A Feminist Revision of Free Speech Theory
 Truth seems to be a phoenix in Western civilization. No matter how many particular truths are destroyed by a new vision or discredited by time, the ideal of truth reasserts itself in the human imagination. Indeed, even the periodic attacks upon the idea of truth itself, though they may dim our enthusiasm for a while, seem unable ultimately to extinguish the human desire for truth. In this Article, I would like to explore the roots of this desire and propose a conception of truth that might meet those needs while avoiding some of the evils associated with more traditional models of truth. My contention is that the rehabilitation of truth can best be accomplished using the insights of feminist theory about the inherent connections between epistemology and ethics. I will outline some of the important functions served by a concept of truth and describe a feminist vision of truth based on a model of moral responsibility that can fulfill those functions. I hope to demonstrate that feminist theory can offer not only a critique of the gendered nature of existing legal theory, but also a foundation for a positive reconstruction of that theory. 1
 My vehicle for this exploration will be an area of law that has -- perhaps more than any other -- been the repository for the stubborn human faith in truth: the free speech clause of the first amendment. The doctrinal development of the free speech clause is, to put it mildly, byzantine. In striking contrast, however, the theoretical frameworks relied upon by both scholars and judges to justify this elaborate doctrinal edifice tend to be remarkably straightforward. One of the oldest, most common, and perhaps most commonsensical, of these frameworks is the truth theory of free speech. In its simplest form, the theory asserts that freedom of speech is either the necessary or at least one of the best means of assuring human discovery of truth. Because the discovery of truth is assumed to be of such great social and individual importance, the protection of free speech in order to achieve this goal is justified despite the admittedly serious costs that speech sometimes generates.
 The truth theory of free speech has a pedigree at least as old as John Milton and has been explicitly relied upon by the Supreme Court in its decisions. In addition, it is a theory that continues to hold the imagination of legal scholars, many of whom have criticized it and some of whom have sought to modify, reinterpret, and defend it. There are, of course, many other theories of free speech that have also been the subject of scholarly inquiry and the basis for judicial decision-making. 2 Any full account of a principle of free speech would need to include the insights of many of these other principles as well. This Article is not intended to be such a full account -- it is simply an attempt to explore this particular theory of free speech in light of recent challenges to traditional models of truth.
 These recent challenges have come from a number of different directions, including hermeneutics, post-modernism, pragmatism and feminism. I will focus on the challenges to traditional notions of truth posed by feminist theorists. While all of these movements share many insights, the emphases and underlying concerns differ somewhat in each. It is my hope that one could reach many of the same results that I will suggest by traveling one of these other paths as well.
 The truth theory of free speech relies on a traditional model of truth that has been largely discredited by feminist and other critics. I will call this traditional model Cartesianism. In the first part of this Article, I will outline the Cartesian model of truth and briefly review the work of Mill and the first amendment case law to substantiate the claim that this model has historically provided the foundation for the truth theory of free speech. For those in other disciplines, this review of Cartesianism may appear to be an attempt to revive a long-dead adversary. In law, however, Cartesianism is still very much alive, in both the courtroom and the legal academy. In many cases, it is, however, an unselfconscious form of Cartesianism. This part of the Article is intended to unmask and identify these Cartesian assumptions in legal texts and reasoning.
 In the next part of the Article, I will turn to the feminist critique of this traditional view of truth. This is ground that I have covered in detail elsewhere, so I will offer only a brief sketch of the argument here. The conclusion of this critique is that the traditional notion of truth is not merely incoherent, but also incorporates some of the worst elements of gender hierarchy and functions often as a mechanism for imposition of power and legitimation of inequality.
 One might assume that, having reached the conclusion that the traditional concept of truth is so deeply flawed, the appropriate next step for a feminist analysis would be to abandon the notion of truth altogether. This is not, however, the step I wish to recommend. I believe that truth is a phoenix for a reason: we must constantly revive it because we cannot do without it. If this need is real, then any theory that attempts to ignore it will ultimately prove unworkable and unlivable. In the third part of the Article, I will examine the functions of truth from a feminist perspective. What is the work that the concept of truth does for human beings, individually and in social groups? Which of these functions is part of the bad old legacy that we must overcome and which, if any, represents a valuable service in human life? I will suggest that truth does indeed fill several important roles in both social and individual life, including: providing a basis for a shared reality on which we can base joint decisions; providing a basis for a deep critique of social conventions; making claims for decisions that work in a practical sense rather than merely serving a symbolic function; and providing a basis for a connection to non-human reality. I conclude that we should, therefore, seek to reconceive truth in such a way as to allow it to fulfill those functions, if we can do so without replicating the damaging aspects of the traditional concept of truth.
 The fourth part of the Article will then offer an outline of a revised conception of truth. This new version of truth is designed specifically to meet the needs described in the previous section while responding to the concerns of the feminist critique. In this new model, truth is constructed rather than found; context-dependent, but not entirely relativist. It is explicitly defined in relation to morality (in particular, moral responsibility) and political ideals (democracy and respect), rather than pretending to some type of transcendent neutrality. It recognizes the constitutive role of language and -- though it retains a concern for reason -- it does not impose itself through a false universalism. Perhaps most importantly, this model of truth incorporates an active, interpretive and responsible role for the knower.
 Finally, in the last part of the Article, I will consider whether this new version of truth could function as the foundation for a theory of freedom of speech. If truth is constructed, context-dependent, and connected to both moral responsibility and democratic politics, what does that tell us about speech? Is there a sense in which speech is still either a necessary or an extremely important means in the process of truth-seeking? This section will argue that speech does, in fact, have a role to play in this new vision of truth. The connections between speech, on the one hand, and responsibility, respect, and reason, on the other, make speech a uniquely valuable method for pursuing truth.
 Thus, it is possible to free the truth theory of free speech. Truth can be understood in ways that meet apparently deep and valid human needs, while avoiding both the incoherencies and the immoralities of the traditional model. And truth, so understood, is still one of the important reasons why we should protect free speech. Feminists, and others, need not abandon either the truth theory of free speech or the concept of truth more generally, despite the important criticisms of the traditional model.
 The model of truth that provides the foundation for the truth theory of free speech is a product of the mainstream tradition in western epistemology: Cartesianism. In the Cartesian view, knowledge consists of true representations of an external and objective reality, which is accessible to individual knowers through the use of their reason, sometimes combined with their sense perception. Such knowledge is true for all knowers, regardless of their particular context and is expressed in language taking a propositional, representational form. 3Cartesianism
 Although philosophers have struggled to shore up Cartesianism by reliance on counter- intuitive notions like individual sense-data, at its foundation the Cartesian vision of truth is quite close to the simple, every-day version adopted by most of us in our daily lives. Truth is a quality of beliefs, preferably beliefs that can be expressed in the form of propositions. A propositional belief is true if and only if it accurately describes reality. 4 Reality, in turn, is understood as objective; that is, as existing independent of human perception or understanding of it. 5 Truth is only about such objective realities. Truth is, therefore, fundamentally independent of moral and political goals, beliefs, or values. For example, there is a fact of the matter -- a truth -- about whether there is a table in front of me and that fact is unaffected by whether or not I believe in equality or happen to find myself in a democratic society. Thus, one manifestation of objectivity is the famous fact/value distinction.
 The primary faculty through which we access truth is reason. Reason is understood as a human faculty fundamentally unconnected to a particular social or physical context. What one can see may depend on where one stands, but what one may logically deduce from a given bit of information does not. Reason is instrumental and abstract. The use of reason generates truth that is true for all people: it is universal or neutral. 6 There cannot be multiple, valid truths; on any given issue, there is only one truth.7
 Finally, this view of truth contains a particular understanding of language as well. Language is the vehicle through which truth is represented. Words paint a picture which corresponds to the external reality, or fails to do so. Language -- in its role as the vehicle for knowledge -- is descriptive rather than performative.8Cartesianism in Free Speech Theory
 This Cartesian model of truth lies at the foundation of the truth theory of free speech. It is because speech is useful (perhaps even essential) to the discovery and spread of this type of truth that it must be protected. I will briefly canvass some of the evidence for this model of truth in the writings of John Stuart Mill and in the free speech decisions by the Supreme Court. Although my interpretation of these materials is hardly uncontroverted, it is, I believe, an interpretation so established and common as not to require extended documentation. 9
 In his famous defense of freedom of speech in On Liberty, Mill relies implicitly on an objective sense of truth. His faith in an objective reality is the foundation for his argument that truth's real advantage over falsehood is that, regardless of how often it is suppressed, there will always be people to rediscover it until it finds a setting in which it can be freely aired. 10 It is because people can access an independent reality -- despite their contrary culture -- that they continue to rediscover these unpopular truths. Mill also embraces reason -- understood as the process of argumentation which challenges and corrects our views -- as the primary mechanism through which truth is apprehended. 11 He is also quite clear about his faith that truth is universal -- in the sense that there is on each issue a single truth which is in principle accessible to all. "As mankind improve, the number of doctrines which are no longer disputed or doubted will be constantly on the increase: and the well-being of mankind may almost be measured by the number and gravity of the truths which have reached the point of being uncontested."12 Finally, Mill's entire argument rests on the assumption that truth will take the form of propositions -- claims to describe reality through language -- which can then be contradicted, disputed, and subjected to logical analysis. Although he recognizes a wider range of subjects for such claims than many other Cartesians, he shares the Cartesian assumption about the form of knowledge. 13
 The Supreme Court has referred explicitly to Mill as a source for its own reliance on truth as a justification for free speech.14 Although Supreme Court opinions rarely, if ever, provide systematic expositions of underlying legal theory, the Court has relied upon the truth theory often enough to supply some evidence of the nature of the truth it has in mind. First, many members of the Court have believed that the truth with which the first amendment is concerned is to be pursued through reason. Speech is often held to have lower value when it is, as the Chaplinsky Court stated, "no essential part of any exposition of ideas, and of such slight social value as a step to truth."15 Justice Brandeis made the connection to reason even clearer in his famous Whitney concurrence, when he attributed to the framers a "[b]elie[f] in the power of reason as applied through public discussion."16
 The Court also seems to assume that the truth we are pursuing concerns an objective reality. For example, the Court has held that a statement must be "provable as false before there can be liability under state defamation law."17 The Court recognizes that not all statements concern "facts" that can be proven true or false. In its reliance on a version of the fact/value distinction, the Court is implicitly incorporating a view of truth as concerned with the representation of an objective reality that is, at least in principle, "provable." Indeed, the Court itself refers to the type of evidence to be used in such a proof as "objective."18 The use of the word "provable" also suggests both a rationalistic bias (proven how? through reason) and universalism (proven to whom? any reasonable person).19
 The universalist view is perhaps most forcefully stated by Justice Holmes in his famous assertion that "the best test of truth is the power of the thought to get itself accepted in the competition of the market."20 The idea that truth will win in a free market of ideas rests on the assumption that people are capable of recognizing and accepting truth regardless of their social circumstances.21 The results of the market will, therefore, reflect the underlying objective reality rather than merely the social forces pressing toward one position or another. The marketplace metaphor captures the faith that all (or at least most) people are able to recognize such objective truths. The difficulty, of course, is that Justice Holmes himself probably did not see his market metaphor this way. Given his pragmatic and skeptical leanings, it is far more likely that Holmes intended a less objective meaning of truth and a less universal level of recognition.22 His own position on the value of free speech may have had much more to do with democracy than with truth. 23
 Despite its creator's likely views, however, the marketplace metaphor has come to be the symbol for the faith in an objective reality, the human capacity to apprehend it, and the centrality of free speech to this enterprise. For example, in Milkovitch, the Court, in the course of asserting that the first amendment does not make a distinction between opinion and fact, quoted with approval from a lower court opinion by Judge Friendly in which he argued that the marketplace of ideas metaphor "points strongly to the view that the 'opinions' held to be constitutionally protected were the sort of thing that could be corrected by discussion."24
 Finally, the Court's approach operates on the assumption that truth will take the form of propositional language: claims to describe reality. It is these claims which are the competitors in the marketplace of ideas. Learned Hand identified this "assumption -- itself an orthodoxy, and the one permissible exception [to the rule against orthodoxies] -- that truth will be most likely to emerge if no limitations are imposed upon utterances that can with any plausibility be regarded as efforts to present grounds for accepting or rejecting propositions whose truth the utterer asserts or denies."25 This preference for truths expressed through propositional language is perhaps clearest in the Court's discomfort with non-propositional forms of communication. The Court's devaluation of symbolic speech -- treating many regulations that impact the message of such speech as though they were no more than time, place, and manner regulations -- is some evidence of this discomfort.26 This lack of sympathy is undoubtedly related to the rationalist bias: such non-propositional forms of speech seem to appeal to faculties other than reason and are more difficult to assess or counter with rational argument. Many members of the Court seem to feel that non-verbal or non-propositional forms of expression simply do not advance the search for truth in a meaningful way.27
 Thus, hints of all of these elements of Cartesianism can be found in the opinions of the Supreme Court interpreting the free speech clause of the first amendment. They are, however, often no more than hints because the Court was not engaged in the self-conscious development of an epistemological foundation for free speech theory. Nonetheless, I believe that it is a fair characterization of the Court's position to say that it is using what might be called "naive Cartesianism" -- an acceptance of the general epistemological framework shared by its society without much attention to its possible deficiencies.
 There are, of course, many forms of feminism and they differ in important ways. Despite their many differences, however, feminist theoretical frameworks across a wide range of positions share certain themes on which I will rely for this critique. 28 One such theme -- perhaps the only one shared by all feminist positions -- is the belief that women are and have been systematically denied equal treatment and respect or oppressed by men and that this situation is wrong and should be corrected. 29 A second theme -- shared by many but certainly not all feminist positions -- is a recognition of social constructionism. In the context of epistemological argument, this takes the form of the social construction of knowledge: the claim that knowledge making is an activity that takes place only within, and is deeply shaped by, a cultural context.30 The feminist critique of Cartesianism links these two themes by attempting to show how an epistemology that sees truth as objective, rationalistic, universal, and propositional works systematically to support gender hierarchy.The Social Constructionist Critique
 First, the social constructionist critique undermines the claim that truth is objective, in the sense of reflecting a reality independent of human perception of it. Instead, the critique argues that truth is deeply, necessarily shaped by the social and personal context of the observer. Facts themselves are the product of a process of selection from experience, selection constrained by the conceptual categories that our culture makes available to us. 31 Moreover, once we have our "facts," we must analyze them. This analysis, however, inevitably involves value choices. To choose between alternative explanations of our facts, we must rely -- implicitly or explicitly -- on value judgements. The truth that is available to us is, in other words, no more independent of moral and political values than it is of other cultural concepts. Finally, even the questions or issues we choose to investigate are shaped by culture generally and value judgments in particular. "[A] problem is always a problem for someone." 32 Which problems are studied will depend upon whose perspectives, concerns, and needs are considered most important. Perhaps this whole argument could be summarized by saying that interpretation is a necessary part of knowledge creation at every stage in the process, and that interpretation is inescapably cultural and evaluative in nature.
 The cultural, contextual, and value-laden nature of our knowledge calls into question the other elements of Cartesianism in addition to its objectivism. The rationalist bias, for example, becomes untenable. If reason is defined as instrumental, and particularly if it is defined in contrast to emotion and value judgment, then it is plain that it is not merely insufficient as a mechanism for acquiring knowledge, but is itself an impossibility. Analysis and understanding cannot be separated from interpretation in this way. The refusal to recognize the inevitable contribution of culture, emotion, and values, does not eliminate their impact, but it may serve to disguise the particular cultural assumptions or value choices being made and thereby immunize them from criticism.
 Once we abandon the false dichotomy between reason and emotion (or perhaps the more familiar version for lawyers -- between reason and politics or values) and give up our claims of access to an objective reality, a broader range of approaches to truth opens up to us. In particular, truth need not take the form of propositions that make provable claims to describe reality. When the reality that truth addresses is contextual and constructed, when the tools that one can use to seek understanding of that reality cover the whole range of human experience and capacity, then the scope and form of truth becomes in principle unlimited. No area of human concern is necessarily excluded (e.g. aesthetics, morality). Nor can descriptive propositions claim any privilege as the ideal form of knowledge.
 The result of contextualizing truth in this way, however, is that it can no longer be confidently assumed to be universal. Knowledge is radically situated. The neutrality that searches for a view from nowhere is unattainable; it is certainly not attainable in every case, it may not be attainable in any case. There may, on a given issue, be not one but many equally valid truths. Truth may be personal or social; it is not necessarily universal.
 Thus, the social constructionist argument challenges every aspect of the Cartesian vision of truth. This argument is shared by feminism with a variety of other philosophical frameworks, including, in legal scholarship, pragmatism and postmodernism in particular. Nonetheless, it is properly identified as feminist (along with these other labels) because it grows out of and incorporates many of the themes and concerns of feminist theory.The Relation to Feminist Themes
 First, the feminist concern for those excluded -- which has its foundation in the exclusion of women, but is not limited to that -- is reflected in the denial of a singular truth and a narrow conception of reason as the primary, if not exclusive, path to that truth. In particular, feminists have been at pains to demonstrate that the exclusion of women and other social groups is not simply a result of bad attitudes on the part of individuals, but is the product of systems of exclusion that are deeply embedded in a wide range of social institutions. 33 The social constructionist analysis of Cartesianism is one example of such a demonstration.
 Second, the social constructionist argument I have described shares the feminist concern to recognize the roles of both social structure and individual agency. As I have argued elsewhere, for a variety of reasons, feminism cannot give up entirely on some notion of autonomy or individual agency. 34 Social constructionism, by emphasizing the mutual constitution of social structure and individual agency, allows for the necessary balance. 35 While all knowledge takes place within a cultural context, that culture is enabling as well as constraining. 36 "If we do not create the conventions that govern and make comprehensible our telling stories, we nevertheless select among them, revising, emending, and even scrapping the materials we have on hand to shape our plots." 37 Indeed, the need for selection and revision that cultural constructionism implies creates the possibility of greater agency than in either the traditional, Cartesian model -- in which the knower bears little or no responsibility for the truth he "discovers" -- or an extreme postmodernist model -- in which agency is neutralized and all knowledge is simply discourse. 38 The concern with responsibility as the expression of agency and the foundation of knowledge is a theme common to much feminist theory and one that is intimately connected to this social constructionist epistemology. 39
 Finally, the relational nature of knowledge that is exposed by the argument from social construction is also consistent with a theme common to much feminist theory. Contrary to the sharp division made by Cartesianism between the knower and the known, feminist theorists have insisted that knowledge is "always in part about the relationship between the knowers and what they know." 40 Indeed, the process of influence, and even of knowledge, is seen to run in both directions: just as the known is defined by the context and characteristics of the knower, the knower's very identity is altered by the process of coming to understand the known. "The reconstruction of knowledge is inseparable from the reconstruction of ourselves." 41 Such a relational view is consistent with trends in feminist moral and political theory, as well as feminist psychology. 42The Relation to Gender Hierarchy
 If one finds this central claim of social construction of reality plausible, then no further argument might be necessary to demonstrate the unacceptability of Cartesianism. The specifically feminist version of this critique offers, however, an additional reason for adopting social constructionism. The feminist critique suggests that the Cartesian framework both incorporates and perpetuates gender hierarchy and is, for that reason, a less acceptable alternative.
 Cartesian epistemology has been used as the foundation for defining the difference between the genders. The Cartesian assumptions have formed the foundation for a series of dichotomies -- for example, reason/emotion, objective/subjective, universal/particular -- that permeate not only philosophy and law, but Western culture more generally. These dichotomies have then been used to construct gender, with the valued half ascribed to men and the dangerous half ascribed to women.43 The Cartesian knower is male; he is associated with objectivity, reason, universality and intellect. The thing known is female; it is associated with particularity, emotion, and physicality. Indeed, nature itself, as the quintessential object of knowledge, is understood as feminine, and women are understood as more closely connected to nature than men. 44 The process of acquiring knowledge is also masculine. The relation between the knower and the known is one of separation, a relation of objectivity. Such separation is required for the autonomy of the knowing subject. And masculinity is, of course, defined importantly in terms of autonomy and separation.45 Femininity, on the other hand, is traditionally defined in terms of connection and dependence, characteristics which make the acquisition of knowledge by women perhaps impossible, certainly unfeminine.
 The Cartesian dichotomies do not merely define men and women as different, they also provide a justification for gender hierarchy. The hierarchy is produced in part by the valuing of one side of these dichotomies and the devaluation of the other that results from associating only one side with the production of knowledge. The side associated with men is privileged, thus helping to justify the privileging of the persons who are assumed to best exemplify those characteristics.
 Gender hierarchy also results from the connection between knowledge and power in the Cartesian view. A central feature of the relationship between the knower and the known is that the knower exercises power or control over the known. The external world, the things to be known, are conceived as passive, not in the sense of being inactive, but in the sense of being reactive rather than self-initiating. They are, therefore, subject to prediction and often to manipulation. Reason, which is the key to autonomy and freedom from determinist particularity, is available to the knower but not to the known. Perhaps this relationship can best be summarized by saying that Cartesianism posits the thing known as an object rather than as a subject in its own right. 46 This result frees the knower to exercise power or control over the thing known for his own ends rather than considering and respecting its ends or establishing a community of shared ends, as one would hope he would feel required to do with another subject. Since knowers are, almost by definition, male, it is men who are thereby authorized to exercise this power. Since women are, almost by definition, part of the natural world to be known, they are among the fit objects of such control.
 There is extensive argument among feminist theorists over the question whether this gender hierarchy is a necessary aspect of a Cartesian epistemology or whether it is only an historical accident that the two ideologies have become joined in this way. 47 For the purposes of my argument, I do not think it matters which position one adopts on this issue. Even if the connection is a matter of historical happenstance, it is too strong and deep to now be severed. Whether or not it was necessary that it be so, it is now the case that part of the meaning and function of Cartesian epistemology is to reinforce existing gender hierarchy in a deep and damaging way. It is particularly damaging because Cartesianism provides a foundation for so many diverse social practices, thus supporting the inequality and oppression of women in a wide range of circumstances. The links between knowledge and masculinity and power contribute to the unequal education of girls, to the discomfort with women holding important political and economic positions, even perhaps to the use and abuse of women in sexual relationships.48
 So, the feminist critique suggests that every aspect of the Cartesian vision of truth is mistaken. The feminist critique also argues that the Cartesian version of truth encodes and contributes to gender hierarchy. In short, it is both epistemologically flawed and morally objectionable.
 Despite these flaws in the traditional model of truth, there are good reasons why we should not, perhaps even cannot, do without a concept of truth altogether. Truth serves a variety of functions that are important and valid -- that is, they are not part of the perpetuation of hierarchy. This section will attempt to describe those functions, in order to create the groundwork for the further argument that a social constructionist model of truth can, in fact, fulfill those purposes.Truth as Shared Reality and the Connection to Democratic Decision-making
 Perhaps the most obvious social or political function of truth is to facilitate joint decision- making by groups. If anything like a democratic process is to be used for decision-making -- whether it is majoritarian or consensus-based, direct or representative, and regardless of the fine points of political theory and voting procedures -- then the members of the group must be able to recognize at least some shared realities. There are at least two ways to see this connection: one is in terms of the cognitive elements of cooperative decision-making (i.e. language and conceptual categories) and the other is in terms of trust.
 First, democratic decision-making requires a shared reality because it depends upon collective use of language and conceptual categories. Perhaps if we see democracy only as a form of preference summing, in which fully formed individuals merely register their desires through a voting mechanism, this shared reality could be limited to some ability to understand the proposal under consideration and the procedures themselves. Even here, however, such a minimal shared reality is necessary if the process is not to degenerate into total chaos. In other words, some minimal intersubjective knowledge must be possible to allow for the use of language, and language use is a precondition of perhaps all forms of social organization, including democratic politics. 49 In any richer account of democracy, which includes some role for deliberation, whether on means alone or also on ends, a correspondingly richer shared reality would be needed. The intersubjective knowledge claims possible in such a richer, shared reality might extend to moral and political claims along with more empirically oriented ones.
 Second, democratic social action may require some minimum amount of trust between the participants. Whatever else may be necessary to such trust -- perhaps practical connections or emotional commitments -- there is certainly an epistemological precondition. One cannot trust someone who does not share, at least to some extent, the same reality. One cannot trust someone who cannot or will not participate in the process of making and accepting claims of intersubjective knowledge. Part of this is simply a matter of respect: if your interlocutor is not participating in this process honestly, then he is manipulating you and treating you with insufficient respect for your subjecthood. This is no basis for trust. But I think honest participation is insufficient. Even a completely honest person cannot be trusted if he inhabits a totally different and inaccessible reality. 50 He cannot be relied upon because we do not understand the reality to which he is responding. Collective social action -- at least democratic action -- requires this reliance. Thus, one way of understanding the need for shared reality in order to support democratic processes is that such a shared reality is necessary to trust and trust is necessary to democracy. 51
 A truth claim can be understood, in part, as a claim about such shared realities. It makes a demand for recognition by the listener that is not present where no truth is claimed. To call something true is to imply that it (generally) can and should be accepted as a belief by the listener. In other words, the function of truth here is to make claims for the possibility of intersubjective knowledge and the existence and desirability of a shared reality.
 If the critique of Cartesianism is taken to the extreme suggested by some postmodernist theorists, it might seem that no such intersubjective knowledge or shared reality is possible. Indeed, with the disintegration of the self, the "subjects" between whom such knowledge is shared become so fragmented that the very notion of intersubjectivity may be called into question. 52 One can see social constructionism as faced with a dilemma here. If the social foundations of knowledge are seen as too static, then social constructionism becomes excessively conservative and cannot allow sufficient room for serious criticism and reform of society. If, on the other hand, the social foundations are seen as too fluid and unstable, then social constructionism becomes unable to support claims for intersubjectivity and community. 53 It is the latter horn of the dilemma that causes this problem for democratic decision making. The point I wish to emphasize here is simply that we cannot accept that horn: the need for an intersubjective knowledge that will allow for a shared reality is fundamental to democratic aspirations and any form of social constructionism that makes it impossible -- i.e. that demands that we completely abandon truth, understood in this way -- is for that reason unacceptable.Truth as a Foundation for Deep Critique
 Nor can we accept the other horn of the dilemma. Another social or political function served by truth claims is that they are sometimes the basis for an appeal to something larger and deeper that can be used to criticize social conventions or culturally sanctioned assumptions. Martha Nussbaum has pointed out that appeals to abstract values, for example, assert norms independent of a particular culture and can be especially important for oppressed groups. 54 Just as the first use of truth -- insisting on the possibility of a shared reality or intersubjective knowledge -- rejected one horn of the social constructionist dilemma (total relativism), this use of truth -- insisting on a grounds for deep and sweeping criticisms of culture -- rejects the other horn (conventionalism). In particular, it rejects conventionalism in the interests of those who are oppressed by the conventions of a given society and seek grounds for criticizing and changing that condition. From a feminist perspective, finding stable grounds for such a critique is essential, because we wish to assert both that we live in a sexist culture and that feminism is possible.55
 Some forms of social constructionism apparently accept the conventionalist position. Richard Rorty, for example, has been quite explicit about this acceptance. 56 He has, as a result been criticized by feminists, among others, for ignoring the impact of such conventionalism on groups suffering under systematic domination within our society. 57 There is no question that the ability to make claims about a larger or deeper reality in order to criticize one's own society is a valuable tool, particularly to those in positions of subordination within that society. The question, of course, is whether such a tool can be fashioned from the materials of social constructionism. As with the other horn of the dilemma, I wish to suggest that a social constructionism that cannot sustain at least some version of this function of truth is, for that reason, unacceptable. Finding an adequate model or metaphor -- a description of the position on this continuum that offers a workable balance between these two extremes, a way of thinking about one's position that helps to retard the slide to either end -- is one of the challenges of social constructionist epistemology and is part of the goal of this Article.Making Decisions that "Work"
 In the social or political realm of group decision making, truth serves another function as well. Even if we are capable of reaching a joint decision, through democratic procedures and with adequate scope for criticism, such decision making loses a great deal of its value if the decisions reached through it don't work. We want the process to generate decisions that move us closer to our goals and that produce as few unanticipated evils as possible. To claim that something is true is, in part, to argue that knowledge of it will help us to reach decisions that will work and, implicitly, that reaching such decisions is a valuable social enterprise. Such truth claims insist that we should care, not only about the symbolic import of our decisions, nor even just about the process through which they are reached, but also about their practical consequences.
 The central issue, of course, is determining what we mean when we say that something "works." This use of truth may be the one that is closest to the hearts of many Cartesians because they understand "working" to mean manipulating an external and objective reality to meet our needs or desires. Thus, the truth that it is necessary to take into account in this process is the knowledge about the nature of that external reality. For example, if I want to build a bridge that will allow me to cross a river, I must take into account the width of the river, the strength and flexibility of my materials, and the vertical distance the river may rise with flooding. If we hope to escape from Cartesianism, then we must come up with a meaning for "working" that does not rely on such an external, objective reality but that nonetheless insists that practical consequences matter. I will draw the connections between this meaning of "working" and an alternative model of truth later in this Article, but here I want to argue two preliminary points: first, that such a standard is conceivable in non-Cartesian terms; and second, that there are important values served by caring about such a non-Cartesian version of workability.
 A concept of workability need not be Cartesian. Indeed, the philosophical tradition most concerned with this notion is the very non-Cartesian tradition of pragmatism. The pragmatic approach suggests that truth is what works, and makes no claims for its relation to any reality independent of human understandings. It "works" if it helps us to cope with our environment, including the internal tensions we may feel as well as the social and political circumstances in which we find ourselves. 58 It "pays to pursue" truth. 59 Truth is what it is good for us to believe.60 Feminism can add to this non-Cartesian account the stipulation that it only "works" if it does not result in systematic, illegitimate hierarchy, like patriarchy. Another way to put this is that the "us" for whom it must work, must centrally include those who are systematically oppressed by such hierarchies.61 Thus, the standard for the social constructionist concept of truth, understood as workability, is fundamentally ethical rather than epistemological and is thoroughly non-Cartesian. 62
 The importance of such a non-Cartesian concept of workability lies precisely in its pragmatism, that is, its focus on the concrete details of human experience. The central claim is that, in making decisions, we must focus not only on the symbolic significance of our actions, but also on the practical impact they will have. The significance of this claim lies in its ethical dimension. It demands that we take seriously the pain and the happiness, the subjecthood, of others. As such, it represents a form of respect for them. This claim also insists that we recognize responsibility towards those others in making such decisions. In other words, it is because failure has consequences for other subjects who matter that we care whether or not a decision "works."
 This focus on "working" helps to explain why, even if we wish to abandon Cartesianism, it would be foolish and self-defeating to abandon the practice of science. The many benefits of science -- for example, medicines to cure disease and bridges that don't fall down -- can be recognized and appreciated without Cartesian assumptions if one includes the concern for the impact on other subjects as part of the function of truth. Indeed, I believe that it is precisely this ethical concern that gives force to the Cartesian commitment to this aspect of truth as well. If I used inadequate materials to build my bridge just because I wanted to make some symbolic point about rejecting the producers of the appropriate materials (even if that point was a valid one), I could be criticized for failing to take account of the practical consequence that people could be hurt by the failure of the bridge. Or if I made decisions about the use of certain medicines on the basis of a process -- no matter how democratic -- that did not take account of the impact of my choices on those suffering from disease, then, again, I could be criticized for ignoring the "truth." In other words, this ethical understanding of what it means to "work" captures the real, although perhaps hidden, concerns behind even the Cartesian use of this meaning of truth. 63 If part of the function of truth is to focus our attention on "working," then, to that extent, the concept is important and should not lightly be abandoned.Truth as a Connection to a Non-Human Reality
 In addition to these social or political functions of truth, the concept seems to play an important personal role in the lives of many people. This function is somewhat more difficult to describe than the previous ones, perhaps because it often takes a form that might be called spiritual. Truth can form the basis for an understanding of oneself as connected to reality in a stable way, so that reality is not set adrift or up for grabs, but anchored. And it can offer a foundation for that stability outside of oneself; a view of reality in which one is not the ultimate creator but simply a part of the larger whole of creation. The relationship to the elements of nature -- not just other human beings -- is central to this view: the sense of being part of, perhaps even in some ways on a par with, the rest of nature.
 Some versions of postmodernist theory have shown little sympathy for this function of truth. Richard Rorty, for example, has suggested that the need for transcendence can be eliminated by a "suitable moral education" which "tries to sublimate the desire to stand in suitably humble relations to nonhuman realities into a desire for free and open encounters between human beings, encounters culminating either in intersubjective agreement or in reciprocal tolerance."64
 In an extremely interesting article, Richard Thomas displays somewhat greater sympathy when he proposes that "[s]uch longings [for transcendence and foundations] are the stuff of religion, and the choice to be secular entails a loss that cannot be recouped through the law, or politics, or reason, or art." 65 While it may be true that we should not seek to found our politics on these sorts of truth claims, we may, nonetheless, wish to secure first amendment protection for them as speech. In other words, if this is one of the sorts of truths which serves an important function for human beings, it should be part of the search for truth protected by the free speech clause. The new concept of truth may not be able to provide a basis for every aspect of this function as experienced by all those who feel these longings, but it should be designed to provide as much as it can.
 Truth, then, serves several important functions in human experience. While the Cartesian model of truth must be rejected, it is worthwhile to explore the possibility of alternative models that might allow us to continue to fulfill these needs without the incoherent and morally objectionable aspects of Cartesianism. Such an alternative might then offer a new vision of the relationship between truth and speech.
 This section will provide an outline of an alternative vision of truth. First, I will describe this new model of truth in order to demonstrate the senses in which it is non-Cartesian. Second, I will connect the model to the feminist themes that I identified as lying behind the critique of Cartesianism. And finally, I will show how this alternative vision allows us -- at least to some extent and in certain ways -- to fulfill the functions of truth that I discussed in the previous section.Non-Cartesian
 This alternative view of truth is non-Cartesian because it rejects all of the central tenets of the Cartesian model of truth: objectivity, rationalism, universalism, and the propositional view of language. This new model sees truth as fundamentally contextual and normative. That is, knowers can engage in the activity of knowledge-making only from a particular position or perspective, one which is radically shaped by their cultural context, including their normative concerns. Such contexts should not be seen as impediments to knowledge, but rather as the preconditions for it. To ask what we could know if we could escape all such perspectives is like asking what we could see if we could escape having any particular sort of eyes.
 One result of this focus on the role of interpretation, is that truth is seen as made by human actors rather than found. There is no passive receptivity to an external reality; rather there is an active process of creation. Another result of this focus is that the traditional dichotomies -- particularly any version of the fact/value distinction -- are made untenable. There are no facts that are not permeated by values because all result from a process of interpretation that is inescapably evaluative in nature. This model also rejects universalism: truth may often be plural rather than singular. The identity, cultural and personal, of the knower is implicated at every point in the process of knowledge creation. As a result, different identities may generate different visions of the truth which are equally valid.
 The rationalist bias of Cartesianism also falls away. First, reason cannot be understood in the abstract and instrumental way that it traditionally has been. Reason, like truth, is highly contextual. Second, reason cannot be contrasted with emotion or politics simply because the latter are seen as evaluative in nature. The opposite of a reasoned judgment is not an emotional one, but an unreasonable one: emotional reactions are often quite reasonable.
 None of this means, however, that reason should lose its place among -- maybe even in the forefront of -- those human capacities that we use to make knowledge. Reason, like truth, is a category worth rehabilitating. When it is seen as contextual, evaluative, and inclusive (i.e. when its opposite is understood as unreasonableness), then reason can play an extremely important role in this alternative model of truth.
 Reason is the process through which we make sense of our experience. "Making sense" is, of course, a highly contextual matter: the sort of argument that can make sense of our experience in a physics laboratory (mass, velocity, vectors, etc.) is not necessarily the same sort of argument that can make sense of our experience watching a baseball game (strikes, balls, and base hits). To try to explain the latter in the same terms as the former, would be to fail to provide good reasons for the relevant experience. The concept of reason captures this concern that some things (arguments, conceptual categories, etc.), in some contexts, help us to make sense of our experience better than others. 66 Understood in this way, the concept of reason has a useful role to play in this alternative model of truth.
 Finally, this alternative model sees language in a way quite different from the traditional Cartesian propositional, descriptive view. Because cultural understandings so deeply shape our truths, and because language is one of the primary vehicles through which such understandings are created, maintained, and transmitted, language is seen as constitutive rather than as descriptive. In other words, language does not draw a picture of some independent reality; language makes reality, or -- perhaps in some senses -- even is reality.The Relation to Feminist Themes
 As I mentioned earlier, this alternative model of truth also draws on and exemplifies several themes widely shared by feminist theories in a variety of disciplines. The first of those themes was a concern for those -- women and others -- who have traditionally been excluded from the social process of knowledge-making. This concern is incorporated into the alternative model of truth in the form of a focus on the relationship between epistemology and politics. "[I]f rhetoric is the means by which reality, or plural realities, are constructed, then democratization of discursive or rhetorical opportunities is inseparable from the project of democratization itself." 67
 The recognition of this connection between epistemology and politics leads to a conception of the process of knowledge-making as itself a political process, and one in which democratic reforms are necessary in order to end the traditional exclusion of less powerful groups. The goal is not some idealized speech situation in which we experience no hierarchy or power differentials. All discourse suffers from distortions and the search for an undistorted discourse is, in the end, not very different from the Cartesian search for a neutral observer. 68 The point is, rather, to subject our process of knowledge production to scrutiny under the same democratic standards that we apply -- or aspire to apply -- to our political process. Indeed, the point is to recognize that the process of knowledge production is ultimately a part of the political process of determining the distribution of power and resources in society. If we are concerned about the exclusion or marginalization of certain groups from that political process generally, we should be equally concerned about it here.
 The argument I have just offered describes the way in which the alternative model of truth opens epistemology up to the sorts of concerns about exclusion that are common to feminist theory. There is another argument that could be made about this connection, however, one suggesting that this alternative model of truth helps to justify or ground the concern about exclusion, rather than merely sharing it or extending it to a new area. This argument points out that once truth is seen as contextual and perspectival, we must recognize our dependence on others, with different perspectives, to help us see beyond our own limited view. 69 Indeed, when the standard for justification is no longer a correspondence to an external reality, but instead defined by reference to concerns and assumptions -- moral and otherwise -- shared by a certain community, then the broader and more inclusive the audience a position can address, the better justified it is. 70 Bringing in the excluded becomes a precondition for making knowledge well.
 The second theme common to much feminist theorizing involves a focus on the issues of agency and responsibility. These concerns are absolutely central to this alternative model of truth. Once one adopts a social constructionist view, the issue of agency becomes critical because one must explain the meaning and conditions for individual agency under social constructionism. As I have written elsewhere, I believe that there is much work that can and should be done to specify the content and conditions -- material, emotion, or cultural -- for agency within a social constructionist framework. 71 I will not explore this general issue further here, but will instead focus on a particular aspect of the issue of agency in the context of social constructionist epistemology: the role of responsibility.
 Responsibility becomes a fundamental part of the process of making knowledge because of the inevitably normative and evaluative nature of the project. Once we recognize that truth is made through a process of selection and that the selection necessarily involves judgment, including moral judgment, we must also recognize that the knower must take personal responsibility for those judgments. 72 This is not to say that these judgments are totally uncaused; they are, of course, deeply shaped by the context and life history of the knower. It is, however, to assert that persons are not simply the passive locations in which social forces collide. There is an active process of interpretation -- of remaking of the cultural materials one is given -- that is an inescapable part of knowledge making. 73
 To accept responsibility for that process of interpretation is to recognize that the frameworks one applies are subject to challenge and criticism, including moral criticism, and must be acknowledged, examined, and defended on those terms. Responsibility, in other words, is a relationship to other people -- those who do or might engage in such a challenge. 74 It is also, however, a relationship to oneself and to the frameworks and moral ideals one holds; one must engage in the practice of knowledge making in a way that does not violate those moral commitments. This understanding of the process of knowledge creation, therefore, connects to our deep sense that responsibility and agency are, at least in part, about integrity.
 One of the more important implications of this focus on agency and moral responsibility, is that it helps to break down the assumption, commonly associated with a social constructionist view, that a particular experience leads inexorably to a certain identity and perspective. 75 Because interpretation is an active process, and because we must accept responsibility for its moral implications and assumptions, it becomes possible for people with one set of experiences to learn from those with another. We cannot, perhaps, ever simply see the world the way they do, but we may see the world or ourselves differently if we actively use their insights as a lens on our own experience. 76
 The third feminist theme that I discussed earlier -- the focus on the relational nature of knowledge -- emphasizes that this process of learning from others is not merely a matter of acquiring information. The new model of truth is relational in the sense that knower and known shape each other. One consequence of this relation is that knowledge implicates the identity of the knower, both in shaping the thing known and in being shaped by it. As a result, when one accepts this responsibility, seeks out excluded perspectives, and tries to learn from them, one changes not just what one knows, but also who one is. The relational aspect of this model of knowledge helps to explain how it is that the process of learning from others leads to the creation (not the discovery) of a new reality, including knowers with new identities, and therefore opens up the possibility of still other new realities that did not appear possible before.Can the New Model Serve the Functions of Truth?
 With this sketch of an alternative model of truth available, we can now turn to consider whether and to what extent this model can serve the purposes of truth outlined in the previous section. The first function of truth is as the basis for claims of intersubjective knowledge, claims concerning a shared reality that makes democratic decision making possible. The second function of truth is as the basis for a critique of existing social institutions and conventions, a critique that can be broad and deep rather than merely a matter of "tinkering." I suggested in the previous section that these two functions marked the acceptable range on a particular continuum: a continuum running from a totally open cultural framework, which would allow for too little shared reality, to a totally closed one, which would allow for too little critique.
 The difficulty, of course, is that however one defines the cultural framework, there are constant pressures from both directions. For example, in working with the cultural category of gender, when one focuses on the force of its constraints on women's lives, one can find oneself sliding toward a position in which the cultural category is so fundamental and so powerful that it is difficult to see how any challenge to it could possibly arise. 77 If, on the other hand, one focuses on the incredible variation within the category and the ability of individuals to reinterpret it, then it becomes difficult to see gender as a coherent category at all or as a source of serious oppression. 78 The challenge is to define or describe a position on that continuum that will retard the slide to either extreme.
 If knowledge is all contextual, one way to see this dilemma is as a problem about the size of the context. If we are looking for a context rich and powerful enough to generate a shared reality on a controversial issue (e.g. abortion) then we can find ourselves drawing the context ever smaller, in order to find enough common ground for a consensus. Such a shrinking context, however, can leave us with a radical relativism in which truth can be determined only from within a given viewpoint and viewpoints are -- at least potentially -- no bigger than a single person. This would, of course, make shared reality for the purposes of political decision difficult or impossible. On the other hand, if we are looking for a context big enough to include deep and broad challenges to cultural assumptions, then we may find ourselves expanding the boundaries of the context beyond the usual cultural limits. Such an expanding context, however, can leave us with too little in common to draw on for any kind of claim at all. That would, of course, make deep criticism meaningless: it would be like criticizing humans for not being whales.
 In other words, the two functions of truth push in opposite directions, but each -- if taken to an extreme -- undermines its own goals. When seen in this way, the problem seems to be finding the right size context: one that is small enough to generate a strong shared reality, but big enough to include the materials for deep criticism of its own assumptions.
 I want to suggest that what makes this problem so intractable is the assumption that contexts are given, as though there were a list of the possible contextual frames for a particular issue and we merely have to choose between them. For example, on issues of gender equality, one might see the possible frames as running from the level of each individual ("you see it your way and I see it mine"), to the ethnic and religious communities that have taken various positi'ons on the issues ("you'll never convince them because they are coming at the issue from within their religious framework"), to the national community that has made certain legal commitments ("it's about the meaning of the equal protection clause"), to the historical tradition of which that national community is seen to be a part ("the history of Western culture"), to the history and/or present circumstances of human beings all over the world. Although arguments are possible within each of these frameworks, nonetheless, the choice of a framework often seems to largely determine whether there is sufficient common ground for a joint decision and whether that common ground includes sufficient materials for a serious challenge to deep cultural assumptions.
 This phenomenon of assuming pre-existing contextual frameworks contributes to the difficulty in two ways. First, it makes it seem as if the ability of any given framework to meet the needs for common ground or deep criticism is set in advance because the nature of the context -- its boundaries and meaning -- are given. Under the new model of truth, this is simply wrong. Each context is created in the act of calling upon it and its resources and possibilities are constantly subject to reinterpretation and revision. This does not mean that simply anything is possible in any framework, of course, but it does mean that we determine what is possible through the very process of defining it rather than knowing its limits in advance. Second, this "list"-of-contexts phenomenon suggests that there are a finite number of contexts on which we could draw and they are determined in advance. Again, not only the substance of a framework, but its very existence as such is a matter of constant creation and recreation. To take the example above of gender equality issues, one of the claims of feminism has been that one context from which we could see these issues is the experience of women, and one of feminism's greatest achievements has been to make that context a reality through mechanisms like consciousness raising and political action. In other words, this context was created by the activity of feminists and wasn't on anyone's list before they acted. Thus, the "list" phenomenon is misleading both about the fluidity of the boundaries and meanings of "existing" contexts and about the possibility of "new" contexts. And it is misleading in a particular way: it fails to recognize the active, interpretive, and relational character of the activity of knowing.
 The point is that we are not at the mercy of some slippery slope down which we will necessarily slide into either too large or too small a context. Contexts don't just sit there waiting for us, they have to be built and constantly rebuilt. And it is the connection to politics and morality that must guide us in interpreting and constructing contexts. Thus, if we find that there is not enough common ground for the social group that must make joint decisions then we must build that common ground. The construction of such common ground is a large part of the purpose and appeal of the narrative turn in legal scholarship. The use of stories allows us to expand our experience without actually changing the circumstances of our lives. And that expanded experience may provide a foundation for a shared reality across significant lines of social difference, such as race, class or gender. 79 Or if we find, on the other hand, that we share a great deal but that there are not sufficient materials within our present context for deep criticism, then we must build a context that includes those materials. As I mentioned above, the recent history of consciousness raising provides a striking example of the process of using the materials of a shared reality -- the experiences and unhappiness of many women in a patriarchal society -- to generate new grounds for critique. Another example might be the borrowing of ideas from continental philosophy by American legal scholars and the effort to work them into forms that are comprehensible and useful within our existing legal culture. 80
 As this last example suggests, such efforts at cultural construction are not guaranteed to be successful, and certainly not within any particular time frame. A useful cultural context is an achievement; it exists only in and through the creative activity of people. And it may take the efforts of many people and significant amounts of time before the context can effectively be drawn upon for the purposes of joint decision-making and/or deep critique.
 Knowledge-making, in other words, is similar to, even continuous with, political action. The focus of this model of truth on knowledge making as active, as political, and as involving moral responsibility is the key to escaping the idea that our contexts are always threatening either to disintegrate or explode. Obviously, this active model does not guarantee that we will have a context for any given issue that is both sufficient for joint decision-making and open to deep criticism; our efforts may fail rather than succeed in any given instance. But it does guarantee that we will not find ourselves sliding uncontrollably toward one end or the other of this spectrum, trapped either in a context that cannot ground a shared reality or one that cannot sustain deep criticism. Contexts are not so given, and we are not so passive, as that picture would suggest.
 Thus, the challenge before us is not a conceptual one (i.e. how to imagine our situation so as to avoid this slide), but a moral, political, and practical one: deciding what sort of context would be good and just for the purposes of a particular issue and then taking the action necessary to create it. Seeing the goal in this way does not, of course, generate simple or determinate answers, but it does make it plain that answers are never simply foreclosed by our epistemological situation. Neither relativism nor conventionalism is ever inevitable. Truth -- understood as either claims on a shared reality or claims to expand the context to include materials for a deep critique -- is possible, meaningful, and desirable within this new model.
 The third function of truth that I discussed in the previous section is to provide a basis for arguments about whether a given course of action will "work." As I argued there, at the heart of this function of truth is a focus on the practical impact of our decisions on all the members of the relevant community. The reason this focus is so important is that it represents an ethical commitment to take seriously the pain and happiness of others. As such, it is a form of respect for their personhood or subjecthood, a refusal to allow symbolic goals or procedural commitments to simply override their concerns.
 I hope it is fairly clear how this meaning of "working," this function of truth claims, is consistent with, even required by, the model of truth offered here. The responsibility that is central to this model requires that we consider and respond to the challenges, moral and political, that others might bring against our decisions. This form of truth claim is precisely that sort of challenge. It insists that, from the perspective of this person or group of people, our decision looks like a failure to respect their personhood. It argues that we have defined the community in too narrow or exclusionary a way if it is not broad enough to include them as morally significant persons. In other words, this model of truth not only allows for the "working" type of truth claim, it places such claims at the center of our attention as raising precisely the issues of moral and political responsibility upon which epistemology rests.
 Both my earlier discussion of the connection between epistemology and politics and this understanding of the truth function of "working" may raise similar concerns for some readers. The difficulty is that allowing -- even demanding -- that we subject our epistemological processes and their resulting truth claims to moral and political criticism may seem to open the door to forms of thought-control that are truly terrifying. It suggests that one could criticize or refuse to fund or perhaps even punish criminally one who dared to make a truth claim that violated a deeply held moral or political commitment of the community. It is "political correctness" with a vengeance.81
 This is a danger that we must take very seriously in light of the history of attempts by governments and powerful groups to control "truth" for political ends. My disagreement is not with the claim that such a danger exists, nor with the admonition that we must forcefully resist it, but with the assumption that a commitment to a Cartesian epistemology is either a necessary or a sufficient foundation for such resistance.
 Cartesianism cannot provide a guarantee against such tyranny and, indeed, it has as often been part of the problem as part of the solution. The history of "research" on sex differences that has been shaped by and used as a justification for gender inequality is a prime example of the role of Cartesianism in enforcing moral and political values in scholarship.82 The reason that Cartesianism has been part of the problem is that it has facilitated the denial that moral and political values were at work in such research. It has allowed scholars to hide behind neutral explanations that did not require them to acknowledge and defend their values openly. A commitment to Cartesianism does not empty truth-seeking of its moral and political content -- nothing can do that -- but it does make that content harder to see and therefore more difficult to challenge.
 Not only is Cartesianism insufficient as a basis for resistance to the danger of thought- control, it is also unnecessary. As I will argue in detail in the next section, the social constructionist model of truth provides a foundation for the protection of freedom of speech. This model recognizes important connections between truth and respect, responsibility, and reason. Speech in turn is a primary -- perhaps a necessary -- mechanism for practicing respect, responsibility, and reason. In other words, there are strong reasons why, if we care about truth understood in a social constructionist way, we would protect speech.
 An opponent of social constructionism might reply that this epistemology has also been misused to support programs of thought-control and can, therefore, claim no advantage over my description of Cartesianism. It is undoubtedly true that social constructionism, like Cartesianism, offers both the possibility of repression and the possibility for resistance to repression. If the argument got me no further than this sort of "moral equivalence," I would be satisfied. If proponents of Cartesianism would approach the claims of social constructionists without constantly raising the specter of totalitarianism, that would improve the dialogue in a meaningful way.
 In fact, however, I do not think there is a moral equivalence. Once one recognizes that both epistemologies are merely tools that can be used either to support or resist oppressive policies, the question immediately arises: what determines in which way they will be used? The answer is, of course, the political and moral commitments of the people using them. In other words, the only ground on which we can ultimately rely in efforts of resistance is composed of moral and political values; it is not epistemological at all. 83 This does not mean, however, that all epistemologies are equal. If our ultimate assurance lies in our values, then an epistemology that directs our attention away from those values is less useful than one that makes those values a central concern. Social constructionism, by highlighting rather than hiding the role of moral and political value judgments, facilitates such recognition and concerted action. This model of truth is, therefore, better protection against the evils that so justifiably worry some Cartesians.
 Finally, the fourth function of truth that I discussed in the previous section is as a basis for a connection to a stable, non-human reality in which human beings are simply one constituent part. The truth claims that fulfill this function may often take a form that many would consider religious: claims about God or about the harmony and moral standing of the natural world. But religious claims -- conventionally understood -- are only a subset of the category I have in mind here. Truth claims based on a secular environmental ethic, or on a commitment to the rights of animals can also serve the same function. Any claim that positions human beings in a reality which they alone do not control, a reality shared with other moral agents or entities, serves this function.
 This function poses the greatest challenge for a social constructionist vision of truth. In this new model, truth is a human creation, deeply shaped by culture and personal context. A large part of the struggle to escape Cartesianism consists of the refusal to associate truth with a transcendent reality, at least if "transcendent" means beyond human culture. Social constructionism cannot take a position on what, if anything, is "really out there." Indeed, from a social constructionist view, this is a meaningless question, since all knowledge-making goes on only "in here" - within human cultures.
 The fact that the question is meaningless, that we cannot ever make knowledge claims about such a transcendent realm, does not mean, however, that a social constructionist view must assert that nothing exists except human beings and their experiences. Indeed, such an assertion would be simply another form of forbidden claim about what is (or is not) "really out there." The claim must be, instead, that all knowledge is relational in the sense that it is about the relationship between the human knower and whatever is known. This does not deny the existence of the thing known, but it does deny that knowledge, as we practice it, is directly about that. Thus, the social constructionist view does not deny the reality of non-human nature. Indeed, in some sense it is very respectful of nature by positing it as a partner in the knowledge- making process, a process which has profound effects on the life and identity of the human knower. But the social constructionist view does imply that what we know of non-human nature (and indeed of another human being as well) is always and only our own relationship to that subject. We can alter and multiply our perspectives, but we can never step outside of our perspective entirely.
 Thus, a social constructionist view does not guarantee that any particular claim about the human relationship to non-human nature is true, but it recognizes such categories of claims as legitimate. From a social constructionist point of view, non-human reality certainly exists, because all existence does or could mean is defined from within a human perspective. A claim that something exists is a claim that we stand in a certain kind of relationship to it and, as such, the claim may be true or false, but it is certainly legitimate and comprehensible. It does not become impossible to talk about our role in some reality large enough to include non-human moral elements just because we have adopted a social constructionist view of truth. We can speak about this subject only from within a human perspective, of course, but that does not mean we are silenced. 84
 To summarize this section: there is an alternative model of truth that avoids the difficulties of Cartesianism. This model relies upon and elucidates common feminist themes and it serves - at least in some ways - the primary functions of truth claims. It achieves these goals by maintaining a balance between its focus on the contextual and relational nature of truth and its focus on the moral and political responsibility of the knower, a responsibility which is grounded in her active, interpretive role. This balance allows the social constructionist model of truth to provide a basis for both a shared reality and a deep critique. This model explains and justifies the function of truth as a claim to pay attention to the practical concerns of others. And this model allows for truth claims that address the human relationship to a non-human reality. The remaining question, of course, is: what does all of this have to do with free speech?
Social Constructionist Model of Truth
 If we adopt this alternative model of truth, is there a sense in which free speech is still necessary to the attainment of truth? Is there, in other words, an alternative truth theory of free speech? In this section, I will outline the role of speech in a social constructionist process of knowledge-making.
 In order to see the connection between speech and this new model of truth, it is necessary to think of speech itself differently from in the traditional version of the truth theory. As Robert Post has so persuasively argued, we must think of speech as a collection of social practices rather than as simply the communication of messages. 85 "Truth-seeking is not merely a matter of sentences and propositions; it also involves habits of mind, priorities of reason, intersubjective orientations, and attitudes that, when taken together, make up what we recognize to be rational exchange or a collective search for knowledge." 86 My argument is that speech is useful or even necessary to each of these elements of the process of knowledge-making. 87Respect
 Perhaps the most fundamental sense in which the social practices we identify as speech serve the purpose of promoting truth (understood in a social constructionist way) is that those practices involve an "intersubjective orientation" that we could describe as respect. In a dialogue, one must recognize the other participant as a moral agent deserving of respect. 88
 As I hope the discussion in the previous section made clear, such respect is a necessary element in the process of knowledge-making when truth is seen in a social constructionist framework. In order to act as a morally and politically responsible knower and in order to adequately take account of the impact of a decision on others (i.e. whether it will "work"), one must engage in the knowledge-making process in a way that respects the other participants (and perhaps some non-participants as well). Engaging in a dialogue with them - i.e. participating in a social practice of speech - is one way of ensuring that one acts with the appropriate respect. Speech is not, of course, the only way to guarantee such respect, but it is an extremely useful mechanism for placing people in the appropriate relationship and creating pressures and incentives for them to stay there.
 This connection between speech and respect goes beyond a merely theoretical relationship: the social practice of dialogue generates some very concrete experiences that provide a foundation for respect. 89 The process of conversation with someone makes it more difficult to think of him as an object; it humanizes one's opponent. It is much easier to demonize an opponent to whom one never has to actually speak. Seeing his efforts to persuade increases one's own willingness to compromise and tends to soften the positions that one takes. Hearing him tell his side of the story makes it easier and more likely that one will make the leap of imagination necessary to see his point of view; in other words, it increases the possibility of empathy, which is conducive to respect. Thus, speech -- both in theory and in practice -- relies upon and generates respect, and respect is a necessary element in the production of truth.Reason
 The connection between speech and reason also makes speech a particularly useful mechanism for knowledge-making. Reason must, of course, be understood in a non-Cartesian sense. Giving reasons is a deeply context-dependent activity in which we try to make sense of our experience. One important part of the process of making sense of experience is the consideration of alternatives to our present perspective. In other words, what it means to be rational -- or to give good reasons -- is, at least in part, to consider the alternatives. 90 The social constructionist model of truth makes clear why such consideration is a requirement for the production of knowledge. If all knowledge is situated, and no perspective can claim the neutrality or objectivity available in a Cartesian framework, then the only mechanism available for insuring that one is not missing something quite important is the consideration of alternative perspectives. As many writers have pointed out, such consideration is a precondition for the only type of objectivity available in a social constructionist epistemology. 91 Such consideration of alternative perspectives is, moreover, part of the moral and political responsibility of the knower in this model. In other words, rationality requires the consideration of alternatives and, in so doing, it promotes the production of truth.
 Speech is connected to rationality because it is an important -- perhaps even a unique -- mechanism for the consideration of alternatives. Speech is the primary way in which we get access to other minds, allowing us to perceive alternatives we would otherwise miss. 92 Indeed, some have argued that communicative activities, uniquely, "contain within themselves the possibility of surveying the alternatives." 93 There is no question that our eyes can be opened to a new way of seeing things by events other than speech: the experience of a serious illness can give one a new perspective on the health care system, for example. But we cannot, and would not wish to, create for ourselves and others all of the sorts of experiences -- some quite painful -- that would be necessary to generate such a shift in perspective. Moreover, some of the changes that would be required -- like growing up with a different skin color -- are simply impossible. Speech is the primary -- perhaps the only -- mechanism through which we can glimpse another perspective without actually changing the circumstances of our own lives. Because of that capacity, speech is the vehicle for reason and an indispensable part of the production of truth.Responsibility
 Just as reason requires that we explore alternative perspectives, the social constructionist model's emphasis on responsibility requires that we open our own perspective to challenge by others on moral and political grounds. Perhaps we could describe the relevant responsibility as a responsibility to make oneself vulnerable to such challenges. The social practices of speech -- both for the speaker and for the listener -- involve such vulnerability. To engage in a dialogue is to make a commitment to a relationship in which one's own identity and values are at stake.
 Again, speech is not the only way in which we can make our identities and values vulnerable to others. But speech is special. For most of these other sorts of vulnerability one must either establish an ongoing relationship with the person to be changed (as in a close personal association) or one must make a substantial change in that person's life circumstances (for example, a total stranger can make a person vulnerable to changes in identity by raping her). Engagement in the social practice of speech can, however, make one vulnerable to changes in identity caused by people whom one has never met and who have altered nothing in the concrete circumstances of one's life. Thus, speech extends our vulnerability in a dramatic way, making us subject to challenge by people who would otherwise find us unreachable because of social and/or physical distance. To make oneself vulnerable to all members of the relevant community in that way is to acknowledge and fulfill part of the responsibility which is a precondition for the production of truth. 94
 Feminists and others who find the Cartesian tradition in epistemology problematic should, nonetheless, refuse to give up on a truth theory of free speech. We need truth for good and powerful reasons: to construct a shared reality for joint decision-making; to argue for deep criticisms of our social conventions and assumptions; to insist on the importance of taking account of the real impact of our decisions on others, and to reassert our relationship with non-human reality. A social constructionist model of truth can accommodate these important functions while retaining a close connection to the themes and commitments of feminism.
 Indeed, the social constructionist model of truth can help us to see that there is a reason why the truth theory of free speech has such intuitive appeal: it captures our sense that truth requires respect, responsibility, and reason and that speech activities promote relationships with these characteristics. Our values are deeply implicated in our truth claims. We can only hope to pursue truth if we recognize and self-consciously address the moral and political assumptions that guide the process of knowledge-creation. Speech is a uniquely valuable mechanism for such reflection because it both relies upon and (hopefully) actualizes the respect, responsibility, and reason that such reflection requires. Truth does and should matter to us; speech really is special when it comes to truth.
 Because truth fulfills such important needs, it can be a powerful engine for mobilizing social and political action. We should not allow those who would defend the existing hierarchies to hold a monopoly on truth. Too often, truth has been used to assuage the longing for certainty,95 a longing that sometimes blinds us to the demands of respect, responsibility, and reason. But "truth" is a word to conjure with, and its magic is not entirely a product of such illegitimate longings. We can, and should, reclaim truth as the symbol for our abiding commitment to live together -- with other people and with the non-human world -- in a way that meets our moral standards. Truth is the ultimate assertion that we are not alone and that our destiny, like our reality, must be created together.
2. The other major contending theories of free speech include theories focusing on the role of speech in the democratic process, see generally Alexander Mieklejohn, Political Freedom (1960); Robert Bork, Neutral Principles and Some First Amendment Problems, 47 Ind. L. J. 1 (1971); Cass Sunstein, Democracy and the Problem of Free Speech (1993); Vincent Blasi, The Checking Value in First Amendment Theory, 1977 Am. B. F. Res. J. 521; the role of free speech in the process of self-realization, see generally C. Edwin Baker, Scope of the First Amendment Freedom of Speech, 25 U.C.L.A. L. Rev. 964 (1978); Martin Redish, The Value of Free Speech, 130 U. Pa. L. Rev.591 (1982); and the negative implications of regulation of speech, see Frederick Schauer, Free Speech: A Philosophical Inquiry 86 (1982); Lillian BeVier, The First Amendment and Political Speech: An Inquiry into the Substance and Limits of Principle, 30 Stan. L. Rev. 299, 301 (1978).
3. For descriptions and criticisms of this epistemological position, see Catharine A. MacKinnon, Toward a Feminist Theory of the State 97 (1989); Alison M. Jagger & Susan R. Bordo, Introduction in Gender\Body\Knowledge: Feminist Reconstruction of Being and Knowing 3 (1989); Jane Flax, Postmodernism and Gender Relations in Feminist Theory in Feminism\Postmodernism 41-42 (Linda J. Nicholson, ed. 1990).
8. See generally The Interpretive Turn: Philosophy, Science, Culture (David R. Hiley, James F. Bohman, & Richard Shusterman, eds. 1991) (providing a variety of viewpoints on "explanatory" v. "interpretive" models of knowledge).
9. See, e.g., David Cole, Agon at Agora: Creative Misreadings in the First Amendment Tradition, 95 Yale L. J. 857, 876-77 (1986); Kent Greenawalt, Free Speech Justifications, 89 Colum. L. Rev. 119, 131-32 (1989); Stanley Ingber, The Marketplace of Ideas: A Legitimizing Myth, 1984 Duke L.J. 1, 15.
15. Chaplinsky v. New Hampshire, 315 U.S. 568, 572 (1942). See also FCC v. Pacifica Foundation, 438 U.S. 726, 746 (1978) (citing Chaplinsky language); Texas v. Johnson, 491 U.S. 397, 431-32 (1989)(Rehnquist, J., diss.).
22. On Holmes' pragmatism, see Thomas C. Grey, Holmes and Legal Pragmatism, 41 Stan. L. Rev. 787, 817 (1989). On his skepticism, see Rogat & O'Fallon, Mr. Justice Holmes: A Dissenting Opinion -- The Speech Case, 36 Stan. L. Rev. 1349, 1368-72 (1984). On his vision of truth, see Steven D. Smith, Skepticism, Tolerance, and Truth in the Theory of Free Expression, 60 S. Cal. L. Rev. 649, 661 (1987).
36. See Joan W. Scott, Experience in Feminists Theorize the Political 34 (Judith Butler & Joan W. Scott, eds. 1992); Steven L. Winter, Human Values in a Postmodern World, 6 Yale J. L. & Hum. 233, 235 (1994).
42. See, e.g., Carol Gilligan, In a Different Voice: Psychological Theory in Women's Development (1982); Nel Nodding, Caring: A Feminine Approach to Ethics (1984); Sara Ruddick, Maternal Thinking (1983).
47. Compare, e.g., Jane Flax, Political Philosophy and the Patriarchal Unconscious: A Psychoanalytic Perspective on Epistemology and Metaphysics, in Discovering Reality: Feminist Perspectives on Epistemology, Metaphysics, Methodology and Philosophy of Science (Sandra Harding & Merrill B. Hintikka, eds. 1983) (suggesting that there is a psychoanalytic necessity behind the connection) with Mary E. Hawkesworth, From Objectivity to Objectification: Feminist Objections in Rethinking Objectivity, supra n.5 at 151 (arguing that the explanations for a necessary connection are unpersuasive).
48. On the education of girls, see generally Myra Sadker & David Sadker, Failing at Fairness: How Our Schools Cheat Girls (1994). On women and political power, see generally Clark, Getting There: Women in Political Office, 515 Annals Am. Pol. Sci. Soc'y 63 (Janet K. Boles, ed. 1991). And on the sexual abuse of women, see generally Catherine A. MacKinnon, Feminism Unmodified (1987).
49. See Robert Post, Recuperating First Amendment Doctrine, 47 Stan. L. Rev. 1249, 1272 (1995). I do not read Wittgenstein as disagreeing with the possibility of intersubjective knowledge, only with the idea that it would take a representational form. In other words, we do not need to share the same "picture", but we do need to share the same way of "going on."
57. See, e.g., Margaret J. Radin, The Pragmatist and the Feminist, 63 S. Cal. L. Rev. 1699, 1726 (1990) ; Joan Williams, Rorty, Radicalism, Romanticism: The Politics of the Gaze, 1992 Wisc. L. Rev. 131, 133-34.
63. Understanding "working" in this way does, however, contradict Cartesian assumptions by confirming the extent to which knowledge claims are permeated by culture and context. For example, if it turned out that the bridge in question was one that I was building in my garden strictly for the purpose of my visual appreciation, then it is not at all clear that my choice did not work. If I could not have appreciated the bridge if it had been made of the symbolically hateful material, then it fulfills its purpose better being built the way it was, even if it is unable to support actual traffic (which it won't have). In other words, "working" is defined, not by reference to some external and objective reality, but by reference to ethical considerations applied within the social and individual context of the decision.
65. Thomas, supra n.52 at 575. Nonetheless, it is clear that he finds these longings disturbing. He connects them rhetorically with fascist and totalitarian impulses. See, e.g., id. at 540 ("the goal of a totalized or universal rational discourse," a "form of `nostalgia for the whole'"), 575 ("the dream of a univocal communal discourse").
67. Thomas, supra n.52 at 550. See also Joseph William Singer, The Player and the Cards: Nihilism and Legal Theory, 94 Yale L.J. 1, 66 (1984) ("'The alternative to "foundations" is not "chaos" but the joint reconstruction of social life . . . the quest of participatory democracy.'") (quoting from Gerald Frug, The Ideology of Bureaucracy in American Law, 97 Harv. L. Rev. 1276, 1386 (1984)).
70. See Lipkin, Indeterminacy, supra n.58 at 630 n. 141. See also Rorty, supra n.60 at 13 (when we reinterpret objectivity as intersubjectivity or solidarity it leads us to ask political questions about the nature and limits of our community).
73. This issue raises, in other words, the old problem of how to conceive of human freedom. I am suggesting that, for a social constructionist, neither simple determinism nor uncaused human activity is a reasonable possibility. We must, instead, break the link between "uncaused" and "fre.e. Once we break that link, we can recognize that caused action may, nonetheless, qualify as free in the only relevant senses: it implicates moral responsibility and it allows for the experience of freedom that our culture, at least, so highly values. For further discussion of these issues, see Williams, A Feminist Reassessment of Civil Society, supra n.34 at 438; Susan H. Williams, Review Essay: Utopianism, Epistemology, and Feminist Theory, 5 Yale J. L. & Feminism 289, 308-09 (1993).
74. See Donna Haraway, Situated Knowledges: The Science Question in Feminism and the Privilege of Partial Perspectives, 14 Fem. Stud. 575, 583 (1988); Schemann, supra n.39 at 184; see also id. at 196.
80. See, e.g., Steven Winter, Indeterminacy and Incommensurability in Constitutional Law, 78 Calif. L. Rev. 1441, 1447-48 (1990) (using the work of French philosopher Maurice Merleau- Ponty to shed light on issues of constitutional interpretation).
81. See Daniel A. Farber & Suzanna Sherry, The 200,000 Cards of Dimitri Yurasov: Further Reflections on Scholarship and Truth, 46 Stan. L. Rev. 647, 653-4 (1994). See generally Daniel A. Farber & Suzanna Sherry, Beyond All Reason: The Radical Assault on Truth in American Law (1997).
82. See Ruth Bleier, Sex Differences Research: Science of Belief? in Feminist Approaches to Science 147 (Ruth Bleier, ed. 1986). See also Alison Wylie, Kathleen Okfuhlik, Leslie Thielen- Wilson, & Sandra Morton, Philosophical Feminism: A Bibliographic Guide to Critiques of Science, 19 Resources for Fem. Res. 2-36 (1990). See generally Ruth Bleier, Science and Gender (1984).
84. See, e.g., Mircea Eliade, The Sacred and the Profane: The Nature of Religion 10 (1959) ("all that goes beyond man's natural experience, language is reduced to suggesting by terms taken from that experience").
87. The following arguments focus on speech in the form of language, which is the most common form in the first amendment case law. Nonetheless, any form of communication that involves reflection on symbolic and/or conceptual systems should qualify as "speech" within this approach. Thus, visual arts and music, along with theater and film, may also include the connections to respect, reason (as understood above), and responsibility that make speech special. The inclusion of these other media of communication is consistent with the case law under the first amendment, which has long recognized non-verbal forms of communication as protected speech.
88. For a full presentation of this argument, see Susan H. Williams & David C. Williams, A Feminist Theory of Malebashing, 4 Michigan J. of Gender & L. 35, 81-97 (1996). The notion that dialogue entails some standard of respect is, at this point, almost a cliche in legal scholarship. See, e.g., Lipkin, Kibitzers, supra n.68 at 110; Seyla Benhabib, Situating the Self: Gender, Community and Postmodernism in Contemporary Ethics 31 (1992).
91. See Sandra Harding, Whose Science? Whose Knowledge? 155-56 (1991); cf. Rorty, supra n.60 at 13 (reinterpreting objectivity as intersubjectivity or solidarity); Singer, supra n.67 at 35 ("All objectivity means is agreement among people.").
94. By using an example such as rape, I hope to highlight the fact that such vulnerability is dangerous, not entirely within our control, and not always a good thing. Speech, of course, usually poses less danger than many other sorts of vulnerability, but even speech can be used in ways that are damaging rather than productive of truth if the vulnerability on one side is not matched by respect and reciprocal vulnerability on the other. Because the function of speech in this alternative truth theory relies on the relationships it assumes and creates rather than on the substantive content of the speech, the shape of First Amendment doctrine might look quite different if we took this model of truth seriously. I hope to pursue the doctrinal implications of this approach in a later article.