Skip to main content

On the very concept of free will


Determinism seems to rule out a robust sense of options but also prevent our choices from being a matter of luck. In this way, free will seems to require both the truth and falsity of determinism. If the concept of free will is coherent, something must have gone wrong. I offer a diagnosis on which this puzzle is due at least in part to a tension already present in the very idea of free will. I provide various lines of support for this hypothesis, including some experimental data gathered by probing the judgments of non-specialists.

This is a preview of subscription content, access via your institution.


  1. This proposal is inspired by discussions with John Maier and by the treatment of weakness of will in May and Holton (2012).

  2. Throughout I will speak of “our” concept of free will and the intuitions “we” have, but this is primarily for convenience and should of course be taken with a grain of salt. It is unclear how far we can generalize from studies using participants largely from Western cultures (although, for a relevant cross-cultural study, see Sarkissian et al. 2010).

  3. For a discussion of options and their relation to freedom, see Maier (forthcoming). The labels “liberty” and “ensurance” are his invention, although he doesn’t employ them in that paper. At any rate, these should be thought of as technical terms, even though they may be used elsewhere in the free will literature with different meanings. In particular, “liberty” has a long history and some have used it to simply designate what we now call “free will” (e.g. Hume and Reid).

  4. Much of this work is done by Eleanor Rosch (e.g. Rosch 1975). See Laurence and Margolis (1999) for a philosophical overview of this experimental work as well how various theories of concepts relate to it.

  5. There is further evidence for the general idea that multiple factors drive ordinary judgments about freedom and responsibility (e.g. Feltz et al. 2009; Knobe and Doris 2010; Weigel 2011). But these factors may not nicely align with ensurance or liberty. The cluster hypothesis is not necessarily in conflict with these proposals, however.

  6. Of course, perhaps a driving force behind the two intuitions is a difference between abstract and concrete scenarios (Sinnott-Armstrong 2008). There is some evidence for this, as it appears to be what remains from Nichols and Knobe, and the basic result has been replicated, including cross-culturally (Sarkissian et al. 2010). This account is in principle compatible with the cluster hypothesis, however. In fact, it may well be that more abstract cases make liberty more salient while concrete cases do so for ensurance. Alternatively, abstractnesss could be a further factor. I remain neutral on the merits of the abstract-concrete theory, although further research could determine whether it is incompatible with a cluster account.

  7. In a pilot study, I attempted to manipulate ensurance by describing the protagonist as having an “uncontrollable urge” but this didn’t fully work. Liberty had a statistically significant effect, but not ensurance, although it did influence responses in the predicted direction. Hence, in the experiment reported here, I attempted to more drastically undermine the protagonist’s ensurance along the same dramatic lines in which she lacks liberty. I’ve opted not to provide the details of this pilot only in the interest of space.

  8. One might worry that when an agent is manipulated this undermines freedom for the same reasons that a deterministic universe does (as proponents of famous manipulation arguments allege). In that case, I wouldn’t have operationalized ensurance in a way that is distinct from the operationalization of liberty (tied as it is to determinism being true). But there is some evidence that, for ordinary folks, free will seems undermined by manipulation in a way that is different from determinism (see Feltz 2013).

  9. For all studies in this paper, participants were users of Amazon’s Mechanical Turk website in the US and were paid for participating. Mturk is a popular and reliable place for soliciting human subjects for Footnote 9 continued scientific research, and it provides an even more diverse sample than simply using university students (see Buhrmester et al. 2011). In each of the studies, the mean age of participants was 30–32 and between 50 and 60 % were male. I took standard measures to encourage and monitor serious participation. The entire project was approved by Monash University’s Human Research Ethics Committee (project number CF12/1686-2012000918).

  10. I also tested whether responses would be different using a slightly different measure: “Jill stole the necklace of her own free will” (we might think of this as Experiment 1b; \(N= 105\)). The results were basically the same as before.

  11. I subjected the data to a 2 (E vs. \(\sim \)E) by 2 (L vs. \(\sim \)L) between-subjects analysis of variance (ANOVA). There was a main effect of Ensurance, F(1, 222) = 58.03, \(p < 0.001\), partial \(\eta ^{2}= 0.207\), and a main effect of Liberty, F(1, 222) = 25.6, \(p < 0.001\), partial \(\eta ^{2}\) = 0.103. There was a significant interaction effect, F(1, 222) = 4.3, \(p = 0.040\), partial \(\eta ^{2}\) = 0.019, such that the differences between responses when Liberty was present or absent was greatest when Ensurance was present.

  12. In ANOVAs where Control and Choice were now treated as dependent variables, there was a predicted main effect of Ensurance on intuitions about control [F(1, 222) = 64.3, \(p < 0.001\), partial \(\eta ^{2}\) = 0.224] as well as on intuitions about choice [F(1, 222) = 39.8, \(p < 0.001\), partial \(\eta ^{2}\) = 0.152].

  13. I collected data from 241 participants, but responses from 13 were not included in the analysis, as they failed the comprehension check (same as in Experiment 1).

  14. I conducted a 2 (E vs. \(\sim \)E) by 2 (L vs. \(\sim \)L) between-subjects ANOVA. There was a main effect of Ensurance, F(1, 224) = 80.9, \(p < 0.001\), partial \(\eta ^{2}\) = 0.265, and a main effect of Liberty, F(1, 224) = 20.36, \(p < 0.001\), partial \(\eta ^{2}\) = 0.083. There was no interaction effect, F(1, 224) = 1.412, \(p = 0.236\).

  15. However, Joshua Shepherd (2012, p. 923) raises the interesting worry that some participants might claim a mental state “had no effect” on an agent’s action only in the sense in which, say, a team’s offense can have no effect on the opposition’s defense. This is compatible with such factors playing a causal role.

  16. Rose and Nichols (2013) have recently provided further empirical evidence that people are reluctant to grant the ability to decide in a deterministic universe. They argue that Murray and Nahmias have incorrectly claimed that bypassing judgments lead to incompatibilist intuitions when the causal order is actually the reverse.


  • Balaguer, M. (2010). Free will as an open scientific problem. Cambridge: MIT Press.

    Google Scholar 

  • Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s mechanical Turk: A new source of inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6(1), 3–5.

    Article  Google Scholar 

  • Chalmers, D. J. (2011). Verbal disputes. Philosophical Review, 120(4), 515–566.

    Article  Google Scholar 

  • Feltz, A. (2013). Pereboom and premises: Asking the right questions in the experimental philosophy of free will. Consciousness and Cognition, 22(1), 53–63.

    Article  Google Scholar 

  • Feltz, A., Cokely, E. T., & Nadelhoffer, T. (2009). Natural compatibilism versus natural incompatibilism: Back to the drawing board. Mind and Language, 24(1), 1–23.

    Article  Google Scholar 

  • Feltz, A., Perez, A., & Harris, M. (2012). Free will, causes, and decisions: Individual differences in written reports. Journal of Consciousness Studies, 19(9–10), 166–189.

    Google Scholar 

  • Fischer, J. M. (1994). The metaphysics of free will: An essay on control. Malden: Blackwell.

    Google Scholar 

  • Hawthorne, J. (2001). Freedom in context. Philosophical Studies, 104(1), 63–79.

    Article  Google Scholar 

  • Knobe, J. (2014). Free will and the scientific vision. In E. Machery & E. O’Neill (Eds.), Current controversies in experimental philosophy. Oxford: Routledge.

    Google Scholar 

  • Knobe, J., & Doris, J. (2010). Responsibility. In J. Doris (Ed.), The moral psychology handbook (pp. 321–354). Oxford: Oxford University Press.

    Chapter  Google Scholar 

  • Laurence, S., & Margolis, E. (1999). Concepts and cognitive science. In S. Laurence & E. Margolis (Eds.), Concepts: Core readings. Cambridge: MIT Press. doi:10.1111/phpr.12038.

  • Maier, J. (forthcoming). The agentive modalities. Philosophy and Phenomenological Research.

  • May, J., & Holton, R. (2012). What in the world is weakness of will? Philosophical Studies, 157(3), 341–360.

    Article  Google Scholar 

  • Mele, A. (2006). Free will and luck. Oxford: Oxford University Press.

    Book  Google Scholar 

  • Mele, A. (2014). Free will and substance dualism: The real scientific threat to free will? In W. Sinnott-Armstrong (Ed.), Moral psychology, Vol. 4: Free will and responsibility. Cambridge: MIT Press.

    Google Scholar 

  • Murray, D., & Nahmias, E. (Forthcoming). Explaining away incompatibilist intuitions. Philosophy and Phenomenological Research. doi:10.1111/j.1933-1592.2012.00609.x.

  • Nahmias, E., Morris, S. G., Nadelhoffer, T., & Turner, J. (2006). Is incompatibilism intuitive? Philosophy and Phenomenological Research, 73(1), 28–53.

    Article  Google Scholar 

  • Nichols, S. (2006). Folk intuitions on free will. Journal of Cognition and Culture, 6, 57–86.

    Article  Google Scholar 

  • Nichols, S., & Knobe, J. (2007). Moral responsibility and determinism: The cognitive science of folk intuitions. Noûs, 41(4), 663–685.

    Article  Google Scholar 

  • Park, J. J. (2013). Prototypes, exemplars, and theoretical & applied ethics. Neuroethics, 6(2), 237–247.

    Article  Google Scholar 

  • Rosch, E. (1975). Cognitive representations of semantic categories. Journal of Experimental Psychology: General, 104, 192–233.

    Article  Google Scholar 

  • Rose, D., & Nichols, S. (2013). The lesson of bypassing. Review of Philosophy and Psychology, 4(4), 599–619.

    Article  Google Scholar 

  • Sarkissian, H., Chatterjee, A., De Brigard, F., Knobe, J., Nichols, S., & Sirker, S. (2010). Is belief in free will a cultural universal? Mind and Language, 25(3), 346–358.

    Article  Google Scholar 

  • Shepherd, J. (2012). Free will and consciousness: Experimental studies. Consciousness and Cognition, 21(2), 915–927.

    Article  Google Scholar 

  • Sinnott-Armstrong, W. (2008). Abstract + concrete = paradox. In J. Knobe & S. Nichols (Eds.), Experimental Philosophy. Oxford: Oxford University Press.

    Google Scholar 

  • Sripada, C. S. (2012). What makes a manipulated agent unfree? Philosophy and Phenomenological Research, 85(3), 563–593.

    Article  Google Scholar 

  • Strawson, P. F. (1962). Freedom and resentment. Proceedings of the British Academy, 48, 1–15.

    Article  Google Scholar 

  • van Inwagen, P. (2000). Free will remains a mystery. Philosophical Perspectives, 14, 1–20.

    Google Scholar 

  • van Inwagen, P. (2008). How to think about the problem of free will. Journal of Ethics, 12(3/4), 327–341.

    Article  Google Scholar 

  • Weigel, C. (2011). Distance, anger, freedom: An account of the role of abstraction in compatibilist and incompatibilist intuitions. Philosophical Psychology, 24(6), 803–823.

    Article  Google Scholar 

  • Woolfolk, R. L., Doris, J. M., & Darley, J. M. (2006/2008). Identification, situational constraint, and social cognition: Studies in the attribution of moral responsibility. In J. Knobe & S. Nichols (Eds.), Experimental philosophy (pp. 61–80). Oxford: Oxford University Press.

Download references


I owe a significant debt to John Maier who inspired me to test the main hypothesis in this paper, the basics of which we both independently developed. For valuable comments on or discussions about this paper, I thank: Lloyd Humberstone, Josh Knobe, Colin Marshall, Alfred Mele, Jonathan Phillips, Patrick Todd, Jason Turner, and the referees for this journal, one of whom kindly identified himself as Adam Feltz. Versions of the paper were presented at the University at Buffalo, Melbourne University, Deakin University, and CSU Wagga Wagga. Many thanks to the attendees for their feedback, especially Wylie Breckenridge, Daniel Cohen, Neil Levy, Edouard Machery, David Ripley, and Laura Schroeter. Work on this paper was supported by Monash University and some of the ideas were developed while participating in a summer seminar for the Big Questions in Free Will project at Florida State University, which was supported by the John Templeton Foundation. The views expressed in this article are solely the author’s and do not reflect those of either funding body

Author information

Authors and Affiliations


Corresponding author

Correspondence to Joshua May.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

May, J. On the very concept of free will. Synthese 191, 2849–2866 (2014).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


  • Freedom
  • Moral responsibility
  • Experimental philosophy
  • Incompatibilism
  • Compatibilism
  • Cluster concept
  • Prototype