maize 13240734.0001.001 in

    8  How to Share an Intention[1]

    Literally sharing a single intention is easier than it seems—and fortunately so, since it seems quite impossible, at least to some philosophers.

    Philosophical puzzlement about how to share an intention doesn’t inhibit us from speaking of shared intention in daily life. When the Dean asks what the Philosophy Department intends to do about a vacancy in its ranks, she seems to be asking seventeen people to participate in a single intention. She is certainly not envisioning that the seventeen members of the department will arrive at seventeen individual intentions that somehow converge. And what rules out this possibility is not just that convergence among these particular people is unlikely; it’s that none of them is in a position to have an individual intention, rather than an individual preference, about how the vacancy is to be filled. How to fill the vacancy is up to the department, and so any intention on the subject must be formed and held by the group as a whole. The trick of sharing a single intention is thus essential to what the Dean has in mind.

    The question is how an intention can subsist in a subject so motley as seventeen philosophers—or even two reasonable people. An intention, after all, is the state in virtue of which someone is said to have made up his mind. How can several different minds submit themselves to a single making up?

    One way to answer this question would be to spell out necessary and sufficient conditions for the sharing of an intention, but my ambitions are more modest. I will be satisfied with finding a single instance that can convincingly be characterized as one in which an intention is shared. How the features of such an instance are to be generalized into an analysis of shared intention will not be my concern.

    ¤

    What has made some philosophers skeptical about literally sharing an intention is that intention is a mental state or event, and minds belong to individual persons. As John Searle puts it, “[T]alk of group minds . . . [is] at best mysterious and at worst incoherent. Most empirically minded philosophers think that such phenomena must reduce to individual intentionality.”[2] Similarly, Michael Bratman says, “[A] shared intention is not an attitude in the mind of some superagent.”[3]

    An apparent voice of dissent comes from Margaret Gilbert, who is willing to speak of a “plural subject”, produced by a “pool of wills”.[4] Surely, a truly “plural subject” ought to be a single subject that isn’t singular—or, if you like, a plural subject that isn’t just a plurality of subjects. That is, it ought to involve two or more subjects who combine in such a way as to constitute one subject, just as two or more referents combine to constitute one referent when subsumed under a plural pronoun. Talk of plural subjects therefore sounds perilously close to talk of group minds or superagents.

    Although Gilbert explains how two or more subjects combine to form what she calls a plural subject, she doesn’t fully explain how the combination qualifies, in its own right, as a subject. The possibility therefore remains that Gilbert, too, is using talk of a plural subject as a mere façon de parler, a convenient way of summarizing facts about a collection of subjects who never actually meld.

    Gilbert says that a plural subject comes into being when “each person expresses a form of conditional commitment such that (as is understood) only when everyone has done similarly is anyone committed.”[5] Yet if this arrangement is just a collection of well coordinated commitments, then it would seem to yield nothing more than a collection of well coordinated subjects. A pool of wills can hardly produce a composite subject if it fails to constitute, in itself, a composite will. And Gilbert doesn’t explain how a single will can be forged from the wills of different individuals.

    I want to fill in this gap in Gilbert’s view, by showing how distinct intentions held by different people can add up to a single token of intention, jointly held. Surprisingly, perhaps, the materials with which to demonstrate this possibility are to be found in Searle’s conception of intention. I believe that Searle’s account of shared intention is not entirely faithful to his own conception of what an intention is. A more faithful application of Searle’s fundamental conception yields the conclusion that talk of literally shared intention is neither mysterious nor incoherent.

    Before applying Searle’s conception, however, I want to examine a different and, in my view, more difficult challenge to the notion of sharing an intention. For I believe that the discussion of shared intention has not yet pinpointed what is problematic about this notion. What’s problematic, I think, is not the idea of sharing a mental item but rather the idea of sharing the particular kind of item that intention is.

    ¤

    As Michael Bratman pointed out in his ingenious paper “Two Faces of Intention”,[6] there must be a difference between the mental state of having a plan and that of having a goal, although either state can be called an intention, in one or another sense of the term.[7] Bratman argues that an agent can rationally have two goals that he knows to be mutually incompatible, in the sense that they cannot both be attained; for he can aim at both and “let the world decide” which one he attains in fact. But an agent cannot rationally plan to produce two outcomes that he knows to be incompatible, since rationality forbids him to have inconsistent plans.

    Thus, for example, you can rationally aim to win a research fellowship while also aiming to receive a visiting professorship in the same year, even though you know that you cannot in fact receive both at once; for you can simply let the relevant institutions decide which, if either, you are to receive. But you cannot rationally plan to get a fellowship while also planning to get a visiting appointment, given your knowledge of their incompatibility.

    As you update your curriculum vitae, you can be described as acting both with the intention of getting a fellowship and with the intention of getting a visiting professorship, in a broad sense of the word ‘intention’ that applies to any goal or purpose with which you act. But Bratman points out that there is a narrow sense of ‘intention’ that’s confined to plans and other, plan-like commitments to act. In this sense, you cannot intend to get both awards if you know that you can’t get both. Indeed, you probably cannot intend, in this sense, to get either, since you’re well aware that whether you get them isn’t up to you.

    This last observation helps to explain why intentions thus narrowly defined are subject to more exacting rational constraints than goals.[8] (I shall follow Bratman in adopting the narrower definition as the default for the word ‘intention’.)

    Your intentions, so defined, are the attitudes that resolve deliberative questions, thereby settling issues that are up to you. If an issue isn’t up to you, then you are not in a position to settle it, and so you face no deliberative question about it. But if an issue is up to you, then you are in a position to settle it, and there is consequently a deliberative question for you as to how it will turn out.[9] In resolving this deliberative question, you will arrive at an attitude that settles the issue both actually and notionally. That is, the presence of this attitude will cause the issue to turn out one way rather than another, thus resolving it in fact; while the attitude will also represent the issue as turning out one way rather than another, thereby resolving it in your mind. This issue-resolving attitude is an intention, in the usage that I have now adopted.

    Having a goal doesn’t resolve anything, in either sense. Your having the goal of getting a fellowship doesn’t settle whether you will in fact get it, nor does it constitute your viewing this issue as settled. Your having a goal merely motivates you to undertake various measures believed conducive to the goal—including measures of deliberation and planning. (Of course, if you have the goal of getting the fellowship, then you can also form the intention of trying to get it, since whether you try is still up to you; but intending to make an attempt at doing something is not the same as intending to do it.)[10]

    The reason why you cannot rationally intend to do things that you regard as incompatible is that you cannot actually settle matters, or rationally regard them as having been settled, in ways that aren’t compatible. The reason why you can rationally have incompatible goals is that having a goal settles nothing. So long as practical issues remain unresolved, they can be pushed toward incompatible resolutions; they just can’t be resolved incompatibly, or rationally regarded as having been so resolved. That’s why intentions and goals are subject to different rational constraints.

    ¤

    These reflections cast doubt, in my opinion, on one attractive strategy for explaining how intentions can be shared. This strategy is to imagine two or more agents as individually holding different token intentions of the same type, by holding intentions formulated in the first-person plural. The idea is that you and I can partake in the same intention if there is something that each of us individually intends that “we” are going to do.[11] The problem with this analysis of shared intention is that it yields, so to speak, too many chiefs and too few braves, or too many cooks and too little broth.

    There is nothing problematic about first-person-plural intentions in themselves. One person can decide or plan the behavior of a group, for example, if he holds authority or control over the behavior of people other than himself. If you will do whatever I tell you to do, then what you’ll do is up to me, and I am in a position to make decisions about it. As your boss or commanding officer or master, then, I am in a position to decide what you and I will do together, and so I am in a position to form intentions about what “we” will do.

    But shared intention is not supposed to be a matter of one person’s deciding or planning the activities of a group; it’s supposed to be a matter of shared intending, in which each member of the group participates equally in forming and maintaining the intention, fully recognizing the others as equal participants. What we are going to do is supposed to be determined by you and me jointly, in this case; and each of us is supposed to regard the issue as being thus jointly determined.

    The problem about sharing this role is that one person’s exercise of discretion over some issue would seem to exclude any other person from exercising discretion over the same issue. That is, if I decide that you are going to do something, then I cannot think that whether you’re going to do it remains up to you; whereas if I want to leave it up to you, I cannot simultaneously regard myself as having decided it. If I am to settle the matter, I cannot think of you as having settled it first or as being in a position to settle it later; whereas if I am to leave you to settle it, I must not preempt you by settling it myself.

    Yet framing an intention as to what “we” are going to do together would seem to entail settling what you are going to do with me as well as what I am going to do with you. And insofar as either participant thus purports to settle what both will do, he would seem to leave no discretion to the other. The model of first-person-plural intentions requires each of us to intend that we will do something, as if he were in fact settling the issue for both of us. Yet how can I frame the intention that “we” are going to act, if I simultaneously regard the matter as being partly up to you? And how can I continue to regard the matter as partly up to you, if I have already decided that we really are going to act? The model seems to require the exercise of more discretion than there is to go around.

    ¤

    What makes shared intention a puzzling phenomenon, then, is that the logical space of decisionmaking is open only to those who are in a position to resolve an issue, and it admits only one resolution per issue. Shared intention therefore involves the sharing of something that ordinarily seems indivisible.

    Suppose that we jointly decide to lift a heavy sofa together. We thereby exercise a kind of joint discretion over the issue of whether the sofa will leave the ground. The interesting question is precisely how two people can jointly exercise discretion over a single issue. The answer cannot be that each of us exercises full discretion over the issue individually, as we would have to do if each of us were to intend that “we” will lift the sofa. Discretion cannot be shared by being multiplied in this way, since no issue can be settled by each of two people at once.

    Of course, each of us can exercise discretion over a part of the sofa-moving operation, I deciding whether to lift my end of the sofa, and you deciding whether to lift yours. But in that case, we wouldn’t really be sharing an intention; we would be holding different intentions that yielded a single result. This result might in fact be intended in the broad sense that encompasses goals. For I might lift my end of the sofa in the hope that you would lift yours, and you might lift your end in the hope that I would lift mine—in which case, the sofa’s leaving the ground would be a goal for each of us. But neither of us would have settled whether the sofa would leave the ground, that outcome being up to neither of us individually; and so neither of us would have intended that outcome in the stricter sense of the word.[12]

    Holding different, coordinated intentions in the pursuit of a common goal may turn out to be the closest that people can come to sharing an intention. If so, then the phenomenon holds no mystery, since the logical space of goals is less exclusive than that of intentions. Even if two people cannot individually settle one and the same issue, they can easily have one and the same outcome as their individual goals. Settling different issues in the pursuit of one goal is thus unproblematic.

    Precisely because it’s unproblematic, however, the sharing of goals is of little philosophical interest. What’s interesting is the possibility that two or more people can somehow share, not the goal of producing a particular outcome, but rather discretion as to whether the outcome will occur.

    ¤

    Gilbert’s account of pooled wills is the only one, to my knowledge, that doesn’t portray the participants either as holding similar intentions that preempt one another or as holding disparate intentions with nothing in common but a goal. It’s the only account in which the participants manage to exercise shared discretion over a single issue.[13]

    The way they accomplish this feat, in Gilbert’s account, is by individually exercising conditional discretion over the issue, in such a way that their conditionally settling the issue separately adds up to their categorically settling it together. As I have said, I think that Gilbert is correct about how this pooling of wills occurs. But she doesn’t fully explain how individual exercises of conditional discretion add up to something that qualifies as a joint exercise of categorical discretion—how a collection of wills becomes one collective will. I want to clarify this feature of Gilbert’s account, by applying Searle’s theory of intention.

    Searle’s theory of intention is appropriate to the task because it picks out the features that make intention an exercise of discretion, which is what seems so difficult to share. According to Searle, an effective intention is a mental representation that causes behavior by representing itself as causing it.[14] That is, when I effectively intend to take a walk, I represent this very representation as causing me to take a walk, and this self-referring representation causes me to take a walk.[15]

    This representation settles the question whether I am going to take a walk, and it settles the question, as an intention should, both actually and notionally. It settles the question notionally because it represents my taking a walk as resulting from this very representation; and it settles the question in fact by causing me to take a walk, just as it represents.

    My power to settle an issue in this fashion is what we’re talking about, I believe, when we say that the issue is up to me. For I am entitled, by virtue of this power, to think of myself either as taking a walk or not, insofar as either thought would cause me to behave accordingly. From my perspective, then, there is no antecedently right answer to the question whether I am going to take a walk—nothing that I must think on pain of being wrong—since the right answer will be whatever I think. And that’s just what it is, I believe, for an issue to be up to me.[16]

    Searle’s analysis thus lays bare how an intention constitutes an exercise of discretion—that is, an agent’s settling of an issue that is up to him. And if the phenomenon described in this analysis can somehow straddle the boundaries between people, then the problem of shared discretion will have been solved.

    ¤

    Searle regards intentions as belonging essentially to individuals because he regards the representations involved as essentially mental and hence as being essentially lodged in minds, which belong to individuals. But I do not think that we can rule out the possibility of literally shared intentions on the grounds that there are no collective minds.

    To begin with, I am not sure that intention is essentially mental. There are of course mental intentions, but perhaps there can also be oral or written intentions—just as there are not only mental but also oral or written assertions. Of course, talk of oral or written intentions sounds odd, but talk of oral or written decisions sounds less odd, and talk of oral or written commitments is not odd at all. If I can commit myself to a course of action by speaking or writing, then there would seem to be a sense in which I am thereby making an oral or written decision; and if I can make a decision by speaking or writing, then there would seem to be a sense in which I can frame an oral or written intention.[17] Indeed, the possibility of such intentions appears to be a consequence of Searle’s theory. All that’s essential to intention, in Searle’s theory, is a representation with a particular content and causal role. Why shouldn’t the relevant content and causal role attach to representations that aren’t in the mind?

    One may want to insist on intention’s being a mental state, of course. But then I would be inclined to say that the existence of collective minds remains an open question. Whether there are collective minds depends on whether there are collective mental states. And if we insist on the proposition that any state constituting an intention is ipso facto mental, then whether there are collective mental states will depend, in part, on whether there are collective intentions. Hence we cannot rule out the possibility of collective intentions on the grounds that there are no collective minds: the direction of logical dependence goes the other way.

    I propose to suspend judgment on whether intentions are essentially mental and whether minds are necessarily lodged in the heads of individuals. What I want to do is to examine whether there can be an item that is literally shared between two or more people while bearing the content and playing the functional role identified by Searle as characteristic of intention. Such an item would be a representation that caused action by representing itself as causing it, but it would be a token representation that was in some sense jointly held by two or more people.

    ¤

    In order to be jointly held, this representation will have to be public. That is, it will have to be an utterance, inscription, or depiction of some kind.[18] And it will have to belong to more than one agent, in some sense of the word ‘belong’. The most obvious candidate is a verbal representation that consists partly in the speech of one agent and partly in the speech of the other, like the statements or anecdotes that one person starts and another finishes.

    I think that the conditional commitments described by Gilbert may in fact combine to form such a jointly held representation. Speaking only roughly for a moment, I would put it like this. When one agent says, “I will if you will,” and the other says, “Well, I will if you will,” their speech-acts combine to produce a single story, just as when you start telling an anecdote and your spouse finishes it for you.[19] Speaking only roughly, I would say that these two utterances combine to form a verbal representation that’s equivalent in content to “We will.” If this representation plays the right causal role, by prompting the behavior that it represents, and if it also represents itself as playing that role, then it will just be an intention—or, at least, it will be everything that an intention is except mental, if anyone still wants to quibble about that term.

    Now, an exchange of commitments with the form “I will if you will” seems problematic, because each agent seems to make his action conditional upon the action of the other agent, with the result that neither will act unless the other does, and hence with the result that action will never get started. But Gilbert cuts through this problem, by understanding this commitment as making the speaker’s action conditional upon the other agent’s commitment rather than his action. “I will if you will,” as interpreted by Gilbert, means something like “I’m willing if you’re willing”; and as I shall explain, each of these statements can constitute the very willing that is required by the other.

    Rather than scrutinize these statements further, however, let me try to construct the right sort of statement from scratch. I’ll start by illustrating how a statement can play the role of an individual’s intention; then I’ll consider how two people’s statements can form an intention that is shared between them.

    ¤

    I think that the statement “I’m going out for a walk” can sometimes be causally responsible for the speaker’s going out for a walk. Before making the statement, the person’s motives for taking a walk may not outweigh his motives against taking a walk, or may not outweigh them enough to produce the kind of exertion required to get him out of the house. But they may be sufficient to produce the statement “I’m going out for a walk,” and this statement may then bring into play an additional motive for taking a walk. The speaker’s love of the outdoors and his desire for exercise may now be significantly reinforced by a further motive—for example, the desire not to have spoken falsely.[20] Having said that he’s going out for a walk, the agent faces two alternatives: either go out for a walk or be in the position of having asserted a falsehood. And taking a walk may well be preferable to having said what turns out to be untrue. The agent may thus have a new motive that tips the balance enough to get the projected walk underway.

    Of course, the announcement “I’m going out for a walk” often plays no role in causing the walk that it announces. If someone speaks these words on his way out the door, chances are that he was already going out for a walk, without any further prompting. But the case that I’m imagining, in which the statement helps to cause its own fulfillment, is a case in which the statement is made from the armchair, by an agent who would have preferred to sit still if he hadn’t raised the price of doing so, by saying that he was going for a walk.

    It’s precisely because an agent expects such a statement to raise the price of inaction that his desire to act can move him to make it. Lethargy may prevent his initial motives from setting his legs in motion, but it won’t prevent them from setting his mouth in motion so as to bring additional motives to bear on his legs. It sounds odd, I know, to speak of getting one’s legs to move by moving one’s mouth. Yet I venture to say that everyone—or at least, anyone who has struggled with lethargy—knows what it’s like to announce his departure in order to counteract the temptations of further delay.

    ¤

    When the agent says “I’m going out for a walk” in such a case, his utterance isn’t meant or understood as a mere report or prediction of his behavior. One might want to distinguish his utterance from a report or a prediction by saying precisely that it expresses an intention to take a walk; and I would agree. But one might then want to question whether this utterance aims or purports to be true, and hence whether it provides a basis on which the speaker might be guilty of asserting a falsehood.

    Yet I think that we can pinpoint the features in virtue of which this utterance expresses an intention rather than a report or prediction, and these features are compatible with—and, in fact, depend on—its purporting to be true. The utterance expresses an intention because it is meant to be understood as not only playing but also implicitly claiming the role of causing its own fulfillment. That is, when the agent says from his armchair, “I’m going out for a walk,” his utterance not only causes him to go out for a walk but also implicitly describes itself as doing so. It is meant to be understood as “I am hereby causing myself to go out for a walk”—a statement that will be false if the agent doesn’t go for a walk, but will be true if the agent is moved to go for a walk in order to make the statement true.[21]

    What distinguishes this statement from a prediction or report is that it doesn’t purport to represent a fact that’s independent of itself. To report or predict a walk would be to represent the walk as something that was already happening or destined to happen, whether it was announced or not. A report or prediction would thus purport to convey an independent fact. By contrast, the agent who expresses an intention by saying “I’m going out for a walk” does not represent the projected walk as something that was going to happen anyway, whether or not he had said so; he represents it as something that is now going to happen precisely because of his hereby saying so. His statement thus differs from a report or prediction in that it doesn’t purport to convey a truth independent of itself.[22] But it still purports to convey a truth.

    When we say that this statement expresses an intention, we might mean that there is some inner state of intention to which it gives vent. But we might alternatively mean—and in this case, I think we should mean—that the statement expresses an intention that consists in the statement itself, much as the utterance “I promise” both expresses and is a promise. The former statement is an intention because it causes its own fulfillment by representing itself as doing so. In its own right, then, it possesses the features that Searle has identified as characteristic of the will.

    ¤

    But how can a statement be an intention? We ordinarily think of an intention as something that an agent not only forms but also maintains, from the time at which he forms it to the time at which he executes it. We therefore think that someone who once intended to do something in the future can sensibly be asked whether he still has that intention. But what I have been calling a statement, in my example, is a speech-act, and hence an occurrence: it’s made at a particular time, and it certainly isn’t still being made throughout the interval between utterance and action. How, then, can we regard a statement as constituting an intention?

    The solution is to recognize that Searle’s analysis naturally encompasses more than intentions conceived as states. Searle analyzes intentions as representations with a particular content and a particular causal role. And representations can be either representational states (like mental images or beliefs) or representational acts (like gestures or assertions), either of which can bear the appropriate content and play the appropriate role. The analysis gives us no grounds for distinguishing between the two, so long as they are representations of the self-describedly self-fulfilling kind.

    Searle’s analysis is therefore best deployed to encompass both states and acts of the will: states of intending, which are what we ordinarily mean by “intentions”, and acts of intending, which we usually call “decisions” instead. So interpreted, the analysis yields a more complicated—but, to my mind, more life-like—picture than one might previously have had of the relation between decisions and intentions.

    One might have thought that all intentions are states, some of which are, though others are not, produced by mental acts on the part of the agent. In this simple picture, decisions are acts of the will only by virtue of having states of the will as their products: the act of producing an intention is not an instance of intention in its own right.

    In the picture that I am suggesting, however, some acts of deciding are intentions in their own right, because they are representational acts with the content and role characteristic of the will.[23] When someone says “I’m going out for a walk,” he may be deciding aloud to go for a walk: his speech-act may in itself be an occurrent intention.

    Having carried Searle’s account this far, however, we should probably carry it one step further. For we can see that representational acts with the content and causal role of intentions may fulfill their role at a temporal distance, by initiating causal processes that will come to fruition long after the acts themselves are complete. Someone’s saying “I’m going out for a walk” may bring into being various conditions that will cause him to go for a walk at a later moment, after his words have faded. These conditions may include, for example, his memory of having recently said that he would go for a walk, and his conception of how much time he has in which to make the statement true.

    These persisting conditions, which secure the causal efficacy of a spoken decision, amount to the agent’s remaining decided, since his decision remains in force so long as they persist. And if someone has decided to go for a walk and remains decided, in the sense that his decision is still in force, then surely he still intends to go for a walk. Extending Searle’s analysis to acts of intending should therefore lead us to extend it even further, to persisting states in the aftermath of such acts.

    Note that these states may not in themselves satisfy Searle’s analysis, since they may not involve representations with the precise content and role that the analysis requires. (The memory that one has announced a forthcoming walk is a mental representation, and it may cause one to go for a walk, but its causing one to go for a walk may not be part of what it represents.) These states may therefore have to qualify as intentions in a secondary sense, by virtue of embodying the causal force of past decisions, which qualified as intentions in the primary sense defined by Searle.

    We thus arrive at a picture in which there are three distinct ways of intending to go for a walk. The intention to go for a walk can consist in a state representing one’s going for a walk (partly because of this state); in an act of representing oneself as going for a walk (partly because of this act); or in the persisting causal force of such an act of deciding.

    I find this picture especially life-like because it corresponds to what I find upon introspection when asked whether I still intend to do something. When asked this question, I sometimes find myself with a standing conception of the action as forthcoming (partly because of this very conception)—a state of mind that I might describe by saying that I was still intent, resolved, or determined to act. Sometimes, however, I find no more than a memory of having said I would act, and a desire to bear myself out—in which case, I might say that I had decided and hadn’t changed my mind. These are two, phenomenologically distinct ways of “still intending”.

    ¤

    Let us return to the statement “I’m going out for a walk.” I have now suggested that this statement can simultaneously express an intention and constitute the intention expressed, much as a statement can express and constitute a promise. This statement can be made to resemble a promise even more closely; for it can be reworded so as to constitute an intention partly by virtue of calling itself one. Just as a statement can be a promise partly because it begins “I hereby promise . . . ,” so a statement can be an intention partly because it begins “I hereby intend. . . .”

    An effective intention, remember, is a representation that causes behavior by representing itself as causing it. To call something an effective intention to take a walk is therefore to represent it as something that causes the taking of a walk. Consequently, something can have the content of an intention to take a walk by calling itself an effective intention to take a walk. Someone can say, “This statement is an effective intention to take a walk,” and his statement will have represented itself as causing him to take a walk—which is just what an intention to take a walk needs to represent. He can therefore say, “I hereby effectively intend to take a walk,” and his utterance will have the content of an intention to take a walk, just as “I hereby effectively promise to take a walk” has the content of a promise.

    The agent can omit the word ‘effectively’ from this announcement, because announcing an intention carries the conversational implicature that it is effective. Announcing an intention that’s ineffective could mislead, and announcing an intention as ineffective would be pointless. The tacit assumption that an utterance is neither misleading nor pointless thus supplies what the agent omits if he fails to say that his intention is effective. The agent can similarly drop the word ‘hereby’ as being implicitly understood in the context. He can then issue an intention to take a walk simply by saying, “I intend to take a walk.”

    One might question the analogy between “I intend” and “I promise” in light of how these utterances differ when spoken insincerely. An insincere utterance of “I promise” still constitutes a promise, whereas an insincere utterance of “I intend” is insincere precisely because the speaker has no corresponding intention. Doesn’t this difference undermine the claim that “I intend” constitutes the intention that it expresses?

    I think not. To be sure, an insincere utterance of “I intend” is insincere because there is no intention corresponding to it. But the intention that’s missing in this case need not be one that, if it existed, would exist independently of the utterance. The intention that’s missing can be one that would consist in the utterance—an intention that the utterance falsely purports to be.

    Keep in mind that Searle’s analysis requires an intention to have both a particular content and a particular causal role. The statement “I intend to take a walk” has the content of an intention, but it may still lack the causal role, of prompting the speaker to take a walk. If the utterance has no tendency to cause this action, then it is false; and if the speaker knows that it hasn’t, then he is lying. His lie consists in saying “I (hereby) intend” while consciously failing to satisfy the causal condition for intending.[24] The insincerity of his utterance—its reporting an intention that he doesn’t have—consists in its purporting to be an intention that it is not.

    Thus, the difference between “I promise” and “I intend” is not that an utterance of “I intend” isn’t an intention; the difference is rather than an utterance of “I intend” is an intention only if it has a tendency to prompt the intended action. An utterance of “I promise” can qualify as a promise whether or not it can prompt the promised result.

    The explanation of this difference is that promises, unlike intentions, are constituted as commitments by social convention, independently of their causal role. The utterance of “I promise” commits the speaker by placing him under a socially defined obligation. But intentions are psychological rather than social commitments. An utterance of “I intend” must commit the speaker in the sense of making him psychologically committed to action.[25] This utterance must therefore play the appropriate role in the speaker’s psychology in order to qualify as a commitment, and hence as an intention.

    This difference between “I intend” and “I promise” does not undermine the broader similarity between them. “I intend”, like “I promise”, can still constitute the intention that it expresses, and it can still constitute that intention partly in virtue of describing itself as such.

    ¤

    A statement can describe itself as an intention without beginning with the words “I intend. . . .” An agent can intend to take a walk by saying “I will take a walk” in a sense that means “I hereby effectively will it.” The emphatic use of “will” in the first person can constitute a self-described willing, whose emphatic tone both acts out and connotes its effectiveness.

    I suggest that the statement “I will if you will” should be understood in this sense. It means, “I hereby frame an effective intention that’s conditional on your framing an effective intention as well”—that is, “I hereby will it, conditional on your willing likewise.”[26] And this statement just is the conditional willing that it describes itself as being.

    This analysis helps us to understand why it seems mildly uncooperative to answer “I will if you will” with “Well, I will if you will.” If I say, “I will if you will,” then I have thereby willed it and said that I’ve willed it—indeed, I’ve willed it precisely by saying so—and your saying “I will if you will” therefore sounds as if you haven’t been listening. Your responding with “Well, I will if you will” perversely calls into question whether I have in fact willed; and so it doesn’t bring the exchange to a satisfactory close.

    The proper response to “I will if you will” is “Then I will.” The word “then” indicates that your intention is conditional on mine, in the same way that mine is conditional on yours, but that the condition has already been satisfied. What you’re saying is, “Given that you have willed likewise, I will it, too.”

    ¤

    Suppose that I say, “I’ll go for a walk if you will,” and you answer, “Then I will.” According to my analysis, each statement describes itself as an effective conditional willing, or intention, to take a walk; and each statement thereby ascribes to itself a conditional causal power—namely, the power of prompting the speaker to take a walk if (or given that) the hearer is found to have willed likewise.

    I now suggest that these statements have the conditional causal powers that they ascribe to themselves, and hence that they qualify as the effective conditional intentions that they claim to be. Each statement will indeed prompt the speaker to take a walk if he finds that the only alternative is to have spoken falsely. Of course, avoiding this alternative won’t entail taking a walk if the condition placed on the statement’s self-ascribed causal power has not been fulfilled. If you haven’t issued an effective conditional intention corresponding to mine, then I can stay home and still have spoken the truth in saying “I’ll take a walk if you will.” But if I find that you have willed accordingly, then I shall be prompted take a walk, since my statement claimed that it would prompt me to take a walk under these conditions, and I don’t want to falsify it. Hence my statement is an effective conditional intention, and so is yours, for the corresponding reasons.

    Note here that the conditional causal power of my statement won’t be activated merely by your issuing the relevant intention; I must recognize that you have issued it. Thus, if my statement “I will if you will” had been fully precise and explicit, it would have carried an additional condition: “I will, if you will and if I recognize you as willing.”[27] But then, if your intention is truly like mine, it will be such as I can hardly help but recognize, since it will explicitly claim to be the effective conditional intention that it is. Your “Then I will” leaves me no way to deny that you have fulfilled my stipulation “if you will”, except by failing to understand or to credit what you say. And since we are engaged in conversation, there is a tacit assumption of mutual understanding and credence.

    The assumption of mutual understanding and credence amounts, on my part, to the concession that, if you issue an intention like mine—that is, a self-proclaimed intention—then I will indeed recognize it for what it claims to be. Hence the further condition on my intention is conceded in advance and can be omitted. I can say “I will if you will” without adding “and if I recognize you as willing,” because if you do will as I have hereby willed, then (I tacitly concede) I shall recognize you as willing, since (we tacitly assume) I’ll understand and credit what you say.

    Your response “Then I will” thus satisfies the only condition that hasn’t been tacitly conceded in my statement; it therefore leaves the truth of that statement to depend, in my eyes, on whether I take a walk; and so it activates the statement’s self-ascribed power to get me walking. And of course, my statement has already satisfied the only condition not tacitly conceded in your statement, thus activating its causal power as well. Together the statements therefore prompt us to take a walk.

    ¤

    What’s more, these statements jointly imply that they will prompt us to take a walk. Each statement not only fulfills the other’s only remaining condition in fact but also discharges it in logic.

    When you say, “I will,” you not only do as I stipulated when I said “if you will”; you also say that you’re doing it; and so your statement licenses the detachment of my stipulation. My statement claims that it will prompt me to take a walk if you will likewise; your statement claims that you are willing likewise; and these statements imply that I shall now be prompted to take a walk. Mutatis mutandis they imply that you will be prompted to take a walk as well. And these very implications, rendered undeniable by the conventions of conversation, are precisely what will prompt you and me to take a walk, on pain of having said things that turn out to be false by virtue of implying falsehoods. Our statements therefore combine to form a joint statement saying, in effect, that they will jointly prompt us to take a walk; and they jointly prompt us to take a walk, as they jointly say. They consequently add up to a single representation that causes our actions by representing itself as causing them—a single token intention that is literally shared between us.

    This case of collective intention is designed to show that Gilbert’s talk of “pooling our wills” can be taken quite literally without becoming either mysterious or incoherent, as Searle might suggest. Indeed, it shows that Searle’s own analysis of intention can account for the possibility of literally pooling our wills, since it implies that two individual intentions, belonging to two different agents, can combine to form a single intention governing the behavior of both. This joining of intentions occurs when two spoken decisions, each of which is logically and causally conditional on the other, combine to form one spoken decision in which the conditions have been discharged.

    ¤

    The main virtue of this account is that it explains how several exercises of individual discretion can add up to a single exercise of collective discretion. The best way to formulate the explanation is by contrasting the case of shared intention with a case of disparate but parallel intentions.

    Suppose that I say, “I’m going out for a walk,” and you say, “I’m going out for a walk, too.” In that case, our statements combine, in the sense invoked above, to form a story to the effect that both of us are going out for walks; and this story can in fact cause its own fulfillment. Why, then, do we not qualify as sharing an intention in this case?

    The answer is that each half of the story settles only half of the story. The first half determines, and represents itself as determining, that I am going for a walk; the second half determines and represents itself as determining that you are. Hence I do not help to settle your behavior, actually or notionally, and you do not help to settle mine. We do not exercise discretion over any of the same issues.

    The situation is different when I say, “I’ll go for a walk if you will,” and you say, “Then I will.” If these statements have the meaning that I have attributed to them, then each of them determines the speaker’s behavior, and represents itself as determining it, only in conjunction with the other’s statement. My statement represents itself as causing my walk only in the presence of your statement; your statement represents itself as causing your walk only in the presence of my statement; and the causal powers of these statements are in fact interdependent, as the statements themselves represent. Hence the behavior of each of us is settled—and is represented as being settled—only by both of us together.

    An interesting feature of this arrangement is that I do not take discretion over your behavior, nor you over mine; rather, you give me conditional discretion over your behavior, and vice versa. Each of us places his behavior under the joint control of both, by issuing an intention that’s conditional on the other’s intention. And each of us exercises the partial discretion that he’s been granted over the other’s behavior by exercising the partial discretion that he has retained over his own. The result is that each of us conditionally settles, and is represented as conditionally settling, one and same set of issues—namely, how both of us will behave—and we thereby categorically settle those issues together.

    ¤

    This case thus instantiates the phenomenon that initially puzzled us: a single making up of two minds. To make up one’s mind, in practical matters, is to become committed to a course of action. Yet each of us becomes committed to taking a walk only by the combination of our utterances, and not by his own utterance alone. Hence each of us has his mind made up by both of us; and we jointly make up both minds simultaneously.

    This joint making-up of minds is not the making-up of a joint mind. And if a commitment’s being oral or written entails that it isn’t mental, our commitment to taking a walk may not be a mental act or state at all.

    Of course, our commitment may have mental consequences—must have them, in fact, if it is to be effective. Each of us must remember what has been said, must be somehow motivated to make it true, and must have some idea of how much time he has in which to do so. As I have suggested, this persisting mental state, which constitutes the persisting force of our decision, is entitled to be called an intention, since remaining decided is a way of intending.

    If one of us has his share of the requisite memories and motives, then he will individually occupy the state of remaining decided to take a walk. But the decision that remains in force with him will be the joint decision, since only the joint decision has the force of a categorical commitment. If both of us have the requisite memories and motives, then both will be under the force of the joint decision.

    This state, in which each of us is individually under the force of a joint decision, deserves to be called a shared intention, but I don’t want to insist on calling it a single intention that is literally shared. In this case, each of us has his own private piece of the intention—maybe even his own private intention. What can be said is that our private intentions in this case consist in the lingering force of a single, shared intention. Although it is our individual minds that remain made up, they were initially made up jointly. Our minds individually remain as they were jointly made up: it’s that kind of shared intention.

    ¤

    My analysis of this case may seem to be circular, however. What I claim to be a single token of intention, in the making-up of our minds, is formed out of two individual speech acts. Don’t these acts need to combine into a joint speech act before they can combine into a joint intention? And doesn’t their forming a joint act depend on their arising from some antecedently shared intention?

    Well, I am committed to the proposition that when we exchange our mutually dependent decisions, it becomes appropriate to say of us that we have jointly decided to go for a walk. But this joint act of deciding is not a shared activity of the sort that requires an antecedently shared intention. It does require that we jointly tell a story to the effect that we will go for a walk (partly because of this story); but we can jointly tell this story without having shared an intention to tell it. We can jointly tell the story by acting with distinct, individual intentions: I, an intention to start a story and leave you to finish it; you, an intention to finish the story that I’ve started. Thus, the shared intention that we produce in telling our story doesn’t require any antecedently shared intentions.

    More importantly, the intentions that lie behind our stories are not intentions in the same sense of the word as the stories themselves.[28] That is, they are not practical conclusions, or exercises of discretion; they are rather aims or goals. If I decide aloud to take a walk, by saying “I’m going to take a walk,” I must utter these words with the intention of making a particular assertion—in fact, a self-fulfilling assertion that will count as a decision—and I may also utter them with the intention of starting a self-fulfilling story that you will finish. But the intention with which I utter these words is not a prior decision to utter them, much less a prior decision, by uttering them, to make a decision. For if I’m going to decide aloud to take a walk, then I don’t have to decide to make that decision; indeed, I had better not decide to make that decision, lest I leave too little remaining to be decided. (Deciding to decide to take a walk is almost already deciding to take it.)

    Rather, deciding aloud to take a walk requires that I utter words for the purpose—that is, with the aim or goal—of making a particular assertion and thereby making a decision. The intention with which I utter these words is simply the purpose, aim, or goal with which I utter them. Hence there is no circularity or regress of intentions: there is simply an act carried out with an intention in one sense of the word, while constituting an intention in a different sense.

    Sharing an intention in the latter sense is what looked difficult, if not impossible. But it’s not difficult at all. Sharing an intention can be as easy as saying “I will.”

    Notes

    1. This chapter was originally published in Philosophy and Phenomenological Research 57 (1997) 29–50 and is reprinted by permission of the International Phenomenological Society. Portions of this chapter were contained in a response to Michael Bratman’s “Shared Intention”, which was delivered at a conference at the University of Chicago in memory of Alan Donagan. I am grateful to Bratman and other participants at the conference for valuable comments and suggestions. (Bratman’s paper appeared in Ethics 104 [1993] 97–113.) An early draft of this chapter was presented at the 1993 meetings of the APA Pacific Division, in a symposium with Bratman and Margaret Gilbert. On that occasion I received helpful comments not only from Bratman and Gilbert but also from: Kent Bach, Annette Baier, Rachel Cohon, Eric Gampel, Ron Laymon, Al Mele, and Harry Silverstein. Thanks also to Jennifer Church, Roger Squires, and an anonymous referee for comments and discussion.return to text

    2. “Collective Intentions and Actions”, in Philip R. Cohen, Jerry Morgan, and Martha E. Pollack (eds.), Intentions in Communication (Cambridge, Mass.: MIT Press, 1990), 404.return to text

    3. “Shared Intention”, 99.return to text

    4. “Walking Together; a Paradigmatic Social Phenomenon”, Midwest Studies in Philosophy 15 (1990) 1–14, 7. See also Gilbert’s On Social Facts (Princeton: Princeton University Press, 1992). Note that Gilbert thinks of shared intention as requiring more than a “pool of wills”. For Gilbert, the pooling of wills is the general condition for producing a “plural subject”, which can be the subject of shared intention, shared belief, or various other shared states and activities. A shared intention, according to Gilbert, requires not just the pooling of wills but the pooling of wills that are embodied, specifically, in commitments to intend; whereas a shared belief requires the pooling of commitments to believe; and so on.

    Although I sympathize with Gilbert’s desire to explain the sharing of intentional states in general, I think that her interest in generality has led to an account of shared intention that is, in a sense, redundant. Anything that qualifies as a pool of wills, I think, already is a shared intention, without those wills’ having to be commitments to intend. I shall therefore be defending Gilbert’s conception of the pooling of wills rather than her conception of shared intention.return to text

    5. Ibid., 7.return to text

    6. The Philosophical Review 93 (1984) 375–405.return to text

    7. On the ambiguity of ‘intention’ see Gilbert Harman, “Willing and Intending”, in Richard E. Grandy and Richard Warner (eds.), Philosophical Grounds of Rationality: Intentions, Categories, Ends (Oxford: Oxford University Press, 1986), 363–80; and Change in View: Principles of Reasoning (Cambridge, Mass.: MIT Press, 1986), ch. 8.return to text

    8. I discuss this issue in my review of Bratman’s first book, Philosophical Review 100 (1991) 277–84, 282–4. Note that the distinction I draw here between goal and intention corresponds to that between boulesis (“wish”) and prohairesis (“choice”), as they are contrasted by Aristotle in Nicomachean Ethics III.ii. 7–11. Aristotle says that choice is confined to matters that are “up to us” (ta ephhemin), such as which means to adopt toward an end; whereas wish is often directed at the end itself, whose attainment is not within our control. As Aristotle puts it, we wish to be happy, and we choose means to happiness, but we cannot choose to be happy; we can only wish it. Note, again, that the English word ‘intention’ is often used for both phenomena. We would say not only that we intend (i.e., choose) to adopt the means but also that we adopt them with the intention (i.e., in the hope or wish) of attaining the end. Hence we use ‘intention’ and its cognates for both prohairesis and boulesis.return to text

    9. For the idea that an agent can form intentions only on issues that are (or at least appear to be) up to him, see Annette Baier, “Act and Intent”, Journal of Philosophy 67 (1970) 648–58. Baier has recently applied this idea to the problem of shared intention in much the same way as I do: ‘Doing Things with Others: the Mental Commons’ in Lilli Alanen, Sara Heinämaa & Thomas Wallgren (eds.), Commonality and Particularity in Ethics (New York: St. Martin’s Press, 1997), 15–44.return to text

    10. David Pears argues that the intention of trying to do something cannot be distinguished from the intention of doing it, since “if I were asked whether I had done what I had intended to do when I had tried and failed, I would give a negative answer” (“Intention and Belief”, in Bruce Vermazen and Merrill B. Hintikka [eds.], Essays on Davidson, [Oxford: Clarendon Press, 1985], 86). See also Hugh McCann, “Rationality and the Range of Intention”, Midwest Studies in Philosophy 10 (1986) 191–211. I think that Pears is falling prey here to the ambiguity of the word ‘intention’. If someone tries to do something, then success is ordinarily his goal, and so it is what he intends, in one sense of the word. But it isn’t what he intends in the sense of his being settled upon it; for if he could simply have settled upon succeeding, then he wouldn’t have needed to think of himself as engaged in a mere attempt. The question to which Pears envisions giving a negative answer is the question whether he has attained his goal. If asked, by a more subtle questioner, whether he has done what he was settled on doing, he might answer in the affirmative: “All that I was in a position to settle on was an attempt, and that I have done.”return to text

    11. This strategy is incorporated into the views of Searle and Bratman. In response to the objection that I raise here (and to similar objections raised by others), Bratman has written a paper demonstrating that an intention about what “we” will do, if held by each agent conditionally on being held by the other, can indeed allocate discretion appropriately, so that each agent’s intention settles matters that are up to him (Bratman, “I Intend that We J”, in G. HolmstronHintikka and R. Tuomela [eds.], Contemporary Action Theory, vol. ii [Dordrecht: Kluwer, 1997], 49–63; reprinted in Bratman, Faces of Intention: Selected Essays on Intention and Agency [Cambridge: Cambridge University Press, 1999], 142–61). Because Bratman quotes and discusses my objection, I have let it stand, despite being convinced that he has successfully answered it. In light of his paper, I would now be inclined to frame a less ambitious objection, to the effect that attributing unconditional “we” intentions does not solve the problem of shared discretion but merely presents it in a new form—albeit a form in which, as Bratman has now shown, it can be solved, by making the intentions mutually conditional. I would add that Bratman’s solution, which relies on making the tokens of a “we” intention mutually dependent, is consistent with the solution (due to Margaret Gilbert) that I defend below. I discuss these issues further in a review of Bratman’s Faces of Intention, Philosophical Quarterly 51 (2001) 119-21.

    Tuomela also speaks of “we-intentions”, but he does not conceive of them as intentions framed in the first-person plural. As conceived by Tuomela, we-intentions arise when each individual in a group intends to do his share in some common activity, and various conditions of belief and common knowledge are satisfied. Although Tuomela sometimes characterizes the intentions in such a case as having the form “We will do it,” they would seem instead to have the form “I will do my part in it.” Raimo, “What are Joint Intentions?”, in R. Casati and G. White (eds.), Philosophy and the Cognitive Sciences (Austrian Ludwig Wittgenstein Association, 1993), 543–7; “We Will Do It: An Analysis of Group Intentions”, Philosophy and Phenomenological Research 51 (1991) 249–77; “Actions by Collectives”, Philosophical Perspectives 3 (1989) 471–96; Kaarlo Miller, “We-intentions”, Philosophical Studies 53 (1988) 115–37.return to text

    12. Of course, I can settle whether the sofa will leave the ground if (or given that) you will lift your end; and you can settle whether the sofa will leave the ground if (or given that) I will lift mine. But the question is precisely how our individually settling these slightly different issues can involve a shared intention by which we jointly settle one and the same issue.return to text

    13. But see n. 11, above.return to text

    14. Intentionality (Cambridge: Cambridge University Press, 1983), ch. 3, 408–9. Actually, Searle draws a distinction between representations and “presentations”, and he identifies intentions with the latter. Insofar as I omit this detail, my application of Searle’s analysis may conflict with its intended interpretation.return to text

    15. I do not think that this account of intention is adequate as it stands. In particular, I think that an adequate account of intention must specify not only the effects of an intention-constituting representation but also its causes. An intention, in my view, is a self-described self-fulfilling representation that is partly caused by a desire for its fulfillment. As for how such a representation can be so caused, see p. 210, below.

    This conception of intention cannot be defended within the confines of the present chapter.return to text

    16. I argue at length for this view in Chap. 3, above, and in Practical Reflection, ch. 5.return to text

    17. For the idea that intentions are essentially commitments, see Bratman, Intention, Plans, and Practical Reason (Cambridge, Mass.: Harvard University Press, 1987), 4 ff.return to text

    18. As Jennifer Church has pointed out to me, I am ignoring the possibility (if it is indeed a possibility) that two or more people might jointly occupy a state that has representational content and yet constitutes neither a mental state of one of them nor a public representation between them. Such a state would truly belong to a “collective mind”, and it is no doubt what Searle and Bratman were thinking of when they disparaged the idea of literally sharing intentions.return to text

    19. One might think that I am begging the question here, since jointly telling a story is a collective action, requiring a collective intention. I discuss this issue below.return to text

    20. Actually, an agent may have many different standing motives that would be engaged by his announcing a future action and would consequently reinforce his preexisting motives for acting as announced. The desire not to have spoken falsely is only one example of such a motive. But one example is enough—enough, that is, to show that a person’s announcements of his future actions can sometimes be self-fulfilling.return to text

    21. For a full account of how self-described self-fulfilling predictions can constitute intentions, see ch. 4 of my Practical Reflection and Chap. 3, above.return to text

    22. Could this be the distinction that Searle intends to mark with the terms “representation” and “presentation”?return to text

    23. The difference between these pictures is analogous to that between two ways of picturing the cognitive act of judgment. A judgment might simply be the mental act of forming a belief—an act that qualifies as cognitive only by virtue of a having a cognitive state as its product. Alternatively, however, some judgments might be intrinsically cognitive acts, acts of mentally representing the world in the manner characteristic of cognition.return to text

    24. The causal condition on intending is difficult to state precisely. We do not want to say that an intention must actually cause the intended action, since there are genuine intentions that are revoked, overridden, or forgotten. In the text I have weakened this condition by saying that an intention must have a tendency to cause the intended action, but talk of tendencies is notoriously imprecise.

    My suspicion is that we cannot give precise conditions that are necessary and sufficient for intending. What we can give, I think, is a precise description of the paradigm case, with a commentary on some of the ways in which actual cases fall short of the paradigm. How short can a case fall while still counting as an intention will then vary with the context. Of course, the paradigm case of intention is one that actually causes the intended action.

    As I have mentioned in n. 15, above, I do not think that the paradigm case of intention can be characterized solely in terms of its content and effects. The paradigm case of intention not only causes the intended action, I think, but is itself caused by the agent’s motives for acting.return to text

    25. Of course, an utterance of ‘I intend’ can sometimes constitute a promise as well as an intention—in which case, it still commits the speaker even if it is insincere.return to text

    26. It is an interesting question whether this statement or intention can be adequately specified. When I intend to act on the condition that you intend likewise, what I intend depends on what would count as a “like” intention on your part; and yet what counts as a “like” intention depends on what exactly I intend. My intention is thus self-referential in a way that leaves its content ungrounded.

    The same problem arises in discussions of cooperative strategies for overcoming prisoner’s dilemmas. In that context, some philosophers have proposed a syntactic solution. See Peter Danielson, “Closing the Compliance Dilemma: how it’s rational to be moral in a Lamarckian world”, in Peter Vallentyne (ed.), Contractarianism and Rational Choice; Essays on David Gauthier’s Morals by Agreement (New York: Cambridge University Press, 1991), 291–322; and J. V. Howard, “Cooperation in the Prisoner’s Dilemma”, Theory and Decision 24 (1988) 203–13. According to this solution, my intention is conditional on your framing an intention that is syntactically similar—that is, framed in the same words. But this solution is clearly unsatisfactory, since it interprets my intention as committing me to act so long as you say “I will”, even if you’re speaking an idiolect in which “will” means “won’t” and “likewise” means “differently”.

    I think that the solution to the problem of ungroundedness, in this instance, is to recognize that “intending likewise” can mean “framing an intention with the same determinate or indeterminate content”. Suppose that I frame the conditional intention of taking a walk if you frame a particular intention; but that I risk failing to specify the latter intention because I attempt to specify it as similar in content to mine. And suppose that you frame the conditional intention of taking a walk if I frame a particular intention; but that you risk failing to specify the latter intention because you attempt to specify it as similar in content to yours. In that case, you and I have intentions that really are similar, precisely because they share the same potential indeterminacies of content. And in that case, their content isn’t indeterminate, after all, since it is perfectly determinate whether either person’s intention has the same potential indeterminacies as the other’s.

    I discuss the form of these intentions further in the Appendix to “Deciding How to Decide” (Chap. 9, below).return to text

    27. Davidson thinks that only some intentions containing an “if” clause are genuinely conditional. “[B]ona fide conditions are ones that are reasons for acting that are contemporary with the intention,” he says; other “if” clauses he dismisses as “bogus conditions” (“Intending”, in Essays on Actions and Events [Oxford: Clarendon Press, 1980], 83–102, at 94). He would therefore accept as genuine the condition “if you will” while rejecting as bogus the addition “if I recognize you as willing”, since the latter doesn’t allude to a reason.

    I think that Davidson’s discussion of conditional intentions is deeply confused. There is of course a difference between requiring the presence of a particular reason for one’s intended action, as in “I’ll do it if I like,” and acknowledging a limitation on one’s ability to perform the intended action, as in “I’ll do it if I can.” But Davidson offers no grounds for calling the former “if” clause a “genuine” condition while dismissing the latter as “bogus”. Each clause helps to specify what one is settling in framing one’s intention, which is precisely what the content of an intention serves to specify. Hence both intentions are conditional in content. For further discussion of Davidson on conditional intentions, see my Practical Reflection, 117–21.return to text

    28. On the ambiguity of the word ‘intention’, see n. 8, above, and the accompanying text.return to text