VOLITIONISM AND VOLUNTARISM ABOUT BELIEF

 

Pascal Engel

Université de Paris IV-Sorbonne

 

 

 

 

1.                  Introduction

 

             In the discussions about belief and the will, two sorts of issues tend to be conflated. One is psychological: can we, as matter of psychological fact, believe something at will, or under the influence of motivational factors? The other is normative: should we, as a matter of normative guidance, believe things for which we have practical, or pragmatic reasons? This ambiguity is present in the title of William James’ famous essay “The will to believe” (James 1897). Does it refer to the fact that we can form beliefs at will or does it contain an advice about how we should form our beliefs? Of course the two questions are connected, for if it is a fact that we can believe at will, it can also be a normative recommendation that we should have more beliefs formed in this way. It may have been James’ view. The two questions, however, are different, for it may be the case that we can form beliefs at will but that there is no good normative reason to do so. This may be Clifford’s position (1879): “It is wrong, always, everywhere, and everyone, to believe anything upon insufficient evidence”. And it may be that it is desirable, or dutiful, to attend to practical considerations when forming our beliefs, and to attempt to form as many beliefs as we can based on such practical or pragmatic reasons, but that as a matter of fact we cannot form such beliefs. If so, the advice would be void, just as if one were to recommend (for the sake of exercising our imagination, say) to form impossible desires (desires which cannot be fulfilled, such as the desire to square the circle). On the latter hypothesis, indeed the falsity of the psychological thesis (that we can form beliefs at will) would entail the irrelevance or vacuity of the normative thesis (that we should form such beliefs).

               The close connection between these two theses is probably responsible for the fact that they are often referred to under the same name, voluntarism. Some writers (e.g. Pojman 1985) talk of volitionalism, or volitionism. In order to distinguish the two theses, I shall talk of volitionism to refer to the psychological thesis, and I shall reserve the label voluntarism for the normative thesis. My aim in this paper is to provide grounds for distinguishing them, as well as to explore some of their relationships. In so doing, I shall attempt to defend a weak or modest version of both theses, while I shall follow the ordinary philosophical wisdom that the stronger versions of volitionism and of voluntarism are false.

 

 

2.                  Belief

 

           Many of these issues are definitional, and turn upon a precise definition of the kind of state or mental attitude that we call “belief”. I take it (see Bratman 1993, Engel  1998) that belief, as ordinarily so called, has the following five main characteristics:

 

(1)                           Beliefs are involuntary and not normally subject to direct voluntary control

(2)                           Beliefs aim at truth

(3)                           Beliefs are normally shaped by evidence for what is believed

(4)                           Beliefs are subject to an ideal of integration or coherence

(5)                           Beliefs come in degrees

 

(1) Is the question at issue, so I shall not comment upon it for the moment.

   (2) says something about the contents of beliefs. As an attitude or as a psychological state, a belief is neither true nor false; but as a propositional attitude, its content is a representation which can be either true or false. While beliefs may of course be false, they “aim at truth” in the sense that truth is their goal, or that it belongs to the game of believing that we should try to have true beliefs and avoid false ones (Dummett 1959). This is often formulated in terms of Anscombe’s (1958, see also Searle 1983) famous notion of “direction of fit”: beliefs, have the “mind to world ” direction of fit (they are such that the mind should adapt itself to the world), unlike desires, which have the “world to mind” direction of fit (are such that the world should adapt itself to the mind, in which case they are not true, but satisfied). This consideration, as one sees, is partly descriptive (the normal role of belief is to register information, which can be true or false) and partly normative (we should aim at true beliefs). Similarly (3) is in part a normative and a descriptive consideration. It is descriptive in the sense that it says that usually and normally beliefs are formed on the basis of evidential considerations, which provide us with most of our reasons for our beliefs. The normative thesis which goes with (3) is that it is rational to form our beliefs on the basis of evidence, and there is good reason to do so. The stronger thesis that it is the only  sort of good reason is known as evidentialism. It says that we should rule out any other sort of consideration than evidential when we evaluate our reasons for believing. Again, (4) is also partly descriptive and partly normative. Beliefs contents are holistic, they acquire their status as beliefs only within a web of inferential relations to other beliefs and to other states (so a creature which would have only one belief would not count as having a belief). But (4) contains also a normative advice: on the whole, when our beliefs do not cohere with others that we have, we tend to make them coherent, and should do so. The requirement of coherence is not necessarily the same as the requirement of being evidence-based – in some respects we can disregard either one in favour of the other[1]. But both requirements can be said to be epistemic, and relevant to our epistemic rationality or to good epistemic reasons to believe. Finally (5) beliefs come in degrees, and this feature is related to the others: one can have more or less evidence for our beliefs, or can more or less integrate them. In such cases, it is said that we can have more or less confidence in our beliefs, and such degrees are called degrees of partial belief, or of subjective probability.

       I disregard two other important  features, which have to do with the dispositional nature of belief, and with whether they can be conscious or not, because they will come up with our discussion of (1).

 

2. Volitionism: a first approach

 

              Let us now turn to one of the main theses at issue, volitionism. On the rough

characterisation above, and as a psychological-descriptive thesis, it is generally the view that we can form beliefs “at will” or that we can “decide to believe”. Hence it denies that the feature (1) of beliefs is necessary. Notice first that it is not bound to be the thesis that all  of our beliefs are formed in this way. For most of our beliefs seem not to be formed in this way. Paradigm cases of beliefs, such as perceptual beliefs, beliefs arrived on the basis of logical inferences, or beliefs based on testimony do not seem, prima facie, to be under the control of the will. The volitionist can grant that, and claim that it is enough to defeat the opposite thesis, anti-volitionism, according to which none of our beliefs can be voluntary, that some of our beliefs are such. But then what are the beliefs in this subset, and what does “voluntary” or “under the control of the will” mean? Neither is volitionism the thesis that we may “want to believe” certain things in the sense of hoping that we could believe them. In a number circumstances, “I want to believe that P” means something like “I wish it were true”. It is out of question that we have such wishes. In this sense the expression simply means that I wish, hope, or desire that P. But “willing to believe” in the sense of “believing at will” means, prima facie, that I can believe successfully that P, or that my will can influence the content of my beliefs and my states of believing that P independently of how the world is.

           First, it can mean that we can have direct control over the formation of some of our beliefs, irrespective of whether we think them true or false. Let us call this direct volitionism. Intuitively this sounds most implausible: can I believe, “just like that”, in the same way as I can snap my fingers, that my trousers are on fire, or that the Dalai Lama is a living god? The view is not only implausible if I have no particular reason, or no reason at all to form such beliefs, but it is also implausible even if I have a strong practical reason to do so. Even if you pay me a large sum of money to believe that the Dalai Lama is a living god, or if I realise that the fate of the world depends upon it, it seems that I cannot believe this “at will”.

         Second, it can mean that we can have an indirect control over the formation of some of our beliefs. Let us call this indirect volitionism. On the face of it, this seems much less implausible than direct volitionism. It is generally admitted that we can form some beliefs voluntarily through various devious and indirect means. Although I cannot, say within the next two seconds, or even within the next hour, come to believe that the Dalaï Lama is a living god ,  I can  certainly come to believe that the Dalaï Lama is a living god if in undergo hypnosis, or if I am subject to more or less long-term indoctrination. Dictators, publicists, or priests would not set up rites and ceremonies whereby people solemnly swear before the flag, register subliminal information, or pray their gods, if they were not confident that belief can be obtained by such indirect means, whether of not the indoctrination is voluntary in the first place.

Indirect volitionism is even more plausible on the view that belief is a sort of habit or disposition. A case at hand would be Pascal’s advice in the second part of his wager argument, when he says (somewhat ironically) that if belief in God cannot be got from his utility considerations, the agent has better go to church, mimic the habits of believers, etc. (“It will dull you”, Cela vous abêtira). These are typically cases of beliefs induced by performing an action (taking a drug, going to the hypnotist, going to the church and taking the holly water, etc.), and they seem to qualify as action just as much as cooking a cake qualifies as an action while it is performed through the performing of various other acts, such as breaking the eggs, melting in sugar, etc. Notice also, that, on the indirect reading, the acquisition of the belief need not be slow, and the difference between the initial act and the result need not involve a long span of time. On this reading, I might acquire the belief that the Dalai Lama is a living god “just like that” but nevertheless through indirect means, if, for instance, I swallow a special pill which I know to have this immediate result, even though on many cases it seems that indirect intentional belief formation has to take some time.

           Third, although it is unclear whether this counts in favour of direct or in favour of indirect volitionism, aren’t there cases of self-fulfilling belief formation, whereby the very fact that one has them makes them true? Such cases as that of the coward who believes that he is courageous and, the frightened diver at the top end of the highest diving board at the swimming pool who does his best to believe that he can dive from it, and as a result of his believing it, succeeds in getting the belief that he will dive, and actually dives successfully, are familiar. Self indoctrination can  be successful.

Fourth, volitionism is often associated to a weaker, and vaguer, thesis: although beliefs might not be under the direct or indirect control of the will, some elements or factors within belief formation involve desires, intentions, or volitions. There is the familiar experience of wishful thinking, of belief arrived at through the influence of desire and wishes, or the formation of beliefs in self-deception and other abnormal or pathological cases. Literature, philosophy and psychiatry  are full of such episodes.  As Locke says, quoting Julius Cesar: “Tell a man, passionately in love, that he is jilted;  bring a score of witnesses of the falsehood of his mistress, ‘tis ten to one but three kind words of hers, shall invalidate all their testimonies. Quod volumus, facile credimus.” (Essay, IV , xx , 12) Or as l’Ariosto says :

 

“Questo creduto fu che' il miser suole

dar facile credenza a quel che vuole”

(Orlando Furioso I, 56)[2]

 

There is also strong evidence from social psychology in studies about cognitive dissonance and similar phenomena . For instance T. Gilovitch (1991 ) reports that

 

“A survey of one million high school seniors found that 70% thought they were above average in leadership ability, and only 2% thought they were below average. In terms of ability to get along with others ,  all  students thought they were above average. 60 % thought they were in the top 10 %, and 25 % thought they were in the top 1%. A study of university professors found that 94 % thought they were better at their jobs than their average colleague.”

 

And a number of famous works by Kahneman and Tversky (1982) suggest that motivation can bias our beliefs. The trouble is that it is not clear what these motivational factors are (so the extent of the word “facile” in the above quotations is not determined). Neither is it clear that such phenomena as cognitive dissonance or bias are solely motivational: there are cognitive biases as well, and it is not easy to sort them out from those which would be purely motivational.

           Finally, it is often claimed that the will enters as a decisive factor in belief formation when we assent to, or when we accept our beliefs. This is Descartes’, or the stoics’ view that assent, or judgement, is voluntary, and takes the main responsibility for error. This view distinguishes the presentation of the ideas of the understanding from their assent by the will in judgement. So it does not deny that there is a passive element in the formation of belief (for the ideas or propositions that we consider are there whether we want it or not), but it grants that an active element comes in we judge, assent, or assert the contents of what is so presented to the mind. In so far as one does not assimilate judgement and assent on the one hand and belief on the other, and as one takes assent to be merely an internal act which accompanies belief, or its external counterpart when assent is expressed in words, this view vindicates only partially volitionism. It would vindicate it fully only if judgement or assent could be identified to belief. But this view seems much too strong: there are many beliefs which never come in front of our minds, and which we can be said to have, although we never have the occasion to assent to them, let alone to manifest them in our behaviour. In fact the existence of such dispositional or tacit beliefs seems to go against volitionism, and to confirm feature (1).

                When we review in this way some of the prima facie phenomenological or experiential evidence in favour, or against, volitionism, it seems that we can make little progress. We are tempted to conclude that some beliefs have voluntary features, and that some have not, but the evidence is inconclusive on either side. The real and interesting arguments must deal with the conceptual features of belief.

 

4. Volitionism further defined

 

Most writers who discuss volitionism  take it to be the following general thesis:

 

(V)     ( a) One can acquire a belief by way of intending to have it;

(b)   Intending to believe that P stands in relation to believing that P in the same way as, for instance, intending to walk stands in relation to walking; or as intending to make a cake stands in relation to making it;

(c)    Hence acquiring a belief can be an action, which has all the ordinary features of an action (it is intentional under the description believing that P, and it brings it about, as a result of this intention, that the agent believes that P.

 

Several comments on this initial definition are in order.

 

First (a) does not say, contrary to what some formulations suggest, that believing, or having a belief is a kind of action. It only says that acquiring a belief, coming to believe, or forming a belief can be an action. The belief itself, which is the result of the act of belief formation, is not itself an action; it is its product. Beliefs are states, or mental events, which have an given content, which can be true of false. Actions may be events, but they do not represent anything, nor are they true or false. So volitionism is not committed to the absurd thesis that belief itself is an act. In particular, it need not deny that beliefs have, in the sense given above (2) a different direction of fit than desires, intentions and other motivational states. Beliefs aim at truth, and are true or false according to whether or not the world fits them or not, whereas desires aim at satisfaction, and are satisfied or not according to whether they fit the world. It is not implied by volitionism that beliefs and desires can exchange their roles and have the same direction of fit, as some writers who argue against the existence of such states as besires  - beliefs which would also be desires-  have argued (cf. Smith 1994 for instance).

              Second volitionism, according to the above formulation  (in particular (b)), implies that coming to have a belief is a basic action, in the usual sense of this phrase in action theory: it can be performed directly without performing some other action. In this sense, it should be like raising one’s arm, or moving one’s legs as a result of the immediate intention of going out for a walk. So, on at least one version of volitionism, coming to believe that P can be the direct and immediate product of the intention to believe that P. This would cover the cases of believing at will “just like that”, and we have seen that this is much implausible. But there is another version according to which this basic action does do reach its results immediately or directly, but mediately and indirectly.  This is parallel to the familiar cases, in action theory, where a basic action is done with the intention of performing, through it, another action which is not basic but the remote product of the basic one. To take up Anscombe’s famous example, a man raising his arm can be, under that description, the action of raising his arm, but on another description, the action of poisoning the well, or the action of killing the inhabitants of the house, etc. Under some descriptions, the further and remote effects of an action can even be events under the description of which the action was not intended, such as poisoning the whole neighbourhood. But volitionism, as stated above in form (V) does not have this consequence. It says that the act of forming a belief has to be intentional under the description believing that P. Hence the intention, and the description of the intended act, have to be consciously entertained. If Oedipus  intentionally forms the intention to believe that Jocasta is sexually attractive, and if, by successfully getting this belief, he turns out to believe that his mother is sexually attractive, this could not count as a case of belief acquired at will (and this formulation is, admittedly, bizarre). But nothing in the previous definition of a basic action prevents the volitionist from claiming  that one can come to believe that P as a result of a basic action that the agent considers as a means to a further end, namely to believe that P. In such cases, the belief can be reached by indirect means. The fact that an action is performed through the doing of something else does not prevent it from being an intentional action, provided the agent was conscious that he could reach this result through this intermediary act. And this seems to be much more plausible that the direct version, for it covers the cases of beliefs acquired through hypnosis, the taking of drugs, or various sorts of habit formation leading in the end to the belief.

Third, if volitionism says that coming to believe can be an action, it must imply that, like any action, it can be performed for certain reasons. Indeed “acting for a reason” is just another name for “acting with an intention”. But unlike ordinary actions, beliefs at will can be performed for two kinds of reasons at least: epistemic reasons, whereby we want to believe that P because we think that P is true, and practical reasons, whereby we want to believe that P because  we think that P is good to believe. In the latter case, I may want to believe that P even though I believe that P is false. For instance I may want to believe that my neighbours are friendly, because I think that having such a belief would enhance my good relations to them, even though I believe that they are not so friendly. Notice that my having these practical reasons for having this belief in no way excludes my having other reasons, epistemic ones, for my willed belief that my neighbours are friendly does not cease to be a belief which I want to have because it is true, or because I want it to be true. But I might also want to believe that P only for practical reasons, irrespective of whether the belief is true or not, and irrespective of my epistemic reasons to have it. For instance I may desire to believe that my neighbours are friendly just because I desire to have this belief, and because I find it useful. And just as one can perform an action for no particular reason, such as scratching one’s nose, it should be consistent with volitionism that I can acquire a belief for no particular reason, whimsically so to say. So depending on the kinds of reasons, or lack of reasons, that we have for our acquisitions of beliefs at will, there will be four versions of voluntarism :

 

(i)                  one can believe at will for epistemic reasons alone

(ii)                one can believe at will for epistemic and practical reasons

(iii)               one can believe at will only for practical reasons (irrespective of the epistemic ones)

(iv)              one can believe at will for no particular reason

 

We are now in position to formulate several versions of volitionism. The stronger version, or strong volitionism,  is this :

 

(SV)  There can be an action, coming to believe at will that P, which is intentional, conscious and direct, and which is motivated by practical reasons, irrespective of epistemic reasons, or by no reason at all.

 

A weaker form is indirect volitionism :

 

(IV) There can be an action, coming to believe at will that P, which is intentional, conscious and achieved by indirect means, and which is motivated by practical reasons or by no reason at all, irrespective of epistemic reasons.

 

An even weaker form is weak volitionism:

 

(WV) There can be an action, coming to believe at will that P, which is intentional, conscious, and achieved either (i) directly or (ii) indirectly, and motivated by

(ii) practical or (iv) epistemic reasons.

 

This will have to be qualified, but let us rest content with this formulation.

Anti-volitionism will be the thesis that :

 

(AV) There cannot be an action, either direct or not, coming to believe that P, which is intentional, conscious, and which is motivated by either practical or epistemic reasons.

 

I want to argue that SV is false, but that WK is true, hence that AV is false.

 

 

5. Conceptual arguments against volitionism

 

               V is usually criticised on conceptual or a priori grounds. The defence takes the form of an argument to the effect that it would be conceptually  incoherent to call a state a belief if it were produced and maintained under the influence of motivational factors. Arguments among this family have been initiated by Williams (1970) and  discussed (among others) by writers such as Elster (1979) , Winters (1979) , O’Saughnessy (1980) and Pojman (1985).

                Williams argues that a state is a belief only if we have reasons to think it true, and that it performs the function or representing an external reality. Beliefs typically are caused by evidence for them and “aim at truth”. It follows that if a person were to have a belief merely at will, irrespective of its truth or falsity and irrespective of the evidence one has for them, the state in question would not be a belief. It would not be a belief, because it would violate the constraints (1)-(5) proposed in section 2 above. It is important here to see that these constraints are epistemic  constraints. They specify what a rational belief is according to criteria which are epistemic or cognitive. And the argument claims that a state which do not obey such criteria is not a genuine belief. Let us call this the conceptual argument. The argument seems to work against the two versions of strong volitionism that I have distinguished above: the direct one and the indirect one. Take the direct one first. If I could form a belief -  say the belief that the Dalai Lama is a living God -  for practical reasons – for instance because you pay me a large sum of money if I successfully perform the action of believing this, and irrespective of its truth of falsity, i.e whether or not I think it is true, or without any epistemic warrant for it, then if I were successful in the performance of this action, I would believe that the Dalai Lama is a living God without rational justification. In the example just given, this belief does not fall short of rational criteria, for it is supposed that it satisfies at least one rationality condition, a utility condition : you would pay me a large amount of money (and let us suppose that I am interested in getting such a sum). So on this hypothesis it would follow that I have some reasons to have this belief: practical reasons. But direct volitionism, as defined above, also says that I could form a belief short of any reason at all, including practical ones. So let us suppose that you do not propose the large sum of money, and I have the intention to form the belief that the Dalai Lama is a living God “just like that”, and whimsically, so to say. Direct volitionism allows for the possibility of such whimsical belief formation. And the conceptual argument  says that it is impossible, because it is in the nature of belief that it should obey epistemic constraints such as (1) –(5).

          The same point applies to beliefs formed indirectly, hence to indirect volitionism, as in the cases were they are formed as a result of the decision to acquire an habit, or to produce in oneself a causal mechanism of some sort which will have the effect of producing the belief. This covers both the cases of belief inducement by drugs or hypnosis, or the Pascalian cases. Some writers (e.g. Losonsky to appear) have argued that these would be bona fide cases of acquiring beliefs at will, provided one distinguishes the direct attainment of belief from the indirect attainment, along the lines of Dretske’s (1988) distinction between structuring causes and triggering causes. If we distinguish a certain event (say one’s arm raising), an event as the product of an internal cause (one’s arm raising, as produced by an internal cause) , and the process that sets up this production of events (the movement of one’s arm being produced by an internal cause), then belief at will might be the outcome of a process of which the event willing to believe is but a part: the product, the belief, is one thing, and the process, the acquisition of the belief, is another thing.  There is, in Dretske’s phrase, the structuring cause on the one hand, and the triggering cause on the other. The belief, on this analysis, is not itself voluntary. No doubt we can form beliefs willingly in this way. But is it so different from arguing that one can believe easily at will that, say, Rome is a beautiful city, by going to Rome, and seeing indeed that Rome is a beautiful city? In such a case too, there is a set up which I know in advance that if I go to Rome I shall believe that it is a beautiful city. So I know in advance that I will have that belief, and that it is true. Hence the belief is not formed independently of its truth; indeed it is formed because I know that it is true and perform the appropriate action to produce it: go to Rome and see Rome. Cook (1987) gives a similar case which he claims to be case of non self deceptive, and of rational believing at will: an ambitious creationist who realises that his professional success will depend upon his becoming an evolutionist, and hence goes to study in Harvard and ends up really believing in the truth of the theory of evolution. “He smiles inwardly at the benighted bumpkin he was”. Such cases  are very close to another one, which is often invoked against the conceptual argument : self-fulfilling beliefs. (Velleman 1989, p.129 sq.). Self-fulfilling beliefs, those that are true just because you have them –believing that one is courageous, believing that one is a good swimmer, a good performer, that one is intelligent , etc. – are also the cause of their own truth. Such beliefs are not formed irrespective of their truth, and their connection to the truth is not threatened by the fact that they are caused by the subject’s wishes. Are these roundabout routes to belief acquisition really counterexamples to the conceptual argument? They are not. If someone knows that there is a certain set up – drug, hypnosis, pill, self-fulfilling belief – which will invariably produce a certain belief, the belief will be true or false irrespective of its being willed. Its truth or falsity will not be created by the subject. It will still be true because of the relationship between its content and the state of affairs which makes it true. It will still have the mind to world direction of fit. One way or another, if there is in me an attitude which is not shaped by a concern for its truth, it is not a belief. Self-fulfilling beliefs, or other set ups, are under our control. There is no problem in willing something which is under our control. But willing something which is not under our control is harder. It is indeed conceptually impossible.

But whatever the force of conviction that it carries, the conceptual argument can be resisted. Why should we stipulate that a mental state which does not obey conditions (1)-(5), or a subset of them, is not a belief? Aren’t there states arrived at by violation of these conditions, such as those which are the product of a process of wishful thinking, or of self-deception, or other more pathological conditions, but which can well qualify as beliefs, even though we recognise that they are in some sense abnormal ? And after all the processes in question seem to be able to produce their results, whether or not they answer the epistemic reasons heralded (Bennett 1990). The conceptual argument has little force against such considerations. After all, psychiatrist do call some beliefs delusive, even though they are obviously not formed according to ideal epistemic conditions specified by (1)-(5). What is wrong with that ?

           In order to be efficient, the conceptual argument has to be reformulated. The crucial condition is that, according to strong volitionism, beliefs could be formed at will for practical reasons whether or not we have good reasons to think them true. And it seems that this condition could not be satisfied (see especially Pojman 1985). If a person were to form ( directly or not) a belief that P, while being conscious that this belief formed irrespective of truth considerations, and hence is not a genuine instance of belief, then this person would be in incoherent condition. He would grant both that he has acquired the belief at will (again directly or indirectly), but, because the belief has not been acquired in the normal way and that he recognises that, he would also recognise that the state that he has thus acquired is not a genuine belief. So he could characterise his in own psychological condition in this way:

 

“I believe that P (as a result of self inducement), but I do not believe that P

(since self induced beliefs are not typical beliefs)”

 

In other words, the person in question would be in sort of position which is very similar to the one which is exemplified by instances of Moore’s paradox:

 

P, but I do not believe that P

 

which carry over into a contradiction if one expands the first conjunct :

 

I believe that P, but I do not believe that P

 

Now it is important to see that this argument (let us call it the incoherence argument), to the effect that the agent would thus be in a sort of self-defeating or incoherent condition, does not tell against the psychological  possibility of there being such abnormal believing conditions. Neither does it tell against the fact that such self induced “beliefs” could still be beliefs, but irrational ones. On the assumption that self-deception is a genuine and real phenomenon, this does not amount to denying that it exists. The argument only implies that an agent who does acquire such wilful beliefs, and is aware (even tacitly) of the usual norms of belief formation, should be in some sort of conflict, which indeed looks very much like self-deception. If one both believes that P, and believes that one’s belief that P stems from a decision to believe, then self-deception is involved at some point. But if self-deception is irrational, then believing in this way is also an irrational form of behaviour. We should also put a restriction on the incoherence argument: if self-deception involves, as some writers claim, beliefs which are not completely conscious, or unconscious, the argument does not work. But remember that volitionism is the thesis that conscious beliefs cannot be got at will. Hence such an analysis of self-deception, or the involvement of unconscious beliefs in self inducement, would not to against it.

              The difference between the incoherence argument and the conceptual argument is that the former does not say that it is impossible to form beliefs at will directly or indirectly because the resulting state would not be a belief, but that given that one is conscious that the beliefs formed in this way would not be a genuine belief (but a “bewill”, if one can coin this phrase), and given that the agent has a rational intention to form this belief, he could not be willing to do so. Something would be wrong, not only within his epistemic reasons for doing this purported action, but also with his practical, not epistemic, reasons. It could, however, be said that this line of thought begs the question against strong volitionism, construed as a merely psychological thesis about the possibility of forming beliefs at will for no particular reason, since it reintroduces into the conditions of formation of a belief that the belief should be formed and held for certains reasons, either practical only or both practical and epistemic. In particular, the incoherence argument does not rule out the formation of such beliefs when the agent is not conscious of having them, or when he has lost memory that he has formed them in this way. And considerations about the formation and the maintenance of beliefs  seem to support this: it often happens that, after we have formed a belief in a particular way, including an irrational one, we loose track of the reasons why we came to have it in the first place. The belief could be further entrenched by the presence of other beliefs, just because of its coherence with them (this would vindicate the feature (4), coherence, of our initial list).

              Again, there would be no reason to deny this possibility, as a psychological possibility (an it may be that it is what happens often in cases of self-deceptive belief formation). But there would still be a problem with it. For in the case of memory loss of the irrational style of formation of my initial belief, it is supposed that this belief is further entrenched by coherence with others. This amount to granting that I could find, at a later stage, reasons to believe it true ( for coherentist epistemic reasons). But if this is so, then it can hardly be said that I am conscious of having formed the belief at will for no particular reasons. On the contrary, I would have acquired in the meantime epistemic reasons to hold it.

            Such a situation, if it could occur, would also highlight an irrational condition of my action. For in the case described, I have no control over whether my memory will retain this belief or not, nor over the reasons why I held the belief in the first place. But the bringing about of an action over which I have no control is not the bringing about of a rational action. This is true also of the indirect cases, such as taking a drug, hypnosis, and the like. When I take a drug, or when I get drunk, I typically loose control of the causes which lead me to the formation of my beliefs. I have only direct control over the basic act which initiates the causal sequence leading to the belief. To take up a distinction proposed by Bennett (1990) when I raise my arm as a result of my intention to raise my arm, my action is motivationnally immediate, since I have direct control over its effect under this description, but it is ontologically mediate since I have no direct control on the movements of my muscles nor of my neurons firing. But in order to have an intentional belief, you should have control over the sequence of events which lead to the belief. But in the case at hand, you don’t have it. I cannot be said to perform an intentional action if I have no control over its effects.

The upshot of all this is the following. If I cannot control whether the conditions of my act of believing something obtain, and if this act is performed for no reason, and if, moreover, I know that I cannot control these conditions and that my belief is formed for no reasons, then it is impossible for me to intend to have the belief. The situation is very similar to that of Kavka’s famous toxin puzzle (Kavka 1983), where I am asked to form an intention to perform an action which I know that I will have no reasons to perform, or which I believe that I shall have no rational intention to perform.

This, I think, is sufficient to refute volitionism under the form  SV and under the form IV. But it does not seem enough to refute weak volitionism, WV. Remember that this says that we could form beliefs at will, either directly or indirectly for some reasons, practical or epistemic. The crucial factor here is the nature of our reasons for forming the belief. The correctness of the view depends upon the normative nature of our reasons to believe. And it is at this point that the issue of volitionism connects up with the issue of voluntarism, as I have characterised it above.

 

6. Voluntarism and volitionism

 

            The issue of voluntarism, remember, is normative because it asks a question about the rationality of our reasons for forming beliefs, and not simply a question about what kinds of reasons we have, as a matter of fact. Given that there seems to be, prima facie, two sorts of reasons for forming beliefs, epistemic and practical, the question is: can it be right, correct, or rational, to have reasons which would be purely practical, or pragmatic, to form a belief? A positive answer would seem to imply that the usual epistemic reasons ( encapsulated in (1)-(5) above) either could not be necessary for the formation of belief, or not sufficient. On the non necessity thesis, the view would amount to saying that we could rationally form beliefs at will solely for practical reasons, irrespective of our epistemic reasons. On the non sufficiency thesis, the view would say that we could form beliefs at will both  for practical and epistemic reasons. The first view, which we may call strong voluntarism,  is most often attributed to Pascal in his wager argument, and possibly to James’ in his essay “The will to believe”. The second one could also be what James meant. In some cases, it might be good to concede something to what James calls “our passional nature” in the search after truth. James says, putting the stress on this thesis : “Our passional nature not only lawfully may, but must, decide an option between propositions, whenever it is a genuine option that cannot by its nature be decided on intellectual grounds; for to say, under such circumstances: “Do not decide, but leave the question open”, is itself a passional decision, - just like deciding yes or no , - and is attended with the same risk of loosing the truth.” (James 1897, p.11) Let us call this weak voluntarism.  But if so, A thesis which would deny both seems to be the strong evidentialist view of Clifford. Clifford says both that it would be “wrong”, in a sort of moral sense, to form beliefs only for practical reasons, but also to let practical reasons compete with epistemic ones and to overcome epistemic reasons.

Can it be rational to have only practical reasons for a belief?  There are some arguments to the contrary such as Zemach’s  (1997)[3]:

 

Let PV (Bp) be the practival value of a belief that p

Let EV (Bp) be the epistemic value of a belief that p

Suppose PV(Bp) >  PV(- Bp) and that EV (Bp) = EV ( - Bp). Then it seems that you should believe that p in such a case. So the maxim is: believe that p iff

PV (Bp) > PV (- B p). Let us call this Pascal’s principle. ( P1)

Pascal ‘s wager can be formulated in this way:

Suppose Bp is the belief that God exists. Pascal says : we should believe that p,

for PV ( - Bp) < PV (Bp) where PV (- Bp ) is much much lower than   PV (Bp).

Suppose we can establish P1 (this implies carrying out infinite calculations to compare the P utilities and the E utilities of all our other belief with respect to p )

The question is : should we believe P1?

Well, if one accepts Pascal’s principle, one has to see whether

PV (B P1) > PV (- BP1) (let us call this P2)

Suppose you calculate (very complex) and  accept P2.

then you have to see whether

PV (Bp2 > PV (- BP2) (let us call this P3)

and so on, with an infinite regress. At one point, if you want to avoid the regress, one has to simply believe for epistemic, not practical reasons. Of course the argument does nothing against the view that belief can be both for practical and  epistemic reasons, but if the thesis that one can believe at will is sound, it must say that the practical reason can override  the epistemic ones. And if this is so, regresses as the above always threaten.

               There may be problems with such arguments. But I shall leave them aside, for even if they do not work, there is a problem with the idea of reasons for belief which would fall short of epistemic considerations. The problem is simply that it is very  difficult, to say the least, to envisage that we could rationally perform an action in the absence of epistemic considerations about their outcome, and about whether we believe that we have the means to achieve this outcome. This fact is almost trivial and contained in the definition of acting for a reason: in order for an action to be intentional the agent must not only desire that P, but also believe, for good reasons, that his doing A will bring it about that P, and that he has the means to do A. So some true beliefs must be involved in the conditions of performing any action at all. And this is where this fact connects up with the question of the possibility of forming beliefs at will: if I were to form a belief at will “whimsically” for no reason at all, I would have no epistemic reason to form it, no belief about whether the belief is true or not. So it could not be a rational action. We encounter here again the same consideration which lead us to espouse the incoherence argument. The idea is that I could not form a believe at will, irrespective of whether I think it true or false. But it is hardly the case in many alleged voluntary beliefs formations : I intend to believe that P because I have some reasons to think P true, or sometimes when I think P false of unjustified. For the first kind of case, suppose that a scientist has no sufficient grounds to believe a scientific hypothesis – for instance that there is a “memory” of molecules in water-  because, say, it is unconfirmed, but that this scientist nevertheless would like to believe that hypothesis. He certainly would like to believe it for epistemic reasons (whether he wants it to be confirmed, or that it could cohere with other beliefs that he has); so his willing it would hardly fall short of epistemic reasons. For the second kind of case, take a self deceived person who thinks that P is true, but who would nevertheless want to believe that not P because he desires to believe it. He has this desire because he believes that P is true, and he intend to believe that P is false. Again, he does not form the latter belief in the absence of considerations for thinking it true or false. So epistemic reasons have to enter into his behaviour. This reinforces the point that I wanted to convey in the incoherence argument: even if we grant that such an act of voluntary belief formation could happen psychologically although it would be an irrational belief formation, we have to admit that epistemic consideration should come into the picture. And if such epistemic considerations enter the picture, they can hardly come in without their normative impact, that it in the absence of consideration about the preconditions for belief (1)-(5).

              If this is correct, then, the strong volitionist thesis, direct or indirect, is defeated. But it does not prevent us from holding a weaker form of volitionism, where belief formation can occur in the presence of epistemic normative constraints. What was wrong in the volitionist picture is the idea that we could get rid of all epistemic reasons for belief. If we reject this, it is still open to us to defend a conjunction of weak volitionism and of weak voluntarism.

              What kind of form could such a combination of theses be? On the issue of voluntarism, it would have to determine, first, whether there can be good normative reasons – whether it is rational – to favour the practical value of our beliefs over their epistemic value. If such arguments as those of Zemach given above are sound, we should rule out the rationality of such a normative recommendation. But as I said above, even such voluntarists as James, in his polemic against Clifford, seems to reject this strong normative thesis. It is in the spirit of pragmatism, whether it is James’ or Peirce’s form, that the search after truth, even when it is conceived as a search for utility, is a search for epistemic utility: we should maximise our true beliefs not simply because they are useful simpliciter, for prudential or practical reasons alone, but because they are epistemicall useful. It involves recognizing that there is an epistemic value or worth, which is a good in itself. On the recognition of this good, James need not be in disagreement with Clifford, when the latter place a sort of moral standing on the search after disinterested truth. Where he disagrees is not in the aim, truth, but in the means. And he allows that the means might involve the voluntary formation of beliefs that we judge epistemically useful. It should be reminded, in this respect, that our beliefs do not come separately, one by one. They come within the curse of inquiry. Inquiry is a long term process of formation of beliefs[4]. And when we keep in mind that there are both epistemic conditions for the formation of beliefs as actions, and epistemic aims at the other end, we need no deny that inquiry is a matter of taking active steps towards the getting of beliefs which should satisfy these aims.*

 

REFERENCES

 

Anscombe, G.E.M. 1958 Intention, Oxford: Blackwell

Bennett , J. 1990 “Why is Belief involuntary”?” , Analysis, 100, 87-107

Bratman, M. 1993 “Practical Reasoning and Acceptance in a Context”, Mind,

102, 1-15

Clifford, W.K.1879 “The Ethics of Belief”, in Lectures and other essays, vol.2,

London: Macmillan

Cook, J. (1987) “Deciding to Believe Without Self Deception”, Journal of Philosophy,

LXXXIV, 8, 441-446

Dretske, F. 1988, Explaining Behaviour, Cambridge Mass: MIT Press

Elster , J.1979 Ulysses and the Sirens, Cambridge: Cambridge University Press

Engel , P. 1998 “Believing, Accepting and Holding true”,  Philosophical Explorations

                         2, 140-151

                 to appear  ed. Believing and Accepting , Kluwer

Frankish  K. 1999“Deciding to Believe Again” this volume

Gilovitch, M. 1991, How We Know What isn’t So, New York: Mcmillan

Helm, P. 1994, Belief Policies,Cambridge, Cambridge University Press

James, W.  1897 “The Will to Believe” in The Will to Believe and Other Essays,

Kahnemann D., Slovic, P. and Tversky, A.eds 1982 Judgment under Uncertainty:

Heuristics and biases: Cambridge: Cambridge University Press

Kavka, G. 1983, “The Toxin Puzzle”, Analysis, 43, 33-36

Losonsky  M. to appear “Wanting to believe”  in Engel to appear

O’Saughnessy, B. 1980 b The Will: A Dual Aspect Theory, vol.1 Cambridge,

Cambridge University Press

Pascal , B.  Pensées , in Œuvres Complètes, ed. Lafuma, Paris: Seuil

Pojman, L. 1985  “Willing to believe”, Canadian journal of philosophy, 15, 1 37-55

Searle , J.1983 Intentionality, Cambridge : Cambridge University Press

Smith  M. 1994, The Moral Problem , Oxford: Blackwell

Velleman, D. 1989, Practical Reflection, Princeton: Princteon University Press

Williams,B.  1973 (1970) “Deciding to Believe” in Problems of the self, Cambridge

Winters,  B. 1979 “Believing at will”, Journal of Philosophy , LXXVI, 243-256

Zemach, E.  1997 Practical reasons for belief?” Nous, , 31: 4, 525-27

 



[1] I am thinking of the issue, internal to epistemology, of the choice between a coherentist or a fundationalist conception of justification, and of the possible combination of the insights contained in both.

[2]  I owe this reference to Paolo Leonardi.

[3] I simplify, but it does  not matter here.

[4]  The notion of a belief policy, in Paul Helm’s sense (Helm 1994) for instance, would be quite relevant to what I mean here: : “A belief policy is a strategy or project or program for accepting or rejecting or suspending judgement as to the truth of propositions in accordance with a set of evidential norms”. Keith Frankish ( this volume) , defends a version of weak volitionism which has many affinities with the one proposed here. But I would not myself characterise my form of weak volitionism as a “direct activism”. Belief policies would rather fall within the range of indirect volitionism.

 

* A version of this paper was read at the conference on Belief and the Will in Nijmegen in May 1999. I thank for their comments and suggestions Anthonie Meijers, Keith Frankish, Jan Bransen, and the participants in that very stimulating meeting. Older versions were given at the conference of the SIFA in Bologna in september 1998, and at a Franco-Italian meeting in Paris in November 1998.  The present version was read in the Departement of Methodology and Philosophy of Science at the University of Athens in November 1999, and at the University of Paris-Sorbonne . I thank for their comments Stelios Virvidakis, Stathis Psillos,  Myrtho Dragonna, , Petros Polyménis, Ray Callender, Jonathan Dancy, Paolo Leonardi, Eva Picardi,  Marco Santambrogio,  Jérôme Dokic, Michel Puech, Stéphane Lemaire, Ivahn Smadja.