Commentary on "Free Will in the Light of Neuropsychiatry"

Is Free Will Incompatible with Neuroscience?

Sean Spence sets forth some interesting approaches to the issue of free will. His concepts are provocative and his marshalling of related quotations is informative.

However, I shall argue with some of his crucial assumptions, in a way that affects the validity of some of Spence’s major conclusions. The latter are that “conscious experience is always temporally postneural activity and thus . . . ‘caused’ not causal. Our ‘control’ is a sense which may be lost. . . . Our sense of agency is apparently illusory.”

I have, of course, no quarrel with the view that even a fully voluntary act is initiated unconsciously (non-consciously). That condition was in fact demonstrated experimentally by us (Libet et al. 1983) when we found that cerebral neural activity (“readiness potential”) precedes the subject’s awareness of his/her intention or wish to act by at least 350 msec. This applied to fully self-initiated acts that occurred without “pre-planning” by the subject of when to move. (Incidentally, those finding have been replicated by others—see Keller and Heckhansen, 1990 with commentary thereon by Libet 1992, and Wong et al. 1988).

However, we also found that conscious intention did precede that actual motor act by about 200 msec. That left open the potentiality for the conscious function to control the outcome of the process that was initiated in the brain unconsciously (see Libet 1985). We have all had the experience of consciously suppressing or vetoing the performance of an urge or intention to act. The 200 msec interval provides sufficient time in which the voluntary process could be stopped, before final activation of the motor area about 50 msec before the act. We were able to show that subjects could veto an act in the 100–200 msec interval before the expected time for the action.

Spence dismisses this free will potentiality in the control or veto of a voluntary process. He says: “It is flawed because the ‘decision’ (in consciousness) to act or not to act is itself the ‘result’ of preceding neural activity.” This argument is based on my own evidence that a period of up to 500 msec of cerebral activity is required to produce a conscious sensory experience (Libet 1993, for review).

But conscious control of a process is not the same as becoming aware of the volitional intent (Libet 1985, 563–564). Control implies the imposing of a change, in this case after the appearance of the awareness of intention to act. Conscious control may not necessarily require the same neural “time-on” feature that precedes the appearance of awareness per se. There is presently no directly applicable evidence against the appearance of a conscious control function without prior neural processes specific to its initiation.

In short, one cannot cavalierly dismiss the possibility of a conscious free will function on the basis of a belief that such a control function must be itself initiated unconsciously by prior specific neural processes. The argument that conscious free will control is simply a useful delusion must still be regarded as a philosophical belief, not a scientifically demonstrated proposition.

Spence goes on to suggest that “Consciousness is necessary for purposeful (intentional) action but it is not the initiator.” But if Spence is consistent in his view that all conscious functions are preceded by specific neural activities that initiate the function, then his putative role for consciousness would have been itself initiated unconsciously. How, then, would consciousness be “performing (purposeful, intentional) actions with subtlety and invention” if it is simply an after-effect of the responsible cerebral process? It would appear that Spence is postulating an active role for consciousness in directing the course of neural activities carrying out purposeful, intentional actions. That is the kind of potentiality I was suggesting must be kept open, but one that Spence dismissed elsewhere as “flawed.”

On a technical point: The Grey Walter experiment is cited as evidence for postulated effects of timing, in the “illusion” of self-initiation of an act. However, that experimental result was simply due to the fact that the motor cortex initiates the command...