Benjamin Libet’s famous experiments, which showed that brain activity precedes the awareness of the intention to carry out simple voluntary motor acts, have forced philosophers to reexamine free will.  Many have taken Libet’s experiments to confirm incompatibilist hard determinist accounts of free will, in which free will is an illusion, or at best compatibilist accounts, in which free will is not what we commonly intuit it to be.  However, others, including Libet himself, have reconciled experimental results with libertarian incompatibilist accounts of freedom.  In this blog post, I present an interpretation of Libet’s experiment which reconciles his experiment with libertarian freedom.  (This is fairly long) 

§1 The Experiment

Libet’s experiments were performed to answer the question:  Does the conscious will to act precede or follow the brain’s action?  The focus of the experiments was on present tensed actions, as opposed to long term deliberations and advanced making of choices about performing motor acts.  Therefore, the actions the subjects performed were simple motor actions—flicking a wrist—that were performed spontaneously, within a given time frame.  

Libet adjusted a cathode ray oscilloscope to act as a timer.  A single dot traveled in a clock-wise motion in forty-three millisecond increments.  The experimental subjects reported the location of the dot when they were aware of the intention to carry out a simple motor movement.  An electroencephalogram (EEG) was used to monitor the brain activity—specifically measuring what is termed a Readiness Potential (RP).  The RP is a slow electrical charge on the scalp that preceded the voluntary motor acts.  

During the experiment, subjects sat in front of the oscilloscope timer with EEG electrodes attached to their scalp.  The subjects were told to flick their wrist within a short time frame.  When the subjects became aware of their intention to act, they were required to report the position of the dot on the oscilloscope.  Researchers compared the RP to the oscilloscope timer’s reported position by the subject. 

The experiments resulted in showing that the conscious will to act followed the brain’s action.  

Specifically, the RP preceded the voluntary motor act by 550 milliseconds on average.  The awareness of the intention to act preceded the voluntary motor act by 200 milliseconds average.  Therefore, the RP preceded the intention to act by 350 milliseconds on average.  According to Libet, “The brain initiates the voluntary process first.  The subject later becomes consciously aware of the urge or wish to act, some 350-400 milliseconds after the onset of the recorded RP produced by the brain.” 

 §2  Prima Facie Implications

Libet defined a free act, as tested by his experiment, as follows:  “The will to act arose endogenously.  That is, there were no external cues for performing the act; no external limitations on when to perform the act; and most importantly, the subject felt she was responsible for the act and also felt that she could control when to act as well as whether or not to act.”  Put more simply, Libet writes, “It has been commonly assumed that in a voluntary act, the conscious will to act would appear before or at the start of the brain activities that lead to the act.”  As Libet was most likely testing a free action as the phrase is commonly used, I use the following descriptive definition:

Free action: x is a voluntary action iff_Df the cause of x is an agent A and A is not caused to cause x and A is aware of the intention to perform x at a time t simultaneous with A’s brain’s initiation of x or prior to A’s brain’s initiation of x and A is not limited to performing x at a certain time t and A could do otherwise than perform x

The definition assumes that A caused the brain’s initiation of x and A is aware of the intention to perform x prior to or simultaneous with the brain’s initiation of x.  Yet, the experiment called this definition into doubt.  The brain’s initiation of x (RP) precedes A’s awareness of the intention to x.  In other words, unconscious brain processes initiate ‘voluntary’ actions before one is consciously aware of her intention to act.  ‘Voluntary’ actions are posterior to unconscious, involuntary cerebral processes in two senses: 

 S1.  A temporally posterior sense:  The awareness of the intention to x occurs at a later time than the unconscious initiation (RP) of it.

 S2.  A causally posterior sense:  x is likely an effect of the RP.

This resulted in part of the definition being experimentally disconfirmed, “A is aware of the intention to perform x at a time t simultaneous with A’s brain’s initiation of x or prior to A’s brain’s initiation of x.”  Libet offers his own response in how his experiment challenges the common assumption of a free act, “. . . free will, if it exists, would not initiate a voluntary act.”   

Daniel Wegner takes Libet’s conclusions further, claiming that the experiment challenges two additional aspects of the definition of x:  The source requirement, “the cause of x is an agent A and A is not caused to cause x,” and the alternate possibilities requirement, “A could do otherwise than perform x.”  A is not free to do or not to do x and A is also not responsible for x.  Instead, Wegner claims that the notion of A acting freely is actually an illusion.  Wegner writes, “The brain creates both the thought and the action, leaving the person to infer that the thought is causing the action.”  (66 TMBT)  As a consequence, Wegner argues that Libet’s experiments are the ace-in-the-hole against free will in favor of hard determinism.  Wegner writes, “The timing of will, finally, seems to seal the fate of that elusive light-bulb.  The detailed analytical studies of the timing of action indicate that conscious will does not precede brain events leading to spontaneous voluntary action but rather follows them.”  (60 IOCW)  If Wegner’s interpretation is correct, we do not have free will.

So the prima facie implications of Libet’s experiments result in free will being an illusion.  I understand the argument as follows:

 If an agent A has libertarian free will then A can perform an action x freely

1. If an agent A has libertarian free will then A can perform an action x freely
2. If A can perform x freely then A is able to do otherwise than x (alternate possibilities)
3. If A is able to do otherwise than x, then x is consciously initiated (source)
4. x is not consciously initiated by A (Conclusion from Libet’s experiment: The RP is not consciously initiated by A)
5. Therefore, A is not able to do otherwise than x (MT 3 and 4)
6. Therefore, A cannot perform x freely (MT 2 and 5)
7. Therefore, A does not have free will (MT 1 and 6)

 §3 Libertarian Interpretation

Yet I argue that the argument is not sound.  In what follows, I argue against premises 4.  Before launching into my argument, I first qualify what I intend to establish in my negation of 4.  Secondly, I define my terms. 

In my proposed interpretation of Libet’s experiment, I will not be arguing for ~(x is not consciously initiated by A).  Instead I argue for the weaker ◊ ~(x is not consciously initiated by A).  That is, I intend to show that while premise 4 is compatible with Libet’s experiment, the experiment does not entail 4.  From this qualified understanding of my negation of 4, I show that the part of the definition of a free act that Libet’s experiments call into question can be defended in a qualified way.  Specifically, I show that ◊(A is aware of the intention to perform x at a time t simultaneous with A’s brain’s initiation of x or prior to A’s brain’s initiation of x).

For my interpretation, I use the terms ‘Awareness’ and ‘Aware’ for intentional states with a mind-to-o direction of fit and an o-to-mind direction of causation.  I use o in place of world (as in mind-to-world), with o standing for the intentional object, as the intentional objects of Awareness can be other mental states or external objects in the world.  When the intentional object is another mental state, I use the terms ‘M-Awareness’ and ‘M-Aware’.  The intentional object of M-Awareness is o_m.  When the intentional object is an external object or event, I use the term ‘E-Awareness’ and ‘E-Aware’.  The intentional object of E-Awareness is o_e.  To say that an agent A is E-Aware of o is to say that A has the mental state of E-Awareness and the content of the state is o and the conditions of satisfaction of A’s E-Awareness of o is o’s existence. 

I use the terms Intention and Intend for intentional states with an o-to-mind direction of fit and a mind-to-o direction of causation.  To say that A intends x is to say that A has an intention to x in which the content of A’s intention is x and the conditions of satisfaction of A’s intention to x is x’s being completed.   The intentional object of A’s intending is o_i.

Now onto the experiment.  In what follows, I first cash out Libet’s experiment, as his experiment is normally understood, in the terms I defined above.  Let t₀ be the time of the completed motor action x of flicking the wrist. 

 t_{-550 msec} = RP

A intends to x.  That is to say, A has an intention to x, and the intention to x is an intentional state of A with world-to-o direction of fit and mind-to-o direction of causation.  The intention to x has a content that can be satisfied and the intention to x is satisfied iff x.  x is the intentional object of A when A has the intention to do x.  As x is the intentional content, x=o_i.

 t_{-200 msec} = Awareness of the intention to act simultaneous with the position of the clock

A is M-Aware of the intention to x and A is E-Aware of the position p of the clock.  That is to say, A’s M-Awareness of the intention to x has a mind-to-o direction of fit and o-to-mind direction of causation.  The intentional object of A’s M-Awareness is the intention to x, therefore the intention to x = o_m.  Further, p is the intentional object of A’s E-Awareness, with A’s E-Awareness standing in a relation of mind-to-o direction of fit to p and o-to-mind direction of causation.  With respect to A’s E-Awareness of p, p=o_e.  

Presumably, A’s M-Awareness of o_e occurs at the same time as A’s E-Awareness of o_m .  That is to say, the experimental subject A notes the position of the clock p at the same time that A is aware of A’s intention to flick A’s wrist.  But that is where the problem lies in the normal conclusions for Libet’s experiment.  Libet presents the scenario as A being aware of the intention to flick a wrist beginning at a time simultaneous to p.  What if A is first aware of A’s intention to x, and second aware of p?  That is to say, what if the lag in awareness of the intention to x is actually a lag in the time it takes A to shift focus from A’s M-Awareness of the intention to x to A’s E-Awareness of p?  I propose that the alleged lag in awareness of the intention to x is actually that—a lag from A being M-Aware of the intention to x to A being E-Aware of p instead of a lag from A’s intending x to A’s being M-Aware of the intention to x.  So the scenario could be as follows:

 t_{-550 msec} = RP

A intends to x and A is M-Aware of A’s intention to x.  That is to say, A has an intention to x, and the intention to x is an intentional state of A with an o-to-mind direction of fit and mind-to-o direction of causation.  The intention to x has a content that can be satisfied and the intention to x is satisfied iff x.  x is the intentional object of A when A has the intention to do x.  As x is the intentional content, x=o’_i.

Simultaneous with A’s intention to x, A is possibly M-Aware of A’s intention to x.  A’s M-Awareness of x is an intentional state with a mind-to-o direction of fit and an o-to-mind direction of causation.  The intentional object of A’s M-Awareness is the intention to x, therefore the intention to x = o’_m.

 t_{-200 msec =} A is E-Aware of p 

A’s E-Awareness of p is an intentional state with a mind-to-o direction of fit and an o-to-mind direction of causation.  With respect to A’s E-Awareness of p, p=o’_e.  

Is this scenario plausible?  I argue that it is.  Specifically, this scenario is one of three interpretations of Libet’s experiment that coheres with the results.  The three interpretations are as follows:

 (i)   A is M-Aware of the intention to x at a time simultaneous with A’s E-Awareness of p

(ii)   A is M-Aware of the intention to x at a time after A’s intention to x and before A’s E-Awareness of p

(iii) A is M-Aware of the intention to x at a time simultaneous with A’s intention to x and before A’s E-Awareness of p

In Libet’s interpretation, (i) is what is assumed.  Yet I argue that (ii) and (iii) are equally plausible, given that it is possible that the delay in consciousness of the intention to act could in fact be a delay from the subjects shift in from introspection to normal perception.  That is, the delay could be when A shifts from being M-Aware of the intention to x to being E-Aware of p.  There could still be a delay from A’s intending x to A’s M-Awareness of A’s intention to x yet with A still being M-Aware of the intention to x prior to A’s E-Awareness of p.  Therefore, there are three equally plausibly interpretations of Libet’s experiment, and one of the interpretations supports the original descriptive definition of free will I offered as well as possibly negating premise 4. 

If you recall, part of the definition of a free action that Libet’s experiment called into question is “A is aware of the intention to perform x at a time t simultaneous with A’s brain’s initiation of x or prior to A’s brain’s initiation of x.”  If (iii) is a possible interpretation of the experiment, then the results of the experiment do not necessarily negate that part of the definition. 

Furthermore, the alleged falsification of that part of the definition also supports premise 4:  x is not consciously initiated by A.  This is because the initiation of x occurs before A is conscious of x.  However, if A is M-Aware of the intention to x simultaneous with A’s intending x, then it is possible that A consciously initiates x.  As there are three possible interpretations of Libet’s experiment, and (iii) coheres with the definition of a free action and free will, I argue that Libet’s experiments do not entail determinism.  Instead, there is still a 1/3 probability that libertarianism is the case.  As libertarianism is probable, ◊ ~(x is not consciously initiated by A).