Heather Bliss (U of Victoria) and Will Oxford (U of Manitoba)

A microparametric approach to syncretisms in nominal inflection Introduction. Throughout the Algonquian family, nouns are inflected for three bivalent features: animacy (animate/inanimate), number (singular/plural), and obviation (proximate/obviative) (Bloomfield 1946). Two basic syncretism patterns are attested: (1) in the single-syncretism system, the obviation contrast is neutralized with inanimate nouns, and (2) in the double-syncretism system, the obviation contrast is neutralized with inanimate nouns as well as plural nouns. (Following Algonquianist convention, we use 3 for animates, 0 for inanimates, and 3′/0′ for obviatives.) (1) Single-syncretism system (Meskwaki, Cree-Innu, Ojibwe, Delaware) ANIMATE (3) INANIMATE (0) SG PL

PROX

OBV

3s 3p

3′s 3′p

PROX

0(′)s 0(′)p

(2) Double-syncretism system (Blackfoot, Mi’gmaq) ANIMATE (3) INANIMATE (0)

OBV SG PL

PROX

OBV

3s

3′s

PROX

3(′)p

OBV

0(′)s 0(′)p

This paper addresses the question of what syncretisms can tell us about the syntax of nominal features. We assume that patterns of syncretism can inform dependency relations amongst features (Aalberse & Don 2011; Aihkenvald & Dixon 1998; Baerman et al. 2009), and that the primary source of crosslinguistic variation lies in the morphology-syntax mapping (Kayne 1996, 2005). Proposal. We propose a syntactic model for capturing syncretisms: if contrast A is neutralized in the context of contrast B, then A scopes over B in the syntactic structure. Under this model, obviation scopes over animacy only in (1) but over both animacy and number in (2). We thus propose the structures in (3). (3) Single-syncretism system (obv > anim) Double-syncretism system (obv > num, anim) [DP number [ɸP obviation [NP animacy ]]] [DP obviation [ɸP number [NP animacy ]]] This structural difference makes several correct predictions about nominal expressions in the two systems. Prediction 1: Content of ɸP. Both types of systems are correctly predicted to have full DPs, coding all three nominal features. Moreover, both types of systems are correctly predicted to permit NPs (i.e., bare nouns), that are only coded for animacy, a lexically-specified feature across Algonquian (e.g., Armoskaite 2011; Wolfart 1973). Crucially, however, the two systems are predicted to differ in the featural content of ɸP. Single-syncretism systems are predicted to permit forms that are coded for obviation but not number while double-syncretism systems are predicted to permit forms that are coded for number but not obviation. The former is exemplified by Innu locative nouns, which can be obviative-marked but are obligatorily number-neutral (4). The latter is exemplified by Blackfoot pseudo-incorporated nouns, which can be plural-marked but cannot be marked as proximate or obviative (5). (4) Single-syncretism system (Innu): [ɸP obviation [NP animacy]] (excludes number) Ishkuatetshi nete { meshkana-ǹ-t / *meshkanaw-a-t } pessish, ... if.be.fire there { road-OBV-LOC / road-PL-LOC } near ‘If there is a fire near the road(s), ...’ (5) Double-syncretism system (Blackfoot): [ɸP number [NP animacy]] (excludes obviation) Nitáíkskimaa { ponoká-iks / *ponoká-wa / *ponoká-yi } I.hunt.IMPF { elk-PL / elk-PROX / elk-OBV } ‘I am elk-hunting.’ Prediction 2: The syntactic function of number. If the syntactic locus of number varies, we predict that number should fulfill different syntactic functions in the two types of systems. Specifically, number is associated with the DP layer in single- but not double-syncretism systems, and is accordingly predicted to fulfill D-like functions in the former but not the latter. This prediction is borne out. The Proto-Algonquian number suffix originated through the bleaching of a suffixal definite article (Goddard 2007), and in the

single-syncretism systems the number suffix continues to be associated with the syntactic function of D (i.e. forming arguments from predicates, cf. Longobardi 1994, Stowell 1989), as evidenced by the fact that argument expressions require number but not dedicated determiners. In Meskwaki, for example, bare NPs must be incorporated (e.g., penēw- ‘turkey’ in penē-hkēwa ‘he hunts turkeys’). When a Meskwaki noun stem functions as a DP argument, number-marking is obligatory: singular penēw-a and plural penēw-aki are possible DP arguments but unmarked *penēw is not. In the double-syncretism language Blackfoot, on the other hand, number-marking is not sufficient for forming argument expressions: a demonstrative is also required (6), indicating that number does not associate with D in Blackfoot. (6) Double-syncretism system: DP requires determiner (Blackfoot) Nitohkóónoayi *(om-iksi) póós-iks. I.find.them DEM-PL cat-PL ‘I found those cats.’ Prediction 3: Other syncretisms. The fact that obviation has scope only over animacy in the singlesyncretism system explains why obviation is neutralized only with inanimate nouns in these languages. However, this structure also predicts an additional possible syncretism in these systems: since number has scope over obviation, we predict to find number syncretisms that are sensitive to obviation in these languages and indeed this is the case. There is a pervasive tendency for single-syncretism languages, including Cree-Innu, Ojibwe, and Delaware, to also (7) Further syncretism in the single-syncretism neutralize the number contrast with animate obviative system (Cree-Innu, Ojibwe, Delaware) nouns, as in (7). In these languages, then, ANIMATE (3) INANIMATE (0) neutralization of the number contrast is sensitive PROX OBV PROX OBV to obviation, exactly as our model predicts. No such SG 3s 0(′)s syncretisms occur in Blackfoot or Mi’gmaq, 3′(s/p) PL 3p 0(′)p languages with the double-syncretism system. Conclusion. On the basis of inflectional syncretisms, we have proposed that Algonquian obviation has scope over animacy only in some languages (single-syncretism) and over both number and animacy in others (double-syncretism). This difference in hierarchical relations among nominal features correctly predicts various other aspects of the patterning of nominals in these languages. Furthermore, the proposal that number marking is syntactically lower in double-syncretism languages such as Blackfoot is consistent with the widely-held view that verbal morphology maps onto lower positions in Blackfoot than in Cree or Ojibwe (cf. Bliss 2013; Bliss et al. 2013; Cook 2008; Déchaine & Wiltschko 2014). Overall, our results illustrate how a scope-based model of syncretisms can clarify microvariation in the syntax of the DP. REFERENCES • Aalberse & Don 2011. Person and number syncretisms in Dutch. Morphology 21: 32750. • Aihkenvald & Dixon 1998. Dependencies between grammatical systems. Language 74: 56-80. • Armoskaite 2011. The destiny of roots in Blackfoot and Lithuanian: UBC dissertation. • Baerman, Brown & Corbett 2005. The syntax-morphology interface: A study of syncretism. Cambridge University Press. • Bliss 2013. The Blackfoot configurationality conspiracy: Parallels and differences in clausal and nominal structures. UBC dissertation. • Bliss, Déchaine & Hirose 2013. A comparison of locative PPs in Blackfoot and Plains Cree. Presented at the 45th Algonquian Conference, Ottawa. • Bloomfield 1946. Algonquian. In Linguistic structures of native America, Viking Fund Publications in Anthropology. • Cook 2008. The syntax and semantics of clause-typing in Plains Cree. UBC dissertation. • Déchaine & Wiltschko 2014. Micro-variation in agreement, clause-typing and finiteness: Comparative evidence from Blackfoot and Plains Cree. In Papers of the 42nd Algonquian Conference, SUNY Press. • Goddard 2007. Reconstruction and history of the independent indicative. In Papers of the 38th Algonquian Conference, UManitoba. • Kayne 1996. Microparametric syntax: Some introductory remarks. In Black & Motapanyane (eds). Microparametric Syntax and Dialect Variation. Benjamins. • Kayne 2005. Movement and Silence. OUP. • Longobardi 1994. Reference and proper names: A theory of N-movement in syntax and logical form. LI 25: 609-65. • Stowell 1989. Subjects, specifiers, and X′-theory. In Alternative conceptions of phrase structure, UChicago Press. • Wolfart 1973. Plains Cree: A grammatical study. Transactions of the American Philosophical Society 63: 1-90.