Evolution of Gender in Indo-European Languages

Abstract

Evolution of Gender in Indo-European Languages Harry E. Foundalis (hfoundal@cs.indiana.edu) Computer Science and Cognitive Science Indiana University, Center for Research on Concepts and Cognition Bloomington, IN 47405 USA Abstract In a recent paper, Lera Boroditsky and Lauren A. Schmidt (2000) examined the degree to which the linguistic category of grammatical gender of nouns influences people’s perception of the cognitive category of biological gender, or sex. Their conclusion was that English speakers’ intuitions about the gender of certain nouns (animals) correlate with the gender assigned to those nouns in languages such as German and Spanish. More important, they found that people’s ideas about the putative biological gender (sex) of objects are strongly influenced by the grammatical gender of those objects in their native language. In this study I sought to reproduce Boroditsky and Schmidt’s results in order to show that the interpretation they supplied is unwarranted, and that the authors conflate the concepts of biological gender (sex) and “formal gender”, which is employed by most Indo-European languages (as opposed to “natural gender”, in English). I compare the intuitions of 20 American monolinguals with the statistics of formal gender as it appears in 14 Indo-European languages. Moreover, I discuss the possible origin and evolution of gender in such languages, and suggest an explanation for the relation between grammatical and biological gender. Introduction The idea that our native language may shape our thought, in part or in whole, is usually associated with the work of Whorf and Sapir, in what is known as “the Sapir-Whorf hypothesis” (Whorf, 1956). This is an intriguing hypothesis because it implies that different cultures — speaking different languages — may perceive the world in different ways. For example, whereas one culture may differentiate objects on the basis of shape, another culture may differentiate them on the basis of material (Imai and Gentner, 1997), and this may be reflected in the corresponding languages. To what extent, then, does language (and culture) force a person’s cognition to perceive the world in one way rather than another? A possible manifestation of this idea was examined by Boroditsky and Schmidt (henceforth B&S), in studying the way grammatical and biological gender interfere with each other in the minds of native speakers of languages such as Spanish and German. B&S support the idea that a speaker whose language assigns the genders masculine and feminine to nouns — whether they refer to people, animals, things, or ideas — is bound to subliminally think of an object as having a corresponding biological gender, male or female. (To avoid circumlocutions, I use the word “sex” to refer to biological gender, reserving “gender” for the grammatical category.) B&S’s proposal rests on the assumption that there is an inherent equating of the concepts of gender and sex in such a speaker’s mind. So, for example, a young learner of an Indo-European language employing “formal gender” could associate a specific category of nouns discernible only through the behavior of neighboring words (say, the feminine nouns) with a perceptual property of entities of the world (say, the femaleness of individuals), even before encountering the words for “feminine” and “masculine”. Although the latter point to a certain relation between gender and sex (which undoubtedly exists), we will see that such an assumption is untenable. First, however, we should briefly review the category of gender as it appears in various languages, in order to understand what it is, and what relation we may expect between the concepts of gender and sex. Although many people are familiar with gender as it appears in Indo-European languages, the notion of gender as understood by linguists is much more general. As a “definition”, I will follow Charles F. Hockett’s description: “Genders are classes of nouns reflected in the behavior of associated words” (Hockett, 1958:231). A characterization like this is general enough to encompass all noun categories that linguists call “genders”, whether they are labeled “masculine”, “feminine”, “neuter”, “common”, or even “class IV”. A language may have two or more classes of nouns that qualify as genders, or it may have none, in which case we say that the language lacks a gender system. Such is the case with several of the major families of Asian languages (e.g., Mandarin Chinese). Tamil, a member of the Dravidian family in south India, divides nouns into “rational” (i.e., people, gods) and “non- rational” (animals, and everything else), and further subdivides rational gender into “masculine” and “feminine” (Corbett, 1991:8–10). Thus, Tamil employs a “natural gender system”, which means that given the semantics of a noun we can predict its gender, and vice- versa. English, a Germanic language, has a natural