“Libertarian Free Will”: How to Avoid Some Confusions

I am not a libertarian about free will.[1] I think there are good objections to libertarianism. But many criticisms of libertarianism rest on confusions, and it is worth calling attention to some of these confusions in order that those of us who reject libertarianism can be sure to avoid criticizing a straw version of the position. Some confusions about libertarianism are so misguided as to be not worth engaging,[2] but the four I discuss below are easy to make and have been made by authors whose work I value and with whom I share many commitments.


Libertarianism about free will is the conjunction of two claims: first, that free will is incompatible with determinism; second, that we have free will.[3]

But what is “free will,” and what is “determinism”? Let’s take them in reverse order. Determinism is the thesis that there is, at any instant, exactly one physically possible future (van Inwagen 1983: 3). Put slightly differently, determinism is a thesis about the laws of nature. For the laws to be deterministic is for it to be the case that, with the state of the world at any instant as a sort of “input,” there is only one “output” concerning the next instant. Applied repeatedly, the result is that there is only one way the world can unfold from that point on.[4]

Now to “free will.” While this term has been deployed in many different ways (some, including van Inwagen, treating it as a mere term of art), it is common now for this term to be defined functionally, such that free will is whatever control is required for an agent (a person who acts) to be morally responsible for an action, omission, or consequence. And the sense of moral responsibility here concerns when praise and blame are deserved, such that the person is morally responsible for (say) a bit of wrongdoing would deserve blame just in virtue of having knowingly acted wrongly.

Upon hearing about free will and determinism, many people are inclined to think of these as contraries, sometimes even describing the problem of free will as a problem of “free will versus determinism.” To think of the problem in that way, however, is to presuppose that free will and determinism are incompatible with each other, and it turns out that the main debate among philosophers working on free will concerns the compatibility question. It would be confused, therefore, to think of the debate as one about whether we have free will or whether determinism is true, since many parties to the debate (myself included) will consider this a false dichotomy. So that is one confusion to avoid, though not one about libertarianism in particular.

Before turning to confusions about libertarianism in particular, it will be helpful to contrast libertarianism with the other main positions in the free will debate. We can divide the positions in the debate into three categories. Compatibilists, as their name suggests, take free will to be compatible with determinism, and compatibilists (with perhaps a few exceptions) affirm that we have free will. Incompatibilists deny compatibilism and then either affirm free will or deny it. Those incompatibilists who affirm free will, as we already saw, are libertarians. Those who deny free will are free will skeptics.

Thus, libertarians agree with compatibilists about our having free will but disagree about the compatibility issue; and libertarians agree with free will skeptics about the compatibility issue but disagree about whether we have free will. Let’s turn now to some confusions.

Confusion #1: Libertarianism and PAP

It is not uncommon for people to talk about “libertarian free will” in contrast to “compatibilist free will.” If these are shorthand expressions for the more unwieldy “control that is required to satisfy the conditions of libertarian accounts of free will” and “control that is required to satisfy the conditions of compatibilist accounts of free will,” then I have no problem with these terms (in fact, I use them myself). But sometimes “libertarian free will” (often abbreviated “LFW”) is used to refer to a particular sort of power or ability to do otherwise than what one actually does.

In fact, it may surprise some readers that, in defining libertarianism, I made no reference to the ability (or freedom) to do otherwise than what one does, nor any reference to the following “Principle of Alternative Possibilities”:

PAP: A person is morally responsible for an action only if they could have done otherwise than that action.[5]

Note that PAP is a principle—one describing the connection between moral responsibility and the ability to do otherwise (or alternative possibilities). One confusion to avoid, then, is describing an agent as having or not having PAP, or PAP as in/compatible with determinism (cf. Bignon 2018: 73).

If one accepts PAP, then one thinks that free will requires a certain sort of leeway. Free will, on this leeway conception, is like a garden of forking paths: at various points on one’s walk through the garden there are two (or more) available options, and one must select which of the available options to pursue.[6] And, given the definition of determinism, one might be tempted to think of this leeway conception of free will as off-limits to the compatibilists and required by libertarianism.

It is possible, however, for one to be a compatibilist and yet to accept PAP and to conceive of free will as requiring leeway.[7] Moreover, it is possible to be a libertarian and to reject PAP and leeway conception of freedom. An alternative to the leeway conception of freedom is a sourcehood conception, according to which what matters for free will is that agents be the proper source of their behavior, which on many views does not require leeway. Most philosophers who endorse sourcehood conceptions of free will are compatibilists (and this is the view I prefer), but some incompatibilists endorse this conception as well.[8]

What this means is that we cannot treat having “libertarian free will” (LFW) as equivalent to “the ability to choose the contrary” (cf. Peterson and Williams 2004: 139; Christensen 2016: 6-7) or to having the freedom to do otherwise (cf. Crisp 2014: 84-85).[9] Some (indeed, most) libertarians are leeway libertarians, but criticisms of that position are not ipso facto criticisms of libertarianism.

Confusion #2: Libertarianism and Causation

According to most forms of libertarianism, directly free actions must not be deterministically caused. Given this requirement, some critics allege that libertarianism requires that directly free actions be uncaused.[10] For it is tempting to think that, for one event to be the cause of another, it must be the case that, given the cause, the effect must occur (i.e., that the cause deterministically bring about the effect).[11]

While there was a time when many philosophers assumed that causation must be deterministic, this is now widely regarded to have been a mistaken assumption. As long as it is possible for there to be probabilistic causation, there is no reason to think that causation cannot be indeterministic. And so one could be a libertarian, maintaining that directly free actions not be determined, but require that directly free actions nevertheless be (indeterministically) caused.

To be fair to critics who have supposed that libertarianism requires uncaused events, there have been some proponents of libertarianism who have espoused non-causal views. But there are also event-causal and agent-causal libertarian views—in fact, the vast majority of libertarians fall into one of these two categories—according to which directly free actions must be (indeterministically) caused.

Confusion #3: Libertarianism and Explanation

The most widely discussed challenge for libertarianism is what may be called the problem of luck. The worry comes in many forms, but the basic idea is that if an agent’s choice is undetermined, as libertarian accounts of free will require, then there is some sense in which it is a matter of luck whether the agent chooses in that way or otherwise. If more than one choice was possible, holding fixed everything that came before, right up to the moment of choice, then the choice itself may start to look like a coinflip, or some other chance occurrence, rather than an exercise of free will. Suppose I am deliberating about whether to tell a lie, out of self-interest, or whether to the truth instead, and suppose that I will make the choice at a certain time t but that it is undetermined which choice I will make then. If I am to satisfy libertarian conditions on free will with respect to my choice at t, then it must be the case that, holding everything fixed right up until t, including all of my reasons for lying and reasons for telling the truth, among other things (e.g., my preferences, desires, etc.), either choice is possible. But if either is possible, then which choice obtains at t appears to be a matter of luck—something outside of my control.

One might be tempted to think, given this challenge, that undetermined actions (including those performed by agents satisfying libertarian conditions on free will) are “ultimately inexplicable” (cf. Wingard 2017).[12] But plausible libertarian views will say that, though free choices are undetermined, they are nevertheless explicable in terms of the agent’s beliefs, desires, intentions, etc.[13] For example, on an event-causal picture, a free choice is indeterministically caused, and thus explained, by the agent’s having reasons to make the choice. True, the agent’s choice is not necessitated (or determined) by their psychological profile at that time (or by that together with the environment and the laws of nature), but it would be mistaken to equate necessitation with “bringing about” or with explanation.

Confusion #4: Libertarianism and the Problem of Luck

But, setting aside the confusion about the explicability of undetermined actions, does the problem of luck nevertheless pose a threat to libertarianism? Some compatibilists (including Wingard) who are eager to press this problem for libertarianism run into the final confusion I want to discuss.[14]

The problem is this: we don’t know whether the world is deterministic or indeterministic.[15] If what is required for freedom and responsibility is that our choices be causally determined, we might escape the worry about those actions being a matter of luck, but our view of ourselves as free and responsible would be vulnerable to the discoveries of the sciences, for we could discover that those very events, our choices, are not deterministically caused. Prior to the 1960s is was very popular for compatibilists to affirm determinism and so to be fine with raising the problem of luck for libertarianism, but with the advent of quantum mechanics and the popularity of indeterministic interpretations of the physics, compatibilists have tended to avoid that precarious position and endorse a kind of “super-compatibilism” instead, taking free will to be compatible not only with determinism but also with indeterminism (even the very kind of indeterminacy that libertarians think is necessary for freedom and responsibility).

Suppose that we come to learn that indeterministic interpretations of quantum mechanics are mistaken and that determinism obtains. Would this allow the compatibilist to return to pressing the problem of luck?[16] If it became widely accepted that determinism obtains, that itself would be a problem for libertarianism, given its commitment to free will and the requirement of indeterminism. Even in this scenario, though, I don’t think compatibilists should raise the luck problem in particular, since they should be willing to grant that even if the world were indeterministic (in just the ways that libertarianism requires) we could still have free will.

If compatibilists want to object to libertarianism, then, they need to look elsewhere than to the problem of luck.[17]

[1] Thanks to Matt Flummer, Dan Miller, and Jordan Steffaniak for comments on an earlier version of this post.

[2] One example is Craig Carter’s recent (5/31/21) blog post on the Credo magazine website titled “The Concept of Libertarian Free Will: Why Christians Must Reject It

[3] One mistake to avoid is equating libertarianism with incompatibilism (cf. Christensen 2016: 10). There are incompatibilists who think that we lack free will and so are not libertarians, so we must keep in mind that libertarianism is a conjunction of incompatibilism plus the thesis that we have free will.

[4] Put this way, determinism is a thesis about the natural world, and that will be my focus here, though it is worth pointing out that there are related theses, including a theological variety sometimes called “divine determinism” according to which, given God’s activity (typically construed as God’s decrees), there is only one possible future.

[5] See Frankfurt (1969) for discussion and putative counterexamples.

[6] It is worth noting that one could be a leeway libertarian and yet deny that we always have the sort of leeway described here. For one thing, it is open to leeway libertarians to maintain that we have free will only some of the time. For another, it is common to distinguish between “direct” and “indirect” freedom, where an indirectly free action is one that depends entirely for its freedom on some earlier free action(s), and leeway libertarians can maintain that only directly free actions require the sort of leeway described here. In fact, this is the view of many prominent libertarians (such as Robert Kane and Peter van Inwagen). Thanks to Dan Miller for suggesting this point.

[7] Classical compatibilists like Thomas Hobbes and David Hume defended this position, and arguably this was Jonathan Edwards’s position (cf. Johnson 2016: 22).

[8] Derk Pereboom considers himself a source incompatibilist, despite his being a free will skeptic. Source libertarians include David Hunt, Eleonore Stump, and Kevin Timpe.

[9] It is worth noting that Crisp is not criticizing libertarianism—indeed, he suggests that one can be a “libertarian Calvinist”—and so the confusion seems innocuous in this case. For discussion and (to my mind, compelling) criticism, see Anderson and Manata (2017) and Preciado (2019: 72ff.).

[10] “On this view [libertarianism], our free choices are absolutely undetermined and uncaused” (Frame 2001: 24, emphasis added).

[11] Relatedly, while libertarianism requires that free actions not be sufficiently caused, it does not require that we be “unmoved movers of our own actions” (Christensen 2016: 27). Some libertarians have conceived of freedom in this way (cf. Chisholm 1964), but this is a minority position among libertarians.

[12] “The problem is that a causally undetermined event, such as an act of choice that is free in the sense required by the incompatibilist, would be ultimately inexplicable. There is, we might say, a certain ‘chanciness’ about such an act. It is in some sense the ‘product’ of chance or happenstance, an event that ‘just happened,’ an act that was ‘just done’” (Wingard 2017). See also Christensen (2016: 30-32).

[13] And so libertarianism itself does not require that “we have the ability to choose contrary to any prior factors that influence our choices, including external circumstances, our motives desires, character, and nature…” (Christensen 2016: 6). Rather (as Christensen goes on to clarify, correctly), libertarians can say that “internal and external influences can serve as reasons for the choices we make,” though “no particular reason or set of reasons is sufficient to determine our choices” (Christensen 2016: 17).

[14] See also Christensen (2016: 30): “Compatibilists argue that every choice has reasons that are absolutely determinative.”

[15] The majority of physicists accept indeterministic interpretations of quantum mechanics, and even if God determines everything, this does not necessarily imply causal determinism, and certainly not that all of our choices are determined by psychological states (as Wingard seems to think is necessary for freedom).

[16] Thanks to Jordan Steffaniak for the question.

[17] For a detailed discussion of this dialectic, and a suggestion about how compatibilists can build on the problem of luck in objecting to libertarianism, see Cyr (Forthcoming).


Anderson, James, and Manata, Paul (2017). “Determined to Come Most Freely: Some Challenges for Libertarian Calvinism,” Journal of Reformed Theology 11: 272-297.

Bignon, Guillaume (2018). Excusing Sinners and Blaming God: A Calvinist Assessment of Determinism, Moral Responsibility, and Divine Involvement in Evil (Eugene, OR: Pickwick Publications).

Chisholm, Roderick (1964). “Human Freedom and the Self,” The University of Kansas Lindley Lecture: 3-15, reprinted in G. Watson (ed.), Free Will (Oxford: Oxford University Press, 2003): 26-37.

Christensen, Scott (2016). What about Free Will? Reconciling Our Choices with God’s Sovereignty (Phillipsburg, NJ: P&R Publishing).

Crisp, Oliver (2014). Deviant Calvinism: Broadening Reformed Theology (Minneapolis, MN: Fortress Press).

Cyr, Taylor (Forthcoming). “Taking Hobart Seriously,” Philosophia.

Frame, John (2001). No Other God: A Response to Open Theism (Phillipsburg, NJ: P&R Publishing).

Frankfurt, Harry (1969). “Alternate Possibilities and Moral Responsibility,” Journal of Philosophy 66: 829–839.

Johnson, Daniel (2016). “Calvinism and the Problem of Evil: A Map of the Territory,” in D. Alexander and D. Johnson (eds.), Calvinism and the Problem of Evil (Eugene, OR: Pickwick Publications).

Peterson, Robert, and Williams, Michael (2004). Why I Am Not An Arminian (Downers Grove, IL: InterVarsity Press).

Preciado, Michael Patrick (2019). A Reformed View of Freedom: The Compatibility of Guidance Control and Reformed Theology (Eugene, OR: Pickwick Publications).

Wingard, John (2017). “Confession of a Reformed Philosopher: Why I Am a Compatibilist about Determinism and Moral Responsibility,” Themelios 42 (2).


Connect With TLL

Podcast Channels

You might also enjoy