Grammatical Gender Affects Bilinguals’ Conceptual Gender: Implications for Linguistic Relativity and Decision Making

We used a non-linguistic gender attribution task to determine how French and Spanish grammatical gender affects bilinguals’ conceptual gender. French-English and Spanish-English bilingual, as well as English monolingual adults were asked to assign a male or female voice to 32 color drawings depicting people, animals, and common objects. FrenchEnglish and Spanish-English bilinguals classified items according to French and Spanish grammatical gender respectively. This effect was replicated for French-English bilinguals on those items whose grammatical gender was opposite in French and Spanish. Unexpectedly, Spanish gender similarly affected classifications by Spanish-English and EnglishSpanish bilinguals, as well as English monolinguals. We discuss how grammatical gender, possible covariates, and the order of L1 and L2 acquisition, affect conceptual gender as well as implications for decision making. GRAMMATICAL GENDER AFFECTS BILINGUALS’ CONCEPTUAL GENDER: IMPLICATIONS FOR LINGUISTIC RELATIVITY AND DECISION MAKING A sizeable literature of empirical research on linguistic relativity, the effect of language on thought, began accumulating in earnest during the 1950s. Predictably, some studies adduce evidence strongly supporting linguistic relativity [e.g., 1-7] while other studies are more dismissive of linguistic relativity claims [e.g., 8-13]. This is hardly surprising, since linguistic relativity means different things to different researchers who have relied on a wide variety of investigative methods. Mindful of the ambiguity of what different researchers have meant by relativity, Fishman [14, 15] distinguishes four levels of linguistic effects on thinking: from superficial effects of specialized lexical terms to more profound effects of grammatical structure on individuals’ behavior. The present study addresses the more substantive effects of language on cognition. Of particular interest is whether differences in grammatical gender between languages affect bilinguals’ conceptual gender classifications. An example of research on how grammatical gender affects categorization is reported by Sera, Berge, and Castillo [6]. They investigated whether monolingual speakers of Spanish and English, two languages with different systems of grammatical gender, also categorize objects differently. Spanish nouns must be classified as masculine (e.g., el fuego) or feminine (e.g., la estrella). English, on the other hand, does not systematically classify nouns into different *Address correspondence to this author at the Angelo State University, Department of Psychology & Sociology, ASU Station 10907, San Angelo, TX 76909-0907, USA; E-mail: [email protected] categories of grammatical gender. Indeed, only some English words even explicitly incorporate a referent’s natural gender with word meaning (e.g., boy, girl, ram, yew; but not writer, painter). In Sera et al.’s procedure, Spanish and English monolingual adults and children, from 5 to 10 years of age, viewed pictures of everyday items in one of two conditions. In the label plus pictures condition, participants were shown each picture and heard the experimenter label it in their native language. In the picture only condition, participants were shown the same pictured items, but they were not specifically labeled. In both conditions, participants were asked to attribute either a male or a female voice to animate the pictured item. Spanish speakers, unlike English speakers, were more likely to classify items according to Spanish grammatical gender. This tendency was most pronounced when the pictured items were explicitly labeled. Furthermore, the tendency to categorize objects in accord with Spanish grammatical gender was observed by the time children were 8 years old. Like Sera et al. [6], the present study investigated the role of grammatical gender on categorization across different languages: French, Spanish, and English. Comparisons of two or more languages is a propitious use of a naturally occurring opportunity to study meaningful language effects on cognition [2]. But unlike previous research, we compared French-English, English-French, Spanish-English, and English-Spanish bilinguals’, performance on the same cognitive task. Using bilinguals who differ in the order (L1 or L2) in which they acquired a grammatical gender system offers a within-participants crosslinguistic perspective. Importantly, a within-participants design comprising bilingual samples is a direct test of language acquisition order effects on conceptual gender. Bilinguals’ Gender Classifications The Open Applied Linguistics Journal, 2008, Volume 1 69 Also in accord with prior research, we focus on a significant language variable: grammatical gender. Conveniently, French and Spanish grammatical gender systems are fundamentally similar, but nonetheless show important dissimilarities. First of all, in both languages nouns have a grammatical gender; they are either masculine or feminine. Secondly, nouns, articles, and adjectives in both languages mark grammatical gender with grammatical morphemes. For instance, a big spoon is rendered, une grande cuillère in French, and una cuchara grande in Spanish. Since neither French nor Spanish are case-based languages, word order does not affect grammatical gender, or semantic meaning. In other words, Spanish and French translations of the spoon are unaffected by sentences such as: “The dish ran away with the spoon.”, “The spoon ran away with the dish.” Because nouns, articles, and adjectives provide multiple sources of grammatical gender information, both languages are said to be gender loaded. Notably, French and Spanish grammatical gender differ in one useful and important aspect. Some nouns that are grammatically masculine and feminine in French are the opposite grammatical genders in Spanish. Hence, a cloud is masculine in French un nuage, but feminine in Spanish: una nube. In contrast, English does not systematically classify nouns into different categories of grammatical gender. Indeed, only some English words even explicitly incorporate a referent’s gender with word meaning (e.g., boy, girl, but not writer, painter). The present study assessed adult bilinguals’ cognitive performance using different items with the gender classification task developed by Sera et al. [6, 16]. We expected French-English bilinguals’ categorization to accord with French grammatical gender and Spanish-English bilinguals’ categorization to accord with Spanish grammatical gender. Moreover, we expected French-English and Spanish-English bilinguals’ classifications to differ most on those items where French and Spanish grammatical gender was opposite. Sera et al. [6] found that English speakers were most likely to assign male voices to artifacts and female voices to natural kinds. Because English lacks a grammatical gender system, we expected English monolinguals’ categorization to follow the same pattern: artifacts-male, natural kinds-female. Of further interest was whether the order of grammatical gender system acquisition (L1 or L2), the language in which the task was conducted, and specifically labeling test items affected performance on the gender attribution task.