Grammatical gender and sex: can a language be fundamentally sexist?

Gender Agenda

Gender Agenda
Issue 2 Michaelmas 2002
The magazine of

Women's Union

What’s sex got to do with it?Grammatical gender and sex: can a language be fundamentally sexist?

Naoise MacSweeney

Language, both in its formal structures and its practical usage, undeniably reflects the
social norms of the culture which shapes and is shaped by it. But how far, then, is it
possible to work backwards from the grammatical framework and prevalent usages of a language
to comment on the cultural expectations of the society which produced it? Can we conclude that
French is a sexist language because it divides its nouns into gender classes? More
importantly, can we go further than this and claim that French society is endemically sexist
because its language subconsciously programmes its speakers to be so? This sexism in grammar
seems all the more significant since some languages have complex and strict rules governing
grammatical gender, and some, such as the Finno-Ugaric languages, have never developed any
gendering at all. Can we take this to mean that the French are sexist, but that Finns and
Hungarians aren’t? Suggestions of this sort are obviously unhelpful in the search for sexual
equality, and arise from grammatical misunderstandings and the arbitrary labelling of the
early linguistic scholars. The origins of grammatical gender were researched extensively in
the mid-twentieth century, but the conclusions of this research have never been widely
publicised. This article is mostly a layperson’s summary of the more technical discussions of
Szemerenyi and Fodor.

Before we go any further, the distinction must be drawn between gender in the grammatical
structures of language, and gendering of usage. In the former, nouns in languages such as
French and German are declined differently according to the gender grouping to which they
belong. This is a fundamental linguistic structure and may, it has been thought, allow
insights into equally fundamental social structures. In the latter, the same limited
vocabulary is utilised by the different sexes in different ways, such as the polite ‘female’
version of Japanese and female adjectival preference in English. This type of gendering of
language is more of a construct of the society contemporary to the analysis, and can be very
telling as to the sexual norms of that society. This social, rather than linguistic, gendering
of language is what concerns most linguists and sociologists today , and is perhaps
fortunately far beyond the scope of both the word limit and author of this article.
The theories of the origins of grammatical gender can be simplified into two polarities – that
it reflects natural sexual distinctions, or that it came about merely as an arbitrary way of
dividing language. The difference between sex and gender becomes important here. While sex can
be described as the biological difference between male and female (man and woman, egg and
sperm, X and Y), gender implies a social construct (actor and actress, waiter and
waitress). Some nouns are indisputably sexed – ‘woman’ and ‘daughter’ are feminine, and ‘man’
and ‘son’ are masculine, but most nouns are gendered – the feminine ‘la chat’ (cat) and
masculine ‘le chien’ (dog) in French. With gendered words, the choice of gender is not always
sexually logical. In German, one of the words for ‘woman’; ‘das Weib’, is not feminine, but
neuter, and the feminine ‘die Maus’ can refer both to the male and the female of the
species. Sometimes words which traditionally would have been applied to a male eg
guard/sentry, are grammatically feminine e.g. ‘la sentinelle’. Words without any natural sex
seem to have been randomly sorted into gender groups, so while the moon in French is feminine
(la lune), in German it is masculine (der Mond). Grammatical gender has got very little, it
seems, to do with sex.


A clay tablet with the undeciphered Minoan hieroglyphic script

There is always an exception to prove the rule. In Akkadian, a Semitic language used in
ancient Mesopotamia (modern Iraq) from the late third millennium BCE and into the second,
gender is used to show sex. The same noun can be either masculine or feminine depending on the
requirements of the speaker. The masculine form of the word ‘sharrum’ means ‘king’, and the
feminine form; ‘sharratum’ means ‘queen’. This more practical function of grammatical gender
is not shared by more modern languages. In French, the masculine word ‘elephant’ cannot be
‘feminised’ to denote a female elephant; ‘elephante’, for example. Instead, the speaker must
resort to the somewhat clumsier ‘elephant femelle’. This disparity between functions of gender
can only highlight the question more: how can we explain the development of grammatical

We can trace the development of noun gendering back through the earliest known
languages. English belongs to a family of languages classed as ‘indoeuropean’, all of which
appear to have descended from a common root language. This ancient root language has been
partially re-created by linguists over the last two centuries, and has been imaginatively
dubbed ‘proto-indoeuropean’. Looking at the different indoeuropean languages through time,
starting with proto-IE and ending up with modern languages, Szemerenyi has traced the
development of gender in the indoeuropean family. Proto-indoeuropean, as well as most
languages deriving from it, contains both fundamental gender distinctions in their grammatical
noun classes and sexual divisions in their vocabulary. The implication is that there has
always been some form of gendering of language. It is thought to have first appeared in the
distinction between the animate and the inanimate. Hittite (used in modern Turkey c.2nd
millennium BCE) had two groups of nouns; animate (both ‘masculine’ and ‘feminine’) and
inanimate. The emphasis here is not on sexual differentiation, but rather distinguishing mere
objects from people and animals. This approach to noun grouping lasted in part for another
millennium, and can be traced in the adjectival patterns of the classical languages of Latin
and Greek. Although Latin and Greek have three possible noun genders (masculine, feminine and
neuter), many Latin and Greek adjectives have only two forms – one for both masculine and
feminine nouns, and another for neuter. Within indoeuropean languages, then, the sexing of
nouns has not always been necessary.

The division of the animate class of nouns into masculine and feminine is seen as a
later development. The early attested indoeuropean language groups of the 1st millennium (Old
Indic, Greek, Latin, Celtic, Germanic and Slavic) all had three noun genders; masculine,
feminine and neuter. It was a later development (and not one shared by all indoeuropean
languages, e.g. modern German which retains three genders) to lose the neuter and to reassign
its nouns into the two remaining classes. The trend in noun grouping in indoeuropean can be
summarised as distinguishing first only between the animate and inanimate, then further
splitting the animate into masculine and feminine, and finally dissolving the neuter class
into the masculine and feminine. This may be seen as reflecting a trend in society – sex
became gradually more of a concern to early Indoeuropeans.Figure – a tablet from the Babylonian ‘Epic of Gilgamesh’, an example of written
Akkadian. From Andrew George.

Grammatical gender is a result of our preoccupation with sex. In most languages, it has
very little to do with ease of sexual definition as it does in Akkadian. The noun groupings of
Hittite prove that it is by no means necessary to language, but rather is only one amongst
many possible ways of noun division. The linguist Wundt has identified five main ways in which
languages divide their nouns:

  1. between superior and inferior value, e.g. Iroquois
  2. between humans and everything else
  3. between animate and inanimate, e.g. American Dakota, Hittite
  4. between the sexes, e.g. the Semitic and Hamitic languages
  5. the three gender system, e.g. Indoeuropean languages

The gendering system we in Western Europe today are most familiar with is, in wider context,
one of the most sexualised systems possible. A good comparison is with some of the native
Australian languages, which can have between two and nine genders, or with native African
languages, which have been reported to have as many as twenty-one noun genders.
Fodor explains the emergence of grammatical gender purely in terms of a morphological
fluke. Since natural gender and sex could not have brought it about, it is to be treated like
any other linguistic phenomena and is not a sexually charged issue. Syntactic agreements,
rather than any meaningful significance, are the explanation. The early grammarians such as
Grim (1831) in the nineteenth century are perhaps to be held responsible for making gender a
sexual issue. They were living in a time and a place where sexual differentiation and
oppression were all-pervasive, even to the extent of affecting their work. The choices of the
labels ‘masculine’, ‘feminine’ and ‘neuter’ are arbitrary ones and the noun classes could have
just as effectively been named ‘one’, ‘two’ and ‘three’. The academic interpretation of
grammatical gender is that it is completely divorced from sex and cannot signify any sexual

The fact that the pioneering grammarians did choose sexual terms to describe noun classes
may be significant in itself. It cannot be denied that the Semitic, Hamitic and Indoeuropean
treatment of grammatical gender is far more sexualised than most other languages, and that
historically, all these cultures have engaged in sexual oppression to varying
degrees. Finally, it is also true that the early grammarians all came from these same
cultural groups which have such an emphasis on sex and gender. We cannot be sure that if the
first scholars (and therefore the people who dictated the fundamental definitions) of grammar
had been native speakers of languages with less grammatical gendering, grammatical gender
would ever have become an issue. Our understanding of language today and the way it works is
inseparably caught up with the way our society works, and sexual differences have always been
an essential part of this. As heirs of an indoeuropean cultural as well as linguistic
heritage, we can and should not try to escape the gendering of language, but rather to
understand it and its origins.




Email us at