The Open Collections website will be unavailable July 27 from 2100-2200 PST ahead of planned usability and performance enhancements on July 28. More information here.

West Coast Conference on Formal Linguistics (WCCFL) (38th : 2020)

Accounting for variation in number agreement in Icelandic DAT-NOM constructions Hoover, Jacob Louis 2020

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


73804-Hoover_Jacob_Icelandic_WCCFL38_2020_handout.pdf [ 138kB ]
JSON: 73804-1.0389856.json
JSON-LD: 73804-1.0389856-ld.json
RDF/XML (Pretty): 73804-1.0389856-rdf.xml
RDF/JSON: 73804-1.0389856-rdf.json
Turtle: 73804-1.0389856-turtle.txt
N-Triples: 73804-1.0389856-rdf-ntriples.txt
Original Record: 73804-1.0389856-source.json
Full Text

Full Text

Accounting for variation in number agreement in Icelandic daॼ–nom constructionsJacob Louis Hoover∗McGill Universityjacob.hoover@mail.mcgill.ca1 IntroductionNormally Icelandic verbs agree with their subjects, which show nominative case.(1) SUBJ.3pl.nom VERB.3pl OBJ.3ॻg.acc✔However, there are constructions in which the subject takes the dative case, and theobject takes nominative. In these constructions, only the nominative object maycontrol agreement.(2) SUBJ.3ॻg.daॼ VERB.3pl OBJ.3pl.nom (Icelandic A)✔Additionally, these daॼ঄nom constructions demonstrate an e੘ect called the PersonRestriction: 1st/2nd person nominatives can’t control agreement.(3) SUBJ.3ॻg.daॼ VERB.2pl OBJ.2pl.nom✘࠮e puzzle: variation and interventionWith 3rd person nominatives, there is variation between speakers in whether or notthere is number agreement. This variation has been described in terms of there beingthree varieties of Icelandic,1 breaking down in the following way:With the dative in the canonical subject position (࠸࠵ࡈࡏVࡏࡂࡃࡁ), the following agree-ment:∗I would like to thank Jessica Coon for guidance and advice, and for her syntax course which launchedthis project. Thanks also to Stefan Keine, and the members of the McGill Syntax-Semantics reading groupfor helpful discussion and comments. This handout can be found at described in detail by Sigurðsson and Holmberg (2008). See also Taraldsen (1995), Holmberg andHróarsdóttir (2003), and Ussery (2017).(4) SUBJ.daॼ VERB.8><>:3pl Icelandic A3ॻg/3pl Icelandic B3ॻg Icelandic C1CA OBJ.3pl.nomOptionality in IcelandicB disappears when the dative intervenes (࠹ࡌࡄࡀࡏVࡏ࠸࠵ࡈࡏࡂࡃࡁ):(5) VERB.{3pl Icelandic A3ॻg Icelandic B,C)SUBJ.daॼ OBJ.3pl.nomThe following table summarizes the variation in number-agreement with 3rd personnominatives:(6)࠸࠵ࡈࡏVࡏࡂࡃࡁ.3ࡄࡀ ࠹ࡌࡄࡀࡏVࡏ࠸࠵ࡈࡏࡂࡃࡁ.3ࡄࡀIcelandic A num agree yes yesIcelandic B num agree optional noIcelandic C num agree no noProposalDi੘erences between varieties can be explained by two independent parameters:1. the order of probing and movement operations, and2. whether there is a [nॽm] feature visible on dative DPs.Why care?• Principled account of an interspeaker syntactic variation phenomenon.• Explain the syncretism ࣴx, wherein derivations expected to a crash due to com-petition become acceptable when competing agreement forms happen to bephonologically identical.1WCCFL 38, 7 March 2020 Variation in Icelandic daॼ–nom number agreement Hoover 22 Data: agreement in ࠸࠵ࡈ–ࡂࡃࡁ constructionsThere is variation in whether or not the verb agrees for number with nom, but whenthere is agreement, it is with the nominative object, never with the dative subject.2(7) Honumhim.daॼlíkalike.3plþeir.they.nom‘He likes them’ (Icelandic A) Sigurðsson and Holmberg, 20083࠮e Person Restriction: no agree with 1st/2nd nom object (all varieties).(8) * Honumhim.daॼlíkumlike.1plvið.we.nomintended: ‘He likes us’(9) * Honumhim.daॼlíkiðlike.2plþið.you.nom.plintended: ‘He likes you(pl)’These constructions are reported to be simply ineࣳable with 1st/2nd personnominatives. They cannot be saved by resorting to a default 3ॻg form.a(10) * Honumhim.daॼlíkarlike.3ॻgvið/þið.us/you.nom.plintended: ‘He likes us/you(pl)’aSigurðsson (1996) reports that there exist speakers forwhomdefault agreementmay be available.Also note, a 3ॻg verb form is available in complex ECM constructions, when the verbal complementis an entire phrase (see §A.2). But this option is not available in simplex constructions like (10).The Person Restriction is in e੘ect for all three varieties, singular and plural, indepen-dent of word order. That is, whether the dative is high (࠸࠵ࡈࡏVࡏࡂࡃࡁ) or low (࠹ࡌࡄࡀࡏVࡏ࠸࠵ࡈࡏࡂࡃࡁ), agreement with non-3rd person nominatives is unacceptable.2There is signiਖ਼cant evidence in the literature establishing that dative subjects are indeed subjectsin these constructions and likewise that nominative objects are indeed objects (as discussed in Zaenen,Maling, and Thráinsson (1985) and others: See Bobaljik (2006) for a summary).3Icelandic examples are from this source, unless otherwise stated.2.1 Diࣳerences between the three varietiesWith a 3rd person nominative object, the Person Restriction does not apply, and thereis 3rd person agreement on the verb. Number agreement di੘ers by variety.4(11) ࠸࠵ࡈ-V-ࡂࡃࡁ constructionsa. Icelandic A: number-agree requiredaðthathenniher.daॼlíkuðuliked.3plþeirthey.nomb. Icelandic B: number-agree optionalaðthathenniher.daॼlíkaði/líkuðuliked.3ॻg/liked.3plþeirthey.nomc. Icelandic C: number-agree disallowedaðthathenniher.daॼlíkaðiliked.3ॻgþeirthey.nom‘that she liked them’When the dative subject remains low (with an expletive in the speciਖ਼er position):(12) ࠹ࡌࡄࡀ-V-࠸࠵ࡈ-ࡂࡃࡁ constructionsa. Icelandic A: agree across daॼÞaðeঀpllíkuðuliked.3pleinum málfræðingione linguist.daॼþessarthesehugmyndir.ideas.nomb. Icelandic B, C: no agree across daॼÞaðeঀpllíkaði/*líkuðuliked.3ॻg/*liked.3pleinum málfræðingione linguist.daॼþessarthesehugmyndir.ideas.nom‘One linguist liked these ideas.’Icelandic A requires number agreement, Icelandic C disallows it (requiring ॻg formon the verb), and Icelandic B shows an intervention e੘ect:Dative intervention eࣳect In Icelandic B,• agreement is optional when the dative subject has moved above the verb (as in11b),• but agreement is blocked if there is a dative intervening between the verb andthe nominative (when the subject remains low, 12b).4These are examples from S&H are of embedded clauses, but this detail is not important: similar ex-amples in matrix clauses are reported in the literature, for instance see Hartmann and Heycock, 2016.WCCFL 38, 7 March 2020 Variation in Icelandic daॼ–nom number agreement Hoover 3࠸࠵ࡈࡏVࡏࡂࡃࡁ.3ࡄࡀ ࠹ࡌࡄࡀࡏVࡏ࠸࠵ࡈࡏࡂࡃࡁ.3ࡄࡀIcelandic A num agree yes yesIcelandic B num agree optional noIcelandic C num agree no noSyncretism exception to the Person RestrictionWhenever the 1st or 2nd person form is phonologically identical to the 3rd personform (due to syncretism in a particular verb’s paradigm), the Person Restriction islifted.For example, in the paradigm for the verb leiðast, 1ॻg and 2ॻg forms happen to besyncretic with the 3ॻg form, and the sentence (13) instead of being ine੘able, is ਖ਼ne.ॻg pl1 leiddumst23leiddist leiddustTable 1: Agreement paradigm for leiðast ‘ਖ਼nd boring’ (paॻॼ)(13) ✔ Henniher.daॼleiddistbored.1+2+3ॻgég/þúI.nom/you.ॻg.nom‘She found me/you boring.’This behaviour in cases of syncretism was ਖ਼rst described as a way for speakers to“both eat their cake and have it too” (Sigurðsson, 1996): not overtly disobeying thePerson Restriction, while getting to use a 1st or 2nd person nominative.Preview: Parameters of variationdaॼ’s visible φ-features ordering at TP boundaryIcelandic A [peॺॻ]  ▷ f#;EPP mvmtgIcelandic B [peॺॻ]; [nॽm]  ▷ f#;EPP mvmtgIcelandic C [peॺॻ]; [nॽm]  ▷# ▷ EPP mvmt  person probe,#  number probe, ▷  ‘precedes’Sketch of an intuitive story: as datives evolve from PP! KP! DP,1. In Icelandic A (the oldest), a di੘erence in ordering has no visible e੘ect.2. Icelandic B is like A, but has grown a [nॽm] feature on datives, and struc-tural ambiguity arises, resulting in optional agree.3. Icelandic C (the newest variety) is like Icelandic B, but an order has beensettled on to resolve the ambiguity: #-probe ▷ EPP3 Ingredients3.1 Probes and goals• Previous work supports splitting φ-feature probing into two independentprobes,– a person probe (),– a number probe (#),which probe separately, and in that order (Béjar and Rezac, 2009; Preminger,2011; Preminger, 2014; Coon and Keine, 2019).Icelandic agreement patterns in particular have been given as evidence in sup-port of a split probe (originally proposed by Taraldsen (1995)5).• I assume a hierarchical structure for features on probes and goals: Valued φ-features on a DP are hierarchically organized (cf. Harley and Ritter, 2002).peॺॻpaॺॼॻpkॺ addॺnॽmplperson features:– 3rd = [peॺॻ],– 2nd = [peॺॻ [paॺॼ [addॺ]]],6– 1st = [peॺॻ [paॺॼ [ॻpkॺ]]]number features:– singular = [nॽm],– plural = [nॽm [pl]]5The original description had number probing before person, but more recent work supports the otherorder.6Either 1st or 2nd person may alternatively be unspeciਖ਼ed beyond paॺॼ. The details here are notimportant for the current analysis. All that’s important here: ਖ਼rst or second person features are a entailthe presence of those of third person, and likewise for number: plural features are a superset of singular.WCCFL 38, 7 March 2020 Variation in Icelandic daॼ–nom number agreement Hoover 4• Probes which have unvalued features, likewise hierarchically organized. Thespeciਖ਼cation of these features is a parameter of variation between languages.For Icelandic:person probe (): [upeॺॻ [upaॺॼ]], number probe (#): [unॽm],Thus the general structure of within the TP in these constructions will look like:(14) [TP … #[unॽm] [upeॺॻ[upaॺॼ]] … [ … DPdaॼ … [ … DPnom …]]]3.1.1 Dative DPs and defective intervention• Agreement with dative case nominals is often limited, crosslinguistically (Alex-iadou, Anagnostopoulou, and Sevdali, 2014; Rezac, 2008).• However, sometimes datives seem to be able to intercept person agree throughtheir own [peॺॻ] feature (Anagnostopoulou, 2003; Danon, 2006).• In Icelandic, dative subject DPs seem to behave externally like 3rd person (andfor B and C, fully like 3ॻg), regardless of intrinsic person or number features.• Roughly following Preminger (2014, §8.3.2) andAtlamaz and Baker (2018), I treatthe dative DP as being wrapped in an shell, the head of which (K) may carry φ-feature(s), maybe inherited from D:KPdaॼKdaॼφ-feature(s)DPDφ-features…I propose that for Icelandic, there is a [peॺॻ] feature on this shell, and, for vari-eties B and C, a [nॽm] feature too.Icelandic A:KPdaॼKdaॼ[peॺॻ]DPDφ-features…Icelandic B,C:KPdaॼKdaॼ[peॺॻ],[nॽm]DPDφ-features…3.2 Agreement meࠫanism: Feature GluttonyI adopt the feature gluttony mechanism for agreement proposed by Coon and Keine(2019), wherein probes are voracious (15). In this account, a probe may become over-valued, having entered into agreement with multiple goals.(15) Agree: (Coon and Keine, 2019, (14))A probe segment [uF] agrees with the closest accessible DP in its domainthat bears [F]. If Agree is established, the hierarchy of segments contain-ing [F] is copied over to the probe, valuing and thus removing [uF].(16) [ probe24uXuY35 … [ DP[X] … [ DP24XY35]]] =) f[X], [X[Y]]g⇛ Gluttony only occurs when the lower DP has something the upper doesn’t.• Mechanism used for spell out as deਖ਼ned by Atlamaz and Baker, after Halle andMarantz (1994):(17) Subset Principle (Atlamaz and Baker, 2018, (61))A vocabulary item’s identifying features must be a subset of the featurespresent at the node where it is to be spelled out.• Crash post-syntax: This account predicts a crash during spell-out only if thereare competing possible phonological forms for the collected bundle of features.A toy example, 3 > 1 conਖ਼guration:(18) [TP … [upeॺॻ[upaॺॼ]] … [ … DP[peॺॻ] … [ … DP[peॺॻ[paॺॼ[ॻpkॺ]]] …]]]=) {[peॺॻ], [peॺॻ[paॺॼ[ॻpkॺ]]]} copied back.By the Subset Principle, two feature bundles are eligible for spellout:a) [peॺॻ]: 3rd person formb) [peॺॻ[paॺॼ[ॻpkॺ]]]: 1st person formWith no way to choose, there is a crash. However, if forms a and b are identical, thenthere is no competition, predicting the syncretism ਖ਼x (13).WCCFL 38, 7 March 2020 Variation in Icelandic daॼ–nom number agreement Hoover 54 Proposal: Feature Gluttony can explain the data4.1 Proposed parameters of variation1. Variability in whether dative subjects have a visible number feature• in Icelandic A, any dative DP has the feature [peॺॻ], and is never speciਖ਼edwith further person features, nor any number features.• in Icelandic B and C: a number feature [nॽm] has become visible on dativeDPs, and therefore they should always act just as if they were singularthird-person DPs.2. Variability in the order of number-probing and movement of the subject• Unspeciਖ਼ed order in Icelandic A and B.• Fixed order in Icelandic C: number probe before subject movement.Proposed di੘erences between varieties (repeated):daॼ’s visible φ-features ordering at TP boundaryIcelandic A [peॺॻ] -probe ▷ f#-probe;EPP mvmtgIcelandic B [peॺॻ]; [nॽm] -probe ▷ f#-probe;EPP mvmtgIcelandic C [peॺॻ]; [nॽm] -probe ▷#-probe ▷ EPP mvmt4.2 Deriving the Person RestrictionThe -probe as unvalued features [upeॺॻ[upaॺॼ]], so in (19)(=9) it copies back fea-tures from both the daॼ and the nom. So, with two possible forms and no way tochoose, there is a crash at spell-out. Unless there’s syncretism! We’ll return to this.(19) * Honumhim.daॼlíkiðlike.2plþiðyou.nom.plIcelandic A#  daॼ[peॺॻ] nom[peॺॻ[paॺॼ]];[nॽm[pl]]x 1. copy back: [peॺॻ]x and [peॺॻ[paॺॼ]]x 2. copy back: [nॽm[pl]]daॼ[peॺॻ] daॼ[peॺॻ] 3. move (EPP)result: {[peॺॻ],[peॺॻ[paॺॼ]],[nॽm[pl]]}=) 3pl and 2pl = !△4.3 Deriving the Dative intervention eࣳect• ࠹ࡌࡄࡀࡏVࡏ࠸࠵ࡈࡏࡂࡃࡁ: With a 3pl nominative, the -probe will only ever agree with the dative;# dative will intervene if it is visible (B, C).(20) =(12a) Icelandic A, no intervention.Þaðeঀpllíkuðuliked.3pleinum málfræðingi[one linguist].daॼþessar[thesehugmyndir.ideas].nomIcelandic AExpl#  daॼ[peॺॻ] nom[peॺॻ];[nॽm[pl]]x 1. copy back: [peॺॻ]x 2. copy back: [nॽm[pl]]result: {[peॺॻ],[nॽm[pl]]}=) 3pl = líkuðu(21) =(12b) Icelandic B/C, dative intervention.Þaðeঀpllíkaðiliked.3ॻgeinum málfræðingi[one linguist].daॼþessar[thesehugmyndir.ideas].nomIcelandic B/CExpl#  daॼ[peॺॻ],[nॽm] nom[peॺॻ];[nॽm[pl]]x 1. copy back: [peॺॻ]x 2. copy back: [nॽm]result: {[peॺॻ],[nॽm]}=) 3ॻg = líkaði࠸࠵ࡈࡏVࡏࡂࡃࡁ.3ࡄࡀ ࠹ࡌࡄࡀࡏVࡏ࠸࠵ࡈࡏࡂࡃࡁ.3ࡄࡀIcelandic A num agree yes ✔yesIcelandic B num agree optional ✔noIcelandic C num agree no ✔noTable 2: Number agreement accounted for so farWCCFL 38, 7 March 2020 Variation in Icelandic daॼ–nom number agreement Hoover 6Optionality in Icelandic B• ࠸࠵ࡈࡏVࡏࡂࡃࡁ: When the dative moves to subject position (࠸࠵ࡈࡏVࡏࡂࡃࡁ construc-tions), there is optional agreement in Icelandic B:(22) =(11b) optional agreement in Icelandic Baðthathenniher.daॼlíkaði/líkuðuliked.3ॻg/liked.3plþeirthey.nom‘that she liked them’This optionality is explained by the order ambiguity (f#-probe;EPP mvmtg).That is, for Icelandic B there are two possible derivations:i. With number greement = líkuðu:Icelandic B#  daॼ[peॺॻ],[nॽm] nom[peॺॻ];[nॽm[pl]]x 1. copy back: [peॺॻ]daॼ[peॺॻ],[nॽm] daॼ[peॺॻ],[nॽm] 2. move (EPP)x 3. copy back: [nॽm[pl]]result: {[peॺॻ],[nॽm[pl]]}=) 3pl = líkuðuii. Without number agreement = líkaði:Icelandic B/C#  daॼ[peॺॻ],[nॽm] nom[peॺॻ];[nॽm[pl]]x 1. copy back: [peॺॻ]x 2. copy back: nॽmdaॼ[peॺॻ],[nॽm] daॼ[peॺॻ],[nॽm] 3. move (EPP)result: {[peॺॻ],nॽm}=) 3ॻg líkaðiThese two derivations are similar to what happens in Icelandic A, (EPP subjectmovement does not disrupt agreement, and the formwill be spelled out 3pl =)líkuðu), and in Icelandic C (which will look precisely the same as the secondoption above).࠸࠵ࡈࡏVࡏࡂࡃࡁ.3ࡄࡀ ࠹ࡌࡄࡀࡏVࡏ࠸࠵ࡈࡏࡂࡃࡁ.3ࡄࡀIcelandic A num agree ✔yes ✔yesIcelandic B num agree ✔optional ✔noIcelandic C num agree ✔no ✔no4.4 Explaining the syncretism exception to the Person Restriction• The syncretism exception: with a 1st/2nd person form that is syncretic 3rd per-son (13, repeated in 23) the Person Restriction is lifted.(23) Henniher.daॼleiddistbored.1+2+3ॻgég/þúI.nom/you.ॻg.nom‘She found me/you boring.’• for Icelandic A:Icelandic A#  daॼ[peॺॻ] nom[peॺॻ[paॺॼ]];[nॽm]x 1. copy back: [peॺॻ]x and [peॺॻ[paॺॼ]]x 2/3. copy back: [nॽm]daॼ[peॺॻ] daॼ[peॺॻ] 3/2. move (EPP)result: {[peॺॻ],[peॺॻ[paॺॼ]],[nॽm]}=) 2ॻg and 3ॻg = leiddist• Icelandic B, the same possible when EPP before#-probe:Icelandic B#  daॼ[peॺॻ],[nॽm] nom[[peॺॻ[paॺॼ]];[nॽm]x 1. copy back: [peॺॻ]x and [peॺॻ[paॺॼ]]daॼ[peॺॻ],[nॽm] daॼ[peॺॻ],[nॽm] 2. move (EPP)x 3. copy back: [nॽm]result: {[peॺॻ],[peॺॻ[paॺॼ]],[nॽm]}=) 2ॻg and 3ॻg = leiddist• for #-probe ▷ EPP (C and optionally B):Icelandic B/C#  daॼ[peॺॻ],[nॽm] nom[[peॺॻ[paॺॼ]];[nॽm]x 1. copy back: [peॺॻ]x and [peॺॻ[paॺॼ]]x 2. copy back: [nॽm]daॼ[peॺॻ],[nॽm] daॼ[peॺॻ],[nॽm] 3. move (EPP)result: {[peॺॻ],[peॺॻ[paॺॼ]],[nॽm]}=) 2ॻg and 3ॻg = leiddistWCCFL 38, 7 March 2020 Variation in Icelandic daॼ–nom number agreement Hoover 75 PredictionsSyncretism in plural(24) syncretism in the plural: judgments vary (reported in Sigurðsson, 1996, (70d))HenniShe.daॼleiddustbored.2+3plþið.you.plRevisit the paradigm for leiðast, ‘to ਖ਼nd boring’ (Table 1, p.3). There is syncretism inthe singular, and also syncretism in the plural. This leads to an interesting prediction:• Varieties should di੘er as follows, for 2ࡄࡀ nominative:• for Icelandic A:Icelandic A#  daॼ[peॺॻ] nom[peॺॻ[paॺॼ]];[nॽm[pl]]x 1. copy back: [peॺॻ]x and [peॺॻ[paॺॼ]]x 2/3. copy back: [nॽm[pl]]daॼ[peॺॻ] daॼ[peॺॻ] 3/2. move (EPP)result: {[peॺॻ],[peॺॻ[paॺॼ]],[nॽm[pl]]}=) 2pl or 3pl = leiddust• Icelandic B, the same possible when EPP before#-probe:Icelandic B#  daॼ[peॺॻ],[nॽm] nom[[peॺॻ[paॺॼ]];[nॽm[pl]]x 1. copy back: [peॺॻ]x and [peॺॻ[paॺॼ]]daॼ[peॺॻ] daॼ[peॺॻ],[nॽm] 2. move (EPP)x 3. copy back: [nॽm[pl]]result: {[peॺॻ],[peॺॻ[paॺॼ]],[nॽm[pl]]}=) 2pl or 3pl = leiddust• but, for #-probe ▷ EPP (C and optionally B): choice between 2ॻg and 3ॻg (not2pl and 3pl). There is syncretism in the singular as well as the plural, so thereis no clash, but the form is di੘erent:Icelandic C#  daॼ[peॺॻ],[nॽm] nom[[peॺॻ[paॺॼ]];[nॽm[pl]]x 1. copy back: [peॺॻ]x and [peॺॻ[paॺॼ]]x 2. copy back: [nॽm]daॼ[peॺॻ] daॼ[peॺॻ],[nॽm] 3. move (EPP)result: {[peॺॻ],[peॺॻ[paॺॼ]],[nॽm]}=) 2ॻg or 3ॻg = leiddist• plural-agreeing form leiddust (2+3pl) should be available for Icelandic A and Bspeakers; non-agreeing form leiddist (1+2+3ॻg) should be available for IcelandicB and C speakers;Prediction: For Icelandic A(25) HenniShe.daॼ*leiddist/leiddustbored.1+2+3ॻg/2+3plþið‘She found you(pl) boring.’Prediction: for Icelandic C(26) HenniShe.daॼleiddist/*leiddustbored.1+2+3ॻg/2+3plþið‘She found you(pl) boring.’• There may be evidence to support this in the data given by Sigurðsson, 1996,which shows a bimodal distribution of judgments, but these judgments havenot been broken down by variety.• Sigurðsson and Holmberg, 2008 discuss this very phenomenon, but don’t use asimplex example, instead, their example is a complex ECM construction, where3ॻg form is available as an alternative for all varieties (see §A.2).⇛ This should be tested in future work. The general prediction: for Icelandic C,the Person Restriction is lifted when there is syncretism between 3rd person andnon-3rd person in the singular (even for plural nominatives).6 Conclusions• A feature gluttony approach to agreement can predict– the person restriction– the syncretism ਖ਼x, speakers can “both eat their cake and have it too”– reported variation in number agreement as being the result of variation in1. the relative order of probing, subject movement2. the visibility of a number feature on dative subjects• This account makes a prediction about di੘erences in the syncretism ਖ਼x betweenvarieties.– Icelandic C will show syncretism ਖ਼x only in the singularWCCFL 38, 7 March 2020 Variation in Icelandic daॼ–nom number agreement Hoover 8ReferencesAlexiadou, A., E. Anagnostopoulou, and C. Sevdali (2014). “Opaque and transparent datives,and how they behave in passives”. In:ࡍe Journal of Comparative Germanic Linguistics 17.1,pp. 1–34. ॽॺl:, E. (2003).ࡍe syntax of ditransitives: Evidence from clitics. Vol. 54. Walter deGruyter.Atlamaz, Ü. and M. Baker (2018). “On partial agreement and oblique case”. In: Syntax 21.3,pp. 195–237.Béjar, S. and M. Rezac (2009). “Cyclic agree”. In: Linguistic Inquiry 40.1, pp. 35–73.Bobaljik, J. D. (2006). “Where’s phi”. In: Agreement as a postsyntactic operation. Ms. Universityof Connecticut.Coon, J. and S. Keine (2019). Feature gluࡉony. ms. McGill, USC. ॽॺl:, G. (Oct. 2006). “Caseless nominals and the projection of DP”. In: Natural Language &Linguistic ࡍeory 24.4, p. 977. iॻॻn: 1573-0859. ॽॺl:, M. and A. Marantz (1994). “Some key features of DistributedMorphology”. In:MIT work-ing papers in linguistics 21.275, p. 88.Harley, H. and E. Ritter (2002). “Person and number in pronouns: A feature-geometric analysis”.In: Language 78.3, pp. 482–526.Hartmann, J. M. and C. Heycock (2016). “Evading agreement: A new perspective on low nom-inative agreement in Icelandic”. In: Proceedings of the 46th Annual Meeting of the North EastLinguistic Society (NELS). Ed. by C. Hammerly and B. Prickett. Vol. 2. GLSA Publications,pp. 67–80.Holmberg, A. and Þ. Hróarsdóttir (2003). “Agreement and movement in Icelandic raising con-structions”. In: Lingua 113.10, pp. 997–1019.Kučerová, I. (2016). “Long-distance agreement in Icelandic: locality restored”. In: ࡍe Jour-nal of Comparative Germanic Linguistics 19.1, pp. 49–74. ॽॺl:, O. (2011). “Asymmetries between person and number in syntax: a commentary onBaker’s SCOPA”. In: Natural Language & Linguistic ࡍeory 29.4, pp. 917–937.– (2014). Agreement and its failures. Vol. 68. MIT Press.Rezac, M. (2008). “Phi-Agree and theta-related Case”. In: Phi theory: Phi-features across inter-faces and modules, pp. 83–129.Schütze, C. T. (2003). “Syncretism and double agreement with Icelandic nominative objects”.In: ॽॺl:ðsson, H. Á. (1996). “Icelandic ਖ਼nite verb agreement”. In:Working papers in Scandinaviansyntax 57, pp. 1–46. ॽॺl:ðsson, H. Á. and A. Holmberg (2008). “Icelandic dative intervention: Person and numberare separate probes”. In: Agreement restrictions, pp. 251–280.Taraldsen, K. T. (1995). “On agreement and nominative objects in Icelandic”. In: Studies in com-parative Germanic syntax. Springer, pp. 307–327.Ussery, C. (2017). “Dimensions of Variation”. In: Syntactic variation in insular Scandinavian 1,p. 165.Zaenen, A., J. Maling, andH.Thráinsson (1985). “Case and grammatical functions:The Icelandicpassive”. In: Modern Icelandic Syntax. Brill, pp. 93–136.A AppendixA.1 Syncretism ࣴx in other languagesA famous example is that of German free relatives, in which the wh-word must showthe case selected for by the matrix verb, as well as the embedded verb. If the twocases being selected for di੘er, the sentence is ungrammatical, but can be saved bysyncretism:(27) Syncretism ਖ਼x in German free relatives Schütze, 2003, (300)a. * IchIzerstöreࡔࡖࡖdestroy[ wer/wenwho.nom/who.accmichme.accärgertࡡࡢࡠ]annoysb. IchIzerstöreࡔࡖࡖdestroy[ waswhat.nom+accmichme.accärgertࡡࡢࡠ]annoys‘I destroy who(ever) annoys me.’See also examples from Polish in Schütze, 2003, and discussion of German copularconstructions in Coon and Keine, 2019.A.2 Complex exceptional case marking constructions in IcelandicIcelandic also has a complex ECM dat-nom construction, with raising verbs7(28) Complex ECM construction, the Person Restrictiona. Honumhim.daॼmunduwould.3plvirðastseem[ þeirthey.nomverabehæडr.competent]‘They would seem competent to him’ (Icelandic A)b. * Honumhim.daॼmundumwould.1plvirðastseem[ viðwe.nomverabehæडr.competent]c. * Honumhim.daॼmunduðwould.2plvirðastseem[ þiðyou.nomverabehæडr.competent]7Such verbs are: डnnast ‘think, feel, ਖ਼nd, consider’; virðast ‘seem’; heyrast ‘(seem to) hear’, ‘sound asi৒’; skiljast ‘(get to) understand’; sýnast ‘seem (to see/look)’; þykja ‘ਖ਼nd, seem, think (that)’; reynast ‘prove(to be …)’ (Sigurðsson and Holmberg, 2008).WCCFL 38, 7 March 2020 Variation in Icelandic daॼ–nom number agreement Hoover 9However, in complex ECM constructions, 3ॻg agreement is also possible (all varieties)(29) Honum mundiwould.3ॻgvirðast [ við/þið/þeirwe/you/they.nom.plvera hæडr. ]This has been explained as optional agreement of the verb with the inਖ਼nitival com-plement (Sigurðsson and Holmberg, 2008). Preminger (2011) uses PLC that non-thirdperson pronouns need to be licensed, but relativized, so that it only applies to a non-3rd person pronoun within a clause.Syncretism in the plural with complex ECM• The examples given for syncretism in the plural by S&H are given in a complexECM setting:(30) Henniher.daॼvirtist/virtustseemed.1+2+3ॻg/2+3plþiðyou.nom.pleiࡉhvaðsomewhateinkennilegirstrangeYou seemed somewhat strange to her.(31) Henniher.daॼvirtist/*virtumstseemed.1+2+3ॻg/1plviðus.nom.pleiࡉhvaðsomewhateinkennilegir.strange‘We seemed somewhat strange to her.’The current proposal’s prediction of disagreement between varieties A and C israther obscured in (30,31) by the availability of the 3ॻg form so, suchmulticlausalexamples are not the most useful for the current proposal.A.3 Long Distance Agreement via object shiࠩHolmberg and Hróarsdóttir, 2003 describe a phenomenon of long distance agreementthat (with a plural nominative across a dative intervener) is possible if dative inter-vener is plural (as in 32a), in Icelandic B (which normally would disallow agreementacross a dative intervener, as in 12b).However, only certain dative interveners may be agreed across and not others (cf.32b, (Kučerová, 2016)).(32) a. Þaðeঀplडnnst/डnnastਖ਼nd.3ॻg/.3plmörgum stúdentummany students.daॼtölvurnarcomputers.def.nomljótar.ugly.nom‘Many students ਖ਼nd the computers ugly’b. Þaðeঀplडnnst/*डnnastਖ਼nd.3ॻg/.3plfáum börnumfew children.daॼtölvurnarcomputers.def.nomljótar.ugly.nom‘Few children ਖ਼nd the computers ugly’Kučerová explains this apparent puzzle in detail, and resolves it as resulting fromobject shift: those datives which always block agreement are precisely those whichcan’t undergo object shift.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items