|Title:||The Person Hierarchy: Primitive or Epiphenomenal? Evidence from Halkomelem Salish|
|Authors:||Jason Brown, Karsten Koch, Martina Wiltschko|
|Comment:||Proceedings of NELS 34|
|Abstract:||It is a common assumption that there is a markedness relation between person rankings and the realization of grammatical relations (cf. Silverstein, 1976; Dixon, 1979): the higher a person is on the person hierarchy, the more likely it is that this person will function as a transitive agent (1) or as a subject (2).
(1) Person-hierarchy á la Silverstein (1976) and Dixon (1979)
1st > 2nd > 3rd Pronoun > Proper Noun > Human > Animate > Inanimate
likelihood of functioning as transitive agents
(2) Hierarchy of grammatical relations:
Subj > Object
There are two apparent person-hierarchy effects in Halkomelem Salish discussed in this paper: split ergativity, and person based gaps in the transitive paradigm. Aissen (1999) argues that Optimality Theory (OT) can be used to account for these person hierarchy effects. Since OT is a formal theory of markedness, it should naturally account for markedness effects that result from the person hierarchy. In Aissen’s account, Silverstein and Dixon’s intuition is formalized by means of harmonic alignment between different hierarchies (Prince and Smolensky 1993). Thus, the hierarchies in (1) and (2) are primitives in an OT account.
In contrast, in a structural approach (i.e. Principles & Parameters), the hierarchy of grammatical relations is structurally derived. If we take this seriously, we might expect that person-hierarchy effects are structurally derived as well. A consequence of this view would be that the person hierarchy is epiphenomenal (Wiltschko 2003c).