How to cite

Acting and Causing: On Pietroski's Causing Actions

by Rowland Stout e-mail

Centre for Philosophy
Dept of Government
University of Manchester
Manchester M13 9PL (UK)

In chapter 1 of Causing Actions, Paul Pietroski endorses Jennifer Hornsby’s (1980) identification of acting with trying, and at the same time takes trying to be a mental event. He regards this mental event as an ‘inner’ event "causally upstream of any bodily motion" (46). But thinking of acting and trying as inner does not seem to me to be warranted by Pietroski’s approach.

The book is an extended argument against neuralism (or against a sort of argument for neuralism), where neuralism is understood to be the identification of mental events with neurophysiological events. So an event of a trying is not supposed to be inner in the sense that a brain event is. And although Pietroski accepts Descartes metaphysical distinction between mental events and physical events, he does not need to extend this to the thought that mental events occupy a special mental realm. So there seems to be no underlying theoretical motivation for denying the natural thought that tryings are usually ‘out there’ in the world. <1>

Pietroski concludes that actions are inner because they are not bodily motions (chapter 1, section 4.1). But he offers no reason for us to accept that these are the only two alternatives. In trying to attract someone’s attention I may wave my arms. Someone else may say that they saw me trying to attract that person’s attention – they saw the event that was my attempt. They did not have to look inside me to see that. The temptation to regard trying to do something as inner may come from considering only very artificial examples of actions like raising one’s arm or pulling a trigger. Trying and not succeeding to raise one’s arm may not be an event that is clearly in the public domain, whereas trying and not succeeding to attract someone’s attention usually is, even though it is not merely a motion of one’s body.

Having internalised trying and consequently action, Pietroski must distinguish inner actions from those events which involve the transformation of the world by agents – such events not being inner. So he has to distinguish the event which is Booth’s action of shooting Lincoln from the event of Booth shooting Lincoln. The event of Booth shooting Lincoln involves in some way Lincoln being shot, but the event which is Booth’s action of shooting Lincoln does not. According to Pietroski, the event of Booth shooting Lincoln includes the action of Booth shooting Lincoln as well as the event of Lincoln being shot being caused by that action.

But this seems to locate agency in quite the wrong place. A "person’s (autonomous) contribution to the causal order" (18) is the mental act of trying, according to Pietroski. The transformation of the world in line with that person’s intentions is not taken to be an aspect of their agency; it is supposed to come afterwards and be caused by their action. On this picture, a person’s contribution to the causal order consists in their sorting out their own mind; the rest is up to nature. But transforming the world in line with one’s intentions should be at the heart of a philosophical account of agency.

Perhaps the problem comes from thinking of causation as being a single sort of relation that links members of the unified class of events. This way of thinking is dignified by the following sort of diagram that features significantly in the book.


Let c be the event of Nora’s deciding to raise her arm now and e be the event of Nora’s arm rising. The o symbol represents the causal link – the causing.

But now consider the event of Nora’s deciding to raise her arm now causing her arm to rise. The italics are supposed to clear up the scope ambiguity. I am interested here in the event that is the causing of Nora’s arm to rise. It does not sound wrong to say that this also causes Nora’s arm to rise. The event of c causing e causes e.

Of course, the idea of causation must be different in the two places it appears in this sentence. It may sound better to say that the process of c causing e results in e (or ‘terminates’ in e to use Pietroski’s term). This means that there are two ways of describing the situation represented by the diagram coe. We may say that the event represented by c causes the event represented by e. Or we may say that the process represented bycoe results in e. <2>

It certainly sounds right to say that Nora’s action of raising her arm causes her arm to rise. But is Nora’s action of raising her arm something that should be represented by coe or something that should be represented by c in coe? Is Nora’s action of raising her arm a process that results in her arm rising? Or is Nora’s action of raising her arm some event that causes her arm to rise by some other process – a process of transforming actions into bodily motions? Once we see that there is this choice, it seems that we must reject the second alternative. <3>

But Pietroski does have an independent argument in chapter 1, section 2 for the conclusion that the event of Nora raising her arm is distinct from the action of Nora raising her arm. And if this conclusion can be made plausible then the internalisation of the action seems to be a natural way to make room for it. The argument as outlined on page 20 (using the grizzly example of the assassination of President Lincoln) is as follows (paraphrased by me):

  1. Actions are events.
  2. The action of Booth pulling the trigger was identical with the action of Booth shooting Lincoln.
  3. The event of Booth pulling the trigger is distinct from the event of Booth shooting Lincoln.
  4. Therefore, the event of the action of Booth pulling the trigger is distinct from the event of Booth pulling the trigger.

I think one should be very suspicious of the first premise of this argument. It is clear that actions and events are described very differently when it comes to tenses. Note the use of "was" in the second premise and the use of "is" in the third. Generally we talk of events using the timeless present tense; but we talk of actions using tensed verbs. And this makes me very suspicious of the thought that actions are events. But I am happy to let this premise go for the sake of argument. I want to challenge the second premise here.

Pietroski’s argument for it echoes the argument used by Davidson (1980, chapter 3) as well as Hornsby (1980). Booth’s action is his personal causal contribution. But after he pulled the trigger there was nothing more for him to do, the rest was up to nature. So his personal causal contribution is the same for pulling the trigger and for shooting Lincoln. Indeed, his personal causal contribution would have been the same if he had merely tried to pull the trigger but his muscles had been temporarily paralysed. So Booth’s personal causal contribution – his action - is causally upstream even from the motion of his finger.

I think that the philosophy of action has had an unhealthy interest in mere bits of behaviour as opposed to ways of behaving. It seems to be primarily concerned with pressings, squeezings, arm raisings, flippings of switches and shootings. There is relatively little treatment of writing books, travelling to work, watching out for the arrival of a train, making a cup of tea, chatting or dancing with someone, carrying a piano downstairs together or playing cards. None of these things can be thought of as bits of behaviour. But even within philosophy of action’s favourite gruesome example of shooting the president, there is room for the distinction between how someone has behaved and their bits of behaviour.

Let me change the assassination from Lincoln’s to Kennedy’s and consider a tale of two assassins. One might be in a book depository and the other might be on a grassy knoll. For simplicity, we shall have them working separately, preferably in different possible worlds just to keep them really apart. Both men intend to shoot the president, and both men try to shoot the president. Both of them do the following. They take aim, squeeze the trigger, watch to see the result, which is the president being shot, and then pack up, and go home.

The difference between them is in what they would have done if, having watched to see the result, they had seen that they had missed the president. Let us suppose that the man on the grassy knoll would have been disappointed, but would still have packed up and gone home. But the man in the book depository would have taken aim and squeezed again, and if necessary again and again. As it happens he did not have to, but this is what he would have done. The man on the grassy knoll was just taking a shot at the president - he was having a go at shooting the president. The man in the book depository was acting to make sure that he shot the president. As such they were acting in different ways even though, as it happened, they produced the same bits of behaviour.

So I am claiming that taking a shot at the president and making sure that you shoot the president are different actions, even when they are both done with the intention of killing the president. One way of seeing this is to consider the extent of the two actions in time. When does the man on the grassy knoll end his action of taking a shot at the president? When he has discharged a bullet in the direction of the president. The moment of quiet watching after this to see the result is not part of that action. When does the man in the book depository end his action of making sure he shoots the president? When the president is shot. The moment of quiet watching to see the results of his first shot is part of this action.

It is a causally relevant part of his contribution. It would not count as such on every philosopher’s view of causation. But Pietroski takes causal relevance to amount to explanatory relevance, and explanatory relevance to be determined in part by the ceteris paribus law that is invoked in the explanation. The moment of quiet watching figures in the law that explains this assassin’s behaviour. If he had missed, he would have had to notice this and shoot again. During the moment of quiet watching to see the results of his shot, the man in the book depository, if asked what he is doing, will say that he is making sure he shoots the president. His personal role in the action is not over just yet. The man on the grassy knoll would say that he is doing nothing much - just seeing if his shot was successful.

Having a go at shooting the president and making sure you shoot the president are different ways of behaving. They are still different ways of behaving even when they involve precisely the same bits of behaviour and they both involve the agent trying to do exactly the same thing – namely to shoot the president. The causal processes that result in these two ways of behaving are different too; they are the realizations of different potentialities. Different laws would have to be invoked to explain the two sets of bits of behaviour.

If this is right it follows that the action of pulling the trigger and the action of shooting the president are not identical if the action of shooting the president is taken to be the action of making sure that the president is shot. Even if the action is taken to be the action of taking a shot at the president, it seems clear that the personal contribution to the causal order is different from that involved in pulling a trigger. This is because taking a shot at the president involves as part of the agent’s causal contribution aiming at the president. If asked what he is doing as he raises his gun to his shoulder and takes aim the assassin on the grassy knoll would say that he is taking a shot at the president. When did he start making his one-off attempt on the president’s life? That is probably a tricky question, but the answer is obviously not that his attempt started at the moment he tried to squeeze the trigger.

My objection to Pietroski here does not go to the heart of his thesis in this book. His attempt to show that mental events need not be identified with neurophysiological events is quite untouched by my argument here, and indeed I find that attempt compelling. If Pietroski is wrong to identify the assassin’s trying to shoot the president with something causally upstream of his bodily motion, then the identification of his trying with a neurophysiological event is even less plausible. So if my objection is right, at the very worst it means that Pietroski did not need to do so much work to get his result.

I have another query that I want to air in this discussion, one which has very little to do with the first one, but I am keen to know Pietroski’s response to it. It concerns his notion of interference. Pietroski’s view of causation makes explanation constitutive of causation. He claims that c causes e if c and e are distinct events and a fact about c explains a fact about e. This sort of principle is often employed by philosophers of explanation who think that causation and explanation are connected but that causation is somehow basic. <4> Pietroski reverses this and takes explanation to be more basic than causation.

Pietroski’s approach to explanation is a variant of Hempel’s covering law approach. One fact explains another if there is a law which one may use to infer the second fact from the first. What makes Pietroski’s approach here distinctive is that he assumes that the law will usually be a ceteris paribus (cp) law. In other words it will say that other things being equal if f1 holds then f2 will hold. This is symbolised as cp[f1 Þ f2].

Pietroski spells out what "other things being equal" amounts to by saying that cp[f1 Þ f2] means that f2 will always hold if f1 holds unless there is interference. He is aware that there is a danger that this would make a cp law vacuously true if one is too liberal about what may count as interference. According to Pietroski, something counts as interference only if it explains something else in addition to the exception.                                             

For example, a phoney magician might try to claim that other things being equal every time they try to move a spoon without moving their own body the spoon moves. And indeed it is true that every time they try to move the spoon it moves unless it is stationary and subject to no physical forces. But the fact that the spoon is stationary and subject to no physical forces cannot be allowed to count as an interfering condition. Pietroski’s way of ruling it out is to say that the spoon’s being stationary and not subject to any physical forces does not explain anything else that is not a logical/analytic/causal consequence of the exceptions to the cp law (125).

I’m not sure that I really understand this condition. It seems that the spoon’s being stationary and not subject to physical force does explain more than the exceptions to the cp law. It explains why the spoon stays still even when the phoney magician is not trying to move it.

Also the condition seems to be too strong. Consider the gravitational law linking acceleration of one body to the mass of another body and the distance between their centres of mass: cp[(The second body has mass m and the distance between the two is d) Þ (The first body will accelerate towards the second at a rate equal to Gm/d2)]. This is an "other things being equal" law, since there will be an exception to it if there is a significant electromagnetic force acting on the body in question. But does electromagnetic force explain anything that is not a logical/analytic/causal consequence of the exceptions to the gravitational law? We can assume for the sake of argument that the electromagnetic force only explains the acceleration of bodies. Then the only phenomena it can explain are those phenomena that are exceptions to the gravitational law. But this does not stop it from being an interfering condition with respect to gravitation.


<1> Indeed Pietroski may not take the metaphor of the inner very seriously himself as is shown when he says that he has no objection to saying that we often see the tryings of others (222).

<2> Saying that a process is represented by coe does not commmit one to thinking of it as having three parts.

<3> I guess that Pietroski would be happy to represent the event as opposed to the action of Nora raising her arm by coe.

<4> Kim (1987) describes this approach as 'explanatory realism'.


Davidson, D. 1980, Essays on Actions and Events, Oxford: Oxford University Press.

Hornsby, J. 1980, Actions, London: Routledge.

Kim, J. 1987, "Explanatory Realism, Causal Realism and Explanatory Exclusion", Midwest Studies in Philosophy 12, 225-39.

Pietroski, P. 2000, Causing Actions, Oxford: Oxford University Press.

top | back to symposium index | Pietroski's reply