It is natural to think that what morality mainly requires of us is to do the right things and avoid doing the wrong ones. The “things” we ought to do are actions. But actions are not merely things we do. A plausible view is that actions (as actual events expressing agency) are doings that have a description under which they are intentional. When people act, they do something; and if what they do is not mere bodily movement of the kind illustrated by snoring while asleep, it is (with a few possible exceptions) action. To see how actions should be conceived, we should clarify not only what actions are, but how and to what extent they are under our control; how they are understandable in terms of our system of beliefs, desires, and intentions; and the manners in which they may be performed.

1 Action as a Kind of Doing

Usually, asking what someone did presupposes that the doing was an action, but snoring is behavior that is also something a person does, yet isn’t action. It could be cited by a light-sleeping host in answering ‘What did he do that so annoyed you?’ But snoring is not a straightforward example of action. Snoring of a kind can be action. An actor can do it in a sleeping scene. In any case, not all doings are actions. Sneezing can occur during a formal lecture and might be described as what someone did that explains the visible annoyance of the speaker. Unlike snoring, however, sneezing is a kind of doing that cannot be intentional: even if an actor impeccably mimics sneezing, the behavior is not genuine sneezing. Still, unlike knee reflexes, some sneezes are not mere movement. Unlike certain reflexes, sneezing, though not action, may be suppressible with effort – thus negatively voluntary and in that way under the control of the will.

What, then, distinguishes actions from doings that do not count as actions? If we consider something a person does an action, we presuppose some answer to ‘Why did you do that?’ where we can offer a description of the doing under which it is voluntary and, at least normally, may intelligibly be viewed as intentional.Footnote ¹ If it is voluntary, as is spontaneously stretching an arm upon waking, it is under the control of the motivational system (and normally suppressible at will), and we can ask why the agent did it. One answer is ‘I just felt like it’. Here one might call it done “on impulse” but not intentional (though also not unintentional, as with stumbling). One might also call it voluntary and incidental rather than purposive – it is normally no part of any plan or project.

It may be debatable whether we should consider such voluntary doings intentional, but we may surely call them actions as opposed to mere movements or, especially, reflexes. It is important that we see voluntary doings as under the agent’s control, since this makes clear that the agent may be held morally responsible, prospectively or retrospectively, for doing the things in question. Compare these two cases. I am sitting on a stone wall above a picnic table with a pitcher of juice just close enough so that, if I impulsively swing my foot forward I will knock it over. In one case, I haphazardly do this while excitedly telling a joke. In another, a prankster hits my knee with a hammer just so as to make the knee reflex produce the same motion. My hosts may count me careless in the first case but not in the second. There the prankster is blameworthy for the spill. Much that we are responsible for is voluntary but not intentional. This includes many of the manners in which we intentionally do things, as the Introduction illustrates with Fred and Jesse.

What we have seen regarding doings suggests that even if there are actions that do not have a description under which they are intentional, it is at least characteristic of action to be rooted in intentionality as an element in the psychological structure of the agent and often indicative of moral character. The point is important for ethics. If action were not generally rooted in intentionality, most typically in intentions but also in desires, emotions, and hopes, it would be difficult to appreciate the adage, “Actions speak louder than words.” As that adage reminds us, although self-descriptions may tell us much about someone’s motivation and character, a pattern of deeds (including speaking, of course) very often more reliably indicates intentions and plans. Intention not only underlies virtually all our actions; it also embodies motivation of a kind that represents some aspect of the will. Good will has great moral importance, and it is at least largely constituted by governing, generally long-standing altruistic intentions.

Many intentional actions are expressions of a “prior intention,” an intention to do some deed before actually doing it. But not every intentional action occurs in this way. It is perhaps easiest to see this when the agent has only a slight hope of success, but aims at something and succeeds. Someone who shoots at a threatening drone with only a slim hope of hitting it may, upon succeeding, intentionally hit it.Footnote ² More common are cases in which one acts instantly in order to prevent something, say blocks a running child who is about to fall down the stairs. One might here posit an intention in action, and that is possible without the existence of the intention prior to the action. What is crucial for a doing to be intentional is its explainability as something the agent does on the basis of a twofold psychological state: wanting to bring about something and having an appropriate belief guiding the action appropriately toward that end.Footnote ³

2 Act-Types and Act-Tokens

So far, action has been characterized in a way that indicates its moral importance as connected with intention in a way that, partly by representing the agent’s will, manifests good or bad character. To understand action sufficiently for moral theory, it is also essential to keep track of two distinctions: one between act-types and act-tokens, the other between basic and non-basic actions. The latter distinction presupposes the former, and I begin with types and tokens.

In typing, I am acting. I’m doing something in time – in space too, if the action is physical. I could have typed the same letters in the same way earlier, and others can also type them. My typing the letter t, here and now, is a concrete event; but there is also the act property, typing the letter t. In typing a t for illustration, I intended to type a t. What I intended, the content of my intention, was typing a letter t (also expressible as ‘to type a t’). A different concrete doing, however – a different act-token – would have fulfilled the intention. One agent can do the “same thing” – token the same act-type – at different times; and multiple agents can do the same thing simultaneously or at different times.

Moral principles characteristically call on us to do things, say keep our promises. These principles implicitly refer to act-types and are internalized only when we form stable intentions regarding the appropriate types. Stable intentions are sometimes created by or expressed in vows, promises, or resolutions, to do the things in question. Our intentions to do things are type-directed and token-realized. Full-scale appraisal of actions performed by actual people must take account of act-tokens – not just what type of thing we do but our doing of things. We can do the right thing (token the right act-type) for the wrong reason, at the wrong time, or in other ways that incur criticism. To be sure, if we say someone did a good deed, e.g. putting out a brush fire, we would normally presuppose that both the type and the token were morally appropriate. That twofold presupposition shows not that the deed we refer to is not a type but rather that the doing of the deed instantiates – tokens – an act-type appropriately good in the circumstances.

3 Basic Action, Levels of Action, and Voluntariness

The second distinction, between basic and non-basic actions, can now be clarified. Consider the point that one types a t by moving a finger but (normally) does not move a finger by doing anything else. We do it at will, in the sense that it is directly voluntary, by contrast with typing a t, which is indirectly voluntary, done by moving a finger. Here distinguishing the directly from the indirectly voluntary becomes crucial.

The voluntariness of act-types is doubly relative: to both agents and times. What is voluntarily done by one person may be involuntarily done by another. The voluntariness of act-tokens, by contrast, is fixed as to both agent and times. My signing a check is by me and at a definite time. In broad terms, to indicate what kind of act-type can be voluntarily tokened (instantiated) by an agent S, “directly” or, in more common terminology, at will, we need a characterization like this: (1) an act-type, A, is directly voluntary for an agent, S, at time, t, if and only if, at t, S can A at will, i.e., without A-ing by instantiating any other act-type (can A basically, in one sense). This formulation concerns agential potentiality. By contrast, direct voluntariness may also characterize actual behavior – act-tokens. Thus (2) A-ing, at t, is directly voluntarily performed (instantiated) by S if and only if, at t, S A-s at will.Footnote ⁴ We normally can do at will things like moving a finger “just because we feel like it,” and without doing something else by which we do it.

In the example of extinguishing a fire by using a hose, ‘by’ is instrumental in a causal sense. But there are other by-relations, including many in which one violates a standard. A wrong might be done by lying or by hitting. This relation between lying and doing a kind of wrong is constitutive, not instrumental. That kind of wrong might be said to be constituted, rather than caused, by lying (a distinction that will be important in Chapters 11 and 12). Another constitutive use of ‘by’ is non-moral: the pleasure of playing a piano piece is partly constituted by playing it, but the relation is not moral, pleasure is not an action, and playing it is not an instrumental means to that distinctive pleasure. The playing is essential to the pleasure of playing, whereas extinguishing a fire is possible by other means than firehoses. Various kinds of constitutive relation are considered in Chapters 2 and 3, but it should now be clear enough how the by-relation is important for understanding action and relevant to ethical theory.

As to mental action, I can at will call up an image of the spruce in my backyard. This is a basic mental act for me, and the act-type is normally under my direct positive control. I can also muse to myself that a proposition (p) seems implausible, and I may thereby (sometimes) bring about withholding belief regarding p. If the withholding is achieved by my thinking about the proposition (which I can do at will), it may evoke a sense of falsehood. This case illustrates indirect negative control. Control also has degrees: I can normally raise an arm easily and at will; whether I can produce withholding a proposition by considering it critically varies greatly with (among other things) the content of the proposition.

These notions of control of our actions are important in ethics for a number of reasons. First, we normally cannot even intend to do what we believe we lack the power to do and thus conceive as not under our control. Second, we normally do not intend, or at least do not plan, to do something when we cannot see or presuppose that we can do it either at will or by doing something at will. (We may presuppose this simply by presupposing that we know how to do it.) Third, we are normally either not morally blameworthy or are less blameworthy than otherwise for not doing something when we lacked at least indirect negative control over it. Fourth, moral rules and moral imperatives must be understood to apply to types of acts that can be intended by those who are to be guided by those rules and imperatives. What you cannot intend to do you cannot put on your itinerary.

Philosophers differ regarding the types of acts we can perform basically (“at will”). On one view, all we ever do is move our bodies; the rest is up to nature. On a Cartesian view, one might say instead that all we ever do is move our wills; the rest is up to nature.Footnote ⁵ Fortunately, the distinction between basic and non-basic action is neutral with respect to the difference between physicalistic (say neurophysiological) and mentalistic (say Cartesian) views of the category of basic human actions. But ethics partly concerns activities of mind, particularly those that manifest motivation and cognition or influence the role of either of these in moral character and intention-formation. For ethics, then, our mental lives are immensely important. A moral directive may concern mental life and call for mental action. The Bible has widely known examples. Take ‘Thou shall not covet’ (Exodus 20; Deuteronomy 5). This seems directed toward the internal dimension of lust – though its external manifestation in trying to get the desired thing is implicitly also proscribed. Clearly, planning to fulfill a lustful desire, even if the planning is only in silent premeditation, is prohibited. Here direct negative control is presupposed: one can – by and large – at will prevent one’s planning such things. If at some point the prospect assails one and is like a roar one cannot quiet, the only resort may be to turn to something like vigorously playing an instrument to drown out the offending voice. This would illustrate indirect voluntary control.

It is often held (plausibly) that we cannot produce desires at will, by contrast with, say, fasting all day, which produces desire by doing something else. This contrast shows how the scope and power of the will is ethically important. Can we, at will or even by some technique, create morally good desires in ourselves? And, when a selfish desire is served by doing something also obligatory, can we, at will, do the deed only for that moral reason? Kant among many others, thought we cannot do this at will but also that actions not done for morally appropriate reasons lack moral creditworthiness. This issue will be examined in Chapter 2, but a prior question is how to conceive what we are obligated to do in the first place.

4 Obligations as Potential Contents of Intentions

The case of lust shows not only that the mental realm is within the reach of moral directives, but also that they presuppose our having a kind of regulatory control of our actions. The control may, as we’ve seen, be indirect and imperfect, but our having it represents our voluntary ability to regulate our conduct. Such control is a power of will. The Biblical example illustrates an important point: that obligations – whether expressed in principles, imperatives, or other directives – are, in content – act-types, including activity-types. The content of obligation is determined by the specified act-type(s): schematically, the obligation is to A, where ‘A’ represents action (including activity); and the type is represented as to be instantiated (though not necessarily in those terms, or even subvocally).

This view does not entail that act-tokens cannot be obligatory in a derivative sense. An act-token, such as saying ‘yes’ if asked to do a job, may (especially in retrospect) loosely be called obligatory, but only by virtue of its tokening some obligatory type, such as agreeing to do a job. Tokens, like types, are individuated very finely. Even if one promised to say ‘yes’ at 10 a.m., in English, and at a certain volume, so long as the specification is not so comprehensive as to cover every possible way in which concrete acts may vary, more than one token can fulfill the obligation. One can do it in writing, in speech, in song, etc. The content of moral obligations is not essentially tied to particular tokens.

Do exceptions occur where an intention is de re, as when it is directed partly toward a particular? I can intend a friend to convey a message or intend my joke to amuse Rosaria. There is no doubt that such intentions concern something concrete. But surely their content is still a type, say to ask her son to convey the message. The content of both intentions is a type, though in the second it is a type “embodying” a singular reference to a person. To be sure, unbeknownst to me, Rosaria may no longer exist, but I leave aside whether this must in every case make a difference in the content (as opposed to reference) of the intention. The important point for ethics here is that if the content of either intention corresponds to that of an obligation, its content is best construed as an act-type.

______

Moral principles, then, and for similar reasons, statements of specific moral obligations, express requirements to do – to instantiate an act-type. Our assertions of obligation normally presuppose that the agent(s) in question can do the obligatory deed. Actual deeds, such as saying ‘yes’, may be considered obligatory in virtue of the agent’s having an obligation to token the obligatory act-type they instantiate. Obligation, however, is far from the only morally important property we must consider. Suppose I say ‘yes’ to doing something, not because I promised to (which I did) when asked to do a favor, but entirely for a profit motive. This is fulfillment of an obligation for the wrong reason – or anyway not a moral one. It fulfills the letter but not the spirit of morality. It may be hypocritical, but it may also fail to be moral even where the promissory obligation is felt but, given what the agent really cares about in assenting, not motivating. What is required to act morally in the fullest sense of that phrase is a matter of great complexity. This will be illustrated in a number of ways in Chapter 5. A prior task is to explore what it is not just to do something that is obligatory, but to do it on a basis that makes our doing it morally creditworthy. This brings us to Chapter 2.