A Cue-based Approach to the Acquisition of Grammatical Gender in Russian This paper aims at explaining a discrepancy in children's use of the semantic gender principle across some exceptional classes of Russian nouns in terms of a cue-based approach to acquisition. We argue that in the process of gender acquisition children rely on certain morphosyntactic cues in the input (cf. Lightfoot 1999). We adopt Westergaard's (2006, 2008) extended cue-based model, which assumes the existence of micro-cues. This means that children focus on linguistically relevant categories and subcategories that express a particular cue and are extremely sensitive to fine distinctions in syntax and morphology. The idea of this selective cuesearching process is used in this paper to account for differences in children's use of the semantic principle with the following subcategories of nouns: male kinship terms in a (papa `daddy'), double gender nouns (plaksa `cry-baby'), hybrids (vrac `doctor'), and female names in ok/ik (Lenok). In all of these there is a mismatch between the morphology of the nouns and the syntactic agreement that it appears with. Consider e.g. moj-ьM pap-a `my daddy' vs. moj-aF mam-a `my mommy': both nouns have morphology typical of feminine nouns, yet they take different agreement forms which correspond to the biological sex of the noun's referent. The results of elicited production experiments with 25 monolingual Russian children (age 2;6-4;0) presented in Rodina (2008) reveal that the semantic principle is least operative with hybrids and female names (see Figure 1). The observed situation is the opposite from what has been found in the caregivers' production in identical experimental conditions. Yet, data from older children (age 5;1-6;5) suggest that the situation improves with age (see Figure 2). In our account qualitative and quantitative properties of the input receive particular attention and the discrepancy is attributed to factors such as input frequency and consistency/saliency of the input information where the cue is expressed. While consistent evidence is sufficient for the establishment of the semantic principle at a relatively early age, in the absence of consistency the amount of input becomes important. The important conclusion that we would like to make in this paper is that in a language where the gender system is organized such a way that the semantic principle is only crucial for very specific classes of nouns, children appear to be conservative learners who use input information to make specific inferences concerning the gender of the nouns in their language and generalize only within a particular class (cf. Roeper 2007). References: Lightfoot, D. 1999. The Development of Language: Acquisition, Change and Evolution. Blackwell, Malden, MA and Oxford. Rodina, Y. 2008. Semantics and Morphology: The Acquisition of Grammatical Gender in Russian. Doctoral Dissertation, University of TromsЬ. Roeper, T. 2007. What frequency can do and what it can't. Frequency effects in language acquisition: Defining the limits of frequency as an explanatory concept. In I. GЭlzow and N. Gagarina (eds.) Frequency Effects in Language Acquisition. Studies on Language Acquisition SOLA Series, ed. I. GЭlzow and N. Gagarina, 23-48. Berlin: Mouton De Gruyter. Not referred to in text. Westergaard, M. R. 2006. Triggering V2: The Amount of Input Needed for Parameter Setting in a Split-CP model of Clause Structure. In A. Belletti, E. Bennati, C. Chesi, E. DiDomenico and I. Ferrari (eds.), Language Acquisition and Development: Proceedings of GALA 2005, 564577. Cambridge: Cambridge Scholars Press.