In the fifth century BC, according to Aristotle's account, Protagoras first created the labels masculine, feminine, and neuter for Greek nouns, and language scholars have been trying to explain the relationship of grammatical gender categories to the world around them ever since.1 Protagoras himself, apparently anxious that the grammatical gender of nouns and the sex of their referents did not always correspond in Greek, is said to have wanted to change the gender of Greek menis 'anger' and peleks 'helmet,' both of which are feminine nouns, to masculine because he felt the masculine was more appropriate given the words' referents (Robins 1971 “1951”: 15–16). Despite Aristotle's subsequent proposal of grammatical reasons for nominal gender classes, the original labels persisted in the descriptions of gender in classical grammars – and, therefore, in all the later Western grammars modeled on them – and these labels have created the pervasive misperception that grammatical gender categories in a language reflect a connection between male and female human beings and masculine and feminine inanimate objects. The terms deceptively imply a link between the categories in the natural gender system of Modern English – in which there is a clear correlation between masculine and feminine nouns and biological traits in the referent – and the categories in the grammatical gender systems of other IndoEuropean languages; in fact, these two types of systems are distinct. The shift of English from a grammatical to a natural gender system is highly unusual and involves a complex set of related grammatical transformations in the language.
Despite their descriptive labels, noun classes in a grammatical gender system, unlike those in a semantic gender system, do not correspond to conceptual categories, no matter how creative the grammarian. In other words, there is no way
1 For more detailed descriptions of the Greek and subsequent Latin treatments of grammatical
gender, see Robins (1971 “1951”) and Vorlat (1975). Vorlat provides the most comprehensive
treatment of early English grammars (1586–1737) currently available, grounding these works in
the classical tradition from which they stem and identifying areas of grammatical conservatism