De Libero Arbitrio—A Thought-Experiment about the Freedom of Human Will

Round 1

Reviewer 1 Report

There is more research on this topic that the author might want to cite, both in cognitive science (such as Wegner) and in philosophy (Mele on the cognitive science. The first couple of chapters of Hilary Bok's book on free will may also be relevant to the author's project, in that she discusses a related type of paradox).

Substantively, the author would need to examine and rule out other possible conclusions from his thought experiment. For example, it might be argued that what this experiment shows, at most, is that the usual link between determination and in-principle predictability breaks down in the special cases the author discusses (so that our inability to predict what the the subject will do isn't conclusive evidence that the choice isn't determined). To rule out this conclusion, the author would need to argue either (i) that we really don't have evidence for determination in the absence of predictability (and thus that we don't have evidence of predictability more generally) or (ii) that we have evidence of non-determination in the absence of predictability (and thus of non-determination more generally). A worry is that (ii), which will be harder to support, may be what the author needs to establish.

Also, is the author claiming that for all we know that the subject might make a different choice in these cases, or that a different choice really is (objectively, metaphysically) possible?

Although the thought experiment is intriguing, I think the author needs to spell out more clearly what he's attempting to show and how it's supposed to follow.

Author Response

This was really a very valuable response and I have included practically all suggestions of the Reviewer into the revised version of the article. Especially his reference to the publication of Alvin I. Goldman makes clear that a principally akin thought-experiment has already been considered in 1970, at this time, however, with the idea of a perfectly informed, accurate deterministic predictive system. I have introduced this remark in the final version of the text.

Also included is the paradoxical complexity which Reviewer 1 rises. Technically, it is a paradox with limited time for a reaction. It was mentioned in the first version of the manuscript, in the final version it has got more space for details.

I have included all remarks of reviewer 1 in the second version of the manuscript, see attachement. Additionally, the manuscript has been improved by a native English reader for required English changes.

Author Response File: Author Response.pdf

Reviewer 2 Report

I really like this thought experiment, and I accept that the author simply presents it as the sort of challenge he imagines it presents, even absent any qualifications. It is a brilliant thought experiment, and should be published. However, because it is so important, I think it needs to be developed in a few ways, as follows, if this venue allows for the level of explication I will suggest; if not, it may be published in its more cryptic current form, which is provocative, in the good sense, of being philosophically thought-provoking.

Here are some of the things I think would, if developed, greatly improve this paper. The set-up, results, and two main conclusions of the Libet experiment ought to be described. The set up involves EEGs attached to subjects' brains, monitoring when their brains indicate a decision and comparing this to when subjects indicate awareness of deciding. The results show that the pre-conscious brain indicates a reliably predictive "readiness potential" signal approximately 1/3rd of a second prior to test subjects' reported awareness of their own having made a decision. The two main interpretations are, first, that because the readiness potential reliably predicts the subsequent decision, Libet concluded that the brain has decided prior to the conscious awareness of a decision, which implies that the conscious awareness of a decision is an epiphenomenon, which undermines the free will notion that volition is informed by conscious election; and second, Libet nonetheless held that this study does not foreclose the possibility that individuals retain a "veto" option, that is, they can over-ride what the brain delivers, if they so choose, and veto the brain's decision just prior to the moment of enacting it. The author should also note that more recent work has established, first, that when test subjects were asked to press either a left hand or right hand lever, analysis of their brain states reliably predicted which hand would be chosen up to six seconds prior, and second, that the readiness potential itself has been debunked as a sort of mathematical noise of the equations used to measure brain activity. These observations should be summarized as part of the context of the present paper. 

The present paper could/should also benefit from trying to link its interpretation of the thought experiment with Libet's claim that the conscious agent retains the veto power mentioned above, since the structure of the thought experiment lends itself to that interpretation. This would be useful because Libet's conclusion about the agent's conscious veto power was not based on the actual results of his experiment, but on the absence of a disproof, in which case the author's thought experiment can at least be presented as counting as a conceptual support for Libet's assertion of agents' veto power. 

Also, it would help readers assess the merits of the thought experiment if the author made the following analogy with an idea which I believe is from philosopher Alvin I. Goldman (1970), A Theory of Human Action (Englewood Cliffs, NJ: Prentice-Hall): If a perfectly informed, accurate deterministic predictive system (informed, say, by knowledge of all the laws and all the facts) predicted that I will perform a certain action (say, I will raise my hand) at time T, and I am aware of that prediction (say, I'm reading it in a metaphorical "Book of Life"), I can simply refuse to raise my hand at T. The author's thought experiment is designed to support the same idea, by having the deterministic predictive system directly monitor his brain and environment, and feed the prediction back to him. As a thought experiment, the author's is pretty good, because it places the bases for the deterministic prediction inside the test subject's own brain, and provides the test subject with what counts as neurological biofeedback together with deterministic knowledge about its own brain, thereby bringing the thought experiment much closer to home for our intuitions about ourselves and our brains, in a way that relates more directly to Libet's veto power claim than Goldman’s thought experiment does with his "Book of Life" thought experiment. 

However, there is a paradoxical complexity to this which the author does not address, and that is the fact that the hypothesis assumes, for the sake of its reductio ad absurdum approach, that the system is reliably deterministically predictive, but if so, then the system will be able to know whether the test subject wishes to rebut its deterministic predictions and will integrate that information into its predictive processing algorithms, in which case it ought to correct for such complications and nevertheless issue correct predictions about test subjects who possess the desire to falsify its predictions; Goldman addresses this. Why should we assume test subjects wish to falsify the prediction? That’s not an empirically established claim, although the author, obviously a proponent of belief in free will, is reasonable to assume it.

This aspect of the problem generates paradoxical complications akin to Newcombe's paradox and other paradoxes, but the author does not address these complications at all. They should be raised and an explanation about how the author responds to them ought to be presented, in support of a more complete presentation of this thought experiment as an argument for free will. 

I do think, overall, however, that this is a brilliant thought experiment, for the reasons stated above. I’m unfamiliar with the word length norms of this publication venue, so I’m not sure if the publication allows for the level of exposition which I’m suggesting would make this article much more defensible. But if not, I recommend publication with minor edits, both proofreading edits (of which there are only a few, English language based edits needed), and a few remarks summarizing my main concerns, if and to the extent that there is room for them.

Author Response

Reply by author: I have included many of the remarks of reviewer 2 in the second version of the manuscript, with the exception of the final passage:

Substantively, the author would need to examine and rule out other possible conclusions from his thought experiment. For example, it might be argued that what this experiment shows, at most, is that the usual link between determination and in-principle predictability breaks down in the special cases the author discusses (so that our inability to predict what the the subject will do isn't conclusive evidence that the choice isn't determined). To rule out this conclusion, the author would need to argue either (i) that we really don't have evidence for determination in the absence of predictability (and thus that we don't have evidence of predictability more generally) or (ii) that we have evidence of non-determination in the absence of predictability (and thus of non-determination more generally). A worry is that (ii), which will be harder to support, may be what the author needs to establish.

Reply by author: I think to rule out the other possible conclusions mentioned by reviewer 2 is not possible within this paper. The two arguments (i) and (ii) concern principal philosophical definitions of “determination” and “predictability”, which sometimes are even taken synonymously. In quantum-mechanics, for example, we can “predict” the results of an experiment if many objects (f.e. electrons) are part of the experiment, however, we cannot predict the results if only one of the quantum-mechanical objects is being investigated. The dispute among physicists about the interpretation of such experiments is ongoing and books have been written about this issue of evidence for determination and predictability.

Author Response File: Author Response.doc

Round 2

Reviewer 1 Report

"Consequentially" in the abstract should be "consequently."