The Contingency of Community

Richard Rorty

If one says, as I did in ‘The Contingency of Language’, that truth is not ‘out there’, one will be suspected of relativism and irrationalism. If one suggests, as I then did in ‘The Contingency of Selfhood’, that we no longer need a distinction between morality and prudence, one may seem to be encouraging immorality.[*] By way of defence, I shall argue here that these distinctions between absolutism and relativism, rationality and irrationality, morality and expediency, are obsolete and clumsy tools – remnants of a vocabulary which we should try to replace. But, as I suggested earlier, ‘argument’ is not the right word. For on my account of intellectual progress as the literalisation of selected metaphors, rebutting objections to one’s redescriptions of some things will be largely a matter of redescribing other things, trying to outflank the objections by enlarging the scope of one’s favourite metaphors. So my strategy will be to try to make the vocabulary in which these objections are phrased look bad, thereby changing the subject, rather than granting the objector his choice of weapons and terrain by meeting his criticisms head-on.

I should like to propose that the institutions and culture of liberal society would be better served by a vocabulary of moral and political reflection which avoids the distinctions I have mentioned than by one which preserves them. I shall try to show that the vocabulary of Enlightenment rationalism, although it was essential to the beginnings of liberal democracy, has become an impediment to the progress of democratic societies. The vocabulary which I adumbrated in the two previous articles, one which revolves around notions of metaphor and self-creation rather than around notions of truth, rationality and moral obligation, is better suited to such societies.

I am not, however, saying that the Davidsonian-Wittgensteinian account of language, and the Nietzschean-Freudian-Bloomian account of conscience and selfhood which I have sketched, provide ‘philosophical foundations of democracy’. For the notion of a ‘philosophical foundation’ goes when the vocabulary of Enlightenment rationalism goes. But the utility of an apologetic remains. By an ‘apologetic’ I mean a way of describing old institutions and practices in a new, more useful way. To offer an apologetic for our current institutions and practices is not to offer a justification of them, nor is it to defend them against their enemies. Rather, it is to suggest ways of speaking which are better suited to them than the ways which are left over from older institutions and practices. To engage in apologetics is more like refurnishing a house than like propping it up or placing barricades around it.

The difference between a search for foundations and an attempt at apologetics is emblematic of the difference between the culture of liberalism and previous forms of cultural life. For, in its ideal form, the culture of liberalism is enlightened, secular, through and through. It is one in which no trace of divinity remains, in the form either of a divinised world or a divinised self. Such a culture, if it became actual, would have no room for the notion that there are non-human forces to which human beings should be responsible. It would drop, or drastically reinterpret, not only the idea of holiness but that of ‘devotion to truth’ and that of ‘fulfilment of the deepest needs of the spirit’. The process of de-divinisation which I described in the previous articles would, ideally, culminate in our no longer being able to make sense of the idea that finite, mortal, contingently-existing human beings might derive the meanings of their lives from anything except other finite, mortal, contingently-existing human beings.

Most of what I said in the first two articles can be seen as footnotes to Isaiah Berlin’s Two Concepts of Liberty. Berlin says there, as I did in ‘The Contingency of Language’, that we need to give up the jigsaw puzzle approach to vocabularies, practices and values: to give up, in Berlin’s words, ‘the conviction that all the positive values in which men have believed must, in the end, be compatible, and perhaps even entail each other.’ My citation of Freud’s remark that we should think of ourselves as each just one more among nature’s experiments, not as the culmination of nature’s design, echoes Berlin’s use of John Stuart Mill’s phrase ‘experiments in living’. (It also, of course, echoes Jefferson’s and Dewey’s use of the term ‘experiment’ to describe American democracy.) Like Berlin, I have been criticising the Platonic-Kantian attempt to do what Berlin called ‘splitting [our] personality into two: the transcendent, dominant controller and the empirical bundle of desires and passions to be disciplined and brought to heel’.

Berlin ended his essay by quoting Joseph Schumpeter: ‘To realise the relative validity of one’s convictions and yet stand for them unflinchingly is what distinguishes a civilised man from a barbarian.’ In the jargon I have been developing, this translates into the claim that the liberal societies of our century have fostered people who recognised the contingency of the vocabulary in which they stated their highest hopes – the contingency of their own consciences – and yet remained faithful to those consciences. I have been claiming that figures like Nietzsche, William James, Freud, Proust and Wittgenstein illustrate what I called ‘freedom as the recognition of contingency’. Such freedom, I would now claim, is integral to the idea of a liberal society.

In order to show how the charge of relativism looks against this background, I want to take up some comments on Berlin’s essay by an acute critic of the liberal tradition, Michael Sandel. Berlin ‘comes perilously close to foundering on the relativist predicament’:

If one’s convictions are only relatively valid, why stand for them unflinchingly? In a tragically configured moral universe, such as Berlin assumes, is the ideal of freedom any less subject than competing ideals to the ultimate incommensurability of values? If so, in what can its privileged status consist? And if freedom has no morally privileged status, if it is just one value among many, then what can be said for liberalism?

In posing these questions, Sandel is taking the vocabulary of Enlightenment rationalism for granted. Or, more exactly, he is taking advantage of the fact that Schumpeter and Berlin themselves make use of this vocabulary, and attempting to show that their view is incoherent. Examining the presuppositions of Sandel’s questions may help make clear what sort of view lies behind the terms ‘relativism’ and ‘morally privileged’. It may thus help show why the term ‘only relatively valid’ is a bad, rationalistic way to characterise the state of mind of the figures whom Schumpeter, Berlin and I wish to praise.

To say that convictions are only ‘relatively valid’ might seem to mean that they can only be justified to people who hold certain other beliefs – not to anyone and everyone. But if this were what was meant, the term would have no contrastive force, for there would be no interesting statements which were absolutely valid. Absolute validity would be confined to platitudes, logical or mathematical truths, and the like: the sort of beliefs nobody wants to argue about because they are neither controversial nor central to anyone’s sense of who she is or what she lives for. All beliefs which are central to a person’s self-image are so because their presence or absence serves as a criterion for dividing good people from bad people, the sort of person one wants to be from the sort one does not want to be. A conviction which can be justified to anyone, which even bad people can be argued into accepting, is of little interest. Unflinching courage will not be required to sustain it.

We must, therefore, construe the term ‘only relatively valid beliefs’ to contrast with statements capable of being justified to all those who are uncorrupted – that is, to all those in whom reason, viewed as a built-in truth-seeking faculty, or conscience, viewed as a built-in righteousness-detector, is strong enough to overcome vulgar superstitions, base prejudices and evil passions. The notion of ‘absolute validity’ only makes sense if we presuppose a self which divides fairly neatly into the part that it shares with the divine and the part that it shares with the animals. But if we allow Sandel such an opposition between reason and passion, or reason and will, we liberals will be begging the question against ourselves. It is incumbent on those of us who agree with Freud and Berlin not to split persons up in this way to drop or restrict the use of the traditional distinction between ‘rational conviction’ and ‘conviction brought about by (mere) causes rather than by reasons’.

The best way of restricting its use is to limit the opposition between rational and irrational ways of changing belief to the interior of a language-game, rather than trying to apply it to momentous shifts in linguistic behaviour, transitions to new language-games. Such a restricted notion of rationality is all that we can allow ourselves if we accept the claim I made in ‘The Contingency of Language’ that what matters in the end are changes of vocabulary rather than changes in belief, changes in candidates for truth-value rather than changes in sentential attitudes. Within a language-game, within a set of agreements about what is possible and important, we can usefully distinguish reasons for belief from causes for belief which are not reasons. We do this by starting with such obvious differences as that between Socratic dialogue and hypnotic suggestion. We then try to firm up the distinction by dealing with messier cases: brain-washing, media hype, and what Marxists call ‘false consciousness’. There is no very neat way to draw the line between persuasion and force, and thus no very neat way to draw a line between a cause of changed belief which was also a reason and one which was a ‘mere’ cause, but the distinction is no fuzzier than most.

However, once we raise the question of how we get from one vocabulary to another, from one dominant metaphoric to another, the distinction between reasons and causes begins to lose its utility. Those who speak the old language and have no wish to change, those who regard it as a hallmark of rationality or morality to speak just that language, will regard as altogether irrational the appeal of the new metaphors, of the new language-game which the avant-garde is playing. The popularity of the new ways of speaking will be viewed as a matter of ‘fashion’ or ‘the need to rebel’ or ‘decadence’. The question of why people speak this way will be treated as beneath the level of conversation – a matter to be turned over to psychologists or, if necessary, the police. Conversely, from the point of view of those who are trying to use the new language, to literalise the new metaphors, those who cling to the old language will be viewed as irrational – as victims of passion, prejudice, superstition, the dead hand of the past, and so on. The philosophers on either side can be counted on to support these opposing invocations of the reason-cause distinction by developing a moral psychology, or an epistemology, or a philosophy of language, which will put the other side in a bad light.

To accept the claim that there is no standpoint outside our own particular historically-conditioned and temporary vocabulary by which to judge this vocabulary in respect of rationality or morality is to give up on the idea that we can reach agreement on good reasons for using new languages, as opposed to good reasons, within old languages, for believing statements within those languages. This amounts to giving up the idea that intellectual or political progress is rational, in the sense of ‘rational’ which entails the satisfaction of criteria which are neutral between the competing parties. But since it seems pointless to say that all the great moral and intellectual advances of European history – Christianity, Galilean science, the Enlightenment, Romanticism etc – were fortunate falls into temporary irrationality, the moral to be drawn is that the rational-irrational distinction is less useful than it once appeared. Once we realise that progress, for the community as for the individual, is a matter of using new words as well as of arguing from premises phrased in old words, we realise that a critical vocabulary which revolves around notions like ‘rational’, ‘criteria’, ‘argument’, ‘foundation’ and ‘absolute’ is badly suited to describe the relation between the old and the new.

Donald Davidson has pointed out that once we give up on the notion of ‘absolute criteria of rationality’, and begin using ‘rational’ to mean something like ‘internal coherence’, then, if we do not limit the range of this term’s application, we shall be forced to call ‘irrational’ many things which we wish to praise. In particular, we shall have to describe as ‘irrational’ what Davidson calls ‘a form of self-criticism and reform which we hold in high esteem, and that has even been thought to be the very essence of rationality and the source of freedom’. He makes the point as follows:

What I have in mind is a special kind of second-order desire or value, and the actions it can touch off. This happens when a person forms a positive or negative judgment of his own desires, and he acts to change these desires. From the point of view of the changed desire, there is no reason for the change – the reason comes from an independent source, and is based on further, and partly contrary, considerations. The agent has reasons for changing his own habits and character, but those reasons come from a domain of values necessarily extrinsic to the contents of the views or values to undergo change. The cause of the change, if it comes, can therefore not be a reason for what it causes. A theory that could not explain irrationality would be one that also could not explain our salutary efforts, and occasional successes, at self-criticism and self-improvement.

Davidson would, of course, be wrong if there were a framework of non-trivial highest-possible-order desires within which self-criticism and self-improvement take place. For then these highest-level desires would mediate and rationalise the contest between first and second-level desires. But Davidson is assuming, rightly in my opinion, that the only candidates for such highest-level desires are so abstract and empty as to have no mediating powers: for example, ‘I wish to be good,’ ‘I wish to be rational,’ ‘I wish to know the truth.’ What will count as good or rational or true will be determined by the contest between the first and second-level desires. Wistful top-level protestations of good will are impotent to intervene in that contest.

If Davidson is right, then Sandel is wrong. For Sandel is assuming that there is a largest-possible framework within which one can ask, for example: ‘If freedom has no morally privileged status, if it is just one value among many, then what can be said for liberalism?’ He is assuming that we liberals ought to be able to rise above the contingencies of history and see the kind of individual freedom which the modern liberal state offers its citizens as just one more value. He is suggesting that the rational thing to do is to place such freedom alongside, for example, the sense of national purpose which the Nazis briefly offered the Germans, or the sense of conformity to the will of God which inspired the Wars of Religion. Then one can scrutinise these various candidates and see whether any of them is ‘morally privileged’. Only the assumption that there is some such standpoint to which we might rise gives sense to his question: ‘If one’s convictions are only relatively valid, why stand for them unflinchingly?’

Conversely, neither Schumpeter’s phrase ‘relative validity’ nor Sandel’s notion of a ‘relativist predicament’ will seem to the point if one grants that new metaphors are causes, but not reasons, for changes of belief, and that it is new metaphors which make intellectual progress possible. If one grants these claims, there is no such thing as ‘the relativist predicament’, just as for someone who does not believe in God there is no such thing as blasphemy. For there is no higher standpoint to which we are responsible and against whose precepts we might offend. There will be no such activity as scrutinising competing values in order to see which are morally privileged. For there will be no way to rise above the language, culture, institutions and practices one has adopted and view all these as on a par with all the others. As Davidson puts it, ‘speaking a language ... is not a trait a man can lose while retaining the power of thought. So there is no chance that somebody can take up a vantage-point for comparing conceptual schemes by temporarily shedding his own.’ Or, to put the point in Heidegger’s way, ‘language speaks man,’ languages change in the course of history, and so human beings cannot escape their historicity. The most they can do is to manipulate the tensions within their own epoch in order to produce the beginnings of the next epoch.

But, of course, if Sandel is right, then Davidson and Heidegger are wrong. Davidsonian and Wittgensteinian philosophy of language – the account of language as a historical contingency, rather than as a medium which is taking on (or might, with luck, take on) the true shape of the true world or the true self – will beg the question against Sandel. If we think that Sandel is right to pose the questions he does, then we shall ask instead for a philosophy of language, an epistemology and a moral psychology which will safeguard the interests of reason, preserve a morality-prudence distinction, and thus guarantee that Sandel’s questions are to the point. We shall want a different way of seeing language, one which treats it as a medium in which to express truths which lurk within the self, truths which formulate the permanent, ahistorical, highest-level desires which adjudicate lower-level conflicts. We shall want to refurbish the subject-object and scheme-content models which Davidson, like Heidegger, thinks are obsolete.

Is there a way to break this stand-off between Sandel and Davidson? More generally, is there a way to break the stand-off between the traditional view, here represented by Sandel, which says that it is always sensible to ask ‘How do you know?’ and the view represented by Davidson, which says that, when the chips are down, we can only ask: ‘Why do you talk that way?’ It is tempting to say that it is philosophy’s job to rise above such stand-offs and adjudicate the matter, but the assumption that there is a highest-level discipline which can do that begs all the questions in Sandel’s favour.

Most philosophers nowadays would not attempt this adjucative, magisterial role. They would admit that there is no one, privileged way to resolve such stand-offs, no single place to which it is appropriate to step back. There are, instead, as many ways of breaking the stand-off as there are topics of conversation. One can come at the issue by way of different paradigms of humanity: the contemplator as opposed to the poet, or the pious man as opposed to the man who accepts chance as worthy of determining his fate. Or one can come at it from the point of view of an ethics of kindness, and ask whether cruelty and injustice will be diminished if we all stopped worrying about ‘absolute validity’ or whether, on the contrary, only such worries keep our characters firm enough to defend unflinchingly the weak against the strong. Or one can come at it by way of anthropology, and the question of whether there are ‘cultural universals’, or by way of psychology and the question of whether there are psychological universals. Because of this indefinite plurality of standpoints, this vast number of ways of coming at the issue sideways and trying to outflank one’s opponent, there are never, in practice, any stand-offs.

You would only get a real and practical stand-off, as opposed to an artificial and theoretical one, if certain topics and certain language-games were taboo: if there were general agreement within a society that certain questions were always to the point, that certain questions were prior to certain others, that there was a fixed order of discussion and that flanking movements were not permitted. That would be just the sort of society which liberals are trying to avoid – one in which ‘logic’ (and, perhaps, philosopher-kings) ruled and ‘rhetoric’ was outlawed. It is central to the idea of a liberal society that, in respect of words as opposed to deeds, persuasion as opposed to force, anything goes. Liberals should not recommend this openmindedness on the scriptural ground that truth is great and will prevail, nor on the Miltonic ground that truth will always win in a free and open encounter – for the latter suggests that we have independent criteria for the presence of truth and the freedom of encounters. We do not. Rather, a liberal society is one which is content to call ‘true’ whatever the upshot of such encounters turns out to be. That is why a liberal society is badly served by an attempt to supply it with ‘philosophical foundations’. For the attempt to supply such foundations presupposes a natural order of topics and arguments which is prior to, and overrides the results of, encounters between old and new vocabularies.

This last point leads me back to my earlier claim that, in order for liberal culture to realise itself by shaping its own vocabulary, it needs an apologetics rather than a set of foundations. The idea that it ought to have foundations was a result of Enlightenment scientism, which was in turn a survival of the religious need to have human projects underwritten by a non-human authority. It was natural for liberal political thought in the 18th century to try to associate itself with the most promising cultural development of the time, the natural sciences. But unfortunately the Enlightenment wove its political rhetoric around a picture of the scientist as a sort of priest, someone who achieved contact with non-human truth by being ‘logical’, ‘methodical’ and ‘objective’. This was a useful tactic in its day, but it is less useful nowadays. For historians of science have made clear how little this picture of the scientist has to do with actual scientific achievement, how pointless it is to try to isolate something called ‘the scientific method’. The Kuhnian rhetoric of praise for great scientists is the same rhetoric which we use to praise Marx or Wordsworth or Freud: we speak of their powerful creative imagination, not of their dogged and methodical devotion to truth.

I said at the beginning of ‘The Contingency of Language’ that the French Revolution and the Romantic movement inaugurated an era in which we gradually came to appreciate the historical role of linguistic innovation – an era whose emblem is the vague but inspiring thought that truth is made rather than found. I also said that nowadays literature and utopian politics are the spheres to which we look when we worry about ends rather than about means. I can now add the corollary that these, rather than science, are the areas to which we should look for an apologetics for liberal society. We need an apologetics for liberalism which revolves around the hope that culture as a whole can be ‘aestheticised’ rather than around the Enlightenment hope that it might be ‘scientised’. Liberal politics is best suited to a culture whose hero is the strong poet rather than the truth-seeking, ‘logical’, ‘objective’ scientist. Such a culture would slough off the Enlightenment vocabulary which Sandel presupposes, and would thus slough off the questions which he puts to Berlin. It would be no longer haunted by spectres called ‘relativism’ and ‘irrationalism’. Such a culture would not assume that a form of cultural life is no stronger than its philosophical foundations. It would drop the idea of such foundations and would regard the justification of liberal society as a matter of invidious comparison with other attempts at social organisation – those of the past and those envisaged by utopians.

To think such an apologetics sufficient would be to draw the consequences of Wittgenstein’s insistence that vocabularies – all vocabularies, even those which contain the words which we take most seriously, the ones most essential to our self-descriptions – are human creations, tools for the creation of other human artefacts such as poems, utopian societies, scientific theories and future generations. To build the rhetoric of liberalism around this Wittgensteinian thought would mean giving up the idea that liberalism could be justified, and Nazi or Marxist enemies of liberalism refuted, by driving the latter up against an argumentative wall – forcing them to admit that liberal freedom has a ‘moral privilege’ which their own values lacked. From the point of view I have been commending, any attempt to drive one’s opponent up against a wall in this way is, in the long run, futile. For sooner or later the wall against which he is driven will come to be seen as just one more vocabulary, one more way of describing things. The wall then turns out to be a painted backdrop, one more work of man, one more bit of cultural stage-setting. An aestheticised culture would be one which would not insist we find the real wall behind the painted ones, the real touchstones of truth as opposed to touchstones which are merely cultural artefacts. It would be a culture which, precisely by appreciating that all touchstones are such artefacts, would take as its goal the creation of ever more various and multi-coloured artifacts.

To sum up, the moral I want to draw from my discussion of Sandel’s claim that Berlin’s position is ‘relativitic’ is that we should learn to brush aside questions like ‘How do you know that freedom is the chief goal of social organisation?’ in the same way as we brush aside questions like ‘How do you know that Jones is worthy of your friendship?’ or ‘How do you know that Yeats is an important poet, Hegel an important philosopher, or Galileo an important scientist?’ We must see allegiance to social institutions as no more matters of knowledge, but also as no more arbitrary, than choices of friends or heroes. Such choices are not made by reference to criteria. They cannot be preceded by presuppositionless critical reflection, conducted in no particular language and outside any particular historical context.

When I say ‘we must’ do this or ‘we cannot’ do that, I am not, of course, speaking from a neutral standpoint. I am speaking from Berlin’s side of the argument, trying to serve as an under-labourer to Berlin by clearing away some of the remaining philosophical underbrush. I am no more neutral, and philosophy can no more be neutral, on political matters of this magnitude than Locke, who originated the ‘under-labourer’ metaphor, could be neutral between Aristotelianism and corpuscularianism. But, here again, when I say that neutrality is not a desideratum, I am not saying this from a neutral philosophical perspective. I am not laying foundations for liberalism by claiming that recent Davidsonian philosophy of language and Kuhnian philosophy of science have demonstrated that the philosophers of the past were mistaken in asking for neutrality. I am saying that Kuhn, Davidson, Wittgenstein and Derrida provide us with redescriptions of familiar phenomena which, taken together, buttress Berlin’s way of describing alternative political institutions and theories. These philosophers help provide an apologetics for political liberalism.

But, of course, political liberalism also helps provide an apologetics for them. There is no natural order of inquiry which prescribes that we must first sort out our ideas about language, then about belief and knowledge, then about personhood and finally about society. One can just as well start from social theory and work back to philosophy of language (as some Marxists, for example, have done). There is no such thing as ‘first philosophy’: that metaphilosophical claim is itself just one more terminological suggestion made on behalf of the same cause, the cause of providing contemporary liberal culture with a vocabulary which is all its own, cleansing it of the residue of a vocabulary better suited to the needs of an earlier day.

I shall try to make this abjuration of philosophical neutrality in the interest of political liberalism more palatable by returning yet again to the Wittgensteinian analogy between vocabularies and tools. I said in ‘The Contingency of Language’ that one problem with this comparison is that the craftsman who designs a new tool can explain in advance why he wants it. By contrast, a new form of cultural life, a new vocabulary, can only explain its own utility post festum. We cannot see Christianity or Newtonianism or the Romantic movement or political liberalism as a tool while we are still in the course of figuring out how to use it. For there are as yet no clearly formulatable ends to which it is a means. But once we figure out how to use the vocabularies of these movements, we can tell a story of progress, showing how the literalisation of certain metaphors served the purpose of making possible all the good things that have recent’y happened. Further, we can now view all the e good things as particular instances of some more general good, the over-all end which the movement served. The attempt at such a view was one of Hegel’s definitions of philosophy: ‘grasping your time in thought’. I construe this to mean ‘finding a new description of all the things characteristic of your time of which you most approve, the things with which you unflinchingly identify, and then using that descriptive vocabulary to redescribe the past events which made these good things possible’.

On that meaning of ‘philosophy’, it follows that, as Hegel went on to say, ‘philosophy paints its grey on grey only when a form of life has grown old.’ Christianity did not know that its purpose was the alleviation of cruelty. Newton did not know that his purpose was the leisure and wealth made possible by modern technology. The Romantic poets did not know that their purpose was the development of an ethical consciousness suitable for the culture of political liberalism. But we now know these things, for we late-comers can tell the kind of story of progress which those who are actually making progress usually cannot. We can view these people as tool-makers rather than discoverers because we have a clearer idea of what was shaped by the use of those tools. The product is us – our conscience, our culture, our form of life. Those who made us possible could not clearly describe the ends to which their work was a means. But we can.

Let me now apply this point to the particular case of the relation between political liberalism and Enlightenment rationalism. This relation was the topic of Horkheimer and Adorno’s Dialectic of Enlightenment. They pointed out, correctly, that the forces unleashed by the Enlightenment have undermined the Enlightenment’s own convictions. What they called the ‘dissolvant rationality’ of Enlightenment has, in the course of the triumph of Enlightenment ideas during the last two centuries, undercut the ideas of ‘rationality’ and of ‘human nature’ which the 18th century took for granted. Horkheimer and Adorno, however, drew the conclusion that liberalism was now intellectually bankrupt, bereft of philosophical foundations, and that liberal society was morally bankrupt. This inference was a mistake. Horkheimer and Adorno assumed that the terms in which those who begin a historical development describe their enterprise are the terms which describe it correctly, and then inferred that a repudiation of that terminology deprives the results of that development of a right to exist. This is almost never the case. On the contrary, the terms used by the founders of a new form of cultural life consist largely in borrowings from the vocabulary of the culture which they are hoping to replace. Only when the new form has grown old, and has itself become the target of attacks from the avant-garde, will the proper terminology of that culture begin to take form. The terms in which a mature culture compares other cultures invidiously with itself, in which it couches its apologetics, are not likely to be the terms which were used to bring about its birth.

Horkheimer and Adorno give an admirable account of the way in which philosophical foundations of social practices (which they view as, typically, linguistic instruments of domination by the rulers) are undercut by Enlightenment scepticism. As they say, ‘ultimately the Enlightenment consumed not just the symbols [of social union] but their successors, universal concepts, and spared no remnant of metaphysics ... The situation of concepts in the face of the Enlightenment is like that of men of private means in regard to industrial trusts: none can feel safe.’ Among the distinctions which have been unable to withstand this dissolution are ‘absolute validity v. relative validity’, and ‘morality as opposed to prudence’. As Horkheimer and Adorno put it, the spirit of Enlightenment dictates that ‘every specific theoretic view succumbs to the destructive criticism that it is only a belief – until even the very notions of spirit, of truth and, indeed, enlightenment itself have become animistic magic.’ This point can be put in my jargon by saying that every specific theoretic view comes to be seen as one more vocabulary, one more description.

Horkheimer and Adorno thought it likely that civilisation could not survive this process. They had nothing helpful to suggest except ‘the hermeneutics of suspicion’ – constant awareness that any new theoretical proposal is likely to be one more excuse for maintaining the status quo. They said that ‘if consideration of the destructive aspects of progress is left to its enemies, blindly pragmatised thought loses its transcending quality and its relation to truth,’ but they had no suggestions for its friends. They had no utopian vision of a liberal culture which would be able to incorporate and make use of an understanding of the dissolvant character of rationality, of the self-destructive character of Enlightenment. They did not try to show how ‘pragmatised thought’ might cease to be blind.

Various other writers, however, have tried to do just that. They have retained Enlightenment liberalism while (partially or completely) dropping Enlightenment rationalism. Berlin is an example, and so are John Dewey, Michael Oakeshott and John Rawls. These writers have helped undermine the idea of a trans-historical ‘absolutely valid’ set of concepts which would serve as ‘philosophical foundations’ of liberalism. But each has thought of this undermining as a way of strengthening liberal institutions. They have argued that liberal institutions would be better-off when freed from the need to defend themselves by an appeal to such foundations – better-off, too, for not having to answer Sandel’s question: ‘In what does the privileged status of freedom consist?’ All four would grant that a circular justification of our practices, a justification which makes one feature of our culture look good by citing still another, or comparing our culture invidiously with others by reference to our own standards, is the only sort of justification we are going to get. On my view, these writers represent the self-cancelling and self-fulfilling triumph of the Enlightenment. Their pragmatism is antithetical to Enlightenment rationalism, but was itself made possible only by Enlightenment rationalism, and now serves as the vocabulary of a mature Enlightenment liberalism.

Let me cite some passages from these authors to remind you of their positions. Dewey echoes Hegel’s definition of philosophy in saying: ‘When it is acknowledged that under disguise of dealing with ultimate reality, philosophy has been occupied with the precious values embedded in social traditions, that it has sprung from a clash of social ends and from a conflict of inherited institutions with incompatible contemporary tendencies, it will be seen that the task of future philosophy is to clarify men’s ideas as to the social and moral strifes of their own day.’ In his Dewey Lectures, Rawls echoes both Berlin and Dewey when he says: ‘What justifies a conception of justice is not its being true to an order antecedent to and given to us, but its congruence with our deeper understanding of ourselves and our aspirations, and our realisation that, given our history and the traditions embedded in our public life, it is the most reasonable doctrine for us.’ Finally, Oakeshott writes, in sentences which Dewey might equally well have written:

A morality is neither a system of general principles nor a code of rules, but a vernacular language. General principles and even rules may be elicited from it, but (like other languages) it is not the creation of grammarians; it is made by speakers. What has to be learned in a moral education is not a theorem such as that good conduct is acting fairly or being charitable, nor is it a rule such as ‘always tell the truth,’ but how to speak the language intelligently ... It is not a device for formulating judgments about conduct or for solving so-called moral problems, but a practice in terms of which to think, to choose, to act and to utter.

This quotation from Oakeshott gives me a springboard for explaining why I think that the distinction between morality and prudence, and the term ‘moral’ itself, are no longer very useful. My argument turns on the familiar anti-Kantian claim, which Oakeshott is here taking for granted, that ‘moral principles’ (the categorical imperative, the utilitarian principle etc) only have a point insofar as they incorporate tacit reference to a whole range of institutions, practices, and vocabularies of moral and political deliberation. They are reminders of, abbreviations for, such practices, not justifications for them. At best, they are pedagogical aids to the acquisition of such practices. This point, common to Hegel and to recent critics of academic moral philosophy such as Annette Baier, J.B. Schneewind, Charles Taylor and Bernard Williams, suggests the question: since the classic Kantian opposition between morality and prudence was formulated precisely in terms of the opposition between an appeal to principle and an appeal to expediency, is there any point in keeping the term ‘morality’ once we drop the notion of ‘moral principle’?

Oakeshott, following Hegel, suggests an answer: we can keep the notion of ‘morality’ just in so far as we cease to think of morality as the voice of the divine part of ourselves and instead think of it as the voice of ourselves as members of a community, speakers of a common language. We can keep the morality-prudence distinction if we think of it, not as the difference between an appeal to the unconditioned and an appeal to the conditioned, but as the difference between an appeal to the interests of our community and an appeal to our private, possibly conflicting interests. The importance of this shift is that it makes it impossible to ask the question ‘Is ours a moral society?’ It makes it impossible to think that there is something which stands to my community as my community stands to me, some larger community called ‘humanity’ which has an intrinsic nature. Such a shift is appropriate for what Oakeshott calls a societas as opposed to a universitas, for a society conceived as a band of eccentrics collaborating for purposes of mutual protection rather than as a band of fellow-spirits united by a goal.

Oakeshott’s answer coincides with Wilfred Sellars’s thesis that morality is a matter of what he calls ‘we-intentions’, that the core meaning of ‘immoral action’ is ‘the sort of thing we don’t do’. An immoral action is, on this account, the sort of thing which, if done at all, is done only by animals, or by people of other families, tribes, cultures or historical epochs. If done by one of us, or if done repeatedly by one of us, that person ceases to be one of us. She becomes an outcast, someone who does not really speak our language, though she at first appeared to do so. On Sellars’s account, as on Hegel’s, moral philosophy takes the form of an answer to the question ‘Who are “we”, how did we come to be what we are, and what might we become?’ rather than an answer to the question ‘What rules should dictate my actions?’ In other words, moral philosophy takes the form of historical narration and utopian speculation rather than of a search for general principles.

This Oakeshott-Sellars way of looking at morality as a set of practices, our practices, makes vivid the difference between morality as the voice of a divinised portion of our soul, and as the voice of a contingent human artefact, a community which has grown up subject to the vicissitudes of time and chance, one more of nature’s experiments. It makes clear why the morality-prudence distinction breaks down when we attempt to transfer it to questions about whether the glue that holds our society together is ‘moral’ or ‘prudential’ in nature. That distinction only makes sense for individuals. It would make sense for societies only if ‘humanity’ had a nature over and above the various forms of human life which history has thrown up so far. But if the demands of a morality are the demands of a language, and if languages are historical contingencies, then to ‘stand unflinchingly for one’s moral convictions’ is a matter of identifying oneself with such a contingency.

Let me now try to connect this point with my earlier claim that the heroes of liberal society are the strong poet and the utopian revolutionary. Such a synthesis will seem paradoxical if one thinks of the poet or the revolutionary as necessarily ‘alienated from society’. But the paradox can be resolved if one thinks of the ideal liberal society as one in which the distinction between the reformer and the (violent) revolutionary is no longer necessary. An ideally liberal society is one in which whatever is both desirable and possible can be achieved by persuasion rather than force, reform rather than revolution, by the free and open encounters of present linguistic and other practices with suggestions for, and examples of, new practices. But this is to say that a liberal society is one which has no ideal except freedom, no goal except a willingness to see how such encounters go and to abide by the outcome. It is a societas rather than an universitas precisely because it has no purpose except to make life easier for poets and revolutionaries while seeing to it that they make life harder for others only by words, and not deeds. It is a society whose hero is the strong poet and the revolutionary because it recognises that it is what it is, has the morality it has, speaks the language it does, not because it approximates the will of God or the nature of man, but because certain poets and revolutionaries of the past spoke as they did.

To see one’s language, one’s conscience, one’s morality and one’s highest hopes as contingent products, as literalisations of what once were accidentally-produced metaphors, is to adopt an identity which suits one for citizenship in such an ideally liberal state. That is why the ideal citizen of such an ideal state would be someone who thinks of the strong poet as the ideal human being. She thinks of the founders and the preservers of her society as such poets, rather than as people who discovered or clearly envisioned the antecedently-existing truth about the world or about humanity. She herself may or may not be a poet, may or may not find her own metaphors for her own idiosyncratic fantasies, may or may not make those fantasies conscious. But she will be commonsensically Freudian enough to see the founders and the transformers of society, the acknowledged legislators of her language and thus of her morality, as people who happened to find metaphors to fit their fantasies, metaphors which happened to answer to the vaguely felt needs of the rest of society. She will be commonsensically Bloomian enough to take for granted that it is the revolutionary artist and the revolutionary scientist, not the academic artist or the normal scientist, who embodies the virtues which she hopes her society will foster.

There are many objections to what I have been saying, but the one which I find most disturbing says that I am treating democratic societies as existing for the sake of intellectuals. I seem to be describing institutions which were constructed in order to prevent cruelty and obtain justice as if they had been constructed to safeguard the freedom of a leisured élite.

My initial reply to this objection is that there are fairly tight connections between the freedom of the intellectuals, on the one hand, and the diminution of cruelty, on the other. We intellectuals of the rich North Atlantic democracies sometimes wonder whether our concern for our fellow intellectuals in places like Cuba, Chile and Poland – our feeling that these dissidents are the exemplary human beings of our time – may not be myopic. These doubts are taken advantage of by apologists for Soviet tyranny, who remind us that the prevention of famine, and the availability of cheap housing and medical care, are more important to most people than the welfare of poets. Apologists for the United States’ support of ‘authoritarian’ governments use similar arguments, suggesting that suppression of the press and of academic freedom is a small price to pay for order and economic stability – a price which all but a few eccentrics are willing to pay. In this situation we are torn between insisting that the welfare of dissident poets ought to be of interest to non-poets, and wondering whether we say this simply because we ourselves find it easier to identify with poets than with peasants.

One way to set aside this sort of self-doubt is to remind ourselves of particular events: of Solidarity, and in particular of the day when they engraved Milosz’s line, ‘A poet remembers,’ on the monument to the strikers murdered by the police in Gdansk. But there is another way. This is to come to terms with the general philosophical claim that what counts as cruelty and injustice is a matter of the language that is spoken. To see a common social practice as cruel and injust is, on the view I have been putting forward, a matter of redescription rather than of discovery. It is a matter of changing vocabularies rather than of stripping away the veil of appearance from an objective reality, of experimentation with new ways of speaking rather than of overcoming ‘false consciousness’. When we ask ourselves how our ancestors, or the inhabitants of an exotic culture, or our past selves could have been so blind to the cruelties they (or we) practised, the right answer is that they (or we) were using a language which was built around this practice, a language different from the one in which we are now condemning it. To protect the poets and the utopian fantasts, the people who do not talk as we do, and who thus necessarily come under suspicion of irrationality and immorality, is the only thing which our society can do to ensure that its language keeps changing. It is, therefore, the only thing that can prevent a society from normalising what will appear, from the point of view of an ideal future, its characteristic patterns of cruelty and injustice.

If the ideal culture of the ideal liberal society ever comes into existence, if Nietzsche, Wittgenstein, Freud and Bloom become part of common sense, then the battle between moralism and Romanticism will be over, because the language in which the issue between them was stated will be obsolete. But in the meantime, we should side with the Romantics and do our best to aestheticise society, to keep it safe for the poets in the hope that the poets may eventually make it safe for everybody else. ‘Imagination,’ in Dewey’s words, ‘is the chief instrument of the good ... art is more moral than moralities.’

[*] ‘The Contingency of Language’ was published in the LRB of 17 April, and ‘The Contingency of selfhood’ in the following issue.