1. Forms of Intention

It will help us to begin with intention-locutions. These indicate some of the variety of intentions themselves, though, as will soon be evident, we should not uncritically assume that every element in the locutions considered reveals an important aspect of intention itself. Some of the locutions are infinitival, e.g. ‘to send a letter’. Some designate umbrella intentions to bring about something, for instance an improvement in relations with certain in-laws. Some designate other indefinite intentions, e.g. to try to A, where A-ing is an act but no act constituting the would-be trying is specified. Some non-actions represented by verbs should also be included, say to forgive and to be a good friend.Footnote ²

A different category of intention-ascriptions includes the objectual, for instance intending some object for a purpose or intending someone to serve on one's committee. I can intend a sharp carving knife for a Thanksgiving turkey. Such cases show that we cannot take intention-ascriptions at face value – as clearly indicating what is intended. The intentional object of this “object intention” is not the knife but some set of acts concerning it. As in other cases of intention-ascription, these (if true) require the agent to have a concept of some action, but – except where one intends an agent to be or do something – need not indicate a particular act or a specific agent. In the knife example, the action, carving, is specified, but no agent is indicated. If I intend a bottle of wine for my hostess, this leaves open both who hands it to her (say my companion or me) and just what, if anything, I intend her to do with it.

A less common and less often discussed intention-ascribing locution is subjunctive, for instance, ‘intending that my students learn punctuation’. Normally one doesn't intend that something be so without having both a related intention to and a tendency to intend to do certain things which, in the case in question, are implicit in the intention or appropriate to fulfilling it. If I intend that my students understand why utilitarianism should not be described as the view that right acts are those that promote the greatest good for the greatest number, I likely have in mind and intend a series of explanatory acts. But suppose that in a prospective conversation with an old friend the latter says ‘I intend that my children be well educated’. In a mood of identification with the friend, one could form a similar intention, say that one's children be freed of anxiety about tuition. At that moment one is of course disposed to form action-intentions, but one need not do so. If, however, this subjunctive intention persists, doing so is expectable.

We should also consider intending to be, for instance to be a friend to someone. This may be best understood in terms of a kind of intending that, say intending that one give the person support in stressful situations. This subjunctive construction is important in part because it can also encompass the content of a person's intentions to be of a certain character, say kind or more understanding. There may be still other intention-locutions best understood subjunctively, and later I will illustrate an important aspect of subjunctive intentions.

All of the cases considered are instances of intention conceived as a practical attitude, in a sense implying an essential connection with action (at least with some act-type). Even if, as with objectual and subjunctive intentions, no particular act is entailed by the concept of the intention in question, there remains a self-evident connection with action: the intention-ascription portrays the person as in some sense aiming at doing something. The ascription may not indicate, and the agent may not know, just what act-type(s) are intended or otherwise (if only in some implicit way) encompassed. Precisely for this reason, objectual and subjunctive intentions cannot be reduced to infinitival ones: they entail a readiness to form the latter but do not require a determinate set of such intentions. The same points hold for desire, to be sure, but desires need not imply the kind of commitment of the will implied by intending. Many of our desires even to do are far from committing us – and may be overwhelmingly opposed by our deepest plans.

2. Sources of Intention

One of the routes to understanding intention is by way of its conceptually relevant sources. These are not just genetically informative, but indicative of elements in its nature. They are also important for understanding moral responsibility. We are responsible for intentions – in the sense that we are subject to praise or criticism for forming or harboring certain intentions – and so should do what we can both to form or nurture those that go with realizing desirable goals and to prevent formation, or at least, harboring, those directed to what is undesirable or wrong. Talk of responsibility for intentions may suggest their formation prior to action, but my concern with sources of intention extends to cases in which (as with actions as immediate responses) intentions are not formed before the intended action.

First, consider volitional routes to intention-formation. We often decide to do something. Deciding to A is a kind of manifestation of will, and as such it entails intending to A, even if not an awareness of intending to A. Resolving to A seems similar. But could there be hollow resolutions, as with some that one might list on New Year's day? Doubtless, but where resolution is volitional and not merely verbal, it is like decision (though ‘decide’ may also have “hollow” uses). The generic notion here is willing, which is familiar in at least some cases of trying as well as in some instances of achieving an aim. Here it suffices to note that we often consider options, decide on one, and thereby form an intention to do something or to try to do it. It of course does not follow, and is not true, that all intentions emerge from decision.

A second route to intention-formation is doxastic. A common case occurs when we form an intention upon forming an instrumental belief given predominant motivation to which the belief is relevant in a certain way. Suppose I am contemplating a prospect I don't see how to realize, say attending a favorite play. Then coming to believe that A-ing, say phoning a friend who has a subscription, will realize it can immediately lead to my forming the intention to A. Belief combined with trying constitutes another route to intention-formation. If one is trying to A, one already hopes or intends to A. If, however, there is some instrumental action, say B, that one intends as a way to bring about A-ing, one might suddenly realize that C-ing will definitely work and come to intend to C instead. Trying to open a jar by tapping its top may not be succeeding when one suddenly remembers an available pipe wrench. Straightaway one forms the intention to apply it.

A third case in which intentions are formed is phenomenologically less prominent. Suppose one has the habit of reaching for the right shoe first when dressing. In the morning, seeing it across the room, one might then intend to get it without even considering the option of reversing one's procedure. Reaching out a hand when introduced to someone is similar, though given certain cues one may be ready to form the intention to abstain. It is interesting to note here that habit need not imply any specific conceptualization of the action in question, say to put on the right shoe first. The intention might be simply to put on that shoe now, where ‘that’ designates the right shoe one is viewing. For some cases, moreover, an objectual intention might do; I might, as I take a pair of shoes from the rack, simply intend the right and left ones for the respective feet without the notion of order's figuring in the content of my intention.

In some of the cases illustrated, there is an event that triggers intention formation: either a belief forms and completes the sufficient conditions for intending, where there is already a desire to do the deed, or a habit is activated by ongoing activity or by circumstances. Perhaps we can also speak of a trigger where an obviously needed response is intended in suddenly encountered circumstances, say where an opportunity to A is suddenly seen. Walking in England, an American starting to cross a street might, having looked for traffic the wrong way, unexpectedly see an approaching car and form an intention to wait for the green light. In this fourth case there are standing beliefs connected with standing motivation to avoid getting hit. Yet the behavior is not mere reflex. Neither is withdrawing a hand from a hot stove, though some instances of the corresponding movement might be.

A fourth kind of case is a sudden attraction to doing something, as where it fits some overall purpose one has or even occurs as part of a plan or prospect one suddenly wants to realize. Planning a party, I may be reminded of Jack, perhaps by glimpsing a letter on the table. I immediately form the intention to invite him to the party.

In all four cases considered, there is some event in consciousness that yields an intention, something in consciousness that, even if not a trigger of intention, is a phenomenally perceptible element. There is a fifth case that does not entail such perceptibility. It occurs where standing motivation to A becomes predominant over competing motivation or, in another kind of case, over inertia. The agent may be torn between attending a play and attending a symphony scheduled for the same night. In time, one of the desires can become stronger than the other in such a way that, on being asked what one will do that night, one can find oneself truly saying one is going to go to the symphony. One might have heard the composer's music and enjoyed it, which might, without one's realizing it, strengthen the desire to attend the symphony; but no such event is necessary for spontaneous motivational predominance.

3. The Range of Objects of Intention

Much about the range that the objects of intention can have is now apparent in outline; but there are important cases not so far identified. Here it will suffice to focus just on objects of intentions to. The ‘to’ is commonly followed by specifications of act-types, including activity-types. The usual paradigms of intention discussed in the literature are single-act intentions. But, in addition to intending, say, to shake someone's hand as a greeting, there are complex sequences we can intend, though the expected conditions for doing a deed need not all figure in the content of intending it.Footnote ³

If we think of the objects of intentions to as limited to such acts and activities, we miss an important element that often comes into the scope of an intention. What we do is done in some way, and sometimes this manner of action is important and intended. To be sure, a manner of an intended action need not itself be intended. Intention is not closed even under believed consequences, including expected manners of performance: I can intend to phone someone at 7 a.m. and believe that this will wake a teenage child, yet not intend to wake the teenager; and we can intend to ask help of someone, believe that we will do it timidly, but not intend to do it timidly.

One might think that acts described by an adverb of manner are simply acts simpliciter and that therefore intending to A M-ly, where M is a manner of acting, reduce to intending to B, for some act-type B, identifiable without such an adverb. Suppose there is an at least approximately equivalent act-type, as in some cases of declining a request decisively and refusing it. It does not follow that to B is the content of the original intention. The intention may have a precise content lost by any but an exactly equivalent substitution. Moreover, in principle any act can be performed in more than one way, and to provide an act-name that absorbs the adverbial modifier (as ‘yelling’ absorbs voicing in a certain loud tone) can be misleading, and even when it is not false, it may easily conceal an important dimension of behavioral assessment. It matters greatly how we do what we do, for instance whether we correct a child's speaking error gently or harshly.Footnote ⁴ Eliminating adverbs of manner in favor of “simpler” act-reports may both reduce clarity and impair normative appraisal.

A further dimension of complexity in the objects of intention is that of ‘scripts’. These might be conceived as encapsulated complex intentional contents. Consider intending to play a piano piece that is memorized so well as to be second nature. When one plays it in fulfillment of the intention, one intends each note (that goes right), though the intention might be objectual; but one could not even describe the whole sequence, nor need one have the thought of each behavioral element in the script as the occasion to act occurs. A simpler case would be a script for leaving a message when one is issuing phone invitations to a party. Intending to invite Liz via her voicemail might encapsulate a series of sentences and even an intonation, speed, and volume. There is no sharp distinction between scripted content of an intention expressible without a detailed conception of an activity and, on the other hand, an unspecific, perhaps generically conceived, activity content, but at least this much is clear. A script is relatively determinate in that it expresses or prescribes a behavioral sequence that, at least in simple cases, can normally be described in some detail in advance, whereas some activities one can intend, such as talking with a friend about a film, are far from routine and are describable beforehand only in outline.Footnote ⁵

Might the scope of intentions also include a purposive element, say a reason for the intended act as well as a conception of the act itself? To see the problem, consider intentions whose objects appear to be acts purposively described. Suppose I promise to help Clarence clean out his messy garage in the spring and by then he inherits a fortune and doesn't need my help. Asked why I'm taking the trouble when Clarence has hired someone to help and perhaps even forgotten my promise, I might say: ‘I intend to help him because I promised to’. Note first that this is ambiguous. It might mean roughly (1) ‘The reason why I intend to help him is that I promised to’ or, by contrast, (2) ‘I intend to help-him-because-I-promised-to’, where this ostensibly specifies the content of the intention and not why one has it. Can an intention have such a content? We can also be said to intend to A in order to B, say to meet a friend in order to plan an event. This is not a case of intending-to-A-for-R but of attributing both an intention to A and a purposive explanation of it. The attribution indicates both what is aimed at and what motivates the aim.

There are perhaps circumstances in which I can intend that my helping be explainable by my promising. This intention might embody an intention to do something that causes that explanation relation to hold. I might, e.g., get my brain manipulated so that my making promises does in fact explain why I do the promised deeds. But do phrases of the form of ‘A-ing for R’, where R is a reason, report any act or activity? Surely this, like ‘A-ing from a sense of duty’ and ‘A-ing for the purpose of helping her’, is a double-barrelled report indicating items in two categories: an action, which is an event, and a reason as a motivational element (e.g. to live up to my duty), which is not an event. We must not let the difference between what is done and why it is done be obscured by our desire to act for the right kinds of reasons combined with the not unnatural thought that, being something we can say we intend, ‘to A for R’ simply designates doing something. We can also say we intend to be upright citizens, and this plainly does not imply that ‘being an upright citizen’ designates doing.

The considerations just noted are not the only possible explanations for misconceptions of locutions like ‘I intend to do it because I promised to’. Ascriptions of intentions, especially ‘further intentions’ (and further aims in general) have considerable power in describing what people are doing. Consider the question ‘What is he really doing?’ This presupposes that (1) he is doing something more than the salient deed referred to, such as helping an elderly woman with her groceries, and (2) that this further action is supported by a significant reason (carried by some intentional element, such as a hope of financial gain, even if not necessarily by intention). An implication of such parlance is that one cannot tell what people are doing (intentionally) without knowing at least one appropriately governing reason for their doing it. Now suppose this is so. This does not entail anything about the question whether reasons can enter the content of the presumed underlying intention. The most important point it shows is that one (objectually) can intend one action to bring about another, say intend assisting with loading groceries to provide an opportunity to pickpocket their owner. This concerns the relation between intended acts, not between an action and a reason for it, and it does not imply that the agent has an intention whose content has the form of ‘to A for R’. Similarly, expressions like ‘intending to B by A-ing’ are informatively double-barrelled: forward-looking (toward a desired result of A-ing) and instrumental. The ‘by’ indicates how – by what action(s) – the agent intends to achieve the result, not a reason for aiming at it. The scope of intention can also include behavioral consequences of the ‘target’, focally intended, act – one might loosely call these penumbrally intended – but it does not follow that it can include the (or a) reason for which the act is to be performed.

This case suggests how the role of intentions in framing act-descriptions can be understood without countenancing intentions with the kind of purposive content we are exploring. One might think that at least for basic acts for which two or more reasons we have are eligible grounds, we can intend to bring about, at will, their being performed for one of these reasons in particular. But is this so? Granted, we can immediately bring about A-ing for R provided that we can A and our only way to A is on the basis of R, as where my only reason to shake your hand is to greet you. But this is really bringing about A-ing for R by A-ing – there is only one reason available to yield the deed: A-ing is performed at will and its performance of course guarantees that the only route to that performance has been traversed. Suppose, however, that I have a reason of self-interest as well as a promissory reason to help someone with a task and each reason is strong enough to yield the deed. Can I at will bring about my doing it for the promissory reason? And does intending this even make good sense? I might do things that justify my expecting that I will do the deed for the preferred reason. But the expectable is not thereby intendable; and intending to cause oneself to A for R does require having an intention whose content is to-A-for-R.

It should be clear, then, that from the point that a further intention or further aim in A-ing, say to achieve G, can yield and often does yield, a better description of what the agent is doing in A-ing, we may not infer that descriptions of the form of ‘A-ing for R’ are act-reporting phrases. That acts can bear many descriptions on the basis of the agent's relevant intentions shows much about the number and scope of underlying intentional attitudes, but does not justify taking any act-description to be a combined report and explanation. We have also seen that one reason not to take ‘A-ing for R’ to be an action-report is that we do not have, and do not take ourselves to have, the needed control over the reason(s) for which we act, and we do not normally intend to do anything we do not take ourselves to be in general able to do. But there is a more important and perhaps better reason.

Questions and explanations are very important in individuation. Given different questions about the same phenomenon, where answering one, say ‘Why did S A?’, leaves open importantly different answers to the second, say ‘For what reason(s) did S A?’, an item cited in properly answering the former should not be identified with an item properly cited in answering the latter. This is especially evident where the answers specify entities in different ontological categories, such as events and desires. Now the question ‘What did S do?’ is not properly answered by ‘S A-ed for R’ – to that one might respond: Never mind why S did it – that is a further matter and for the jury to decide. Similarly, ‘For what reason did S A?’ is properly answered by ‘For R’ – that S A-ed is presupposed, and various reasons are eligible. Such erotetic individuation is important. When a statement constitutes a proper and direct answer to one question and not to a second, and the counterpart point holds for a second statement regarding the second question, this is a strong indication that the two statements report different kinds of thing.Footnote ⁶

4. The Itinerary Conception of Intention

Both the constitution and the scope of intention can be clarified by fruitful analogies. Our belief systems are in some sense our maps of the world. But if, as intellectual beings – beings having theoretical reason – we have a map of our world, do we not, as practical beings, have an itinerary? Do we not, for instance, have plans and, usually, efforts afoot to realize them? Unpacking the itinerary metaphor might begin with reflection on the questions – posed, say, to a visitor, ‘What do you hope to do today?’ and ‘What do you actually intend to do today?’ Both presuppose stored “directional” information such as our internal maps express – routes indicated by our cartographic beliefs. One, the intention question, should elicit expectations that are roughly predictive and a basis for evaluation of the day's activities as successful or not. The hope question should elicit a kind of wish list, which may be much wider than the former list. An itinerary presupposes a map – or at least a need for one. When we have an intention to move to a destination on our map, we tend to have some subsidiary, instrumental intentions (though dispositions to form them will do). Whether we do have them is variable. The point is that without intention we could have a splendidly informative map but no itinerary.Footnote ⁷

We can follow an itinerary for various reasons; and intending to go to a destination, though it requires a conception of getting there, does not require a conception of why one is or should be going there (though we normally have some such conception, even if we need not and are sometimes quite wrong about our motivation). This is to be expected if it is true that intentions do not have contents of the form of ‘intending to-A-for-R’. Indeed, we commonly distinguish between where one is going and why one is going there. Of more direct relevance is that two people can share an itinerary though they are going to the various shared destinations for different reasons. Intentions carry the direction of behavior in their content and tend to energize and guide it. But that content should not be taken to include the reason (or a reason) for the intended action, and, if it can include it, then the content of such intentions is at odds with the capacities we take ourselves to have. It would be as if we could have destinations on our itinerary that we do not and cannot reasonably expect to reach. Such aspirational goals could be objects of hope, but that is quite different. Objects of hope need not even be acts.

It might seem that the itinerary view of intention is too intellectualistic, requiring deliberation or at least reflection for intention-formation and precluding unconscious intentions. It does neither. An itinerary may be implicit. In certain cases, we would not form an intention without considering the prospect of its realization. Where there are competing desires, for instance, we may compare the prospects of satisfying each. But as we saw in considering sources, forming intentions does not require such comparison. Moreover, on the plausible view that intending to A (as opposed to hoping to A) entails some degree of expectation that one will A, so long as one has appropriate access to such expectations, the intentions undergirding them can play the role in our planning for the future – for our pursuit of our itineraries – that they normally do play. Granted, one can have intentions one would disavow; but even that, as with self-deception, does not entail that one does not take account of the relevant expectations in planning.

These points leave open not only the possibility of self-deception but also the belief-desire theory of intention, but our concern is with scope and other matters that are neutral with respect to that theory as against competitors. The more important point here is that the itinerary conception also helps us to see that from an act's being expected as a result of intending something, it does not follow that it is intended. Much that we expect to do in reaching our destination is of no concern to us and in no sense a destination. I expect to trample grass on a walk through a field, but this is not something I want and is incidental to my itinerary. That does not absolve me of responsibility for it: like collateral damage in targeting a munitions dump, it is foreseeable and I may be criticizable for doing it. Similarly, even seeing that one will do something, say vote for a candidate, for a selfish reason, does not imply an intention to-do-it-for-that-reason. Indeed, one might prefer to do it for the reason that the candidate is best. A meritorious reason we have may be important and we may be embarrassed at its not determining a vote. Causing “collateral damage” may be similar: one might hope to intend only to hit the target but be aware that the collateral damage has greater military value and act for that reason instead. One would be morally responsible either way; but the actions reflect differently on character depending on their grounding reason(s).

We have already made room for activities as objects of intention, but have not considered how much detail regarding them must figure in the content of intentions to perform them. The same holds for scripts. The degree of detail is variable and I leave it open. An important point about that openness is that, just as one can form specific beliefs and intentions in thinking about how one will follow a geographic itinerary, one can do the same for an intentionalistic itinerary. In either case, such reflection sometimes leads to a change in a destination on the itinerary or a change in planned routing to reach it, or in both. This must be kept in mind as we consider how intentions figure in realizing some major moral standards.

5. The Problem of Intended Conduct

Often people speak of conduct where they are referring to a single action or a single activity, usually one that is an object of evaluation. But the term can have wider reference. Consider the shameful conduct of a confidence man. In order to steal her purse, he helps an elderly woman load her car, gently placing her packages. This is immoral conduct, even if the arrival of her husband prevents the theft. The helpful deeds are motivated by concealed immoral desire, and their gentleness is hypocritical. Conduct can also merit criticism even when its motivation is admirable and it also fulfills an obligation, but is marred by a constituent act performed in the wrong way. Suppose that, annoyed by an unpleasant phone call, the boss, Jim, gruffly says to a new intern who has been slightly overusing lights, ‘Turn off the lights when you leave’. Jim may be obligated to save power and saying this might be called for, but if Jim does it harshly and glaringly upon briskly closing the door, the intern is likely to feel hurt, and Jim is criticizable for this aspect of his conduct: acting hurtfully toward a junior employee.

These cases illustrate that ‘conduct’ can designate a three-dimensional behavioral phenomenon specified by act-type (ordering that the lights be turned off), motivation (to save electricity), and manner of performance (gruff). Once we clearly focus on conduct as a three-dimensional behavioral notion, we can better account for both the appeal and the elusiveness of Kant's categorical imperative in its humanity formula – a directive in which almost everyone finds something plausible and morally important.Footnote ⁸ It calls for treating persons as ends in themselves and never merely as means. With this in mind, recall our confidence trickster. His assisting his prey has the appearance of treating her as an end: he helps with loading the car. But his conduct, being partly constituted by a desire to rob her, is morally repugnant. As to the angry boss, his action – directing that the lights be turned off – and his motivation for his directive are appropriate, but his conduct is unacceptable as unduly harsh. He fails to treat the intern as an end.

If, as seems clear, treatment of persons as ends is a kind of conduct – since it involves doing an appropriate thing for an appropriate reason and in an appropriate manner – then we face a problem. If treatment, as having a motivational element, is not action, how is it commandable or otherwise an appropriate object of an injunction? Must not the commandable be intendable? Must we either give up the idea that intentions (at least when infinitival) are act-directed or take complexes of act-plus-its-motivator to be action after all? On the itinerary theory, we might conceive the problem as that of how treatment of persons can be on our itinerary. In outline, the answer is that it can be a destination without being an action. Even if there is no act of treating-someone-as-an-end, there can be acts of bringing this about. These take various forms and can be both commanded and contents of intention.Footnote ⁹ The notion of treating persons as ends an important challenge to understanding intention, and we should consider at least three approaches to a solution.

Take first the case of the boss and intern. I have already cited a boss-intern relationship in which the former should treat the latter as an end and not merely as means, say to an income. Suppose that the boss later realizes the objectionable conduct, and imagine that he is rather Kantian. How should we conceive his resulting resolution to treat the intern as an end? Certainly the intention to treat the intern as an end is not completely fulfilled unless any action constituent in the treatment – its vehicle – is motivated substantially by caring about the person's good intrinsically: for the sake of that person. The difficulty is to understand what counts as the content of the intention to treat as an end. A similar problem is to account for the content of the intention to love someone else as oneself, where motivation is implicitly part of the letter of the command. The challenge in both cases is to account for the intention in a way that does justice to its practicality by taking its content to be broadly a matter of action while also specifying conditions that normally yield, and can be conceived by the agent at least to tend to yield, the success of the relevant action(s) in achieving the conduct aimed at: roughly, to be brought about.

A natural way to approach the problem is to reflect on the point that – as is illustrated by intending to try – an intention can have a higher-order behavioral object. Treating others as ends is something one can surely intend to bring about. What one intends to bring about will in this case be conduct and thus entails acting for a reason, but it should not be thought that this implies that the intention to bring about something embodying action for a reason is thereby a case of intending to bring-about-for-a-reason. Bringing about is a kind of doing, and its objects can be simply acts or even conduct, conceived in terms of acts, their underlying motives, and their manner of performance. Applied to the imperative to treat others as ends, we might call this an altruistic-pattern interpretation. On this interpretation, the intention that is formed in fully accepting the imperative is roughly this: to do things having an altruistic pattern (though the agent need not use exactly this concept), where the agent has an expectation or at least hope of performing the relevant deeds for appropriate reasons and in appropriate ways.

In line with this interpretation, we might widen the kind of behavioral content the object of intention can have and take it to include at least mainly bringing about states of affairs of any sort within a broad range, such as, for a teacher, gently explaining imperfections in a student's paper. The behavioral focus of the intention is the acts that one sees as the vehicles of such treatment, acts like pointing out errors and omissions. But there is also an accompanying penumbral aim that differs from person to person in a way the focal aim does not. The penumbral aim is directed toward what one takes to support the motivational and manner requirements of conduct. For me, such supporting elements might be reminding myself of both the constructive aim of my relationship to students, recounting the good points to be mentioned, putting out of mind annoying deadlines that might sour my mood. For you, it might be more a matter of bearing in mind the needs and vulnerability of the student, which you know will strengthen your tendency to act toward the student for the right kinds of reasons – the end-regarding kinds of reasons that “aim” at some good for the student. These elements are less likely to figure in considering the content of one's intention, but might well be mentioned if one were to explain in detail how one will carry it out. To be sure, in explaining this, one might form new intentions. But just as elements in a script need not figure in forming or expressing an intention to do the thing in question, say recite a memorized poem, many elements that are, as it were, tacitly intended need not figure in forming or expressing an intention to bring about a complex end.

On the altruistic pattern view, then, the content of the intention to treat someone as an end is roughly the higher-order aim to bring about or to sustain (or both) the kind of pattern that characteristically manifests such end-regarding treatment. Since bringing about is action (though often not basic action, if it ever is), this interpretation accommodates the idea that the objects of intention are acts of some kind, and it also allows us to deny that ‘intend to treat as an end’ expresses an intention whose object is conduct conceived explicitly in the three-dimensional way I have described. In this way, we can leave open what is in the penumbral area of the intention. This openness helps in accommodating individual differences among agents.

Once it is seen that the content of an intention may be higher order and include in its content not just determinate focally intended acts but also penumbrally intended acts, we can see how a second interpretation might be developed. On this interpretation – call it the conjunctive interpretation – the expression ‘to treat others as ends’ can be taken to indicate (elliptically) not a single act-type but an open-ended conjunction of actions and activities of a beneficent kind.Footnote ¹⁰ The difference is that on the altruistic pattern interpretation a behavioral umbrella notion like bringing about expresses the content of the intention, whereas, on the conjunctive interpretation, ‘to treat as an end’ is conceived more concretely as a shorthand for a perhaps indefinite list of specific kinds of deeds, and the umbrella notion need not figure in the content. Something like it must figure, however, if the intention is to reflect the full content of treating as an end; the agent might have to have, for example, some at least penumbral aim to develop or sustain a tendency to be properly motivated in deeds toward the person(s) in question. No specific vocabulary is required, but the intentions should include deeds directed toward causing, or making likely, doing the intended things for the right kind of reason. Moral agents often do, and certainly should, have a sense of the appropriate range of actions here. That they usually do is illustrated by the ready understandability of the morally criticizable treatment of the elderly lady by the confidence trickster. Compare the intention to be a better friend. This could bear an altruistic pattern or conjunctive interpretation, depending on the circumstances.

On both interpretations, intentional content may be quite vague, but vagueness is ineliminable in natural languages and, on some views, a property of concepts themselves. Vagueness may be expected to figure in any kind of content broad enough and flexible enough to capture our normal representations of our aims and our projected routes to attaining them – our itineraries. Still, even allowing for vagueness, not every agent thinks in terms of such an umbrella concept, at least in framing or considering the content of an intention. In practice, roughly the same conduct might be expected from people who internalize the end-regarding injunction on either interpretation. Moreover, for different people, or for the same person at different times, one or the other interpretation the agent's intentional state may be preferable.

These two understandings of the intention to treat another as an end are each applicable in different cases. To be sure, if we consider promises as raising similar questions about the scope of the intention formed through sincerely making them, then the second, conjunctive interpretation may accord better with the way many who promise to love would explain what they intend. Think of someone sincerely promising to love an adopted child. Such persons would describe themselves as intending to do nurturing deeds, to provide resources, to comfort during illness, and so forth, but more would be implicit in such a promise. Even if we can intend only acts and activities, either understanding of the content of promising – and correspondingly of the injunction to treat persons as ends – makes it possible to see how we can intend what we must if we are properly said to intend to treat persons as ends. That phrase is not an act-description, but the intention referred to does have the kind of intimately related behavioral content I have described, and such behavior can be commanded in a literal sense in which intention-formation is acceptance of a command and actions are its execution. Moreover, either kind of treatment-directed intention I have described is available to Kantian ethics in relation to the humanity formula and to the theory of deciding, promising, resolution, and other intentionalistic elements important for ethics and the theory of action.

The umbrella and altruistic pattern interpretations of intending to treat another person as an end are not the only options. There is a third interpretation, appealing to subjunctive intentions. These can be carried out by actions realizing either of the first two kinds of intention. They may provide for an interpretation of treating a person as an end superior to either of the other interpretations taken by themselves. As we have seen, intentions that may have indefinite content, at least regarding the acts one takes to realize the state of affairs one intends to bring about. Such subjunctive intentions always imply some intentions to do, including open-ended intentions to bring about something, but they do not require as detailed an agential conception of the acts in their scope, if indeed they require any specific act-type in that role. They are best conceived as act-directed but not act-specific, and they are instrumentally unified but not necessarily in any particular way. Intending that seems irreducible to intending to, but this leaves open a supervenience relation such that (allowing for differences due to self-reference) no two people who are alike in their intentions to do and dispositions to form these can differ in their intentions that, their subjunctive intentions. This supervenience relation may perhaps hold for intentions to be as well.

We could take injunctions to treat persons as ends to require for their fulfillment that one be, in a certain way, altruistic toward the person in question and that one achieve treating persons as ends, where these construals are understood in the open-ended way I have indicated, concerning act-type, motivation, and manner of action. Intentions that have indefinitely wide content, and the intention that one treat others as ends may imply intentions with the behavioral content indicated on either of the first two readings of this intention. Given these options, we can retain the idea that one cannot be commanded to do what one cannot intend, but we must clarify this by indicating the kind of intention and the elements and range of its content.

It should be stressed that none of these options for understanding an ethics that is like Kant's in emphasizing conduct as well as acting on principles implies that acting-for-a-particular-reason is itself an act-type or that only treatment-types (such as treating someone as an end) are genuinely objects of moral obligation. The assumption that, for Kant, they are the only genuine objects of obligation – perhaps a natural assumption given his taking his various formulations of the categorical imperative to be equivalent – may be one reason why he is sometimes taken to be committed to conceiving acting for a particular reason, as well as acting on a specific maxim, to be types of action.Footnote ¹¹ But doing justice to Kant's overall view does not require this interpretation of him.

It may be that some people can form intentions to do what is required to treat a person as an end only where they conceive the commitment as directed toward bringing about a certain range of deeds in a certain way, whereas others have a more concrete understanding of their commitment in terms of an open-ended list of kinds of acts that they can readily perform and that they take to be normally motivated in some appropriate way. Suppose that, in either instance or both – i.e. on the altruistic pattern or the conjunctive interpretation of the injunction to treat persons as ends – a case can be made that the content of commands and promises need not, even when properly formulated, be acts. They certainly need not be in cases wherein intentions that are the crucial elements whose content indicates commitment to the commands or promises. One can, for instance, intend that all one's acts toward a distressed child be gentle. Nonetheless, fulfilling commands and promises – at least fulfilling their spirit – is done by our actions under certain motivational and behavioral conditions.

On the assumption I favor, that what we are commanded to do is intendable, even if only implicitly, the content of the intention formed in accepting a command is constituted by act-types (including higher-order act- and activity-types), and we can still give a reasonably clear interpretation of the imperative to obey the humanity formula, to keep promises to love, and the like. But on any plausible interpretation of the injunction to treat persons as ends, we must suppose that complete fulfillment, as opposed to mere behavioral conformity, is not achieved not just by actions of an appropriate act-type, or even by those actions performed in appropriate ways. It is achieved only by the relevant deeds being based on end-regarding motivation and being done in an appropriate way; in short, by what I call conduct. Bringing about good conduct requires a measure of self-understanding and self-control. Achieving it may be difficult. The difficulty varies with different people and with different kinds of situation. Reducing that difficulty is one aim of moral education, though there is no formula for doing so and, in my judgment, no adequate substitute for good role modeling.

Conduct is not constituted by action alone, but most of our interpersonal actions manifest conduct, and action is its perceptible element. Given that conduct has action as its vehicle and as its most commonly manifest element, it is no surprise that conduct – including the loving conduct intended in the commandment to love one's neighbor as oneself – can be commanded and required. If we can readily perform the relevant action(s), and especially if we have considerable control of the manner of our performance, we can at least begin to fulfill the requirement the command imposes by readily achievable doings. Moreover, we apparently can, at will, sometimes extinguish certain anti-social feelings, such as resentment or, at will, do for someone a deed that, beforehand, we are reasonably sure we will in fact do for an end-regarding reason when the occasion for action comes.

We now have the core of an account of the problem of how we can intend to fulfill conduct requirements, such as those of the humanity formula, even though we cannot at will bring it about that we have the right motivation, much less that if we act consistently with that motivation, our action will be for the reason expressed in the motivational content. Perhaps the first point to make here is that we sometimes can cause ourselves to do a deed for the right reason. We can, e.g., bear in mind not only doing the good we take to be possible in, say, treating someone as an end, but also the attractive elements, even if their realization is far off, in living up to the standard, and the rewards of enhancing the other's well-being. Such mental activity helps some people more than others – something one would expect to vary with empathy and with other-regarding elements in which people differ. Still, the point is not to provide a formula for treating persons as ends or even explain success in this, but, within a plausible theory of intention, to make sense of the attempt to achieve that conduct. By countenancing umbrella intentions, conjunctive ones that may include the former, and subjunctive intentions, we make room for numerous indirect strategies, variable from person to person, for increasing the likelihood of various kinds of conduct: treating persons as ends, acting from virtue, and perhaps even loving one's neighbor as oneself – or at least the elements in this that are constituted by conduct rather than emotion.

6. Responsibility for Action, Intention Formation, and Conduct

It is clear that we bear moral responsibility for our actions: on their account, we are subject to praise or blame and, more generally, to holistic appraisal. But once it is seen that conduct, including treatment of persons as ends or, negatively, merely as means, is an important behavioral type not reducible to an act-type, it is plain that our moral responsibility extends, in quite definite ways, beyond action. Here I want to consider moral responsibility in two cases already discussed: intention formation and treatment of persons.

Regarding intention formation – or preventing its formation – consider a case in which someone is obligated to make a promise to care for an ageing aunt, but makes a lying one, i.e., one accompanied by no intention to do the things promised. May we say that the person nonetheless ought to intend to keep the promise, and may we criticize the person for not having that intention? This seems correct, but there is no doubt about something related to it: one should abstain from promising to A if one cannot, or foreseeably will not, thereby form the intention to A. In this case, there is also reprehensibility for the act of promising. Note, however, that some people might not be able to tell that they will not intend to A, even when they make a promise to. This might imply some responsibility for failure to intend; but again, arguably the responsibility is for failure to recognize one's motivational incapacity or, perhaps, failure to revoke the promise after making it in the absence of the required intention.

A large question that arises here is the status of the traceability thesis: the view that moral responsibility ultimately traces to action, including omission, and in that sense to the will.Footnote ¹² This thesis is consistent with my main point: that, whatever its ultimate basis, responsibility applies to intention in certain cases. If there is any doubt about this point about the scope of responsibility, consider two possibilities. One is forming an intention to do something morally wrong, as where an agent makes a plan to murder. A second is forming an incontinent intention, as where, against one's better judgment, one forms the intention to have another glass of champagne but never makes any attempt because every glass is taken before the waiter gets near. Self-criticism is appropriate here, whatever one's view about whether one could have done something to prevent this unfortunate failure.

The case of adversely criticizable intention is highly pertinent to the negative injunction of Kant's humanity formula: we are to avoid treating others merely as means. Such treatment is conduct and, as such, implies a kind of motivation, here merely instrumental motivation. But this conduct also requires action, and from this kind of action one can by and large abstain. Indeed, we should be able to tell when we are even tempted to treat someone merely instrumentally, or, especially, when we have formed an intention to treat someone so, and we should try either to abstain from any act that would be a vehicle of such treatment or at least to prevent the treatment by eliciting in ourselves some end-regarding motivation. The former is a kind of negative control of conduct, the latter a kind of positive control – or at least positive influence. The latter may not always be possible, but it sometimes is.

Indirect control of conduct is clearly possible. We can, for instance, remind ourselves of what we care about; call to mind things about others that evoke empathy; and focus on our moral and professional commitments. Even such indirect and partial control of intention, motivation, and conduct is highly limited. But if we cannot always cause ourselves to form the right kinds of intentions and to avoid forming the wrong kind, we have some sense of how this is to be done. Philosophical reflection, combined with psychological investigation, can assist in this. People differ in their capacities and certainly in their opportunities to use these resources. Moral education is partly aimed at helping in such endeavors. There is, however, for most of us at least some of the time, a significant measure of moral luck in both our formation and our realization of intentions important for good lives.

____

Intentions have much more diverse content than is generally recognized. They may be not only to do, but also to be, and that something be the case. They may also be toward an ordinary object not specifiable in an infinitive or propositional clause at all: we may intend an object for a task, where the ‘for’-clause carries the relevant intentionalistic content. Intentions are crucial in determining our life's itinerary; and, partly for that reason, how good or bad we are is to a significant degree a matter of what we intend. This applies not only to what we intend as a result of decision, but also to intentions that arise from combinations of desires and beliefs that, singly or in combination, may influence the will without our reflecting on the intended object or even being aware of our intending it. These major roles of intention are among the grounds for its moral importance, but it is also morally important as essential in making sincere promises and in carrying out directives. Both our promises and such important injunctions as Kant's categorical imperative and the commandment to love one's neighbor as oneself are properly accepted only if the addressee forms some appropriate intention – indeed, at least one intention whose realization is appropriate to yielding the sort of conduct describable as treating persons as ends.

The problem that emerges on analysis is that the kind of treatment central in the humanity formula, the love commandments, and promises to love is not an act-type but more complex: a kind of conduct. One of its dimensions is motivational, and this presents a problem insofar as we do not have direct voluntary control of our motivation, nor is either desiring something or acting on a reason the kind of thing that is an object of intention. We saw that it is only by appeal to umbrella intentions, conjunctive intentions, or subjunctive intentions that we can make good sense of the intention to achieve the morally required conduct. I have described the sorts of elements that constitute the content of such intentions. Those intentions are highly diverse, and their cultivation and maintenance requires not only determining what actions to perform, but also monitoring oneself and performing many kinds of acts aimed at eliciting appropriate motivation, reinforcing certain habits, resisting certain temptations, and controlling the manner of one's actions. The scope of our moral responsibility is thus very wide and extends to internal as well as external behavior. Even in the realm of thinking and planning, there is not only what we do, but also why we do it and how we do it. Responsibility for conduct reaches the domain of thought, including intention and even emotion, as well as the realm of overt action. There is much to be said for the view that a major moral responsibility of all of us is to treat others as ends. Such conduct is neither easily analyzed nor easily achieved. It cannot be achieved or even well understood apart from understanding intention and the range of its contents. This paper is meant to both enhance that understanding and to clarify some major dimensions of moral responsibility for conduct.Footnote ¹³