Reclaiming Truth

Catherine Elgin has usefully diagnosed a "bipolar disorder" that continues to incapacitate philosophy and much of contemporary social theory and that inflicts its unwitting sufferers with a perpetual oscillation between equally unhappy alternatives. As she puts it:

Unless answers to philosophical questions are absolute, they are arbitrary. Unless a position is grounded in agent-neutral, determinate facts, it is right only relative to a perspective that cannot in the end be justified.2

Following Elgin, I will define the absolute position as one committed to the belief in determinate truths, as oppose to relative or pluralist ones, and to the possibility of discerning truth in a way that is agent-neutral, or better, agent-transcendent. Both those espousing absolutism and arbitrariness share this conceptualization of truth. The difference is simply in whether or not they are fatalistic in regard to its acquisition.

Many who want to cure philosophy and contemporary social theory of this pathology and transcend the dualism of the absolute and the arbitrary argue that we need to leave behind truth talk, and many who take this line of argument see themselves as following in the pragmatist tradition. By truth talk I mean here not simply the use of the word "true" but the idea that we can assert more than assertability about our most justified or likely claims, that truth is therefore substantive rather than redundant, that it is not collapsible to or a mere extrapolation from procedures and concepts of justification. In short, truth talk brings in the world. Those who reject truth talk say that it unnecessarily creates absolutist requirements and makes everything non-absolute look like it can have nothing to do with truth and must therefore be arbitrary. Those who argue in this way sometimes say we should aim for edification or for understanding or for utility; others say we can retain truth as long as we empty it of content and thus disarm it.

I will argue in this paper that the cause of transcending the bipolar disorder of the absolute and the arbitrary is not served well by dispensing with truth talk. Elgin herself argues against truth talk at times, as does Rorty, which should indicate that the repudiation of truth talk can be made for very different reasons. Rorty wishes to dispense with a metaphysical description of what we know in favor of an aesthetic one,3 while Elgin merely wishes to forego the application of representational models to every arena of inquiry. Many philosophers are motivated to move away from truth precisely because of their concern with representation, as if to be true a statement must represent a bit of transcendent reality, where the latter is defined by Dummett as "recognition transcendent," that is, transcendent of any human being's ability to recognize its truth status. The problem for Elgin is with representation's hegemony over conceptions of truth, while for others the problem is with the very notion of representation. How can we recognize that which is recognition-transcendent as recognition-transcendent? We obviously cannot, but we can recognize meaning, verifiability conditions, instrumental utility, and the normative appropriateness of social practices, so some suggest that we remove the world-condition from truth—since it would seem to require the status of a claim to be recognition-transcendent—and define it in one or more of these terms.

Some philosophers have also pointed out that to characterize truth as a correct representation of an independent, unmediated world has the nefarious political effect of allowing the one who possesses such a truth to transcend the human world of mediation, and thus the give and take of discussion among fallible inquirers. Simone de Beauvoir was one of the first to describe this problem. In The Second Sex she explains that

in his hands, as [woman] knows, masculine reasoning becomes an underhand form of force; men's undebatable pronouncements are intended to confuse her. The intention is to put her in a dilemma: either you agree or you do not....in yielding to him, he would have her yield to the convincingness of an argument, but she knows that he has himself chosen the premises on which his rigorous deductions depend. As long as she avoids questioning them, he will easily reduce her to silence; nevertheless he will not convince her, for she senses his arbitrariness. And, so annoyed, he will accuse her of being obstinate and illogical; but she refuses to play the game because she knows the dice are loaded. 4

It might be possible, of course, to open up the game more democratically, and thus to retain a determinate concept of truth without loaded dice, but many have been skeptical at the feasibility of doing this since any characterization of truth that transcends justification would, they think, remove the motivation to listen to alternative or newly developed justificatory considerations.

These various arguments against truth talk can be loosely grouped under four categories: semantic, metaphysical, epistemological, and political, though in many cases of a particular critique, such as Dummett's or Derrida's, more than one rubric is involved. The semantic argument is based on the view that truth talk does no work, or no good work, in the language. It adds nothing substantive to the content of a claim, and any substance it does add is dubious at best. The metaphysical argument characterizes why it is that giving a substantive content to truth is considered dubious: because it offers to characterize a relationship between thought and reality, for example, which cannot be characterized without begging the question. The epistemological argument refers not to what can be stated intelligibly but what can be known, and many agree with Dummett that truth cannot be known because truth, unlike verifiability or assertability, is recognition-transcendent. The concern of the political critique is that, precisely because it postures as recognition-transcendent, truth talk is a kind of discursive violence, a speech act whose goal is to close down discussion.

The semantic and metaphysical arguments largely motivate, I believe, the epistemological and political arguments. Political concerns, in and of themselves, would not be sufficient to turn away from truth unless one thought that truth was at least suspicious-looking for other reasons as well.5 The epistemological arguments—that we cannot know the truth—depend heavily on how we understand what the truth is that we are supposed to know, and thus depend upon its semantic and metaphysical characterization. Thus I believe that the main grounds of critique are metaphysical or semantic or, what is often the case, some combination of these two.

This is a large conversation with many interlocutors. To make my project manageable, I will look at just two of the participants in this discussion: Richard Rorty and Hilary Putnam, who are today surely the main competitors for the title of head pragmatist. Both Rorty and Putnam repudiate absolutism and the possibility of recognition-transcendence, and thus both have adopted some of the main premises on which the repudiation of truth relies, but they have come to different conclusions nonetheless about the viability of truth and representation. To compare their positions, I will take up a specific example of a recent feminist argument in the discipline of history, in order to consider just how plausible, or relevant, any of the arguments pro and con truth talk appear in relation to this example. The example comes from Nancy Armstrong and Leonard Tennenhouse's excellent, recent study of personal life and the emergence of the English middle class in The Imaginary Puritan: Literature, Intellectual Labor, and the Origins of Personal Life.

Philosophers too often pick relatively easy cases, such as simple perception, or claims in the natural sciences that have a lot of empirical evidence and appear neutral, such as the existence of atoms or electrons. (I know that philosophers take pride in turning relatively easy cases into unsolvable conundrums, but even when turned into conundrums these kinds of cases are still of a different order than the case I will be discussing; in this kind of case there certainly is, as Peirce would say, genuine doubt). The question of truth is much more difficult, and arises more ordinarily of its own accord without the meddling of philosophers, in more complex, multi-variable explanatory accounts or theories in the social sciences. In cases where empirical evidence is at least a part of the argument, but the grounds for justification are highly interpretive, can we ever claim truth? Even if we think not, it is not so easy simply to dispense with this arena of inquiry as inappropriate to truth talk, since it spans received knowledge from evolutionary biology to Napoleon's defeat at Waterloo. Moreover, there is much at stake in these debates in the social sciences, much more than in the question of how to characterize electrons ontologically.

I chose the particular example I will discuss for two main reasons. First, because it is explicitly feminist, and thus useful because some will be suspicious about its truth status just on those grounds: how can a claim be both agent-transcendent and politically motivated? Yet every large claim in the social sciences necessarily begins with some assumptions, and the choice of assumptions almost always reflect some broad political values. It has become especially clear in the domain of historical narrative—the revisions of which continue to elicit debate even in legislative chambers—that political values inform the choice of narrative, as between, for example, a story of "discovery," an "encounter," or an "invasion." Nor can we simply add such various accounts together to achieve the truth; they often directly conflict. Thus, arguably, feminist arguments simply make explicit what is there all the time.

My second reason for choosing this particular example is that the feminist historians I will discuss are on the side of dispensing with truth. Inspired by deconstruction, Armstrong and Tennenhouse refuse to describe their claims as more truthful or likely to be true about the actual historical past. They approve of Foucault's approach, whose "histories no more presume to say what things, people, words, thoughts, or feelings are now than they do to say what these things used to be" and who therefore wanted simply "to demonstrate how they were written into existence in one way rather than another."6 In other words, the truth claims made by historians can be only about representations, without shedding any reliable light on the actual content or what is being represented. They also agree with Geoff Bennington's view that "The claim to be able to discern the real continuities and thus to ground those fantasies at least partially in Œtruth' depends simply on the illusion of an intelligentsia as subject of science to stand outside and above reality and those fantasies."7 Thus, their retreat from truth is motivated precisely by the view that in the absence of the absolute, truth claims about the actual historical past have to be let go in toto.(6-7) I believe and will argue that such a rendering of their argument is unnecessarily belittling of it; they are in fact arguing over the historical truth.

Let me turn now to what will have to be a brief and truncated rendition of the example. In a series of powerful critiques, Armstrong and Tennenhouse have analyzed two apparently contradictory historical accounts of the formation of the family in seventeenth and eighteenth century Britain. Though both of the accounts that Armstrong and Tennenhouse critique argue for different accounts of family history, they both privilege a normative rendition of the nuclear family with a fairly traditional gendered division of labor, one in which children "need their mothers and obey their fathers," and they both assume that such families are both natural kinds and natural goods because "a small number of individuals who are together for a long time without outside interference tend to care for one another as for themselves."8 In other words, these accounts both take the affective ties that emerge from that sort of family as "exempt from history." 9

The first account that Armstrong and Tennenhouse critique is Peter Laslett's highly influential history of the British family in his The World We Have Lost (which began in the 1950's as radio broadcasts, was published as a book in 1965, and went on to become "one of the most frequently cited books on the topic.")10 According to Laslett, "Time was when the whole of life went forward in the family, in a circle of loved, familiar faces, known and fondled objects, all to human size. That time has gone forever."11 Laslett's thesis is that in the pre-industrial family of early modern England, in which work and family were combined in one unit and one location, there were fewer people interacting regularly together, they spent more time together, and as a result "enjoyed a closer emotional bonding than was the case during the modern period."12 Moreover, "Englishmen...felt they were parts integrated into an organic whole"13 with the result that neither modern alienation nor class antagonism existed. Armstrong and Tennenhouse explain that

By an almost invisible logic of internalization, [Laslett] reasons that even Œthe head of the poorest family was at least the head of something.' That each of them was on top of some little heap of humanity apparently made it possible for heads of households to identify with people higher up on the social scale in a way that became impossible once the workplace was detached from the home.14

Laslett goes so far as to characterize pre-modern England as a "one-class society."15 One of the most important implications of his claim, and what came under much debate later, is that the political upheavals in England of the 1640's and 50's had no impact on the basic way most people lived or understood themselves; only the destruction of the nuclear family through industrialization could bring significant change, and the changes it brought about for people's emotional and personal life was, for Laslett, all to the bad.

Armstrong and Tennenhouse also look at Lawrence Stone's equally influential history of personal life in his book The Family, Sex and Marriage, 1500-1800, which argues, against Laslett, that family ties that were volitional rather than founded as economic units made for a much happier life. Stone also argues that privacy and size made an enormous difference in the capacity to develop happy relationships, and it was only after what he names the "open lineage family"—Laslett's ideal type—becomes replaced by the "closed domesticated nuclear family"—Stone's ideal type—that the household became the site of personal happiness. In regard to the open lineage family, prevalent in the sixteenth century, Stone bemoans the fact that "relations within the nuclear family, between husband and wife and parents and children, were not much closer than those with neighbors, with relatives, or with friends."16 The closed domesticated nuclear family, by contrast, was the product of what he calls "affective individualism," in which the privacy surrounding the family somehow constituted privacy for individuals within the family, wherein each could develop personal autonomy.

Stone also takes issue with Laslett's preferred family because of its treatment of children. In the early modern period, the use of wet nurses and the widespread tendency to hire children out ("about two out of every three boys and three out of every four girls were living away from home" from just before puberty until their marriage17 ) made it virtually impossible to have a "single mothering and nurturing figure." Stone sees this as the "denial" of maternal affection and he uses this fact to explain both the passionate religious enthusiasms of the period as well as its high degree of casual violence and antagonism, on the grounds that the natural emotion rightfully found in mother-child relations had to be deflected into other channels.18

Where Laslett paints a regressivist story in which we have lost a world of happiness and equality, Stone offers a progressivist history in which the chances for personal happiness have been enhanced. They differ in the value they confer on privacy, and on the optimism or suspicion by which they regard families based on economic relationships. But Armstrong and Tennenhouse argue that, despite these important differences, both Laslett and Stone make naturalistic assumptions about the impact of family structure on affective life, and they both privilege traditional gender roles within the family, including especially the role of the mother as almost the exclusive nurturing figure.19 Thus, Armstrong and Tennenhouse charge both Laslett and Stone with romanticizing and revering the traditional family and neglecting to historicize their own beliefs and preferences about personal life. This cultural terrain is, as Armstrong and Tennenhouse point out, "as close as one comes to sacred ground in a modern secular culture."20

Armstrong and Tennenhouse's critiques are first and foremost based on their claim that there is, to put it mildly, questionable evidence for Laslett's and Stone's various claims about the affective history of the family. They make some of the very traditional empirical charges that historians use to challenge each other's accounts, that claims are based on generalizations from evidence that is insufficient, too limited in its scope, and too amenable to contrary interpretations. But the most interesting aspect of their critique for our purposes is that they charge Laslett and Stone with using history to offer support for contemporary ideological convictions espoused in present day pop psychology as well as embedded deeply into our collective common sense. Thus, they argue that historians cannot use their own emotional proclivities or current beliefs and practices in regard to personal life as any kind of ground to theorize the affective lives of people long since dead. They argue, in other words, that interior life itself needs to be historicized. I take it that what it would mean to historicize interior affective life is not just that one will recognize that personal life has undergone changes and to chart those changes—both Laslett and Stone do that—but to recognize the possibility that our needs and wants, the conditions necessary for our personal happiness, and the texture of our emotional bonds, can change. And not just change in the sense of shrinking or atrophying or developing and flourishing—which would presume a single, unified process and character to human life—but actually change in content and causal effect. If this is right, one might well be led to think that truth-claims the historical interpretation of affect are simply impossible.

But even more than this, I find their critique interesting and useful in this context because, while reading through their book, one cannot help but develop the firm conviction that Armstrong and Tennenhouse are also working with assumptions, and that it is these assumptions that play a critical role in their ability to perceive the weaknesses in Laslett's and Stone's accounts. In other words, all of their criticisms cannot be put in the form of a Pyrrhonic skeptical question which takes an agnostic position equally to all claims. Some of their assumptions they make explicit, others they don't (and I think their argument would be more persuasive if they did). But it raises the obvious question of whether their arguments are any more legitimate than those they critique. If all historians must work with some assumptions when they try to make sense out of the din of history, and if at least some of these assumptions cannot be proven by uncontroversial empirical methods, then perhaps the deconstructionists are right and we need to read history exactly as we read literature.

What are the assumptions made by Armstrong and Tennenhouse themselves? I think there are at least three we can gather just from their critique of Laslett and Stone. The first is that the traditional gendered division of labor in the family is not a manifestation of human nature. This is suggested in part by their demand that interior life be historicized, which of course assumes that interior life can be historicized. This is not a claim grounded in the actual existence of sufficient evidence, but a metaphysical claim about the flexibility of the human self. Even if it is entered here just as a hypothesis that warrants investigation, it is a truth claim, or a claim that the hypothesis might well be true.

That assumption seems to me to be presupposed as a logical necessity by Armstrong and Tennenhouse's overall argument. Other assumptions have weaker relations to their argument, but still seem to play a guiding role in the path they take through this material. For example, one might reasonably suppose that Armstrong and Tennenhouse want to work with the assumption that women can have the same general wants and needs as men. It is this assumption that would cast doubt on the claim that a patriarchal form of the family, in which the roles and power of father and mother are neither equal nor reciprocal, would be an optimal form of the family from the point of view of personal happiness. Laslett relates without comment that in the days of yore, England was an association between the male heads of wealthy families, and that the father ruled the family in more than name only. He does not consider this prima facie evidence for the possibility that the women in these families will experience unhappiness; Armstrong and Tennenhouse clearly do.

A third assumption that Armstrong and Tennenhouse make is that the closed domesticated biologically related form of the family that Stone prefers is not necessarily the best form of family in terms of its effects on society. Stone argues that there are a number of social and political advantages to such a family, in creating the possibility of individual autonomy that will then find its way into anti-authoritarian political movements, for example. Armstrong and Tennenhouse remark that, in criticizing what he balefully calls the exchange of children, Stone "apparently cannot imagine...that the presence of other children in the family might have extended the sense of closeness to a community beyond the biological family."21 This is a possibility Armstrong and Tennenhouse clearly see as a potential social good. This is a truth claim.

Some of these assumptions even look dangerously close to being generalizations, such as the assumption that women have the same basic wants and needs as men. Given their demand for the historicizing of everything, surely Armstrong and Tennenhouse cannot countenance a cross-historical generalization of this sort. But here it should be noted that the demand that we historicize everything does not entail that we will then find that absolutely everything changes; it is simply a demand that we not assume simply on the basis of current sentiment what can and cannot change. We should, in other words, hold nothing back from the cultural historians' examination.22

All of these historians, Armstrong and Tennenhouse no less than Laslett and Stone, are working with assumptions and even a political orientation. But not all assumptions have the same kind of epistemic impact. Thus, we can agree, along with Putnam, with James's claim that all knowledge is mediated without having to then agree that any given mediating influence is equal in its epistemic content to any other. One of the ways assumptions can operate in the production of historical narrative is to make some things appear and others disappear. Because Laslett privileges patriarchy, the particular point of view women may have had on the families he idealizes don't come into view, at least not fully or with prominence. In fact, he doesn't even mention them, nor is gender thematized in The World We Lost. Stone takes as a given that a central, nurturing maternal figure—not paternal—is necessary for children's well-being. This assumption operated to preempt asking certain kinds of questions, from which other possibilities might have come into view. Armstrong and Tennenhouse, on the other hand, clearly have women in mind when they offer some of their critical analysis about the way in which Laslett and Stone have naturalized a traditional gendered division of labor in the family.

I am not making an argument for a prima facie privileging of any and all feminist assumptions, for one thing because feminists often disagree but also because feminists can simply be wrong. One such controversy relevant here is precisely over the sort of individualism that Stone champions. One might well think, at first blush, that individualism is in the interest of women, but many have rejected this claim. The individualist ideology of freedom and happiness assumes that all associations must be volitional for there to be just or happy relationships, which is a model of intersubjective relations based on public associations in voluntary organizations. Families are not like that; in fact, neither are communities. We are born into relations with specific others; we give birth to others and thus bear a necessary emotional relationship to them. These are never volitional—we may choose to become a parent but we cannot in general choose who we will become a parent to. Traditional liberal individualist notions of human relationships have been unable to effectively evaluate and analyze such non-volitional relationships; thus they have tended to ignore them, following the Hegelian dogma that family relations belong to the sphere of nature, not the sphere of culture. That kind of claim is definitely not in the interest of women, since it exempts familial relationships from political critique and suggestions for change, but feminist ethicists also have argued persuasively that the kinds of non-volitional relationships born out of families and communities can enhance autonomy, and can also be subject to political and moral judgement. Early feminist theorists who made these very individualist assumptions—valorizing volitional relationships over non-volitional ones in all cases, for example—have been critiqued, quite persuasively.23

Thus I am not championing feminist assumptions in all cases. But at the very least, the assumption that women count, that we may have an independent point of view on things, that we may have the same wants and needs as men, and that our optimal life situation is probably not to be found in a condition of life long subordination, are assumptions proven useful in illuminating new aspects of the historical record unseen before the recent period. To argue for an epistemic equality between these assumptions and blatantly patriarchal ones—such that we can forego listening to what women say because they don't know their own needs, for example—is surely ludicrous. In this light, I find Putnam's project very interesting, which he has recently (re-)stated as the project to show how a realist commitment, which he takes to be opposed to James, can be squared with his acceptance of James' claim that perception is never unmediated, without the two beliefs together leading to skepticism. It seems to me that the there are many "real world examples" such as the one I just discussed which manifest the possibility of squaring these two claims.

But what about truth? As I said, I picked the Armstrong and Tennenhouse example because they retreat from truth. Although they make truth claims throughout the book, when asked to give a kind of meta-characterization of the epistemic status of their arguments, Armstrong vigorously denies the referential character of her claims.24 She is just offering us a narrative, to be judged by its effects in the present on discourses and practices. She might be able to agree with the claim that her arguments have assertability, but she will not claim anything approaching truth about the past. She is, in effect, a Rortyan.

But a narrative can be true or false: narratives tell a story about the world. Even fictional narratives offer accounts about true things indirectly: true ways in which human beings can respond to each other, can be affected by a given experience, can fall into trouble, or pull themselves out of trouble. Although we may compare narratives by what they each allow us to see or appreciate anew, and we may grant that multiple and even conflicting narratives can be informative about a given event, the value of a narrative generally rests on the quality and depth of its relation to the world. In this sense, a narrative is very different from a conversation, which does not require a relation to the world for it to be good or meaningful; conversations can resemble lovemaking, play, or chess matches (and philosophy conversations often resemble the latter), with or without a relation to the world.

Rorty's repudiation of truth is based on concerns he has with all four of the kinds of arguments I listed above: semantic, metaphysical, epistemological, and political. In general, Rorty has argued that truth talk merely gets in the way of conversation, posing a requirement that is as unnecessary to conversation as it is likely to lead the conversation off to a dead end. And Rorty of course portrays himself as carrying on the pragmatist tradition by this argument. To be accurate, Rorty does not argue against any use of the word "true" but against a specifically philosophical concern with the word or the concept. In itself, the elimination of a metaphysical project to understand the meaning of truth does not preclude us from calling some historical accounts true, depending, of course, on how one construes that metaphysical project. But the question does arise when Rorty eliminates talk of representation because then the world-content of a historical narrative would be dropped out. By his account, we can call Armstrong and Tennenhouse's account true but we cannot really claim that it represents any truths about the way things really were in pre-modern Britain, in so far as we understand ourselves not to be merely participating in the contestations among historians over how to construct historical narratives but in so far as we are seeking to know the real nature of the past.

Now it may seem as if this is pushing Rorty's anti-metaphysics too far. We can make ordinary claims, after all, and claims about the past are ordinary claims. To say that truth is a primitive is not to deny its existence. But for Rorty, unlike for example Donald Davidson, to say that truth is a primitive preempts even the possibility of claiming an extra-epistemic meaning to truth, or its relation to an objective world not of our making.25 Truth is "what is good for us to believe," full stop, and the gap between justification and truth, or justified belief and true belief, is simply the gap between the "actual good and the possible better."26 By contrast, one could bring use in at critical points along the way of inquiry, as Elgin does for example, without it preempting the possibility of giving a world-content to truth. Use here plays the role of mediator, which can reveal or direct us toward certain aspects of reality. It does not serve the cause of transcending the absolute and the arbitrary to present use and objectivity as mutually exclusive choices.

In contrast to Putnam, Davidson, Elgin and others of the pragmatist tradition, Rorty seems to retain rather than argue against the binaries that have structured both foundationalist and postmodern treatments of knowledge and, in particular, the binary between representation and construction. This becomes apparent, for example, in his reading of Davidson. Rorty argues that Davidson's coherence theory of truth amounts to a kind of constructivism—a belief in the making rather than the discovering of truth. And then he argues that "since Œmaking true' is the inverse of Œrepresenting, ‘this doctrine makes it impossible for Davidson to talk about language representing the world—standing to it as scheme to content.’27 He argues further that Davidson "marries" truth and meaning to each other in such a way that the theory of truth (or truth/meaning) that results "will be of no use to a representationalist epistemology, nor to any other sort of epistemology [because it is] an explanation of what people do, rather than of a non-causal, representing relation in which they stand to non-human entities."28 I will set aside the question of whether this is a persuasive reading of Davidson for the moment, to simply note Rorty's contrast between these two types of explanations of truth in so far as it is a conceptualization of the human-world relation: we have the choice of either a non-causal, that is, uninterpreted, relation, or a making relation. I am not inclined to defend the non-causal account, but to explore the possibilities of a third way to describe the relation, a way which in fact shows that "making" and "representing" are not mutually exclusive truth operations.

This is Putnam's latest project, or latest formulation of what his project has been all along. In contrast to Rorty, Putnam does not dispense with truth talk in the sense of a relation with the world, nor even of realism. Though he shares with Rorty the view that a metaphysical project of elucidating the interface between thought and reality is nonsense, he does not go as far as Rorty in dispensing with all forms of metaphysical talk. The differences between Rorty and Putnam are especially interesting to look at because both are more Jamesian than Peircean, especially in their critique of scientism in philosophy and their tendency to psychologize philosophical quandaries.

In his latest book, The Three-Fold Cord, Putnam takes us once again beyond his previous views, or rather, takes his earlier self to be his greatest foil. He argues now against metaphysical realism, internal realism, and pragmatic realism (all positions that he once held) and argues for a form of natural or direct realism. Direct realism is naive realism (what we believe to be true by our best lights is true about the world) but it has a second-order naivete, having rejected initial naivete and then moving back to the substance of the naive position after having tried, I suppose, sophistication. It's similar to Nietzsche's notion of the adult playing at playing like a child, thus retaining both the status of sophistication with the benefits of frivolous innocence. The difference between the adult playing like a child and the child playing is that the adult knows that s/he is playing like a child, knows the alternatives, and has made a choice.

As my sarcasm may indicate, I don't think Putnam's second-order naivete works as naivete. One cannot, after all, return to a carefree bliss in the Garden of Eden once one has seen what lies just beyond the gates. Putnam's realism, thus, and his notion of truth, retains some level of its previous sophistication, and thus has a content. Let me explain what I mean.

Putnam argues that direct or naive realism correctly holds that "the world is as it is independently of describers."29 As I mentioned earlier, one of his aims in this new book is to show how that realist commitment can be squared with the fact that perception is always mediated. Thus, he wants to counter the skeptical conclusions argued for by those who, like Dummett, have realist commitments in their account of what is required for truth but acknowledge that neither human inquiry nor language can transcend its clay feet and thus meet the requirements. As I read it, Putnam's strategy has two stages: (1) to argue against, once again, one of the primary ways these clay feet have been conceptualized—in terms of the "interface" idea in which sense-impressions, qualia, mental representations, or some such are put between human beings and the external world, and (2) after having vanquished this idea, to retrieve the meaningfulness of the concept of representation without it being entangled in the assumption of an interface.(59) Putnam argues, persuasively in my view, that the concept of representation must be retrieved if we are to retain the possibility of veridical discourse, which is precisely discourse that goes beyond conversation to make claims about the world that are in fact true.

Putnam thinks that it is the "interface" idea that keeps mediated inquiry from plausibly achieving relations with the world; without the interface, representation is free to refer to the world rather than to our image of the world. And thus we can return to a naive realism. But it is not really a naive realism that he returns to for, according to Putnam, representations are not thing-like entities at the interface of human beings and the world but practices. And it is because they are practices that we can understand the mediated nature of perception without becoming anti-realists. He uses Wittgenstein's duck-rabbit and Cora Diamond's discussion of two picture faces that have the same expression as examples of the way in which representations can be both real, or accurate as representations, and mediated. In Wittgenstein's example, a single picture can be seen equally well as a duck or a rabbit. In Diamond's example, two pictures of faces represent the same expression despite the fact that it is impossible to point to features of the faces that they have in common and that engender the expression. In each of these examples, one cannot point to anything different about the drawings themselves, anything materially different about them, to explain either the distinction we make on the one hand or the similarity we find on the other.

Seeing an expression in the picture face is not just a matter of seeing the lines and the dots; rather, it is a matter of seeing something in the lines and the dots—but this is not to say that it is seeing something besides the lines and the dots.30

By this analogy, Putnam suggests, we can conceptualize the relation of human inquiry to the world. The world "by itself" does not cause us to see a duck or a rabbit, and yet the shapes are there in the world and not merely in our minds. We can affirm simultaneously the fact that the world does not force us to choose duck or rabbit and that our claim to see a duck represents a truth about the world, and not just about human perception or human practices, though it may also be about those things.

This, however, is hardly a naive realism. In its substance, it is still the internal realism that Putnam developed in his middle period and has been denying ever since, in that it produces a combination between the aboutness claim of realism and the ontological relativity thesis of pragmatism. It works this out by making a claim about the world that can explain, not how it is possible to have truth at all (which is the metaphysical project Putnam rejects along with Rorty), but how it is possible to have many truths. It is, then, a realism in its claim about the content of truth claims but an internal realism since it holds that human practices must be taken into account to understand which truths will be accepted, or how the world will be seen, at any given moment.

The swing between the absolute and the arbitrary is caused by a conception of truth as determinate and agent-transcendent. But truth is neither of these things. Even in regard to historical argument about the past, where extrapolations are large, complex, and always positional, we aim at the truth, and we can be more or less successful. The mistake is to think that in aiming at the truth we can hit it or miss it, as if truth is an "it." Thinking of truth as an "it" is what makes us think we cannot claim truth. But truth is as dense and multivalent as lived reality, which is, after all, what it is about.

FOOTNOTES

1. I am indebted to Marianne Janack for very helpful discussions about the arguments of this paper. I am also grateful to Nancy Armstrong for her feedback on an earlier version.

2. Catherine Elgin, Between the Absolute and the Arbitrary (Ithaca: Cornell University Press, 1997), p. 1.

3. Whether his criterion is ultimately aesthetic or political in the end does not matter, since Rorty is among those who equate the two, thus arguing, for example, that a criterion of openness serves poetic vision and political judgement in equal measure, and that, in fact, the two are co-constitutive.

4. Simone de Beauvoir, The Second Sex translated by H.M. Parshley, (New York: Random House, 1952), p. 612, my emphasis.

5. Alternatively, one could argue that the political arguments against truth show that there are extra-epistemic reasons for some to hold onto truth, and this provides a prima facie case that we should take a careful look at the metaphysical basis for truth talk. One could read Lyotard and Derrida in this way.

6. Nancy Armstrong and Leonard Tennenhouse, The Imaginary Puritan: Literature, Intellectual Labor, and the Origins of Personal Life Berkeley: University of California Press, 1992, p. 4. In their Introduction, Armstrong and Tennenhouse explain the epistemological assumptions of their study. They claim that "To overturn history, one simply has to demonstrate that words come chronologically as well as ontologically before the things they are presumed to represent and the differences that already exist among those things. Those of us who are willing to entertain this possibility have had little difficulty finding evidence to substantiate the inversion of traditional historical priorities"(p. 4).

7. Armstrong and Tennenhouse, pp. 6-7; quoting from Geoff Bennington, "Demanding History," in Post-structuralism and the Question of History edited by Derek Attridge, Geoff Bennington and Robert Young (Cambridge: Cambridge University Press, 1987), p. 25. They actually differ with Bennington's skepticism about the epistemic basis of historical narrative, but only because they want to redefine narrative as a process of discursive self-referring, a "function of the surface," indistinguishable from writing. Thus, rather than complain about history's groundlessness, they shift the historian's focus to the imaginary itself. Paradoxically, they then make the claim that they can give the history, in precise detail, of the emergence of this imaginary!

8. Armstrong and Tennenhouse, p. 84.

9. Armstrong and Tennenhouse, p. 71.

10. Armstrong and Tennenhouse, p. 71.

11. Armstrong and Tennenhouse, p. 72.

12. Armstrong and Tennenhouse, p. 75.

13. Armstrong and Tennenhouse, p. 73.

14. Armstrong and Tennenhouse, p. 72.

15. Armstrong and Tennenhouse, p. 73.

16. Armstrong and Tennenhouse, p. 76.

17. Armstrong and Tennenhouse, p. 81.

18. Armstrong and Tennenhouse, pp. 81-82.

19. Armstrong and Tennenhouse, p. 84.

20. Armstrong and Tennenhouse, p. 85.

21. Armstrong and Tennenhouse, p. 81.

22. This is not to say that all of their feminist assumptions must be put to the test of history, since I am denying that this is possible. Some feminist claims, such as that women have the same wants and needs as men, can be challenged and debated through historical record, but others, such as that women's own views should always be consulted in assessing the past, cannot be coherently challenged. That is, one can accept it or reject it, and give reasons, but it is doubtful that the reasons given on one side will be intelligible to the other‹such as that women simply don't know their own interests or cannot interpret the world around them. And it is the latter sort of feminist assumption—that women's own views should always be consulted in assessing the past—that I observe working in Armstrong and Tennenhouse's arguments.

23. A clear overview of these issues can be found in Virginia Held, Feminist Morality: Transforming Culture, Society, and Politics (Chicago: University of Chicago Press, 1993).

24. This occurred at a presentation of her work at the Pembroke Center, Brown University, February 2001.

25. See e.g. his "Solidarity or Objectivity?" in Objectivity, Relativism and Truth: Philosophical Papers Volume 1 (New York: Cambridge University Press, 1991), p. 22; and Donald Davidson, "A Coherence Theory of Truth and Knowledge" in Reading Rorty edited by in Alan Malachowski, (Cambridge, Mass.: Basil Blackwell, 1990), p. 123.

26. Rorty, "Solidarity or Objectivity?" p. 124.

27. Rorty, "Representation, Social Practice, and Truth," in Objectivity, Relativism and Truth, p. 153. See also his "Pragmatism, Davidson, and Truth," same volume.

28. Rorty, "Representation, Social Practice, and Truth," in Objectivity, Relativism and Truth, p. 154. 29. Hilary Putnam, The Three Fold Cord: Mind, Body, and World (New York: Columbia University Press, 1999), p. 6.

30. Putnam, p. 63.