Causation and Free Will: Compatibilist vs. Incompatibilist Intuitions

Estimated read time (minus contemplative pauses): 17 min.


Today, I finished reading Carolina Sartorio’s 2016 book Causation and Free Will. The last section of the book’s final chapter, which I’ll briefly respond to in a moment, provides a clear instance of a phenomenon I worriedly rambled on about in another recent post on free will (I’ll link that post below).

In a nutshell, and admittedly as an outsider, my worry is that in recent decades the free will debate has developed into a swelling catalog of intricately tweaked and cross-referenced thought experiments in the service of a contest between basic intuitions. The result, or so I worry, is an impasse between philosophers who could otherwise be making much-needed contributions to a world that doesn’t wait around for the resolution (or dissolution) of arcane academic disputes.

On the other hand, I think it’s important to recognize where the (potential) lines of impasse exist. Sartorio presents a neat opportunity for thinking about this with a pair of examples that just might pinpoint precisely where the relevant intuitions diverge in the free will debate (despite those involved possessing roughly the same cultural and educational backgrounds).

Before getting to the examples, a few words on Sartorio’s book, beginning with some of her own words, taken from the book’s final pages (note that, below, I’ll briefly explain the most important concepts mentioned here):

I have laid out a compatibilist view of freedom according to which freedom is just a function of actual causes, and I have defended it from the main source-incompatibilist challenges. The view is inspired by certain insights about Frankfurt cases, and it is the result of taking seriously two simple and natural ideas. First, freedom is just a function of actual sequences; nothing else matters. Second, actual sequences are just what they appear to be: causal histories. Taken together, these two ideas result in the view that freedom is a function of actual causes, and nothing else (p 171; all citations here are from the Kindle Edition).

Sartorio grounds this “actual-sequence” view on certain conceptions of causation and reasons-sensitivity that, among other things, allows for absence causation (e.g., part of what caused you to turn left was an absence of a road block), and for sensitivity to both present and absent reasons. The upshot is that (morally salient) choices and actions issuing from “causal histories that reflect the agent’s sensitivity to reasons” (p 172) result in the sort of freedom that makes that agent morally responsible for those choices and actions.

This is an oversimplification of Sartorio’s carefully laid out view, in which she addresses alternative conceptions she herself might (reluctantly) endorse, counterarguments, subtle variations in the literature (aspects of which she may or may not find appealing), and in which she at times expresses worries about her view and acknowledges the need for refinement or further development. That said, while I’m not convinced that the sort of freedom Sartorio lays out makes agents morally responsible, I found the book thought-provoking. In particular, her exploration of absence of reasons is fascinating; it’s an area I’ve given some thought to, but clearly there’s still lots more thinking to do.

To be clear, even if I don’t believe Sartorio’s findings ground moral responsibility, I believe they ground legal responsibility in the sense, for example, of assessing whether a person poses a serious danger to society. I view the moral-legal distinction as critical. Moral responsibility suggests that wrongdoers deserve to be punished (i.e., made to suffer in proportion to the harm they have caused—or might cause—in others, etc.), while being legally responsible implies that, while removal from society, and indeed punishment, of the wrongdoer may be unavoidable, such things should not be done in a spirit of retribution. This crucial distinction is what motivates the conversation for me.

Now, before getting to the meat of this post, a quick overview of compatibilism, incompatibilism, and Frankfurt cases.

Compatibilism comes in flavors. You’ll see terms like “classical compatibilism” (which means different things to different people) and semi-compatibilism (which is roughly where I see Frankfurt-influenced views landing), but I’ll use the term “compatibilism” broadly as: the assertion that determinism is compatible with the sort of freedom required for holding agents morally responsible for their choices and actions. Some might parse this as “determinism and free will are compatible” or “determinism and moral responsibility are compatible, regardless of what we think about free will.” But my broad conception is good enough for this post.

Determinism” also needs defining, but I’ll just let the general idea emerge from the discussion going forward. (My aim here is just enough of a gloss in order to discuss Sartorio’s examples.)

Incompatibilism is essentially the opposite of compatibilism: the assertion that the sort freedom required for moral responsibility is incompatible with determinism, or, for that matter, with indeterminism such that factors other than the agent herself—e.g., randomness or nature-nurture—are the ultimate cause of her actions. Incompatibilism is often hashed out as: an agent is free when she has has access to alternative possibilities; or: an agent is free only if she could have chosen to do otherwise than she actually chose to do.  This is sometimes called “leeway incompatibilism.”

There is also a “sourcehood” approach, which seems to be especially common among those who take Frankfurt cases seriously (see below). To quote Derk Pereboom: “…an agent’s basic desert-involving moral responsibility for an action would be explained primarily by the action’s having a causal history in which she is the source of her action in a specific way.”1 This is the sort of incomptabilism Sartorio is primarily concerned with addressing.

I’m a free will skeptic with sympathies for the leeway and sourcehood views (among others). I also tend to stress the psychological dimensions of free will (thus my use of “chosen to do” here, which could be substituted with “intended” or “decided” etc.), but these details aren’t terribly important for this post.

Frankfurt cases (as they’ve come to be known) were introduced by Harry Frankfurt in his still-highly-influential, 1969 essay, “Alternate Possibilities and Moral Responsibility.”2 This 2016 book contains that paper along with several responses: Moral Responsibility and Alternative Possibilities: Essays on the Importance of Alternative Possibilities. As far as I can tell, this paper marks the turning point that’s led the free will debate to the aforementioned impasse.

Frankfurt cases are meant to challenge the principle of alternative possibilities by showing that it’s possible for an agent to lack the ability to do otherwise than she does, yet still be morally responsible for what she does. Suppose Alexandra has decided to shoot Tony. An evil neuroscientist who’d also like to see Tony shot has a mechanism in place so that, at the slightest sign that Alexandra will change her mind, the mechanism will ensure that Alexandra continue on with the decision to shoot Tony. As it turns out, Alexandra gives no such sign—she shoots Tony for her own reasons, and in fact does everything exactly as she would have done had the neuroscientist not been waiting in the wings. So the mind-control mechanism is not triggered. The claim is that, because Alexandra shot Tony entirely on her on, just as she would have had the neuroscientist not been watching from afar, she is morally responsible for shooting Tony (just as she would be were the neuroscientist not watching from afar)—even though she could not have done otherwise, as she had no alternative but to shoot Tony.

In my post, “Frankfurt Cases & Moral v. Legal Responsibility,” I elaborate on why I, and others, don’t find such cases convincing. In short, I claim that Alexandra could have triggered the neuroscientist’s intervention. (I take it the vast majority of people agree that, had the nueroscientist used mind control on Alexandra, she could not be rightly held morally responsible for shooting Tony.)

Furthermore, the triggering signal issued by Alexandra would have to occur outside of her awareness (that is, maintaining the notion that she could not have done other than she did—in particular, that she could not have consciously done something to trigger the exculpating mind-control mechanism—requires, as far as I can tell, that Alexandra not only be denied the option of deciding not to shoot Tony, but that she be denied the option of having any serious doubts about shooting him). This means the triggering signal must be something that doesn’t involve conscious thought; in other words, it must be something outside of Alexandra’s control (maybe it’s a certain pattern of neuronal firings that precedes serious doubt).

Given this, I claim that Alexandra’s decision to shoot Tony is outside of her control even when the neuroscientist is not present (just as it would be outside of her control were the neuroscientist to have implemented his mind-control mechanism).

This brings us to Sartorio’s examples, which feature in a discussion that starts on page 156 in a section called “Manipulation Arguments.” Manipulation arguments, simply put, claim that a (sufficiently) manipulated agent lacks the sort of freedom required for moral responsibility. Sartorio recognizes that such arguments pose a serious challenge to her view. To address the challenge, she introduces the following variations on Alfred Mele’s well-known “zygote” example (as presented in this 2006 book, which I haven’t read: Free Will and Luck):

Diana: Mary is about to use in vitro fertilization to conceive a child. A goddess, Diana, intervenes by inducing a certain genetic mutation in the material in the dish, which she knows will result in the conception of a child, Ernie, with certain innate genetic dispositions. Diana knows that, partly due to those innate dispositions, thirty years later Ernie will perform a terrible act: he’ll murder his uncle to inherit a fortune. Diana induces the relevant genetic mutation because she wants Ernie to murder his uncle, has full knowledge of the state of the world and the deterministic laws, and can predict exactly how the world will evolve after she intervenes in that way. Still, when Ernie murders his uncle thirty years later, he satisfies all of the standard compatibilist conditions on freedom with respect to that act (his act is not the result of compulsion or coercion, he is responsive to reasons, he identifies with the desires that move him to act, he takes responsibility for the relevant mechanisms, etc.) (pp 158–159).

From this, Sartorio extracts a simple argument (p 159):

Diana Argument (Simple Version):
(1) Ernie’s murdering act is not free.
(2) Ernie meets all of the standard compatibilist conditions when he commits the murdering act.
(3) Therefore, all standard forms of compatibilism fail.

My central concern here is with the intuitive plausibility of (1). It seems not only plausible to me, but obviously true. Sartorio also finds it plausible, but with a caveat:

Incompatibilists would note that (1) is intuitively very plausible, and I think they’re right about this. I, at least, do find it intuitively plausible, and my guess is that many others would too. However, I will suggest that there are reasons to be skeptical of the reliability of our intuitions in scenarios of this kind (p 160).

To bear out her skeptical point, she shares another example:

Lightning Strike: This time there is no goddess Diana. Instead, a flash of lightning strikes the laboratory while the in vitro procedure is taking place, and this results in the same genetic mutation in the material in the dish as the one induced by Diana in the Diana scenario. Everything else is the same as in the Diana scenario (pp 160–161).

According to Sartorio, her intuitions “change drastically” in this case. That is, in contrast with the Diana scenario, here she’s “not at all tempted to regard Ernie’s act as not free” (p 161). Furthermore, she goes on to note that, while she recognizes this is strictly an intuitive response, her guess is that “many others would have a similar reaction” (p 161). Indeed, she tends to use words like “we” when discussing this intuition; for example:

If I am right and our intuitions about such cases [as the Lightening Strike case] differ radically from our intuitions about standard (intelligent-design) manipulation cases, then this can be used to cast serious doubt on the manipulation argument (p 166).

Why do we tend to feel that Ernie’s freedom and responsibility are undermined when the source of the genetic manipulation is intelligent design, but not when it’s blind forces? This is an interesting question, and it would be nice to have an answer to it (167).

This strikes me as utterly bizarre, even when I try to hold onto the compatibilist intuition. What difference could it possibly make if Ernie’s act is predetermined by the calculations (or miscalculations; see below) of a conscious agent or by some naturally occurring fluke? I do believe Sartorio is arguing in good faith, but her shift in intuition strikes me as so strange that I have to do a little work to ignore a little voice in my head insisting that she must be putting us on in order to stay true to her overriding claim that people are (at least sometimes) morally responsible for their actions.

My aim here isn’t to convince anyone that my intuition is superior, but I will make the following small effort (you’ll need to read Sartorio’s book to get the full thrust of her own persuasive efforts). Imagine you commit a heinous act tomorrow. You then learn that some lightening struck and so on, as described above, essentially programming you to perform—without fail—precisely that act. Would you really feel morally responsible for the act? I’m not asking if you’d feel guilty. I imagine that, were I to have an unexpected seizure while driving, thus resulting in someone’s death, I’d experience horrible guilt. But legally and morally, most (contemporary Westerners) would agree that, given my seizure-free history and lack of reasons to expect the seizure, I am not responsible for the seizure’s consequences.

I could elaborate on the parallels between the lightening and seizure cases while trying to reconcile their differences (e.g., one isn’t responding to reasons during a seizure, which, furthermore, directly precedes the consequence in question rather than happening several years before). But I’ll leave my attempts at persuasion here, as, again, my goal today isn’t to support my intuition, but rather to marvel at the fact that my intuition needs any defending.

An intuition Sartorio and share is that our moral intuitions should be the same in both the Diana and Lighten Strike cases (even if hers isn’t). In fact, this is an important feature of her argument. That is, she determines that the common intuition in the Diana case must be misguided, and thus her Lightening Strike intuition must be the correct one. The upshot, then, is that Ernie must be morally blameworthy in both cases. She fleshes this out more than I’ll get into here, as the argument strikes me as non-starter if the goal is to convince someone like me, whose intuition does not tend to diverge from Diana to Lightening Strike in the first place (again, it seems dazzlingly obvious to me that Ernie is not morally responsible in either case). Still, I will share one aspect of her argument, as it again emphasizes the dissonance I’m encountering not only here, but in many writings on free will.

One Sartorio attempts to motivate skepticism about her initial Diana intuition with an example from John Fischer, who has argued that, while the Diana case does stoke the intuition that Ernie is not free, and this may nudge us towards incompatibilism in less fantastical cases, the opposite occurs if we’re first given an ordinary case and then given a fantastical one.3 Sartorio relates Fischer’s example as follows:

Thus, imagine, first, that John and Mary conceive a child, Ernie, in a deterministic world in the normal way. Imagine that, as an adult, Ernie performs a certain act A, while meeting all the standard compatibilist conditions for free agency. We have the intuition that Ernie’s A-ing is free. Now change the story slightly: imagine that John and Mary conceived their child when they did hoping that the child would grow to be the kind of person who performs act A, believing that he would, and intending to create someone who would do such a thing. We clearly think that John and Mary’s mental states and intentions are irrelevant to Ernie’s freedom. But it’s easy to see, then, that the same would have to be true of a scenario where Diana, not John and Mary, created Ernie. As a result, we’re more likely to have a compatibilist-friendly intuition about the Diana scenario (p 163).

First, as an incompatibilist, I don’t believe that Ernie’s A-ing is free (in the morally relevant sense) in a deterministic world, even under ordinary conditions. This already makes the example seem like a non-starter for anyone but, I suppose, a worried or conflicted compatibilist. But let’s assume for a moment that I can assume the compatibilist intuition (indeed, it’s closer to my current sensibilities than is the libertarian intuition I grew up with).

Second, I agree that, when John and Mary hope really hard to conceive a child who grows up to A, the compatibilist intuition about Ernie’s A-ing survives. Recall that the assumption here as that they are hoping in the moment of conception. The scenario here is not that John and Mary, for example, somehow train Ernie from an early age to A. This scenario is about Ernie as zygote—pre-Ernie, in a sense. And so, I agree that “John and Mary’s mental states and intentions are irrelevant to Ernie’s freedom.”

Third, I agree that Diana’s intentions are irrelevant to Ernie’s freedom. This is true even though Diana’s goddess superpowers guarantee that whatever she intends will come to pass. One might be tempted to think that Diana’s infallible superpowers are what seem, even if erroneously, to get Ernie off the hook. Maybe I’d feel this way were I an authentic compatibilist rather than a pretend one. Maybe this is where we strike my incompatibilist core (you know, as we dig down into my intuitions). Let’s say If Diana carefully rearranged the zygote’s particles so that Ernie would grow up to B, but she made a mistake and made him to grow up to A. I’d still see Ernie’s A-ing as as unfree. What matters to me is not Diana’s intentions, but that Ernie’s particles were arranged in such a way that guarantees he’ll A. How the particles got that way was not—simply could not have been—up to Ernie.4

Now, Diana might be implicated in a way that one should not (in my opinion) implicate the parents, who have no such causal power (unless, say, they hired or otherwise prayed to Diana for her intervention).5 Nor, of course, does it make sense to say that lightening can be held responsible—morally or legally—for whatever lightening does or causes to happen. Perhaps the compatibilist will more or less agree with these sentiments. The divergence, for sure, is that I do not—cannot—view Ernie’s A-ing as free in any of these cases.

Indeed, I struggle to get across just how alien Fischer and Sartorio’s intuition is to me. (And, again, I ask that you imagine yourself in these situations in your contemplation of them.)

Finally, Sartorio speculates—that is, divulges her “best guess” (p 167)—about why intuitions may differ about the Diana and Lightening Strike cases. My thought at the start of all this was that maybe Sartorio would blame her (self-described) intuitional misstep on the fact that Diana is a conscious agent to whom blame seems to transfer (while the only conscious agent in the Lightening scenario is Ernie). I figured, however, that this might be too simple an explanation, and thus the sort of fallacious thought one could somewhat readily shake off; still, it is more or less the chief idea Sartorio entertains:

It seems to me that something like a psychological “dilution of responsibility” effect may be at play in these cases. … This psychological effect, assuming it exists, might explain why we tend to think that Ernie is not responsible in the Diana case but he is in the Lightning Strike case. For, whereas there is no one else to blame for Ernie’s act in the Lightning Strike case, there is in the Diana case: Diana herself (pp 167–168).

Sartorio also considers the possibility of a similar phenomenon involving “dilution of control,” but then reminds us that, while figuring out why our intuitions are misguided in the Diana case would be nice, it’s not strictly necessary in order to take down the manipulation view; all that’s required is recognizing that our intuitions are misguided in the Diana case, for whatever reason.

To conclude, I’ll emphasize that my aim here isn’t to say Sartorio has made poor arguments. (Quite the contrary; immersing myself in Sartorio’s book has been time well spent!) My point, rather, is to highlight just how divergent our intuitions are, and to note how much hinges on this fact. If we can’t agree on cases as clear cut as the above, how can we ever hope to come to an agreement about the arguments built from those intuitions? And this is just within the relatively narrow field of academic philosophy. How might this play out across cultures and over time?

This leads me to admit that, as with many other philosophical quandaries (e.g., what are the necessary and sufficient conditions for knowledge?), I find myself confused about whether the aim of the discussion is to uncover intuitions so that we may construct a world that’s intuitively satisfying, or, rather, to uncover facts about the world so that our systems of social interactions reflect facts, rather than mere intuitions, provided there are indeed facts of the matter to be uncovered (surely majority intuition can misrepresent the world; and I assume that wrongdoers either do or do not really deserve to be punished).

Or maybe intuitions are just the best guides we have, and we can only hope that, sufficiently interrogated, they’ll steer us as close as we can get to the truth. In which case, the impasse—humming along in the background while the world goes about its business—serves as a frustrating reminder that we’ve still got a lot of work to do.


Here’s one more intuition I find curious, brought up multiple times in the book:

For imagine that we had some causal access to all of the deterministic causes of our acts. Imagine, for example, that we could travel back in time, to any time in the past, and have some causal influence on those events in the past. In that case determinism wouldn’t pose a problem to our freedom, at least not in the same way. For we could then, at least in principle, have control over those causes, and, if we could have control over them, then we could, at least in principle, have control over our acts (the inevitable result of those causes) (p 2).

Were I to travel back in time under deterministic conditions, that action—even the mere decision to pursue that action—would have been determined in the same way anything else I do is determined. And whatever I do once in the “past” would likewise be determined. And so on. I fail to see how traveling into the past could make any difference to my freedom.

Enjoy or find this post useful? Please consider pitching in a dollar or three to help me do a better job of populating this website with worthwhile words and music. Let me know what you'd like to see more of while you're at it. Transaction handled by PayPal.

Further Reading


  1. Pereboom, Derk. Free Will, Agency, and Meaning in Life (p 4). OUP Oxford. Kindle Edition.
  2. Frankfurt, Harry (1969) “Alternate Possibilities and Moral Responsibility.” Journal of Philosophy. 66. 23, pp. 829–39.
  3. Fischer, John Martin (2011) “The Zygote Argument Remixed.” Analysis. 71. 2, pp. 267–72.
  4. Echoes of Galen Strawson’s “impossibilist” view, in which one would have to be able to self-create in order to be morally responsible for an act. A view I’m sympathetic to. See: Strawson, Galen (1994) “The Impossibility of Moral Responsibility.” Philosophical Studies. 75, pp. 1–2, 5–24.
  5. This isn’t to suggest that I think Diana is morally blameworthy. On my view, even goddesses and gods lack such freedom.

Share your thoughts: