Velleman (2007) points out a methodological assumption common in functionalist psychology, namely the assumption that our attitudes or cognitive faculties have a function not just in the sense that they have a causal role but in the sense that they have a purpose, something they are designed to do and thus ought to do. Functions in this latter sense are commonly called teleofunctions. This methodological assumption needs not entail a belief in some intelligent designer. Instead, it can be cashed out by appealing to evolutionary theory and to natural selection as a blind designer. Typically, the evolutionary story goes like this: a trait has the teleofunction of producing effect E just in case producing this effect conferred some benefit that contributed to the reproductive success of organisms endowed with the trait and, thereby, to the propagation of the trait itself. This methodological assumption, when it guides our inquiry into intentions, leads us to take the question what intentions are for, i.e., what purpose are they meant to serve, as necessarily meaningful and demanding an answer.
Velleman discusses two teleological stories meant to answer this question. He links the first story to Bratman’s theory of intentions (1987) and the second to Anscombe’s theory (1963). I start with the story inspired by Bratman’s theory.
We are, in Bratman’s words, planning agents regularly making more or less complex plans for the future and guiding our later conduct by these plans. This planning ability appears to be if not unique to humans at least uniquely developed in the human species. People can, and frequently do, form intentions concerning actions not just in the near but also in the distant future. Why should we bother forming future-directed intentions? What purposes can it serve? What benefits does it bring us? What features of future-directed intentions allow them to serve these purposes?
Bratman offers two complementary answers to that challenge. The first stems from the fact that we are epistemically limited creatures: our cognitive resources for use in attending to problems, gathering information, deliberating about options and determining likely consequences are limited and these processes are time consuming. As a result, if our actions were influenced by deliberation only at the time of action, this influence would be minimal as time pressure isn't conducive to careful deliberation. Forming future-directed intentions makes advance planning possible, freeing us from that time pressure and allowing us to deploy the cognitive resources needed for successful deliberation. Second, intentions once formed commit us to future courses of action, thus making the future more predictable and making it possible for agents to coordinate their activities over time and to coordinate them with the activities of other agents. Making deliberation and coordination possible are thus the two main benefits that accrue from a capacity to form future-directed intentions.
What makes it possible for future-directed intentions to yield these benefits is, according to Bratman, the fact that they essentially involve commitments to action. Bratman distinguishes two dimensions of commitments: a volitional dimension and a reasoning-centered dimension. The volitional dimension concerns the relation of intention to action and can be characterized by saying that intentions are “conduct-controlling pro-attitudes” (Bratman 1987, p. 16). In other words, unless something unexpected happens that forces me to revise my intention, my intention today to go shopping tomorrow will control my conduct tomorrow. The reasoning-centered dimension of commitment is a commitment to norms of practical rationality and is most directly linked to planning. What is at stake here are the roles played by intentions in the period between their initial formation and their eventual execution. First, intentions have what Bratman calls a characteristic stability or inertia: once we have formed an intention to A, we will not normally continue to deliberate whether to A or not. In the absence of relevant new information, the intention is rationally required to resist reconsideration: we will see the matter as settled and continue to so intend until the time of action. Intentions are thus terminators of practical reasoning about ends or goals. Second, during this period between the formation of an intention and action, we will frequently reason from such an intention to further intentions. For instance, we reason from intended ends to intended means or to preliminary steps. When we first form an intention, our plans are typically only partial, but if they are to eventuate into action, they will need to be filled in. Thus intentions are also prompters of practical reasoning about means. Third, because intentions are commitments to action, our intentions should be jointly executable. Finally, taken together the volitional and the reasoning-centered dimensions of commitments help explain how intentions can promote coordination. They provide support for the expectation that agents will act as they intend to and these expectations are central in turn to both inter- and intra-personal coordination. In particular, this is what motivates the rational agglomerativity requirement on intentions, i.e., the requirement that my intentions be jointly executable.
The benefits that accrue from a capacity for intentions are, ultimately, pragmatic benefits. As Bratman puts it, future-directed intentions “enable us to avoid being merely time-slice-agents” (1987, p. 35). Instead of constantly starting from scratch in our deliberations and simply weighing current belief-desire reasons, intentions allow us to become temporally extended agents. They provide a background framework that allows us to expand the temporal horizon of our deliberation while at the same time narrowing its scope to a limited set of options. In so doing they contribute in the long run to our securing greater desire-satisfaction than simple desire-belief practical reasoning would.
Velleman (2007) sees three main problems with Bratman’s pragmatic account of what intentions are for. The first problem concerns the status and role of present-directed intentions. On Bratman’s account, a future-directed intention requires a present-directed intention to convey its motivational force and guide the action once the time to act is seen to have arrived. Bratman identifies no further role or function of present-directed intentions beyond conveying the motivational potential of future-directed intentions. At the same time, he insists that intentional actions, whether or not they are preceded by future-directed intentions, always involve present-directed intentions. This leaves us with a potentially large class of spontaneous intentional actions that involve present-directed intentions but are not preceded by future-directed intentions. These intentions do not incorporate the results of any prior deliberation, they don't set the stage for any further planning and they don't provide a basis for any coordination. The first worry raised by Velleman is thus that these intentions do not seem to serve any of the pragmatic purposes that, on Bratman’s account, constitute the raison d'être of intentions. Second, Velleman points out that a similar worry arises for the intentions involved in various cases of planning. He illustrates his point with a voting example. He argues that while there may be good reasons for my starting to think about my vote in advance, such as giving me sufficient time to deliberate, there doesn't seem to be any good reason for settling in advance of my arrival in the voting booth whom I will vote for. On the contrary, settling in advance seems to carry potential costs, by making me resistant to reconsideration, without procuring any benefits, since the actual act of casting my ballot doesn't require any particular prior preparation. Thus, at least in these cases where no further planning is needed once one has settled on a course of action, it is unclear what purpose settling in advance could serve.
Velleman’s third worry relates to Bratman’s view that intention need not imply belief. Bratman indeed maintains that “there need be no irrationality in intending to A and yet still not believing one will”, but that, in contrast, “there will normally be irrationality in intending to A and believing one will not A” (1987, p. 38). According to Velleman, this view of Bratman’s leaves much of his functional account of intentions unmotivated. In particular, it becomes unclear why in intending to A, an agent should be rationally required to identify means of A-ing or to rationally constrain her subsequent practical reasoning by ruling out options inconsistent with her A-ing, if she is agnostic whether she will in fact carry out her intention. Similarly, it becomes unclear why we should impose an agglomerativity requirement on intentions. As Velleman points out, it is unclear why intentions should be jointly executable if the agent can be agnostic as to whether they will be executed.
In my view, Velleman’s third worry is exaggerated. Firstly, while Bratman indeed maintains that an intention to A does not require belief that one will A, he insists at the same time that an intention to A normally supports the belief that one will A. Secondly, Bratman also makes the point that agnosticism about whether one will act as intended does not directly undermine coherent planning but makes it more complex, leading us to form conditional intentions and plans for both failure and success to act as intended. Of course, the viability of such a move depends on agnosticism being the exception rather than the rule; otherwise, we would have an unmanageable proliferation of conditional branching in our plans.
Velleman’s first and second worries run deeper. If the only purposes of intentions are the pragmatic functions Bratman identifies, then there appear to be many instances where intentions don't serve these purposes or where serving them is actually counterproductive. This may be taken to indicate that Bratman’s account is incomplete and that he has overlooked some of the functions intentions serve. This line of thought can be pursued in two different directions. On the one hand, we may try to identify further pragmatic functions that intentions, including present-directed intentions, could serve; on the other hand, we may look for non-pragmatic functions that intentions could serve. As we will now see, Velleman explores the second option, turning to Anscombe’s theory of intentions in search of an answer. In contrast, what I will do myself later in this paper is explore the first option, giving it a social twist.
Velleman argues that Bratman’s account of intentions misses an important function of intention, a function that is a central theme in Anscombe’s theory of intention. In her book Intention (1963), she argues that intentions provide us with a special kind of self-knowledge and claims that this knowledge is special in two ways. It is knowledge of our own intentional actions, i.e., knowledge not just of what one is attempting to do, but of what one is actually doing, and it is knowledge without observation. Much philosophical ink has been spilled on how exactly these two claims should be interpreted. Following Falvey (2000), Velleman favors a reliabilist interpretation of these claims. According to this interpretation, knowledge of one’s own intentional actions is non-observational because it is given by the content of our intentions and intentions in turn normally constitute (practical) knowledge of our own intentional actions because they reliably cause the facts that make them true. Note also, that on this reliabilist reading, Anscombe’s claim is not that the content of our intentions provides us with infallible knowledge of what we are doing. To say that there normally exists a reliable connection between our intentions and actions is not to say that there cannot be cases when this connection does not obtain. However, as Velleman emphasizes, on Anscombe’s account, failures of reliability undermine not just the epistemic status of intentions, they also undermine the intentionality of actions. If my intending to A does not reliably cause my A-ing, then, on the one hand, my intending to A will not amount to knowledge that I am A-ing and, on the other hand, my A-ing when it happens will be an accident rather than an intentional action. According to Anscombe, intentional actions are those “to which the question ‘Why?’ is given application” (1963, p. 9) and having practical knowledge is knowing a description of what one is doing, has done or is proposing to do that answers the question “Why?". Thus, the basic epistemic function of intentions is to provide us with a form of self-knowledge and self-understanding qua intentional agents.
According to Velleman, acknowledging this epistemic function of intentions does much to alleviate the worries raised by Bratman’s practical account. With respect to the first worry – that present-directed intentions serve no purpose – one can now argue that while they might serve no practical purpose they still serve an epistemic function. With respect to the second worry – that on many occasions making one’s mind in advance serves no pragmatic purpose –, one can now reply that in matters that are important to one’s self-conception, uncertainty about one’s future behavior is both uncomfortable and undesirable and that forming an intention allows us to gain self-knowledge and avoid this mental discomfort. With respect to the third worry – that absent a strong enough connection between intention and belief, it is unclear why intentions should be subject to the practical rationality requirements emphasized by Bratman–, Anscombe’s theory regarding the epistemic function of intentions lets us see how the epistemic role of intentions could support their pragmatic functions.
The story as told so far suggests that we should think of the epistemic and pragmatic functions of intentions as complementary. However, as Velleman points out, it still leaves us with two possible hypotheses or creation myths about the origin and ultimate purpose of intentions. On the pragmatic creation myth, the ultimate purpose of intentions would be pragmatic and their epistemic function would be subservient to their pragmatic functions, but may occasionally exemplify re-purposing: “That is, intention might have been designed to embody self-knowledge for the sake of facilitation coordination, but it might then be used on occasion, for the sake of self-knowledge alone, when coordination isn't necessary” (Velleman 2007, p. 208). By contrast, on the epistemic creation myth, the ultimate purpose of intentions may be to embody self-knowledge, and the pragmatic functions of intentions might have emerged simply as a fortuitous by-product of self-knowledge.
While Velleman has more sympathy for the epistemic than for the pragmatic creation myth, he thinks both should ultimately be rejected. In the next section, I'll consider his reasons for rejecting them, discuss the alternative story he proposes, and advance my own reasons for being skeptical about this story.