I. The excitement of theory

How could anyone find A Theory of Justice—that long, elaborate tome with barely a concession to the attention span of the distractible reader—exciting? “Why,” asked Brian Barry, “was this extremely long, poorly organized and stylistically undistinguished book such a smash hit?”Footnote ¹ H. L. A. Hart, for instance, had claimed, “No book of political philosophy since I read the great classics of the subject has stirred my thoughts as deeply.”Footnote ² Robert Nozick, no admirer of Rawls's central theses, expressed his admiration in aesthetic terms: Rawls's book was “a fountain of illuminating ideas, integrated together into a lovely whole … it is impossible to finish his book without a new and inspiring vision of what a moral theory may attempt to do and unite; of how beautiful a whole theory can be.”Footnote ³ Stuart Hampshire declared it “a permanent refutation of the reproach that analytical philosophy cannot contribute to substantial moral and political thought.”Footnote ⁴ Bernard Williams pronounced it

In each case, we find a hint of what Rawls provided that felt novel. Nozick and Hampshire speak in the voice of the newly unshackled, slightly giddy with their newfound freedom, their sense of what was now permitted and possible. When the journal Ethics commissioned a review of the five-volume collection of secondary literature on Rawls in 2003, the reviewer titled it “The House That Jack Built.”Footnote ⁶ The house in the metaphor was the vast Rawlsian corpus, but it could equally have referred to the post-1971 tradition of anglophone political philosophy itself. The house seemed to have come into being ex nihilo, with no immediate precedent in philosophy. This view of Rawls, if not exactly initiated by him—he was thought by his acquaintances to be a genuinely modest man—was certainly not discouraged by his notorious habit of leaving most mentions of his contemporaries to footnotes, with only the canonical greats allowed into the main text.Footnote ⁷ Rawls described his project as an attempt “to present a conception of justice which generalizes and carries to a higher level of abstraction the familiar theory of the social contract as found, say, in Locke, Rousseau, and Kant.”Footnote ⁸ The remark can be read as doing two things at once: making a simple avowal of his anachronistic intellectual allegiances to a no-longer-influential tradition, but also—and less modestly—inviting comparisons of his project not to the work of his contemporaries, but to the canonical texts of modern social and political thought.Footnote ⁹

Recent, historicizing, scholarship on Rawls, including several of the essays in the present issue, has put paid to the assessment of Rawls's achievement as wholly without precedent.Footnote ¹⁰ But the reactions I quote above came from contemporaries of his, writing shortly after the publication of A Theory of Justice and well placed to distinguish genuine originality from the product of historical amnesia. What is that sense of freshness and possibility these remarks express if not a recognition of the de novo quality of Rawls's achievement? This is the question I shall be trying to answer in this article. Hampshire's remark about Theory constituting “a permanent refutation” of a certain “reproach” offers a natural invitation to the intellectual historian. Why might anyone have thought that something called “analytical philosophy” could not “contribute to substantial moral and political thought”? This calls both for a clearer statement of what “analytical” philosophy was being understood to be, and for an understanding of what might be “substantial” political thought. Philosophers after 1971 told a range of stories about the bad old days before Rawls. Like other postrevolutionary accounts of life under the ancien régime, these stories should not be accepted at face value. One might see them as reports of the front from those who had served in the trenches—sincerely described, if somewhat imperfectly remembered. Alternatively, one might see them as the first drafts of later entrenched legitimation myths that it is now time to debunk. Either way, my questions call for, or at least permits, historicization in a certain style, where the sources do not have to be unpublished to be novel. These questions can fruitfully be approached by extrapolating from published—and in some cases extremely well-known—texts answers to questions that we are only now beginning to ask.

When Peter Laslett in 1956 wrote in his introduction to an anthology of philosophical essays on politics and society that “for the moment, anyway, political philosophy is dead,”Footnote ¹¹ he was taking a (perhaps deliberately) narrow view of what philosophy was, viz. academic, analytic, anglophone. It is certainly hard to imagine a French author making a similar assessment in 1956. The authors in Laslett's volume cleaved to different conceptions of philosophy but seemed both to concur in Laslett's judgment and to see nothing regrettable about it. Michael Oakeshott's conception of philosophy was in its origins idealist and the conception of politics Burkean: “political philosophy … has no power to guide or to direct us in the enterprise of pursuing the intimations of our tradition.”Footnote ¹² With others, more closely identified than Oakeshott with what was by then becoming the “analytic” mainstream of philosophy, the motivations seemed to be grounded in a distinct form of skepticism about the powers of philosophy. T. D. Weldon, for instance, wrote, “Philosophers are to make no claim to improve either the theories or the methods of those who are engaged on scientific or political activities.”Footnote ¹³ In The Vocabulary of Politics published a few years previously, Weldon had condemned much of the earlier tradition of political philosophy for having “formulated questions to which no empirically testable answers could be given, and such questions are nonsensical.”Footnote ¹⁴ Even one of the more optimistic appraisals in Laslett's volume of the possibilities of political philosophy, by W. B. Gallie, was distinguished mostly by its minimalism: political philosophy “might help to dispel that thin but bewildering fog which seems to surround so much current ethical discussion and to make its terms and distinctions so unearthly and unreal.”Footnote ¹⁵ If philosophy had to restrict itself to the clarification of concepts and the formulation of hypotheses that might then be tested by one or another kind of scientist, it was unlikely that anything it produced on the subject of politics was likely to be of much interest to someone interested in politics, rather than in concepts.

A later account, from 1975, came from a radio broadcast by Bernard Williams, tasked with explaining to listeners of BBC Radio 3 the state of contemporary political philosophy, as part of a series titled Current Trends in Philosophy. “Twenty years ago,” Williams began, “the prevailing view in English-speaking philosophical circles was that political philosophy would never flourish again.”Footnote ¹⁶ Twenty years ago was 1955, when Williams himself was a young “don,” giving lectures and tutorials at Oxford, still the hegemonic center of anglophone philosophy. It was roughly the time that Rawls himself had visited the city, on a Fulbright fellowship. Williams put the “moribund condition of political philosophy at that time” down to several causes, some philosophical and others political. The primary philosophical causes he took to be the “prevailing theory that statements of value were sharply to be separated from statements of fact or theory,” and reluctance on the part of philosophers “to come out with statements of value.” Outside the academy, the state of contemporary politics in Britain and the USA, “one which prematurely saluted the end of ideology,” did not provide the “context of political urgency” that made for “living political philosophy.”Footnote ¹⁷

By 1975, Williams continued, this had all changed. Few philosophers believed in sharply distinguishing fact from value; philosophers were much less reluctant to make judgments of value, and with the Vietnam War came the immediate demand to reflect on questions about the justice of war, the right of the state to demand (military) service, and the moral status of civil disobedience and conscientious objection. Williams also mentions the politics of economic redistribution, feminism and environmentalism.Footnote ¹⁸ A Theory of Justice, he noted, was in the making “for a long time before recent political developments, but it is by no means unmarked by them.”Footnote ¹⁹

The story I defend here is not altogether revisionary. I have no quarrel with those who assert the novelty of Rawls's work in contrast to some of his context, nor with those who decried the moribund state of anglophone political philosophy in the academy before Rawls. But the picture stands to be complicated a little, the details brought into sharper relief. I shall not be trying to say that the bad old days were not so bad after all. Rather, I shall be trying to explain why political philosophy before Rawls was in the state it was in—in terms that do not assume that those who practiced it were stupid. I am looking for an account—articulable both in general theoretical terms and as features of the context—that would make the ways of the ancien régime at least intelligible. Why, in other words, didn't more people at the time take the same dim view of the status quo as would come to be orthodox later on? And can we understand this without simply appealing to considerations a sensitivity to which is itself a product of the philosophical revolution?

We can begin with a potted, but orthodox, account of the tradition of “analytic” philosophy. On any account of analytic philosophy, that tradition emerged out of foundational work in logic and the philosophy of language in the late nineteenth century, in particular in the work of Gottlob Frege in Germany, and later (and with more immediate influence) in the early work of Bertrand Russell in Britain, centrally concerned with giving an account of linguistic meaning.Footnote ²⁰ Russell's insights were both developed and subjected to deep critique by Ludwig Wittgenstein, an Austrian émigré who studied with Russell at Cambridge. Wittgenstein's central preoccupation with the distinction between sense and nonsense was taken up, partly in virtue of his influence and partly as an organic development, in Vienna after the First World War, and developed by its central philosophical figures, Moritz Schlick and Rudolf Carnap, into a more narrow and polemical doctrine, generally labeled “logical positivism,” which restricted meaningful discourse to two things: the empirical statements of the (natural and social) sciences and statements true in virtue of the meanings of the terms involved.Footnote ²¹ Viennese positivism was a proselytizing doctrine and won converts (or fellow travelers) in the English-speaking world through a combination of frenetic publication, the organization of international conferences and the efforts of such anglophone writers as A. J. Ayer and Susan Stebbing (in Britain) and W. V. O. Quine (in the USA). The publication of Ayer's Language, Truth, and Logic in 1936 was a particularly important moment in allowing positivist claims to become the focus of academic debate in Oxford, beginning to emerge even in the 1930s and certainly in the postwar decade as a hegemonic center of anglophone philosophy.Footnote ²²

Positivism won few converts in Oxford, but it very much set the terms of the conversation, in particular with its strident opposition to any body of discourse that went beyond the straitened horizons that positivism permitted.Footnote ²³ The general term for such discourse was “metaphysics,” and positivism was nothing if not the rejection of metaphysics. Among the things that counted as objectionably metaphysical were the entirety of such previously respectable pursuits as theology, aesthetics and ethics, and, by extension, the kind of normative political philosophy grounded in ethics.Footnote ²⁴ When T. D. Weldon, an Oxford philosopher, denied that philosophers were in a position to “improve either the theories or the methods of those who are engaged on scientific or political activities,” the implicit picture of philosophy bore the impress of several decades of positivist influence. There were only two kinds of sense to be spoken about politics: empirical and conceptual. The empirical investigations were best left to social scientists, which left philosophers to analyze the concepts (not, of course, to synthesize new ones).

My main claim is that we can make the best sense of Rawls's reception by seeing him, in the first instance, as a postwar thinker approaching basic questions in moral and political philosophy in the light of two sets of concerns. The first were secular and skeptical concerns that have a long history in Western intellectual history about the authority of moral and political claims. The second were the more immediate concerns of his philosophical (anglophone, analytic) contemporaries about the appropriate ambitions of philosophical reflection. Rawls, I will argue, employed, subverted and developed the styles and methods of his philosophical tradition into something that could—per impossibile—yield a “substantial” contribution to political thought that could also explain its own authority. Rawls shared with the analytic tradition a certain conception of clarity and strove for a prose unsullied by excursions into the high-flown, rhetorical or edifying. He also shared the methodological “naturalism” of its orientation in that he had a conception of what ethical, and normative political, questions were about, i.e. their subject matter, that offended against nothing in a fully scientific view of both human beings and their world. However, he addressed, and consciously bracketed off, the primarily semantic focus of the analytic tradition, with its emphasis on the analysis of language and concepts, and replaced it with a focus on epistemology, specifically on the notion of justification. In this, he was effectively returning philosophy to its pre-linguistic innocence, treating as fundamental not the Frege–Russell question (“How do words mean?”) but the older, Cartesian question (“What can we know?” “What ought we to believe?”)Footnote ²⁵

Rawls could treat such basic ethical questions as “What makes some action wrong?” as not calling for an account of the meaning or reference of “wrong,” but rather pointing towards a different question, “Why would I not be justified in doing this?” and connecting that question—about what might be called “normative” justification—with a more general question in what might be called “epistemic” justification: “What methods yield justified beliefs in this area?”Footnote ²⁶ He could answer his own questions in terms that at least addressed, and very possibly allayed, doubts about the authority and scope of those answers, at least as they pertained to the questions of moral and political philosophy. He could, in other words, explain who should take his proposals seriously and why. Finally, his sustained affiliation to the stylistic (if not doctrinal) strand of the analytic tradition in philosophy made his work an ideal exemplar of a certain kind of work that other philosophers had seen the need for but had not been able to carry out. Rawls actually did what many people by then agreed needed to be done.

What made Rawls's ability to answer these questions of interest to philosophers in the “analytic” tradition was the fact that philosophers in their tradition had allowed themselves to be backed into a corner by setting themselves an impossible task: to devise an account of ethical thought and language that would allow (some of) it to be true, and true in virtue of corresponding to some timeless ethical reality that was not put there by human beings. A succession of influential arguments, some summarized above, seemed to have shown that this could not be done. The idea of an ethical reality was at worst a fantasy, at best something very different from what it had seemed to be, and either way, not the sort of thing about which philosophical speculation was likely to uncover much that was both true and interesting. It was no surprise that the vision of philosophy and its ambitions in the domain of the ethical and political came out looking severely circumscribed. What Rawls seemed to have achieved was a restatement of the question, one that made ethics possible, scientifically respectable, potentially relevant to the sorts of questions that the politically active thinkers of the age were asking anyway, and able to claim rational authority for its conclusions. Moreover, he made a case for it as part of an inquiry that was recognizably philosophical, even to philosophers hostile to his kind of project, and yet continuous with inquiries in the (empirical) social and human sciences. In short, political philosophy was possible; it could be interesting; it could be done by philosophers, but not by philosophers working alone.

II. The death and rebirth of naturalism

Stuart Hampshire's review of A Theory of Justice concludes with a provocative historical claim, that the book is best seen “in the tradition of Adam Smith and Mill and Sidgwick: in the best tradition of British moralists, revived now at Harvard.”Footnote ²⁷ The reference to Harvard is probably deliberate, the point being to contrast Rawls's ambitions with an orthodoxy that had its home at Oxford. As Hampshire writes earlier in his review, “Twenty years ago, in the heyday of what came to be called Oxford philosophy, and while Wittgenstein's later philosophy was becoming widely known, ethics was still considered an intellectually barren subject, in which no major discoveries were to be expected and no new lines of inquiry could be opened.”Footnote ²⁸ What had come between the tradition of British moralists and Rawls was, Hampshire thought, the revival of an ancient form of philosophical skepticism about morality (recalling, for instance, the “immoralist” figures in Plato's dialogues), and therefore about the moral side of politics. Hampshire continues (and it is worth quoting him in full):

Rawls, by contrast, from his earliest papers on ethics onward, “represented a return to the classical, nonskeptical tradition, and yet [his] method of argument could not be criticized as lacking logical rigor and precaution.”Footnote ³⁰ He was, then, “one of the few analytical philosophers who, by the example of their inquiries, have effectively undermined this skeptical orthodoxy.” Hampshire, certainly, seemed to treat this return to the “classical, nonskeptical” tradition—and the marriage of its substantive (rather than merely conceptual or formal or clarificatory) ambitions with the vaunted rigor of analytic methods—with relief rather than consternation. Political philosophy could, Rawls seemed to be showing, be robustly ambitious without falling prey to the vices Weldon and others had identified, i.e. by concerning itself with questions and hypotheses that were not empirically testable and therefore (as they saw it) nonsensical. And on this point, the conditions were ripe for such a reception, even within the Oxford tradition that Rawls was seen to be repudiating. Hampshire's potted history was conventional wisdom. A slightly less potted version of the story was told, concisely and polemically, in Geoffrey Warnock's brief monograph from 1967, Contemporary Moral Philosophy.Footnote ³¹

Warnock's introduction set out the state of things in the late 1960s with an influential clarity. The background to the twentieth-century British tradition of moral philosophy was the received utilitarianism of Bentham and Mill, theories whose basic metaphysical commitments were what might be called “naturalistic.” Bentham's formulation of his hedonism begins, indeed, with an invocation of nature, who has—in his famous phrase—“placed mankind under the governance of two sovereign masters, pain and pleasure”; in Mill, the naturalism takes the form of a thoroughgoing “psychologism” about a variety of phenomena, including (to pick the most controversial) mathematics. The psychologism in Mill was the subject of sustained critique in what would come to be thought the founding text of the “analytic” tradition, Gottlob Frege's Foundations of Arithmetic (1884); in Britain, the founding text of twentieth-century moral philosophy was, as the later tradition came to have it, G. E. Moore's Principia Ethica (1903), which convicted the nineteenth-century utilitarians of what he (influentially, if confusingly) termed the “naturalistic fallacy.”

The charge of a fallacy has been famously difficult to pin down.Footnote ³² Warnock thought the charge to be twofold. First, one erred in trying to define the undefinable. Second, one erred in trying to define something nonnatural in terms of the natural. Of these, the second seemed to be the problem that deserved the label—as is involved, for instance, in defining the putatively nonnatural property of goodness with the natural property of being conducive to pleasure. But even here, it was found unclear just what the charge was. “The trouble is that [Moore] does not satisfactorily explain the terms ‘natural’ and ‘non-natural,’ or seek to show why qualities of the one kind are not definable in terms of those of the other; so that there is really nothing here for critical discussion to take hold of.”Footnote ³³

Warnock thought the consequences of Moore's view no less unfortunate. It “emptied moral theory of all content by making the whole topic undiscussably sui generis.”Footnote ³⁴ The idea of “nonnatural” qualities had no takers in the style of philosophy that evolved over the next two decades in Vienna. The “logical positivism” of the interwar period—to which Hampshire attributes the regnant skepticism about moral philosophy—could be seen as a way of developing the distinction between the natural and nonnatural in a semantic register, i.e. as a claim about the distinct meanings of two kinds of judgment, rather than as a claim about two different kinds of property or object. Judgments about “goodness” and “rightness” were, on the positivist conception, not to be “regarded as genuine judgments” at all. Genuine judgments had to have “statable meanings and discoverable grounds,” which moral judgments seemed not to have.Footnote ³⁵ What was left, then, was to give an account of what was distinctive about these judgments in terms of their effects (to evince emotion, to persuade others to share one's attitudes, etc.).Footnote ³⁶ The view was expressed with confidence and polemical passion in a chapter of A. J. Ayer's Language, Truth, and Logic, which brought what came to be called the “emotive” theory of ethics to interwar Britain.

A subsequent stage in the evolution of these views brought a different strand of positivist thinking on ethics to Britain. This was the “prescriptivist” view, suggested by remarks from such figures as Rudolf Carnap, but developed at length in R. M. Hare's The Language of Morals (1952). In this view, moral judgments are best understood as being, or being like, prescriptions, commands or imperatives—“Murder is wrong” meaning, roughly, “Do not murder.”Footnote ³⁷ I shall come back to Hare and the wartime period shortly; for now, what concerns us about his project is his concern to distinguish it both from Moore's style of “nonnaturalism” and from Ayer's “emotivism.” Hare took the naturalists’ mistake to consist in the fact that they “leave out the prescriptive or commendatory element in value-judgements, by seeking to make them derivable from statements of fact”—for instance, by deriving “Torture is wrong” from “Torture causes pain.”Footnote ³⁸ He drew on distinctions then being made, though mostly in informal conversation, by his charismatic Oxford colleague J. L. Austin, between—as Warnock summarized it—“what is sought to be achieved by issuing a moral utterance” and “what is actually done in issuing it—not what effect is aimed at, but what ‘speech-act’ is performed.”Footnote ³⁹

Warnock thought that Hare had a better account of the semantics of moral discourse than the positivists; among other things, it made for a better way of distinguishing between moral discourse and (say) manipulative propaganda.Footnote ⁴⁰ But the difficulties it raised had to do with its conception of the moral as a matter not of the distinctive content of moral judgments but of their form, roughly, that they can be universalized—a Kantian idea that Hare detached from any more extensive Kantian metaphysics. Warnock noted that this was far from being an orthodoxy by the time he was writing in the late 1960s, citing Philippa Foot—another Oxford colleague of Hare's—as foremost among the authors who rejected it, preferring an account of morality as distinguished by a concern with human well-being.Footnote ⁴¹

Foot's two papers from the late 1950s, “Moral Beliefs” and “Moral Arguments,” stated plainly that in insisting on thinking of moral judgments as distinguished by something about their content (rather than their form), she was indeed trying to reintroduce the despised naturalism of the nineteenth century as a serious philosophical contender. The opening paragraph of “Moral Beliefs,” presented at a meeting of the Aristotelian Society in 1958, puts the point in historical terms: “To many people it seems that the most notable advance in moral philosophy during the past 50 years or so has been the refutation of naturalism; and they are a little shocked that at this late date such an issue should be reopened.”Footnote ⁴² But Foot certainly did wish to reopen it, though it was an open question for her just what sort of naturalism was to be reinstated. To say that ethical facts are one set of natural facts among others was not automatically to affirm (say) the utilitarian view of which natural facts they were.

Foot was frequently seen at the time as allied with two other Oxford colleagues, also fellows at women's colleges: G. E. M. (“Elizabeth”) Anscombe and Iris Murdoch, who shared—as Murdoch summarized—an interest in “the reality that surrounds man—transcendent or whatever,” and a denial that “the human being was the monarch of the Universe, that he constructed his values from scratch.” Their shared naturalism was neither a semantic thesis about the meanings of moral statements nor a metaphysical thesis about the nature of moral properties. It seems to have been more of a methodological orientation: to see human beings and their ethical practices in the context of a more general “anthropology” of human life. They rejected, in other words, the view represented by Hare in his insistence on the formal characterization of moral principles with, in principle, no substantive restrictions on their content.

Rawls's first serious encounter with the energy and creativity of philosophy at postwar Oxford seems to have been in the person of J. O. Urmson, a fellow at Christ Church, Oxford, appointed to a visiting professorship at Princeton in the academic year 1950–51.Footnote ⁴³ Urmson, a few years away from publishing his well-regarded study of the emergence of what would come to be called “analytic” philosophy, told Rawls of events and figures at Oxford.Footnote ⁴⁴ Such philosophers as Gilbert Ryle, J. L. Austin, Isaiah Berlin, H. P. (Paul) Grice, P. F. (Peter) Strawson were in the most productive and original stages of their careers– and also such figures, already mentioned, as H. L. A. Hart, Stuart Hampshire and Bernard Williams. It was Urmson who urged Rawls to apply for the Fulbright fellowship that brought him to Christ Church, Oxford the next year.

At Oxford, Rawls encountered the work of H. L. A. Hart—in the process of developing the ideas that would appear a few years later in that succinct classic of legal positivism The Concept of Law—and attended seminars run by Isaiah Berlin and Stuart Hampshire.Footnote ⁴⁵ Rawls was in Oxford at the time of the publication of R. M. Hare's The Language of Morals (1952) and Elizabeth Anscombe's facing translation of Wittgenstein's Philosophical Investigations (1953). Wittgenstein had already had some influence on Rawls through the teaching of Norman Malcolm, an American student of Wittgenstein's at Cambridge who had taught Rawls at Princeton as an undergraduate, and was part of the faculty at Cornell when Rawls was a visiting graduate student there in 1947–8.Footnote ⁴⁶ Malcolm was joined there by such figures as Max Black and Arthur Murphy, all of whom took an interest in the nature, justification and uses of induction as a philosophical method.Footnote ⁴⁷ An oft-quoted remark of Wittgenstein's spoke to the idea that justification had ultimate limits: “Once I have exhausted the justifications, I have reached bedrock, and my spade is turned. Then I am inclined to say: ‘This is simply what I do.’”Footnote ⁴⁸

The influence of Wittgenstein, or rather a certain strand of Wittgensteinian ideas, on Rawls has now been widely recognized, despite the apparent freedom of A Theory of Justice from any Wittgensteinian traces.Footnote ⁴⁹ What is worth stressing is the fact that the character of his influence on Rawls was distinct from its form in Oxford. At Oxford, Wittgenstein was received into a milieu that had already been developing—in the work of J. L. Austin and Gilbert Ryle even before the war—a view that found in language the principal theme and subject of philosophy. There was, it is true, a diversity of “linguistic” styles at Oxford and the one associated with Austin—somewhat ambiguously termed the “ordinary-language” school—certainly had its affinities with Wittgenstein, even if there had been little direct influence from Cambridge, except in the occasional publication by John Wisdom.Footnote ⁵⁰ Austin's method, at its most distinctive, consisted in the careful and pointed assembly of examples of distinctions marked in ordinary (English) speech as ways of showing some tempting theoretical construction to be overambitious or inattentive to features of the phenomena it sought to explain, or, presumptuously, supplant.Footnote ⁵¹ What this had in common with Wittgenstein was overshadowed by the distinctly different tone of the Oxonians, more deliberately jokey, ironic and ever anxious to deny any therapeutic or salvific aspirations.Footnote ⁵²

To the extent that the Oxford philosophers of the period sought to “do” things with Wittgenstein's ideas, it was to some, more purely linguistic, strands of those ideas rather than others that they turned. Their uses of Wittgenstein were not like Rawls's. They do not seem to have found in Wittgenstein the basis of an inductive, or for that matter communitarian, idea of justification based on the norms and practices of some actually existing community. The Oxonians’ attention to language came with a severe, and often rigorously enforced, circumscription of the ambitions of philosophy. Roughly: philosophy was analysis; the object of analysis was concepts and what Oxford philosophers liked to call the “logic” of concepts, i.e. the rules governing their interrelations and manifestations in language use.Footnote ⁵³ Philosophical critique could proceed by showing that some substantive thesis, for instance, one that asserts the existence of a subjective inner life that is in principle unknowable to anyone except the subject, is in violation of the logic of mental-state attributions, which require such attributions to be, in principle, public.Footnote ⁵⁴

But the question raised by these forms of argument was this: what was the authority of logic as thus conceived? What compelled the acceptance of such principles? Where were they grounded? The question was especially pressing in the realm of ethics. And the slow return to prominence of naturalistic views of ethics suggested a happy rapprochement: the logic of moral concepts could be understood in naturalistic, psychological terms, as one set of facts among others about human beings.Footnote ⁵⁵ Rawls had already reached such a view before he arrived at Oxford. Lecturing at Princeton on emotivism, he had said, “The only way to refute the emotive theory, if it is stated carefully, is to show that there are invariantly acceptable ethical criteria: hence reasoning is possible.” He continued, “We shall look at ourselves. We shall ask what are the criteria which we use.”Footnote ⁵⁶ Once we answer that question, he seemed to be saying, there was nothing further left to ask or say.

III. Justification and the longue durée

I have been stressing the extent to which A Theory of Justice is, among other things, a work about justification. The questions it takes up are partly conceptual (what is justification?), partly moral (how does one justify?), and partly political (what can serve to justify the exercise of coercive power by the institutions over people who don't all agree on what is good and right?). Rawls traced the motivations for his theory to the preoccupations he acquired after his wartime experience in the Pacific during the Second World War. His essay “On My Religion,” written in 1997 and published posthumously, tells the story in broad, almost mythic, strokes animated by a few striking details. A young man, earnest and devout (his undergraduate senior thesis at Princeton is on “the meaning of sin and faith”), goes to war. Out on the front, he confronts mendacity in the person of a Lutheran preacher saying to the American servicemen in his congregation “that God aimed our bullets at the Japanese while God protected us from theirs. I don't know why this made me so angry, but it certainly did.”Footnote ⁵⁷ The violence and randomness of war was concentrated in his memories in the death of a friend. One of those meaningless and quotidian contingencies had it that he, rather than Rawls, was sent on a certain mission that ended with him killed by a Japanese mortar shell. And as news of the German concentration camps began to filter through to soldiers in the field after the Allied army had begun to liberate them after the German surrender had Rawls wondering about the possibility– and presumably, the propriety– of prayer: how in good conscience to pray for oneself, one's family, one's country, to a God who “would not save millions of Jews from Hitler?”Footnote ⁵⁸

“To interpret history as expressing God's will,” he wrote, moving from the particular judgment to the general principle, a move characteristic of his later philosophic method, “God's will must accord with the most basic ideas of justice as we know them to be. For what else can the most basic justice be?”Footnote ⁵⁹ The shape of his most distinctive convictions about justice is foreshadowed here, no doubt by design. The account of a sudden loss of faith in light of clearly articulable concerns is too simple to be accepted as the simple autobiographical truth. But there are reasons to take it seriously, if not strictly literally. For one thing, even if it is a retrospective attempt to impose a semi-mythic narrative on what was in fact a messier experience of moral and religious transformation, it would be revealing as just that. Moreover, it is possible that it does genuinely capture a process, if simplified into its barest, almost parable-like, outlines.Footnote ⁶⁰

It isn't surprising to find a Christian previously drawn to Protestant theology now seeking a philosophy adequate to life in a world in which there may be no God but where there are still other human beings. To put it differently, there isn't all that long a distance to cover from a senior thesis titled “A Brief Inquiry into the Meaning of Sin and Faith” to a paper titled “Outline of a Decision Procedure for Ethics,” and it is the passage of twenty years and the ironing out of countless small details that separates that paper from A Theory of Justice. The pattern of turning from theology to ethics, from faith to justification, is extremely familiar in the history of philosophy. The most obvious, perhaps the original example, is the shift in emphasis at the center of Plato's Euthyphro. At that stage of the dialogue, Euthyphro's “voluntarist” thesis—as it would come to be called—that things are holy if and because they are loved by the gods, has been dismissed as much less plausible than the alternative, that the gods love holy things because they are (as it were, already) holy. Socrates breaks the awkward, aporetic, silence with a suggestion: holiness is surely only one part of that larger thing, justice, the part concerned with what we owe to the gods, as opposed to the part concerned with what we owe to fellow human beings.Footnote ⁶¹ For those interested in the patterns of the longue durée in Western intellectual history, this might well constitute the original moment of secularization and provide a structural model in which to understand all the others. Gods either add nothing to a secular picture of ethics, or they add the wrong sort of thing.Footnote ⁶²

To continue this broad-strokes history: an argument with a similar structure appears in Immanuel Kant, with the stress on the difficulty of reconciling the spirit of voluntarism, with its demand of obedience to an omnipotent god, with something internal to morality, viz. the need for the fundamental moral motivations being unrelated to, and independent of, the prudential concerns represented by the threat of punishment in the afterlife.Footnote ⁶³ Kant provides tests of moral principles logically independent of any theological postulates, which he takes to count in the principles’ favor. As read by Rawls, the novelty of Kant's moral philosophy lay not in its tests of generality and universalization, which was not original to him anyway, but in his idea of autonomy and his vision of a morality chosen by rational agents with a conception of themselves as free and equal legislators. “The force of the self's being equal is that the principles chosen must be acceptable to other selves. Since all are similarly free and rational, each must have an equal say in adopting the public principles of the ethical commonwealth.”Footnote ⁶⁴ The justificatory test for a moral principle is, in other words, not the will (or caprice) of the gods, but that of other people.Footnote ⁶⁵

The pattern of secularization, as I have been calling it, and its tendency to prompt a “justificatory turn,” recurs in Henry Sidgwick (there are almost as many index entries for him in A Theory of Justice as for Kant). Sidgwick too lost his faith, though this time as a result of the interrogations to which he was subjected in discussions with fellow members of the Cambridge Conversazione Society (“the Apostles”). Again, the consequence of unbelief was a turn towards questions of method: if things were not to be justified because God commanded or otherwise commended them, what “method of ethics” could one devise to justify them? The work that came of these reflections, The Methods of Ethics (whose first edition was published in 1874) was haunted by the ghost of a theistic voluntarism its author could no longer affirm.Footnote ⁶⁶

The turn from theology to secular ethics and politics, I have been suggesting, is unsurprising, part of a familiar and intelligible historical pattern. But approaching the basic questions of ethics in the light of this pattern can make the focus of analytic philosophers of the mid-century on questions of metaphysics (what sorts of facts have to exist for ethics to be rationally defensible?) and semantics (what is the meaning of a moral statement?) seem thoroughly puzzling. But this fact makes sense once we consider that the analytic project up to that point consisted in applying the putative insights of science and other parts of philosophy to ethics. In other words, in looking at ethics from the outside in. What seemed revolutionary in Rawls was his insistence on proceeding in precisely the opposite direction. Rawls did not ask whether and how ethics fit into a larger picture of philosophy and the universe. Rather, he started from within ethical thought and practice—and from within ordinary ideas of justification—and sought a conception of philosophy and the universe that could accommodate these practices. He proceeded, in short, from the inside out.

IV. The charge of subjectivism

The significance of Rawls's inversion can be understood more clearly in comparison to the most prominent nonnaturalist still standing, R. M. Hare.Footnote ⁶⁷ Hare, born in 1919, had first been drawn to philosophy from a concern with pacifism—which, after much reflection, he rejected.Footnote ⁶⁸ Among the Oxford philosophers of his generation, Hare had by far the worst war. Taken prisoner by the Japanese while serving in Singapore, he was one of thousands of forced laborers on the infamous “Death Railway” from Siam to Burma. The Language of Morals, his 1952 book—written a few years after his return to Oxford, the completion of his degree and his election to a fellowship at Balliol College—began with the firm stipulation: “Ethics, as I conceive it, is the logical study of the language of morals.”Footnote ⁶⁹

The Japanese prison camp where he had spent most of his war, he told Ved Mehta, the New Yorker journalist profiling him, in an account of his wartime experiences, was a place

Hare's use of the almost extra-social community of a prisoner-of-war camp as revealing the basic structure of ethical thought contrasts strongly with the emphasis that his exact contemporary Iris Murdoch placed on the “reality that surrounds man.” Morality, as Hare saw it, was not about the particularity of individuals—apt enough after an experience in which individuals had been reduced to rank and serial number. On the contrary, to think morally was to see oneself as just one person among others: a person's being that person was less important than his being a person.Footnote ⁷¹

Over the course of his career, Hare would come to argue that by attending closely to the logical structure of moral language, one could derive the correct theory of how to live and act. His theory amounted, in its later forms, to a kind of utilitarianism: to act so as to maximize the satisfied preferences of all agents.Footnote ⁷² And he thought it no objection to his argument that it made ethics into one kind of logic among others. That was, in fact, the point: if a theory had anything to offer the ordinary person, it was authoritative guidance about how to live, and what had higher, and more neutral, nonpartisan, secular, apolitical authority than logic?Footnote ⁷³ Hare's philosophical outlook, deductive rather than inductive, prescriptivist rather than naturalistic, was never likely to find Rawls a congenial figure.

When A Theory of Justice was first published, Hare was one of the small number of its reviewers to declare himself deeply unimpressed. The two-part review, published in 1973 in the Scottish journal Philosophical Quarterly, was generally marked by a frustration about Rawls's carrying on in blithe disregard of his—Hare's—firm strictures about what may and may not be done in philosophy. Curiously, it seemed barely aware of the fact that the book was a piece of political philosophy (the words “politics” and its cognates appeared only twice, and glancingly in both cases, in the review). Hare's attention throughout was on the moral and methodological basis of Rawls's theory.

The condescension of Hare's concluding paragraphs was striking:

To be clear, Hare was an outlier in his opinions even among his Oxford contemporaries (note the remarks from Hart, Hampshire and Williams quoted earlier). Of Hare's many objections to the non-stylistic qualities of Rawls's book, the most revealing is his repeated contention that Rawls's project was, despite his apparent protestations to the contrary, subjectivist. He complained that Rawls “is making the answer to the question ‘Am I right in what I say about moral questions?’ depend on the answer to the question ‘Do you, the reader, and I agree in what we say?’”Footnote ⁷⁵ Hare thought it an enormous mistake of Rawls's to “leave questions of meaning and definition aside and to get on with the task of developing a substantive theory of justice.”Footnote ⁷⁶ “There is,” Hare said, “in fact a vast hole in his 600-page book which should be occupied by a thorough account of the meanings of these [moral] words, which is the only thing that can establish the logical rules that govern moral argument.”Footnote ⁷⁷

Hare's concern here goes a long way down. He is certainly skeptical of appeals to “intuition,” but these lines suggest a more basic discomfort with inductive argument in moral philosophy.

The authority of his own system, if it had any, had been the authority of logic: something that allows for deductive arguments that proceeded from a combination of formal premises and empirical fact towards substantive conclusions. Rawls's method, by contrast, was in its turn skeptical that such formal premises could be found at all, a fortiori that they may be found in moral language. In the earlier parts of this article, I have tried to place Rawls in a tradition—expressed in a relatively programmatic, promissory form within Hare's Oxford milieu by such figures as Foot and Anscombe—of ethical naturalism. This point is more or less explicit even in the original edition of Theory:

The reference to “most classical British writers” here serves, among other things, to mark Hare as the outlier in his insistence on a more austere conception of moral philosophy. There is, of course, a simple response to arguments like Hare's, quite clearly stated in A Theory of Justice, namely that the project of the book is described in terms that explicitly and deliberately put aside the question whether “our” considered judgments are right or true. The point is, rather, that those judgments are ours, and that fact had to be relevant to the project of devising a theory for our ethical life.Footnote ⁷⁹ In other words, Hare's objection to the “subjectivism” he saw in Rawls's theory was based on a misunderstanding of the project, a misunderstanding connected to his neglect of its political subject matter.

As Rawls would later make explicit, more needed to be said, partly in historical, cultural and sociological terms, about who exactly “we” are: are we the set of all rational beings? Or are we some much smaller subset thereof (e.g. a community of people shaped by certain historical experiences—modernity, capitalism, the American War of Independence, and so forth).Footnote ⁸⁰ At what point exactly can we say, and not just from laziness, that our spade is turned; this is how we carry on? The answer evidently depends on who “we” are and how exactly it is that “we” do carry on.

I have contrasted, in this section, two theories, two approaches to moral and political philosophy, one conceived and the other motivated, under conditions of war—indeed, the same war against the same enemy. Both sought secular bases for ethics; both found inspiration in the “analytic” methods and techniques developed in the interwar decades. One of them, Hare's, embodies the hope that these methods might take us away from the merely contingent, the merely parochial, away from what is merely “ours” towards what is simply true. The other, Rawls's, more at ease with contingency, embodies the belief that the discovery of what is ours is achievement enough.

I started this paper with the question of what was found so exciting about A Theory of Justice and am now in a position to offer a more succinct answer. Rawls appeared to have found a way to unite formalism in style with naturalism in substance. His work was taking advantage of the possibilities, by then recognized even in Oxford, of a naturalistic metaethics rooted in facts about human beings. But it managed to make the elusive link between such a metaethical picture and the “first-order” questions that the analytic tradition had treated as lying outside the scope of philosophy. In Rawls, the analytic tradition's style of writing and presentation, with all its appearance of scientific detachment, turned out to be perfectly capable of addressing what was, by anyone's account, a traditional project of philosophy: the project of human self-understanding, and the use of reason to solve a human problem.Footnote ⁸¹