Slovenian Verbal Valency: Syntax, Semantics, and Usage

In comparison to the principles of phrasal syntax, lexical items are idiosyncratic in their distributional and semantic characteristics. Because of this, the leading generative approaches to the modelling of the mental lexicon follow the Lexicalist Hypothesis, which assumes that the lexicon is a subcomponent of grammar with combinatorial rules for word formation and the manipulation of lexical meaning that are separate from those that govern morphosyntactic derivation and semantic composition. As Slovenian is a synthetic language with a rich inventory of morphologically related words that differ from one another in their selectional properties, the Lexicalist Hypothesis is assumed in the great majority of contemporary Slovenian linguistics. However, a major theoretical issue of this is that the subtle distinctions in syntagmatic processes between closely related words are explained away as the arbitrary characteristics of a highly complex lexical inventory, whereas they should be accounted for in a more systematic manner to ensure a parsimonious model of Slovenian grammar. Pre-syntactic word formation rules employed in mainstream descriptive Slovenian linguistics also result in a grammatical model of Slovenian in which the generation of syntactic structures built on top of lexically derived words is constrained significantly more than is empirically motivated. By focusing on Slovenian verbal valency, the postdoctoral project will develop a new account of the lexical-syntactic interface in Slovenian using an approach to word formation that takes place exclusively as part of the syntactic-semantic computation. Such a purely derivational approach to the modelling of the syntax and semantics of Slovenian verbs will accurately account for the complexities of actual usage. Beyond linguistics, the project will contribute to the greater understanding of the cognitive processes that underlie the syntactic and semantic computation of Slovenian predicate-argument structure, which lies at the very core of Slovenian grammar. The concrete object of study will be the impersonal construction in Slovenian, which will be researched from the intralinguistic perspective, where it will be compared to other Voice alternations in Slovenian, and from the cross-linguistic perspective, where it will be compared to other Slavic impersonal constructions. The project will propose that the reflexive clitic in the Slovenian constructions syntactically corresponds to a Voice head that introduces the agentive entailment in semantics. It will be claimed that this explains several major but hitherto unrecognised valency-related restrictions concerning pronominal arguments in Slovenian. The analysis of such restrictions will be used to present a new account of the syntax and semantics of Slovenian argument structure in which (1) case assignment takes place in a configurational framework, (2) accusative but not nominative corresponds to structural case valuation in syntax, and (3) case valuation is disentangled from the process of phi-feature agreement between the nominative subject and the verbal elements. The project will also examine the interaction between the lexical aspect and thematic structure of intransitive impersonal constructions, on the basis of which it will propose a new syntactic account of the long-standing lexical puzzle as to why impersonal constructions are incompatible with unaccusative verbs. Empirically, the investigation will employ a novel mixed methods approach that will combine a large-scale corpus analysis of the 4-billion-token metaFida corpus of Slovenian with carefully controlled acceptability judgement tasks performed on corpus data. The project will show that richly annotated corpora are a reliable source of infrequent language information that will shed new light on fine-grained characteristics of the aspectual system and verbal argument structure in Slovenian, and unveil how the two systems interact compositionally.