(FORTUNE Magazine) – As everybody knows, men and women tend to hold different kinds of jobs. We increasingly view the existence of these differences as a major social problem, and we take for granted affirmative-action programs designed to reduce the differences--to get more women into jobs traditionally held by men, and vice versa. The background thought of these programs is that men and women are segregated because of discrimination. The central assumptions are that employers force or pressure women to take less-skilled, lower-paying jobs and that society encourages young girls, not to mention young boys, to view these arrangements as natural. Are the assumptions valid? If you turn to Sex Segregation in the Workplace (National Academy Press, $22.95), edited by Barbara F. Reskin, you will find that many scholars believe they certainly are valid. The book assumes sex segregation to be at bottom like racial segregation--both being defeats for the American ideal of fairness. The authors view sex segregation not just as a fact to be explained but as an injustice to be redressed. In her introductory chapter, Reskin expresses the hope that the book will help us move toward a desegregated world in which men and women earn equal wages, this apparently being the ultimate test of fairness. The book's 18 chapters are mostly based on papers presented at a workshop held in 1982 under the sponsorship of the National Research Council, an arm of the National Academy of Sciences. As you would expect, given these auspices, the participants were serious scholars, mostly sociologists and economists, and the end product manifests prodigies of research. To be sure, it suffers from certain defects often associated with social science writing: you can expect jargon, repetition, arcane statistics, heaps of undigested data, and the familiar academic commitment to the passive voice, all presented with no index. Still, you do end up learning a lot about sex segregation.

A principal finding is that sex segregation declined sharply during the Seventies but was still remarkably high in the early Eighties. One way to measure the segregation is to ask what proportion of women (or men) would need to change jobs to ensure that all occupations were fully integrated. The estimate for 1981 was 61%, down from 68% in 1972. The authors' projections do not foretell so large a gain in the Eighties. As they point out, even that 61% estimate understates the reality of sex segregation. The measurement problem is that the available data tend to describe broad categories of workers, and these often tend to mask the true separation of men and women at work. Both men and women seem well represented in a broad category like ''health workers,'' but the data for narrower categories like ''physicians'' and ''nurses'' would instantly make it clear that they're holding quite different jobs. Segregation also tends to be underestimated because the national totals don't tell us enough about what's happening inside companies. The sales personnel at company A may be overwhelmingly female and their counterparts at company B may be overwhelmingly male, but if the two sales forces are the same size, they're together creating an illusion of integration among salespeople. In general, occupations tend to be far more balanced in the national data than they are within companies. The authors have relentlessly, often skillfully, dug out such details; however, they do not work so hard at trying to understand what the details might signify. Just why is it--to pick one rather obvious question not pursued by the authors--that things always look more integrated when you view the occupation broadly, and more segregated when you view it narrowly? One possible answer is that individual preferences and suitabilities don't really play a role in career choices until you define the career narrowly. No particular traits are associated with ''health workers,'' for example, but different kinds of people may well be drawn to doctoring and nursing.

THE POSSIBILITY that true sex differences may affect employment practices is essentially unexamined in Sex Segregation in the Workplace. The authors simply assume, and do not feel obliged to demonstrate, that men's and women's careers would be more or less indistinguishable but for male sexism and arbitrarily imposed sex roles. That assumption may be the proper place to start an inquiry, but it is hardly the only view worth considering. Some contributors occasionally seem to understand that persistent and widespread sex discrimination is an economic anomaly. If women workers were interchangeable with men, how could they be paid less in a competitive economy? If their labor truly came at bargain prices, there would be competition for it and the price would rise until the bargain was wiped out. One way around this logical difficulty is to search for noneconomic explanations. So there are repeated references in the book to men's desire to maintain ''patriarchy'': that is, the upper hand in society as a whole. By blocking success at work for women, men keep them subservient at home. The argument is that employers support the principle of male control over the best jobs because they recognize that any breakdown of the principle means that ''upper-level managers, entrepreneurs, and capitalists would soon find , their own male privileges under siege.'' This view seems excessively conspiratorial and in any case is not supported by the authors' abundant data. Much of the economics profession explains discrimination against working women as a kind of penalty they suffer for their repeated moves in and out of the labor force--absences related to child raising and other family responsibilities. The authors of Sex Segregation in the Workplace take a hard look at this explanation and essentially reject it. So far as pay is concerned, their regressions tell them that the bulk of the wage gap is not explained by differences in men's and women's work histories. They argue that the wage penalty associated with departures and returns tends not to be severe. (''The period following the return is characterized by rapid wage growth, and the net loss . . . is small.'') Furthermore, the authors find no evidence that all this coming and going is related to sex segregation. They find that women with ''discontinuous work careers'' are no more likely than other women to end up in segregated jobs. In rejecting these popular explanations for our sex-segregated labor force, the book will leave many readers wondering what explanation they're supposed to think of as correct. The authors acknowledge that their research has not satisfactorily explained the phenomenon. They apparently prefer this explanatory vacuum to an obvious alternative. The alternative is the possibility that women and men differ in ways that significantly affect them as workers. What if they differ innately--that is, for reasons arising out of their differing natures--and not just because of the ways they're molded by society? If that were true, it would no longer be self-evident that the sexual division of labor is unjust and should be eliminated. Men and women do in fact differ, on the average, in many ways that affect their job roles. They differ, of course, on physical measures like upper-body strength. More important, psychologists have repeatedly found differences, many of which are almost certainly innate to some extent, in a broad range of traits: in motor and perceptual skills, in patterns of intellectual ability (men tend to be superior in mathematics and reasoning, women in language use), in values placed on personal association (as distinct from abstract obligations), in attitudes toward love and the family, in ambitiousness and assertiveness, and more. It would be extremely odd if none of these were relevant to people's jobs. , To be sure, these differences are typically minor and are measurable only in large populations; anything is possible in an individual case. But minor sex differences in broad averages can translate into major differences in occupational patterns. Suppose, for example, that a job had five qualifications, each of which could be met by 60% of men and 40% of women. An unbiased employer who successively screened out applicants failing to meet each qualification would end up with a labor force that was 8 to 1 male (because .6 to the fifth power is about eight times .4 to the fifth power). ANYONE FAMILIAR WITH the psychological literature on sex differences might be inclined to argue with the assumptions and conclusions of Sex Segregation in the Workplace (none of whose authors are psychologists). For example, the book describes surveys of the job aspirations of 14- and 15-year-olds. It turns out that the boys mainly look forward to doing what has traditionally been thought of as men's work, the girls to doing women's work. The authors take these findings to be evidence that the forces of sex-role socialization have already taken their toll. But this conclusion conflicts awkwardly with another finding: after eight more years of exposure to those socializing pressures, the young men and women (now in their early 20s) aspire to somewhat less segregated occupations. What's missing in all the literature on socialization are data about some period in a young person's life when he or she is oblivious to the associations between jobs and gender. Even nursery-school children want the sexually segregated jobs appropriate to their own sex. As soon as a child can be meaningfully questioned about career aspirations, the sex differences are there. In children under 10, boys are more aware of and concerned about the income associated with different occupations than girls are. In wondering about the future of sex segregation, we can assume that some of the traditional job differences will continue to be wiped out by technology. The average male's broad back will continue to play a reduced role in occupational segregation of the sexes, and perhaps all of the demonstrated physical and intellectual differences between them will ultimately prove irrelevant too. It seems less plausible that the differences in personality and motivation will become irrelevant. At any rate, it is surely wrong, and surely unscientific, to assume as an article of faith that sex segregation in the workplace is entirely unnatural.