Will you or will you not?

Most philosophers of action prior to the twentieth century seemed to have thought that actions were caused by volitions. Ryle’s attack on the notion of volition in The Concept of Mind seems to have killed off the notion. But, like a lot of other previously unfashionable notions, volitions have been making something of a comeback in recent years. Some (relatively) recent defenses of volition: Ginet’s On Action (1990); Lowe’s Subjects of Experience (1996), Pink’s Psychology of Freedom, as well as a whole series of publications by our own Jing Zhu. Some attacks on volitions: Adams and Mele (1992), ‘The Intention/ Volition Debate’ and Berent Enç’s How We Act (2003).

Why think volitions are required? Ryle’s main argument against volitions was a regress argument (embedded in a dilemma: either volitions are voluntary or they are not. If they are not, then how can the actions they cause be voluntary? If they are, then don’t we need to postulate a volition as a cause of our volitions?). Volitionists tend to reply with a regress argument of their own: typical human actions can be reconstructed as action-chains, in which each link is caused by the one prior to it. There must be a first link, on pain of infinite regress, and that first link had better be an action (if it is not, it is hard to see how the subsequent links attain this status). Volitionists suggest that this first link is an act of will.

Danto uses a similar regress argument, for the claim that action chains begin with basic actions (rather than volitions). Why should we accept the volitionists claim that action chains begin instead with volitions? At this point, volitionists tend to wheel in their second argument, the argument from paralysis. Suppose that S’s arm is paralysed, unbeknownst to her. S is blindfolded and asked to move her arm. Then her blindfold is removed, and she is surprised to see that her arm has not moved. She has failed to perform a basic action. But she has done something; she tried to move her arm, and this trying equates to a volition.

I’m a little puzzled by this argument. It seems to me that it provides no real support for either side. Everyone should agree that actions begin in the brain. We can accept that certain brain events took place in S (and perhaps not only brain events, but also nerve firings). Volitionists say that these events are acts; others might say that they are the forming of proximal intentions. Are they really disagreeing about anything , apart from what to call these events? Debates about what to call something are best settled by stipulation.

Let me turn now to Jing Zhu’s arguments for the existence of volitions. One thing Zhu invokes repeatedly is action slips. Sometimes, when we intend to do or say something, we end up doing or saying something else; sometimes to our great embarrassment (and sometimes with disastrous consequences). Zhu claims that action slips show that we need volitions to keep our actions on course. I think they show the exact opposite. Action slips occur when one over-learned action sequence is replaced by another, similar one. What they show is the extent to which ordinary voluntary actions do not resemble the kind of effortful and mindful acts that volitionists have in mind. They are typically much more automatic than that. The monitoring of actions which usually keep them on track is carried out below the level of conscious awareness, and there is a limit on the extent to which we can make it conscious (how do you manage to monitor your talking? The only way to do it consciously is to slow down, to the extent necessary for you to rehearse every word before you say it. It goes without saying that we almost never do this. We just trust that the scripts we rely on will keep us on track, and think about what we are talking about, rather than about how we talk. The exact phrases I use in speaking and writing can come as a surprise to me).

The only way to rescue volitions, in the light of implausibility of their playing a role in the 97% of our actions which are more or less automatic, is to restrict them to a small class of especially complex and novel actions, à la James. Zhu endorses something like this approach, in ‘Locating Volition’. But (a) if volitions are restricted to a small class of voluntary actions, then all those arguments about how we need to postulate volitions to explain actions are no longer permissible. And (b) it is very hard to isolate a class of actions which require volitions without circularity or stipulation. Zhu suggests that actions elicited by the environment do not require volitions, whereas actions internally generated do. But it is plausible to think that all actions are both internally generated and environmentally elicited. We might go along with James, and think volitions are only necessary in cases of conflict. But breakdowns of the mechanisms of action inhibition (in utilization behavior, imitation behavior, and so forth) in which people fail to inhibit their responses show the extent to which almost all actions involve conflict. If volitions are necessary anywhere, they will be necessary in routine actions too. And they don’t seem to be necessary in routine actions.

Comments

[Jing Zhu]

Neil:
Thanks for bringing up this topic. I offer some comments below.
1. A explicit volitionalist response to the regress argument, along the lines suggested in your post, was developed by Hugh McCann:
“Volition and basic action”, Philosophical Review, 83 (1974), 451-473
“Trying, paralysis, and volition”, Review of Metaphysics, 28 (1975), 423-442.
Both reprinted in his 1998 book The Works of Agency (Cornell University Press).
2. I tend to agree with you that the discussions around the regress argument provide no real support for either side. Thus I attempted to take another route (in “Understanding volition”, Philosophical Psychology, 17 (2004), 247-273). By reflecting on the debates in philosophy and psychology between mentalism and behaviorism and about mental imagery, where similar regress arguments had once been heavily invoked, I suggested that we should not dismiss profound conceptions on the nature of mind and block potentially promising research programs on face value from merely a priori conceptual analyses, but rather take it as a gadfly to stimulate further empirical inquiries. My own efforts have been an attempt to develop a more empirically inspired account of volition (see “Locating volition”, Consciousness and Cognition, 13 (2004), 302-322; “Reclaiming volition: An alternative interpretation of Libet’s experiment, Journal of Conscious Studies, 10.11 (2003), 61-77).
3. A unifying conception of volition I suggested is to view volitions as special kinds of mental action which bridge the gaps between deliberation, decision, and action. Volitions, conventionally understood as acts of will, can be construed as: (1) decision-making (will-setting); (2) action initiator; and (3) executive control. I have tried to make sense of this account (in “Locating volition” and “Reclaiming volition”) by employing some recent findings in cognitive neuroscience.
4. Regarding action slips, what is the difference between a slip of the tongue and the voluntary imitation of the slip just made or the mindful and effortful correction of the slip? A nature answer would be that the latter involve some volitional processes (this point was suggested by Bernard Baars). In this case, volitions are involved as executive control, implicating selective attention, mental effort and willful self-control.
5. Regarding the majority-minority argument which attempts to decrease the significance of volition in our daily life, there are two responses. First, even if a great deal, even the vast majority of human thinking, feeling, and behavior operate in automatic fashion with little or no need for conscious, intentional control, the evidence does not guarantee that conscious will is an illusion or epiphenomenon. The conscious, mindfully controlled mental processes may account for only a small part of the total amount of our mental life, like the tip of iceberg, but this part is nonetheless significant. Conscious, controlled activities are essential in such tasks as planning, decision-making, error-correction or “trouble-shooting”, learning and acquiring new skills, and overriding strong habitual responses or resisting temptations. The ability of conscious control of one’s thought and action can provide substantive adaptive advantages that help a species deal with the environmental and social pressures. Conceptually, it seems possible that an agent is just awake and being aware of what are happening upon him, with no need of the works of volition. But in this case, the agent can hardly be said to have a will. Second, it is quite arguable that many effortless thoughts and actions which are more or less automatic, such as decisions by recognization and some habitual behaviors, are developed from once volitional processes through learning and exercise. This is somewhat similar to the case of what Robert Kane calls the conditions of Ultimate Responsibility.

[Neil]

Thanks for the reply, Jing. One worry I have is the role of consciousness on your view. You seem to want to identify volitions with a conscious process. But not all controlled processes are consciousness processes. The great majority are not. My actions are voluntary (as you seem to agree) so long as they have the right kinds of causes, whether I am conscious of those causes or not, whether I have consciously deliberated or not.So what role are volitions playing? Much of the activities in which I am mindful of what I do is unskilful (that's why it is closely monitored), conversely, much of my more automatic action is skilful and important. Second, even when I am conscious of what I do, it is far from clear that my consciousness contributes to my control. Consciousness lags behind the subpersonal mechanisms which correct for error (think of Jeannerod's work on error correction). And what I am conscious of is often my decisions and judgments, and not my reasons for these, and when I am conscious of the reasons, the assigning of weight to considerations is done below the level of consciousness... As Enc points out, in a foundationalist account the foundation has to have the feature which it is transfered to higher links in the chain. But the foundations of conscious processes are nonconscious processes which do a lot of the really important work. None of this is meant to deny the reality of the will, in some sense of the word. If by will we mean what enables us to make rational decisions and act on them, and what is missing in various pathologies, then we have a will. But volitions seem to be more than the workings of such a will (which can be largely subpersonal), and I'm unconvinced we need them.

[Clark Goble]

Quick tangental comment. Why does everyone automatically assume infinite regress is bad? It seems like some philosophers, most notable C. S. Peirce, allow for true continuity (i.e. infinity) and sometimes real higher order infinities. (Kelly Parker has a great book on this in Peirce) Peirce explicitly uses an infinite regress to argue against mind/body dualism in explaining how mind can affect matter. But I really can't think of too many folks outside of Peirce who do this. Typically philosophers follow a more Aristotilean approach that assumes infinities are bad and we need some variation of an "unmoved mover" or first cause.

[Heath White]

The paralysis argument for volitions strikes me--and maybe I have read this somewhere--as parallel to two other arguments. One is the argument for sense data based on hallucination. If I hallucinate a dagger, I'm seeing something; _what_ is it that I see? The answer we're supposed to give is, a sense-datum. If I see a sense-datum when there is no dagger, then I see a sense-datum when there is a dagger too.

The other argument is for assertions as expressions of belief, based on the possibility of silent belief. There seem to be lots of things I think but don't say; I can think that p without saying that p. So when I say that p, we should conclude that I believe that p as well. And even though there is no conscious "believing" process going on 95% of the time I am saying anything, we should hypothesize that the belief is in there anyway.

Now, I think these two arguments have very different merits. But I wanted to draw attention to the structural similarity. Jonathan Dancy has an article, "Arguments from Illlusion" which draws out this structure in a number of areas.

[Neil]

Clark,

I have no defence for the view that infinite regresses are always bad. But we naturalists think they're bad in philosophy of action (at very least). We think that actions begin in the brain, and if there is an infinite regress of neural processes (or if neural processes themselves depend upon extra-neural processes, which go on forever) then action can't begin at all.

Heath,

Bernet Enc has an interesting discussion of the parallel between the paralysis argument and the argument from hallucination, in how we act. I'm not sure that the argument from assertions as expressions of belief is really parallel, though - where's the dysfunction? In any case, I don't think Enc's argument - the parallel sense-datum account is implausible, therefore so is the volitionist account - is very strong. Whatever actions are, they begin in the brain, and neither side gets any leverage from pointing this out.

[Jing Zhu]

Neil:

Regarding consciousness and volition, there are indeed a host of difficult questions. David Rosenthal once suggested that the conception of volition should be extended to cover unconscious mental processes. But I think this proposal is implausible. Volitional actions are the paradigmatic of all sorts of voluntary action, essentially relevant to our sense of free agency. So it seems conceptually inappropriate to adopt the term “nonconscious volition”, for this gratuitously diminishes our valuable philosophical and scientific vocabulary.

There may be different ways to unpack the claim that “the foundations of conscious processes are nonconscious processes which do a lot of really important work”. Consider a large business company (e.g., Microsoft), there is usually a high level of administration that makes plans and important decisions, launches key projects, and oversees the perform of the whole organization. There are also middle and low level administration and varieties of working forces, which take over the majority of labor to carry out the company’s plans. In a sense of the word, you can say the lower levels are the foundations of the higher level; without them, nothing can be really achieved. But it would be misleading to claim that the processes occurring at the lower levels, which take the lion’s share of the labor, are the norm or essence of a large company’s activities. Volitions are implicated in setting a will and carrying out a will. They are most similar to what the CEOs are doing.

[Neil]

One word, Jing: homunculus. The 'CEO' (what is it: Shallice and Norman's supervisory attentional system?) is either built out of lower level processes, or we get exactly the same problems in explaining its operations that it was designed to solve.