7—
Anti-Kantian Complaints Revisited

This inquiry began with a catalog of complaints commonly adduced against neo-Kantian enterprise in political theory concerning its aspirations toward moral neutralism, its treatment of political community, its deontological aspirations, and its naive assumptions about the ideological dimensions of political theory and argument. Now it is time to reassess these complaints in light of the arguments discussed in the last five chapters to see how effectively the alternative views we have been considering respond to the initial complaints. This reassessment is a prelude to my outlining an alternative view of political theory, termed critical naturalism, discussing its central moral and political implications.

I. The Rejection of Moral Neutralism

A major motivation behind the turn to context and history has been the conviction that the neo-Kantian aspiration toward moral neutrality is misplaced. The claim that neutralist projects invariably fail has been around for some time,^[1] but most of the authors we have been discussing want to argue in addition that these projects are pernicious. For Bloom and MacIntyre, although the aspiration toward neutralism is philosophically bankrupt, there is an anthropological sense in which it has triumphed, generating an "emotivist" or "subjectivist" culture in which no moral position can be defended because every position is presumed to be equally (in)valid. For MacIntyre, the triumph of neutralism has meant the victory of alienation through the destruction of communitarian practice, whereas for Bloom it has brought the supremacy of a mindless egalitarianism

[1] For my account, see Shapiro (1986: 200, 214, 240, 249–51, 256, 282–84, 291–301).

in terms of which morality cannot even be talked about because every moral view is equally discredited. Pocock's construction of a republican alternative also rests on a view of "the paradigm of liberalism" that is at best agnostic about purposes, narrowly preoccupied as it is "with what can be distributed, with things and rights." Indeed, it is just because liberalism's negative libertarian language appears to reduce politics to matters of law and right that "the republican or political conception of virtue exceeded the limits of [liberal] jurisprudence" (Pocock 1981b: 359, 357–58). Walzer's impatience with a moral philosophy like Thomas Nagel's, explicitly conducted from "no particular point of view" (Walzer 1987: 5), and his defense of a method that through its "connectedness" is explicitly partial and parochial reflect an analogous antipathy for the neutralist aspiration.

Such attacks on neutralism are often closely connected to the claim that the underlying philosophical defect of the neo-Kantian arguments is that they are foundational theories of a particular sort. For it is not just that we have been alienated from our substantive moral commitments by the doctrine of neutrality. Since the seventeenth century our expectation that all our commitments must be justified from first principles places the neutralist aspiration in a paradoxical position. On the one hand neutrality is defended on the grounds that there are multiple and competing conceptions of the good. On the other, at least since Hobbes, it has been an aspiration in the Western tradition of political theory that a political system that takes account of these differences must be scientifically derived from first principles, shown to operate in our objective interests. Yet all such projects have failed, so the argument goes, and as a result the felt need to justify neutrality from first principles compounds our moral alienation. On this view subjectivist moral theories have triumphed on the back of the relativizing influence of failed modernist expectations, undermining the possibility of the kinds of substantive moral conviction that allegedly prevailed in a premodern past.

If the aspiration toward neutrality is thought to be pernicious because it weakens our faith in all substantive moral values, the quest for apodictic foundations to it has had the arguably more serious consequence of undermining our social and political institutions. Thus Rorty's (1984a: 16) claim that it is failed attempts to defend liberal democracy as the objectively best system that have undermined our faith in it and MacIntyre's argument that the Enlightenment Project of justifying morality necessarily had to fail because once we came to believe that the purposes that gave moral deliberation its point had themselves to be morally justified, the setup for disenchantment with our institutions was inevitable. Bloom's

claim, embedded in the subtitle to his book, that it is democracy that has been failed by higher education, suggests an analogous view Although Bloom's precise diagnostic argument is difficult to pin down, he appears to think that there were beliefs and values, embodied in the texts of the tradition, that our predecessors accepted as valid in ways of which we have since become incapable. Pocock's concluding plea, in The Machiavellian Movement , for an eschewal of all moral absolutes, also suggests a rejection of the Philosophical expectations characteristic of the Enlightenment Project.^[2]

Now it is not essential to identify the preoccupation with foundational argument with modernity, and some contextualists—notably Walzer—do not embrace the view that contemporary foundationalism is especially pernicious. But Rorty, MacIntyre, and Bloom are clearly all committed to variants of the claim that there is something distinctively pernicious about the Enlightenment foundationalist project. My analysis revealed that as matters of both history and philosophy this view runs into serious difficulties. For Bloom and MacIntyre, I showed that the different historical stories they tell are misleading in their treatments of both past and present. Their accounts of cultural decline since the Enlightenment rest on pictures of a premodern era in which there was agreement about core purposive values that is today held to be missing, but neither author offers any evidence to support these contentions—relying instead on the blanket notion of traditional society as a residual category, indicative of moral cohesiveness. The concept of traditional society has long been under assault by historians, anthropologists, and sociologists, and it is doubtful that this idealization captures much that is accurate about any actual or historical human society.^[3]

My discussion of Bloom revealed that there has never been (nor should we expect) the consensus within the tradition that his account presupposes. Although MacIntyre concedes this (holding it to be a feature of every tradition that there be continual arguments over the meanings of its constituting terms), he fails to see that this undercuts his critique of modernity, which took as its point of departure the prevalence of irresolvable moral disagreement. As well as failing to make the case that there were not pervasive moral disagreements in the ancient and medieval worlds, he fails to explain how the contemporary disagreements that so trouble him—over abortion, affirmative action, distributive justice, and nuclear

[2] See Pocock (1975a: 552).

[3] For a powerful critique of this notion, see Shils (1981: 19, 287–330), who lays to rest romantic conceptions of traditional society in much sociology and anthropology. See also Rudolph and Rudolph (1967) and Gussfield (1975: 37–39).

war—could be resolved on his neo-Aristotelian view. Not only do these backward-looking critiques of modernity present a false view of a comparatively homogeneous past while overestimating the power of subjectivist views of morality today, but they mislead more seriously through inappropriate comparison. In both cases the basic comparison is between a premodern elite culture (be it aristocratic, religious, or scholastic) and contemporary mass culture. If Bloom was to compare instead the lives of medieval peasants, with all their attendant insecurities of war, economic vulnerability, and predictable early death, with the contemporary mass culture for which he has so much disdain, he might be less quick to discern decline or to be surprised that the contemporary masses do not read the Great Dead. After all, the masses of Renaissance and early modern Europe never read at all. Likewise with MacIntyre, if the condition of the many was compared in both cases, he might be less quick to conclude that the modern occidental masses are alienated in ways in which their predecessors were not. None of this amounts to a defense of contemporary liberal culture, of course, but it does suggest that some basic realism about the terms in which debate about it should be cast is lacking in these arguments. In short, the contentions that neutralism has triumphed in the anthropological sense overestimate the moral homogeneity of the past as well as the fragmentation of the present, generating specious arguments for simplistic linear depictions of modern decline.

Pocock deals with the problem of neutralism in the terminology of negative and positive liberty, assuming throughout that liberalism's core deficiency is its reduction of politics to instrumental questions about right and law and its inability, lacking the language, even to articulate questions about virtue, ends, or the good. Powerful as his account may be as a corrective to neo-Kantian political theory's explicit embracing of neutrality and to a Hartzian stereotype of the hegemony of interest group liberalism from the American beginning, we have seen that Pocock's view seriously misleads. As a philosophical matter it fails to take into account the truth that every theory of politics operates with conceptions of the right and of the good. The interesting questions are about their contents, and these are ignored via the mindless opposition of the gross concepts: liberal versus republican. Pocock's republic ranges ahistorically, we saw, over virtually every kind of political system, from ancient communities to modern national states, and over socioeconomic systems as different as ancient, feudal, and capitalist; it cannot be surprising that so little is said about the internal nature of the republic given the generality of this conceptual range.

There is, indeed, an irony to be grappled with by those who would embrace

Pocockian republicanism as an alternative to "the paradigm of liberalism" in virtue of the latter's chimerical aspiration toward a scientifically justified moral neutralism. For the central preoccupation with stability as the core organizing concern of republican theory in post-Renaissance humanist thought militates against moral argument and critical reflection about many political ends. In the history of republican theory after Machiavelli, on Pocock's telling, the basic problem of politics is irreducibly instrumental: to deal with the inevitable entropy of republics in a world of secular time by searching for sources of institutional stability. We saw how the virtues were continually redefined for this purpose and, indeed, that from the days of the Florentine debate at least, the question of what best conduces to republican stability occupied central attention. Whether it was Guicciardini's claim that aristocratic prudence was the best means of survival in an uncontrollable and hostile world or Machiavelli's argument that fortune should be subdued through decisive action and imperial expansion, the goal of staving off entropy, corruption, and decay remained supreme. This is why it was so important for Machiavelli to establish that imperial expansion did not necessarily invite instability, as had been the conventional republican wisdom. In its English variants, too, we saw that Harrington and his successors were centrally concerned to ensure stability by bringing political institutions into conformity with the social distribution of wealth. In America similar preoccupations were present from the start; much of the Federalist/anti-Federalist debate revolved around questions of what best conduces to institutional stability. Yet stability is an instrumental good: its value depends on what kind of society is being preserved or undermined, and in the usages Pocock describes it takes much for granted that we would find morally controversial. The treatments of outsiders and the elitist view of the role of the many, we saw, left much to be desired just because of the supremacy ascribed to internal stability. This instrumental treatment of substantive values is not a new form of moral neutralism, but it exhibits analogous defects. By keeping many substantive questions off the political agenda entirely or dealing with them only in instrumental terms, the outlook remains indifferent to these questions as moral questions. Whether political participation is inherently desirable, whether and on what basis exclusion of outsiders or territorial expansion can be justified, and whether a market-based system of production and exchange is desirable independently of its effects on stability are not questions that are addressed, or even presented, from the civic republican standpoint. Even the much-vaunted ideal of independence is defended, ultimately, on the grounds that only an independent citizen can contribute to the preservation of the republic. Indeed, the value of stability

is never itself explained. In a postfascist and post-Stalinist age, it is natural to inquire under what conditions stability should be thought to be desirable.

The difficulties involved in making and defending substantive political choices in the absence of a neutral or "god's eye-view" standpoint also infect the arguments of Rorty and Walzer. In Rorty's case we saw that he has no mechanism for dealing with conflicting or overlapping community values, with those people who are systematically excluded from or harmed by prevailing social practices, and he seems wholly innocent of the verity that dominant norms and practices play integral parts in the reproduction of power relations. Although Walzer is sensitive to these issues, he also fails, in the end, to explain how to choose among conflicting values and conflicting interpretations of the same values. Historicizing these choices as MacIntyre attempts to do also turned out not to be a solution because an account of the genesis and empirical particularities of a practice will not, ultimately, answer questions concerning the rules for its best organization or its desirability as a practice to begin with. This last, we saw, has been a difficulty for Aristotelianism from the beginning, yet MacIntyre fails to face up to and confront its implications.

Now, as Walzer argues so forcefully in Interpretation and Social Criticism , the defense of the contextualist view is not that it is superior as a philosophical system to the paths of discovery and invention of general normative principles but rather that it is inescapable. The paths of discovery and invention turn out on closer inspection not simply to fail in their own terms but to involve interpretation; like it or not we are judges all in Walzer's metaphorical sense. Yet in the last analysis this becomes a complex restatement of the problem rather than a solution to it, for to say that there are no neutral choices and that social criticism must inevitably be contextually informed is not to say what form it should take or to tell us how to choose among competing interpretations of the norms prevailing in our cultures, traditions, and practices. To take a recent illustration, when Walzer criticizes the Israeli policy of beating Palestinians on the West Bank and the Gaza strip in January 1988, arguing that negotiations should be attempted instead, partly on the grounds that this is "the democratic, humane, and,indeed Jewish way,"^[4] he is invoking an interpretation of Jewish values that many Jews would themselves reject on unimpeachable

[4] Letter to the editor of the New York Times from Irving Howe, Arthur Hertzberg, Henry Rosovsky, and Michael Walzer, Tuesday, January 26, 1988,p. A24, italics added. I am not suggesting that these are the only arguments these authors make against Israel's iron fist policy; several independent (and powerful) arguments are adduced.

biblical authority.^[5] Engaging in connected criticism, in short, does not tell the critic what to say.

A great contribution of the interpretivist view is its insistence that, as moral dilemmas arise in the situations of everyday life, they cannot be detached from those situations to be argued over from a neutral standpoint. The sheer complexity of social life, as Aristotle noted long ago, makes reference to the specific context essential to all practical moral argument.^[6] However, we cannot assume, as the old legal realists did and as Rorty and Walzer seem to suppose in different ways that we can, that close-to-the-ground analyses of political options will naturally reveal the one best course for all concerned. We know too much today about how little undisputed knowledge there can be in the human sciences to accept this sanguine pragmatism and contextual ethics, and we must take seriously the possibility that where people's natures and interests differ sufficiently, and alternative courses of action are open in a given situation, there may be no best or optimal solution.^[6] Understanding moral dilemmas in all their contextual complexity may often recast them for us in ways that we could not possibly have foreseen in the abstract, may reveal courses of action to have probable consequences that we might not have anticipated, and may unmask possibilities that we might not otherwise have considered, but we cannot assume that it will make moral and political choices for us. Although an understanding of contextual complexity is essential to good moral argument, it can never be all there is to moral argument. In the absence of neutral principles for choosing among alternative courses of action, some alternative that does more theoretical work than mere reference to contextual meaning has to be supplied.

II. The Appeal to Community

The idea of community has been invoked against the neo-Kantians for reasons of both substance and method. Methodologically the objection is to their atomistic individualism, which generally takes one of two forms. In the first formulation it is argued that their use of the contractarian

[5] To take one of hundreds of possible illustrations, see 1 Samuel, chapter 15, where the story is told of Saul's falling out of God's favor by failing to carry out his directive to destroy the Amalekites who had been Israel's enemies, "man, and woman, infant and suckling, ox and sheep, camel and ass." Because he had shown some compassion, sparing Agag and some of the animals, God sent Samuel to remove his kingship "because I feared the people, and obeyed their voice" (1 Sam. 15: 3–24).

[6] See Aristotle (1977: 199–200).

[7] For a useful critical discussion of the "situation sense" of the legal realists, see Ackerman (1983: 27, 73, 93–110).

tactic of imagining what institutions prepolitical people could, would, or should create wrongly assumes that the whole is not more than the sum of its parts. The resolutive-compositive method inaugurated into modern political theory by Hobbes, and appealed to in different ways by Rawls, Nozick, Buchanan and Tullock, Wolff, and others, treats all social relationships as incidental to the achievement of individual goals; indeed in this tradition the need for a state at all is frequently taken to arise only because of specific problems of "market, failure." The other form of the methodological complaint (often coupled with the first) is that because there is no natural or precontractural person, the neo-Kantians import into their construals of him characteristics of the societies in which they live, treating those characteristics as if they were natural.^[8] This complaint often generates pejorative charges of atomistic individualism from a different source. Critics of contemporary society who regard its foundation on market values as alienating, for instance, frequently criticize theorists who appeal to contractarian devices as mechanisms for respecting Kantian autonomy on the grounds that they attribute these values to human beings generally.^[9]

The substantive objections flow out of the methodological ones. Some critics of the neo-Kantians, such as Bloom (1975a: 648–62), take them at their word when they claim either to have no conception of the good or to be indifferent among competing rational conceptions and object to that.^[10] Most, however, go the route of Sandel in discerning that the Rawlsian-style thin theories may be thicker than their proponents admit but in not liking the particular density they discern. For Sandel, liberalism's commitment to the primacy of justice, and to the priority of right this brings with it, makes it fundamentally unsatisfying. Divorcing public from private lives, it recognizes only in the latter the constituting roles of our cultural attachments. The deontological liberal holds that "while we may be thickly-constituted selves in private, we must be wholly unencumbered in public. It is there that the primacy of justice prevails," and we become submerged in a circumstance that ceases to be ours." By placing the cultural

[8] This complaint, as old as the hills, is traceable—in its modern form at least—to Montesquieu's critique of Hobbes on the grounds that "is it not obvious that he attributes to mankind before the establishment of society what can happen but in consequence of this establishment?" (Montesquieu 1949: 4). See also Rousseau's attack on Hobbes for attributing to natural man "a multitude of passions which are the product of society" (Rousseau 1964: 129).

[9] For my account regarding Rawls and Nozick, see Shapiro (1986: 151—305), regarding Posner, see Shapiro (1987: 999–1047), and regarding Buchanan and Tullock and Wolff, see Shapiro (1989b).

[10] The discussion of the remainder of this section partly draws on, partly elaborates on, Shapiro (1989a: 60–66).

self beyond the reach of politics, liberalism "makes human agency an article of faith rather than an object of continuing attention and concern, a premise of politics rather than its precarious achievement" (Sandel 1982: 11–14, 183). Thus Sandel claims that extending the intimacies of our cultural commitments into an explicit theory of the political good would provide a more satisfying public community than the impoverished deontological vision.

Sandel's view—that justice belongs characteristically to the world of strangers reified in the deontological vision he rejects—both misdescribes the private communities we live in and rests on expectations for politics that cannot withstand analysis. First, it is false that the idea of justice does not operate in our private communal lives. In the family—that paradigm of the private community, held together in the ideal by bonds of intimacy and affection—justice and the sense of it play indispensable roles. The pertinent currency of family life is not voting rights or economic well-being but caring and affection. Although these are not properly distributed on the basis of merit, desert, efficiency, or Rawls's difference principle, they have their own systems of equities that, when violated, generate powerfully felt injustice and sometimes conflict. The child who knows herself to be loved or respected less than her siblings or the wife who has been abused will certainly bring to bear notions of injustice because they will feel the pertinent economy of love and affection has been violated. The idea that existing private communities provide a yardstick for a kind of postpolitical politics assumes a benign view of those communities that is not defended by Sandel or by any of the advocates of community discussed earlier in this book. Just as there are some happy and stable families blessed by abundance, good health, and fortune, many are not this lucky, and the language of justice comes into play whenever there is scarcity of, and conflict over, the goods internal to the relevant community.

What Sandel understands to be the limits of justice are actually limits, to the politicization of communities. By declaring a community to be beyond politics, part of the private sphere, we are rendering it immune from political criticism. So when Nozick (1974: 32) implies, for instance, that sexual leering may use women in ways that violate the categorical imperative but that this is not a political form of using, he is assuming a comparatively narrow definition of politics.^[11] Yet the accepted boundaries of politics are constantly shifting as the result of political struggles. When the law changes from denying the possibility of marital rape by conclusive presumption to creating such a crime by statute, a significant movement

[11] On leering as a form of harassment, see di Leonardo (1981: 51–57).

of this kind has occurred. The public/private dichotomy in this and other prevalent formulations misses such poignant complexities.^[12] Those who invoke the idea of justice in traditionally private communities like the family are generally those on the short end of power or distributive relationships. The general rule (to which there are doubtless exceptions) is that the dominated try to politicize to delegitimize, whereas the dominators try to depoliticize to legitimize.^[13] It is mere romanticism to suppose that the relations of scarcity, power, and domination that render battles over politicization necessary either do not exist in the private sphere of the present or would not exist if we could return to some idyllic public sphere in our collective past.

This is not to imply, with Foucault, that just because all social structures—historical, actual, and possible—involve relations of power, domination, and scarcity that they are all alike. Some may be preferred to others on various grounds, and arguments about those grounds should be among our central concerns. But such arguments cannot even be engaged in, there is no linguistic space for them, as long as the terms of debate are allowed to be a variant of "whether-or-not-community." Thus we have seen that moral appeals to the primacy of community and republic are consistent with every politics from far left to far right, from rural to urban, from ancient to feudal to modern capitalist and socialist, and from city-state to continental nation. Yet the communitarian appeal is not only vacuous, it is pernicious; it creates the misimpression that a second-order appeal to community will actually resolve moral dilemmas when it will not. This is so not simply because communities differ so much from one another but because, as we saw in our discussions of Rorty, Walzer, and MacIntyre, pervasive disagreements over the interpretations of norms within communities leave most politically interesting disputes unresolved. Even when there is consensus, we saw, there may be circumstances where this is more troubling than its absence. In short, the second-order appeal to community is no more plausible as a basis for critical standards

[12] On the changing law of marital rape, see note, "To Have and to Hold: The Marital Rape Exemption and the Fourteenth Amendment," Harvard Law Review 99 (1986): 1255–73. Problems relating to changing conceptions of the limits of politics are usefully taken up in Connolly (1981: 63–89).

[13] The great exception to this might be argued to be fascist and totalitarian regimes, which often depend on highly politicized ideologies and the continuous mobilization of populations in their legitimating myths. See Friedrich and Brzezinski (1956: 17–27, 85–115). But I would argue that if such regimes endure for any length of time, they invariably try to depoliticize and thus normalize political life. One recent, if anecdotal, illustration is the decision in the Soviet Union to dismantle the huge political slogans that are among the best-known features of the skylines of its cities. See the New York Times , May 21, 1988, pp. 1, 4.

than is the second-order appeal to autonomy, which latter is characteristic of the neo-Kantian arguments that the communitarians reject.^[14]

More romantic still is the attempt by other communitarians to base the postpolitical community not on some existing or allegedly historical community but on a community yet to be created. In these formulations the necessity for a system of justice can be abolished only when the causes of present conflict and scarcity go with it. This was Marx's view: that a communist society would differ from all its predecessors in that a superabundance of wealth would obviate the conflict generated by scarcity. Goods would be distributed on the basis of need, and government displaced by mere administration. It is often thought that such a view could be persuasive if a theory of needs could be developed that distinguished them from wants; wants may be infinite, as the bourgeois economists argued, and scarcity with respect to them therefore inevitable, but needs are not. A well-developed theory of needs could provide an archimedean point for limiting the induced wants of the market, making an economy of needs not subject to the limitations of scarcity at least a possibility. Yet these formulations assume a static and unrealistic view of human needs.^[15] Such lifesaving technologies as dialysis machines and artificial hearts satisfy

[14] For elaboration, see Shapiro (1986: 273–305).

[15] At times Marx embraced a more dynamic view of needs, as in the argument in Capital that what counts as subsistence is socially and historically conditioned. See Marx (1974, I: 164–72). But he never came to terms with the difficulties this raised for his account of communism as a permanent state of superabundance. One contemporary Marxist, G. A. Cohen, indicates awareness of the problem when he says that Marxists' confrontation with liberalism "is avoidable only as long as Marxists continue to maintain that abundance will ensure complete compatibility among the interests of differently endowed people, and abundance on the required scale now seems unattainable. A lesser abundance, which enables resolutions of conflicts of interests without coercion, may well be possible" (Cohen 1986: 117). But he offers no account of how this might be done, and in his more recent work advocating equalization of "access to advantage" he supplies no account of how and by whom different disadvantages are to be weighted in the event (presumably often unavoidable once endemic scarcity is conceded) that equalizing access along some dimension diminishes it along a different one. See Cohen (1989a, 1989b). On these issues, see Rae et al. (1986: 104–29). Even advocates of needs as instruments of social policy-making like David Braybrooke concede that the concept breaks down in the face of demands for medical resources (among other areas), although the difficulty is more serious for his theory than Braybrooke imagines. Braybrooke (1987: 295) perceives conflicts between the need for medical care and other needs and notes that in such circumstances choices will have to be made that bring to bear considerations other than those of meeting needs. But he fails to notice that even within the realm of medical care needs must invariably outstrip resources, so that the problem is not only one of deciding how different goods should be ranked, but of deciding how to allocate particular goods to competing claimants whose conflict with respect to the relevant good is zero-sum.

needs, not wants, on just about any credible definition, yet the potential for such innovation is limitless. Once technological change is taken into account, human needs are infinite, scarcity and concomitant conflict inevitable, and the languages of politics and distributive justice inescapable. It is sheer fancy to suppose that there are any possible conditions of human association in which the need for a system of justice will evaporate, given these facts about the human condition. The limits of justice have nothing to do with liberalism, contra the implication of the title of Sandel's book and the assumptions of many Marxists, for political conflict is endemic to all human social interaction.^[16]

In sum, to say that all morality is teleological does not begin to establish that there may not be different and conflicting goods for different persons in the same, different, and overlapping communities, as we saw in our discussions of Rorty and Walzer. For MacIntyre, too, once Aristotle's assumptions about natural harmony are jettisoned, there is no reason to believe that the imperatives of different subordinate practices can be subsumed under the telos of a unified human life for an individual, let alone for an entire society or world. Although we can empathize with MacIntyre's disillusionment with contemporary liberalism and Marxism and perhaps even with the quest to re-create ancient communities that it prompted, he has not begun to indicate why they would be superior forms of political communities, let alone how they might be realized in today's world.

The treatment of community in Pocock's analysis raises different difficulties. In addition to the instrumental preoccupation with stability, our examination of the history of republican argument revealed the lack of an adequate account of boundaries or membership, which is a serious defect because communities are always mechanisms of exclusion as well as mechanisms of inclusion. Inherent in all these views, then, is a benign, if not outright romantic, view of communities and the values they reproduce, whether this be of communities that prevail in the contemporary world, those that are imagined to have existed in some premodern past, or those that are to be created in some transformative future.^[17] There is often a troubling indifference to the impact of the community on other communities and individuals except to the extent that such impact becomes instrumental to its stability and an almost studied blindness to the ways

[16] Many contemporary Marxists have eschewed the utopian ambitions of Marx's politics. See, for example, Sirianni (1981: 33–82), Laclau and Mouffe (1985), and Isaac (1987a: chapter 6). The question for these authors has to be whether there is anything distinctively Marxist about their politics once the classical Marxian telos is abandoned.

[17] Walzer is an exception to this last charge. His account of dominance is discussed further in section IV of this chapter.

in which the values dominant in a community are instrumental in the reproduction of relations of domination and control.^[18]

III. The Rejection of Deontological Aspirations

A third dimension of disaffection with the neo-Kantians derives from skepticism toward the attempts of Rawls and his successors to defend principles of political and economic organization while bracketing considerations Kant dubbed anthropological . Rawls's philosophical project can be made sense of only in terms of his contrast between deontological and teleological theory, the point of which was to distinguish purposive views of ethics that depend on particular conceptions of the good from theories that claim agnosticism about them. If a theory of justice could be shown not to rely on any particular conception of the good but to be neutral among competing rational conceptions, the procedure of defending it without reference to consequentialist considerations, indeed without reference to causal matters of any kind, might appear plausible. Thus in teleological theories "the good is defined independently from the right, and the right is then defined as that which maximizes the good." By contrast, a deontological argument like Rawls's "either does not specify the good independently from the right, or does not interpret the right as maximizing the good" (Rawls 1971: 24, 30). Deontological arguments are attractive to neo-Kantians because they appear uniquely consistent with the preservation of individual autonomy, and teleological reasoning in political theory threatens them because it appears to undermine that autonomy. Doctrines like Sidgwick's utilitarianism permit sacrificing the interests of some individual to maximize the general good on consequentialist grounds, a result Rawls is determined to avoid.^[19] Likewise with Nozick's rejection of patterned or end-state consequentialist principles of justice in favor of procedural or historical ones, the central claim is that there cannot be genuine liberty in society unless it is liberty to undermine, through voluntary transaction, whatever distributive patterns happen to prevail.^[20] Analogous in spirit is Dworkin's attempt categorially to distinguish questions of principle from those of policy, even allocating these to different branches of government.^[21] This aspiration toward working out principles of justice in anticonsequentialist terms in the realm of "ideal theory" and

[18] For one useful discussion of how dominant conceptions of virtue in revolutionary America functioned to repress women, see Bloch (1987: 37–58). On the generally hierarchical and often repressive character of early American Puritan communities, see Haskins (1960: 9–93).

[19] See Rawls (1971: 22–33, 179–92).

[20] See Nozick (1971: 161–64).

[21] See Dworkin (1985: 81–89).

then applying them to particular situations to see how they measure up gives the neo-Kantian enterprise its deductivist and transcendental whiff, even though its proponents have never actually adduced transcendental arguments in support of their views. The aspiration is revealingly summed up by Rawls (1971: 121) when he remarks that the eventual goal (that he does not yet claim to have attained) is to develop a political theory that is "strictly deductive," a kind of "moral geometry."^[22]

Skepticism toward ideal theory of this sort is as old as political philosophy, figuring centrally, for example, in Aristotle's insistence that human social life is too complex and unpredictable for us ever to suppose that principles of justice might adequately be legislated for all circumstances.^[23] For this reason he introduced the concept of equity in terms of which the law must be modified as applied to unforeseen circumstances and problems, a little like modern notions of common law reasoning. But where for Aristotle equity was limited either to effectuating what would have been a legislator's intent under changed conditions or to bringing the law into conformity with the requirements of "absolute justice" (dictated in turn by the requirements of the prevailing regime in its uncorrupted form), for most contemporary anti-Kantians all critical argument is exclusively immanent.^[24] There is no architectonic structure of ideal justice; all we have is the system of norms and categories that make up our contemporary reality. Thus Rorty invokes Wittgenstein's notion of a language game to argue against all metanarratives. Without explicitly appealing to linguistic behaviorism, Walzer and MacIntyre both argue in different ways for the inevitability of immanent criticism. For Walzer it is not simply his desire to remain "in the cave, on the ground" to develop connected criticism that can have some possibility of influencing those to whom it is directed—its very terms must be derived from the values accepted in a culture. Distributions of goods can thus be just or unjust only "relative to the social meanings of the goods at stake" (Walzer 1983: 9). Prevailing

[22] For further discussion, see Shapiro (1986: 151–305; 1989a: 53–55).

[23] "Our account of this science [politics] will be adequate if it achieves such clarity as the subject-matter allows; for the same degree of precision is not to be expected in all discussions, any more than in all the products of handicraft" (Aristotle 1977: 64–65). On Aristotle's hostility to abstract moral argument, see (ibid.: 69, 93, 96, 103, 111, 189–90, 199).

[24] The discussion of equity and its relation to absolute justice can be found in Book Five, sections vii and x of the Nicomachean Ethics (Aristotle 1977: 189–90, 199–200), and that of the reform of regimes in Book Three, chapters 6, 9, 11, and 12 of The Politics (Aristotle 1984: 94–95, 97–104). I concede that there are other credible interpretations of Aristotle's understanding of absolute justice (linking it to a higher law theory), but I do not take up the burden of defending my interpretation against these here because the point made in the text is consistent with these stronger interpretations as well as with mine.

usage defines both descriptive meaning and normative ideal. For MacIntyre, too, practices are ongoing, rule-governed activities, typically outliving any given generation of participants; it is just because we want to excel in terms of these ongoing practices that we inevitably accept their constituting rules, even if we modify them as we employ them.

Although Bloom would be the last person explicitly to embrace contextual criticism, we saw that the critical standards he invokes rest on no more than an act of faith (however implausible) that there is consensus in the tradition on the ultimate questions of political morality. The irony is that his moral absolutism rests, ultimately, on no more than an appeal to an idealized description of an intellectual context. In the republican tradition as portrayed by Pocock, too, we have long since departed the ideal republic; the realm of prudence is one of cautious adjustment to a dangerous and often unfathomable world, and even projects of Machiavellian mastery must eventually succumb to the corrosive effects of time. There are no moral absolutes, and genuine freedom lies in acceptance of this fact. "There is a freedom to decline moral absolutes," as Pocock sums up a moral of his study, "even those of the polis and history, even that of freedom when proposed as an absolute" (Pocock 1975: 552).

For many of these authors, then, the critique of transcendentalism, like that of neutrality, flows from a basic disaffection with the foundational argument characteristic of the Enlightenment.^[25] This may take a linguistic form, as it does explicitly for Rorty and implicitly for Walzer and MacIntyre, so that we are locked, in Frederick Jameson's phrase, in "the prison house of language." Or it can be an historical argument, as part of Rorty's and MacIntyre's arguments are and as Pocock's is exclusively. In either case the abandonment of Kantian aspirations turns into that of all foundational argument, and we must now sort out the conflation of arguments that brings about this metamorphosis. For although the critiques of both transcendentalism and deontology are well taken, no good argument has ever been made against foundational argument as such.

Let us begin with the assault on transcendentalism. In most formulations considered in this book, it is not implausible as a theoretical argument but is a red herring because none of the authors commonly dubbed neo-Kantian relies on transcendental arguments. Yet by reacting against the transcendentalist whiff of the neo-Kantian arguments rather than what is actually wrong with them, the new contextualists replicate some

[25] Bloom is clearly an exception here. Although there is the irony just alluded to deriving from his embracing of the tradition as a source of critical standards, he would disassociate himself from the rejection of foundationalism in any of its forms—even if he has failed to indicate even in broadest outline what a good foundational argument would be.

of their most serious defects. The central weakness of the neo-Kantian arguments is not to be found in the attempt to reason about principles without reference to ordinary usage. As Rawls, Dworkin, Ackerman, and their defenders never tire of reminding their critics, prevailing beliefs and practices are integral to the procedures of justification they employ. To be sure, there are legitimate questions that I and others have raised concerning whether Rawls's argument from the original position can be persuasive without assuming the truth of the Kantian interpretation of his principles, whether Dworkin's recent move toward contextual justification in Law's Empire is consistent with his steadfast rejection of it in earlier writings, and whether Ackerman's dialogic method actually plays any philosophical role in the derivation of his principles of social justice. There are further questions about why, if writers like Rawls and Walzer are both claiming to derive their principles from beliefs prevailing in American culture, they reach such different conclusions. But the point to notice here is that the move toward contextual justification need not, and typically does not, dictate a move away from deontological theory in the Rawlsian sense of being justifiable or unjustifiable independently of consequentialist considerations.

I do not mean to suggest that the anti-Kantians do not employ reasoning that Rawls would describe as teleological; they do. They openly commit to the supremacy of conceptions of the good in their accounts of right and justice. But we must not be diverted by a second red herring that lurks in the deontological/teleological dichotomy itself. At the heart of what is implausible in the neo-Kantian arguments is the inability of their authors to see that the distinction between deontological and teleological theory cannot be sustained. Their failure to see, or to concede, that consequentialist considerations invariably enter political theory give their arguments a quality of preaching to the converted. Only readers persuaded in advance of the truth or desirability of their particular assumptions about human nature, about what is good and bad for human beings, and about the causal structure of human interaction will be impressed by their claims that their theories, if implemented, would generate just societies that preserve Kantian autonomy. In effect this means that they rely on undefended and submerged assumptions about human nature and causal theories of how the world works to justify their principles, either naively believing or cynically pretending that these theories were not deeply contested and ideologically charged.^[26] Consequentialist argument was engaged

[26] Here are a few examples: Rawls assumed general laws of psychology and economics to be the neoclassical assumptions about risk aversion and economic rationality and a host of Keynesian macroeconomic assumptions to render his account of chain connection (and hence his difference principle) plausible. Nozick assumed without argument that market-based appropriation works to the benefit of all and that the only natural monopoly is the coercive force exercised by his minimal state. These and related issues are taken up at length in Shapiro (1986: 151–270, 289–92).

in, as it inevitably must be in arguments over the design of political institutions, but little serious attention was given to making it plausible. Rather than dealing with these issues by assumption and methodological sleight of hand, theorists who want to persuade anyone other than those with whom they already agree should get their hands dirty in empirical controversy and argue for the particular assumptions about individuals and society that they seek to advance.^[27]

What is so frustrating about much of the turn away from neo-Kantian theory is that for all the appeal to historical specificity by which it is often motivated, there has been little serious grappling with empirical complexity and causal argument. In many of the accounts we examined, the neo-Kantians' second-order commitment to an abstraction called autonomy was simply displaced by a second-order commitment to an abstraction called community . Both in diagnosis of our circumstances and in prescriptive argument, we discovered that in different ways Rorty, MacIntyre, and Bloom all pay scant attention to the empirical credibility of large parts of their arguments. Their causal diagnoses rest on undefended, and prima facie quite implausible, historical arguments about how a few philosophical texts created the problems of modernity, indeed created modernity itself, arguments that can only be stated through tacit reliance on undefended and outmoded notions of traditional society. This causal idealism creates the dangerous misimpression, most explicit in Rorty's account, that the problems they identify could be wished away by undoing the intellectual mistakes embodied in those texts. In their explicit normative claims the arguments of Bloom, MacIntyre, and Sandel were all seen to rest on undefended, and again less than plausible, counterfactuals about cohesive social and political communities that either once existed or would exist but for liberalism's alienating tyranny. Virtually nothing was said about the nature of these communities, about how they did or could work, about how conflicts over core values within them would be resolved (when it was conceded that such conflict could occur), or about what economic and political institutions, forms of ownership, and systems of obligation would prevail. The republican tradition described by Pocock doffs a genuflecting cap to the importance of a "sociology of liberty," but we saw that

[27] For elaboration, see Shapiro (1986: 241–42, 289–92).

the way in which this tradition evolved, on his telling, made this sociology both undetermined and naive—undetermined in that its core terms were easily appropriated by all sides in the great debates of the past several centuries over the structure of economic and political institutions and naive in that there was generally insufficient awareness that, far from ensuring the republican ideal of independence, a regime of private property and market institutions could and would be deeply threatening to it.

Walzer alone among the communitarians has tried to lay out the substance of a theory of the good and give illustrations of how it would operate in practice and of how different goods would be allocated in different realms of social life. Even in his case, however, we found no answers to the questions of how conflicting interpretations of shared meanings would be settled, of how, a market system could be accepted while its globalizing internal dynamic could be "walled in," and of what basic changes in the underlying political economy might actually reduce dominance rather than merely redistribute its incidence.

Why do so many authors who have rejected the neo-Kantian enterprise pay so little attention to contextual complexity and causal practicality, given the terms of their critiques? A large part of the answer is that by casting the turn away from neo-Kantian theory as a rejection of foundationalism as such, many of the contextualists wrongly think that they need no longer articulate, let alone argue for, their foundational commitments. We should instantly be suspicious of the claim that "contextualism" is a meaningful alternative to "foundationalism" when we find one of its most lucid defenders arguing that he wants to "clear the ground" to defend the antifoundationalist view.^[28] Just as our inability to generate one type of physical foundation that will support every kind of building on every possible terrain does not mean that builders and architects can henceforth build without foundations, just as the absence of a single skeletal structure common to all mammals does not mean that we could suddenly decide to do without skeletons, so not finding a secure basis for all knowledge in deductive introspection does not mean that henceforth we could—even if we chose to—ignore foundational questions in political theory and philosophy. The failure of that peculiar amalgam of transcendental ambitions and deductive arguments that preoccupied the central figures of the Enlightenment may well by now be undeniable, but to think that in discarding it we can now get on with political theory without foundations is simply false. Yet it is the belief that we can, more or less articulated in different cases, that enables many of the authors we have been

[28] See Herzog (1985: 27).

examining to believe that mere appeal to a postulated consensus is sufficient to justify core political commitments.

This endorsing of consensus generates a curious convergence between those who embrace and those who reject the neo-Kantian enterprise in political theory. Labels to the contrary notwithstanding, where the likes of Rawls and Nozick argue that their principles of justice are legitimate because they would be freely chosen by people like us under certain specified conditions, so theorists like Rorty, Walzer, MacIntyre, and Sandel all seem to think that the values they defend, and the institutions that allegedly flow from them, command, commanded, or would—under the right conditions—command agreement. Writers like MacIntyre and Walzer, who are aware that requiring this kind of agreement is a tall order and explicitly disclaim it, end up appealing, we saw, to variants of just this claim in their positive contentions. In no case is evidence for the assumed consensus on the interpretation of prevailing values actually adduced, and we saw that there are good reasons to doubt that it does or could exist. Yet the various appeals to consensus that grow out of the critique of foundationalism appear to render plausible the practice of not defending particular foundational commitments with either argument or evidence. This plausibility comes at the price of the lost interest of those who do not agree with the conclusions in advance. For, as by now should be obvious, the view that foundational questions can be ignored is in fact a variant of the claim that they can be settled.

We can agree with the pragmatist critique of foundationalism if we interpret it to mean that there is no isolated or isolable discipline of epistemology that, once the philosophical specialists have gotten it all straight, so to speak, will generate secure foundations of knowledge for all time. But where we are bound to part company with Rorty, though not with Dewey or Quine, is in recognizing that this does not entail that foundational questions can ever be ignored, abandoned, or otherwise regarded as settled. Indeed the mere statement of the thesis that metanarratives should be abandoned indicates that people like Rorty and Jean François Lyotard (1984: 31–41) are co-opted by the thesis they are attacking; it implies that first- and second-order questions can be neatly distinguished into separate classes, that we can specify what we are going to abandon. It is quite contrary to the spirit of pragmatism to assume this or even to assume that our most basic commitments are epistemological. It makes more sense to begin with the intuition that everything is connected to everything else and that epistemological or any other questions may well be central to many first-order debates about morals, politics, or anything else, though this is not necessarily so, of course, because nothing is necessarily

so. Dewey did not advocate abandoning philosophy as Rorty does; rather he took this different and more subtle view: "Better it is for philosophy to err in active participation in the living struggles and issues of its own age and times than to maintain an immune monastic impeccability" (Dewey 1929: I, iii). In this view we should think of philosophy, politics, and life as parts of an interconnected whole, and arguments, no matter where they start, must follow the connections wherever they lead. An argument that begins about whether or not a particular agent is unaware of her interests in a given situation might well become an argument about what it means to know what an interest is and then an argument about what knowledge is. There is no reason, in principle, why this should not be so; often, though not always, it will be exactly what is required. Rorty's resistance to this logic and its implications, we saw, is at the heart of his political and philosophical complacency.

I have sought to show in the last five chapters that in political theory arguments invariably lead protagonists to certain commitments about human nature and social interaction. Every view of politics rests on a conception of human psychology and a set of assumptions about the pertinent causal structure of the social world and about how that structure may facilitate and limit human purposes. The pragmatist tradition of Dewey and Quine offers perhaps the most viable third way between the neo-Kantians and their contextualist critics, but I argued in chapter 2 that it must rest ultimately on a thesis about human survival, not on linguistic consensus. More generally I argued that the only defensible reformulations of the anti-Kantian arguments makes a realist foundational commitment inevitable, that Rorty's Quinean account of how we adapt beliefs to changing circumstances requires exactly the philosophical realism that he claims to reject. Walzer makes us see that a credible analysis of contextually based critical standards must rest on a prior account of how they function in the reproduction of power relations. If he does not himself supply a developed account, he makes clear the imperative for causal argument that strives at least partly to escape, as it explains, the hermeneutic process. My analysis of MacIntyre's neo-Aristotelian account of human practices revealed a comparable need for a causal account of how they interact with one another. MacIntyre's account is also useful because his discussion of the irreducible narrative structure of human experience provides the beginnings of a psychology for political theory that is richer and more plausible than the utilitarian assumptions that inform most liberal and much Marxist political theory. MacIntyre, we saw, also provides one of the few powerful modern defenses of the view that moral argument is inevitably teleological, which in turn implies that any plausible political theory is bound

to articulate and defend a theory of the human good.^[29] Yet my discussions of the historical arguments of Bloom and MacIntyre revealed that the appeal to history does not obviate the need to defend one's assumptions about human nature and social causation; displacing internalist contextualist argument with internalist historical argument turns into exchanging one idealist demon for another. If our examination of Pocock's civic republican tradition generated the conclusion, contra Rawlsian and Hartzian conventional wisdoms, that every political theory embodies a conception of the human good, it also revealed that no political theory can be better than its causal account of the conditions for that good's realization.

IV. The Ideological Dimensions of Political Theory

The neo-Kantians have been subjected to a barrage of criticism concerning the ideological dimensions of their arguments—either for innocence concerning the ideological uses to which these arguments can be put or for more or less self-consciously dressing up ideologically motivated claims in spurious philosophical garb. Although appeals to context and history are frequently among the best tools for debunking ideologically loaded philosophical arguments, the harvest of the last five chapters in this regard has been notably poor. With the exception of Walzer not only do these authors fail to speak to the ideological dimensions of their own arguments, but they replicate much of the ideological innocence that we hoped to be leaving behind when we rejected the neo-Kantian enterprise.

Pocock's analysis of the republican tradition is disappointing from the

[29] Two different senses of the term teleological must be distinguished for present purposes. In contemporary political theory teleological arguments are contrasted by Rawls and others with deontological arguments, as already discussed in this section, the point being to distinguish purposive theories of ethics, which depend on particular conceptions of the good, from theories that claim agnosticism about particular purposes. Thus in teleological theories "the good is defined independently from the right, and the right is then defined as that which maximizes the good." By contrast, a deontological argument "either does not specify the good independently from the right, or does not interpret the right as maximizing the good" (Rawls 1971: 24, 30). This may be called the analytical sense of teleological. The historical sense of teleological assumes considerably more, paradigm cases being Hegel's teleological view of history and, on my interpretation, Marx's historical materialism. On these views there are purposes immanent in social processes that give history a direction and suggest that it will eventually reach its goal and come to an end. Historical teleologists are usually (perhaps even necessarily) analytical teleologists, but the converse does not hold. Thus, although I have argued that commitment to analytical teleology is inescapable, it will be clear from the argument in the next two chapters that I regard the historical teleological view as untenable.

standpoint of comprehending the relationship between ideology and political theory for reasons of both substance and method. Methodologically his analysis flowed out of the claims of the new Cambridge historians to be rereading the history of ideas as the history of ideologies, but in practice this turned out to be mere internal description of evolving republican arguments as subjectively comprehended by participants without reference to the roles played by these arguments in the changing socioeconomic practices that they were used to undermine, transform, or legitimate. The result is a history that, although an illuminating description in its own terms, tells us virtually nothing about the actual operation of republican ideology and lacks critical edge for its evaluation. Substantively, Pocock's much-trumpeted promise to supply us with a history of our intellectual culture that is not dominated by the problematic of "the rise of bourgeois modes of thought" turned out to be much ado about remarkably little, partly because his straw conception of the liberal paradigm generates a misleading opposition of republicanism to liberalism and partly because—protestations to the contrary notwithstanding—much of his account of the evolution of the republican argument turned out to be the story of its accommodation with capitalist practices and bourgeois social life.

Rorty's argument more than any other exemplifies the innocence involved in an internalist preoccupation with language. By uncritically endorsing the values he takes to be enshrined in our culture, he becomes inexorably wedded to a parochial conservatism, however genial. Of course immanent criticism need not necessarily be ideologically conservative.^[30] Because of the tensions and contradictions internal to all ideologies, the practices they legitimate will frequently be open to criticism in their own terms. Yet it is notable that Rorty fails to engage in arguments of this kind, explicitly taking the view that philosophical argument has nothing to contribute to political debate and that philosophical positions are generally without political significance.^[31] Yet we saw that his philosophical

[30] We saw this in our discussion of Walzer in chapter 2. For a more general defense of the view that hermeneutics need not be conservative, see Warnke (1987: 73–141). Although I agree with Warnke's argument that immanent criticism need not necessarily be conservative or reinforce the status quo (as it does for Rorty), I cannot make the leap of faith to the view she appears to share with Gadamer, that over time it should be expected to foster "practical reason or an increased capacity to discriminate" (ibid.: 174). That seems to me to be only one of the possible outcomes.

[31] Thus although Rorty recently declared himself to be a fan of poststructuralism (where this is understood to involve the denial of all forms of objectivism and the recognition that culture is irreducibly rooted in language), he maintains that this has no political implications and that he is a "political democrat who doesn't think philosophy does much for the wretched of the earth" Chronicle of Higher Education, November 25, 1987, p. A8.

commitments generated a willful blindness to the problems of conflicting and subordinate values in our own culture and an undefended assumption that our political institutions preserve Enlightenment values.

Although Bloom and MacIntyre are, for different reasons, troubled by the institutions of modern liberal democracy and would doubtless be even more troubled by Rorty's suggestion that philosophy has nothing to say about its defects, their own arguments are hopelessly deficient from the standpoint of understanding the relationship between political theory and ideology. The problem is not merely that their idyllic assumptions about a vaguely specified past to which normative appeal is made supplies the basis for their critical judgments about the present but that these assumptions are invariably embedded in implausible causal accounts of Modern Decline. Bloom takes over from Strauss the unargued-for assumptions that the ideas of a few philosophers at the turn of the seventeenth century brought about the first wave of Modern Decline and that the subsequent waves down through the present have been the causal result of intellectual movements within universities. MacIntyre, we saw, literally identifies modernity with changes in the history of philosophy at many points in his argument without ever defending this conflation. Rorty, too, identifies what he takes to be the philosophical mistakes of Descartes, Locke, and Kant as causally shaping the intellectual culture of modernity and undermining "our" faith in our institutions. Not only is this an implausible view of the relationship between ideas and political and cultural change, it militates against serious attention to the actual causal possibilities in the world of contemporary politics. Bloom, we saw, is wholly uninterested in addressing questions about how to alter the modernity against which he is determined to rant. MacIntyre has nothing to say about how his small, self-governing communities are to be created or sustained in the contemporary world, and Rorty's quaint insistence that we can just choose to abandon the Cartesian paradigm—perhaps after a period of therapeutic historical consciousness-raising—ignores the ways in which political and philosophical ideas are woven into the complex processes by which cultural and political relations are reproduced and transformed.

Walzer is the only author we have examined who actively engages the problem of ideology's relation to political theory. His discussions of dominance and monopoly supply a causal account of the dynamics of ideological change, and his conception of connected criticism is intended in part to debunk ideological claims. Yet neither view is without its difficulties. Plausible as Walzer's account of ideological innovation might be, it has a descriptivist quality that undermines his search for a critical edge. If every dominant group becomes imperialistic with respect to the good that it best

controls, although it might be predicted that some excluded group will beat back this dominance if it can and eventually impose a new dominant good, it is not obvious why, from Walzer's standpoint, this is desirable. Although he wants to avoid Foucault's position, which reduces all social relationships equally to mechanisms of domination and control, he has made the case neither that some dominant goods are superior to others as dominant goods nor that there could be a society in which no good is dominant. We saw that although his view is complicated by the claim that in a liberal culture opposition to dominance as such is part of the dominant culture in terms of which immanent critical argument is conducted, in the end this resolves neither of the difficulties just mentioned. Because Walzer explicitly refrains from making the claim that our culture is morally superior to others, while embracing the view that moral arguments are never settled definitively, opposition to dominance and the affirmance of a pluralist conception of the good become no more than weapons in this ongoing battle. They may be among the weapons that we are bound to use in our political struggles, but they do not accurately describe any culture nor could they.

More serious are the difficulties attending Walzer's account of connected criticism. We saw that although there is sometimes a good deal of tactical sense behind his argument that only connected criticism can be expected to influence the behavior of those at whom it is directed, this is not always so. In addition, the world of overlapping memberships and conflicting identifications makes the problem of not denying one's emotional bonds a good deal more complex than Walzer's discussion of Sartre and Camus implies. But, more important, we have repeatedly seen that the critic's connectedness (or lack of it) does not begin to tell her what to say. As a consequence, the idea of connected criticism fails as a bridging device between Walzer's descriptive account of the dynamics of ideological change and his own normative arguments. Thus although Walzer's argument is the most powerful of the new contextualists' because he alone among them has both developed a substantive account of the good community (rather than simply invoking the abstract value of community) and grappled seriously with the problem of articulating critical standards from within a contextualist view, much remains to be done. To grapple is not to triumph, and the seriousness and tenacity with which someone of Walzer's sophistication has struggled with this problem without resolving it should hint to us that the limits to mere context as a source of critical standards have by now been fully exhibited.