it under illuminating and powerful generalizations – but also for strength, in the sense of informativeness, and for simplicity, elegance, and avoidance of the ad hoc. The method is sometimes called inference to the best explanation, though philosophical explanations are constitutive rather than causal. The first edition is quite consistent with the abductive aspect of philosophy, which is implicit in the chapter on evidence in philosophy, but somehow it remained in the background.2 The omission was brought home to me when I gave a week-long colloquium based on the book at the University of Göttingen in 2009, invited by the students: I found myself answering question after question with reference to the role of abduction in philosophy, and wondering why I had not said more about it in the book itself. For the abductive aspect of philosophy was nothing new to me. During my doctoral studies at Oxford in 1976–1980, my closest friend amongst my fellow graduate students in philosophy was Peter Lipton, whose DPhil thesis later turned into his classic treatment Inference to the Best Explanation.3 The relevance of the topic to assessing philosophical theories was salient to me even then. In my book Vagueness, the overall case for classical logic was fundamentally abductive (e.g. 1994a: 186). In this second edition, the additional Section 9.2, “Abductive Philosophy,” fills this gap in the first edition; Section 9.5 briefly responds to some criticisms of the approach.4
Another omission was the methodology of model-building. I had started thinking seriously about it thanks to having the economist Hyun Song Shin as a colleague at University College Oxford in 1990–1994, trained in philosophy too, with a degree in PPE (Philosophy, Politics, and Economics) from Oxford. We shared an interest in epistemic logic, on which we published two joint papers (Shin and Williamson 1994, 1996). Our collaboration gave me fascinating experience of the differences in research culture between two disciplines when dealing with the same phenomena, in this case knowledge and ignorance of one’s own or another’s knowledge and ignorance. As an economist, he was used to a model-building approach, on which models are assumed from the outset to involve drastic simplifications of the reality under study, so that a mere discrepancy between model and reality is not news, and just pointing it out is not considered a significant intellectual contribution. Rather, what displaces a model is a better model. He once remarked to me, of Gettier’s seminal paper (1963) refuting the analysis of knowledge as justified true belief by counterexamples, that in economics it would have been considered unpublishable. As a philosopher, used to treating counterexamples as the gold standard, I was shocked. Did these economists not care about truth? On second thoughts, however, I realized that the model-building methodology was just as oriented towards truth as the potentially naïve falsificationism of conjectures and refutations by counterexamples (thought experiments), though in a subtler and less direct way. One of our joint papers used an explicitly model-building methodology, and it was employed in an increasingly prominent role in some of my own publications from that period on.5 “Must Do Better,” the Afterword to the first edition, recommends the use of mathematical models to test philosophical ideas (293, this volume), though without discussing such methods in detail. Later reflection on the nature of progress in philosophy convinced me that, like progress in natural science, much of it takes the form of building better and better models of the phenomena under study, rather than discovering exceptionless universal laws, and that failure to recognize the model-building methodology is one of the reasons for widespread overestimation of the difference between philosophy and natural science. In that respect, the additional Section 9.3, “Model-Building in Philosophy,” goes far beyond the first edition, while Section 9.6 briefly considers a proposed alternative.6
A recent side interest, which played no role in the first edition, has been the surprisingly effective dialectical role of moral and political considerations in philosophical debates which seem to have nothing specifically to do with the moral or political – for example, over general relativism, general skepticism, and general internalism in epistemology. The story of how I first came to notice this phenomenon tempts me into a digression.
As a graduate student at Oxford, I used to attend meetings of the Radical Philosophy group, associated with the journal Radical Philosophy. In practice, what was philosophically radical about it was its rejection (and often ignorance) of analytic philosophy, in favor of just about anything which then counted as “continental” – they discussed Nietzsche, Saussure, Althusser, Derrida, the more arid parts of Foucault’s corpus, and so on, with varying degrees of reverence. The “analytic”-“continental” distinction cut at an obvious joint in the sociology of philosophy, however artificial it may have been in other respects. I experimented with those alternative traditions because I felt oppressed by the style and assumptions of the kind of analytic philosophy then most fashionable in Oxford, and hoped that I might find different ideas for use in my own work. I didn’t get much out of the experiments, though I enjoyed reading Nietzsche and Saussure. I came to realize that those who led the discussion often understood the obscure texts they talked about no more clearly than I did, although they certainly had a far more extensive acquaintance with them than mine, and were willing to “go on in the same way” as their authors. On the rare occasions when I asked a question or made an objection, they never seemed in danger of getting the point. There were one or two exceptions, fully open to rational discussion of ideas from both sides of the divide – one was Michael Rosen, now at Harvard. After I had left Oxford for my first proper teaching job, at Trinity College Dublin, I felt liberated to discover that what had really oppressed me about the then-predominant style of Oxford philosophy was not that it was too analytic but that it was not analytic enough. However, one of the things I did learn from my Oxford experience of Radical Philosophy was this: within such an intellectual world, much of the resistance to the relativist-sounding extremes of Post-Modernism came from Marxists and others on the far Left, who feared relativism as a threat to their political hopes. How far will those who view the case for revolution from a relativist stance commit to the revolutionary cause? In that world, objections to relativism from common sense, natural science, or logic had much less credibility. Later, while in Dublin (1980–1988), I was intrigued to hear from a talk by Richard Kearney (now at Boston College) of Richard Rorty describing absolutism about justice as much harder to give up than absolutism about truth. I was never tempted to give up either, but I could imagine how someone more concerned with morality and politics than with logic might feel that way.
I did nothing with those thoughts at the time, but they stayed with me. Much more recently, in responding to Paul Boghossian’s epistemological internalism, I found myself objecting that it counts as justified (though false) a consistent neo-Nazi’s belief that he ought to kill members of a target group, and wondering whether such a view would also count as justified (though wrong) his acting on that belief (Boghossian and Williamson 2020). That got me thinking more carefully about why emotive cases are dialectically effective, and whether invoking them is some kind of cheat. That is an obvious danger, especially in the current philosophical climate, where morally or politically wrong-footing one’s opponent is all too often used as a convenient excuse for not engaging properly with their arguments or objections. Nevertheless, I came to the conclusion that it is legitimate to use such examples in order to make vivid the practical consequences of a philosophical theory, especially one which had seemed to have none. The justification of belief and the justification of action should not be treated as orthogonal issues: the considerations for and against internalism are similar in the two cases, and after all the distinguishing mark of a belief is the agent’s willingness to act on it. The additional Section 9.4, “Morally Loaded Cases in Philosophy,” encapsulates my reflections on these issues.
The Preface to the first edition starts by expressing my long-held view that the self-images then salient for contemporary philosophy failed to fit its actual development over the preceding decades. The book aimed to help put that right. I had also long been aware of a related strangely growing gap in the historiography of analytic philosophy. When I started as an undergraduate at Oxford in 1973, historical narratives of analytic philosophy tended to stop the story around 1960. Naturally, I expected that, as time went on, the lag between the time of writing and the end of the period written about in narratives of analytic philosophy would remain roughly constant. It did not happen. Thirty years later, historical narratives of analytic philosophy still tended to stop the story
