Believing, holding true, and accepting
Philosophical Explorations, I, 2. 1998, 140-151
1. Suppose you leaf through the pages of a book on Taoism, written by a renowned expert, and that you do not know nothing about the Tao, or Chinese philosophy, or even the Chinese language, and you read this:
"Zhuangzi's principle is that you cannot get a shi without having a cheng "
You do not know who this Zhuangzi is, not even if it is the name of a man; you don’t know either what a shi or a cheng is. But on the basis of the author's authority, you believe that the sentence “Zhuangzi's principle is that you cannot get a shi without having a cheng” is true. You do not understand this sentence, for you do not know what it means, therefore you don't believe that Zhuangzi's principle is that you cannot get a shi without having a cheng. Nevertheless you believe the sentence to be true. But how can you believe that a sentence is true without believing what it says? That seems paradoxical, because it leads to a version of Moore's paradox. You have:
"Zhuangzi's principle is that you cannot get a shi without having a cheng" is true, but I do not believe that Zhuangzi's principle is that you cannot get a shi without having a cheng
but by the famous Tarskian equivalence you have:
"Zhuangzi's principle is that you cannot get a shi without having a cheng" is true iff Zhuangzi's principle is that you cannot get a shi without having a cheng.
And, substituting, you get a Moore's paradox :
Zhuangzi's principle is that you cannot get a shi without having a cheng, but I do not believe that Zhuangzi's principle is that you cannot get a shi without having a cheng
But on reflection, there is no Moorean paradox here. For there are two distinct propositional attitudes. One is the familiar attitude of believing that p, in the sense of believing the content or the proposition expressed by the sentence 'p'. The other one is the attitude of holding true the sentence 'p'. And it certainly is possible to have the latter without having the former. This happens quite often. We could never learn any new language, if we could not hold true certain sentences which are presented to us, before learning their meanings. For instance, my method for learning Mandarin asks me to hold true ‘Wo hui shuo Fawen ’, although I will discover only later what it means. There are many things that we believe on the basis of some authoritative statements, but that we do not understand , or at least do not understand fully. I am not saying that it should happen all the time, and that we should only believe what famous authorities say, although that's obviously what authorities -- such as the Pope, or Nobel laureates, or the Dalai Lama, or the Ministry of Information -- want us to believe. But it quite often happens, especially when learning rules. As Wittgenstein says, when I obey a rule, I follow the rule blindly. In this respect, holding true is like following a rule. Some writers, like Cardinal Newman, even say that this happens with religious belief (Newman wants to allow that we can believe a lot of things which we do not understand, such as the dogma of the Holy Trinity).
So the conditional
(a) if X holds true 'p' then X believes that p
is not always true. But its converse seems to be true:
(b) if X believes that p, then X holds true 'p'
at least according to the view that believing that p involves assenting to a sentence meaning that p. From the viewpoint of (b) assenting can be either an inner mental act, accomplished in foro interno, or a public speech act of asserting a sentence in a public language. So (b) can be spelled out in two ways:
(b') if X believes that p, then X assents mentally to 'p'
(b'') if X believes that p, then X asserts 'p'.
Certainly if I believe that you can't have a shi without having a cheng, then I will normally assent to the sentence "you can't have a shi without having a cheng" and normally assert this sentence, provided, of course, that I understand what the sentence means. But this proviso is precisely the reason why a number of philosophers object to the theory that believing that p implies assenting to an actual or to a potential sentence which means that p. For, they say, you may not know what 'p' means, and you could believe that p without ever assenting to any sentence 'p ', actual or potential. This is precisely the objection that has been put forward against all so-called quotational theories of propositional attitudes reports, according to which to believe that p is to be in relation to a certain sentence. What these writers object to is the so-called disquotational principle:
(c) if X sincerely (honestly) assents to 'p' then x believes that p
(see e.g. Marcus 1986). So, according to these authors, believing that p cannot be a matter of assenting to a sentence. Some authors take believing to be an assent to entities other than sentences. For instance, some say:
(d) if X believes that p, then X assents to some mental representation meaning that p.
(e) if X believes that p, then X assents to a token sentence 'p' in X's language of thought.
But some writers do not like the idea of assenting to any item at all. For these writers believing must involve a relationship to some other entity than a sentence. Some say that:
(f) if X believes that p, then X is related to a proposition that p
and others say that:
(g) if X believes that p, then X is related to the fact that p.
But there are also authors who do not think that believing implies any assent, or any special relationship to an entity. Some hold the dispositional view of belief:
(h) if X believes that p, then X is disposed to act in such a way that p comes out true.
However, (h) can, according to one version, come back to the view that believing is assent to a sentence 'p', indeed by asserting that p in some public language. But according to other versions, there is no conscious assent to any sentence or any mental representation.
This is, in particular, Ramsey's version of the so-called dispositional view of belief. For Ramsey, it is not necessary, for believing that p, that one has any conscious mental episode of assent to any mental or linguistic item. According to him, we have many beliefs during our life of which we are never conscious, or which never ‘come to our mind’. This does not mean that these beliefs are unconscious in some deep Freudian sense, but only that they may never be activated or become explicitly entertained by the individuals who have them. Nor does this mean that such beliefs cannot be mental episodes, only that they may not be. For instance, I may now believe that it is raining, and think or entertain the thought that it is raining, and be consciously attending to my belief. But I also may stop thinking that it is raining without stopping believing that it is raining. Or I may have always believed that Hong Kong is smaller than Mexico City, but the belief will probably have never occurred to me. Spelling out fully the dispositional theory of belief is a delicate matter. Some philosophers, like Peirce, say that beliefs in this sense are habits, just like the habit I have of beginning with the left leg when I put on my trousers. Others identify them with dispositions to act which are not necessarily habits. Others talk of ‘tacit’ or ‘implicit’ beliefs, opposing them to explicit beliefs which prompt certain acts of assent. Are beliefs in this sense dispositions, states, or what? I shall not answer these difficult questions here. I shall simply assume the following, more or less obvious things about beliefs.
(1) Beliefs are involuntary, and are not normally subject to direct voluntary control. I cannot believe at will that my trousers are on fire or that the Dalai Lama is a living God, even if you pay me a large amount of money to believe so. Thus I agree with Bernard Williams (1971) that the phrase ‘deciding to believe’ is a sort of contradiction in terms, which makes Williams James's famous title The will to believe at least odd. Williams links this feature to two others, which are related :
(2) Beliefs aim at truth: this means that when one's belief is shaped not by a concern for its truth, but by what one wants to be the case, the subject is more or less a wishful thinker or possibly a self-deceiver.
(3) In relation to this, beliefs are shaped by evidence for what is believed, and the degree to which a belief is reasonable is proportional to the degree of evidence that one has for its truth. This is why, prima facie, it is not reasonable to believe contrary to the evidence that one has. This is also why cases of wilful belief are not in general cases of belief at will, but of manipulation of the evidence that leads to having a belief (for instance, by use of drugs, or hypnosis).
(4) An agent’s beliefs are subject to an ideal of integration or agglomeration. Other things being equal, we should try to make our beliefs coherent or consistent, and to fit them together within some larger view. Some philosophers say that in this sense belief is a ‘normative’ notion; however, one does not need to express oneself in such grand terms to see what this means. But this is another sense in which our beliefs aim at being rational, or reasonable.
(5) Belief is context independent: at a given time a subject believes something or does not believe this something, but he does not believe that p relative to a context and not relative to another. It would be absurd, or a case of conceit, to say that on Wednesdays I believe that Hong Kong is a noisy town, and on Sundays I do not believe this. I certainly can have both beliefs if they answer to different pieces of evidence (for instance, there are a lot of cars on Wednesdays which do not come to town on Sundays), but it is an odd thing to say as a general truth about Hong Kong. Either I believe that Hong Kong is noisy, or I don't, period.
These five features, you will notice, apply to beliefs as well as to propositions, or to sentences. A sentence is true or false independently of whether we like it or not; depending on the evidence for its truth, sentences can form consistent sets, and their truth, although often contextual, can at least be envisaged independently of context. And similarly for propositions, whatever they are supposed to be. So it is not surprising that one can be tempted to assimilate the contents of beliefs with sentences or with propositions understood as entities more or less similar to sentences, and tempted to take believing to be some form of assent to sentences or to propositions, some form of holding true a mental or a public sentence. But although I do not deny that a number of our beliefs are tied to such mental states of assent and to linguistic acts of assertion, they need not be so. In this respect, I agree with Ramsey that not all of our beliefs need to be conscious. Assents form only the tip of an iceberg of beliefs as more or less persistent states or dispositions to act. So we have to distinguish beliefs as assents from beliefs proper. The main difference is this (and this is the sixth feature of belief):
(6) beliefs come in degrees, and these degrees are subjective probabilities. In this respect I agree with the Ramseyian, Bayesian view of beliefs. A belief corresponds to a probability function which is determined by a person's preferences regarding gambles. According to the Bayesian, one has subjective probability function p only if there exists a utility function u such that one's preferences maximise expected utility relative to p and u. In other terms, belief is not primarily a qualitative but a quantitative concept. This is why it is not completely right to say, as I said above, that either you believe that Hong Kong is noisy, or you don't. This disjunction applies to partial belief, belief to a degree between 0 and 1 and not only to full belief, belief to a degree 0 or 1.
Now Bayesians, when they defend the quantitative notion of belief, tend to reject the qualita tive notion. For instance, Patrick Maher writes: "The term belief is traditionally used to refer to propositions that a person accepts, or holds as true. Belief in this sense is closely associated with assent, and like assent it is a qualitative concept: a person either believes something or does not believe it." (Maher 1986: 363) And Maher goes on to argue that belief as assent or as acceptance of a proposition is ‘irrelevant’ to rational action (see also Jeffrey 1972). The Bayesian says this because, for him, to determine whether an action is rational you need only ascribe to an agent partial beliefs and degrees of desirabilities. The person need not believe, in the sense of assent to, or accept a given proposition, in order to act upon it. She only needs to partially believe that p. Against this many writers object that we can act rationally or deliberate when we take a given course of action without assigning precise degrees to our beliefs. They point out that if we were to attempt to assign such degrees, our computations would be so enormously complicated that we would risk combinatorial explosion. They also argue that we generally act and reason on the basis of propositions that we believe, or don't believe, or that we accept or not, period. (Harman 1986, Ch. 3).
How can we settle this dispute? I think the proper course to take is to split, as I did before, the concept of belief into two different kinds, one being belief proper (which has the six characteristic features of belief), the other being assent or acceptance (cf. de Sousa 1971, or Dennett 1978). On this view, assent or acceptance is one among other manifestations of belief proper. But indeed assent or acceptance is related to belief, in the following sense: in general we assent to propositions to which we assign a high degree of probability (at least superior to 0.5), that is to propositions which we take to be likely to be true on the basis of our evidence. So from this point of view, even if assent or acceptance is a different mental state from belief, it is still closely associated with it. Indeed in this sense, assent or acceptance is irrelevant, or epiphenomenal, with respect to rational action and with respect to reasoning or inquiry in general.
But the critics of the Bayesian, or of the dispositional conception of belief have something else in mind than just a difference between belief proper and assent on the other hand as based on belief. They want to argue that there are propositions to which we give our assent or that we accept independently of what we believe, in the dispositional-subjective probability sense. And they want to allow the possibility that we might accept things in spite of what we actually believe. So the strategy of splitting belief into two kinds has to be refined by refining the concept of acceptance.
2. A number of writers (in particular Stalnaker 1984, Van Fraassen 1989, Lehrer 1990, Cohen 1992, Ullman-Margalit and Margalit 1992, Bratman 1993) have argued that one must distinguish belief, understood in more or less the sense that I have given above, from acceptance. But the distinction they have in mind is not the distinction between on the one hand belief as a disposition or a permanent state, and on the other acceptance as a mental act of assenting to a sentence (holding true) or to a proposition. For them, accepting that p may be tied to such a mental act, but it need not be. Cohen's definition of acceptance is that
"to accept that p is to have or to adopt a policy of deeming, positing, or postulating that p, i.e. of including that proposition or rule among one's premisses for deciding what to do or think in a particular context, whether or not one feels it to be true that p." (Cohen 1992, p.4)
As these last words indicate, Cohen takes belief to be a ‘feeling that p is true’. I do not see what it means exactly. He says that belief is not a disposition or a habit, but a mental state of feeling, which is presumably conscious, unlike dispositions. Feeling that is a propositional attitude more or less similar to mental act of assent, so it does not coincide with what I have called belief proper. But we do not need to agree with Cohen on this notion of belief to see the contrast that he wants to make. He claims that acceptance differs from belief in the sense that one can accept, or posit, or suppose that p irrespective of what one believes on a matter. The reasons for accepting that p may be prudential, rather than evidential, and they are often tied to a context of practical deliberation. To take one of his examples, an attorney can accept, for the purposes of pleading before a jury, that his client is innocent, although he believes that his client is guilty. In this case, one accepts that p although one believes that not p. Acceptance in this sense is very close to what Kant calls pragmatic belief and which he describes with the following example:
"The physician must pursue some course in the case of a patient who is in danger, but is ignorant of the nature of the disease. He observes the symptoms, and concludes, according to the best of his judgements, that it is a case of phthisis. His belief is, even in his own judgement, only contin gent; another man might, perhaps, come nearer the truth. Such a belief, contingent indeed, but still forming the ground of the actual use of means for the attainment of certain ends, I term pragmatic belief. " (KRV, Transcendental doctrine of method, quoted in Margalit & Ullman 1992, p.173)
Here, as Kant says, the physician may not believe -- in the sense of having good reasons for it -- that it is a case of phthisis. But he nevertheless accepts it, for pragmatic reasons: it’s safer to err on the side of prudence. As Kant's example show, acceptance that p in this sense can coexist with a relatively high degree of belief that p, but it is a distinct mental act, indeed a sort of deci sion, that one takes, whatever one believes. One can frame a lot of examples like this. I may, on going out today, believe to a high degree that it will rain, because there are a lot of heavy clouds in the sky, and because the weather forecast predicts rain. But I don't believe it fully, because in this region it quite often happens that there are signs of rain but no rain comes. I may also find the prospect of carrying an umbrella not very pleasant; in fact, I dislike carrying an umbrella or anything. Nor do I like the prospect of getting wet. So in that case I may just accept, or take it for granted, that it will rain, and on the basis of this acceptance, carry my umbrella. It simplifies my deliberation. Although I believe that it will rain, it is not this belief which leads me to action. Or take an example of Bratman's. A person who plans to build a house is given various estimates of the cost by subcontractors (plumbers, carpenters, etc.). But she is unsure whether she currently has all the financial resources to do the whole thing at once. In the face of this uncertainty, she simply accepts the highest of the estimates, and determines whether she has enough money to start the project. Although she does not in fact believe that the house will cost so much, it is useful to accept that it will cost a certain amount in order to decide whether or not to start construction. As Bratman says, "in contrast, if you offered me a bet on the actual cost of the project -- the winner being the person whose guess is closest to the actual total -- I would reason differently.
Acceptance in this sense differs from belief. If we compare it to the various features of beliefs that we have given, we see that it lacks a number of them, or even has contrary ones.
(1') Acceptance is voluntary or intentional, unlike belief. In a certain sense, it is a form of decision to believe. But it is not a decision to believe in the sense that a belief could be brought out by our will, because in all the cases I have presented the agent does have certain, sometimes contrary, beliefs distinct from what he accepts. So we can say that the belief content, or the proposition which is accepted, is not the same or need not be the same as that which is the object of acceptance.
(2') Acceptance aims not at truth, but at utility or success. In this sense it is a pragmatic notion, not a cognitive or theoretical one. On can accept certain things that one believes to be false. Nevertheless, acceptance is also deeming true certain things. It goes with the attitude of acceptance that it take what I accept as true, although I may believe it is false.
(3') Acceptance, for the same reason, need not be shaped by evidence or evidential reasons.
(4') Acceptance is not regulated by an ideal of rational integration in the same sense as belief. I may, for pragmatic or perhaps social reasons, accept certain things which do not cohere with my other beliefs. For instance, it is in general a good social policy to accept that the speaker who is reading a paper is not boring, although you may not believe it, and repress yawns. Nevertheless, acceptance is subject to some sort of practical, rational integration, in the sense, emphasised by Cohen, that my policies of deeming true certain things have to be coherent.
(5') Unlike belief, acceptance is context-dependent. As I said, I believe (to a degree) that p independently of a context. But my acceptances are contextual: I may withdraw them in other contexts.
(6') Unlike belief, acceptance is an all or nothing matter. My beliefs have degrees, but in the deliberation contexts in which I accept that p, my acceptance is qualitative, not quantitative.
We now have a concept of acceptance that is distinct from the one we started with. Earlier I have distinguished acceptance from belief on the basis of the idea that acceptance is a form of conscious assent, the mental act of consciously assenting that, or in a more traditional terminology, judging that p. This can be either a voluntary act (for instance, it is one, based on] Descartes’ conception of judgement as a voluntary act of the will, where the will assents to the ideas presented by the understanding), or a non-voluntary one (from Hume’s point of view of belief as a certain feeling of assent -- and perhaps from Cohen’s point of view, assent is not an effect of the will). I do not want to discuss this issue here, although I side more with Hume than with Descartes on this point. From the viewpoint I propose, belief proper is distinct from assent, and from holding true a sentence. Let us call acceptance as assent acceptance1. But acceptance1 is distinct from the second species of acceptance that I have presented, which we may call acceptance2. Acceptance1 is related to belief in the sense that is prompted by belief; it is indeed the ordinary manifestation of belief, even if it is accomplished in foro interno. But acceptance2 is, as we saw, both distinct from belief, because it can occur without the corresponding belief, and from acceptance1, because it may not be tied to any assent. Acceptance2, in Cohen's or Bratman's sense, is a policy of taking for granted that p. The policy may not be reflexive or conscious. As Cohen says "acceptance [i.e or us acceptance2 ] may be tacit” (1992, p.12). Nevertheless, Cohen does not sufficiently distinguish acceptance1 from acceptance2, although he does distinguish accepting2 and performing some speech act of assertion. Ullman and Margalit (1992) have a nice word for acceptance2 : they call it holding as true, in order to distinguish it from holding true or assent.
I think it is quite useful to have this threefold distinction between belief, acceptance1 as assent, and acceptance2 as pragmatic acceptance. I agree with Cohen that failure to make both the twofold and the threefold distinction leads to all sorts of confusions on various topics (self-deception, purposive explanation, belief vs knowledge, etc.). Most important, it allows us to see that between the two extreme attitudes of entertaining an idea and knowledge, there lies a multiplicity of attitudes on the spectrum ( Ullman & Margalit 1992, p.169). There are others. For instance, there is the attitude of holding ‘come what may’ a statement, in the sense in which Quine says in his critique of the analytic-synthetic distinction:
"A statement can be true come what may [my italics], if we make drastic enough adjustments in the system." (From a logical point of view, p.13)
Here Quine is referring to an attitude, which is not simply practical (although it is to some extent practical) but also theoretical, of maintaining a statement in spite of good reasons we might have for not doing so. Acceptance, in this more theoretical sense, may also coincide with Van Fraassen's (1981, 1989) concept, when he says, in an instrumentalistic vein, that we may accept a scientific theory as empirically adequate without believing that it is true. This is closely related to acceptance2. But I shall not examine Van Fraassen's view here. Another attitude along the spectrum of doxastic attitudes that we have described, which Ullman and Margalit mention, is Wittgenstein's notion of ‘holding fast’ in On Certainty. I hold fast, for instance, that I was not born five centuries ago, or that the earth existed long before my birth, etc. Nor would it be appropriate to say that it is something I believe, or even that is something I know. In some sense it is something that I accept, or take for granted, but not in the more specific sense of acceptance2.
3. On Cohen and others’ conception of the distinction between belief and acceptance (acceptance2, but I shall drop the subscript in what follows, unless specified), the main point is that acceptance does not entail belief and is independent of it. Hence their complaint that many philosophers have lumped the two notions together. But why have the philosophers tended to lump them together? Because, I want to argue, acceptance is much more closely tied to belief than Cohen and others think.
The most obvious connection between the two attitudes it that, as Cohen (1992, p.17) says, "having a belief that p can often be taken to be at least a prima facie reason for accepting p, even though it may well not be the only, or the best, or not even a sufficient reason." We tend to accept something because in the first place we believe it. But what actually makes acceptance distinct from belief is that belief is not our only reason for accepting. There are two kinds of cases that we have to distinguish here. One is the case where we have a belief, but which is an insufficient but necessary reason for acceptance. This is what happens in the example of the physician given by Kant, in the example of the umbrella, and in the example of the person who plans to build a house. And there are the cases where our beliefs are neither sufficient nor necessary for acceptance, for instance the attorney case. But is it correct to say, as Cohen and Bratman do, that belief can be neither a necessary nor a sufficient reason for acceptance?
Take the first kind of case. Would I accept that it will rain if I did not already have the belief that it will rain? No, for in the case described, I have this belief. My acceptance that it will rain is, so to say, supervenient upon my belief: I would not accept it if I did not believe, and if I believed differently, I would not accept it. So here belief is necessary. If I believed that it will not rain, could I accept that it will rain? Cohen and Bratman say it is possible. The first kind of case collapses into the second. That would be a case similar to the attorney's case. But then I would accept, against all odds, that it will rain. It would be very strange, and indeed to some extent irrational, to go out with an umbrella on one of these sunny autumn days that we enjoy in Hong Kong, when all the weather forecasts predict dry and sunny days. I might insist that is safe to take my umbrella, but my sense of safety may be exaggerated. It is said that Schopenhauer always insisted, when he sat in a house, to be seated close to the door, in case of fire. That may have proved to be prudent, but it might better be described as a case of mania. Take now the attorney case. He believes that his client is guilty, but accepts that he is not. This is rational, given his professional role. But why does he accept that the client is innocent? Precisely because he believes that he is not. He would not have to take this act of acceptance in the first place if did not believe this. I want to suggest here that there is a similarity with self-deception. A self-deceiver is usually described as someone who believes that p on the basis of reasonable evidence, but nevertheless believes that not p, because of some unpleasantness of the belief that p, or because in some way he desires that it would be true that not p (Davidson 1985). Cohen criticizes (1992, Ch. 5) this description, and claims that what happens to the self-deceiver is that he believes that p, but accepts that not p. I think it is correct, and that, for a least a class of cases, it dispels the self-deception paradox: the subject is not irrational in the sense that he does not have contradictory beliefs; it is rather that he has two different states of mind, one with respect to p, the other with respect to not p. The reason why it is irrational, nevertheless, is that there should be a connection between what one believes and what one accepts. In the attorney case, something like this happens, although it is not a case of irrationality. It is, in a sense, similar to Ulysses’ attitude in the presence of the sirens (Elster 1979): he ties himself to the mast in order not to react in the way he fears he will react, namely by following these charming creatures. Similarly, we accept certain things because we are not sure of where our beliefs can lead us. But these acts of acceptance are hardly independent of our beliefs. Indeed, our beliefs are among the causes of (reasons for) these acts, the other reasons being desires, fears, etc. and other evaluative attitudes.
Finally, is acceptance unrelated to belief in the sense that the former, but not the latter, has degrees of subjective probability? Take here Bratman's example of the person who plans to build a house. Bratman's point is that acceptance is, like what he calls plans, quite independent of what I believe and of my subjective probabilities. He wants to say that there can be rational deliberation informed by plans but not by beliefs. Acceptance is just one attitude coherent with this planning process. But here again, we can agree that there can be acts of acceptance distinct from beliefs in the light of evidence, without agreeing that they are independent of subjective probabilities. For Bratman says that the prospective builder accepts that the cost will be a certain amount, say £ 20,000. But (as Clarke 1994 remarks), what he accepts here is not that the house will cost £ 20 000 . What he accepts is £ 20,000 but that the probability of the estimated cost of £ 20,000 exceeds the estimated cost that he makes. And he makes this estimation just in the way a good Bayesian would do it: by calculating the prior estimated probability, and by taking the conditional probability of the cost given this first estimate. He accepts finally that the probability that the total cost of 20,000 exceeds his prior estimate. But he certainly makes this estimate on the basis of his subjective probabilities. So acceptance here is not identical to a subjective probability, although it is based on one. So, contrary to what we said in point (6'), acceptance is tied to subjective probabilities. Cohen recognizes this when he says that although acceptance is not a matter of degree, acceptance-worthiness is (Cohen 1992, p.114). This does not threaten, however, the idea that Cohen wants to put forward, i.e.that acceptance is a qualitative attitude, voluntary, contextual, whereas belief is an quantitative, involuntary, non-contextual attitude. But it does threaten the idea that acceptance does not entail belief.
If this is correct, then, there is an important distinction to be made between the various kinds of attitude related to belief: belief itself, holding- true, assenting (what I have called acceptance1) and pragmatic acceptance (acceptance2). The distinctions are important not only for themselves, but also for a wide variety of philosophical issues, which I have only mentioned in passing. But we should not forget that these attitudes are, nevertheless, all closely related to belief. Given that belief is primarily a cognitive attitude, which relates to how we get information from the world, besides being also an attitude which plays a role in our actions, we should not forget that however ‘pragmatic’ or ‘action-oriented’ our acceptances can be, they are tied to our beliefs, and therefore to our cognition. This sets a limit to pragmatism, understood as a doctrine according to which the pursuit of knowledge is determined by our actions and their success. However much we will to believe, i.e. to accept, there is still much to believe without willing it.*
Bratman, M. (1987) Intentions, Plans, and Practical Reason, Cambridge
Mass: The MIT Press
(1993) ‘Practical Reasoning and Acceptance in a Context’, Mind 102, 1-
Clarke, D. (1994) ‘Does Acceptance entail belief?’, American Philosophical
Quarterly, 31, 2, 145-155
Cohen, J. (1989) ‘Belief and Acceptance’, Mind, 98, 367-369,
(1992) An Essay on Belief and Acceptance, Oxford: Oxford
Davidson, D. (1980) Essays on Actions and Events, Oxford: Oxford
(1984) Inquiries into Truth and Interpretation, Oxford:
(1985) ‘Deception and Division’, in Le Pore (1985)
Dennett, D. (1978) Brainstorms, Cambridge Mass: The MIT Press,
de Sousa, R. (1971) ‘How to Give a Piece of your Mind’, Review of Meta
physics , 35, 52-79
Elster, J. (1986) Ulysses and the Sirens, Cambridge: Cambridge University
Engel, P. (ed.)To appear , Believing and Accepting , Dordrecht: Kluwer
James, W. (1901)The Will to Believe, New York: Longmans Green
Lehrer, K. (1990) Theory of Knowledge Westview, Col.: Westview
Le Pore, E.(ed) (1985) Truth and Interpretation: Essays on Davidson,
Maher, P. (1986) ‘The irrelevance of belief to rational action’, Erkenntnis,
Marcus, R. (1986) ‘Some Revisionary Proposals about Belief’, in her
Modalities, Oxford: Oxford University Press1993
Stalnaker, R. ( 1984)Inquiry, Cambridge Mass: MIT Press
Ullman-Margalit, E. and Margalit, A. (1992) ‘Holding true and Holding as
True’, Synthese, 92, 167-187
Van Fraassen, B.(1980) The Scientific Image, Oxford: Oxford University
(1984) Laws and Symmetry, Oxford: Oxford University
Williams, B. (1971) ‘Deciding to Believe’, in his Problems of the self,
Cambridge: Cambridge University Press
 Chad Hansen, A Taoist Theory of Chinese Thought, Oxford: Oxford University Press, 1992, p. 287
* Versions of this paper have been read at the Universities of Hong Kong, Osaka, Dundee, Edimburgh, Stirling, King’s College London, at the Centre de Recherches en Epistémologie Appliquée, Paris, at a Davidson conference in Leuven in 1995 and at a conference on “believing ancd accepting” that I organized at the University of Caen in 1995 (proceedings to appear in Engel (ed) to appear . I thank for their remarks the participants in these seminars and conferences, and in particular Laurence Goldstein, Michio Kobayashi, Neil Cooper, Roger Young, Tim Williamson, Mark Sainsbury, David Papineau, Richard Sorabji, Jean Pierre Dupuy, David Clarke, Jonathan Cohen, Peter Railton, Donald Davidson and Philip Pettit.