Conjugation classes and syntax-driven morphology

Patrycja Jabłońska

This talk will assume a neo-constructionist approach to argument structure (cf. Marantz (1997), Borer (2003), and Ramchand (2003) inter alia), where the information associated with a particular item in the lexicon is maximally impoverished (cf. Marantz's Encyclopaedia) and the notion of Lexical Semantic Representation is dispensed with. The flexibility of argument structure configurations associated with particular verbs is instead derived from various syntactic environments a given root can be embedded in.

Thus, I will argue that the morphology occurring in Polish verbal forms in between the root and the Tense marker, traditionally responsible for assigning a root to a given conjugation class, constitute category defining heads (verbalizers in the sense of Marantz (1997)).

I will build on the idea in Déchaine (2003), where the level and the way of root merger can vary from language to language, with Salish merging the root very low and English merging it as high as possible in a given structure. This parametrization accounts for the necessarily augmented causative verbs in Salish and the labile character of the causative/inchoative verbs in English (cf. (1)):

I will argue that this variation in root merger is also relevant on an intra-language level. To wit, in a language that has verbalizers (like Polish), they can come in different flavors and impose a limit to what a given root can name. The typology of verbalizers is presented below:

Table 1: Flavors of verbalizers

Note that all of the conjugation classes I am making claims about are quite productive in Modern Polish (e.g. -i-class derives novel deadjectival and denominal verbs, and -ej-derives highly colloquial forms e.g. z-dziadz-ie-ć (pref-root-V-inf 'become old and gaga', cf. also Cetnarowska (2000). All the remaining classes (except for -owa-) are no longer productive and I won't make any claims about them. In case the verbalizer is of the high type and imposes root merger at the level of o (in the sense of Chomsky (1995)), the root will necessarily name both the causing and the caused subevent. On the other hand, when the verbalizer is of the low type, the root will only name the Become subevent and no further event augmentation will be possible.

The following predictions are made and borne out:

• there are certain easily observable valency -conjugation class correlations: Pattern 1. (in (2))transitive -i-stem into intransitive -ej-stem; Pattern 2.(in (3))transitive -i-stem into inchoative -n-stem.

• high verbalizers can never be used intransitively (i.e. in the inchoative use). Since e.g. -istems are widespread in the lexicon, there has to be a way of circumventing the transitivity problem.Iwillargue that this is the function fulfilled bythe reflexive marker (in (4)), roughly comparable to the Antipassive morphology in ergative languages:

• certain assumptions about the reflexive marker, namely the fact that it is merged in Spec,vP and signals coreference of the initiator of the causing subevent with the subject of the caused subevent will result in the necessarily bieventive nature of the reflexive verbs. That, in turn, accounts for their non-unaccusative properties, e.g. possibility of forming impersonal passive -NO/TO:

• certain semantically motivated assumptions concerning the nature of so-called Derived Imperfective, namely operating on the processual part of the macro-event, account for the nonunaccusative properties of secondary imperfectivized verbs (compare (6) with (7)):

Thus, it turns out that what used to be considered purely inflectional burden on the language learner has strictly definable syntactic consequences and in fact facilitates language acquisition. (Some) conjugation classes can be conceived of as lexical features on roots ([v] or [V]), which restrict the flexibility of argument structure alternations.

References:

Borer, Hagit. 2003. Structuring Sense. An Exo-Skeletal Trilogy. Ms., USC.

Cetnarowska, Bożena. 2000. Resultative adjectives in Polish. Acta Linguistica Hungarica 47:47–79.

Chomsky, Noam. 1995. The Minimalist Program. Cambridge, Massachusetts: MIT Press.

Déchaine, Rose-Marie. 2003. Morphology as the intersection of phonology and syntax: evidence from roots. Talk presented at CASTL Workshop on Morphology, Tromso 30-31 May.

Marantz, Alec. 1997. No Escape from Syntax: Don't try morphological analysis in the privacy of your own lexicon. Penn Working Papers in Linguistics 4:201–225.

Ramchand, Gillian. 2003. First Phase Syntax. Ms, University of Oxford.

Home | Abstracts