Genuine Belief and Genuine Doubt in Peirce

Penultimate Draft – Please quote from published version or check with author Forthcoming in European Journal of Philosophy Genuine Doubt and Genuine Belief in Peirce Along with the pragmatic maxim to which it supposedly leads, the doubt-belief theory of inquiry stands as one of Peirce’s signature innovations. This paper hopes to complicate and then “reclarify” perhaps the clearest component of the doubt-belief theory. Peirce makes it quite clear, not just that doubt and belief are different, but that they oppose one another. One could capture that opposition by saying that where belief is doubt is not. But that slogan admits of a weaker and a stronger reading. According to the weaker reading, to the extent that one believes, one does not doubt. A stronger reading insists that if one believes, then one does not (genuinely) doubt. The weaker reading permits and the stronger reading forbids the possibility that one can be in a state of doubt and of belief with respect to the same proposition at the same time. The stronger claim is standardly attributed to Peirce, for significant textual and philosophical reasons. As a result, the suggestion that one can combine doubt and belief about the same issue sounds more than a little odd to anyone familiar with the literature on Peirce. I will argue that this standard construal of the doubt-belief theory is excessively strong. Ultimately, I think that we get a more flexible and more plausible version of Peirce’s approach to inquiry if we grant him the possibility that doubt and belief can combine. But I cannot undertake to show that here. I have my work cut out for me in rendering unheretical the suggestion that Peirce himself allows genuine belief to be adulterated by genuine doubt. The paper begins by sketching the contrast between doubt and belief that figures centrally in Peirce’s classic paper, “The Fixation of Belief.” It acknowledges some important textual evidence on behalf of the strong and standard interpretation of this contrast, though it will later try to neutralize that evidence. Turning to the secondary literature on Peirce, I argue that it, not unreasonably, tends to presuppose the strong reading but that, as a result, it often slides between weak and strong readings of the contrast, often confusing sufficient conditions for belief with necessary ones. I further argue that the thesis commonly attributed to Peirce according to which one needs completely undoubted beliefs in order to inquire, even if true, does nothing to establish the incompatibility of genuine belief and genuine doubt. The paper then briefly links the doubt-belief theory of “Fixation” to the papers on probability that occupy the bulk of the Illustrations of the Logic of Science. “Fixation,” the first paper of the series, and “The Probability of Induction,” the fourth, share a striking emphasis on when belief is appropriately felt. “Probability,” in the course of an innovative and sophisticated discussion of measuring belief, anticipates J.M. Keynes’ notion of (gross) weight of evidence. The deep connections between evidential weight and stability of belief make it compulsory to render “Fixation” and “Probability” consistent. But, as Jonathan Adler has argued, an emphasis on the importance of evidential weight carries with it the prospect of receiving evidence that undermines confidence without undermining belief. I support, on Peircean grounds, Adler’s contention that these acknowledged epistemic reasons for reflection, checking, and hesitation deserve to be called reasons for doubting. The prospect of a Peircean approach to inquiry that is richer and more flexible than the typical reading, which gives Peirce only “pure belief” and “pure doubt” to work with is thus, I claim, opened up. The Doubt-Belief Contrast Section III of “The Fixation of Belief” constitutes the ur-text for contrasting doubt and belief. As is well known, Peirce distinguishes doubt from belief in three ways. First, “there is a dissimilarity between the sensation of doubting and that of believing,” a dissimilarity that Peirce links to desires (W3 247). The sensation of doubting indicates a desire to ask a question while that of believing indicates a desire to pronounce a judgment. Presumably, Peirce should be granted some flexibility about the strength of the desires in question and about appropriate conditions for attending to the relevant sensations. In addition, says Peirce, “there is a practical difference” between doubt and belief. “Our beliefs guide our desires and shape our actions…The feeling of believing is a more or less sure indication of there being established in our nature some habit which will determine our actions. Doubt never has such an effect” (ibid). When we have a confident belief, we (tend to) act unhesitatingly and rely on the belief in question for even momentous decisions. Doubt, on the other hand, induces us to hesitate and to be disinclined to risk much on any proffered course of action. It is worth emphasizing that, in the course of drawing this contrast, Peirce explicitly acknowledges that belief and doubt come in degrees: “So it is with every belief, according to its degree” (ibid). It is not clear whether talk of degrees of belief applies directly to the feeling of belief and derivatively to the willingness to act indicated thereby, or whether Peirce intends to measure belief directly in terms of willingness to act. Either way, though, Peirce makes it plain that believers can feel and proceed more or less confidently, and this will matter in what follows. Peirce’s third contrast between doubt and belief does the most work in the argument of “Fixation”: “Doubt is an uneasy and dissatisfied state from which we struggle to free ourselves and pass into the state of belief; while the latter is a calm and satisfactory state which we do not wish to avoid, or to change to a belief in anything else” (ibid). Interestingly, this contrast is the only one of the three in which Peirce does not mention the feelings or sensations of doubting and believing. In his first contrast, Peirce seems to identify the feeling of believing or doubting with the desire to offer a judgment or ask a question. In his second contrast, he draws an explicit distinction between the feeling and a state indicated by the feeling (roughly, a willingness to premise the content of the feeling in deliberations). Peirce seems to be following Bain here. As characterized by Max Fisch, Bain holds that “the belief is the readiness, not the feeling which avouches it” (Fisch 1986, p. 84). Fisch offers a valuable characterization of doubt and belief in terms of Peirce’s categories at pp. 95-6. On this third contrast, Peirce is oddly silent. Is he saying that doubt feels uneasy and unsatisfactory or that the feeling of doubt indicates that one is in a state that prompts its own removal, or both? Does doubt feel uneasy or just indicate that one is in an unstable state? In either case, Peirce seems here to be emphasizing, not how doubt and belief feel, but that belief is a settled and doubt an unsettled state. Though many questions of detail remain, it is abundantly clear that Peirce opposes belief to doubt, and that this opposition structures his account of inquiry. Furthermore, while Peirce cautions that “doubt” and “belief” are words that are “too strong for my purpose” since, “as the words are commonly employed” they “relate to religious or other grave discussions,” ordinary usage agrees with Bain and Peirce in treating belief and doubt as opposites of one another. But the vague claim that belief and doubt preclude each does not entail the specific claim that nobody can simultaneously believe and doubt that P. If we grant, for the sake of argument, that being (or feeling) hungry and being (or feeling) sated are opposed to one another, all that follows is that one is no longer hungry to the extent that one is sated, and vice versa. Complete satiation then, would preclude hunger, but nothing so far commits us to denying that one could be full enough to count as sated and yet still also genuinely be hungry (for the wonderful dish you just smelled, for instance). It admittedly sounds odd to predicate genuine hunger and genuine satiation of someone, but at least part of the awkwardness seems to be pragmatic rather than semantic. Even if it could be true that someone was both satisfied and hungry, it would rarely be an appropriate thing for her to assert. I am arguing for no more than a conceptual possibility here, at least for now. From what Peirce has said so far, it follows that complete or maximal belief precludes all doubt, but it does not yet follow that anything worth calling a genuine belief precludes all doubt about the matter in question. The standard interpretation does not gratuitously ascribe such a strong version of the doubt-belief contrast to Peirce. In Section V of “Fixation,” in which Peirce argues that the unscientific methods of fixing belief are ineffective at doing so, Peirce argues that: [T]here are some people, among whom I must suppose that my reader is to be found, who, when they see that any belief of theirs is determined by any circumstance extraneous to the facts, will from that moment not merely admit in words that that belief is doubtful, but will experience a real doubt of it, so that it ceases to be a belief. (W3 253, emphasis added). This passage seems clear and compelling. The presence of a real doubt is sufficient to disqualify a state from counting as a belief. Standard readings of Peirce’s approach to inquiry seem to be on solid ground in denying that genuine belief and genuine doubt about a given claim can be realized together. In the face of this evidence, I will have to try to show that Peirce distanced himself from this claim and that it sits poorly with other key claims offered in the Illustrations. Before turning to that task, however, I want to highlight the role played by the strong construal of the belief/doubt contrast in the secondary literature concerning Peirce and inquiry. The Secondary Literature Presumably on the strength of the passage noted above, the strong version of the contrast tends to be taken for granted in the literature. Many characterizations of Peirce on inquiry can be read in either the strong sense or in the weaker sense I have adumbrated. The centrality of the strong reading emerges on the relatively rare occasions when such statements get disambiguated, since the clarification nearly always proceeds in the direction of the stronger reading. In an early and influential treatment of the doubt-belief theory, for instance, Israel Scheffler writes that “[w]hen opinion is settled, and ‘real and living doubt’ overcome, genuine inquiry cannot arise. When no actual doubt affects any given proposition, this proposition may be employed as a premiss, no matter how it might be imagined to be doubtful by hypothetical considerations” (Scheffler 1974, p. 60). At first glance, this certainly sounds like a statement that endorses or presupposes the strong reading of the doubt/belief opposition; the appeal is to a state of belief entirely free from doubt, after all. As I understand matters, however, Scheffler’s claim is neutral between the two readings. I certainly grant that a proposition’s being entirely free from doubt is a sufficient condition for concluding inquiry into the matter in question and for treating the proposition as a premise in one’s deliberations. The issue is whether such a condition is necessary. Nothing Scheffler says here precludes the possibility that one could believe a proposition, cease inquiring into its truth, and rely on it for various purposes, without the proposition being free of all actual doubt. The stronger version of the opposition may well be correct, for all I have so far said, but it should not be conflated with a weaker and more plausible claim. Elizabeth Cooke provides a nice example of how quickly the secondary literature tends to move from quite mild to quite demanding construals of the doubt-belief theory. “Practically speaking, the individual takes all her beliefs to be true. To believe p is to believe p to be true. C.J. Misak makes the point that, in essence, all our beliefs function as certain; they are practically certain” (Cooke 2006, p. 23). Notice how quickly this passage moves from a truism, viz. that to believe a proposition is to believe it to be true, to an extremely strong claim, viz. that nothing is a belief unless it “functions as certain.” The assumption that Peircean belief is a state that involves a very high degree of confidence and tenacity is clearly greasing the wheels for the slide from the truism to the strong and interesting claim. And when we turn to the cited passage from Misak, we again see how stealthily claims about Peircean belief get inflated. The part of Misak’s text in question does not license a claim as strong as the one that Cooke attributes to Misak and to Peirce. Misak quotes Peirce’s 1897 claim that “practically speaking many things are substantially certain” (CP 1.152). She makes it clear that practical certainty involves freedom from any serious doubt, and she concludes that “[p]ractically, we must treat some hypotheses as certain” (Misak 1991/2004, p. 54). But this falls far short of Cooke’s claim that anything we genuinely believe must function as a practical certainty for us. Again, a clearly sufficient condition of belief is being treated as a clearly necessary one. Though it is not evident in the passage cited by Cooke, some slippage between weaker and stronger construals of the doubt-belief contrast can in fact be found in Misak’s work. “An inquirer has a body of settled belief; a set of beliefs which are, in fact, not doubted” (Misak 1991/2004, p. 47). This formulation, like those discussed above, requires only the weaker version of the contrast, but invites interpretation in terms of the stronger version. It may well be true that Peirce holds that an inquirer must have a set of beliefs which are not doubted. And it certainly seems plausible that such beliefs should count as settled. But the passage suggests a much stronger claim, viz. that only beliefs which are, in fact, not doubted, count as settled. Again, I have not yet argued against this claim, but only against a conflation of it with something less demanding. Other passages continue the creep toward the strong version of the contrast. Peirce, Misak says, “uses ‘infallible’ to make the point that we do not doubt what we believe” (Misak 1991/2004, p. 52). This could be read weakly to say that “to the extent that we doubt, we do not believe,” but it is more naturally taken to say that if we believe, we do not doubt. Similar considerations apply to Misak’s claim that “the inquirer must regard her settled beliefs as infallible, in the sense that she does not doubt them for the purposes of inquiry; science has ‘established truths’ to be used as premises in further deliberation” (Misak 1991/2004, p. 54). A weaker reading can allow that some settled beliefs must remain free from doubt and that this is sufficient for their status as “established truths” in the context of deliberation. But proponents of the less demanding version of the contrast will deny that only beliefs treated as infallible will count as settled or can be used as premises. Since it is clear that Peirce thinks of belief as a settled state, it seems reasonably clear that Misak is here attributing the strong reading of the belief-doubt contrast to Peirce. More recent work by Misak fits this pattern as well. Again, it is useful to begin with an ambiguous formulation and witness its quick inflation to a more demanding construal of the contrast. She begins unobjectionably. “Inquiry begins with the irritation of doubt … it ends once we get another stable doubt-resistant belief” (Misak in Olsson 2006, p. 19). The issue at hand concerns how demandingly “stable” and “doubt-resistant” are to be taken. The strong reading of the doubt-belief contrast insists that beliefs are not stable unless they are practically certain and free of all genuine doubt. It also requires that a belief be doubt-free in order to count as doubt-resistant. Misak’s commitment to the standard, strong version of the contrast emerges a couple of pages later when she characterizes Peirce as rejecting the Cartesian insistence “that global doubt is the way to approach the task of justifying beliefs that are current and hence, not subject to doubt” (Misak in Olsson 2006, p. 21). The inference marked by “hence” clearly requires the premise that anything currently believed is not subject to doubt. It is no accident that Misak’s commitment to the strong version of the contrast emerges most clearly when she is juxtaposing Peirce’s work with that of Isaac Levi. As Misak puts it, “Levi is now the bearer of Peirce’s torch against Cartesian epistemology” (Misak in Olsson 2006, p. 20). Levi has developed the strong, standard reading of the opposition between doubt and belief more explicitly and more forcefully than anyone else. Both as a matter of Peirce interpretation and in his own development of the doubt-belief model of inquiry, Levi has emphasized the status of an agent’s current beliefs as resources for inquiry, and he thinks that those beliefs must be free from doubt in order properly to serve as resources. Erik Olsson’s introduction to a volume on Levi’s pragmatism provides a representative distillation of Levi-inspired approaches to Peirce. “All of our beliefs,” Olsson says, “insofar as they are genuine convictions of ours, are things we accept as true without a moment’s hesitation. There is, on our part, no ‘real and living doubt’ that they are true” (Olsson 2006, p. 2). “There is no point in inquiring further” once we have attained a state of belief, “because there is no serious possibility that things are otherwise than we believe them to be” (ibid). Misak nicely brings out Levi’s decision-theoretic development of Peirce: “Peirce’s notion of genuine or live doubt is transformed by Levi into the notion of serious possibility” (Misak in Olsson 2006, p. 21). The opening sentences of Levi’s The Fixation of Belief and Its Undoing show how quickly and decisively Levi moves from a generic to a demanding characterization of Peirce’s approach to inquiry: “As inquiring agents modify their evolving doctrines, they come to believe where they were initially doubtful. Conjectures or hypotheses are thereby converted into settled assumptions free from serious doubt and, therefore, counted as certainly true” (Levi 1991, p. 1). The second sentence, as Levi clearly understands, goes far beyond anything contained in the first. The notion of a settled belief is construed demandingly so as to be identified with “settled assumptions free from serious doubt and, therefore, counted as certainly true.” Levi, in effect, requires that a settled belief be an entirely settled belief, one free of any trace of (serious) doubt. With this understanding of how settled settled beliefs are, it follows that the believer should feel perfectly certain that each of her current beliefs is true. Levi has developed and defended this position in impressive detail. But it does not pass unnoticed by Levi and his followers what a strong position they are attributing to Peirce, and it should not go unnoticed by us. The commonsense opposition of belief to doubt and the plausible idea that inquiry seeks to settle opinion have been amplified into a position that sounds alarmingly like the method of tenacity described by Peirce in Section V of “Fixation.” As Peirce says there of the proverbial ostrich, why should it raise its head to see if it feels perfectly sure that there is no danger? (W3 249). Levi develops his own position around the Peircean term “full belief.” The “assumption” that an inquiring agent “judges it impossible” that her current state of full belief contains error is taken by Levi to be essential to the doubt-belief model of inquiry. That model, says Levi “requires” taking the assumption that the inquirer’s current state of full belief is entirely error-free “very earnestly” (Levi in Olsson 2006, p. 332). Even if an approach to inquiry makes belief and doubt the central states and even if it opposes them to one another, the approach will not, for Levi, count as Peircean unless it makes crucial use of maximally settled beliefs. “Full beliefs,” says Levi, “carry maximum degree of belief and their complements carry maximum degree of disbelief” (Levi 2012, p. 170). It is crucial to recognize that Levi does not simply oppose full beliefs and partial beliefs in the sense of degrees of belief. He has long been interested in a wide range of measures of cognitive commitments, including degree of belief, degree of probability, degree of confidence, informational value, and weight of evidence. See Levi 1967, for starters. And he allows for, though he does not make much use of, a notion of mere or plain belief that is qualitative but which does not serve as a standard of serious possibility (Levi 2012, p. 176). Philosophers like Jonathan Adler have argued that practical constraints involving norms of assertion, reactive attitudes like resentment, and our inability to track evidence for our beliefs all compel us to adopt a qualitative attitude of believing. It would be costly if we hedged all of our assertions and reactions with qualifiers about the extent of our entitlement to what we’re claiming. Belief figures here as a qualitative, threshold notion rather than a degreed one, but the threshold falls well short of certainty or maximal belief. It is telling that Adler describes such cases as “constraints on us to fully believe,” though he clearly has in mind a less demanding conception of full belief than does Levi. For Adler, transparency is crucial to full belief; one fully believes when “[i]n belief’s everyday roles, prominently as guides to action, one sees through one’s attitude to the world without seeing that attitude” (Adler 2002, p. 11). For Adler, “full belief, though demanding, is ordinary, common belief – the notion that is expressed in assertion” (Adler 2002, p. 231). And Adlerian full belief is indeed demanding. Adler and Levi agree that if one acknowledges counterevidence to a proposition, one no longer fully believes that proposition. Such a standard threatens to render unintelligible a position like Thomas Kuhn’s according to which scientists remain dogmatically confident in the truth of some constitutive claims of paradigms even while acknowledging serious problems with those claims. Adler and Levi also agree that full belief involves “the all-out acceptance of a proposition,” and that acceptance in this sense marks the end of inquiry, rather than merely issuing in a modification of one’s degree of belief (Adler 2002, p. 40, p. 232). Nevertheless, as we will see in more detail below, Adler’s full belief is compatible with genuine doubt and so is a much weaker notion than Levi’s. It refers to an attitude that is unqualified, but potentially much less confident than what Levi has in mind. Levi, in other words, goes beyond even Adler’s strenuous construal of full belief in thinking that such unqualified acceptance must also involve the adoption of a standard of serious possibility. Levi’s phrase “full belief” derives from Peirce, but not from “Fixation.” Peirce used the term in 1898, but not in 1877, and there are more than a few reasons for doubting how representative Peirce’s remarks about doubt, belief, action, and science in these Cambridge Conference lectures are of his considered opinions. For discussion, see Misak 2004 and Migotti 2005. Levi’s interests are not primarily scholarly, and he characterizes himself as offering a “reconstruction” of Peirce. Levi emphasizes that he is offering a reconstruction of Peirce’s notion of truth in Olsson 2006, p. 328. I conjecture that Levi would not object to a characterization of his broader presentation of Peirce as a reconstruction. Among other complications ignored here, Levi emphasizes that he thinks of beliefs in terms of commitments, rather than feelings. It is an intelligent and influential reconstruction, however, and for that reason, I will concede the term “full belief” to him. So I will stipulate it to be a conceptual truth that full belief precludes all genuine doubt, since that is how Levi (and many other Peirce scholars) understand the term. Two points are crucial for current purposes, however. First, it is not as obvious as it is often taken to be that Peirce relies on a notion of full belief in Levi’s sense. Misak’s suggestion that Levi “transforms” the 1877 notion of living doubt into a standard of serious possibility accurately represents matters as I see them. The fact that Levi chose a striking term from a quite different Peircean text (one in which Peirce says seemingly unrepresentative things about belief) as his designation for his reconstructed Peircean notion invites confusion about Peirce’s intentions in “Fixation.” More importantly, though, the strong version of the doubt-belief contrast is not established even if one concedes an important role for Levi’s notion of full belief in “Fixation.” From the alleged fact that some beliefs must be free of all genuine doubt for inquiry to proceed, it does not follow that anything genuinely believed must be free of all genuine doubt. Accordingly, I do not here address the question of whether Peirce relies, in “Fixation,” or elsewhere, on a notion as strong as Levi’s full belief rather than a weaker but still genuine and unqualified notion of belief. I claim only that Peirce deploys a less demanding conception of belief than Levi’s either instead of or in addition to full belief. I will be suggesting that “Fixation” allows for the possibility that an inquirer can reasonably terminate inquiry into the truth of a proposition and can rely on the proposition as a premise in some or all of her deliberations without having the proposition count for her as certainly true and without it being free from all genuine doubt. I take no stand here on the quite different issue of whether some propositions must, for the Peirce of “Fixation,” be fully believed in Levi’s sense. I have identified a strand in the secondary literature that I think is genuine and powerful. I do not want to exaggerate the intensity or extensiveness of that strand, however. Under pressure from other Peircean texts, notably the 1890’s writings in which Peirce discusses full belief, we find Cooke, for instance, making room for a weaker notion of belief than we have seen so far. While “practice deals with beliefs held as infallibly true,” belief in such a strong sense has no place in scientific inquiry (Cooke 2006, p. 66). Still, “the scientist working in the here and now must have something like a belief in order to get scientific inquiry going” (Cooke 2006, p. 68). She goes a bit farther later on the same page, calling these “low-grade beliefs,” since the scientist is willing to act on them. But the grip of the strong version of the doubt-belief contrast, combined, no doubt, with remarks from the 1890’s like “[t]here is thus no proposition at all in science which answers to the conception of belief” (EP2 33) induces Cooke to resolve any ambivalence in the direction of the standard view. In wrestling with the same vexing 1890s lectures, Mark Migotti seems to arrive at a view more like the one I am proposing. He grants to Peirce considerable variation in the “sheer degree of credence, which might be defined as the inverse of the degree to which one would be surprised if proven wrong” with which a belief is held (Migotti 2005, p. 49). And he further credits Peirce with flexibility about the manner in which a belief is held; one can believe in either a provisional or a committed way. “Since the two dimensions cut across one another, it is possible to have either a high or a low degree of credence in either the theoretical or the practical manner” (ibid). Migotti is tolerably clear that he is attributing belief, not just degrees of belief, to Peirce: he speaks of weak practical belief and strong theoretical belief, for instance. He is not at all explicit about belief coexisting with doubt about the same question, but that is a tempting way to understand his suggestion that one can believe that the ice is thick enough to support one’s weight, while also being in such a state that one would not be all that surprised to be surprised about the matter. I do not know whether Migotti would endorse my proposal here, but I hope to secure for Peirce’s classic presentation of the doubt-belief theory the kind of flexibility that Migotti suggests for the view of the 1890’s. In doing so, I will suggest an additional dimension of confidence and stability, one that I think looms large in the Illustrations of the Logic of Science. “Fixation” and the Doubt-Belief Contrast I have so far argued that commentators sometimes slide rather readily from weak to strong formulations of the doubt-belief contrast. And I have insisted that the strong construal of that contrast is much less intuitively compelling than is the weak construal. But I have so far offered no argument at all against the claim that the strong version of the contrast is correctly attributable to Peirce. Indeed I have acknowledged the existence of one unambiguous passage from “Fixation” which seems to license the move from the weak claim that doubt and belief tend to preclude each other to the strong claim that genuine belief precludes all doubt. Peirce does indeed license an inference from the presence of doubt to the absence of belief. But, as clear as this passage is, it is not clear that it represents Peirce’s considered view. As part of his last attempt to prepare “Fixation” for republication (as part of Essays on Reasoning circa 1910), Peirce amended this crucial passage. He remains comfortable with his supposition that his reader will experience a real doubt of a proposition if she comes to see that her belief is determined by circumstances extraneous to the facts. But he distances himself from the implication that warrants attributing to him the strong version of the contrast. He no longer says that if a real doubt is experienced, the belief in question “ceases to be a belief.” Instead, he claims only that it “ceases, in some degree at least, to be a belief,” and this looks much friendlier to the weak than to the strong version of the contrast. See DeWaal 2014, p. 66. Similar considerations apply to another problematic passage, one which insists that reflection on the fact that a belief has been formed capriciously will cause a belief to be rejected, “[b]ut it will only do so by creating a doubt in the place of that belief” (W3 248). This formulation does seem to suggest that one either believes or doubts. But it is worth noting that Peirce omits this sentence from the French version of “Fixation” composed at around the same time, and he qualifies it in 1903 by adding “Unless, indeed, it leads us to modify our ideas.” See De Waal 2014, p. 55 and also CP 5.375n. We have, then, some evidence that Peirce distanced himself from the formulations that most forcefully endorse the strong version of the belief-doubt contrast. Other passages in “Fixation” encourage, without mandating, a reading according to which belief and doubt can coexist about the same matter. As was noted above, in setting up the belief-doubt contrast, Peirce claims that belief guides action “according to its degree” (W3 247). He stated this even more forcefully in the manuscripts from which the main sections of “Fixation” were ultimately drawn. In one of those manuscripts (from 1872), he notes that “[b]elief and doubt may be conceived to be distinguished only in degree” (W3 21). The idea that belief comes in degrees, especially when conjoined to the suggestion that belief and doubt might blend into one another, sounds like an embarrassment for the standard reading, as indeed I think it is, but we must not be hasty here. Even Levi’s full beliefs can admit of degrees in some sense, all while remaining maximally certain, so Peirce’s talk of degrees of belief is not directly fatal to the standard view. Perhaps Peirce means something more like degrees of vulnerability to undermining than he does degrees of certainty. For some discussion of this distinction, see the exchange between Levi and Hansson in Olsson 2006. This proposal will get some attention at the end of this paper. For now it is enough to insist that both the published passage from 1877 and the unpublished one from 1872 suggest that confidence and action-guidingness can come in degrees, and that this presents at least the beginnings of a problem for the idea that anything believed is not at all doubted. I suggest that the classic formulations from “Fixation,” taken as a whole, offer as much support for my modest reading as they do for the more ambitious standard reading. The claim that “as soon as a firm belief is reached we are entirely satisfied” seems to favor the view that belief (or firm belief, at least) excludes all doubt (W3 248). But it permits a reading according to which “firm” is read rather ambitiously, and genuine beliefs can possess degrees of firmness. Degrees of firmness could mean degrees of confidence, degrees of resistance to revision, or any of a number of other possibilities. This matter will be addressed somewhat in the next section. In a fuller treatment, some discussion of the value of information would also be needed. Such a reading receives support from the claim on the same page that “[i]f the premises are not if fact doubted at all, they cannot be more satisfactory than they are.” This passage strongly suggests that premises could be satisfactory even if they could also be more satisfactory than they are. See also W3 16 (1872): “Any settlement of opinion, … if it is full and perfect, is entirely satisfactory and nothing could be better.” This seems to allow for imperfect, but genuine and valuable cases of settled opinion. Even Peirce’s characterizations of belief and doubt themselves fit the weak reading somewhat better than they do the strong one, I think. Peirce suggests that “nervous associations” are the “analogue of belief, in the nervous system,” while doubt can be thought of along the lines of “the irritation of a nerve and the reflex action produced thereby” (W3 247). It is perfectly plausible, though, that a nervous association can persist, along with its action-guiding functions, through the presence of an inhibiting stimulus. We can grant that the stimulus weakens the habit, but it seems implausible to insist that it must annihilate the habit, and to grant that is almost to concede the possibility that I am suggesting. Similarly, when Peirce associates belief with the feeling involved in wanting to pronounce judgment and doubt with the feeling involved in wanting to ask a question, we should remind ourselves that we quite ordinarily find ourselves genuinely wishing to do both of these things about the same topic. More generally, don’t we sometimes feel entitled to rely on an opinion for theoretical or practical purposes while also feeling, and so not just verbally acknowledging, a need to be clearer and more confident than we are? I read “Fixation” as permitting the flexibility that Migotti finds in Peirce’s work of the 1890s. Even such central formulations from “Fixation” as “with the cessation of doubt [inquiry] ends” (W3 247) fit comfortably into my reading. Peirce is clearly here providing a sufficient condition for the end of inquiry; he does not say that only with the cessation of doubt does inquiry end, much less that only with the cessation of all genuine doubt does inquiry end. We saw above how the temptation to conflate a sufficient condition with a necessary one eases the slide from weak to strong versions of the doubt-belief contrast. I have acknowledged that there is textual evidence in support of the strong version of the doubt-belief contrast. I maintain, however, that the textual evidence speaks equally strongly on behalf of the weak version of the contrast. And I think that evidence from elsewhere in the Illustrations, when combined with global considerations of charity and plausibility, favors attributing to Peirce the weaker version of the contrast. Stability, Weight of Evidence, and the Feeling of Belief “Fixation” and its companion piece, “How to Make Our Ideas Clear” have dominated discussion of the Illustrations. The doubt-belief theory and the pragmatic maxim for the clarification of concepts are important and intriguing, and gained fame when William James essentially made these papers the founding documents of pragmatism. But in the context of the series of papers in Popular Science Monthly of which they represent the first two installments, probability, not pragmatism, is the star. Indeed, neither “pragmatism” nor “pragmatic maxim” appears anywhere in the series. Peirce drew on his manuscripts from 1872 for the best known and most influential sections of “Fixation,” but he added two new introductory sections to that material, the first of which uses Clausius, Maxwell, and Darwin to set the table for a series which will include “The Doctrine of Chances,” and “The Probability of Induction.” With his Popular Science Monthly audience in mind, Peirce occasionally apologizes for the “tedious” way in which his now-famous papers have indulged in matters abstruse and metaphysical and indicates that it is only with the papers on probability that he will cross “the threshold of scientific logic” (W3 275). Peirce might or might not have had probability in mind when he first drafted the main sections of “Fixation,” but he almost certainly had it in mind as he prepared the paper for publication. See DeWaal 2014, especially p. 10. Taken together (Peirce wrote them as a single paper), “The Doctrine of Chances” and “The Probability of Induction” offer a pragmatic elucidation of the notion of probability and a novel assessment of the dispute between conceptualist (i.e. Laplacian or Bayesian) and materialist (i.e. frequentist) approaches to probability. One striking feature of these papers, given current purposes, is their emphasis on the feeling of belief. “Belief is certainly something more than a mere feeling; yet there is a feeling of believing, and this feeling does and ought to vary with the chance of the thing believed, as deduced from all the arguments” (W3 293). By “chance,” Peirce means the ratio of favorable to unfavorable cases. An event that has a .5 probability of occurring has an even, or 1 to 1, chance of occurring. Peirce focuses on chances rather than probabilities because probabilities are restricted to the range between 0 and 1, while chances can take on any value. Measuring belief intensity via chances has the consequence that a given item of evidence will have the same impact on one’s feeling of belief whether the prior state of belief is very high, very low, or anywhere in between. For discussion, see Reference Suppressed. Peirce quite resourcefully develops an account of appropriately felt belief on behalf of conceptualism. He proposes that the logarithm of the chance “as a thermometer for the proper intensity of belief” (ibid) because that magnitude has several desirable features, chief among them that “belief ought to be proportional to the weight of evidence, in this sense, that two arguments which are entirely independent, neither weakening nor strengthening each other, ought, when they concur, to produce a belief equal to the sum of the intensities of belief which either would produce separately” (W3 294). Alan Turing and I.J. Good See Good 1985. later and independently arrived at this conception of evidential weight, though without Peirce’s emphasis on the feeling of belief. Little can be said here about this prescient approach to inference and evidence, which Peirce calls the method of balancing reasons: “Take the sum of all feelings of belief which would be produced separately by all the arguments pro, subtract from that the similar sum for arguments con, and the remainder is the feeling of belief which we ought to have on the whole” (ibid). Our interest centers on Peirce’s reasons for rejecting the Bayesian approach he has just improved. Peirce objects to conceptualism in the way one would expect him to, viz. by rejecting indifference principles and radically subjective approaches to obtaining values for prior probabilities. But his first and most important objection anticipates not Bertrand, et al., but John Maynard Keynes. Though he used the phrase “weight of evidence” in developing the method of balancing reasons and doesn’t use it in his objection to that method, Peirce deploys the notion of weight of evidence that Keynes would later make famous in his 1921 Treatise on Probability. He here has in mind, not net weight of evidence, as in his anticipation of Good and Turing, but gross weight of evidence. We can imagine two different inquirers, one of whom has a small amount of evidence all of which points in the same direction. The other has a much larger body of evidence which, on the whole, lacks a clear valence. There is a sense in which the first inquirer is in a better situation than the second one: her evidence, such as it is, speaks clearly to the question of what she should believe. But there is a quite different sense in which the second inquirer occupies a more favorable evidential position. Her evidence is relatively silent about, say, the color of the next bean to be drawn from the bag. But her evidential situation is more settled than that of the first inquirer. Her opinions about the distribution of beans in the bag will not get undermined by a small run of black beans, for example. Important details remain, and it is unclear whether Peirce and Keynes have quite the same notion in mind. Peirce connects gross weight of evidence to the diminution of probable error, while Keynes denies such a connection. For discussion, see Kasser 2015. But both thinkers treat this sense of weight of evidence as a measure of confidence and of stability. Even thinkers who do not spend much time with Peirce use strikingly Peircean language to describe the effects of gross or Keynesian weight of evidence. Brian Skyrms, for example, says that weight of evidence tends to increase the “resilience” of probability judgments, while James Joyce uses the Peircean term “stability” to characterize the effects of increasing weight. See Skyrms 1977 and Joyce 2005. Both Peirce and Keynes seem to have missed one tricky point about evidential weight, however. Weight seems not to increase monotonically; new information can come in that lowers my confidence and increases my risk of error, as when I learn new relevant evidence that also makes me aware of new depths of my ignorance. I thereby increase my total evidence and increase my uncertainty by learning that there’s a new relevant information of which I’m ignorant. See Runde 1990 and Weatherson 2002. Interestingly, Keynes himself was ambivalent about the importance of the notion that he and Peirce formulated. See Adler 2002, p. 252. Peirce reveals no such ambivalence, however. He had argued, in “The Doctrine of Chances” that the “real and sensible difference between one degree of probability and another, in which the meaning of the distinction lies, is that in the frequent employment of two different modes of inference, one will carry truth with it oftener than the other” (W3 280). Probability judgments, he thinks, refer to real facts but ones that make essential reference to an indefinitely large number of inferences. For reasons that need not concern us here, Peirce is not assuming frequentism, though the passage does invite such a reading. Discussion of this point can be found in Kasser 2015. His objection to the conceptualist project stems not from an assumption of materialism about probability, but from the alleged impossibility of an adequate conceptualist treatment of gross weight of evidence which, Peirce thinks, also measures the probable error of the probability estimate. Now, as the whole utility of probability is to insure us in the long run, and as that assurance depends, not merely on the value of the chance, but on the accuracy of the evaluation, it follows that we ought not to have the same feeling of belief in reference to all events of which the chance is even. In short, to express the proper state of our belief, not one number, but two are requisite, the first depending on the inferred probability, the second on the amount of knowledge on which that probability is based. (W3 295) We need not concern ourselves with whether conceptualists can answer Peirce’s challenge. It is the Keynesian notion of evidential weight itself that will provide an important ingredient of the case against the standard reading of Peirce on belief and doubt. Peirce has explicitly decomposed the appropriate feeling of belief into two distinct magnitudes, one of which represents the balance of one’s reasons, and the other of which represents one’s total amount of evidence, Or, perhaps, something like one’s estimation of the fraction of the total evidence that is in one’s possession. Keynes seems to have vacillated between these (and perhaps other) conceptions of weight, and perhaps Peirce did as well. See Runde 1990 and Weatherson 2002. and which cries out for interpretation in terms of Peircean notions like stability. And he is emphatic about the significance of gross weight of evidence. It is true that when our knowledge is very precise, when we have made many drawings from the bag, or, as in most of the examples in the books, when the total contents of the bag are absolutely known, the number which expresses the uncertainty of the assumed probability and its liability to be changed by further experience may become insignificant, or utterly vanish. But, when our knowledge is very slight, this number may be even more important than the probability itself; and when we have no knowledge at all this completely overwhelms the other….(W3 295-6). There can be little doubt that Peirce thought of this discussion in “The Probability of Induction” as consistent and continuous with what he wrote in “Fixation.” We have seen that, though “Fixation” tends to operate with a threshold conception of belief, it clearly makes room for a degreed notion as well. The decomposition of belief along two dimensions does not, of course, refute the standard reading. Genuine belief, we may suppose, simply requires that both thresholds be reached, and then degrees can perhaps be admitted both above the threshold and below it. Since Peirce clearly wants to allow that methods like tenacity could, in principle, fix genuine belief, he’s clearly not, in “Fixation” treating responsiveness to evidence as a success term. The tenacious believer might not have or need anything worth calling evidence, but does have a source of confidence that she regards as adequate for reaching the appropriate threshold. Full belief requires something close to maximal confidence along both dimensions, and if that is the only notion of genuine belief one wants to attribute to Peirce one simply sets the threshold all the way up at the top. But a deeper examination of gross weight of evidence reveals why it threatens the strong reading of the doubt/belief contrast. Weight of Evidence and the Compatibility of Belief and Doubt Jonathan Adler takes the distinction Peirce has just labored to draw and deploys it to some very un-Peircean sounding ends, especially if one has the standard reading of Peirce in mind. Adler describes himself as defending the compatibility of full belief and genuine doubt, but I have conceded the term “full belief” to Levi and his followers, so I will present Adler as defending the compatibility of genuine belief with genuine doubt. The balance or net weight of evidence has a valence; changes in the balance of evidence reflect changes in the credibility of the statement in question. Gross or Keynesian weight of evidence, however, has no directionality; the amount of evidence that bears on P is always identical to the amount that bears on ~P. Adler agrees with Peirce that these are independent aspects of the believer’s situation that cannot be integrated into a single judgment. For this reason, evidence that gross evidential weight is low must be treated very differently from evidence that a judgment is mistaken. To modify one of Adler’s examples, if you receive a letter of recommendation on behalf of a job candidate and later learn that the recommender is a careless letter-writer, it would be a mistake to lower your opinion of the candidate. Carelessness could as readily lead to an underestimation of the candidate as an overestimation. If you still retain adequate evidence concerning the candidate, you should maintain your belief. Evidence that undercuts one’s weight of evidence cannot be assimilated to counterevidence. Nevertheless, your evidence is less substantial than it had been, and your feeling of belief, to return to Peirce’s way of thinking about such cases, ought to reflect this diminution. Your belief should be retained, but with diminished confidence. We can readily distinguish between being asked our opinion about a question like “Which city has a larger population, Tokyo or Mexico City?” and being asked how confident we are in that opinion. The less confidently held opinion might still be genuinely believed. Similarly, Adler thinks that you have a new and genuine reason to doubt your previous belief, and to lower your confidence in it, but not new evidence against the truth of that belief. See Adler, p. 250. The standard reading of Peirce can correctly point out that an opinion need not be a belief; it can be more like a guess or hunch. But I think it need not and should not deny that genuine beliefs can be held with varying levels of confidence. I am more confident that the population of Mexico City exceeds that of Paris than I am that the population of Beijing does, but I believe both propositions. We have seen textual evidence that Peirce allows for beliefs to be held more or less strongly, and even Levi allows room for “mere belief” as opposed to “full belief.” Now that we have seen Peirce explicitly emphasizing the importance of non-directional components of appropriate feelings of belief, the standard view should allow confidence to vary independently of belief. While that might constitute a bit of a concession, it need not be a costly one. What the standard view needs to deny is that a recognition of diminished confidence can, for Peirce, be identified with a genuine doubt. Adler thinks that when you learn that your evidence isn’t as weighty as it had been, you acquire a good reason for doubting, but not (necessarily) a good reason for no longer believing. The standard view needs to block that possibility. Transposing into the Peircean key, Adler thinks that when you learn of the unreliability of the letter writer, your belief remains and remains justified (based on the rest of the applicant’s dossier) and, since belief is a settled state, your belief remains settled. But your belief is now less settled, more easily undermined than it had been. You feel no need to reopen inquiry into the truth of the proposition, but you are aware that you are now more open to reopening inquiry than you had been. As Migotti puts it, you would now be less surprised to be surprised than you otherwise would have been. See Migotti, p. 50. You remain willing to act on the proposition in question, but less assuredly than before. I hope that these no longer seem like manifestly implausible things for a Peircean to say about belief and doubt. But the standard view has become standard for a reason, and it is not without resources with which to back its insistence that if the proposition is doubted, it is not believed. The most promising path for the standard view involves granting Adler that diminished confidence entails being increasingly disposed to doubt, but denying that it involves actually doubting. Beliefs that are held less confidently or with decreased confidence, on this view, are not thereby less believed. Acknowledged reasons for doubting make a belief less stable, but they do not amount to doubts unless they make the proposition in question less believed. Even if we grant the standard view this move, some of the cleanliness and simplicity that motivated it will have been sacrificed. An inquirer can, on the view now under consideration, genuinely believe a proposition (e.g. that X is a careless writer of recommendation letters), genuinely believe that it provides an evidential reason for diminished confidence in a different proposition (e.g. that Y is a strong job candidate), self-consciously maintain the undermined belief less confidently, be less willing to rely on the undermined belief, etc. and yet not have experienced anything worth calling a doubt of the undermined proposition. I do not think that this position is unintelligible, and I am willing to consider it compatible with the standard reading, but a standard reading so modified has freed itself from much of what I found objectionable therein. It acknowledges that there are states in between genuine and paper doubt, for instance, that matter to understanding and motivating inquiry. It will cease explaining inquiring behavior solely in terms of pure belief and pure doubt. While acknowledging realistic continuities among “beliefish” and “doubtish” states is a start, I suspect that more concessions from the standard view are in order. Perhaps it can say that some beliefs are undoubted despite the believer’s awareness of their flimsiness and instability. Levi is right to insist that I can recognize some of my beliefs to be more underminable than others without thereby doubting any of them. Even so, the view is starting to sound unPeircean. Peirce’s insistence that belief is a stable state and his insistence on the centrality of gross weight of evidence to appropriately felt belief and to stability of belief make it natural to insist that perceived instability tends to undermine belief. If we are to attribute to Peirce a genuinely stability-centered epistemology, it will be difficult to keep doubt and acknowledged instability as far apart as the standard view requires. But the claim that belief and doubt are incompatible will really struggle with cases in which our evidence is weakened rather than weak. When we obtain evidence that our evidence has diminished, we have not received counterevidence. But this new evidence has acknowledged force in lowering confidence, and increasing caution. Those sound like reasons for doubting. We now feel and have a modest incentive to reopen inquiry, but not a reason to think that something we believed is false. At least in some such cases, it seems sensible that one should maintain one’s belief while being newly aware of weaknesses in it. And it seems that Peirce would like to avoid being interpreted so as to foreclose this possibility. Such an interpretation of Peirce will force us to reconsider what constitute “positive reasons for doubting” and what relationship those have to “recalcitrant experiences.” The standard reading there presses us toward treating the exemplary case of a belief being shaken by clear counterevidence as the only case. Here as before, I think that Peirce shows more flexibility than do his interpreters; the method of authority, for instance, does not get undermined because authorities make clearly false predictions. Discussion of this argument can be found in Kasser 2011. Such additional Peircean heresies will have to await another occasion. It is also worth noting that I have made significant concessions in the direction of the standard view. I have granted, arguendo, to Adler and Levi that genuine belief is incompatible with acknowledged counterevidence and have instead focused on epistemic underminers that do not amount to counterevidence. But there is room to develop a more “ambitiously modest” reading of Peirce on belief and doubt that finds room for him to reconcile belief and acknowledged counterevidence. Such a reading would make room for Kuhn’s not obviously outlandish suggestion that confident belief and acknowledged counterevidence can coexist. I do not by any means claim to have refuted the standard reading of Peirce. Levi, in particular, is aware of Adler’s arguments and similar ones from thinkers like Hansson, and he professes to be untroubled by them. See Levi in Olsson 2006, pp. 335-340 and Levi 2012, Chapter 9. He recognizes that a proposition can be sufficiently certain for a given practical purpose, but is not, in general, interested in that kind of certainty. And he sees little need for a notion of “sufficiently but not maximally certain for theoretical purposes.” On this score, however, Levi is defending himself, not Peirce. Both Levi and his critics tend to associate Peirce with Levi’s side in such debates. I suggest that, while Levi’s powerful view is inspired by Peirce’s doubt-belief theory of inquiry, it may be that Peirce anticipated, to some degree at least, not only Levi but also Adler and Hansson. I see Peirce as placing the same premium upon settled belief that Levi does, but as availing himself of a more nuanced conception of stability than does Levi. I hasten to add that Levi’s conception of such stability-centered notions as corrigibility is quite nuanced indeed. He just does not feel much of a need to avail himself of some of these nuances in explaining inquiry. What I do hope to have established here is that the standard view ought not to be taken for granted. It is okay with me if proponents of the standard view still believe it, especially if they now have their doubts about it. I am grateful to Mark Migotti, Robert Lane, and an audience at the Society for the Advancement of American Philosophy 2016 annual meeting (especially Robert King, Nicholas Guardiano, and Joseph Urbas) for valuable comments and discussion. WORKS CITED Adler, Jonathan. Belief’s Own Ethics. Cambridge: MIT Press, 2002. Cooke, Elizabeth. Peirce’s Pragmatic Theory of Inquiry: Fallibilism and Indeterminacy. London: Continuum, 2006. DeWaal, Cornelis. (Ed.). Peirce’s Illustrations of the Logic of Science. Chicago: Open Court, 2014. Fisch, Max. Peirce, Semeiotic, and Pragmatism: Essays by Max Fisch. Edited by K. Ketner and C. Kloesel. Bloomington: Indiana University Press, 1986. Joyce, James. “How Probabilities Reflect Evidence.” Philosophical Perspectives 19: 153-178. 2005. Kasser, Jeff. “How Settled are Settled Beliefs in ‘The Fixation of Belief’?” Transactions of the C.S. Peirce Society, Volume 47 Number 2, pp. 226-247, 2011. Kasser, Jeff. “Two Conceptions of Weight of Evidence in Peirce’s Illustrations of the Logic of Science,” Erkenntnis, DOI 10.1007/s10670-015-9759-5, 2015. Levi, Isaac. Gambling with Truth. New York: Alfred A. Knopf, 1967. Levi, Isaac. The Fixation of Belief and Its Undoing: Changing Beliefs Through Inquiry. Cambridge: Cambridge University Press, 1991. Levi, Isaac. Pragmatism and Inquiry. Oxford: Oxford University Press, 2012. Migotti, Mark. “The Key to Peirce’s View of the Role of Belief in Scientific Inquiry.” Cognitio 6:43-55. 2005. Misak, Cheryl. Truth and the End of Inquiry: A Peircean Account of Truth. Oxford: Oxford University Press, 1991/2004. Misak, Cheryl. “C.S. Peirce on Vital Matters” (In C. Misak (Ed.), The Cambridge Companion to Peirce (pp. 150-174) Cambridge: Cambridge University Press, 2004. Olsson, Erik, (Ed.). Knowledge and Inquiry: Essays on the Pragmatism of Isaac Levi. Cambridge: Cambridge University Press, 2006. Peirce, Charles S. Writings of Charles S. Peirce: A Chronological Edition. Edited by the Peirce Edition Project. Bloomington: Indiana University Press, 1982-. References have the format W(Volume Number), page number. Runde, J. “Keynesian Uncertainty and the Weight of Arguments.” Economics and Philosophy, 6: 275-292. 1990. Skyrms, Brian. “Resiliency, Propensities, and Causal Necessity.” Journal of Philosophy, 74: 704-713. 1977 Weatherson, Brian. “Keynes, Uncertainty, and Interest Rates.” Cambridge Journal of Economics, 26: 47-62. 2002. PAGE \* MERGEFORMAT 1