The acquisition of verbal inflection in child grammars in a variability model of early morphosyntactic development: A biolinguistic perspective.

摘要

This dissertation investigates the acquisition of early verb inflection in child Slovenian from morphosyntactic and morphophonological perspectives. It centers on the phenomenon of root nonfinites, particularly the patterns of omission and substitution errors in verb inflection marking.;It argues that every acquisition model needs to account for the following robust developmental phenomena: initial telegraphic speech, optionality, variability, graded (rather than absolute) morpheme order, and gradualness.;It suggests that current biolinguistic models can be enhanced by supplementing evidence from UG-based studies with that from usage-based and neuropsychological accounts. This proposal is tied to Chomsky's (2005) hypothesis that language comprises three factors---the innate language faculty, experience, and computational efficiency, and that language acquisition relies considerably on this third factor.;The study motivates the Inflectional Hierarchy Complexity Hypothesis (IHCH), suggesting that while children's early morphosyntax is adultlike regarding the availability of functional categories and the concatenation operation, children's morphophonological spell-out is unreliable due to computational bottlenecks. It is hypothesized that grammars initially contain verbs which may be disguised as adult-like finite forms with minimal or zero morphology (e.g., 3rd singular present [3S] reported in early Romance). Such verbs are arguably vPs and are morphophonologically spelled out with greater success than tensed forms which, in turn, are more successful than person-based forms. That is, a probabilistic hierarchy of bare stems > tensed forms > person agreement verbs is predicted, where the forms to the left are postulated to be spelled out statistically better than those to the right.;Based on lemmatization and frequency counts and syntactic and morphophonological analyses of existing and new child Slovenian data, the study confirms the IHCH. It is shown that early Slovenian verbs are mainly complex bare verbs (CBVs) that carry only the conjugation class morphology with no person/number inflection---homophonous with 3S forms---and past participles that lack auxiliaries (arguably TNSPs). Person-related inflection is supplied statistically less reliably than participle inflection which, in turn, is less reliable than the inflection on CBVs. This hierarchy is observed even at age 2;5+ when overt inflection is supplied 80%-90% of the time and most morphophonological properties have been acquired.