Erkenn (2013) 78:451–468 DOI 10.1007/s10670-011-9351-6 ORIGINAL ARTICLE
Regulative Assumptions, Hinge Propositions and the Peircean Conception of Truth Andrew W. Howat
Received: 18 April 2011 / Accepted: 9 November 2011 / Published online: 4 December 2011 Springer Science+Business Media B.V. 2011
Abstract This paper defends a key aspect of the Peircean conception of truth—the idea that truth is in some sense epistemically-constrained. It does so by exploring parallels between Peirce’s epistemology of inquiry and that of Wittgenstein in On Certainty. The central argument defends a Peircean claim about truth by appeal to a view shared by Peirce and Wittgenstein about the structure of reasons. This view relies on the idea that certain claims have a special epistemic status, or function as what are popularly called ‘hinge propositions’. Let us not pretend to doubt in philosophy what we do not doubt in our hearts. (Peirce, SC, p. 141) So I am trying to say something that sounds like pragmatism. (Wittgenstein, OC §422)1 Charles Sanders Peirce considered this claim (or one very like it) to be some part of our ordinary conception of truth2: (T)
1
If a hypothesis H is true, then (if inquiry into H were pursued long enough and well enough, then H would be believed).3
Throughout, OC denotes Wittgenstein (1969) On Certainty.
2
See esp. Misak (2004), who argues that Peirce also endorses the reverse claim—i.e. If (if inquiry into a hypothesis H were pursued long enough and well enough, then H would be believed) then H is true. I do not defend this claim here. Note I have also changed Misak’s wording by substituting ‘long enough and well enough’ for ‘as far as it could fruitfully go’.
3
Like Peirce I prefer to think that of propositions as the bearers of truth and falsity. I’ve used ‘hypothesis’ here just to emphasise that this is a claim about propositions into whose truth we might inquire. The relevance of this emphasis should become clear as we proceed.
A. W. Howat (&) Department of Philosophy, California State University, Fullerton, Fullerton, CA 92834-6868, USA e-mail:
[email protected]
123
452
A. W. Howat
(T) is not a conceptual analysis of TRUTH in the conventional sense; it does not purport to define or to provide necessary and sufficient application-conditions for our concept TRUE.4 Instead, (T) is a so-called ‘pragmatic elucidation’ of TRUTH. This means (T) tells us what practical or experiential consequences we can expect when TRUE applies to a hypothesis (namely that our inquiries concerning that proposition will have a certain kind of outcome).5 Despite its unusual status, many critics (even those broadly sympathetic to Peirce or to pragmatism) consider (T) to fly in the face of non-negotiable intuitions about the objectivity of truth. This generates a standard objection to (T)—that it wrongly undermines the idea that whether or not a proposition is true is entirely independent of whether or not anyone does, will or ever would believe it. In what follows, I provide a new way to defend the Peircean view from this objection, by focusing on the special epistemic status of (T). In §1 I set out the problem of lost facts—the standard objection to (T). I also explore a familiar riposte—the appeal to regulative assumptions (the ARA). I then suggest a novel way to develop and defend the ARA—by claiming that (T) is what Wittgenstein called a ‘hinge proposition’ (§2). I then defend this claim from some objections (§3).
1 An Objection to (T) and a Response One way to defend and elaborate (T) further is by addressing the most telling and difficult objection to it—the ‘problem of lost facts’. For H1 read anything that is plausibly both (a) a true proposition and (b) such that no amount of diligent inquiry will ever plausibly lead to a stable belief in its truth. For example, H1 might be the proposition that ‘‘Churchill sneezed 52 times in 1945’’6: 1. 2.
3. 4. 5. 6. 7.
If a hypothesis H is true, then (if inquiry into H were pursued long enough and well enough, then H would be believed). (i.e. assume (T) for reductio) H1 is true. (Supported by the broadly realist intuition that there is (e.g.) some definite number of times Churchill sneezed in 1945. Prima facie it does not matter that we are unable to determine the correct number.) If H1 is true, then (if inquiry concerning H1 were pursued long enough and well enough then H1 would be believed). (from 1 and 2) It is not the case that (if inquiry concerning H1 were pursued long enough and well enough then H1 would be believed). (i.e. the relevant fact is ‘lost’) So H1 is not true. (from 1, 3 and 4) H1 is true and H1 is not true. (i.e. assuming (T) generates a contradiction, from 1 to 5) So it is not the case that if a hypothesis H is true, then (if inquiry into H were pursued long enough and well enough, then H would be believed). (i.e. by RAA the Peircean conception of truth is false, from 1 to 6)
4
Small caps denote the concept as distinct from the thing itself.
5
Many critics overlook this vital point, e.g. Wright (1992).
6
The example is Misak’s (2004, p. 139), and originally inspired by Smart (1986).
123
Regulative Assumptions, Hinge Propositions
453
Peirceans such as Cheryl Misak typically avoid tackling this argument head-on. Instead, her reply (similar to one explored by Chris Hookway) is one I call the Appeal to Regulative Assumptions.7 (ARA)
We Peirceans do not assert (T), nor are we committed to saying we can have adequate justification to do so. Our claim is that (T) is a regulative assumption of inquiry, i.e. one must assume (T) in order to inquire rationally into the truth of any hypothesis. If one did not assume (T), then there would be no point inquiring.
What is the basis for this move? Peirce argues that what we aim for when we inquire is an especially stable sort of belief (EP 1.114). This is because inquiry is motivated by doubt, which Peirce describes as a kind of irritation—the irritation of not knowing something, or of suspecting one’s existing beliefs to be inaccurate or incomplete. Peirce’s thought is that when we inquire, our goal is to eliminate this nagging doubt once and for all. To do this, we seek to replace the doubt with a settled belief, one that will remain stable in the face of all future evidence and inquiry.8 Once we have a belief with this feature, we have inquired ‘long enough and well enough’.9 Taking this into account, the ARA says this: if one did not assume that properly conducted inquiry will yield a belief that will remain forever stable, then there would be no point inquiring, since the need or desire for such a belief is what motivates inquiry in the first place. The ARA, as usually articulated in work on Peirce’s conception of truth, is incomplete and vulnerable in several ways. It invokes a new technical notion (the ‘regulative assumption’) that requires some intuitive or theoretical precedents to prevent the reply appearing ad hoc. I explore this vulnerability and propose such a precedent in §2 and defend it from the most obvious and immediate objections §3.
2 Regulative Assumptions as Hinge Propositions I take it the most important component of the ARA is the claim that while we lack justification to believe (T), we are nevertheless entitled to assume that (T) is true (and must do so, on pain of irrationality). It is clear that we lack justification for (T) because of the problem of lost facts. It is quite possible, if not likely, that there are some true propositions that we cannot know, however long or carefully we inquire. The Peircean does not want to be in the position of flatly denying this. The ARA tells us that we are nevertheless ‘entitled to assume (T)’ because it is a ‘regulative assumption’ of inquiry. As typically formulated, it does not tell us much about what this means, what is meant by ‘entitlement’, nor how it differs from 7
See esp. Misak (2004) and Hookway (2002).
8
Peirce argues that our seeking the permanent fixation of belief will rule out things like wishful thinking (the ‘method of tenacity’) and belief that is forced or coerced (the ‘method of authority’).
9
It’s worth noting now (as I do again later), that Peirce’s view does not commit us to the view that one must be able to recognise when one has inquired long enough and well enough, that is, to be in something like epistemically ideal circumstances in order for us to know the truth of a proposition. This neutralises a popular objection to Peirce’s conception of truth, see e.g. Wright (1992).
123
454
A. W. Howat
traditional notions of epistemological warrant or justification. The appeal to a special kind of epistemological status for (T) requires some motivation or support that is independent of its supplying a reply to the problem of lost facts. There are a number of helpful suggestive remarks made by Misak and by Peirce himself we can consider as our starting point, alongside some of the history of the notion of regulative assumptions.10 Misak (2004, p. 68) describes a regulative assumption as one ‘such that, without making it, the participants in a practice could make no sense of that practice’. Peirce offers this rather nascent explanation of how such assumptions operate: When a hand at whist has reached the point at which one player has but three cards left, the one who has the lead often goes on the assumption that the cards are distributed in a certain way, because it is only on this assumption that the odd trick can be saved. That is indisputably logical… (CP 2.113) The notion of a regulative assumption or principle has its roots in Kant’s Critique of Pure Reason, in which they are distinguished from ‘constitutive principles’.11 For Kant, the distinction seems to be one between two types of status we can accord certain special principles. Kant argues that we ought to employ transcendental principles, those capturing the conditions of the possibility of various things (such as inquiry or reason), in a regulative manner in order to avoid certain philosophical mistakes: …transcendental ideas never allow of any constitutive employment. When regarded in that mistaken manner, and therefore as supplying concepts of certain objects, they are but pseudo-rational, merely dialectical concepts. On the other hand, they have an excellent, and indeed indispensably necessary, regulative employment, namely that of directing the understanding towards a certain goal…12 Jonathan Bennett infers the following from this and other passages about the nature of regulative principles: A ‘regulative’ principle seems to be one which is neither provable nor disprovable, but is a useful guide to scientific inquiry… Kant says that regulative principles ‘contribute to the extension of empirical knowledge, without ever being in a position to run counter to it’… he says that regulative principles are defensible only as ‘heuristic principles’ which can be ‘employed with great advantage in the elaboration of experience.’13
10 An anonymous referee helpfully recommends Misak (forthcoming) as a source of further clarity. Regrettably the author was unable to acquire a copy prior to the completion of this paper. 11 Peirce saw his own philosophical project as an attempt to ‘correct and develop Kant’s philosophical vision’. Hookway (1998, p. 270). The relevant passages in Kant’s Critique are B221-2, B536-37, B672 and B675. 12
B672.
13
Bennett (1974, p. 271).
123
Regulative Assumptions, Hinge Propositions
455
All of the above suggests we might thus construe the ARA as the claim that (T) is a transcendental principle capturing a condition on the possibility of inquiry. We could then answer the problem of lost facts by conceding that (T) may only be considered regulative rather than constitutive. In positive terms, this would mean the ARA involves asserting that (T) is indispensably necessary to inquiry and that it is neither provable nor disprovable. The denial of constitutive status to (T) would, according to Bennett, involve denying ‘that what entitles us to accept [it] is our ‘insight’ into reality’ as it is in itself, and the possibility that (T) could ever ‘run counter to’ empirical knowledge.14 Whilst this does help to clarify matters somewhat, the Kantian distinction between regulative and constitutive principles is not a legitimate or plausible way of explaining Peirce’s point about truth and inquiry. There are two reasons for this. First, Peirce rejects the distinction altogether, along with the very coherence of its basis—Kant’s notion of reality-in-itself.15 Thus, it would be no concession at all for Peirce to say that (T) should regarded as ‘merely’ regulative—since it is unclear what ‘higher’ epistemic status could thereby be denied to it. Second, Peirce explicitly rejected the transcendental approach, along with the sort of indispensability argument that goes with it. It may be indispensable that I should have $500 in the bank – because I have given checks to that amount. But I have never found that the indispensability directly affected my balance in the least… A transcendentalist would claim that it is an indispensable ‘presupposition’ that there is an ascertainable true answer to every intelligible question. I used to talk like that, myself; for when I was a babe in philosophy my bottle was filled from the udders of Kant. But by this time I have come to want something more substantial.16 So we know that Peirce wanted to deploy one half of a Kantian contrast—the regulative principle—though to deny the coherence of the other half—the constitutive principle. We know he did not want to make use of a straightforward transcendental or indispensability argument. So we still lack a clear, independently motivated understanding of what special epistemic status we can accord to (T) that is distinct from straightforward (justified or unjustified) belief.17 This is why I
14
Ibid. p. 278.
15
CP 3.215 ‘Kant’s distinction of regulative and constitutive principles is unsound.’
16
Both quotations CP 2.113, emphasis added.
17
One answer I do not consider here is Peirce’s suggestion that ‘all that logic warrants is a hope not a belief’. Hookway (2002), for example, prefers to construe regulative assumptions as justified hopes. This approach has clear exegetical advantages. It also accommodates various claims prevalent in Peirce’s later work: that belief has no role to play in scientific investigation; that at best we may hope that our hypotheses are true, because we lack the reflective justification required for belief; that belief is only appropriate in vital matters (see e.g. CP 1.635). Despite its exegetical strength, I have doubts about the likely success of this strategy and propose to criticize it in future work. My main concerns are (a) if we understand belief as a disposition to act (as Peirce does), then there is no pragmatically legitimate way to draw the distinction Peirce needs here between a rational/justified hope and a belief, and (b) I am sceptical about the prospects of identifying a distinctive set of norms governing hopes versus beliefs to which we can appeal in saying a hope is ‘justified’ or ‘rational’.
123
456
A. W. Howat
suggest that the best way to articulate and defend Peirce’s view is by appealing to Wittgenstein’s notion of a hinge proposition. …the questions that we raise and our doubts depend upon the fact that some propositions are exempt from doubt, are as it were like hinges on which those turn. That is to say, it belongs to the logic of our scientific investigations that certain things are in deed not doubted… But it isn’t that the situation is like this: We just can’t investigate everything, and for that reason we are forced to rest content with assumption. If I want the door to turn, the hinges must stay put. (OC, §§341–343) There is, I think, a very strong case for equating these two notions—regulative assumptions and hinge propositions—or, better, for explaining the former in terms of the latter.18 What inspires this case is the observation that Peirce and Wittgenstein’s views about knowledge and doubt exhibit a sort of symmetry. In On Certainty (OC), Wittgenstein famously objects to a set of epistemological claims (made by G.E. Moore) on the grounds that one cannot properly claim to know certain propositions.19 In Some Consequences of Four Incapacities (SC), Peirce famously objects to a different set of epistemological claims (made by Descartes) on the grounds that one cannot properly claim to doubt certain propositions. The inspiration for this paper is that we can explain these parallel objections by appeal to a single, shared view about inquiry and the structure of reasons: that some propositions are immune from both rational support (and thus genuine claims to knowledge) and rational criticism (and thus genuine claims to doubt), by virtue of the epistemological role they play in our lives and practices. My argument is that this shared view provides the pragmatist with a properly motivated and substantive strategy for defending the Peircean conception of truth (or at least the aspect of it embodied in (T)) from the problem of lost facts. What, then, is the Wittgensteinean hinge strategy and how might it help motivate the ARA? G.E. Moore famously attempted to defeat epistemological scepticism by enumerating various things of which he could be certain, such as his having two hands. In On Certainty, Wittgenstein explores and criticises this strategy in great detail. His chief complaint seems to be that Moore’s claims rest on a false picture of the structure of reasons.20 Claiming that one ‘knows’ something of which one is certain constitutes a misuse of the expression ‘‘I know’’; the very certainty of some propositions, such as ‘I have two hands’ means that we cannot properly claim to know them.21 One says ‘‘I know’’ when one is ready to give compelling grounds. ‘‘I know’’ relates to a possibility of demonstrating the truth. […] But if what he believes is of such a kind that the grounds that he can give are no surer than the 18 It may be best to think of regulative assumptions as a special sub-class of hinge propositions, distinguished from others by their role in regulating a practice or activity. 19
Please excuse the deliberate pun—both meanings of ‘certain’ are fitting in this case.
20
This is Pritchard’s (2011) way of putting it.
21
OC §6 and §11.
123
Regulative Assumptions, Hinge Propositions
457
assertion, then he cannot say that he knows what he believes. (OC, §243, my emphasis) I suggest that the Peircean claim (T) serves as a hinge proposition for us when we inquire, and that this is the best way to explain and to justify the ARA. In other words, that (T) is a hinge proposition helps to explain and to provide a non-ad-hoc account of the special epistemic status enjoyed by ‘regulative assumptions of inquiry’, amongst other propositions.22 Consider the way Wittgenstein and his contemporary followers describe hinge propositions. Their key features are: 1. 2. 3.
No knowledge: ceteris paribus one cannot properly claim to know hinge propositions, because No grounds: there are no grounds for hinge propositions that are more certain than the propositions themselves. Indubitable: Since they play a ‘supporting role’ in our belief system, hinge propositions are ceteris paribus indubitable.23 To doubt a hinge proposition, one would have to doubt most or all of one’s other beliefs at the same time (and thus, Wittgenstein suggests, be subject to a ‘mental disturbance’).24
I will say a little about each of these features by way of illustration, though the account that follows is necessarily brief, since my intent is not to provide a fully satisfactory interpretation or defence of the hinge strategy (in particular, I take no view here about whether or not it is a satisfactory answer to epistemological scepticism). Instead, I propose to show that if the hinge strategy is right in its ascription of a special epistemic status to some of our beliefs, then our belief in (T) has that status, and this fact allows us to answer the objection embodied by the problem of lost facts. All three features of hinge propositions are very closely related, so it is difficult to elaborate upon them separately. Pritchard (2011) claims that the key to the hinge strategy is what it says about the structure of reasons: ‘In short, the suggestion is that the very possibility that one belief can count as a reason for or against another belief presupposes that there are some beliefs which play the role of being exempt from needing epistemic support’. This observation on its own is relatively familiar, and could easily be seen as the motivation for traditional epistemological foundationalism (i.e. the postulation of certain self-evident truths by appeal to which all other beliefs are justified). However, the hinge strategy differs from foundationalism in a number of ways (more on this later). In particular, the strategy also comes with some striking and distinctive anti-sceptical views about doubt, as explained here by reference to the sceptical challenge to Moore’s proposition that ‘I have two hands’ (p):
22 It may be less confusing simply to drop the term ‘regulative assumption’ altogether in favour of hinge proposition, since the former has unhelpful Kantian overtones. 23
See Pritchard (2011, p. 7) for more on this point.
24
See OC §71–72 and Pritchard (2011, p. 11).
123
458
A. W. Howat
For example, suppose that I try to motivate the doubt [concerning p] on the grounds that I cannot at that moment see my hands. Presumably, however, I am less certain that my eyesight is functioning correctly than that I have two hands. Accordingly, on the Wittgensteinian picture of the structure of reasons the former cannot rationally be used as a ground for doubt regarding the latter. This is because the lesser certainty of the ground for doubt will inevitably ensure that it is more rational to doubt the ground for doubt than to doubt the target proposition. That is, I’ve more reason to doubt what I see (in this case that I don’t just now see my hands) than for doubting that I have two hands on the basis of what I (don’t) see. Doubt of that which is most certain is thus necessarily groundless, and hence incoherent (e.g. OC, §§4; 122–123).25 These views give rise, in turn, to the idea that doubting a hinge proposition makes one subject to a mental disturbance. This may sound extreme, but seems less so once placed within the context of sceptical doubts about propositions like p. Wittgenstein explains the idea by reference to an imagined scenario. A friend of mine has been living, say, at an address on Ryle Road, for 10 years. One day he comes to doubt the truth of the proposition that he has ever lived on Ryle Road. Concerning this kind of scenario, Wittgenstein writes: …I should not call this a mistake, but rather a mental disturbance, perhaps a transient one… But what is the difference between mistake and mental disturbance? Or what is the difference between my treating it as a mistake and treating it as a mental disturbance? Can we say: a mistake doesn’t only have a cause, it also has a ground? I.e. roughly, when someone makes a mistake, this can be fitted into what he knows aright. (OC §§71–74) The thought is that my friend’s doubt is just too big to be thought of as a mere mistake. Imagine if he were correct—he could not possibly assimilate this discovery into all of the things he previously took himself to know. How could he come to understand that and how he had made such a ‘mistake’, when his discovery will equally have undermined pretty much everything else he takes himself to know? As 25
Pritchard (2011, p. 6). I have one reservation about this, inspired by Peirce (via Bernstein 2010, p. 43), which seems to represent an objection to the above proposal. The objection is this: grounds that are individually less certain than the conclusion they support can provide grounds if they are more certain when taken collectively. For example, imagine you receive 20 distinct eye-witness reports about a shooting that took place on a busy shopping street. Each individual report may be hazy, shaky and vague in lots of ways. Yet if every single report agrees on the identity of the shooter, then when taken collectively, those individually uncertain reports are quite sufficient, I take it, to justify a belief about the identity of the shooter. Now Moore’s anti-sceptical move is clearly not analogous to the eye-witness case. His claim to know he has hands is based solely on immediate perceptual awareness. Taken this way, Wittgenstein’s objection stands. The objection does however raise an interesting possibility: that one might claim (legitimately, by Wittgenstein’s own standards) to know some hinge proposition, on the basis of many individually weak, but collectively strong pieces of evidence. The hinge strategist would presumably have to make some sort of concession (how significant, I don’t know), if, when taken together, those many weak claims are genuinely more certain than the conclusion. As far as I can tell, this is a bare possibility, and not something we need worry about unless or until someone can make a case for what is uncontroversially a hinge proposition being open to support from a large number of collectively more certain, individually weak pieces of evidence.
123
Regulative Assumptions, Hinge Propositions
459
Johanson (1994) puts it, if hinge propositions turned out to be false ‘everything would fall apart.’26 Supposing for the time being that these brief remarks about the hinge strategy makes the idea of a regulative assumption somewhat clearer,27 consider whether it seems plausible to ascribe these key features to the Peircean claim (T). Can one properly claim to know that (T) is true? Peirce presumably thinks not, otherwise he would not feel compelled to argue for its having a special epistemological status—something like a regulative assumption. He recognises that one might try to employ an indispensability or transcendental argument to establish these claims, but as we saw previously, professes to finding such arguments wanting (and thus presumably inadequate as epistemic justification for any belief in (T)). Just before the passage quoted above about playing whist, Peirce writes: I am not one of those transcendental apothecaries, as I call them - they are so skillful in making up a bill - who call for a quantity of big admissions, as indispensible Voraussetzungen of logic […] I reduce the indispensability of their postulates far more than that, namely, all the way from universality to the single case that happens to have come up; and even then, I do not admit that indispensability is any ground of belief. It may be indispensable that I should have $500 in the bank—because I have given checks to that amount. But I have never found that the indispensability directly affected my balance, in the least.28 This suggests though, that one could at least try to give grounds for (T). Of course, this will only be a genuine possibility if there are grounds one might give— so the key question is actually whether (T) plausibly has the second feature (No Grounds). Is it true that there are no grounds for (T) that are stronger than (T) itself? Is this what’s really behind its being impossible properly to claim knowledge of (T)? I suspect the answer here is yes. For in order to attain or to provide justification for (T), one would have to engage in a process of inquiry. That is, one would have to attempt to settle any doubts about the truth of (T) by replacing them with stable belief.29 But ex hypothesi, to inquire is already to presuppose the truth of (T). Thus, the attempt to provide epistemological justification for (T) will inevitably beg the question. Therefore, there are no grounds for (T) more certain than (T) itself. This claim is liable to sounding like a sort of sleight of hand, so consider an example. Imagine someone suggests that one might support (T) by means of an inductive argument. The argument appeals to past success in various scientific endeavours, to various hypotheses in which we now have stable beliefs after long and diligent inquiry. Suppose for a moment that this argument might be plausible 26
p. 172.
27
For those who feel it does not, the subsequent discussion of objections to the proposal will, I hope, supply greater illumination. 28
CP 2.113.
29
I suggest later that one cannot even raise legitimate or genuine doubts about the truth of (T).
123
460
A. W. Howat
(in other words, ignore the fact that most scientific theories have turned out to be false in at least some respects). Could it provide grounds for (T) stronger than (T) itself? No, of course not. For the argument is precisely an attempt to resolve a doubt about the truth of (T)—it is the product of an inquiry. For the argument to work, we already have to assume (T) is true—that this process of inquiry and argument is going to lead us to a right answer about whether or not (T) is true.30 Finally, is (T) indubitable in normal circumstances? Well, if ceteris paribus one seriously doubted that inquiring long enough and well enough led one to the truth, one would have good reason to doubt most, if not all of one’s other beliefs. This is because inquiry, as Peirce understands it (and as it figures in (T)) need not entail some highly formal, intellectual or academic enterprise. It is merely the process of trying to resolve genuine doubts by replacing them with stable beliefs. This is, arguably, a key feature of learning itself.31 If one had reason to doubt everything one had learned, then it seems reasonable to suggest that ‘everything would fall apart’ in the appropriate sense. Moreover, if someone had such a doubt, we could reasonably describe them as being subject to a ‘mental disturbance’, so long as their doubts were genuine, rather than merely affected.32 This is particularly so since such a person would presumably on the one hand experience doubt about almost everything but on the other refuse to inquire in order to settle those doubts, given that they doubt such inquiry could possibly achieve that goal.
3 Objections 3.1 A Peircean Objection Some Peirceans may worry that there is no precedent for ascribing this kind of view to Peirce himself, or that the view is inconsistent with what Peirce wrote on this topic. One particularly problematic example is Peirce’s thought that the sorts of indubitable propositions with which he is concerned were necessarily vague, and could not be precisely formulated.33 If that is so, then the whole enterprise of formulating and defending (T)—a precisely formulated and circumscribed claim— is at odds with Peirce’s true intentions. There are two reasons I do not find this objection compelling. First, my primary interest here is to find a way to defend the (uncontroversially Peircean) idea that truth is somehow epistemically-constrained, from a standard objection (the problem of lost facts). In doing so I would hope to remain true to the spirit, if not the letter of Peirce’s actual views, but this interest is secondary to 30 Another relevant observation on the No Grounds feature: the burden of proof—to come up with grounds for (T) stronger than (T) itself—would again seem to be firmly on the critic here, and the critic is unlikely even to want to meet this burden, since they think (T) is false anyway. 31
See e.g. Nesher (2001).
32
More on this in Sect. 3.3.
33
See e.g. CP 5.498, CP 5.505 or Johanson (1994, p. 175).
123
Regulative Assumptions, Hinge Propositions
461
finding a compelling defence of (T). If we must sacrifice elements of Peirce’s original view, so be it.34 Second, I believe there are some reasons to think that the proposal is broadly consistent with the spirit of Peirce’s actual views. Consider the influence of the Scottish common sense tradition on Peirce’s views. Hookway (1985, p. 229, emphasis added) finds this influence in three of Peirce’s doctrines in particular: 1. 2. 3.
justification ‘will have to come to a halt somewhere’ and rest upon some opinions which are accepted without grounds or justification.35 the beliefs which provide ‘the bedrock of truth’ are indubitable, and beyond rational support and criticism. ‘they must be regarded as the very truth’—they remain secure foundations, despite being without grounds and beyond support and criticism.
Moreover, as stated above, this may also be an attractive way to understand and defend Peirce’s famous attack upon Descartes’ methodological ‘paper’ doubts in the Meditations. The thought is that Descartes, like Moore, fails to recognise that when we inquire we must hold some claims to be true, despite a lack of grounds or evidential support. Thus, Descartes may pretend if he wishes that his method of doubt leads him to reject all knowledge except the cogito, but in doing so he is engaged in an inquiry; to be engaged in inquiry is to presuppose certain things, such as—that suitably diligent inquiry will yield the outcome one seeks (that is, Descartes presupposes (T)). One must presuppose these things not only in spite of, but in part because of the lack of any evidential support. 3.2 A Wittgensteinean Objection My response to worries about correctly interpreting Wittgenstein echoes the one above. My concern is not to arrive at a (or the) position Wittgenstein advocated in On Certainty (supposing he ever advocated positions at all, which many scholars believe he did not). My concern is to explain and defend the idea that (T), or a claim very like it, is a regulative assumption of inquiry. Whether or not this explanation and defence is consistent with Wittgenstein really said or meant is not my concern here. That said, there is at least one interpretive issue that Wittgensteineans will likely raise that does pose a problem for my argument. Put very plainly, the worry is that (T) doesn’t sound like a hinge proposition. More precisely, the thought is that (T) is fundamentally dissimilar to the examples Wittgenstein gives of hinge propositions, not with respect to the three key features identified above (about which, more later) but with respect to another feature I haven’t identified. This is: that for a claim to be a hinge proposition it should be obvious that it’s something we believe to be true. The objection follows swiftly and easily. (T) doesn’t fit the bill of a hinge 34
I’m inclined to agree with Hookway’s argument that ‘when Peirce attempted to work out the details of his philosophical picture for his logic text in the early 1870s, he was forced to confront some problems which he could not answer.’ Hookway (2002), p. 18. Unlike Hookway—as noted in footnote 17—I believe Peirce’s hope-based solution was not the right approach to solving the problem. 35
This is strikingly similar to Pritchard’s (2011) point about the structure of reasons.
123
462
A. W. Howat
proposition because it’s not at all obvious whether or not we believe (T)—the very fact that this claim requires elaboration and defence already proves that it cannot be a hinge proposition.36 My answer to this objection comes in two parts. First, I think we can believe things without it being initially obvious to us that we believe them. Second, I think some of these things are, like hinge propositions, beyond reasonable doubt, even if we must first reflect upon them in order to discover this. Therefore, it is entirely possible that there are hinge propositions such that (a) we don’t initially realise we believe them and (b) we don’t initially realise they cannot coherently be doubted. (This response, by the way, fits nicely with Peirce’s idea that ‘one cannot pronounce a belief indubitable until one has subjected it to very thorough examination’.)37 The simplest way to justify this answer to the objection is by observing that whether or not a proposition plays a certain role in the structure of reasons is, in some fairly robust sense, an objective matter. ‘Robust’, that is, in the sense in that it is something about which we can be mistaken, or initially uncertain.38 Let me make a more elaborate defence of these two claims however, lest this basic response be insufficiently persuasive. First, how is it possible that we believe some propositions without realising that we believe them? Consider belief in the truth of generalisations. Following, Lewis (1969) and Pettit (1998), let us say that S believes a generalisation in sensu composito if and only if S can or does readily assent to a particular sentence along the lines of ‘All Fs are Gs’. S believes the same generalisation in sensu diviso when S satisfies a weaker condition, viz. ‘‘S is universally disposed, on recognising something as F, to assent to the particular claim ‘That is a G’.’’ It is legitimate, I claim, to attribute a belief in some generalisation G in sensu diviso to a subject S, even if S does not know what to say or think about G (that is, even if S does not believe it in sensu composito). For example, asked whether he believes a generalisation expressing modus ponens (‘All inferences of the form… are valid.’), Lionel may not know what to say (Lionel, let’s assume, lacks training in formal logic). This is perfectly consistent with the thought that Lionel already believes the generalisation in the diviso sense. That is, assuming that whenever Lionel is presented with a material conditional along with an affirmation of the antecedent, Lionel infers the consequent, or takes that inference to be legitimate, it’s quite coherent and apt to say that he already believes the generalisation of modus ponens. If the above reasons are sufficient to show that there are some beliefs we may not initially realise we possess, could any of them fit the bill as hinge propositions? That is, could any of those beliefs be such that on reflection, we see them as beyond reasonable doubt? Think again about modus ponens. Once Lionel has the right technical vocabulary, and a firm grasp on the meaning of the generalisation, he comes to realise that yes, he does believe that sort of inference is valid. Now we ask 36
I’m grateful to Steve Burwood for pressing this point.
37
See Johanson (1994, p. 174), CP 5.451, and CP 5.509-10.
38
It is odd, but not contradictory I take it, to speak of being uncertain about whether or not one is certain. Consider its plainly well-formed opposite—I’m certain that I am uncertain about p.
123
Regulative Assumptions, Hinge Propositions
463
Lionel if he is certain—or if he can entertain the thought that this sort of reasoning isn’t valid. Can we thereby induce a genuine (even a coherent) doubt? I confess I do not see how one could bring about such a doubt. More specifically, I do not see how one could bring about a doubt without employing reasoning that in some way already depends upon the validity of modus ponens. 3.3 Foundationalism, Conservatism, Dogmatism Given Peirce’s powerfully anti-Cartesian streak, it is important to distinguish the appeal to common sense embodied by the ARA, here construed as a forerunner of the Wittgensteinean ‘hinge strategy’, from any sort of familiar, potentially regressgenerating epistemological foundationalism. Nor need it be vulnerable to charges of conservatism or dogmatism. Peircean scholars are keen to emphasise that Peirce’s ‘common-sensism’ is critical. Bernstein explains this point thus: [Peirce] is anti-foundationalist when foundationalism is understood as the doctrine that claims that there are basic or incorrigible truths that are not subject to revision. But he is not denying - indeed he is affirming - that all knowing has a foundation in the sense that there are tacitly held beliefs, which we don’t doubt and take to be the bedrock of truth. Peirce would certainly endorse Wilfrid Sellars’ famous remark: ‘For empirical knowledge, like its sophisticated extension science, is rational, not because it has a foundation, but because it is a self-correcting enterprise which can put any claim in jeopardy, though not all at once.’ (Sellars 1997, p. 79) Peirce would add that this is true for any inquiry - including logical, mathematical and philosophical inquiry.39 This raises two new questions. These questions put pressure on the idea that there is any coherent via media between the position that (T) is an incorrigible foundation, and the position that (T) is either false or at least dubitable. Pressure towards the foundationalist side comes from the worry that our belief in (T) cannot be both fallible and indubitable. How are we to make sense of this seemingly paradoxical combination of traits (§3.4)? Pressure towards accepting that (T) is false or at least dubitable comes from the question: if our belief in (T) is genuinely fallible, why doesn’t the problem of lost facts show that it is false, or at least dubitable? (If it does so, then again, it cannot be a hinge proposition). 3.4 Pressure Towards Foundationalism: Belief in (T) Cannot be Fallible In what sense does Peirce allow that our belief in (T) is fallible? The objection is that by construing (T) as a hinge proposition, by making it immune from doubt or rational criticism, one makes any claim to fallibility concerning it impossible or incoherent. One must be claiming hinge propositions are incorrigible, in which case, one is committed to old-fashioned foundationalism after all. There is something right about this objection—(T) is not fallible in the ordinary sense of the word. 39
Bernstein (2010, p. 34).
123
464
A. W. Howat
There is also something wrong about this objection—there is still a perfectly respectable notion of fallibility that does apply to (T). Our ordinary way of testing for fallibility, I take it, is simply by considering the possibility of error. We try to imagine a set of circumstances where a particular belief might prove to be mistaken. We can readily imagine a situation where we discover, for example, that the bus we thought departed at a quarter past the hour, actually departs at 10 past the hour—we simply imagine our having consulted the wrong timetable, having misread or misremembered the correct one, or having been misinformed by a stranger. This allows us to say with confidence that our belief that the bus departs at quarter past the hour is fallible, in the ordinary sense. (T) is not fallible in this ordinary sense. The problem of lost facts notwithstanding (we’ll talk about this in the next section), we cannot imagine a set of circumstances where we discover that inquiring long enough and well enough does not yield a stable belief in some hypothesis H. Why not? Simply because on Peirce’s account, one cannot know for sure when one has inquired long enough and well enough. Inquirers’s reaching this point depends upon whether or not they have the sort of stable belief that will survive all further inquiry and experience. Inquirers cannot know for sure, at any given moment, that their beliefs have this feature.40 So if (T) is not fallible in the ordinary sense, in what sense can it be fallible? Why isn’t Peirce’s claim just an old-fashioned foundationalist one as the objection alleges? As Hookway observes, this objection to Peirce’s fallibilism is similar to one often levelled at Quine: Thus readers of the final sections of Quine’s ‘Two Dogmas of Empiricism’ sometimes seem to take him to be committed to the idea that we can be aware of the real possibility that we may turn out to be wrong to accept, for example, elementary truths of arithmetic or logical laws. And it is then natural to respond that, for many such propositions, the possibility of error cannot be seriously entertained.41 Hookway’s response, his way of explaining the force of both Peirce and Quine’s point about fallibilism in a global sense (fallibilism applying to our beliefs in even elementary truths of arithmetic) is to appeal to Hilary Putnam’s characterisation of fallibilism, as the view that ‘[t]here is never a metaphysical guarantee to be had that such-and-such a belief will never need revision.’42 In order for a belief in (T) or in any other proposition to be infallible in this sense, it must belong ‘to an epistemic kind of which, of necessity, all members are true’.43 It will belong to this class in virtue of possessing a certain metaphysical feature—a feature that guarantees that beliefs or judgements about that proposition cannot be mistaken. On this picture, the fallibility of our belief in (T) ‘consists in the absence of a distinctive metaphysical explanation of its infallibility. And it is compatible 40 Note that this feature of Peirce’s account is also overlooked by some of his critics. See in particular Wright (1992). 41
Hookway (2007, p. 10).
42
Putnam (1995, pp. 151–181).
43
Hookway (2007, p. 11).
123
Regulative Assumptions, Hinge Propositions
465
with this that, as things stand, we cannot imagine or conceive how we might be mistaken.’44 That is, it is compatible with this sort of fallibilism in the metaphysical sense, that our belief in (T) is not fallible in the ordinary sense. 3.5 Pressure Towards Doubting (T) My thesis is that (T) is a hinge proposition, and that this move will make the Peircean claim about truth safe from the problem of lost facts. I anticipate two further, closely related objections to this proposal, which both put pressure on the idea that (T) is genuinely indubitable, a feature it must have to fit the hinge strategy. These objections are perhaps the most important of all. First: the whole point of the Problem of Lost Facts is that it shows that (T) is false—in which case (T) can’t be a hinge proposition (since such propositions are immune from rational criticism). Second: even if the Problem doesn’t show (T) to be false conclusively, it is at least perfectly adequate grounds to doubt (T), and so (T) still can’t be a hinge proposition (since such propositions are supposed to be indubitable, at least ceteris paribus). I think the first objection is the weakest, because the quest to prove that there are some true propositions that we cannot, in principle, know faces two major and rather similar obstacles. One is that it must avoid begging the question by presupposing something about the nature of truth (i.e. that truth can be beyond the scope, in principle, of even the best inquiries). The second is the danger of becoming trapped in a sort of pragmatic paradox, by presupposing the truth of (T) in formulating the objection. This looks especially hard to avoid if one accepts that in trying to prove that (T) is false, by formulating the problem of lost facts, one is transparently engaged in inquiry (that is, the attempt to resolve a doubt raised by Peirce’s claims about the capacity of inquiry). Inquiry is the activity that ex hypothesi presupposes or relies upon the truth of (T), the very thing one is trying to undermine. Another way to frame the same point: the hinge strategy suggests that certain propositions are only immune from support/criticism given (and that means when) they play a particular role in our reasoning process. Is (T) playing this role when we reason that the problem of lost facts shows that (T) is false? We need a reason to think not, and I suspect this will be difficult to come by. Now the critic might think that if presupposing (T) leads to a paradox, then so much the worse for (T). This is hardly a decisive move however, when one takes into account other paradoxes (that of rulefollowing, for example), which only rarely prompt calls for the relevant beliefs to be dropped altogether. The second objection is more difficult to overcome since it relies on a weaker claim—not that the problem of lost facts establishes the falsity of (T), but that it merely raises a genuine doubt as to its truth. If we can doubt it (ceteris paribus), it cannot be a hinge proposition. Notice that we cannot simply re-run the strategy we used in the case of the first worry. We cannot show that in the process of raising doubts about (T), the critic is bound to be engaged in a process of inquiry that already presupposes its truth. For the critic is not looking to replace doubt with 44
Op. Cit. p. 12.
123
466
A. W. Howat
belief, but rather to do the opposite—to unsettle the pragmatist’s belief and create a doubt. There is however, a weaker version of the same strategy that may help. There are two ways we might develop this strategy (these ways may be complementary rather than mutually exclusive). The first way is to say that the doubt raised by the problem of lost facts is not a genuine doubt (§3.5.1). The second is to appeal to the ceteris paribus clause in the hinge strategy (§3.5.2).
3.5.1 The Doubt Is not Genuine Peirce is famous for his attack upon Descartes’ methodological doubt, which he claimed was merely feigned: A person may, it is true, in the course of his studies, find reason to doubt what he began by believing; but in that case he doubts because he has a positive reason for it, and not on account of the Cartesian maxim. Let us not pretend to doubt in philosophy what we do not doubt in our hearts. (SC, p. 141) What does it take then, for a doubt to be ‘genuine’? Well, primarily, it must not (since it cannot) arise from a mere act of will. Peirce suggests that genuine doubt is rather a product of confrontation with some surprising event or experience. That event or experience will be external to us—it will not be subject to the will. A second key aspect is that genuine doubt is distinguished by the way it interferes with a belief. For Peirce, beliefs are habits of action, and thus, a genuine doubt is one that interferes with the habit of action embodied in the relevant belief. We might take this to be precisely what Wittgenstein means when he says that some things are in deed not doubted (OC §342). Returning then, to the problem of lost facts, there are two aspects to the resulting response. First, does the doubt about the capacity of inquiry to arrive at the truth arise from a confrontation with some external, surprising event or experience? No. The doubt comes about through entertaining a mere possibility. Does the doubt interfere with belief, which is to say, with our habits of action? No, for if our critic genuinely doubted the truth of (T) then he would not inquire. The critic thinks he doubts (T), but the moment the critic goes back to doing inquiry (which, I think it’s fair to say, he must), it becomes clear that his doubt was feigned rather than genuine. His actions will belie the claim to genuine doubt. Similarly, a man who puts his money on a horse to win stating that (a) he genuinely doubts (which is to say, has grounds for doubting) that it will win and (b) that he desires to back a winner is getting something wrong, is behaving irrationally. He might explicitly profess doubt, but then he needs to explain his action somehow (and our hypothesis, that he does not really doubt after all, would make for a pretty good rationale).45 The burdens on the critic then are these: first, how could the ‘doubt’ arising from the problem of lost facts come about through confrontation with some external, surprising event or experience? Second, how is he going to explain most of what 45
I’m heavily indebted to Joe Morrison for this example & response.
123
Regulative Assumptions, Hinge Propositions
467
he’s up to in his mental and rational life if he really, genuinely does doubt that suitably diligent inquiry leads one to the truth? 3.5.2 The Ceteris Paribus Clause In order for (T) to qualify as a hinge proposition, it is only necessary that it be indubitable ceteris paribus. Consider again p, the proposition that ‘I have two hands’. Someone who has just regained consciousness after a large explosion could, presumably, genuinely and coherently doubt whether or not p was true in their case in that situation. But these are hardly normal circumstances with respect to the relevant proposition, nor with respect to one’s overall epistemological position. Are the postulated circumstances of identifying lost facts normal circumstances with respect to the question of whether inquiry will yield a stable conclusion? It would help, in answering this question, to have a schematic way to distinguish ceteris paribus dubitability from indubitability. We can do this by appeal to our two previously established notions of fallibility (§3.4). If our belief in some proposition p is fallible in the ordinary sense (we can imagine circumstances, that is a surprising event or experience, in which the belief turns out to be mistaken), then it is dubitable ceteris paribus. If our belief in p is not fallible in the ordinary sense, as I maintain is the case with (T), then it is not indubitable ceteris paribus. As I have already argued, (T) is not fallible in the ordinary sense, only in the global, special sense that there is no metaphysical guarantee that our belief in it is true. It follows that our belief in (T) is not, therefore, ceteris paribus dubitable. Since it is not a circumstance involving a surprising event or experience, where we can clearly conceive of discovering our error in trusting inquiry to deliver us a stable belief, the problem of lost facts does not raise any genuine doubt about the truth of (T). I began with a Peircean claim about our concept TRUTH. This formulation is probably only part of the pragmatist story about that concept, nevertheless, it is a very important part. It tells us what the practical upshot is of some proposition’s being true. It tells us the consequences we commit ourselves to when we assert that something is true. It tells us what we can expect from a true hypothesis should we set out to test it with argument and experiment. It serves as a guide for inquirers, instructing them to seek a stable belief, one that is sufficiently well grounded to endure come what may, which is not apt to be unsettled by confrontation with other opinions or new and contrary evidence. (T) is also the part of the Peircean story that has, historically, generated the most resistance and dissent, and more than its fair share of outright caricature and mockery.46 I have tried to show that this resistance, while understandable, can be undercut with closer and more careful scrutiny, at least, if one takes the Wittgensteinean hinge strategy to reveal something true and significant about the structure of reasons. The advantage of appealing to this strategy, I have suggested, is that it gives the Peircean claim that (T) has some special epistemic status a principled, substantive and non-ad-hoc precedent and explanation. This explanation 46
See in particular, Russell (1946).
123
468
A. W. Howat
is not peculiar to (T), nor even peculiar to claims that, like (T), regulate our practices (which is to say, presupposing them is a necessary part of engaging in those practices). This explanation is a completely general one about how reasons as such work—how reasons ground knowledge claims, how reasons are involved in the raising and resolution of doubts, and so on. It says that there are hinges upon which our inquiries turn, which must remain fixed for us to engage in the enterprise, and that a belief in the capacity of inquiry to deliver what we seek from it—a stable belief—is one such hinge. These hinges are fixed—they ‘stand fast’ for us—not because they are self-evident or infallible foundations for our other beliefs, but because ‘it belongs to the logic of our scientific investigations that certain things are in deed not doubted’. This does not mean such claims cannot be questioned, that our belief in them is infallible (thought they may be so in an ordinary, everyday sense). This does not mean that apparent counter-examples cannot be raised to such propositions. But there is an important and principled difference, between a genuine doubt—one arising from confrontation with recalcitrant experience—and doubt we merely affect, as the product and preserve of a purely theoretical exercise. Acknowledgments I am grateful to Prof. Chris Hookway, Dr. Joe Morrison, and two anonymous referees for invaluable commentary and conversation about this paper. I am also grateful to audiences at the University of Hull and the University of Sheffield for lively and useful discussion of the ideas contained herein.
References Bennett, J. (1974). Kant’s dialectic. Cambridge: Cambridge University Press. Bernstein, R. (2010). The pragmatic turn. Malden, MA: Polity Press. Hookway, C. J. (1985). Peirce. New York, NY: Routledge. Hookway, C.J. (1998). Peirce, Charles Sanders. In E. Craig (Ed.), Routledge encyclopedia of philosophy (Vol. 7, pp. 269–282). London: Routledge. Hookway, C. (2002). Truth, rationality and pragmatism: Themes from Peirce. Oxford: Clarendon. Hookway, C. (2007). Fallibilism and the Aim of Inquiry (Inaugural Address). In Proceedings of the Aristotelian society, Supplementary volume LXXXI, pp. 1–22. Johanson, A. E. (1994). Peirce and Wittgenstein’s on certainty. In G. Debrock & M. Hulswit (Eds.), Living doubt: Essays concerning the epistemology of C. S. Peirce (pp. 171–186). Dordrecht: Kluwer. Lewis, D. (1969). Convention. London: Harvard University Press. Misak, C. (2004). Truth and the end of inquiry: A Peircean account of truth. Oxford: Clarendon. Misak, C. (forthcoming). Pragmatism & indispensability arguments, 2011 Presidential Address to the Charles S. Peirce Society, Transactions of the Charles S. Peirce Society. Nesher, D. (2001). Peircean epistemology of learning and the function of abduction as the logic of discovery. Transactions of the Charles Sanders Peirce Society, 37(1), 23–57. Peirce, C. S. (1992). The essential Peirce, vol 1 (EP vol./page no.)). Bloomington and Indianapolis: Indiana University Press. Pettit, P. (1998). Practical belief and philosophical theory. Australasian Journal of Philosophy, 76(1), 15–33. Pritchard, D. (2011). Wittgenstein on scepticism. In M. McGinn (Ed.), The Oxford handbook of Wittgenstein. Oxford: OUP. Putnam, H. (1995). Words and life. Cambridge, MA: Harvard University Press. Russell, B. (1946). History of western philosophy. London: Allen & Unwin. Smart, J. J. C. (1986). Realism v. Idealism. Philosophy, 61, 295–312. Wittgenstein, L. (1969). On certainty (OC). In G. E. M. Anscombe & G. H. von Wright (Eds.), On certainty (P. Denis & G. E. M. Anscombe, Trans.). Oxford: Blackwell. Wright, C. (1992). Truth and objectivity. London: Harvard University Press.
123