Skip to main content Accessibility help
Hostname: page-component-684899dbb8-8hm5d Total loading time: 0.246 Render date: 2022-05-25T09:32:42.574Z Has data issue: true Feature Flags: { "shouldUseShareProductTool": true, "shouldUseHypothesis": true, "isUnsiloEnabled": true, "useRatesEcommerce": false, "useNewApi": true }


Published online by Cambridge University Press:  03 September 2019

Kristin M. Mickelson*
Philosophy, University of Gothenburg


This essay begins by dividing the traditional problem of free will and determinism into a “correlation” problem and an “explanation” problem. I then focus on the explanation problem, and argue that a standard form of abductive reasoning (that is, inference to the best explanation) may be useful in solving it. To demonstrate the fruitfulness of the abductive approach, I apply it to three standard accounts of free will. While each account implies the same solution to the correlation problem, each implies a unique solution to the explanationproblem. For example, all libertarian-friendly accounts of free will imply that it is impossible to act freely when determinism is true. However, only a narrow subset of libertarians have the theoretical resources to defend the incompatibilist claim that deterministic laws (qua deterministic) undermine free will, while other libertarians must reject this traditional incompatibilist view.

Research Article
Copyright © Social Philosophy and Policy Foundation 2019 

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)



Many thanks to Social Philosophy and Policy, the Lund-Gothenburg Responsibility Project (LGRP), the Swedish Research Council, and the Department of Philosophy, Linguistics, and Theory of Science (FLoV) at the University of Gothenburg for their support. I am also indebted to the organizers and participants of the 2017 “Moral Responsibility: The Next Generation” conference—especially for extensive feedback from Michael McKenna, Carolina Sartorio, Robert Kane, and Keith Lehrer. I am also deeply grateful to my LGRP and FLoV colleagues for their insights and encouragement, with special thanks to Gunnar Björnsson, Christian Munthe, Robert Hartman, Jon Erikson, Ragnar Francén, Per Milam, Sofia Jeppsson, Ben Matheson, Anna-Sofia Maurin, and Alex Skiles.


1 Basic desert is associated with a backward-looking, retributive notion of moral responsibility. For clarification, see Pereboom, Derk, Free Will, Agency, and Meaning in Life (Oxford: Oxford University Press, 2014), chap. 6.CrossRefGoogle Scholar

2 Notable exceptions include works by Strawson, Galen (e.g., “Free Will” in Craig, E., ed., Routledge Encyclopedia of Philosophy [London: Routledge, 2011]:,Google Scholar Campbell, Joseph (e.g., “Free Will and the Necessity of the Past,” Analysis 67, no. 2 [2007]: 105–11), andCrossRefGoogle Scholar Levy, Neil (cf. Hard Luck: How Luck Undermines Free Will and Moral Responsibility [Oxford: Oxford University Press, 2011].CrossRefGoogle Scholar

3 Abductive (inference to the best explanation) reasoning is not a standard tool within the free will debate. One candidate exception to this is Derk Pereboom’s Four-case Argument, which Pereboom classifies as a best-explanation argument (Pereboom, Free Will, Agency, and Meaning in Life, 74–80). However, Pereboom develops the Four-Case Argument primarily as a slippery-slope “no difference” generalization argument which involves ruling out only those proposed explanations which would (if true) block the generalization to the conclusion that free will cannot exist when determinism is true. As discussed below, the abductive approach described in this essay provides reason to think that Pereboom’s proposed explanation is not the “best” given his own background commitments regarding free will.

4 This characterization of incompatibilism (i.e., as a partial solution to the explanation problem and not as a solution to the correlation problem) can be found, for example, in McKenna, Michael and Pereboom, Derk, Free Will: A Contemporary Introduction (New York: Routledge, 2016), 151;CrossRefGoogle Scholar Vihvelin, Kadri, Causes, Laws, and Free Will: Why Determinism Doesn’t Matter (Oxford: Oxford University Press, 2013), 35;CrossRefGoogle Scholar Sartorio, Carolina, Causation and Free Will (Oxford: Oxford University Press, 2016), 147; see alsoCrossRefGoogle Scholar Levy, Neil, Hard Luck: How Luck Undermines Free Will and Moral Responsibility, 12 n. 1.Google Scholar

5 William James, “The Dilemma of Determinism,” originally an address to Harvard Divinity Students (in Lowell Lecture Hall, Harvard University), published in The Unitarian Review (September 1884); reprinted in James, William, The Will to Believe and Other Essays in Popular Philosophy (New York: Longmans Green and Co., 1907), 145–83.CrossRefGoogle Scholar

6 As Susanne Bobzien explains, the Stoics thought of fate as “the organizing principle of the world . . . the organization is such that—in some way—whatever occurs had been organized, hence fixed, to do so before it occurred. And this organization or fixing in advance is also eternal: what occurs was always organized and fixed to occur” ( Bobzien, , Determinism and Freedom in Stoic Philosophy (Oxford: Clarendon Press, 1998), 4889.Google Scholar

7 I speak of “laws” out of tradition, and do not assume that laws have the ontological status of propositions. Rather, I use “laws” to pick out (roughly speaking) the factors which account for the diachronic evolution of the universe in virtue of which determinism is true.

8 The terms “strong” and “weak” are borrowed from John Perry, though he applies the weak/strong distinction to accounts of the laws rather than the laws themselves ( Perry, , “Compatibilist Options,” in Campbell, J., O’Rourke, M., and Shier, D., eds., Freedom and Determinism (Cambridge, MA: MIT Press, 2004), 231–54.Google Scholar

9 See, for example, my “Mastering the Manipulation Argument” (manuscript).

10 Further refinement is needed to identify the precise nature of the actual-sequence leeway permitted by probabilistic laws/causation, though it is standard to favor the view described by Armstrong, David (A World of States of Affairs [Cambridge: Cambridge University Press, 1997], 238) rather than, say,CrossRefGoogle Scholar Hausman, Daniel (Causal Asymmetries, [Cambridge: Cambridge University Press, 1998], 205).CrossRefGoogle Scholar

11 Proponents of weak laws may argue that weak laws may be classified as either deterministic or probabilistic as well, assuming their preferred definitions of the terms “deterministic” and “probabilistic.” However, weak laws cannot be deterministic or probabilistic as the terms are defined here, for weak laws (as I characterize them) are not, ontologically speaking, the kind of things which could make it the case that a future event has an antecedent objective probability of 1, 0, or anything in between.

12 Weak laws are impotent.

13 This apt phrase is borrowed from Smilansky, Saul (Smilanksy, “Compatibilism: The Argument From Shallowness,” Philosophical Studies 115, no. 3 [2003], 257–82).CrossRefGoogle Scholar

14 This makes Jamesian determinism a rather radical proposal, so it is unsurprising that many philosophers prefer to use “determinism” and “deterministic laws” to refer to less potent, arguably more “scientifically respectable” laws (Sehon, Scott, “A Flawed Conception of Determinism in the Consequence Argument,” Analysis 71 [2011]: 3038).CrossRefGoogle Scholar If my arguments in this essay are successful, it suggests (pace Sehon) that there is an important dialectical role to be played by both conditional and unconditional deterministic strong laws (even if the latter ultimately turn out to be metaphysically impossible).

15 The distinction between (un)conditional strong laws is superficially similar to the distinction between so-called “exceptionless” or “strict” laws on the one hand and “ceteris paribus” laws on the other (Pereboom, Living Without Free Will, 85). However, since weak laws may be exceptionless despite being impotent, a new distinction is needed to track the specific potency distinction between strong laws that is drawn here. I explain my preference for the (un)conditional divide in “Leeway and the Laws of Nature” (manuscript).

16 This way of taxonomizing the laws also allows us to set aside the standard determinism/indeterminism divide and distracting verbal disputes over the proper definition of “determinism”).

17 Notably, a world with weak laws may lack actual-sequence leeway on other grounds (e.g., the nature of time, the nature of God, or the nature of logic).

18 Again, “laws” is used here out of tradition, but the focus is on the fundamental features of the natural world that account for its evolution over time. As such, it might be preferable to speak of the overall evolution of the world in terms of strength, probability, and potency (e.g., Jamesian determinism describes strong, deterministic, unconditional global evolution).

19 A similar correlation question may be formulated for each type of laws we have identified, but the dialectic of interest in this essay is the one that springs directly from answering “no” to the question stated here.

20 Because incompossibilism is often assumed to entail the incompatibilist view that strong (unconditional/conditional) deterministic laws undermine (or otherwise account for the lack of) free will when such laws obtain, the term “incompatibilism” is often used (and often equivocally) as the name for incompossibilism as well as incompatibilism. In the light of the arguments given in this essay, this ambiguous use of “incompatibilism” seems as problematic as using the term “causation” to refer to both causation and mere correlation. The taxonomy of free-will views suggested in this essay is motivated and developed in greater detail in my “(In)compatibilism,” in Campbell, Joseph, ed., A Companion to Free Will (Wiley-Blackwell, forthcoming).Google Scholar

21 Campbell, Joseph, “Free Will and the Necessity of the Past,” Analysis 67 (2007): 105–11.CrossRefGoogle Scholar

22 Sartorio, Carolina, “The Problem of Determinism and Free Will is Not the Problem of Determinism and Free Will,” in Mele, Alfred, ed., Surrounding Free Will: Philosophy, Psychology, Neuroscience (New York: Oxford University Press, 2015).Google Scholar

23 Matters become complicated when we consider what is required to satisfy the law-overriding condition when weak laws obtain, but such complications must be glossed over here; I address them explicitly in “Hard Times for Hard Incompatibilism” (manuscript).

24 Pereboom, Free Will, Agency, and Meaning in Life, 5–6.

25 Strawson, “Free Will.”

26 Pereboom, Living Without Free Will, 129.

27 Pereboom, Living Without Free Will, 85–86 and 128. To simplify the discussion, I make the following assumption: If all normal humans in the actual world lack the intrinsic properties required to override laws, then it is part of human nature to lack such properties (that is, a possible being with intrinsic properties required to override strong laws would technically not be a normal human).

28 Ibid., 85.

29 One upshot of this: Pereboom need not appeal to the Four-case Argument to explain why humans would be unfree when determinism is true. Moreover, combining the view that we are subject to the laws by nature with Pereboom’s commitment to possibilism, it seems that the Four-case Argument delivers the wrong diagnosis of what keeps a normal human from exercising free will when determinism is true: normal humans fail the law-overriding condition on free will when determinism is true, but not (even in part) because of the causal relations that hold when determinism is true.

30 On this approach, Pereboom seems committed to the view that it is impossible for someone who is subject to the laws of nature to exercise free will—which is a way of expressing the modal commitments of his set of views without explicitly mentioning the law-overriding condition.

31 Depending on one’s views about conditional strong laws, it may be that some conditional strong laws permit the relevant sort of law-overriding while some do not. In that case, we would have to change to conditional strong laws of the specific sort that permit law-overriding. I will gloss over the issue here, as it would merely involve more steps of the same basic strategy to address the matter.

32 Kadri Vihvelin’s distinction between “narrow” and “wide” abilities is useful here (Vihvelin, Causes, Laws, and Free Will: Why Determinism Doesn’t Matter, 11): one has the narrow ability to peform an action A at a particular time and place when one has all of the requisite intrinsic properties for doing A, including the “necessary skills and the psychological and physical capacity to use those skills”; one has the wide abiltiy to do A so long as one has the narrow ability to A and “nothing external stands in her way.” If humans lack the intrinsic properties required to override strong laws of nature, then we cannot act freely because we lack the basic narrow ability to override the laws. By contrast, a superhuman (for example, Super-Norm) who has the requisite properties for law-overriding would have the narrow ability to override overridable laws no matter what the laws of nature are like, but unconditional strong laws would deprive her of the wide ability to override the laws (that is, laws of this potency would stand in the way of her exercising her narrow law-overriding abilities). This is another way of highlighting that normal humans and superhumans may, in precisely the same environment, lack free will on different grounds.

33 Notably, these results are not mere artefacts of assuming that possibly strong laws are unconditional. Even if it turns out that unconditional strong laws are metaphysically impossible because no laws could be that potent, the law-overriding conception of free will does not imply that we (or anyone else) lack free will because the laws are deterministic. If, necessarily, strong laws are conditional and it is also the case that humans have the requisite properties to be law-overriders, then all strong laws would permit the leeway required for us to override them and, so, would pose no special threat to human free will. On the other hand, if humans do not have the properties required to override overridable strong laws, then this conception of sourcehood implies that our modest human properties keep us from exercising free will, irrespective of whether the strong laws are deterministic or otherwise. Again, this line of reasoning leads us to the conclusion that human free will is incompossible with unconditional deterministic strong laws, but also to theanti-incompatibilist conclusion that their incompossibility is not even partly due to the fact that the laws are deterministic.

34 Pereboom, Living Without Free Will, 89–90.

35 Pereboom supports the conclusion that deterministic strong laws (more precisely deterministic causation by factors beyond one’s control with his Four-case Argument (Pereboom, Living Without Free Will, 112–17). The Four-case Argument supports its explanatory conclusion with an inference to the best explanation, but it now seems that Pereboom’s proposed explanation is mistaken given his background commitments regarding the requirements on free agency. I provide a detailed argument for the claim that the explanatory conclusion of the Four-case Argument is in tension with Pereboom’s response to Strawson’s Basic Argument—and, more generally, that hard incompatibilism is an untenable view—in my “Hard Times for Hard Incompatibilism” (manuscript).

36 Kane, Robert, The Significance of Free Will (Oxford: Oxford University Press, 1996).Google Scholar

37 It seems that a godlike being could exploit the low potency of conditional deterministic laws to insert probabilistic relations in “just the right place” (whatever that means, given the details of the event-casual account) for a person to exercise free will, cf. Jim Stone, “Free will as a Gift from God: A New Compatibilism,” Philosophical Studies 92, no. 3 (1998): 257–81. Kane, in conversation (2017), is open to such an interpretation of his theory, which has the interesting consequence that his event-causal libertarianism would support the traditional incompatibilist view that the specific laws (causal relations or the like) which make determinism true pose a unique threat to free will (qua being the toxic combination of both deterministic and unconditional). This is a potentially interesting result, given that most philosophers would argue that unconditional strong laws are, because of their potency, metaphysically impossible. Assuming unconditional strong laws are impossible, a proponent of Kane’s view might hold that strong laws of nature pose no insurmountable obstacle to free will whether they are deterministic or probabilistic. For further discussion of this position, see my “Mastering the Manipulation Argument” (manuscript).

38 The claim here is that, on the assumption of incompossibilism, if there is no adequate libertarian-friendly account of free will, then we have no motivation for believing that actual-sequence leeway is relevant to free will. However, a compossibilist may hold that the lack of actual-sequence leeway that obtains when determinism is true is required for free will. So, assuming that compossibilism is true, there is room to argue that actual-sequence leeway is freedom-relevant (in the sense that certain types of actual-sequence leeway is freedom-undermining). The same points hold, mutatis mutandis, for anthropocentric versions of incompossibilism and compossibilism.

39 To be clear, the claim here is that the availability of actual-sequence leeway is of no help to someone who satisfies all of the necessary and jointly sufficient conditions on free will except for Strawson’s especially demanding starting-point condition. Figuratively speaking, I am treating Strawson’s source condition as a “bottleneck” condition on free agency, a condition which cannot be (or is at least rarely) satisfied by possible agents who satisfy all other necessary conditions on free will.

40 Strawson is an anti-incompatibilist impossibilist (insofar as he rejects the incompatibilist’s partial solution to the explanation problem), but it follows a fortiori from the fact that Strawson is an impossibilist that he is also an anthropocentric impossibilist and an anthropocentric incompossibilist.

41 Strawson speaks of “constitutive luck,” by which he means luck in the way that one is constituted (at least in crucial mental respects) at the time of action; Neil Levy prefers the term “endowment luck” (cf. Levy, Hard Luck: How Luck Undermines Free Will and Moral Responsibility). Notably, the notion of constitutive/endowment luck is tightly tethered to the problem of moral luck (cf. Nagel, ThomasMoral Luck,” in Mortal Questions [Cambridge: Cambridge University Press, 1979]: 2438),Google Scholar but constitutive luck may exist even if constitutive moral luck does not. I illuminate the dialectical connections between the problem of free will and determinism, the problem of constitutive luck, and the problem of moral luck in “Free Will, Self-Creation, and the Paradox of Moral Luck,” Midwest Studies in Philosophy 43 (forthcoming).

42 Worries related to constitutive/endowment luck might be better (or more fully) fleshed out, say, in terms of the metaphysics of personal identity.

Cited by

Save article to Kindle

To save this article to your Kindle, first ensure is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about saving to your Kindle.

Note you can select to save to either the or variations. ‘’ emails are free but can only be saved to your device when it is connected to wi-fi. ‘’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

Find out more about the Kindle Personal Document Service.

Available formats

Save article to Dropbox

To save this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you used this feature, you will be asked to authorise Cambridge Core to connect with your Dropbox account. Find out more about saving content to Dropbox.

Available formats

Save article to Google Drive

To save this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you used this feature, you will be asked to authorise Cambridge Core to connect with your Google Drive account. Find out more about saving content to Google Drive.

Available formats

Reply to: Submit a response

Please enter your response.

Your details

Please enter a valid email address.

Conflicting interests

Do you have any conflicting interests? *