Williams spent the bulk of his career at four academic institutions: Oxford, Cambridge, University College London, and the University of California, Berkeley. Early in his career at Cambridge, Williams became known internationally for his attempt to reorient the study of moral philosophy to history and culture, politics and psychology, and, in particular, to the Greeks. Described as an "analytic philosopher with the soul of a humanist," he saw himself as a synthesist, drawing together ideas from fields that seemed increasingly unable to communicate with one another. He rejected scientific and evolutionary reductionism. For Williams, complexity was beautiful, meaningful, and irreducible.
Williams was renowned for being sharp in discussion, with Oxford philosopher Gilbert Ryle once saying of him that he "understands what you're going to say better than you understand it yourself, and sees all the possible objections to it, all the possible answers to all the possible objections, before you've got to the end of your sentence." He also became known as a supporter of women in academia, seeing in women the possibility of that synthesis of reason and emotion that he felt eluded analytic philosophy. The American philosopher Martha Nussbaum said he was "as close to being a feminist as a powerful man of his generation could be."
He met his future spouse, Shirley Brittain-Catlin, the daughter of political scientist and philosopher George Catlin and novelist Vera Brittain, while he was on leave in New York, where she was studying at Columbia University. At the age of 22, after winning a prize fellowship at All Souls College, Oxford, Williams returned to England with Shirley to take up the post, although not before she had had a brief relationship with four-minute-miler Roger Bannister. They were married in 1955. Shirley Williams, as she became known, was later elected as a Labour Member of Parliament, then crossed the floor as one of the so-called Gang of Four to become a founding member of the SDP, a centrist breakaway party.
For 17 years, the couple lived in a large house in Kensington with the literary agent Hilary Rubinstein and his wife. During this time, described by Williams as one of the happiest of his life, the marriage produced a daughter, Rebecca, but the development of his wife's political career kept the couple apart, and the marked difference in their personal values — Williams was a confirmed atheist, his wife a devout Catholic — placed a strain on their relationship, which reached breaking point when Williams had an affair with Patricia Law Skinner, then wife of the historian Quentin Skinner. The marriage was dissolved in 1974, and Williams and Skinner subsequently married; their marriage produced two sons.
Williams was appointed Knightbridge Professor of Philosophy at Cambridge in 1967, and subsequently vacated the chair to serve as Provost of King's College from 1979 until 1987.
He left England in 1988 to become Deutsch Professor of Philosophy at the University of California, Berkeley, citing the relative prosperity of American academic life and the so-called "brain drain" from England of academics moving to the U.S. He told a British newspaper at the time that he could barely afford to buy a house in central London on his salary as an academic. He told The Guardian in November 2002 that he regretted his departure became so public: "I was persuaded that there was a real problem about academic conditions and that if my departure was publicized this would bring these matters to public attention. It did a bit, but it made me seem narky," adding "it's harder to live with a family out there [i.e. the U.S.] than I supposed."
Late in life, he held simultaneous positions as Deutsch Professor at Berkeley, and White's Professor of Moral Philosophy at Oxford, eventually becoming a Fellow of All Souls College once again.
In the 1970s, he chaired the Committee on Obscenity and Film Censorship, which reported in 1979 that: "Given the amount of explicit sexual material in circulation and the allegations often made about its effects, it is striking that one can find case after case of sex crimes and murder without any hint at all that pornography was present in the background." The Committee's report was evidently influenced by the liberal thinking of John Stuart Mill, a philosopher greatly admired by Williams, and who used Mill's principle of liberty to develop what he called the "harm condition," whereby "no conduct should be suppressed by law unless it can be shown to harm someone."
Williams concluded that pornography could not be shown to be harmful and that "the role of pornography in influencing society is not very important ... to think anything else is to get the problem of pornography out of proportion with the many other problems that face our society today." The committee reported that, so long as children were protected from seeing it, adults should be free to read and watch pornography as they saw fit.
Apart from pornography, he also sat on commissions examining the role of gambling, drug abuse, and private schools. "I did all the major vices," he said.
Although Williams' disdain for reductionism sometimes made him appear a moral relativist, he argued that moral concepts could be "thick" or "thin"; the former – but not the latter – are "world-guided" in that they are related to real features of the world such that, in principle, disputes about them can be resolved objectively.
A much discussed thought experiment of Williams', directed against utilitarianism, centres on Jim, a scientist doing research in a South American country led by a brutal dictator. Jim finds himself in the central square of a small town facing 20 rebels, who are captured and tied up. The captain who has defeated them says that, if Jim will kill one of the rebels, the others will be released in honour of Jim's status as a guest. But if he does not, they will all be killed. Simple act utilitarianism says that Jim should kill one of the captives in order to save the others. For most consequentialist theories, there is no moral dilemma in a case like this; all that matters is the outcome.
Against this, Williams argued that there is a crucial moral distinction between a person being killed by me, and being killed by someone else because of what I do. The utilitarian loses that vital distinction, he argued, thereby stripping us of our agency and our humanity, turning us into empty vessels by means of which consequences occur, rather than preserving our status as moral actors and decision-makers with integrity. Moral decisions must preserve our psychological identity, he argued.
An advocate of utilitarianism would reply that the theory cannot be dismissed as easily as that. The Nobel Prize-winning philosopher of economics Amartya Sen, for example, argued that moral agency, issues of integrity, and personal points of view can be worked into a consequentialist account; that is, they can be counted as consequences too.
Sen and Williams also disagreed on the degree to which so-called "rule-utilitarianism" — a version of utilitarianism that promotes not the act, but the rule, that leads to the greatest happiness of the greatest number — could solve the fundamental problems of utilitarian theory. For example, to solve parking problems in London, Williams wrote, a utilitarian would have to favour threatening to shoot anyone who parked in a prohibited space. If only a few people were shot for this, illegal parking would soon stop; thus the utilitarian calculus could justify the shootings because of the happiness the absence of parking problems would bring to millions of Londoners. Any theory that has this as a consequence, Williams argued, should be rejected out of hand, no matter how intuitively plausible it feels to agree that we do judge actions solely in terms of their consequences.
Sen and others argued that rule utilitarianism would ask what rule could be extrapolated from the parking example. The rule that we ought to "shoot those who commit parking violations" is unlikely, in the long run and considering all its consequences, to maximize good outcomes. For Williams, however, making the argument was itself a reason to reject rule-utilitarianism; we should already know that threatening to shoot people over parking offences is wrong, he argued, and any system that requires us to make that calculation is a system we should reject.
One of the main alternatives to utilitarian approaches is the moral philosophy of the 18th-century German philosopher Immanuel Kant. Williams' work throughout the 1970s and 1980s outlined the basis of his attacks on the twin pillars of utilitarianism and Kantianism. Martha Nussbaum wrote that his work "denounced the trivial and evasive way in which moral philosophy was being practised in England under the aegis of those two dominant theories."
Kant's Critique of Practical Reason and Groundwork for the Metaphysic of Morals expounded a moral system based on what he called the categorical imperative, the best known version of which is: "Act as if the maxim of your action were to become, by an act of will, a universal law of nature." This is a binding law, Kant argued, on any rational being with free will. You must imagine, when you act, that the rule underpinning your action will apply to everyone in similar circumstances, including yourself in future. If you cannot accept the consequences of this thought experiment, or if it leads to a contradiction, you must not carry out the act. For example, if you want to kill your wife's lover, you must imagine a law that says all wronged husbands have the right to kill their wives' lovers; and that will include you in future, should you become the lover of someone else's wife. In other words, Kant argued, you must universalize your experience.
Williams argued against the categorical imperative in his paper "Persons, character and morality." Morality should not require us to act selflessly, as though we are not who we are, in the circumstances in which we presently find ourselves. We should not have to take an impartial view, or a Christian view, of the world, he argued. Our values, commitments, and desires do make a difference to how we see the world and how we act; and so they should, he said, otherwise we lose our individuality, and thereby our humanity.
Philosophers have tried to argue that moral agents can have "external reasons" for performing a moral act; that is, they are able to act for reasons external to their inner mental states. Williams argued that this is meaningless. For something to be a "reason to act," it must be "magnetic"; that is, it must move us to action. But how can something entirely external to us — for example, the proposition that X is good — be "magnetic"? By what process can something external to us move us to act?
Williams argued that it cannot. Cognition is not "magnetic." Knowing and feeling are quite separate, and a person must feel before they are moved to act, he wrote, and as such he argued that reasons for action are always internal. If I feel moved to do something good, it is because I want to. I may want to do the right thing for a number of reasons. I may have been brought up to believe that X is good, and I may wish to act in accordance with my upbringing, or I may want to look good in someone else's eyes, or perhaps I fear the disapproval of my community. The reasons can be complex, but they are always internal and they always boil down to desire, he wrote.
The debt to Nietzsche is again clear, most obviously in the adoption of a genealogical method (a term in academic philosophy which essentially means studies based on the historical behaviour and thought of Mankind) as a tool of explanation and critique. Although part of Williams' intention was to attack those he felt denied the value of truth, the book cautions that to understand it simply in that sense would be to miss part of its purpose: rather, it "presents a ... challenge" to both "the fashionable belief that truth has no value" and "the traditional faith that [truth's] value guarantees itself." The Guardian wrote in its obituary of Williams that the book is an examination of those who "sneer at any purported truth as ludicrously naive because it is, inevitably, distorted by power, class bias and ideology."