Case and Agreement in Creek
- Amy Goodwin Davies, University of Edinburgh
- Glendinning Lecture Theatre 2D67, UWE Bristol
Dixon (1994: 8-9) notes that three primitive syntactic relations (Subject of a transitive (A), Object of a transitive (O), Subject of an intransitive (S)) pattern differently cross-linguistically.
In a nominative-accusative alignment A and S are grouped together, whereas in an ergative-absolutive alignment S and O are grouped together. In Split-S and Fluid-S systems intransitive verbs are divided into two sets: Subjects of intransitive unergative verbs (Sunerg) and subjects of intransitive unaccusative verbs (Sunacc) (Dixon 1994: 71).
The morphosyntactic alignments of case and agreement systems are matched in many languages. According to Dixon (1994), Corbett (2006) and Woolford (2010) the typological literature reports only one of two possible types of case-agreement mismatch; an ergative case system paired with nominative- accusative agreement. “Both case-marking and cross-referencing affixes can be accusative, or both can be ergative; but if there is a split, then bound forms will be accusative and free forms will be ergative...” (Dixon 1994: 95). In discussion of the cross-linguistic patterns Dixon (1994) subsume both ergative-absolutive alignments and unergative-unaccusative alignments under the term ‘ergative’.
Creek is a Muskogean language spoken in Oklahoma and Florida (Martin 2011: 17). In 2000, there were estimated to be 3900 speakers of Creek1, however fewer than 12 were thought to be monolingual speakers (Martin 2011: 17).
In Creek the morphosyntactic alignment on free nominals is nominative-accusative, whereas the agreement markers exhibit a Split-S morphosyntactic alignment. This ‘mismatch’ is theoretically interesting because it presents a challenge to case-based theories of agreement (Woolford 2010: 9). Previous cross-linguistic surveys of case and agreement morphosyntactic alignments have not reported this pattern (Dixon 1994) (Corbett 2006). To account for the cross-linguistic data, several influential analytical approaches to case and agreement do not generate this ‘mismatch’ ((Woolford 2010), (Bobaljik 2008)).
In discussion of previous analyses of this pattern I discuss Choctaw, another Muskogean language which exhibits the same ‘mismatch’ in morphosyntactic alignments. I also discuss work by Patel (2007), which suggests that the universality of Dixon (1994)’s generalisation should be reconsidered.
I describe and analyse the case and agreement patterns of Creek within a Minimalist framework. My analysis proposes that the relationship between case and agreement is epiphenomenal, and arises because both case and agreement are dependent upon the same structural configurations. I relate my proposal to the current Minimalist literature on this topic, and discuss its implications.
1The data discussed is primarily sourced from ‘A Grammar of Creek (Muskogee)’ by Jack B. Martin (2011) and an online collection of narrative texts made available at http://lingspace.wm.edu/lingspace/creek/texts/.
In a nominative-accusative alignment A and S are grouped together, whereas in an ergative-absolutive alignment S and O are grouped together. In Split-S and Fluid-S systems intransitive verbs are divided into two sets: Subjects of intransitive unergative verbs (Sunerg) and subjects of intransitive unaccusative verbs (Sunacc) (Dixon 1994: 71).
The morphosyntactic alignments of case and agreement systems are matched in many languages. According to Dixon (1994), Corbett (2006) and Woolford (2010) the typological literature reports only one of two possible types of case-agreement mismatch; an ergative case system paired with nominative- accusative agreement. “Both case-marking and cross-referencing affixes can be accusative, or both can be ergative; but if there is a split, then bound forms will be accusative and free forms will be ergative...” (Dixon 1994: 95). In discussion of the cross-linguistic patterns Dixon (1994) subsume both ergative-absolutive alignments and unergative-unaccusative alignments under the term ‘ergative’.
Creek is a Muskogean language spoken in Oklahoma and Florida (Martin 2011: 17). In 2000, there were estimated to be 3900 speakers of Creek1, however fewer than 12 were thought to be monolingual speakers (Martin 2011: 17).
In Creek the morphosyntactic alignment on free nominals is nominative-accusative, whereas the agreement markers exhibit a Split-S morphosyntactic alignment. This ‘mismatch’ is theoretically interesting because it presents a challenge to case-based theories of agreement (Woolford 2010: 9). Previous cross-linguistic surveys of case and agreement morphosyntactic alignments have not reported this pattern (Dixon 1994) (Corbett 2006). To account for the cross-linguistic data, several influential analytical approaches to case and agreement do not generate this ‘mismatch’ ((Woolford 2010), (Bobaljik 2008)).
In discussion of previous analyses of this pattern I discuss Choctaw, another Muskogean language which exhibits the same ‘mismatch’ in morphosyntactic alignments. I also discuss work by Patel (2007), which suggests that the universality of Dixon (1994)’s generalisation should be reconsidered.
I describe and analyse the case and agreement patterns of Creek within a Minimalist framework. My analysis proposes that the relationship between case and agreement is epiphenomenal, and arises because both case and agreement are dependent upon the same structural configurations. I relate my proposal to the current Minimalist literature on this topic, and discuss its implications.
1The data discussed is primarily sourced from ‘A Grammar of Creek (Muskogee)’ by Jack B. Martin (2011) and an online collection of narrative texts made available at http://lingspace.wm.edu/lingspace/creek/texts/.