IE *weid- as a Root with Dual Subcategorization Features in the Homeric Poems.

Abstract

This paper is organized as follows: the first section sketches the theoretical background involved in the case study of Old Greek éidon/óida. As is well known, the aorist éidon takes only an accusative DP-object, while the perfect óida can take either a genitive or an accusative DP-object. Sections 2-5 I aim to prove that the diachronic development of the root *weid- in early Greek must be take into consideration to explain the synchronic phenomenon of dual subcategorization features. This root proves indeed to be polysemous and is split into two different meanings which are lexicalised by means of different bridging contexts and different morphological developments. In section 6 the peculiar evolution of óida from a genitive/accusative to an accusative syntax is considered from a typological point of view. The genitive/accusative syntax is held to be a residual trace of a previous active-stative system which had characterised the Proto-Indo-European before it adopted a nominative-accusative system. Section 7 examines the available evidence about the inherent status of genitive and accusative in Homeric Greek. In section 8 the modular conception of grammar is discussed by considering the role played by the morpho-semantic properties of the root when dealing with a theta-related case system. Sections 9-10 are concerned with a minimalist representation of the syntactic structure of the early Greek verb. Against the Late Insertion principle, a model endowed with three functional heads (TP, AspP, vP) is put forward which requires the root features to be visible in syntactic derivation before Spell-out. Thus, the syntactic Object Case feature at that stage of the language is taken to have been interpretable in LF (Object-in-situ hypothesis). Section 11 examines the Greek sentence structure after the nominative-accusative type had been stabilized: once case had turned to a purely formal feature, it became uninterpretable; so, the semantic properties of the root were not able to determine the inherent case assignment any longer, and tense tended to incorporate all temporal distinctions related to the verb (including aspect).