UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Topic time : the syntax and semantics of SqwXwu7mish temporal adverbials Currie, Elizabeth J. 1997

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1997-0520.pdf [ 4.8MB ]
Metadata
JSON: 831-1.0088237.json
JSON-LD: 831-1.0088237-ld.json
RDF/XML (Pretty): 831-1.0088237-rdf.xml
RDF/JSON: 831-1.0088237-rdf.json
Turtle: 831-1.0088237-turtle.txt
N-Triples: 831-1.0088237-rdf-ntriples.txt
Original Record: 831-1.0088237-source.json
Full Text
831-1.0088237-fulltext.txt
Citation
831-1.0088237.ris

Full Text

TOPIC TIME: THE SYNTAX AND SEMANTICS OF SQWXWU7MISH TEMPORAL ADVERBIALS by ELIZABETH J. CURRIE B.A.(Hons), Queen's University at Kingston Ontario, 1986 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES (Department of Linguistics) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA October 1997 © Elizabeth Jean Currie, 1997 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree . that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of / _ I ' ^ ^ r i i i sH C-s The University of British Columbia Vancouver, Canada Date Q c k o ^ r / ^ , i ^ ^ DE-6 (2/88) Abstract The goal of this thesis is to explain the syntax and semantics of phrasal temporal adverbials in SqwXwu7mish, a Coast Salish language. This thesis proposes that SqwXwu7mish temporal adverbials function either as a main predicate or as a temporal argument corresponding to Reichenbach1s (1947) reference point. When these adverbials are the main predicate, they get an event time reading; when they are at the right edge of the sentence, they get a reference time reading. Thus, SqwXwu7mish adverbials support the claim by de Swart (to appear) that adverbials in focus structure get an event time reading, while adverbials in topic structure get a reference time reading. Furthermore, this thesis argues that adverbs in SqwXwu7mish are not adjuncts but arguments, based on their restricted distribution. Therefore, SqwXwu7mish adverbs do not simply modify the reference time, they denote is as a temporal argument, Topic Time (Klein 1994). This argument is located in the specifier of the spatiotemporal predicate Aspect within the framework proposed by Demirdache and Uribe-Etxebarria (to appear, a,b). ii Table of Contents Abstract ii Table of Contents iii Acknowledgement v Dedication vi 1 . Introduction 1 1.1 The Proposal 1 1.2 Definitions of Tense 2 1.3 Reichenbach (1947) 7 1.4 The Notion of Reference Time 8 1.4.1 Hornstein and Thompson 10 1.4.2 Stowell, Dowty, and Kamp and Reyle 13 1.4.3 Klein's Topic Time 16 1.4.4 Summary 17 1.5 The Syntax of Tense and Aspect 18 1.6 The Syntax of Temporal Adverbials 20 1.7 The Semantics of Temporal Adverbials 22 1.7.1 Locating Adverbs 22 1.7.2 Aspectual Adverbs 23 1.7.3 Quantificational Adverbs 24 1.8 Outline of the Thesis 24 2 . SqwXwu7mish Morphology and Syntax 28 2.1 Word Order in the Main Clause 28 2.2 Tense and Aspect Marking 31 2.2.1 Realis and Irrealis 32 2.2.2 The Durative and the Deictic 33 2.2.3 Past, Present and Future 34 2.2.4 Local Directional 3 6 2.2.5 Summary 37 2.3 Determiners and Temporal Adverbs 38 2.3.1 The SqwXwu7mish Determiner System 39 2.3.2 Today, Yesterday and Tomorrow 40 2.3.3 Morning, Year 40 2.3.4 Calendar Names 41 2.3.5 Quantificational Adverbs 42 2.4 Complex Sentences 42 2.4.1 Coordination 42 2.4.2 Conjunctive Clauses 43 2.4.3 Nominalized Clauses 44 2.4.4 Zero Relative Clauses 46 2.5 Summary 46 3 . The Interpretation of Adverbs 47 3.1 Adverbs and Topic/Focus Structure 48 3.2 Adverb-Quantifier Interaction 50 3.2.1 Adverbs and Quantified Subjects 50 3.2.2 Adverbs and Only 53 iii 3.2.3 Cardinality Quantifiers 54 3.2.4 Summary 57 3.3 Topic Time in Narrative 57 3.4 The Perfect in English and SqwXwu7mish 61 3.5 Aspectual Adverbs 63 3.5.1 English Aspectual Adverbs 63 3.5.2 SqwXwu7mish Aspectual Adverbs 65 3.6 Conclusions 67 4 . The Distribution of SqwXwu7mish Adverbs 68 4.1 Two Temporal Adverbs 68 4.2 SqwXwu7mish Adverbs Prefer the Right Periphery 70 4.3 Adverbs and Subordinate Clauses 7 5 4.4 Functional Adverbs 77 4.5 Summary 81 5 . The Syntax of the Topic Time Argument 82 5.1 Topic Time, Temporal Adverbials and Aspect 82 5.2 The Syntax of Tense and Aspect 84 5.3 Adapting the Framework to SqwXwu7mish 86 5.3.1 Simple Tenses 86 5.3.2 The Category of Temporal Arguments 88 5.3.3 Aspect Recursion in SqwXwu7mish 90 5.3.4 The Predicate Tense in SqwXwu7mish 91 5.4 Analyzing the Pronoun-Verb Construction 93 5.4.1 Topic Time "Yesterday" 93 5.4.2 Zero Topic Time 95 5.4.3 Topic Time "Sunday" 96 5.4.4 Habitual Topic Time 98 5.4.5 Conclusions 99 5.5 Word Order 100 5.6 Subordinate Nominalized Clauses 104 5.6.1 The Syntax of Subordinate Clauses 105 5.6.2 The Syntax of Nominalized Clauses 106 5.6.3 Topic Time Adverbs and CP Nominalized Clauses 107 5.6.4 Topic Time Adverbs and DP Nominalized Clauses 111 5.6.5 Conclusions 114 5.7 Temporal Adverbs and Quantification 115 5.8 Conclusions 118 6 . I m p l i c a t i o n s , Further I s sues and Conclus ions 12 0 6.1 Impl ica t ions of the Analysis 120 6.2 Fur ther I s sues 122 6.3 Conclusions 125 Abbreviations. 127 Key to SqwXwu7mish Orthography 128 References 129 iv Acknowledgement First and foremost, I am indebted to the SqwXwu7mish Elders Language Group, YJ, LJ, DW, LB, and especially to EL, for her cookies and juice, her encouragement, and her wonderful sense of humour. Ha71h n-sqwalwen. Thank you also to SqwXwu7mish Nation, for granting permission to conduct this research, and in particular to Peter Jacobs and SeeQwaLia, who took charge of the protocol, as well as to the SqwXwu7mish Nation Education Department. I am eternally in the debt of my supervisor, Hamida Demirdache; if this thesis exists, it's because she never gave up on it or me; to Henry Davis, for his incisive comments and for his course on Salish Syntax, which made a BIG difference; and to Peter Jacobs, for setting up and sitting through every elicitation session, and discussing the results; any remaining inaccuracies in the data are not for lack of trying. The fieldwork for this thesis was funded by SSHRCC grant #410-95-1519. With my committee, I thank Rose-Marie Dechaine for her role as my supervisor before she escaped to Africa. The faculty and post-doctoral fellows in the Department of Linguistics at U.B.C. have been just grand, in particular Dr. Dale Kinkade, the great Salishanist and fine baker; Michael Rochemont, my introductory syntax and semantics professor; and Doug Pulleyblank, who made my various difficulties seem almost simple. Special thanks to Dr. Laurel Brinton for allowing me to participate in her course on Tense and Aspect. The Girona International Summer School in Linguistics 1996 helped sub-consciously, as it turns out. Thank you to professors Angelika Kratzer, Richard Kayne, James Higginbotham, Alain Rouveret, and especially, Myriam Uribe-Etxebarria; thanks also to the organizers and to my fabulously fun fellow participants, especially my "roomie" Isabelle Roy. The participants of the Northwest Linguistics Conference sat through a preliminary outline of this work that got the ball rolling. Fellow Graduate Students at U.B.C, in particular Lisa Matthewson, Leora Bar-el, Eleanor Blain, Monica Sanchez, Susan Blake, Tomio Hirose,Nicole Horseherder, Myles Leitch, Nike Ola, Taylor Roberts, Kimary Shahin, Lisa Chang, Cora Li, Sandra Lai, and Vanessa Valerga, have been great, and as for Carmen da Silva, the "key to the department", thanks for letting me behind the counter. Thanks to my dear friends Anne Robertson and her men, for keeping me anchored in the rest of the world; Denise Oliver and Doug Klassen, for debriefing sessions after Salish overdoses in Victoria; and Linda Andrews, for those three hour brunches. Finally, thanks to my father, the late William M. Currie, for being proud of my academic aspirations; my younger brother, James Currie, for being both an impediment and an inspiration; and especially, my mummy, Audrey Currie, for listening to me complain, cooking dinner for me, and extending my line of credit just one more time. v This thesis is dedicated to Auntie Eva and to Francisco, for his future. 1 . Introduction All languages use temporal adverbials to express time, but many lack grammaticalized tense. Indeed, tenses "are primarily parasitic on time adverbials (...) and cannot be properly understood without an understanding of their interaction with time adverbials" (Dowty 1979:323). The status of these adverbials and their interaction with tense is therefore central to the understanding of temporal interpretation in human language. 1. 1 The Proposal The goal of this thesis is to explain the syntax and semantics of phrasal temporal adverbials in SqwXwu7mish. Based on their interpretation and distribution, this thesis proposes that SqwXwu7mish temporal adverbials function either as a main predicate or as a temporal argument corresponding to Reichenbach's (1947) reference point. SqwXwu7mish temporal adverbs which are focused, that is, the main predicate, get an event time reading, whereas SqwXwu7mish adverbs that appear at the right edge of the sentence get a reference time reading. Furthermore, temporal adverbials in SqwXwu7mish provide evidence for the claim by de Swart (to appear) that reference time is specified by adverbs in the topic structure of the sentence, while the event time is specified by adverbs in the focus structure. This thesis further claims that adverbs in SqwXwu7mish do not modify the reference time, they denote it as an argument 1 called Topic Time (following Klein 1994). This argument is located in the specifier position of the spatiotemporal predicate Aspect within the framework proposed by Demirdache and Uribe-Etxebarria (to appear). This introductory chapter sets out necessary background on tense, temporal interpretation, and temporal adverbs. Section 1.2 examines definitions of tense; underlying these definitions are the three times formalized in Reichenbach's model, which is outlined in Section 1.3. The status of Reichenbach's "third" point, the reference time, has been widely debated, and this controversy is examined more closely in Section 1.4. Section 1.5 outlines Demirdache and Uribe-Etxebarria 's proposal that both Tense and Aspect are predicates which take temporal arguments corresponding to speech time, event time, and reference time. The syntax of temporal adverbials is discussed in Section 1.6, then the semantic classification of adverbs is presented in Section 1.7. The final section of this chapter outlines the thesis as a whole. 1. 2 Definitions of Tense Traditional definitions of tense combine tense and aspect, thereby classifying the English past perfect progressive, Kitty had been eating, as a tense rather than as past tense and perfect and progressive aspect. Typological treatments define tense as a grammatical category which expresses "location in time " (Comrie 1985:1) or, more 2 specifically, which "characterizes the location of an event with respect to a point in time" (Chung and Timberlake 1985:256). Such approaches seek to provide a descriptive framework to account for the range of tense, aspect and mood systems cross-linguistically. In truth conditional semantics, tense has been analyzed as a sentential operator which specifies the time at which the truth of the utterance is evaluated1. Generative grammarians seek to define tense in terms of the interaction of its semantics and syntax; tense is either a referential expression (Enc 1987) or a predicate (Zagona 1990, Stowell 1993, 1997). Both conceptions of tense are important for the claim of this thesis that temporal adverbs, which are referential expressions, are the arguments of a temporal predicate. Enc (1987:638) argues that "all temporal expressions should be treated on a par with NPs" (1987:639), that is, as referential expressions; for example, temporal adverbs can appear in argument position as in (1): (1)a.Tell me why I don't like Monday. b.Bob was worried about yesterday. Specifically, she proposes that tenses are "referential expressions denoting intervals". The interval denoted by tense is a temporal argument selected by the verb, therefore tense is a "referential argument of the verb" (Stowell, to appear:13) along with the NP subject and object. The •'-Arguments against the analysis of tense as a sentential operator can be found in Enc (1986, 1987) and Hornstein (1990). 3 proposal that tense refers in the same way as a nominal2 allows Eng to assume that tense carries a referential index. Assuming that tense is in INFL and has the meaning "past" or "present", a past tense is a relation which "must denote an interval that is prior to some other interval" (Eng 1987:641), while present denotes an interval that is the same as some other interval. This other interval, which in the matrix clause is the speech time, is "denoted" by COMP. When COMP is this "specifier of tense", it also has a temporal index and it governs tense in INFL. The index on a matrix COMP is always 0: in the present tense, the tense denotes an interval that is the same as the speech time denoted by COMP, and COMP and INFL are co-indexed; in the past tense, the tense denotes an interval which is before the interval denoted by COMP: (2)a.John died. b.[COMPo [NP [PASTi VP ]]], i*0 COMP and INFL cannot be co-indexed and co-referential in this case. The relation between COMP and tense is determined by "Anchoring Conditions", which state that tense is anchored "if it is bound in its governing category, or if its local Comp is anchored"; COMP is anchored "if it is bound within its governing category" or else "if it denotes the speech time" (Eng 1987:643). These principles are similar to those of the binding theory, yet are still distinct in order to 2Partee (197 3) compares tenses to pronouns, arguing that both require antecedents either in the sentence or the discourse. 4 characterize the difference between temporal and nominal reference. For Zagona (1990) and Stowell (1993, to appear a,b), tense is not itself a referential expression but a predicate which relates referential expressions, the two temporal arguments of the clause. The external argument, called "reference" time, denotes the speech time in the matrix clause; the internal argument is "the time of the event or state denoted by the verb phrase" (Stowell, to appear b:9). In Stowell's theory, the temporal predicate locates the internal event time argument in relation to the external "reference" time argument by its meaning: past means "after", present means "within", and future means "before". Syntactically, the Tense predicate is the head of the maximal projection TP; the external "reference" time argument is projected in the specifier of TP, while the internal event time argument is the complement of T, as proposed by Zagona (1990) . Stowell proposes that the category of these two time denoting arguments is a functional category ZP, "Zeit Phrase": the internal argument of tense, denoting the event time, is therefore a ZP which contains the VP; either Z or an operator in the specifier of ZP binds a variable, itself a ZP, in the specifier of VP3. The external argument of tense is a non-overt ZP-PRO, which in main clauses denotes the time of utterance and in subordinate clauses is controlled by the 3The binding of the variable in the specifier of VP by Z is analoguous to the binding of the variable in the specifier of NP by D under the DP hypothesis. 5 event time of the main clause. This yields the structure in (3), from Stowell (to appear b): (3) TP PRO-ZP T' / \ T - ZPi / \ Opi Z1 / \ Zi VP ZP VP The mechanics of this system will be further explored in terms of the extension proposed by Demirdache and Uribe-Etxebarria, presented in Section 1.5. To summarize, tense has been defined as a grammatical category which locates events in time, a sentential operator, a referential expression, and a predicate. Each of these treatments assumes at least the existence of the time of the event and the time of the utterance itself, which in matrix clauses is now; in addition, some assume the existence of a third time which mediates between the event time and the utterance time. This idea of temporal interpretation as a relation of three points has its origins in grammars of the seventeenth and eighteenth centuries and decriptive work of the nineteenth century (Klein 1994) and is formalized in Reichenbach's 1947 proposal. 6 1.3 Reichenbach (1947) From an example of the past perfect, Peter had gone, Reichenbach argues that tense is composed of three time points, the "point of speech" S, the "point of the event" E and the "point of reference" R, and that the third point, R, is instantiated in all tenses, even where its effects are not manifest. He claims that these three points can be ordered before, after or simultaneous to each other in order to calculate thirteen logically possible tenses; by considering the relation of the speech time to the reference time as "past," "present," and "future" and the relation of the event time to the reference time as "anterior," "simple," and "posterior," these possibilities are reduced to nine "fundamental forms". For example, the forms of the past are shown in (4); the comma indicates that the points are contemporaneous, while the underline indicates they are not: (4) E R S E,R S R E S R S,E R S E Reichenbach's Name anterior past simple past posterior past Traditional Name past perfect simple past -Example She had done. She did. She would do. The a d d i t i o n of a temporal adverb or "time determination" to th i s system is "referred, not to the event, but to the reference point of the sentence" (Reichenbach 1947:294). In the example I had met him yesterday, 1 Reichenbach argues t h a t yesterday i s the r e f e r e n c e p o i n t which i s "the c a r r i e r of the time p o s i t i o n " : (5)a . E=meet S < _ _ _ [ | _ _ _ ] | > R=yesterday b. E=meet S < | [ ] |___> R=yesterday If the meeting occurred yesterday, the event point and the reference point coincide, as in (5a.); if the meeting occurred prior to yesterday, these two points are distinct, as in (5b.). Kamp and Reyle (1993:523) describe this work as being "the first to recognize what we have been calling the 'anaphoric dimension' of the tenses of natural language," where the second tense in a pair of sentences or a subordinate clause is interpreted "as 'anaphoric' to the time that the previous tense introduced" (Kamp and Reyle 1993:495). They credit Reichenbach's model with accounting for this temporal anaphora through the idea of the reference point. 1.4 The Notion of Reference Time The status of this third time is controversial, however. In this section, different conceptions of Reichenbach's reference time are outlined. Hornstein (1990) draws on Reichenbach's ideas to develop the principles of a theory of tense that account for a wider range of data; to do so, he argues explicitly for the equal 8 status of the three points, speech time, S, event time, E, and reference time, R. Thompson (1994a,b) assumes the existence of these three times and argues for their association with syntactic heads, based on the interpretation of adverbs. These two analyses are discussed in Section 1.4.1. In Section 1.4.2, the status of Reichenbach's reference point according to Dowty (1979), Stowell (to appear a), and Kamp and Reyle (1993) is discussed. Dowty and Stowell both consider this notion to be a tool for narrative tense sequencing; Stowell then adopts the term "reference time" for the external argument of the predicate tense, which is the speech time in a matrix clause, and suggests that narrative ordering is effected by "a null Topic time-denoting phrase" (Stowell to appear a:13). Kamp and Reyle (1993:594-595) point out that Reichenbach1s concept of reference time tried "to do too many things at once," encompassing both narrative sequencing and the interpretation of the perfect; they therefore distinguish a "reference point" for the narrative progression of time reference and "a temporal perspective point" as "the point from which the described event is viewed". Finally, Klein proposes that a possible interpretation of Reichenbach1s reference time is the notion of "topic time", which he defines as the "time span to which the speaker's claim on this occasion is confined" (Klein 1994:4; original italicized). The boundaries of topic time may be 9 specified by an adverbial. This is presented in Section 1.4.3. This thesis adopts Klein's conception of Reichenbach's reference time, and in particular his use of the name "topic time". The justification for naming reference time adverbs as topics will be presented in Chapter 3, Sections 3.1 and 3.2. 1.4.1 Hornstein and Thompson Hornstein's (1990) syntactic revision of Reichenbach's proposal seeks to define and constrain what is a possible tense in order to account for the interaction of tense and adverbs, which will be briefly discussed here, and to explain Sequence of Tense effects in English. To argue for the inclusion of the reference point, R, Hornstein points out that if tense is a relation between only two points, the speech time, S, and the event time, E, then there are four possible tenses. While this would be an advantage for learnability, it would fail to account for a six tense system like English^, with its perfect tenses. The past perfect and the future perfect both interpret the event time's relation to the speech time in terms of a third point: (6)a. John had left the office. b. John will have left the office. c. John had left the office at six o'clock. The past perfect in (6a.) has the temporal interpretation that John's leaving the office, E, is before both the speech 4 The six tenses of English are considered to be the past, present, future, past perfect, present perfect, and future perfect. 10 time, S, and before another moment, R, which is itself before S but after E, as in (7a.): (7)a. < E R S > b. < S E R > The future perfect in (6b.) is represented by (7b.), whereby John's leaving, E, is after the speech time, S, and before another point in time, R, which is itself after both S and E. The sentence in (6c.) is ambiguous between the reading that John left at six, where the adverb identifies E, and that John left before six, where the adverb specifies R; this is schematized in (8): (8)a. < E R S > I 6 p.m. b. < E R S > I 6 p.m. Hornstein argues that R gives the same effect as the semantic analysis which embeds operators, and the combination of S, E, and R correctly predicts what is a possible tense in English, which the operator analysis does not. Hornstein states that the existence of both an event point, E, and a reference point, R, explains why the number of temporal adverbs in English is restricted to two per clause: (9)a. A week from tomorrow, John will leave in a month. b.*From tomorrow, John, in a week, will leave in a month. c.*In a week, John, from tomorrow, will leave in a month. Sentence (9a.) is grammatical because the adverbial cluster a week from tomorrow is analyzed as a single constituent, which 11 occupies one adverbial position, while in a month occupies the other; the sentences in (9b.) and (9c.) are ungrammatical because there is no place in the syntax for a third adverb. Thus, the presence of R along with E and S correctly predicts the judgments in (9) since only E and R can be modified by adverbs, leaving the third adverb uninterpretable.5 Hornstein claims that adverbs may associate to R or E depending on their position in the sentence: sentence initial adverbs associate to the reference point, R, and sentence final adverbs associate either to the reference point, R, or or to the event point, E. He goes on to present several arguments for the analysis that R must mediate between between S and E, in order to constrain the total possible number of tenses available in human language. Thompson (1994:3) proposes that these two time points are associated with syntactic heads, subject to the "Tense Structure Mapping Condition" that constrains the mapping to a one-to-one relation. A simplified version of her structure is shown in (10): (10) TP T AspP • / \ Asp VP V E 5This argument extends to sentences with adverbial clauses, but these are not presented here. 12 She takes Hornstein's claim that speech time, S, and reference time, R, are ordered by tense morphemes, and locates the speech point, S, in the head of TP. Furthermore, she argues that the English perfect auxiliary have "orders the Event point with respect to the Reference point" and that the reference point, R, is therefore located in Asp, the head of AspP. Finally, the event time, E, is located in V, the head of VP. Adverbs that modify the R point are therefore adjoined to AspP, and those that modify E are adjoined to the VP. Therefore, for both Hornstein and Thompson, the existence of the reference point with status equal to the speech time and event time explains the tenses of English and their interaction with temporal adverbials. 1.4.2 Stowell, Dowty, and Kamp and Reyle Stowell (to appear,a) claims that Reichenbach' s reference time "provides a formalization, for the purposes of tense interpretation, of the traditional notion of 'point of view'. This has considerable intuitive appeal, but mainly in the case of the perfect tense constructions." In other words, he argues that reference time exists but that it does not have the same status as the event time and speech time arguments of the predicate Tense. He proposes to account for the English perfect as a "complex tense predicate" composed of the verb have and its complement Perfect Phrase, PrtP, which is headed by the past morphology of the participle. He then adopts the term "reference time" as a label for the 13 external argument of the tense predicate, which is "speech time" in matrix clauses and which in subordinate clauses is controlled by the event time of the matrix clause. Stowell also acknowledges the possible presence of Reichenbach's reference time in narrative contexts, but offers an alternative analysis. In a footnote, he suggests that temporal ordering of events in a narrative can be achieved if the event time of one sentence acts as the antecedent of a "topic time" phrase, which in turn binds the event time (or rather, its variable) of the following sentence. Dowty's criticism of Reichenbach1s approach is that the simple past and the present perfect have the same truth conditions under his analysis. He suggests that reference time may capture a pragmatic difference between the two, and concludes that the idea of a reference time "has its proper place in a theory of narration, i.e. of the way indefinitely identified times in a sequence of sentences in a narrative are understood to be ordered, perhaps with the aid of common information not included in the sentences themselves" (Dowty 1979:332) . Kamp and Reyle (1993) point out that in fact Reichenbach's proposal treats the time responsible for ordering sentences in narrative and the time that characterizes the perfect, located between the utterance time and the "described eventuality", as if they were the same. 14 They provide an example6 of "extended flashbacks" to show the necessity for distinguishing them: (11) Fred arrived at 10. He had got up at 5; he had taken a long shower (at x) , had got dressed (at y) , and had eaten a leisurely breakfast (at z) . He had left the house at 6:30. (p. 594) In this discourse, all the perfect constructions require both a "point of view" time which is set at the beginning, Fred's arrival at ten, and a time that moves with the narrative, to ensure for example that the event of Fred's getting dressed follows the event of his shower and precedes the event of his eating breakfast: (12) < I I---I---I I I l---> 5 x y z 6:30 1 0 n o w Kamp and Reyle c a l l the time needed for the n a r r a t i v e sequencing the "reference time" and the time necessary for i n t e rp re t i ng the perfect the "temporal perspect ive po in t" . They argue further for th i s l a t t e r notion with an example of the future in the past : (13) Mary got to the s ta t ion at 9:45. Her t r a i n would a r r ive at 10:05. In t h i s example, the temporal perspective point, R, i s Mary's a r r i va l at the s ta t ion , which is pr ior to the speech time, S, and to the time of the event, E, which i s a lso before the speech time: (14) < R E S > • I I I 9 : 4 5 1 0 : 0 5 now "The ho ld ing , i t a l i c s and p a r e n t h e t i c a l t imes a r e added here t o c l a r i f y the i s s u e s of the example 's temporal i n t e r p r e t a t i o n . 15 Thus, this analysis refines the central notion of Reichenbach's reference time, R, as the time from which the event is viewed, and accounts for an implicit notion in Reichenbach's reference point, the time in narrative that shifts from event to event. 1.4.3 Klein's Topic Time Reichenbach's reference time is interpreted by Klein (1994:4) as "topic time", TT, the time for which the claim or assertion of the utterance is made. Topic time is related by tense to the time of utterance and by aspect to the time of the situation or event described. It is specified by a phrasal or clausal temporal adverbial, as in (15), from Klein (1994:40) : (15)a.What did you notice when you entered the room? b.A man was lying on the floor. c.A woman was bending over him. d.She was taking a purse from his pocket. e.She turned to me. The topic time set by the adverbial clause in the first speaker's question, when you entered the room, is the same for each of the second speaker's answers in (15b.), (c.) and (d.). However, the time for which the claim she turned to me is made in (15e.) follows the time the second speaker entered the room; in other words, the topic time of sentences (15a.-d.) is TTl, while that in (15e.) is TT2. Note that if the question had asked when you were in the room, all the responses in (15) would fall within the same topic time. The topic time is therefore fixed in one of two ways: by "anaphoric maintenance," whereby the topic time of a previous 16 utterance remains the topic time of the present utterance but is not explicitly restated; or by "explicit specification, " when the topic time of the present utterance specifies a time span which is not the topic time previously specified. The topic time, TT, may contain or be contained in the time described by the event or state, but its boundaries are undefined unless specified by an adverbial. In English, the adverb which fixes the topic time is sentence initial or post subject, or, more generally, part of the topic structure of the sentence. 1.4.4 Summary The comparative interpretations of Reichenbach's notion of reference time presented in this section are summarized in the table in (16), which specifies the terminology used by each analysis and the times which each propose- as equal in status7: (16) Reichenbach Hornstein Thompson Dowty Stowell Kamp and Reyle Klein Speech Time speech time speech time reference time (=matrix speech time) utterance time time of utterance Reference Time reference time -temporal perspective point topic time Event Time event time event time event time described eventuality t ime o f situation Narrative Sequencing reference time reference time topic time reference time 'This thesis does not address the problem of the formalism needed to sequence times in narrative. 17 Hornstein and Thompson argued that the existence of Reichenbach's reference time explained the tenses attested in English and the interpretation of temporal adverbials. Klein's interpretation of this notion as topic time, the time for which the claim is made or "to which the assertion is confined" (Klein 1995:687), included the claim that this time is a topic of the sentence. These claims will be extended to temporal interpretation and adverbials in SqwXwu7mish in Chapter 3 of this thesis, which now adopts Klein's term, topic time, TT, for the third time interval. 1. 5 The Syntax of Tense and Aspect The equal status of the speech time, event time, and reference or topic time is important to the syntax of tense and aspect proposed by Demirdache and Uribe-Etxebarria (to appear a,b). Building on the proposals by Zagona (1990) and Stowell (1993, to appear a,b) that Tense is a predicate with two time denoting phrases as its arguments, Demirdache and Uribe-Etxebarria argue that grammatical Aspect is also such a predicate. They base their proposal on the claim (Klein 1994, 1995) that tense and aspect are semantically parallel because they both order two times8. Under this proposal, the external argument of the predicate Tense is a speech time argument UT-T and the internal argument is a topic time "Lexical aspect is not considered in Demirdache and Uribe-Etxebarria (to appear a,b); a short sketch of the interaction of lexical aspect and temporal interpretation in SqwXwu7mish is provided in the concluding chapter of this thesis. 18 argument, TOP-T9. This TOP-T is the external argument of the predicate Aspect and the internal argument is an event time argument, EV-T. Each of these arguments is a ZP or Zeit Phrase, adapting Stowell's (1993, to appear a,b) proposal. The arguments are ordered by the prepositional heads of the temporal predicates: Past Tense and Perfect Aspect, which mean "after"; Future Tense and Prospective Aspect, which mean "before"; and Present Tense and Progressive Aspect, which mean "within". The representation of this approach is shown in (17): (17) TP / \ U T - T T ' / \ T ASP-P / \ T O P - T ASP' / \ ASP VP / \ EV-T V V The external temporal argument of the predicate Tense, the speech time, is generated in the specifier of T; the external argument of Aspect, the topic time, is generated in the specifier of Asp; and the event time argument, the temporal argument of the verb, is generated in the specifier of the VP. 9While Demirdache and Uribe-Etxebarria label this argument AST-T or assertion time, following Klein (1995), this thesis follows Klein (1994) and adopts the label topic time, TOP-T, as discussed in the previous section. Both labels refer to the same thing, that is, the time for which the claim or assertion is made. 19 In order to accomodate the thematic subject of the verb, which is generated in the specifier of VP and raised to the specifier of TP for Case assignment or feature checking, Demirdache and Uribe-Etxebarria (to appear b:9) assume that T and Asp play "dual roles". Thus, both are lexical heads that "have argument structure that they project into the syntax", and functional heads that "enter (...) into feature checking relations with arguments that have been displaced to their checking domain" (to appear b:9-10). This dual function of syntactic heads is achieved with two specifier positions, one for each function (Chomsky 1995 and references therein) . Although for Demirdache and Uribe-Etxebarria, the V does not have a functional role, the event time argument and the thematic subject of the verb are both accomodated through this use of multiple specifiers. 1. 6 The Syntax of Temporal Adverbs Temporal expressions in language can be adverbs and adverbial phrases, nouns and noun phrases, prepositional phrases, and subordinate clauses (Binnick 1991). English adverbs include always, eventually, henceforth, now, soon, and then; they can appear in phrases such as just now and not too soon. Temporal nouns such as Sunday, today, tomorrow and yesterday can also be used in noun phrases such as the day before yesterday and last Sunday, along with next year, tomorrow morning, and every day. Temporal prepositional phrases may be formed with temporal nouns, such as before 20 tomorrow, until Sunday and during 1994, or with nouns denoting an event, such as after the party and before the deadline. Finally, temporal clauses include after the party is over, before the library closes, and when Catherine arrives. These time adverbials can also form complex "clusters" in English, such as for a few minutes before dinner this evening. English temporal adverbials are attached to either the sentence or the verb phrase (Thompson 1994a,b, Hitzeman 1995). When they are attached to the sentence, they may be sentence-initial as in (18) or sentence final as in (20). When they are contained within the verb phrase, adverbs may appear before the verb, as in (19a.), after the verb or auxiliary, as in (19b.), or after the verb and its complements, as in (20): (18) Last night, the moon was full. (19)a.Anna sometimes takes tea. b.She was recently criticized for her behaviour. (20) Alexander had written a poem about it yesterday. The sentence final adverb in (20) is therefore ambiguous between attachment to the sentence or to the verb phrase. The relatively unrestricted attachment site of the adverb in English has been assumed to be an argument for its status as an adjunct10. This will be pursued in terms of SqwXwu7mish adverbs in Chapter 4. 10Cinque (in progress) argues against the analysis of adverbs as adjuncts. He proposes that adverbial phrases are "the overt manifestation of (the specifiers of) functional projections." 21 1.7 The Semantics of Temporal Adverbs To this point, adverbs have been identified simply as a class; there is a great deal of variety within this class, however. For the purposes of this thesis, a distinction of three semantic types of temporal adverbs (Kamp and Reyle 1993, Binnick 1991 and others) is assumed: i) locating adverbs, which include deictics, such as yesterday, tomorrow, now, anaphoric adverbs such as a week ago, context-dependent calendar names such as on Sunday, last Sunday and on the preceding Sunday, and purely referential adverbs such as on September 16, 199711; ii) aspectual or temporal measure adverbials, which include in an hour and for an hour; and iii) quantificational adverbs or temporal quantifiers, such as always, twice, and never. 1.7.1 Locating Adverbs A purely referential adverb such as September 16, 1997 acts^ "as a proper name, which rigidly designates one particular date" (Kamp and Reyle 1993:614) across all possible worlds; it is not bound by the speech time. Deictics like yesterday are defined as expressions which refer to individual times or intervals of time (Dowty 1979) and which appear to involve "an existential quantifier over times" (Enc 1986:76). These "discourse-oriented" adverbs are temporally interpreted "relative to the moment of speech" H H . Davis (p.c.) points out that this range of temporal reference parallels the range of nominal referring expressions. Kamp and Reyle (1993) argue that context sensitivity in temporal reference is much more varied than in nominal reference. Partee (1973) describes the degree of parallelism as language specific. See Section 6.1. 22 (Hornstein 1990:30), in contrast to referential adverbs and to anaphoric adverbs, such as a week ago. These may be interpreted in relation either to the speech time or to a reference time such as yesterday. Calendar names like Sunday refer to more than one interval of time, the particular interval being determined by the context (Kamp and Reyle 1993); they cannot refer to the utterance time, and the day referred to will be the closest to the utterance time in the direction indicated by the tense. Locating adverbs, which are the most varied and interact most closely with the tense-aspect system (Klein 1994) , are the main concern of this research. 1.7.2 Aspectual Adverbs Kamp and Reyle (1993) state that aspectual or measure adverbs, such as in an hour and for an hour, appear to behave in the same way as locating adverbs but in fact function differently semantically. They argue that these adverbs refer to the duration of the event but do not locate it in terms of the speech time; they interact with both, the grammatical aspect of the sentence and the aspectual class of the predicate. The latter is illustrated in (21): (21)a.Peter wrote a letter in an hour/for an hour. b.Vanessa worked for an hour/*in an hour. c.Francisco arrived in an hour/*for an hour. Telic durative verbs are compatible with both in an hour and for an hour as in (21a.); atelic durative verbs can occur with for an hour only, as in (21b.); while punctual verbs can occur with in an hour only, as in (21c). 23 1.7.3 Quantificational Adverbs Quantificational or frequency adverbs locate events and states in time, but differ from locating adverbs "in that their discourse referents do not act as representatives of particular times, but as bound variables." (Kamp and Reyle 1993:612); in other words, they choose from a set of possible times rather than one particular time. For example, in (22), always does not refer to a single time but ranges over the set of springs in Eva's lifetime and picks every member in the set: (22) Eva always gathers cedar in spring. Klein (1994) further categorizes this type of adverb into definites, such as twice and ten times, and indefinites, such as always, often, and occasionally. NP's with every or most and PP's with after every or during most are also possible quantificational adverbs. 1 . 8 Outline of the Thesis The proposal to be defended in this thesis, that reference or topic time has equal status with event time and speech time (Reichenbach 1947, Hornstein 1990, Thompson 1994) and that SqwXwu7mish adverbs are the overt expression of a topic time argument of the predicate Aspect, is supported by their semantics and syntax. In Chapter 2, the grammar of SqwXwu7mish is surveyed, with an emphasis on temporal interpretation. Word order in main clauses is established with both nominal and pronominal 24 arguments, and subject and object pronominal paradigms are provided. The system of aspectual, temporal, and modal clitics is described; the function of each clitic is defined, and its interaction with other clitics in the system is outlined. The interaction of the determiner system with temporal adverbs is examined. Finally, complex sentences with the types of subordinate clauses found in the data in this thesis are presented. Chapter 3 examines the semantic evidence for the existence of reference or topic time. Adverbs are shown to specify the topic time when they are part of the topic structure of the sentence; in order for them to denote the event time, adverbs must be focused, that is, syntactically the main predicate. Furthermore, the interaction of adverbs with quantifiers shows that the topic time adverb must take semantic scope over the quantification. The existence of topic time is also supported by evidence from narrative; the topic time adverbial overlaps with part but not all of the interval denoted by a stative verb. Finally, the existence of the perfect in SqwXwu7mish is established, then the interaction of the perfect and the future with the aspectual adverb in two weeks is shown to support the claim that there are three time points. Given that adverbs denote the topic time, their representation in the syntax is examined in Chapter 4. Only one position is available to adverbs in SqwXwu7mish: more than one temporal adverb cannot occur in a clause; the topic 25 time adverb's preferred position is the right edge of the clause; and an adverb on the right of a subordinate clause cannot be construed with the main clause. These distribution facts argue against the traditional analysis of adverbs as adjuncts and follow from the claim that these adverbs are the reference or topic time, assuming a unique reference time (Partee 1985). This claim is further supported by cleft constructions that contrast left adjoined "functional" adverbs such as maybe and always, which are never main predicates, with locating adverbs and thematic arguments, which are frequently main predicates. The explanation proposed in this thesis for the topic time interpretation of SqwXwu7mish adverbs and for. their argument-like syntactic behaviour is that they are the topic time external argument of the predicate Aspect. This analysis is presented in Chapter 5; in order to carry out the analysis, modifications to the syntax of Tense and Aspect developed by Demirdache and Uribe-Etxebarria (to appear a,b) are proposed. The argument status of the adverbs allows an explanation of how SqwXwu7mish pronoun-verb constructions, apparently uninflected for tense, are interpreted temporally. Word order is derived by leftward movement of the verb and its arguments, in accordance with Kayne (1994); the adverbial temporal argument does not move because it has no features to be checked (Thompson 1994). The construal of adverbs in complex sentences is shown to distinguish between clausal and phrasal nominalized clauses in SqwXwu7mish. Finally, the 26 interaction of topic time adverbs with quantification argues for a non-structural analysis of scope in Salish (Demirdache and Matthewson 1997). This analysis has implications for the information strucutre of the Salish sentence and for the status of adverbs. Furthermore, this analysis of SqwXwu7mish adverbs requires many assumptions about the nature of tense and the structure of the clause in Salish. More work needs to be done on the tense-aspect system in SqwXwu7mish, on the interaction of lexical aspect with temporal interpretation, and on the temporal effects of different determiners on the interpretation of adverbs in particular and nouns in general. These implications and issues are the focus of the concluding chapter. 27 2 . SqwXwu7mish Morphology and S y n t a x SqwXwu7mish i s a S a l i s h l anguage spoken i n t h e B u r r a r d I n l e t and Howe Sound a r e a a round Vancouver , B r i t i s h Columbia . U s i n g c o m p a r a t i v e v o c a b u l a r i e s , Swadesh (1950) c l a s s i f i e d t h i s l a n g u a g e a s a member of t h e Sou th G e o r g i a Branch of Coas t S a l i s h , a l o n g w i t h Halq 'emeylem, S t r a i t s , and Nooksack. The s k e t c h of t h e grammar of SqwXwu7mish i n t h i s s e c t i o n p r e s e n t s word o r d e r f a c t s , t h e t e m p o r a l and a s p e c t u a l c l i t i c s y s t e m , t h e d e t e r m i n e r sys t em, p h r a s a l t e m p o r a l a d v e r b i a l s , and complex s e n t e n c e s . These t o p i c s a r e p r e s e n t e d i n t e r m s of t e m p o r a l i n t r e p r e t a t i o n i n t h e l a n g u a g e . 2 . 1 Word Order i n t h e Main Clause S a l i s h l a n g u a g e s a r e p r e d o m i n a n t l y p r e d i c a t e i n i t i a l . I n SqwXwu7mish main c l a u s e s , t h e p r e d i c a t e , be i t a v e r b , noun o r a d j e c t i v e , i s n o r m a l l y t h e f i r s t e l e m e n t i n t h e s e n t e n c e when t h e s u b j e c t and o b j e c t a r e n o m i n a l s and t h e t ime of t h e u t t e r a n c e i s t h e p a s t o r p r e s e n t 1 2 : ( 1 ) a . s w i 7 qa t a sm6q'wa7 man DET c r a n e Crane was a man K u i p e r s 1967:169 -^Abbrev ia t ions used in to g los s the da ta in SqwXwu7mish: l s g . s u b j = f i r s t person s i n g u l a r sub jec t agreement; l p l . s u b j = f i r s t person p l u r a l sub j ec t agreement ; l p o s s = f i r s t person p o s s e s s i v e ; 3 s u b j = t h i r d person s u b j e c t ; 3 p o s s = t h i r d pe r son p o s s e s s i v e agreement ; 3 p l = t h i r d p e r s o n p l u r a l ; c a u s = c a u s a t i v e ; DET=determiner; DEM=demonstrative; d i r = d i r e c t i o n marker ; DR=durative; FOC=focus marker; FUT=future; I N T = i n t e r r o g a t i v e ; i n t r a n s = i n t r a n s i t i v i z e r ; I R R = i r r e a l i s ; LOC=locat ive ; NEG=negation; nom=nominal izer ; o b l = o b l i q u e ; PT=past; PR=present ; REC=rec ip roca l ; r e d = r e d u p l i c a t e d morpheme; R L = r e a l i s ; R E L = r e l a t i v e ; TD=temporal d e i c i t i c ; t r a n s = t r a n s i t i v i z e r ; WH=wh-word. 28 (I)b.ha71h ta lam7 good DET house The house is good Kuipers 1967:169 c.(na) k'wach-nexw-as kwelhi slhanay' ta miXalh (RL) see-trans-3subj DET lady DET bear the lady saw the bear EL 27-3-97 23 As the subject may precede or follow the object in transitive sentences, word order is VSO or VOS. Furthermore, the speaker may emphasize the third person subject by moving it in front of the verb for the order SVO: (2) kwelhi slhanay' k'wach-nexw-as ta miXalh DET lady see-trans-3subj DET bear the lady saw the bear EL 27-3-97 24 When there is a single overt nominal with a transitive verb, it is interpreted as the object; this is the "One Nominal Interpretation" (Gerdts 1988}: (3) na ch'£m7t-as ta sqwmay7 RL bite-3subj DET dog He bit the dog. *The dog bit him. Kuipers 1967:172 For a past or present interpretation, pronominal elements for the first and second person precede the predicate: (4)a.chen swi7qa lsg.subj man I am a man Kuipers 1967:171 b.chen q'way lsg.subj hungry I'm hungry YJ 12-12-96 c.chen ilhen kwi chel'aqlh lsg.subj eat DET yesterday I ate yesterday YJ 13-3-9 6 29 (4)d.chen tl'iq-s ti siten ta sXalhnat lsg.subj arrive-caus DET basket DET Sunday I brought the basket on (any) Sunday EL 17-4-97; EL 19-6-97 01 When these elements follow the predicate, they are interpreted as future with eventive verbs and adjectives, and as present with stative verbs and nouns: (5)a.tayeq' chen mad lsg.subj I'm gonna get mad EL 8-5-96 b.slhanay' chen woman lsg.subj I'm a woman EL 8-5-96 c.lhq'i7-s chen know-caus lsg.subj I know it (already) EL 8-5-96 d.ts'its'ap' chen work lsg.subj I'm going to work EL 7-6-96 When the second person pronominal occurs after the predicate, it gets an imperative interpretation: (6) ch'aw-at-s chexw help-TR-lsg 2sg.subj Help me! Jacobs 1992:17 The indicative subject, object, and possessive agreement paradigms are summarized in the table in (7): (7) P O S S E S S I V E SUBJECT OBJECT 1st 2nd 3rd sg. n-7a--s pi. -chet 7a--yap -s-wit sg. chen chexw -0, -as pi. chet chap, chayap -0-wit, -as-wit sq. -s, -msh -umi _013 Pi. -umulh -umiyap -wit 1 3Note t h a t ze ro agreement has not been marked on t h e d a t a in t h i s t h e s i s . 30 Along with the pronominal elements, certain "functional" adverbs may precede the predicate: chiyalh 'soon', way'ti 'maybe' and lhiq' 'always'. These do not constitute a semantic class, but may be classified as auxiliaries: (8)a.chiyalh chen ts'its'ap' soon lsg.subj work I' m gonna work soon EL 1-5-97 18 b.way'ti chen ts'its'ap1 maybe lsg.subj work I'm not sure/I think I will work EL 1-5-97 19 c.lhiq' chen wa ts'its'ap' always lsg.subj DR work I'm always working (volunteered form) EL 1-5-97 23 The distribution of these adverbs will be examined in Section 4.4. 2 . 2 Tense and Aspect Marking The system of c l i t i c s in SqwXwu7mish which serve temporal or aspectual functions usually a t tach to the l e f t of the main p red ica t e , e spec ia l ly in the f i r s t and second person. These are summarized in (9), approximately following Jacobs (1992); each i s given a gloss and an approximate meaning14: l ^The d i s t r i b u t i o n and i n t e r p r e t a t i o n of t h e s e c l i t i c s i s on ly s u p e r f i c i a l l y unders tood and needs fu r the r r e s e a r c h . 31 (9) na RL realis q IRR irrealis wa DR durative aspect kw TD temporal deictic 7i PR present tense t PT past tense eq' FUT future tense 2.2.1 Realis and Irrealis The clitic na performs a similar function for the third person, but it is optional; it can combine with ch-pronominals for an emphatic sense, like that of English do: (10)a.chen lhchiws lsg.subj tired I am tired Kuipers 1967:157 b.na lhchiws RL tired He is tired Kuipers 1967:157 c.na chen tsut RL lsg.subj say I did say it Kuipers 1967:157 The clitic q "marks a predicate as irreal in the widest sense of the word" (Kuipers 1967:160); it can have the meaning "when" or "if", as in (11a.) and (b.), and is always used with negation, as in (lie): (ll)a.kwayl q-ts'its'ap'-as tomorrow IRR-work-3subj He will work tomorrow, lit. Tomorrow is when he will work Kuipers 1967:189 b.ha71h n-sqwalwen q-lhemxw-as good lposs-heart IRR-rain-3subj I would be glad if it rained Kuipers 1967:190 32 (ll)c.haw-q tin&7-an NEG-IRR from.there-lsg.subj I'm not from there Kuipers 1967:190 As q is always in second position in main clauses, in the past conditional sentences in (12), the q follows na in the main clause and precedes it in the "if" clause: (12) 7u-q-na ch'awat-way'-wit, na-q-huy-nexw-as-wit INT-IRR-RL help-REC-3pl RL-IRR-finish-trans-3sg-3pl If they had helped each other, they would have finished it. Kuipers 1967:191 2.2.2 The Durative and the Deictic The durative marker wa15 can indicate an action that is habitual, ongoing or iterated; it is normally translated by the English progressive: (13) chen wa shuk'w-um lsg.subj DR bathe-intrans I am taking a bath/ I am in the habit of bathing Kuipers 1967:159 It appears between the pronominal and the verb; Kuipers lists a limited number of examples, usually imperatives, where the durative precedes the pronominal. Otherwise, speakers judge wa before the pronoun to be ungrammatical. The realis clitic na may be followed by the pronominal chen and the durative wa: (14)a.na chen wa i l hen RL l s g . s u b j DR eat I am e a t i n g YJ 13-3-96 b .na wa i l hen RL DR ea t he i s ea t ing ( r igh t now) EL 6-3-96 1!:>There i s no f u l l verb form of wa in SqwXwu7mish, u n l i k e languages in Northern I n t e r i o r S a l i s h (P. Jacobs, p . c ) . 33 The "temporal d e i c t i c " kw i s often t r a n s l a t e d as "a l ready" or "now", and i s used for the SqwXwu7mish equivalent of the per fec t , which wi l l be demonstrated in Section 3.4: (15)a.chen kw lhexwlhsha7-awanexw lsg.subj TD thirty-years I was thirty years old/ I had reached the age of thirty Kuipers 1967:159 b.na-kw tl'iq ta Peter RL-TD arrive DET Peter Peter's here now, Peter has arrived EL 13-8-97 This clitic follows either the first or second person pronominal element or, in the third person, the realis clitic na. It precedes the main verb, the local-directional clitics nam' and mil and, in a single example, the durative wa. 2.2.3 Past, Present and Future The three clitics 7i, t, eg' are glossed as tense markers, for lack of a better understanding of their distribution and interpretation. The meaning of 7i is "proximity to the situation of speech" (Kuipers 1967:158): it can mean "at this moment" as in (16a.); it can refer to "anticipated events", as in (16b.) and ( c ) , in which case it is often used with "chiyalh, soon" (Kuipers 1967:158); or it may have an immediate past sense with men 'just', as in (16d.): (16)a.7i chen ilhen PR lsg.subj eat I'm gonna eat right now DW 20-3-96 b.chiyalh 7i chen ilhen soon PR lsg.subj eat soon I'm gonna eat DW 20-3-9 6 34 (16) c.chiycilh chen 7i ilhen pretty soon lsg.subj PR eat pretty soon I'm going to eat EL 6-3-96 d.men-yalh n-s-7i tl'iq just-finally lposs-nom-PR arrive I've just arrived, lit. Just" recently (is) my arriving Kuipers 1967:158 The data in (16) show that 7i can precede or follow ch-pronominals; if wa is added, 7i must precede the durative. In negative sentences, it can be preceded by g. One of the problems with translating 7i as the present is that it can also cooccurs with t16; £ usually receives a simple past or habitual past interpretation, attaching to the first person agreement or, in the third person, to the realis clitic na: (17)a.chen-t mi-s lsg.subj-PT come-caus I brought it YJ 13-3-96 b.chen-t ts'its'ap' 1sg.subj-PT work I used to work YJ 3-4-96 c.na7-t ilhen RL-PT eat She/he's finished eating DW 20-3-96 The "past" clitic precedes wa and follows na and chen: (18)a .chen- t wa i lhen l sg . sub j -PT DR eat I was e a t i n g YJ 13-3-96 1 6 P . Jacobs ( p . c . ) has e l i c i t e d grammatical sentences such a s : 7 i chen- t wa esqwuy PR lsg-PT DR s i ck I was s i ck (and I s t i l l am.) from examples found in Hi l l -Tout (1900). 35 (18)b.na chen-t wa ts'its'ap' RL lsg.subj-PT DR work I used to work YJ 3-4-96 c.na7-t wa ilhen RL-PT DR eat He was eating EL 6-3-96 Neither 7i nor t cooccurs with eg', which gets a distant future interpretation. It follows the pronominal element, the clitic mil 'come', or the main verb itself: (19)a.chen-eq' ilhen lsg.subj-FUT eat I'm going to eat YJ 6-6-96 b.ilhen chen-eq' eat lsg.subj-FUT I'm gonna eat EL 7-6-96 c.mi-chen-eq' t'uk'w come-lsg.subj-FUT home I'll come home Kuipers 1967:207 d.mi7-eq" t'uk'w come-FUT home He'11 come home Kuipers 1967:207 e.ilhen-eq' eat-FUT (She, he's) gonna eat DW 20-3-96 This contrasts with the near future/intention meaning of the verb-pronoun examples in (5). 2.2.4 Local Directional The "local-directional" clitics are nam', which is used for a near future with motion away from the place the utterance is made, and mi (7), which means "become" or describes a motion towards the speaker. These clitics are in 36 complementary distribution with the durative clitic via, but they do occur with na in the third person: (20)a.nam' chen wi7xw-em go lsg.subj drop-intrans I'm gonna fall YJ 3-4-96 b.chen nam' wi7xw-em lsg.subj go drop-intrans I'm falling YJ 13-3-96 c.na nam' ilhen RL go eat He went to eat EL 6-3-96 (21)a.chen mi n^ch'-i lsg.subj come change-intrans My expression changed Kuipers 1967:162 b.mi nach'-i come change-intrans His expression changed Kuipers 1967:313 c.mi sh£m come low tide (He) emerge(s) from the water, come(s) to the surface Kuipers 1967:162 It must be noted that the full verb forms of these clitics do cooccur with the durative. 2.2.5 Summary The ordering and cooccurrence restrictions outlined above are summarized in the table in (22); it reads from left to right, with a check indicating that the clitic in that column can follow the clitic in that row, and an asterisk indicating that the clitic in that column must precede the clitic in that row. An asterisk and a check indicate conflicting data. The symbol "ch" shows the interaction of 37 the indicative subject pronoun with these clitics. An empty section indicates that the evidence exists but has not been worked out. (22) c h n a a wa kw 7 i t ea ' nam m i CH XXX V * * / V * V *H * V V NA * XXX V * V V * * Q V V XXX * * WA V V XXX V V V KW V V * XXX V 7 I V V * * XXX * * T V V • * V XXX * EQ ' V * / V V * * XXX * V NAM V V * V V XXX * M I V V * V V * XXX The large number of empty cells in this summary show the need for explicit elicitation of these orders and combinations. 2.3 Determiners and Temporal Adverbs Determiners in SqwXwu7mish encode spatial and temporal information; they introduce temporal adverbs17 and in part fix the interval referred to by the adverb. 1 7 Coas t S a l i s h languages p r e f e r DP adve rbs , wh i l e I n t e r i o r S a l i s h languages p r e f e r CP adverbs (H. Davis, p . c ) . 38 2.3.1 The SqwXwu7mish Determiner System Kuipers (1967) analyses SqwXwu7mish determiners as encoding distinctions between gender, proximity or distance, presence or absence; they may also be strong (demonstrative) or weak (non-demonstrative), definite or indefinite: (23) DEFINITE INDEFINITE NON-PRESENT WEAK WEAK STRONG masc fern PROXIMAL ta lha ti tsi DISTAL tay alhi kwa kwelhd kwetsi kwelhi kwi kwes This classification requires revision, however; Matthewson (1996) showed that determiners in Salish do not encode definiteness or specificity. P. Jacobs (p.c.) proposes an alternative system of classification, which contrasts visible and non-visible with invisible, and determiners (DET) with demonstratives (DEM): ( 2 4 ) POTENTIALLY V I S I B L E I N V I S I B L E VISIBLE NON-VISIBLE DET DEM masc fern masc fern Proximal ti tsi tiwa tsiwa Distal ta lha tay alhi kwa kwelha kwetsi kwelhi kwi kwes kwiya(wa) kwesawa The distinctions of visible, non-visible and invisible, and of proximal and distal, refer not only to space but also to time: the invisible determiner has a modal meaning, non-visible and distal visible determiners appear to be 39 inherently past, while proximal visible means close to the present. The details of such an analysis are currently under investigation. It is known, however, that changing the determiner on an NP can change the temporal interpretation of the utterance (Demirdache 1997 for St'at'imcets); as most temporal adverbs in SqwXwu7mish take determiners, its deictic contribution interacts with that of the adverbial in establishing the temporal interpretation of the sentence. 2.3.2 Today, Yesterday and Tomorrow Part of the meaning of deictic locating adverbs such as today, yesterday and tomorrow, is their relation to the time of utterance; thus, they always take the same determiners in SqwXwu7mish. The "invisible" determiner kwi occurs with the day that is just past, while the visible proximal ti appears with the day that contains now, the utterance time. (25)a.kwi chelaqlh "yesterday" DET yesterday b.ti stsi7s "today" DET today The day that is coming, tomorrow, is either the DP kwayl used on its own as a predicate with a subordinate clause, or is a conjunctive clause introduced by g: (26) q-kwayl-es IRR-day-3sg tomorrow, when it is the next day 2.3.3 Morning, Year Adverbials like natlh 'morning' and siyl&nem 'year' have no inherent relation to the utterance time so the determiner identifies which interval of time is being referred to. 40 Thus, ti identifies the morning closest to now, ta identifies a morning past, and kwa identifies specifically yesterday morning: (27) ti natlh this morning ta natlh that morning kwa natlh yesterday morning The effect of the determiner with siyeldnem 'year' is somewhat different: while ti identifies the year that contains now, kwi identifies the year prior to this one, and ta, along with hem'i 'coming', identifies the year to come: (28) ti siyelanem this year kwi siyelanem last year ta hemi siyelanem next year The visible proximal ti clearly identifies the interval closest to the utterance time, and kwi appears to locate the interval in the past. 2.3.4 Calendar Names Calendar names also take determiners. Again the proximal visible ti identifies the Sunday closest to now, but in this case it is an interval in the future. The distal and the non-visible determiners identify a Sunday past, while the invisible determiner kwi appears only with way'ti 'might, maybe', in contrast to the past effect in the examples above. (29) ti sxalhnat this Sunday ta sXalhnat that Sunday kwetsi sXalhnat any, a certain Sunday kwi sXalhnat a possible Sunday 41 The effects of determiners on temporal interpretation in SqwXwu7mish remain a topic for further research. 2.3.5 Quantificational Adverbs Quantificational adverbs or cardinal quantifiers do not, however, take a determiner since they never appear in argument position. The examples in (30) are composed of the number form for objects (which are separate from those for animals and persons) and a suffix meaning "times": (30)a.7an7us-alh "twice" two -times b. 7upen-alh "ten times" ten -times These adverbs appear only as main predicates. 2 . 4 Complex Sentences Complex sentences in SqwXwu7mish may be coordination structures, or they may consist of a predicate and a subordinate structure. Subordination in SqwXwu7mish includes conjunctive clauses and nominalized clauses, which contrast future and non-future, as well as zero relative clauses. 2.4.1 Coordination Two main clauses are "coordinated" with the clitic complex 7i-kw-na, which makes a "simple reference to time" with the meaning "and, then, now, when" (Kuipers 1967:212); P. Jacobs (p.c.) states that this would be better analyzed as 7i-kw 'and already' followed by the realis clitic na, and this analysis is adopted here. These constructions are 42 analyzed as coordination structures because of the main clause agreement chen following 7i-kw-na. (31)a.0 kwi chel'aqlh 7i-kw-na-chen tl'iq-s ti siten FOC DET yesterday and-TD-RL-lsg.subj arrive-caus DET basket Yesterday I brought the basket EL 1-5-97; EL 8-5-97 44 b.nilh kwi chel'aqlh 7i-kwe-na k'wach-nexw-as kwetsi miXalh FOC DET yesterday and-TD-RL see-trans-3subj DET bear She saw the bear yesterday EL 2-4-97 32 In (31a.), the adverb is an argument with a zero focus marker, whereas it occurs with the overt focus marker nilh in (31b.). These sentences function as cleft constructions, and are labelled by Kroeber (1991) as "and-fronting". 2.4.2 Conjunctive Clauses18 These clauses include those introduced by the "irrealis" complementizer g, which are future or non-future, and those introduced by lh, a sort of wh-complementizer. In future conjunctive clauses, the subject agreement is attached to the predicate, as in (32a.), while in non-future conjunctive clauses the agreement precedes the predicate and is attached to the irrealis marker, as in (32b.). The functions of these clauses include conditional and temporal adverbial uses, as well as complement of the negation predicate haw 'not',as in (32c.) and (d.): (32)a.chen ts'its'ap' [q-kwayl-es] lsg.subj work [IRR-tomorrow-3sg] I will work tomorrow (lit. when tomorrow it) YJ 13-3-96 1 B The c l a s s i f i c a t i o n of 2 h - c l a u s e s as c o n j u n c t i v e c l a u s e s i s t h e sugges t ion of P. Jacobs ( p . c ) . 43 (32)b.chen ta7aw'n [q-'axw huya7] lsg.subj think [lRR-2sg leave] I thought you left Jacobs 1992:30 chaw [q-tina7-an] not [iRR-be from-lsg.subj] I'm not from there Jacobs 1992:28 d.haw [q-an k'wach-nexw-0] not [IRR-lsg.subj see-trans-3obj] I d idn ' t see him Jacobs 1992:29 Clauses with lh and conjunc t ive agreement a r e "subordinate subject clauses, (occurring) almost exclusively with predicates referr ing to a quantity of t ime(s)" (Kuipers 1967:196) . As s t a t ed above, the lh i s a sor t of wh-complementizer, s imi lar to when, where and about which in English, but i t never corresponds to the subject or object of the clause i t in t roduces . In the example in (33), the l a t e r a l f r i c a t i v e lh has combined with the g l o t t a l stop of the agreement to produce the g lo t t a l i zed l a t e r a l a f f r i c a t e t l ' : (33) nilh kwi sXalhnat tl'-as nam' huya7 FOC DET Sunday WH-3subj go leave He went away on Sunday lit. it was Sunday when he went away EL 1-5-97 43 In addition, lh clauses with possessive agreement refer principally to clauses meaning "the reason why", "the place where", although these are not relevant for this thesis. 2.4.3 Nominalized Clauses Nominalized clauses are normally introduced by the determiner kwi, followed by the nominalizer procliticized to either the predicate or the first aspectual clitic (Jacobs 44 1992). The subject of these clauses is marked by possessive agreement. Past time is indicated by the "realis" clitic na for the first person singular and second person singular and plural, and by the attachment of the third person possessive agreement to the nominalizer, as in (34); future time is indicated by the absence of the clitic na in the first person singular and second person singular and plural, and by the attachment of the third person possessive agreement to the predicate of the nominalized clause, as in (35): (34)a.ha71h [kwi-n-s-na t^lnexw-an ti] good [DET-lposs-nom-RL find out-lsg.subj DET] I'm glad to have found this out, lit. It is good, my having found this out Kuipers 1967:186 b.chen lhq'i7-s [kwi-s-es-kw huya7-0] lsg.subj know-caus [DET-nom-3poss-TD leave-3sg] I knew that he had already left Jacobs 1992:41 c.kwi chelaqlh [s-es ts'its'ap' ta Peter] yesterday [nom-3poss work DET Peter] Yesterday Peter worked. EL 19-6-97 ( 3 5 ) a . e n - s t l ' i 7 [kwi-n-s nam' t ' uk 'w] l p o s s - d e s i r e [DET-lposs-nom go go-home] I want to go home, l i t . My d e s i r e i s t h a t I go home P. Jacobs 1992:4719 b.kwayl [kwi-s t s ' i t s ' a p ' - s t a Pe ter ] tomorrow [DET-nom work-3poss DET Peter] Pe te r w i l l be working tomorrow EL 19-6-97 These c l a u s e s f u n c t i o n as e i t h e r s u b j e c t s of nominal p r e d i c a t e s , ob jec t complements of verba l p r e d i c a t e s , c l a u s a l temporal a d v e r b i a l s , or purpose c l a u s e s . i y T h e g l o s s of t h i s example does not follow Jacobs , who a n a l y s e s t h e d e t e r m i n e r , f i r s t pe rson p o s s e s s i o n , and nomal izer as an i n f l e c t e d complementizer . 45 2.4.4 Zero Relative Clauses A direct object focused with a zero marker in the main predicate takes a zero relative clause as its complement: (36) 0 kwetsi siten [na tl'iq-s-t-an] FOC DET basket [RL arrive-caus-trans-lsg.subj] It was the basket I brought EL 8-5-97 05 This structure will be discussed again in Section 4.4. 2 . 5 Summary Temporal interpretation in SqwXwu7mish is effected by word order, temporal and aspectual clitics, determiners and temporal adverbials. This thesis concentrates on the role that temporal adverbs play in main clauses in determining the topic time of the utterance. 46 3 . The Interpretation of Adverbs The interpretation of adverbs in both English and SqwXwu7mish argues for the existence of a reference or topic time, in addition to the event and speech times, as will be shown in this chapter. Furthermore, the time denoted by the reference or topic time is specified by locating adverbs in the topic structure of the sentence, while adverbs in the focus structure denote the event time (de Swart, to appear). Section 3.1 shows that English preverbal adverbs are interpreted as the topic time, but postverbal adverbs are ambiguous between a topic time and an event time reading (Hornstein 1990, Hitzeman 1995, Thompson 1994a,b). In contrast, SqwXwu7mish postverbal adverbs have unambiguous topic time readings; for SqwXwu7mish adverbs to get an event time reading, they must be part of the focus structure, that is, the main predicate of the sentence. In Section 3.2, the interaction of adverbs with quantifiers is examined. When adverbs in both SqwXwu7mish and English get the topic time reading, they must take semantic scope over other quantificational elements in the sentence. Narrative evidence in Section 3.3 shows that the topic time reading of the adverb overlaps with part of the time specified by the stative predicate. The English perfect and its approximate equivalent in SqwXwu7mish are shown to exemplify Reichenbach's claim that tense is composed of three times in Section 3.4. Finally, in Section 3.5 the interpretation of the perfect and the simple 47 future with the aspectual adverb for two weeks (Hitzeman 1995, Thompson 1994) further supports the argument for the equal status of speech time, event time, and topic time. 3 . 1 Adverbs and Topic/Focus Structure In her examination of the relation between position and meaning, de Swart (to appear) points out that the examples in (1) "describe the same facts" but differ in terms of the topic/focus structure: (1) a. At six o'clock, Jane left. b. Jane left at six o'clock. Sentence (la.) is the preferred answer to the question in (2), although (lb.) is a possible answer: (2) What happened at six o'clock? Sentence (lb.) is the only possible answer to (3): (3) When did Jane leave? Thus, when the sentence initial adverb gets a topic time reading as in (la.), it is the topic introduced in the question in (2); the sentence final adverb may also be the topic with special intonation. When the sentence in (lb.) answers the question in (3), the topic is the information introduced in the question, Jane leave. The sentence final adverb is the focus of this sentence, and specifies the event time of the verb. In SqwXwu7mish, the sentence corresponding to the order in (l)b. above answers the question in (2), What happened. yesterday?; deictic adverbs cannot simply be preposed in SqwXwu7mish, as will be shown in the next chapter. 48 (4)a.na chanem kwi chel'aqlh? RL do DET yesterday What happened yesterday? EL 2-4-97 28 b.na k1wach-nexw-as kwelhi slhanay' kwetsi miXalh RL see-trans-3subj DET lady DET bear the lady saw the bear EL 2-4-97 29 c.na k1wach-nexw-as kwetsi miXalh alhi slhanay' kwi chel'aqlh RL see-trans-3subj DET bear DET lady DET yesterday this lady (right here) saw the bear yesterday EL 2-4-97 30 Therefore, the postverbal adverb in SqwXwu7mish gets the topic time reading unambiguously. The answer to the "when" question is a sentence with nilh, a focus marker "used to refer to a thing, time, place or fact on which the addressee has already focused his attention, or which he can readily identify as a result of situational or contextual factors" (Kuipers 1967:144). This cleft or focus construction is shown in (5)b: (5)a.na 7encha 7i-kw-na k'wach-nexw-as kwetsi miXalh tsi slhanay' RL where and-TD-RL see-trans-3subj DET bear DET lady When did this lady right next to me see the bear ? EL 2-4-97 31 b. nilh kwi chel'aqlh 7i-kw-na k'wach-nexw-as kwetsi miXalh FOC DET yesterday and-TD-RL see-trans-3subj DET bear She saw the bear yesterday lit., It was yesterday that she saw the bear EL 2-4-97 32 The event time reading of this adverb is due to its focused position. The SqwXwu7mish data therefore parallels the English data semantically but not structurally: the adverb on the 49 right periphery can only be the topic time, whereas the sentence intial adverb must be focused syntactically as the main predicate to specify the event time. 3.2 Adverb-Quantifier Interaction In order to examine the semantics of scope relations, de Swart (to appear) presents English data in which existentially quantified time adverbs interact with other quantifiers. She argues that when the adverb gets the topic time reading, it must take wide scope over other quantifiers, and when the adverb identifies the event time, it must be within the scope of other quantifiers. This generalization holds for SqwXwu7mish, where the absence of ambiguity between topic time and event time readings of temporal adverbs clarifies these semantic scope relations. 3.2.1 Adverbs and Quantified Subjects The examples in (6) show the interaction of a universally quantified subject (V) with an existentially quantifed (3)temporal adverb in English. The sentence in (6a.) can only mean, "on the same Sunday, all the students went hiking either as a group or individually, and there was either one or many events of hiking"; the adverb gets the topic time reading and must take scope over the quantifier OV). (6)a.0n a beautiful Sunday in spring, every student went walking in the hills. b. Every student went walking in the hills on a beautiful Sunday in spring. 50 In (6)b, however, the reading of the adverb is ambiguous: the adverb can take scope over the universal quantifier to give the topic time reading (3V) obtained for (6a.); or the universal quantifier can take scope over the adverb to get an event time reading (V3). The event time reading is either "distributive", where "each student goes hiking individually on a Sunday but not necessarily the same Sunday, " or "collective", in which case there is "one group event of going hiking on one certain Sunday." The collective event time reading and the topic time, reading are therefore indistinguishable on the surface, but the existential quantifier of the adverb and the universal quantifier of the subject have different scope relations. . As seen in Section 3.1, syntactically focused adverbs in SqwXwu7mish receive an event time reading; the adverb in (7) has an event reading which may be distributive or collective, and is therefore in the scope of the universal quantifier: (7) na-malh-eq'sXalhnat 7i nam' 7i-7imash 7i7Xw ta sta7exwlh RL-well-FUT Sunday PR go redup-walk all DET children (volunteered form) It will be Sunday when they (all the children) go for a walk. EL 13-8-97 Sentence final SqwXwu7mish adverbs are not ambiguous. The only meaning available in (8) is that "all the children go on the same Sunday, " which is the topic time reading of the adverb where it takes wide scope over the universal quantifier: 51 (8)a.7i7Xw ta sta7exwlh nam' 7i-7imash na7 ta sXalhnat all DET children go red-walk RL DET Sunday All the children/every child went walking on Sunday ie they all go on the same Sunday EL 2-4-97 16 b.7i7Xw ta sta7exwlh nam' 7i-7imash na7 t kwetsi sXalhnat all DET children go red-walk RL obi DET Sunday All the children went for a walk on a Sunday (they all go on the same Sunday) EL 10-4-97 31 A sentence final topic time adverb cannot get a distributive reading. In the examples in (9), the speaker uses again and different to differentiate Sundays and children, but in (9a.) "the same group of children go on a (single) different Sunday" and in (9b.) "a different group of children go on a (single) different Sunday": (9)a.s-es-men qiy'at na nam' 7i-7imash ta 'Sta7exwlh na7 t kwetsi sXalhnat nom-3poss-just again RL go red-walk DET children RL obi DET Sunday And then they went again on a (different) Sunday EL 10-4-97 . 32 b.s-es-men qiy'at na nam' ta nach' sta7exwlh t kwetsi nach'sXalhnat nom-3poss-just again RL go DET different children obi DET different Sunday So different children went for a walk on a different Sunday EL 10-4-97 33 Thus, the in terac t ion of universally quantified subjects with e x i s t e n t i a l l y quantif ied temporal adverbials shows tha t the top ic time reading i s obtained with the sentence f i na l adverb; the event time reading, e i t h e r c o l l e c t i v e or d i s t r i b u t i v e , i s obta ined only when the adverb i s syntac t ica l ly the main predicate . 52 3.2.2 Adverbs and Only De Swart (to appear:9) adopts the definition of only as a quantifier which has as its domain "the set of contextually relevant alternatives"20: (10)a.On Sunday morning, Julia only [goes to church]. b.Julia only [goes to church] [on Sunday morning]. In (10a.), only does not take scope or quantify over on Sunday morning. In this sentence, only quantifies over the set of properties of individuals, one of which is "going to church"; this gives the reading that "the only thing Julia does on Sunday is go to church, " which is the topic time reading of the adverb. In (10b.), the quantifier takes scope and quantifies over goes to church and on Sunday morning. Therefore, this sentence can have either the same interpretation as (10a.) or the interpretation where only quantifies over the set of properties of times, one of which is the focused adverbial Sunday morning. The latter gives the reading that "the only day she goes to church is Sunday, " which is the event time reading of the adverb. The closest equivalent to only in SqwXwu7mish is men 'just', which Kuipers (1967:163) labels "determinative" or "limitative". This clitic appears just before the first stressed element in the sentence or separated from it by the clitic wa. In the example in (11), the adverb ta sXalhnat 'on Sunday' gets the topic time reading, as in (10)a. above: de Swart attributes this definition to Rooth (1992). 53 (ll)a.na wa nam' ta Peter ta k'wemaylh-aw'txw ta sXalhnat RL DR go DET Peter DET pray-house DET Sunday Peter goes to church on Sunday EL 2-4-97 01 b.men huy kwi-s-es wa nam' ta k'wemaylh-aw'txw ta sXalhnat just finish DET-nom-3poss DR go DET pray-house DET Sunday Peter goes to church on Sunday (he doesn't do anything else that day) EL 14-8-97 Therefore, the quantifier men 'only' is taking scope over "go to church" but cannot be taking scope over the topic time adverb. The event time reading occurs when the adverb is syntactically focused as a main predicate with men huy: (12)a.men huy ta sXalhnat just finish DET Sunday He just goes on Sunday EL 2-4-97, 03 b.men huy ta sXalhnat tl'-as nam' just finish DET Sunday REL-3subj go It's only Sunday that/when he goes ie.the only day he goes to church is Sunday EL 2-4-97, EL 14-8-97 05 To get this event time reading, the quantifier must take scope over both the adverb and "go to church," as in (10b.). These examples show that in SqwXwu7mish the adverb at the right edge is unambiguously the topic time, and is not in the scope of the quantifier; furthermore, the adverb must be syntactically focused to obtain the event time reading. 3.2.3 Cardinality Quantifiers The generalizations observed with quantified subjects and only parallel the interaction of locating adverbs such as on Sunday with cardinal quantifiers like twice and ten times. 54 In SqwXwu7mish, when the cardinal adverb is the main predicate and the temporal adverb is at the right edge of the clause, the temporal adverb gets a topic time reading as expected, as in (13c): (13)a.na ts'its'ap ta Peter na7 t kwetsi sXalhnat RL work DET Peter RL obi DET Sunday-Peter did work on a Sunday EL 2-4-97 11 b.na 7upen-alh s-es (wa) ts'its'ap1 ta Peter RL ten-times nom-3poss (DR) work DET Peter Peter worked/was working ten times lit. It was ten times that Peter worked/was working EL 10-4-97 18 c.na 7upen-alh s-es ts'its'ap' ta Peter (t) kwetsi sXalhnat RL ten-times nom-3poss work DET Peter (obi) DET Sunday Peter worked ten times/for ten hours on a Sunday EL 10-4-97 19 As with quantified subjects and only, the cardinal quantifer is not quantifying or taking scope over the right edge topic time adverb. This gives the topic time adverb a collective meaning. For the event time reading of the temporal adverb, it must be the main predicate ten Sundays, as in (14): (14) na 7upen sXalhnat kwi-s-es ts'its'ap' ta Peter RL ten Sunday DET-nom-3poss work DET Peter Peter worked on ten Sundays/Ten Sundays Peter worked EL 10-4-97 22 The quantifier in this case must take wide scope over the adverb, which gets only the distributive event time reading. Furthermore, unlike locating temporal adverbs, cardinal quantifiers can only be the focus of the sentence as the main predicate, and can only identify the event time; they cannot occupy the right edge of the sentence as topic time, as shown in the examples in (15): 55 (15)a.*na ts'its'ap' ta Peter na7 t 7an7us-alh t kwetsi sXalhnat RL work DET Peter RL obi two-times obi DET Sunday EL 2-4-97 12 b.*na ts'its'ap' ta Peter na7 t kwetsi sXalhnat 7an7usalh RL work DET Peter RL obi DET Sunday twice EL 2-4-97 13 c.*na ts'its'ap' t kwetsi sXalhnat ta Peter 7upen-alh RL work obi DET Sunday DET Peter ten-times EL 10-4-97 30 d.*na ts'its'ap' na 7upen sXalhnat ta Peter RL work RL ten Sunday DET Peter Conversely, the cardinal quantifier, like other temporal adverbs, cannot cooccur in a clause with another temporal adverb. If another temporal adverb such as kwi siyeldnem 'last year' is added, however, it occurs in the subordinate clause and t kwetsi sXalhnat 'on Sunday' appears in the main clause with the cardinal quantifier predicate. As expected, the locating adverb in this position gets the topic time reading, as indicated by the speaker's emphasis on "just that one Sunday": (16)a.na qeX-alh s-es wa ts'its'ap' ta Peter kwi siyel'anem RL many-times nom-3poss DR work DET Peter DET year Peter was working a lot last year context for next record EL 10-4-97 26 b.na 7an7us-alh t kwetsi sXalhnat s-es ts'its'ap' kwa Peter kwi siy'elanem RL two-times obi DET Sunday nom-3poss work DET Peter DET year Peter worked twice on a Sunday last year EL 24-4-97 11 56 The topic time reading of the adverb in (16) shows once more that the existentially quantified topic time adverb takes scope over other quantifiers in the sentence. 3.2.4 Summary The interaction of temporal adverbs with quantified subjects such as all the children, quantificational adverbs such as only, and cardinal quantifiers such as ten times, shows that the right edge adverb with the topic time reading has wide scope over these other quantifiers. In order for other quantifiers to take scope over the existentially quantified adverb, the adverb must be focused, that is, the main predicate, where it has an event time reading. Furthermore, the topic time adverb gets a collective reading with universal and cardinal quantification, while the event time adverb gets a distributive reading in these cases. 3 . 3 Topic Time in Narrative As noted in the introductory chapter, narrative provides examples of the topic time as it is maintained from sentence to sentence. A stative sentence "describes a condition prevalent at the time indicated by the reference time given in the preceding context, but it need not do so when the sentence contains its own adverb of temporal location." (de Swart to appear:2). This is illustrated in the following example from de Swart, where the event of the wife's dying occurs after 57 the time of her going to bed and before the topic time specified by the next morning: (17) The doctor came home and found his wife waiting for him. They had a drink and went to bed. The next morning, she was dead. In fact, the adverbial overlaps with part but not all of the state of being dead, in keeping with its interpretation as the reference or topic time. A similar narrative in SqwXwu7mish is illustrated in (18) : (18)a .wa nexwt i7 kwiya miXalh na7 t kwa 7 a t s ' q DR come.by. DET b e a r RL ob i DET o u t s i d e The b e a r was hang ing around o u t s i d e EL 24-4-97 03 b . c h e t nam' 7 e X i t s kwi txw-na7nat l p l . s u b j go l i e down DET d i r - n i g h t We went t o bed t h a t n i g h t EL 24-4-97 04 I n ( 1 9 a . ) , t h e main p r e d i c a t e i s s t a t i v e and t h e s e n t e n c e f i n a l a d v e r b g e t s a t o p i c t ime r e a d i n g c o n s i s t e n t w i t h i t s p o s i t i o n i n t h e t o p i c s t r u c t u r e ; i n (19b . ) t h e v e r b g e t s an e v e n t i v e i n t e r p r e t a t i o n and t h e main p r e d i c a t e a d v e r b g e t s an e v e n t t i m e r e a d i n g c o n s i s t e n t w i t h i t s l o c a t i o n i n t h e focus of t h e s e n t e n c e 2 1 : ( 1 9 ) a . h a w - q t a miXalh (7 i ) t i n a t l h NEG-IRR DET b e a r (PR) DET morning The b e a r was gone t h i s morning EL 13-8-97 b . 0 t i n a t l h 7 i -kw-na haw-q t a miXalh FOC DET morning and-TD-RL NEG-IRR DET b e a r The b e a r d i e d t h i s morning EL 24-4-97 09 T h i s morning t h e b e a r d i ed /*was dead EL 19-6-97 21The t r a n s l a t i o n s of these examples a re those provided by the speaker . The p r e d i c a t e haw-q i s unusual , and fu r the r t e s t i n g with o the r s t a t i v e p r e d i c a t e s i s r e q u i r e d . 58 (19)c.na natlh 7i-kw-na q'wuy ta miXalh RL morning and-TD-RL die DET bear The bear died in the morning EL 24-4-97 08 The bear died in the early morning EL 19-6-97 In (19c), the adverb in main predicate position gets an event time reading with the eventive verb q'wuy 'die'. Thus, the syntax determines the interpretation of both the adverbial and the verb. The sentence in (20), which appears to have the same structure as (19b.) and (c), presents an apparent problem for the generalization that adverbs in main predicate position get an event time reading. In this sentence, the adverbial natlh 'morning', without the determiner but preceded by the realis clitic, is the main predicate but it gets a topic time interpretation as in (19a.), and the predicate haw-q gets a stative interpretation. The speaker explains that the bear "went in the night or the morning, but when you looked he wasn't there." (20) na natlh 7i-kw-na haw-q ta miXalh RL morning and-TD-RL NEG-IRR DET bear In the morning, the bear was dead/ was gone EL 24-4-97, EL 19-6-97, EL 13-8-97 07 The difference between (20) and (19b.) is the determiner on the adverb, while the difference between (20) and (19c.) is the aspectual class of the verb. Assuming that, in (19b.), there is a zero focus marker and the adverb with its determiner is an argument, while in (20), the adverb with the realis clitic na is a predicate, the topic time reading of the adverb in (20) appears to be a result of the equation of the adverb with a stative predicate. Clearly, the 59 interaction of stative predicates and topic time adverbs requires further investigation. The explanation of the examples in (19) and (20) relies on the analysis that the determiner on the adverb identifies an argument and the absence of the determiner on the adverb identifies a predicate (Jelinek 1993, Matthewson 1996). Furthermore, the adverb without the determiner is ungrammatical at the right edge of the main clause, as shown in (21): (21) *haw-q ta miXalh na natlh NEG-IRR DET bear RL morning EL 13-8-97 This example therefore argues for the argument status of the right edge adverb22. The narrative data in this section supports the argument that the topic/focus structure determines the interpretation of the adverb. In addition, the aspectual class of the verb interacts with the interpretation of the adverb in ways that remain unclear. Furthermore, the presence or absence of a determiner on the adverb identifies it as an argument or as a predicate. ^2The sentence in (i) is very common in narrative, but presents a problem for the present analysis. The adverb at the right edge of this nominalized clause gets the event time reading and the verbal predicate gets the eventive interpretation: (i) s-es-men haw-q ta miXalh ti natlh nom-3poss-just NEG-IRR DET bear DET morning The bear died this morning. EL 13-8-97 60 3.4 The Perfect in English and SqwXwu7mish The basis for the proposal of a reference or topic time is the English perfect (Reichenbach 1947, Hornstein 1990, Thompson 1994). If the perfect exists in SqwXwu7mish, it must also argue for the existence of topic time. Analyses that deny the existence of the topic or reference time do not dispute this interpretation of the perfect; they simply dispute the theoretical status of a third time. To distinguish the present perfect from the simple past, the former is analyzed as having an event time, E, prior to a topic time, T, which coincides with the time of utterance, S, as in (22a.); the past perfect describes an event, E, which is before a topic time, T, prior to the speech time S, as in (22b.); while the future perfect identifies an event, E, before a topic time, T, which is after the speech time, S, as in (22c.): (22)a.Peter has arrived (just now). < E T,S > T = now b.Peter had arrived at 2 p.m.(it is now 6 p.m.) < E-----T S > T = 2 p.m. c.Peter will have arrived at 2 p.m. ( it is now 10 a.m.) < S E T > T = 2 p.m. The following data establishes that in SqwXwu7mish, a past perfect meaning can be obtained in sentences with na 'realis', kw 'temporal deictic' and an overtly specified topic time; in (23) the topic time is given by the clausal temporal adverbial introduced by when: 61 (23) na-kw huya7 kwelha Vanessa 7i-kw-na chen tl'iq RL-TD leave-DET Vanessa and-TD-RL lsg.subj arrive She (Vanessa had) already left when I got there. EL 14-8-97 The existence of the perfect in SqwXwu7mish argues for the existence of topic time. The present perfect, which is also marked by ha-kw, is shown in (24a.); it contrasts with the simple past, which is marked by the morpheme t, as in (24b.): (24)a.na-kw tl'iq ta Peter RL-TD arrive DET Peter Peter's here, now EL 13-8-97 b.na7-t tl'iq ta Peter RL-PT arrive DET Peter Peter got here EL 13-8-97 The future perfect could not be elicited in parallel to the sentences in (23) and (24), and therefore does not appear to exist in SqwXwu7mish. The expression of an equivalent remains an area for further research. The English present perfect has at least two readings: the "pure perfect" reading, where the action of the verb coincides with now; and the simple past reading, where the action of the verb occurred some time in the past. The temporal deictic kw is incompatible with the adverb today, as shown in (25), which suggests that it is a pure perfect: (25)a.*na-kw tl'iq ta Peter ti stsi7ts RL-TD arrive DET Peter DET today EL 13-8-97 b.na tl'iq ta Peter ti stsi7ts RL arrive DET Peter DET today Peter got here today EL 13-8-97 62 The expression of the perfect in English and SqwXwu7mish is summarized in the table in (26): (26) Present Perfect Past Perfect Future Perfect S,T,E E T,S E T S S E T English she has v-ed (now) she had v-ed (at X, when..) she will have v-ed (at X) SqwXwu7mish na-kw v na-kw v + when clause * Note that the when-clause is obligatory to distinguish the past perfect from the present perfect. 3 . 5 Aspectual Adverbs Turning to the interaction of durative adverbs with the perfect and with the future provides further evidence of a topic time. For both perfect and future, the topic time and the speech time are the same interval, now; the durative adverb always specifies the duration of the event. Section 3.5.1 illustrates how the sentence can have either a "topic time (dependent) " reading for which the event time of the adverb coincides with the speech and topic time, or an "event time (dependent) " reading for which the event time does not coincide with the speech and topic time. Section 3.5.2 shows that the equivalent data in SqwXwu7mish is unambiguous. 3.5.1 English Aspectual Adverbs Thompson (1994) and Hitzeman (1995) present the example of a durative adverb and present perfect stative verb that has both "topic time" and "event time" readings (from Dowty 1979, Kamp and Reyle 1993): 63 (27) John has been in Boston for two weeks. 1. John has been in Boston for two weeks, and he is still in Boston, (topic time reading) 2. John has been in Boston for two weeks, some time in the past., (event time reading) In contrast, the fronted durative adverb with a present perfect stative verb can have only the "topic time" reading that coincides with the topic and speech time: (28) For two weeks, John has been in Boston. 1. For two weeks, John has been in Boston and he is still there, (topic time reading) 2. *For two weeks, John has been in Boston a long time ago. (*event time reading) In fact, in both -the present perfect and the future, the topic time coincides with the speech time. Therefore, Hitzeman (1995) argues, the topic time/event time reading ambiguity of the perfect exists in the future, as in (29): (29) Martha will be in her office for an hour. 1. She will be in her office for an hour from the time of utterance (topic time reading) 2. She will be in her office for some unspecified hour in the future (event time reading) The reading that coincides with the topic time is the only one available when the adverb is sentence initial. Further evidence for the two readings in the future results from the addition of the adverb one day next week: (30)a.Martha will be in her office for an hour one day next week. b.*For an hour, Martha will be in her office one day next week. 64 Because an hour one day next week cannot be from the time of utterance, it forces the event time reading in these cases. 3.5.2 SqwXwu7mish Aspectual Adverbs In SqwXwu7mish, the equivalents of the English data do not exhibit the same ambiguity. The durative adverb is always the main predicate, specifying the duration of the event23. This section will show that the presence of the perfect or the absence of future morphology with the adverbial predicate results in the "topic time" reading by relating the time specified by the durative adverb to now; the presence of future morphology with the adverbial predicate gives the "event time" reading by locating the time of the adverb away from now. The sentence in (31) shows that the adverb with the perfect morphology kw gets only the "topic time" reading, "two weeks before now": (31)a.na-kw 7an7us sXalhnat s-es wa na7 ta Peter RL-TD two week(Sunday) nom-3poss DR there DET Peter Peter's been there already two weeks Peter was at SqwXwu7mish for two weeks, just until today EL 24-4-97, EL 13-8-97 15 b.na-kw 7an7us sXalhnat kwi-s-es wa na7 ta Peter RL-TD two week DET-nom-3poss DR there DET Peter Peter's been there already two weeks EL 24-4-97 16 In contrast, in both examples in (32), Peter is going but is not yet in Squamish. The adverb is modified by eg', the "future" morpheme, to specify two weeks in the future which 23See Section 1.5 for Kamp and Reyle's (1993) characterization of these adverbs. 65 do not coincide with the contemporaneous speech and topic times, giving the "event time" reading: (32)a.na-eq' 7an7us sXalhnat kwi-s-es-eq' na7 kwa Peter na7 t kwa SqwXwu7mish RL-FUT two week DET-nom-3poss-FUT there DET Peter there obi DET Squamish Peter will be up at Squamish for two weeks EL 24-4-97, EL 13-8-97 13 b.na-eq" 7an7us sXalhnat kwi-s na7-s kwa Peter na7 t kwa SqwXwu7mish RL-FUT two week(Sunday) DET-nom there-3poss DET Peter there obi DET Squamish Peter will be up at Squamish for two weeks (he's not there yet) EL 24-4-97 14 When the aspectual adverb is not modified by eg' but the nominalized clause agreement marks the future, as in (33), "two weeks" do coincide with the topic and speech time now for the "topic time" reading: (33) 7an7us sXalhnat kwi-s na7-s ta Peter na7 t kwa SqwXwu7mish two week(Sunday) DET-nom there-3poss DET Peter there obi DET Squamish For two weeks, Peter will be there in Squamish EL 24-4-97, EL 13-8-97 18 Thus, aspectual adverbs in both SqwXwu7mish and English can specify an interval that coincides with the speech time and the topic time, which are contemporaneous in both the perfect and the future. This is summarized in (34): (34) "topic time dependent" "event time dependent" kw-adverb+[PAST] V * adverb+[FUT] V * eq' -adverb+[FUT] * V This argues further for the existence of topic time. 66 3.6 Conclusions This chapter presented evidence to argue for the presence of topic time in all utterances. The right edge adverb in SqwXwu7mish gets a topic time reading and is part of the topic structure of the sentence, whereas the main predicate adverb gets an event time reading and is in the focus structure. The interaction of adverbs with quantificational elements confirms these conclusions; quantifiers can take scope and quantify only over the event time adverb, not over the topic time adverb. With universal and cardinal quantification, topic time adverbs get a collective reading while event time adverbs get a collective or a distributive reading. Narrative evidence provides evidence of a topic time adverb with a stative predicate and shows that this adverb must occur with a determiner at the right adge of the sentence. Finally, the perfect has been shown to exist in SqwXwu7mish, which, along with contrasts between readings of aspectual adverbs with the perfect and the future, argues for the equal status of speech time, event time, and topic time. If the sentence final adverb is unambiguously the topic time and its determiner marks it as an argument, what is its syntactic position? The following chapter examines the distribution of the adverb to support the claim that it is not an adjunct but the argument of a temporal predicate. 67 4 . The Distribution of SqwXwu7mish Adverbs SqwXwu7mish phrasal adverbs do not move freely in the sentence. This chapter demonstrates that their syntactic position is restricted in the following ways: two temporal adverbs may not co-occur in the same clause; the preferred position of the adverb is on the right periphery of the clause; and the adverb on the right edge of a subordinate clause cannot be interpreted as specifying the event time of the main clause. Furthermore, locating adverbs are shown to behave differently than "functional" adverbs such as always and maybe. These restrictions argue against the analysis of SqwXwu7mish adverbs as adjuncts. 4 . 1 Two Temporal Adverbs The sentence in (1) is ungrammatical because the number of temporal adverbs in a clause is restricted to one: (l)a.*chen s&tshit ta Peter ti siten ta sXalhnat kwi siy'elanem lsg.subj give DET Peter DET basket DET Sunday DET year I gave Peter the basket on a Sunday last year EL 17-4-97 17 b.*kwi siy"elanem chen s&tshit ta Peter ti siten ta sXalhnat DET year lsg.subj give DET Peter DET basket DET Sunday Last year I gave Peter the basket on a Sunday EL 17-4-97 18 This restriction extends to cardinal quantifiers, as shown in (2) : (2) *na ts'its'ap' ta Peter na7 t 7an7us-alh t kwetsi sXalhnat RL work DET Peter there obi two-times obi DET Sunday Peter worked twice on a Sunday EL 2-4-97 12 68 Although the two adverbs denote only one time, overlapping in reference, this is a strong restriction.24 In order for the two adverbs to express a single time, as in the English cluster on a Sunday last year, Sunday occurs in the main clause with the predicate there25, and last year appears in a nominalized clause, as shown in (3), or in a conjoined clause as in (4)26: (3) na na7 t kwetsi sXalhnat s-en satshit ta Peter kwi siy'elanem RL there obi DET Sunday nom-lposs give DET Peter DET year Twas a Sunday last year that I gave Peter (the basket) EL 17-4-97 19 (4) na na7 t kwetsi sXalhnat 7i-kw-na chen satshit ta Peter ti siten kwi siy'elanem RL there obi DET Sunday and-TD-RL-lsg.subj give DET Peter DET basket DET year I gave Peter the basket on a Sunday last year EL 17-4-97 21 These facts argue for the existence of a s ingle pos i t ion for the adverb in the clause. This makes the analysis of adverbs as adjuncts improbable; if they were adjuncts , temporal adverbs should be able to appear in s t r ings , a l b e i t ordered s t r ings , in a single clause. 2 4 T h i s obse rva t ion i s by Hamida Demirdache ( p . c ) . 25The e x i s t e n c e of the ob l ique on the adverb i s not a problem for t h e a n a l y s i s of adverbs as arguments. Enc (1987) observes t h a t , in Engl i sh , p r e p o s i t i o n a l adverbs such as on Monday have the same d i s t r i b u t i o n as NP adve rbs such as yesterday. She a n a l y s e s t h e s e p r e p o s i t i o n s a s " seman t i ca l ly vacuous" and sugges ts t h a t temporal PPs a r e s y n t a c t i c a l l y NPs. 2^ S e e S e c t i o n 2 . 4 . This i s a case of c o o r d i n a t i o n , r a t h e r t han s u b o r d i n a t i o n , because of the main c l ause agreement chen in t h e second c l a u s e . 69 4.2 SqwXwu7mish Adverbs Prefer the Right Periphery The three positions of the English adverb are exemplified in (5): (5)a.Yesterday, the lady saw the bear. b.?The lady, yesterday, saw the bear. c.*The lady saw yesterday the bear. d.The lady saw the bear yesterday. Sentences (5a.) and (5d.) are equal in acceptability; sentence (5b.) requires a particular intonation to be understood; and sentence (5c.) is ungrammatical because the adverb intervenes between the verb and its complement. Thus, in English the position of the adverb is marginally restricted. In contrast, the position of SqwXwu7mish adverbs is confined to the right periphery27. The examples in (6) show that the adverb cannot be positioned before the subject and verbal predicate in a main clause: (6)a.*kwi chelaqlh chen (-t) ts'its'ap' DET yesterday lsg.subj(PT) work EL 19-12-96 b.*ti sti7ts chen ts'its'ap' DET today lsg.subj work EL 19-12-96 c.*q-kwayeles chen ts'itsap' IRR-tomorrow lsg.subj work EL 19-12-96, EL 6-2-97 d.*kwayeles chen ts'its'ap' tomorrow lsg.subj work I'll be working tomorrow EL 20-2-97 27Henry Davis (p.c.) states that to his knowledge, this restriction has not been observed in any other Salish language. 70 The preferred position of adverbs in single clause constructions is at the end of the sentence, as shown in (7) and (8). (7)a.chen ilhen kwi chel'aqlh lsg.subj eat DET yesterday I ate yesterday YJ 13-3-9 6 b.chen ilhen 7i7Xw sqwayl lsg.subj eat every day I eat every day DW 20-3-96 c.chen ilhen kwayeles lsg.subj eat tomorrow I'll be eating tomorrow DW 20-3-9 6 d.chen ilhen kwetsi sXalhnat lsg.subj eat DET Sunday It was a Sunday that I ate EL 19-6-97 (8)a.chen tl'iq-s ti siten kwi chel'aqlh lsg.subj arrive-caus DET basket DET yesterday I brought this basket (from close by) yesterday EL 27-3-97 03 b.chen tl'iq-s ti siten ta sXalhnat lsg.subj arrive-caus DET basket DET Sunday I brought the basket on (any) Sunday EL 17-4-97; EL 19-6-97 01 c.chen tl'iq-s ti siten t kwetsi sXalhnat lsg.subj arrive-caus DET basket obi DET Sunday I brought the basket (here) on Sunday EL 17-4-97; EL 8-5-97 03 The adverb is dispreferred in the position between the verb and its object, as shown in (9): (9)a.*/Vchen tl'iq-s kwi chel'aqlh ti siten lsg.subj arrive-caus DET yesterday DET basket I brought this basket (from close by) yesterday EL 27-3-97; EL 17-4-97 04 b.*chen tl'iq-s ta sXalhnat ti siten lsg.subj arrive-caus DET Sunday DET basket EL 17-4-97; EL 8-5-97; EL 19-6-97 06 The speaker often rejects such sentences on the basis that "you're bringing yesterday" or "you're bringing the Sunday!". 71 Sentences with a full DP subject before or after the verb show a similar pattern: the adverb is grammatical at the right periphery, as in (10b.) and (lib.); the adverb intervening between the subject and object is ungrammatical, as in (10c.) and (lie); and the adverb appearing between the verb and either a postverbal or preverbal subject is ungrammatical, as in (lOd.) and (lid.): (lO)a.(na) k'wach-nexw-as kwelhi slhanay' ta miXalh (RL) see-trans-3subj DET lady DET bear The lady saw the bear EL 27-3-97 23 b.k'wach-nexw-as kwelhi slhanay' ta miXalh kwi chel'aqlh see-trans-3subj DET lady DET bear DET yesterday Yesterday the lady saw the bear EL:" start from the end" EL 27-3-97 28 c.*k'wach-nexw-as kwelhi slhanay' kwi chel'aqlh ta miXalh see-trans-3subj DET lady DET yesterday DET bear Yesterday the lady saw the bear EL 27-3-97 29 d.*k'wach-nexw-as kwi chel'aqlh kwelhi slhanay' ta miXalh see-trans-3subj DET yesterday DET lady DET bear The lady saw the bear yesterday EL 27-3-97 32 (11)a.kwelhi slhanay' k'wach-nexw-as ta miXalh DET lady see-trans-3subj DET bear The lady saw the bear EL 27-3-97 24 b.kwelhi slhanay' k'wach-nexw-as ta miXalh kwi chel'aqlh DET lady see-trans-3subj DET bear DET yesterday The lady saw the bear yesterday EL 27-3-97 25 c.*kwelhi slhanay' k'wach-nexw-as kwi chel'aqlh ta miXalh DET lady see-trans-3subj DET yesterday DET bear The lady saw the bear yesterday EL 27-3-97 26 72 (11)d.*kwelhi slhanay' kwi chel'aqlh k'wach-nexw-as ta miXalh DET lady DET yesterday see-trans-3subj DET bear The lady saw the bear yesterday EL 27-3-97 27 In ditransitive constructions, again the adverb is grammatical at the right edge, as in (12): (12)a.chen s&tshit ta Peter ti siten kwi chel'aqlh lsg.subj give DET Peter DET basket DET yesterday I gave this basket to Peter yesterday EL 8-5-97 EL 27-3-97; EL 8-5-97 07 b.chen satshit ta Peter ti siten ta sXalhnat lsg.subj give DET Peter DET basket DET Sunday I gave Peter the basket on Sunday WOR; *kwi sXalhnat EL 17-4-97 09 When the adverb intervenes between the direct and indirect object, as in (13), the sentence is marginal: (13)a.*/Vchen satshit ta Peter kwi chel'aqlh ti siten lsg.subj give DET Peter DET yesterday DET basket I gave this basket to Peter yesterday EL 27-3-97; EL 17-4-97; EL 8-5-97 08 b.chen satshit ta Peter ta sXalhnat ti siten lsg.subj give DET Peter DET Sunday DET basket I gave it to Peter on a Sunday EL 17-.4-97 15,23 When the adverb appears between the verb and its objects, as in (14), the sentence is again marginal or ungrammatical: (14)a.*/Vchen satshit kwi chel'aqlh ta Peter ti siten lsg.subj give DET yesterday DET Peter DET basket I gave this basket to Peter yesterday EL 27-3-97; EL 17-4-97;EL 8-5-97 09 b.*chen satshit ta sXalhnat ta Peter ti siten lsg.subj give DET Sunday DET ,Peter DET basket I gave this basket to Peter on Sunday EL 17-4-97; EL 8-5-97 22 Finally, as seen in Chapter 3, the adverb may' be syntactically focused as a main predicate, on its own or as an argument with nilh or the zero focus marker. This is 73 shown in the coordination structures in (15). In (15a.), the adverb on its own is a predicate (Jelinek 1993, Matthewson 1996) which is coordinated with another clause by the clitics 7i-kw; in (15b.), the adverb, with the oblique and the determiner, is an argument with a zero focus marker; and in (15c), the adverb with the oblique and the determiner is an argument with the overt focus marker nilh: (15)a.sXalhnat 7i-kw-na chen tl'iq-s ti siten Sunday and-TD-RL lsg.subj arrive-caus DET basket I brought the (your) basket on Sunday (volunteered form) EL 17-4-97; EL 1-5-97; EL 8-5-97 04 b.0 t kwetsi sXalhnat 7i-kwe-na chen mi-tl'iq-s ti siten FOC obi DET Sunday and-TD-RL lsg.subj come-arrive-caus DET basket I brought the basket on a Sunday, lit. It was any Sunday that I brought the basket EL 17-4-97;EL 8-5-97 05 c.nilh t kwetsi sXalhnat 7i-kwe-na chen satshit ta Peter kwetsi siten FOC obi DET Sunday and-TD-RL lsg.subj give DET Peter DET basket It was a Sunday I gave Peter the basket EL 8-5-97 26 d.nilh sXalhnat 7i-kwe-na chen mi-tl'iq-s ta siten FOC Sunday and-TD-RL lsg.subj come-arrive-caus DET basket I already brought the basket on a Sunday EL 8-5-97 24 The absence of the determiner on the adverb with the overt copula in (15d.) is predicted to be ungrammatical, but is assumed to be a case where the determiner is dropped. To summarize, the position of the adverb in SqwXwu7mish is preferred at the right edge of the main clause. Adverbs before the verb are disallowed, while adverbs between the verb and its complement are dispreferred. Such restrictions 74 on the attachment of the adverb further argue against their characterization as adjuncts. 4 . 3 Adverbs and Subordinate Clauses Thompson (1995) and Hitzeman (1995) argue for English that adverbs which specify the event time are adjoined to VP, whereas adverbs which specify the topic time are adjoined at sentence level. This explains why English adverbs which appear at the right periphery of the sentence are ambiguous between event time and topic time, while adverbs at the left edge are unambiguously the topic time. A parallel analysis can be extended to adverbs at the right edge of subordinate clauses in comparison with adverbs at the right edge of main clauses, as shown in (16). Addressing only adjunction to S for simplicity28, in (16a.), the adverb at the right edge of the subordinate clause is ambiguous between construal with the subordinate clause and construal with the main clause; in (16b.), the adverb at the right edge of the main clause is unambiguously construed with the main clause: (16)a.I said that Peter would be going away yesterday. 1. I said at a time in the past that the event of Peter's leaving was yesterday. 2. I said yesterday that the event of Peter's leaving would occur. 28In the simple past, where the topic time and the event time coincide, the distinction of readings resulting from adjunction to V and readings resulting from adjunction to S is impossible with a deictic adverb like yesterday (see Hitzeman 1995). 75 (16)b.I said yesterday that Peter would be going away. 1. *i said that the event of Peter's leaving was yesterday. 2. I said yesterday that the event of Peter's leaving would occur. Again, the asymmetry between these sentences can be explained by the adjunction site of the adverb. When the adverb appears at the right edge of the subordinate clause, it may be adjoined either to the higher S, for reading (2.), or to the lower S for reading (1.); when the adverb appears at the right of the main clause, it can only be adjoined to the higher S. The equivalent sentences in SqwXwu7mish argue against the analysis of adverbs as adjuncts, however. The adverb on the right of a subordinate clause cannot be construed with the main clause because there is only one position available for the adverb in each clause. (17) *chen tsut kwi-s nam' huya7 ta Peter kwi chel'aqlh lsg.subj say DET-nom go leave DET Peter DET yesterday I said yesterday that Peter would be going away29 EL 27-3-97 33 The adverb yesterday cannot be attached to the lower CP because it is incompatible with the future construction nam' huya7 'going to leave'; the ungrammaticality of this sentence shows that the adverb cannot be attached to the upper CP or its interpretation with the main clause would be possible. The volunteered correction of the sentence moves the adverb to the right of the main verb as shown in (18): 2 9Note t h a t t he t r a n s l a t i o n s of the examples in t h i s s e c t i o n a r e those p r o v i d e d by t h e speake r ; they show sequence of t e n s e e f f e c t s (see Stowell 1993 for a d e t a i l e d explana t ion of these in E n g l i s h ) . 76 (18) chen tsut kwi chel'aqlh kwi-s nam' huya7 ta Peter lsg.subj say DET yesterday DET-nom go leave DET Peter I said yesterday that Peter would be going away EL 27-3-97 34 In order for the adverb to be interpreted with the main clause from the right edge of the sentence, the agreement for the third person possessive, -es, is added to the nominalizer of the subordinate nominalized clause30: (19)a.chen tsut [kwi-s-es nam' huya7 ta Peter] kwi chel'aqlh lsg.subj say [DET-nom-3poss go leave DET Peter] DET yesterday I said yesterday that Peter's going away EL 27-3-97 35 b.chen t su t [kwi-s-es huya7 ta Peter] kwi chel'aqlh lsg.subj say DET-nom-3poss leave DET Peter DET yesterday I said yesterday that Peter went away I said that Peter went away yesterday EL 27-3-97 36 The sentences in (18) and (19) appear to be l ike the English examples in (16); the adverb at the r ight periphery in (17) should a l so be i n t e rp re t ab l e in the main c lause , but the sentence i s ungrammatical. Thus, there can be only, one p o s i t i o n per c lause for the adverb in SqwXwu7mish, in con t ras t with the three in English. An account for the contrast in these four examples i s proposed in Section 5.6. 4 . 4 Functional Adverbs31 Unlike the locat ing adverbs discussed in the previous sec t ions , ce r t a in "functional" adverbs in SqwXwu7mish must occupy the posi t ion before the subject c l i t i c and verb as in 3 0 P . Jacobs ( p . c . ) says t h a t t h e r e i s possess ive marking on nam' in (17) and (18) which has been dropped in rap id speech. 3•'•This s e c t i o n was the sugges t ion of Rose-Marie Dechaine. 77 (20); (21) shows that they cannot occupy the position at the right periphery preferred by adverbs such as yesterday and on Sunday. (20 )a .ch iya lh chen t s ' i t s ' a p ' soon l s g . s u b j work I 'm gonna work soon EL 1-5-97 18 b.way ' t i chen t s ' i t s ' a p ' maybe l s g . s u b j work I'm not s u r e / I th ink I w i l l work EL 1-5-97 19 c . l h i q ' chen wa t s ' i t s ' a p ' always l s g . s u b j DR work I'm always working (volunteered form) EL 1-5-97 23 d . 7 i - w a y ' t i l h i q ' chen t s ' i t s ' a p ' PR-maybe always l s g . s u b j work Maybe I w i l l work (volunteered form) EL 1-5-97 21 (21)a .*chen t s ' i t s ' a p ' chiyalh l s g . s u b j work soon EL 1-5-97 24 b.*chen t s ' i t s ' a p ' way' t i l s g . s u b j work maybe EL 1-5-97 26 c.*chen wa t s ' i t s ' a p ' l h i q ' 3 2 l s g . s u b j DR work always I 'm working always EL 1-5-97 27 There fo re , f u n c t i o n a l adverbs a re o p e r a t o r s , m o d i f i e r s , or q u a n t i f i e r s ; Kuipers (1967) c a l l s them " sen t ence -ad junc t s . " -^The adverb lhiq' ' a lways ' has a l e s s r e s t r i c t e d d i s t r i b u t i o n and can appear a f t e r the sub jec t c l i t i c , as shown in ( i ) : i . chen l h i q wa t s ' i t s ' a p ' l s g . s u b j always DR work I 'm always working (I never stop) answers the question:"Why d o n ' t I ever see you?" EL 1-5-97 28 78 Converse ly , f unc t i ona l adverbs cannot be main predica tes , e i ther on the i r own as in (22) or with nilh as in (23) : (22)a.*chiyalh 7i-kwe-na chen tl'iq-s ti siten soon and-TD-RL lsg.subj arrive-caus DET basket EL 1-5-97 53 b.*way'ti 7i-kwe-na chen tl'iq-s ti siten maybe and-TD-RL' lsg.subj arrive-caus DET basket EL 1-5-97 51 c.*lhiq' 7i-kwe-na chen tl'iq-s ti siten always and-TD-RL lsg.subj arrive-caus DET basket EL 1-5-97 48 (23)a.*nilh chiyalh tl'-as nam' huya7 FOC soon WH-3subj go leave EL 1-5-97 37 b.*nilh way'ti tl'-as nam' huya7 FOC maybe WH-3subj go leave EL 1-5-97 39 c.*nilh lhiq" tl'-as nam' huya7 FOC always WH-3subj go leave EL 1-5-97 41 This is in direct contrast to locating adverbs such as yesterday, tomorrow and Sunday, which have been shown to be main predicates when they determine the event time of the utterance; in (24), the adverbs are main predicates with nominalized clause complements (see Section 2.4): (24)a.(na) sXalhnat kwi-s-es ts'its'ap' ta Peter (RL) Sunday DET-nom-3poss work DET Peter Peter worked on Sunday EL 19-6-97 b.kwi chelaqlh s-es ts'its'ap' ta Peter yesterday nom-3poss work DET Peter Yesterday Peter worked. EL 19-6-97 c.kwayl kwi-s ts'its'ap'-s ta Peter tomorrow DET-nom work-3poss DET Peter Peter will be working tomorrow EL 19-6-97 79 (24)d.kwayl(-eq') 7i nam1 huya7 kwa Peter tomorrow(-FUT) and go leave DET Peter Tomorrow Peter will be going away EL 1-5-97 34 The focus marker nilh is often used with these adverbs as predicates, here with conjunctive clause complements: (25)a.nilh kwi sXalhnat tl'-as nam1 huya7 FOC DET Sunday WH-3subj go leave He went away on Sunday lit. it was Sunday when he went away EL 1-5-97 43 b.nilh kwi chel'aqlh tl'-as nam1 huya7 FOC DET yesterday WH-3subj go leave It was yesterday (when) somebody went away EL 1-5-97 30 In this way, locating adverbs behave in the same way as arguments such as the direct object, which can be the main predicate without a determiner, with the zero focus marker and with nilh, as shown in (26) : (26)a.siten [ti na tl'iq-s-t-an] basket [DET RL arrive-caus-trans-lsg] It's a basket that I brought volunteered form EL 8-5-97 03 b.0 kwetsi siten [0 na tl'iq-s-t-an] FOC DET basket [rel RL arrive-caus-trans-lsg] It was the basket I brought EL 8-5-97 05 c.nilh ti siten [0 na tl'iq-s-t-an] FOC DET basket [rel RL arrive-caus-trans-lsg] This basket I brought EL 8-5-97 07 The complement of these predicates is a zero relative clause. This comparison of temporal adverbs with functional adverbs shows a strong distinction: functional adverbs are preferred at the left edge of main clauses and cannot be main predicates, whereas locating adverbs occur at the left edge 80 of a- sentence only as main predicates. Their ability to be main predicates aligns temporal adverbs with other arguments, such as direct objects. In fact, it is a property of Salish (Davis 1996:1-2) that "open-class" categories can be predicates on their own and arguments when preceded by a determiner (Jelinek 1993, Davis 1996). 4 . 5 Summary The restrictions on the position of the adverb support the claim that they are not adjuncts. If they were adjuncts, multiple temporal adverbs should be able to occur in the clause, adjoined to different positions. Therefore, the distribution of SqwXwu7mish adverbs forces their analysis as arguments; more precisely, the distribution and interpretation of SqwXwu7mish temporal adverbs argues for their analysis as time-denoting arguments, selected by a temporal predicate. 81 5 . The Syntax .of the Topic Time Argument SqwXwu7mish adverbials identify the topic time, or, if they are focused syntactically as main predicates, they identify the event time, as shown in Chapter 3. Chapter 4 presented data that argued for a restriction of one adverb per clause, with a preferred position at the right edge of the sentence. This supported the claim that the adverb in SqwXwu7mish is an argument, not an adjunct. This chapter will argue that SqwXwu7mish adverbs are the topic time argument in the specifier of the spatiotemporal predicate Aspect (Demirdache and Uribe-Etxebarria, to appear). This analysis is motivated on conceptual grounds in Section 5.1. Section 5.2 reviews the syntax of Tense and Aspect proposed by Demirdache and Uribe-Etxebarria, then Section 5.3 lays out the modifications necessary for the application of this framework to SqwXwu7mish. Section 5.4 proposes an analysis of how the temporal interpretation of pronoun-verb constructions is derived; the derivation of word order is examined in 5.5. In Section 5.6, the construal of adverbials in complex sentences is addressed. Finally, in Section 5.7, the interaction of topic time adverbs and quantification supports a non-structural theory of scope (Demirdache and Matthewson 1997). 5.1. Topic Time, Temporal Adverbials and Aspect The proposal that it is Aspect which determines the relation between the reference or topic time and the event 82 time is explored in analyses such as Giorgi and Pianesi (1991), Thompson (1994), and Klein (1994, 1995). Giorgi and Pianesi (1991) propose two Tense heads: the higher Tl establishes a relation between speech time, S, and reference (or topic) time, R, and the lower T2 establishes a relation between R and event time,E. These relations are governed by the "Biunique Mapping Principle", whereby there is a "biunique correspondence" between overt "temporal morphemes and T-relations" (Giorgi and Pianesi 1991:191). Thus, under this analysis "the syntactic primitives of tense are the relations between temporal points (..) and not the points themselves" (Thompson 1994a:fn.l). Thompson argues that such a proposal fails to account for the interpretation of adverbs, as adverbs never modify more than one time point. As discussed in Section 1.4, she claims that each of the three time points associates to a head: speech time, S, to Tense, reference (or topic) time, R, to Aspect, and event time, E, to V. Tense morphemes thereby determine the S,R relation while the presence or absence of have determines the relation of R to E. The ambiguity in adverbial interpretation between a reference or topic time reading and an event time reading can be accounted for if the topic time reading is a result of adjunction of the adverb to Aspect and the event time reading due to adjunction to V. Klein (1994:99) also argues that tense relates the time of utterance, TU, and the reference or topic time, TT, while aspect is a way "to relate the time of situation (TSit) to 83 the topic time: TT can precede TSit, TT can follow TSit, TT can contain TSit, or TT can be partly or fully contained in TSit". Crucially, the topic time, TT, may be specified by adverbs that denote either a single point in time or an interval33. Thus, the topic time denoted by the adverb is related to the event time by aspect. This thesis proposes that if SqwXwu7mish adverbs are arguments which denote topic time, and if aspect is a relation between event time and topic time, then temporal adverbs are a topic time argument of the temporal predicate Aspect. The syntax of Tense and Aspect proposed by Demirdache and Uribe-Etxebarria (to appear a,b), with adaptations for SqwXwu7mish, provides the framework for such an analysis. 5.2 The Syntax of Tense and Aspect34 This section briefly outlines the syntax of Tense and Aspect presented in Section 1.5, then examines Demirdache and Uribe-Etxebarria's analysis of the English Past Progressive. To capture the semantic parallelism that both tense and aspect order two times, Demirdache and Uribe-Etxebarria argue that syntactically Tense and Aspect are prepositional-type predicates which order two time-denoting arguments with the meaning "before," "after," or "within". Each argument is a 33As discussed in Chapter 1, Klein's topic time differs from Kamp and Reyle's (1993) Temporal Perspective Point in that it can denote an interval, and it serves as the temporal frame for the assertion of the sentence. 3^ S e e Chapte r 1, S e c t i o n 1.5 for a more d e t a i l e d e x p l a n a t i o n of Demirdache and U r i b e ' s framework. 84 time denoting phrase: the speech time argument, UT-T, is the external argument of Tense, generated in the specifier of Tense; the reference or topic time argument, TOP-T, is the internal argument of Tense and the external argument of Aspect, generated in the specifier of Aspect; and the event time argument EV-T is the internal argument of Aspect and the outer external argument of the VP, generated in the specifier of the VP. Tense and Aspect make use of multiple specifier positions (Chomsky 1995) to fulfill their "functional roles". The functional specifier of TP takes over the function of the AgrS specifier, while the specifier of AspP is the equivalent of the specifier of AgrO. These multiple specifier positions accomodate both the temporal arguments of Tense and Aspect and the arguments of the verb, which are base-generated in the VP and moved up for Case assignment or feature checking. The authors' analysis of the Past Progressive in the utterance in (1) illustrates how their framework functions: (1) Henry was building a house. (2) TP / \ U T - T T ' / \ T ASP-P af ter / \ TOP-T ASP' / \ ASP VP w i t h i n y > v EV-T V V 85 Progressive Aspect is a predicate meaning "within", which orders the topic time within the event time, the time of building; Past Tense is a predicate meaning "after", which orders the speech time after the topic time. Therefore, the analysis in (2) "focuses a subinterval within the interval defined by the event of building. This subinterval is itself located in the past" (Demirdache and Uribe-Etxebarria to appear b). This is illustrated by the time line in (3): (3) TOP-T UT-T EV-T 5. 3 Adapting the Framework to SqwXwu7mish Four issues must be addressed in order for the SqwXwu7mish data presented in the previous chapters to be analyzed within this framework: the analysis of simple tenses; the syntax and semantics of the Zeit Phrase, ZP; the status of recursive aspect; and the content of the Tense predicate in SqwXwu7mish. 5.3.1 Simple Tenses This thesis proposes to account for the simple tenses, Past and Future35, that are not addressed in Demirdache and Uribe-Etxebarria (to appear a,b), by proposing an Aspect predicate with the meaning "around". This is motivated with an example of the Simple Past. 3-"The present is set aside,, although the present reading of an eventive verb in SqwXwu7mish gets either a durative or a generic reading in translation to English. 86 For a sentence in the Simple Past, the speech time, UT-T, is after the topic time, TOP-T, as above; however, the event time, EV-T, is viewed as a whole, therefore it is contained in the topic time, TOP-T, as shown in (4): (4) EVT-T UT-T < [__[ ]_] | > TOP-T The Progressive relationship between EV-T and TOP-T schematized in (3) is the inverse of that of the Simple Tense shown in (4); therefore, the meaning of Aspect in the simple tenses is the inverse of the meaning of Progressive Aspect. As the topic time, TOP-T, in the Progressive is ordered "within" the event time, EV-T, by the meaning of Aspect, the topic time in a Simple tense is ordered "around" the event time. This is illustrated by the example in (5) and the analysis in (6): (5) Henry built a house. (6) TP UT-T T' / \ T ASP-P after / \ TOP-T ASP' / \ ASP VP around / \ . EV-T V / \ V In the Simple Future, the utterance time, UT-T, is ordered before the topic time, TOP-T, by the predicate Tense; the predicate Aspec orders the topic time around the event 87 time, EV-T. The diagram of these relations shown in (7) is the mirror image of that in (4) : (7) UT-T EVT-T < | [__[ ]_] > TOP-T The addition of "around" to the other possible meanings of Aspect, "within," "before," and "after," appears to violate the "strict parallel" between Tense and Aspect used by Demirdache and Uribe-Etxebarria to motivate their analysis36. Conversely, it completes the system following Klein's observation, stated in Section 5.1 above, that the topic time can be before the event time (Prospective Aspect), after the event time (Perfect Aspect) , that all or part of the topic time can be within the event time (Progressive Aspect), and that the topic time can "contain" or be around the event time (Completed Aspect) as described in this section. Furthermore, it is consistent with the analysis of temporal adverbials such as yesterday and the next day as the topic time argument, which will be taken up in Section 5.4. 5.3.2 The Category of Temporal Arguments The syntactic category of the temporal argument is the functional category ZP (Zeit Phrase), a "time-denoting phrase" which is a "referential category analoguous to DP". For Stowell (to appear a:4), the head, Z, binds a variable, itself a ZP, in the specifier of the VP it contains. The present proposal analyzes the temporal topic 36H. Demirdache (p.c.) suggests that the "Reportative Present", which is used for telic events contained in the utterance time, is the Tense parallel to the Aspect in simple tenses. 88 time argument adverb as a ZP which contains an NP; the head Z is the determiner of the adverb which binds a variable in the specifier of the NP. The structure of the topic time ZP ta sXalhnat 'that Sunday' is shown in (8a.), with the structure of the DP ta miXalh 'the bear' next to it in (8b.) for comparison: (8)a. ZP b. DP Z' D' / \ Zi NP Dj NP ta / \ ta ZP N' ej N' • i / \ N N sXalhnat miXalh Stowell (1993:9) claims that as N is a "predicative category analoguous to V (...) and that D is a referential category, a property that enables the category DP to refer" to individuals, Z is also a referential category which allows ZP to refer to times. The structure proposed in (8a.) simply extends this analogy to NP temporal adverbs, where the Z gives the adverb its temporal reference. This proposal has several advantages for the analysis of SqwXwu7mish temporal adverbials. First, it encodes the difference between phrasal adverbials, which are ZP's, and clausal adverbials, which are CP's, to account for differences in their distribution and syntactic behaviour37. 3 7H. Davis (p.c.) states that in a neighbouring Interior Salish language, St'at'imcets, temporal adverbials are CP's which attach freely in the manner of adjuncts. 89 Second, it explains how changing the determiner on the adverb, as seen in Section 2.3, can change the temporal reference of the adverb to past, present or future: (9) ti sXalhnat this Sunday ta sXalhnat that Sunday kwetsi sXalhnat any, a certain Sunday kwi sXalhnat a possible Sunday When the determiner ti, for example, is under D, it refers to individuals which are near and visible, and when ti is under Z, it refers to times which are near. Third, this proposal captures the parallel between nominal and temporal reference, while formalizing the way in which they differ. In SqwuXwu7mish, these determiners refer to individuals and to times differently: for example, kwi indicates things that cannot be seen, people who are dead, Sundays that may happen, and days and years that are past. In addition, the category ZP explains why adverbs, unlike DP nominals, do not get case. This section has focused on the analysis of topic time adverbs. The external argument of Tense is assumed to be a ZP which is PRO, following Stowell (1993, to appear a,b). Event time arguments are assumed to be ZP variables, which are interpreted in terms of their relation to the topic time argument as determined by the predicate Aspect. 5.3.3 Aspect Recursion in SqwXwu7mish Demirdache and Uribe-Etxebarria (to appear b) show how their proposal can derive Aspect recursion in English without additional mechanisms. Conversely, there cannot be Aspect recursion in SqwXwu7mish if there is only one topic time argument position. The data presented in Chapter 4 argued 90 for a single topic time position in order to account for the restriction of one temporal adverb in a clause and for the construal of adverbs in subordinate clauses. Indeed, the fact that most aspectual clitics in SqwXwu7mish (wa the durative, nam' the near future, and mi7 "become, come toward") are in complementary distribution argues against the recursion of Aspect38. Although the temporal deicitic clitic kw, which was shown to give the equivalent of the perfect, may cooccur with wa as in (10), the translations show that kw and wa are not both aspectual in this case; if they both were aspectual, the meaning would be the perfect of a progressive: (10) na-kw wa tsits'ap' RL-TD DR work He has started to work, ie., he is working now, he is already working (Kuipers 1967:158) As this requires a more thorough examination, this thesis will adopt the conclusion that there is no recursive Aspect in SqwXwu7mish. Therefore, a clause can have at most one Aspect projection, one AspP specifier, and one topic time argument, TOP-T. 5.3.4 The Predicate Tense in SqwXwu7mish Tense in SqwXwu7mish is analyzed in this thesis as lacking morphological content or feature specification, given the pronoun verb constructions in (11): The main verb forms of nam' and mi7 may cooccur with wa, however. 91 (11)a.chen ilhen lsg.subj eat I eat YJ 13-3-96 I'm eating (these days) DW 20-3-96 I ate (context:"What were you doing this morning?") EL 6-3-96,17-4-96 b. chen ilhen kwi chel'aqlh lsg.subj eat DET yesterday I ate yesterday YJ 13-3-96, LJ 29-3-96 c. chen ilhen kwayeles lsg.subj eat tomorrow I'll be eating tomorrow DW 20-3-96 d. chen ilhen kwetsi sXalhnat lsg.subj eat DET Sunday It was a Sunday that I ate EL 19-6-97 e. chen ilhen 7i7Xw sqwayl lsg.subj eat all day I eat every day DW 20-3-9 6 These sentences get their temporal interpretation either from an adverb or from the discourse, although the pronoun and verb on their own are at times uninterpretable. Furthermore, the interpretation of the bare pronoun and verb is sensitive to the aspectual class of the predicate: activity verbs get a past or present reading, statives a present reading, and achievement and accomplishment verbs a past reading39. In order to account for similar constructions in Haitian, Ddchaine (1993) proposes that Tense is a Functional head, a syntactic position without morphological or semantic content. Since the predicate Tense must have semantic content to project its arguments, its meaning is assumed to be unspecified and derived via the relation of its arguments: the internal argument of Tense, the topic time, is either a See Chapter 6 for examples. 92 temporal adverbial or a non-overt pro, which is bound by the topic time of the previous utterance; the external argument of Tense is either now in main clauses or controlled by the main event time in subordinate clauses. Thus, the temporal interpretation of these sentences is determined by the relation of the utterance time, UT-T, to the inherent temporal features of the topic time, TOP-T, and by the topic time itself, which is related by Aspect to the event time. 5.4 Analyzing the Pronoun-Verb Construction This section proposes an analysis of the sentences in (11) to show how the topic time determines the temporal interpretation of the sentence. 5.4.1 Topic Time "Yesterday" The surface form of (lib.), represented in (12), is derived as follows: the first person subject pronoun is assumed to be generated in the specifier of the TP (Davis to appear); the verb raises through Asp to Tense. The adverb does not move because it does not need to move, having no features to be checked (Thompson 1994a). The temporal interpretation of (lib.) is derived without morphological Tense in the following manner. The main clauses utterance time means "now"; the topic time denotes a time, "yesterday", that is before now. The topic time determines the time of the event by being in an ordering relation with it via Aspect: the topic time contains the event time, which therefore must also be before now. 93 (11)b.chen ilhen kwi chel'aqlh lsg.subj eat DET yesterday I ate yesterday YJ 13-3-9 6 (12) CP TP DP TP chen UT-T = ZP T' PROnow / \ T ASP-P ilhenj / \ TOP-T = ZP ASP' kwi chelaqlh / \ ^ ASP VP ej-around / \ EVT-T = ZP VP 0 pro V ej This derivation gives the interpretation: "there is a time yesterday, and now is after that time, and that time includes the time of my eating." This is represented on a time line in (13): (13) TOP-T yesterday < [ [__] ] | > eat now EVT-T UTT-T The same meaning is derived in a different formalism by the "semantic representation" constructed in Discourse Representation Theory (Kamp and Reyle 1993), shown in (14): 94 (14) n e t x yesterday(t) e C t t < n I (x) e: x eat This is interpreted informally as "there is a time yesterday and there is an event contained in that time and that time is before now; x is 'I' and the event is 'I eat'." The symbol "C" gives the meaning "e is temporally included within t"; as above, the event is contained in the time specified by the temporal adverbial. This proposal explains how the adverbial can be the sole determinant of the sentence's temporal reference. Furthermore, this proposal allows the topic time to determine the time of the event by being in an ordering relation with it via Aspect, without recourse to morphological Tense. 5.4.2 Zero Topic Time The temporal interpretation of (11a.), which has a null pro topic time, has a similar interpretation as shown in (15): 95 (11)a.chen flhen lsg.subj eat I eat, I'm eating (these days), I ate (15) CP C C TP / \ DP TP chen / / / \ UT-T = ZP T' PROnow T ASP-P i l h e n j / \ . TOP-T = ZP ASP' prod / \ ASP VP e j -around y \ EVT-T=ZP VP 0 pro V eJ The topic time argument pro in this case is bound by the topic time of the previous sentence or question; this analysis captures the fact that the topic time is often not overtly specified but is maintained from sentence to sentence. This utterance is therefore uninterpretable if no topic time is available from the discourse. 5.4.3 Topic Time "Sunday" When the topic time argument is a calendar name such as Sunday, it contains the time of the event of eating, in the same way as yesterday. Calendar names, however, are "context-dependent" (Kamp and Reyle 1993); they can pick out more than one time interval, and the relevant interval is 96 then determined by the context. In English, a demonstrative such as this or an adjective such as last may identify the referent of Sunday. In SqwXwu7mish, the determiner in Z must identify the referent of Sunday as "present," "past," or "possible" (see Sections 2.3.1 and 2.3.4). In the example in (lid.), the determiner kwetsi identifies the Sunday as "past," thereby locating the topic time before the utterance time without recourse to morphological Tense. (ll)d.chen ilhen kwetsi sXalhnat lsg.subj eat DET Sunday It was a Sunday that I ate EL 19-6-97 (16) CP C / \ C TP / \ DP TP chen //\ s UT-T = ZP T' PROnow / \ T ASP-P ilhenj y / \ TOP-T = ZP ASP' kwetsi sXalhnat y / \ ASP VP ej-around /\^ EVT-T = ZP VP 0 pro V ej The temporal interpretation of this sentence is: "there is a time, a certain Sunday in the past, and now is after that Sunday, and the event of eating is included in that time interval." This is derived by Aspect ordering of a certain 97 (past) Sunday, the topic time TOP-T, around the event time of the verb, EV-T, and by the past meaning of the topic time locating it before the matrix utterance time, which is now. This analysis shows that Tense may be morphologically null and semantically unspecified precisely because of the temporal information supplied by the topic time argument and the temporal reference of the determiner. 5.4.4 Habitual Topic Time In English, the habitual reading of an eventive verb is obtained with the generic reading of the present tense. Eng (1991:7) argues that these result from the presence of a generic operator, a "phonologically null operator capable of binding variables unselectively, " which binds the time argument of the verb. In SqwXwu7mish, the habitual reading can be obtained with the pronoun, verb, and a universally quantified adverb, as in (lie). Again, it is not Tense but the habitual reading of the quantified adverb itself which locates the utterance time within the topic time. Thus, the utterance time now is a moment in a day which is part of the set picked out by all days. Similarly, the topic time surrounds the event of eating via the Aspect predicate "around" to locate that event on a day which is part of "all days". 98 ( l l ) e . c h e n l l h e n 7i7Xw s q w a y l l s g . s u b j e a t a l l d a y I e a t e v e r y d a y , l i t . a l l d a y s DW 2 0 - 3 - 9 6 (17) CP C / \ C TP DP TP c h e n UT-T = ZP T" PROn o w / \ T ASP-P i l h e n j / \ TOP-T = ZP ASP' 7i7Xw s k w a y l / N ^ ASP V P e j - a r o u n d / \ EVT-T = ZP VP 0 p r o V eJ Thus, the difference between present moment readings and habitual readings can be ascribed to the specification of the overt universal topic time argument, without recourse to an null generic operator as in Enc (1991). 5.4.5 Conclusions In conclusion, this proposal successfully explains the paradigm in (11) : the pronoun and verb are uninflected for tense but their temporal interpretation is determined by the topic time argument. This argument is specified by temporal adverbials, either in the discourse or in the sentence. Assuming the utterance time is always now in main clauses, it is the semantics of the ZP topic time adverb, its determiner, 99 and the presence of quantifiers which determine the relation of the topic time to the utterance time and thereby account for the absence of morphological tense or semantic specification in the Tense predicate. Finally, the content of the Aspect predicate orders the event time of the verb in relation to the time determined by this topic time argument. 5 . 5 Word Order This sect ion proposes an analysis to account for the preferred r ight peripheral posit ion of the topic time adverb. The der iva t ions of SVO-Adverb and VSO-Adverb orders 4 0 are presented. Assuming the Universal Order hypothesis (Kayne 1994:35), the VSO order i s analyzed as "deriving from SVO order by leftward V-movement"; rightward movement and r igh t adjunction are prohibited. The SVO-Adverb order of the preverbal nominal subject and nominal d i r ec t object in (18) i s unproblematic. The nominal subject i s base generated in the speci f ier o'f VP and moved to the specif ier of the TP, while the object, generated in the complement posi t ion of VP, i s ra ised to the thematic spec i f i e r of AspP41. These movements are assumed to be for Case assignment or feature checking. The verb r a i s e s via head to head movement through Aspect to Tense, as marked by the index "k", with the effect that the adverb in i t s base 4 0The d e r i v a t i o n of the VOS order i s not addressed here because i t d id not a r i s e in t h e da ta e l i c i t e d . P. Jacobs ( p . c . ) s t a t e s t h a t , wi th the except ion of one speaker , VOS i s gene ra l l y d i s p r e f e r r e d . 4^As noted p r e v i o u s l y , t he func t iona l s p e c i f i e r of TP i s comparable t o t h a t of AgrS, and the s p e c i f i e r of AspP to t h a t of AgrO. 100 generated position, the inner specifier of AspP, surfaces on the right periphery: (18) kwelhi slhanay' k'wach-nexw-as ta miXalh kwi chel'aqlh DET lady see-trans-3subj DET bear DET yesterday The lady saw the bear yesterday EL 27-3-97 25 (19) CP C C TP / \ DP TP kwelhi slhanay'i / \ ^ UT-T = ZP T' PRO / \ T ASP-P k'wach-nexw-ask / \ DP ASP-P ta mixalhj / N ^ TOP-T = ZP ASP" kwi chelaqlh y/\^ ASP VP ek-around //N^ EVT-T = ZP VP 0 V ek ej The sentence in (20) exemplifies the verb intial order with two overt nominal arguments: (20) k'wach-nexw-as kwelhi slhanay' ta miXalh kwi chel'aqlh see-trans-3subj DET lady DET bear DET yesterday Yesterday the lady saw the bear EL 27-3-97 28 This order may be derived in one of three ways (see Davis 1997b): by subject adjunction to the right of the VP; by the movement of the verb to T, with the subject staying in its base generated position for the VSO order; or by the movement 101 of the verb to T then to C and the subject to the specifier of TP, as has been analyzed for Celtic languages. This thesis argues for the traditional analysis of movement to C. The VSO order cannot be derived by subject adjunction to the right, if Kayne's (1994) Universal Order hypothesis is assumed. If the verb moves to T and the subject and object remain in their base-generated positions, the adverb would appear between the verb and its subject, which is its most unacceptable position. Therefore, if one assumes Universal Order (Kayne 1994) and predicative Tense and Aspect (Demirdache and Uribe-Etxebarria, to appear), the only account for these facts is to move the verb through Asp and T to C to derive the verb initial order42. The C may have content: in a main clause there is an optional na which cliticizes to the left of the verb; in subordinate clauses, such as nominalized clauses introduced by kwi-s, the verb and the complementizing elements are also cliticized (Kuipers 1967). This movement is indicated in (21) by the index "k"; the subject is raised to the specifier of TP, as marked by "i", and the object to the specifier of AspP, as marked by the index " j " : (20) k'wach-nexw-as kwelhi slhanay' ta miXalh kwi chel'aqlh see-trans-3subj DET lady DET bear DET yesterday Yesterday the lady saw the bear EL:" start from the end" EL 27-3-97 28 ^Adding a Voice projection (Davis to appear) to Demirdache and Uribe-Etxebarria's framework would provide an alternative landing site for the subject, above the topic time adverb, which would allow the verb to raise only to T. This option is outside the scope of the present work. 102 (21) CP C / \ C TP k'wach-nexw-ask / V DP TP kwelhi slhanay'i / N ^ UT-T = ZP T' PRO / \ T ASP-P ek / \ DP ASP-P ta miXalhj / \ TOP-T = ZP ASP' kwi chelaqlh / \ ^ ASP VP e^-around / \ ^ ei VP EVT-T = ZP V 0 ek ej The temporal interpretation of both sentences in this section is identical to that of (lib.), as shown in the time line in (22): (22) TOP-T yesterday < [ [__] ] | > see now EVT-T UTT-T The utterance time now is after yesterday, the inherently past topic time adverb, while the event time of seeing is contained in the topic time. Word order provides another argument for the claim that there is a single position available for the temporal adverbial, high in the tree in the topic structure. Moving 103 the verb and its arguments up accounts for the fact that the surface position of the topic time adverb is the right edge. 5.6 Subordinate Nominalized Clauses This proposal has argued that the adverb at the right edge of a single clause can occupy only one position in the syntax, the specifier of the predicate Aspect. The adverb at the right edge of a bi-clausal sentence exhibits asymmetries in grammaticality, as illustrated in (23) and (24)43: (23)a.*chen tsut [kwi-s nam' huya7 ta Peter kw-i chel'aqlh] lsg.subj say [DET-nom go leave DET Peter DET yesterday] *I said yesterday that Peter would be going away EL 27-3-97 33 b.chen tsut kwi chel'aqlh [kwi-s nam' huya7 ta Peter] lsg.subj say DET yesterday [DET-nom go leave DET Peter] I said yesterday that Peter would be going away EL 27-3-97 34 (24)a.chen tsut [kwi-s-es nam' huya7 ta Peter] kwi chel'aqlh lsg.subj say [DET-nom-3poss go leave DET Peter] DET yesterday I said yesterday that Peter's going away EL 27-3-97 35 b.chen tsut [[kwi-s-es huya7 ta Peter] kwi chel'aqlh] lsg.subj say DET-nom-3poss leave DET Peter DET yesterday I said yesterday that Peter went away I said that Peter went away yesterday EL 27-3-97 36 The sentence in (23a.) is bad because the adverb at the right edge of the sentence can only specify the topic time of the embedded clause, and the past adverb kwi chelaqlh is incompatible with the future auxiliary and verb nam' huyal. In order for the adverb to specify the topic time of the See also Section 4.3 for discussion of this data. 104 matrix verb, it must be moved into the main clause, as in (23b.). In contrast, the adverb at the right edge of the sentence in (24a.) can be interpreted with the matrix verb; in addition, the adverb in (24b.) can be construed either with the matrix verb, or with the past tense nominalized clause. This thesis proposes to account for this intriguing contrast by analyzing the nominalized clause in (23) as a CP whose position is fixed, and the nominalized clause in (24) as a DP, which moves. Section 5.6.1 examines the construal of temporal interpretation in embedded clauses, following Stowell (1993, to appear); Section 5.6.2 introduces nominalized clause and the debate over their syntactic category in Salish. In Section 5.6.3, the word order of the CP nominalized clauses in (23) is analyzed and its effect on temporal interpretation assessed. Finally, in Section 5.6.4, the word order and temporal interpretation of the DP nominalized clauses in (24) is addressed. 5.6.1 The Syntax of Subordinate Clauses Stowell (1993, to appear) proposes that the external argument of the Tense predicate is a temporal ZP-PRO which is determined by Control Theory. In the matrix clause, where it denotes the utterance time, this temporal argument is not c-commanded by another ZP and is therefore not controlled; it has the default meaning "now", the moment of speech. In the subordinate clause, this ZP-PRO is controlled by the closest c-commanding ZP, the event time argument of the main clause. 105 This orders the event time of the embedded clause with respect to the event time of the matrix clause, allowing an account of the "past shifted" reading in the sentence in (25) : (25)a.Peter said [that Vanessa left.] b. EV-Tl UT-Tl say now < | |___> I EV-T2 EV-T1= leave UT-T2 <__, | > The event of Peter's saying is ordered after the event of Vanessa's leaving: the matrix event time, EV-Tl, is the same as the "speech" time of the embedded clause, UT-T2, which is ordered after the embedded event time, EV-T2, by the embedded past tense. 5.6.2 The Syntax of Nominalized Clauses As described in Section 2.4, nominalized clauses are introduced by the determiner kwi, and the predicate or leftmost aspectual clitic of these clauses carries the nominalizer s-, along with the possessive agreement that marks the subject (Jacobs 1992). Jacobs describes their function as sentential complements (either the subject of a nominal predicate or the object of a transitive predicate) or adverbial complements. Davis and Mattewson (1996) propose that the nominalizer, s-, in St'at'imcets is the head, F, of its own functional projection, FP, or Finite Phrase. This head is situated 106 below the determiner and above the verb of the nominalized clause, as shown in (26): (26) DP (Spec) D' (Spec) F' F VP This structure is adopted here for nominalized clauses in SqwXwu7mish. Davis and Matthewson argue that the categories T and D are collapsed in Salish, and furthermore that the deteminer on nominalized clauses is a D, not a complementizer, C. This thesis has glossed the kwi which introduces nominalized clauses in SqwXwu7mish as a determiner for descriptive purposes; however, the next section will show that the only way to explain the difference in the construal of the adverb in the data above is to analyse the kwi as functionally determined: C in (23a.) and as D in (24a.). 5.6.3 Topic Time Adverbs and CP Nominalized Clauses Returning to the facts in (23), repeated here, in sentence (a.) the past tense adverb clashes with the future tense of the nominalized clause, therefore the adverb must be the topic time of the embedded clause. The adverb cannot be the topic time of the main verb because the sentence is ungrammatical. Furthermore, if the adverb at the right edge of .the sentence were the topic time of the main clause, the embedded clause would have to move leftward to derive the 107 surface word order. The ungrammaticality of (23a.) shows that moving the. clause is not possible; if it were, the sentence would be grammatical. The embedded clause is therefore assumed to be a CP, and its inability to move is schematized in (27): (27) DS: V Adv Cp[ FUT ] > SS: * V CP [ FUT ] Adv The order in (27) cannot be generated because CP does not need case under the Case Resistance Principle. In order for the adverb to be interpreted with the main verb, it must appear in the main clause as in (b.): (23)a.*chen tsut [kwi-s nam' huya7 ta Peter kwi chel'aqlh] lsg.subj say [DET-nom go leave DET Peter DET yesterday] *l said yesterday that Peter would be going away EL 27-3-97 33 b.chen tsut kwi chel'aqlh [kwi-s nam' huya7 ta Peter] lsg.subj say DET yesterday [DET-nom go leave DET Peter] I said yesterday that Peter would be going away This data argues for a single position in each clause for the topic time temporal adverb: the adverb in (23a.) must be the topic time of the embedded clause, while the adverb in (23b.) is the topic time of the matrix clause, as illustrated by the structure in (28): 108 (28) TP S\ DP TP chen /N^ UT-T = ZP T' PROnow T / \ ASP-P tsutj //\-TOP-T1 = ZP ASP kwi chelaqlh y ^ \ (CP) ASP ej-around EVT-T1 = 0 <-VP / \ ZPt VP ^ pro <r V eJ < • CP <J c / \ C FP kwi /N^ F' F TP s-nam'k-huya7n / N ^ DP TP ta Peterm / N ^ EV-T1 = ZPt T' PRO / \ T ASP-P / \ EV-T1 = TOP-T2 = ZPt ASP' 0 / \ ASP VP ek/n-towards / / s \ EVT-T2 = ZP VP 0 em V e n The pronominal subject of the main clause is generated the specifier of TP; the main verb tsut 'say' moves through Asp 109 to T, marked by the index "j". The subject of the embedded clause, fca Peter, raises to the specifier of the embedded TP, marked by the index "m" . As discussed in the previous section, the embedded verb must raise high in the clause to precede the subject; here it is analyzed as moving through Asp, where nam' attaches to it, then through T to F, the head projected by the nominalizer which attaches to the auxiliary and verb. The complex under F could feasibly raise to C to attach to kwi, in parallel to the V to C raising analysis for VSO main clauses. The temporal interpretation of this complex sentence is effected via the relations of the temporal arguments, as represented on the time line in (29): (29) T0P-T1 yesterday < [ [__] ] | > say now EV-Tl UT-T I EV-Tl = T0P-T2 say say <_[ ]_>[___]_> go away EV-T2 The empty external argument of Tense in the subordinate clause is controlled by the event time of the matrix predicate, EV-Tl. The embedded topic time, T0P-T2, is itself controlled by EV-Tl, therefore both refer to the time of the action say; in this way, the embedded event time, EV-T2, is directly ordered with respect to the matrix event time, EV-Tl. Thus, Peter's going away may be located in three possible relations to now, as indicated by the speaker's 110 translation "would": his going away may be "yesterday, after the event of saying," or else "today, before the utterance time," or else "today, after the utterance time". 5.6.4 Topic Time Adverbs and DP Nominalized Clauses The adverbial at the right edge of the sentence can be interpreted with the main verb only if third person possessive agreement is attached to the nominalizer44, as in (24) : (24)a.chen t s u t [kwi-s-es nam' huya7 t a Pe te r ] kwi c h e l ' a q l h l s g . s u b j say [DET-nom-3poss go leave DET Peter ] DET yes te rday I s a id yes te rday t ha t P e t e r ' s going away EL 27-3-97 35 b .chen t s u t [ [kwi -s -es huya7 t a Pe ter ] kwi c h e l ' a q l h ] l s g . s u b j say DET-nom-3poss leave DET Peter DET yes t e rday I s a id yes te rday [ tha t Peter went away] I s a id [ tha t Pe ter went away] yes te rday EL 27-3-97 36 The t o p i c time adverb in (24a.) i s s t i l l incompat ib le wi th the fu tu re a u x i l i a r y of the nominalized c l a u s e , but i n t h i s case the adverb can be i n t e r p r e t e d with the mat r ix verb from the r i g h t edge of the sen tence . The d e s c r i p t i v e d i f f e r e n c e between ( 2 3 a . ) , which i s bad, and (24a . ) , which i s good, i s the posses s ive agreement on kwi. This a n a l y s i s proposes t ha t possess ive agreement on kwi makes the nominalized c lause a DP, which can move, whi le the absence of pos se s s ive agreement on kwi makes i t a CP, which cannot move. The ca tegory of the nominalized c l ause in the 4 4 P . Jacobs ( p . c . ) says t h a t t he t h i r d person agreement in t h e c l a u s e s in (23) i s an -s on the a u x i l i a r y nam' which i s e l i d e d ; t h e p resence of p o s s e s s i v e agreement on the kwi i t s e l f i s then what d e t e r m i n e s i t s ca tegory as D. I l l sentence in (24a.) is therefore analyzed as a DP between the matrix verb and the adverb, as schematized in (30): (30) DS: V Adv DP [ FUT ] > SS: V V DP [ FUT ] Adv Furthermore, with a past nominalized clause in (24b.), the adverb at the right edge can be construed either with the embedded verb or with the main verb. This ambiguity is explained by the analysis of the adverbial as the topic time argument: its interpretation with the matrix verb, the preferred reading, is a result of the adverb's status as the matrix topic time argument and the movement of the DP nominalized clause; its interpretation with the embedded verb is the result of the adverb's position as topic time of the nominalized clause, which also moves string vaccuously. The position of the adverb in the four sentences is summarized in (31) : (31) 23a. * V Cp[ FUT Adv] 23b. V Adv Cp[ FUT ] 24a. DS: V Adv DP[ FUT ] --> SS: V DP[ FUT ] Adv 24b. V DP[ PAST Adv] or V DP[ PAST ] Adv The structure of (24a.) is given in (32). In this case, the adverb is the topic time of the main clause, while the DP nominalized clause is the object, complement of the main verb. It moves to the specifier of Aspect in the same way as any object DP. Within the DP clause, the subject ta Peter moves to the specifier of the embedded TP, as indicated by the index "m"; the verb huyal moves to Asp, then the verb and the auxiliary nam' move through T to F, as they did in (28). 112 (24)a.chen tsut kwi-s-es nam' huya7 ta Peter kwi chel'aqlh lsg say DET-nom-3poss go leave DET Peter DET yesterday I said yesterday that Peter's going away. (32) TP / \ DP chen UT-T = ZP TP / \ m / PROnow / \ s - e s T ASP-P t su t j / / \ s • » > DPk ASP-P T / \ T 0 P - T 1 = ZP ASP kwi c h e l a q l h / X ^ f ASP e j - a r o u n d I s EVT-T1 = (DPk) / \ D ' / \ D FP kwi /\^ T 1 r F TP -nam 'p -huya7 n / \ DP TP VP / \ ZPt VP 0 / \ p r o V / \ e j ek ta Peterm / \ UT-T1 = ZPt PRO T T ' / \ ASP-P en / \ U T - T l = TOP-T2 ep /n-= ZPt ASP' 0 / \ ASP VP -towards / \ EVT-T2 = ZP ' em V en 113 The English translation of the sentence analyzed in (32) has an unambiguous future meaning. This falls out from the movement of the clause; the external argument of the embedded Tense predicate and the embedded topic time are controlled by the utterance time of the matrix clause to give the reading that now is "towards" the event of Peter's going away. The matrix event time yesterday orders the event time of the matrix verb say in the past. This is represented on the time line in (33): (33) TOP-Tl yesterday <___[ [__] ] | > say now EV-Tl UT-Tl I UT-Tl = T0P-T2 now now <_[ ]_>[___]_> go away EV-T2 5.6.5 Conclusions The analysis of SqwXwu7mish adverbs as the topic time argument predicts that they can occupy a single position in the clause. This accounts for the existence of constraints on the interpretation of adverbs in subordinate clauses as seen in (23) and (24). The distinction of CP nominalized clauses, which have possessive agreement on the auxiliary and which do not move, and DP nominalized clauses, which have their agreement on the determiner kwi and which do move, explains the constraints on the interpretation of the adverbs. This distinction also accounts for a difference in the temporal interpretation of the two types of clause. 114 5.7 Temporal Adverbs and Quantification In Section 3.2, it was claimed that adverbs which denote the topic time must take scope over other quantifiers. Under the analyses proposed in this chapter, however, the adverb would always be c-commanded by and structurally in the scope of the other quantifiers. In order to account for the topic time reading of the adverb, a non-structural analysis of scope is adopted (Demirdache and Matthewson 1997). The data from Section 3.2 is repeated in (34) and (35): (34)a.na 7upen sXalhnat kwi-s-es ts'its'ap' ta Peter RL ten Sunday DET-nom-3poss work DET Peter Peter worked on ten Sundays/Ten Sundays Peter worked EL 10-4-97 22 b.na 7upen-alh s-es ts'its'ap' ta Peter (t) kwetsi sXalhnat RL ten-times nom-3poss work DET Peter (obi) DET Sunday Peter worked ten times on a Sunday EL 10-4-97 19 (35)a.na-malh-eq' sXalhnat 7i nam' 7i-7imash 7i7Xw ta sta7exwlh RL-well-FUT Sunday PR go red-walk all DET children . It will be Sunday when they go for a walk. EL 13-8-97 b.7i7Xw ta sta7exwlh nam' 7i-7imash na7 ta sXalhnat all DET children go red-walk RL DET Sunday All the children/every child went walking on Sunday ie they all go on the same Sunday EL 2-4-97 16 The (a.) sentences, in which the adverb is focused syntactically as a main predicate for the event time reading, get the distributive reading of the quantifier, whereas the (b.) sentences, in which the adverb at the right edge gets a topic time reading, get the collective reading. In addition, the event time reading of the adverb in (35a.) may have a 115 collective event reading which is identical to the topic time reading in (35b.). In this way, temporal adverbs in SqwXwu7mish pattern like thematic arguments in St'at'imcets45. Demirdache and Matthewson (1997) show that a transitive sentence with a plural subject in St'at'imcets never gets a distributive reading, as in (36): (36) [ q u s - e n - i t - a s ] np[s-Rosa muta7 s-Tanya] Dp[i kalh61hs-a mixalh] qus-an-it-as [ s-Rosa muta? s-Tanya ] [?i ka4a4s-a mixa*l s h o o t - t r - p l - e r g [nom-Rosa and nom-Tanya] [DET three-DET bear] Rosa and Tanya shot three bears. (RW) This sentence can get a collective reading, that is, Rosa and Tanya used one gun together to shoot three bears, or a "cumulative" reading, that is, Rosa and Tanya each had a gun and shot three bears between them, but it cannot have the distributive reading whereby each girl shot three bears for a total of six. In (36), therefore, three bears is referentially independent. In order to get a distributive reading, along with a collective and a cumulative reading, the object must be the main predicate of the sentence. The authors use these facts to argue for a non-structural analysis of scope in St'at'imcets. The ZP topic time argument is referentially independent, like its nominal counterpart the DP, and must take wide scope due to the deictic properties of the determiner kwetsi. This 4 5 H. Demirdache ( p . c . ) i s e n t i r e l y r e spons ib l e for t h i s o b s e r v a t i o n . 116 explains the scope relations between the cardinal quantifier and the topic time adverb in (34) . The topic time adverb should take scope over the quantifier to give the collective reading in (34b), but as the analysis in (37) shows, the cardinal adverb is the main predicate and the topic time adverb is in the nominalized clause that is structurally its complement. The adverb on a Sunday does not therefore take structural scope over the cardinal; in fact, the cardinal quantifier appears to take scope over the adverb. (34)b.na 7upen-alh s-es ts'its'ap ta Peter kwetsi sXalhnat RL ten-times nom-3poss work DET Peter DET Sunday Peter worked ten times/for ten hours on a Sunday (37) NP N FP 7upen-alh / \ . F' / \ F TP s-esm-ts'its'ap'n / \ DP TP ta Peterm / A\ s UT-T = Z P T 1 PROnow y / \ T ASP-P en / \ ASP-P / \ TOP-T = ZP ASP' kwetsi sXalhnat // s\ ASP VP ©n-around / \ EVT-T = ZP VP 0 em V / en 117 The reading of the sentence can be paraphrased as:"the set of times that Peter worked on Sunday has a cardinality of ten" (H. Demirdache, p.c). The nominalized clause, a bare FP with no determiner kwi, is analyzed as the unaccusative subject46 of the cardinal predicate 7upenalh. Within the FP, the subject ta Peter is generated in the specifier of the embedded clause then raised to the specifier of the TP; this DP is coreferential with the possessive agreement generated with the nominalizer in F, marked accordingly by the index "m" . The verb, indexed "n", raises via head movement to attach to this "inflected nominalizer" in F. The temporal interpretation of this sentence is as follows: because the utterance time of the nominalized clause has no event time to bind it, it gets the default interpretation "now"; now is before the past time specified by the topic time adverb kwetsi sXalhnat; this topic time is ordered "around" the event time of working by the predicate Aspect. Therefore, this sentence gets the interpretation: "Peter's working is in the past on a Sunday, and it has a cardinality of ten". 5.8 Conclusions This chapter has argued for the proposal that SqwXwu7mish adverbials are the topic time argument of the predicate Aspect within the syntax of Tense and Aspect developed by Demirdache and Uribe-Etxebarria (to appear). 4 6H. Davis (p.c.) says that in general there are no sentential subjects in Salish, and that this is a "defective propositional complement." P. Jacobs (p.c.) counters that sentential subjects certainly do exist in Coast Salish. 118 This proposal provides a straightforward account of the distribution and interpretation of a range of data discussed in Chapters 3 and 4. In constructions with a pronoun and verb uninflected for tense, the adverb locates the event in time (section 5.4). The restricted position of the adverbial at the right periphery and its status as topic time argument argues for the derivation of verb initial word order in SqwXwu7mish as movement to C (Section 5.5). The inability of right edge adverbs to be construed as the topic time of the main clause from the right edge of a complement subordinate clause is explained by the single position per clause available to the adverb (Section 5.6). Finally, the interaction of topic time adverbs with quantified subjects and cardinals supports a non-structural analysis of scope (Section 5.7). The implications of these conclusions for the analysis of the syntax of adverbs are taken up in the concluding chapter. 119 6 . Implications, Further Issues and Conclusions 6.1 Implications of the Analysis This analysis has implications for the analysis of adverbs cross-linguistically, for the study of Salish in terms of the Common Ground Parameter (Matthewson 1996), and for the parallelism of nominal and temporal reference. The analysis of SqwXwu7mish adverbs as the topic time argument of Tense and Aspect predicates has implications for the analysis of adverbs cross linguistically. It contravenes the assumption that adverbs in English are simply adjoined to the event time (Stowell to appear a), modifying either the ZP variable in the outer specifier position of VP, as in (la.), or the ZP directly dominating the VP as in (lb.): b. ZP Adv ZP / \ Z' / \ Zi VP / \ ZP V ei / \ V Indeed, Cinque (in progress) argues from Romance data that adverbs are not adjuncts but the heads of adverb phrases which "fill the unique Spec position of a distinct maximal projection." Demirdache and Matthewson (1995) argue that overt nominals cannot be topics. For Matthewson (1996), this follows from the semantic property of Salish determiners that 120 they are non-presuppositional. Presupposition requires access to the common ground of the discourse, that is, the beliefs and assumptions of both the speaker and the hearer. From this, Matthewson (199 6:124) proposes the Common Ground Parameter, shown in (2): (2) Determiners may access the common ground: Yes: {English,...} No: {Salish, ...} Therefore, a nominal which has no access to the common ground cannot be a topic. Matthewson claims that the only examples of presupposition in Salish are syntactic structures which induce it, such as cleft constructions and syntactic nominalizations. However, this thesis has argued that phrasal temporal adverbs in SqwXwu7mish, which have determiners, are canonically topics. Indeed, the "key to the interpretation of time adverbials resides in recognizing their presuppositional character" (de Swart to appear:l). The key to reconciling the analysis in this thesis with the Common Ground Parameter lies in the relationship of nominal and temporal reference. Partee (1973) claims that temporal adverbs can serve as discourse or sentential antecedents of tense in a manner analoguous to nominals and pronouns. She points out that in English, this analogy is not strict since temporal adverbials are always accompanied by tense morphology, whereas noun phrases normally do not appear with coreferential pronouns. She notes that language X, which marks subject and object on the verb with pronominal agreement, "whether or not the 121 subject (or object) is overtly expressed," and also marks tense in every sentence, has "more parallel (...) tense and pronoun systems." Language Y, which like English does not use pronoun agreement and overt nominals together, "could also have more parallel tense and pronoun systems if it omitted the tense morpheme in clauses containing an explicit time adverbial" (Partee 1973:604). These possibilities are summarized in (3): (3) tense + adverb *tense + adverb pronoun + nominal X SqwXwu7mish pronoun + *nominal English Y Under the present proposal, SqwXwu7mish is the fourth possibility, a mirror image of English; it is a language which obligatorily marks pronominal agreement on the verb with or without an overt nominal, yet which does not grammatically mark tense with a temporal adverbial.47 6.2 Further Issues Temporal interpretation in SqwXwu7mish is the focus of topics requiring further investigation. The analyses proposed in this thesis assumed that Tense is a predicate which lacks morphological content and which is semantically unspecified. The inherent temporal meaning of the topic time argument or its determiner and the default meaning of the utterance time enable SqwXwu7mmish sentences ^Furthermore, if SqwXwu7mish temporal adverbs are arguments as proposed, they should have a pro counterpart which would be subject to the Binding Theory (H. Davis, p . c ) . 122 to be temporally interpreted without recourse to Tense. This assumption remains to be evaluated in terms of the SqwXwu7mish temporal system of clitics outlined in Chapter 2. It is clear that the meanings of the morphemes 7i, t and eg' are not "present", "past" and "future" respectively, as previously described (Kuipers 1967, Jacobs 1992). Furthermore, the position classes within the temporal/aspectual clitic system require elucidation. This research will help to define what is a1 possible "tense" is in SqwXwu7mish. Inner aspect or "aktionsart" needs closer examination. The temporal interpretation of statives differs from that of eventives: states usually contain the reference time, while events precede, follow or are contained in it (Kamp and Reyle 1993). Stative verbs in SqwXwu7mish pronoun-verb constructions are never interpreted in the past, only in the present, while achievement and accomplishment verbs get only a past reading: (4) a. chen, lhq'i7-s lsg.subj know-caus I know it EL 8-5-96 I know it at this very moment YJ 13-3-96 I know DW 20-3-96 b.chen wi7xw-em lsg.subj drop-intrans I fell YJ 13-3-96; YJ 3-4-96 c.chen mi-s lsg.subj come-caus I brought (it) YJ 13-3-96 123 However, the addition of future time adverbs, as illustrated in (5), causes all the sentences to receive future interpretations: (5)a.chen es-lhq'i7-s ti taXw sqwayl lsg.subj stat-know-caus DET noon I'll know at noon (spoken in the morning) EL 8-5-96 b.chen hiy'am kwayl-es lsg.subj arrive tomorrow I will arrive tomorrow YJ 6-6-96 c.chen mi-s kwayl-es lsg.subj come-caus tomorrow I'm going to bring it tomorrow YJ 6-6-96 Thus, the ability of the temporal adverbial to determine the topic time is not sensitive to the aspectual class of the verb, whereas the "zero" topic time provided in the discourse clearly is. The role of the determiner system in temporal interpretation in SqwXwu7mish requires further investigation in terms of the effect of the determiner on the interpretation of nouns and of adverbials. The determiner has been shown to provide the time of sentences with bare verbal stems in Bella Coola (Davis and Saunders 1974) and to fix the "predication time" of a noun and/or of the whole sentence in St'at'imcets (Demirdache 1997). As seen in SqwXwu7mish, the determiner can determine the temporal interpretation of a locating adverb48, as in (6) : 48In Section 5.3.2, the analysis of temporal adverbs as ZP's offered a possible explanation for these data, but it remains to be tested. 124 (6) ti siyelanem "this year" kwi siyelaiiem "last year" ta hem'i siyelanem "next year" The determiner ti is interpreted as "present" with the adverb; with nominals, it is interpreted as "this X right here." The determiner kwi is often used to locate the time specified by an adverb as being in the past, but is also used in conjunction with the modal operator way'ti 'maybe'. The determiner ta, translated as "that" with nominals, can have a past interpretation with some adverbs yet is used with the adjective hem'i 'next' for the future in (6). This use of the determiner with the adverb is not attested in all Salish languages. Finally, this analysis has examined only phrasal temporal adverbials. In Interior languages such as St'at'imcets, temporal adverbials are clausal and have a freer distribution (H. Davis p . c ) . Adverbial subordinate clauses constitute a large • number of interesting constructions in SqwXwu7mish and their analysis will have important implications for the proposal developed here. 6.3 Conclusions This thesis has argued for the status of Reichenbach's third time, reference time, as a third temporal argument of the temporal predicate Aspect. Furthermore, the claim that adverbs in Squamish are the overt realization of this "topic time" argument has been supported by the semantic and syntactic properties of these adverbs. Squamish adverbs are 125 part of the topic structure of the sentence unless syntactically focused as main predicates; they do not act like adjuncts with respect to word order or cooccurrence, nor like "functional" adverbs in terms of cleft constructions. Their analysis as temporal arguments helps to explain how temporal interpretation is achieved in a language without grammaticalized tense. 126 Abbreviations AUX=auxiliary; lsg.subj=first person singular subject agreement; lpl.subj= first person plural subject agreement; lposs=first person possessive agreement; 2sg.subj=second person singular subject agreement; 3subj=third person subject; 3sg=third person singular; 3poss=third person possessive agreement; 3pl.=third person plural; caus=causative; DET=determiner; DEM=demonstrative; dir=direction marker; DR=durative; FOC=focus marker; FUT=future; INT=interrogative; intrans=intransitivizer; IRR=irrealis; LOC=locative; NEG=negat ion; nom=nominali zer; obl^oblique; PT=past; PR=present; REC=reciprocal RL=realis; REL=relative; red=reduplicated morpheme; TD=temporal deicitic; trans=transitivizer; WH=wh-complementizer. 127 Key t o SqwXwu7mish Orthography orthography P P' m t t' ts ts' ch ch' s sh tl' lh 1 k k' phonemic script P > P m t t c > c c c s V s > 4 1 k * o r t h o g r a p h y p h o n e m i c s c r i p t kw k'w xw q q' qw q'w h w y 7 a e i u x XW ,w V xw q q qv qv h w y ? a a e o X X V 128 References Binnick, R. I. 1991. Time and the Verb: A Guide to Tense and Aspect. Oxford University Press, Oxford. Bybee, J., R. Perkins and W. Pagliuca. 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. University of Chicago Press, Chicago. Chomsky, N. 1995. The Minimalist Program. MIT Press, Cambridge MA. Chung, S. and A. Timberlake. 1985. 'Tense, Aspect, and Mood' in T. Shopen, ed. Language Typology and Syntactic Description Volume III, 202-257, Cambridge University Press, Cambridge. Cinque, Guglielmo. in progress. Adverbs and the Universal Hierarchy of Functional Projections, ms. Comrie, B. 1985. Tense. Cambridge University Press, Cambridge. Crystal, D. 1987. The Cambridge Encyclopedia of Language. Cambridge University Press, Cambridge. Currie, E. 1996, 1997. SqwXwu7mish Fieldnotes. Davis, Henry. 1996. 'Remarks on Nominalization (in St 'at' imcets) . ' Paper presented at MIT, Cambridge MA. Davis, H. 1997a. 'Against the Clitic String.' Paper presented at the 5th Salish Morphosyntax Workshop, University of Victoria, Victoria B.C. Davis, H. 1997b. 'Remarks on Word Order and Scrambling in St'cit' imcets. ' Paper presented at the 5th Salish Morphosyntax Workshop, University of Victoria, Victoria B.C. Davis, H. to appear. 'On Agreement in St' cit ' imcets (Lillooet Salish)' in Proceedings of Langue et Grammaire II, Universite de Paris VIII. Davis, H. and L. Matthewson. 1996. 'Subordinate Clauses and Functional Projections in St'at'imcets' in Papers for the 31st International Conference on Salish and Neighbouring Languages, 59-73. University of British Columbia, Vancouver B.C. 129 Davis, P. and R. Saunders. 1974. 'Bella Coola Nominal Deixis.' Paper presented to the Ninth International Conference on Salish Languages, University of British Columbia, Vancouver B.C. D^chaine, Rose-Marie. 1993. Predicates Across Categories: Towards a Category-Neutral Syntax. PhD. dissertation, Graduate Linguistic Student Association, University of Massachusetts, Amherst MA. Demirdache, Hamida. 1997. 'Predication Times in St'£t'imcets (Lillooet•Salish).' Paper presented at the Texas Linguistic Society. Demirdache, Hamida, Dwight Gardiner, Peter Jacobs, and Lisa Matthewson. 1994. 'The Case for D-Quantification in Salish: "All" in St'at'imcets, Squamish, and Secwepemctsin' in Papers for the 29th International Conference on Salish and Neighbouring Languages, 145-203. Salish Kootenai College, Pablo Montana. Demirdache, Hamida and Lisa Matthewson. 1995. 'Quantifier-Raising and Focus-Topic Structure in St'at'imcets Salish.' Paper presented at the Linguistic Society of America Annual Meeting, New Orleans, Louisiana. Demirdache Hamida and Lisa; Matthewson. 1997. 'On Ways of Expressing Scope: Distributivity in Salish.' Paper presented at the University of British Columbia Colloquium. Demirdache, Hamida and Myriam Uribe-Etxebarria. to appear a. 'The Syntax of Temporal Relations: a Uniform Approach to Tense and Aspect' in Proceedings of the Sixteenth West Coast Conference on Formal Linguistics, University of Washington, Seattle WA. Demirdache, Hamida and Myriam Uribe-Etxebarria. to appear b. 'The Primitives of Temporal Relations' in R. Martin, D. Michaels, and J. Uriagereka, eds. Papers in Honour of Howard Lasnik. MIT Press, Cambridge MA. Dowty, David. 1979. Word meaning and Montague Grammar. D. Reidel Publishing Company, Dordrecht Holland. Eng, Murvet. 1986. Tense without Scope: An Analysis of Nouns as Indexicals. Indiana University Linguistics Club, Bloomington Indiana. Eng, Murvet. 1987. 'Anchoring Conditions for Tense' in Linguistic Inquiry, Volume 18, Number 4, 633-657. 130 Enc, Miirvet. 1991. 'On the Absence of the Present Tense Morpheme in English1. ms. , University of Wisconsin, Madison. Erteschik-Shir, N. 1993. The Dynamics of Focus Structure. ms. , Ben Gurion University of the Negev, Beer Sheva Israel. Gerdts, Donna. 1988. Object and Absolutive in Halkomelem Salish. Garland, New York. Giorgi, Alessandra and Fabio Pianesi. 1991. 'Toward a Syntax of Temporal Representations' in Probus 3.2, 187-213. Haegeman, L. 1991. Introduction to Government and Binding Theory. Blackwell, Cambridge MA. Higginbotham, James. 1985. 'On Semantics' in Linguistic Inquiry, Volume 16, Number 4, 547-593. Higginbotham, James-. to appear. 'Semantic Computation' in Natural Language Semantics. Hill-Tout, C. 1900. 'Notes on the Skqo'mic of British Columbia, a Branch of the Great Salish Stock of North America,' Report 70 BAAS 1900, 472-549. Hitzeman, Janet. 1995. 'A Reichenbachian Account of the Interaction of the Present Perfect with Temporal Adverbials' in Proceedings of the North East Linguistics Society 25, University of Pennsylvania, Philadelphia PN. Hornstein, Norbert. 1990. As Time Goes By: Tense and Universal Grammar. MIT Press, Cambridge MA. Jacobs, Peter William. 1992. Subordinate Clauses in Squamish: A Coast Salish Language. MA Thesis, University of Oregon, Eugene OR. Jager, Gerhard. to appear. 'Topic, Scrambling, and Aktionsart' in Proceedings of CONSOLE III. Jelinek, Eloise. 1993. 'Definiteness and Second Position Clitics in Straits Salish,' ms., University of Arizona. Johnson, Marion. 1981. 'A Unified Temporal Theory of Tense and Aspect' in P. Tedeschi and A. Zaenen, eds., Syntax and Semantics Volume 14: Tense and Aspect, Academic Press, Toronto ONT, 145-175. Kamp, Hans and Uwe Reyle. 1993. From Discourse to Logic. Kluwer, Dordrecht. 131 Kayne, Richard . 1994. The Antisymmetry of Syntax. MIT P res s , Cambridge MA. K l e i n , Wolfgang. 1994. Time in Language. Rou t l edge , London. K l e i n , Wolfgang. 1995. 'A T i m e - R e l a t i o n a l A n a l y s i s of Russian Aspect ' in Language, Volume 71 , Number 4, 669-695. Kroeber, Paul . 1991. Comparative Syntax of Subordination in Salish. PhD D i s s e r t a t i o n , U n i v e r s i t y of Chicago , Chicago. Kuipers , Aer t . 1967. The Squamish Language: Grammar, Texts, Dictionary. Mouton &Co, The Hague. . La r son , .R icha rd 1988. 'On the Double Object C o n s t r u c t i o n ' in Linguistic Inquiry, Volume 19, 335-91. Larson, Richard and Gabr i e l Sega l . 1995. Knowledge of Meaning: An Introduction to Semantic Theory. MIT P r e s s , Cambridge MA. Lasnik, H. and J . Uriagereka. 1988. A Course in GB Syntax. MIT Press , Cambridge MA. Mat thewson, L i s a . 1996. Determiner Systems and Quantificational Strategies: Evidence from Salish. Ph.D. d i s s e r t a t i o n , U n i v e r s i t y of B r i t i s h Columbia, Vancouver B.C. P a r t e e , Barbara . 1973. 'Some S t r u c t u r a l Analogies between Tenses and Pronouns in E n g l i s h ' i n The Journal of Philosophy, Volume 70, Number 18, 601-609. P a r t e e , Barbara . 1985. 'Binding I m p l i c i t V a r i a b l e s i n Quant i f ied Contexts ' in Chicago Linguistics Society 25. P a r t e e , Barbara . 1991a. 'Topic, Focus and Q u a n t i f i c a t i o n ' in SALT I, Cornel l Un ive r s i ty . P a r t e e , Barbara. 1991b. 'Adverbial Quan t i f i c a t i on and Event S t r u c t u r e s ' in Proceedings of the Seventeenth Annual Meeting of the Berkeley Linguistics Society, Volume 17, 439-456. Reichenbach, Hans. 1947. Elements of Symbolic Logic. The Free Press , New York. van Riemsdijk, H. and E. Wil l iams. 1986. Introduction to the Theory of Grammar. MIT Press , Cambridge MA. 132 R e i n h a r t , Tanya . 1983 . Anaphora and Semantic Interpretation. The U n i v e r s i t y of Chicago P r e s s , Chicago. Robins , R.H. 1990. A Short History of Linguistics. Longman, New York. Rooth, M. 1992. 'A Theory of Focus I n t e r p r e t a t i o n ' i n Natural Language Semantics 1, 75-116. Shaer, Benjamin. 1996. Making Sense of Tense: Tense, Time Reference and Linking Theory. PhD D i s s e r t a t i o n ( d r a f t ) , McGill Un ive r s i ty , Montreal . Speas, Margaret . 1991. 'Func t iona l Heads and I n f l e c t i o n a l Morphemes' in The Linguistic Review 8. Stowel l , Tim. 1993. 'Syntax of T e n s e ' . ms, U n i v e r s i t y of C a l i f o r n i a , Los Angeles. S towel l , Tim. to appear a. 'The Phrase S t r u c t u r e of Tense ' in L. Zaring and J . Rooryck, e d s . , Phrase Structure and the Lexicon, Kluwer, Dordrecht. S towel l , Tim. t o appear b . 'What do the Presen t and Past Tenses Mean?" in P-M. B e r t i n e t t o and M.Squar t in i , e d s . , Proceedings of the Cortona Tense-Aspect Meeting, Scuola Normale Super iore , P i sa . Swadesh, M. 1950. ' S a l i s h I n t e r n a l R e l a t i o n s h i p s ' i n International Journal of American Linguistics, Volume 16, 157-167. de Swart, H e n r i e t t e . to appear. ' P o s i t i o n and Meaning: Time Adverbia ls in Context ' in P. Bosch and R. van der Sandt, e d s . , Proceedings of the Conference on Focus & Natural Language Processing. Thompson, E l l e n . 1994a. 'The S t r u c t u r e of Tense and t h e Syntax of Temporal Adverbs ' in Proceedings of the Thirteenth West Coast Conference on Formal Linguistics. Thompson, E l l e n . 1994b. 'A Minimal i s t Approach t o t h e Syntax of Temporal Adverbia ls ' in University of Maryland Working Papers in Linguistics 2, 221-237. Zagona, Karen . 1989. 'On t h e Non-Isomorphism of Morphological Tense and Temporal I n t e r p r e t a t i o n ' in C. Ki r schner and J . DeCesaris , e d s . , Studies in Romance Linguistics, John Benjamins, Amsterdam & P h i l a d e l p h i a , 479-494. 133 Zagona, Karen . 1990. 'Times as Temporal Argument S t r u c t u r e . " ms. Paper p resen ted a t the Time in Language conference, MIT, Cambridge MA. Zagona, Karen. 1992. 'Tense-binding and the Cons t rua l of P r e s e n t Tense ' in C. Laeufer and T. Morgan, e d s . , Current Issues in Linguistic Theory, Volume 74: Theoretical Analyses in Romance Lingusitics, John Benjamins, Amsterdam & Phi lade lph ia , 385-398. 134 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0088237/manifest

Comment

Related Items