DOXASTIC CORRECTNESS Pascal Engel University of Geneva Dpt of philosophy Rue de Candolle , CH 1211 Geneva
[email protected]
Abstract Normative accounts of the correctness of belief have often been misconstrued. The norm of truth for belief is a constitutive norm which regulates our beliefs through ideals of reason. I try to show that this kind of account can meet some of the main objections which have been raised against normativism about belief: that epistemic reasons enjoy no exclusivity, that the norm of truth does not guide, and that normativism cannot account for suspension of judgment.
I.
Introduction
No one denies that beliefs can be correct or incorrect, and that truth is the central dimension along which to evaluate them. But apart from these platitudes, there is little agreement about the nature of the relation between belief and truth. Along with a number of others (Wedgwood 2002, Boghossian 2003, Shah 2003), I think that the best account of belief correctness is the normative one: a belief is correct because it is governed by a certain constitutive standard or norm, which is truth. Normativism about belief, however, must meet prima facie at least three challenges: it must show that the normative character of belief is a genuine feature which governs in some way our belief formation, it must tell us why a deontic account of the normative properties of belief is better than an evaluative account, and it must show how the norms for belief are related to knowledge and to our epistemic goals. I cannot take up all these challenges. My specific question here is: what form should the normative account take? I shall try to show that the norm of correctness is better conceived as an ideal of reason, and that this conception can meet some of the main criticisms raised against belief normativism: that the norm provides no guidance, that I cannot account for the relationship between evidence and truth, and that it cannot account for states of withholding of judgment.
2
II.
Correctness
A number of actions, such as tying one’s necktie, or playing Bach flute sonata in G minor are said to be correct or not. As Thomson (2009) reminds us, correctness of an item is relative to a kind: a correctness fixing kind is what fixes the standards that a K has to meet in order to be correct qua K. For example the kind map of England fixes the standards that have to be met for something to be a correct map of England, and the kind spelling of chiaroscuro sets the standard that a spelling of that word has to meet in order to be correct qua spelling of that word. “Correct” in this sense is an attributive adjective like “good”: that X is correct qua K does not entail that it is correct, period. But is it clear that correctness is a normative property? The standard for a tune is fixed by a set of notes, the standard for a map is fixed by the similarity between the map and the territory represented, the correct spelling is fixed by a certain pronunciation of the word. These are descriptive properties, not normative ones. They indicate what condition a certain kind has to satisfy in order to meet the given standard. The normative concept of correctness is distinct from this descriptive one. It concerns the way, or the operation which an agent has to perform in order to meet the descriptive condition. To take another example (Sosa 2011), an archer’s shot is correct –successful - when it hits the target. But the shot can be more or less competent, apt or adroit. When we evaluate it, we do not attend only to the successful hitting – which could be just lucky- but also to the way it is performed, which can be good or bad. A certain performance can be correct in the first sense – meet its condition of satisfaction - without being correct in the second – being a valuable performance, and vice versa. Thomson calls the first kind of correct “external” and second “internal”, and distinguishes e-correctness from i-correctness. Only the second seems to be normative per se. It would, be wrong, however, to conclude that the only normativity involved in correctness is the internal one, and that “correct” just means “good” or “valuable”. For a performance to be correct, it is not enough that it is i-correct, in the sense of done well or aptly. It has also to be the right kind of performance, and to meet the standard which is fixed by the descriptive condition. So for a kind K to be correct it has to meet both the e- and the i-correctness condition. It would be wrong to reduce correctness to either one of these two dimensions. How can we transpose this distinction to mental attitudes, such as belief, desires, or intentions? The notion of e-correctness seems to be close to the familiar notion of direction of fit of attitudes. Cognitive and conative attitudes have opposed directions of fit. They also have different e-correctness conditions. Alternatively we could use the schoolmen’s vocabulary: different attitudes have different formal objects1. Just as a future evil is the formal object of fear, or someone else’s good the formal object of envy , truth is the formal object of belief, but also of all doxastic attitudes such as guesses, judgments, conjectures, hypotheses, presumptions, convictions, doubts, suspicions etc. There are grounds to take the e--correctness condition of desire to be 1
Kenny 1963, Teroni 2007. This notion was rediscovered by John Searle under the name of the “satisfaction conditions” of mental states: the satisfaction condition of the belief that p is that p is true, which we can distinguish from the correctness condition: a belief is correct iff it is true.( Mulligan 2008)
3
goodness. Things are less clear for intentions or for imaginings: is there a specific correctness condition or formal object for intention or for imagination? There are three interesting asymmetries here. The first one is that although imaginings do not have obvious e-correctness conditions, they have i-correctness conditions: one can imagine more or less well, more or less aptly. Truth is not clearly the e-correctness condition of imagination. Sometimes imaginings can come out true, but in many cases they needn’t be to be good imaginings. The same holds for guesses, conjectures, judgments or suspicions: one can guess, conjecture, judge or suspect more or less well – hence more or less correctly in the internal sense- or with more or less good reasons, without necessarily hitting the truth. In contrast does it make sense to say that one can believe more or less well? Here it cannot be the belief which is good or apt, for belief, when it is thought as a state endowed with content, is not a performance. Could we say that the believing can be correct in the internal sense? Perhaps we could, but only in the sense in which the subject has more or less reasons for her belief. Believing for bad reasons, or on the basis of insufficient evidence is poor believing and thus i-incorrect, although it can be correct in the external sense (just as the product of a bad guessing can be a true lucky guess). The second asymmetry is between factive states like knowing, seeing, perceiving, realizing, and the like which are e-correct –true - by definition and non-factive states, i.e all doxastic attitudes akin to belief. The former cannot fail to be e-correct, whereas the latter can. This invites the thought that the normativity of certain attitudes has to do with the capacity of the agent to reach the aim or target at which it is directed (one inevitably meets here the metaphor of beliefs “aiming” at truth). And this can be done with more or less success. In this sense knowing, as a state, is not normative, whereas conjecturing or guessing can be: one can guess or conjecture well or badly, but one cannot no know more or less well, since by definition, if one knows, on has already reached the target. The third asymmetry is that e-correctness is not a matter of degree, whereas i-correctness. If a belief is ecorrect in the sense of true, it cannot be true to a certain degree. One can, however, have more or less, better or worse evidence for believing, just as one can guess or conjecture more or less well. 2 In what sense, then, is the correctness condition of belief is normative? If we attend to e-correctness only, the answer seems to be that it is not normative at all. The standard for belief, like any standard, is measured against a purely descriptive property- here the truth of the content of the belief. A deflationist about truth, who claims that truth is a purely logical device, can argue from this that the fact that a belief is correct if true is a mere truism, which carries no normative weight (see e.g. McGrath 2002). The deflationist accepts that we can move from “His belief that p is true” to “He believed correctly that p”, but he denies that “correctly” here means more than “is true”. But this is false. If “correctly” meant the same thing as “true”, then it could be applied to the content of other attitudes than belief to which “true” applies. When ones desires, hopes, or imagines that p one desires, hopes or imagines that p is true, but this is not amount to saying that one desires, hope or imagine correctly (Wedgwood 2006: 158). “Correct” means something distinct from “true” because it 2
Thus I disagree with Wedgewood(to appear) when he takes correctness to be matter of degree. Credences have degrees of i-correctness , but beliefs ,on my view, do not have degrees of e-correctness.
4
refers to the existence of a standard or norm, and implies that the attitude in question is conform to it, and that the content of the attitude is true in virtue of the way it is formed, or acquired, or entertained. Thus the correctness feature is the product of two factors, the external and the internal one. This, however, does not yet tell us in what the normativity of belief consists in. Does it consist, as Sosa proposes, in the aptness and adroitness with which believings are performed and in their being successful? Or does it consist, as a number of functionalist accounts of normativity (Papineau 1999, Millikan 1993) in the good performance of a certain function? If this is so, the normativity of belief is nothing intrinsic: it is but a descriptive feature of the teleological structure of a cognitive device (Dretske 2000). A genuinely normative account takes a different form. In my view, it rests on three claims. (i) Correctness is a property of our concept of belief. The concept of belief involves the standard or norm that a belief is correct if and only if it is true (Boghossian 2003, Shah 2003). It is a constitutive and conceptual norm, which a subject needs to possess to be able use the concept of belief. (ii) Correctness is determined by the way our attitudes fit a certain feature. As an early proponent of the normativism says: “Whatever else one does with a truth, believing the proposition which expresses it is the first and most fitting thing to do with it "3 To be correct is to be the fitting object of the attitude of belief. This is consonant with the “fitting attitude” or “buck passing” account of norms and values in ethics, according to which the value of something depends on the rightness of the attitude that one has towards it. Now “buck passers” about value claim that the primary normative concept is the concept of reason. So we should expect a normative account of correctness to rest upon the reasons that we have for our beliefs. Does that entail that correctness is, so to say, in the eyes of the beholder, and not an objective feature? The account that I intend to propose is neutral about whether one should be an expressivist or a cognitivist about epistemic norms. But it must reject at least the irrealist or nihilist view according to which there are no reasons at all for our beliefs. 4 (iii) A norm for belief is a criterion by which one evaluates a belief as correct or right. Clearly many forms of believing do not conform to the norm of truth, which neither explains all the features of our belief formation nor guides it. This invites he thought that the norm is useless or shallow. But this is to misconstrue the role of constitutive norms. Their role is rather than of an idealization: ideally one ought to believe what is true, although we do not necessarily live up to that ideal (Kornblith 2001). The truthnorm for belief tells us what right belief requires, but it does not prescribe that we ought to believe what is true and only what is true. Epistemic ideals, like moral ideals, do not necessarily tell us how to think or act. If an account of correctness for belief takes this form, I want to argue, it can answer three main challenges which have been raised against the normative account. 3
A.Phillips-Griffith 1963: 182. “The connexion between belief and truth is that belief is appropriate to truth; it is proper only when it is of what is true, and only intelligible, therefore, when it is of what could be true” (ibid). See also Mayo 1963. This kind of account can be traced back to neo-Kantian philosophers and to Husserl and his disciples (Mulligan 2009, 2012). It was transmitted, within the English speaking tradition, through the work of J.N.Findlay (1968). 4 Scanlon 1998, Skorupski 2011, Shah 2010 gives an argument against the irrealist view.
5
III.
The right kind of reason
The normativist of the correctness of belief presupposes that the only kind of reason that there can be for believing something are epistemic reasons, namely that a belief is true and well supported by evidence. But why suppose that these are the right kind of reason and that they enjoy this “exclusivity” in being the only kind of reason that one can have to believe anything? Why is it incorrect to believe what is not true or not justified, if one’s doing so is beneficial? Might not we believe for non-evidential reasons, such as pragmatic ones? After all there might be various ways for a belief to be correct, depending on the kind of interest and context. A dandy can hold a given belief, which is otherwise absurd, because it is elegant, or Camp, or Kitsch. I can desire to believe that I am healthy, because it will bring me comfort, even though I have excellent reasons to doubt that I am in good health. What is wrong or incorrect, in such beliefs? The problem is the analogue of the “wrong kind of reason” problem for the fitting attitude analysis of value: if an evil demon threatens to torture me if I fail to value a cup of mud, I may have a good reason to value the cup, although it is not a reason of the right kind (Rabinowicz-Rasmussen 2004). There are many ways to trace the distinction. The kind of reason that we have for having an attitude is “state given” or attitude-related while the kind of reason to which the attitude responds is object or content given (Parfit 2011,I, 420-432 ). The partisan of the exclusivity might try to answer by denying that attitude-related reasons really are reasons: to actually have a reason to is one thing, to have a reason to desire to - and to bring it about that one s - is another. But this seems to beg the question against the view that it could be equally correct to believe for epistemic reasons and to believe for prudential reasons. Alternatively, the exclusivist about reasons for belief can point out that there are criteria or distinctive marks which differentiate epistemic from pragmatic reasons. One is that epistemic reasons are always more immediate and available, hence more easy to answer, than those which one has because of some pragmatic aim or goal. A second mark is that epistemic reasons are always “transparent” in contexts of epistemic deliberation: the question whether to believe that p seems to be normally answered through considering whether p , whereas pragmatic reasons for belief do not lend themselves to this transparency test ( Shah 2003, Hieronymi 2005). A third mark is that if there were various ways for a belief to be correct, it would be possible in some sense to weigh different criteria of correctness. But typically we do not say that pragmatic or prudential reasons for belief are correct, and epistemic reasons are always those that we recognize as the correct ones.5 Yet another mark is that there seems to be a central core of rationality which always resists against any claim of rationality on the part of pragmatic considerations. There are various ways of spelling out these differences, but they seem either to be question begging or to fall victim of counterexamples purporting to show that the criteria which are supposed to apply to object or content given reasons only also apply to state given ones. Schroeder (2012) argues that the distinction between the right kind and the wrong kind of reason of cannot be explained through the state given/ objet given distinction, for there are cases 5
This is a transposition, for correctness, of the claim, defended by Owens (2003) that truth cannot be an aim or a goal of belief, for if it were it would be possible to weigh this goal against others.
6
where, both for intentions and for beliefs, there can be reasons for a state or attitude which bear all the marks of the right reasons. Most of his examples involve cases where who has to defer a decision for which one awaits more information, or cases of withholding one’s judgment that p when the evidence is lacking about whether p or not-p. For instance when the doctor makes a diagnosis on my illness, but tells me that he waits for more evidence from the lab to deliver his final opinion, it is reasonable for me to suspend judgment. In such cases, it is correct to adopt he attitude of lacking belief, and for the right (evidential) reasons, although the reasons are clearly about a state or attitude (withholding) rather than an object or a content. I come back to withholding below, but it seems to me that there is a simple answer to Schroeder’s argument (which he himself gives): when I suspend judgment, I have reasons to adopt this attitude or state, hence my reasons are state given, but it does not follow that I cease to have content given epistemic reasons for my suspending judgment. On the contrary I withhold just because I lack evidence about whether p or not p. Schroeder succeeds in showing that some right reasons for belief can be attitude or state given, but he does not show that the right reasons for belief in these cases can be pragmatic. In order to show that the property of correctness of a belief can be transferred to merely pragmatic reasons, one would have to consider the familiar cases adduced in the literature on doxastic voluntarism and on pragmatism – e.g of people who have good reasons to believe that they are not in danger although they know that they are to the effect that it can be rational to believe for pragmatic reasons or to cause oneself to believe something for such reasons. I cannot deal with these case here, but none of them how that although it can be good, or beneficial, or instrumentally rational to believe that one is not in danger, it is right or correct to believe that when one is in danger and aware of this fact. Another line of argument against the right/wrong reason distinction consists in arguing that, contrary to the third mark above, it can be possible to weigh the aim of belief, or, to adapt the argument to the present normativist account, the various criteria of correctness for belief. Steglich-Petersen (2009) and Reisner (2009) argue that there can be cases, where the correctness norm (or aim, the difference here does not matter) for belief , i.e truth and evidence can interact with other norms for belief, of a prudential or pragmatic kind. Their examples are similar to Schroeder’s: cases where one has to defer one’s judgment when more incoming information is expected ( as the doctor case above), or cases where one prefers not to seek more evidence because of the bad consequences that the having of a belief could bring (for instance when one prefers not to investigate further into the foreseen evidence that one’s brother is guilty of a horrible crime). But it is dubious that these cases illustrate situations where one would have to balance between an epistemic norm or aim and a pragmatic norm of aim for believing, for in none of them the truth or the evidential norm ceases to govern the beliefs: it is just that one defers, for pragmatic reasons, the formation of a belief which, if it were acquired, would be governed by the truth norm or the evidential norm. The epistemic norms of truth and of evidence do not prescribe that one ought to form a belief anyway, however poor or overwhelming one’s evidence can be, but that if one decides to go ahead and form a belief, it will have to conform to these epistemic norms (Bisset and Noordhof 2012). On can adapt here Sosa ‘s image (2011: 32): the pilot of a
7
boat on a river can, at various moments, decide to deviate from the trajectory that the current imposes on his boat, but he is not free, once he takes such decisions, to change the direction of the current. The epistemic correctness condition of belief is always the default mode of our belief formation.
IV. Guidance by ideal There are two candidates for the correctness condition of belief. One is the external, or as it is sometimes called, objective, norm of truth. The other is the internal norm, which associates correct belief to our evidential reasons for believing, which is sometimes called the subjective norm. They seem to be closely related, for it seems that one cannot follow the first without following the second: the best way to know the truth is to rely on one’s evidence, and evidence is evidence for truth. But that is easier to say than to argue. When Boghossian (2003:38-39) characterizes the truth-norm for belief as constitutive in this way: “The truth is what you ought to believe, whether or not you know how to go about it, and whether or not you know if you have attained it” he seems to separate off the first norm from the second. For how can one follow the norm that one ought to believe the truth, if one has not idea of how one could know it? It is easy to imagine cases where they conflict, for instance when a doctor has good evidence to prescribe to a patient a drug, which, unbeknownst to him, will actually kill the patient because of some condition that the doctor can’t have access to (Gibbard 2005). In such cases, the objective truth norm seems to prescribe you one thing, whereas the evidential subjective norm prescribes you to believe the opposite. If we suppose that the capacity of a believer to be guided in his beliefs by a norm implies that he has evidential reasons, the truth-norm seems be simply blind (Hattiangadi 2010). This point is reinforced by what is perhaps the most forceful argument against normativism, which can be presented as a dilemma (Glüer and Wikforss 2009). On the one hand, if the truth-norm is constitutive, in the sense that it spells out a conceptual necessity for belief, it has no normative force, for conceptual necessities have no power to explain our belief formation or to guide us in our beliefs. They lack another ordinary feature of norms, “normative freedom” (Railton 2003: 322-351): a norm which is such that under any description you are supposed to be governed by it and have no choice to follow another norm does not seem to be norm at all. On the other hand, if the truth norm is taken at face value and formulated in deontic terms: (NT) One ought to believe that p if an only if p it seems to fail whatever way it is interpreted and to give us perfectly implausible, useless or trivial prescriptions. Implausible: if read from right to left (NT), if p you ought to believe that p, it does not make sense to ask someone to believes any truth. Useless: if read from left to right, you ought to believe p only if p, then it gives a merely negative advice about when the norm does not hold, and no positive one, about when it holds. Trivial, because (NT) is “useless as a directly guiding principle, for it
8
can gain purchase one one’s conduct only through a belief that p is true, but if one believes that p is true, then one believes that p already” (Sosa 2001:54, Glüer and Wikforss 2009:44). One can weaken the norm (NT) by reading it as a permission rather than an obligation (Whiting 2010), but one would encounter similar difficulties. I shall not detail here the other objections to the effect that on any reading (NT) cannot give what one normally expects from a norm (Bykvist and Hattiangadi 2007, Mc Hugh 2012). The answer to these various versions of the ”no-guidance” argument which a normativist has to give is simply to bite the bullet and take the first horn of the dilemma: the truth norm for belief is indeed a conceptual norm, which does not give us any prescriptive – or even permissive –guidance. The norm indeed just tells us that we ought to believe the truth, “whether or not we know how to go about it”. A constitutive norm is neither a prescription in the sense of an imperative of the form “If condition C holds, do A”, nor a directive in the sense of an ought-to-do (Thomson 2008:207-230). It is an ideal of reason, in the sense that it tells you what you ought to ideally believe, namely the truth, and thus is belongs to the category of the ought-to-be rather than to the category of the ought to do (Kornblith 2001: 238, Millar 2005:76, Chrisman 2008:20). By definition, an ideal is meant to describe an abstract situation which hold only “in principle” or a kind of conduct which only certain imaginary beings endowed with powers which are distinct from ours could follow (logical saints, believing all the consequences of their beliefs, perfectly rational agents). The status of the truth-norm for belief , as well as that of the evidential norm is of this sort: it tells us what believing requires, but neither what kind of beliefs one must have before applying the norm, nor what kind of belief on must have once one has applied it. It is blind to the actual psychology of the agents. In this sense, it need not explain or guide our belief formation. The ordinary notion of an ideal however, is not completely conform to this model, for it can mean, as C.S Peirce (1938: II, 211-213) called it, a “rule of conduct”, or an ultimate goal. This is not the normativist’s idea, since for him the truth norm is not an aim or a goal in the teleological sense. Moreover, ideals are also supposed to be in some sense responsive to human capacities, and in this sense to be able to guide actions and beliefs. When they are conceived as optimal conditions, and when these conditions to not apply to a given situation, one is supposed to be able to revise the optimum and to sub-optimise. How is this compatible with the view that the truth norm is a criterion of rightness which is by definition rigid? Must not ideal oughts imply cans? My interpretation of the truth norm as an idealisation can be formulated in terms of a distinction which Wedgwood (to appear) has adapted to these discussions from the epistemological context, between “ex ante” or “propositional” justification and “ex post” or “doxastic justification”. On the ex ante reading of the correctness condition A belief is correct if and only if it is true, or One ought to believe what is true only what is true, it is not entailed that you have the belief in question, nor that you know whether is available to you in any psychological sense. On the ex post reading, the advice only makes sense retrospectively, in the light of the beliefs that you actually have and which are available to you. According to Wedgwood, the appropriate reading of the truth norm should always be the ex post, or retrospective reading, because, according to him correctness for belief has to be indexed to the degree of confidence
9
that the agent has – presumably proportional to the evidence at his disposal- towards the belief-content. Wedgwood adopts this reading because he takes correctness to be, like belief, a matter of degree, lying between perfect or maximal correctness, which is truth, and minimal or absolute incorrectness, which is falsity. In between are, according to him, degrees of correctness. On the contrary, the idealisation account which I favour takes the truth-norm to have systematically the ex ante or prospective reading. To take the retrospective or doxastic reading of the correctness condition supposes that, as Wedgwood makes clear, that full or outright belief is a limit case of credences or degrees of belief (Wedgwood 2012), and that correctness for belief is identified with degrees of confidence located on a scale between true full belief as1 and false full belief as 0. I have three objections to this proposal. First it has the implausible consequence that there are degrees of correctness. But correctness, as we use ordinarily this notion is not a matter of degree. “Correct” as we saw, means more than simply “true”, but truth, pace the theorists who want to substitute to it the notion of “verisimilitude”, is not a graded notion. When we say that a belief is correct we do not say that it is true to the maximal degree, and when we say that a belief is incorrect we mean that it is false, not that it is correct to a degree (indeed politicians who want to make disclaimers sometimes say that their previous declarations were “incorrect” in order to minimize their alleged falsity and to suggest that they were not as wrong as their critics say they are, but the very idea of a disclaimer is to reject an accusation as false). There is actually a notion of degree of correctness, which is accuracy, and which is perfectly adapted to the notion of degree of confidence (Joyce 2009). Among doxastic states, guesses, suppositions, or conjectures can be more or less accurate (a guess as to who the winner is can miss by a mile, or a conjecture can be more or less correct), but it sound odd to say that a belief is accurate. Second, to assimilate correctness to a graded notion is to conflate external or objective correctness with internal or subjective correctness, and to implicitly make it a wholly epistemological notion. But the evidential norm, although it is indeed an epistemic norm constitutive of belief, cannot be the primary and the more general norm, for even if a belief is based on sufficient an perhaps overwhelming evidence, it is still not correct. It is unclear, to say the least, how much evidence is needed to form a correct belief, but the level of evidence cannot be the only criterion of correctness. Third, to take correctness as a graded notion misrepresents the role of belief in action. There are strong arguments, which I shall not try to rehearse here, for claiming that the role of belief in action implies that full beliefs only, and not credences, have an impact on our actions and play a role in our practical reasoning. To take full, coarse grained, belief as an idealisation of the fine grained notion of degree of belief, is to misconstrue the notion of an ideal here. The correctness norm of truth for belief is indeed an ideal, but it is not an ideal which is abstracted away from the notion of correctness of credences. Bayesians often say that their notion of degree of confidence is the real one, and that coarse grained full belief is but an idealisation, but we can see things the other way round, and say that the degrees of belief are idealisations from the full belief notion (Holton, to appear). 6 6
There is perhaps less disagreement between the view proposed here and Wedgwood , for he claims actually that there are two kinds of credences, those which fit our theoretical purposes, and those which fit our
10
The division that is proposed here between the ex ante or external notion of correctness as truth, which is essentially non epistemic, and the ex post or internal notion, will appear unsatisfactory because it does not seem to solve our problem, for it does not tell us how the two are related, nor how the norm can make a difference. I agree with the critics of normativism that the norm of truth has to make a difference. Ideal ought cannot be completely alien to can. But I deny that it is the kind of difference that they expect. Two main options are open here for the normativism. The first consists in rejecting the idea that the correctness norm is truth, and to defend the view that it has to be knowledge. This has the advantage of reconciling the external and the internal face of the correctness conditions for belief, and to explain why our most plausible reasons, so to say, to form belief, are that we have sufficient evidence for them. The proposal has the advantage of integrating the two aspects of correctness (Smythies 2010). It is an option which I consider to be open, and which I have accepted elsewhere. Nevertheless taking the correctness of belief to consist in believing what one knows still does not explain how the norm can make a difference. Unless the kind of knowledge needed is specified, the norm seems to be as idealised as the truth norm (and actually I agree with Wedgewood that they are two faces of the same coin : a belief one cannot be ideally correct in the sense of being truth without being ideally known (2002: 289). But that does not prevent us, as a second option, from giving a plausible story of how the norm can in some sense influence our actual believing. Just as a law, in some legal systems (such as the French one), needs, in order to be applied to actual cases, be supplemented with “decrees of application” which specify the circumstances, we can say that the ideal norm can be supplemented with psychological decrees of application. I can see three ways in which it can be done. The first one is the transparency feature of belief, emphasised by Shah (2003) and others: when one asks oneself whether to believe that p, the immediate answer comes with believing whether p. In order for the norm to be in place, one does not need to consider antecedently, as the objection from triviality above suggests, the belief in order to apply the norm. But as soon as a belief is considered, and in so far as one conceives of it as a belief, it is implicit that it has to be conform to the truth norm. The test has been contested as being generalizable to all kinds of believing, but it fits at least the specific contexts of belief deliberation or of inquiry, which are contexts were we explicitly set to ourselves the aim of truth. The second comes through the experience that I have described above, of being faced with two possible ways to go about a given belief, which seem to appear equally correct, or when a rival aim - comfort, prudence, or aesthetic objective – presents itself. As I suggested above, in such situations the truth norm may not prevail, but it is always present by default, including when it is overcome7. The third way come with the fact that someone who has the concept of belief is committed to the ideals of reason which go with it. The commitment to the truth norm for belief does not entail that one has a reason to believe, or a justification. But it entails that I should be prepared to give one. This is where the evidence practical purposes. On Wedgwood’s view, the second requires outright or full belief, and the first degree of belief. But I do not see why the notion of full belief can serve both purposes. 7 Steglich-Petersen 2010 suggests that the truth norm, in such cases, guides in a teleological way. But we do not need to suppose that truth is a goal here, only that it is the default mode norm.
11
requirement comes in: someone one incurs the truth commitment has to recognize that the requirement obliges him to give a justification. That does not mean that he has to possess actually the justification (in the doxastic mode), but only that he is bound to have one (in the ex ante or propositional mode). Arguably most of these decrees of application are negative, an epistemic ideals do not have the power to motivate us positively, as moral ideals. But it would be wrong to say that there cannot be also, in the epistemic case, motivation by ideal.
V Suspension of judgment
The last objection to normativism which I would like to consider, although briefly, is the one which Mayo (1963: 144) presented a long time ago to the early version of normativism given by Philipps-Griffith (1963): on the view that belief is the fitting attitude than one must have towards a truth. The truth-norm seems to entail that full belief is the only correct attitude towards a true proposition, and full disbelief the only correct attitude towards a false proposition. But this entails that someone who withholds or suspends judgment about p not have the correct attitude to P, since the norm prescribes only two correct attitudes: belief that p when p is true disbelief that p when P is false. Here the point is that withholding or suspension of judgment is not only the lack of an attitude or the failure to have an attitude. It is a genuine attitude, alongside with belief and disbelief (Friedman to appear). This also seems to be an objection to the prescriptive reading of the norm (McHugh 2012). I gave above a reply to that objection in the context of the discussion of state-or attitude related as opposed to object or content-related reasons. The agent who suspends judgment is not actually having only an attitude-related reason in the sense of a reason which pertains only to the kind of attitude that he has. He has also an object-related reason, for the suspension of judgment bears on the object or content of judgment. In other words it is a right reason to neither assent nor dissent, and the stay in that state, because of the very content of one’s belief or judgment. It is false that the norm of truth allows only two doxastic attitudes. If one considers whether p is true, and does not have enough evidence for either or not-p, the norm does not prescribe to believe p or to believe not-p It prescribes witholding belief. But isn’t’ withholding then under the governance of the evidential norm or the norm of i-correctness r of justification? It is, but it is also under the governance of the truth-norm, for there is no possibility to be governed by the truth-norm unless one follows the evidential norm. Something like What Smythies (2012: 19) calls the “linking principle” - one has justification to ϕ if and only if one has justification to believe that it is correct to ϕ. Now “correctness here cannot mean “having evidence” or “having justification”, for otherwise the principle would be tautological. Correctness means what it ought to mean in the first place in the case of belief: truth, in the straightforwardly non epistemic sense.
12
REFERENCES Boghossian, P. 2003 ‘The Normativity of Content’ Philosophical Issues 13, 31–45 Bykvist, K. and Hattiangadi, A. 2007 ‘Does Thought Imply Ought?’ Analysis 67, 277–85. Glüer, K. and Wikforss, Å. 2009 ‘Against Content Normativity’ Mind 118, 31–70 Chrisman, M. 2008 “Ought to Believe”, The Journal of Philosophy, vol. CV, 7, 2, 346-370 Dretske, F. 2000, ‘Norms, history and the mental’, in Perception, Knowledge and Belief Cambridge, Cambridge University Press .Engel P. to appear , “In Defense of Normativism about the aim of Belief”, in Chan, T. The Aim of Belief, Oxford: Oxford University Press Fassio, D. 2011 “Belief, Correctness and Normativity”, Logique et analyse, 54, 216 Findlay, J.N.1968 Values and Intentions, London, Allen and Unwin Friedman, J. to appear “Suspensed Judgment”, Philosophical Studies Gibbard, A. 2005 Truth and correct belief. In E. Sosa, & E. Villanueva (Eds.), Normativity: Philosophical issues, 15, 338–350 Glüer, K. and Wikforss, Å. 2009 ‘Against Content Normativity’ Mind 118: 31–70 Hattiangadi , A. 2010 “The Love of Truth” Studies in History and Philosophy of Science 41: 422–432 Hieronymi, P. (2005). The wrong kind of reason. The Journal of Philosophy, 102, 437–457. Holton,R. to appear 'Intention as a Model for Belief' in, Vargas and Gideon Yaffe (eds.) Rational and Social Agency: Essays on the Philosophy of Michael Bratman (Oxford: Oxford University Press) Jarvis, B. 2012 “Norms of Intentionality: Norms that don’t guide”, Philosophical Studies157:1–25 Joyce, J.M. 2009 “Accuracy and Coherence: Prospects for an Alethic Epistemology of Partial Belief”, in F. Huber and C. Schmidt-Petri, eds, Degrees of Belief ; Berlin:Springer,263-297 Kornblith , H. “Epistemic Obligation and the Possibility of Internalism”, in Fairweather, A. and Zagzebski, L. Virtue Epistemology, Oxford: Oxford University Press Mayo, B.1963 “Belief and constraint” Proceedings of the Aristotelian Society, 64, 139-156 McHugh 2012 “The Truth Norm or Belief” Pacific Philosophical Quarterly 93: 8–30 Millar, A. 2004 Understanding People, Oxford: Oxford University Press Mulligan,K.2007 “Intentionality, Knowledge and Formal Objects”, Disputatio,II, 23: 205-228 2012 “Acceptance, Acknowledgment, Affirmation, Agreement, Assertion, Belief, Certainty, Conviction, Denial, Judgment, Refusal & Rejection”, ed. Textor, M. Judgement and Truth in Early Analytic Philosophy and Phenomenology London: Palgrave/Macmillan. Papineau, D. 1999 “Normativity and judgment “Proceedings of the Aristotelian Society 73:16- 43 Parfit, D. 2011 On What Matter , vol. 1, Oxford, Oxford University Press Peirce, C.S. 1938 Collected Papers, vol I, ed. Burk, Harvard, Harvard University Press Phillips-Griffith, A.P.1963 “On Belief”, Proceedings of the Aristotelian Society,63: 167-186 Rabinowicz, W., & Rønnow-Rasmussen, T. 2004. “The strike of the demon: On fitting proattitudes and value”. Ethics, 114, 391–423. Railton, P. 2003 Facts, Values and Norms, Cambridge, Cambridge University Press Reisner, A. 2009 “The Possibility of pragmatic Reasons for Belief, Philosophical Studies 145:257–272
13
Shah, N 2003 “How Truth Guides Belief“, Philosophical Review, 112, 447–482 Schroeder, M. “The Ubiquity of State-Given Reasons”, Ethics, Vol. 122, 3. 457-488 Skorupski, J. 2011 The Domain of Reasons, Oxford: Oxford University Press Smythies, D. 2010 “The Normative Role of Knowledge”, Nous 46.2: 265-88 Sosa, E. 2011 Knowing full well ,Princeton, Princeton University Press Steglish-Petersen, A. 2009 ‘Weighing the Aim of Belief’,Philosophical Studies, 145:395-405. 2010 “The Truth Norm and Guidance: a Reply to Glüer and Wikforss”, Mind, 119,475,749-755 Sullivan-Bissett , E. and Noordhof, P. to appear, “A defence of Owens’ exclusivity objection to Beliefs having aims, Philosophical Studies Thomson J.J. 2008 Normativity, Open court, La Salle, Ill Wedgwood, R. 2002 ‘The Aim of Belief’ Philosophical Perspectives 16, 267-97 2007 The Nature of Normativity, Oxford, Oxford University Press 2012 “Outright Belief”, dialectica 66 (3):309–329 To appear “The right thing to believe, in Chan, T. The Aim of Belief, Oxford: Oxford University Press Whiting, D. 2010 “Should I Believe the Truth?”Dialectica 61, 213-24
