a. [\u273d Clare, \u2714 DOGINV] \u273d nit\u00eam s\u00eakihik Clare\u2019a. ni-t\u00eam s\u00eak =ih -ikw Clare-a 1-dog scare=by.neut-INV Clare -XT = (?) My dog? scared Clare?. \u2260 # ClareOBV scared my dogPROX. (Presented S2) \u201cObviative\u201d referents obligatorily induce \u2013ikw when acting on anyone but \u201cInanimate\u201ds. 298 (153) Clare \u1403\u1422\u1475\u1424 \u1411\u148b\u1426\u1401\u1424 \u14c8\u1438\u1418 \u1401 \u14a5\u1422\u1472\u140f\u146f\u1420. [\u2714 Obvinv] iskw\u00eaw w\u00eecih\u00eaw n\u00e2p\u00eawa \u00ea-miskawikot. iskw\u00eaw w\u00eec=ih -\u00ea -w n\u00e2p\u00eaw-a \u00ea- mi=iskaw -ikw-t woman help=by.neut -DIR-3 man -XT C1-find=by.body.TA-INV-3 \u2018The womanPROX helped the manOBV when heOBV found herPROX.\u2019 (Presented S2) (154) Clare \u1403\u1422\u1475\u1424 \u1411\u148b\u1426\u1401\u1424 \u14c8\u1438\u1418 \u1401 \u14a5\u1422\u1472\u14a5\u1528\u1420. [\u2714 Obvinv] iskw\u00eaw w\u00eecih\u00eaw n\u00e2p\u00eawa \u00ea-miskamiyit. iskw\u00eaw w\u00eec=ih -\u00ea -w n\u00e2p\u00eaw-a \u00ea- mi=isk -am-yi-t woman help=by.neut -DIR-3 man -XT C1-find=by.body.TI-TI -DS-3 \u2018The womanPROX helped the manOBV when heOBV found itIN.\u2019 (Presented S2) Simple \u201cAnimate\u201d referents only trigger inverse coding when they interact with first and second persons, or are animals; in all other contexts, they do not occur as the argument of -ikw. (155) \u14c2\u1411\u148b\u1426\u1403\u146f\u14c8\u1423 \u14c8\u1438\u1424 \u1401 \u14a5\u1422\u1472\u140f\u146f\u152e\u157d. [\u2714 Aninv] niw\u00eecihikon\u00e2n n\u00e2p\u00eaw \u00ea-miskawikoy\u00e2hk. ni-w\u00eec=ih -ikw-w n\u00e2p\u00eaw \u00ea- mi =iskaw -ikw-y\u00e2hk 1-help=by.neut-INV-3 man C1-find=by.body.TA-INV-1pl \u2018The manAN helped me when hePROX found us.\u2019 (Presented S2) \u201cProximate\u201d referents also only trigger inverse coding if they interact with first and second persons. (156) \u14c2\u1411\u148b\u1426\u1403\u146f\u14c8\u1423 \u14c8\u1438\u1424 \u1401 \u1411\u148b\u1426\u140b\u141f \u1405\u1455\u141a\u14ef\u14a5\u14f4. [\u2714 Proxinv] niw\u00eecihikon\u00e2n n\u00e2p\u00eaw \u00ea-w\u00eecih\u00e2t otaw\u00e2simisa. ni-w\u00eec=ih -ikw-w n\u00e2p\u00eaw \u00ea- w\u00eec =ih -\u00e2 -t ot-aw\u00e2s=im -is -a 1-help=by.neut-INV-3 man C1-find=by.body.TA-INV-3 3- child=DSJ-DIM-XT \u2018The manPROX helped me when hePROX was helping hisPROX childrenOBV.\u2019 (Presented S2) First and second person interactions never trigger the inverse, under any context, in Plains Cree, which is the stable Algonquian pattern (Goddard 2007) except for Blackfoot (cf. Frantz 1976, 1991). The only way that an inverse verb form can code a perspective-possessing referent is if that perspective is itself embedded, either in the Speaker\u2019s (157a), or in the \u201cProximate\u2019s\u201d (157b). 299 (157) a. \u2013ikw ARGUMENT EMBEDDED IN SPEAKER\u2019S PERSPECTIVE \u14c2\u140f\u1426\u1455\u14ab\u1420 Clare \u1401 \u141a\u1438\u14a5\u146f\u152e\u1541. niw\u00eehtam\u00e2k Clare \u00ea-w\u00e2pamikoy\u00e2hk. ni-w\u00eeht=amaw -ikw Clare \u00ea- w\u00e2p=am -ikw-y\u00e2hk 1- tell=APPLIC-INV Clare C1-see =by.eye.TA-INV-1PL \u2018ClareAN told me that sheAN saw us.\u2019 (Presented S2) Speaker Clare C tell(x,S) Clare see(S,Clare) 300 b. \u2013ikw ARGUMENT EMBEDDED IN \u201cPROXIMATE\u2019S\u201d PERSPECTIVE Clare \u140f\u1426\u1455\u14ab\u1420 \u1411\u1489\u141a\u1472\u14c7 \u1401 \u141a\u1438\u14a5\u146f\u141f. Clare w\u00eehtam\u00e2k w\u00eec\u00eaw\u00e2kana \u00ea-w\u00e2pamikot. Clare w\u00eeht=amaw-ikw w-w\u00eec\u00ea =w=\u00e2=kan -a \u00ea-w\u00e2p=am-ikw-t Clare tell=APPLIC-INV 3-accompany=3=dir=nom-XT C1-see =by.eye.TA-INV-1PL \u2018Clare\u2019sPROX friendOBV told herPROX that sheOBV saw herPROX.\u2019 (Presented S2) Speaker y Clare C EXT(y,C) tell(y,Clare) Clare y tell(y,Clare) friend(y,Clare) y see(y,Clare) Crucially, this second perspective-possessing referent (e.g. Clare\u2019s friend), whose perspective the inverse argument\u2019s (e.g. Clare\u2019s) perspective is embedded in, occupies the topic op position of the matrix clause. Summarizing, \u2013ikw almost always codes with a perspective-less argument, and the only way that \u2013ikw can introduce a perspective-possessing argument is if it is embedded in the perspective of its antecedent. This is exactly the inverse of the properties of the topic operator\u2019s referent. 301 1ST \/ 2ND PROX OBV INAN Animal Unfamiliar Generic Topic op \u2714 \u2714 ?15 \u2718 \u2718 \u2718 \u2718 Inverse pro \u2718 \u2718 (\u2714) \u2714 \u2714 \u2714 \u2714 \u2714 Table 4.10. Topic operator and Inverse pro are opposites in their properties The argument of the inverse, and the Topic operator are two sides of the same coin. 4.4.223. Conclusion: \u2013ikw as referential dependency From the preceding discussion, we see that the theme sign \u2013ikw is sensitive to both conditions on referential dependency, and that it is sensitive to these independent of obviation constructions. In section 4.4.221, we saw that the suffix \u2013ikw introduces an argument that is either c-commanded by the topic operator in CP, if that position is filled, or linearly-dependent on an antecedent, if that position is empty. Then, in section 4.4.222, we saw that the referent associated with the argument introduced by \u2013ikw is always interepted within the perspective of this antecedent. This means that \u2013ikw satisfies both halves of referential dependency, and thus its relation to obviation becomes transparent; when the subject of a verb is third person and contextually extentional it is simultaneously \u201cObviative\u201d and a candidate for for \u2013ikw. In this context, the properties of obviation and \u2013ikw overlap. 4.5. Conclusion The evidence presented here suggests that referential dependency in Plains Cree is constructed out of two components: (i) A SYNTACTIC COMPONENT: Linear precedence or c-command (ii) A SEMANTIC COMPONENT: Perspectival embedding Together, these two mechanisms define how it is that one referent can become the context for the interpretation of another. This evidence also suggests that the \u201cObviative\u201d referential category is constructed out of components that code various kinds of referential dependency. Some of these elements have been shown to obey the two conditions on referential dependency, in particular nominal ordering 15 Since there are no examples of the inverse with an \u201cObviative\u201d antecedent in the Independent Order (i.e. -ikoyiwa) in any text that I have access to, and my consultants do not produce or recognize it, I am here agnostic on its ability to occur as the topic operator, although the lack of evidence for its existence is certainly suggestive. 302 (\u00a74.3.1), the suffix \u2013yi\u2013 (\u00a74.3.2), possessor constructions (\u00a74.4.1), the direct theme sign \u2013\u00ea\u2013 (\u00a74.4.21), and the inverse theme sign \u2013ikw (\u00a74.4.22). Further, the last four of these domains show referential dependency effects even when they are not used in \u201cObviative\u201d constructions. This independence of their referential dependency from their \u201cObviative\u201d functions provides confirmation of the constructional view of obviation taken in chapter 3 (\u00a73.3); \u201cObviative\u201d reference is opportunistically constructed out of semantically and structurally compatible pieces. I have not been able to consider all of the components of obviation (e.g. \u2013im, \u2013a, \u2013ihi: \u00a73.3). It is left up to further investigation to consider if these other components of obviation also show the same behaviour. If they do, the current analysis can be extended to full coverage of \u201cObviative\u201d forms. If they do not, they may be telling us something interesting about the different kinds of components that construct obviation (i.e. some are referential dependencies, some are plurality coding, and some are repair strategies when referential dependencies fail). Either way, we will have learned more about Plains Cree than we knew before. 303 Chapter 5 Eliciting obviation and animacy 5.1. The place of methodology in linguistic analysis As the physicist Werner von Heisenberg said, \u201cwhat we observe is not nature in itself but nature exposed to our method of questioning\u201d (Heisenberg 1958). The Plains Cree data presented in the previous chapters did not grow on those typewritten pages; it came from someone and was gotten by someone. This state of affairs is a necessary part of all linguistic work \u2013 all language always has a human source, and only humans (to our knowledge) work on it. This presents a complex set of methodological challenges to anyone seeking a systematic account of anything in human language. In particular, we must consider properties of both the human conducting the investigation (i.e. the linguist) and the human that the linguistic forms have come from (i.e. the consultant).1 This chapter considers these questions in the context of the data and analysis presented in this thesis. I first consider the broad question of what fieldwork is, and what we ask it do for us (\u00a75.2), dividing fieldwork into elicitation and text collection. I then turn to the specifics of elicitation methodology, considering how the \u201cmethod of questioning\u201d impacts the picture of Plains Cree\u2019s referential system that we develop (\u00a75.3\u20135.6). In doing this, it is hoped that the reader will be provided with the tools necessary to evaluate the data that has been crucial to the argumentation of the previous chapters. Although \u201cobjectivity\u201d is not possible in linguistic analysis, transparency is. 5.2. Fieldwork in linguistics The linguistic data in this thesis has been gotten in two ways, both of which fall under the domain of \u201cFieldwork,\u201d meaning someone has literally gone out to meet language speakers and collected the linguistic material from them. Traditionally, this fieldwork has been divided into two kinds: (i) text collection, and (ii) elicitation. These two kinds of fieldwork agree on their purpose \u2013 to collect linguistic data suitable for systematic analysis of the language, but differ as to the directness of the method. A useful analogy is fishing: text collection drags nets through the water, catching all fish in its path, while elicitation baits its hook for a particular fish. 1 As is shown below, sometimes these two roles are conflated; the linguist is the consultant, and the consultant is the linguist. 304 Fieldwork that consisted of text collection is indirect in its method; it seeks to collect the needed linguistic forms by collecting a large body (called a \u201ccorpus\u201d) of language material, and then sorting through it later. This approach has the advantage of catching many things in its net \u2013 numerous pieces of which were never imagined to exist by the linguist, giving the linguist a wide knowledge of the language and its many features. Much like fishing with a net, however, this method also drags in a great deal of extraneous, murky material, making it an intimidating undertaking for the novice linguist. Further, there is never any negative data \u2013 lack of evidence for a form does not constitute proof that it does not exist. This limits the kinds of linguistic analysis that can be done (cf. Matthewson 2004). Fieldwork that consists of elicitation is direct in its method; it seeks to collect a linguistic form by targeting it specifically. This approach has the advantage of providing negative evidence; the linguist can (theoretically) prove that a form does not exist in the language. However, it has the disadvantage of being unnaturally narrow; the linguist does not see much of the language under study, and most often only finds out information about the forms being sought. Further, the relation between the intent of the elicitation (the target) and the resulting data is not easy to discern. Following the fishing analogy, when targeting a particular fish, it is crucial to know the relation between the bait and the fish caught; just because you\u2019ve baited your hook for trout doesn\u2019t mean that whatever you pull from the water is a trout. Since elicitation fieldwork has been a significant component of the data collection methods for this thesis, it would be wise to consider the relation between the intended target and the result. This is the topic of the following discussion. 5.3. Obviation and animacy data fluctuates in elicitation Obviation and animacy phenomena do not behave in standard elicitation tasks as they do in corpus material. This is true of both the form of obviation and animacy and their content\/context. For example, transitive clauses as in (1a) often lack overt obviation marking on the nominal construed as denoting an obviative referent (1b), which is at variance with obviation\u2019s formal properties (\u00a73.3). 305 (1) a. \u201cOBVIATION\u201d CODING PRESENT ON iskw\u00eaw \u2018WOMAN\u2019 \u2026 \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u2026 \u00ea-w\u00e2pam\u00e2t n\u00e2p\u00eaw iskw\u00eawa \u00ea-w\u00e2p=am -\u00e2 -t n\u00e2p\u00eaw iskw\u00eaw-a c1-see=by.eye.TA-DIR-3 man woman-XT \u2018\u2026 (that) the manPROX sees the womanOBV.\u2019 (Volunteered S2) b. \u201cOBVIATION\u201d CODING ABSENT ON iskw\u00eaw \u2018WOMAN\u2019 \u2026 \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1424 \u00ea-w\u00e2pam\u00e2t n\u00e2p\u00eaw iskw\u00eaw \u00ea-w\u00e2p=am -\u00e2 -t n\u00e2p\u00eaw iskw\u00eaw c1-see=by.eye.TA-dir-3 man woman \u2018\u2026 (that) the man? sees the woman?.\u2019 (Volunteered S2) Likewise, sentences that contrast in obviation patterns are sometimes said to be equivalent in meaning, suggesting that obviation does not have any content at all (2), which is at variance with obviation\u2019s contentful and contextual properties (\u00a73.4-5). (2) \u201cOBVIATIVE\u201d AND \u201cPROXIMATE\u201d INTERACTIONS RULED EQUIVALENT a. \u2026 \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u2026 \u00ea-w\u00e2pam\u00e2t n\u00e2p\u00eaw iskw\u00eawa \u00ea-w\u00e2p=am -\u00e2 -t n\u00e2p\u00eaw iskw\u00eaw-a c1-see=by.eye.TA-DIR-3 man woman-XT \u2018\u2026 (that) the manPROX sees the womanOBV.\u2019 (Volunteered S2) = 2b (S1,S2,S3,S4) b. \u2026 \u1401 \u141a\u1438\u14a5\u146f\u141f \u1403\u1422\u1475\u1424 \u14c8\u142f\u1418 \u2026 \u00ea-w\u00e2pamikot iskw\u00eaw n\u00e2p\u00eawa \u00ea-w\u00e2p=am -ikw-t iskw\u00eaw n\u00e2p\u00eaw-a c1-see=by.eye.TA-INV -3 woman man-XT \u2018\u2026 (that) the manOBV sees the womanPROX.\u2019 (Volunteered S2) = 2a (S1,S2,S3,S4) This fluctuation means that any linguist conducting elicitation that targets obviation and animacy phenomena must think carefully about the nature of elicitation, and its relation to these phenomena. In this chapter, I consider these issues in detail. I begin by considering the kinds of tasks typically done in elicitation work (\u00a75.4), the kinds of confounds they can introduce, and the way that obviation and animacy behave in each of these tasks. I divide these tasks into five types: 306 (i) Translation tasks (\u00a75.4.1) (ii) Judgment tasks (\u00a75.4.2) (iii) Utterance-in-context tasks (\u00a75.4.3) (iv) Correction tasks (\u00a75.4.4) (v) Analysis tasks (\u00a75.4.5) From this data, I conclude that elicitation is a particular kind of performance task, requiring the speaker to construct imaginary discourses to judge others\u2019 sentences (\u00a75.5). In the subsequent sections, I turn to the contextual properties of obviation and animacy in elicitation. In section 5.5.1, I show that the unmarked, animate referent is (almost) always preferred in elicitation. In section 5.5.2, I show how to contextually support the occurrence of obviatives in elicitation. Finally, in section 5.5.3, I consider data that suggests that contextless obviation is semantically similar to inanimate forms. With these generalizations in hand, the reasons for behaviour of animacy and obviation in elicitation become transparent: elicitation lacks context. If obviation codes a contextual, rather than an inherent, property of the referent (\u00a73.4), lack of context will cause a lack of obviation. If animacy relies on the construction of a perspective (\u00a72.5), the lack of a perspective in discourse means that consultants will either fall back on statistically probably forms or lose animacy contrasts entirely. Within the context of the current model of obviation, the necessity of context for the disambiguation of obviative and animacy forms is expected. Because they code distinctions in extentionality and perspective anchoring, these referential types are purely intensional. Being intensional, contrasts in meaning will only appear when we attempt substitution tests (cf. Church 1941, etc.). In contextless elicitation, speakers systematically rely on extensional semantics, whether they speak Cree or English, and thus intensional differences are not offered. Only when a context is supplied that destroys substitutability does the equivalence of forms break down. Thus, using tasks that alternately enrich and impoverish context provide crucial support for the model of extentionality that is developed in this thesis. 307 5.4. A taxonomy of elicitation tasks and their effects on obviation and animacy Linguists that conduct elicitation-type fieldwork typically ask the speaker to perform a limited number of tasks. These include what Cook and M\u00fchlbauer (2007) have labeled Translation tasks (\u00a75.4.1) and Judgment tasks (\u00a75.4.2). Some linguists expand this repertoire to include Utterance- In-Context tasks (\u00a75.4.3). In addition, linguists may also take data into consideration that comes from Correction tasks (\u00a75.4.4) and Analysis tasks (\u00a75.4.5). The amount of each of these tasks that any elicitation session has is a function of the linguist and the consultant. For the linguist, the relevant factors are (i) competency in the target language, (ii) specificity of the target forms (e.g. \u201caspect\u201d versus k\u00ee-), and (iii) the kind of forms targeted (e.g. agreement vs. kinship terms). Beginning work on an unfamiliar language, translation tasks constitute the majority of elicitation time. If the linguist becomes more competent in the language, judgement tasks and utterance-in-context tasks become much more frequent. Consultants likewise bring different skills; some are comfortable only with translation tasks, some prefer judgement tasks, while others will only do complex utterance-in-context tasks and analysis tasks resulting from these. In my own work, the majority of tasks have been judgement tasks (\u2248 50%), with utterance-in-context tasks (\u2248 40%) and analysis tasks (\u2248 10%) providing numerous crucial insights. Translation has only been used to find lexical items and forms that I do not know, and to produce basic minimal pairs for grammatical exposition. Towards a more complete taxonomy of methods, the following discussion includes methods that I have never used, but have seen other linguists use. Relevant to the present discussion, obviation and animacy behave differently in these different environments. In what follows, I consider the interaction of these phenomena with each task type. 5.4.1. Translation tasks Translation tasks involve the transfer of some form from the shared language to the target language (Cook & M\u00fchlbauer 2006). In a typical translation task, the speaker is asked how to say some sentence in the target language. 308 (3) LINGUIST: How do you say, \u201cI am tired\u201d in Cree? CONSULTANT: umm, nin\u00eastosin. ni-n\u00eastosi-n 1-tired-LP Another kind of translation task works in reverse to this: the speaker is asked what some sentence that has been constructed in their target language means. (4) LINGUIST: What does it mean if I say \u201cnin\u00eastosin\u201d? ni-n\u00eastosi-n 1-tired-LP CONSULTANT: You said \u201cI\u2019m tired.\u201d Here, the target language is used to produce an utterance in the shared language. 5.4.11. Confounds in translation: Non-correlation and priming Translation tasks can be thought of as a mapping operation, from one language to the other. This mapping is only partial, in that it does not necessarily pick out identical elements. At issue is the difference between translation and equivalence. If a form in one language translates into a form in another language, this does not mean that the two forms are equivalent, merely that they have related meanings. For example, when I asked a native German speaker2 to do a translation task that targets the german \u2018conversational perfect,\u2019 using English as the shared language, the exchange in (5) resulted. (5) LINGUIST: How do you say \u201cI saw a movie yesterday\u201d in German? CONSULTANT: uhh, Ich bin gestern ins Kino gegangen. uhh ich bin gestern in das Kino ge-gang-en hes 1 am yesterday in the cinema prtcpl-go.pst-prtcpl Here, the English verb phrase \u201csaw a movie yesterday\u201d is translated into a German verb phrase that has three components that differ from the English sentence: (i) the lexical meaning of the verb (gehen \u2018go\u2019 vs. see), (ii) the use of a prepositional phrase (ins Kino \u2018into the Cinema\u2019) instead of a determiner phrase as the direct object (a movie), and (iii) a perfective verb form (bin \u2026 gegangen) instead of a simple past verb (saw). At best, what this shows is that the German 2 Thanks to Sonja Thoma (p.c.) for discussion of these forms. 309 perfective construction occurs where English often uses its past tense form; it does not demonstrate total equivalence between the forms. In the sample German data just seen, the speaker was able to respond to the English sentence with a natural German sentence, but this is not always the case. In a translation task, the context of the target-language sentence is often the shared-language sentence (Cook & M\u00fchlbauer 2006). This means that, unless the speaker manages to abstract away from the task, priming becomes a significant problem. This causes the target language to more closely resemble the shared language. For example, in the German sample above, a speaker that had priming intereference would produce the following dialogue instead. (6) LINGUIST: How do you say \u201cI saw a movie yesterday\u201d in German? CONSULTANT: uhh, Ich sah ein Film gestern. uhh Ich sah ein Film gestern hes 1 see.pst a Film yesterday Here, the German speaker has produced a form that much more closely mirrors the English form: (i) the German Imperfekt has been used, which has the closest formal equivalence to the English simple past (sah vs. saw), (ii) the word order has shifted placing the adverb gestern \u2018yesterday\u2019 in the position of \u2018yesterday\u2019 in the English sentence, (iii) the choice of a lexical item that more closely approximates English \u2018movie\u2019 (Film instead of Kino), and (iv) the use of a direct object, like English (sah ein Film vs. \u2018see a movie\u2019), instead of a prepositional phrase (ins Kino). While this is a well-formed German utterance, it is not said in the same context as the English utterance. Instead, it means something strange, as though the Speaker had gone and sat in a cinema, staring at the screen listlessly. Thus, the English framing has interfered with the production of an utterance that accurately reflects the grammar of German. 5.4.12. Obviation in translation tasks As Cook and M\u00fchlbauer (2006) point out, translation tasks are one of the most common places for obviation to disappear, particularly when a speaker is first acclimating to the performative demands of elicitation. Thus, examples like (7) are not uncommon, particularly in the first few months of a consultant\u2019s elicitation work. 310 (7) LINGUIST: How do you say \u201cA man saw a woman\u201d in Cree? CONSULTANT: Umm, n\u00e2p\u00eaw w\u00e2pam\u00eaw iskw\u00eaw n\u00e2p\u00eaw w\u00e2p=am -\u00ea-w iskw\u00eaw man see=by.eye.TA-DIR-3 woman Here, the consultant has produced a sentence that differs from expected Plains Cree patterns in several ways: (i) the word order matches English (SVO) instead of the more common Plains Cree pattern of VSO (cf. Wolfart & Reinholtz 1996), (ii) the verb is in the independent order of inflection instead of the more often-used conjunct order (w\u00e2pam\u00eaw vs. \u00ea-w\u00e2pam\u00e2t), and (iii) the nominal iskw\u00eaw \u2018woman\u2019 lacks the suffix \u2013a (\u00a73.2.1). Based on the model of obviation proposed in this thesis and our understanding of what is entailed in a translation task, this fact should not be surprising. While Plains Cree nominal reference is organized around extentionality (chapters 2 & 3), it appears that English nominal reference is not; the English sentence \u201cthe man saw the woman\u201d could be interpreted within the perspective of either refernt, depending on the discourse context (cf. Kuno 1988). Thus, both obviative and unspecified (in this context, proximate) nominals in Plains Cree map onto the unmarked English nominal. This means, when asked to translate an English sentence into Cree, a speaker could either invent a reason to treat a referent extentionally (since the English provides no information), which takes a great deal of energy and cannot be maintained for a two-hour session, or simply translate the unspecified English into the unspecified Cree. This produces sentences that directly mirror the English form, both in word order and in absence of obviation contrasts (cf. Cook & M\u00fchlbauer 2006). 5.4.12. Animacy in translation tasks Translation tasks do not appear to interfere with animacy coding, either noun-verb agreement or nominal marking. Consultant do not translate an English sentence into a Cree form that is lacking in animacy agreement. (8) LINGUIST: How do you say \u201cI saw that dish\u201d in Cree? CONSULTANT: niw\u00e2paht\u00ean anima wiy\u00e2kan ni-w\u00e2p=aht -\u00ea -n an -ima wiy\u00e2kan 1- see=by.eye.TI-TI-LP DIST-IN.SG dish 311 Here, the form offered by the consultant matched its animacy across the forms: (i) the verb stem uses the TI final \u2013aht\u2013 instead of the animate form \u2013am\u2013, (ii) the theme sign used is \u2013\u00ea\u2013 for inanimate and obviative objects (\u00a73.3.3) rather than \u2013\u00e2\u2013 used for animate, and (iii) the demonstrative used has the inanimate suffix \u2013ima rather than the animate \u2013a. In all respects, this coding matches text data and natural speech, which means there do not appear to be any clause- level coding discrepancies between translation tasks and text material. There are two likely reasons for this stability: (i) word-level and clause-level agreement patterns are robust in elicitation, and (ii) elicitation usually works with nominals that fall easily into English \u201cnatural\u201d animacy. First, the Cree system of agreement is extremely stable in elicitation; once an agreement pattern is picked, it always follows across all contexts. For example, the clause-typing system never gets mismatches (i.e. no mixing of independent and conjunct mode), for any speaker, under any conditions (9). (9) CLAUSE-TYPING IS STABLE IN ELICITATION a. \u14c2\u141a\u1438\u1426\u144c\u1423 niw\u00e2paht\u00ean ni-w\u00e2p=aht -\u00ea -n 1- see=by.eye.TI-TI-LP \u2018I see itIN.\u2019 (Presented S2) b. \u1401 \u141a\u1438\u1426\u1455\u14ab\u1423 \u00ea-w\u00e2pahtam\u00e2n \u00ea-w\u00e2p=aht -am-\u00e2n C1- see=by.eye.TI-TI -1 \u2018\u2026(as) I see itIN.\u2019 (Presented S2) b. \u273d \u1401 \u141a\u1438\u1426\u144c\u1423 \u273d \u00ea-w\u00e2paht\u00ean \u00ea-w\u00e2p=aht -\u00ea-n C1- see=by.eye.TI-TI -LP --- (Presented S2) c. \u273d \u14c2\u141a\u1438\u1426\u1455\u14ab\u1423 \u273d niw\u00e2pahtam\u00e2n ni-w\u00e2p=aht -am-\u00e2n 1- see=by.eye.TI-TI -1 --- (Presented S2) The presence of these robust, word-internal agreement patterns in elicitation supports the maintenance of clause-level agreement. 312 Second, elicitation tends to rely on canonical forms, because these have well-established linguistic properties and high levels of acceptibility. If the linguist is trying to target only one component of a structure, other factors need to be kept constant. Non-targeted forms will thus be picked from a canonical set, to reduce the possibility of their intereference. For example, a linguist targeting the formal properties of animacy coding will restrict nominal forms to things that are known to be canonically animate in Plains Cree (dogs, cats, women, men, children, relatives), or things that are known to be canonically inanimate (dishes, shoes, hats). That is, linguists are intentionally picking forms that are easily mapped from English into Cree. Thus, the English sentence already provides information about animacy in these cases, making the Cree consultant\u2019s translation task straightforward. 5.4.2. Judgment tasks The other most common task that linguists ask speakers to perform is the judgment task (Cook & M\u00fchlbauer 2006). Defined broadly, tasks of this kind require that the consultant consider some form (or forms) in the target language and rule them appropriate or not. While this requirement on judgment remains constant, the method of getting these judgments is quite varied. The most common way to obtain a judgment from a speaker is to produce a form and ask if it is acceptable (these have been labelled \u201cPresented\u201d throughout this thesis). (10) LINGUIST: Is it a good sentence if I say \u201cAd\u00e9 d\u00e1 n\u00ed ap\u00e1\u201d? Ad\u00e9 d\u00e1 n\u00ed ap\u00e1 Ad\u00e9 break LOC arm CONSULTANT: Yes. That\u2019s good. Here, the linguist has uttered a Yor\u00f9b\u00e1 (Kwa, Nigeria) form and asked a native speaker of Yor\u00f9b\u00e1 if it is acceptable.3 The consultant hears the form, considers it, and judges it to be well- formed in their language. Of course, this method requires some competene in the target language on the part of the linguist. Linguists who have not achieved this minimal competency often resort to writing them on paper and ask the consultant to judge it. This is sometimes carried as far as to place (hypothesized) morpheme breaks in the examples, dropping (hypothesized) surface phonology. 3 Thanks to Oladi\u00edp\u00f2 Aj\u00edb\u00f3y\u00e9 (p.c.) for discussion of this example. 313 (11) LINGUIST: Is this a good sentence? PAPER: nit-isi-\u00eehkaw-isw-n (Spoken as [\u0294n\u0329tsi\u00e7k\u00e0s\u028an]) CONSULTANT: Uhh\u2026 [attempting to read paper]. Yes, I think. This tactic of presenting a speaker with highly analyzed forms has significant methodological problems. First, it assumes literacy on the part of the consultant \u2013 an assumption that should not be made for Plains Cree, where literacy in the first language is between 1% and 5% (SIL 1982). Assuming that the consultant is one of this 1-5%, presenting them with analyzed forms can cause other problems: (i) It makes significant assumptions about the relation between surface phonology and underlying forms (e.g. nit- is the underlying form of the surface [\u0294n]?). (ii) It assumes that speakers have direct, conscious access to underlying morphology (e.g. a speaker can quickly reconstruct \u2013\u00eehkaw\u2013isw\u2013 to \u2013\u00eehk\u00e2so\u2013). (iii) It builds analysis directly into the discovery procedure, which introduces biases into the data collection process. (iv) By appealing to written forms, prescriptive, rather than descriptive judgments become central. (v) It often violates the cultural traditions of the speech community, both in terms of a preference for oralcy and a preference against dissecting language forms (cf. Lightning 1996, Valentine 2001). With so many confounds, it is impossible to know why a Speaker would prefer or disprefer a form presented in this way. Another method for obtaining judgments that I have seen used is to record sentences from one speaker and play them to another speaker for judgments. This has the advantage of not requiring language competency on the part of the linguist, and controls for accidental mispronunciations, but also introduces a number of confounds. In many cultures, including Plains Cree (Lightning 1996) but also English-speaking cultures (e.g. my own), offering judgments on other people\u2019s speech is usually taboo. Even if the consultant is willing to violate their own cultural pattern in order to perform the task, speakers are often not willing to rule utterances produced by someone else as \u201cbad,\u201d especially if the recording of the other speaker is perceived to be highly competent or higher in status than the consultant (e.g. an elder in the community). This makes getting negative data \u2013 a crucial part of elicitation \u2013 difficult. 314 5.4.21. Confounds in judgment tasks: Metalinguistics The difficulty with judgment tasks lies in their metalinguistic nature. In these kinds of tasks, the speaker is being asked to consider a form for acceptability. As Cook and M\u00fchlbauer (2006) point out, this can lead a speaker to lose the connection between the forms being considered and the context they are being considered in. Losing context, the speaker may fail to control the forms correctly, allowing contextually-determined components to fluctuate without the consultant noticing it. Thus, a sentence that could be uttered in a real speech context can sometimes be ruled \u201cbad\u201d because of a failure to identify the felicitous context. Consider the English sentence in (12). (12) The window, my brother broke. I asked a class of approximately 35 native English speakers if this was a good sentence.4 More than half of the group aggressively claimed it was not, identifying it as a \u2018sentence fragment.\u2019 When asked to repeat the target, they uttered (12) as one intonational unit with pitch declination throughout and main stress on \u2018brother,\u2019 as in (13). (13) [\u0111\u04d9 w\uf09cndo mai br\u04d9\uf08f\u0111\uf092 brok] In English, a form uttered this way is interpreted as a nominal and a relative clause (i.e. \u201cthe window that my brother broke \u2026\u201d ] , which is not a complete sentence. Crucially, the form that I had intended them to understand had two intonational units, with two intonational peaks in it, as in (14). (14) [\u00f0\u04d9 w\uf09c\uf08fnd\uf098o mai br\u04d9\uf08f\u00f0\uf092\uf098 brok] Thus, the speakers that had claimed the utterance was ill-formed had not controlled their intonation sufficiently the speakers had failed a judgment task; they had not kept all the variables identical between my production and their own. This means that they were judging a target form that was not the target form of the linguist. The fact that more than half of these students failed to do this should give anyone who relies heavily on judgment tasks pause.5 4 Thanks to two tutorials (p.c.) of Ling100 students from the Fall Semester of 2007. 5 This problem is not restricted to linguistics novices, of course. I have listened to many professional-level discussions about English judgments, even among native speakers, wherein one or more of the linguists was unable to control their intonation. 315 5.4.22. Judgment tasks and obviation With respect to obviation, judgment tasks produce interesting and telling results. Recall from Chapter 3 that nominal forms that code obviation have two kinds of variables, with respect to their morphosyntax. On the one hand, they are identical to inanimate plurals (15), and on the other, they are ambiguous for singular or plural interpretations (16). (15) \u201cOBVIATIVE\u201d IS IDENTICAL TO \u201cINANIMATE\u201d PLURAL a. \u1405\u14c8\u142f\u14aa \u140a\u14c2\u1426\u1403 \u1473 \u14a5\u152a\u14ef\u1528\u141f. on\u00e2p\u00eama anihi k\u00e2-miyosiyit. o-n\u00e2p\u00ea=m -a an -ihi k\u00e2-miyw=isi -yi -t 3-man =DJ-XT DST-XT C2-good=STAT-DS-3 \u2018That is herPROX husbandOBV who is good-looking.\u2019 (Presented S2) b. \u1405\u14aa\u1422\u146d\u14ef\u14c7 \u140a\u14c2\u1426\u1403 \u1473 \u14a5\u1429\u141a\u14ef\u146d. omaskisina anihi k\u00e2-miyw\u00e2siki. o-maskisin-a an -ihi k\u00e2-miyw=\u00e2si-k-i 3-shoe -XT DST-XT C2-good=stat-0-PL \u2018Those are herAN shoesIN that are nice.\u2019 (Presented S2) (16) \u201cOBVIATIVE\u201d IS AMBIGUOUS FOR NUMBER a. \u2026 \u142f\u152d\u1420 \u1401\u14a5\u152a\u14ef\u1528\u141f \u140a\u14c2\u1426\u1403 \u1403\u1422\u1475\u1418. \u2026 p\u00eayak \u00ea-miyosiyit anihi iskw\u00eawa. p\u00eayakw \u00ea-miyw=si-yi-t an -ihi iskw\u00eaw-a one c1-good=STAT-DS-3 DST-XT woman-XT \u2018\u2026(as) that one womanOBV is pretty.\u2019 (Presented S2) b. \u2026 \u14c3\u14f1 \u1401 \u14a5\u152a\u14ef\u1528\u141f \u140a\u14c2\u1426\u1403 \u1403\u1422\u1475\u1418. \u2026 n\u00eeso \u00ea-miyosiyit anihi iskw\u00eawa. n\u00eesw \u00ea-miyw =si -yi -t an -ihi n\u00e2p\u00eaw-a two C1-good=STAT-DS-3 DST-XT man -XT \u2018\u2026(as) those two womenOBV are pretty.\u2019 (Presented S2) For a speaker to make a judgment on the acceptability of an obviative form, then, they must hold both of these variables constant; that is, they must control for the potential interpretation of the nominal as denoting an inanimate, plural entity. Sometimes, consultants do not succeed in doing this, as the example in (17) shows. 316 (17) LINGUIST: Can I say \u201catim nipah\u00eaw min\u00f4sa\u201d? atim nip =ah -\u00ea -w min\u00f4s-a dog sleep=by.neut-DIR-3 cat -XT CONSULTANT: Yes. \u2026 Wait, no. You wanted to say that the dog killed a cat? min\u00f4sa is plural. You need to say min\u00f4s. atim nipah\u00eaw min\u00f4s. atim nip=ah -\u00ea -w min\u00f4s dog sleep=by.neut-DIR-3 cat Here, the speaker has initially accepted the sentence, but, upon further consideration, she has ruled the obviative form bad, based on a perceived desire to encode \u201csingularity.\u201d Misinterpreting the suffix -a as intended to denote \u201cplural\u201d (an Analysis task, see \u00a75.4.5 below), which it is in other contexts (cf. \u00a73.3.1), she rules the sentence bad. Thus, the judgment task has invited the consultant to engage in metalinguistic consideration of forms \u2013 a process that resulted in the loss of context necessary to support the felicity of the form. This data demonstrates the constructional nature of obviation (cf. \u00a73.3). In voluntarily taking on the metalinguistic task of considering forms in (17), the consultant has repeated the nominal min\u00f4sa \u2018cat\u2019 out of its sentential context. In contextless environments, this suffix most often codes inanimate plurality (\u00a73.3.3). Without its formal context (i.e. a verb bearing animate agreement), there is no way to determine that the suffix \u2013a was being used to construct an obviative referent (cf. \u00a73.3.1, Wolfart 1973). Thus, the fluctuation in the data caused by a metalinguistic loss of context actually provides crucial insight into the nature of the construction of obviation in Plains Cree. 5.4.23. Animacy in judgment tasks In judgment tasks, animacy is erratic. A typical elicitation session for animacy begins with a few clear judgments, where the consultant is able to clearly and easily express the target meaning contrasts. However, this quickly becomes chaotic after only a few questions. (18) SAMPLE ANIMACY JUDGMENT SET a. \u2026\u14aa\u1422\u146d\u14ef\u14c7 \u1401 \u146d\u14ce\u146d. maskisina \u00ea-kinw\u00e2ki. maskisin-a \u00ea-kinw =\u00e2 -k-i shoe -XT C1-long=STAT-0-PL \u2018\u2026 (as) the shoesIN are long.\u2019 (Presented S2) 317 b. \u273d \u2026\u14aa\u1422\u146d\u14ef\u14c7\u1420 \u1401 \u146d\u146f\u14ef\u148b\u1420. \u273d maskisinak \u00ea-kinosicik maskisin-ak \u00ea- kinw=si -t-ik shoe -PL C1-long=STAT-3-PL Intended: \u2018the shoesANIM are long.\u2019 (Presented S2) \u2026 i. \u2026\u14aa\u1422\u146d\u14ef\u14c7 \u1401 \u146d\u14ce\u146d. maskisina \u00ea-kinw\u00e2ki. maskisin-a \u00ea-kinw =\u00e2 -k-i shoe -XT C1-long=STAT-0-PL \u2018\u2026 (as) the shoesIN are long.\u2019 (Presented S2) j. \u2026\u14aa\u1422\u146d\u14ef\u14c7\u1420 \u1401 \u146d\u146f\u14ef\u148b\u1420. maskisinak \u00ea-kinosicik maskisin-ak \u00ea- kinw=si -t-ik shoe -PL C1-long=STAT-3-PL \u2018\u2026 (as) the shoesAN are long.\u2019 (Presented S2) Here, the initial set of judgments were unambiguous for the consultant: the inanimate form was good and the animate form was bad. However, after more animacy-related elicitation tasks were done, all forms become equally acceptable, and the earlier clarity of judgments completely disappears. All forms are said to now \u201cmean the same thing.\u201d At that point, the task must be set aside until a later session. If Animacy were a lexical property of nominals (\u00a72.3), this kind of data would be extremely puzzling. In a lexical model, either a nominal is inanimate or it isn\u2019t. Why should the data fluctuate so much? By contrast, the current model expects this kind of fluctuation; animacy is a property of a nominal in a context. Judgment tasks are typically very context-poor, and the context can fluctuate (\u00a75.4.2). Without a stable context, judgments on animacy cannot be stable either. 5.4.3. Supporting judgment tasks: Utterance-in-context tasks One way to help deal with the potential confounds of judgment tasks is to work harder to support the task with a richer context. Thus, the tasks Cook and M\u00fchlbauer (2006) label Utterance-in- context tasks can be thought of as an enriched form of standard Judgment tasks. There are numerous ways to construct the needed context, and I here consider a few them. One way to construct context for a judgment task is to simply provide more linguistic material to frame the target form. For example, if we are targeting a judgment on the 318 acceptability of subject-topicalized structures in English, the addition of preceding linguistic context can be used to support test the target form. (19) a. ? The window, my brother broke. b. Fiona broke the vases in the living room, but the window, my brother broke. Here, the addition of extra linguistic material significantly improves comprehension of the form. In the class of 35 students I tested, all of them agreed that this was well-formed, even if they had claimed the topicalized form alone was bad. Another way to construct context is with pictures or other visual aids. This works particularly well when targeting forms that have spatial semantics encoded in them. For example, the Plains Cree preverb nitaw- codes that the event is moving away from the speaker. We can test this by drawing a picture and asking the consultant which form is appropriate.6 (20) Craig \u1401 \u14c2\u1455\u140f \u1472\u1475\u148b\u14a5\u141f \u1403\u1422\u1475\u1418 \u1472 \u14c2\u1472\u14a7\u148b\u1420. Craig \u00ea-nitawi-kakw\u00eacim\u00e2t iskw\u00eawa ka-nikamocik. Craig \u00ea- nitawi-kakw\u00eac=im -\u00e2 -t iskw\u00eaw-a ka- nikamo-t-ik Craig C1-go.off -try =by.mouth.TA-DIR-3 woman-XT FUT-dance -3-PL \u2018CraigPROX went and asked the girlOBV to dance.\u2019 (Presented S2) Here, supporting a judgment task with a picture makes the judgment clear both for the linguist and the consultant. Because of this clarity, some linguists rely heavily on this picture-elicitation method, employing animated computer animations to aid elicitation tasks (cf. Burton 2005, Munro 2006). 6 For a discussion of visual-based elicitation techniques that target spatial reference, see Munro (2006). 319 An easy way to provide context for a judgment task is to use the speech situation that is actually happening, since this provides many of the crucial ingredients to speech; times, places, and people. Taking advantage of the spacial context, it is easy to get consistent judgments on an element like nitaw- in Plains Cree. (21) CONTEXT: A linguist and a consultant are sitting in Vancouver, British Columbia. The consultant\u2019s cousin is visiting from Edmonton, Alberta. a. \u142f \u146d\u152a\u146b\u1424. p\u00ea-kiyok\u00eaw p\u00ea- kiyok\u00ea -w c1-come-visit-3 \u2018She has come to visit.\u2019 (Presented S2) b. # \u1401 \u14c2\u1455\u140f \u146d\u152a\u146b\u141f. # nitawi-kiyok\u00eaw nitawi-kiyok\u00ea-w go- visit -3 \u2018She has gone to visit.\u2019 (Presented S2) A task of this kind can be easily constructed for anything that relies on speech time or place. Another way that a linguist can get a consultant to perform utterance-in-context tasks is by asking the consultant to imagine a situation they describe before introducing a judgment task. For example, if I want to test the familiarity conditions on bare nominal forms in Yor\u00f9b\u00e1 (Kwa, Nigeria), I can construct situations and ask for judgments in these situations.7 (22) LINGUIST: If I just walked into the room, without warning, and said Ay\u00f2 \u00f3 d\u00e1 igi, is that a good form? Ay\u00f2 \u00f3 d\u00e1 igi Ayo HTS break stick CONSULTANT: Yes, I would understand. You just said that Ay\u00f2 broke a stick. LINGUIST: Suppose I walked into the room and said Ay\u00f2 \u00f3 d\u00e1 or\u00ed instead? Ay\u00f2 \u00f3 d\u00e1 or\u00ed Ay\u00f2 HTS break head CONSULTANT: No, that does not make sense. You have just said \u201cAy\u00f2 broke head.\u201d But whose head has he broken? LINGUIST: Suppose I came in and reported to you about a horrible fight that Ay\u00f2 had been in, where he beat up many, many people. Could I tell you about it saying that sentence? 7 Thanks to Oladi\u00edp\u00f2 Aj\u00edb\u00f3y\u00e9 (p.c.) for discussion of these forms. 320 CONSULTANT: \u273dpause\u273d Yes. Yes, if you were coming in and said that there was a battle, and Ay\u00f2 had been fighting, then this sentence would be good. It would mean that he \u201cbroke heads.\u201d Just anyone\u2019s heads. The heads of the unknown people he was fighting with. Thus, asking the consultant to imagine a set of situations allows for a more sophisticated set of judgments to be obtained. One final way that consultants can perform utterance-in-context tasks is to use corpus materials. In such a situation, the linguist can use both well-formed discourses and manipulated discourses as contexts for obtaining judgments about a particular target form. For example, consider the following pair of discourses. (23) \u201cMy dear Mr. Bennet,\u201d said his lady to him one day, \u201chave you heard that Netherfield Park is let at last?\u201d Mr. Bennet replied that he had not. \u201cBut it is,\u201d returned she; \u201cfor Mrs. Long has just been here, and she told me all about it.\u201d Mr. Bennet made no answer. \"Do you not want to know who has taken it?\" cried his wife impatiently. \"YOU want to tell me, and I have no objection to hearing it.\" This was invitation enough. (Pride and Prejudice by Jane Austen: Chapter 1, page 1) (24) \u201cMy dear Mr. Bennet,\u201d said a woman to her husband one day, \u201chave you heard that Netherfield Park is let at last?\u201d Her husband replied that he had not. \u201cBut it is,\u201d returned she; \u201cfor Mrs. Long has just been here, and she told me all about it.\u201d He made no answer. \"Do you not want to know who has taken it?\" cried the woman impatiently. \"YOU want to tell me, and I have no objection to hearing it.\" This was invitation enough. (Altered form) Here, I have taken a section from the beginning of a well-formed discourse (23) and altered the pronominal and relational forms (24). Presenting first one and then the other to a native speaker 321 of English, judgments can be sought about the effect on a discourse that relational forms have. While these are both well-formed discourses with different properties, I could also manipulate the discourse to cause it to be ill-formed. (25) \u201cMy dear Mr. Bennet,\u201d said she to him one day, \u201chave you heard that Netherfield Park is let at last?\u201d # A man replied that he had not. \u201cBut it is,\u201d returned she; \u201cfor Mrs. Long has just been here, and she told me all about it.\u201d # Her husband made no answer. \"Do you not want to know who has taken it?\" cried the girl\u2019s mother impatiently. \"YOU want to tell me, and I have no objection to hearing it.\" This was invitation enough. (Altered form) This is the form of elicitation used most extensively in the current thesis. 5.4.31. Obviation in utterance-in-context tasks Utterance-in-context tasks significantly improve the stability of obviation\u2019s occurrence (Cook & M\u00fchlbauer 2006). This task is particularly good at getting disambiguation in forms that are, with less context, ruled \u201cidentical.\u201d (26) a. \u2026\u1401 \u141a\u1438\u14ab\u141f \u14ce\u142f\u1424 \u1403\u1422\u1475\u1418. \u2026\u00ea-w\u00e2pam\u00e2t n\u00e2p\u00eaw iskw\u00eawa. \u00ea-w\u00e2p=am -\u00e2 -t n\u00e2p\u00eaw iskw\u00eaw-a c1-see=by.eye.TA-DIR-3 man woman-XT \u2018\u2026(as) the manPROX saw the womanOBV.\u2019 (Presented S2,S3,S4) b. \u2026\u1401 \u141a\u1438\u14a5\u146f\u141f \u1403\u1422\u1475\u1424 \u14ce\u142f\u1418. \u2026 \u00ea-w\u00e2pamikot iskw\u00eaw n\u00e2p\u00eawa. \u00ea-w\u00e2p=am -ikw-t iskw\u00eaw n\u00e2p\u00eaw-a c1-see=by.eye.TA-INV-3 woman man -XT \u2018\u2026(as) the manOBV saw the womanPROX.\u2019 (Presented S2,S3,S4) With pairs like this that lack clear context, consultants invariably say that these forms \u201cmean the same thing,\u201d and are usually unable to offer more explicit discussion of the difference. I then constructed a drawing and context to support the judgment task. 322 (27) In this picture, the man is behind a hill, out of the Speaker\u2019s sight, and the woman is standing next to the Speaker. With this more explicit context set up, the consultant\u2019s (S2) judgments suddenly became clear; the form with an obviative agent was required (26b), and the form with a proximate agent (26a) was infelicitous. In explaining this, the consultant relied heavily on the necessity that the woman must have told the Speaker about it, given the situation; the Speaker could not have known what this man was doing without the woman\u2019s help. This is in line with the analysis of obviation\u2019s content and context considered in section 3.4. and 3.5. Utterance-in-context tasks that employ texts are particularly good at getting clear, insightful obviation judgments. As one example, I located a passage from a text (k\u00e2- pimw\u00eaw\u00eahahk 1998) in which two unmarked referents are used within the same clause \u2013 an occurrence that is fairly rare in Plains Cree. I first presented this entire discourse (approximately 2 pages of Cree) to three consultants in an unchanged form, to make sure it was acceptable. The target sentence of this discourse is shown in (28). (28) TARGET SENTENCE: TWO \u201cANIMATE\u201d FORMS IN SAME CLAUSE \u1401\u1418\u146f \u1401\u147f \u1473 \u146d\u1422\u146d\u14f2\u14a5\u141f \u14c5\u1426\u1456\u140f\u1429 \u1405\u1432\u146d\u1422\u1475\u140f\u1423; \u00eawakw \u00eakwa k\u00e2-kiskis\u00f4mit n\u00f4ht\u00e2wiy op\u00eekiskw\u00eawin; \u00eawakw \u00eakwa k\u00e2-kisk =isi=\u00f4m -it n-\u00f4ht\u00e2wiy o-p\u00eekiskw\u00eawin resum then C2-know=stat=by.mouth.TA-3>1 1-father 3-word \u2018he then reminded me of my father\u2019s words;\u2019 After all of the consultants determined that the entire discourse was coherent and were able to interpret and translate it, I altered the crucial passage so that one of the referents was obviative, rather than unmarked. 323 (29) FATHER IS \u201cOBVIATIVE\u201d \u1401\u1418\u146f \u1401\u147f \u1473 \u146d\u1422\u146d\u14f2\u14a5\u141f \u14c5\u1426\u1456\u140f\u152d \u1405\u1432\u146d\u1422\u1475\u140f\u14c2\u1528\u1424; \u00eawakw \u00eakwa k\u00e2-kiskis\u00f4mit n\u00f4ht\u00e2wiya op\u00eekiskw\u00eawiniyiw; \u00eawakw \u00eakwa k\u00e2-kisk =isi=\u00f4m -it n-\u00f4ht\u00e2wiy-a o-p\u00eekiskw\u00eawin-yi-w resum then C2-know=stat=by.mouth.TA-3>1 1-father -XT 3-word -DS-3 \u2018hePROX then reminded me of my father\u2019sOBV words;\u2019 Upon considering the altered form, all three consultants responded that they preferred the first form. One consultant, S4, then went on to explain that the reason for the preference was the the speaker was emphasizing that the words of his mentor and his father were identical, and that he had fully-remembered both. According to S4, if he had marked one of the referents as obviative, he would not have remembered that person\u2019s counselling well. In the terms of the current analysis of obviation\u2019s content (\u00a73.4) and context (\u00a73.5), we can understand this as a judgment on extentionality and accessibility (\u00a73.5.4); the obviation of n\u00f4ht\u00e2wiy \u2018my father\u2019 makes that referent contextually extentional (\u00a73.4), and that referent\u2019s perspective (i.e. intentions) becomes inaccessible directly to the Speaker. Thus, the textual task provides for rich context and rich judgments, and gives data that is crucial to our understanding of obviation in Plains Cree. 5.4.32. Animacy in utterance-in-context tasks Animacy judgments become much clearer in utterance-in-context tasks. For example, in a particular session, I had tried a series of largely context-less judgment tasks, and animacy marking had begun to fluctuate (\u00a75.4.22). After some other elicitation was done, the consultant was getting up to leave, and noticed a realistic-looking plastic snake that was on a nearby table. She flinched and then laughed, explaining that she had thought it was a real, live snake. Seizing this opportunity, I asked her how to describe the event before and after finding out it was a plastic snake. (30) LINGUIST: Before you found out it was plastic, could you say \u201cnik\u00ee-w\u00e2pam\u00e2w awa kin\u00eapik\u201d? ni-k\u00ee- w\u00e2p=am -\u00e2 -w aw -a kin\u00eapikw 1- PREV-see=by.eye.TA-DIR-3 PRX-AN.SG snake CONSULTANT: Yes, you would say it that way, because I thought it was real! I thought it was going to bite me. 324 LINGUIST: What about now? Would you now say \u201cnik\u00ee-w\u00e2paht\u00ean \u00f4ma kin\u00eapik\u201d? ni-k\u00ee- w\u00e2p=aht -\u00ea -n aw=ima kin\u00eapikw 1- PREV-see=by.eye.TI-TI-LP PRX-AN.SG snake CONSULTANT: \u273dlaughing\u273d Well, yes, because it\u2019s not a real snake. It\u2019s made out of plastic. Thus, the availability of a rich discourse context supports a clear, definite judgment about animacy marking. This further supports the current analysis of animacy as a discourse-dependent mechanism in Plains Cree (\u00a73.4-5). 5.4.4. Correction tasks Another task that recurs during elicitation work is what Cook and M\u00fchlbauer (2006) label a \u201cCorrection task.\u201d In a task of this kind, consultants are presented with an unacceptable form and then asked to correct it. Consultants often volunteer a corrected form when presented with an unacceptable form. For example, if a linguist wants to test the presuppositions of Turkish (Turkic, Turkey) cleft constructions, the consultant may voluntarily fix the forms offered.8 (31) LINGUIST: What if I walked into the room, out of the blue, without any context, and I said kad\u0131n\u0131n g\u00f6rd\u00fcg\u00fc adamd\u0131? kad\u0131n -in g\u00f6r-d\u00fcg-\u00fc adam-di woman-nom see-part-acc man-past \u2018It was the man that the woman saw.\u2019 CONSULTANT: I think in that case, I would say, if I were you, kad\u0131n adam\u0131 g\u00f6rd\u00fc kad\u0131n adam-i g\u00f6r-d\u00fc woman man-acc see-past \u2018A woman saw a man.\u2019 Here, the consultant has voluntarily, and explicitly, fixed the form offered to a form that would be acceptable in the context described. This has provided crucial insight about the nature of these cleft constructions; they are dispreferred in environments where the referents are not already in the discourse. 8 Thanks to Kerim Demirci (p.c.) for this data and discussion of it. 325 Sometimes, consultants are explicitly asked to provide a better form than the one offered. For example, in the Turkish cleft examples, an explicit prompt for a corrected form can produce sentences with more sophisticated forms and better acceptability. (32) LINGUIST: Would \"kad\u0131n\u0131n g\u00f6rd\u00fcg\u00fc adamd\u0131\" be more appropriate if I was arguing? kad\u0131n -in g\u00f6r-d\u00fcg-\u00fc adam-di woman-nom see-part-acc man-past \u2018It was the man that the woman saw.\u2019 For example, if I wanted to say \"Hay\u0131r, hay\u0131r! kad\u0131n\u0131n g\u00f6rd\u00fcg\u00fc adamd\u0131!\" Hay\u0131r, hay\u0131r kad\u0131n -in g\u00f6r-d\u00fcg-\u00fc adam-di no no woman-nom see-part-acc man-past \u2018No, No! It was the man that the woman saw.\u2019 CONSULTANT: The man is not defined and I don\u2019t know him LINGUIST: Is there a better way to say it than these sentences? CONSULTANT: You should say kad\u0131n\u0131n g\u00f6rd\u00fcg\u00fc \u015fey adamd\u0131. kad\u0131n -in g\u00f6r-d\u00fcg-\u00fc \u015fey adam-di woman-nom see-part-acc EVID man-past \u2018The thing that the woman saw was the man.\u2019 Thus, correction tasks may either be a part of the linguist\u2019s elicitation plan, or an unplanned event offered by the consultant. Either way, the correction tasks provide important information about target forms. 5.4.41. Confounds in correction tasks: Un-noticed corrections Consultants sometimes make non-explicit corrections. This is particularly true if the linguist does not have much competency in the target language, especially the morphophonology of it. For example, I have seen exchanges like the following. (33) LINGUIST: Is it a good form if I say [nitkhis\u04d9thuskhiyan]? nit-k\u00ees- atosk\u00ea-y\u00e2n 1- finish-work -1 CONSULTANT: Uhh, no. No. I wouldn\u2019t know what you\u2019re trying to say. LINGUIST: How about if I change it and say [nitphun\u04d9thuskhiyan]? nit-p\u00f4n-atosk\u00ea-y\u00e2n 1- stop-work -1 326 CONSULTANT: Okay. [epun\u04d9toskeyan] is good. \u00ea- p\u00f4n-atosk\u00ea-y\u00e2n C1-stop-work -1 Here, the linguist is trying to test the acceptability of the two preverbs k\u00ees- and p\u00f4n- by applying them to the same verb form. However, they have unintentionally produced a series of Plains Cree forms that are unacceptable for other reasons: (i) they have mixed the independent order and the conjunct order (nit- plus -y\u00e2n), (ii) they have employed the wrong form of the person prefix (nit- instead of ni- before the preverb p\u00f4n-), (iii) they have heavily aspirated the stops when in onset position (something that Cree never does), and (iv) they have produced vowel qualities that are never heard in the natural language (e.g. [nit-] for [n\uf09ct-] or [nt-] or [nd-]). The first time these forms were presented, the consultant ruled them bad. However, after it was clear that the linguist was unable to \u201cdo better,\u201d the consultant appears to have found a compromise. The linguist would be mistaken were they to conclude that such data said anything about the relative acceptablility of the preverbs k\u00ees- and p\u00f4n-. In the previous example, the linguist was lucky enough to hear the form that the consultant had settled on as acceptable (e.g. \u00ea-p\u00f4n-atosk\u00eay\u00e2n \u2018I stopped working\u2019), which fixed all of the errors in the linguist\u2019s proffered form. This is not always the case; correction can often be done silently. This usually happens when the consultant has been pushed with bad forms for too long and finally finds something improved in a proffered form. Consider the continuation of the exchange in (33), which exemplifies this kind of scenario (34). (34) LINGUIST: Is it a good form if I say [nitkhis\u04d9thuskhiyan]? nit-k\u00ees- atosk\u00ea-y\u00e2n 1- finish-work -1 CONSULTANT: Uhh, no, that\u2019s not good. I don\u2019t know what you\u2019re trying to say. LINGUIST: How about if I change it and say [nitphun\u04d9thuskhiyan]? nit-p\u00f4n-atosk\u00ea-y\u00e2n 1- stop-work -1 CONSULTANT: Okay. [epun\u04d9toskeyan] is good. \u00ea -p\u00f4n-atosk\u00ea-y\u00e2n c1-stop-work -1 LINGUIST: Can I add [khi-] in there? Something like [nitphunikhi\uf0b0\u04d9thuskhiyan]? nit-p\u00f4n-k\u00ee- atosk\u00ea-y\u00e2n 1- stop-prev-work -1 CONSULTANT: No, that is not understandable. 327 LINGUIST: How about if I said [nikhiphunikhi\uf0b0\u04d9thuskhiyan]? nit-k\u00ee- p\u00f4n- atosk\u00ea-y\u00e2n 1- PREV-stop-work -1 CONSULTANT: Yes. That\u2019s good. You\u2019re saying \u201cI stopped working.\u201d Here, the linguist has repeatedly produced forms that are obviously not possible in Plains Cree. The linguist was corrected a few times, but has then unintentionally introduced another confound: the preverb k\u00ee- is in the wrong place in the preverbs; it should be ordered before p\u00f4n- \u2018stop,\u2019 not after it. When the linguist fixes this problem, placing it before p\u00f4n- instead of after it, and also has the correct form of the person prefix (ni- vs. nit-), the consultant identifies the improvement and responds positively, despite the impossibility of the form. Accomodating the linguist\u2019s bad forms, the consultant reconstructs a correct form in her head (\u00ea-k\u00ee-p\u00f4n-atosk\u00eay\u00e2n or nik\u00ee-p\u00f4n-atosk\u00e2n). (35) a. ACCOMMODATED FORM \u14c2\u146e \u1434\u1423 \u140a\u1455\u1422\u146b\u152e\u1423. nik\u00ee-p\u00f4n-atosk\u00eay\u00e2n ni-k\u00ee- p\u00f4n- atosk\u00ea-y\u00e2n 1- PREV-stop-work -1 --- (?) (Presented S2) b. ACTUAL FORM \u14c2\u146e \u1434\u1423 \u140a\u1455\u1422\u1473\u1423. nik\u00ee-p\u00f4n-atosk\u00e2n ni-k\u00ee- p\u00f4n- atosk\u00ea-y\u00e2n 1- PREV-stop-work -1 \u2018I stopped working.\u2019 (Presented S2) c. ACTUAL FORM \u1401 \u146e \u1434\u1423 \u140a\u1455\u1422\u146b\u152e\u1423. \u00ea-k\u00ee-p\u00f4n-atosk\u00eay\u00e2n \u00ea- k\u00ee- p\u00f4n- atosk\u00ea-y\u00e2n C1-PREV-stop-work -1 \u2018I stopped working.\u2019 (Presented S2) This means that a consultant\u2019s affirmation of a form is not always a straightforward case of grammaticality; the context of the judgment must be carefully considered. 328 If a linguist unintentionally establishes with a consultant that the linguist is both incompetent in the language and unable or unwilling to improve, the elicited data can diverge from the target forms. The consultant will cease focussing on the production of natural forms that would be ideal for the given context and instead accommodate the linguist\u2019s disfluency by making the utterances merely functional. Lost will be all the more sophisticated structure of the language, and elicitation results will be a pared-down, simplified form of the target language. If complex linguistic structure is the target, then, the development of basic competency in the target language is a necessity. In tandem with improved competency, there are strategies for making these silent corrections overt. This can be done by simply asking the consultant to repeat the \u201cgood\u201d form. If the form has been silently corrected, this will become immediately apparent; native speakers have a difficult time producing these bad forms. By constantly asking for the repetition of the form, then, the linguist will gain feedback, both on their own production skills, and on the forms that the consultant is judging acceptable. 5.4.42. Obviation in correction tasks When we consider obviation data with respect to correction tasks, an interesting pattern emerges. Consultants only rarely offer forms that are corrected for obviation coding. (36) LINGUIST: Is it a good sentence if I say \u201cClare w\u00e2pam\u00eaw Sucun\u201d? Clare w\u00e2p=am -\u00ea -w Shujun Clare see =by.eye.TA-DIR-3 Shujun CONSULTANT: Yes, that\u2019s a good sentence. You\u2019re saying \u201cClare sees Sucun.\u201d Instead, obviation mistakes are passed over silently, and more attention is paid to other features, such as morphophonology. This suggests that a form that is lacking in obviative marking can usually be fixed silently by the consultant, which would mean that obviation information is usually reconstructable from other contextual information. This is expected within the current framework, because obviative coding is constructed from several distributed parts of the grammatical system (\u00a73.3): (i) the stem-level agreement of verbs, (ii) the morpheme \u2013yi\u2013, (iii) the suffix \u2013im, (iv) the theme-sign system, and word ordering. All of these provide redundant cues to the identity of the missing obviative marking (i.e. reinforcement, as in Bateson 1972). Understanding obviative-less forms, then, is not too difficult for a Cree speaker. 329 If this is the case, and the consultant is silently reconstructing obviative forms, we would expect this reconstructed obviation to show up when the consultant is asked to repeat the obviative-less sentence. In fact, this is exactly what happens. (37) LINGUIST: Is it a good sentence if I say \u201cClare w\u00e2pam\u00eaw Sucun\u201d? Clare w\u00e2p=am -\u00ea -w Shujun Clare see =by.eye.TA-DIR-3 Shujun CONSULTANT: Yes, that\u2019s a good sentence. You\u2019re saying \u201cClare sees Sucun.\u201d LINGUIST: Could you say it, so I can hear how it\u2019s said? CONSULTANT: Sure. Clare w\u00e2pam\u00eaw Sucuna. Clare w\u00e2p=am -\u00ea -w Shujun-a Clare see =by.eye.TA-DIR-3 Shujun-XT Thus, the obviative marking (here, the suffix \u2013a) reappears when the consultant repeats the sentence. This underlines the importance of getting constant feedback from the consultant. 5.4.43. Animacy in correction tasks Based on memory and records, only one consultant (S4) has ever corrected an animacy form, in any context. This particular instance was when the consultant was instructing me on the canonical words for parts of the body. (38) CONSULTANT: The word for this \u273dpointing to legs\u273d is nsk\u00e2ta. nsk\u00e2ta \u00f4hi. ni-sk\u00e2t-a. ni-sk\u00e2t-a aw-ihi 1-leg -PL 1- leg -XT PRX-XT \u2018\u2026legs. TheseIN are my legsIN.\u2019 LINGUIST: nsk\u00e2tak \u00f4ki ni-sk\u00e2t-ak aw-iki 1-leg -PL PRX-PL \u2018TheseAN are my legsAN.\u2019 CONSULTANT: Nam\u00f4ya, nsk\u00e2t\u00e1 nam\u00f4ya ni-sk\u00e2t-a NEG 1-leg-XT \u2018No, my legsIN.\u2019 Here, I had purposely said the \u2018wrong\u2019 form, out of curiousity to see what the speaker would do. My form had the animate plural suffix \u2013ak, and the corresponding animate demonstrative \u00f4ki, rather than the consultant\u2019s proffered \u2013a and \u00f4hi. The speaker responded by repeating the correct 330 form, with an added prominence (high pitch, increased amplitude) on the suffix \u2013a, so that I could attend to the relevant feature.9 Crucially, the form that was corrected was not ungrammatical. This is demonstrated by further questioning. (39) CONSULTANT: The word for this \u273dpointing to legs\u273d is nisk\u00e2ta. nisk\u00e2ta \u00f4hi. ni-sk\u00e2t-a. ni-sk\u00e2t-a aw-ihi 1-leg -PL 1- leg -XT PRX-XT \u2018\u2026legs. TheseIN are my legsIN.\u2019 LINGUIST: nisk\u00e2tak \u00f4ki ni-sk\u00e2t-ak aw-iki 1-leg -PL PRX-PL \u2018TheseAN are my legsAN.\u2019 CONSULTANT: Nam\u00f4ya, nisk\u00e2t\u00e1 nam\u00f4ya ni-sk\u00e2t-a NEG 1-leg-XT \u2018No, my legsIN.\u2019 LINGUIST: Could I ever say nisk\u00e2tak \u00f4ki? ni-sk\u00e2t-ak aw-iki 1-leg -PL PRX-PL \u2018TheseAN are my legsAN.\u2019 CONSULTANT: \u00eah\u00e2. Yes. You could, you could. If your legs were kicking you, out of your control, you would say nisk\u00e2tak for them. ni-sk\u00e2t-ak 1- leg -PL \u2018My legsAN\u2019 Thus, the consultant, in offering the correction, has a particular context in mind. The correction is not a judgment between a good form and a bad form, then, but between a felicitous form and an infelicitous one relative to an implicit context. 9 Goddard (p.c.) points out that this kind of final stress pattern often occurs when eliciting paradigmatic variation; the consultant will stress the element that changes with each iteration. 331 5.4.5. Analysis tasks In an analysis task, the consultant considers the structure of their own grammar, either the properties of forms or concatenations of forms. Thus, this is the kind of task done when the consultant is themselves the linguist. Analysis tasks are often done in tandem with other kinds of elicitation tasks, or are embedded within an elicitation task. For example, a judgment task, wherein the consultant is asked to consider the acceptability of the form, will often lead the consultant to analyze the form, particularly when giving a negative judgment on the form. From this analysis, a correction is offered. (40) LINGUIST: Is it a good sentence if I say \u201c\u00e2h\u00e2siw \u00ea-kitocik\u201d in Cree? \u00e2h\u00e2siw \u00ea-kito -t-ik crow c1-call-3-pl CONSULTANT: No. The \u201c\u00e2h\u00e2siw,\u201d there\u2019s only one of them, but the \u201c\u00ea-kitocik\u201d says there is more than one. You need to put [\u04d9k] on the end of \u201c\u00e2h\u00e2siw\u201d if you want to say that. That makes it mean more than one crow. \u00e2h\u00e2siwak \u00ea-kitocik. The crows were making noise. \u00e2h\u00e2siw-ak \u00ea-kito-t-ik crow -PL C1-call-3-PL Thus, an analysis task is here an intermediate stage within two other kinds of tasks (here, a transition from a judgment task to a correction task). 5.4.51. Confounds: Decontextualizing forms, pseudo-linguistic explanations One confound of analysis tasks has to do with decontextualization. When a consultant considers a form in their language, the form is de facto removed from its context. It has ceased to be a part of the language structure and become an object of perception in itself. This means that whatever contextual properties it has will recede, and whatever formal properties the element has (i.e. those pertaining strictly to its form, rather than its function) will become foregrounded. If the element that is being analyzed in this fashion has little contextual functionality, such a process will not alter it significantly, but if the form has rich contextual properties, these will be greatly diminished or disappear under scrutiny. This is something that linguists must keep in mind when considering data from analytic tasks performed by consultants. Another confound for analysis tasks pertains to the tools the consultant uses for analysis. If the consultant has no linguistic training, the number of grammatical terms that the consultant 332 has will be very limited. Typically, there are two sets of terminology that consultants appeal to: time for verbs and plurality for nouns. Consider the example in (41). (41) LINGUIST: Is it okay if I say niw\u00e2pam\u00e2w ana n\u00e2p\u00eaw. ni-w\u00e2pam-\u00e2-w an =a n\u00e2p\u00eaw 1-see-dir-3 DIST=sg man CONSULTANT: Yes. niw\u00e2pam\u00e2w ana n\u00e2p\u00eaw. I see that man. LINGUIST: What if I changed it a little bit and I said niw\u00e2paht\u00ean ana n\u00e2p\u00eaw? ni-w\u00e2p=aht -\u00ea-n an =a n\u00e2p\u00eaw 1-see=by.eye.TI-TI-LP DIST=SG man CONSULTANT: Hmm. Yes. It\u2019s still good. niw\u00e2paht\u00ean ana n\u00e2p\u00eaw. LINGUIST: What is the difference between the two ways of saying it? CONSULTANT: Well, the first form is if you\u2019re seeing the person now, I guess, and the second form is if you saw them in the past.10 Here, the consultant has appealed to a contrast in tense readings to explain a contrast in animacy coding on the verb. This is not ironclad proof that the animacy agreement is temporal in nature, however; almost all verbal contrasts are explained as contrasts in tense. Thus, the consultant is here appealing to tense because she has little else in the way of grammatical categories to employ. This makes the linguistic training of consultants particularly relevant to elicitation work, a topic that I specifically address below. 5.4.52. Obviation in analysis tasks When consultants analyze obviative forms, they almost invariably remove the obviative suffix. Almost invariably, the pattern of this kind of removal follows the sequence in (42). 10 This data was actually collected from (S2). I have no idea what it means, and have been unable to find parallel data anywhere else. On other occasions, she rejects these offered TI forms. This makes me think it is an elicitation confound of some kind. Notice, in particular, that the TI form was offered by the linguist, rather than volunteered by the consultant. 333 (42) LINGUIST: Is it a good sentence if I say \u00ea-w\u00e2pam\u00e2t aw\u00e2sis min\u00f4sa \u00ea-w\u00e2p=am -\u00e2 -t aw\u00e2sis min\u00f4s-a c1-see=by.eye.TA-DIR-3 child cat -XT CONSULTANT: Yes. \u00ea-w\u00e2pam\u00e2t aw\u00e2sis min\u00f4sa. min\u00f4s\u00e1. Hmm. Usually, we say min\u00f4s. I don\u2019t know what the [\u04d9] is doing at the end. LINGUIST: So you don\u2019t need the [\u04d9] on min\u00f4s? CONSULTANT: No, min\u00f4s is fine. In this exchange, the consultant has begun by accepting an obviative form, but then has considered the obviative-referring nominal min\u00f4sa \u2018the cat(s)OBV.\u2019 To consider this form, she repeats it without the rest of the sentence, in what amounts to a citation form. This time, she places English-type stress on the suffix \u2013a (increased amplitude). She then rules the obviative- marked form as ill-formed. Crucially, the consultant has here pulled the nominal form min\u00f4sa out of its sentential context and considered it by itself. In so doing, she has removed the reason for the obviative suffix to be affixed; the nominal was obviative-marked because of the context it was used in. Removed from this context, there is no reason to mark the nominal with obviation, and thus the obviative disappears. This is as expected by a model that treats obviation as a property of contexts and constructions (cf. Chapters 3 & 4). 5.4.53. Animacy in analysis tasks In elicitation settings, Cree speakers often analyze animate forms as being \u201calive\u201d in some way. (43) LINGUIST: Is it a good sentence if I say \u201cniw\u00e2pam\u00e2w awa asiniy\u201d in Cree? ni-w\u00e2p=am -\u00e2 -w aw =a asiniy 1- see =by.eye.TA-DIR-3 PRX=AN.SG rock CONSULTANT. niw\u00e2pam\u00e2w awa asiniy. Yes. That\u2019s good. LINGUIST: What about if I say \u201cniw\u00e2paht\u00ean \u00f4ma asiniy?\u201d ni-w\u00e2p=aht -\u00ea -n aw=ima asiniy 1- see =by.eye.TI-TI-LP PRX=AN.SG rock CONSULTANT: niw\u00e2paht\u00ean \u00f4ma asiniy. Hmm. I prefer the other way [niw\u00e2pam\u00e2w awa asiniy]. This way sounds funny. LINGUIST: Okay. 334 CONSULTANT: \u273dpause\u273d Do you think rocks are alive? Do you think they have power of some kind? I mean, they stay together. They don\u2019t fall apart, through time. \u273dpause\u273d What about airplanes? They have powers, I guess\u2026 Here, the consultant has been asked to consider which form is preferred for animacy marking. She rules that the animate form awa asiniy \u2018thisAN rock\u2019 is preferred. She then begins to consider the animate-marked asiniy, and wonders aloud if her grammatical coding of it as animate means that it is \u201calive\u201d in some sense. She then thinks of other forms that she codes this way (e.g. pimh\u00e2kan \u2018airplane\u2019) and extends the \u201cliving\u201d property to them by analogy. Thus, an analysis task has led the consultant to interpret the animate form as denoting a \u201cliving\u201d thing. With the model of animacy proposed by Goddard (2002) and adopted in this thesis, we can understand this \u201canimate\u201d interpretation as arising from setting up a contrast between animate and inanimate forms. Recall from Chapter 2 that \u201canimate\u201d acquires an intentional reading in the context of an \u201cinanimate\u201d form, via blocking. In this kind of elicitation, we have done exactly this; the animate form and inanimate form of the same nominal are considered, and the animate form is selected. Thus, analysis contexts lead consultants to consider alternatives, and alternatives inherently involves the notion of blocking. Blocking with animacy ends up conveying that some referents are inherently intentional, even though they are not normally used that way in the language. Elicitation, then, has provided an interesting environment to study the effects of animacy and blocking, and supports the model of animacy proposed in this thesis. 5.5. Elicitation and impoverished context In all of the elicitation data seen here, the recurrent theme is one of context impoverishment. In translation tasks, the context for the target sentence often becomes the shared-language sentence. In judgment tasks, the consultant often fails to transfer the target form with its context, something that is fixed by utterance-in-context tasks. Finally, in analysis tasks, the process of analysis itself can often decontextualize a form, altering judgments on its well-formedness. While it may be tempting to conclude that elicitation is the wrong place to test contextually-sensitive forms, since it has so little context, in fact there is evidence to suggest that the context-impoverishment of elicitation provides crucial insights. So long as we are aware that elicitation often lacks context, we can use this to our advantage, considering what the target phenomenon does under this strain. 335 In the following sections, I consider how obviation in particular behaves in context-less constructions. I argue that the behaviour of obviation in elicitation environments demonstrates the contentful, marked nature of obviation. The obviative form bears semantic content, and, as such, it has conditions on its use in context. Elicitation, being context-poor, does not usually provide enough specification for these conditions of use to be met, and thus, the speaker does not produce obviative forms. Only when the linguistic environment that exists in elicitation is altered do obviative constructions begin to appear; the consultant learns to accommodate the linguistic context of elicitation. 5.5.1. The proliferation of simple animates in elicitation Probably anyone who has done elicitation on an Algonquian language will attest to the fact that obviative forms are not as easily obtained as simple animates. For example, asking a Plains Cree speaker how to translate \u201cthe dog is barking\u201d will invariably result in the use of a simple animate referent (44). (44) \u140a\u144e\u1466 \u1401 \u14aa \u14a5\u14ef\u146d\u14a7\u141f. atim \u00ea-ma-misikimot. atimw \u00ea- ma- misikimo-t dog c1-RED-bark -3 \u2018The dogAN is barking\u2019 (Translation S3) While this is a form that could, in theory, occur in natural speech, the preference for simple animate forms goes beyond the boundaries established by textual data (cf. Cook & M\u00fchlbauer 2006). Crucially, these unexpected animates occur in contexts where obviation is expected to occur. For example, simple animate referents can be possessed by third persons (45), which is not supposed to happen (cf. Wolfart 1973, Cook & M\u00fchlbauer 2007, Chapter 4, etc). (45) a. OFFERED FORM \u14c2\u141a\u1438\u14ab\u1424 \u140a\u1418 \u1405\u1434\u14f0\u14a5\u1422. niw\u00e2pam\u00e2w awa op\u00f4s\u00eems ni-w\u00e2p=am -\u00e2 -w aw=a o-p\u00f4s\u00ee=m -is 1-see =by.eye=TA-DIR-3 prx=an.sg 3-cat=DSJ-DIM \u2018I saw this person\u2019sAN catAN.\u2019 (Translation S1) 336 b. EXPECTED FORM \u14c2\u141a\u1438\u14a5\u14ab\u1418 \u140a\u1418 \u1405\u1434\u14f0\u14a5\u14f4. niw\u00e2pamim\u00e2wa awa op\u00f4s\u00eemsa ni-w\u00e2p=am -im -\u00e2 -w-a aw=a o-p\u00f4s\u00ee=m -is -a 1-see =by.eye=TA-DSJ-DIR-3-XT prx=an.sg 3-cat=DSJ-DIM-XT \u2018I saw this person\u2019sPROX catOBV.\u2019 Simple animate forms can also occur as both arguments of a transitive verb (46), which is unexpected based on grammatical descriptions of textual data (cf. Wolfart 1973). (46) a. OFFERED FORM Tomio \u14c2\u1438\u1426\u1401\u1424 \u140a\u1418 \u146d\u14c0\u1431\u146f. Tomio nipah\u00eaw awa kin\u00eapik. Tomio nip=ah -\u00ea -w aw =a kin\u00eapikw Tomio kill=by.tool.TA-DIR-3 PRX=AN.SG snake \u2018TomioAN killed this snakeAN.\u2019 (Translation S1) b. EXPECTED FORM Tomio \u14c2\u1438\u1426\u1401\u1424 \u1406\u1426\u1403 \u146d\u14c0\u1431\u147f. Tomio nipah\u00eaw \u00f4hi kin\u00eapikwa. Tomio nip=ah -\u00ea -w aw =ihi kin\u00eapikw-a Tomio kill=by.tool.TA-DIR-3 PRX=XT snake -XT \u2018TomioPROX killed this snakeOBV.\u2019 Further, simple-animate forms can occur in clauses embedded under third person subjects (48), which is also not usually seen in texts.11 (47) a. OFFERED FORM \u1456\u14c2\u14ef \u1401 \u14ef \u141a\u1438\u1426\u1455\u1541 Misti \u1401 \u14ef \u14ef\u143b\u1426\u144c\u141f \u141a\u1438\u1422\u144e\u1466? t\u00e2nsi \u00ea-si-w\u00e2pahtahk Misti \u00ea-si-sipw\u00eaht\u00eat W\u00e2pastim? t\u00e2nsi \u00ea- si-w\u00e2p=aht -am-k Misti \u00ea- si- sipw\u00ea=ht\u00ea -t W\u00e2pastimw how C1-RR-see=by.eye.TI-TI -0 Misti C1-RR-away=by.walking-3 W\u00e2pastimw \u2018How did MistiAN see W\u00e2pastimAN leave?\u2019 (Translation S1) 11 Proper names also seem to resist obviative-marking in texts. See, for example, the full extent of obviative-less proper names in Emme Minde (1997). 337 b. EXPECTED FORM \u1456\u14c2\u14ef \u1401 \u14ef \u141a\u1438\u1426\u1455\u1541 Misti \u1401 \u14ef \u14ef\u143b\u1426\u144c\u1528\u141f \u141a\u1438\u1422\u144e\u14b7? t\u00e2nsi \u00ea-si-w\u00e2pahtahk Misti \u00ea-si-sipw\u00eaht\u00eayit W\u00e2pastimwa? t\u00e2nsi \u00ea- si-w\u00e2p=aht -am-k Misti \u00ea- si- sipw\u00ea=ht\u00ea -yi-t W\u00e2pastimw-a how C1-RR-see=by.eye.TI-TI -0 Misti C1-RR-away=by.walking-DS-3 W\u00e2pastimw-XT \u2018How did MistiAN see W\u00e2pastimAN leave?\u2019 (Translation S1) Altering elicitation techniques to avoid translation likewise nets largely simple forms. Thus, presenting a speaker with forms (\u2018judgment\u2019 tasks), asking for Cree sentences in a context (\u2018context\u2019 tasks), or asking for their consideration of forms (\u2018analysis\u2019 task) results in the same patterns. We can try to account for the gaps by relying only on context-less \u201cgrammatical\u201d explanations, but the attempt will fail. It cannot be a simple morphological gap, where affixal material on nominals is lost: (i) speakers readily produce other nominal affixes such as the suffix -ak (48), (48) \u140a\u141a\u14ef\u14f4\u1420 \u14ab\u1450\u1418\u1420. aw\u00e2sisak m\u00e2towak. aw\u00e2sis-ak m\u00e2to-w-ak child -PL cry -3-PL \u2018The childrenAN cried.\u2019 (Translated S1) and (ii) the verbal equivalents to obviation (e.g. the suffix -yi-) is also missing (51). It likewise cannot be an issue of syntactic agreement between words: (i) plurality is coded readily on both nouns and verbs, in both sets of clausal agreement (48), and (ii) more complex chains of adjunct extraction agreement (e.g. \u2018relative roots,\u2019 see Cook 2005) are intact, even across multiple clauses (47). By contrast, the current model offers a way to understand this patterning. In chapter 2, I argued that the simple animate form is unmarked with respect to content (\u00a72.4); it bears no special meaning. By contrast, the obviative referents are specified as contextually extentional (cf. \u00a73.4-5), which means that they fundamentally require a context to be evaluated in. Without a context, obviative constructions will not be called for. Elicitation, being naturally context-poor (cf. Matthewson 2004, Cook & M\u00fchlbauer 2006, Calecott & Koch 2007), would be exactly such an environment. 338 5.5.2. Obtaining \u201cobviative\u201d forms in elicitation While simple animate forms are freely volunteered by consultants, the situation is much different with obviative forms. Obviative forms are never volunteered for simple declarative sentences. In fact, getting an obviative form at all in elicitation requires work, since, as we have just seen (\u00a75.5.1), obviatives do not even appear where the grammar expects them to be. Presenting an obviative form to a consultant out of the blue will usually12 end in rejection (49). (49) OUT-OF-THE-BLUE \u201cOBVIATIVE\u201d IS REJECTED a. \u1401\u1422\u1475\u1424 \u1401 \u14a5\u152a\u14ef\u141f. iskw\u00eaw \u00ea-miyosit. iskw\u00eaw \u00ea- miyo=isi -t woman C1-good=STAT-3 No context = \u2018A woman is beautiful\u2019 (Judgment S2) b. # \u1401\u1422\u1475\u1418 \u1401 \u14a5\u152a\u14ef\u1528\u141f. # iskw\u00eawa \u00ea-miyosiyit iskw\u00eaw-a \u00ea- miyo=isi-yi-t woman-XT C1-good=STAT-DS-3 (Judgment S2) COMMENT (S2): \u201cWho are you talking about??\u201d While the \u201cAnimate\u201d form in (49a) is readily accepted, the speaker rejects the obviated form because she does not know who the referent is. In attempting to repair such a sentence, the speaker will have to invent a context; for example, the form in (49b) was repaired by creating a context in which we were discussing the girlfriend of a man we knew. Without such a context, the sentence is completely infelicitous. While subordinate clauses often lose their obviation in elicitation (see 47), there is one kind of subordination that seems to consistently produce it. When then matrix verb denotes an act of speaking, the referents of subordinate clauses will regularly be obviated. Thus, the same speaker who produced the obviation-less example in (47) ruled the parallel quotative form in (50) \u2013 which lacked obviation \u2013 ungrammatical at nearly the same time. 12 Sometimes consultants will accept these forms, but their judgments are always based overtly on \u2018being able to understand what you mean,\u2019 i.e. the truth-conditions are clear (some woman is pretty), even though the contextual structure is wrong. When questioned, they will affirm that they themselves would never utter such a sentence in the (non-)context considered here. 339 (50) QUOTATIVE CLAUSE MUST HAVE \u201cOBVIATIVE\u201d \u273d \u1456\u14c2\u14ef \u1401 \u14ef \u1403\u1458\u141f Misti \u1401 \u14ef \u14ef\u143b\u1426\u144c\u141f \u141a\u1438\u1422\u144e\u1466. \u273d t\u00e2nsi \u00ea-si-tw\u00eat Misti \u00ea-si-sipw\u00eaht\u00eat W\u00e2pastim. t\u00e2nsi \u00ea- isi-it =w\u00ea-t Misti \u00ea- isi-sipw\u00ea=ht\u00ea -t W\u00e2pastimw how C1-RR-thus=AI-3 Misti C1-RR-away =by.walking-3 W\u00e2pastimw Intended: \u2018What did MistiAN say was the way W\u00e2pastimAN walked away?\u2019 (Judgment S1) Correcting this offered form, the speaker produced (51), which obviates the embedded referent and its clause. (51) QUOTATIVE CLAUSE CORRECTED FOR OBVIATION \u1456\u14c2\u14ef \u1401 \u14ef \u1403\u1458\u141f Misti \u1401 \u14ef \u14ef\u143b\u1426\u144c\u141f \u141a\u1438\u1422\u144e\u1466. t\u00e2nsi \u00ea-si-tw\u00eat Misti \u00ea-si-sipw\u00eaht\u00eayit W\u00e2pastimwa. t\u00e2ns \u00ea- isi-it =w\u00ea-t Misti \u00ea- isi-sipw\u00ea=ht\u00ea -yi -t W\u00e2pastimw how C1-RR-thus=AI-3 Misti C1-RR-away =by.walking-DS-3 W\u00e2pastimw Intended: \u2018What did MistiAN say was the way W\u00e2pastimAN walked away?\u2019 (Correction S1) Based on the textual evidence above, we have already concluded in Chapter 3 that obviation is highly sensitive to quotative environments. Thus, introducing a verb of speaking in elicitation provides the right kind of contextual specification to meet the obviative\u2019s conditions of use; building context supports the occurrence of obviative constructions. 5.5.3. Contextless obviation = \u201cinanimate\u201d Recall from Chapter 2 that inanimate nominals are distinguished by their inability to have intentions; they denote \u2018mindless\u2019 referents (extentional). Thus, the inanimate-marked nominal in (52) denotes an entity that will never be able to think, feel, or speak without a supernatural intervention (at which time, it must be shifted to animate \u2013 see Chapter 2). (52) INHERENTLY EXTENTIONAL REFERENT = \u201cINANIMATE\u201d \u14c2\u146e \u141a\u1438\u1426\u144c\u1423 \u1406\u14aa \u146d\u14c0\u1431\u146f. nik\u00ee-w\u00e2paht\u00ean \u00f4ma kin\u00eapik. ni-k\u00ee- w\u00e2p=aht -\u00ea-n aw=ima kin\u00eapikw 1-PREV-see=by.eye.TI-TI-LP PRX=in.sg snake \u2018I sawIN thisIN snake\u2019 (Judgment by S2) Context: Speaker sees a snake, then realizes it is just a rubber toy. If obviative-referring nominals share semantic properties with inanimate nominals (i.e. both refer to extentional referents, differing only in their contextuality; \u00a72.4, 3.4) , we expect that, in certain situations, obviative constructions will give rise to a \u2018mindless\u2019 interpretation. This is exactly 340 what happens when an obviative referent is not adequately embedded in a discourse. This is easily done in elicitation, where contexts are often too impoverished. For example, when I attempted to construct a canonical reportative context, the speaker rejected it on the grounds that I was making one of the human referents mindless (53). (53) HUMAN REFERENT MADE \u2018MINDLESS\u2019 BY OBVIATION i. \u273d\/# Clare \u1401 \u146d\u152a\u1472\u141a\u141f Martha\u1418 \u1401 \u1431\u1433\u14c2\u1528\u1420. \u2026 \u273d\/# Clare \u00ea-kiyokaw\u00e2t Marthawa \u00ea-piponiyik. Clare \u00ea-kiyok=aw-\u00e2 -t Martha-a \u00ea- pipon -yi-k Clare c1-visit=TA -DIR-3 Martha-xt c1-winter-DS-0 Intended: \u2018ClarePROX went to visit Marthaobv during the winterobv. \u2026 ii. \u2026 \u201c\u14c2\u14a5\u1429\u140d\u1528\u1426\u144c\u1423 \u1406\u14aa,\u201d \u1401 \u1403\u1458\u1528\u141f \u1405\u14f0\u14a5\u14f4. \u2026 \u201cnimy\u00eayihten \u00f4ma,\u201d e-tw\u00eayit os\u00eemsa. ni-miyw\u00eayiht-\u00ea-n aw-ima \u00ea-itw\u00ea-yi-t o-s\u00eemis-a 1-like-ti-lp prox-sg C1-say-dep-3 3-young.sis-obv \u2026 \u201cI\u2019m happy about this,\u201d herPROX younger sisterobv said.\u2019 (Presented S2) Comment (Speaker 2): \u201cIf we use \u2018\u00ea-tw\u00eayit,\u2019 it would be more like a recording, something coming off of an answering machine. It doesn\u2019t sound like it\u2019s coming from a person. Maybe if it was a big committee.\u201d Here, the speaker takes the obviation of the referent os\u00eemsa \u2018her younger sister\u2019 (\u2018Martha\u2019 in the context) to mean that the referent is not human or is not speaking as a human. Based on the data we saw in chapter 2 (e.g. \u00a72.4), this is an inanimate interpretation; the referent is only speaking as a mechanical entity produces noise (i.e. \u2018mindless speaking\u2019). Data of this kind, where obviative speakers are rejected based on claims of \u201cmindlessness\u201d is likely an artifact of the elicitation environment. This is confirmed by going back and being more careful in the construction of the context of the utterance (e.g. saying the sentence altogether, in one clean utterance, rather than hesitating between sentences). When this is done, the sentence in (53) is judged perfectly acceptable by both (S2) and (S3). Thus, the ungrammaticality or infelicity was a product of a failure to link the discourse-dependent structures together (i.e. a failure to embed the obviative in the proximate\u2019s perspective; \u00a73.5), not a failure in obviation to co-occur with human referents. 341 5.6. The consultant as linguist Many of these elicitation tasks surveyed in this chapter crucially rely on the analytic and linguistic ability of the consultant as well as the linguist. This raises the issue of how much knowledge a consultant should have of the targeted material in an elicitation situation. There are essentially two positions on this that have been taken by linguistic fieldworkers: (i) the \u201cexperimentalist\u201d approach, and (ii) the \u201ccooperative\u201d approach. In this section, I consider what the differences are between these approaches, and how those differences affect the kind of fieldwork carried out. 5.6.1. The \u201cexperimentalist\u201d position: Consultant as data bank One view is that the consultant should be as na\u00efve as possible, with absolutely no idea of what the linguist is doing during the elicitation session. This is based on the idea that, were the consultant to understand something about the grammatical targets, it would bias the data. Based on its parallelism to experimental work, I label this model the \u201cexperimental\u201d position. Linguists who take this approach cannot allow for much discussion or reaction during an elicitation session, because they belief it could lead to biases in the data. Likewise, discussion among the linguists present, about what forms should next be considered, cannot be done, since this would undo the principle of keeping the consultant na\u00efve as to what is being targeted. This means that \u201cexperimentalist\u201d linguists must rely on careful planning sessions ahead of time. Since the lists of forms to be obtained are pre-planned, they are usually systematic in their organization. For example, a pre-planned list for a Plains Cree session targeting possession forms may look like (54). (54) ELICITATION PLAN FOR OCTOBER 18, 2007 1. min\u00f4s 2. nimin\u00f4s 3. nimin\u00f4sim 4. kimin\u00f4s 5. kimin\u00f4sim 6. omin\u00f4s 7. omin\u00f4sim 8. omin\u00f4sima \u2026 342 These lists are then presented to the consultant one after the other. The consultant\u2019s role is to provide the grammaticality judgment for each form; the consultant is a data bank for the grammaticality of forms. 5.6.2. The \u201ccooperative\u201d position Another approach to elicitation is to train the consultant in linguistics as far as the consultant desires it. On such a view, the more the consultant understands the context of fieldwork, the better the data they will offer. In line with work on conversational analysis done by Schlegloff and Sacks (e.g. Schlegloff & Sacks 1992), the consultant is taken to have numerous intuitive linguistic analytic abilities which are crucial in understanding the properties of the grammar. The primary goal is to develope a cooperative relationship between the linguist and the consultant (who is a natural, intuitive linguist, and a soon-to-be-trained linguist); I label this the \u201ccooperative\u201d position. Linguists who take this position will often discuss the data with the consultant, explaining what forms are being targeted, what has been puzzling about them up to now, and, as far as the consultant can understand the linguists\u2019 jargon, what the linguist thinks may be happening. Sometimes, this work is done with a consultant who is themselves a fully- trained linguist. Going into an elicitation situation from this perspective means that the linguist can never be sure what kind of data will become the center of discussion, since it depends on more than their own goals. Thus, making engagement of the consultant in the linguistic analytic process a primary goal means that there is necessarily less ability for the linguist to follow a planned elicitation schedule. Instead, the linguist comes with an idea of what kind of forms they are curious about, perhaps even a detailed plan, and often begins the elicitation session by discussing them with the consultant. However, this plan is always only tentative. This has the result of producing a relaxed, informal setting. The consultant is as involved in the linguistic work as they want to be, and the conversation is unimpeded, and can take whatever direction it happens to. As fieldwork of this kind progresses, the consultant becomes more and more of a trained linguist, and thus becomes a colleague in the work, instead of a data bank. Most people are interested in their own language. Anybody that speaks a language has had to consider linguistic issues in considerable detail, and thus not only are most speakers capable of considerable insight about their native language, but are quite happy to talk about it and think about it more. In such a situation, the consultant often \u201ctakes their work home,\u201d 343 thinking about targeted forms when not even in the elicitation setting. This has the significant advantage of engaging the consultant directly in the process of constructing context to support forms. 5.6.3. The two types of fieldwork and obviation data In this thesis, it has been shown repeatedly that the forms under consideration are highly sensitive to discourse context: (i) the use of animate-marking is sensitive to the knowledge of the Speaker and the contextual ability of the referent to hold a perspective (Chapter 2), (ii) the use of obviative constructions is sensitive to the contextual knowledge of the Speaker about a referent and that referent\u2019s (non)ability to possess a perspective (Chapter 3), and (iii) the construction of referential dependency requires a context in which the dependent referent can be embedded inside the perspective of its antecedent (Chapter 4). Since these two approaches to fieldwork produce significantly different speech contexts, it should be no surprise that they often produce different kinds of data with respect to these constructions. Here, I consider how they effect one of these areas \u2013 obviation. Experimentalist elicitation tactics obtain only obviation patterns that can be produced with reference to extremely local information. Minimally, this means that third-person-possessed forms can reliably be gotten. (55) WORD-LEVEL OBVIATION \u1405\u14a5\u14c5\u14ef\u14aa omin\u00f4sima o-min\u00f4s=im-a 3-cat=DSJ-XT \u2018his\/herPROX catOBV\u2019 Often, speakers can produce and control sentence-level obviation forms. (56) SENTENCE-LEVEL OBVIATION \u1403\u1422\u1475\u1424 \u141a\u1438\u14a3\u1424 \u140a\u141a\u14ef\u14f4. iskw\u00eaw w\u00e2pam\u00eaw aw\u00e2sisa. iskw\u00eaw w\u00e2p=am -\u00ea -w aw\u00e2sis-a woman see=by.eye.TA-DIR-3 child -XT \u2018The womanPROX saw the childOBV.\u2019 344 However, the more complex discourse-driven forms are often absent, for example, all the multi- clause and logophoric judgments seen in Chapter 3. Crucially, because the linguist has not provided the consultant with any linguistic training, the consultant cannot provide a check on the linguist\u2019s conclusions. The linguist, believing that these forms have been gotten in a proper experimental way, without bias on the part of the consultant, concludes that this is an accurate representation of obviative forms. In a cooperative approach, conversations like (57) are not uncommon, which happened when I talked to a consultant specifically about obviative forms. (57) LINGUIST: I\u2019m curious about the funny \u2018-a\u2019 that shows up on the end of nouns some times, like in \u201cn\u00e2p\u00eaw w\u00e2pam\u00eaw min\u00f4s\u00e1.\u201d n\u00e2p\u00eaw w\u00e2p=am -\u00ea -w min\u00f4s-a man see=by.eye.TA-DIR-3 cat -XT CONSULTANT: Oh, yes. n\u00e2p\u00eaw w\u00e2pam\u00eaw min\u00f4s\u00e1. Hmm. min\u00f4sa. You see, you have to use it that way when it\u2019s in a sentence. It\u2019s because it\u2019s in a sentence. LINGUIST: But what about if I say \u201cmin\u00f4s w\u00e2pisksiw\u201d? There\u2019s no \u2013a there, is there? min\u00f4s w\u00e2piski=si -w min\u00f4s white=STAT-3 CONSULTANT: Yeah, there isn\u2019t. min\u00f4s w\u00e2psksiw. Hmm. \u273dpause\u273d It\u2019s because the one before had another noun in it. Otherwise you wouldn\u2019t know who was the one being seen, maybe. n\u00e2p\u00eaw w\u00e2pam\u00eaw min\u00f4s. No, you still know who\u2019s being seen. Hmm. LINGUIST: Some linguists have thought it had to do with how the speaker thinks about that character, the one with the -a on it. They call that thing an \u201cobviative marker.\u201d And they say that it\u2019s on there because that character is less- important, somehow. CONSULTANT: Well, yes, the cat is definitely less important than the man in that example, so I see what they\u2019re saying. LINGUIST: Is it always less important? What about if I say \u201caymh\u00eaw\u00eeniw \u00ea-mamsk\u00f4m\u00e2t m\u00e2ntowa\u201d? ayamih\u00eawiyiniw \u00ea- mamisk\u00f4=m -\u00e2 -t m\u00e2nitow-a priest C1-talk.about=by.mouth-DIR-3 god -XT \u2018The priestPROX is talking about GodOBV.\u2019 CONSULTANT: Oh! Hmm. 345 LINGUIST: Is it better to say it the other way around? \u201cm\u00e2ntow \u00ea-mamsk\u00f4mkot aymh\u00eaw\u00eeniwa\u201d? m\u00e2nitow \u00ea-mamisk\u00f4=m -iko -t ayamih\u00eawiyiniw -a god c1-talk.about=by.mouth-INV-3 priest -XT \u2018GodPROX is talked about by the priestOBV.\u2019 Is that\u2026 CONSULTANT: \u273dinterrupting\u273d Okay. The sentence is good. \u00e2ymh\u00eaw\u00eeniw \u00ea-mamsk\u00f4m\u00e2t mantowa. But it\u2019s not that the priest is more important than God. It\u2019s just that the priest is talking at the time. He\u2019s just the one you\u2019re hearing about it from. Thus, by educating the consultant about the form, and the claims that are made about this form, the consultant has produced an insightful judgment about obviation. Some of the most crucial data in this thesis could never have been gotten from an uninformed speaker. One of the speakers I have learned a great deal from (S4) was completely uninterested in traditional elicitation at all, much less the formal approach of experimentalists. Instead, we13 took a cooperative approach and discussed forms with him. For example, when considering the evidential properties of clauses and obviation, we found a passage in a text that was of interest for the particular forms we were puzzled about (Ahenakew\u2019s telling of her husband\u2019s bear vision), explained that we were looking for, and then went through the text line by line with him, to his great interest. When we were about halfway through it, a conversation ensued in which the spaker produced a crucial set of obviative forms as a paradigm, with an explanation of what they meant and what contexts to use them in. (58) a. \u1401\u146f\u14ef \u1404\u1458\u1424 \u140a\u14c7 \u1438\u141a\u1472\u1423. \u00eakos \u00eetw\u00eaw ana paw\u00e2kan. \u00eakosi it =w\u00ea-w an =a paw\u00e2kan thus thus=AI -3 DST=AN.SG dream.spirit \u2018That\u2019s what the dreamAN spirit said\u2019 (Volunteered S4) b. \u1401\u146f\u14ef \u1401 \u1403\u1458\u141f \u140a\u14c7 \u1438\u141a\u1472\u1423. \u00eakos \u00ea-tw\u00eat ana paw\u00e2kan \u00eakosi \u00ea- it =w\u00ea-t an =a paw\u00e2kan thus C1-thus=AI -3 DST=AN.SG dream.spirit \u2018That\u2019s what the dream spiritAN said\u2019 (Volunteered S4) 13 This work was done jointly with Clare Cook. 346 c. \u1401\u146f\u14ef \u1401 \u1403\u1458\u1528\u141f \u140a\u14c2\u1426\u1403 \u1438\u141a\u1472\u14c7. \u00eakos \u00ea-tw\u00eayit anihi paw\u00e2kana \u00eakosi \u00ea- it =w\u00ea-yi-t an =ihi paw\u00e2kan -a thus c1-thus=AI-DS-3 DST=XT dream.spirit-XT \u2018That\u2019s what the dream spiritOBV said\u2019 (Volunteered S4) In my opinion, data of this kind is as reliable and insightful as a linguist can hope to get in elicitation settings. 5.7. Conclusion: Elicitation as performance, not competence All of the elicitation data, when properly understood, supports the claim that animacy and obviation are context-dependent. As we have surveyed the kinds of events that occur in elicitation environments, we continually come back to the same conclusion: elicitation is a task. Elicitation is something that consultants and linguists do; it is an action. As an action, it became relevant to ask if it is like other linguistic actions or not. In answer, we have seen repeatedly that elicitation is not like other speech tasks that a consultant does in their use of their language. Thus, elicitation is a task, and elicitation is a new task. Chomsky (1965:3) has made reference to what he considers a fundamental division in linguistics: performance vs. competence. Performance can be defined as \u201cthe production of specific linguistic forms in specific situations,\u201d while competence means \u201cthe fluent native speaker\u2019s knowledge of the language\u201d (Radford 1988:3). Competence is thought to underly performance, and is considered the proper realm of linguistic study (cf. Chomsky 1965:3). In studying a particular language, then, a linguist seeks to cut away all of the performance, and expose the the underlying knowledge that a fluent speaker has. Considered in light of this division, one can conclude that elicitation data is performance data. That is, it reflects the ability of the consultant to perform a certain set of tasks in a context- poor environment. As performance data, its relation to the abstract properties of the language system (\u201cCompetence\u201d) are not always immediately obvious. Since it is performance data, elicitation data can mislead the linguist if it is taken to provide straightforward evidence about competency. Instead, the nature of elicitation tasks must be considered carefully, with awareness that different tasks produce different results. With this approach established, there is no longer bad data, only misunderstood data. 347 Chapter 6 Conclusion and Further Research 6.1. Conclusion: Constructing intentionality in Plains Cree In this thesis, I have shown that Plains Cree systematically distinguishes between four classes of referents with respect to perspective possession (intentionality): (i) \u201cINANIMATE\u201d : Referents that never have perspectives (\u00a72). (ii) \u201cANIMATE\u201d : Referents that are unspecified for this property (\u00a72). (iii) \u201cOBVIATIVE\u201d : Referents that have no contextual perspective (\u00a73), (iv) \u201cPROXIMATE\u201d : Referents that have a contextual perspective (\u00a74). While class (i) and class (ii) are coded by dedicated forms in the language (\u00a72.3), and thus represent primitives of Plains Cree grammar, class (iii) is a constructed category, being built out of forms that code referential dependency (\u00a74). Class (iv) differs from the other three in having no grammatical exponence at all, being built solely from contextual contrast. In modelling the semantic properties of these classes, I have used the notion of a \u2018perspective\u2019 (K\u00f6lbel 2002), which situates a proposition as being true with respect to someone\u2019s perspective (\u00a71.3). The possession of a perspective, then, results in what is commonly called \u201cIntentionality.\u201d Plains Cree constructs intentionality from the linguistic forms that code dependency and perspectival embedding, discriminating perspective-holders (Intentional referents) from those not allowed to possess a perspective (Extentional referents). Returning to the explanation of this system given by language speakers, k\u00e2-y\u00f4sk\u00e2tahk \u00f4ma n\u00eahiyaw\u00eawin (\u2018Plains Cree is a soft language\u2019), it is hoped that the current description makes a small step towards understanding how a Plains Cree speaker goes about \u201cmaking truth happen\u201d for their listener. The question of why a speaker makes truth happen is not mine to answer, belonging instead to the speakers themselves to ask and to answer. In the following sections, I consider directions for future research in Plains Cree suggested by the current work, and also present approaches to the typology of referential typing in Cree and Athabaskan. 348 6.2. Approaches to \u2013a via the verbal suffix \u2013i: Plurality and obviation When the suffix \u2013a restricts an \u201cAnimate\u201d referent, it is insensitive to the number of the nominal (1), but when this suffix restricts an \u201cInanimate\u201d referent, it can only be interpreted as non- singular (2). (1) \u201cOBVIATIVE\u201d \u2013a CAN BE SINGULAR OR PLURAL a. \u142f\u152d\u1420 \u1401 \u14a5\u152a\u14ef\u1528\u141f \u140a\u14c2\u1426\u1403 \u14c8\u142f\u1418. p\u00eayak \u00ea-miyosiyit anihi n\u00e2p\u00eawa p\u00eayak \u00ea- miyw=si -yi -t an=ihi n\u00e2p\u00eaw-a one c1-good=STAT-DS-3 dst=XT man -XT \u2018ThatINAN one manINAN is goodANIM.\u2019 (Presented S2) b. \u14c3\u14f1 \u1401 \u14a5\u152a\u14ef\u1528\u141f \u140a\u14c2\u1426\u1403 \u14c8\u142f\u1418. n\u00eeso \u00ea-miyosiyit anihi n\u00e2p\u00eawa n\u00eesw \u00ea-miyw=si -yi -t an=ihi n\u00e2p\u00eaw-a two c1-good=STAT-DS-3 dst=XT man -XT \u2018ThoseOBV two menOBV are goodOBV.\u2019 (Presented S2) (2) \u201c\u201cINANIMATE\u201d\u201d \u2013a CAN ONLY BE PLURAL a. \u273d \u142f\u152d\u1420 \u1401 \u14a5\u1429\u141a\u14ef\u146d \u140a\u14c2\u1426\u1403 \u14aa\u1422\u146d\u14ef\u14c7. \u273d p\u00eayak \u00ea-miyw\u00e2siki anihi maskisina p\u00eayak \u00ea-miyw=\u00e2si -k-i an =ihi maskisin-a one c1-good=STAT-0-PL DST=XT shoe -XT Intended: \u2018ThatIN one shoeIN is niceIN.\u2019 (Presented S2) b. \u14c3\u14f1 \u1401 \u14a5\u1429\u141a\u14ef\u146d \u140a\u14c2\u1426\u1403 \u14aa\u1422\u146d\u14ef\u14c7. n\u00eeso \u00ea-miyw\u00e2siki anihi maskisina n\u00eeso \u00ea-miyw=\u00e2si -k-i an =ihi maskisin-a two c1-good=STAT-0-PL DST=XT shoe -XT \u2018ThoseIN two shoesIN are niceIN.\u2019 (Presented S2) If \u2013a is only one suffix, how does it get these two seemingly different readings? An approach to answering this question is suggested by surveying the forms of \u201cInanimate\u201d plural marking in the verb system. The verbal system uses a morpheme both to mark \u201cInanimate\u201d plurality and subjunctive contexts. 349 (3) a. \u201cINANIMATE\u201d NON-SINGULAR \u1401 \u14a5\u1429\u141a\u14ef\u146d \u00ea-miyw\u00e2siki \u00ea-miyw=\u00e2si -k-i c1-good=STAT-0-PL \u2018They are niceIN.\u2019 (Presented S2) b. HABITUAL PAST1 \u1401\u146f\u14ef \u14ab\u14c7, \u146d\u1528\u146d\u14ed\u1439\u152e\u146d \u1401\u147f \u14ab\u14c7, \u1432\u1528\u1422 \u140b\u1455 \u140f\u152d \u1432\u1528\u1422 \u146e \u141a\u1438\u1423, \u2026 \u00eakosi m\u00e2na, kiy\u00eekis\u00eap\u00e2y\u00e2ki \u00eakwa m\u00e2na, p\u00eeyis \u00e2ta wiya m\u00e2na k\u00ee-w\u00e2pan, \u2026 \u00eakosi m\u00e2na, iy- k\u00eekis\u00eap\u00e2y\u00e2-k-i \u00eakwa m\u00e2na, p\u00eeyis \u00e2ta w-iya so usual IC-morning -0-PL then usual final although 3-body m\u00e2na k\u00ee-w\u00e2pan usual PREV-be.dawn \u2018And so in the morning, finally, dawn would break, \u2026\u2019 (Ahenakew 2000: \u00a71.9) c. IRREALIS \u14a5\u1429\u141a\u14ef\u146d miyw\u00e2siki miyw=\u00e2si -k-i good=STAT-0-PL \u2018If it\/they are nice.\u2019 (Presented S2) Here, the same suffix -i is affixed in the same place to the same person marking. In the first example (3a), it occurs in a changed conjunct clause (signalled by the prefix \u00ea-, see Wolfart 1973, Cook 2007), and means that there is a non-singularity of referents. In the second case, this marker co-occurs with a different kind of changed-conjunct clause (signalled by the infixation of -iy- in the stem, see Wolfart 1973, Cook 2007), and denotes a habitual set of events (Cook p.c.). In the third case (3c), it occurs with a simple conjunct clause (signalled by the absence of clause- typing prefixes, see Wolfart 1973, Cook 2007), and means that the clause is irrealis (i.e. that it denotes an event that has not happened). CLAUSE TYPE MEANING Changed \u00ea- Non-Singular Referent Changed -iy- Habitual Realis Event Simple Irrealis Event Table 6.1. A Summary of the occurrences of -i 1 The use of text data here is because I have been unable to elicit internal change from any consultants. Speakers have control of it in natural speech, but it disappears in eliciation (cf. Blain 1997). 350 That is, the verbal system uses a syntactically and phonologically identical morpheme to mark non-singular referents, habitual realis events, and irrealis events. 6.2.2. The \u201cplural\u201d reading In the case of non-singular reference, it is easy enough to see that the morpheme -i is responsible for the addition of a plural-type interpretation. Consider the pair in (4). (4) a. NO \u2013i = \u201cSINGULAR\u201d \u1401 \u14a5\u1429\u141a\u14ef\u1420 \u00ea-miyw\u00e2sik \u00ea-miyw=\u00e2si -k c1-good=STAT-0 \u2018ItIN is nice.\u2019 (Presented S2) b. \u2013i = \u201cPLURAL\u201d \u1401 \u14a5\u1429\u141a\u14ef\u146d \u00ea-miyw\u00e2siki \u00ea-miyw=\u00e2si -k-i c1-good=STAT-0-PL \u2018TheyIN are nice.\u2019 (Presented S2) Here, the first example lacks -i, and receives an interepretation of a singular referent. The second example has changed only by adding -i, and now has an interpetation of relating to more than one referent. 6.2.2. The habitual reading With the internal-change form of clauses, the use of the subjunctive appears to denote past habitual contexts. For example, in the example in (3b), both the clause itself and all the surrounding clauses are marked with m\u00e2na, which denotes a habitual event (cf. 5). (5) m\u00e2na = \u201cHABITUAL\u201d \u14c2\u1472\u14a7\u1424 \u14ab\u14c7. nikamow m\u00e2na. nikamo-w m\u00e2na sing -3 usual \u2018S\/he usually sings.\u2019 (Presented to S2) 351 The context for (3b) is a story about Alice Ahenakew\u2019s childhood. In this story, which runs for a page of text, this particle m\u00e2na occurs on almost every clause (27 out of 36). To further cement the habitual properties of this clause, a speaker that has overt control of internal change (S4) explicitly responded that it gives a habitual interpretation to the clause it is applied to. This comment was gotten by finding an example of internal change in a text and asking the speaker to comment on it. (6) \u1401\u146f\u1455 \u1406\u14aa \u146d\u1526 \u1405\u1426\u148b\u1426\u1403\u146f\u152d\u1541 \u140a\u1422\u146d\u1429. \u00eakota \u00f4ma kiy\u00ea-ohcihikoyahk askiy-~ askiy. \u00eakota aw =ima iy-ka- ohcih-ikw-yahk askiy there PRX=in.sg IC-FUT-fight-INV -21pl land \u2018It will be then that theyAN fight us over the land.\u2019 (K\u00e2-pimw\u00eaw\u00eahahk 1998:\u00a77.7) Here, the speaker (S4) said that the sentence meant that the referents will come to fight over and over again. Importantly, this example lacks the suffix -i, and denotes a future event. Thus, with changed conjunct clauses, it seems that -i is only used in past realis contexts. A habitual event is easily defined as an event that has happened more than once. That is, a habitual event is a non-singularity of events. When Alice Ahenakew uses the habitual construction in (3b), then, she is referring to a non-singularity of past events; it was morning more than once during the events she was discussing. Thus, the morpheme -i can here be said to denote a non-singularity of realis events, rathern than a non-singularity of referents. 6.2.3. The irrealis reading The third context of use for the suffix -i is in simple conjunct clauses. Simple conjunct always has an irrealis interpretation (Wolfart 1973, Cook 2007). The simple conjunct itself is exceedingly rare in connected speech and unattainable in elicitation (Wolfart 1973, Cook 2007). Most often, it occurs with the addition of the suffix -i. This suffix can be affixed to any kind of person agreement, and thus is clearly not marking some kind of \u201cINANIMATE\u201d referent. (7) \u2013i IS INSENSITIVE TO ANIMACY a. \u14c2\u1472\u14a7\u152d\u1426\u146d\u2026 nikamoy\u00e2hki\u2026 nikamo-y\u00e2hk-i sing -1PL-SBJ \u2018If\/when we sing\u2026\u2019 (Presented to S2) 352 b. \u14c2\u1472\u14a7\u148b\u2026 nikamoci\u2026 nikamo-t-i sing -3-SBJ \u2018If\/when s\/heAN sings\u2026\u2019 (Presented to S2) c. \u146d\u14a5\u141a\u1426\u146e\u2026 kimiwahki kimiwan-k-i rain -0-SBJ \u2018If\/when itIN rains\u2026\u2019 (Presented to S2) When -i affixes to this kind of clause, it consistently induces an irrealis interpretation. For example, this kind of clause is required with dependent temporal clauses modifying future events (8). (8) DEPENDENT TEMPORAL CLAUSE IN FUTURE CONTEXT REQUIRES ka\u2013 \u14c2\u1472 \u1403\u1450\u1426\u1456\u1423 \u140a\u1455\u140d\u140f\u1472\u14a5\u146f\u1541 \u14c3\u14f1 \u146e\u14ef\u1473\u146d. nika-toht\u00e2n ataw\u00eawikamikohk n\u00eeso-k\u00eesik\u00e2ki. ni-ka- itoht\u00e2n ataw\u00ea=w=kamikw-ihk n\u00eesw-k\u00eesik\u00e2-k-i 1- FUT-itoht\u00e2-n buy =3=place -loc two- be.day-0-SBJ \u2018I will go to the store on Tuesday.\u2019 (Presented to S2) The event denoted by the verb has not happened, but may happen in the future, under certain conditions. In (9), the Speaker is saying that, should Tuesday come, they will go to the store. The Speaker\u2019s going is dependent on Tuesday\u2019s arrival. For all possible contexts in which Tuesday arrives, the Speaker will go to the store. Thus, the addition of -i always induces a conditional sense; \u201cIf or when Tuesday happens, I will to the store.\u201d This kind of irrealis, then, has to do with a multiplicity of potential contexts. There are many routes the world could take before Tueday gets here, and the Speaker does not know which one will happen. It could snow all week, or there could be a war, or the Speaker could break her leg painting the house. It is even possible that next Tuesday never comes at all \u2013 the world could end before then, or the Speaker could die. Thus, this kind of irrealis has to do with a non-singular event. There are, in a very fundamental way, many Tuesdays that could happen. 6.2.4. Conclusion Thus we see that Plains Cree exploits a single morpheme to code both non-singularities of referents and non-singularities of events in its verbal domain. This correlation between referential non-singularity and eventive non-singularity should not be surprising to any linguist 353 that looks at irrealis constructions in natural language (i.e. averidicality; Farkas (1992), Giannakidou (1999)). For example, Kratzer (1981) models parts of the modality system of German and English by building on the \u201cpossible worlds\u201d analysis of the subjunctive to refer to pluralities of possible events. Based on entirely independent criteria in unrelated languages, Emmon Bach (1986) has concluded that it is necessary to bring the descriptive and theoretical tools of nominal plurality into the discussion of verbal event semantics. Returning to our original concern, the picture developed for the obviative\/\u201dInanimate\u201d parallel no longer looks so exotic. If we conceive of habitual and irrealis contexts as ones that introduce eventive plurality, then the recruitment of plural coding to express this is expected. In the nominal domain, the obviative form codes a referent whose intention is unknown (i.e. averidical; Giannakidou 1998), which suggests an analysis in which the suffix \u2013a codes either a plurality of referents or a plurality of perspectives. Form Plurality \u201cInanimate\u201d Nominal Referent \u201cANIMATE\u201d Nominal Perspective \u00ca- or K\u00c2-CLAUSE Referent IC-CLAUSE Realis Event BARE CLAUSE Irrealis Event Table 6.2. A summary of contexts for plurality In the cases of the subjunctive and obviative, the non-singular quantification is being used to denote Speaker non-knowledge (averidicality). That is, the addition of -i to irrealis clauses denotes a multiplicity of potential events, while the addition of -a to nominals denotes a multiplicity of potential perspectives. 6.3. \u2013yi\u2013 across Cree Plains Cree is the westernmost member of a group of languages or dialects that form an unbroken chain across Canada. While all of these languages possess forms that are straightforwardly related, the organization of these forms varies significantly. All of these languages possess forms related to those considered in this thesis. However, these formal parallels should not be taken to indicate functional parallels. As Cook (2004) has shown for Athabaskan, several languages may possess forms that are obviously related without those forms sharing much in functional properties. This means that it is not automatic that the properties of the Plains Cree system would be mirrored in every other Cree language. In this section, I begin 354 to sketch out the parallels and differences for one morpheme: the disjoint subject marker \u2013yi\u2013. I focus on three other Cree languages: (i) Moose Cree (around James Bay), (ii) Montagnais (Qu\u00e9bec), and (iii) Innu-Aimun (Labrador). Together with Plains Cree, these make up a sample of the Cree spoken in all of the four quadrants of its range. 6.3.1. All Cree languages have a reflex of \u2013yi\u2013 Plains Cree is categorized as one of the \u2018y\u2019 languages in the Cree group; this is because systematically has [y] where other Cree languages have other segments. Relevant to the current discussion, the morpheme \u2013yi\u2013 translates into other Cree languages as \u2013li\u2013, or \u2013ni\u2013 depending on the language. It appears that all Cree languages have a form of the Plains Cree suffix \u2013yi\u2013. 6.3.2. \u2013yi\u2013 occurs with \u201cobviative\u201d possessors This suffix occurs with obviative possessors an all four languages (1-4). (9) PLAINS CREE \u1405\u144c\u14a5\u1528\u1418 ot\u00eamiyiwa o-t\u00eam -yi-w-a 3-horse-DS-3-XT \u2018his\/herOBV horseOBV.\u2019 (Presented S2) (10) MOOSE CREE \u14c8\u14ef\u14a1 \u14c7\u1426\u1401\u14d5\u14a5\u146f\u14d5\u1417 \u1405\u1456\u14f4 \u1405\u1490\u1419\u1511\u14a5\u1511\u14d6\u1424. n\u00e2sic nah\u00ealimikoliwa ot\u00e2nisa ocaw\u00e2\u0161imi\u0161il\u00eew. n\u00e2sic nah=\u00ealim -iko-li -w-a o-t\u00e2nis -a o-caw\u00e2\u0161imi\u0161-li-w-a very fond=by.mind.TA-INV-DS-3-XT 3-daughter-XT 3-child -DS-3-XT \u2018HerPROX daughter\u2019sOBV children are very fond of herOBV.\u2019 (Ellis 2000:332) (11) MONTAGNAIS Pon \u028ak\u028as\u026anwa pon \u028a-k\u028as-\u026an-w-a Paul 3-son-DS-3-XT \u2018Paul\u2019sOBV sonOBV\u2019 (Martin 1991:80) 355 (12) INNU-AIMUN Tshi-tshissenim\u00e2u\u00e2 M\u00e2n\u00ee tsheku\u00e2n kuet aimi\u00e2t P\u00fbna utshim\u00e2minua? Tshi-tshiss=enim -\u00e2-u-\u00e2 M\u00e2n\u00ee tsheku\u00e2n kuet aimi-\u00e2-t 2- know=by.mind.TA-DIR-3-XT Mary what try call -DIR-3 P\u00fbn-a u-tshim\u00e2=m-in-u-a Paul-xt 3-chief=DSJ-DS-3-XT \u2018Do you know why MariePROX called Paul\u2019sOBV bossOBV?\u2019 (Branigan & MacKenzie 2002) 6.3.3. \u2013yi\u2013 occurs with intransitive subjects The suffix also occurs with the subject of intransitive verbs in all four languages. (13) PLAINS CREE \u14c2\u1472\u14a7\u1528\u1418. nikamoyiwa. nikamo-yi-w-a sing -DS-3-XT \u2018s\/heOBV sings.\u2019 (Presented S2) (14) MOOSE CREE \u1490\u1472\u1419\u1511\u1511\u14d5\u1417 cakaw\u00e2\u0161i\u0161iliwa cakaw\u00e2\u0161i\u0161-li -w-a be.few -DS-3-XT \u2018TheyOBV are few.\u2019 (Ellis 2002:328) (15) MONTAGNAIS ap\u04d9t\u04d9h\u026anwa ap\u04d9t\u04d9h-\u026an -w-a useful -DS-3-XT \u2018S\/heOBV is useful.\u2019 (Martin 2000:88) (16) INNU-AIMUN \u2026ekue kutapan\u00eeunit\u00ee. \u2026ekue kutapan\u00eeu -ni -t-\u00ee. \u2026and go.underwater-DS-3-SBJ \u2018\u2026and itOBV went underwater.\u2019 (Hasler 2000:99) 6.3.4. \u2013yi\u2013 occurs with transitive subjects Finally, the suffix occurs with the subject of transitive verbs in all four languages. 356 (17) PLAINS CREE a. \u2026\u141a\u1438\u14ab\u1528\u141f. \u2026\u00ea-w\u00e2pam\u00e2yit. \u00ea-w\u00e2p=am -\u00e2 -yi -t c1-see=by.eye.TA-DIR-DS-3 \u2018\u2026s\/hePROX sees him\/herOBV.\u2019 (Presented S1) b. \u2026\u141a\u1438\u1426\u1455\u14a5\u1528\u141f. \u00ea-w\u00e2pahtamiyit \u00ea-w\u00e2p=aht -am-yi -t c1-see=by.eye.TI-TI -DS-3 \u2018\u2026s\/heOBV sees itIN.\u2019 (Presented S1) (18) MOOSE CREE a. \u14c7\u1426\u1401\u14d5\u14a5\u146f\u14d5\u1417 nah\u00ealimikoliwa nah =\u00ealim -iko-li -w-a fond=by.mind.TA-INV-DS-3-XT \u2018s\/heOBV children is very fond of him\/herOBV.\u2019 (Ellis 2000:332) b. \u14c7\u1455\u140c\u14d5\u1426\u1455\u14a5\u14d5\u1417 nataw\u00ealihtamiliwa nataw=\u00ealiht -am-li -w-a want =by.mind.TI-TI-DS-3-XT \u2018s\/heOBV wants itIN.\u2019 (Ellis 2000:332) (19) MONTAGNAIS a. wapam\u026akonwa wap=am -\u026akw-n -w-a see=by.eye.TA-INV-DS-3-XT \u2018s\/he is seen by him\/her.\u2019 (Martin 2000:95) b. wapat\u04d9m\u026anwa wapat-\u04d9m-\u026an-w-a see -TI -DS-3-XT \u2018S\/heOBV sees itIN.\u2019 (Martin 1991:79) (20) INNU-AIMUN a. N\u00e2n\u00e2tu\u00e2kamenua mishtikua. n\u00e2n\u00e2tu\u00e2kam -e -ni-u-a mishtikw -a break.in.two.dup -DIR-DS-3-XT tree -XT \u2018TheyOBV were chewing down treesOBV.\u2019 (Hasler 2002:133) 357 b. \u2026tu\u00e2shkunamuenishapan\u00ee. tu\u00e2shkun -am-u-eni-shapan\u00ee knock.down.sticks -TI -3-DS -EVID \u2026 heOBV must have taken away too many sticksIN. (Hasler 2002:140) 6.3.5. \u2013yi\u2013 is insensitive to animacy It is insensitive to animacy in all the languages, occurring equally well with both animate and inanimate arguments. (21) PLAINS CREE \u2026\u1401\u146e\u14ef\u1473\u1528\u1420. \u2026\u00ea-k\u00eesik\u00e2yik. \u00ea-k\u00ee\u0161ik\u00e2-yi-k c1-day -DS-0 \u2018\u2026one dayIN.\u2019 (Ellis 2000:328) (22) MOOSE CREE \u2026\u1401 \u146e\u14ef\u1473\u14d5\u1420. \u2026 \u00ea-k\u00ee\u0161ik\u00e2lik. \u00ea-k\u00ee\u0161ik\u00e2-li-k c1-day -DS-0 \u2018\u2026one dayIN.\u2019 (Ellis 2000:328) (23) MONTAGNAIS ap\u04d9t\u04d9h\u026anno ap\u04d9t\u04d9h-n-\u026an-w useful -0-DS-3 \u2018ItIN is useful.\u2019 (Martin 2000:88) (24) INNU-AIMUN Anite tshim\u00e2ten\u00fb m\u00eetshu\u00e2pin\u00fb pess\u00eesh, \u2026 anite tshim\u00e2te -in-\u00fb m\u00eetshu\u00e2p-in\u00fb pess\u00eesh there stand -DS-3 house -DS close.by \u2018There stood a houseIN close by, \u2026\u2019 (Hasler 2002:175) Thus far, the systems appear to all be running in close parallel. The morpheme has the same distribution, marks the same argument, and occurs in the same position across all four languages. However, as we saw in Chapter 3 (\u00a73.3.5), some Cree languages use the suffix \u2013yi\u2013 on referentially-dependent inanimates (cf. \u00a73.3.5, Cowan 1985). In fact, it appears that all the systems do this but Plains Cree. 358 (25) PLAINS CREE * \u148c\u14ab\u14c2\u1528\u1424 * c\u00eem\u00e2niyiw c\u00eem\u00e2n-yi -w canoe-DS-3 \u2018---\u2019 (Presented S3) (26) MOOSE CREE \u148c\u14ab\u14c2\u14d5\u1424 c\u00eem\u00e2niliw c\u00eem\u00e2n-li -w canoe-DS-3 \u2018canoeIN\u2019 (Ellis 2000) (27) MONTAGNAIS akop\u026ano akop -\u026an -o blanket-DS-3 \u2018blanketIN\u2019 (Martin 2000:49) (28) INNU-AIMUN N-u\u00eetshi\u00e2u Pien uieuesht\u00e2t ishk\u00eet\u00fbnnu \u2026 N-u\u00eetshi-\u00e2-u Pien uieuesht\u00e2-t ishk\u00eet\u00fb -n -u 1-help -DIR-3 Peter fix -3 Ski-Doo-DS-3 \u2026 \u2018I helped Peter fix the Ski-DooIN \u2026\u2019 (Branigan & MacKenzie 2002:392) The use of \u2013yi\u2013 on non-possessed nominals in these other Cree languages could suggest that these languages use it to code disjunction from other Referring-expressions (i.e. nominals) in the clause, rather the argument of some preceding predicate (see Saxon 1986 for the Athabaskan language Dogrib, \u00a76.3 below). If this is the case, it would mean that the disjunction properties of \u2013yi\u2013 in these languages are A (argument) disjunction rather than A' (adjunct). This means that these other Cree languages would show behaviour more like Navajo and other Athabaskan languages (see. \u00a76.3 below). Related to this difference in distribution of the affix is a difference in the distribution of the clause types. In Plains Cree, the independent order of clausal morphology is highly restricted (cf. Wolfart 1973, Cook 2007, 2008), occurring only rarely in developed discourse. What is striking, from a Plains Cree perspective, when looking at some of the other Cree languages is their heavy use of independent-order clauses. Consider the following passage from Swampy Cree. 359 (29) i. kwayask \u00eatikw\u00ea awa k\u00ee-ayiw k\u00ee-n\u00e2p\u00eawiw w\u00eesahk\u00eac\u00e2hk, \u2026 kwayask \u00eatikw\u00ea aw=a k\u00ee- ayi-w k\u00ee- n\u00e2p\u00eaw=i -w w\u00eesahk\u00eac\u00e2hk proper evid prx=an.sg PREV-be -3 PREV-man =AI-3 W. \u2018He was then a good man, W\u00ees\u00e2hk\u00eac\u00e2hk was, \u2026\u2019 ii. \u00eakw\u00e2ni \u00eatikw\u00ea m\u00eena k\u00ee-ati-ay-ispaniniw ta-ay\u00e2t ta-w\u00eekihtot. \u00eakw\u00e2ni \u00eatikw\u00ea m\u00eena k\u00ee- ati- ay- ispani -ni -w ta- ay\u00e2 -t and EVID also PREV-along-RED-happen-DS-3 FUT-have-3 ta-w\u00eek=ih=to-t FUT-live=by.neut=RECIP-3 \u2018\u2026and it seems it came time for him to be, for him to marry.\u2019 (Sanderson 2007) Here, the Speaker has strung together multiple independent-order verbs in a row. Further, the second clause contains an independent-order verb marked with \u2013yi\u2013 (\u2013ni\u2013), a situation that is highly uncommon in Plains Cree. In comparison to Plains Cree, then, some Cree languages appear to make much more use of independent order clauses in general and independent order clauses marked with \u2013yi\u2013 in specific. The differences in clausal architecture, then, could be forcing differences in the domain of dependency for \u2013yi\u2013. In systems where the independent order is highly specialized to the text-level of discourse (i.e. Plains Cree; Cook 2008), \u2013yi\u2013 will either be highly restricted (if it codes a cross-predicate dependency: Plains Cree) or be allowed to code a clause-internal dependency (perhaps Innu-Aimun?). If the independent order is less specialized, and occurs in broader discourse contexts than just text-level (perhaps Swampy Cree?), then the suffix \u2013yi\u2013 can be manipulated to either code cross-predicate (A') or clause- internal (A) dependencies. Further research is needed to see if these speculations are borne out. 6.4. A comparison of the Plains Cree system to Athabaskan Athabaskan languages possess a set of prefixes that distinguish between two kinds of third persons (cf. Saxon 1986, Horseherder 1998, Cook 2004). Navajo Dogrib D\u00ebne S\u0173\u0142in\u00e9 1 shi- se- se- 2 ni- ne- ne- 3\u2019 bi- we- be- 3\u2019\u2019 yi- ye- ye- Table 6.3. Navajo, Dogrib and D\u00ebne S\u0173\u0142in\u00e9 pronominal affixes 360 When we compare the functions of these prefixes to the Plains Cree system described in this thesis, some striking resemblances emerge. For example, the alternation between yi- and bi- correlates with a change in the argument structure of the verb (cf. Hale 1973), as in (30), and sometimes yields passive English translations (cf. Reichard 1951). (30) \u201cPASSIVE\u201d TRANSLATION OF BI- IN NAVAJO a. 'ashkii at\u2019\u00e9\u00e9d yiyii\u0142ts\u00e12 boy girl 3-3.Pf.see \u2018The boy saw the girl.\u2019 b. at\u2019\u00e9\u00e9d 'ashkii bii\u0142ts\u00e1 girl boy 3-3.Pf.see \u2018The girl was seen by the boy.\u2019 (Horsherder 1989:12) This bears a strong functional resemblance to the Plains Cree inverse suffix \u2013ikw (\u00a74.4.22), as the minimal pair he suggests. (31) \u201cPASSIVE\u201d TRANSLATION OF \u2013ikw IN PLAINS CREE a. \u2026\u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u14ef\u1422 \u1403\u1422\u1475\u1418. \u2026\u00ea-w\u00e2pam\u00e2t n\u00e2p\u00easis iskw\u00easisa. \u00ea- w\u00e2p=am -\u00e2 -t n\u00e2p\u00ea=sis iskw\u00ea =sis -a C1-see =by.eye.TA-DIR-3 man=DIM woman=DIM-XT \u2018\u2026the boyPROX saw the girlOBV.\u2019 (Presented S2) b. \u2026\u1401\u141a\u1438\u14a5\u146f\u141f \u14c8\u142f\u14ef\u1422 \u1403\u1422\u1475\u1418. \u2026\u00ea-w\u00e2pamikot n\u00e2p\u00easis iskw\u00easisa. \u00ea- w\u00e2p=am -ikw -t n\u00e2p\u00ea=sis iskw\u00ea =sis -a C1-see =by.eye.TA-DIR-3 man=DIM woman=DIM-XT \u2018\u2026the boyPROX was seen by the girlOBV.\u2019 (Presented S2) Further, the prefix yi- occurs when the predicate\u2019s argument is distinct from some other third person (cf. Saxon 1986). (32) DISJOINT REFERENCE OF ye- IN DOGRIB Patrick yem\u01eb ts'\u00e0htla.3 DA.mother 3.PF.visit \u2018Patrick visited her (someone else\u2019s) mother.\u2019 (Saxon 1986:103) 2 Glossing is that of Horseherder (1998). \u2018Pf\u2019 = \u2018Perfect\u2019 3 The glosses are Saxon\u2019s (1986). \u2018DA\u2019 stands for \u2018disjoint anaphor.\u2019 \u2018PF\u2019 stands for \u2018perfective.\u2019 361 This bears a strong functional resemblance to the Plains Cree suffix \u2013yi\u2013 (\u00a73.3.4, \u00a74.3.2). These resemblances mean that a consideration of the parallels between the systems would be useful. When we compare these systems in detail, a number of generalizations emerge. We see the following similarites: (i) Both Athabaskan yi\u2013 and Plains Cree \u2013yi\u2013 code disjunction in the third-person system. (ii) Both Athabaskan yi- and Plains Cree \u2013ikw are used in transitive contexts. (iii) Both Athabaskan yi- and Plains Cree \u2013yi\u2013 are used in the third person system. (iv) All forms are sensitive to animacy distinctions. We see the following differences: (i) The Athabaskan system is pronominal but the Plains Cree system is not. (ii) Plains Cree \u2013ikw does not code disjunction but Athabaskan yi- does. (iii) Plains Cree \u2013yi\u2013 is insensitive to transitivity but Athabaskan yi- is. (iv) Ordering relations in Plains Cree are between predicates but Athabaskan appears to be sensitive to argument ordering. (v) Plains Cree is strictly sensitive to A\u2019 (adjunct) dependencies while Athabascan is either mixed or A (argument) oriented. (vi) Plains Cree \u2013ikw is insensitive to person while Athabascan yi- is. These generalizations are summarized in the table below. \u00a7 PC \u2013yi\u2013 PC \u2013ikw\u2013 NV yi\u2013 DR ye\u2013 C1 Pronominal? No No Yes Yes C2 Disjunction Yes No Yes Yes C3 Transitivity 1-ARG 2-ARG 2-ARG 2-ARG C4 Dependency A\u2019 A\u2019 A & A\u2019 A Table 6.4. Comparing Plains Cree \u2013ikw and \u2013yi\u2013 to Athabascan yi- From this comparison, we see that the two systems distribute the work differently, but appear to reach similar results. Both systems appear to build structures in which third persons are disambiguated via disjunction and subordination, but the means to these ends are different. This use of third person dependencies and subordination means that, if the picture of referential dependency developed in this thesis holds across languages, the reference system of Athabaskan should be tracking ex\/intentionality. Based on the discussion of Horsherder (1998) for Navajo, there is suggestive evidence that this is in fact the case; Horseherder (1998:18) argues that one of the primary semantic properties of the yi-\/bi- alternation is agentivity or control. In considering this alternation in Navajo, Horseherder argues that the choice has to do 362 with whether or not the subject is conceived of as agentive. Something that is conceived of as agentive is something that is capable of having intentions \u2013 the referent must be capable of having mental states in order to act towards achieving some end. However, this association of yi- \/bi- with agenttivity leads to some familiar paradoxes \u2013 lightning and rain are agentive, because they have movement, Horseherder concludes. If this argument is turned on its head, in parallel to the animate argument made in Chapter 2 (\u00a72.1), then the yi-\/bi- alternation could be understood as coding non-agentivity. 6.4.1. Athabaskan yi-\/bi- are pronominals, Plains Cree\u2019s are not The set of prefix alternations in Athabaskan languages are typically described as pronominal in nature (cf. Saxon 1986, Sandoval & Jelinek 1989, Horseherder 1998, Cook 2004, etc.). For example, consider the forms in (33) from Navajo. (33) NAVAJO yi- AND bi- a. 'ashkii at\u2019\u00e9\u00e9d yiyii\u0142ts\u00e1 boy girl 3-3.Pf.see \u2018The boy saw the girl.\u2019 b. at\u2019\u00e9\u00e9d 'ashkii bii\u0142ts\u00e1 girl boy 3-3.Pf.see \u2018The girl saw the boy.\u2019 (Horsherder 1989:12) Here, the prefixes bi- and yi- occupy the pronominal position on the verb. By contrast, the relevant components of Plains Cree\u2019s grammar are clearly not pronominal in nature. Both \u2013yi- and \u2013ikw stack with pronominal prefixes (34). (34) CO-OCCURRENCE OF PC \u2013yi\u2013 AND \u2013ikw WITH PRONOMINALS a. \u2026\u1401 \u14c2\u1472\u14a7\u1528\u141f \u2026\u00ea-nikamoyit \u00ea-nikamo-yi -t c1-sing -DS-3 \u2018\u2026s\/heOBV sings.\u2019 (Presented S2) b. \u2026\u1401 \u14ed\u146d\u1426\u1403\u146f\u141f. \u2026\u00ea-s\u00eakihikot. \u00ea-s\u00eak =ih -ikw-t c1-scare=by.neut-INV-3 \u2018\u2026s\/heOBV scared him\/herPROX.\u2019 (Presented S2) 363 Further, the two forms both are required to co-occur with pronominal marking. (35) PC \u2013yi\u2013 AND \u2013ikw MUST HAVE PRONOMINAL MARKING a. \u273d \u1401 \u14c2\u1472\u14a7\u1528 \u273d \u00ea-nikamoyi \u00ea- nikamo-yi C1-sing -DS \u2018---\u2019 (Presented S2) b. \u273d \u1401 \u14ed\u146d\u1426\u1403\u146f \u273d \u00ea-s\u00eakihiko \u00ea-s\u00eak =ih -ikw c1-scare=by.neut-INV \u2018---\u2019 (Presented S2) Likewise, both forms are not sufficiently specified for person distinctions; they both can occur with either animate or inanimate referents. (36) \u2013yi\u2013 IS INSENSITIVE TO ANIMACY a. \u2026\u1401 \u146d\u14c4\u14ef\u1528\u141f. \u2026\u00ea-kinosiyit. \u00ea- kinw=si -yi -t C1-long=STAT-DS-3 \u2018\u2026s\/heOBV is tall.\u2019 (Presented S2) b. \u2026\u1401 \u146d\u14ce\u1528\u1420. \u2026\u00ea-kinw\u00e2yik. \u00ea- kinw=\u00e2 -yi -k C1-long=STAT-DS-0 \u2018\u2026itIN is tall.\u2019 (Presented S2) (37) \u2013ikw IS INSENSITIVE TO ANIMACY a. \u2026\u14ed\u146d\u1426\u1403\u1420. \u2026s\u00eakihik. s\u00eak =ih -ikw scare=by.neut-INV \u2018s\/hePROX was scared by him\/herOBV.\u2019 (Presented S2) b. \u2026\u14ed\u146d\u1426\u1403\u146f\u1424. \u2026s\u00eakihikow. s\u00eak =ih -ikw-w scare=by.neut-INV-0 \u2018s\/hePROX was scared by itIN.\u2019 (Presented S2) 364 While the suffix \u2013yi\u2013 is restricted to third persons, the inverse \u2013ikw is not. (38) \u2013ikw OCCURS WITH SPEECH ACT PARTICIPANTS \u14c2\u141a\u1438\u14a5\u1420 niw\u00e2pamik ni-w\u00e2p=am -ikw 1- see =by.eye.TA-INV \u2018s\/he saw me.\u2019 (Presented S2) From this data, it is necessary to conclude that the Plains Cree system uses non-pronominal structures to code these referential distinctions. Instead, the affixes used appear to be argument- structure related \u2013 a generalization that makes sense when we consider the argument-structure independent nature of most pronominal forms in Plains Cree (cf. D\u00e9chaine & Reinholtz 2008, etc.). 6.4.2. Plains Cree \u2013yi\u2013 and Athabaskan yi- code disjoint reference As I argued in Chapter 3 (\u00a73.3.5), the suffix \u2013yi\u2013 in Plains Cree has already been argued to encode disjunction; the subject of the predicate is disjoint from some previous 3rd person referent. (39) PRESENCE\/ABSENCE OF \u2013yi\u2013 CODES CO\/DISJOINT REFERENCE a. \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u1401 \u14c2\u1472\u14a7\u141f n\u00e2p\u00eaw \u00ea-w\u00e2pam\u00e2t iskw\u00eawa \u00ea-nikamot. n\u00e2p\u00eaw \u00ea- w\u00e2p=am -\u00e2 -t iskw\u00eaw-a \u00ea-nikamo-t man C1-see=by.eye.TA-DIR-3 woman-XT C1-sing -3 \u2018The manPROX saw the womanOBV when hePROX was singing.\u2019 (Presented S2) b. \u1401 \u141a\u1438\u14ab\u141f \u14c8\u142f\u1424 \u1403\u1422\u1475\u1418 \u1401 \u14c2\u1472\u14a7\u1528\u141f n\u00e2p\u00eaw \u00ea-w\u00e2pam\u00e2t iskw\u00eawa \u00ea-nikamoyit. n\u00e2p\u00eaw \u00ea- w\u00e2p=am -\u00e2 -t iskw\u00eaw-a \u00ea-nikamo-yi-t man C1-see=by.eye.TA-DIR-3 woman-XT C1-sing -DS-3 \u2018The manPROX saw the womanOBV when sheOBV was singing.\u2019 (Presented S2) Here, the presence\/absence of \u2013yi\u2013 on the predicate nikamo- \u2018sing\u2019 tells us who is singing; if it is absent, the singer is the same referent that was the previous proximate referent. If it is absent, it is not this referent - it is the woman, or someone else. This also works with possessed nominals, as the pair of examples in (40) demonstrate. 365 (40) PRESENCE\/ABSENCE OF \u2013yi\u2013 SIGNALS CO\/DISJOINT REFERENT ON NOMINALS a. \u14c8\u142f\u1424 \u141a\u1438\u14ab\u1424 \u1403\u1422\u1475\u1418 \u1405\u144c\u14a5\u1528\u1418 n\u00e2p\u00eaw w\u00e2pam\u00eaw iskw\u00eawa ot\u00eamiyiwa n\u00e2p\u00eaw w\u00e2p=am -\u00ea -w iskw\u00eaw-a o-t\u00eam -yi-w-a man see =by.eye.TA-DIR-3 woman-XT 3-horse-DS-3-XT \u2018The manPROX saw the woman\u2019sOBV horseOBV.\u2019 (Presented S2) b. \u14c8\u142f\u1424 \u141a\u1438\u1426\u1455\u1466 \u1403\u1422\u1475\u1418 \u1405\u14aa\u1422\u146d\u14ef\u14c2\u1528\u1424 n\u00e2p\u00eaw w\u00e2pahtam iskw\u00eawa omaskisiniyiw n\u00e2p\u00eaw w\u00e2p=aht -am iskw\u00eaw-a o-maskisin-yi-w man see =by.eye.TI-TI woman-XT 3-shoe -DS-3 \u2018The manPROX saw the woman\u2019sOBV shoeIN.\u2019 (Presented S2) Here, the possessed form that carries \u2013yi\u2013 cannot be interpreted as the horse or shoe of the man \u2013 it must be someone else\u2019s. The Athabascan prefix yi- has also been analyzed as disjoint reference (cf. Saxon 1986). When this prefix is added to a possessed nominal, it codes that the possessor is not the verb\u2019s subject. (41) DOGRIB ye- CODES DISJOINT POSSESSOR Patrick yem\u01eb ts'\u00e0htla.4 DA.mother 3.PF.visit \u2018Patrick visited her (someone else\u2019s) mother.\u2019 (Saxon 1986:103) Likewise, when a third person object is acted on by a third person subject, yi- is used to code that the object is disjoint from the subject. (42) DOGRIB ye- CODES DISJOINT OBJECT Patrick yets'\u00e0htla DA.3.PF.visit \u2018Patrick visited her.\u2019 (Saxon 1986:108) This analysis of disjunction has been extended to many other Athabaskan languages as well (cf. Horsherder 1998 for Navajo, Cook 2004 for D\u00ebne S\u0173\u0142in\u00e9). 4 The glosses are Saxon\u2019s (1986). \u2018DA\u2019 stands for \u2018disjoint anaphor.\u2019 \u2018PF\u2019 stands for \u2018perfective.\u2019 366 6.4.3. Plains Cree \u2013yi\u2013 and Athabascan yi- differ in transitivity The suffix \u2013yi\u2013 in Plains Cree is entirely insensitive to the transitivity of the predicate it attaches to. It can occur equally well with single-argument predicates (43a) or two-argument predicates (43b). (43) \u2013yi\u2013 IS INSENSITIVE TO TRANSITIVITY a. \u2026\u1401 \u14c2\u1472\u14a7\u1528\u141f. \u2026\u00ea-nikamoyit. \u00ea-nikamo-yi -t c1-sing -DS-3 \u2018\u2026s\/heOBV sings.\u2019 (Presented S2) b. \u2026\u141a\u1438\u14ab\u1528\u141f. \u2026\u00ea-w\u00e2pam\u00e2yit. \u00ea-w\u00e2p=am -\u00e2 -yi -t c1-see=by.eye.TA-DIR-DS-3 \u2018\u2026s\/hePROX sees him\/herOBV.\u2019 (Presented S1) c. \u2026\u141a\u1438\u14a5\u146f\u1528\u141f. \u2026\u00ea-w\u00e2pamikoyit. \u00ea-w\u00e2p=am -ikw-yi-t c1-see=by.eye.TA-INV-DS-3 \u2018\u2026s\/heOBV sees him\/herOBV.\u2019 (Presented S1) Since this suffix codes disjoint subject properties (cf. \u00a73.3.5), this insensitivity to transitivity makes sense \u2013 both one and two-argument predicates have a subject. By contrast, the Athabascan prefix yi- codes objects, possessors, and preposition-like elements (cf. Saxon 1986, Horseherder 1998, etc.). Consider the following examples from Navajo. (44) a. POSSESSED NOMINAL yim\u00e1 3.mother his\/her mother\u2019 b. VERB OBJECT \u2026yizta\u0142. 3-3.kicked \u2018\u2026he kicked him.\u2019 367 c. PREPOSITION \u2026yi\u0142\u2026 3.with \u2018\u2026with him\u2026\u2019 (Horseherder 1998) Here, yi- can be associated with an object (15a), a possessor (15b), and preposition (15c), but not with a subject (Saxon 1986, Horseherder 1989, Cook 2004). Considered from a Plains Cree perspective, the morphosyntactic parallel between objects of transitive verbs and possessors in Athabascan is striking. In Plains Cree, there is a systematic correspondence between possession and transitive verbs; possessors behave exactly like verbal subjects. Athabascan, on the other hand, appears to treat objects and possessors in parallel; non- subject possessors are like non-subject objects (cf. Saxon 1986, Horseherder 1998). It is possible that this difference relates to the different positions of possessed nominals within clause structure. In Plains Cree, we have seen that possessed nominals are equivalent to verbs in terms of their ability to trigger dependency-marking, and not like other nominals (e.g. \u2013yi\u2013 and \u2013ikw : \u00a74.3.2, \u00a74.4.22). This suggests that their syntactic relations to the verb are like that of another verb, rather than a nominal. Thus, a chain of a possessed nominal and a verb is more like a serial verb or clause chain construction. If, in Athabaskan, the possessed nominal is dependent on the verb's argument structure (e.g. yi- marks that the possessor is not the subject of the verb), then the nominal is dependent on the verb. Plains Cree, then, can use possessed nominals to drive verbal morphosyntax, whereas Athabascan uses verbs to drive possessed nominal morphosyntax. 6.4.4. A and A' dependencies In Plains Cree, I have argued that the suffix \u2013yi\u2013 looks beyond the clause for their dependency relation. In Chapter 4 (\u00a74.2.2), I argued that the disjoint subject marker \u2013yi- depended for its reference on some preceding predicate. (45) a. PRED1 PRED2 6 6 y x PRED2-yi 368 b. PRED1 PRED2 PRED 3 6 5 6 y x PRED3-yi c. PRED1 PRED2 6 6 y PRED3 6 x PRED3-yi Crucially, -yi- is not used to mark disjunction from other arguments of the same predicate. Thus, forms like (46), where the subject is disjoint from the third person object, do not use \u2013yi\u2013. (46) DISJOINT OBJECT \u2260 \u2013yi\u2013 \u2026\u141a\u1438\u14a5\u146f\u141f. \u2026\u00ea-w\u00e2pamikot. \u00ea- w\u00e2p=am -ikw-t C1-see=by.eye.TA-INV-DS-3 \u2018\u2026s\/hePROX sees him\/herOBV.\u2019 (Presented S1) This means that \u2013yi\u2013 does not code argument (A) disjunction, but rather codes disjunction from some non-argument which is gotten from previous structure (i.e. A'; Ross 1967). Likewise, in Chapter 4 (\u00a74.4.22), I argued that the inverse suffix \u2013ikw depended on the content of the Topic op position \u2013 in the spec of CP. (47) CP 3 op 3 C IP \u00ea- 3 pro 3 -ikw- VP 3 ROOT 3 pro 3 FINAL This is, again, a classic case of an adjunct (non-argument; A') (cf. Ross 1967). Turning to Athabaskan, the A\/A' properties appear to vary, but all systems minimally appear to have A dependencies. Horseherder (1998) considers the A\/A' properties of yi-. She 369 argues that yi- in Navajo has two requirements: (i) it must be disjoint from the subject of the verb, and (ii) it must be co-referent with an A' antecedent. Consider the pair of sentences in (48). (48) a. yi\u2013 IS DISJOINT FROM SUBJECT haayit\u2019eego Frank Joe yika\u2019elwod \u2026 Q.in.what.way Frank Joe 3.for\/after.run \u2018How did Frankf help Joej?\u2019 b. yi\u2013 IS CO-REFERENT WITH PREVIOUS REFERENT \u2026 yi-l\u00ed\u00ed\u2019 yaa yizloh \u2026 yi-horse 3.for 3-3.Pf.rope \u2018\u2026 hef roped hisj horse for himj\u2019 (Horseherder 1998:29) Here, Horseherder argues that the use of yi- in the second sentence is well-formed because the possessor and object of the verb are disjoint from the subject of the verb (an A dependency) and co-referent with a preceding referent (a topic, A' dependency). Horseherder (1998) then suggests that Dogrib lacks this A' dependency, but it appears that the data is not conclusive (cf. Horseherder 1998:52). Summarizing, Plains Cree is coding disjunction from an A' (Adjunct) position, while Athabaskan is coding disjunction from an A (Argument) position, but co-reference with an A' position. This difference likely has widescale repercussions for the organization of the two systems\u2019 syntax 370 Glossary ANIMATE (AN): One of the two grammatical classes of nominals in Algonquian languages. Shown by the form of demonstratives, plural suffixes, and verbal agreement. In this thesis, it is argued that nominals treated in this way refer to discourse referents that have no specified inherent or contextual intentional properties. CHANGE (C1, C2, IC): A process that affects the phonological shape of an Algonquian verb structure, signalling one of several kinds of clausal dependencies. This change consists most often in vowel alterations (similar in process to germanic Ablaut). In Plains Cree, change is shown in two places: PREVERBS (where the first vowel of the preverb is typically altered, e.g. k\u00ee- ~ k\u00e2-, w\u00ee- ~ w\u00e2-, or, if there is no preverb, the change occurs by itself, as a single vowel; \u00ea-), or the verb stem (where change is shown by the alteration of the first syllable of the verb stem; e.g. w\u00e2pam- ~ wiy\u00e2pam-). Change is always limited to the CONJUNCT form of the verb. A CONJUNCT form lacking change is referred to as SIMPLE CONJUNCT. CONJUNCT (CONJ): One of the two paradigms (ORDERs) of verbal inflection found in most Algonquian languages (except Blackfoot). Characterized by suffixation of person marking. This paradigm is thought to be the older verb form (Goddard 1967). CROSS-REFERENCE: A relation between a noun and a verb signalled by marking on one or both members. Similar to the syntactic notion of \u2018agreement,\u2019 but making no specific structural claims. (see Bloomfield 1962) DIFFERENT SUBJECT (DS): A suffix, taking the shape -yi-, that occurs on predicates when the subject of the predicate is disjoint from some previous subject. This previous subject must either linearly precede or c-command the predicate. Discussed in Chapter 3 and Chapter 4. 371 DIMINUTIVE (DIM) : This is a suffix, -si(s), that occurs on both nouns and verbs and codes a reduction in degree. It can chage the denotation of nominal predicates, as well (e.g. dog \uf0e0 puppy vs. dog \uf0e0 small dog). DISJOINT ARGUMENT (DSJ): A suffix, taking the shape -im-, that affixes to stems. It codes that the argument it is related to is disjoint from some other referent. In the case of transitive verb forms, it codes that the verb\u2019s object is disjoint from another referent. When affixed to transitive predicates, it codes that the predicate\u2019s subject is not the object (i.e. it adds an argument). Discussed in Chapter 3 and Chapter 4. DISTAL (DST) : Using the root an-, this is one of two spatial categories coded by demonstratives. It codes that the referent is far away from the Speaker, but also codes complex discourse properties that are not fully understood. It is often used as a substitute for English definiteness with apparently no spatial deictic meaning. DIRECT (DIR): One of the THEME SIGNs used to signify that the agent-patient relation follows the pattern established in the previous discourse context. The name comes from a notion of DIRECTION. DIRECTION: A reference to one of two models set up to describe the occurrence of different theme signs on Algonquian TRANSITIVE verbs. On the first model, direction is the order that morphemes are to be read on an INDEPENDENT ORDER, TRANSITIVE ANIMATE verb; if the form is \u2018direct,\u2019 the morphemes are to be read from left to right to get the agent-patient relation, while, if the form is \u2018inverse,\u2019 the morphemes are to be read from right to left to get the agent-patient relation. Since then, another idea has developed, wherein the notion of direction refers to a person hierarchy; the \u2018direct\u2019 form follows the hierarchical ordering, while the \u2018inverse\u2019 form goes in the opposite direction. EVIDENTIAL (EVID) : One of the rich set of particles and preverbs in Plains Cree that code the Speaker\u2019s relation to the proposition expressed. The most common ones in this data are \u00easa and \u00eatikw\u00ea, which are thought to mark kinds of indirect knowledge. 372 EXTENTIONAL (XT) : The suffixes \u2013a and \u2013ihi, which are here analyzed as coding that the referent is extentional (i.e. either inanimate or obviative). FINAL: The final element in an Algonquin stem, selected from a closed class. This element gives the stem its categorical properties as well as some of its argument- structure. FUTURE (FUT): A preverb that codes the verb as non-assertional, typically futurate in meaning, but also used in conjunction with other components to code different kinds of Plains Cree\u2019s modality system. INANIMATE (IN): One of the two grammatical classes of nominals in Algonquian languages. Shown by the form of demonstratives, plural suffixes, and verbal agreement. In this thesis, it is argued that referents coded in this way are judged to never have INTENTIONs. INCHOATIVE (INCH) : The verb final \u2013ipayi\u2013. This final occurs in environments that are often inchoative in meaning (e.g. kinosipayiw \u2018s\/he got tall \/ became tall.\u2019), but it also occurs in contexts where the verb\u2019s actor is simply out of control of the event (e.g. s\u00eahk\u00eapayiw \u2018s\/he\/it rides away \/ is conveyed away\u2019). It\u2019s meaning, then, is likely more abstract than \u201cinchoative\u201d and the gloss here should be considered a mere convenience. INDEPENDENT (IND): One of the two paradigms (ORDERs) of verbal inflection found in Plains Cree. Characterized by circumfixal person marking. INITIAL: The first required element in an Algonquian stem. This element carries the idiosyncratic lexical material of the stem. It is roughly equivalent to the more common linguistic term \u2018root.\u2019 373 INITIAL CHANGE (IC): CHANGE that affects the initial syllable of a verb or PREVERB. The \u2018elsewhere\u2019 kind of CHANGE, based on frequency of appearance. Thus, k\u00e2- and \u00ea- are typically glossed as specific instances of CHANGE because they are heavily used, whereas a more general gloss is necessary for the -iy- segment in w\u00eey\u00e2pam\u00e2t-, since it occurs perhaps once in three books\u2019 worth of Plains Cree. In some sense, \u00ea- and k\u00e2- may represent a diachronic change, whereas initial change appears to be synchronic. INVERSE (INV): One of the THEME SIGNs used to signify that the agent-patient relation runs contrary to the pattern established in previous discourse context. The name comes from a notion of DIRECTION. LOCAL PERSON (LP): A grammatical person category in Algonquian languages (cf. Bloomfield (1962), Hockett (1958), etc.). The collection of both the Speaker and the Hearer. Roughly equivalent to Speech Act Participant. MEDIAL: An optional element that occurs between the INITIAL and FINAL elements of an Algonquian stem. MODE: A paradigm that is distinguished from all other paradigms in its ORDER by the addition of some extra element. This extra element is, in the CONJUNCT order, most often a prefix (e.g. \u00ea- vs. k\u00e2-), while it is most often a suffix in the INDEPENDENT order, (e.g. -n vs. -tok\u00ea). NOMINALIZATION (NOM) : A suffix that converts verbs into nominals. The suffix \u2013win, which is itself possibly a set of suffixes, which is used to derive abstract nominals out of verbs (e.g. masinahik\u00ea- \u2018write\u2019 \uf0e0 masinahik\u00eawin \u2018writing\u2019). The suffix \u2013kan, which is also possible compositional, is used to derive a concrete nominal out of a verb, typically the element referent related to the verb or the outcome of the verb\u2019s action (e.g. masinah- \u2018inscribe\/write\u2019 \uf0e0 masinahikan \u2018book \/ paper\u2019). 374 OBVIATIVE (OBV): A referential category in Plains Cree. It refers to a discourse referent that is judged by the Speaker to not have any intentions in a given context. ORDER: The set of all paradigms that share the same personal inflections. Plains Cree has two such verbal orders; the INDEPENDENT Order, and the CONJUNCT Order. A sample paradigm, for intransitive forms, is shown here. SINGULAR PLURAL Intransitive Inflection CONJ IND CONJ IND 1 STEM-y\u00e2n ni-STEM-n STEM-y\u00e2hk ni-STEM-n\u00e2n 2\/1 - - STEM-yahk ki-STEM-naw 2 STEM-yan ki-STEM-n STEM-y\u00eak ki-STEM-n\u00e2w\u00e2w ANIMATE STEM-t STEM-w STEM-cik STEM-wak OBVIATIVE STEM-yit STEM-yiwa STEM-yit STEM-yiwa INANIMATE STEM-k STEM-w STEM-ki STEM-wak G1. Summary of verbal morphology PLURAL (PL): A morpheme that denotes non-singularity. Examples are nominal \u2013ak (animate plural, nouns & independent order), \u2013iki (animate plural, demonstratives), and \u2013ik (animate plural, conjunct order). PREVERB: A dependent, closed class of elements that affix loosely (open Sandhi) to the left of the verb stem. Preverbs may occur alone or be stacked, but have a definite order when stacked. Most of these elements mark some kind of temporal, aspectual, or modal property. Note that other elements may also occur in the preverb domain, either as-is or with the addition of the preverb marker -i. PREVIOUS EVENT PREVERB (PREV): A preverb, taking the shape k\u00ee-, that is used to code that the verb\u2019s event is previous to some other discourse-specified event. See Cook (2007) for discussion and diagnostics for this form; it is neither a tense marker nor an aspectual marker. 375 PROXIMAL (PRX) : Using the root aw-, this is one of the two spatial deixis categories of modern Plains cree. It codes that the referent is near the Speaker, but is also used to pick out topical referents, among other poorly understood effects. PROXIMATE (PROX): An simple ANIMATE referent that stands in discourse contrast to the OBVIATIVE referent. This dissertation argues that, in this context, the proximate referent has a special relation the Speaker of the discourse. The Speaker knows about the INTENTIONs of this referent, and has received information from this referent. (Wolfart 1978) REDUPLICATION (RED): A prefix that copies the first consonant of the some grammatical form; usually either a preverb or a verb stem. The copied syllable either has the shape Ca- or Cah-, with the difference signalling a difference in aspectual properties. The Cah- kind is often referred to as \u201cheavy reduplication\u201d REFLEXIVE (REFL) : A verbal suffix with the form \u2013iso, that makes the agent act upon themselves (e.g. w\u00eecih- \u2018help\u2019 \uf0e0 w\u00eecihiso- \u2018help oneself\u2019). RELATIVE ROOT (RR) : One of a closed class of elements that code different things in different positions. When standing alone, they typically code prepositional-like properties (e.g. ohci \u2018from\u2019 and isi \u2018towards\u2019). When inserted into the perverbs, they either code a relation to some element external to the verb structure (e.g. \u00eakosi \u00ea-isi-w\u00e2pam\u00e2t \u2018that\u2019s how he saw her\u2019), or, under negation, a tense-like effect. They are frequently used as the root of verbs, especially verbs that relate to propositions (e.g. it\u2013 in it\u00eayihtam \u2018think thus of it\u2019). STATIVE (STAT) : A convenient gloss for a common pair of AI finals (-isi \/ -\u00e2), which often occur in stative environments (e.g. kinosi- \u2018be tall\u2019). Its exact nature is complex, however, and the gloss here is only meant as a convenience, not an final analysis. 376 STEM( [\u2026] ): The part of an Algonquian word that does not change when the rest of the inflection changes. Since changes in person interactions result in changes to some of the transitivity marking, the Algonquianist notion of \u2018stem\u2019 is not equivalent to the shape posited by formal linguists, who posit that the verb structure includes all transitivity marking. (Bloomfield 1962, Wolfart 1973, Hirose 2000, D\u00e9chaine 2003). SUBJUNCTIVE (SBJ): A MODE of the CONJUNCT ORDER, distinguished by the lack of CHANGE (\u00ea- or k\u00e2-) and the addition of the suffix -i. When it occurs with INTIAL CHANGE, the subjunctive means that the verbal event has occurred multiple times (roughly like English \u201cwhenever\u201d). When it occurs without INITIAL CHANGE, it means that the verbal event is only hypothetical (roughly like English \u201cif\u201d or \u201cwhen\u201d). It is for this last usage that it is named. THEME SIGN: A closed class of morphemes that occur immediately after the final on TRANSITIVE Algonquian verbs. When combined with the information provided by person marking, these suffixes identify agents and patients. TRANSITIVITY CLASSES (TA, TI, AI, II, AIT): A classification scheme for Algonquian verbs, based on the morphological marking they show. This marking identifies what kind of nominals the verb relates to. The classification assumes an SVO ordering; AI = Animate (Subject) Intransitive (Verb), TA = Transitive (Verb) Animate (Object). The classification is only a convenience, since forms exist that function as transitive verbs but lack transitive morphology (sometimes called AIT forms; Animate Intransitive Transitive). UNSPECIFIED SUBJECT (USC): A construction in Plains Cree that suppresses reference to the Agent of a transitive verb. The resulting form has only a single ARGUMENT, with only a generic or unspecified existential REFERENT being understood to be the agent. a possessor. 377 References Aboh, Enoch. 2004. Morphosyntax of Complement-Head Sequences: Clause Structure and Word Order Sequences in Kwa. Oxford: Oxford University Press. Ahenakew, Alice. 2000. \u00e2h-\u00e2y\u00eetaw isi \u00ea-k\u00ee-kisk\u00eayihtahkik maskihkiy \/ They knew both sides of medicine: Cree tales of curing and cursing told by Alice Ahenakew, H.C. Wolfart & Freda Ahenakew (eds). Publications of the Algonquian Text Society \/ Collection de la Soci\u00e9t\u00e9 d\u2019\u00e9dition des textes algonquiens, Winnipeg: University of Manitoba Press. Ahenakew, Freda & H.C. Wolfart. 1983. Productive reduplication in Plains Cree. Actes du 14e Congr\u00e8s des Algonquinistes, ed. by William Cowan, pp. 369-377. Ottawa: Carleton University Press. Ahenakew, Freda & H.C. Wolfart. (eds) 1997. kwayask \u00ea-k\u00ee-p\u00ea-kiskinow\u00e2pahtihicik. Their Example Showed Me the Way, Told by Emma Minde. Edmonton: University of Alberta Press. Ahenakew, Freda. 1987. Cree language structures: A Cree approach. Winnipeg: Pemmican Publications. Aikhenvald, Alexandra. 2004. Evidentiality. Oxford: Oxford University Press. Aissen, Judith. 1997. On the syntax of obviation. In Language, Vol. 73, pp. 705\u2013750. Austin, J. L. 1962. How to do things with words. Cambridge, MA: Harvard University Press. Baker, Mark. 1996. The polysynthesis parameter. Oxford studies in comparative syntax. New York: Oxford University Press. Banfield, Ann. 1982. Unspeakable Sentences. Boston: Routledge & Kegen Paul. Barker, Chris. 1995. Possessive Descriptions. California: CSLI. Barwise, Jon. 1981. Scenes and other situations. In Journal of Philosophy, Vol. 78, pp. 369-397. Bateson, Gregory. 1972. Steps to an Ecology of Mind: Collected Essays in Anthropology, Psychiatry, Evolution, and Epistemology. Chicago: University Of Chicago Press. Battistella, Edwin L. 1990. Markedness: The evaluative superstructure of language. New York: SUNY Press. Belletti, A. and Luigi Rizzi. 1988. Psych-Verbs and Theta-Theory. In Natural Language and Linguistic Theory, Vol. 6, pp. 291-352. Blain, Eleanor M., & Rose-Marie D\u00e9chaine. 2002. Evidential marking across the Cree dialect continuum. Paper read at the 7th Workshop on the Structure and Constituency of the Languages of the Americas, Edmonton, University of Alberta. 378 Blain, Eleanor. 1987. What\u2019s going on with the weather in Plains Cree? Unpublished manuscript, Winnipeg, MB: The University of Manitoba. Blain, Eleanor M. 1995. Emphatic wiya in Plains Cree. In Papers of the 26th Algonquian Conference, David H. Pentland (ed.), pp. 22-34. Winnipeg: University of Manitoba. Blain, Eleanor. 1997. Wh-Constructions in N\u00eahiyaw\u00eawin (Plains Cree). Ph.D. Dissertation. University of British Columbia. Blain, Eleanor, Clare Cook; Rose-Marie D\u00e9chaine; & Jeffrey Muehlbauer. 2006. Plains Cree evidentials: major and minor modes. Paper read at the annual meeting of the Society for the Study of the Indigenous Languages of the Americas, Albuquerque, NM. Bliss, Heather. 2005. Formalizing point of view: The role of sentience in Blackfoot\u2019s direct\/inverse system. Master\u2019s Thesis. Calgary: The University of Calgary. Bliss, Heather. 2005. Topic, Focus, and Point-of-view in Blackfoot. In Proceedings of the 24th West Coast Conference on Formal Linguistics, John Alderete, Chung-hye Han & Alex Kochetov (eds), pp. 61-69. Somerville, MA: Cascadilla Proceedings Project. Bloomfield, Leonard. (ed.) 1930. Sacred stories of the Sweet Grass Cree. Bulletin \/ National Museum of Canada, no. 60. Anthropological series, no. 11. Ottawa: F. A. Acland. Bloomfield, Leonard. 1946. Algonquian. In Linguistic structures of Native America, ed. by Harry Hoijer et al., pp. 85-129. New York: Viking Fund Publications in Anthropology 6.. Bloomfield, Leonard. 1958. Eastern ojibwa: grammatical sketch, texts, and word lists. Charles. F. Hockett (ed). Ann Arbor, MI: University of Michigan Press. Bloomfield, Leonard. 1962. The Menomini Language, ed. by Charles F Hockett. New Haven, CT: Yale University Press. Bloomfield, Leonard. 1975. Menomini lexicon, ed. by Charles F. Hockett. Milwaukee, WI: Milwaukee Public Museum Publications in Anthropology and History 3. Bouchard, Denis. 1995. The semantics of syntax. Chicago: The University of Chicago Press. Borer, Hagit. 2005. Structuring Sense. Oxford: Oxford University Press. Branigan, Phil and Marguerite MacKenzie. 2002. Altruism, A\u2019-movement, and object agreement in Innu-Aimun. Linguistic Inquiry 33(3). pp. 385-407. Branigan, Phil & Marguerite MacKenzie. 2007. The morphosyntax of verbal orders in Innu- Aim\u00fbn. Paper presented at the 12th Workshop on the Structure and Constituency of the Languages of the Americas, Lethbridge: AB. Brentano, Franz. 1874. Psychologie vom empirischen Standpunkt, ed. Oskar Kraus, 2 vols. Hamburg: Felix Meiner, 1955. 379 Bruening, Benjamin. 2001. Syntax at the edge: Cross-clausal phenomena and the syntax of Passamaquoddy. Ph.D. thesis, Massachusetts Institute of Technology. Burton, Strang. 1995, Six Issues to Consider When Choosing a Husband, Doctoral Dissertation, Rutgers, New Brunswick, New Jersey. Burton, Strang. 2005. Multimedia text methodology. Paper presented at SILS 12, University of Victoria. Carden, Guy. 1986. Blocked forwards co-reference; unblocked forwards anaphora: Evidence for an abstract model of co-reference, in Proceedings of the Chicago Linguistics Society 22, pp. 262-276. Carden, Guy, & Thomas Dieterich. 1981. Introspection, Observation, and Experiment: An Example where Experiment Pays Off, in Proceedings of the 1980 Biennial Meeting of the Philosophy of Science Association, Volume 2, ed. Peter Asquith and Ronald Giere, pp. 583-597. East Lansing, MI: Philosophy of Science Association. Cheung, Candice & Richard Larson. 2006. Chinese psych verbs and covert clausal complementation. Paper presented at the Chicago Workshop on Chinese, Chicago. Chisholm, Roderick. 1967. Intentionality. In The Encyclopedia of Philosophy, Paul Edwards (ed.), Vol. 3: 201-204. New York: Macmillan and Free Press. Chomsky, Noam. 1982. Some concepts and consequences of the theory of government and binding. Cambridge, MA: MIT Press. Chomsky, Noam. 1995. The minimalist program. Cambridge, MA: MIT Press. Chung, Sandra & William Ladusaw. 2003. Restriction and saturation. Cambridge, MA: MIT Press. Clarke, Sandra. 1982. North-West River (Shesh\u00e2tsh\u00eet) Montaignais: A grammatical sketch. Ottawa: National Museum of Man (Mercury Series, Canadian Ethnology Service Paper #80). Clements, George N. 1975. The logophoric pronoun in Ewe: Its role in discourse, in Journal of West African Languages, Vol. 10, pp. 141-77. Cook, Clare. 2003. Syllabification in Plains Cree. Unpublished manuscript, University of British Columbia Cook, Clare. 2003b. Relative roots in Plains Cree and Passamaquoddy. Paper given at the 35th Algonquian Conference; London, Ontario. Cook, Clare. 2006. k\u00ee- as a CP-level anaphoric disjoint marker. Paper given at the 2006 SSILA conference, Albuquerque, NM. 380 Cook, Clare. 2007. Distinguishing modes in Plains Cree, in Proceedings of 38th Algonquian Conference, H.C. Wolfart (ed.), pp. 47-80. Winnipeg: University of Manitoba Press. Cook, Clare. 2008. The syntax and semantics of clause-typing in Plains Cree. Doctoral Dissertation. Vancouver: University of British Columbia. Cook, Clare & Jeff M\u00fchlbauer. 2006. The behaviour of obviation in elicitation, in Actes du 37\u00e8me congr\u00e8s des algonquianistes, H.C. Wolfart (ed.), pp. 103-129. Winnipeg: University of Manitoba. Cook, Clare & Jeff M\u00fchlbauer. 2007. Constructing presence in Plains Cree. In Proceedings of the Workshop on Structure and Constituency in Languages of the Americas XII, S. Chin & H. Fusheini (eds.). University of British Columbia Working Papers in Linguistics. Vancouver. Cook, Eung-Do. 2004. A Grammar of D\u00ebne S\u0173\u0142in\u00e9 (Chipewyan). Winnipeg: Algonquian and Iroquoisan Linguistics. Costa, David J. 2001. Shawnee noun plurals. In Anthropological Linguistics 43:255-287. Cresswell, M. J. 1985. Structured meanings: The semantics of propositional attitudes. Cambridge, MA: MIT press. Dahlstrom, Amy. 1986. Plains Cree morphosyntax. PhD dissertation. University of California \u2013 Berkeley. Dahlstrom, Amy. 1991. Plains Cree morphosyntax. New York: Garland. Dahlstrom, Amy. 1995. Motivation vs. predictability in Algonquian gender. Papers of the 26th Algonquian Conference, David H. Pentland (ed.), pp. 52-66. Winnipeg: University of Manitoba. Dahlstrom, Amy. 1995. Topic, Focus and other word order problems in Algonquian. The Belcourt Lecture, Winnipeg: Voices of Rupert\u2019s Land. Darnell, Regna, & Anthony L. Vanek. 1976. The semantic basis of the animate\/inanimate distinction in Cree. In Papers in Linguistics 9.3\/4:159-180. Daviault, Diane. 1994. L\u2019algonquin au XVIIe si\u00e8cle. Une \u00e9dition critique, analys\u00e9e et coment\u00e9e de la grammaire du P\u00e8re Louis Nicolas. Sainte-Foy, Qu\u00e9bec: Presse de l\u2019Universit\u00e9 du Qu\u00e9bec. D\u00e9chaine, Rose-Marie. 1997. Nominal predication in Plains Cree. Papers of the 28th Algonquian Conference, David H. Pentland (ed.), pp. 105-135. Winnipeg: University of Manitoba. D\u00e9chaine, Rose-Marie. 1999. What Algonquian morphology is really like: Hockett Revisited. Proceedings from the Workshop on the Structure and Constituency in Native American Languages, Leora Bar-el, Rose-Marie D\u00e9chaine, & Charlotte Reinholtz (eds.), pp. 25-72. Cambridge MA: MIT Occasional Papers in Linguistics 17. 381 D\u00e9chaine, Rose-Marie. 2003. The structure of vP in Plains Cree. Paper given at the Seventh Workshop on the Structure and Constituency of the Languages of the Americas, Brandon, MB. D\u00e9chaine, Rose-Marie. 2007. Order in a free word order language: the conceptual necessity of discourse configurationality. Paper given at the Linearization Workshop. Nagoya, Japan. D\u00e9chaine, Rose-Marie & Charlotte Reinholtz. 1997. The direct\/inverse as split case. Paper given at the Second Workshop on the Structure and Constituency of the Languages of the Americas, Brandon, MB. D\u00e9chaine, Rose-Marie & Charlotte Reinholtz 2008. Case theory meets linking theory: Algonquian direct\/inverse as split case. Paper given at the Thirteenth Workshop on the Structure and Constituency of the Languages of the Americas. Kingston: ON. D\u00e9chaine, R-M., & M. Wilstchko. 2002. Decomposing pronouns. In Linguistic Inquiry 33:409- 442. DiSciullo, Anna-Maria. 2005. Asymmetry in morphology. Cambridge, MA: MIT Press. Dryer, Matthew S. 1996. Passive vs indefinite actor construction in Plains Cree. Papers of the 27th Algonquian Conference, ed. by David H. Pentland, pp. 54-79. Winnipeg: University of Manitoba. Edwards, Mary. 1954. Cree: An Intensive Language Course. Edmonton: Northern Canada Evangelical Mission. Eliot, John. 1666. The Indian grammar begun: Or an essay to bring the Indian language into rules. Ellis, C. Douglas. 1983. Spoken Cree: West coast of James Bay. revised edition. Edmonton: Pica Pica Press. Farkas, Donka F. 1992. On the semantics of subjunctive complements. In Romance Languages and Linguistic Theory, P. Hirschbuhler & K. Koerner (eds.), pp. 69-104, Amsterdam: John Benjmans. Fillmore, Charles. 1975. Lectures on Deixis, 1971. Indiana University Linguistics Club. Fodor, Janet. 1992. Islands, learnability and the lexicon. In Island constraints, Theory, Acquisition, and Processing, Helen Goodluck & Michael Rochemont (eds.), pp. 109-180. Dordrecht: Kluwer. Frantz, Donald. 1970. Frantz, Donald G. 1970. Toward a generative grammar of Blackfoot (with particular attention to selected stem formation processes). Ph.D. thesis. University of Alberta. Frantz, Donald G. 1991. Blackfoot grammar. Toronto: University of Toronto Press. 382 Featherstone, Samuel. 2004. Bridge verbs and V2 verbs \u2013 the same thing in spades? In Zeitschrift f\u00fcr Sprachwissenschaft, Vol. 23(2), pp. 181-209. Berlin: Walter DeGruyter. Frege, Gottlob. 1892. \u00dcber Sinn und Bedeutung, Zeitschrift f\u00fcr Philosophie und philosophische Kritik, 100, pp.25-50. Garrett, Edward. 2001. Evidentiality and assertion in Tibetan. Ph.D. thesis, University of California \u2013 Los Angeles. Giv\u00f3n, Talmy. 1979. On understanding grammar. New York: Academic Press. Goddard, Ives. 1967. The Algonquian independent indicative. In Bulletin 214 Contributions to anthropology: Linguistics I, pp. 66-106. Ottawa: National Museum of Canada. Goddard, Ives. 1974. Remarks on the Algonquian Independent Indicative. International Journal of American Linguistics 40:317-327. Goddard, Ives. 1984. Obviation in Fox narrative discourse. In Papers of the 15th Algonquian Conference, William Cowan (ed.), pp. 273-286. Ottawa: Carleton University. Goddard, Ives. 1991. Obviation in Fox discourse. In International Journal of American Linguistics 58: Goddard, Ives. 2002. Grammatical gender in Algonquian, in Proceedings of 33rd Algonquian Conference, H.C. Wolfart (ed.), pp. 195-231. Winnipeg: University of Manitoba Press. Goodwin, William. 1875. Syntax of the moods and tenses of the Greek verb. Boston: Ginn Brothers. Grafstein, Ann. 1984. Disjoint reference in a free word order language. In Theoretical perspectives on Native American languages, ed. by Donna B. Gerdts. Green, Georgia. 1989. Pragmatics and Natural Language Understanding. Hillsdale, NJ: Lawrence Erlbaum Associates. Grice, H. P. 1989. Studies in the way of words. Cambridge, MA: Harvard University Press. Grimshaw, Jane. 1990. Argument structure. Cambridge, MA: MIT Press. Hag\u00e8ge, C. 1974. Les pronoms logophoriques. In Bulletin de la Soci\u00e9t\u00e9 de Linguistique de Paris 69, pp. 287-310. Hahn, E. Adelaide. 1952. The Moods in Indirect Discourse in Latin, In Transactions and Proceedings of the American Philological Association, Vol. 83, pp. 242-266. Haiman, J. & Pamela Munro. 1983. Switch Reference and Universal Grammar. Amsterdam: John Benjamins. 383 Hale, Kenneth L. 1973. A note on subject-object inversion. In Issues in Linguistics: Papers in Honor of Henry and Renee Kahane, B. Kachru (ed.), pp. 300-399. Urbana, IL: University of Illinois Press. Halle, Morris & Alec Marantz. 1993. Distributed morphology and the pieces of inflection. The view from building 20, ed. by K. Hale and J. Keyser, pp. 111-176. Cambridge, MA: MIT Press. Hasler, Laurel. 2002. Obviation in two Innu-Aimun Atanukana. Master\u2019s Thesis, Memorial University of Newfoundland. Heim, Irene. 1982. The Semantics of Definite and Indefinite Nouns Phrases. Ph.D. Dissertation, Amherst, MA: University of Massachussetts. Heisenberg, Werner. 1955. Das Naturbild der heutigen Physik. Hamburg: Rowohlt. Higginbotham, James. 1983. Logical Form, Binding, and Nominals. In Linguistic Inquiry 14, pp 305-420. Hintikka, J. 1962. Knowledge and Belief. Ithaca, NY: Cornell University Press. Hirose, Tomio. 1999. Diminutive palatalization in Plains Cree. Unpublished manuscript, University of British Columbia. Hirose, Tomio. 2000. Origins of predicates: Evidence from Plains Cree, Doctoral Dissertation, The University of British Columbia, Vancouver, British Columbia. Hockett, Charles F. 1966. What Algonquian is really like. International Journal of American Linguistics, Vol. 32, pp. 59-73. Horseherder, Nicole. 1998. A binding-theoretic analysis of Navajor possessor yi-. Master\u2019s Thesis, University of British Columbia. Howse, Joseph. 1865. A Grammar of the Cree Language: With which is combined an Analysis of the Chippeway Dialect. London: Trubner & Co. Huang, Shuanfan. 1981. On the (almost) perfect identity of speech and thought: Evidence from Chinese dialects. In Proceedings of the 14th International Conference on Sino-Tibetan Languages and Linguistics, pp. 171-186. Hume, David. 1743. An Enquiry into human understanding. Hunter, Emily, and Betty Karpinski. 1999. Introductory Cree: Part 1. Edmonton: School of Native Studies, University of Alberta. Jackendoff, Ray. 2002. Foundations of language. New York: Oxford University Press. Jakobson, Roman. 1971. Selected Writings Vol II: Word and Language. The Hague: Mouton. 384 Jaszczolt, Katarzyna. 2004. Default semantics: Foundations of a compositional theory of acts of communication. Oxford: Oxford University Press. Jelinek, Eloise. 1984. Empty categories, case, and configurationality. In Natural Language and Linguistic Theory 2: 39-76. Junker, Marie-Odile. 2003. East Cree dependent nouns and disjoint reference, in Algonquian and Iroquoian Linguistics 28, pp 11-13. Junker, Marie-Odile. 2004. Focus, obviation, and word-order in East Cree, in Lingua, Vol. 114, pp. 345\u2013365. K\u00e2-N\u00eepit\u00eaht\u00eaw, Jim. 1998. ana k\u00e2-pimw\u00eaw\u00eahahk okak\u00easkihk\u00eamowina \/ The counselling speeches of Jim K\u00e2-N\u00eepit\u00eaht\u00eaw, Freda Ahenakew & H.C. Wolfart (ed. & tr.). Publications of the Algonquian Text Society \/ Collection de la Soci\u00e9t\u00e9 d\u2019\u00e9dition de textes algonquiens, Winnipeg: University of Manitoba Press. Kamp, H. 1981. A theory of truth and semantic representation. In Formal methods in the study of language, J. Groenendik, T. Janssen and M. Stokhof (eds.), pp. 277-322. Amsterdam: Mathematics Centre Tracts. Kamp, H. 1990. Prolegomena to a structural theory of belief and other attitudes. In Propositional Attitudes: The Role of Content in Logic, Language, and the Mind, C.A. Anderson & J. Owens (eds.). Stanford, CA: CSLI. Karttunen, Lauri. 1977. The syntax and semantics of questions. Linguistics and Philosophy 1: pp. 1-44. Kayne, Richard. 1994. The Antisymmetry of Syntax. Cambridge, MA: MIT Press. King, Jeffrey. 2001. Complex Demonstratives. Cambridge, MA: MIT Press. K\u00f6lbel, Max. 2002. Truth without objectivity. London: Routledge. Koopman, Hilda. 2000. The syntax of specifiers and heads. Leading Linguist Series. Routledge. Korzybski, Alfred. 1958. Science and sanity: An introduction to non-Aristotelian systems and General Semantics. Institute of General Semantics. Kratzer, Angelika. 1977. What \u2018Must\u2019 and \u2018Can\u2019 Must and Can Mean. Linguistics and Philosophy 1: 337\u2013355. Kratzer, Angelika. 1981. The Notional Category of Modality. In Words, Worlds, and Contexts. New Approaches in Word Semantics, H.-J. Eikmeyer and H. Rieser (eds.), pp. 38\u201374. Berlin: de Gruyter. Kratzer, Angelika. 1991. Modality. In Semantik. Ein internationales Handbuch der zeitgen\u00f6ssischen Forschung, A. von Stechow and D. Wunderlich (eds.), pp. 639\u2013650. Berlin: de Gruyter. 385 Kratzer, Angelika. 2007. Situations in natural language semantics. In The Stanford Encyclopedia of Philosophy. Palto Alto, CA: CSLI Publications. Kripke, Saul. 1963. Semantical analysis of modal logic. In Zeitschrift f\u00fcr Mathematische Logik und Grundlagen der Mathematik, Vol. 9, pp. 67-96. Kripke, Saul. 1980. Naming and Necessity. Cambridge, MA: Harvard University Press. Kuno, Susumu. 1972. Pronominalization, relativization, and direct discourse. In Linguistic Inquiry, Vol. 3., pp. 161-195. Kuno, Susumu. 1987. Functional Syntax: anaphor, discourse, and empathy. Chicago: University of Chicago Press. Lacombe, Albert. 1874. Dictionnaire de las Langue des Cris. Montreal. Lakoff, George. 1970. Counterparts and the problem of reference in transformational grammar. In Report # NSF-24 to the National Science Foundation, S. Kuno (ed), pp. 23-27. Cambridge, MA: Harvard Computational Laboratory. Lakoff, George. 1971. Crossover phenomena. New York: Holt, Rinehart, and Winston. Lakoff, George. 1987. Women, fire, and dangerous things. Chicago: University of Chicago Press. Landman, Fred. 1986. Towards a theory of information. PhD Thesis, Universiteit van Amsterdam. Langacker, Ronald. 1969. Pronomilization and the chain of command, in Modern Studies in English, D. Reibel and S. Schane (eds.), pp. 160-186. New Jersey: Prentice Hall. Lasersohn, Peter. 2005. Context Dependence, Disagreement, and Predicates of Personal Taste. In Linguistics and Philosophy 28: 643\u2013686. LeClaire, Nancy & George Cardinal. 1998. Alberta Elders' Cree Dictionary. Edmonton: University of Alberta Press. Leman, Wayne 1996: A Cheyenne Version of \u2018The Rolling Head\u2019. In: Nichols, John D. and Arden C. Ogg (eds.): nikotw\u00e2sik iskw\u00e2ht\u00eam, p\u00e2skiht\u00eapayih! Studies in Honour of H. C. Wolfart. Algonquian and Iroquoian Linguistics, Memoir 13. Winnipeg: Voices of Rupert\u2019s Land. Levin, Beth & M. Rappaport Hovav. 1996. Unaccusativity: At the syntax-lexical-semantics interface. Cambridge, MA: MIT Press. Lewis, David. 1979. Attitudes De Dicto and De Se. Philosophical Review 88: 513\u2013543. Li, Charles & Sandra Thompson. 1991. Mandarin Chinese: A functional reference grammar. Santa Barbara, CA: University of California Press. 386 Lightning, Walter. 1992. Compassionate mind: implications of a text written by elder Louis Sunchild. MA thesis. University of Alberta. Link, Godehard. 1983. The Logical Analysis of Plurals and Mass Terms: A Lattice-theoretical Approach. R. B\u00e4uerle, C. Schwarze, A. von Stechow, eds., Meaning, Use, and Interpretation of Language, 302-323. Berlin, New York: Springer. Lochbihler, Bethany. 2007. Mapping inversion: A morpho-syntactic analysis of the inverse system in Ojibwa. Paper given at the 2007 CLA. Saskatoon, SK. Long, Michelle. 1999. Obviation and clausal relations in Plains Cree. MA thesis, University of Manitoba. Lyons, John. 1982. Deixis and Subjectivity: Loquor, ergo sum? In Speech, Place, and Action, Robert Jarvella & Wolfgang Klein (eds.), pp. 101-124, New York: John Wiley & Sons, Ltd. MacKenzie, Marguerite. 1980. Towards a dialectology of Cree-Montagnais-Naskapi. PhD thesis. University of Toronto. Matthewson, Lisa. 2004. On the methodology of semantic fieldwork. In The International Journal of American Linguistics 70:4, pp. 369-415. McCawley, James. 1988. The syntactic phenomena of English. Chicago: The University of Chicago Press. McLendon, Sally. 1978. Ergativity, case, and transitivity in Eastern Pomo. In International Journal of American Linguistics 44(1): 1-9. Minde, Emma. 1997. kwayask \u00ea-k\u00ee-p\u00ea-kiskinow\u00e2pahtihicik \u2018Their Example Showed Me the Way,\u2019 F. Ahenakew and H.C. Wolfart (eds), Edmonton: University of Alberta Press. M\u00fchlbauer, Jeffrey. 2002. Adjuncts to AI verbs in Menominee. Paper presented at the 34th Algonquian conference, Kingston, OT. M\u00fchlbauer, Jeffrey. 2003. Nominal ordering in Plains Cree texts. Unpublished manuscript, Vancouver: University of British Columbia. M\u00fchlbauer, Jeffrey. 2004. Specifying possessor relations in Plains Cree. Paper presented at the annual meeting of the Canadian Linguistics Association, Winnipeg. M\u00fchlbauer, Jeffrey. 2006. Genericity from the perspective of discourse. Paper presented at the annual meeting of the Society for the Study of the Indigenous Languages of the Americas. M\u00fchlbauer, Jeffrey. 2006. Pitch as accent in Plains Cree. In Actes du 37\u00e8me congr\u00e8s des algonquianistes, H.C. Wolfart, ed., pp. 229-268. Winnipeg: University of Manitoba. 387 M\u00fchlbauer, J. 2007. Evidence for three distinct nominal classes in Plains Cree. In Natural Language Semantics 15:2, pp. 167-186. M\u00fchlbauer, Jeffrey. 2007. Decomposing obviation in Plains Cree: The case of the morpheme \u2013yi\u2013. Presented at the annual meeting of the Candian Linguistics Association, Saskatoon, Saskatchewan. M\u00fchlbauer, Jeffrey. 2007. The speaker\u2019s knowledge and obviation in Plains Cree. In Proceedings of the 38th Algonquian Conference, H.C. Wolfart (ed.), pp. 343-389. Winnipeg: University of Manitoba. M\u00fchlbauer, Jeffrey. 2007. The syntax and semantics of animacy in Plains Cree. Unpublished manuscript, Vancouver: University of British Columbia. Munro, Pamela. 2006. Chickasaw positional verbs. Presented at SSILA, Albuquerque, New Mexico. Ogg, Arden. 1991. Connective particles and temporal cohesion in Plains Cree Narrative. Master\u2019s thesis, University of Manitoba. Okimasis, Jean & Solomon Ratt. 1999. Cree, Language of the Plains: Nehiyawewin, Paskwawi- Pikiskwewin. Regina, SK: University of Regina Canadian Plains Research Center. Oshima, David. 2007. Syntactic direction and obviation as empathy-based phenomena: a typological approach. In Linguistics: an interdisciplinary journal of the language sciences, Vol. 45-4, pp. 727-763. Partee, Barbara. 1989. Binding implicit variables in quantified contexts. In Papers of the Chicago Linguistics Society, Vol. 25, pp. 342-356. Partee, Barbara H. & Vladimir Borschev. 2003. Genitives, relational nouns, and argument- modifier ambiguity. In Modifying Adjuncts, E. Lang, C. Maienborn, & C. Fabricius- Hansen (eds.), pp. 67-112. Berlin: Mouton de Gruyter. Pelletier 2007 Genericity talk at UBC Pentland, David. 1996. Verbs of Precipitation in Algonquian. In nikotw\u00e2sik iskw\u00e2ht\u00eam, p\u00e2skiht\u00eapayih!, A. Ogg & J. Nichols (ed.), Winnipeg, Algonquian & Iroquioan Linguistics. Perlmutter, David. 1983. Studies in relational grammar. Chicago: University of Chicago Press. Pesetsky, David. 1995. Zero Syntax: Experiencers and Cascades. Cambridge, MA: MIT Press. Piriyawiboon, Nattaya. 2007. Reconsidering the obviative. In Proceedings of the 12th Workshop on the Structure and Constituency of the Languages of the Americas, Seok Koon Chin & Fusheini Hudu (eds.), pp. 100-113, Vancouver: University of British Columbia Working Papers in Linguistics. 388 Piriyawiboon, Nattaya. 2007. The elimination of the obviative. Generals Paper, Toronto: University of Toronto. Platzak, C. 1999. The subject of Icelandic psych-verbs : A minimalistic account. In Working papers in Scandinavian syntax, Vol. 64, pp. 103-115. Postal, Paul. 1969. Anaphoric islands. In Papers from the fifth regional meeting of the Chicago Linguistic Society, pp. 205-239. Postal, Paul. 1971. Crossover phenomena. New York: Academic Press. Postal, Paul. 1974. On Raising: One Rule of English Grammar and its Theoretical Implications. Cambridge, MA: MIT Press. Proulx, Paul. 1984. Proto-algic I: Phonological Sketch. In The International Journal of American Linguistics, Vol. 50, pp. 165-207. Proulx, Paul. 1988. The demonstrative pronouns of Proto-Algonquian. In The International Journal of American Linguistics, Vol. 54, pp. 309-330. Reinhart, Tanya. 1983. Anaphora and Semantic Interpretation. Chicago: University of Chicago Press. Reinholtz, Charlotte. 1997. Evidence for determiner phrases in Cree. Paper presented at the 2nd Workshop on the Structure and Constituency of the Languages of the Americas, University of Manitoba. Reinholtz, Charlotte, & H.C. Wolfart. 1996. Sentences [Sketch of Cree, an Algonquian language]. Handbook of North American Indians, Ives Goddard (ed.), v. 17: Languages, pp. 391-398. Washington: Smithsonian Institution. Rhodes, Richard A. 1976. The morphosyntax of the Central Ojibwa verb. Ph.D. thesis, University of Michigan. Rhodes, Richard A. 2001. Text strategies in M\u00e9tchif. In Actes du 32e Congr\u00e8s des Algonquinistes, John D. Nichols (ed), pp. 455-469. Winnipeg: University of Manitoba. Ritter, Elizabeth. and Sara Rosen. 2005. Agreement without A positions: Another look at Algonquian. In Linguistic Inquiry, Vol. 36, pp. 648-660. Ritter, Elizabeth & Martina Wiltschko. 2007. Alternatives to tense in Blackfoot and Halkomelem. Paper presented at the 12th Workshop on the Structure and Constituency of the Languages of the Americas, Lethbridge, AB. Rizzi, Luigi. 1986. Null Objects in Italian and Theory of pro. Linguistic Inquiry Vol. 17(3), pp. 501\u2013557. Rizzi, Luigi. 1990. Relativized Minimality. Cambridge, MA: MIT Press. 389 Rizzi, Luigi. 1997. The Fine Structure of the Left Periphery. In Elements of Grammar. Liliane Haegeman (ed.), pp. 281-337. Dordrecht: Kluwer. Ross, John. 1967. Constraints on variables in syntax. Ph.D. Dissertation, Massachussetts Institute of Technology. Ross, John Robert. 1970. On declarative sentences. In Readings in English transformational grammar, Roderick Jacobs and Peter Rosenbaum (ed.), pp. 222\u2013272. Waltham, MA: Ginn and Company. Russell, Bertrand. 1918. The philosophy of logical atomism. In The Monist. Russell, Kevin. 1991. Obviation as discourse structure in a Swampy Cree \u00e2cimowin. Papers of the 22nd Algonquian Conference, ed. by William Cowan, pp. 320-335. Ottawa: Carleton University Press. Russell, Kevin. 1996. Does obviation mark point of view? nikotw\u00e2sik iskw\u00e2ht\u00eam, p\u00e2skiht\u00eapayih! Studies in honour of H.C. Wolfart, ed. by John D. Nichols & Arden C. Ogg, pp. 367-382. Algonquian and Iroquoian Linguistics Memoir 13. U Manitoba Press: Winnipeg, MB. Ruwet, Nicolas. 1972. Th\u00e9orie syntaxique et syntaxe du fran\u00e7ais. Paris: Le Seuil. Ruwet, Nicolas. 1991. Syntax and human experience. Chicago: University of Chicago Press. Safir, Ken. 2004. The syntax of (in)dependence. Cambridge, MA: MIT Press. Saxon, Leslie. 1986. The syntax of pronouns in Dogrib (Athapaskan): Some theoretical consequences. Ph.D. dissertation, University of California, San Diego. Searle, John. 1965. What is a Speech Act? In Philosophy in America, Max Black (ed.), Ithaca, N.Y.: Cornell University Press. Schlenker, Philippe. 2003. A plea for monsters. In Linguistics and Philosophy 26; pp. 29-120. Schulz, Dora & Heinz Griesbach. 1965. Grammatik der deutschen Sprache. Sells, Peter. 1987. Aspects of Logophoricity. Linguistic Inquiry 18, pp. 445-479. Smith, Carlotta. 2004. Modes of discourse. Cambridge: Cambridge University Press. Stalnaker, Robert. 1974. Pragmatic presuppositions. In Semantics and Philosophy, M. Munitz and P. Unger (eds.), pp 197-213, New York University Press, New York, New York. Stephenson, Tamina. 2007. Towards a Theory of Subjective Meaning. Ph.D. Dissertation. Cambridge, MA: MIT. Steriade, Donca. 1995 Underspecification and markedness. In The Handbook of Phonological Theory, John Goldsmith (ed.), pp. 114-174, Cambridge, MA: Blackwell Publishers. 390 Stirling, Lesley. 1993. Switch-reference and discourse representation. Cambridge: Cambridge University Press. Tarski, Alfred. 1944. The semantic conception of truth and the foundations of semantics. In Philosophy and Phenomenological Research 4. Thwaites, Reuben G. 1896-1901. The Jesuit Relations and allied documents. Cleveland: Burrows Brothers. Valentine, J. Randolph. 2001. Nishnaabemwin reference grammar. Toronto: University of Toronto Press. Whitecalf, Sarah (1993) Kin\u00eahiy\u00e2wiwininaw N\u00eahiyaw\u00eawin \/ The Cree Language is Our Identity: The La Ronge Lectures of Sarah Whitecalf. H. C. Wolfart and Freda Ahenakew (ed. & tr.). Winnipeg, Manitoba: University of Manitoba Press. Williams, Edwin. 1980. Predication. Linguistic Inquiry 11. pp. 203-238. Williams, Edwin. 1981. Argument Structure and Morphology. In Linguistic Review, Vol. 1, pp. 81-114. Williams, Edwin. 1997. Blocking and anaphora. Linguistic Inquiry 28:577-628. Wolfart, H.C. 1971. Plains Cree internal syntax and the problem of noun-incorporation. In the Proceedings of the International Congress of Americanists, Vol. 38.3, pp. 511-518. Wolfart, H.C. 1973. Plains Cree: A Grammatical Study. Philadelphia, PA: American Philosophical Society Transactions. Wolfart, H.C. 1978. How many obviatives? Sense and reference in a Cree paradigm, in Linguistics Studies of Native Canada, Eung-Do Cook & Jonathan Kaye (eds.), pp. 255- 279, UBC Press: Vancouver, BC. Wolfart, H.C. 1996. Sketch of Cree, an Algonquian language. In W.C. Sturtevant, ed., \u201cHandbook of North American Indians 17, pp. 390-439. Washington D.C.: Smithsonian Institution. Wolfart, H.C. 2000. Introduction & Notes. \u00e2h-\u00e2y\u00eetaw isi \u00ea-k\u00ee-kisk\u00eayihtahkik maskihkiy \/ They knew both sides of medicine: Cree tales of curing and cursing told by Alice Ahenakew, ed. by H.C. Wolfart and Freda Ahenakew, H.C. Wolfart & Freda Ahenakew (eds.). Winnipeg: University of Manitoba Press. Wolfart, H.C., and Freda Ahenakew. 1993. kin\u00eahiy\u00e2wiwininaw n\u00eahiyaw\u00eawin: The Cree Language is Our Identity. The La Ronge Lectures of Sarah Whitecalf. University of Manitoba Press. Wolfart, H.C., and Freda Ahenakew. 1998. The Counselling Speeches of Jim K\u00e2-Nip\u00eet\u00eaht\u00eaw. Manitoba: University of Manitoba Press. 391 Wolfart, H.C., and Freda Ahenakew. 1998. The student\u2019s dictionary of literary Plains Cree: Based on contemporary texts. Winnipeg: Algonquian and Iroquoisan Linguistics. Wolfart, H.C., and Freda Ahenakew. 2000. They Knew Both Sides of Medicine: Cree Tales of Curing and Cursing Told by Alice Ahenakew. Winnipeg: University of Manitoba Press. Wolfart, H.C. & F. Ahenakew: 1998, The Student\u2019s Dictionary of Literary Plains Cree, Winnipeg: Algonquian and Iroquoian Linguistics. Wolfart, H.C., & Freda Ahenakew. 2000. They Knew Both Sides of Medicine: Cree Tales of Curing and Cursing Told by Alice Ahenakew. Winnipeg: University of Manitoba Press. Wolfart, H.C. & Janet Carroll. 1973. Meet Cree: A practical guide to the Cree language. Edmonton: University of Alberta Press. Wolvengrey, Arok. 2001. N\u0113hi\u00fdaw\u0113win: Itw\u0113wina, Canadian Plains Research Center, Regina, Saskatchewan. Wolvengrey, Arok. 2003. The function and word order of Plains Cree demonstratives. In Alonquian and Iroquoisan Linguistics Vol. 28, pp. 22-25. Wolvengrey, Arok. 2007. Placing Plains Cree locatives: On the road to a functional analysis. Paper presented at the 12th Workshop on the Structure and Constituency of the Languages of the Americas, Lethbridge, Alberta. Zanuttini, Rafaella. 1997. Negation and clause structure: A comparative study of Romance languages. Oxford: Oxford University Press. ","attrs":{"lang":"en","ns":"http:\/\/www.w3.org\/2009\/08\/skos-reference\/skos.html#note","classmap":"oc:AnnotationContainer"},"iri":"http:\/\/www.w3.org\/2009\/08\/skos-reference\/skos.html#note","explain":"Simple Knowledge Organisation System; Notes are used to provide information relating to SKOS concepts. There is no restriction on the nature of this information, e.g., it could be plain text, hypertext, or an image; it could be a definition, information about the scope of a concept, editorial information, or any other type of information."}],"Genre":[{"label":"Genre","value":"Thesis\/Dissertation","attrs":{"lang":"en","ns":"http:\/\/www.europeana.eu\/schemas\/edm\/hasType","classmap":"dpla:SourceResource","property":"edm:hasType"},"iri":"http:\/\/www.europeana.eu\/schemas\/edm\/hasType","explain":"A Europeana Data Model Property; This property relates a resource with the concepts it belongs to in a suitable type system such as MIME or any thesaurus that captures categories of objects in a given field. It does NOT capture aboutness"}],"GraduationDate":[{"label":"Graduation Date","value":"2008-11","attrs":{"lang":"en","ns":"http:\/\/vivoweb.org\/ontology\/core#dateIssued","classmap":"vivo:DateTimeValue","property":"vivo:dateIssued"},"iri":"http:\/\/vivoweb.org\/ontology\/core#dateIssued","explain":"VIVO-ISF Ontology V1.6 Property; Date Optional Time Value, DateTime+Timezone Preferred "}],"IsShownAt":[{"label":"DOI","value":"10.14288\/1.0066442","attrs":{"lang":"en","ns":"http:\/\/www.europeana.eu\/schemas\/edm\/isShownAt","classmap":"edm:WebResource","property":"edm:isShownAt"},"iri":"http:\/\/www.europeana.eu\/schemas\/edm\/isShownAt","explain":"A Europeana Data Model Property; An unambiguous URL reference to the digital object on the provider\u2019s website in its full information context."}],"Language":[{"label":"Language","value":"eng","attrs":{"lang":"en","ns":"http:\/\/purl.org\/dc\/terms\/language","classmap":"dpla:SourceResource","property":"dcterms:language"},"iri":"http:\/\/purl.org\/dc\/terms\/language","explain":"A Dublin Core Terms Property; A language of the resource.; Recommended best practice is to use a controlled vocabulary such as RFC 4646 [RFC4646]."}],"Program":[{"label":"Program (Theses)","value":"Linguistics","attrs":{"lang":"en","ns":"https:\/\/open.library.ubc.ca\/terms#degreeDiscipline","classmap":"oc:ThesisDescription","property":"oc:degreeDiscipline"},"iri":"https:\/\/open.library.ubc.ca\/terms#degreeDiscipline","explain":"UBC Open Collections Metadata Components; Local Field; Indicates the program for which the degree was granted."}],"Provider":[{"label":"Provider","value":"Vancouver : University of British Columbia Library","attrs":{"lang":"en","ns":"http:\/\/www.europeana.eu\/schemas\/edm\/provider","classmap":"ore:Aggregation","property":"edm:provider"},"iri":"http:\/\/www.europeana.eu\/schemas\/edm\/provider","explain":"A Europeana Data Model Property; The name or identifier of the organization who delivers data directly to an aggregation service (e.g. Europeana)"}],"Publisher":[{"label":"Publisher","value":"University of British Columbia","attrs":{"lang":"en","ns":"http:\/\/purl.org\/dc\/terms\/publisher","classmap":"dpla:SourceResource","property":"dcterms:publisher"},"iri":"http:\/\/purl.org\/dc\/terms\/publisher","explain":"A Dublin Core Terms Property; An entity responsible for making the resource available.; Examples of a Publisher include a person, an organization, or a service."}],"Rights":[{"label":"Rights","value":"Attribution-NonCommercial-NoDerivatives 4.0 International","attrs":{"lang":"en","ns":"http:\/\/purl.org\/dc\/terms\/rights","classmap":"edm:WebResource","property":"dcterms:rights"},"iri":"http:\/\/purl.org\/dc\/terms\/rights","explain":"A Dublin Core Terms Property; Information about rights held in and over the resource.; Typically, rights information includes a statement about various property rights associated with the resource, including intellectual property rights."}],"RightsURI":[{"label":"Rights URI","value":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/","attrs":{"lang":"en","ns":"https:\/\/open.library.ubc.ca\/terms#rightsURI","classmap":"oc:PublicationDescription","property":"oc:rightsURI"},"iri":"https:\/\/open.library.ubc.ca\/terms#rightsURI","explain":"UBC Open Collections Metadata Components; Local Field; Indicates the Creative Commons license url."}],"ScholarlyLevel":[{"label":"Scholarly Level","value":"Graduate","attrs":{"lang":"en","ns":"https:\/\/open.library.ubc.ca\/terms#scholarLevel","classmap":"oc:PublicationDescription","property":"oc:scholarLevel"},"iri":"https:\/\/open.library.ubc.ca\/terms#scholarLevel","explain":"UBC Open Collections Metadata Components; Local Field; Identifies the scholarly level of the author(s)\/creator(s)."}],"Title":[{"label":"Title ","value":"k\u00e2-y\u00f4sk\u00e2tahk \u00f4ma n\u00eahiyaw\u00eawin : the representation of intentionality in Plains Cree","attrs":{"lang":"en","ns":"http:\/\/purl.org\/dc\/terms\/title","classmap":"dpla:SourceResource","property":"dcterms:title"},"iri":"http:\/\/purl.org\/dc\/terms\/title","explain":"A Dublin Core Terms Property; The name given to the resource."}],"Type":[{"label":"Type","value":"Text","attrs":{"lang":"en","ns":"http:\/\/purl.org\/dc\/terms\/type","classmap":"dpla:SourceResource","property":"dcterms:type"},"iri":"http:\/\/purl.org\/dc\/terms\/type","explain":"A Dublin Core Terms Property; The nature or genre of the resource.; Recommended best practice is to use a controlled vocabulary such as the DCMI Type Vocabulary [DCMITYPE]. To describe the file format, physical medium, or dimensions of the resource, use the Format element."}],"URI":[{"label":"URI","value":"http:\/\/hdl.handle.net\/2429\/940","attrs":{"lang":"en","ns":"https:\/\/open.library.ubc.ca\/terms#identifierURI","classmap":"oc:PublicationDescription","property":"oc:identifierURI"},"iri":"https:\/\/open.library.ubc.ca\/terms#identifierURI","explain":"UBC Open Collections Metadata Components; Local Field; Indicates the handle for item record."}],"SortDate":[{"label":"Sort Date","value":"2008-12-31 AD","attrs":{"lang":"en","ns":"http:\/\/purl.org\/dc\/terms\/date","classmap":"oc:InternalResource","property":"dcterms:date"},"iri":"http:\/\/purl.org\/dc\/terms\/date","explain":"A Dublin Core Elements Property; A point or period of time associated with an event in the lifecycle of the resource.; Date may be used to express temporal information at any level of granularity. Recommended best practice is to use an encoding scheme, such as the W3CDTF profile of ISO 8601 [W3CDTF].; A point or period of time associated with an event in the lifecycle of the resource.; Date may be used to express temporal information at any level of granularity. Recommended best practice is to use an encoding scheme, such as the W3CDTF profile of ISO 8601 [W3CDTF]."}]}