1.

I have read Bernard Williams’ ‘Philosophy as a Humanistic Discipline’ maybe two dozen times over the years. It’s a sign of an ambivalent appreciation. I am admiring of Williams’ eloquence, and sympathetic to the essay’s spirit. That philosophy is in an important sense a ‘humanistic’ discipline – that, at its best, it does not ape the natural sciences, even while taking account of them; and that, again at its best, philosophy demands a certain historical sensitivity and self-awareness – seems to me correct, even if that’s more a report of my personal preference than a systematic methodological claim. And yet, ‘Humanistic Discipline’ always leaves me with a feeling of nagging disquiet, as if Williams has got away with something he shouldn’t have, even though his intentions were no doubt sincere in so doing. Perhaps the best way of putting it is that I find myself resisting, to borrow a term from psychoanalysis, Williams’ attempt to diagnose and liberate his readers (in this case, me) from a certain kind of anxiety.

Elsewhere I have called this ‘genealogical anxiety’ (Srinivasan Reference Srinivasan2019): the anxiety that the standing of our beliefs, values, concepts or practices, once their contingent origins are revealed, will somehow be discredited – rationally requiring us, at the limit, to jettison them. Genealogical anxiety is prompted by what Williams, in his final book Truth and Truthfulness, calls ‘shameful’ histories or (more broadly) genealogies. Williams takes the term ‘shameful’ from Nietzsche, who uses the phrase pudenda origo (‘shameful origin’) (see Nietzsche 1997 [Reference Nietzsche, Clark and Leiter1881]: §§42, 102, and Nietzsche Reference Nietzsche, Colli and Montinari1980a: 2[189]), and whose Genealogy of Morals is the paradigm ‘shameful history’ or (as I prefer) ‘critical genealogy’.Footnote ¹ Nietzsche’s Genealogy presents a story of the origins and development of modern bourgeois morality that purports to reveal it as the result not of divine revelation, rational insight or innate moral goodness, but of a complex and bloody interplay of contingent historical forces. It is a matter of substantial exegetical debate what precisely Nietzsche takes himself to be doing in offering this genealogy. Is he simply attempting to exercise a certain persuasive power over his reader, loosening the psychic grip of conventional morality in order to clear the way for his own favoured worldview? Or does Nietzsche take himself to be intervening in ‘the space of reasons’, providing his reader with a rationally compelling argument, the upshot of which is that they must abandon their moral beliefs and values? If so, how is this argument supposed to work – and does it work? These questions can be posed about critical genealogies or shameful histories more generally. Can they make a difference in the space of reasons? And if so, which beliefs, concepts, values and practices are impugned by their genealogies, and why?

One central task of ‘Philosophy as a Humanistic Discipline’ is to give us reassuring answers to these questions – or, as Williams puts it, to steady us in those moments ‘in which historical understanding’ seems to ‘get in the way’ of the ‘philosophical enterprise’ (Williams Reference Williams2000: 489). In such moments, history alienates us from our ways of thinking and living, prompting us to see them as mere artefacts with which we are coincidentally associated, rather than as the proper objects of philosophical analysis. (Consider: if a concept like justice or freedom is merely a quirk of history or culture, then what would be the point in trying to explicate, as philosophers seek to do, their meaning and demands ‘from the inside’? Shouldn’t we instead treat these concepts and associated practices as would an ethnographer, ‘from the outside’?) Williams aims to diagnose and cure genealogical anxiety, without simply dismissing historical understanding as philosophically irrelevant. Put otherwise, Williams aims to show philosophers how to be genealogically aware without being genealogically anxious.

2.

Let me begin with an overview of how precisely Williams proposes to achieve this aim. Williams’ first and foundational move is to distinguish between two types of case: first, genealogical anxiety as it concerns ‘our’ (i.e. Euromodern) ‘scientific’ outlook; and second, genealogical anxiety about ‘our’ (again Euromodern, and specifically liberal) ‘ethical and political’ outlook. For Williams, there is a fundamental asymmetry between these two types of outlook, which in turn aligns with his distinction between descriptions of the world from an ‘absolute’ perspective (or the world ‘as it is anyway’), and those descriptions that are essentially ‘perspectival’ and ‘local’.Footnote ² This asymmetry, Williams claims, has important implications for how we should think about genealogical anxiety.

The problem of genealogical anxiety does not arise for our scientific beliefs, concepts, values and practices, Williams says, precisely because the history of these items is not ‘shameful’ but rather ‘vindicatory’. It is only in the case of our ethical-political outlook that historicist inquiry raises the problem of genealogical anxiety. For in this case history does not and cannot yield a ‘vindicatory’ genealogy, but only a shameful one. History reveals that our ethical judgments do not enjoy objectivity: they do not describe the world ‘as it is anyway’, but (only) as it appears from our historically- and culturally-contingent perspectives.Footnote ³ Thus I cannot claim to know that it is a grievous wrong to undertake mass killing with the intent to destroy, in whole or part, a national, ethnic, racial or religious groupFootnote ⁴ – at least not once I have recognised the historical and cultural contingency of the Euromodern outlook from which that judgment emerges.Footnote ⁵

Why not? For a current outlook to admit of historical vindication, it must be able, Williams says, to make:

I will call this Williams’ internalist condition on vindication.Footnote ⁶ There are several features of it that Williams leaves underspecified, some of which are of decisive importance when it comes to assessing Williams’ proposed solution to genealogical anxiety. Let O2 be some current outlook whose susceptibility to vindication we are assessing, and O1 be ‘the earlier outlook’ to which we are comparing it. Williams strongly implies that the ‘reason’ in question must be internal to the relevant outlooks: what matters is that O1 and O2 provide reason (presumably of the all-things-considered variety) to their respective holders to see the historical transition from O1 to O2 as an improvement. (This is why I call the condition ‘internalist’.) Now, Williams frames the internalist condition in explicitly historical terms: the relevant contrast is between some current outlook O2 and some historically prior outlook O1. But Williams, as he makes clear in ‘Humanistic Discipline’, is in fact interested in more than just the question of whether we can reasonably see our outlooks as superior to their historical rivals – interested in more, that is, than the problem of historical contingency. He is also interested in the problem of cultural contingency: whether we can reasonably see our outlook as superior to those of our contemporaries with substantially different cultural formations (Williams Reference Williams2000: 491–492). Given this, it makes sense to expand Williams’ condition to include comparisons between our current outlook and other contemporaneous rival outlooks, held by people with different cultural formations. We would then have a condition not simply on historical vindication, but more broadly on genealogical vindication, and it seems clear that this would be in keeping with Williams’ own aims.

What is far less clear is whether, for Williams, the susceptibility of O2 to genealogical vindication requires that there must have previously existed (or currently exist) some rival outlook O1 such that holders of both O1 and O2 possess (all-things-considered internal) reason to recognise the transition from O1 to O2 as an improvement; or, more strongly, that any previous (or current) rival outlook O1 is such that both the holders of O1 and O2 would have (all-things-considered internal) reason to recognise a transition from O1 to O2 as an improvement. This uncertainly is produced by Williams’ apparently inconsistent treatment of the internalist condition. Consider his application of it to science:

I take it that Williams has in mind something like the following. Not only do our current scientific theories (Einsteinian physics, say) explain away the evidence that seemed to support the older theory (e.g., Newtonian physics), they do so by showing the older theories to be a ‘special’ case of the new theories: that it is an entailment of the new theory that the old theory is (approximately) true as it concerns a certain set of familiar cases. The proponents of Newtonian physics would thus have reason, internal to their Newtonian scientific outlook, to see the shift from Newtonian to Einsteinian physics as a progressive one. In turn, Williams says, this means that our scientific outlook, or the various theories that comprise its elements (including Einsteinian physics, neo-Darwinian evolutionary theory, etc.), admits of genealogical vindication.

In so arguing, Williams implicitly invokes the weaker version of the internalist condition. After all, history is replete with examples of non-scientific – theological or otherwise supernatural – outlooks whose proponents would not have reason, internal to those outlooks, to see their outlook’s displacement by the contemporary natural scientific outlook as an ‘improvement’.Footnote ⁷ Williams does not consider these rival historical outlooks when assessing whether the scientific outlook is susceptible of genealogical vindication, choosing instead to compare current scientific theories only to their close historical ancestors: ‘later’ scientific theories to ‘earlier’ scientific theories.Footnote ⁸ This suggests that the internalist condition is to be understood weakly: what matters is that there has existed or currently exists some rival theory O1 that satisfies the relevant desiderata.Footnote ⁹

Now compare Williams’ application of the internalist condition to the case of ethical-political outlooks. Williams writes:

For Williams, history shows us that that ‘our’ ethical-political outlook, with its commitments to e.g. ‘equality and equal rights’, does not meet the internalist condition. For there are rival ethical-political outlooks – for example, the outlook of ‘the ancien régime’ – whose proponents have no ‘reason to recognize the transition’ from their outlook to ours ‘as an improvement’. While ‘we’ can tell a story about what these historical subjects were, ethically speaking and by our own lights, ‘missing’, and, so too, how it is that ‘we’ came to see what they failed to see, this is not a story that those subjects, with their different historical formations, would be under any internal rational pressure to accept.

Put another way, ‘our’ ethical-political outlook does not relate to that of the ancien régime as Einsteinian physics relates to Newtonian physics. But why think this marks a deep disanalogy between ethical-political outlooks and scientific outlooks in general? There are presumably ethical-political outlooks that do relate to each other as ancestor and heir scientific theories relate: consider, for example, the transition from a pro-imperialist liberalism to an anti-imperialist liberalism. Here we have a transition in outlook such that, Williams would presumably say, proponents of both outlooks have reason, internal to their respective outlooks, to agree that it constitutes ‘an improvement’. But clearly this fact is insufficient, on Williams’ view, to secure the vindication of anti-imperialist liberalism. What is needed is not just the existence of some rival ethical outlook O1 such that its proponents would see the transition from it to O2 as an improvement, but that all rival ethical outlooks are such that its proponents would see this transition as an improvement. Thus we have an endorsement of the following condition on genealogical vindication:

In order to justify what he calls, in Ethics and the Limits of Philosophy, the ‘fundamental difference … between the ethical and the scientific’ (Williams 2006 [Reference Williams1985]: 135), Williams applies the strong internalist condition to ‘our’ ethical-political outlook with full force, while weakening it in its application to science. Either that, or Williams doesn’t see (or perhaps refuses to acknowledge) that there are outlooks that rival our scientific outlook in a way that is structurally analogous to the rivalry between the ethical-political outlook of the ancien régime and our own. I will return to this issue, and its implications for Williams’ general approach to genealogical critique, below.

For now, let me proceed with Williams’ argument. We have arrived at what is for Williams a crucial point: our ethical-political outlook (indeed, any ethical-political outlook) cannot be genealogically vindicated, for it cannot meet the internalist condition, strongly construed.Footnote ¹² Any apologia one offers for one’s ethical-political outlook, in relation to a sufficiently alien rival outlook, will be question-begging, appealing to assumptions and modes of argument that the rival outlook rejects. Thus ‘while you can of course say that’ the representatives of the ancien régime ‘were wrong’ – ‘who is to stop you?’ Williams asks – ‘the content of this is likely to be pretty thin: it conveys only the message that the earlier outlook fails by arguments the point of which is that such outlooks should fail by them.’ It is a ‘good question’, Williams concludes, ‘whether a tune as thin as this is worth whistling at all’ (Reference Williams2000: 488).

In other words: the claim that ‘our’ ethical-political outlook is objectively true and an object of knowledge, and its rivals objectively false and objects of ignorance, is not one we are rationally entitled to make.Footnote ¹³ We have in Williams an endorsement of what I will call genealogical scepticism.Footnote ¹⁴ In endorsing such scepticism about our ethical-political outlooks, Williams shares common cause with Richard Rorty, one of Williams’ primary interlocutors in ‘Humanistic Discipline’. For Rorty, the radical contingency of all our outlooks means that we cannot claim for them the status of objective truth or knowledge. This is an especial problem, for Rorty, as it concerns our ethical-political outlooks. For how are we to go on together, in collective social life, if we take ourselves to not know about good, value, virtue, duty, right, and so on? Rorty’s proposed answer to this predicament is found in his notion of ‘irony’ (Rorty Reference Rorty1989). Williams describes the Rortyan solution as follows: ‘that qua political actors we are involved in the [liberal] outlook, but qua reflective people (for instance, as philosophers) we stand back and in a detached and rather quizzical spirit see ourselves as happening to have that attachment’ (Williams Reference Williams2000: 490). For Rorty, to take up an ironic stance towards a given ethical-political outlook is to evince both a first-order commitment to it, and a higher-order alienation from it. Through irony we are able to register the rationality of genealogical anxiety, while at the same time avoiding, in the practical world we share with others, nihilism. As Williams puts it, for Rorty the ‘ironist commits himself to things while knowing that that is all he is doing; he believes in things while knowing, in a sense, that there is nothing to believe in’ (Williams Reference Williams1989).

Williams does not at all like Rorty’s recommendation of irony. For Williams, it rests on a deep confusion about our relationship to ‘our’ ethical-political outlook. Williams and Rorty agree that there can be no genealogical vindication of this outlook; here they are both genealogical sceptics. But this, Williams insists, should not trouble us, making irony a needless solution to a non-problem. He writes:

Genealogical scepticism is not really a problem. Our feeling that it is a problem rests on a mistaken understanding of what we should want from our ethical-political outlook: namely, that it (like the natural scientific outlook) can be vindicated from ‘an absolute point of view’, one that does not itself carry the traces of our contingent history.

But we should not want this, Williams thinks, and so should not be disappointed we cannot have it. This is because we are not, in fact, ‘unencumbered intelligences selecting in principle among all possible outlooks’. Rather, we are ethically and politically encumbered beings, who find ourselves always already operating within some outlook, indeed constitutively entangled with it. To sometimes adopt, as Rorty suggests we do, a posture of dispassionate alienation in relation to our outlook is not only, Williams suggests, psychologically unfeasible but, more importantly, a mistake of ontology or philosophical anthropology, about the kinds of creatures we are. For ‘We and our outlook are not simply in the same place at the same time’, but rather formed by the same contingent history.

Indeed for Williams, the real question is not whether our ethical-political outlook can be genealogically vindicated (it cannot), but whether it can, even so, be rightfully seen as ours. If our ethical-political outlook is ‘a contingent development that happens to obtain here and now, can we fully identify with it? Is it really ours except in the sense that we and it happen to be in the same place at the same time?’ And to this question Williams answers forcefully: yes. History – and genealogy more generally – shows that our ethical-political outlook cannot aspire to objective truth or knowledge. But this need not and should not disrupt our practical relationship to our ethical-political outlook. For our practical relationship to our ethical-political outlook turns on a claim (one vindicated by history) about its constitutive relationship to ourselves, and not a claim (one impugned by history) about its epistemic merits. We and our outlook are co-constituted; and that is enough.Footnote ¹⁵

3.

Whence my resistance to Williams’ proposed diagnosis and cure? Williams would presumably say that it is due to my lack of understanding. After all, if I ‘really’ and ‘deeply’ understood his point I would be ‘free’ from genealogical anxiety. That such anxiety remains for me a problem implies, by Williams’ logic, that I have failed to adequately understand his argument.Footnote ¹⁶ But, having done all I think I can on that front, I am left to anatomise, as best as possible, my misgivings.

I begin with history. ‘To some extent’, Williams says, the problem of genealogical anxiety is

Williams locates the ‘deep roots’ of genealogical anxiety (at least in the tradition of ‘European thought’) in the ‘early nineteenth century’, and identifies it with a particularly modern malaise: ‘historicist weariness and alienation’. But, as Nietzsche knew – and so presumably Williams knew as well – genealogical anxiety appeared on the scene, within the context of ‘European thought’, well before the early nineteenth century.Footnote ¹⁷ We see it, for example, in Greek thought from at least the sixth century B.C.E., in the writings of the Ionian physiologoi, whose experiments with genealogical thinking ushered in, by the time of Plato, critical genealogies of religious belief, birth and death, common-sense ontology, social hierarchy, morality and sensible qualities.Footnote ¹⁸

So too do we see critical genealogy in European thinkers of the seventeenth and eighteenth centuries, including Levi ben Gershon (Gersonides),Footnote ¹⁹ Thomas Hobbes (2012 [Reference Hobbes and Malcolm1651]: Ch. 12), Baruch Spinoza,Footnote ²⁰ Gottfried Wilhelm Leibniz (1985 [Reference Lebniz1710]: ‘Preface’ & §§136–138),Footnote ²¹ John Toland (2013 [Reference Toland and Leask1704]: Letter 2),Footnote ²² Charles Louis de Secondat, Baron de Montesquieu (2004 [Reference de Secondat and de Montesquieu1721]: Letter 59, 233–234) and David Hume (2007 [Reference Hume and Beauchamp1757]). There are also, unsurprisingly, examples of critical genealogical thinking in non-European thinkers before the nineteenth century, including Gautama Buddha,Footnote ²³ Abu Hamid al-Ghazâlî (Reference Al-Ghazali2000: §6) and Tominaga Nakamoto (Reference Nakamoto and Chan2023). To point this out is not merely to advance a historical quibble. In downplaying the antiquity, endurance and cross-cultural ubiquity of genealogical anxiety – in making it into a mere symptom of late Euromodernity – Williams makes it easier (psychologically, if not logically) to dismiss it. Indeed, I think this is an object lesson in why philosophers should attend more carefully and deeply to history: the central point of ‘Humanistic Discipline’.

What is more, Williams’ reference to ‘historicist weariness and alienation’ obscures the way in which critical genealogies have so often been put forward in a spirit not of weariness or alienation, but of revolutionary and emancipatory zeal. This is evident already in the ancient Buddhist and Greek uses of critical genealogy; and that same spirit of ‘creative destruction’ can be found in many other critical genealogists, including Hume, Hobbes and of course Nietzsche himself; as well as in many twentieth and twenty-first century practitioners of critical genealogy, including the Frankfurt School Critical Theorists, Michel Foucault, feminist genealogisers of sex and gender like Simone de Beauvoir and Judith Butler, debunkers of Eurocentrism like Edward Said and Dipesh Chakrabarty, critics of liberalism like Charles Mills and Carole Pateman, and politically-oriented intellectual historians like Quentin Skinner.

Taken together, these two historical points suggest that genealogical anxiety has a stronger pull than Williams suggests: that it is not the sole property of ‘modernity’ (except in the trivial, and so not especially enlightening, sense in which any culture that entertains critical genealogies is eo ipso ‘modern’), and that it does not necessarily bespeak ‘weariness’ or ‘alienation’, but instead can be bound up with a sense of freedom, possibility and novelty. If our task is (to use a Williamsian phrase) ‘to make sense’ of genealogical anxiety as part of the common human estate, then we would do well to keep these points in mind. Otherwise, we make the problem that such anxiety poses too easy for ourselves.

My more substantively philosophical worry, as I have already intimated above, is that Williams appears to me to fail – just as he accuses Rorty of failing – to go ‘far enough in recognizing contingency’. In insisting that our scientific outlook can be genealogically vindicated, but that our ethical-political outlook cannot, Williams does not see, or refuses to reckon with, the full power of genealogical anxiety and the problem of contingency that provokes it. For what Williams says of our ethical-political outlook is plausibly true of our scientific outlook as well. ‘We’ have historical counterparts whose worldviews were such that they would have no reason, internal to their outlook, to think that the shift to the scientific worldview of Euromodernity is an instance of progress. (Perhaps we don’t even need to go back in history to find such counterparts: consider contemporary creationists and climate change deniers.) In failing or refusing to see this, Williams is guilty, ironically, of a kind of scientism about science, an overly credulous relationship to the self-conception of the natural sciences as a form of inquiry free from any historically- or culturally-conditioned perspective.

To make my case, I want now to turn to Wittgenstein, another of Williams’ primary interlocutors in ‘Humanistic Discipline’. Williams calls Wittgenstein, with justification, ‘the least scientistic of philosophers’ (Williams Reference Williams2000: 493). Part of Wittgenstein’s anti-scientism, I want to suggest, consists in his willingness to see how the scientific outlook might, as much as any other type of outlook, be an object of genealogical anxiety. As a result, Wittgenstein takes seriously the very reply to genealogical anxiety that Williams was quick to dismiss as a ‘thin tune’ not ‘worth whistling’. If Wittgenstein is right to think of scientific and ethical outlooks as equally susceptible of genealogical critique, and Williams wrong to insist otherwise, then it will turn out that this ‘thin tune’ is the only one, for better or worse, we have available to whistle.

4.

In notes written between 1946 and 1949, Wittgenstein observes:

Here Wittgenstein is canvassing a genealogical scepticism of a very general kind. Of any concept of which someone is convinced that it is ‘absolutely correct’, and that having instead a rival concept would mean a loss of knowledge (‘not realizing something that we realize’), he should simply imagine his own natural history having been different to what it in fact is, ‘and the formation of concepts different from the usual ones will become intelligible to him’. And, with that, Wittgenstein implies, the possibility that he has a wrong concept – one that does not limn reality, that does not deliver knowledge – will likewise become intelligible to him.

And yet Wittgenstein immediately continues:

Choosing a style in which to paint might appear to us to be a matter of voluntary and arbitrary choice: there are various painterly possibilities arrayed before us, and the skilled painter can select among them. But this appearance, Wittgenstein here suggests, is mistaken. For ‘our style of painting’ is grounded in ‘pretty and ugly’ – that is, our aesthetic sensibility, what we find pretty and ugly. The painter’s apparent ‘choice’ to paint in ‘our’ way rather than all the other possible ways is not in fact arbitrary; it reflects how ‘we’ see the world. In fact, this too is a matter of contingency: the Egyptians found the ‘Egyptian’ style of painting pretty, whereas ‘we’ (so Wittgenstein says) do not. Wittgenstein’s point is that even here – when it is most prima facie compelling – the appearance of ‘arbitrary choice’ is misleading. The apparent counterfactual possibilities are not in fact possibilities for us, because they would not conform to our form of life. And if this is true of aesthetic styles, then so it must be a fortiori true of conceptual ‘choice’.

In short, Wittgenstein provokes genealogical anxiety about our concepts only to liberate us from it, showing how it rests on a mistaken picture of the relationship between ourselves and our concepts. In so doing, he anticipates Williams’ rejoinder to Rorty in ‘Humanistic Discipline’. For Williams, as we have seen, it is ‘history’ that allows us to see the true relationship between ourselves and our ‘outlooks’, namely a relationship of co-constitution. Whereas for Wittgenstein, release from genealogical anxiety is found not in historical self-consciousness, but in greater attention to our practices as we now find them; when we stop ‘thinking’ and start ‘looking’ (Wittgenstein 2009b: §66), we see that the ‘possibilities’ that prompt genealogical anxiety are not, for us, really possibilities at all. Nonetheless, in both Williams and Wittgenstein we find a therapeutic approach to genealogical anxiety, one that diagnoses it as resulting from a false picture of the relationship between ourselves and our ways of seeing the world.

But Wittgenstein does not stop here – that is, where Williams stops. He may agree with Williams that, as a practical matter, we are unable to simply ‘choose’ to use concepts that are alien to our own. And certainly one of Wittgenstein’s abiding purposes in his later work was to draw our attention to this (putative) feature of human psychology. (Thus Williams is right to treat Wittgenstein as a key source of the idea that ‘there [is] an end to justifications, that at various points we run into the fact that “this is the way we go on”’ – Williams Reference Williams2000: 493.) But, as an epistemological matter, Wittgenstein appears to reject what we might call Williams’ ‘divide and rule’ strategy, according to which our ethical-political concepts are susceptible to the debunking force of genealogical critique, but our scientific concepts are not. We see this already in the notes quoted above, in which Wittgenstein speaks of ‘concepts’ generally, not specifically ethical-political concepts. And we see this elaborated in the notes that Wittgenstein wrote in the final months of his life, published later as On Certainty.

On Certainty is an extended engagement with G.E. Moore’s ‘Proof of an External World’ and ‘A Defence of Common Sense’, in which Moore, as a retort to the idealist,Footnote ²⁵ had claimed to ‘know’ a number of ‘truisms’, including that he has hands, that his body has never been far from the surface of the earth, and that the earth has existed for a long time before his body’s birth. For Wittgenstein, Moore’s claim to know his various ‘truisms’ is impotent as a refutation of idealism. But, he says, Moore is nonetheless right to pick out these truisms as playing a special role in our cognitive lives – as having ‘a peculiar logical role in the system of our empirical propositions’ (Wittgenstein Reference Wittgenstein, Anscombe and von Wright1969: §136 – further section citations are to this work). Moore’s truisms, Wittgenstein says, constitute ‘certain types of case in which I rightly say I cannot be making a mistake’ (§674). We cannot be mistaken about Moorean truisms not because we have some ‘proof’ of their truth (§§204, 295), or because they are ‘intrinsically obvious or convincing’ (§144), but rather because they do not (despite their surface appearance) function for us as empirical hypotheses, subject to counterevidence and thus open to doubt (§§167, 401, 402).

If not empirical hypotheses, what are these truisms about which we ‘rightly say we cannot be mistaken’? They are, Wittgenstein says, akin to mathematical and logical propositions (§§401, 447, 448, 651–653). They, and ‘not only propositions of logic’, form ‘the foundation of all operating with thoughts’ (§401). They are parts of our ‘frame of reference’ (§83), ‘world-picture’ (Weltbild) (§167), ‘the ‘scaffolding of our thoughts’ (§211), ‘the system’ in which ‘arguments have their life’ (§105). They are like the ‘axis around which a body rotates’ (§152) or the ‘hinges’ on which a door turns (§§341, 343). To attempt to subject our truisms to doubt would require us to abandon wholesale our most basic understanding of the world – as a shared world of stable objects, multiple selves and natural regularities which we access through our senses and capacity for scientific rationality – without which we cannot engage in the very practices of doubting, inquiring, inferring and judging that we are trying to undertake. Just as the artist cannot simply ‘choose’ to paint in the ‘Egyptian style’, it is not practically open to us to reject our form of life with its constitutive truisms. This includes those (apparently empirical) truisms that are elements of what Williams calls our ‘scientific outlook’.

Thus what Williams says of our ethical-political outlook, Wittgenstein says also of our scientific outlook. And he is motivated to say so precisely because he thinks, pace Williams, that our scientific outlook, just like our ethical-political outlook, is susceptible of genealogical critique. On Certainty presents several cases in which Wittgenstein imagines himself having an (apparent) disagreement over a truism with some other person or community. In many of these cases, Wittgenstein’s response is to insist that they are not cases of actual disagreement. Thus if ‘Moore were to pronounce the opposite of those propositions which he declares certain, we should not just not share his opinion: we should regard him as demented’ (Wittgenstein Reference Wittgenstein, Anscombe and von Wright1969: §155). The apparent ‘disagreement’ is not indicative of the possibility of a mistake, but just the possibility of losing one’s grip on reality.

But On Certainty also includes cases in which Wittgenstein encounters humans from cultures remote from his own; humans who as a result of their distinctive genealogical formations have world-pictures that are profoundly alien to his. In such cases, Wittgenstein does not simply dismiss his imagined interlocutors as madmen. Imagining a ‘primitive’ culture that consults an oracle rather than the laws of physics, Wittgenstein writes:

Similarly, Wittgenstein imagines Moore encountering a king who had been ‘brought up in the belief that the world began with him’:

These passages and others like themFootnote ²⁶ suggest that, for Wittgenstein, different historical and cultural genealogies can produce (genuinely) conflicting world-pictures.Footnote ²⁷ What’s more, any attempt to compare the epistemic merits of these conflicting world-pictures – as more or less true, correct, accurate, rational, evidentially-supported, coherent, consistent, etc. – will itself have to appeal to epistemic standards that are world-picture-relative.

In other words, ‘our’ truisms are elements of a world-picture – what Williams calls the ‘scientific outlook’ – that do not meet the internalist condition on genealogical vindication.Footnote ²⁸ Just as the ‘representatives of the ancien régime’ are under no internal rational pressure to regard our ethical-political outlook superior, the king is under no internal rational pressure to regard Moore’s scientific outlook as superior to his own. Insofar as one is going to abandon one world-picture and take up another – move, say, from a world of magic and oracles to a world of physics – one will do so not because something internal to one’s former world-picture rationally recommends adoption of the latter, but because one is ‘persuaded’ to see the world differently. Thus traveling from one world-picture to another can be akin to ‘conversion’: a matter of leaping from one way of seeing things to another.

Many readers of On Certainty take these and similar passages to show that Wittgenstein endorsed ‘relativism’ or (in my terms) genealogical scepticism, i.e. the view that there is no rational comparison to be made between rival world-pictures, and that we are unable to know of any world-picture that it is objectively epistemically superior.Footnote ²⁹ But is Wittgenstein really a genealogical sceptic? His relativistic readers assume that insofar as Wittgenstein isn’t a genealogical sceptic, it must be because he thinks that there exist universal standards of rational assessment that are shared by all possible world-pictures, regardless of their historical or cultural particularities. In other words, they assume that a vindicatory genealogy of a given world-picture must show it to be superior in a way that would satisfy the holders of a rival world-picture. This is precisely Williams’ internalist condition on genealogical vindication. Indeed this assumption – that a vindicatory genealogy must meet Williams’ internalist condition, or at least that Wittgenstein thought so – is shared also by Wittgenstein’s anti-relativistic readers. Thus they argue that, despite appearances, Wittgenstein conceded the possibility of only a single, universal world-picture,Footnote ³⁰ or that he posited world-picture-neutral epistemic principles.Footnote ³¹

But why agree to the internalist condition at all? Why not think that we can know that our world-picture is the objectively correct (or more correct) one, if indeed it is; and that we can rationally conclude this, even though such reasoning would essentially rely on elements of our world-picture?

This question becomes pressing once we reject Williams’ ‘divide and rule’ strategy, and the scientism on which, I’ve suggested, it rests. Once we ‘go far enough in recognizing contingency’, we see, as Wittgenstein saw, that our ‘scientific’ outlook is in the same boat as our ethical-political outlook. Insofar as the latter does not meet the conditions for genealogical vindication, the former doesn’t either. Genealogical contingency then would imply a genealogical scepticism of a very broad kind: indeed, precisely the kind of broad scepticism that Wittgenstein’s relativistic interpreters ascribe to On Certainty, and the kind of broad scepticism that Williams wants to resist (Williams 2006 [Reference Williams1985]: Ch. 8).

But to read Wittgenstein as a simple genealogical sceptic or relativist is to read him with insufficient care. Imagining a community who believe that humans can and do travel to the moon,Footnote ³² Wittgenstein (Reference Wittgenstein, Anscombe and von Wright1969: §286) writes: ‘We say: these people do not know a lot that we know. And, let them be never so sure of their belief–they are wrong and we know it’. Note that what Wittgenstein is not saying is that between ‘us’ and the moon-travellers there is some shared epistemic standard according to which we can all agree as to who is right and who is wrong. Rather, even in the absence of such a shared standard, he says, ‘we’ are able to ‘know’ that it is our world-picture that is correct, and theirs mistaken.

We, Wittgenstein implies, reason as follows: ‘Unlike the moon-travellers, we have had the good fortune of being enculturated in a scientific rather than magical worldview, and it is in virtue of this good fortune – this good genealogical luck – that we are able to get onto the truth. Our scientific world-picture explains why their magical world-picture is false. And so the admitted fact that we would have believed in the possibility of moon-travel had we born in their culture and not our own does nothing to undermine our ability to know that we are right, and they are wrong.’

Of course, the moon-travellers can tell an analogous story, using their magical world-picture to explain how it is that they came to have a correct world-picture, and us a faulty one. But the difference is – let us assume, for do we doubt it? – that ‘we’ really are genealogically lucky with respect to our moon-beliefs, and the moon-travellers are not. And this difference, Wittgenstein appears to suggest, makes all the epistemic difference.

The difference in question is, crucially, an ‘external’ one: a difference not in the internal coherence and evidential consistency of our rival world-pictures,Footnote ³³ but in how our world-pictures objectively match up with reality. (A reality – and there’s the rub – which we can only access through our rival world-pictures.) Indeed the underlying epistemological account that Wittgenstein is invoking here is an ‘externalist’ one. According to the externalist, what matters for epistemic justification and rationality is whether our judgments are in fact reliably hooked up to the world, however it may seem to us ‘from the inside’. As Wittgenstein (Reference Wittgenstein, Anscombe and von Wright1969: §505) reminds us: ‘It is always by favour of Nature that one knows something.’Footnote ³⁴

5.

To see just how externalism responds to genealogical scepticism, consider what we might call evidence of truth-indifference: that is, evidence that one’s belief-forming method does not reliably track the truth. The genealogical sceptic thinks that cases involving Williams’ ancien régime or Wittgenstein’s moon-travellers provide us with evidence of truth-indifference. For, the genealogical sceptic says, such cases show us that our belief-forming method is simply believing-in-accordance-with-whatever-world-picture-our-contingent-genealogy-has-bestowed-on-us. This method is not a truth-tracking one, for – depending on the contingencies of history and culture – it could just as easily have resulted in, say, the belief that some humans are due more ethical consideration than others; or that humans visited the moon prior to 1969. In the face of such genealogical revelations, the sceptic says, ‘our’ beliefs lose whatever justification they may have previously enjoyed.Footnote ³⁵

The externalist counters the genealogical sceptic by insisting that genealogical evidence of truth-indifference is damaging for one’s epistemic justification only if that evidence is not misleading.Footnote ³⁶ Suppose that the Euromodern scientific world-picture is unique in giving ‘us’ an unprecedently reliable grasp of empirical reality. If that is so, then the belief-forming method that ‘we’ use is importantly distinct from, and epistemically superior to, the alternatives. The genealogical evidence, insofar as it suggests that the method we use is just ‘believing-in-accordance-with-a-genealogically-contingent-outlook’, is misleading. Our outlook isn’t any old outlook, and not only for the reason, as Williams puts it, that the same history ‘has both made us, and made the outlook as something that is ours’. For our outlook, the externalist defender says, has a special relationship to reality.

On the externalistic view, if this is indeed true – if ‘our’ outlook or world-picture indeed constitutes a distinctly truth-tracking basis for judging – then misleading genealogical evidence to the contrary can be rationally dismissed. ‘We’ are able to justifiably hold onto our modern scientific and ethical beliefs, even while recognising that, had we been born in a different time or place, we would have rejected them. Indeed, the externalist says, we are able to know the truth of such beliefs; and we are able to rationally dismiss alternative world-pictures as unreliable, and the judgments based in them false. And all this is true precisely because ‘our’ outlook is in fact a broadly truth-tracking basis for judging.

In saying all this, isn’t the externalist licensing a form of circular reasoning on ‘our’ part? Indeed. It is only by assuming the correctness of our world-picture or outlook that we can reason our way to the conclusion that we are genealogically lucky, and that holders of rival world-pictures or outlooks are genealogically unlucky and so ignorant (even if blamelessly so).Footnote ³⁷ Specifically, the externalist conception of epistemic rationality and knowledge permits what is known as ‘bootstrapping’: reliance on a method of belief-formation to confirm the reliability of that very method.

Bootstrapping is a form of circular reasoning, and for that reason is felt by many to be epistemically illicit. Imagine using the deliverances of your watch to confirm that your watch is in fact keeping time: ‘My watch says it is 4pm, and indeed – for so says my watch! – it is 4pm…’ and so on. Surely, we want to say, this isn’t a good instance of reasoning; isn’t a way of coming to know that one’s watch is keeping time. Indeed, to engage in bootstrapping is, for Williams, to whistle a tune so thin as not to be worth whistling. The watch example gives succour to this indictment. You can go around saying that you know your watch is keeping time by checking it against itself – ‘who is’ as Williams says ‘to stop you?’ – but if this is all that is demanded of reasoning and knowing, then reasoning and knowing may well turn out to be less valuable achievements than we previously thought. Indeed, we might come to think they are barely achievements at all.

But two points need to be made about the externalist licensing of bootstrapping in the face of genealogical scepticism. First, it is only when one’s outlook or world-picture is in fact a reliable basis for judging that one is, according to the externalist, permitted to reason in this way. The moon-travellers who ‘reason’ to the conclusion that they are the genealogically lucky ones have done something epistemically illicit; for they are using an unreliable world-picture to ‘confirm’ its supposed reliability. The same could be said of the representatives of the ancien régime who reason their way to the conclusion that we, with our modern superstitions about equality, are the genealogically unlucky ones. Externalism thus doesn’t sanction willy-nilly bootstrapping. It only permits bootstrapping that will lead to knowledge.

Second, there is a difference between what we might call reflexive and non-reflexive bootstrapping. In reflexive bootstrapping, one uses one’s outlook to explain away the misleading evidence of truth-indifference. For example, in defending the scientific outlook, ‘we’ might draw on the outlook itself to explain its own historically and culturally contingent development: why it is that humans progressively if unevenly abandoned the supernatural in favour of naturalistic explanation. A major proponent of such reflexive vindication of the scientific outlook was Wittgenstein’s contemporary and Vienna Circle member Otto Neurath. Neurath insisted that the sciences broadly understood – especially sociology, psychology and (Marxist) history – could elucidate the contingent social and material forces that drove scientific development. Rather than undermine the epistemic credentials of the natural sciences, such scientific explanation of science would reflexively vindicate them. (To this project of reflexive vindication Neurath gave the term ‘unified science’.) Similarly, Neurath (a committed socialist) conceded that Marxism was an ideology, but insisted that it was an ideology that – unlike bourgeois ideology – could be consistently applied to itself, without having ‘to rely on metaphysics or other ultimately unscientific procedures to ensure theoretical cohesion’.Footnote ³⁸ For this reason, Neurath said, Marxism was both an ideology and a science.

Contrast reflexive bootstrapping with a case of non-reflexive bootstrapping. Suppose that the scientific outlook had simply no explanation for why it is that so many people, at other times and places, don’t share it. One could still bootstrap one’s way to the conclusion that the scientific outlook is reliable, for example by thinking: ‘Our outlook says that humans didn’t travel to the moon in the 1950s. And indeed humans weren’t traveling to the moon then. So our outlook must be reliable (and so on)’. I take it that this would be a far less satisfying reply to the genealogical sceptic, no better than my confirming the reliability of my watch by using my watch – a thin tune indeed.Footnote ³⁹ If so, this suggests that it is not bootstrapping per se that is impermissible, but rather only bootstrapping that is non-reflexive.Footnote ⁴⁰

Granted, the ‘reflexivity’ test does not rule out non-scientific outlooks. Many religious and supernatural outlooks are characterised by a form of reflexivity. For example, we can imagine Wittgenstein’s moon-travellers performing certain rituals which, on their view, give them a special receptivity to evidence of moon-travel, which in turn explains why ‘we’ do not know what they are able to know. Less counterfactually, we can think about those Christian apologists who defuse the epistemic threat presented by the cultural contingency of religious belief by arguing that the Christian (unlike, say, the Muslim) forms her religious beliefs via a distinct and superior method – the Internal Testimony of the Holy Spirit, say – and thus is able to know the Christian truths.Footnote ⁴¹ In both these cases, the relevant outlook is recruited to explain why it is taken up by some but not others, in a way that is consistent with that outlook being known to those who are lucky enough to believe it.

In any case, it is not my aim here to defend bootstrapping, or an externalism that licenses it, as a response to genealogical scepticism. Rather, my aim is to show how such a response functions: how it at once concedes Williams’ assumption that between rival outlooks there can be no neutral assessment of their epistemic merits, while also insisting that this does not preclude the possibility of knowledgable belief on the part of the holders of a correct outlook. Williams, as we have seen, dismisses the externalist response out of hand. (The best that can be said of it is that there is no one ‘to stop you’ from making it.) But he feels free to do so precisely because he does not take seriously the structural similarities between scientific and ethical-political outlooks, falsely assuming that the former can satisfy his internalist condition for genealogical vindication. It is little surprise that Wittgenstein, who saw our scientific judgments as on all-fours with our other judgments, took the externalist response more seriously.Footnote ⁴² If the externalist response to genealogical scepticism is a thin tune, it is one we should grow accustomed to whistling. For it may be the only tune we have.

6.

I want to close with some brief comments on Williams’ understanding of the place of reasoning, argument and persuasion in ethical change, again by placing him in conversation with Wittgenstein.Footnote ⁴³ For Williams, vindication of a favoured outlook must appeal only to reasons that are (also) ‘internal’ to any rival outlook. Otherwise, a transition from a rival outlook to the favoured outlook would not count as a rational transition. As Williams puts it, the triumphant outlook might be said to have ‘won’, but not to have ‘won an argument’. This, for Williams, is precisely what distinguishes the transition between, say, Newtonian and Einstein physics from the transition between pre-modern and modern ethical-political outlooks. The first transition, between scientific theories, is correctly seen as one involving rational persuasion, while the latter transition, between ethical-political outlooks, cannot be so seen.

What then, for Williams, does a transition between ethical-political outlooks, of the kind seen in the shift from the outlook of the ancien régime to that of Euromodernity, involve? It presumably involves what Wittgenstein calls ‘conversion’. In conversion, as opposed to ‘argument’, what is appealed to is not one’s current beliefs and commitments, but rather one’s capacity for wholesale perspectival transformation, one’s susceptibility to being brought to see the world anew (and, from the convertor’s perspective, aright). Conversion involves what Williams calls a ‘crisis of legitimacy’ rather than (as in the Newtonian-Einsteinian example) a ‘crisis of explanation’ (Williams Reference Williams2000: 488). A crisis of explanation calls for the progressive, rational development of a new outlook out of an old one. A crisis of legitimacy calls for the wholesale abandonment of one outlook for the sake of a rival – in short, a conversion.

One might well wonder how to square Wittgenstein’s comments on conversion, quoted above, with his apparent openness to externalism as a response to genealogical scepticism. In suggesting that conversion is the only mechanism by which one abandons one world-picture in favour of another, isn’t Wittgenstein admitting that even a correct world-picture cannot be rationally acquired or held? Isn’t he showing himself, in other words, to be like Williams an internalist about genealogical vindication?

Not necessarily. As Alice Crary observes, Wittgenstein’s remarks about conversion ‘only … express [a relativistic] outlook given the assumption that the persuasive methods Wittgenstein mentions … are properly understood as non-rational modes of communication’ (Crary Reference Crary, Moyal-Sharrock and Brenner2005: 284). Implicit in the relativistic reading of Wittgenstein’s comments about conversion is what Crary calls a ‘narrow’ conception of rationality, according to which a person’s being rational cannot essentially depend ‘on her possession of any particular subjective endowments’, including attitudes, interests, affects and values (Crary Reference Crary, Moyal-Sharrock and Brenner2005: 290). On the ‘narrow’ conception, rationality is equally available to all minimally competent subjects, regardless of the genealogical contingencies of their formation; or, as Williams (2006 [Reference Williams1985]: 155) puts it, to ‘any adequate investigator, of whatever constitution’. Thus if having a true judgment about something requires not simply ‘reasoning’ one’s way to that judgment – in the narrow sense of taking a series of deductive and/or inductive steps on the basis of the subjectively available evidence – but moreover being (as Wittgenstein says) ‘brought to look at the world in a different way’, then a fortiori that judgment cannot be ‘rationally’ acquired or held.

Indeed it is this narrow conception of rationality that Williams implicitly endorses in putting forward the internalist condition on vindication. For Williams, any form of persuasion that does not narrowly appeal to beliefs or commitments that one’s interlocutor already possesses is ipso facto an attempt at non-rational argument.Footnote ⁴⁴ To Williams’ credit, he does not insist – as do, for example, many contemporary Critical Theorists, with their particular rendering of ‘immanent critique’ – that actual, historically-realised ethical-political change can be seen as a progressive development of the internal logic of past human societies. In my view, Williams is right to see much of human history as involving not just crises of explanation but moreover crises of legitimation. As involving, that is, not just the working out of internal logics, but also ideological ruptures, conceptual revolutions, and the wholesale adoption of new ways of seeing and being. Nonetheless, why think that the changes prompted by such crises of legitimation could not themselves be rational? Why, in other words, have a narrow, internalistic understanding of rationality? Why not think that being rational – and, indeed, ‘winning’ an ethical-political argument – can involve more than following minimal, outlook-transcendent procedures of good inference and calculation? Can involve, that is, possessing or acquiring the contingently-granted endowments that allow one to see the world aright?Footnote ⁴⁵

Take a concrete case. Seeing Palestinians as persons worthy of ethical concern rather than extremist threats to be eradicated sometimes requires, plausibly, more than just narrow rationality. The problem with many defenders of Israel’s genocide in Gaza is not simply that they are making bad inferences or being inconsistent or are neglecting the evidence. (Consider the IDF soldiers who enact the genocide on the ground; if anything they have more ‘evidence’ of Palestinian humanity than non-Palestinians living in the West do.) What at least some of these génocidaires lack are the right affects, the right forms of ethical attention and sensitivity, the right ways of seeing. Their genocidal world-picture exhibits an impressive and terrifying coherence; and in its weaponised deployment of ‘antisemitism’ it has a powerful pseudo-explanation for why so many of us see things in a contrary light. The holders of the genocidal world-picture can, as Wittgenstein says, be ‘converted’ to a different view, can ‘be brought to look at the world in a different way’. (Indeed, they might be able to undertake a process, like many religious believers before them, of self-conversion, through deep affective engagement with Palestinian humanity.) On the wide conception of rationality, such a conversion could be an instance of rationality, not a departure from it. And it is precisely to such a wide conception of rationality that the externalist rejection of genealogical scepticism appeals.Footnote ⁴⁶

In resisting a wide, externalist conception of rationality, Williams is again guilty of a kind of scientism about science. As Kuhn observed in The Structure of Scientific Revolutions (Reference Kuhn1962), scientific rationality itself must be understood widely if it is to exist at all. Kuhn’s central thesis in Structure is that the shared values that are constitutive of the scientific enterprise – empirical adequacy, consistency, simplicity, elegance, plausibility, fecundity, etc. – do not uniquely determine, in any particular case, theory-choice; and that this fact was essential in explaining transitions from ‘normal’ to ‘revolutionary’ science. Thus: ‘though values are widely shared by scientists and though commitment to them is both deep and constitutive of science, the application of values is sometimes considerably affected by the features of individual personality and biography that differentiate the members of the group’ (Kuhn, 1996 [Reference Kuhn1962]: 185, emphasis added). In other words, the shift from an old scientific paradigm to a new one cannot always be shown to be progressive by the lights of the proponents of the old paradigm. What is sometimes further required is that the proponents of the old paradigm have the right ‘features of individual personality and biography’, allowing them to embrace the new paradigm – to see the world aright. If, as Williams seems to think, rationality is a wholly internalistic business, then it is not only ethical-political change that must involve arational conversion. Science must too.Footnote ⁴⁷

There is much more to say here, especially about just why it is that Williams is so keen to idealise science as an uncomplicatedly objective, rational and progressive enterprise, and to deny any analogies between it and the ethical-political enterprise of figuring out how best to collectively live. One clue lies in Williams’ assertion that, when it comes to ethics, externalistic ‘beliefs or modes of argument may turn back to those old devices of cognitive reassurance such as “intuition”’ (Williams Reference Williams2000: 492). Evidently that prospect is, for Williams, a regressive one. And for good reason: claims to special ethical perception are all too easily used in programmes of domination and oppression, and perhaps are metaphysically suspect to boot. But one does not have to posit a magical faculty of ethical-political ‘intuition’, any more than one needs to posit a magical faculty of ‘scientific insight’. Just as scientific knowledge is the result of a suite of familiar capacities, including sensory perception, observation, hypothesis-generation, deductive and inductive inference, and so on, ethical-political knowledge might be a matter of familiar (and widely if not evenly shared) capacities like curiosity, attentiveness, empathy, constancy, judgment, courage and imaginativeness.Footnote ⁴⁸

However – and this is my ultimate point – ethical knowledge, whatever else it is, is surely also a matter of matters not ‘internal’ to us; a matter of factors not wholly within our control, of the distribution of good luck. I mean specifically the good genealogical luck of being brought up in a place and time, and inhabiting a social position and material conditions, that endow one with a decent enough ethical-political outlook, as well as the capacities required to improve it; and then of having the right sorts of experiences and encounters (up to and including ‘conversions’) that allow one to improve one’s outlook, to come to see others and our relations to them as they really are.Footnote ⁴⁹

Indeed this was, in an important sense, Williams’ ultimate point, too. What Williams pejoratively called the ‘morality system’ is built on the idea that we can identify a ‘moral value’ that ‘transcends luck’, a value that lies ‘not only in trying rather than succeeding, since success depends partly on luck, but in a kind of trying that lies beyond the level at which the capacity to try can itself be a matter of luck’. Williams is right to dismiss ‘moral value’, so understood, as a consoling but damaging ‘illusion’ (Williams 2006 [Reference Williams1985]: 217–218). But here too he does not go far enough. Owning up to our human lot involves acknowledging the outsized role that luck plays in all our affairs, including our aspirations to ethical knowledge. Indeed this is one of the things we learn when we attend, as Williams would have us do, more carefully to history. In turn, acknowledging the place of luck in human affairs means resisting the impulse to try to eradiate it by philosophical fiat. It is this impulse that drives Williams’ repudiation of ethical objectivity. By debasing ethical knowledge, he hopes to secure it against the machinations of genealogical contingency. But in doing so Williams arguably betrays his own understanding of the human situation. And he makes ethical knowledge a thin tune indeed.