Michael Bratman’s work has convinced many of the falsity of the so-called Simple View, the view that an intention to A is necessary if an agent is to A intentionally. In his celebrated Video Game thought experiment an agent has mutually exclusive ends—ends that he knows cannot be jointly realized. Bratman argues that given a rational constraint on intentions, Strong Consistency, which requires that the agent’s intentions are jointly realizable, it is irrational for the agent to intend both ends. Nevertheless, the agent seems to perform an intentional action when he intentionally brings about one of his ends and is not irrational in doing so. Hence, it seems that an intention is not necessary for intentional action, and the Simple View is false (Bratman 1999: 113–116).

Objections to the Video Game thought experiment have been raised and answered, most recently in a discussion between Di Nucci (2009, 2010) and McCann (2010, 2011). This paper presents an objection that goes beyond this debate. Close examination of the case shows that the game-player is, after all, best understood as having mutually exclusive intentions. So, the example doesn’t seem to be a clear case of intentional action, A, without the intention to A. I believe that Video Game does not so much force us to abandon the Simple View as bring into sharp relief the puzzle of how we can rationally have and act on mutually exclusive intentions. Such mutually exclusive planning is not uncommon and a response to this puzzle is offered.

Although I am more sympathetic to the Simple View than many, my aim here is not to defend it. My twin aims are to object to Bratman’s putative counter-example (Sect. 2) and to suggest a solution to the puzzle of how an agent can rationally engage in mutually exclusive planning (Sect. 3). Briefly, the suggested solution turns on a claim about the nature of the subjective authority of intention. I argue that intention rationally requires of the agent that she engage in a reflexive attitude of evaluation. She can, however, impose rationally justified limitations on the self-evaluative attitudes that the intention requires of her. If, for example, she knows that her control over the fulfillment of an intention is limited in certain conditions, she is not required to evaluate herself negatively for failing to fulfill the intention in those conditions. It is plausible to suppose that it is irrational to have mutually exclusive intentions when having them guarantees that you are a failure by your own lights. However, in cases like Video Game the agent is exempt from the requirement to evaluate himself negatively when one of the intentions fails because the other is fulfilled. Hence, he is not guaranteed to fail by his own lights and isn’t obviously irrational. Ultimately, this calls Bratman’s Strong Consistency into question.

1 The Video Game Thought Experiment

Bratman invites us to consider the following thought experiment, which I have been calling Video Game. Imagine that I am ambidextrous and playing two video games simultaneously. The aim of each game is to

… guide a “missile” to a certain target … Let us now suppose that the two games are known to me to be so linked that it is impossible to hit both targets. If I hit one of the targets, both games are over. If both targets are about to be hit simultaneously, the machines just shut down and I hit neither target. … I know that although I can hit each target, I cannot hit both targets … I decide to play both games simultaneously; I proceed to try to hit target 1 and also to try to hit target 2. I give each game a try. (Bratman 1999: 114)

The example is a little involved, so let’s consider it in a little more detail. Suppose that you are playing two video games in which the aim is to hit a target. The games are connected so that you cannot hit both targets simultaneously, and if you hit one, the games shut down, preventing you from hitting the other. You are ambidextrous and adopt the strategy of trying to hit both. You are also skilled. Suppose that you hit one of the targets, T1. It seems that you hit it intentionally. If the Simple View is true, then you intended to hit T1. But if you intended to hit T1, you intended to hit T2—after all, you decided to try for both. According to Bratman, if you are rational, and you seem to be, you cannot have both intentions: you believe that you cannot hit both and “It should be possible (other things equal) for me to successfully execute all my intentions in a world in which my beliefs are true.” (Bratman 1999, 113) This is what Bratman calls the requirement of Strong Consistency on intentions. It is motivated by the role of intentions in co-ordination. According to the planning theory of intention that Bratman develops, an essential function of intentions is to allow agents to co-ordinate the fulfillment of competing desires. But if an agent believes that she has intentions that are inconsistent with one another—that cannot be fulfilled together—then these intentions do not play this co-ordinating role. They are in unreconciled competition, after all. Bratman maintains both that the game-player is rational, and that when he hits a target, he does so intentionally. Consequently, he cannot have an intention to hit T1 and an intention to hit T2, and we must have a case of intentionally hitting a target, e.g. T1, without the intention to hit T1. The Simple View is false.

Two of Bratman’s assumptions seem correct: the agent is rational in trying to simultaneously hit both targets, and if he intends to hit T1, he intends to hit T2. The following assumptions don’t seem to be consistent and generate the pressure to drop the Simple View:

  1. 1.

    Intentional Action: When the game-player hits a target, he hits it intentionally.

  2. 2.

    Simple View: Doing something intentionally involves intending to do that thing.

  3. 3.

    Strong Consistency: A rational agent should be able to successfully execute all of her intentions (in a world where her beliefs are true).

For defenders of the Simple View, 1 and 2 stand or fall together. But Bratman aims to show that we can prise these apart. He accepts Intentional Action and then focuses on the tension between Simple View and Strong Consistency. Given the assumption that the agent is rational, Video Game does not give us grounds for abandoning Strong Consistency. It is the Simple View that must be abandoned—1 and 3 are true, and 2 is false.

In this paper I grant that Intentional Action and Bratman’s other assumptions about the game-player are true, I leave open whether Simple View is true or false, and I will challenge Strong Consistency. The reason for the latter two stances is as follows. I argue that Video Game does not give us grounds for denying the Simple View (Sect. 2). In fact, as I mentioned, I think that Video Game presents, in a very nice way, the puzzle of how mutually exclusive planning is rationally possible. A proposal for solving this puzzle is suggested in Sect. 3. The positive proposal (Sect. 3) aims to show that the competition between the intentions to hit T1 and to hit T2 can be eased and that they can be rationally pursued in tandem with one another. Whether or not the Simple View is false, Video Game does not force us to give it up, and a careful look at the example gives us a reason to doubt Strong Consistency.

2 Video Game Does Not Give Us a Reason to Reject the Simple View

The argument that Bratman’s case does not give us grounds for rejecting the Simple View rests on exploring the question of what makes Intentional Action true if the agent does not have an intention. It cannot be that the agent foresees hitting the target or that he hopes or desires to hit it. Desiring and hoping for outcomes is insufficient for intentional action, and foresight is not promising. One may foresee that air molecules will be moved whenever one moves one’s body in the course of intentional action, but it doesn’t follow that one does this intentionally. Briefly, hoping, desiring, and foreseeing are contingently connected to agential control, the exercise of which is essential to intentional action. Intention, by its nature, involves the agent in exercising executive control over her conduct and her deliberation, so it is intimately connected to intentional action. In general, then, it is not as easy to discard intention in the psychological characterization of intentional action as one might think. But this will not be the claim here. The main claim is the more modest one that in the particular case of the game-player, we cannot make sense of his psychology without appeal to intention.

Given how influential the case has been, the argument here may seem naïve. If it were sound, surely Bratman and others would have realized that this putative counter-example to the Simple View must involve intention. There has not, however, been sufficiently close attention paid to the details of the case in the debate. Added to that, the steady stream of skepticism about the Simple View from work by philosophers, such as Gilbert Harman, Al Mele, and Michael Bratman, to the more recent barrage of studies in experimental philosophy has arguably and understandably dampened enthusiasm for exploring the appeal of the Simple View.

Three suggestions that attempt to make sense of the game-player’s psychology without attributing intentions to him will now be considered. Each is rejected as inadequate. Then I argue that we have excellent reason to believe that no such attempt could be successful.

2.1 Disjunctive Intention

Di Nucci offers an interesting way to understand the game-player’s mental states:

An alternative … would be to redescribe the agent’s intentions as an exclusive disjunction: the player intends to ‘hit t1 or t2’ but not ‘t1 and t2’: [(A v B) & not (A&B)]. It is not irrational for the player to intend to hit either of the targets but not both; in fact, that is just what the player is attempting to do. But the disjunctive intention … cannot be reduced to the only intention that can save SV, t1, because the disjunctive intention is true even when t1 is false (Di Nucci 2009, 75).Footnote 1

Intentions involve the agent prescribing a course of action to herself. The content of the intention is as follows: hit t1 or t2, but don’t hit t1 and t2! If this is right, the game-player prescribes a course of action for himself in which he is not to hit both of the targets. But it is unclear why the agent would, rationally speaking, prescribe such a course of action for himself. The game-player knows that not hitting both targets is inevitable and beyond his control, because the game shuts down as soon as one target is hit, so prescribing such a course of action for himself is fruitless. Consequently, any account of the agent’s mental states that does not involve this prescription is to be preferred. Second, this intention leaves open such questions as which target to attempt to hit at any given time or whether to attempt to hit both simultaneously. Of course, sub-plans can be developed to settle such issues. If the sub-plans are to capture the scenario as Bratman depicts it, then the agent should try simultaneously to hit both targets. However, if we characterize the game-player’s intention as Di Nucci suggests, then it is not clear that it can consistently generate a sub-plan in which the agent is to try simultaneously to hit both targets. The problem again seems to result from the second part of the exclusive disjunction: the prescription that the agent not hit both targets. For if this prescription is to be taken seriously, how can the agent square it with the prescription to try simultaneously to hit both targets?

Although it is appealing to think of the game-player as intending to hit either T1 or T2, it remains unclear how this proposal is to be worked out in a way that preserves the rationality of the game-player and captures the scenario as Bratman presents it. How is it, for example, that the game-player does not intend to hit both T1 and T2 if we don’t specify that he satisfies the second part of the exclusive disjunction? Even if there is some way to spell this proposal out, I believe that my positive proposal in Sect. 2 is at least as plausible a construal of the game-player’s psychology. The proposal is, moreover, grounded in a view of intentions that has independent plausibility.

2.2 Trying

Bratman says that the game-player may be trying to hit the targets rather than intending to hit them (Bratman 1999, 117).

The trying at stake might involve effortful activity towards an end where the agent knowingly suffers from incompetence. For example, a novice at tennis who attempts to get the ball over the net may, taking her lack of skill into account, focus on the easier goal of swinging in order to hit the ball rather than the goal of getting the ball over the net. Let’s suppose that the content of her intention is something like: swing the racket when the ball comes near! Strictly speaking, she does not intend to get the ball over the net. Perhaps, she is, nevertheless, trying to get the ball over the net. This raises the question of what makes it true that she is trying to get the ball over the net. A plausible assumption is that this is because she represents this as her ultimate goal. But, it matters what kind of mental state is involved in this representation. If it is a planning state, then there is trouble for Bratman’s challenge to the Simple View. However, let us set this question aside and grant that she is trying to get the ball over the net even without intending to do so.

Where it may be rational for the tennis-player to set aside the explicit goal of getting the ball over the net because of her lack of skill, it is not similarly rational for the game-player to set aside the goal of hitting the targets. Indeed, given his skill, it seems rational for him to intend to hit both targets. So, it seems that the scenario is not best understood in terms of this kind of case.

Although the game-player is confident about his skill, the difficulty of the task may make it rational for him not to intend to hit the targets. For example, Bratman maintains that in trying to hit the targets the game-player may merely intend to shoot at them (Bratman 1999, 117). Let us again set aside the worry about the nature of the mental state that represents the game-player’s ultimate goal and simply grant that in such a case he is trying to hit the targets without explicitly intending to because of the difficulty of the task.

The guiding role of intention keeps behaviour on course, so that, for example, the agent will make compensatory adjustments to his behaviour whenever they are needed. But if the intention is not to hit the targets, but to shoot at them, then hitting the target as a result of movements guided by the intention is accidental, and hitting is lucky. This calls into question whether we can say that he hits the target intentionally.

Bratman stipulates that the game-player plays with skill, so it may seem mistaken to say that hitting the target would be a matter of luck. But if his intention is merely to shoot at the target, his skill is being employed for that end and not for the end of hitting the target, and hitting would still be lucky. This alternative way of making sense of the game-player’s psychology does not preserve the power of the challenge to the Simple View.

A different kind of case involves the agent simply acknowledging that he is likely to fail because the task is very difficult. However, if this applies to the game-player, it does not follow that he does not intend to hit the targets, for this seems to involve simply having a belief that one may fail or adding to the content of one’s intention the acknowledgment that one may fail. This kind of case does not show how intentions to hit the targets are absent. Again, the scenario is not best understood in terms of this kind of trying.Footnote 2

2.3 Motivational Potential

Bratman maintains that an intention may have “motivational potential” that licenses the intentional tokening of an act-type that is not intended, but is, for example, merely desired. This depends on certain relationships—as yet unspecified—holding between intention and other psychological states. It may be thought that this will provide us with an “intention-free” way to understand the psychology of the game-player’s hitting. But, first, “motivational potential” remains a theoretical placeholder and this approach a promissory note, so it is unclear how much scrutiny this claim will bear. Second, the case of the game-player is meant to motivate the promissory note, it is not meant to presuppose its appeal. So, this is, dialectically speaking, a problematic move. Finally, whatever the fate of the promissory note for other cases of intentional action, it is not clear that it provides an account of the game-player. Let’s suppose that the game-player desires to hit the targets while intending only to shoot at them, that his desire causes his intention, and that his intention causes him to hit the target. Let’s also suppose that there is, in addition, some sort of relation between desire and intention that licenses the intentional hitting of the target. This relation between desire and intention must do justice to the psychological profile of the game-player. But as discussed earlier, if the game-player only intended to shoot at the targets, it seems to be a bit of luck if he hits one of them. At least, as long as the desire is not directly conduct-controlling, then hitting is lucky. However, if the desire is directly conduct-controlling, and moreover, controlling of the prolonged skillful activity of the game-player, then it no longer seems lucky. But the desire now also seems to play the role of an intention—it plays an executive role and is conduct-controlling. If the promissory note is developed in a way that allows us to make sense of the psychology and actions of the game-player, it is not clear that a meaningful challenge to the Simple View is preserved, because we must rely on intentions or intention-like states to make sense of him. This is a point to which I will return below. For now, I think that there is a good reason to be sceptical of this avenue of defense.

The foregoing attempts to explain how we can deny the Simple View and still make sense of the game-player’s psychology are unsuccessful. I will now argue that we have every reason to doubt that any attempt to characterize the game-player’s psychology without appeal to intention could ever be successful.

The game-player tries for both T1 and T2 as a strategy to maximize his chances of hitting one of the targets. To act on this strategy, he must, at a minimum, represent in his mind the goal of hitting T1 and the goal of hitting T2. In addition, these goal-representing states must control his conduct. After all, his conduct must be guided by the strategy if he is to act on it. We can also suppose that his decision to try for both targets forecloses further means-end deliberation—he settles on this approach. In fact, this supposition is required if we are to view him as having adopted a strategy. In short, the agent’s mental states play the role in planning agency that intentions do: they are goal-representing states that control conduct and deliberation. Given all of this, how could we coherently deny the Simple View for this case?

It is worth bearing in mind that most philosophers characterize intentions in functional terms, in terms, for example, of settling on a course of action, rationally requiring deliberation about means, rationally constraining practical deliberation about other ends, and so forth. There is no attempt to characterize intention in terms of a distinctive phenomenology or in other intrinsic features of the mental state. Given that any mental state that plays the distinctive functional role will count as an intention, and given that the goal-representing states that the strategic and skillful game-player must rely on to guide his conduct will play the functional role of intention, we do not have good grounds for denying that the game-player intends to hit T1 and to hit T2.

It is noteworthy that although there are a host of examples in the literature that putatively challenge the Simple View, Bratman’s game-player appears to be in a different league to others. It is a very clear case of intentional action and intention is not just absent, but apparently, rationally impossible. The problem that I see is that the game-player appears to be too strategic, too focused, and his skill is too obviously harnessed for the goals of hitting T1 and T2 to allow us to deny that he has intentions to hit T1 and T2. In an attempt to avoid the problem one might change the example so that the game-player is less strategic, focused, etc. The challenge is to make these changes while preserving the special status of the example. One worry is that the clear intuitions that everyone has that the case is one of intentional action may also evaporate if such changes are made. It is not implausible to suppose that it is these aspects of the game-player’s demeanour that point, not only towards the presence of intention, but also to the presence of intentional action. This is not to say that it is impossible to develop the example differently, but it is difficult to see how this could be done.

Bratman’s case seems to force a choice between Strong Consistency and the Simple View. But Strong Consistency only wins out if we ignore the issue of how to make sense of the game-player’s psychology in the absence of intention. There are excellent reasons to think that we cannot make sense of the psychology of the strategic and skillful game-player in the absence of intention, so this example is not the decisive counter-example to the Simple View that it is often taken to be.

3 The Rationality of Mutually Exclusive Planning

Whatever the fate of the Simple View, the case of the game-player poses a challenge. Assuming that he is rational, and that if he has an intention to hit one target he has an intention to hit the other, we must somehow resolve the tension in the following three claims:

  1. 1.

    When the game-player hits a target, he hits it intentionally.

  2. 2.

    The game-player’s intentionally hitting the target involved his intending to hit the target.

  3. 3.

    Strong Consistency: A rational agent should be able to successfully execute all of her intentions (in a world where her beliefs are true).

I take it that the arguments of Sect. 2 make it implausible to suppose that (2) can be given up for Video Game. I will argue that Strong Consistency should be re-evaluated.Footnote 3

The question that we face is whether the game-player can rationally intend to hit T1 and intend to hit T2 when he knows that he cannot hit both. More generally, is mutually exclusive planning—intending mutually exclusive ends—necessarily irrational? Suppose that in a very entrepreneurial spirit, I start two new businesses at the same time. At the outset, each business takes up about twenty hours of my time per week. But my intention is to expand each business to a certain point. At that point, each business would take about forty hours of my time per week. I know that I will only be able to give forty hours of my time in total to my work in any given week. Given this and other circumstances, I know that if my intentions come to fruition, one of the businesses will have to be sold or discontinued. But as I work on developing each business, I intend to bring it to a point at which I know it will demand forty hours of my time. Are my intentions for each business irrational as they cannot be satisfied together? It is not obvious that they are. It is worth noting that we frequently and fruitfully engage in mutually exclusive planning. We apply to mutually exclusive college programmes, we pursue mutually exclusive careers or romantic relationships, at least for a time. Such planning is not obviously irrational. And if the arguments of Sect. 1 are on track, it is not as easy to deny that the agents have intentions as it may, at first, seem. More often than not, such mutually exclusive planning is strategic. We may, like the game-player, have excellent reason to pursue two projects quite wholeheartedly even when we know that one must ultimately eclipse the other. A common feature of such cases is epistemic limitation: I don’t know if one or both or none of my businesses or applications or relationships will be successful, and this plays a role in my mutually exclusive planning.

I will confine myself to a discussion of the game-player in what follows. As it is a clear case of the general phenomenon, establishing that the game-player can rationally intend exclusive ends is an important step in coming to a better understanding of the general phenomenon. As it stands, there is considerable scepticism about whether such planning is rationally possible, but if I am right, this broad skepticism is, at least sometimes, misplaced.

3.1 The Subjective Authority of Intention

I have assumed throughout that the planning theory of intention, as that is developed by Michael Bratman, is fundamentally correct. Some of the features of intending that make them planning states are being settled on the intended course of action, the intention’s exerting rational pressure to engage in appropriate means-end reasoning, and the intention’s constraining one’s deliberation about further ends.

How do intentions play this controlling role? Intentions don’t operate by brute force, they are bound up with what the agent thinks is rationally required of her. Given this, it must be that when an agent forms an intention, the intention has a kind of authority over the agent. If it did not have such authority, it wouldn’t play the kind of controlling role that it plays in planning agency.

There are two things that we should note about the authority of intention. First, a given intention may or may not have objective rational authority over an agent. That is, it may or may not be based on reasons and it may or may not be a rationally justified requirement on the agent who has it. However, if it is to play its role in planning—controlling the agent’s deliberation and action in the relevant ways—it must have subjective authority. That is, the agent must take it that she should comply with the intention. In fact, if the agent genuinely intends to A, but fails to act on her intention without, in her own eyes, having any grounds for this, then she is normatively required to regard herself as a failure for not complying with the intention.Footnote 4 It is plausible to suppose that this reflexive evaluative dimension of intention is constitutive of the subjective authority of intention—what it is for something to seem to have authority over one if not that one regards oneself as evaluable depending on whether or not one complies with it? To put the same points in a different way that relies on treating intentions as commitments, the authority of intention involves the subjective authority of a commitment over the one who has committed. Qua commitments, intentions rationally ground self-evaluations by agents. For ease, I will call this the Self-Evaluation aspect of intention.

Second, it is widely accepted that intentions for the future contain prescriptions about how or when or in what circumstances the agent must act and intentions in action guide the agent in keeping her action on track as she performs it. Given the Self-Evaluation aspect of intention, it is plausible to suppose that intentions may play their guiding role by guiding the agent. Specifically, the content of an intention allows the agent to initiate or guide the action appropriately because it contains information about when the agent is criticizable and when she is not. To take a simple case of intention in action, let’s suppose that John’s intention in action is to wave his hand back and forth to attract the attention of Jane. In intending to wave he regards himself as criticizable, not just if he does not manage to produce a wave, but if the waving does not attract Jane’s attention. Because of this, he keeps waving until that happens, and indeed, if Jane seems not to notice, he may exaggerate the wave to ensure that she does etc.

It is tempting to say that the content guides the movement, not by informing the agent of when she is criticizable, but directly. This is more parsimonious, certainly, and there will be phases of action and whole intentional actions where the agent is not consciously consulting her intentions. In driving, I perform automatic actions that don’t involve this—I don’t consciously attend to what I am doing, let alone consciously think about the circumstances in which I am criticizable. But it doesn’t seem plausible to sideline the agent for all of her intentional actions. In a case like John’s it is not implausible to suppose that, and this is done with lightning speed, John is first guided by a prior intention to catch Jane’s attention, and then, having noticed Jane’s failure to see his waving, he is again guided by the intention in deciding to wave more vigorously. If the contents of intentions never guided the action by guiding the agent, it would be harder to explain how agents do register a sense of achievement when they have attained an end, or a sense of disappointment when they have failed, and it would be harder to explain how agents can recount in quite fine-grained detailed what precisely they are doing and under what conditions they will continue to pursue or abandon their ends. It would also not be easy to explain how failure to attain some intended end prompts explicit deliberation about whether to keep trying, whether to adopt an alternative means, or whether to modify the end etc.

Planning agency can be an intellectually demanding expression of agency and involves explicit deliberation and explicit consultation with what the agent has prescribed for herself in adopting her intentions. The game-player is a case in point—faced with the challenges of the video game and his own limitations, he strategizes about how to approach the difficulties, settles on a plan of action, and puts it into effect. Such exercises of agency seem to require that the agent knows when to stop trying and when to keep trying, and it seems that information in his intentions about when he is criticizable and when he is not allows him to do this.

3.2 Limitations on the Subjective Authority of Intention

So far, I have emphasized some features of the subjective authority of intention—intention involves self-evaluation, or more accurately, normatively requires self-evaluation. But none of this should lead us to lose sight of the fact that intentions are tacitly or explicitly conditioned in ways that delimit the scope of their authority. One kind of limiting condition derives from the agent’s background beliefs, principles, policies, values etc. The subjective authority of John’s intention to catch Jane’s attention is limited by his belief that one should not wave too vigorously in public and the worry that Jane is avoiding him since their difficult interaction when last they met.

The authority of an intention also depends on the agent’s continuing ability to execute the intention. The authority of my intention to run a marathon is tacitly conditional upon my not becoming paralyzed in the meantime. Although I do not consider this remote possibility when forming the intention to run, if I did become paralyzed and didn’t run, I would not be criticizable. The circumstance is one in which my intention has no authority, because, rationally speaking, I cannot be required to do what I am unable to do.

Limitations on the authority of intention pertain not just to the conditions in which it is rationally required of the agent that she act in accordance with it, but also to whether failure to complete the relevant task implies that she is criticizable. For example, before I go to bed, I may form the intention to lock the front door. I realize that my partner may have already locked it, but I don’t want to wake them to ask. However, if I act on my intention but fail to lock the door because my partner has already done so, I am not criticizable. My intention is best understood as being to lock the door in the event that my partner has not locked it. In this case, unlike the case of the marathon runner, the agent doesn’t know whether she is able to successfully execute her intention, but given her lack of knowledge and her desire to have the door locked, it is rationally required that she try to lock it.

If the examples are a guide, the scope of the authority of an intention can be rationally limited in at least two different ways:

  1. a.

    Action Limitation: A rational limitation on the range of conditions in which an intention requires action from an agent.

  2. b.

    Evaluation Limitation: A rational limitation on the range of conditions in which an intention normatively requires a negative self-evaluation from the agent for not fulfilling the intention.

3.3 Limitations on the Authority of The Game-Player’s Intentions

We can use the distinction between Action Limitation and Evaluation Limitation to understand how the game-player can rationally intend to hit T1 and intend to hit T2. Specifically, we can understand the game-player as limiting the authority of each intention along the lines of Evaluation Limitation. For example, his intention to hit T1 involves the following thought:

I will hit T1, but I am not necessarily criticizable if I don’t hit T1—in circumstances in which I hit T2, I wouldn’t be able to hit T1, so I am not criticizable in such circumstances.

The same thought is involved, mutatis mutandis, in his intention to hit T2. Like the agent who intends to lock the door in the event that their partner has not locked it, the game-player intends to hit T1 in the event that he doesn’t hit T2 (and to hit T2 in the event that he doesn’t hit T1). Let’s suppose that the game-player hits T1 and doesn’t hit T2: the intention to hit T2 does not ground a negative self-evaluation even though the agent acts wholeheartedly on the intention and fails to fulfill it. The authority of the intention to hit T2 is rationally limited as follows: if the agent doesn’t hit T2 because he has hit T1, he is not criticizable.

Construed in this way, we can also see how the intention plays an effective role in guiding the agent’s action. Intentions guide the agent by indicating when she is criticizable and when she is not, when she should stop trying and when she should not. Because T2 cannot be hit in circumstances where T1 is hit and because the agent has achieved his overall goal—hitting one of the targets—the game-player is not rationally criticizable for not hitting T2 and can stop trying to do so.

It is worth stressing that the game-player does not form conditional intentions. That is, my proposal is not that the intention to hit T1 only becomes operative when the game-player doesn’t hit T2 and vice versa. This would be rational, but does not reflect how Bratman describes the case: the game-player acts on both intentions simultaneously. Interpreting my proposal as involving the claim that the intention to hit T1 and hit T2 are conditional intentions would involve mistaking Evaluation Limitations for Action Limitations. On my view, the intention to hit T1 and the intention to hit T2 require action in the same range of circumstances—if, for example, there are Action Limitations on the intention to hit T1, the same limitation will apply to the intention to hit T2 etc. It is not that action on T2 is required only when action on T1 fails and vice versa, as conditionality in intentions implies. However, the conditions in which the intentions normatively require certain self-evaluations do differ. The intention to hit T1 normatively requires negative self-evaluation when T1 is not hit AND T2 is not hit and no negative self-evaluation if T1 is not hit BUT T2 is hit. The intention to hit T2 normatively requires negative self-evaluation when T2 is not hit AND T1 is not hit and no negative self-evaluation when T2 is not hit BUT T1 is hit. The intentions are subject to different Evaluation Limitations but the same Action Limitations, whereas conditional intentions are subject to different Action Limitations.

Although it is not my aim to closely examine Strong Consistency here, given the way that Bratman and others have set up the argument against the Simple View, it is important to at least briefly consider whether it has been violated. My tentative conclusion is that it has indeed been violated and that what I have said about the Self-Evaluation aspect of intention gives us good reason to re-evaluate it. My proposal aims to show that two intentions that cannot be fulfilled together do not have to be in the kind of irreconcilable competition that precludes their role in rational co-ordination of the agent’s action and deliberation.

We can think of the puzzle of the game-player as being the puzzle of how he can both sincerely intend to bring about an end, such as hitting T1, but also take a course of action that will thwart the achievement of that end (in acting on the intention to hit T2). How can such an agent not be criticizably irrational when they themselves put the achievement of an intended end out of reach, or, put another way, how can they sincerely intend an end that they simultaneously work against?

The answer to the puzzle is, first, that the agent acknowledges limits on his ability to successfully execute both of his intentions at the outset, and absolves himself of criticizability when the limits of his control have been reached—specifically, when success in one intention precludes success in the other. Second, the limitation on criticizability is rationally justified by the fact that thwarting the achievement of one of his ends is the rational price to pay for the only success he is able to achieve in the context: hitting one target is the best outcome and hitting both is impossible.

Given the Self-Evaluation aspect of intention, the irrationality of holding and acting on competing intentions derives from the fact that a rational agent should not make demands on herself that she knows that she cannot meet as this would involve the agent regarding herself as criticizable for not doing what she knows she couldn’t have done. But if I am right, the game-player is not guilty of any such irrationality—he places rational limitations on his negative self-evaluation.

In addition, intentions that are impossible to fulfill together give no clear guidance about the circumstances under which the agent is rationally relieved of the burden of trying to fulfill the intention. For, intentions that impose impossible requirements have no rational authority on when they require action from the agent and when they do not. And, of course, such intentions do not allow for effective planning, because although, qua intentions, they should constrain further plans, it is not clear what rational authority they could have to impose such constraints if the agent knows that she cannot fulfill them.

The Self-Evaluation aspect of intention allows us to spell out how the game-player’s intentions allow him to co-ordinate his actions. As mentioned, sacrificing the success of one intention is the rational price to pay in the circumstances for the chance to hit one of the targets. Not hitting T1, for example, is not criticizable as long as T2 is hit, and because of this the game-player is relieved of the burden of trying to hit T1 in such circumstances in spite of his failure to fulfill his intention. Having incorporated Evaluation Limitations into his intentions, he has clear guidance about when to abandon attempts to fulfill an unfulfilled intention. The intentions to hit T1 and T2, although they cannot be fulfilled together, can nevertheless be rationally held and acted on simultaneously. The upshot is that Strong Consistency appears to be too strong.

Having said that, it is not my primary aim to argue against Strong Consistency. It may be, for example, that the way I have described the cross-linkage between the intentions to hit T1 and to hit T2 means that they are not two distinct and competing intentions, but one intention. There may be, in this, a way to defend Strong Consistency. And indeed its deeper motivations and implications deserve more space than I give them here. Given my twin aims in this paper—a challenge to Video Game and addressing the puzzle of mutually exclusive planning—fully assessing Strong Consistency is work for another paper.

4 Concluding Remarks

My aims in this paper are twofold. The first is to argue against the view that Video Game seals the fate of the Simple View. Put simply, the case is hard to make sense of unless we attribute an intention to hit T1 and an intention to hit T2 to the game-player. If the task is to force a choice between the Simple View and Strong Consistency, this case is not up to that task. The second aim is to use the case as a window into the rationality of mutually exclusive planning. Given this phenomenon’s prevalence, its seeming reasonableness, and its apparent incorporation of intentions, it behooves a robust philosophical psychology of action to more fully explore its rational viability. The Self-Evaluation aspect of intention is, I believe, a good place to start. Exploring the nature and limitations of the subjective authority of intention promises to shed light on the ways in which agents may render competing intentions rationally compatible. The relatively neglected topic of the subjective authority of intention promises to shed some light, not only on mutually exclusive planning, but on rational planning agency more generally.