Harry Frankfurt (1929-2023): A Tribute (UNLOCKED)

Harry Frankfurt died last month. His work fundamentally changed the map of the philosophical debate on free will and moral responsibility.

Last Sunday, we finished a two-part deep dive into the philosophy of time. Harry Frankfurt died in between Pt. 1 and Pt. 2. If you don’t quite know who that is but the name sounds familiar, there’s an excellent chance that what you’re remembering is his 2005 book On Bullshit. A philosopher writing a bite-sized book, accessible but rigorous, applying the tools of his trade to analyzing how “bullshitting” was different from “lying,” was catnip to a certain kind of reader, and that thing was on the New York Times bestseller list for 27 weeks. I remember seeing Jon Stewart interviewing him about it on the Daily Show. But he was a celebrity among academic philosophers a very long time before his brief stint as a celebrity in the wider world, and in what follows I want to convey some sense of why.

One more note before I get started:

As far as I know, Professor Frankfurt’s one foray into political philosophy was On Inequality (2015). In it, he argues that the way we talk about inequality often confuses the real issue of concern. He thinks poverty is objectionable but inequality per se is not. The object of our condemnation should be the floor being too low.

Raising it may have reducing inequality as a side effect, and he understands that there might be something particularly offensive about the conjunction of great wealth with great poverty—but once you’ve dealt with the “great poverty” part, he doesn’t think there’s anything intrinsically important about how much space is left between the economic floor and the ceiling.

If the paragraph you’re reading right now isn’t the first one you’ve ever read from me on politics or political philosophy, you probably know I don’t agree with any of that. Another reason I’m not a fan of On Inequality is that some of it reads to me like a brilliant philosopher being, frankly, too much of a philosopher. There are relevant facts about the material world that I don’t think get nearly enough attention.

I promise that I’ll do a deep dive into my issues with On Inequality in some future installment of this Substack, but for now I just want to acknowledge that it’s there before ignoring it for the rest of the essay. I never met the man, but what I’d regard as Frankfurt’s best work had a real impact on me. Today, at least, I’d like to spend my time praising him.

---

The first time I heard Frankfurt’s name was the same year On Bullshit hit the shelves. I was getting my MA at Western Michigan University and one of the speakers at our “brown bag” lecture series brought up the argument for which Frankfurt’s best known among academics.

When I first heard it, my reaction was irritation. I don’t remember whether I sparred with the presenter at all in Q&A but I can remember feeling sure that Frankfurt couldn’t be right.

Within a year or two, I can remember talking about the same argument at a dive bar in South Miami called The Seven Seas. I was a PhD student at the University of Miami by then, and several of us would go to “the Seas” every Thursday night to do karaoke. That particular Thursday, somewhere between the sixth or seventh plastic cup of rum and coke of the evening and probably after watching our friend Brian do a very enthusiastic rendition of AC/DC’s “Dirty Deeds Done Dirt Cheap,” a couple of us started talking about Frankfurt. Somewhere over the course of all those months of letting his argument marinate in the back of my head, my position had shifted. I now thought Frankfurt was obviously right.

His target is what he calls the “Principle of Alternate Possibilities” although my impression is this principle was so deeply assumed that few previous philosophers would have even thought to give it a name. This is the principle that Person P can only be responsible for Action A if P could have done otherwise.

Note that there are nearby formulations that might sound synonymous but which turn out to be importantly different from the PAP. The claim that P could only be responsible for doing A if doing A was “up to them” or “under their control,” for example, is definitely not what Frankfurt has in his sights.

In fact, his point is precisely that these formulations are not synonymous. We’ll get to that. Let’s back up first and think about the broader issue.

---

We’ve been over free will and determinism a couple different times and in a couple different ways here, but here’s a slightly different way into the subject:

What are the conditions that have to hold for (at least some) human beings to (at least sometimes) be morally responsible for their actions—to be genuinely, objectively at fault if they do something wrong, or genuinely, objectively praiseworthy for doing something right? Intuitively, there seem to be at least two—a “knowledge condition” and a “control condition.” Imagine a cult practicing a human sacrifice ritual. There are at least two kinds of participants that we might not think are responsible, or at least not fully responsible, for their participation. One would be participants who were drugged or hypnotized or otherwise manipulated by the cult leader in such a way that they weren’t exercising the kind of control over their behavior we normally think cognitively standard adult humans exercise. Another would be participants who were brainwashed by the cult from infancy to believe that human sacrifice rituals were morally good, and who thus might plausibly not know any better.

A “hard determinist” disbelieves in moral responsibility not because she doubts that sometimes people do things they know are wrong, but because she thinks we live in a deterministic universe and as such she doesn’t believe anyone anywhere meets the control condition. The claim that we live in a deterministic universe, after all, is just the claim that everything that ever happens—including everything that ever happens in human brains—is inevitable given all the chains of physical cause-and-effect that led up to it. The hard determinist may still believe that it’s pragmatically useful to treat people as if they were responsible for their actions, but she doesn’t see how anyone in this kind of universe could be objectively responsible.

It’s easy to nod along to that—but let’s think harder about that as such.

What exactly is the argument?

One way of understanding it goes like this:

P1: If we live in a deterministic universe, nobody can do otherwise than what they actually do.

P2: If nobody can do otherwise than what they actually do, no one is in control of their actions in the right way to be morally responsible for them.

P3: We live in a deterministic universe.

C: No one is ever in control of their actions in the right way to be morally responsible for them.

The classical way for compatibilists—people who think that objective moral responsibility can exist in a deterministic universe—to resist this argument is by denying P1. They can’t deny that there’s a sense of “can do otherwise” in which P1 is true, of course, but they deny that it’s the same sense in which P2 is true. Of course, it’s impossible to do otherwise given every single fact about everything that’s ever happened, but “classical compatibilists” deny that this is the sense of “can do otherwise” that’s important for moral responsibility.

To many ears, this move is going to sound like silly wordplay. Personally, I think there’s a (non-crazy) case to be made for it, but let’s assume for the sake of argument that this is an argumentative dead end. Do we then need to reject compatibilism?

Not necessarily. Frankfurt pursues a different compatibilist strategy. He denies P2—which is, of course, what he identifies as the Principle of Alternate Possibility.

---

That key Frankfurt argument I found so implausible when I first heard it around a seminar table in Kalamazoo, Michigan in 2005—but that I’d managed to completely internalize and accept by the time I was talking about it in the Seven Seas a year or two later—was that we should reject P2 because there are cases in which we’d reasonably think some person P would be responsible for some action A even though there’s no interesting sense in which P could have done other than A.

In the original paper where he lays this out, he gives not so much a counter-example to the PAP as a template for formulating counter-examples:

Suppose someone—Black, let us say—wants Jones¹ to perform a certain action. Black is prepared to go to considerable lengths to get his way, but he prefers to avoid showing his hand unnecessarily. So he waits until Jones is about to make up his mind what to do, and he does nothing unless it is clear to him (Black is an excellent judge of such things) that Jones is going to decide to do something other than what he wants him to do. If it does become clear that Jones is going to decide to do something else, Black takes effective steps to ensure that Jones decides to do, and that he does do, what he wants him to do. Whatever Jones's initial preferences and inclinations, then, Black will have his way.

What steps will Black take, if he believes he must take steps, in order to ensure that Jones decides and acts as he wishes? Anyone with a theory concerning what "could have done otherwise" means may answer this question for himself by describing whatever measures he would regard as sufficient to guarantee that, in the relevant sense, Jones cannot do otherwise….

He goes on to specify that this is what would have happened if Black felt he had to take steps. But in the actual case, he knows he won’t have to, so he doesn’t. Jones does what Black wants him to do for no reason other than reasons intrinsic to Jones’s own psychology.

Frankfurt’s point is:

If Jones did what Black wanted because Jones had no alternate possibilities open to him—that is to say, Black threatened him, and Jones was cowed by the threat—we might say Jones wasn’t responsible, but if whatever would have happened if Black had felt the need to take steps, didn’t happen, if the possibility that it would have happened was unknown to Jones and played no role in Jones’s decision-making, it seems massively implausible to say that this un-actualized hypothetical somehow neutralizes Jone’s responsibility for his actual actions.

---

I know that this quick and abstract run-through the argument might not quite get the idea across. But I want to pause here to underline that Frankfurt’s argument fundamentally changed the map of one of the longest-running debates in the history of Western philosophy. There’s a whole sprawling literature on “Frankfurt cases,” which are basically versions of the original Black/Jones example tweaked in some way, or analyzed from a new angle, to either try to refute Frankfurt’s original argument against the PAP or to defend it against critiques. Does Black only act when he has some sign that Jones is going to try to do otherwise? If so, does this “flicker of freedom” count as “doing otherwise” in itself? How about if Black has a brain scanner, and he’s able to see if a neural pathway is being entered which Jones would have to go down before forming an intention to do otherwise, and thus is able to intervene before even a “flicker” of resistance? Etc. etc. etc. etc. etc.

There’s decades of this stuff. You can’t write about free will without it. Trying to canvass that literature would turn this 3,000-word Substack essay into a 300-page academic monograph, but trust me, it exists.

To try to make it all a bit more vivid, here’s the sort of version I often give to Intro to Philosophy classes:

Black is both a mad scientist and a dentist. Jones comes in to get his wisdom teeth removed. While Jones is under, Black—who does take out the wisdom teeth, he’s not a complete monster—also performs a little impromptu brain surgery. Black implants a chip in Jones’s head which, if activated, will take over Jones’s brain and make him do what Black wants. He also puts a loaded gun and a map with directions to a playground full of children and puppies on Jones’s nightstand. When Jones wakes up, one of two things happen:

(1) Black (who’s in the next room with a long-distance brain scanner) can tell if Jones is in the neural state he’d have to be in on the way to forming the intention not to do it, so he activates the chip, and the chip causes Jones to form the intention to commit the massacre Black wants him to commit.

(2) Jones never enters into this neural state that would cause Black to activate the chip. Instead, he’s overjoyed when he sees the gun and the map, because he’s always wanted to massacre a playground full of children and puppies. He proceeds to do so.

Assume (2) happens. The chip is never activated, and it plays zero role in the chain of cause-and-effect in Jones’s brain leading up to the massacre. Is the fact that he wouldn’t have been able to do otherwise relevant in any way to the extent to which Jones is at fault for committing the massacre? Is he somehow less responsible than he would be if not for the chip? It sure doesn’t seem like it.

Frankfurt’s position is sometimes called “semi-compatibilism” because if “free will” means “the kind of control necessary for moral responsibility,” Frankfurt is a compatibilist about determinism and “free will.” On the other hand, if it means “freedom to do otherwise,” Frankfurt’s argument doesn’t even attempt to establish that this sort of free will exists. The vast majority of authors on all sides of the debate prior to Frankfurt’s 1969 paper assumed that these were one and the same. Now, that assumption is—at the very least—no longer safe.

---

So: If “free will” in the way that matters for moral responsibility—“Jones massacred those children and puppies of his own free will”—isn’t freedom to do otherwise, what is it?

That’s a massive question. Frankfurt’s own answer is about acting on “higher-order” desires—meaning that you’re not just doing what you want to do but what you want to want to do. Later semi-compatibilists have refined this and added more conditions, focusing not just on the structure of the agent’s mental states at the time of making a decision but the causal history of the decision. John Martin Fisher, for example, talks about “reasons-responsiveness”—roughly, whether an agent has the kind of cognitive mechanism capable to a reasonable (though in most cases very imperfect) extent of understanding and being moved by reasons for and against a course of action.

Many people think “real” free will has to involve indeterminism but Fankfurt has many ways of searchingly pushing back. In another of his classic papers, Freedom of the Will and the Concept of a Person, he points out that one of the things we want from a concept of free will (in the control-condition-for-responsibility sense) is a relevant distinction between the cognitive mechanisms of standard-issue adult humans we think are capable of moral responsibility and those of, say, small children or non-human animals or adults with certain kinds of sufficiently profound cognitive defects that we wouldn’t consider them to be responsible for their actions. And “freedom from determinism” just seems irrelevant to that whether it’s present or not. Some anti-determinists have tried to hold onto a "robust” conception of free will by postulating that a miracle happens in the human brain every time we engage in free actions, such that what happens is somehow independent of all chains of cause-and-effect leading to it. Fair enough, Frankfurt says, but if that’s happening in human brains, why not squirrel brains? And if it is happening in squirrel brains, would we really say squirrels are therefore morally responsible?

---

Your mileage, dear reader, may vary, but to me this is fairly exciting stuff for a few reasons. Part of it is that it intersects with my political interests. I’m keenly aware that many of my comrades don’t see the relationship between the concepts the same way I do, but my own view is that free-will denialism is deeply reactionary. Seeing people as fleshy robots incapable of meaningfully autonomous action fits very poorly with the idea that it’s normatively important that we get to individually and collectively determine our own destiny—to my mind, the beating heart of any left project worthy of the name—and all too well with the kind of technocratic paternalism that sees human beings as walking talking containers of good and bad consequences who simply need to be managed for their own good. If you see things that way, Frankfurt—notwithstanding his being the author of On Inequality—did a valuable service to left politics, helping to reconcile free will and human dignity with a realistic and robustly materialist understanding of our universe.

Putting politics aside, though, these kinds of basic questions about freedom, morality, and the human person are the kinds of things that made me interested in philosophy in the first place. Harry Frankfurt’s work helped me think much harder about it all than I might have otherwise, and I’m grateful. Rest in peace.

1. In the paper, this is actually the fourth example involving a “Jones” character so it’s “Jones₄“ rather than “Jones” throughout the original quote but it seemed like it would be annoying and confusing to reproduce that.