Explaining Normative Import
Explaining Normative Import
Abstract and Keywords
This chapter discusses how to explain the relation between ascriptions of belief and intention, and statements ascribing the corresponding implication or means-end commitments. Dispositionalism about propositional attitudes must explain the link in terms of principles connecting a non-normative attitude ascription with a normative commitment. It is argued that such explanations are unsatisfactory because they cannot do justice to the peculiar character of our relation to our own current intentions and beliefs. Under dispositionalism, it would be possible to adopt a purely contemplative attitude to our current intentions and beliefs just as it is possible to adopt such an attitude to our character traits. Such an attitude to our current beliefs and intentions is not intelligible and we can make better sense of how we stand in relation to those beliefs and intentions if we assume that ascriptions of beliefs and intentions are implicitly normative.
1. The way ahead
Concepts of belief and intention are concepts like believing that summer is coming to an end and intending to learn more about European history. In this and the following chapter I argue and defend the view that such concepts are normative concepts. The claim that I believe P brings into play the concept of believing P and on that account implies that I have incurred a commitment to believing any implication of P on which I give a verdict. Similarly, the claim that I intend to Φ brings into play the concept of intending, and on that account implies that I have incurred a commitment to doing whatever is necessary for me to Φ. This view is controversial on at least the following two counts. First, it is not widely accepted that the concepts of believing and intending are normative; if my view is to emerge as being plausible and interesting, I need to explain how it can be that they are. That requires me to explain how the psychological dimension to believing and intending relates to the normative dimension and why acknowledging both dimensions is important for personal understanding. I address these matters later in this chapter when I introduce the notion of a psychological commitment, and then further in Chapter 7. Second, some of those who accept that the concepts of belief and intention are normative may deny that believing and intending implicate commitments in the sense introduced in Chapter 3. One might think that those concepts are normative simply because rationality is a normative concept and because believing and intending are inextricably linked to rationality. Those who think that the concept of rationality is normative are not bound to suppose that its normativity should be explicated in terms of normative commitments. The central role that I accord to (p.101) normative commitments should turn out to be plausible and to do some work.
In framing the issues, I shall assume a certain view about the character of concepts of belief and intention. I take these concepts to apply to people. If they apply to things other than people these things will be organized systems, like animals. When we ascribe a belief to a person, we apply a concept of belief to that person and ascribe the corresponding belief‐property.59 Thus, in applying the concept of believing that summer is coming to an end we ascribe the belief‐property believes that summer is coming to an end. A parallel claim holds for concepts of intention and corresponding intention‐properties. Talk of beliefs or intentions as psychological states is shorthand for talk about the state of someone's believing something or of someone's intending something. In other words, it is about belief‐properties and intention‐properties.60 It is not about properties possessed by structures in the brain, nor is it about particulars located at some place in the brain.61
Given the distinction between concepts and properties, it is clear that distinct concepts can pick out the same property. Thus, to take a minimally controversial example, I can pick out the property of being water as I have just done by deploying the concept of being water. But if I had just written the word ‘water’ on my whiteboard I could also pick out the property as that which is referred to by the concept expressed by the word I have just written on my whiteboard. A more interesting, and also more controversial, example of distinct concepts picking out the same property is provided by the concepts of being water and of being H2O. (The example is used in precisely this connection in Gibbard 2002, though the thinking has its routes in Putnam's 1975a theory of natural kind terms.) The idea is that the property of being water is identical with the property of being H2O, yet the concept of being water and the concept of being H2O are distinct. Someone might have the concept of water and lack the concept of H2O, knowing nothing about hydrogen, oxygen, and (p.102) their combination to form molecules of H2O. Irrespective of whether this approach to the concept of being water and its corresponding property is right, it is at least intelligible that it should be thought that specifying the property in terms of the concept of H2O brings out the nature of water in a way that specifying the property in terms of the concept of being water does not. Pointing at water, we can reasonably ask, ‘What is this substance?’ When we are told that it is H2O, we seem to learn about its nature—about what any substance would have to be if it were to be this substance.
Even if this is the right way to think about some concepts and their corresponding properties, it is certainly not the right way to think of all of them. The concept of a teacher is the concept of an individual who occupies a specified role. It is certainly possible to pick out this property without using the concept of a teacher. I can run through the kind of exercise illustrated above in connection with the concept of water—I write the word ‘teacher’ on the board and pick out the property via reference to the word. But there is no underlying nature to being a teacher—a nature that is not specified by the ordinary role‐specifying concept of a teacher. There can be all kinds of wisdom about what it takes to be good teacher—wisdom that a person with a grasp of the concept of a teacher is not bound to have. But if so, that is wisdom about how to fill the role well, given the duties and responsibilities that define it. It is not about an underlying nature common to all teachers.
The stance taken here is that concepts of belief and intention are more like the concept of a teacher than they are like the concept of being water, understood on the model sketched above. There is no more to believing that summer has come to an end or to intending to learn more about European history than is specified by, respectively, the concept of believing that summer is coming to an end and the concept of intending to learn more about European history. So we should not expect there to be true identity statements specifying underlying natures for the beliefs and intentions to which these concepts apply, which are not captured by the concepts.
The discussion in this and the following chapter is an attempt to expand on what is implied by ascriptions of beliefs of the form, ‘A believes that p’ and ascriptions of intentions of the form, ‘A intends that he/she Φs’. These are ascriptions that attribute certain (p.103) propositional attitudes, specified in terms of their category (belief or intention) and their content. In the next section I discuss a view according to which concepts of propositional attitudes are non‐normative psychological concepts. I shall then consider how, in terms of this view, we might attempt to explain the normative import of ascriptions of beliefs and intentions.
Each of us has a vast range of beliefs, desires, intentions, hopes, fears, and so on, yet only a very few of those are, as we say, before our minds at any given time. While focusing on some task in hand, we need not think of other things we plan to do. At any given time, much that we believe lies dormant, having no immediate effect on what we are thinking or doing. None the less, our various propositional attitudes have the potential to impact on thought, desire, action, or feeling under appropriate circumstances. This suggests that we might think of having a propositional attitude as a matter of having various dispositions, for instance to form beliefs, to perform actions, or to acquire feelings and desires. (Here as elsewhere I shall concentrate on belief and action.) I have a standing belief that there is a farmers' market in town on the second Saturday of each month. To have this belief might involve my having a disposition such that, for instance, were I set on buying the sorts of things I generally buy at this market I would be liable to go there. Of course, that could hardly be the only disposition associated with the belief. My having this belief has the potential to manifest itself in many different ways. For instance, were I to be asked when the farmers' market is held I would be liable to answer by saying something to the effect that it is held on the second Saturday of every month.
According to the theory I consider in this chapter, propositional attitudes, or at least those propositonal attitudes that are standing states, are essentially dispositional. The dispositions are taken to be akin to the paradigm cases of dispositions in philosophical literature. These include dispositions such as solubility in water, flexibility, and fragility. Roughly speaking, a thing has the disposition of flexibility provided that, if it were put under suitable pressure, it would bend. (p.104) A widely held view of such dispositions is that they are higher‐order properties or states. On this view, a thing has the disposition of flexibility if it has the property of possessing some property that accounts for its bending if put under suitable pressure. That property—the lower‐order property—is the ground of the disposition.
Suppose it occurs to me at a certain moment that I should be at a meeting. If that is an attitude it seems, even so, to be an occurrence and so, one might think, is not something dispositional. Whether it should be regarded as a distinct attitude, rather than a recollection of something one believes, is a moot point. In any case I shall bypass problems like this. The view I am about to describe may be understood to be a view about propositional attitudes that are standing states, even though they may be short‐lived. The attitudes that are my main concern—beliefs and intentions—are naturally understood as being states of this sort. Acquiring a belief and abandoning a belief are events, but that is no reason to think that the belief in question is anything other than a standing state.
What I shall call dispositionalism about propositional attitudes is the view that the attitudes (at least those conceived as standing states) are dispositional and that our concepts of the attitudes are dispositional concepts. In addition, the relevant dispositions are conceived as being analogous to the paradigm cases of dispositions, such as flexibility, fragility, and the like.
There are notorious problems about trying to specify dispositions relevant to the characterization of propositional attitudes in terms of conditionals stating what would happen if a certain condition were fulfilled. People who believe conditionals and come to believe their antecedents are not bound to come to believe their consequents. Whether they do so depends on a wide range of factors, such as whether or not they are interested in the truth or falsity of the consequent, and whether the play of thoughts running through their minds is such as to prompt belief in the consequent. It seems a hopeless task to specify the relevant factors except in the vaguest terms. (Some attempts in this direction are made in Millar 1991: chapter 2.) And even when this is done, a realistic formulation of the antecedent of the conditional will almost certainly have to include a hedging clause to the effect that there should be no countervailing conditions. (Problems about laws containing hedged clauses are (p.105) explored in Schiffer 1991. Fodor 1991 is a response to this.) In what follows I shall cut through these issues by specifying dispositions in terms of what subjects are liable to do under a certain condition. A subject is liable to Φ in condition C if, when C obtains, he or she is liable to Φ. The idea is that being in C raises the chance of the subject's being liable to Φ. The cases that interest us are ones in which, for those who believe the relevant conditionals, it would be no surprise if a subject in C were to Φ. This may be so even though it is not the case that in C it is likely that the subject will Φ. In some cases the strongest claim that could justifiably be made is that in C the subject might well Φ. I choose this approach not just because it is conveniently rough‐and‐ready. I suspect that it is not in general possible to provide realistic specifications of dispositions associated with propositional attitudes in more precise terms and that including hedging clauses gives an illusion of definiteness.
I turn now to a somewhat more precise characterization of dispositionalism. Propositional attitudes are individuated by content and category. Thus, the belief that Stirling has a castle and the belief that Edinburgh has a castle are different attitudes because they have different contents, though they belong to the same category—belief. A belief that it will be cold on Tuesday and a desire that it will be cold on Tuesday are different attitudes because they belong to different categories—belief in one case and desire in the other. I shall say that a set of dispositions is characteristic of an attitude, individuated in the way just explained, provided that one would have that attitude if and only if one had those dispositions. With these notions to hand, I can be more precise about what dispositionalism is. I take it to be the conjunction of the following claims:
(A) For any propositional attitude (individuated as above) there is a set of dispositions that are characteristic of that attitude.
(B) For any propositional attitude (individuated as above) to have that attitude is to have the dispositions in the set characteristic of that attitude.
(C) Dispositions are conceived by analogy with the philosopher's paradigm cases, for instance flexibility and fragility.
(A) tells us nothing about either the metaphysics of propositional attitudes or the concepts of propositional attitudes. It merely states (p.106) that propositional attitudes are paired with sets of characteristic dispositions. (B) gives us a metaphysics. It tells us that possessing, for instance, a belief‐property like believing that Stirling has a castle consists in having the appropriate dispositions. In keeping with my earlier remarks about propositional attitude concepts and properties, I take it that to possess the property believing that Stirling has a castle is nothing other than to fall under the concept believing that Stirling has a castle. So (B) reflects a view about concepts of attitudes: the idea is that, when we think of an attitude as the attitude of believing that Stirling has a castle, we do so in dispositional terms. In more general terms, for any propositional attitude Ξ, the concept of Ξing that p is just the concept of possessing such‐and‐such dispositions—those that are characteristic of Ξing that p. There is an issue as to how a concept of an attitude is supposed to represent the dispositions characteristic of the attitude. I address this in Section 6.
(C) is crucial. The view that I shall eventually defend is compatible with (A) and (B) provided that the notion of a disposition is broadly conceived. It differs from the dispositionalism characterized above in specifying the dispositions characteristic of propositional attitudes in terms that are normatively loaded. (C) highlights the fact that the dispositionalism presently under discussion treats dispositions in the standard non‐normative way.
It is important that dispositionalism incorporates a view about concepts of attitudes because, so understood, it yields an account of how ascriptions of attitudes to agents bear upon expectations about, and explanations of, what these agents will think and do in virtue of having these attitudes. The idea is that we are sometimes in a position to anticipate what people are liable to think and do if they have certain attitudes because,
(i) in virtue of having a grasp of relevant attitude concepts, we associate these attitudes with their characteristic dispositions, and
(ii) we are in a position to anticipate that circumstances likely to lead to the manifestation of some of these dispositions (triggering circumstances) will obtain.
(iii) we are in a position to judge that what the agent has come to believe or do results from the manifestation of dispositions linked to attitudes we ascribe to the agent.
This last condition will be satisfied if we are in a position to judge that the manifestation of relevant dispositions results in the subject's being liable to form the belief or perform the action to be explained, and that the subject's being so liable led to the formation of the belief or the performance of the action.
In Chapter 1, Section 3, I presented a familiar reason for thinking that rationalizing explanations are causal explanations. One might wonder whether this view is compatible with the dispositionalist view of propositional attitudes. If a fragile cup falls off the shelf and breaks on the stone floor, it manifests its disposition to break on being put under suitable stress. But what caused the manifestation, it might be said, is not possession of the disposition—the higher‐order property. Rather, it is possession of the property that is the ground of the disposition. It is this property that figures in the causal process leading from the triggering cause—the cup's being put under suitable stress—to its breaking. The cup's fragility is not a further property with a role in that process. It is simply the property of having some property or other that has such a role. So if one is keen to regard propositional attitudes as states that can figure in causal processes leading to beliefs, actions, or feelings, one might prefer to suppose that to have a propositional attitude is to have the various properties that are the grounds of the relevant dispositions.62 I do not myself think that we need to go down this path. It would be overly fastidious to take the fact, if it is a fact, that dispositions are not causes to be a reason to deny that rationalizing explanation is causal under dispositionalism. If rationalizing explanation is dispositional explanation, then, certainly, it is more like explaining the breaking of the cup in terms of its fragility than it is like explaining the breaking of the cup in terms of its hitting the floor. But the dispositional explanation of the cup's breaking is not a non‐causal explanation. It presupposes that there was some triggering cause, even though it need not specify this cause. It also presupposes that there was a causal process associated with the (p.108) disposition, leading from the triggering cause to the breaking of the cup. The dispositional explanation is therefore a way of telling part of the story of the processes leading to the breaking of the cup. It is open to the dispositionalist about propositional attitudes to think of rationalizing explanation in similar terms. Some may wish to withhold the notion of a cause from dispositions. My point is that, even if there is reason to do so, this would be no objection to treating dispositional explanations as causal explanations, albeit explanations that provide limited insight into the causation of what is to be explained (compare McLaughlin 1995: 123–4).
Dispositionalism seems to me to be the most plausible alternative to the view for which I shall be arguing. But it does face some serious problems when it comes to explaining the normative import of ascriptions of propositional attitudes.
3. Dispositionalism and the explanation of normative import
At least part of the normative import of ascriptions of beliefs and intentions is captured by statements about (normative) commitments incurred through having the belief or intention ascribed. These statements are statements that would be true if the corresponding ascriptions were true. One way of accounting for the normative import of an ascription is in terms of the assumption that the ascription is normative. If the ascription were normative, then statements that capture its normative import would follow from it without further ado. Towards the end of the previous chapter I described a strategy for explaining the normative import of ascriptions of beliefs and intentions to rational agents without invoking the assumption that these ascriptions are normative. I envisage the dispositionalist adopting this strategy. In this and the following two sections, I argue that the strategy is unsuccessful.
To recall: the strategy was to explain the normative import of ascriptions of beliefs and intentions to rational agents by invoking principles connecting beliefs and intentions, conceived in non‐normative terms, with normative commitments. To this end the following two principles might be invoked, where ‘x’ ranges over rational agents: (p.109)
The Means–End Commitment Principle: For any x, ϕ, if x intends to ϕ then x incurs a commitment to doing what is necessary if x is to ϕ.
The Implication Commitment Principle: For any x, π, if x believes π, then for any θ, if π implies θ, then x incurs a commitment to believing θ, if x gives any verdict on θ.
I can best approach the problem I aim to identify by considering an analogous problem in the sphere of ethics. To illustrate this, I draw again on an example used earlier. This was the case in which my neighbour appears at my door in distress and calling on me for help. I regard this as a reason for me to help. Let us suppose that my doing so reflects my acceptance of a principle to the effect that the fact that one's neighbour is in distress and calling on one for help is a reason for one to help. This is a reason‐specifying principle connecting the non‐normative with the normative. Suppose now that I am pressed to say what it is about my neighbour's being in distress and calling on me for help that gives me a reason to help. Although my acceptance of the principle contributes to the explanation of my reaction to the situation, I cannot address this enquiry simply by alluding to the principle. If there is a puzzle about how the consideration about my neighbour constitutes a reason for me to help, there is as much of a puzzle over what makes the connecting principle acceptable.
The query about what makes the consideration in question a reason for me to help need not be motivated by scepticism about practical or moral reason. It crops up in normative moral philosophy, for instance in discussions about what makes killing wrong when it is wrong (see, e.g. Norman 1995). We do not advance this enquiry simply by citing a plausible principle to the effect that the fact that an action would be a killing is, in the absence of special circumstances, a reason against doing it. In keeping with this, the main types of response to the issue are all concerned with providing a deeper explanation of why such a principle holds. Consequentialists look for an answer in terms of the consequences of individual acts of killing for the general good, or in terms of the consequences of there not being widespread acknowledgement of a strong prescription against killing. Theorists inspired by Kant look for an answer in terms of the view that persons are ends in themselves who, as such, are worthy of respect. If we were to follow Scanlon's distinctive kind of contractualism, we would try to show (p.110) that some appropriate covering principle could not reasonably be rejected by anyone motivated by a concern to live with others on terms that could be justified to them provided that they shared such a concern (see Scanlon 1998: chapter 3). All of these strategies explain the wrongness of wrong killing in terms of a failure to have due regard to something that is held to be good in itself or worthy of some special kind of concern. For the consequentialist it is the general good. For Kantians it is persons viewed as ends in themselves. For Scanlon it is the ideal of living with others on terms they could not reasonably reject. (Scanlon 1998: 154 explicitly regards this as an ideal.) In each case the normative import of the consideration that an action would be a killing is traced to a basic assumption about what is valuable or worthy of special concern. The same general strategy would apply to the consideration about my neighbour. The connecting principle in this case is not self‐explanatory. If it is acceptable, its acceptability will be grounded in considerations about what is of value or worthy of regard.
The question now is, How is the dispositionalist to explain why having a belief or an intention incurs a normative commitment of the sort I have considered? Invoking the Implication Commitment Principle and the Means–End Commitment Principle will not do the trick. We need to know why these principles are acceptable. It can easily seem that there is no real issue here. Dispositionalism about belief and intention looks plausible, and the principles in question seem acceptable. My brief discussion of principles in ethics connecting the non‐normative to the non‐normative is intended to dislodge such complacency. It will emerge that there is an issue about whether the principles would be acceptable if the dispositionalist view of belief and intention were correct. To lead into the issue, I first draw attention to a problem for dispositionalism that is often overlooked and which initially might seem to be quite independent of the problem of explaining normative import.
4. How we relate to our current intentions and beliefs
Imagine that I have an intention to take a holiday. According to dispositionalism, this is a matter of having the dispositions that are (p.111) characteristic of that attitude. If the dispositionalist were right, it would be easily intelligible that I should adopt a purely contemplative stance towards this intention. By that I mean a stance in which I simply register that I have the intention, without taking that fact to have any normative import applying to me. For the dispositionalist, my taking this stance would not reflect any deficiency in my understanding of what it is to have the intention in question or in my appreciation of the implications of my having this intention. For the case would be analogous to that of the distressed neighbour. Suppose that I had not taken the consideration that my neighbour is in distress and is calling on me for help to be a reason for me to help. This might reveal a degree of insensitivity on my part. But it would not in itself show that there is some deficiency in my understanding of what it is for my neighbour to be in this state. Viewed from the dispositionalist perspective, the case in which I adopt a purely contemplative stance to my own intention would be exactly parallel. There is the consideration that I intend to take a holiday, which I accept. According to the dispositionalist, this consideration is non‐normative and so does not, in and of itself, imply that I incur a commitment to doing what is necessary to carry out the intention. It is possible that I should adopt the purely contemplative stance, just as it is possible that I should have viewed the situation of my neighbour as having no normative import for me. Dispositionalists might well agree that if I did adopt such a stance I would be foolish or irrational. But it is not open to them to suppose that my adopting the stance must reflect a deficient understanding of what it is for me to intend to take a holiday or a deficient appreciation of what is implied by the consideration that I have this intention.
In taking this view of the matter, dispositionalists treat intentions as being analogous to character traits. This is not surprising since character traits are dispositional. Suppose that I am prone to irritability. In certain kinds of situations, in which many others remain unruffled, I become irritable. Knowing this, I might adopt a practical stance towards this trait. I could decide that I should do something about it or that it is best left alone. (Maybe I think that trying to deal with it would make me even more prone to irritability.) But I could adopt a purely contemplative stance in which I simply register that I am easily irritated. I could be indifferent or just curious about how I will react (p.112) in situations I am about to encounter. In any case, the matter would be of merely theoretical interest.63 While such an attitude might be foolish and imprudent, it is surely possible, and readily intelligible.
In conceiving of intentions as being in this way like character traits, the dispositionalist misrepresents the character of our normal relationship to our current own intentions. The peculiarities of this relationship can be brought out by reflection on self‐ascriptions of current intentions. Suppose that I formed my intention to take a holiday by taking a definite decision to do so. I might have thought to myself, and perhaps voiced the thought, ‘I intend to take a holiday.’ My thought here represents me as intending to take a holiday, but it does more than that. I do not merely report on a state that I am in, as I might report on the state of my health or on some character trait that I have. For there seems to be an intimate connection between thinking the thought and recognizing that it manifestly has normative import for me. To adopt a contemplative stance to my intention would fly in the face of this.
Decisions are not always made by thinking thoughts that explicitly self‐ascribe an intention. I might have thought, ‘I'm going to take a holiday.’ The content of this latter thought would simply be that I shall take a holiday. But in a context in which I am making up my mind, I would in thinking that thought implicitly self‐ascribe an intention. In that case too it is hard to make sense of a purely contemplative stance towards my intention. It would fly in the face of the fact that my having the intention manifestly has normative import for me.
Suppose that later, as I pass a travel agent, I recall that I intend to take a holiday. The recollection, like the thought by which I made up my mind, does more than merely represent me to have an intention: it is a renewal—an endorsement—of a previously formed intention. It is not the same as thinking that I intended to take a holiday. My actual thought is about a current intention—about something I currently intend. Again, it is hard to see how in having the thought I could fail to acknowledge its normative import.
These considerations are problematic from the dispositionalist perspective. If intentions, and indeed beliefs, were dispositional in (p.113) the way character traits are, then self‐ascriptions of current beliefs and intentions would merely represent one to be in a dispositionally characterized state of mind. In that case, adopting a purely contemplative stance towards our current intentions and beliefs would be readily intelligible. I am suggesting that such a view distorts the phenomenology of belief and intention.
Jane Heal has suggested that the peculiar character of self‐ascriptions of some mental states is best explained in terms of the idea that they have a constitutive role in relation to the states they ascribe (Heal 1994a; 2002). Her principal focus, it should be said, is on beliefs, but she thinks it plausible that this general approach works also for intentions (Heal 2002: 18). Applied to the cases I have been considering, the idea would be that my self‐ascription of an intention to take a holiday not only represents me as having that intention, but constitutes my having the intention. So there is an analogy between the self‐ascriptive thoughts in question and performatives. A performative proper is an utterance that in appropriate circumstances makes true what it represents to be so. Saying to my father ‘I'll visit this weekend, I promise’ both represents me as promising to visit my father and makes it the case that I have so promised.64 On Heal's approach, the self‐ascriptions under consideration would be taken to have a performative‐like character on the grounds that they make true what they represent to be so. On this account, the thought that is my making up my mind to take a holiday represents me as intending to take a holiday and makes it true that I do so intend. An attractive feature of the account is that it offers an explanation of the peculiar authority that conscious self‐ascriptions of intentions seem to have. In the absence of circumstances rendering the performance, in Austin's words, ‘unhappy’, like being befuddled by alcohol or deranged, a conscious self‐ascription would, according to the account, be authoritative just because it constitutes the obtaining of the very state of affairs it represents to be so. This is a strikingly imaginative way to deal with the peculiarities of self‐ascriptions of intentions, but I am not convinced that it works. Two problems arise.
(p.114) The account raises a question to which it is difficult to give a clear and convincing answer: how can merely thinking a thought conjure up the motivational potential characteristic of intentions? This matters because it is plausible that the thinking constitutes an intention only if, when the conditions are ‘happy’, it guarantees that one has the intention and therefore the motivation that goes with it. How can the self‐ascriptive thinking, even in ‘happy’ conditions, guarantee that? It is true that when I decided to take the holiday I must have already had, or conceived then and there, a desire for a holiday, for otherwise I would not have made the decision. So it might be suggested that my deciding merely channels existing motivation towards taking a holiday and creates some additional motivation, the latter being the added motivation associated with my having settled on a course of action. On this view, the question posed rests on a misunderstanding. The thinking that is the deciding—the making up of one's mind—does not conjure up the motivation: it merely directs it towards action and adds to it. The problem now, I think, is that this response merely shifts the problem. How can the mere thinking of the thought suffice to bring it about that the motivation I already had to take a holiday comes to be channelled towards action and supplemented by additional motivation? We are to suppose that the thinking is like a conduit for existing motivation. So long as the water is high enough prior to the conduit's being opened, it flows through and its force is somehow increased. The model makes it look as if, because the thinking was the making up of one's mind, it explains how the existing motivation is channelled and supplemented. It remains unclear how thinking the thought is supposed to do this.
The second problem—perhaps a more serious one—concerns whether the constitutive account can cope with self‐ascriptions that are false. I imagined that when I thought ‘I am going to take a holiday’, or ‘I intend to take a holiday’, I was making up my mind to take a holiday. In that case, in thinking the thought I formed the intention. But it could have been that it only seemed to me that I had made up my mind. Perhaps on such matters I am liable to vacillate and shilly‐shally. The phenomenology—what it struck me that I was doing—would not settle the matter because it need not differentiate between the case in which I truly made a decision and the case in which I did not. (p.115) Against this, it might be suggested that there is no real problem. If I am a vacillator I will be liable to make decisions and then quickly unmake them and then remake them again. Certainly, we should not set the standards for having made a decision so high that we cannot accommodate quick changes of mind. The borderline between having and not having made up one's mind may not be sharp. But nor should we lose track of the fact that making up one's mind about what to do, or for that matter about whether something is so, is settling on what to do or think. If I am a vacillator, I may not have made up my mind even though it seems to me that I have. From the inside, so to speak, a case of my not having made up my mind need be no different from one in which I have. That is why a self‐ascription can be false. Now, the constitutive account that appeals to the performative‐like character of self‐ascriptive thoughts can accommodate some cases of false self‐ascription—those in which the conditions under which the ascription is made are unhappy. But the edge is taken off this account if the conditions that can render a self‐ascriptive thought unhappy are to include cases in which, with a clear mind, a person seems to him‐ or herself to form an intention, while lacking adequate motivation. On the approach suggested by Heal, unhappiness conditions were to be conditions like being drunk or deranged. But so far as intention is concerned, this is a misleading picture. It is relatively easy, even with a clear head, to suppose that one intends to do something when one does not. People can think to themselves that they intend to give up smoking, or reduce their intake of alcohol or fatty food, and be mistaken. If I really made up my mind to take a holiday, it was not just in virtue of explicitly or implicitly thinking the self‐ascriptive thought, but in virtue of thinking the thought in the setting of the wider motivational state I was then in. The self‐ascription could not have constituted the intention, since it could not have guaranteed that I had really made up my mind, even in the absence of incapacitating conditions like drunkenness or derangement.
Self‐ascriptions of intention are peculiar. There is more to a self‐ascription of a current intention than a mere report of a state of mind. The peculiarity lies in the fact that adopting a purely contemplative stance towards our own current intentions is barely intelligible. I put this down to the fact that having a belief or an intention manifestly has normative import.
(p.116) Similar considerations apply to belief. When I ascribe a belief to myself I represent myself to be in state that manifestly has normative import for me. But here too it seems best to avoid the constitutive account. One can think one has settled in one's mind that something is so when one has not. A constitutive account can accommodate this fallibility, but at the price of making out that the cases in which the self‐ascriptions are false are relatively rare. But mistakes about what we think we believe can easily happen when strong emotions are engaged, as when people are trying to make up their minds whether someone close to them is trustworthy. In such situations the phenomenology is not bound to discriminate between entertaining the thought that something is so and reaching the conclusion that it is so. Whether one really has reached the conclusion will depend on the degree to which one is motivated to use the content of the thought as an assumption in subsequent thinking. That might not be clear to the subject until later.
In the light of these considerations, we need to return to the issue of how the dispositionalist attempts to explain the normative import that ascriptions of beliefs and intentions have. The sense of there being a problem here can easily slip away. If I intend to do something, then, of course, I have made it my goal to do that thing. And if I have made it my goal to do that thing, then, surely, I incur a commitment to doing what is required to achieve that goal. That is true, but if the dispositionalist is right, to have made something my goal is just to be disposed to take such steps as I think will bring about that thing. It has yet to be explained why my being so disposed should be thought to have any normative import. My being prone to irritability can contribute to providing a reason for me to do something about this trait, since it is both personally disadvantageous and unpleasant or even harmful to others. So we can explain why there is reason for me to control my proneness to irritability in terms of the desirability of not disadvantaging myself and of avoiding being unpleasant or harmful to others. The dispositionalist thinks that intentions are like character traits and so must explain the normative import of ascriptions of intentions in a similar fashion. The mere having of an intention, or a belief, under dispositionalism does not explain why it is commitment‐incurring.
(i) why, if we intend to Φ, we should avoid retaining that intention while never getting around to doing what is necessary if we are to Φ, and
(ii) why, given that we believe P, we should avoid retaining that belief while giving verdicts other than belief on implications of P.
There are two lines of objection to the dispositionalist's enterprise. The first draws attention to its oddness. The second raises problems for attempts to carry it out.
The oddness lies in the very idea that we need to explain the Implication Commitment and the Means–End Commitment Principles in terms of some value that accounts for why it should matter that we do the things that would discharge the commitments incurred by a belief or intention. This is a road that in my view leads to a dead end. If we go along it, and imagine ourselves contemplating our own current intentions or beliefs, it will look as if we need to posit a concern to realize something of value to explain why we take ascriptions of belief and intention to have normative import. (Note once again the analogy with the ethical case.) But it should, I think, strike as odd that such an explanation is needed. It brings too many concerns into the picture to account for a problem that looks unreal.
The other line of objection to the proposed strategy for explaining normative import is that there are problems in attempting to carry it out. To fail to discharge the commitments incurred by a belief or by an intention is to fail to satisfy an ideal of reason (Chapter 3, Section 2). Recall that the Implication Ideal is
For any π, θ, if θ is implied by π, then avoid believing π while giving a verdict on θ other than belief.
And the Means–End Ideal is
For any ϕ, avoid intending to ϕ while never getting around to doing what is necessary if you are to ϕ.
Granted that these are ideals of reason, it might seem that we could explain the normative import we attach to ascriptions of beliefs and intentions in terms of the value we place on conforming to those ideals. Indeed, the dispositionalist might invoke this idea to explain the oddity of adopting a contemplative stance to one's own current (p.118) beliefs and intentions. The oddity would be taken to lie in representing oneself as intending to do something or as believing something and yet being so irrational as to be indifferent to whether or not one satisfied the relevant ideal. This approach does not address the present problem. Valuing conformity to the ideals of reason certainly makes sense. But given dispositionalism, it becomes an issue whether reason does hold up to us the relevant ideals. If beliefs and intentions are as they are represented under dispositionalism, it becomes problematic that the supposed ideals are indeed ideals of reason. For if intentions and beliefs are dispositional in the way that proneness to irritability is dispositional, why should it be that, in so far as we are rational, we aspire to satisfy the Implication Ideal and the Means–End Ideal?
These considerations have a bearing upon another move that dispositionalists might make. It might be suggested that the Implication Commitment Principle and the Means–End Commitment Principle are necessary truths, even though ascriptions of beliefs and ascriptions of intentions are non‐normative. If this is meant to provide the sought‐for explanation of why we should think these principles true, it hardly succeeds in this task. Merely stipulating that the principles are necessary truths illuminates nothing. We need an explanation for why they should be thought to be necessary truths. My claim is that there is no reason to suppose that they are necessary truths if ascriptions of beliefs and intentions are non‐normative. Once you take seriously the idea that the ascriptions are non‐normative, then it becomes a mystery why, if they are true, a claim to the effect that the subject incurs a normative commitment is true as well.
5. Intentions, beliefs, and psychological commitment
A crucial feature of the view I shall defend is that the Implication Commitment Principle and the Means–End Commitment Principle are constitutive principles—partial specifications of what it is to believe something or intend something. They are also conceptual principles. Grasping the concept of intention commits one to accepting the Means–End Commitment Principle; grasping the concept of belief commits one to accepting the Implication Commitment Principle. That there should be constitutive normative principles is not (p.119) puzzling. I shall comment on two such principles that serve as analogues for the Means–End and Implication Commitment Principles. The first is the
Promise Commitment Principle: For any x, ϕ, if x promises to ϕ then x incurs a commitment to ϕing.
The other is the
Dean Commitment Principle: For any x, if x is Dean then x incurs a commitment to carrying out the duties of Dean.
These are not principles connecting the non‐normative to the normative in the way the Means–End and Commitment Principles are supposed to do, on the dispositionalist reading to which I have been objecting. They are constitutive principles. The first partially specifies what it is to promise something. The second partially specifies what it is to be Dean. They are also conceptual principles. Grasping the concept of a promise commits one to accepting the Promise Commitment Principle. Grasping the concept of Dean commits one to accepting the Dean Commitment Principle. I take the Means–End Commitment Principle and the Implication Commitment Principle to be analogous to the Promise Commitment Principle and the Dean Commitment Principle, at least in the respect that they are constitutive and conceptual.65
Part of the reason for taking seriously the idea that the Means–End Commitment Principle and the Implication Commitment Principle are constitutive and conceptual principles is that if they are then we can avoid the problems posed by the dispositionalist explanation of the normative import of ascriptions of beliefs and intentions. If these principles are constitutive and conceptual, then the normative import of ascriptions of beliefs and intentions is explained by the (p.120) normativity of the ascriptions. Can we shed some light on how this can be so?
Promising in and of itself gives rise to commitments on the part of the one making the promise. Being Dean incurs a commitment to carrying out the duties of the Dean. In both of these cases an explanation of why a normative commitment is incurred is available in terms of the notion of a practice and the commitments incurred by participating in a practice (see Chapter 3, Sections 4–5). Promising is a move in the activity of giving and reacting to promises. Participating in the practice in and of itself incurs a commitment to following its governing rules. Among the rules is that, roughly speaking, one does what one promises. So the commitment to following the rules generates a commitment to doing the thing promised. The Promise Commitment Principle is constitutive because it expresses the constitutive link between making a promise and incurring a commitment to doing the thing promised. It tells us something about the nature of promising and the concept of promising. Parallel claims hold for the concept of a Dean. But where believing and intending are concerned, the style of explanation that invokes practices will not serve the purpose. As I noted in the previous chapter, the commitments incurred by intentions and beliefs are grounded in ideals of reason that are not linked to practices.
We gain some insight into the normativity of intention by comparing intentions with desires. An intention, by my account, in and of itself incurs a normative commitment to doing whatever is necessary to do the thing intended. In this respect intention contrasts with desire. If I want to take a holiday, then I have some motivation to take a holiday, but I need not be psychologically committed to doing so. My desire consists in my finding the prospect of taking a holiday attractive—perhaps so attractive that in suitable circumstances I would decide to do so. But it could be that, even though I have this desire, I am not motivated to do what is necessary to satisfy it, because of the time and inconvenience of making suitable arrangements, or because other matters are more pressing. If this is so, and yet my desire to take a holiday persists, I need not be in breach of any ideal of reason. There can be good reasons not to satisfy a desire, which are not reasons to try to get rid of the desire. But if I intend to take a holiday, then I have settled on a course of action—made up my (p.121) mind to pursue it. It is in this sense that I am psychologically committed to taking a holiday and thus doing whatever is necessary to that end. If, despite being so committed, I fail to do whatever is necessary, then I will not have done what I am psychologically committed to doing—the intention will not be fulfilled. Desires and hopes are often not fulfilled. The prospect of their not being fulfilled does not in and of itself demand anything of the subject. But this is precisely where intention contrasts with desire and hope. An intention is a state of mind such that the prospect of its not being fulfilled demands a response on the part of the subject—either to give up the intention or to ensure that it will, after all, be carried out. The reason for thinking that intentions are intrinsically commitment‐incurring is that their psychological role is shaped by the subject's responsiveness to this demand.
This, I think, sheds some light on the status of the Means–End Ideal which grounds the normative commitments incurred by intentions. We can understand why this ideal is an ideal of reason in terms of the character of intentions. Though it might initially seem odd, there is reason to think of intention as having a constitutive aim distinct from the aims in which particular intentions consist. Intentions are clearly directed at doing the thing intended, but they are not directed at doing the thing intended come what may. There is, as it were, psychological work for an intention to do even if there is no prospect of its being carried out or reason not to carry it out. That work is to ensure that the agent does not let things drift but gives up the intention. An intention that is not going to be fulfilled calls for a change of mind. This can be captured in terms of the idea that the constitutive aim of intention is to ensure either that the intention is carried out or that the agent changes his or her mind. The explanation for why it is an ideal of reason that when we have an intention we either carry it out or give it up is that to do neither of these things is incompatible with intention's constitutive aim.66
Conceived as psychological commitments, intentions impact on thought and action in ways that are shaped by the normative commitments they incur. Consider two scenarios. In the first I am exploring a town I have never visited before. I drift along turning (p.122) this way and that. My exploration of the town is, of course, intentional, and when I turn to go along a particular street my doing so is usually intentional. (It might not be if I'm daydreaming.) But I am not carrying out an intention to follow a particular route. I just go where I feel like going at each point at which there are options. In the second scenario, I am again visiting a strange town, but I intend to visit the castle and my choice of route is directed to this end. In the first scenario there is no way I have to go—which is to say, no way I am committed, in the normative sense, to going. In the second scenario there is. This difference is a normative difference, but it is reflected in a difference in my psychology in the two cases. Whether or not I succeed in doing what I intend, my thoughts and actions in the second scenario are guided by the means–end commitment that the intention incurs. I will have discharged this commitment provided that I do the necessary or give up the intention. If I do the necessary by way of carrying out the intention, I shall have been guided by the means–end commitment. If I give up the intention, because of a change of mind or because I see that I am not going to be able to carry it out, I shall again have been guided by the commitment. For this to be so it is not necessary that I operate with any theory of intention. All that is necessary is that I am aware of what I intend and, as we say, of what I have to do to carry it out. Some might respond at this point by questioning whether intention is inextricably tied to such awareness. I think it is harder to make a case for thinking that intention is not so tied than to make out a corresponding case for desire. Intentions are not simply forces that steer us towards the thing intended. An intention's characteristic work is done via the agent's knowledge that he or she has that intention. This is at least part of the explanation for why it is so natural, when explaining that one did not mean to do something, to say that one did not realize one was doing it. If I have been joking in a manner that has upset a friend, my saying that I did not realize that I was upsetting the friend would generally be taken to indicate that I did not intend to do so. At least typically the things we do intentionally are the things we do knowing what we mean to do. I explore the matter further in the next chapter.
It will come as no surprise that I regard belief too as a psychological commitment by analogy with intention. My believing P is a psychological commitment to using P as an assumption in my thinking, (p.123) should the need to do so arise. If I believe P then the prospect of giving a verdict other than belief on an implication of P is not something I am in a position to view with indifference. It calls for a reaction.
It is in the nature of the psychological commitment in which believing P consists that the way it impacts on my thought and action is shaped by the implication commitment that it incurs—the normative commitment to believing any implication of P on which I give a verdict.67 The normative commitment can have this shaping role only if the subject has appropriate reflective capacities—including a capacity to reflect on the implications of things believed. It might be doubted that beliefs essentially involve such capacities. My point is that, since beliefs in the realm of personal understanding—the beliefs we ascribe to one another—implicate reflective capacities through the exercise of which the impact of those beliefs is shaped, we should acknowledge a distinction in psychological kind between those beliefs and any belief‐like states the impact of which is not so shaped.
Just as means–end commitments are grounded in the Means–End Ideal, so implication commitments are grounded in the Implication Ideal. But how should we explain the status of the Implication Ideal? Consider first a weaker ideal of reason—the ideal of avoiding inconsistency in one's beliefs. We can account for this ideal in terms of the idea that belief constitutively aims at truth—an idea discussed at some length in Chapter 2, Section 2. Suppose now that I believe P and that P implies Q. I might avoid believing P and believing ¬Q, and so avoid inconsistency, yet still fail to satisfy the Implication Ideal. For I might believe P and withhold both belief and disbelief from Q. (In that case I give a verdict other than belief to Q, despite believing P.) The view that belief constitutively aims at truth does not explain what has gone wrong in the situation envisaged. It looks as if we cannot account for what is wrong here in terms of the idea that belief aims at truth. The situation is one in which I have a picture of how things are on which both P and Q are true, and yet I give a verdict on Q that is out of kilter with its being part of that picture—the very picture that I use to steer my thoughts, actions, desires, and feelings. What is wrong here is that, although I believe P, I do not exploit P in (p.124) a way that is relevant to the issue as to whether or not Q is true. With that in mind, I suggest that the Implication Ideal has to do with our need to exploit correctly what we already believe when this is called for. If one fails to satisfy the Implication Ideal, then something that one believes is not playing the role that believing it accords it. Belief not only aims at truth; it aims at the correct exploitation of what is believed when the need arises. This is reflected in our actual psychology, since the realization that one is giving a verdict other than belief to an implication of what one believes is liable to prompt a readjustment. Of course, the right adjustment need not be to believe the implication. It might be to give up belief in what implies it.
The notion of a psychological commitment helps to account for the difference between self‐ascriptions of current beliefs and intentions on the one hand, and self‐ascriptions of current character traits on the other. (Recall that the failure adequately to account for this is a problem for dispositionalism.) In self‐ascribing a character trait, for instance proneness to irritability or aggressiveness, I merely represent myself to have the trait and thus the dispositions in which the trait consists. Even if the self‐ascription is true, it does not settle whether I should do anything, or whether I am committed to doing anything, about the matter.68 It does not in and of itself give rise to any normative commitments, and the psychological impact of having the trait is not necessarily shaped by any such commitments. By contrast, when I represent myself as having a current belief or intention, I thereby represent myself to have incurred a normative commitment. That is why it is not possible for me to adopt a contemplative stance towards my own current beliefs or intentions. Barring barely intelligible befuddlement, I cannot but acknowledge that the belief or intention has normative import for me.
It should be stressed that the problem for dispositionalism of accommodating the peculiarity of our relationship to our own current intentions and beliefs does not depend on the assumption that ascriptions of beliefs and intentions are normative. The problem arises because it is not readily intelligible that we should think of ourselves (p.125) as currently intending to do something, or as believing something, and yet adopt a contemplative stance towards our having this belief or intention. The trouble is that, on the dispositionalist account of the nature of beliefs and intentions, this would be readily intelligible.
The general stance I have adopted sheds some light on Moore's paradox as applied to thought. (See Moore 1942; 1944. For suggestive discussions from which I have benefited, see Heal 1994a and Moran 2001.) There is something odd about thinking, ‘I believe that p, and it is not the case that p’, even though the thought is not self‐contradictory. The thought represents me as believing that p and therefore as being normatively committed to believing whatever is implied by the proposition that p. It therefore represents me as being normatively committed to believing that p. But the second conjunct of the thought is the proposition that it is not the case that p. The oddity lies in the fact that my accepting the second conjunct is at odds with what the first conjunct represents me to be committed to. The analogue for intention of the paradox for belief just discussed is, ‘I intend to Φ, but I shall not do what is necessary if I am to Φ.’ This yields to an analogous treatment. Accepting the second conjunct is setting oneself not to do what the first conjunct represents one to be committed to doing.
To sum up, part of the case for thinking that ascriptions of beliefs and intentions are normative is that this conclusion accounts for the peculiarity of our relationship with our own current intentions and beliefs. The other part of the case concerns the inadequacy of the dispositionalist account of the normative import of ascriptions of beliefs and intentions. Here the argument was that under dispositionalism the Implication Commitment Principle and the Means–End Commitment Principle become problematic. Though they are in fact acceptable, the dispositionalist lacks a satisfying account of why they are so. Further support for my own view is provided by the light shed on Moore‐type paradoxes.
6. The problem of representing the dispositions characteristic of beliefs and intentions
In this concluding section of the chapter I aim to provide further support for the view that ascription of beliefs and intentions are (p.126) normative by discussing another problem for dispositionalism. We have seen that, according to dispositionalism, to believe something or intend something is to have certain non‐normatively specifiable dispositions. There is a problem for this view that has received inadequate attention. This is the problem of representing the dispositions. How is an ascription of a belief or intention supposed to represent the dispositions that are characteristic of the attitude ascribed?
When illustrating dispositionalism about the attitudes, we tend to use examples that turn on logical concepts expressible by, for instance, ‘everything…’, ‘something…’, ‘if…then…’, ‘either… or…’, ‘…and…’, and ‘it is not the case that…’. That is because the case for dispositionalism about propositional attitudes looks strongest when the focus is on candidates for plausible dispositions that mirror the logical properties of these logical concepts. Grasp of the disjunction‐concept, for instance, might be taken to implicate a disposition such that, roughly speaking, if one believes a disjunction and comes to believe the negation of one of its disjuncts, one is liable to come to believe the other disjunct. Departing from such examples, we might try dealing with concepts, like those of a bachelor or of an uncle, that have a tolerably, though not indisputably, clear inferential role. Then we might try to say something plausible about colour concepts like being red. To have that concept might be thought to implicate a disposition to come to believe that something before one is red when one has a visual experience such that there seems to be something before one that is red.69 Beyond such examples it becomes quite unclear what to say.
The problem that concerns me now arises even if we bypass worries about hedging clauses (see Section 2) and settle for rough specifications of dispositions in terms of what the subject is liable to do under certain conditions. Suppose that we learn that Tom, a keen gardener, thinks that his flower beds lack adequate nutrients. How does that inform us about Tom's dispositions to thought or action? The picture suggested by the dispositionalist theory is something like this. We are given an ascription to Tom of a belief B. In virtue of our grasp of what it is to have that belief, we take Tom to have a number (p.127) of dispositions. That puts us in a position to anticipate how Tom is liable to think and act and not to be surprised when he thinks and acts in particular ways. Notice that this view requires not just that in having B it is determined, in the sense of fixed, that one has certain dispositions. It requires that we should be able to read off specifications of the dispositions from the ascription of B. There is a question, though, as to how we are supposed to be able to do this.
How Tom will think or act as a result of having B will obviously depend on what else he thinks, what he wants, what circumstances he encounters, how he is feeling at the time, and many other factors. How are we supposed to have a grip on the potential of B to hook up with all of this? Examples turning on logical concepts are easier to handle, because in these cases the potential that a belief has to hook up with others seems to have a definite shape. It looks plausible that, if I believe that either Mary is arriving by plane or she is arriving by train, then I am disposed in such a way that, roughly speaking, were I to come to believe she is not coming by plane I would be liable to believe that she is coming by train. In this sort of case there is some plausibility to the idea that we can read off the disposition from the category and content of the attitude in question. But how is the potential for Tom's belief B to combine with other beliefs and attitudes to produce thought or action supposed to be written into the category and content of B? Again, the fact that the category of attitude is belief tells us a lot. It tells us that B has the potential to link up with other beliefs to produce further beliefs, and that it has the potential to link up with intentions to produce actions. But it is not so clear in this case how the content is supposed to tell us which sorts of beliefs and what sorts of intentions it can combine with. There can easily seem to be no problem, because it is not difficult to think of scenarios in which we might well form certain expectations about what Tom will think or do in view of what he takes to be the state of his flower beds. Knowing him to be a keen gardener who likes to keep his flower beds in tip‐top condition, we might expect him to apply fertilizer. The question though is how we are able to do this. I see no prospect of spelling this out in non‐normative terms that are psychologically realistic. It is clearly crucial that B concerns flower beds and the property of lacking nutrients, considered as applying to flower beds. So B is apt to combine with beliefs that Tom has about (p.128) flower beds, and about what it is for flower beds to lack nutrients, to produce some further belief. It would be nice from a theoretical point of view if we could demarcate this potential with the help of rules for drawing implications from the content of B that turn on the concepts of flower beds and of lacking nutrients. But there is no reason to expect such a project to be feasible. I do not appeal here to a general scepticism about conceptual truths. The point is simply that there is no reason to suppose that all of the concepts brought into play by B have a non‐normatively specifiable conceptual role which makes it possible to spell out, in non‐normative terms, its potential to contribute to the basis of other beliefs.
How should we then characterize the dispositions to thought and action characteristic of believing P? The answer, I think, is implicit in the previous discussion:
In virtue of believing P subjects are so disposed that, should the need arise, and given suitable prompts, they are liable to use P as an assumption in reasoning, constrained by the basic implication commitment incurred by the belief, and by the derivative commitments incurred when it combines with other attitudes, including other beliefs.
For my purposes, what needs to be stressed is that, while on this approach beliefs are dispositional in nature, the dispositions characteristic of them are normatively specified. For theoretical purposes we could delve further into the nature of the disposition just specified—conceiving of it as implicating dispositions linked with particular concepts brought into play by the belief and other attitudes with which it combines. (I shall turn to this matter in Chapter 6.) But these dispositions will themselves be normatively specified. The moral is that we represent the dispositions characteristic of a belief via our grasp of the commitments incurred by the belief. They are, roughly, speaking, dispositions to discharge the commitments. That is why it is no surprise that, when trying to think up examples of the dispositions characteristic of beliefs, by way of illustrating dispositionalism, we first think of the normative commitments of the belief we have selected and try to gerrymander a non‐normatively specified disposition to fit. We might hit on a specification of a disposition that looks just about plausible. If it is plausible, that will be because behaving as one would if the disposition were manifested would (p.129) be a way of discharging a commitment incurred by the belief. What gets overlooked in these exercises is that, absent our grasp of normative commitments incurred by a belief, and we have no way of representing the potential of the belief to affect thought and action.
There is, of course, no suggestion, that the use of the assumption P by someone who believes P will always be properly constrained by the relevant commitments. The person who has the belief may fail to realize that some proposition is an implication of the content of the belief and on that account may fail to discharge the implication commitment with respect to that implication. It is also possible for people to be mistaken about what is implied by something they believe. What it takes to discharge a belief‐relative commitment may not discharge a commitment that is not belief‐relative (Chapter 3, Section 2).
It might seem a simple matter to specify the dispositions characteristic of intentions in non‐normative terms. For surely those who have an intention are so disposed that, roughly speaking, given convenient opportunities and assuming they do not lose track of their intention, they will be liable to do what they believe to be necessary to carry it out. That is right so far as it goes, but it does not suffice to account for the explanatory and predictive use we can make of ascriptions of intentions. What we are trying to do is characterize the potential of an intention with a specified content to combine with beliefs and other intentions to produce action. In a case in which content of the intention is a disjunction, say, to travel to London by plane or by train, it is clear that the disjunctive form of the content serves as an index of the potential relevance of certain beliefs to carrying out the intention. But, as in the case of belief, once we move away from the logical form of contents, the matter is much less clear. The content of an intention does not have parameters for all of the possible variable factors that might affect the intention's potential to combine with beliefs to produce action. About all we can say, in advance of consideration of actual or possible scenarios in which the intention comes into play, is that the intention will impact on thought or action in ways that make some kind of sense. What makes some kind of sense will reflect the commitments the intention and the subject's beliefs incur.
(p.130) It might be suggested that the points about which I have been making a fuss simply reflect a familiar and widely accepted thesis in the current philosophy of mind—the thesis that the mental, at least so far as it concerns propositional attitudes, has a holistic character. It has long been recognized that the way in which any given attitude impacts on thought and action depends on what other attitudes the agent holds, and therefore that there is no hope of specifying the psychological (functional) roles of the attitudes one by one. A familiar response to this is to suppose that the role of any given attitude can be specified only with the help of a total theory of the formation, maintenance, and adjustment of attitudes (Lewis 1972; Loar 1981; see Schiffer 1987 for trenchant criticism). The problems I have been throwing up, it may be said, arise from trying to say something in general terms about the roles of particular attitudes taken individually. This fails to come to grips with the issues I have been raising, but it helps to clarify those issues. I have not been trying to specify the role of attitudes one by one, but rather to say something about the potential of attitudes to articulate with other attitudes and thereby to impact on thought or action. There are two problems for the contrasting approach now under consideration. The first is that, even assuming that there is some theory that, when applied, generates non‐normative specifications of the roles of (p.131) particular propositional attitudes, it seems implausible that we actually draw upon the assumptions of such a theory in our commonsense thinking. Recall that we are trying to become clearer about how attitude‐ascriptions can provide a basis for our expectations about, and explanations of, people's thought and action. The assumptions of the imagined theory must be available to us, and it must be plausible that we rely upon them. As things stand we have very little idea of what the theory is, which is why we lack any full‐blown attempt to spell it out, beyond illustrations of how the logical concepts and some observational or sensation concepts might serve to characterize psychological roles.70 The second problem puts in question whether there really is a theory that generates non‐normative specifications of the roles of particular propositional attitudes. There is little reason to be confident that a non‐normative specification of the potential of an attitude to impact on thought and action is obtainable from its content and category. The trouble is that, when considering how the belief or intention is likely to affect thought and action if these scenarios obtain, or were to obtain, we have no recourse but to think of what it would make sense for the subject to do in the circumstances. In other words, we are constrained to think about what will or would happen by thinking about normative matters. The normative commitments that have been so central to the preceding discussion are crucial for such thinking.
7. Back to explanatory irrelevance
Earlier (Chapter 1, Section 5) I described a line of thought deriving from Harman according to which moral judgements and principles are explanatorily irrelevant. The idea is that, if we want to understand why a person makes a moral judgement about some situation, we need never invoke the truth of the judgement or the truth of any principles that underpin it. We need only advert to non‐normatively specifiable features of the situation and psychological facts about the person making the judgement. Those who take this view about moral judgements might well be tempted to extend it to cover normative judgements generally. Why, they may ask, should we think of thought and action as being shaped by the commitments incurred by beliefs and intentions, rather than by beliefs of subjects to the effect that they have incurred those commitments? On this approach there is no need to suppose that normative commitments come into the picture other than as believed.
I do not, of course, deny that the commitments incurred by our beliefs and intentions do their work via our appreciation that we have incurred those commitments. But it is important for the overall picture I am presenting that ‘appreciate’ here is a success verb. What we appreciate to be so is so. Suppose that you want to hail a taxi. You wait at the kerb of the pavement until you see one coming. When one appears you raise your hand in the usual way. You do so because you believe that a cab is approaching. In this case we do not (p.132) hesitate to suppose that you have this belief because a cab is approaching. Indeed, this explanation for your belief is highly relevant to understanding what you are doing. Why should you be raising your hand in these circumstances, if not to hail an approaching cab? Now suppose that we know that you believe that if the stock market falls further then interest rates will be lowered. It turns out that the stock market falls further but interest rates are not lowered. When you affect not to be surprised by this, we take pleasure in reminding you that you are, or have been, committed to thinking that interest rates would be lowered. We can say this confidently not because you have told us of what your beliefs commit you to, or because you have voiced any beliefs as to what you are committed to—you had no need to tell us. We know what you thought and thus what you were committed to, and in this case we expect that, because you were so committed, you will appreciate that you were and will agree that you were wrong in thinking as you did. Just as in the taxi case the fact that the cab is approaching is independent of, and explains, your believing that it is, so in this case the fact that you are committed to believing a certain implication of things you believe is independent of, and explains, your believing that you are so committed. That you incur the commitment is as much a fact you encounter as is the fact that the cab is approaching. The commitment only makes a difference to what you think if you are aware of having incurred it. But it would be just as wrong to ignore the explanatory role of the commitment as it would be to ignore the explanatory role of the fact that the cab is approaching. I discuss the explanatory role of normative considerations more fully in Chapter 7.
(59) Some theorists think of properties in a more restricted sense. Here properties are the semantic values of concepts. A thing's possessing the property of being G is what makes it true that the thing is G.
(64) The classic texts are those of J. L. Austin (see Austin 1961: ch. 10; 1962). Austin himself believed that performative utterances are neither true nor false. It is certainly right that if my utterance made a promise, then what I said was not false. But since it makes it true that I promised and I said that I promised, the most natural way of describing the proceedings is to say that the utterance makes true what it represents me to have done.
(65) John Searle (1969: 33–44) made use of the idea of a constitutive rule in his early work on speech acts. His constitutive rules are akin to the rules that govern what I call practices. On Searle's view they commonly take the form of imperatives or state what counts as what, for example what counts as making a promise. In her second Locke Lecture (in draft), Christine Korsgaard talks about constitutive standards. She takes the hypothetical imperative, understood as the principle, ‘If you will an end, you must will the means to that end’, to be a constitutive standard for action in the sense that it is a ‘standard that we must at least be trying to follow if we are to count as acting at all’. The various commitment principles I have picked out are not constitutive standards in this sense. They are not principles one can follow or fail to follow. You can have an intention or belief and fail to discharge the commitment it incurs, but that is not the same as having a belief or intention and not incurring the corresponding commitment. What I called requirements of rationality in Ch. 3, Sect. 2, are, I think, akin to Korsgaard's constitutive standards.
(66) It is important to recall that failure to satisfy such an ideal is not necessarily irrational. Care must be taken over the relation between ideals of reason and requirements of rationality, if the latter are conceived in such a way that to fail to satisfy a requirement is ipso facto irrational. See Ch. 3, Sect. 2.
(68) There might be an issue arising from writings of Frankfurt (1998) about whether we can in some sense identify with a character trait. I do not think, however, that identifying with a trait could plausibly be regarded as treating the having of the trait as a normative matter. Interesting discussions of identification may be found in Buss and Overton (2002).
(70) It does not help to treat the theory as being tacit, as do Stich and Nichols (1992), since the issue is to explain how we are able to treat certain considerations as reasons for expecting agents to think and act in certain ways, and not simply to explain how behaviour is moulded in sub‐personal ways by sub‐personal data‐structures.