engleski

Semantics of Croatian Neuter Class Nouns

Gender is one of classes Croatian nouns are obligatory assigned to, as is declension type. Croatian nouns are marked for one of the three gender subcategories, i.e. three classes of nouns which can be distinguished syntactically by the agreement they take (Corbett 1991), although the picture is more complicated than that, especially when extralinguistic category of sex is included (which affects agreement on syntactic level): some nouns are hybrid, some heterogeneous, some have double gender, common gender etc. (cf. Tafra 2001, 2007, Pišković 2011, Marković 2012). The declension type (a, e, i) is in principle independent of gender, as is the ending of the citation form. However, uneven distribution of inflectional paradigms between nouns with different gender, distribution of phonological ending of the citation form (e.g. closed or open syllable), as well as the type of a final suffix, may suggest their interdependence at some level. This paper intends to find possible connections by analyzing the third gender class whose behavior is far more predictable starting from any of the above mentioned properties: the citation forms of neuter nouns mostly end with -o or -e, only sometimes with -a, and they belong to a-declension (where NAV Cases share same phonological shape in each number, the rest they share with masculine nouns, except for animacy in Asg). Masculine and feminine nouns make 80-92% of all nouns in data researched so far, both for tokens or lemmas counted (in spoken or written language, adult or child language, e.g. Jelaska, Kovačević 2001, Jelaska, Kovačević, Anđel 2002, Fuček 2005, Cvikić, Jelaska 2007), neuter nouns make significantly smaller proportion of 8-20%. Major gender types exibit morphologically and phonologically much more complex picture, neuter nouns are the simplest with regard to the relationship between the form and the declension types. The other two gender classes consist of many underived as well as derived nouns, neuter category consists of comparably small set of underived nouns that are categorized around semantic prototypes and much bigger and open group of derived nouns with suffixes that share semantics of underived prototypes. While it does not seem to be true that Croatian is basically semantically empty as European gender in general (as e.g. Allan 1977 says), the question is if there is semantic motivation of assignment to classes beyond connection between the sex of animates gender classification has limited or not. In order to test the hypothesis that derived nouns share properties of prototypical underived members of the category neuter nouns, this paper explores semantics of Croatian neuter nouns in two sources of data (both excluding proper nouns). One source are the most frequent six thousand Croatian words according to the only Croatian published frequency list (Moguš, Bratanić, Tadić 1999), based on one million tokens, as up to this vocabulary size word frequency may play an important role in vocabulary development. There are 17% neuter noun lemmas out of 2794 nouns of all genders. The other source of data are neuter nouns found in more than six hundred thousand tokens (Vojnović 1991) of one Croatian translation of the Bible. In the analyzed Croatian Bible (618 716 tokens) neutral tokens make 16% of all noun tokens, while neuter lemmas make 20 % of all nouns (4904 lemmas). Although those two sources of data are of a different size and origin, and the second data consists of some nouns that are culture specific, they share many neuter nouns. The semantics of the most common derived neuter nouns will be analyzed and compared, especially with connection to their suffixes, such as -lo for objects or -ište for places, in order to find the semantic core and offer some explanation as to marginal extensions.

noun class; gender; neuter; semantic core; Croatian language

nije evidentirano