628: Kant thinks that someone else’s testimony can provide me with knowledge, and generally has the same epistemic weight as experience. Gelfert calls this the symmetry thesis. These facts about testimony give rise to a presumptive principle: “Testimony is to be believed, unless we know that certain defeating conditions obtain.”
628n3-4: See 16:501 and 16:509 for some remarks from Kant on testimony.
631: It’s possible to have an experience according to which X is the case, while X is not the case. This is a fortiori true for testimony: if it’s possible for your own experience to be misleading, it’s possible for someone else’s experience to be misleading too, so if you believe X on the basis of their experience, you could be wrong. However, testimony has two additional problems that experience doesn’t have: while experience may be non-veridical (therefore, so too can testimony), the speaker on whose testimony you rely can be incompetent or insincere.
631-32: According to Meier, from whose logic textbook Kant lectured, for someone’s testimony to be reliable, she has to be “a competent witness as someone who possesses “sufficient powers to not only acquire the right experience but also to designate it in the right way” (Auszug, §207). … Competence eludes formal definition because it is essentially a skill that varies with context … competence on this account is not conceived of as entirely within the witness but also takes into account the  circumstances under which experiences are made. In order for someone to be a competent witness, what is required is not only ability but also opportunity.
“In addition to observational skills and favorable circumstances, a third element is required for a competent witness, namely “the adroitness to declare his experiences so that one can well understand what he means”. … A general inability to express oneself clearly and accurately, in this context, is far more detrimental than an occasional lapse … as it introduces a systematic error. … Thus, competence, when analyzed in terms of Tüchtigkeit, is not a statistical point about someone’s being right more often than wrong, but requires the presence of an adequate “rule-guidedness” of reasoning and observation.”
See AA 14.2, 896 and 898 for some of the quotations for Kant on this matter.
So, for someone’s testimony to be reliable, she has to be a competent witness. To be a competent witness, she must have the ability to observe things, the opportunity to observe them, and the ability to clearly and accurately express what she has witnessed. For her to have these two abilities (the opportunity is not an ability), her reasoning and observation must be rule-guided.
633: So, Kant believes it’s possible for testimony to convey knowledge. OK, but when is it possible? Nowadays, epistemologists say that you have a right to believe someone’s testimony just so long as there are no relevant defeaters to it. But that’s not Kant’s tack. Instead, “Kant’s presumptive principle does not so much establish a “presumptive right” as a “presumptive obligation” to accept another’s testimony.”
634: Very interestingly, Kant thinks it’s immoral to be too suspicious of people: “we … have a presumptive obligation not to distrust others. The mere fact that this is not a perfect obligation … does not render it any less fundamental. It does, however, call for a specification of the social and institutional conditions that circumscribe its applicability.”
634: “Testimonial belief requires an element of faith … that is, an exercise of trust. … in accepting someone’s testimony we do not just adopt a new belief, we trust him for the truth.” Note that Kant thinks that we have an obligation to believe people who make promises to us. Since someone else’s testimony is a kind of promise to us that what the person claims to have witnessed is something he really witnessed, underlying Kant’s principle of presumptive acceptance of testimony is “a duty to fidelity combined with a duty not to be wantonly suspicious”.
635: The thought behind Kant’s claim that we have a duty to trust others’ promises is that if we didn’t, then we wouldn’t be able to have promising, which would have negative consequences for social life. Similarly, if we couldn’t accept anyone’s testimony (unless we verified it ourselves), then this would make public life impossible. Kant says such incredulity violates human dignity.
635: This is a nice summation of Kant’s position so far: “Kant’s epistemology of testimony differs markedly from other positions in that it balances the recipient’s perspective, which is characterized by the hearer’s desire to have adequate justification for his testimonial beliefs, with the testifier’s perspective and his legitimate expectation to be believed. … The idea seems to be that, as human beings, we are all in the same epistemic predicament of sometimes giving and sometimes receiving testimony; hence, if we desire to be believed by other people — as we all, at times, do — we ought to also believe the word of others. Of course Kant is not suggesting that we should always believe everyone … But some degree of trust is necessary if our epistemic position is to be tenable and sustainable.”
637: Kant distinguishes between testimony materialiter (testimony about contingent historical events) and testimony formaliter (testimony about universal truths of reason). In the first case, we rely on others’ witness to give us knowledge about contingent events that we can’t experience ourselves because we weren’t situated in the right way. In the second case, we should rely on our own reason rather than that of others (this makes the role of a church–to publicly propound certain moral principles so that people don’t rely just on their own judgment–even more curious). “The idea … is that as rational beings we are all on a par with one another: Hence, in matters that are the domain of reason (such as morality and universal truths), each of us can, and indeed has a duty to, think things through on their own, whereas in contingent empirical matters … we ought to rely on other people’s experiences”.
638: Kant is an anti-inferentialist about the epistemic authority of testimony: it has epistemic weight, not because you infer to its weight from the epistemic weight of direct experience, but rather because “the adoption of testimony works by “tying together the experiences of others with our own” and spotting an inconsistency, or a violation of an independently known universal truth of Reason, when one occurs. Such transpersonal extension of the domain of experiences hinges on our own experiences being ocntinuous with those of others (at least as far as knowledge claims are concerned)”.
639: If Kant thinks we should not be suspicious of others’ testimony, how can he also think that enlightenment is thinking for oneself? To answer this question, you have to look at Kant’s notion of enlightenment as it appears in the Critique of Judgment.
640: In CJ, Kant discusses a sensus communis, which “is a form of “common” sense not so much because of its merely being widely accepted but because it is something which is genuinely shared by a community. … He identifies three maxims that are constitutive of sensus communis: “1. To think for oneself; 2. To think in the position of everyone else; 3. Always to think in accord with oneself”.” Kant thinks that accepting 1 is essential to being enlightened, because it is only by accepting 1 that you can avoid being prejudiced. The worst prejudice, on Kant’s view, “”is to imagine nature as not being subject to those rules which the understanding has constitutively imposed on it … by means of its own essential law: i.e., superstition“. Kant goes on to introduce the notion of enlightenment as a “merely negative element”: “Liberation from superstition is called enlightenment“.” I really need to get clear on what Kant understands by “superstition”.
640-41: One of the prejudices that Kant warns against is “the “logical prejudice” of “authority” … of the speaker” (640), which is when one accepts a proposition of reason on the basis of someone’s say-so rather than on the basis of thinking for oneself. “The very fact that propositions of reason are such that every rational being can come to know them by properly exercising their own capacity to reason excludes them from testimony” (641).
641: In these pages (641-43), Gelfert makes a series of fascinating points that are of use to my paper on the church. Let’s take them in the order in which they appear: “While [Kant’s demand to think for oneself] imposes limits on our reliance on the word of others, these limits are by no means overly narrow. In particular, they leave room for a presumptive acceptance of empirical testimony; indeed, as Kant acknowledged earlier, such testimony provides us with knowledge that we could not otherwise obtain. It is, thus, perfectly rational to rely on such testimony, provided this reliance is not “uncritical” (in the sense described earlier): “Historical belief is reasonable … if it is critical”. [16: 501; Refl. 2763]. Being “critical” … means that we must not adopt testimonial beliefs in a dogmatic fashion that ignores that testimony, just like experience, is always fallible. Staying attuned to the fallibility of knowledge claims, while avoiding any slippery slope leading to skepticism, is what “being critical” is all about. Thus, we must be ready to revise our  (testimonial) beliefs, for example when a comparison of our own judgments with those of others gives us reason to suspect that we are in error. This is what justifies an attitude of presumptive acceptance in the first place.” This attitude of revisability is what, perhaps, makes autonomous religion so hard for people to accept.
642: “The significance of one’s ability to “think in the position of everyone else” is clear: it indicates the opposite of being prejudiced, namely being “broad-minded”. As such, it is more of an attitude than a skill or (natural) capacity.” Gelfert then quotes the CJ, §40 for evidence of this.
643: Gelfert quotes from the Bauch Logic to make a point that relates interestingly to the propensity for deception described in the Theodicy essay. First, the quote from the Bauch Logic: “We do not only have a propensity to participate [in society] but also to communicate. Man only learns something so as to be able to communicate it to others. He does not trust his own judgment, unless he has told it to others. Everything is unimportant to us if we cannot communicate it to others.” (BL, 55) About this, Gelfert writes, “[Kant] refards communicability (and, as we shall see shortly, actual communication) as a normative “touchstone” of truth. In the section of the Critique of Pure Reason, titled “On having opinions, knowing, and believing” (A 820, B 848f.), Kant makes communicability a fundamental criterion by which to distinguish cases of mere opinion (on the basis of persuasion) from believing something to be knowledge (on the basis of testimony)”.
644: “Communicability, as Kant conceives of it, is not a merely derivative phenomenon but is constitutive of reliable judgments concerning the truth of matters of fact. The point of communicability … consists precisely in furnishing us with a means of distinguishing between knowledge, beliefs, and opinions, by allowing others to scrutinize, challenge and correct our own judgments. In the Bauch Logic, Kant illustrates this “validating” function of communicability using a very apt metaphor: “Man always wishes to test his judgment on others; other people’s judgments are to be regarded as contrôleurs of our own judgment“. [BL, 55] Other people’s judgments, of course, are typically available to us through their testimony only, which is why an attitude of presumptive acceptance — in the spirit of the “critical method” outlined earlier — is called for.” Gelfert says that communicability is crucial for ensuring that our judgment concerning matters of fact goes right; but it seems to me that it would also have an important role to play in judgments of reason as well, as we need to know whether other people can reach the same judgmen ts to know whether our judgment of reason is counterfeit or the real McCoy.
644-45: “It is important to realize that Kant’s conception of communicability as a “touchstone”, or “criterion”, of truth requires both that communication of one’s own thoughts is possible under the given circumstances, and that we display a general readiness to critically accept other people’s testimony, since it is through testimony that we learn about other people’s judgments. If either of the two requirements fails to be met, communication fails to lead to epistemic progress.” (644) So, in order for communication to be useful to the race, people have to have communicable thoughts, and they have to display a willingness to accept others’ testimony. The reason, then, we have to worry about censorship is this: “It is only to the extent that we are deprived of the opportunity to receive critical feedback from other people that restrictions  on expressing our own judgments are unjust. … if self-expression derives its justification first and foremost from the possibility of criticism by others, then clearly the normative role of communicability cannot be separated from an attitude of presumptive acceptance of testimony, to the extent that the latter is necessarily the prime source of learning about other people’s judgments.” So, self-expression is important because if we’re allowed to express ourselves, we can improve our own ideas and the ideas of others through mutual challenge. Heteronomous religion is dangerous, then, because it pressures us to limit self-expression. Autonomous religion is useful because it can encourage certain ideas that are useful for the spread of truth — e.g., it can encourage us to take other people’s testimony seriously, at least as a presumption. (Of course, this raises the question: what about taking others’ testimony seriously — is this an idea we merely must accept, or can we challenge this as well? Presumably we can challenge it, and it can defend itself, as it is an idea of reason.)
645: Guyer claims Kant is a methodological solipsist — he thinks each individual should simply focus on his own thinking in order to know the limits of thought. If Guyer is right, then it doesn’t make sense why Kant would take testimony seriously. However, Kant is a methodological solipsist only when it comes to “searching for transcendental conditions that secure the possibility of experience in general.” Kant has another project, though, and about this he is not a methodological solipsist. This other project is that of “analyzing the conditions that make knowledge possible in the given situations we find ourselves in. … [this] project analyzes the conditions of knowledge for us as finite social beings.”
646: Kant “acknowledges that testimony can not only communicate knowledge but, given the right circumstances, can also create it.” Gelfert cites What Does It Mean to Orient Oneself in Thinking, p. 54. I don’t know what version it is, but it’s where Kant talks about Rome. The point seems to be that if lots of people have testified to the existence of something, then you can say that you know that whatever they’re testifying to exists. (E.g., if Rome no longer exists, but there are lots of records of people in the past saying they went to Rome, then you can know they went to Rome — you don’t have to say, simply, “well, I believe there was a Rome”, or that “there is very good evidence that there was a Rome” — no, you can say you know it. “Of course, if we had reason to doubt the veracity of the announcement [of the death of a great man], we could put our judgment to the test by inquiring further and comparing our own judgments with those of others, but as things stand, the rules by which such announcements do occur, seldom give reason to begin a detailed investigation.” (647)
647-48: In these very interesting and relevant pages, Kant points out that testimonial practices vary by time and place (“specific testimonial practices are culturally determined”). Consequently, the testimony of people in one place/time may provide us with knowledge while those of people of a different time/place don’t (see 14.2, 898f.). His thinking is that there need to be certain institutions in place before you can rely as much on people’s testimony (“The ancient world lacked certain standards of accuracy that, in modern times, are enforced through collective projects such as science … and communication (“institution of the postal service”): “Not until the beginning of the last century did people begin to realize that it is necessary, to tell the whole truth, and every one thus had to be wholly accurate in his reports, and if someone swerved from the truth only a little in his writings: then he would be embarrassed and dishonored”). If these institutions are in place, though, then you can believe people’s testimony (“it is the presence of collective endeavors such as science and modern systems of communication (alongside the postal system he also specifically mentions the printing press and the emergence of newspapers) which warrants trust in the truthfulness of testimony, because it is through their presence that they set standards of accuracy and rules of communication to which people on the whole aspire.)” (All quotations taken from 647).
647-48: “the presence or absence of  practices that ensure that the criteria laid down for the sensus communis — first and foremost, a willingness to revise one’s own judgment in the light of other people’s testimony — are fulfilled, is indeed a contingent matter. … the mere fact that its presence is a contingent matter does not entail that the sensus communis cannot play a justificatory role. On the contrary, it is the very standard of justification that we must adhere to in our judgments”. Obviously, this is all quite relevant for how seriously we take miracle-accounts.
- 16:501, 508, and 509; Rflx. 2763, 2780 (pp. 628, 634-36, 638, 641)
- Vienna Logic: 14.2, 893, 896, 898 (pp. 630-33, 647)
- Dohna-Wundlacken Logic: 14.2, 750
- Blomberg Logic:14.1, 246 (635)
- CJ, B157-160 (see pp. 640)
- Cataldi Madonna, 1992: 38 about Kant’s skepticism regarding a logic of probability (as opposed to a mathematics of probability).
- Bauch Logic: 59, 72, 139ff. (see pp. 636-37, 640-41)
- Pölitz Logic, 14.2, 560f. (p. 637)