Yesterday I promised to give some examples of the complexity of findings about language and gender, where published claims sometimes contradict one another, and where the various things that "everybody knows" are not always confirmed by experiment. This happens in every area of rational inquiry, but it's especially common in cases where generalizations are associated with strong feelings. In this case, we're talking about the nature of men and women as biological and social categories, and the way individual men and women interact in both private and public spheres. There aren't many topics that generate stronger feelings than this one.
Strong feelings tend to generate contradictory research for two obvious reasons. First, systematic observation sometimes fails to confirm evocative anecdotes, which may be evocative because they resonate with stereotypes rather than because they genuinely confirm experience. Second, even systematic observation can be misleading, if you don't make the right observational distinctions or don't control for the context in an appropriate way. When the emotional stakes are high, people should in principle be especially careful not to overinterpret or overgeneralize their findings, but in practice, the opposite is often true.
Tag questions are grammatical structures in in which a declarative is followed by an attached interrogative clause or 'tag', such as
In her influential (1975) work Language and Women's Place, Robin Lakoff depicted a typical female speech style, allegedly characterized by the use of features such as hesitations, qualifiers, tag questions, empty adjectives, and other properties, which she asserted to have a common function: to weaken or mitigate the force of an utterance. Thus tag questions "are associated with a desire for confirmation or approval which signals a lack of self-confidence in the speaker."
Lakoff's description of female speech style was based on her remembered impressions rather than on any systematic, quantitative observation. When subsequent researchers went out and counted things, they often found it difficult to confirm her observations. For instance, some studies found that men actually used more tag questions than women did.
Thus Cameron et al. (1988) looked at tag questions in a 45,000 word sample from a British corpus of transcribed conversations, called the "Survey of English Usage" (SEU). There were nine sections of 5,000 words each; three of all-male conversation, three of all-female conversation, and three of mixed-sex conversation. In this corpus, there were 60 tag questions used by men, and only 36 by women. This is a significant sex difference, but in the opposite direction!
When they looked more closely at the function of the tag questions in this corpus, a further sex difference appeared -- which on closer examination seems not primarily to be a sex difference at all.
Holmes (1984) distinguishes two functions of tag questions: modal vs. affective. Modal tags "request information or confirmation of information of which the speaker is uncertain":
But you've been in Reading longer than that, haven't you?
Affective tags "are used not to signal uncertainty on the part of the speaker, but to indicate concern for the addressee":
Affective tags are further subdivided into two kinds: softeners like the first example above, which conventionally mitigate the force of what would otherwise be an impolite demand, and facilitative tags like the second example, which invite the listener to take a conversational turn to comment on the speaker's assertion.
When the tag data in the SEU study are categorized in this way, it turns out that in the category of modal tags -- that is, the tags that genuinely express uncertainty -- are much more likely to be used by men, while the affective tags are only somewhat more likely to be used by men:
|Modal tags||9 (25%)||24 (40%)|
|Affective tags||27 (75%)||36 (60%)|
Suspecting that something besides sex/gender was involved here, the authors of this study turned their attention to another corpus. This database consisted of
nine hours' recorded unscripted talk from three broadcast settings: a medical radio phone-in where the participant roles were ... doctor and caller/client; classroom interaction recorded for ... educational TV, in which the salient roles were those of teacher and pupil; and a general TV discussion programme, in which the roles were ... presenter and audience.
In each case, one of the participants can be identified as "powerful" -- "institutionally responsible for the conduct of the talk", and typically also endowed with greater social power and status in the context of the conversations -- doctor vs. patient, teacher vs. student. The data was sampled so that men and women were equally represented in the "powerful" and "powerless" roles. All tag questions were identified and classified according to Holmes' categories. The results:
|Modal tags||3 (5%)||9 (15%)||10 (18%)||16 (29%)|
|Affective tags (facilitative)||43 (70%)||0||25 (45%)||0|
|Affective tags (softeners)||6 (10%)||0||4 (7%)||0|
First, in this database -- unlike in the SEU data -- there is no significant overall difference in tag usage between the sexes.
Second, men continue to use modal tags relatively more often, and affective tags relatively less often.
The most striking difference by far, however, is not the sex/gender effect but the power effect: it is only the people who are in charge of the conversations -- the "powerful" speakers -- who use affective tags.
The results of the tag question study can be interpreted in several different ways. One view would be that Lakoff's general orientation is confirmed, even though she was wrong about the facts: affective tags are used by people who feel that they are in control of a conversation; the greater use of tag questions overall by men in the SEU data means that the men in those conversations felt more powerful. Another interpretation of such data has been that women's higher proportion of affective tags, which are used to manage the flow of conversation, means that women are saddled with a higher proportion of "interactional shitwork."
Yet another interpretation might be that Lakoff was wrong: men are actually more insecure about their opinions (whence men's greater usage of modal tags), and less interested in controlling the conversational actions of others (whence powerful men's lower usage of affective tags).
Overall, the interpretation of gender differences in language use -- and the extent to which such differences are emphasized in the first place -- seems to have a strong political component. Certainly the more abstract interpretations that are sometimes given to observed differences -- for instance, the conclusion that women are more cooperative and men more competitive in conversation -- are highly political. In evaluating such interpretations, it is well to remember how widely they can vary.
An older set of sexist stereotypes about gender differences in communication are expressed in Rudyard Kipling's 1911 poem The female of the species. Kipling depicts the stereotypical man as an equivocator, "whose timid heart is bursting with the things he must not say." Men in conversation are therefore ready to compromise and to discuss all sides of an issue, and tend to be diverted by humor, doubt and pity. A woman, on the other hand, "who faces Death by torture for each life beneath her breast / May not deal in doubt or pity -- must not swerve for fact or jest." For a woman, "her contentions are her children," and anyone who disagrees will be met with "unprovoked and awful charges -- even so the she-bear fights." The conclusion, for Kipling, is that women should be excluded from politics:
So it cames that Man, the coward, when he gathers
With his fellow-braves in council, dare not leave a place for her.
We should have learned since Kipling's time that this rush of feeling, in response to the well-crafted expression of a social stereotype, is not to be trusted. To quote Penny Eckert and Sally McConnell-Ginet again:
Women's language has been said to reflect their (our) conservatism, prestige consciousness, upward mobility, insecurity, deference, nurturance, emotional expressivity, connectedness, sensitivity to others, solidarity. And men's language is heard as evincing their toughness, lack of affect, competitiveness, independence, competence, hierarchy, control [...] When we recombine all these abstractions, we really do not know what we have. Certainly we don't seem to find real women and men as sums of the characteristics attributed to them.
Posted by Mark Liberman at May 9, 2004 06:50 AM