Between Autonomy and Authority: Kant on the Epistemic Status of Testimony Joseph Shieber* Lafayette College 1.

Introduction

One of the most striking aspects of contemporary society is the extent to which all aspects of life are affected by the results of specialization: we pay people whose only function is to perform quality controls on widgets, or to design new flavors of coffee, or—at least in the American League—to hit a hurtling leather sphere with a large ashen cudgel. Perhaps it is little wonder, then, that we pay others to think for us as well. In a passage so striking that it must be quoted in its entirety, this feature of modern capitalist life was already described by one of its earliest and most penetrating theorists, Adam Smith. Smith writes: In opulent and commercial societies … to think or to reason comes to be, like every other employment, a particular business, which is carried on by a very few people, who furnish the public with all the thought and reason possessed by the vast multitude that labour. Let any ordinary person make a fair review of all the knowledge which he possesses concerning any subject that does not fall within the limits of his particular occupation, and he will find that almost every thing he knows has been acquired at second hand, from books …. A very small part of it only, he will find, has been the produce of his own observation or reflections. All the rest has been

* For invaluable advice on an early version of this paper, I thank Ernest Sosa and David Matheson. Thanks are also due to my colleagues Owen McLeod, George Panichas, and Julie Yoo, who gave excellent and detailed comments on a more recent version, and to an anonymous referee for this journal, who gave outstanding and highly specific comments that substantially improved this paper, both in its readability and content. I also thank audiences at Connecticut College, Florida State University, Hofstra University, and Lafayette College for their comments on various drafts of the work presented here.

-1-

purchased, in the same manner as his shoes or his stockings, from those whose business it is to make up and prepare for the market that particular species of goods. It is in this manner that he has acquired all his general ideas concerning the great objects of religion, morals, and government, concerning his own happiness or that of his country.1 It is striking to note that Smith does not question that what the “ordinary person” acquires through his or her reading of books is knowledge. Indeed, although the context from which this passage is drawn suggests that Smith is troubled by the ethical implications of this aspect of advanced capitalist society, others of his Scottish contemporaries celebrated it as one that leads to the enlightenment of its citizenry.2 The last two decades have seen a rebirth of interest in the question of whether, and, if so, how, one can acquire knowledge from one’s communications with others. It is therefore understandable that this interest has been accompanied by a rediscovery of debates concerning the status of testimony in previous periods within the history of philosophy. Thus, it is no wonder that a number of works have taken as their subject the views of historical figures, particularly Hume and Reid, on the status of testimony as a source of knowledge and justified belief. Given the current intense interest, however, it is perhaps quite puzzling that Kant’s views on the status of testimony have been largely ignored.3 In this essay I will attempt to fill this lacuna within the burgeoning literature on testimony. I will do so in three stages. In the first (Section 2), I will discuss the reasons leading to the ignorance of Kant’s views on testimony within the context of the

1

From Adam Smith’s “Early Draft” of The Wealth of Nations, quoted in Buchan [2003], 229. 2

Thus, Dugald Stewart rejoices that “[t]he progress of knowledge must be wonderfully aided by the effect of … a free commerce of ideas all over the civilized world; effects, not proportioned merely to the increased number of cultivated minds, thus engaged in the search of truth, but to the powers of the increased number, combined with all those arising from the division and distribution of intellectual labor.” (Quoted in Buchan [2003], 231.) 3

One sign of this is Kant’s absence from the seminal—and otherwise wide-ranging—historical survey of the discussion of testimony in the Western philosophical tradition in Coady [1992].

-2-

larger debate. I will suggest that this widespread ignorance has two causes: the sense that Kant had little, if anything, to say on the status of testimony and the sense that what he had to say would place him in the radical position of denying that testimony is a source of knowledge and justified belief. Following this discussion, in the second stage (Sections 3-5), I will present evidence from Kant’s work, and particularly from his lectures, that he did address the status of testimony, and that he did consider testimony to be a source of knowledge. This discussion, however, raises as many questions as it answers. Since the only sustained passages from Kant’s writings in which he discusses testimony are in the lectures, we do not have from Kant a discussion of the epistemic status of testimony that demonstrates the systematic rigor of his published work. One result of this is that we are left without the sort of detailed discussion of testimony that would allow us to assess how it is that Kant took testimony to underwrite knowledge—the issue that stands at the focus of the contemporary discussion of the epistemic status of testimony. Such a detailed discussion of testimony within the context of Kant’s broader systematic thought will form the body of the third stage of this essay (in Sections 6-11). In it, I will argue that Kant’s system does contain the resources necessary to assess the way in which testimony serves to underwrite knowledge. Indeed, I will argue that an adequate assessment of Kant allows us to recognize him as the natural precursor of one strand of thought that considers testimony to be a basic, non-inferential source of epistemic license— that is, as a precursor to one strand of what, in the contemporary literature, has come to be known as Anti-reductionism with respect to testimony. 2.

Schmitt on Kant on Testimony

One of the obstacles to assessing Kant’s thought on testimony is that, except for a few suggestive remarks in the Kritik der Urteilskraft, there is no extended discussion that impinges upon the question of the epistemic status of testimony in any of the three Kritiken. Thus, one might easily be led to the impression that Kant simply has no interest in the question of the epistemic status of testimony—or, at least, that he remained silent on the question of its epistemic status. Alternatively, one might take Kant’s prizing of autonomy as evidence that Kant would have rejected testimony as a source of

-3-

knowledge or justified belief. This is the route taken by Frederick Schmitt, the one contemporary philosopher writing on testimony to attempt to take note of Kant’s position on the subject.4 Schmitt refers to Kant’s digression on the nature of the sensus communis in §40 of the Kritik der Urteilskraft. There, Schmitt suggests, Kant seems to provide support for the idea that “testimonial evidence, however conclusive, is not the sort on which an intellectually autonomous subject would rely.” (Schmitt [1987], 46) Schmitt notes that Kant characterizes one of the three maxims of sensus communis, “to think for oneself” as “the very motto of enlightenment.”5 This, Schmitt suggests, should be taken as evidence for Kant’s having thought that countenancing testimony would violate an agent’s autonomy. Indeed, as Schmitt points out, even Kant’s “maxim of enlarged thought”, which requires one “to think from the standpoint of everyone else”, merely requires that subjects abstract away from their individual perspectives by considering the standpoints of others. This way of bringing others into one’s deliberations, however, does not in fact require that one take the testimony of those others into account. Rather, Kant’s description of the maxim of enlarged thought is perfectly compatible with one’s adopting the standpoint of others merely through an exercise of imaginative projection on one’s own part into the roles of the others. Clearly, however, such an exercise of imaginative projection requires no role for the thoughts of others in one’s own deliberations. Thus, Schmitt concludes that “Kant remains an individualist in the same sense in which Locke is an individualist” (Schmitt [1987], 47)—a reference to Locke’s famous claim that “we may as rationally hope to see with other men’s eyes as to know by other men’s understandings.” (Locke, quoted in Schmitt [1987], 43) Indeed, certain passages from “Was ist Aufklaerung” provide even stronger evidence that Kant should be understood as following in a Lockean individualist line. There, Kant notes that Aufklaerung, enlightenment, “is man’s leave-taking of his self-imposed immaturity [Unmuendigkeit]. Immaturity is the inability to use one’s 4

In Schmitt [1987], 46-47.

5 Schmitt further suggests that this point is strengthened by considering Kant’s suggestion, in “Was heisst: Sich im Denken orientiren?”, that thinking for oneself “means nothing other than asking oneself, in each case in which something is to be assumed: whether one would find it practicable to make the basis of one’s assumption, or the rule that follows from one’s assumption, to a universal principle of one’s use of reason.” (Kant [1900], VIII, 146-7; translation mine) We will consider this passage in further detail in section 4.

-4-

understanding without the direction of another.” (Kant [1900], VIII, 35; translation mine) Kant continues, in a passage that seems like a biting echo of Adam Smith’s remarks, It is so comfortable being immature. If I have a book that understands for me, a priest who has a conscience for me, a doctor who considers my diet for me, and so on, then I don’t need to trouble myself. I don’t need to think, if I can simply pay; others will already undertake that tiresome business for me. (Kant [1900], VIII, 35; translation mine] The implication that Kant makes devastatingly clear through the tone of this passage—e.g., “habe ich ein Buch, das fuer mich Verstand hat”—is, it would seem, that reliance on others for one’s own beliefs, the sort of reliance detailed at length by Smith in the passage with which we began, is unacceptable for the autonomous, enlightened, agent. 3. Testimony

Additional Textual Evidence for Kant’s Views on

Rather than attempting to puzzle out Kant’s views on testimony on the basis of implication, it would be far preferable to consider Kant’s explicit views on the question of the epistemic status of testimony.6 There are two central sources in which Kant addresses the question of the status of testimony. The first, published in Kant’s lifetime, is the essay, “Was heisst: Sich im Denken orientiren?” There, Kant notes that the testimonially derived belief serving as the foundation of all of our historical beliefs is a form of knowledge. He writes: Historical belief, e.g., concerning the death of a great man, as reported in some letters, can become knowledge if the government of the man’s country reports on his burial, his will, and so on. That therefore something historical may be taken to be true—that is, may be believed—simply on the basis 6 For citations of the material on testimony contained in Kant’s lectures, and for a discussion of the background of Kant’s writings on testimony, I am indebted to Oliver Scholz [2001], particularly pp. 834-8. For an exhaustive discussion of the textual evidence bearing on Kant’s views on testimony—not available to me when I wrote this paper—cf. Gelfert [2006], who argues that Kant’s views were in fact more univocal in their support of testimony than I suggest here.

-5-

of witness reports, for example, that there is a city, Rome, in the world, and that even one who was never there can say, “I know”, and not merely, “I believe”, that Rome exists, is clearly related to this. (Kant [1900], VIII, 141; translation mine) Thus, this passage suggests that, rather than taking Kant to be a Lockean with respect to the status of testimony, one should rather read him as having recognized the possibility of acquiring knowledge on the basis of testimony. The other source of material in which Kant considered the epistemic status of testimony, and the source in which he did so in far greater systematicity, may be found in Kant’s lectures on logic. Kant lectured on logic by reading and commenting on Georg Friedrich Meier’s Auszug aus der Vernunftlehre. In §206 of the Auszug, Meier notes: Through the experience of other people we become certain by means of faith. Someone who presents something real as true, so that another takes it to be true, is called a witness (testis), and his action a witness report (testimonium, testari). To believe (credere) means to accept something on the basis of a witness report. Belief (fides, fides historica) is the applause that we give to something on the basis of a witness report.7 Thus, Meier’s account of the sources of certainty includes testimony; believing someone is a source of certainty. Kant criticizes Meier’s account for two flaws. The first flaw has to do with the fact that Meier, as a follower of Wolff,8 failed to appreciate what Kant took to be the central distinction between opinion, knowledge, and belief. For Kant, opinion is a taking-true (Fuerwahrhalten) that is neither subjectively nor objectively adequate. On the other hand, knowledge, for Kant, is a taking true that is both subjectively and objectively adequate; to know that p is to be convinced (subjective) that p with certainty (objective). Between those two extremes is belief—objectively inadequate but subjectively adequate. Thus, when one believes, one is convinced,

7

Cited in Scholz, 835; translation mine.

8

Cf. Scholz, 835.

-6-

but lacks certainty.9 This is the basis for Kant’s first criticism of Meier: “So-called historical belief can … actually not be called belief and as such be set apart from knowledge, since it itself can be knowledge.” (Kant [1900], IX, 69; translation mine) Thus, in opposition to Meier, Kant suggests that testimony may yield not only conviction based on mere faith, but may in fact convince in a way that is objectively adequate—i.e., in a way that yields knowledge. Kant’s second criticism of Meier has to do with Meier’s insufficiently distinguishing between one’s believing a speaker and one’s believing what is said. The former may be considered a form of faith (fides), whereas the latter, as we just saw, may attain the status of knowledge. Thus, Kant defines “historical belief” as “when we believe something, because we believe another person …—or; when one takes something to be true, merely on the basis of a witness report.” (Kant [1900], XXIV.1, 600-1; translation mine) Thus, Kant recognizes the possibility of acquiring knowledge on the basis of testimony. Indeed, he notes that “we can accept an empirical truth with the same certainty on the basis of the witness report of another, as if we had arrived at it through facts of our own experience. In the former case of empirical knowledge there is something misleading, but also in the latter.” (Kant [1900], XXIV.2, 896; translation mine) Furthermore, Kant recognizes that “one can often believe another’s witness report more than one’s own experience.” (Kant [1900], XXIV.1, 560; translation mine) In fact, on this point we can appeal to a considered element of Kant’s critical thought. For, in arguing for the claim that judgments of taste are not open to proof, Kant appeals, as the contrasting case, to examples drawn from the realm of testimony.10 Thus, in the Kritik der Urteilskraft, §33 (my translation), Kant writes: One clearly accepts … that others in any case may see and observe for him, and that what many have seen in a certain

9

It is crucial to note that the term “certainty” (Gewissheit) for Kant does not denote merely subjective necessity. Indeed, Kant’s description of certainty as involving necessity at all may be misleading for contemporary readers—e.g., for Kant, one may have empirical certainty. (Cf. Kant [1900], IX, 71) That certainty is not subjective for Kant may be seen in that certainty, for Kant, involves a standard of correctness. Thus, Kant speaks of standards of proof in the context of his discussion of rational certainty, and of appropriate grounds in his discussion of empirical certainty. (Cf. Kant [1900], IX, 72) 10

On this, see Hopkins [2000], pp. 209-36.

-7-

way may serve as sufficient grounds for a theoretical or even logical judgment for one who believes to have seen it differently … Note particularly, in this passage, Kant’s confident assertion that others “may see and observe for” us; contrast this in particular with Locke’s disparagement of the “hope to see with other mens’ eyes.” Thus, Kant recognizes that testimony (a) at least sometimes underwrites knowledge, (b) is subject to only the same possibilities for error that affect other sources of empirical knowledge, and (c) is sometimes a better source of evidence than one’s own experience. So much for the notion of Kant as a Lockean individualist, as Schmitt contends.11 4.

Some Subtleties of Kant’s Stated Views

Before we turn to a question of the implications of Kant’s views on testimony for the contemporary debate between Antireductionists and Reductionists with respect to the epistemic status of testimony, however, it will be useful to note a number of subtleties of Kant’s view with respect to testimony. Indeed, we must modify the extent to which we take Kant explicitly to have rejected a Lockean position with respect to the possibility of acquiring testimonial knowledge. It is crucial to recognize that Kant’s acceptance of the possibility of acquiring knowledge by means of testimony applies only to empirical knowledge.12 In the lectures, Kant notes that three sorts of knowledge cannot be acquired testimonially: mathematical, philosophical, and moral. Thus, he writes: One can, of course, believe mathematical rational truths on the basis of witness reports, in part because error is in this case not easily possible, in part because it can be easily discovered, but one certainly cannot know them in this way. Philosophical rational truths are, however, not even able to be believed, they must only be known; for philosophy allows in itself no mere convincing. And particularly what

11

And leaving aside the question of whether Locke himself ought usefully to be considered a Lockean individualist in Schmitt’s sense. Cf. fn. 14, below, and my [2009]. 12

Cf. Kant [1900], IX, 71.

-8-

applies to the objects of practical rational knowledge in morals … even so little could a mere belief occur in the case of these. (Kant [1900], IX, 68-70; translation mine)13 Given this, then, it would seem that one explanation for Schmitt’s reading of the brief digression on the sensus communis in the Kritik der Urteilskraft as supporting a Lockean account of testimony is that such a reading is in fact correct—only, of course, when applied to the possibility of acquiring knowledge of mathematical, philosophical, or moral truths via testimony. In the case of testimony regarding empirical facts, Kant considers a variety of factors affecting the reliability of witnesses.14 He notes that reliable witnesses will have competence and integrity.15 Furthermore, he suggests that they will possess the ability rationally to consider their experiences, a reliable memory, and the ability to express themselves clearly.16 He distinguishes between direct witnesses (Augenzeugen) and indirect witnesses (Hoerenzeugen), who merely transmit the testimony of another, and suggests that indirect witnesses are only half as believable as direct witnesses, since one would have to examine the reliability both of the direct as well as the indirect witness.17 5. Difficulties in Establishing a Stable Kantian Position on Testimony The question remains, however, as to whether Kant held that testimony, in some cases, could ground knowledge without additional inductive support for the competence or veracity of the testifier, as contemporary Anti-reductionists with regard to testimony hold, or whether he held that testimony always requires additional inductive support regarding the qualities of a testifier that make her a reliable testifier, as contemporary Reductionists contend.

13

Cf. Kant [1900], XXIV.1, 244.

14

Though, it must be noted, Kant’s considerations with respect to these questions do not advance far beyond Locke’s rules of thumb for weighing witness testimony. Cf. Locke [1975], IV, xvi, §10. For more on Locke on testimony, see my [2009]. 15

Kant [1900], IX, 72.

16

Kant [1900], XXIV.1, 244.

17

Kant [1900], XXIV.1, 244-5.

-9-

The difficulty with Kant’s pronouncements on testimony in the student notes of his Vorlesungen on logic is that the text provides support for both the Anti-reductionist and the Reductionist positions. Thus, Kant writes: As far as the believability and honesty of witnesses … is concerned, everyone will be held to be honest and candid until the contrary, namely, that they stray from the truth, has been proven. (Kant [1900], XXIV.1, 246; translation mine) This “Rule of Fairness”, as Kant calls it, is precisely the strategy of the Anti-reductionist: testimony is presumed to confer epistemic license, unless and until there is reason to doubt it. Of course, one might claim that Kant’s statement at this juncture is simply a description of the attitudes that the folk normally hold with regard to testimonial evidence.18 However, the context in which this statement occurs would indicate, on the contrary, that Kant is positively disposed toward this strategy. For Kant here discusses not simply strategies actually employed by the folk, but strategies that he himself endorses. Furthermore, the labeling of this strategy as one following a “Rule of Fairness,” would indicate that Kant was offering an endorsement of the strategy. Unfortunately, Kant immediately follows this claim with a diametrically opposed statement, namely that “in order to achieve right knowledge of the truth, it is nevertheless required that one examine whether someone speaks truth or lies.” (Kant [1900], XXIV.1, 246; translation mine) Later, Kant states: An open-eyed belief is one that is bound up with an examination of the witness, and is grounded in the same. A blind belief, however, is one that accepts a witness report without examination and exploration of the reliability of the same. The former leads to truth, the other, however, is the way to error, and thus harmful, whereas the former is helpful. (Kant [1900], XXIV.1, 249; translation mine) Thus, the same text also contains strong textual evidence in favor of a Reductionist position: that testimony only confers epistemic license when coupled with an inductive argument in support of the reliability of the witness.

18

This point was suggested to me by an anonymous referee for this

journal.

-10-

This ambivalence in the textual evidence with respect to the two sides in the contemporary debate between Anti-reductionism and Reductionism reflects, I would suggest, a lack of integration of Kant’s thinking on the epistemic status of testimony into his overall system. In the final, third, stage of this paper, we will attempt to sketch one line of thinking that links Kant’s thought on testimony to his larger systematic thought. 6. Testimony

A Systematic Integration of Kant’s Views on

In section 2 we mentioned the difficulties arising from the fact that Kant’s pronouncements regarding the epistemic status of testimony do not appear in any of his major published works. A potential criticism of our attempt, in sections 3-5, to tease out Kant’s views on testimony largely from the record of the student Mitschriften of his lectures is that the views contained in the Mitschriften ought not be considered part of Kant’s developed, systematic thought. This criticism, if apt, would provide further reason to ignore Kant on testimony—or to grant his views on testimony no more interest than one would the views of any historical figure. In this, the third stage of our discussion, I will argue that Kant’s systematic thought provides us with the resources to make sense of his acknowledgement of the epistemic status of testimony. Indeed, I will suggest that a proper understanding of this systematic argument for the epistemic status of testimony will allow us to recognize Kant as a precursor of one of the main forms of Antireductionist thought on the contemporary scene. The key to a systematic integration of Kant’s views on testimony lies in his discussion of lying. In this discussion, Kant reveals his vision of the role of linguistic communication in rational agency. Once this vision is manifest, however, it is clear that Kant must accord testimony positive epistemic status. In his discussion of lying in the Metaphysik der Sitten, Kant makes clear his utter contempt for the practice. He writes: Lying is disposal and at the same time destruction of one’s human dignity. A person, who himself does not believe what he tells another person (even if it were only a merely idealized person), has even less worth than if he were a mere thing, because another person can make some use of the property of a thing to be of utility, because it is real and given. However, the

-11-

report of one’s thoughts to someone through words that contain (intentionally) the opposite of what the speaker is thereby thinking is a use that is completely contrary to the natural function of the speaker’s ability to communicate his thoughts, and consequently is a renunciation of his personhood and a merely misleading appearance of a person, not a person himself. (Kant [1900], VI, II, §9; translation mine) He continues: A person as a moral being (homo noumenon) cannot use himself as physical being as a mere means (language machine), which is not tied to the inner purpose (of the communication of thought). Rather, he is … obligated to himself to honesty. (Kant [1900], VI, II, §9; translation mine) Thus, according to Kant, the purpose of language is the communication of thought, a purpose that lies at the very foundation of rational agency. Lying, in that it is a use of language radically at odds with the fundamental purpose of language, is a violation of rational agency. Thus, Kant’s argument here against the right to lie is that lying, by undermining the possibility of a shared language, thereby undermines the very basis of rationality.19 Given this argument, however, we might attempt to mount a “quick and dirty” parallel argument that one must not only be honest in one’s dealings with one’s interlocutors, but also that one must trust in their honesty.

19

An alternate account of Kant’s rejection of lying sees that rejection as founded upon one’s respect for others as persons; lying to others would be a violation of their autonomy. Cf. Korsgaard [1996], particularly “The Right to Lie” and “Two Arguments Against Lying,” and Buss [2005]. Note that this argument would also admit of a parallel argument for the presumption of trust; not to trust one’s interlocutor would be to treat that interlocutor as not fully a person. Although this is not Korsgaard’s project, she provides some materials for the beginnings of such an argument in her [1996], 141-2. Following out such a project to its conclusion would, I suggest, allow one to see Kant as prefiguring a contemporary Anti-Reductionist strategy first employed by Ross [1986]. Though I was unaware of it at the writing of this paper, this strategy of linking the requirement that one trust testifiers to one’s obligation to respect others as persons has been invoked in Gelfert [2006], particularly on pp. 633-36 and 649.

-12-

Indeed, one might argue, following Lewis [1975], that the existence of a shared language requires not merely practices of truthfulness, but also those of trust. Such an argument, however, would require an extension of Kant’s own argument. Rather than pursue that line here, we will pursue a reading of Kant according to which one’s entitlement to trust the communications of others stems from the status of others as rational. If this is the case, however, then the role of communication in transmitting reasons from one epistemic agent to others dictates that there be an epistemic presumption in favor of treating the communications of others as truthful.20 Given this understanding of the role of truthfulness in ensuring the purpose of linguistic communication, and the strictures on rational agents in respecting the role of truthfulness, the following argument would seem to be available to Kant: (1)

If H understands an assertion, then H is prima facie entitled to believe that the source of that assertion is rational.

(2)

If H is prima facie entitled to believe that the source of an assertion is rational, then H is prima facie entitled to believe that the source is a source of truth.

(3)

If H is prima facie entitled to believe that the source of an assertion is a source of truth, then H is prima facie entitled to believe the assertion.

Thus, (4)

If H understands an assertion, then H is prima facie entitled to believe that assertion.

We will briefly consider premises (1) and (3) of this argument before turning to a detailed discussion of premise (2) in the following sections. Premise (3) requires little in the way of defense. Certainly if one is prima facie entitled to believe a certain source to be a source of truth, then one is clearly prima facie entitled to believe the deliverances of that source. The case for premise (1), however, is a bit more involved.

20 Cf. Burge [1993], 470-1. As will become apparent, subsequent discussion of the Kantian line in support of the epistemic status of testimony is, in a number of points, indebted to Burge’s discussion in his [1993].

-13-

According to premise (1), it is a sufficient condition for H to be prima facie entitled to believe a source to be rational that H understand the source to have made an assertion. One need not appeal to Quinean or Davidsonian reasons to motivate premise (1). Indeed, an appeal to inference to the best explanation should suffice. The best explanation for H’s having understood a source S to have made an assertion, all things being equal, is that S intended to make that assertion. The best explanation, in turn, for S’s having the capacity to intend to make an assertion and to make the assertions she intends is, however, that S is rational.21 As the best explanation for H’s understanding S to have made an assertion is thus that S is rational—and that S made the assertion she intended to make—it is a sufficient condition for H to be prima facie entitled to believe a source to be rational that H understand the source to have made an assertion. Thus, premise (1) is true. Of course, one objection to premise (2) ought to be immediately apparent: sometimes it can be rational for an agent to lie. However, Kant may deal with this difficulty in the following way. Lying introduces a disunity among the functions of reason—the liar employs his/her rational agency in a way that aims at the dissolution of that very agency. Indeed, this is perhaps the central motivating force behind the passage quoted above, in which Kant argues against lying by emphasizing that “the report of one’s thoughts to someone through words that contain (intentionally) the opposite of what the speaker is thereby thinking is a use that is completely contrary to the natural function of the speaker’s ability to communicate his thoughts, and consequently is a renunciation of his personhood ….” (Italics mine.) Given this fact, however, one need not question the integrity of a rational source unless one has specific reasons for doubting the rational unity of that source’s functions of reason.22

21

The sense of “rational” here is intended to be robust enough to rule out extreme cases of irrationality—including mental illness. For more on why the best explanation for one’s having the capacity to intend to make an utterance would rule out one’s being irrational in this sense, see the discussion of the way in which non-truth-directed, or strategic, uses of language must be parasitic on truth-directed uses, in Section 9 below. The argument presented there also suggests, mutatis mutandis, that understanding intentions to assert on the part of irrational agents would also be parasitic on the prevalence of rational intentions to assert. 22

Recall Kant’s invocation of the “well-known ‘Rule of Fairness’”, in Section 5 above. Cf. Burge [1993], 475.

-14-

We will discuss the response available to Kant in addressing the problem of the rationality of insincere assertion below, in sections 8-10. For now let us pass over problems related to the sincerity of speakers. Even if we ignore the question of the sincerity of speakers, a further worry regarding the argument would involve questioning the reliability of assertoric communications. We will turn in the following section to an answer to this challenge available to Kant. 7.

The Reliability of Assertion

Even accepting that one’s source is not lying, why ought the rationality of the source of a putative assertion be a reason for one to take the source’s assertion to be true? Surely honest mistakes cannot be considered irrational. Thus, the Kantian would need some way of linking the rationality of sources to the sources’ accuracy in representing states of affairs in the world in order to motivate premise (2) of the argument introduced in section 6. The answer I wish to appeal to here is the Kantian one that plays on considerations about what it is to take one to have beliefs at all.23 As we have noted, for the time being we are conceding the link between assertion and sincerity—that is, we are, arguendo, conceding that rational assertion presupposes honest assertion. Given this assumption, however, the conclusion that there is a necessary connection between rational assertion and truthful assertion follows due to considerations governing the nature of belief. It is widely held that the overarching norm governing belief itself is that of believing that p only if p is true.24 This is the point of

23 An alternate strategy would be the Davidsonian one of suggesting that there is a necessary connection between understanding one’s interlocutor and imputing true beliefs to that interlocutor. Cf. Burge [1993], p. 487. Note that there Burge takes his argument to be an extension of the Principle of Charity in that he assumes not only that rationality implies that rational epistemic agents have, largely, true beliefs, but also that there is a presumption that those whose assertions we seem to understand do not need to be interpreted. Thus, Burge writes that “[t]he default position is that understanding can be presumed until something goes wrong”. 24

As Schwarz [1970], p. 62, has pointed out in another context, Kant would not express the norm in this way. Rather, the Kantian argument, as presented in (a) – (c) below, would proceed from evidence

-15-

distinguishing belief from desire in terms of its direction of fit; beliefs aim at fitting the world, whereas desires aim at bringing it about that the world fits them. Certainly, however, if one should believe that p only if p is true, then one should believe that p only if one has adequate evidence that p is true.25 Thus, Kant suggests that “belief requires a marriage [connubium] of other sources of evidence … in order to be evident and convincing. If a source of evidence is incapable of giving an adequate proof of the truth of something, then one must seek the help of other sources of evidence and seek to bring forth a greater level of evidence.” (Kant [1900], XXIV.1, 249-50; translation mine) In the case of belief, however—even more so, perhaps, than in the case of assertion—it would seem that the level of adequacy of evidential support at which beliefs should aim is support sufficient to underwrite knowledge. Indeed, the only case in which mere belief is appropriate, according to Kant, is one in which its objects are of a sort unsuited to knowledge. In all other cases—including the sorts of empirical propositions that would be the subjects of the instances of testimony here under consideration—knowledge is the only stable option, as the other, mere opinion, is neither subjectively nor objectively adequate, according to Kant. (Cf. Kant [1900], IX, 66-7) Thus, it is a norm of belief that one should believe that p only if one has evidence sufficient to underwrite knowledge that p. (Cf. Bernard Williams, [1978], 32-46) Thus, the argument proceeds as follows: a.

If H is prima facie entitled to believe that the source of an assertion that p is rational, then H is prima facie entitled to believe that the source of that assertion, S, believes that p.

b.

If H is prima facie entitled to believe that S believes that p, then H is prima facie entitled to believe that S is a source of truth (with respect to the question whether p). 26 Therefore,

sufficient to underwrite knowledge as the norm of belief. However, the – for Kant – anachronistic order of presentation in this paragraph will be one more palatable to the contemporary reader. 25

On this point, cf. Williamson [2000], 245-55.

26 Again, it is important to stress the difference between this argument and the Davidsonian one that, if one is to interpret S as having contentful beliefs at all, one must interpret S as having largely true belief. It is this strategy that Burge [1993] follows, e.g., at p. 471: “content is

-16-

c.

If H is prima facie entitled to believe that the source of an assertion that p is rational, then H is prima facie entitled to believe that the source is a source of truth.

This, then, is the defense of premise (2) against the claim— canvassed at the outset of this section—that the link between rationality and a prima facie presumption of truthfulness is too weak. Indeed, the force of Kant’s assertion, in the passage on lying quoted above, that “the report of one’s thoughts to someone through words that contain (intentionally) the opposite of what the speaker is thereby thinking is a use that is completely contrary to the natural function of the speaker’s ability to communicate his thoughts,” is precisely that, barring an instance of lying, the presumption is that speakers speak their minds. Combining this with the presumption that rational believers are governed by a norm compelling them to strive to believe only what they know, however, yields the result that rational speakers will strive to assert only what they know. This, in turn, supports the presumption, cited in premise (2) of the argument presented in section 6, that rational speakers are sources of truth. 8.

Lying and the Unity of the Functions of Reason

We ought, however, now to consider the objection, based on the possibility of insincerity on the part of speakers, to premise (a) of the argument in section 7—and, mutatis mutandis, to premise (2) of the initial Kantian argument introduced in section 6. Kant’s reply to this objection is quite succinct. As we noted above, he states that, because lying introduces a disunity among the functions of reason, one need not question the integrity of a rational source unless one has specific reasons for doubting the rational unity of that source’s interests. This reply, however, might seem initially implausible. If Kant is to appeal to the unity of reason, it would seem that he must demonstrate that all of our rational norms—including those involved in practical rationality—are directed at truth. While it seems much more promising to argue that the norms of theoretical rationality are united in aiming at truth, it would certainly seem difficult to motivate the idea that all of our norms are truth-directed. Although Kant’s suggested strategy seems, thus, unpromising, we may defend the notion of a default entitlement to trust in the sincerity of others by employing considerations gleaned from the norm of veracity that, according to Reid, is essential to the constitutively dependent, in the first instance, on patterned connections to subject matter …. So presentations’ having content must have an origin in getting things right”.

-17-

proper functioning of our language faculty.27 If Reid is correct, unless one has a reason to lie, one’s rational course of action in assertion is to say what one believes to be true. This suggests the following argument in support of such an entitlement to trust: d.

Without special reasons for lying, the rational course of action is to say what one believes.

e.

The simplest explanation for a given action is the one that does not involve unnecessarily attributing special reasons to the agent for performing that action.

f.

In a given assertoric speech situation, absent reasons believed by a hearer as to why a speaker should lie, the simplest explanation for a speaker’s assertion is that the speaker is saying what one believes.

g.

In interpreting the behavior of another, one is entitled to assume the simplest explanation for that person’s actions.

Thus, h.

In a given assertoric speech situation, absent reasons known by a hearer as to why a speaker should lie, the hearer is entitled to assume that the speaker is saying what she believes.28

Thus, this argument suffices to answer the objection, based on the possibility that one could be rational and nevertheless lie, to premise (2) in section 6, If H is prima facie entitled to believe that the source of a putative assertion is rational, then H is prima facie entitled to believe the source is a source of truth. For this argument demonstrates that the prima facie presumption of rationality includes a prima facie presumption of honesty. There is a further criticism of the Kantian premise (2) that is, however, unique to the case of one’s trust in the communications of others. According to this line of argument, the reasoning from (d) to

27

Cf. Chap. VI, Section XXIV, “Of the analogy between perception, and the credit we give to human testimony,” in Reid’s [1769], pp. 335-6. See also my [1999]. 28 Cf. Burge [1993], p. 468, where Burge notes that his reply to the criticism of premise (2) is that “one is entitled not to bring one’s source’s sincerity or justification into question, in the absence of reasons to the contrary”.

-18-

(h) fails because, depending upon how we cash out the argument, either (d) is false or (d) and (e) do not yield (f). 9. The Challenge of Non-Truth-Directed Linguistic Communication While it is true that it is irrational to be insincere, barring special reasons for one’s insincerity, it is not irrational for one to make use of the form of assertoric utterances without intending to provide factual statements. There are many situations —e.g., jokes, story telling—that seem to involve assertoric utterances that are not intended as transmissions of factual information, but are nevertheless not insincere.29 Furthermore, there are prominent uses of language, such as irony or metaphor, that, if accepted as literally understood, would often yield false beliefs.30 Finally, there may be reasons stemming from evolutionary psychology that tell against the notion that the primary aims of even seemingly assertoric utterances are truth-directed. The existence of many of the uses of language considered above may be taken as evidence that language evolved not because its primary functions involved an essential truth-aim, but rather due to the function of language in fostering cooperation and furthering other endeavors.31 The fact that such rational uses of seemingly assertoric utterances do not involve an aim toward truth would seem to block the argument from the rationality of sources to the entitlement for hearers to presume that those sources are sources of accurate factual statements.32 The point of this new objection is to accept, at least for the sake of argument, that there is an entitlement accrued to sources of content—i.e. our perceptual faculties—and to those sources— memory, induction, deduction—that preserve content, and promote 29 Compare Paul Faulkner’s similar criticism in his recent [2000], 585-7. One natural reading of Anne Bezuidenhout’s [1998], 261-98, would involve taking her to advance the same point against Burge. 30

Often, but not exclusively. Consider, e.g., “No man is an island”.

31

On this point see Dan Sperber’s [2001].

32 Although the present context is that of the discussion of the argument from (d) to (h), it will be clear to the reader that the present objection also threatens the previous two arguments—specifically premises (2) of the argument in section 6 and (a) of the argument in section 7. Thus, the discussion here is essential in clearing away a further—and perhaps the most serious—obstacle to accepting the Kantian argument in favor of an a priori prima facie presumption in favor of the acceptance of testimony.

-19-

coherence of content, within one cognitive agent. The objection, however, continues by emphasizing a substantive distinction between (in Thomas Reid’s formulation) the solitary and the social powers of mind.33 Even if one accepts that considerations of the proper functions of the solitary powers requires one to accept that those powers involve a significant truth-connection, considerations regarding the variety of uses to which language may be put seem to require that one accept that even seemingly assertoric utterances do not involve a similarly strong truth-connection. Unlike the solitary powers, the social powers do not, even when functioning properly, require a truth-connection, as they could still be directed at satisfying other aims, such as the achievement of social cohesion. Thus, so this line of criticism goes, the argument from (d) to (h) fails in that the motivation for accepting (f) no longer has bite. Let us take the first two of these points—those dealing, e.g., with jokes, fiction, irony, and metaphor—together, and then turn to a discussion of the point concerning the evolutionary role of linguistic communication. The points concerning jokes and metaphor emphasize the fact that premises (d) and (f) are stated in terms of lying, but that there are many cases in which one can utter a literal falsehood without thereby being insincere. The quick response to this objection is simply to reformulate the argument, replacing all formulations such as “without special reasons for lying” in (d) with “without special reasons for doing otherwise”—i.e., otherwise than speaking the literal truth. The point of making such a change is that all “strategic” uses of language, to borrow an expression from Habermas, are parasitic on truth-directed uses. (Cf. Thomas McCarthy [1978], 287) This includes sincere uses of assertoric utterances to express literal falsehoods, as in cases such as jokes or uses of metaphor. Indeed one measure of both the sincerity and rationality of those making strategic use of language is that they adequately signal their intention to employ language strategically. It is the goal of rational communicators in telling, say, a joke to be understood, and they would fail in that endeavor were their interlocutors to understand them as aiming at communicating literal truths. One need only call to mind a single instance in which a joke fell flat because it went unrecognized as a joke, and to recall the awkwardness both on the part of the joke-teller and his intended recipient, to recognize that the intention of the joke-teller must include that he be understood as 33

Cf. Chap. VIII, “Of Social Operations of Mind,” in Reid’s [1790], pp.

77-80.

-20-

not simply telling the truth. Thus, in cases of jokes, communicators have an interest in signaling that they are joking—and their recipients can trust in them, to the extent that they are rational, to so signal their intention to communicate strategically. These same considerations apply, mutatis mutandis, to the other strategic uses of language mentioned above. 10. Truth-Directedness and The Evolution of Communication Thus, the first two points fail to have any bite against the argument from (d) to (h), understood as underwriting a prima facie entitlement not to take communicators as communicating strategically, absent reasons to the contrary. This brings us to a consideration of the third point regarding the evolution of linguistic communication. Here one must, perhaps, concede that, as recent evidence suggests, language evolved not primarily for the transmission of truths—say, for the purposes of hunting or foraging—but rather as a social mechanism more efficient, in fostering group cohesion, than chimpanzee behaviors involving mutual grooming. (R. I. M. Dunbar [1992]) Such findings are supported by the fact that we spend a vast amount of time speaking about each other. Thus, researchers have found that social interactions and personal experience make up the content of about 70 percent of conversation time, and about half of that time is spent talking about those not present. Thus, the tendency to gossip, and an interest in gossip, would seem to serve a function of fostering social coordination over large groups of individuals. Suppose, then, that we concede that the transmission of truths is not the primary evolutionary goal of linguistic communication. In so doing, we must neither concede that the transmission of truths is not the primary goal of linguistic communication, nor that such transmission is not an evolutionary goal of linguistic communication. That the transmission of truths is, in fact, the primary goal of linguistic communication—and that other goals are parasitic on our broad success in achieving that primary goal—has already been established in our discussion of the first two points against the argument from (d) to (h). As to whether the transmission of truths is an evolutionary goal of linguistic communication, it would seem extremely plausible that it must be. Indeed, if our considerations concerning the necessity of norms of truthful communication for the very existence of a shared language are correct, then the only way that a practice of

-21-

linguistic communication could develop at all would be if that practice had the transmission of truths as one of its goals. Furthermore, consider the satisfaction of the goal of fostering social coordination through the communication gossip about oneself and others. It would seem difficult to imagine a situation in which social coordination could be fostered through the dissemination of widespread falsehoods about oneself and others. Mutual trust and cooperation would seem to require that a large proportion of the gossip with which we occupy ourselves be—broadly speaking— accurate gossip. Thus, the objection to the argument from (d) to (h) on the basis of considerations of the evolution of linguistic practice also fails to tell against the argument. 11.

Conclusion

Thus, contrary to the contemporary opinion concerning the lack of discussion of testimony by Kant, as evidenced by the sparse existing exegesis of Kant’s views on the epistemic status of testimony, we have seen that there is ample material to motivate a Kantian discussion of the epistemic status of testimony. Although the space available here provided the opportunity for the discussion of merely one line of thought, from Kant’s more systematic works, in support of an Anti-reductionist view of the epistemic status of testimony, it is to be hoped that such a discussion will prompt further investigation of Kant’s philosophical positions as they relate to questions pertinent to the contemporary discussion of the epistemology of testimony.

-22-

Bibliography Bezuidenhout, A. [1998] Is Verbal Communication a Purely Preservative Process? Philosophical Review 107:261-88. Buchan, [2003] Crowded With Genius. New York: HarperCollins. Burge, T. [1993] Content Preservation. Philosophical Review 102: 457-88. Buss, S. [2005] Valuing Autonomy and Respecting Persons: Manipulation, Seduction, and the Basis of Moral Constraints. Ethics 115:195-235. Coady, [1992] Testimony: A Philosophical Study. New York: Oxford University Press. Dunbar, R. I. M. [1992] Neocortex Size as a Constant on Group Size in Primates. Journal of Human Evolution 22: 219-33. Faulkner, P. [2000] The Social Character of Testimonial Knowledge. The Journal of Philosophy 97: 581-601. Gelfert, A. [2006] Kant on Testimony. British Journal for the History of Philosophy, 14: 627-52. Hopkins, Robert [2000] Beauty and Testimony. In A. O’Hear, ed., Philosophy: The Good, The True and the Beautiful. Cambridge: Cambridge University Press. Kant, I. [1900] Gesammelte Schriften. Berlin: Preussische Akademie der Wissenschaften. Kant, I. [1974] Metaphysik der Sitten. Frankfurt a.M.: Suhrkamp Taschenbuch Verlag. Korsgaard, C. [1996] Creating the Kingdom of Ends. Cambridge: Cambridge University Press. Lewis, D. [1975] Languages and Language. In K. Gunderson, ed., Minnesota Studies in the Philosophy of Science, vol. 7. Minneapolis: University of Minnesota Press. Locke, J. [1975] An Essay Concerning Human Understanding. Oxford: Clarendon Press. McCarthy, T. [1978] The Critical Theory of Juergen Habermas. Cambridge, MA: MIT Press. Reid, T. [1769] An inquiry into the human mind, on the principles of common sense. 3rd ed. London: Cadell and Longman. Reid, T. [1790] Essays on the intellectual and active powers of man.

-23-

Dublin: Byrne and Milliken. Ross, A. [1986] Why Do We Believe What We Are Told? Ratio 28: 6988. Scholz, O. R. [2001] Autonomie angesichts epistemischer Abhaengigkeit: Kant ueber das Zeugnis anderer. In V. Gerhardt, R.-P. Horstmann, R. Schumacher, eds. Kant und die Berliner Aufklaerung: Akten des IX. Internationalen KantKongresses. Berlin: Kant-Gesellschaft e.V. Vol. 5, 829-39. Schmitt, F. [1987] Justification, Sociality, and Autonomy. Synthese 73: 43-85. Schwarz, W. [1970] Kant’s Refutation of Charitable Lies. Ethics, 81:62-7. Shieber, J. [1999] “Thomas Reid on Cartesianism With Regard to Testimony: A Non-Reductivist Reappraisal,” Reid Studies, 2: 59-69. Shieber, J. [2009] “Locke on Testimony: A Reexamination,” History of Philosophy Quarterly, 26: 21-41. Sperber, D. [2001] An Evolutionary Perspective on Testimony and Argumentation. Philosophical Topics Williams, B. [1978] Descartes: The Project of Pure Enquiry. London: Penguin. Williamson, T. [2000] Knowledge and Its Limits. Oxford: Oxford University Press.

-24-