What language barrier?
It is a truism that men and women do not communicate in the same way. But is there really any evidence to support this Mars-and-Venus theory? Oxford language professor Deborah Cameron investigates in the first of three extracts from her new book
Click here for the table on Gender differences in verbal/communicative behaviour adapted from Hyde, ‘The Gender Similarities Hypothesis’.
The Guardian, Monday 1 October 2007
Writers in this vein are fond of presenting themselves as latter-day Galileos, braving the wrath of the political correctness lobby by daring to challenge the feminist orthodoxy that denies that men and women are by nature profoundly different. Simon Baron-Cohen, the author of The Essential Difference, explains in his introduction that he put the book aside for several years because “the topic was just too politically sensitive”. In the chapter on male-female differences in his book about human nature, The Blank Slate, Steven Pinker congratulates himself on having the courage to say what has long been “unsayable in polite company”. Both writers stress that they have no political axe to grind: they are simply following the evidence where it leads, and trying to put scientific facts in place of politically correct dogma.
Yet before we applaud, we should perhaps pause to ask ourselves: since when has silence reigned about the differences between men and women? Certainly not since the early 1990s, when the previous steady trickle of books began to develop into a raging torrent. By now, a writer who announces that sex-differences are natural is not “saying the unsayable”: he or she is stating the obvious. The proposition that men and women communicate differently is particularly uncontroversial, with cliches such as “men never listen” and “women find it easier to talk about their feelings” referenced constantly in everything from women’s magazines to humorous greeting cards.
The idea that men and women “speak different languages” has itself become a dogma, treated not as a hypothesis to be investigated or as a claim to be adjudicated, but as an unquestioned article of faith. Our faith in it is misplaced. Like the scientists I have mentioned, I believe in following the evidence where it leads. But in this case, the evidence does not lead where most people think it does. If we examine the findings of more than 30 years of research on language, communication and the sexes, we will discover that they tell a different, and more complicated, story.
The idea that men and women differ fundamentally in the way they use language to communicate is a myth in the everyday sense: a widespread but false belief. But it is also a myth in the sense of being a story people tell in order to explain who they are, where they have come from, and why they live as they do. Whether or not they are “true” in any historical or scientific sense, such stories have consequences in the real world. They shape our beliefs, and so influence our actions. The myth of Mars and Venus is no exception to that rule.
For example, the workplace is a domain in which myths about language and the sexes can have detrimental effects. A few years ago, the manager of a call centre in north-east England was asked by an interviewer why women made up such a high proportion of the agents he employed. Did men not apply for jobs in his centre? The manager replied that any vacancies attracted numerous applicants of both sexes, but, he explained: “We are looking for people who can chat to people, interact, build rapport. What we find is that women can do this more … women are naturally good at that sort of thing.” Moments later, he admitted: “I suppose we do, if we’re honest, select women sometimes because they are women rather than because of something they’ve particularly shown in the interview.”
The growth of call centres is part of a larger trend in economically advanced societies. More jobs are now in the service than the manufacturing sector, and service jobs, particularly those that involve direct contact with customers, put a higher premium on language and communication skills. Many employers share the call-centre manager’s belief that women are by nature better qualified than men for jobs of this kind, and one result is a form of discrimination. Male job applicants have to prove that they possess the necessary skills, whereas women are just assumed to possess them. In today’s increasingly service-based economy, this may not be good news for men.
But it is not only men who stand to lose because of the widespread conviction that women have superior verbal skills. Someone else who thinks men and women are naturally suited to different kinds of work is Baron-Cohen. In The Essential Difference he offers the following “scientific” careers advice: “People with the female brain make the most wonderful counsellors, primary school teachers, nurses, carers, therapists, social workers, mediators, group facilitators or personnel staff … People with the male brain make the most wonderful scientists, engineers, mechanics, technicians, musicians, architects, electricians, plumbers, taxonomists, catalogists, bankers, toolmakers, programmers or even lawyers.”
The difference between the two lists reflects what Baron-Cohen takes to be the “essential difference” between male and female brains. The female-brain jobs make use of a capacity for empathy and communication, whereas the male ones exploit the ability to analyse complex systems. Baron-Cohen is careful to talk about -“people with the female/male brain” rather than “men and women”. He stresses that there are men with female brains, women with male brains, and individuals of both sexes with “balanced” brains. He refers to the major brain types as “male” and “female”, however, because the tendency is for males to have male brains and females to have female brains. And at many points it becomes clear that in spite of his caveats about not confusing gender with brain sex, he himself is doing exactly that.
The passage reproduced above is a good example. Baron-Cohen classifies nursing as a female-brain, empathy-based job (though if a caring and empathetic nurse cannot measure dosages accurately and make systematic clinical observations she or he risks doing serious harm) and law as a male-brain, system-analysing job (though a lawyer, however well versed in the law, will not get far without communication and people-reading skills). These categorisations are not based on a dispassionate analysis of the demands made by the two jobs. They are based on the everyday common-sense knowledge that most nurses are women and most lawyers are men.
If you read the two lists in their entirety, it is hard not to be struck by another “essential difference”: the male jobs are more varied, more creative, and better rewarded than their female counterparts. Baron-Cohen’s job-lists take me back to my schooldays 35 years ago, when the aptitude tests we had to complete before being interviewed by a careers adviser were printed on pink or blue paper. In those days we called this sexism, not science.
At its most basic, what I am calling “the myth of Mars and Venus” is simply the proposition that men and women differ fundamentally in the way they use language to communicate. All versions of the myth share this basic premise; most versions, in addition, make some or all of the following claims:
1 Language and communication matter more to women than to men; women talk more than men.
2 Women are more verbally skilled than men.
3 Men’s goals in using language tend to be about getting things done, whereas women’s tend to be about making connections to other people. Men talk more about things and facts, whereas women talk more about people, relationships and feelings.
4 Men’s way of using language is competitive, reflecting their general interest in acquiring and maintaining status; women’s use of language is cooperative, reflecting their preference for equality and harmony.
5 These differences routinely lead to “miscommunication” between the sexes, with each sex misinterpreting the other’s intentions. This causes problems in contexts where men and women regularly interact, and especially in heterosexual relationships.
The literature of Mars and Venus, in both the self-help and popular science genres, is remarkably patronising towards men. They come off as bullies, petulant toddlers; or Neanderthals sulking in their caves. One (male) contributor to this catalogue of stereotypes goes so far as to call his book If Men Could Talk. A book called If Women Could Think would be instantly denounced; why do men put up with books that put them on a par with Lassie or Skippy the Bush Kangaroo (“Hey, wait a minute – I think he’s trying to tell us something!”)?
Perhaps men have realised that a reputation for incompetence can sometimes work to your advantage. Like the idea that they are no good at housework, the idea that men are no good at talking serves to exempt them from doing something that many would rather leave to women anyway. (Though it is only some kinds of talking that men would rather leave to women: in many contexts men have no difficulty expressing themselves – indeed, they tend to dominate the conversation.)
This should remind us that the relationship between the sexes is not only about difference, but also about power. The long-standing expectation that women will serve and care for others is not unrelated to their position as the “second sex”. But in the universe of Mars and Venus, the fact that we (still) live in a male-dominated society is like an elephant in the room that everyone pretends not to notice.
My father, like many men of his generation, held the belief that women were incompetent drivers. During my teenage years, family car journeys were invariably accompanied by an endless running commentary on how badly the women around us were driving. Eventually I became so irritated by this, I took to scouring passing traffic for counter-examples: women who were driving perfectly well, and men who were driving like idiots.
My father usually conceded that the men were idiots, but not because they were men. Whereas female idiocy was axiomatically caused by femaleness, substandard male drivers were either “yobbos” – people with no consideration for others on the road or anywhere else – or “Sunday drivers”: older men whose driving skills were poor because they used their cars only at weekends. As for the women who drove unremarkably, my father seemed surprised when I pointed them out. It was as if he had literally not noticed them until that moment.
At the time I thought my father was exceptional in his ability to make reality fit his preconceptions, but now I know he was not. Psychologists have found in experimental studies that when interpreting situations people typically pay most attention to things that match their expectations, and often fail to register counter-examples.
It is not hard to see how these tendencies might lead readers of Mars and Venus books to “recognise” generalisations about the way men and women use language, provided those generalisations fit with already familiar stereotypes. An anecdote illustrating the point that, say, men are competitive and women cooperative conversationalists will prompt readers to recall the many occasions on which they have observed men competing and women cooperating – while not recalling the occasions, perhaps equally numerous, on which they have observed the opposite. If counter-examples do come to mind (“What about Janet? She’s the most competitive person I know”), it is open to readers to apply the classic strategy of putting them in a separate category of exceptions (“of course, she grew up with three brothers / is the only woman in her department / works in a particularly competitive business”).
In relation to men and women, our most basic stereotypical expectation is simply that they will be different rather than the same. We actively look for differences, and seek out sources that discuss them. Most research studies investigating the behaviour of men and women are designed around the question: is there a difference? And the presumption is usually that there will be. If a study finds a significant difference between male and female subjects, that is considered to be a “positive” finding, and has a good chance of being published. A study that finds no significant differences is less likely to be published.
Most people, of course, do not read academic journals: they get their information about scientific research findings from the reports that appear in newspapers, or from TV science documentaries. These sources often feature research on male-female differences, since media producers know that there is interest in the subject. But the criteria producers use when deciding which studies to report and how to present them introduce another layer of distortion. And sometimes headlines trumpet so-called facts that turn out, on investigation, to have no basis in evidence at all.
In 2006, for instance, a popular science book called The Female Brain claimed that women on average utter 20,000 words a day, while men on average utter only 7,000. This was perfect material for soundbite science – it confirmed the popular belief that women are not only the more talkative sex but three times as much – and was reported in newspapers around the world.
One person who found it impossible to believe was Mark Liberman, a professor of phonetics who has worked extensively with recorded speech. His scepticism prompted him to delve into the footnotes of The Female Brain to find out where the author had got her figures. What he found was not an academic citation but a reference to a self-help book. Following the trail into the thickets of popular literature, Liberman came across several competing statistical claims. The figures varied wildly: different authors (and sometimes even the same author in different books) gave average female daily word-counts ranging from 4,000 to 25,000 words. As far as Liberman could tell, all these numbers were plucked from thin air: in no case did anyone cite any actual research to back them up. He concluded that no one had ever done a study counting the words produced by a sample of men and women in the course of a single day. The claims were so variable because they were pure guesswork.
After Liberman pointed this out in a newspaper article, the author of The Female Brain conceded that her claim was not supported by evidence and said it would be deleted from future editions. But the damage was already done: the much-publicised soundbite that women talk three times as much as men will linger in people’s memories and get recycled in their conversations, whereas the little-publicised retraction will make no such impression. This is how myths acquire the status of facts.
Do women and men really speak so differently?
In 2005, an article appeared in the journal American Psychologist with the title The Gender Similarities Hypothesis. This title stood out as unusual, because, as we have seen, the aim of most research studies is to find differences rather than similarities between men and women. Yet, as the article’s author Janet S Hyde pointed out, on closer inspection, the results of these studies very often show more similarity than difference.
Hyde is a psychologist who specialises in “meta-analysis”, a statistical technique that allows the analyst to collate many different research findings and draw overall conclusions from them. Scientists believe that one study on its own does not show anything: results are only considered reliable if a number of different studies have replicated them. Suppose that the question is: who interrupts more, men or women? Some studies will have found that men interrupt more, others that women do, and others may have found no significant difference. In some studies the reported gender difference will be large, while in others it will be much smaller. The number of people whose behaviour was investigated will also vary from study to study. Meta-analysis enables you to aggregate the various results, controlling for things that make them difficult to compare directly, and calculate the overall effect of gender on interruption.
Hyde used this technique to review a large number of studies concerned with all kinds of putative male-female differences. In Table 1, I have extracted the results for just those studies that dealt with gender differences in linguistic and communicative behaviour.
To read this table you need to know that “d” is the formula indicating the size of the overall gender difference: minus values for “d” indicate that females are ahead of males, whereas plus values indicate that males are ahead of females.
Click here for the table on Gender differences in verbal/communicative behaviour adapted from Hyde, ‘The Gender Similarities Hypothesis’.
So, for instance, the table tells us that when the findings of different studies are aggregated, the overall conclusion is that men interrupt more than women and women self-disclose more than men. However, the really interesting information is in the last column, which tells us whether the actual figure given for d indicates an effect that is very large, large, moderate, small, or close to zero. In almost every case, the overall difference made by gender is either small or close to zero. Two items, spelling accuracy and frequency of smiling, show a larger effect – but it is still only moderate.
There were a few areas in which Hyde did find that the effect of gender was large or very large. For instance, studies of aggression and of how far people can throw things have shown a considerable gap between the sexes (men are more aggressive and can throw further). But in studies of verbal abilities and behaviour, the differences were slight. This is not a new observation. In 1988 Hyde and her colleague Marcia Linn carried out a meta-analysis of research dealing specifically with gender differences in verbal ability. The conclusion they came to was that the difference between men and women amounted to “about one-tenth of one standard deviation” – statistician-speak for “negligible”. Another scholar who has considered this question, the linguist Jack Chambers, suggests that the degree of non-overlap in the abilities of male and female speakers in any given population is “about 0.25%”. That’s an overlap of 99.75%. It follows that for any array of verbal abilities found in an individual woman, there will almost certainly be a man with exactly the same array.
Chambers’ reference to individual men and women points to another problem with generalisations such as “men interrupt more than women” or “women are more talkative than men”. As well as underplaying their similarities, statements of the form “women do this and men do that” disguise the extent of the variation that exists within each gender group. Explaining why he had reacted with instant scepticism to the claim that women talk three times as much as men, Liberman predicted: “Whatever the average female versus male difference turns out to be, it will be small compared with the variation among women and among men.” Focusing on the differences between men and women while ignoring the differences within them is extremely misleading but, unfortunately, all too common.
Do women really talk more than men?
If we are going to try to generalise about which sex talks more, a reliable way to do it is to observe both sexes in a single interaction, and measure their respective contributions. This cuts out extraneous variables that are likely to affect the amount of talk (like whether someone is spending their day at a Buddhist retreat or a high school reunion), and allows for a comparison of male and female behaviour under the same contextual conditions.
Numerous studies have been done using this approach, and while the results have been mixed, the commonest finding is that men talk more than women. One review of 56 research studies categorises their findings as shown here:
Pattern of difference found / Number of studies
Men talk more than women / 34 (60.8%)
Women talk more than men / 2 (3.6%)
Men and women talk the same amount / 16 (28.6%)
No clear pattern / 4 (7.0%)
· Source: based on Deborah James and Janice Drakich, ‘Understanding Gender Differences in Amount of Talk’, in Deborah Tannen (ed.), Gender and Conversational
The reviewers are inclined to believe that this is a case of gender and amount of talk being linked indirectly rather than directly: the more direct link is with status, in combination with the formality of the setting (status tends to be more relevant in formal situations). The basic trend, especially in formal and public contexts, is for higher-status speakers to talk more than lower-status ones. The gender pattern is explained by the observation that in most contexts where status is relevant, men are more likely than women to occupy high-status positions; if all other things are equal, gender itself is a hierarchical system in which men are regarded as having higher status.
“Regarded” is an important word here, because conversational dominance is not just about the way dominant speakers behave; it is also about the willingness of others to defer to them. Some experimental studies have found that you can reverse the “men talk more” pattern, or at least reduce the gap, by instructing subjects to discuss a topic that both sexes consider a distinctively female area of expertise. Status, then, is not a completely fixed attribute, but can vary relative to the setting, subject and purpose of conversation.
That may be why some studies find that women talk more in domestic interactions with partners and family members: in the domestic sphere, women are often seen as being in charge. In other spheres, however, the default assumption is that men outrank women, and men are usually found to talk more. In informal contexts where status is not an issue, the commonest finding is not that women talk more than men, it is that the two sexes contribute about equally.
If it does not reflect reality, why is the folk-belief that women talk more than men so persistent? The feminist Dale Spender once suggested an explanation: she said that people overestimate how much women talk because they think that, ideally, women would not talk at all. While that may be rather sweeping, it is true that belief in female loquacity is generally combined with disapproval of it. The statement “women talk more than men” tends to imply the judgment “women talk too much”. (As one old proverb charmingly puts it: “Many women, many words; many geese, many turds.”)
The folk-belief that women talk more than men persists because it provides a justification for an ingrained social prejudice. Evolutionary psychology is open to a similar criticism: that it takes today’s social prejudices and projects them back into prehistory, thus elevating them to the status of timeless truths about the human condition.
Champions of the evolutionary approach often say it is their opponents whose arguments are based on prejudice rather than facts or logic. They complain that feminists and other “PC” types are unwilling even to consider the idea that sex-differences might have biological rather than social causes. Instead of judging the arguments on their merits, these politically motivated critics just denounce them, and those who advance them, as reactionary and bigoted.
But their stories have a basic flaw: they are based not on facts, but on myths.
© Deborah Cameron. Extracted from The Myth of Mars and Venus published, by Oxford University Press in hardback at £10.99. To order a copy for £9.99 with free UK p&p go to guardian.co.uk/bookshop or call 0870 836 0875.