Belief is involuntary

L. Jonathan Cohen defines belief that p briefly as a disposition to feel it true that p.Footnote ² While at first glance it might seem a bit odd to explicate belief as a truth-related feeling, the characterization is often taken to capture something that defines believing. For example, Alvin Plantinga speaks about propositions believed or known as having certain felt attractiveness, rightness, or appropriateness in them. There is a sort of perceived correctness in propositions one believes as opposed to propositions one does not believe. Plantinga holds that this phenomenology is familiar to us all though hard to describe.Footnote ³

Plantinga appears to refer to the feeling that belief purportedly involves as if it were a universal psychological fact. However, it might be queried whether any empirical research has been done on the subject. Some might deny having any feelings attached to their occurrent beliefs even after careful introspection. But given that there are such persons and that they do nonetheless have beliefs, the view of belief as including some kind of feeling is open to dispute. Perhaps there is not much more essential to the phenomenology of belief besides the aiming at truth, while some persons may also experience certain feelings when they believe.Footnote ⁴

One central issue discussed regarding belief is whether it is voluntary. Direct doxastic voluntarism is the thesis that believing at will is possible in full awareness. According to this view, one can consciously will to believe that p and hence immediately acquire the belief that p.Footnote ⁵ But if belief involves a feeling, as some think they do, they seem to not be under direct control of the will, since feelings are not usually voluntary. On the face of it, this appears to be William Alston's claim. He holds that the feeling beliefs include is immediate and something one experiences rather than thinks out, and in his view this is connected with belief's involuntary character.Footnote ⁶ However, this line of argument loses much of its plausibility, if it is denied that belief essentially involves a feeling.

On the other hand, Alston notes that by speaking of feeling in connection with belief he is not suggesting that it is some kind of sensation or emotional state. Instead, he takes ‘feel’ to be a broad term that encompasses a great variety of experiential states. But then Alston's simple argument against direct doxastic voluntarism can be reinstated without the premise that belief involves a feeling in any typical sense of the term. Thus, it might be claimed that the experience of belief is, to use Alston's expression, ‘a matter of one's being struck by (a sense of) how things are rather than deciding how things are’.Footnote ⁷ This, if accurate, seems to support the view that belief is involuntary. Either way, Alston-type arguments point to the conclusion that believing directly at will is, to utilize Alston's terms again, ‘psychologically impossible’.Footnote ⁸ That is, it is an ability we do not in fact have. This leaves open the possibility that if our mental structure were different, believing at will could be feasible.

Some have, however, defended the stronger thesis that volitional believing is logically or conceptually impossible: it is an ability we could not even possibly have. Bernard Williams has presented an influential argument for this conclusion. The argument can be put as follows. Beliefs aim at truth, that is, to believe that p is to hold that p is true. But willing to believe that p can never directly make p true. So, if a person could believe that p at will, she would recognize that she believes that p irrespective of p's truth. But this is inconsistent with belief's truth-oriented nature. Hence, believing at will involves a conceptual inconsistency.Footnote ⁹

For reasons parallel to that of Williams, Louis Pojman has also argued that volitional believing implies a conceptual confusion. As a noteworthy corrective to this view he argues that because the possibility of forming a belief at will in ‘less than full consciousness’ cannot be altogether excluded, believing at will is incoherent only in full awareness. Pojman maintains that to say ‘I believe that p, but I believe it only because I want to believe it’ has the same incoherence attached to it as Moore's paradoxical ‘I believe that p but it is false that p’.Footnote ¹⁰ According to Pojman, consciously acquiring or sustaining a belief by volition is not possible.Footnote ¹¹

I find Williams's and Pojman's reasoning compelling, and we will return to similar considerations later. It is appropriate to note that Williams-type ‘logical arguments’ against volitional believing do not presuppose the possibly debatable view that belief is a feeling-like state. Instead, the arguments merely require that belief aims at truth. This assumption suffices for my purposes. Furthermore, note that even if direct control over beliefs is not possible, our decisions can indirectly have an effect on what we believe, for example, by taking inquiries into propositions' truth.Footnote ¹²

Belief and evidence

Since beliefs aim at truth, they presumably require a close link to evidence which supports the truth of the proposition believed. It seems plausible to maintain that a person would not find a proposition true unless she thought she had truth-related reasons for finding it true. In other words, a person would not apparently believe a proposition unless she thought she had evidence to support the truth of the proposition believed. In a somewhat similar way, Richard Swinburne holds that a person could not believe a proposition while acknowledging that all the public evidence rendered it improbable and she had no other evidence.Footnote ¹³ Consequently, the preliminary claim is that evidence is a necessary condition of belief. It is perhaps worthy of separate notice here that for similar reasons as those why belief is not voluntary, what a person sees as evidence for a given proposition's truth is not a voluntary decision either.Footnote ¹⁴

We can here take ‘evidence’ as a person-relative and broad notion: whatever from a believer's viewpoint intuitively seems to support the truth of p is her adequate evidence for p. Thus, evidence can among other things consist of arguments, memories, first-hand experiences, and reports by others. For example, a person may believe that she ate dinner on the evidential basis of having memories of it, just as she may believe that fairies exist because she has in her view a convincing argument for that conclusion. This liberal notion of evidence should not bother us too much, since here it is not important what adequate evidence ipso facto is. We are interested in the distinction between having purportedly adequate evidence and not having evidence. The notion of evidence just introduced should enable this division: ordinarily a person thinks that some propositions have adequate evidential support while others lack it.Footnote ¹⁵

On the supposition that there is a close link between belief and evidence, it seems perplexing to assert that:

1. (1) I believe that p, but I only have evidence for not-p.

Similarly confusing appears to be the more modest statement:

1. (2) I believe that p, but my evidence for p is undecidable.

(1) is a case of believing against evidence, and (2) amounts to believing in the face of unsettled evidence.Footnote ¹⁶ The problem in (1) and (2), as stated so far, is just this: given the preliminary claim that a necessary condition for belief that p is evidence for p's truth, it follows that believing while not having evidence is problematic.

So, there appears to be something obscure in (1) and (2). However, this is a claim John Bishop has disputed. While he thinks that (1) might be possible,Footnote ¹⁷ he focuses on cases of the sort in (2) and holds that at least they can have instantiations. To be precise, Bishop argues that a person can believe that p while she recognizes, in accordance with the relevant norms she accepts as applicable, that she lacks adequate evidence for p's truth.Footnote ¹⁸ Thus, the believing takes place in full consciousness (awareness, recognition): the believer is in some obvious sense aware that she believes and does so without adequate evidence. I take this awareness to be implied by the first-person point of view of statements (1) and (2).

Bishop argues for the possibility of (2) in connection with his view on religious faith. He is thus mainly interested in showing that theistic propositions can be believed while evidence for them is ambiguous, that is, while the evidence is consistent with both the truth and the falsity of the propositions in question.Footnote ¹⁹ However, if Bishop's considerations are correct with respect to religious beliefs, which he regards as claims about mind-independent realities, they must also apply to other factual beliefs, as Bishop concedes. Accordingly, his view is here regarded as concerning beliefs about facts, in general.Footnote ²⁰

Passionally caused beliefs

Bishop defends the possibility of consciously believing without adequate evidence by drawing our attention to causes that give rise to beliefs. In line with the reasoning of the previous section, he admits that a typical cause of belief-attitude is deliberate or undeliberate consideration of evidence.Footnote ²¹ But this, Bishop holds, is not the only way to form beliefs. He argues that there is also what William James labels as passional causes and which Bishop prefers to call non-evidential causes. I will here use ‘passional’ and ‘non-evidential’ equivalently. In Bishop's view the term ‘passional’ includes all kinds of causes that do not provide the believer with ‘grounds (relative to the assumed correct evidential practice) for holding the propositions believed to be true’.Footnote ²² These causes Bishop sees as vindicating the possibility of consciously believing in a situation of evidential deficiency.

Bishop argues that a wide variety of passional causes need to be acknowledged, such as emotions, wishes, desires, positive evaluations, affections, and affiliations. While he admits that lack of evidence does tend to weaken any inclination there is to believe, he holds that this tendency is not ‘a universal psychological law’. Instead, he argues that sufficiently strong passional causes can sustain belief even though the believer recognizes that the truth of the proposition believed lacks adequate evidential support.Footnote ²³ I assume that in such cases the believer comprehends that her belief is passionally caused and sustained; the believer can say that passional factors are the reason why she believes as she does. Bishop holds, for example, that through being moved by encounter with a theistic religion a person may be passionally caused to believe that God exists, and this belief can be sustained even if the believer comes to realize that the existence of God is not established on the evidence. Consequently, for Bishop (2) is not problematic as long as belief that p is sustained by a passional cause.Footnote ²⁴

In his article concerning Bishop's early views on passionally caused beliefs and related topics, Andrei Buckareff has raised some doubts against the possibility of believing by passional causes.Footnote ²⁵ Though Bishop has replied to Buckareff's criticism,Footnote ²⁶ I believe that Buckareff's fundamental assertions are still sound. My critique is thus motivated by his treatment of the issue, and hence it is appropriate to start with a brief recapitulation of Buckareff's main points against Bishop's view.

Using reasoning similar to that offered in the previous section, Buckareff argues that belief aims at truth and this means, in part, that ‘belief that p is acquired and maintained in response to evidence for the truth of p’.Footnote ²⁷ Buckareff holds that the truth-directedness of belief counts against the conceptual possibility of consciously acquiring and sustaining a belief by sheer passional causes. He argues as follows. Suppose p's truth is not supported nor excluded by one's evidence. If in this situation one could come to believe that p either by a decision to believe or by a passional cause alone, then one could do so with no regard for the truth of p or, as I would say, with no regard for what seems to one to be the truth of p. But then a person could consciously believe that p while at the same time failing to believe that p because her evidence does not support the truth of p. Buckareff holds that we are here ‘somewhere in the vicinity of Moore's paradox’.Footnote ²⁸

Buckareff's reasoning bears some resemblance to Williams's criticism against the possibility of believing at will, but what exactly is Buckareff claiming? One of his contentions appears to be that believing on insufficient evidence would yield inconsistent beliefs, and sustaining such beliefs is logically puzzling. With respect to (2) this means something like the following. If my evidence for p is undecidable, then it seems that I don't deem p true, and so I actually lack the belief that p, that is, I don't believe that p (and I don't believe that not-p). But then by (2) I am claiming both to believe that p and not-believe that p, and this is inconsistent.Footnote ²⁹

I am inclined to think, like Buckareff seems to do, that consciously believing on insufficient evidence would yield inconsistent beliefs, at least after reflection, and that sustaining such beliefs in full awareness is conceptually problematic.Footnote ³⁰ However, I don't think that having inconsistent beliefs is the most fundamental problem in non-evidential believing.Footnote ³¹ Instead, in my view the central problem is intrinsic to the notion of non-evidential believing. The problem arises solely from the truth-directed nature of belief: given that to believe that p is to deem p true, then consciously believing that p for non-truth-related, passional reasons is conceptually troubling. By all appearances, something like this seems to be Buckareff's claim, too. I will try to vindicate this impression in the next section.

In addition to the above-mentioned criticism, Buckareff offers some further remarks about the subject. In his view non-epistemic reasons may in a certain way explain why a person believes as she does. Passional causes can, for example, make the evidence a person has more vivid or salient than it actually is. But this, Buckareff says, does not mean that from the person's perspective she will esteem her belief as being caused and sustained by non-epistemic reasons. Nor does it mean, he continues, that if beliefs could have passional causes they could be sustained upon introspection.Footnote ³²

As a response to Buckareff, Bishop maintains that it is indeed impossible intentionally and directly to acquire a belief one sees as lacking appropriate evidential support. That is, Bishop rejects direct doxastic voluntarism. But, he maintains, passional believing does not involve such self-inducing of belief from some non-epistemic motive. According to Bishop, the fact that one cannot directly decide to believe from passional motives does not entail that beliefs cannot have passional causes.Footnote ³³ Bishop offers some additional reasoning to support his view, too, and it will be presented in due course. However, I doubt whether it can rebut Buckareff's challenge.

Conscious and unconscious beliefs

Presumably beliefs can be acquired consciously as well as unconsciously. If I am unsure whether it is the case that p, I check it from a source I take to be reliable and belief follows. This is conscious belief-acquisition. Then again, some beliefs are acquired plausibly rather unconsciously. For example, on my way to work I didn't have to contemplate the weather in order to gain the belief that it is dull. Perhaps some beliefs can also be sustained in less than full awareness. While my current belief about the weather is something I have now reflected upon, I usually hold and apply such beliefs without, say, carefully thinking about their origins or evidential status or how they cohere with my other beliefs. Many beliefs we hold may be of this type.

What is the place of passionally caused beliefs in the above scheme? To begin with, it is not obvious that acquiring a belief by a passional cause could occur consciously. Note that we are not talking about intentionally deciding to believe because of a passional motive. That would amount to direct doxastic voluntarism. Instead, we are interested in whether conscious reflection on passional factors could cause a person to believe in the same way as considering evidence does. To me this seems to be the scenario Buckareff was picturing in his argument against passional believing. So, in spite of what Bishop appears to think, Buckareff is not criticizing him of endorsing direct doxastic voluntarism. But then Buckareff's critique cannot be refuted just by pointing out that passional believing is not a case of acquiring beliefs directly at will. We are still somewhere in the vicinity of Moore's paradox regarding passionally caused beliefs. What is more, it seems psychologically odd that thinking about one's passional reasons for believing a certain proposition could alone cause one to believe that proposition.

Bishop remarks that ‘to deny that beliefs may sometimes have passional causes is … purely a rationalist dogma’.Footnote ³⁴ This might still be partly true, since persons may have acquired beliefs by non-evidential causes in less than full consciousness. Perhaps such beliefs can also endure if they are never properly reflected upon. I presume that this sort of passional believing is possible. Maybe it is a rather common phenomenon, too. But despite this, I doubt that passionally caused beliefs could be sustained when their origin is realized. Consider the following argument for this conclusion.

Suppose that my evidence for p is inconclusive and that I have in less than full consciousness acquired belief that p by a passional cause. Suppose then that I happen to reflect consciously this passionally caused belief. But now I see that I believe irrespective of what seems to me to be the truth of p, since I acknowledge that passional causes do not indicate that p is true and that my evidence for p's truth is inconclusive. But believing irrespective of what seems to be the case is inconsistent with belief's truth-oriented nature, and this is something I also realize. Thus, I do not and cannot believe that p anymore. So, consciously to sustain passionally caused belief is not only psychologically peculiar, but it involves a conceptual inconsistency. Here, too, the argument offered bears resemblance to that of Williams and Pojman against direct doxastic voluntarism.

In one connection Bishop claims that to believe consciously that p is to find oneself with the attitude towards p that it is true, but to find oneself with this attitude is only typically but not necessarily to find p's truth evidentially supported. ‘To believe is, indeed, to believe true’, he holds, but ‘it is not necessarily to believe evident’.Footnote ³⁵ Bishop notes that this distinction is in conflict with Jonathan Adler's assertion that it is intrinsic to the concept of belief that one cannot believe a proposition whose truth one recognizes as evidentially insufficiently supported.Footnote ³⁶ If, following Bishop, it is denied that belief as a truth-directed attitude is necessarily tied to evidence, can there after all be room for non-evidential believing?

I have granted that believing by non-evidential causes may occur in less than full consciousness. So, at least in such cases the link between belief and evidence is not a necessary one. However, I am inclined to think otherwise with regard to conscious belief. In this case it seems that believing is unavoidably tied to evidence. In fact, there does not appear to be any other option. As argued, consciously believing irrespective of what seems to be the case is inconsistent with belief's truth-oriented nature. But this is just to say, in other words, that consciously believing irrespective of truth-indicating evidence is inconsistent with the view that to believe is to believe true. Accordingly, it seems that consciously only ‘evidential believing true’ is possible.

To move into a different point, I have implied that the arguments offered against passional, non-evidential believing are similar to Williams's and Pojman's logical argument against directly acquiring or sustaining a belief at will. This likeness is not a coincidence. The crucial consideration here is that volition arguably is one possible non-evidential cause of belief (and it does not seem to differ qualitatively from any other non-evidential cause). But then there seems to be no reason why logical arguments against direct doxastic voluntarism would not apply, mutatis mutandis, to any type of non-evidential believing. Is not this actually the procedure put into practice in this paper? So, if conscious believing at will is conceptually incoherent, so is any type of conscious non-evidential believing. Incidentally, Pojman appears to make this inference when he presents his argument against direct doxastic voluntarism. On one occasion he concludes that ‘one cannot believe in full consciousness “that p and I believe that p for other than truth [that is, evidential] considerations” ’.Footnote ³⁷ This is my conclusion, too.

Regarding the aforesaid point it is appropriate to recognize, conversely, that if someone argues that conscious believing by non-evidential factors is possible, then she would have to admit that it is also possible consciously to believe by sufficiently strong volition, since volition appears to be a non-evidential cause of belief. Consequently, it seems that the idea of consciously believing by non-evidential causes implies the disputed thesis of direct doxastic voluntarism or something close to it, at the very least. With respect to this, Bishop is actually not consistent: although he argues for the possibility of passional believing, he denies the possibility of believing at will by referring to Williams's argument.Footnote ³⁸

The main conclusion of this article may go against our everyday experiences. It may certainly look as if many people fairly consciously believe by non-evidential causes. However, such counterexamples must be merely apparent, if non-evidential believing is not possible. Indeed, in my view supposed counterexamples can be explained away without reference to the thesis of believing by non-evidential causes. To conclude, I will offer a few considerations of how such explanations might go.

First of all, a person may obviously claim to believe that p by non-evidential causes while she really does not believe that p. The person may be confused about her beliefs, and this may amount to some kind of self-deception. The person may also have incorrectly identified her attitude as belief when it is more aptly described, say, as hope or as acceptance, which are distinct attitudes from belief.Footnote ³⁹ In common parlance there may be a tendency to use the term ‘belief’ to refer to various propositional attitudes that actually should be distinguished from belief and each other.

On the other hand, a person may seem to believe passionally because she believes for reasons some of us would deem passional, but which the believer herself sincerely sees as evidential. Bishop's early example of passional believing is a bereaved son, whose emotions cause him to believe that his dead father is still somehow with him.Footnote ⁴⁰ If this means that the son has some kind of feeling of his father's presence and sees this as evidence for his actual presence, then the son may believe, although some of us would probably think that the son believes merely because of a passional reason. Then again, if the son merely wishes that his father were still with him, it is hard to see how this wish alone could cause and sustain corresponding belief. While reasons of the strangest sort can perhaps lead to the acquisition and sustenance of persons's beliefs, when reflected they must from the believer's viewpoint indicate the truth of the proposition believed. Otherwise the person would not believe.