Abstract
Much recent work in neuroscience aims to shed light on whether we have free will. Can it? Can any science? To answer, we need to disentangle different notions of free will, and clarify what we mean by ‘empirical’ and ‘testable’. That done, my main conclusion is, duly interpreted: that free will is not a testable hypothesis. In particular, it is neither verifiable nor falsifiable by empirical evidence. The arguments for this are not a priori but rather are based on a posteriori consideration of the relevant neuroscientific investigations, as well as on standard philosophy of science work on the notion of testability.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Libet (2002, 292) gives a list.
Two cases of this already: first, several writers argue that free will should be seen as something that comes in degrees (e.g. O’Connor 2009). But I will focus on the simpler issue of whether there exists free will to any degree at all, arguing that even that is untestable. Second, it is disputed by some that we must require free behavior and random behavior to be distinguishable. If we do not then free will is much harder to test (Sect. 3 below), but generally I will assume that we do.
This does assume that the unaccounted for portion cannot legitimately be put down to unknown causes other than free will—an important condition. I return to this point at the end of Sect. 4 below.
As with ‘empirical’, by ‘scientific’ and cognate terms I mean third-person investigations.
In practice, in the eyes of many even a proven absence of this sort would not be deemed enough for verification. Most scientists would undoubtedly further demand a detailed positive theory of free will, which made novel empirical predictions, could be related to other scientific theories, and so on. Perhaps imposing this further demand is equivalent to imposing an a priori commitment to methodological naturalism. But my concern in this paper is that empirical verification of free will may still be impossible even without such a commitment.
A proven absence of prior physical causes also does not rule out some non-physical cause (other than free will). But, although this is undoubtedly a logical possibility, I do not think that either side of the free will debate takes it seriously, in which case it is not dialectically significant. If we did take it seriously that would only make verification still harder, so ultimately supporting this paper’s thesis.
Even then, to establish causation we would need to rule out the readiness potential and subsequent physical action being merely two independent effects of a common cause. But for our purposes we may ignore that possibility, since it would still imply the action having been caused by some prior unconscious event.
That said, this interpretation of the neurological evidence is not unanimous. Recently, Schurger et al (2016) have argued that the build-up in the readiness potential is more akin to background noise than to any signal or ‘decision’ to act. If so, falsification is even further away than will be argued in the text.
Various other types of empirical evidence have been claimed to threaten free will too, such as some results from social psychology or the symptoms of some clinical mental disorders. But these too are unconvincing falsifiers—see, e.g. O’Connor (2009, 179–180) for discussion.
Perhaps a breakthrough in the metaphysical interpretation of quantum theory would generate a consensus that the universe as a whole is deterministic, thereby rendering incompatibilist free will empirically falsified without any need for neuroscience at all. But: first, this currently seems an unlikely deus ex machina. Second, it would not adjudicate between compatibilism and incompatibilism, thus still leave open which of those is correct, and thus still leave open whether we have free will. And third, at least one incompatibilist, namely O’Connor (2000, chapter 6), denies that it would falsify free will anyway, on the grounds that a deterministic quantum mechanics at the micro-level still would not rule out emergent—and possibly indeterministic—causal powers at higher levels.
Mele is one exception (2009, 157–158). Roughly speaking, he argues that falsification would follow from any demonstration that our psychological decision-making system is deterministic. I think, in turn, that the best evidence for this system being deterministic would be perfect prediction at the appropriate level. Thus, I take this to endorse our criterion of perfect neuroprediction.
On the other hand, elsewhere O’Connor seems in effect to urge the opposite attitude, i.e. to endorse Testable: “It is an open empirical question… whether and when the basic capacity to choose, to which philosophers give most of their attention, is present and regularly exercised in a way necessary for true freedom of choice” (2009, 185, my italics).
This line of argument could also be used against the verification condition for free will suggested in Sect. 2. The thought is that even if an absence of fully determining physical causes of volitional actions were established, nevertheless those actions could still be ascribed to chance. If so, that would render free will unverifiable as well as unfalsifiable.
Suppose no one ever has any understanding of why they are doing what they are doing—as suggested, for instance, by Gazzaniga (2011). As an anonymous referee points out, this claim would, if accepted, falsify free will without perfect neuroprediction being necessary, and regardless of whether determinism is true. But for it to establish Testable, Gazzaniga’s claim would need (as per Sect. 5) to be accepted by all sides, and I do not think that prospect is likely—for instance, non-naturalistic defenders of free will certainly haven’t accepted his claim so far.
Neither this nor the previous reason implies that the universe itself is indeterministic. The indeterminism is purely with respect to our best models.
Indeed, arguably empirical testing is always a comparative matter, being between a proposition and an alternative. Because this alternative may not always be the proposition’s simple negation, strictly speaking testability is therefore a property of a problem, i.e. of a proposition plus contrast, rather than of a proposition singly (Sober 1999). In our case though, the competing propositions are indeed ‘we have free will’ and its negation, so we may elide this technicality without loss, as in the paper’s title. The points developed in the text, about testability’s relativization to auxiliary assumptions, are unaffected.
In Bayesian terms, each side’s degree of belief in whether free will exists will not be altered significantly.
The degree of belief in free will is not related linearly to the accuracy of prediction. Thus, 60% accuracy clearly does not imply 0.6 degree of belief in (lack of) free will, and neither need 99% accuracy imply 0.99 degree of belief.
Two nuances: First, if compatibilists are viewed as being committed to the causal efficacy of conscious decision-making, then perhaps empirical investigation could threaten even compatibilist free will—although that would merely confirm that it is testable. Second, I am ignoring science fiction scenarios, such as aliens secretly manipulating our every action so that what we ordinarily think are uncoerced decisions are in fact anything but. Even then, the untestability of compatibilist free will would further require that this alien conspiracy could never be discovered by us. But in any case, ultimately, as the text explains, the paper’s overall thesis stands or falls regardless of whether compatibilist free will is testable or not.
References
Baumeister, R. (2008). Free will, consciousness, and cultural animals. In J. Baer, J. Kaufman, & R. Baumeister (Eds.), Are we free? Psychology and free will. New York: Oxford University Press.
Chater, N., & Oaksford, M. (Eds.). (2008). The probabilistic mind: Prospects for bayesian cognitive science. Oxford: Oxford University Press.
Chisholm, R. (1976). Person and object. LaSalle: Open Court.
Clarke, R. (1993). Toward a credible agent-causal account of free will. Nous, 27, 191–203.
Clarke, R. (2003). Libertarian accounts of free will. Oxford: Oxford University Press.
Duhem, P. (1954). The aim and structure of physical theory, trans. from 2nd ed. by P. W. Wiener, Princeton, NJ: Princeton University Press. Originally published as La Théorie Physique: Son Objet et sa Structure (Paris: Marcel Riviera & Cie., 1914).
Fried, I., Mukamel, R., & Kreiman, G. (2011). Internally generated preactivation of single neurons in human medial frontal cortex predicts volition. Neuron, 69, 548–562.
Gazzaniga, M. (2011). Who’s in charge? Free will and the science of the brain. New York: Harper Collins.
Ginet, C. (1990). On action. Cambridge: Cambridge University Press.
Ginet, C. (1997). Freedom, responsibility, and agency. Journal of Ethics, 1, 85–98.
Kane, R. (1996). The significance of free will. New York: Oxford University Press.
Kane, R. (2008). Three freedoms, free will, and self-formation. In N. Trakakis & D. Cohen (Eds.), Essays on free will and moral responsibility (pp. 142–162). Newcastle: Cambridge Scholars Publishing.
Laudan, L. (1990). Demystifying underdetermination. In C. W. Savage (Ed.), Scientific theories Minnesota studies in the philosophy of science (Vol. 14, pp. 267–297). Minneapolis: University of Minnesota Press.
Libet, B. (1985). Unconscious cerebral initiative and the role of conscious will in voluntary action. Behavioral and Brain Sciences, 8, 529–566.
Libet, B. (2002). The timing of mental events: Libet’s experimental findings and their implications. Consciousness and Cognition, 11, 291–299.
Maoz, U., Mudrik, L., Rivlin, R., Ross, I., Mamelak, A., & Yaffe, G. (2014). ‘On reporting the onset of the intention to move’, chapter 10 in Mele. New York: Oxford University Press.
Mele, A. (2009). Effective intentions: The power of conscious will. New York: Oxford University Press.
Mele, A. (2013). Unconscious decisions and free will. Philosophical Psychology, 26, 777–789.
Mele, A. (Ed.). (2014). Surrounding free will: Philosophy, psychology, neuroscience. Oxford: Oxford University Press.
Mylopoulos, M., & Lau, H. (2014). Naturalizing free will, chapter 7 in Mele. pp 123–141.
Nahmias, E. (2014). ‘Is free will an illusion? Confronting challenges from the modern mind sciences. In W. Sinnott-Armstrong (Ed.), Moral psychology: Freedom and responsibility (Vol. 4). Cambridge, MA: MIT Press.
Nahmias, E. (2015). Why we have free will. Scientific American, 312, 77–79.
Newsome, W. (2014). Neuroscience, explanation, and the problem of free will. In W. Sinnott-Armstrong (Ed.), Moral psychology: Free will and moral responsibility (Vol. 4). Cambridge, MA: MIT Press.
Nozick, R. (1981). Philosophical explanations. Cambridge, MA: Belknap Press.
O’Connor, T. (1995). Agent Causation. In T. O’Connor (Ed.), Agents, causes, and events: Essays on indeterminism and free will (pp. 173–200). New York: Oxford University Press.
O’Connor, T. (2000). Persons and causes: The metaphysics of free will. New York: Oxford University Press.
O’Connor, T. (2009). Conscious willing and the emerging sciences of brain and behavior. In N. Murphy, G. Ellis, & T. O’Connor (Eds.), Downward causation and the neurobiology of free will (pp. 173–186). Berlin: Springer.
Pereboom, D. (2003). Living without free will. Cambridge: Cambridge University Press.
Pink, T. (2004). Free will: A very short introduction. Oxford: Oxford University Press.
Quine, W. V. (Ed.). (1951). Two dogmas of empiricism. In From a logical point of view. 2nd edn. (pp. 20–46). Cambridge, MA: Harvard University Press.
Quine, W. V. (Ed.). (1955). Posits and reality. In The ways of paradox and other essays. 2nd edn. (pp. 246–254). Cambridge, MA: Harvard University Press.
Roskies, A. (2006). Neuroscientific challenges to free will and responsibility. Trends in Cognitive Science, 10, 419–423.
Roskies, A. (2014). Monkey decision making as a model system for human decision making, chapter 12 in Mele, pp. 231–251.
Schurger, A., Mylopoulos, M., & Rosenthal, D. (2016). Neural antecedents of spontaneous voluntary movement: A new perspective. Trends in Cognitive Sciences, 20(2), 77–79.
Sober, E. (1999). Testability. Proceedings and Addresses of the American Philosophical Association, 73, 47–76.
Soon, C. S., Brass, M., Heinze, H. J., & Haynes, J.-D. (2008). Unconscious determinants of free decisions in the human brain. Nature Neuroscience, 11, 543–545.
Soon, C. S., He, A. H., Bode, S., & Haynes, J.-D. (2013). Predicting free choices for abstract intentions. Proceedings of the National Academy of Sciences, 110(15), 6217–6222.
Acknowledgements
I would like to thank an audience at Birkbeck philosophy department, Anna Alexandrova, and anonymous referees at this and another journal for useful feedback.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Northcott, R. Free Will is Not a Testable Hypothesis. Erkenn 84, 617–631 (2019). https://doi.org/10.1007/s10670-018-9974-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10670-018-9974-y