Sul rapporto testa/dipendente nel sintagma nominale. Il caso dell’ittita e del latino (con cenni al greco)
This article starts from the different typologies in agreement described by [Nichols 1986] concerning the relationship between head and dependent(s). She discovered that four types are possible: head-marking, dependent-marking, double marking and zero marking. Although Nichols analyses this relationship within the NP and also within the sentence, in this article we will examine this feature only as far as NP is concerned. Moreover, we will consider a small group of languages, which are all genetically related: the Indo-European family. In fact, despite their common origin and the fact that they all show a dependent-marking type, these languages organize their strategies in agreement in different ways. For instance in Hittite, inanimate nouns (like watar) agree with an adjective declined in inanimate gender (i.e. suppi) only if they are subjects of intransitive verbs. If theyare subjects of transitive verbs, the nouns take another ending (-anza) and agree with an adjective in animate gender (like suppis). We shall avoid discussing questions concerning the alignment of Hittite: our aim is only to consider that such a syntactical feature plays an important role in agreement between head and dependent in Hittite and in other Anatolian languages, but it doesn’t appear in the other Indo-European languages. On the other hand, Latin doesn’t distinguish masculine from feminine gender morphologically : we can find statistical tendencies, but nominal declensions are pure formal schemes based on phonological criteria. So the first declension offers a large majority of feminine nouns (like rosa), but it also includes a small number of masculine nouns (like agricola): we shall be able to discern the gender information only in a NP, thanks to information carried by the adjectival/pronominal dependent (for instance bonus agricola). However, not all adjectives have separate forms for masculine and feminine. Greek has more or less the same structure as Latin, but in the first declension there is a case where a subtype is introduced in order to distinguish the masculine gender morphologically: so we find nominatives in -ā, -ē for feminine, whereas the masculine has -ās, -ēs. These few examples show that these languages take into account different features in order to constitute agreement classes. It is a first step to an analysis of this syntactical feature in other representative languages of the Indo-European family.
KEYWORDS: nominal heads, head/dependent relationship, agreement, Hittite, Latin