A Reply to an Objection Against the Agent-Causal Version of Libertarian Free Will

            In the following, I spell out an objection to the agent causality theory of libertarian free will. I then address the objection. But, first, I should say just a little about the theory itself. Proponents of agent causality contend that, pace typical views of causality, causal relata are not always or exclusively just events. It might be the case that the only things that can be caused are events, but this does not imply that all causes must be events. It is also possible, they will insist, that things can sometimes act as causes as well. In particular, agents are such things that can act as causes. Take a case in which someone moves his hand. In this such case, an event in your brain, e, sends a signal to your muscles telling them to move in such and such a way. Moreover, e is not caused by any other events, because it is caused by something that is not an event, i.e., the person moving his hand. As it applies to free will, the basic motivation behind the theory is not hard to figure out. Proponents of the view want to argue that agents, in the sense of free agents, act such as to cause e, but are not caused to cause e. That is, free agents are directly responsible for causing e.

            Now, the objection I have in mind is this: imagine the following case. Yasmine asks Imran to pass the salt, and Imran immediately responds and does so. If there is such a phenomenon as agent causation, this, surely will count as an example of it, i.e., an example of a case in which Imran causes his own action. But it also seems clear that in this case there is an event outside of the agent that also causes the relevant action. For Yasmine's asking Imran to pass the salt seems like an event; it is most certainly not a thing, and thus not possibly an agent.

            My preliminary to my response to this objection is that it is not immediately clear that 'cause' is used in the exact same sense in the case (a) in which Yasmine asks Imran to pass the salt and in the case (b) in which Imran passes the salt to Yasmine. In any case, it is clear that the proponent of agent causality would want to claim that

(EC) For any persons x and y and any time t and any act a, if x can act upon y at t in such a way as to causally influence y to do a at t, then x is a cause of y's doing a at t. 

For it seems obvious that many things in the world act upon other things in such a way as to causally influence these things in various ways, and in causally influencing the thing in question, they are said to be causes of the effects which the thing brings about. In the present case, the proponent of agent causality would have to say that the agent x acts upon y by producing an event, e, that causes y to do a. Here, it is not enough simply to say that x causes y to do a, because the intermediary through which x causes y to do a is e. And if e does not cause y to do a, it is not the case that x causes y to do a. But, clearly, one wants to say that x causes y to do a; so one must say that e causes y to do a. 

            So, here, there are two responses, I think, that the proponent of agent causality could make. First, he could insist that at least some causal relata are such that they take place directly between things, i.e., in this case, x acts directly on y at t such as to cause y to do a. Now this response does not strike me as an implausible one, and, ideally, it is probably the one I favor. It seems clear, however, that not all philosophers would accept the stronger principle this reformulation implies, so, without the time to defend this move in depth, I will here work towards a more modest strategy. This leads me to the second potential response a proponent of agent causality could make, which is to argue that the critic's claim that the event, e, is a cause of y's doing a relies on an equivocation.

            Towards beginning to work towards this second response, one thing to point out is that it is clearly not true that x's causing y to do a is a determining cause of y's doing a. Taking into consideration one factor that philosophers consider relevant in assessing whether something is a determining cause, one should note that (1) it is not the case that x's causing y to do a does not entail, with a proposition that is true at t, that y does a. Nor is x's causing y to do a sufficient with the laws of nature and the previous states of the universe prior to the present time t, to cause y to do a. So the sense of 'cause,' as it is used in the example, seems clearly different than that which would pose potential difficulties for y's freedom with respect to whether y does a at t. But, of course, the objection was not that e's causing y at t inhibited y's freedom; it was that e's causing y to do a at t should not be true at all. What should be true is that y causes a full stop, with no further causal relata entering into the picture.

            But, as I suggested, the proponent of agent causality should here insist that there is an equivocation between 'cause' in each of the two cases, between e's causing y to cause a and y's causing a, as an agent. What sense, then, if this is true, is e a cause of y's causing a? Here, a plausible suggestion is that e causally influences y such that y causes a. Causal influence can here be taken in one of three senses:

(a) x's causally influencing y through an event, e, at a time t is necessary for y's doing a.

(b) x's causally influencing y through an event, e, at a time t is sufficient for y's doing a.

(c) x's causally influencing y through an event, e, at a time t is neither necessary nor sufficient for y's doing a.

In a situation like (a), x's causally influencing y through e at t is necessary in the sense of being a necessary condition for y's doing a at t. Examples of what a might include in (a) are the following: x's giving b a boost to reach something y could not reach at t without x's help; x gives y a piece of information at some time t that y requires for doing a at t; x asks y to do a at t such that y could not have done a at t in the absence of x's asking him.

            The last option here, of course, is supposed to represent the kind of case in which Yasmine asks Imran to pass the salt. A moment's reflection, however, will reveal that Yasmine's asking Imran to pass her the salt cannot be a necessary condition for Imran's passing the salt. For, though it would be admittedly bizarre, Imran might decide that Yasmine's food needs more salt, and accordingly, and without her request, passes her the salt. Perhaps, more plausibly, Imran realizes that Yasmine has not included any salt in her food -- perhaps she simply forgets. Being the gentleman that he is, Imran accordingly passes Yasmine the salt. For, for at least the sort of case directly under consideration, it is not true that Yasmine's asking Imran to pass the salt is a necessary condition for his passing it.

            In a situation like (b), x's causally influencing y, through an event, e, at a time t is by itself a sufficient condition for b's doing a at t. This would include such activities as x's hypnotizing y at some time prior to t and then commanding y to do a at t. Or, alternatively, it might include x's using a machine by which x effectively assumes control of b's body, and thus of all of b's bodily actions. Or, perhaps, it might include x's uttering a specific word in y's presence with the full knowledge that if x affects y thusly at t, y will do a at t. In any of these cases, it is obvious, y is in no way free with respect to his doing at at t. But this seems to satisfy the condition I discussed earlier in which a's action joined with the truth of the proposition:

(LF): if x makes is true at some time t that y does a at t, then y does a at t,

seems to logically entail that y does a at t. But this does not, obviously, give an accurate account of what takes place with respect to Yasmine's requesting of Imran that he pass her the salt. For it is clearly not true that Yasmine's requesting of Imran that he pass her the salt is not a sufficient cause of Imran's passing Yasmine the salt. No theory should have this consequence if one is to regard it as a plausible account of free agency; nor is there anything that compels the proponent of agent causality to say that Yasmine's asking Imran to pass the salt is a sufficient condition of his passing her the salt.

            That leaves the remaining situation one has to consider as (c). And in (c), x causally influences y, through an event, e, such that it is true that x causes y to do a at t, but x's causing y to do a at t, is neither a necessary nor a sufficient condition of y's doing a at t. But in what sense, then, is it true that x causes y to do a? One way in which is seems possible to say that 'x causes y to do a' is true in the sense of (c) is that, in affecting y through e, x gives y a reason in virtue of which y does a at t. In this sense, I submit, there is nothing that would force one to say that x acts on y, through e, in the sense of x's causing y to act is anything like the sense of 'cause' involved in either sense (a) or sense (b). x's giving y a reason to do a at t, in other words, is just not the sort of cause that satisfies either (a) or (b). For, as most philosophers I think will agree, a genuine cause in the sense that one generally uses the term will include either (a) or (b). In no case, I submit, will it include (c), unless one uses 'cause' in a very broad sense to include something like 'explanation.'

            And it does seem that reasons for acting count as causes in the broad sense of a cause. Here, perhaps, one might bring Aristotle's distinction between causes in the sense of explanations to the fore. 'Cause' as I have in mind here would seem to fit nicely with what Aristotle would term a 'final cause'  in the sense of the object that an agent wills in acting for some good. I find this distinction between causes in Aristotle's sense here to be helpful, but I don't think anything of ultimate importance hinges on it with respect to what I want to argue here. So, again, lacking causal efficacy in the sense of either (a) or (b), there is a definite sense in which a cause in the sense of (c) is not a cause in the same sense in which (a) and (b) are causes. Yasmine's asking Imran to pass the salt is neither necessary, in the sense of being a necessary causal condition, nor a sufficient cause of Imran's passing the salt. And, in this sense, 'reason' in the sense of a 'reason for acting' seems to fit the bill well for something that does not fall into either category of (a) or (b).

            Granting that this is right, however, there still seems one last obstacle to the proposal I want to make. Typically, proponents of agent of causality have wanted to affirm that an agent's wiling a is sufficient for the agent's doing a. But this faces an obvious objection. If an agent's willing a is sufficient for the agent's doing a, then there is not anything that could obstruct the agent from doing a if it was in his power to will a. And this is clearly false. Nor is it enough for the proponent of agent causality to amend the account by stipulating that the agent's willing a is typically sufficient for the agent's doing a, as this simply translates to the claim that the agent's willing a is not sufficient for the agent's doing a.

            One possibility here is that, borrowing from J.L. Mackie's more general account of causality, the proponent of agent causality can say that the agent's willing a is a necessary condition that, with all other necessary conditions for the agent's doing a at t, is sufficient for the agent's doing a at t. And this last suggestion seems to me prima facie plausible. But taking this account seriously, one would have to revise one's judgment and then treat e's causally affecting y as a necessary condition for y's doing a at t. Does this, then, not concede that e is a cause of y's doing a in the sense of (a)?

            I do not believe so. All that is needed for the proponent of agent causality to say is that there is some reason that is a necessary cause for y's doing a at t. That the reason turns out to be Jasmine's asking Imran to pass her the salt is incidental to Imran's passing her the salt. That is, it is still not a necessary condition for Imran's passing her the salt, and so the original judgment I made about the event in which Jasmine asks Imran to pass her the salt can still be classified under (3) and not (1) So, having argued thus, I conclude there is a good case to be made that the sense in which 'cause' is being used with respect to the original objection that Yasmine's asking Imran to pass the salt rests on an equivocation between 'cause' in the sense of (a) or (b) and a distinct sense of 'cause' in the sense of (c). And, as I have spelled it out and I hope have made clear, the two senses are distinct.