Abstract
In order to tell what a person ought to do in a certain situation, we must know what she can do, or in other words, what actions are performable by her, in this situation. The concept of performability is thus of great importance for C (as for all moral theories, it seems). Different interpretations of this concept, may yield very different C-moralities. One might perhaps mean several things by saying that a certain person can (intentionally) perform a certain particular action.1 I shall assume, though, that there is a unique sense or use of ‘can’ which is relevant in the present context. This ‘can’ is that which is implied by the ‘ought’ appearing in formulations of C. The purpose of this chapter is to find an acceptable analysis of this ‘can’, an analysis that will in the subsequent chapters be used to clarify and defend certain arguments and theses concerning alternative actions and consequentialist principles. 2 The analysis I shall endorse is based on the (not very original) idea that a statement to the effect that a person can perform a certain action is equivalent to a statement to the effect that he would perform that action if he wanted to do so.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Sometimes when we use expressions of the form ‘P can do a’, we are talking about generic actions. This is the case, for instance, when we say that ‘Brown can play the piano’, although we know that Brown will for the rest of his life be stranded on a desert island without pianos. However, I shall be exclusively concerned with particular actions.
Phrases like ‘P can do a’, ‘P is able to do a’, and ’a is performable by P will be treated as synonymous. Thus, I shall interchangeably speak of analyses of ‘can’, ability, and performability.
This type of analysis is sometimes said to originate with G.E. Moore, who proposed an analysis in terms of choosing. (Moore(2), Chapter VI.) Jan österberg has informed me, however, that the idea goes back to St. Augustine.
See e.g. Lehrer(2), p. 64 and VAN Inwagen, p. 117.
Cf. VAN Inwagen, p. 118.1 am not sure that choosing or deciding should be regarded as generic actions, either. Keith Lehrer argues that they should, since “a question about what a person is doing at a given moment may be naturally answered by replying that a person is [...] choosing a sugar ball, or deciding which road to take” (Lehrer(2), p. 64). On the other hand, it is also natural to say things like ‘we must decide what to do before we act’, or ‘she has not done anything yet, but she has made her choice’. The linguistic evidence is thus inconclusive. (Lehrer agrees that wanting is not a generic action, since it “is peculiar to answer the [above] question by replying that a person is [...] wanting a sugar ball”. (Lehrer(2), p. 64.))
For a defence of this idea, see Goldman(I), Chapter 4, Section 4, and Goldman(2).
If my reasons for preferring ‘want’-analyses to other conditional analyses are sound, they partially answer the objection, made by Keith Lehrer and others, that most conditional analyses are equally plausible, which makes it improbable that any one of them is correct. (See Lehrer(2), p. 68. Lehrer attributes this objection to Richard Taylor.)
Lehrer(2), pp. 36–37.
Anscombe(2), pp. 154–157.
Anscombe is aware of the need for time-references, but she introduces them in a rather confused way. She does not seem to distinguish between the time of the ability to perform an action, and the time of the action itself. (See Anscombe(2), p. 157.)
according to the analysis just suggested, El and Fl state a sufficient, but not a necessary condition for P’s ability at t to do a. More accurately, El and Fl should read: There is a time t’, not earlier than f, such that if P at t’ chooses to do a, then P will do a’. Since nothing in the present context hangs on this qualification, I have used the simpler formulations in the text.
Strictly speaking, it may be impossible at t to intentionally perform an action starting at t. It takes some time to turn a want or decision into action. Analyses of ‘immediate performability’ should therefore be seen as analyses of the ability at t to perform an action that starts slightly later than t. This qualification will be left understood.
The second clause should be read as ‘if P at t would want to do a, and not want more strongly to do some action incompatible with a, then P would do a.’ People sometimes have conflicting desires, and we would not say that a person cannot do a certain action that he to some degree wants to do just because he wants more strongly to do something else. This qualification will be left understood in the analyses discussed below.
Goldman(1), pp. 87–91. Goldman attributes the distinction between occurrent and standing wants to William P. Alston.
Lehrer(1), p. 32, Van Inwagen, pp. 115–119.
It might be retorted that Brown could not eat of the candies until he realized that it was the only way to save his life. But this would be to suggest, in effect, that what a person can do in a certain situation, is determined by his actual, rather than his possible, motivations in that situation. This is an implausible view, since it implies that one cannot do what one is not motivated to do. (Those who believe that Premiss (1) may be true even if Brown were to succumb to Black’s threat, could use a simpler argument against CA1. Brown’s giving in to Black’s threat falsifies Premiss (2), whether or not it falsifies Premiss (1). But if it does not falsify (1), (1) and (4) may both be true, without (2) or (3) being true. The nearest world in which Brown wants to eat a candy may be a world where Black threatens him. In that case, (2) is not needed in order to refute CA1.)
He would not, of course, have an intrinsic desire for eating a red candy. But he would have an extrinsic desire for doing so, since this would be a means to something he values, viz. staying alive, or getting a lot of money.
Unless, of course, we make the question-begging assumption that CA1 is correct.
A back-tracking counterfactual says that if the present were different from what it actually is, then the (non-immediate) past would also be different from what it actually is. (Cf. Chapter 4, note 10.) David Lewis is suspicious about the truth of back-tracking counterfactuals in general. (Lewis(3).) In polemic against Lewis, Jonathan Bennett defends back-tracking. (Bennett.)
I do not claim that it is necessary to have a full understanding of the concepts involved in a certain action, in order to be ‘able to want’ to do that action. But one must have some degree of understanding. The situation is similar concerning beliefs. If I claim that whales are a kind of fish, my erroneous belief might still be a belief about whales. I may possess enough of the concept of a whale, e.g. that it is a sea-living animal shaped rather like a fish, to be ‘able to believe’ things about whales. But if I claim that whales are small furry animals living in tree-tops, my belief is hardly a belief about whales. It is rather a belief about the word ‘whale’. What I mistakenly believe is that this word refers to a small furry, tree-living animal. (If the animals I connect with the word ‘whale’ are in fact squirrels, my belief is perhaps a belief about squirrels. I may mistakenly believe that squirrels are called ‘whales’.)
For a similar requirement on performability, see Bergström(7), p. 100.
It may be thought that the assumptions (i)-(v) above could all be true even though P’s wanting to do a is not in any sense incompatible with his being in state c. Suppose that John can run a mile under 4.30, given that he is not very tired. At t, however, he is exhausted after a hard 20-mile run. Hence, he cannot at t run a mile under 4.30. Furthermore, John understandably does not like to run a mile at top speed when he is tired after a long run. Now, someone might reason like this. In the past, John has never wanted to run a mile under 4.30, when he has been very tired. We can therefore regard it as a kind of psychological rule, although not a ‘law of nature’, that John does not want to do such hard runs when he is very tired. Hence, if John at t would want to run a mile under 4.30, he would not be very tired. Since he would not then be very tired, he would in fact be able to run a mile under 4.30. Thus, if John at t would want to run a mile under 4.30, he would be able to do so, and would therefore in all probability do it. Hence, CA2 is false. This line of reasoning is surely erroneous. The claim that John would not be tired, if he would want to run a mile under 4.30 is absurd. This shows, I believe, that the fact that a person actually never wants to do actions of a certain type when he is in a certain state, is not enough to justify the judgement that he would not, on a certain occasion, be in this state if he wanted to do an action of this type. Such a conclusion is only plausible when being in this state is at least physically or psychologically incompatible with wanting to do this type of action.
Sobel(3), p. 198. See also 7.4.
That conditional analyses generally have this fault has been noticed by Lehrer. (Lehrer(2), pp. 69–70.) Goldman’s definition of ‘basic action’ (Goldman(l), p. 65) suffers from the same weakness.
Another problem with CA5 is discussed near the end of this section.
Remember that an arbitrarily long sequence of actions may be an ‘action’, in my terminology. (See 2.1.)
Goldman suggests that only rather short actions can be caused by a single occurrent want: “For example, even though I frequently walk from my office to a certain luncheon spot, I cannot execute the entire three-block walk simply as the result of a single occurrent want. I can, perhaps, take ten or twenty steps at a time without concentration, but after each sequence of steps of this sort I must ‘monitor’ my progress to see how far I have come and then form an occurrent want for the next stage of the project.” (Goldman(l), p. 90.) To me, it seems an exaggeration to claim that a three-block walk cannot be caused by a single occurrent want. No doubt, however, actions that are so caused are never very long.
It is not only that it would be absurd to blame you for the deaths of the hostages, if you do not open the safe. This, it might be said, only shows that you have no ‘subjective’ obligation to open the safe. I think it is clear that you have no ‘objective’ obligation, either. This example is a fleshed-out version of a case found in Feldman, pp. 24–25. Feldman in fact claims that you can open the safe. Cf. Goldman’s distinction between an ‘epistemic’ and a ‘non-epistemic’ sense of ability. (Goldman(I), p. 203.) Goldman would say that you can open the safe in the non-epistemic sense of ‘can’, but not in the epistemic sense. I believe that it is the epistemic sense which is relevant in normative contexts.
It is not uncommon that a person can do a certain thing only by forming an ‘indirect’ or ‘secondary’ intention to do something else. A clearer example of this phenomenon is that of a right-handed golfer with a heavy slice, who has to aim far to the left of the flag. If such a golfer aims at the flag, she will certainly miss. Nevertheless, we would conclude that she can hit the green, provided that she will do so by aiming a certain distance to the left of the flag, and that she knows (or at least truly believes) that so aiming will do the trick. (This example is borrowed from SOSA, p. 111.)
Cf. note 20.
For some exceptions, however, see Chapter 7.
Rabinowicz has pointed out that if P is unconscious, or for some other reason unable to act at t 0 , CA8 implies that she can perform no future actions at t 0. (No compound action starting at t 0 then satisfies CA7 for P at t 0 .) This flaw could be overcome, by making the analysis somewhat more complicated. I will not state this more complicated analysis, however, since I will in 7.7 be concerned only with actions that are non-immediately performable in a situation of moral choice. In all such situations, of course, the agent is able to act. (Similarly, PL defines a possible life for an agent only in situations where he is able to act.)
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1995 Springer Science+Business Media Dordrecht
About this chapter
Cite this chapter
Carlson, E. (1995). Performability. In: Consequentialism Reconsidered. Theory and Decision Library, vol 20. Springer, Dordrecht. https://doi.org/10.1007/978-94-015-8553-8_5
Download citation
DOI: https://doi.org/10.1007/978-94-015-8553-8_5
Publisher Name: Springer, Dordrecht
Print ISBN: 978-90-481-4571-3
Online ISBN: 978-94-015-8553-8
eBook Packages: Springer Book Archive