Introduction and Outline of the Book
The principal aim of this book is to analyse what it means to act for a reason in such a way that we intentionally do what we have a reason for doing and intentionally attain the end for which we do this action, as specified by the reason. This analysis will also cover the simpler case in which we perform actions intentionally without having any reasons to perform them, for their own sake. It is, however, of interest to see how reasons fit in with intentional actions, since by far most of them are performed for reasons. By contrast, it will transpire that the analysis needs to be adapted to suit situations in which we let something be the case, or allow it to be the case, by refraining from acting.
I regard it as a virtue of the analysis of intentional action here presented that its analysans does not appeal to any concepts that are distinctive of the domain of action theory, like the concept of an intentional action itself. This analysis does not appeal to any primitive concept of an act(ion), whether in the guise of a unique type of agent-causation, which is irreducible to causal connections between facts or events, or in the guise of irreducible mental acts, like acts of will, volitions, decisions, or tryings. Nor does it appeal to any unanalysed attitudes essentially related to intentional action, like intentions and desires (to do something). Instead, the intentionality of actions will ultimately be understood in terms of (physical) states of agents causing facts because this will fit how agents think of them.
The direction of fit between thought and fact is here the opposite to what it is when something (propositional) is thought to be a fact: it is causing something to be a fact which is designed to fit the content of a thought rather than—as in standard cases of belief—thinking something which is designed to fit a fact. Notwithstanding these opposite directions
of fit, the pivotal notion of this analysis is that of a (propositional) thought fitting a fact, and this is a notion of a more general application than the context of action theory. Thus, it could reasonably be claimed that the analysis put forward here is a reductionist analysis of intentional action which does not leave any action-theoretical residue. As such, it promises to be more informative as well as simpler than analyses which harbour such residues.
Regarding reasons for action, I suggest in Chapter 2 that they can be put in a conditional form, in which the antecedent features a description of the action that the reason is a reason for performing, and the consequent specifies an outcome of this action that in the eyes of the agent counts in favour of performing it, in other words, something that qualifies as an end or goal of the action. Of course, I do not claim that reasons for action are always phrased in accordance with this mould in everyday life—that would be patently false. But it seems to me inescapable that when a reason for doing an action is made fully explicit, it must contain both a reference to the action for which it is a reason, or which it is about, and a mention of a consideration which counts in favour of doing this action. In a reason-conditional such as ‘If I scratch my head, this will make the itch that I am feeling disappear’: my scratching my head is the act for which this conditional is a reason, or the act which the reason is about, and its consequence that my itch disappears is what counts in favour of doing this act.
If we take it that something can be an action only if some action is intentionally performed when it is performed, we have an account in which a reference to the analysandum of an analysis of acting intentionally crops up in the analysans, since the antecedent of a reason-conditional is a description of an action. This would imply, implausibly, that we could not perform an intentional action, unless we were already in possession of the notion of an intentional action. More plausibly, it is the other way around: we acquire this notion by acting intentionally; thus, we cannot be self-consciously aware of acting intentionally until we have had some experience of acting intentionally.
In order to circumvent this awkwardness, I shall contend in 1.1 that there is a broader or wider concept of action which does not involve any intentionality. According to this concept, you act just in case you—that is, some facts to the effect that something happens to you— cause
something to happen, and these facts need not imply that what you cause is intentional under any description. The cause could be a spasm, or a stimulus—such as an object approaching your eyes—that elicits an act on reflex. It is this broad, less specific concept of action that figures in the antecedent of the reason-conditional. In the case of an action like scratching my head, and most other kinds of action, this claim is plausible, for whether my scratching relieves the itch does not depend on whether or not it involves intentionality.
What is specified in the antecedent is more precisely a type of act. Particular exemplifications or instantiations of this type will be equipped with many details which do not belong to the type. For instance, a particular scratching of my head will be conducted by some of my fingers moving at a precise speed in precise ways. It may also have the relational property of resulting from an intention. These features do not belong to the type if they are not relevant to the occurrence of what is specified in the consequent of the conditional.
Although it is relatively uncontroversial, it should be explicitly stated that the concept of propositional thinking, or thinking that something is true, that is invoked by the analysans is not an act. There are some related phenomena that may be acts, such as thinking of a proposition, thinking about whether a proposition is true, or saying something to yourself silently in your mind. These may be acts which can be intentionally executed, but thinking that something is true, such as a conditional proposition, is something that happens to you rather than something that you do. However, thinking that p is true can be the outcome of the act or activity of thinking about whether or not p is true. Alongside the concept of an action or act, thinking is the subject matter of Chapter 1, more precisely 1.2. Here it will emerge, as will be further discussed in 5.2, that intentional mental acts require a somewhat different treatment than intentional physical acts.
The reason-conditional should not be understood as a material implication; for instance, the falsity of the antecedent should not be assumed to suffice to make the conditional true. Rather, a reason-conditional presupposes that the truth of the antecedent is in some sense possible and asserts something about this possible situation, namely that it ensures or makes highly probable the truth of the consequent. So, in order for it to be true that if I scratch my head, this will result in the disappearance
of the itch, it must in some sense be possible for me to scratch my head in a broad sense of acting. This sense of possibility must be clarified, along with the purport of the conditional. It is in part merely an epistemic possibility, having to do with the fact that it is in principle impossible for us to predict our own decisions.
Although the truth of the antecedent must ensure or make highly probable the truth of the consequent in order for an ‘if-then’ conditional to be true, the truth of such a conditional is compatible with the consequent being true even if the antecedent is not, if its truth then is significantly less than highly probable. So, when I am scratching my head for the reason that my itch will then disappear, I might think that there is some possibility that my itch will disappear even if I do not scratch my head, provided that I think that this is markedly less likely than that my scratching will make it disappear.
The possibility of it disappearing even in the absence of scratching is excluded if I am thinking that the itch will disappear only if I am scratching my head. However, if I do not then think that it is also true that if I am scratching my head, the itch will disappear, or that there is something else that I can do which along with scratching my head will make up a sufficient condition in the circumstances for relieving the itch, I do not have any reason to scratch my head. These issues about reason-conditionals and their possibility presupposition form the topic of discussion in Chapter 2.
I shall argue in 3.1 that in order for me to act for the reason that if I am scratching my head, my itch will be relieved, I must want or desire the consequent to be true, but this attitude of wanting or desire is not part of the content of my reason. Its content is simply what is expressed by the conditional; the reference to desire is instead a part of the characterization of this conditional as a reason for me to act.
In this book, I talk primarily about reasons for which we act. Facts to the effect that we have these reasons in mind explain the actions that we perform, whereas these reasons themselves—that is, the contents of some of our conditional thoughts—justify these actions in our eyes. These reasons may not in fact justify the actions. For one thing, the reasonconditionals may be false (the itch may not in fact disappear). For another thing, the favouring or motivating factor may not be a state of affairs
that is good all things considered for us (the itch may serve the function of keeping me alert, which is crucial in the situation), in which case it is a mistake to desire it all things considered. If the reason-conditional is false, our reason is merely an apparent reason, not a real reason that really justifies our action, or even contributes to its justification; this requires the truth of the reason-conditional. An apparent reason for an action is a reason that at least in the eyes of the agent contributes to justifying the action, but in the absence of its truth it does not do so in reality.
Something may be a real, justifying reason for us, even though we are not aware of its existence, and have no desire with respect to its consequent. Some claim that real reasons are independent of our desires to the extent that they need not specify anything that we would desire even if we were aware of it. I have elsewhere argued that real reasons do not possess such desire-independence (see esp. Persson 2005: chs. 9 and 10; 2013: ch. 12), but I can steer clear of this debate here. My present contention is only about apparent reasons: it is to the effect that in order for it to be true that we act for a reason(-conditional), we must want its consequent to be true.
The sense of ‘wanting’ or ‘desiring’ involved must also be explicated. I shall advance a dispositional analysis of what I term ‘decisive’ desiring in Chapter 3. A decisive desire is a desire that emerges victoriously when you balance the strength of all your known desires bearing on the action alternatives at your disposal. It must also be a desire whose object is an act that you are fairly sure that you can perform. Forming a decisive desire is in effect to make a decision which, I believe, consists in the formation of an intention; thus, in my view, a decisive desire can be equated with an intention.
I might realize that if I am scratching my head, I shall cause not only the itching to stop, but also a scratching sound in my head, although this is an effect to which I am indifferent. If this is so, this state of affairs does not provide an additional reason for me to scratch my head; nor would it be an end or purpose of my performing this action. This fact by itself might make us reluctant to say, when I am scratching, that I am causing this sound intentionally, though we might go along with saying that I am causing it knowingly, consciously, or wittingly. Whereas foreseeing that my scratching would not relieve the itch would make me abstain from
the scratching, foreseeing that it would cause no sound in my head would not.
This implies that we need not want what we believe to be consequences of something we want. Thus, if I want to scratch my head, because I think that it will relieve my itch, and think that it will cause a sound in my head as well, it does not follow that I must want the latter. This would follow, I hypothesize, only if we make the unrealistic assumption that causing the sound is a sort of means for me to do something that I want, like the disappearance of the itching (or, even more unrealistically, my scratching of my head). The sound being a means implies, I shall contend in 4.1, that I can ascertain that it occurs prior to, and thus independently of, ascertaining the disappearance of the itching, so that I could use the occurrence of the sound as a sign that the itch will vanish.
Means that we can apply intentionally to accomplish ends are necessarily epistemically prior to them in the present sense. Causes are often epistemically prior to their effects; that is why causes are cut out to be the most frequent type of means. This type of means may be called manipulative means: they are means to do something, as opposed to epistemic means, which are means to tell whether something is, has been, or will be the case. It follows from what has been said that manipulative means that are intentionally applied to accomplish an end can also function as epistemic means to tell that this end is being accomplished, though the reverse is not true.
When we employ something as a means to an end, we can distinguish between that to which we do something as a means and that which we use as a means. Thus, when we scratch our heads, we use, for instance, our fingernails as a means to do something to the skin of our heads. It may be asked whether it is this distinction, or the distinction between manipulative and epistemic means, that the well-known doctrine that it is morally worse to harm somebody as a means than as a side effect turns on. These issues about means-end reasoning and the notion of means compose the subject matter of Chapter 4.
So much for the business of clarifying the notion of having a reason for action for which we act. But we may have a reason for action for which we act, and yet fail to do anything intentionally. When I have a reason for scratching my head for which I act, I may fail to act intentionally because the action I am performing is not scratching my head but,
say, scratching your head. But even if I succeed in scratching my head, my act need not be intentional because the connection between my reason and the resulting action is deviant or non-standard: perhaps because, unbeknownst to me, somebody has put some sort of extensions on my fingertips. To spell out the connection which makes the resulting actions intentional is the task of Chapter 5.
Realizing, to my surprise, that what I am doing is not what I thought that it would be—namely scratching my head—but scratching your head would probably make me stop or modify what I am doing. We know how to perform many simple acts, such as scratching our heads—though, as a rule, this is knowledge-how that we cannot verbalize—and I now did something because I thought that it would be putting into effect my knowledge of how to scratch my head. If I am receiving sensory feedback—in particular, by sight and proprioception—showing that what I am doing is not scratching my head, I shall in all likelihood modify my behaviour. I call this the correspondence control model of intentional action, CORCON: we execute the action of scratching our heads because we take it to be what we have learnt to be scratching our heads, an event which we believe will achieve our end of removing an itch; and we continue doing what we are doing only if we receive sensory feedback confirming that it is indeed scratching our heads (and that it relieves the itching). If we receive such confirmation, we are intentionally scratching our heads.
This is in outline the account of the intentionality of basic action offered in 5.1, but there is also the intentionality of non-basic actions to account for, like the relieving of the itching. The distinction between basic and non-basic actions can be drawn with respect to both action in a broad sense and intentional action. Non-basic actions are actions that we perform by (means of) performing some other actions; basic actions are actions that we do not perform by performing other actions.1 In the case of actions in the broad sense, basic actions can consist in actions that cause the results of other actions: for instance, the squeezing of a trigger causes a shot to be fired and a victim, Vic, to be killed, and as a consequence we have performed the non-basic actions of firing a shot and killing Vic by squeezing the trigger. By contrast, the results of our
1 This distinction was introduced by Arthur Danto (1963), (1965), and (1973).
basic actions are not themselves caused by any other actions of ours. So, our action of squeezing the trigger qualifies as a basic action only if its result is not caused by any other action of ours. It is not evident that this is so.
We perform such non-basic actions as firing a shot intentionally by squeezing the trigger intentionally if we perform the latter action intending by means of it to fire a shot, and successfully implement this intention, while the action of squeezing the trigger is an intentional basic action if our successfully implemented intention to perform it is not an intention to perform it by means of any other action. It will be seen that it is harder, but fortunately less important in the present context, to identify basic actions in the broad sense than intentional basic actions. It is the notions of intentional basic and non-basic actions that I shall attempt to clarify in 5.1 and 5.3, respectively.
CORCON implies that it is essential that basic actions are, as I shall put it, ‘contactual’, as opposed to non-basic actions which can be ‘consequential’. Your contactual actions consist in your causing some change as regards either just your body, or your body and a material thing in contact with it, such as an instrument of action, which you use as a means, or something you act on, or do something to, with (some part of) your body or the means used. Your consequential actions consist in changes that you bring about—by means of contactual actions—which extend beyond your body and what is in contact with it. Thus, stabbing and strangling victims are contactual acts, while harming and killing them are consequential acts.
When the consequences of your contactual actions include someone else causing something, that is, acting, these actions may be included in your consequential actions, but in case the actions caused are contactual they will not be your contactual actions. Thus, if you cause me to have a spasm which makes me pull a trigger and fire a shot that kills Vic, you may be said to have killed Vic, or caused her death. But you could not be said to have pulled the trigger, for pulling the trigger is a contactual act which entails that the agent’s body is in contact with the trigger.
For obvious reasons, the concept of contactual acts does not apply to mental acts, which may be exemplified by attending to or visualizing something, but the distinction between intentional basic and non-basic
actions applies to mental acts: for instance, you can intentionally make yourself excited or embarrassed by intentionally visualizing an erotic scene. However, an analysis of intentional basic and non-basic physical actions cannot be carried over to mental actions without modification, as will be explained in 5.2. One reason for this is that the mental episodes which, according to CORCON, participate in making basic physical acts intentional would interfere with the execution of mental acts. But it will be seen that this type of restriction is of less importance, since basic mental acts cannot have as long a duration as basic bodily acts, and consciousness does not have parts with which we can simultaneously perform different acts. By contrast, our bodies do have such parts which enable us simultaneously to do such things as to clap our hands and tap our feet to the beat of music, and these acts need to be coordinated. In general, our repertoire of basic mental acts is much more limited than our repertoire of basic bodily actions.
Most of us can reliably perform such acts as visualizing the faces of familiar people as intentional basic acts, but occasionally we unexpectedly fail to accomplish this. What we can do in such circumstances, and when we try to call to mind the images of things more unusual, is to utilize the fact that images can stand in relations to other images which enable us to associate from images that are present to us to absent images. These relations, e.g. resemblance, are relations that images have to each other in virtue of their contents. I shall argue in 6.2 that conscious episodes, qua such, are not causally connected, though they are presumably correlated with physical events (in the brain) that are causally connected but, as such, conscious episodes are only contentually related. We can facilitate the occurrence of content-related associations by interrupting competing associations, and minimizing distracting sensory input through closing our eyes and being silent and immobile. But these means are less reliable than the ones usually at our disposal with respect to non-basic physical actions, so the latter can be more diversified and far-reaching than non-basic mental actions.
When you refrain from acting and, thereby, let or allow it to be a fact that p, you do not cause p to be a fact. For it to be true that you let it be the case that p, you must be aware of the fact that you could cause something, q, to be a fact that will prevent p from being a fact, which it
otherwise would be. But, owing to your not finding sufficient reasons to form such a desire, you form no decisive desire to cause (it to be a fact that) q, with the result that it becomes the case that not-q and p. You do not have to form a decisive desire not to cause q, as you would have to do if you had to take action to prevent yourself from continuing a causal process already under way, such as your sliding or running downhill. It should be stressed, however, that in order for you to qualify as refraining from causing q, your forming no decisive desire to cause q must have a specific explanation to the effect that you do not find any sufficient reasons to form this desire. If you do not form such a desire because, say, you lose consciousness or are distracted by something else, you obviously do not refrain from causing q, and let p be the case, though it is true that you do not cause q, and p becomes a fact because of that. These are matters to be discussed in 6.1.
Although I shall not argue for it here (but see Persson, 2013: 3.2), it is my view that whenever you simply let something be a fact or the case, this is true because you refrain from or omit some preventive action that you could have committed, and never because you commit some action. You do not simply let something be a fact if, say, you remove an obstacle and then let it remain out of the way with the result that something happens that would not otherwise have happened.
When you let p be a fact, you cannot decisively desire it to be a fact, since decisively desiring something to be a fact is decisively desiring to cause it to be a fact. But you could adopt some other non-conative attitude to it, such as wishing or being glad that p becomes a fact. Since there is no decisive desire or intention that p be the case when you let it be the case, I do not think that it qualifies as intentional, strictly speaking. Letting p be the case by refraining from causing q is rather something that you engage in knowingly, consciously, or wittingly because you anticipate that p will be the case when you do not form a decisive desire to perform the preventive action of causing q.
Whereas you do not cause anything if you simply let it be the case, when you knowingly or intentionally perform an act , that is, cause something to be a fact, you must of course be the cause of something. According to the analysis of the concept of desiring that I advance in Chapter 3, this concept refers to a state of your organism—in all
probability a state of your brain—which is causally operative along with a thought that it presupposes to the effect that you can perform the relevant action. In 6.2 I argue, however, that, strictly speaking, mental episodes, like thoughts, are not causes; instead, it is the neural processes with which they are correlated that are causes. As already remarked, mental episodes are related to each other and to bodily events in terms of their contents.
This may be clearest in cases in which we infer a proposition from some other propositions that we think true because it logically follows from them, but similarity between what mental images are images of, and the fact that they are images of things that have regularly accompanied each other, are other examples of the contentual relations that we exploit when we engage in those psychological associations in which our intentionally imagining or thinking of something consists. Furthermore, according to CORCON, when we act intentionally, the relation between what we in fact cause and what we think that we cause is also contentual: the former corresponds to and makes true the content of the latter. The direction of fit is here opposite to what it is in the case of thinking something true on the basis of perception, which is involved in the sensory feedback that we normally receive when we act. Although the explanatory relations on the mental level are contentual, I believe that they presuppose the truth of causal explanations on a physical level, between the neural states with which the mental states are correlated. When letting p be a fact by refraining from acting explains why p is a fact, it is clear that the explanation cannot be causal.
It has been seen that we do not experience our decisions—whether they be decisions to do or believe—as being caused and that they are in principle unpredictable by us. 6.3 appeals to these facts together with the fact, established in 4.1, that practical decisions cannot be contentually constrained in the way beliefs can be by premises that entail them, or by perceptions that verify them, to offer a debunking explanation of the strong sense that we have that these decisions are up to us or determined by us. Such an illusory sense of self-determination is presumably what lends credence to libertarian conceptions of free will. Finally, 6.4 examines whether it must be true that we can refrain from (not) doing what we ought (not). My conclusion is that this is not necessarily so, though when we cannot refrain, we could at least have eschewed responsibility
for this failure. But normally we can refrain in a sense of ‘can’ which is compatible with determinism. Responsibility is, however, a side issue which cannot be fully explored in the present book, but it is important to indicate that there is a sense in which we are responsible. Otherwise, it could not be true that we act intentionally for reasons and do what we ought or have most reason to do.
Acting and Thinking
1.1 Acting and Causing
Philosophers differ over whether the concept of action necessarily involves intentionality. For instance, Donald Davidson claims that you perform an action just in case there is something that is intentional under some description, that is, just in case there is something that you do intentionally (1971: 6–7). On the other hand, Judith Thomson for one thinks that inanimate objects can act in the same sense as animate beings (1977: 252–4). I side with her: it seems to me evident that there is a broad or unspecific concept of action which is applicable to us and inanimate objects alike. We act in this sense even when we do something on reflex—such as blinking or withdrawing a hand from something painful—automatically—such as most instances of breathing, swallowing, and tapping our feet to the beat of music—or involuntarily, as can happen when we sneeze, laugh, or pull a trigger because of a spasm.
Suppose that I break a window when I am withdrawing my hand on reflex; then it seems obvious to me that in breaking the window, I am acting, doing something to the window just as, for instance, a rock does when it breaks it. I am behaving in a way that could be intentional were it intended. To act in this wide sense, we do not even have to move our bodies: for instance, we can score a goal by a ball hitting us and bouncing into the goal when we are standing still, unaware of what is going on.
Brian O’Shaughnessy rightly argues (2008: ch. 10) that the extension of the concept of an action that is intentional under some description is wider than it might initially appear, for instance, that when you do something absent-mindedly, such as bringing a pen instead of a key to a lock, what you do might be intentional under some rather vague description, like bringing to the lock a thing of a rough key-like shape and size that you keep in a certain pocket (2008: 350–1). He is, however, correctly doubtful about whether this manoeuvre succeeds in making
everything that strikes us as an act or activity intentional under some description, say, the tongue movements that we might indulge in when we concentrate on balancing something (2008: 357–62). The classification of something as an act does not hinge on settling such vexing cases if we accept a broader sense of action which does not require anything like intentionality, and which is applicable even to inanimate things.
As a matter of course, I also follow Thomson in regarding acting in this sense as causing something to happen: when I or the rock is breaking the window, we cause it to break. In the case of the rock, it is uncontroversial that when it causes something, it does so in virtue of some of its properties, for instance, it causes the window to break because it is hard and hits it with a certain momentum. It is, however, controversial whether it is also true of us that whenever we cause something, this is elliptical for some facts about us causing this. Roderick Chisholm (1966), Richard Taylor (1966), and a string of later philosophers deny this, believing in an irreducible form of agent-causation. Having argued against this approach elsewhere (Persson, 2005: 424–8), I shall not do so now; instead, I shall simply proceed on the assumption that whenever we are described as causing something, it is more precisely certain facts or events about us that are doing the causing. In the present context, what is primarily of interest is what these facts or events must be for an action to be intentional.
It is, however, hardly true, conversely, that whenever something about us causes something, we could be said to cause it and, thus, to be acting. For instance, our hearts pump our blood by beating, but this surely does not imply that we pump it around. What is missing is, I suggest, that we do not perceive or are conscious of this piece of causation: of course, we perceive and are conscious of the beating of our hearts, but not that this causes our blood to be pumped around. This claim assumes, as I shall soon argue, pace Hume, that we can perceive instances of causation. The tenability of this constraint of consciousness is, however, less relevant for present purposes than that attributions of causation to ourselves are analysable in terms of event- or fact-causation.
Turning now to the specification of the effect caused in action, it should be noted that some verbs—‘bend’, ‘cool’, ‘dissolve’, ‘harden’, ‘melt’, ‘move’, ‘sink’, ‘soften’, ‘warm’, etc.—have both a transitive and an intransitive use. In these cases, specifying the effect caused in action is often simple.
Thus, when you move a finger in the transitive sense, you cause a finger to move in the intransitive sense. More precisely, you cause it to be a fact that a finger of yours is moving. For I shall assume that, strictly speaking, effects as well as causes are facts—‘dynamic’ facts to the effect that something happens or changes—rather than events. An event is roughly the instantiation of an indefinite set of properties in a certain spatial region or in a substance at a certain time, whereas a fact is the instantiation of a specific property somewhere or by something at a certain time. For example, the fact that your finger is now moving consists in your finger now instantiating the property of moving. Therefore, an advantage of casting facts in the roles of cause and effect is that these relata will be more precisely described than if events are cast in these roles (but it is not crucial for the purposes of this book whether facts or events are cast in the role of causal relata). Another, more practical advantage is that I am able to use propositional variables, like ‘p’, ‘q’, etc. for what is caused. Obviously, in order to make the notion of a fact more precise, the notion of a property must be explicated, since the identity conditions of facts hinge on the identity conditions of properties. I shall not argue for it, but I am inclined to adopt a fine-grained approach to property-identity, according to which not even the logical equivalence of two predicates suffices for it to be true that they attribute the same property to the subject. For instance, being an equiangular triangle and being an equilateral triangle strike me as different properties, although it is necessarily true that something has one of these properties if and only if it has the other.1 But in the present context, it is enough to insist, less controversially, that the material equivalence of two predicates is not sufficient for them to express the same property. This point is essential for my claim in 6.2 that the properties that mental predicates express are distinct from the properties that any neural or other physical predicates express.
1 Derek Parfit accepts (2017: 130) Frank Jackson’s claim (1998: 125–7) that these are the same property even in Parfit’s ‘description-fitting’ sense of property, preferring instead to rely on the strength of his example of being the only even prime number and being the positive square root of 4 to demonstrate that we need a narrower criterion of property identity than necessary co-extensionality. But although a triangle which is equiangular and equilateral has only one shape, its ‘sides are distinct from angles’, as Jackson himself notes (1998: 127). To my mind this is sufficient for its equiangularity and equilaterality being distinct properties; it implies, for instance, that we have to use different procedures to check whether triangles have these properties. However, Parfit’s example is enough to make the point that necessary co-extensionality does not suffice for property-identity.
It is not always so straightforward to spell out what you cause in an action, as in the case of your moving a finger. For instance, it is harder to specify what you cause when you eat—harder but not impossible: you cause your lower jaw to move in such ways that the food in your mouth is chewed, and so on. Effects that are entailed by the performance of actions are sometimes called their (intrinsic) results (following von Wright, 1963: 39–41); thus, a finger movement is the result of the action of moving a finger, and death a result of killing. An action can have several results, some of which are entailed by others, as a finger movement is entailed by an obscene finger movement.
According to my conception, an action is an (at least partially causal) process, which can be quite protracted temporally. It starts with the occurrence of the cause, and does not terminate until its result occurs (cf. Thalberg, 1977: ch. 5; Thomson, 1977; and Ginet, 1990: ch. 3). Thus, imagine that you kill Vic by giving her a very slow-working poison. Then your action starts when some facts about you cause some movements of your body which constitute giving her the poison, but it does not end until the poison has caused her death, perhaps months later. This means that you may be dead when your action of killing Vic is completed. Some find this corollary—that agents can continue to act after their demise— implausible. This provides them with a reason to identify the action of killing somebody with some bodily or other basic action performed in order to perform the killing (cf. Anscombe, 1957: 37–47 and Davidson, 1971). These theories, however, carry an implication which seems at least as implausible, namely that Vic can be killed before she dies.
A distinction between what might be called contactual and consequential actions might help us deal with this conundrum. Contactual actions consist in causing some change as regards either just the agent’s body, or the agent’s body and/or some material thing in contact with it, such as an instrument that the agent uses, or something that the agent does something to with the instrument, or acts on. The ‘contact’ must be of the robust, material sort, not the sort of contact that you can have with the aid of radio transmitters, remote controls, etc. Thus, if you cut Vic with a knife, this is a contactual action which consists in your acting on Vic’s body with the instrument of a knife. Consequential actions consist in changes that you bring about by means of contactual actions, and whose results extend beyond your body and what is in contact with
it. For example, while stabbing and strangling victims are contactual acts, harming and killing them are consequential acts.
Consequential actions are necessarily non-basic actions—they are executed by (means of) the execution of contactual actions—but contactual actions are not uncontroversially basic. Some—notably, Davidson (1971)—claim that basic actions are restricted to bodily actions, so that when a contactual action such as pulling a trigger is executed, the basic action performed is just moving a finger in a certain manner, and that the pulling of the trigger is something that you do by moving the finger. But, analogously, it might be claimed that moving a finger is done by contracting some muscles; thus, the basic action performed would rather be contracting these muscles. However, these muscle contractions are effects of electric impulses being sent along efferent pathways, so perhaps you contract the muscles by releasing these electric impulses, and so forth.
This goes to show, as remarked in the Introduction, that it is tricky to identify basic actions in the case of actions in a wide sense. That is a reason why I have introduced the less contentious and more observational concept of a contactual action. It is less contentious because saying that you perform one contactual action such as moving a finger does not rule out that you simultaneously perform many other contactual acts, be they less capacious, like contracting certain muscles, or more capacious, like moving some external object in contact with your finger. And the concept is observational because the contact in question must be of a ‘robust’ sort (though we should not expect precision as regards what qualifies as such).
Now, on the one hand, it seems that ‘killing’ describes a consequential action and means roughly the same as ‘causing the death of’. This would explain why we would not say that we have killed Vic until what we have done has caused her death, and her death has occurred. On the other hand, we would not say that we are killing someone when we have performed the contractual act which will cause their death, but which has not yet caused it. But nor would we say that we have killed. Instead we might say that we have not yet killed, but have performed an action that will kill or cause death. Apparently, we would not say that we are killing, unless it is true that we are then performing a contactual action that will cause death.
