The first recorded use of the term “sexual harassment” was in a letter written in March 1975 by lecturers at Cornell University, circulating the case of a woman Carmita Wood who had been refused welfare benefits to support herself and two children. Why had she needed benefits? She had been forced to leave her previous job, the letter went on, after eight years’ service, “because of a pattern of sexual harassment by a male superior which caused her tension and anxiety so severe that she developed physical side effects.” Her employer had refused to investigate the complaint and she had resigned. On seeking unemployment benefits, she was told that the reasons for leaving her job were merely “personal” and “non-compelling”.
Wood maintained that Boyce McDaniel, a physicist at Cornell University’s Laboratory of Nuclear Studies where she also worked as an administrator, had placed his hand on her bottom, and on other occasions leaned against her while she sat at her desk, pinned her body between his and her desk, stared repeatedly at her body, attempt to kiss her, and at one year’s Christmas party, rubbed his hands up and down her back while lifting her clothes and exposing her back. McDaniel went on leave. As his return approached, Farley began to suffer pains in her arm and hand, which a psychotherapist told her been related to the stress of his unwanted attentions. She complained to the Executive Officer of the Nuclear Laboratory, Henry Doney, who encouraged her to “try not to get into those situations”. Wood resigned and applied for benefits. Employees of the New York State Department of Labor saw her case as a chance for humour, joking when they heard of her symptoms that “So you’re saying, in effect, that he was a pain in the neck?” When she was mentioned to them that McDaniel was 57 years old, she was told “Well, he’s young enough to be interested, anyway” (Strebergh, 2009, pp 218-226; also Baker, 2008, p28).
Looking at this behaviour, why was McDaniel wrong to behave as he did? What moral principle does sexual harassment breach? Here, I’ll consider two main approaches to this question, the ideas of the liberal philosopher Edmund Wilson, in which sexual harassment is portrayed as an inappropriate form of communication, and of the feminist lawyer and activist Catherine MacKinnon, who emphasises its gendered aspect, before posing a third alternative, which sees the typical form of sexual harassment as harassment in the workplace, i.e. an act that is about gender and also about enforcing managers’ power over their workers.
Edmund Wall: harassment as miscommunication
One of the most often cited definitions of harassment was first proposed twenty years ago by the liberal American philosopher Edmund Wall. In his definition there are four stages to sexual harassment. First, the person doing it chooses not to ask for the consent of the person to whom the act is done (“X does not attempt to obtain Y’s consent”). Second, the content of the harassment is the communication of a sexual interest in the person who is being harassed (“X communicates … sexual interest”). Third, the person to whom it this is communicated does not consent to it. Fourth, the person to whom the sexual harassment is done is upset (“emotionally distressed”), either because of the act or because of the lack of consent.
This is a careful definition, and one that has been nicely constructed, so as to make it capable of informing legislation. It also has problems. The first step “does not attempt to obtain … consent” adds nothing. If you think about it as a courtroom advocate might, i.e. with each phrase constituting a separate test that would have to be satisfied before sexual harassment was established, and each requirement (if it was not satisfied) also providing a potential defence to a harassment claim, then there is no good reason for the first step at all. Otherwise by merely first asking “Is it ok if I lift your clothes off?”, McDaniel would not have committed sexual harassment when he then lifted Wood’s clothes, even if her answer to him had been an unequivocal “No, please don’t”.
Wall’s definition of sexual harassment already has a requirement of consent, which is its third stage. Why does he add in this extra requirement, which at first sight seems to do no more than duplicate it? I suppose that he must have had a sense of the difficulties someone bringing a case has in proving an act was unwanted. You can imagine the exchange in court: “I did not want McDaniel to remove my clothes.” “But did you tell him beforehand not to remove them?” “No, I did not.” “If you really didn’t want him to do take your clothes off, surely you would have asked him beforehand not to remove them?” “Maybe; but you need to understand where we were, at a work party with all our other colleagues. I never expected him to start removing my clothes…”
Wall may have thought that the women’s position would be protected by requiring a man first to have expressly requested her consent, and then not to have had it, when he proceeded. But if that was his idea that is not what the first limb actually does, and it would need further rewriting to have this effect. It simply seems never to have occurred to him that a person might hear a “No” or several Noes and still continue. However, the deliberate ignoring of No signals is endemic in harassment cases. The typical rape case, like the typical rape case, involves a man ignoring repeated negatives: “I do not want to meet you outside work”, “I do not want to be taken by you to a bar”, “I do not want you to buy me drinks”. The man may think he is being gallant in overcoming the woman’s resistance. If the woman appears to assent it is from fear of the power he holds over her as her employer.
For Wall the essence of sexual harassment is the communication of desire for a sexual encounter (“sexual interest”). Why did Boyce McDaniel place his hand on Carmita Wood’s bottom? For Wall, the answer is straightforward, because McDaniel was sexually attracted to her. When it comes to rape, there has long been a debate about whether the purpose of the act is really sex or power (see, for example, Bourke, 2007 pp12-13). The soldiers who visit wartime brothels appear to be satisfying their desire for sex; the soldiers who rape the women of a defeated country seem to be doing something significantly different – humiliating them as another step in defeating them. Although it is a different act, we should admit at least the possibility that something like the same thing happens in sexual harassment, namely that some of its about wrong is about more than unrestrained sexual desire.
In Drucilla Cornell’s theory of sexual harassment, what matters is not the man’s sexual desire, but the woman‘s desire (which Cornell terms, her “imaginary domain”) and her interest in men or women but not in her manager. “The chance of sexual freedom” requires the woman’s sexual capacity to be unlimited. But her ability to love freely will be diminished if her self-respect is undermined. And her dignity will be undermined if the woman is repeatedly sexually harassed (Cornell, 1995, p170).
What if a legal code said that in order for there to be sexual harassment, the man doing it must have actually been sexually interested in the woman? Respondents would be able to resist harassment complaints by saying, “well, now you mention it, I was never attracted to my victim”. If they were believed, this would be a defence and they would be cleared, and this would make the law ridiculous. It is far better to focus on the effect of conduct on the person who is on the receiving end – why it was upsetting behaviour – than to have to speculate about why the perpetrator of the harassment focussed on her.
Stage three of Wall’s definition is – rightly – the woman’s consent. But the clarity of this part of his definition is then confused by his fourth stage, which gives the reason for the lack of consent as either failure to communicate or the act itself. Of course, in the vast majority of cases it is the act which is the issue: in Wood’s case, the touching, the staring, the attempted kiss, the removal of clothes. “Emotional distress” is in fact merely the symptom of lack of consent. In so far as Wood is saying that in sexual harassment cases the victim might potentially be upset purely because they did not consent, this is confused and circular reasoning.
At the heart of Edmund Wall’s theory of sexual harassment is the idea of a wrong communication that violates the autonomy of the person who receives the message. In his words, “What is inherently repulsive about sexual harassment is not the possible vulgarity of X’s sexual comment or proposal, but X’s failure to show respect for Y’s rights. It is the obligation that stems from privacy rights that is ignored” (Wall, 1992, p74). Wall seems to assume that most harassment is like a love letter which has by ill-fortune been sent to the wrong recipient; or a heartfelt kiss for someone who feels no attraction back. You can imagine the dialogue: “I wouldn’t have minded being kissed, if it was my boyfriend who I love who had kissed me, rather than my boss”. In Wall’s theory most sexual harassment is essentially good and desirable conduct which but for its context would be welcome by its recipient. But much harassment – jeers, stares, whistles, aggressive propositioning – is behaviour that would be upsetting whoever had done it.
Going back to the behaviour described by Carmita Wood, when a man pins a woman with his body, stares repeatedly at her, removes her clothes in front of her work colleagues, this just does not feel like the sort of behaviour that the man would actually do if he was hoping to become or was already the woman’s lover. It might be about lust, but not about any serious attempt at courtship. It is about marking a different kind of relationship, one shaped by the man’s greater power in the workplace.
Catherine MacKinnon: harassment as discrimination
Published just four years after the Cornell lecturers had written their letter describing Carmita Wood’s case, Catherine MacKinnon’s book Sexual Harassment of Working Women provided it, and the other cases of sexual harassment which had then come to light, with a theoretical explanation. Earlier drafts of the book had in fact been used as advocacy in Wood’s case; and an appendix contained draft particulars of claim which other lawyers might use in this emerging area of the law.
For all the stern words that MacKinnon’s many critics on the left have said about her other choices as an activist (eg Smith, 1991), it is worth remembering that her intervention helped to give the social wrong of sexual harassment its content, and that in this way she has helped to develop a redress for unwanted and humiliating behaviour where previously women had no rights at all.
The theme of her argument was that sexual harassment was not merely a wrong (in legalese, a “tort”) that might happen to men or women inside the workplace or without; it was a specific form of sexual discrimination, and, as such, claims of sexual harassment should be determined under the Civil Rights Act 1964 which made it unlawful for an employer to discriminate against an employee in terms of that employee’s sex.
Central to her argument is the idea that women face harassment in the workplace in far greater numbers than men. “Work is critical to women’s survival and independence”, MacKinnon wrote, “Sexual harassment exemplifies and promotes employment practices which disadvantage women in work (especially occupational segregation) and sexual practices which intimately degrade and objectify women.” (MacKinnon, 1979, p7). This analysis is true in fact, and necessary to MacKinnon’s argument at a time when the dominant conception of discrimination held that it was primarily about unjustifiable diversions from the principle of equal treatment. In this understanding, where a man employed two people, one male and female, and harassed the latter, his behaviour was wrong because it meant treating the woman worse. If a man sexually harassed all his employees, male and female, according to this dominant view in the courts (known colloquially as “the bastard’s defence”), then there was no differential treatment. The conduct might be unwanted, but men and women were being treated equally badly, and neither could complain.
MacKinnon’s book opens with a flurry of statistics showing that men are treated better than women in work, that women are segregated into lower-status jobs, that men are more likely to be managers and are better paid. Many other writers, feminists, Marxists and at all points in between, have suggested that there are a number of jobs which are defined as female and where segregation proceeds from the assumption that women simply make better, more skilled workers: childcare assistants in nurseries, primary school teachers, nurses, cooks. Here women’s presumed greater understanding of child-care serves to guide them into jobs which are segregated in gender terms, enabling women to be paid far less than men for the work they do.
Sexual harassment of working women takes the argument further, suggesting that the characteristic form of women’s employment under modern industrial society, is secretarial work, i.e. work in which women serve adult men. “Part of the reason the woman is hired us to be pleasing to be a male boss whose notion of a qualified worker merges with a sexist notion of the proper role of women … work relationships parallel traditional home relationships between husband and wife”. Sexual harassment, it follows, is structured into women’s experience of work as adjuncts to men: “the very qualities which men find sexually attractive in the women are the real qualifications for the jobs for which they hire them” (MacKinnon, 1979, p17, 23).
I am not at all sure that MacKinnon’s refinement actually strengthens the arguments that others had made previously. The next 25 years of capitalist development have witnessed a continuing expansion of “women’s work” in jobs related to childcare, for example, through a massive increase in the total number of teachers. At the same time, computerisation has tended to push back onto men tasks such as typing which were previously considered women’s work. The pattern varies from sector to sector, but overall the skilled office jobs which were once seen as naturally men’s work, are no longer exclusively the preserve of either gender.
A better, lawyer’s, point which MacKinnon grasped is that harassment could not be opposed simply as a breach of the equality principle. The first reported complaint of sexual harassment in the US court system, Corne and DeVane v Bausch & Lomb had foundered in part on the principle that the conduct was not discrimination. Brought by two female clerical workers who complained of repeated unwanted verbal and physical sexual advances, one of the reasons given by Judge Frey for dismissing the claim it was that “if the conduct complained of was directed equally to males there would be no basis for suit” (MacKinnon, 1979, pp60-1), in other words, because a man might be sexually harassed by woman, logically, this behaviour was misconduct was not exclusive to men, and since it could be done either to men or women, no woman who had been harassed could ever have a case.
In Barnes v Train (Costle) it was accepted that harassment might be gender-derived, “But for her womanhood, from that aught that appears,” the court held, “her participation in sexual activity would never have been solicited” (MacKinnon, 1979, pp67). The granting of the principle that the reason for a man’s harassment of a woman could be her gender did not dispose of the matter, unfortunately. An employer’s counter-offensive then began directed to saying that the reason why the conduct took place was not a woman’s gender but her sexual attractiveness to the particular person who had carried out the harassment, and that this characteristic – her “sexuality” for want of a better word – although related to her gender was not reducible to it.
MacKinnon anticipated this argument, and attempted to address it. She noted that in the context of pregnancy, the US Supreme Court had had difficulty in accepting that discrimination against pregnant women was discrimination on grounds of sex. Other legal systems we can note in passing have faced the same problem. If you assume that discrimination is only about treating similar groups of people differently, then who is the similar person to whom a pregnant worker should be compared? Is it man, is it a man with an illness which will take him out of work for 9 months before he returns, or is it a woman who is not pregnant? The comparison you choose is likely to determine in advance whether you find that the pregnant woman’s treatment resulted in her being treated worse than others, and indeed whether or not you decide the difference was caused by her sex. When thinking about sexual harassment, she grasped, it is not enough to just think about gender.
“The relationship of sexuality to gender is the critical link in the argument that sexual harassment is sex discrimination”, MacKinnon ventured. Sexuality, she explained, is a large part of the way in which people learn to live in different genders. It is expressed in sex roles: i.e. the sexual segregation of workplaces and the ubiquity of stereotypes that women are weak, good with their fingers, bad with numbers, bad at long hours, too emotional, etc (also Laws and Schwartz, 1977). Sexual expression in defined sex roles leads to dysfunctional sexual rituals characterised by assumptions of male activity and female passivity. At the heart of sexual harassment it follows is “male ego aggrandizement”, the conversion of millions of female workers into secretaries. MacKinnon concluded: “sexual harassment expresses one social meaning that sex roles create in the sex difference; gender distributes power as it divides labor, enforcing that division by sexual means” (MacKinnon, 1979, pp151, 158).
In general, this account of “ego aggrandizement” seems to capture something missing from Edmund Wall’s theory. Sexual harassment can be, as Wall seems to imagine it, a relative humdrum form of sexual advance and negotiation, made unwelcome only because the woman involved is unwilling. But it can also be behaviour which would be in any context disturbing. MacKinnon explains this type of harassment by connecting it to the man’s desire to impose his will on the woman, to make himself feel important at her expense. MacKinnon was right to see this side of harassment, this is indeed a part of its story.
Yet what MacKinnon’s account misses is a sense of hierarchy which connects gender and class. In passing, she notes an early decision of the US courts, Tomkins v Public Sector Electric & Gas Co, in which Judge Stern held that part of the pain of sexual harassment is role often played in it by supervisors: “the power inherent in a position of authority is necessarily coercive … every sexual advance made by a supervisor would be made under the apparent cloak of that authority. Any subordinate knows that the boss is he boss whether a file folder or a dinner is at issue”. Stern “firmly grasped the nettle of hierarchy”, MacKinnon writes (MacKinnon, 1979, p71), but her account does not develop this point, does not explain that it is the manager’s freedom to manage which leaves harassment ordinarily unpunished. It is as if MacKinnon could see every hierarchy, except that of class.
Sexual harassment as a means to reinstate managerial hierarchy
There are other kinds of harassment which are close to sexual harassment, and sometimes given the same label, but are logically distinct from it. One is behaviour in which the reason for the behaviour, or the reason for its degrading effect, is not the victim’s sexuality but her gender. Carmita Wood’s story illustrates these distinctions; while the motivation of McDaniel appears to have been at least partly sexual (in Wall’s terms, he had a “sexual interest” in her), the employees of the New York State Department of Labor were humiliating (“harassing”) her too without a sexual motive. This was outside the workplace, but the men were similarly in a relationship of power to the women (they would decide whether she would receive benefits or not). And the hurt of their behaviour was their sexual interest in her, but their demeaning of her as a woman who had complained about a man.
There are at least two possible types of sexual harassment: first, unwanted conduct which is done to humiliate a woman as a woman, and second, unwanted conduct which is done to humiliate a woman as an object of desire. Wall seems to miss not merely the first of these categories, but the capacity of this second type of harassment to cause deep and lasting hurt, and even the potential for these categories to overlap (e.g. where a man humiliates a woman, ostensibly by stating his attraction to her, but also to mark and reinforce some element of his power over her in the workplace).
The further you get away from the workplace, this less this behaviour looks like what employment lawyers think of sexual harassment. For example, in Britain the main legislation which criminalises harassment outside the workplace or the family, and which enables people in civil cases to get orders from the court stopping harassment, was drawn up to prohibit stalking. This behaviour may be unwanted conduct, and it may even have a sexual content, but at least potentially, it can be done by people with significantly less power or status than their targets. By contrast, when we think about Carmita Wood, a significant part of the hurt that she was caused is that the act took place in the workplace (i.e. in public, and where most of us spend more of our waking lives than anywhere else), and was done by her superior (i.e. a person who was ordinarily responsible for her, and for her safety in the workplace, and who had power over her including the power to dismiss her).
The liberal philosopher Edmund Wall seeks to isolate sexual harassment as much as he can from the gender and class relationships within which it takes place. It is of considerable importance to him that his theory is neutral in terms of both gender and class . Wall states that a definition of sexual harassment must be found which takes account of the possibility for harassment “if the employee propositioned the employer” (Wall, 1992, p74, his emphasis). This is one of those occasions where the magical capacity of a philosopher for logical abstract reasoning lets him down.
Proceeding by analogy: when thinking about the harassment of one person on the grounds of race, it is important to grasp that the racial categories of the perpetrator and the victim are irrelevant. Just to give a single example, if a white person (X) commands another white person (Y) to harass a third, black person (Z), and Y refuses X’s order and does nothing, then Z has never been harassed. Should X dismiss Y, the only victim of the discriminatory behaviour will be a white person. X and Y are the same race, but Y has undoubtedly suffered from X’s intention to racially harass.
When thinking about sexual harassment some of the same logic applies. We know that the vast majority of complaints of sexual harassment are made by women about men. The behaviour can sensibly be described as “male” behaviour, in that it connects to ways in which men are socialised into gender in capitalist society. Acknowledging that men are more likely to harass does not mean assuming that men have a right to harass or that it is an unfortunate but pervasive male privilege but to blame deep-rooted patterns of how people are taught to be masculine for what is dehumanising and offensive behaviour (Segal, 1987, pp104-5). From the reality that most sexual harassment is perpetrated by men we can reasonably infer that most sexual harassment is aimed at women. But there is no reason of principle why the source of some of the behaviour might not be a woman mistreating a man, a man harassing a man, or a woman whose behaviour violates the dignity of another woman. These will be a minority cases, but just like rape, which can be done by either men or women, and to men, they are not impossible.
For Wall, it is important that the gender definition of sexual harassment is “mirrored” (i.e. that it allows for harassment to take place either by men or women). This is perfectly understandable and right in principle. He is wrong however when he assumes that in class terms harassment should also be “mirrored”. It simply does not seem to have occurred to him that the majority of employees in the United States and Britain work for companies, i.e. the employer is not an actual person, there is no “Mr Ronald MacDonald CEO” waiting to be harassed by a sexually interested fast food employee, “Jane Doe”. Nor does he pause to ask why it is that the overwhelming majority of sexual harassment complaints are made by junior employees complaining about harassment from co-workers or managers above them rather than by senior managers complaining about the behaviour of those below.
In a workplace, there are at least three potential relationships between the person doing the unwanted conduct and the person to whom it is done. It is possible that the harassed person could (1) be a superior (an owner or a senior manager) or (2) at a position of relative equality with the perpetrator, or (3) a more junior employee. There are no figures, either in the US or Britain, for how many victims of harassment are in each category. But we do know that the harassment of superiors is so rare as to be almost unknown.
Of all the thousands of sexual harassment cases in the courts each year a proportion are appealed, and are then reported, and appear on free, public access sites such as (in the UK) Bailii. They can be read and anyone can get a rough sense of how many of each type of case there are from year to year. Of the appellate cases of sexual harassment, at least in the UK, a vanishingly small number of cases are about junior employees (of either gender) harassing more senior people. In any year, the number of women who report rape outnumber the number of men who report it by around 10 ten 1. We can reasonably assume that at least 1 in 10 victims of sexual harassment at work are men and the proportion may even be higher than that (harassment is in all likelihood closer to domestic violence, where depending on who is asked and how they are asked, the ratio of female to male reports can be as close as to 2 to 1). But any experienced employment lawyer would confirm: not one in 50 cases of sexual harassment concern senior managers complaining about their harassment by junior employees. Harassment of a manager is so rare not because it is a crime against women (in terms of its gender dynamics, the sexual harassment of superiors is no different from the harassment of subordinates). Harassing a manager is wrong, because it offends the hierarchy and results in summary dismissal.
Most sexual harassment is done by managers of more junior workers. These are the types of cases which really manifest “ego aggrandizement”, in MacKinnon’s terms, in that they are about powerful men manipulating the fact that since they are managers they will not be dismissed if there is a dispute between them and their more junior employee. Their power means that they can behave badly and get away with it. The harassment may skip between gender-related harassment (e.g. asking a relatively senior co-worker, but one more junior in the hierarchy than the perpetrator to “be mum and make tea”), harassment taking a sexualised form (eg McDaniel’s mistreatment of Wood) and conduct which has no obvious reference in either gender or sexuality (dismissing the woman from the room, shouting at her to her face, using aggressive but not sexualised language or behaviour). Most workers would recognise such sexual harassment as simultaneously an abuse and a deliberate use of managerial power. The purpose of the communication is to say to the woman: I am your manager, I stand in place of your employer, I can manipulate you directly and treat you in an unwanted sexual manner. You will have to accept my behaviour, because it has the implied support of the company. And in accepting my behaviour without having any response, you are confirming my power over you.
The harm of sexual harassment is not lessened because its victim is employed by a charity, say, or a campaign or union or some other good cause. If anything, campaigns and parties and kindred groups need to be more attentive to the risk of sexual harassment as they are likely to be characterised by very high levels of identification between their employees and the organisation, the sort of values which in turn promote the institutional disbelief of complainants, whose complaints are easily assumed to be attacks on the cause rather than on a harassing manager. Charities, unions and their like also face the “problem” of large numbers of volunteers, in roles akin to that of the employee, but even more junior. From the perspective of men who harass people in workplaces, these are even easier targets, as they may lack institutional support within the organisation to complain of harassive behaviour. It is not “bourgeois morality” to look askance at employers who have or seek sexual relationships with workers and volunteers who are many years their junior, or play less prestigious roles within the organisation, or lack the social movement capital of their seniors. It is a necessary concomitant to an understanding of the way power operates, even within the left.
Sexual harassment is a device by which managers confirm their power over their subordinates. Its hurt derives in part from the sexual setting within which this power is expressed; sex being such an important part of our common humanity. But it derives equally from the victim’s feeling of powerlessness, from her knowledge that if she complains her employer will in all likelihood turn its every resource against her. This is sexual harassment in its purest, most defining form; it is one of the nastiest ways in which those with power in the workplace feel free to control and humiliate their more junior employees.
Backhouse, C. and Cohen, L. (1978), The Secret Oppression: Sexual Harassment of Working Women. Toronto: Macmillan.
Baker, C. N. (2008), The Women’s Movement against Sexual Harassment. Cambridge: Cambridge University Press
Banyard, K. (2010), The Equality Illusion: The Truth about Men and Women Today. London: Faber & Faber.
Cornell, D. (1995), The Imaginary Domain: Abortion, Ponrography and Sexual Harassment. London: Routledge.
Laws, J. L. and Schwartz, P. (1977), Sexual Scripts: The Social Construction of Female Sexuality. Hinsdale, llinois: Dryden Press.
MacKinnon, C. (1979), Sexual harassment of Working Women. New Haven: Yale University Press
MacKinnon, C. (1989), Towards a Feminist Theory of the State. Cambridge, Massachusetts: Harvard University Press.
Segal, L. (1987), Is the Future Female? Troubled Thoughts on Contemporary Feminism. London: Virago.
Sergrave, K. (1994), The Sexual Harassment of Women in the Workplace 1600 to 1993. Jefferson, North Carolina: Mcfarland and Company.
Smith, S. (1990), ‘Feminists for a strong state?’, International Socialism Journal 51, pp. 71-90.
Strebergh, F. (2009), Equal: Women Reshape American Law. London: W. W. Norton and Company.
Wall, E. (1992), ‘The Definition of Sexual Harassment’, in E. Wall (ed), Sexual Harassment: Confrontation and Decisions. New York: Prometheus Books.