Philosopher Dan Dennett has recently taken Sam Harris to task over his arguments for rejecting the concept of free will. Here are my thoughts on this intellectual fallout.
Sam Harris is perhaps the most prominent defender of the claim that we lack free will, a conclusion that he thinks is inescapable if we’re intellectually honest with ourselves and accept the best scientific picture of the world we have.
It’s not a new claim, and Harris doesn’t advocate any new arguments for his position on free will, usually known as hard determinism. (It’s called hard determinism because it’s based on the acceptance of determinism — the thesis that every event that happens has a preceding cause — combined with a denial of free will, which is seen to be incompatible with determinism.)
Harris’s hard determinism is a minority position within contemporary philosophy. An alternative, yet also minority, position is libertarianism. Like the hard determinist, the libertarian believes that determinism and free will are incompatible. However, unlike the hard determinist the libertarian accepts the reality of free will, and therefore rejects determinism.
The dominant view within academic philosophy rejects the idea that determinism and free will are incompatible, and so goes by the name of compatibilism. Compatibilists accept that the universe is deterministic, but argue that this does not pose a threat to free will — or moral responsibility or praise or blame. (A final position, not much defended, is that the universe is not deterministic and yet we still lack free will.)
Dan Dennett is one of contemporary philosophy’s most sophisticated defenders of compatibilism. He’s been arguing for this position for more than 40 years, along with many other philosophers. He’s also a friend of Harris’s, and both were members of the Four Horseman of the New Atheism that emerged in the mid-2000s (along with Richard Dawkins and the late Christopher Hitchens). Nonetheless, Dennett has recently delivered a blistering review of Harris’s thinking on free will, as spelled out in The Moral Landscape and at marginally greater length in his short book Free Will.
Dennett describes Harris book as housing a “veritable museum of mistakes, none of them new and all of them seductive” — a diagnosis with which I wholeheartedly agree. My intention here is not simply to repeat or summarise all of the many specific charges that Dennett brings against Harris — you can read Dennett’s admirably clear essay here — though I will reiterate a number of key points. What I want to do is to use Dennett’s critique as a lens through which to look at Harris’s general style of argumentation as expressed in his books and essays. In some ways, this is a prelude to a longer, more extensive and comprehensive critique of Harris’s ideas on morality, human behaviour, religion, and free will that I’m currently writing. (Though I think he’s misleading on all these topics, he certainly picks the big issues to engage with!)
Here I’ll stick with Harris’s treatment of free will. The bottom line of Dennett’s searing critique is simple: Harris has not read, thought about or understood the relevant literature on free will, and therefore fails to engage with compatibilism in any substantial way. Harris sets up the supposed conflict between free will and determinism vividly, but there’s nothing original about doing so: it’s the starting point for most discussions of free will, familiar to anyone who has thought about free will seriously.
However, in contemporary philosophy (dominated by compatibilists) this apparent conflict is usually taken as a jumping off point for thinking about what we mean by freedom, choice, and moral responsibility — and then asking whether and how these can be compatible with determinism.
Rather than engaging with these compatibilist arguments, Harris simply sweeps them aside. As Dennett says, “[Harris’s] opinion of [compatibilism] is breathtakingly dismissive: After acknowledging that it is the prevailing view among philosophers (including his friend Daniel Dennett), he asserts that ‘More than in any other area of academic philosophy, the result resembles theology’. This is a low blow, and worse follows: ‘From both a moral and a scientific perspective, this seems deliberately obtuse’.”
Dennett goes on:
I would hope that Harris would pause at this point to wonder — just wonder — whether maybe his philosophical colleagues had seen some points that had somehow escaped him in his canvassing of compatibilism. As I tell my undergraduate students, whenever they encounter in their required reading a claim or argument that seems just plain stupid, they should probably double check to make sure they are not misreading the “preposterous” passage in question. It is possible that they have uncovered a howling error that has somehow gone unnoticed by the profession for generations, but not very likely.
As Dennett concludes, “Harris and others need to do their homework if they want to engage with the best thought on the topic.” In the absence of this required reading and thought, Harris, in my opinion, is simply not worth listening to about free will.
Yet even if you do think he’s worth reading on this topic, you might find that it’s not clear exactly what he’s saying. As Dennett notes in his essay, Harris seems to be confused about the position he’s defending. “[T]he points [Harris] defends later in the book agree right down the line with compatibilism; he himself is a compatibilist in everything but name!”.
For example, Harris gives the impression that while he rejects compatibilism he nonetheless accepts the reality of moral responsibility in a deterministic universe.
If Harris really believes that moral responsibility (and choice/praise/blame?) are compatible with determinism, then why doesn’t he go the whole hog and say that whatever features of the universe and human behaviour that allow moral responsibility to exist are the very features that constitute free will? As Dennett says, the only reason Harris seems to offer is that this is not what regular folk mean by free will. But so what?! The whole point about compatibilism is to change the way the talk about free will — and in a way that makes the apparent conflict between determinism and free will evaporate!
As Dennett also observes, it’s not even clear that Harris is right about what regular folk think about free will. One of the things that struck me when reading Harris on free will (and what people think about it) was his complete failure to mention a single empirical study on what people’s intuitions are about free will, moral responsibility and determinism. (You can get a flavour of what these studies have found in my article on this very topic from a few years back.) The reason this omission is so egregious is that Harris like to present himself as Mr Empirical — yet he ignores relevant empirical data when it suits him. (I include another, rather embarrassing, example of Harris’s strange relationship with empirical evidence in footnote 1.)
Leaving these omissions, oversights and selective uses of evidence aside, when it comes to spelling out his position Harris is often far from clear — despite his reputation for rigorous, perspicuous prose. Admittedly, Harris does turn a good phrase here and there, and does clearly express some sensible points. It’s when he comes to arguing, especially against others, that his thinking and writing breaks down.
To see this in action, let’s go back to the idea that determinism doesn’t undermine moral responsibility. I, along with other compatibilists, agree — and so does Harris, apparently. So is he actually a crypto-compatibilist, as Dennett suggests?
I don’t think so. For the compatibilist, one of the key reasons we care about free will in the first place is that it undergirds moral responsibility. Compatibilists agree that the free will supporting moral responsibility cannot be of the magical libertarian kind. (I say ‘magical’ because the libertarian has to somehow argue that we can harness indeterminism to make our own free choices, which is a tall metaphysical task.) So the compatibilists have developed accounts of free will that are consistent with determinism but also preserve talk of moral responsibility. (The devil, of course, is in the detail — but this is what we need in a serious discussion of free will and moral responsibility, and it’s precisely what is lacking from Harris’s account!)
Harris, in contrast to compatibilists, doesn’t really defend moral responsibility in any substantial way — and so there’s little reason to believe that he’s actually a kind of compatibilist at all. He remains a hard determinist, but one who wants to be free of the charge of leaving moral responsibility at the door. This can be seen from a section in The Moral Landscape entitled ‘Moral responsibility’ (starting on p106; it’s also repeated in Free Will, starting on p48). This is supposed to show why accepting that free will is an illusion does not undermine talking about moral responsibility. However, instead of offering an argument for preserving moral responsibility within a deterministic framework, Harris provides a pragmatic argument that punishing people for their misdeeds (by sending them to prison, for example) is consistent with the rejection of free will and responsibility.
It’s an argument I accept. Some things, and some people, are dangerous, and pose a threat to the safety and well-being of others. Avalanches and hurricanes, along with thieves and murderers, fall into this category. To the extent that we would like to prevent, or at least reduce, the deaths of skiers and mountain climbers, we would like to be able to control and prevent avalanches (and other natural disasters, where’s there’s no agent to be held morally responsible at all). As Harris says, “[I]f we could incarcerate earthquakes and hurricanes for their crimes, we would build prisons for them”
Obviously the idea of prisons for natural disasters sounds a bit silly, but the point Harris is making is clear enough: if we could, we would to like to prevent bad things happening to people, irrespective of what causes the bad outcomes. This logic applies to criminals as much as to natural disasters. If someone likes raping, torturing and killing people, then when it comes to protecting other people from their depredations it doesn’t matter whether they’re act of their own free will or not. That is, irrespective of metaphysical debates about free will, a sadistic killer is a sadistic killer, and there’s an obvious case to be made for removing them from society so they can’t do the harm they ordinarily would.
Just think, if an out-of-control lawnmower was imperiling the lives of children playing in a park, we’d disable it — and this has nothing to do with free will. Likewise, if a person posed a threat to the lives of children we’d want to put them in some kind of secure unit (so that we can keep them away from children, or to rehabilitate them so that they no longer pose a threat to children).
That’s all well and good. But this argument has nothing to do with moral responsibility. That is, recognising that some entity (person, machine or mountain) will lead to bad outcomes is enough to justify ‘punishing’ that entity, but the language of morality need never enter the picture. After all, who holds mountains morally responsible for the fatal avalanches that roll down their slopes?
There’s also an irony in Harris’s supposed defence of moral responsibility. One of his main objections to the compatibilist’s account of free will is that it diverges from what ordinary people (non-philosophers) mean by free will. But so does Harris’s account of moral responsibility, which implies that anyone (or anything) that causes bad things to happen, and which we might therefore like to ‘punish’ or ‘imprison’ (criminals, natural disasters etc), is morally responsible for the outcomes they/it produce.
To be fair, Harris does try another tack to establish what moral responsibility means. In Free Will, in another section reproduced almost exactly from The Moral Landscape, Harris writes:
Yesterday I went to the market; I was fully clothed, did not steal anything, and did not buy anchovies. To say that I was responsible for my behaviour is simply to say that what I did was sufficiently in keeping with my thoughts, intentions, beliefs, and desires to be considered an extension of them. If I had found myself standing in the market naked, intent upon stealing as many tins of anchovies as I could carry, my behaviour would be totally out of character; I would feel that I was not in my right mind, or that I was otherwise not responsible for my actions. (p49)
It’s good to know that if Harris ever acts out of character and does something morally bad he’ll absolve himself by virtue of its unusualness and the fact that it’s not part of his typical repertoire of behaviour. But really. If I go out today and for the first time in my life kill someone, that’s pretty out of character — so, following Harris, can I say, in my defence, that I am therefore not responsible for the killing? Of course not. (The above quote from Harris is yet another example of an argument that mixes up a bunch of issues, and confuses rather than clarifies. For instance, Harris talks about responsibility, but is this moral responsibility, or something less contentious like causal responsibility (that is, being the agent/entity that set in chain the series of events leading to a particular outcome)? I’ll deal with this, and other instances, in my forthcoming critique.)
In this case — and many others like it — it’s not clear what’s going on in Harris’s head. Perhaps he really does have sophisticated arguments about why his rejection of free will and “deep responsibility” (as he calls it elsewhere) still leaves room for moral responsibility. However, in the absence of any sign that he does in fact have such arguments, and given his misreading of compatibilism as it’s been developed over the past 40 years, another alternative seems more likely: Harris is simply confused about his position. I take it as prima facie evidence that he’s confused when he implies that a pragmatic argument for punishing people (which relies on causal responsibility) is the same as showing that we can talk sensibly about moral responsibility. (I say ‘imply’ because Harris doesn’t bring these different strands out clearly in his writing.) In which case, it’s not really that Harris is a compatibilist in all but name, as Dennett suggests; rather, he’s a hard determinist who inconsistently holds that moral responsibility can exist in the absence of anything that we could reasonably call free will.
I can understand why Harris would want to offer his readers this concession. He surely realises that some people, even if they accept his rejection of free will, may feel a bit uneasy in jettisoning moral responsibility — and with it, all meaningful talk of praise, blame and so on. So he throws his readers a bone: “Don’t worry, it’s all OK, moral responsibility remains intact, and I’ll show why!” Except he doesn’t. He shows us something else — namely, that punishment can have a justification that doesn’t depend on the weird metaphysics of libertarian free will. (And on this point if he thinks he’s telling compatibilists anything new, he’s profoundly off base.)
This is just one example where Harris’s thinking goes off the rails. Dennett provides many more. But exposing these errors of thought and reasoning isn’t straightforward. In many ways, reading and responding to Harris’s arguments (and not just on free will) is a trickier business than first appearances might suggest. I’m reminded of what the political philosopher Alan Ryan once said about evaluating the arguments of John Stuart Mill in his essay On Liberty:
John Stuart Mill is — surprisingly — a difficult writer. He writes clearly, nontechnically, and in a very plain prose that Bertrand Russell once described as a model for philosophers. It is never hard to see what the general drift of an argument is, and never hard to see which side he is on. He is, nonetheless, a difficult writer because his clarity hides complicated arguments and assumptions that often take a good deal of unpicking.
This kind of unpicking is, frankly, an often tedious exercise. Yet it’s part of what philosophers do, and Dennett applies himself to the task with all the gusto we’d expect. In order to make sense of what Harris’s drift is, and what the arguments supporting it are, Dennett has to assemble fragments of argument distributed across many pages, and break down the sentences in which these arguments are expressed into clauses so that each can be analysed separately. Such a forensic analysis, while necessary, is hard work. Reading Dennett’s critique of Harris ideas I found myself thinking, “What percentage of people who have read and were persuaded by Harris’s arguments about free will are going to even look at Dennett’s critique — and of these, how many will have the tolerance for a long, detailed explanation of where Harris has gone wrong?”. The fact is, correcting profound misunderstandings can take much more space and effort than articulating a baseless philosophical position in the first place.
Even when these dispersed statements are brought together, it’s often unclear what they add up. After quoting a particular claim by Harris, Dennett writes: “I have thought long and hard about this passage, and I am still not sure I understand it, since it seems to be at war with itself.” I do not believe that this points to Dennett’s inability to understand difficult arguments, but to the confusions inherent in Harris’s arguments. Trying to make sense of these arguments, and then judging the merits of the various readings they admit of, is a tiring task — for both author, and reader.
In some ways, responding to Harris is like debating a creationist in front of an audience without much scientific background. I don’t mean to imply for a second that Harris is any sort of creationist — he’s as Darwinian as they come, as am I — nor that he’s as ignorant of science as a creationist, as he isn’t. What I mean is that Harris, like the creationist, can make short, snappy points that are compelling to the audience but, despite their falsity or stupidity, can be quite difficult to counter in a short space. Imagine a creationist stands up and says, “But if we evolved from monkeys, how come there are still monkeys about?”. How do you counter such an epic misunderstanding? First you probably have to explain the difference between monkeys and apes, and point out that apes are our closest evolutionary cousins, not monkeys. Then you have to explain how populations of animals evolve and diverge from a common ancestor and so on until you’ve dispelled the idea that entire groups of animals evolve from one form to the other. If your audience doesn’t get this when they walk into the lecture hall, there’s a good chance you won’t be able to correct them by the time they walk about. (So when Harris points out that Dennett’s critique of Free Will is almost as long as the book itself, that’s why.)
Although Dennett does not consider Harris’s broader body of work in his review, the same pattern of argumentation found in Free Will can be found in his other writing. He frequently misunderstands or misrepresents the ideas of others — and then deploys bad arguments against these misrepresentations, as well as in trying to ground his own views (whatever they are!). Harris frequently spends time criticising implications of ideas that don’t actually follow from the ideas in question, and also follows up his own arguments with a variety of non sequiturs. He takes his readers down long, irrelevant detours that confuse issues rather than clarify them — all the while finding time to pepper his arguments with gratuitous insults to the opposition.
The last point may seem trivial, but it’s reflective of a certain style of intellectual engagement. Harris’s friend Dennett is, morally and scientifically, “deliberately obtuse”, peddling philosophy that “resembles theology” (ouch!). When it comes to other people, he gets even less polite. In disagreeing with a position neuroscientist and novelist David Eagleman took on religion and atheism, Harris wrote “This posture will win him many friends, but it is intellectually dishonest” (sort of like when you say that determinism is true, free will is an illusion, deep responsibility is a myth — but moral responsibility is real!). A couple of months later, Harris offered not quite an apology but a minor mea culpa while talking about his invitation to debate Eagleman: “A few people chastised me for issuing insults along with my invitations (point taken)”. Another example: Harris disagrees with anthropologist Scott Atran’s views on the causes of suicide terrorism, and thus describes Atran as “preening” and “delusional”. Harris has also fallen out with fellow New Atheist and self-proclaimed rationalist blogger PZ Myers, referring to Myers’s “odious blog” and describing Myers as “unscrupulous” and the “shepherd of Internet trolls”.
It’s funny, in a way, to watch someone like Harris, who likes to claim the moral and rational high ground, acting like this. But it’s a real problem. It reflects the fact that Harris’s style of argument is more about beating people down than engaging in any sort of dialogue that would help him, and others, get straight on deep, complex issues. That might make for good spectator sport, but it’s not a serious intellectual endeavour.
And who is this audience who so enjoys Harris’s polemics? I’m not sure, but the very fact that he makes the best-sellers lists makes me fairly confident that it includes a lot (a majority?) of people who are not especially familiar with the topics he covers. As Harris must anticipate, this gives him free reign to play fast and loose with the ideas he’s supposed to be arguing for and against. That’s the only way I can make sense of Harris’s insouciance towards scholarship when tackling deeply scholarly topics.
In the case of free will, this lack of concern about the academic background to his chosen subject is well documented by Dennett. But it also applies to his treatment of moral philosophy, the ostensible topic of an entire full-length book by Harris. In a footnote to The Moral Landscape he offers these two justifications for ignoring much of academic moral philosophy: he came up with his ideas on his own and not from others, and academic discussions of morality are boring. That’s fine, in a sense, but one of the major criticisms of the ideas put forth in The Moral Landscape — including the ideas about free will — is that Harris does not display a deep understanding of the topics he’s elected to talk about! Really, you have to do your homework in these fields, even if you’re not going to adopt the idioms of these fields.
The above issues are ones I’ve contended with in my evolving critique of Harris’s thoughts on free will, morality, and religion. Over and over again, it’s extremely difficult to pin down exactly what Harris’s argument or thesis is at any given point. I get the drift of what he’s saying and I can tell what side he’s (pro-science and rationality/anti-religion etc). However, his arguments are not spelled out clearly or concisely. As with the statements that Dennett had to analyse, they are typically distributed across many short, and often ambiguous, sentences. Frequently, where we need an argument what we get from Harris is a rhetorical question (which, as Dennett points out, is a tactic that can be used to conceal the fact that you don’t have an argument).
While working on this critique, I’ve often had the following train of thought. “Harris reaches a big audience, so it seems that many people are interested in reading about, and thinking about, issues that are close to my heart, such as free will, morality and religion. But how many of them are interested in reading a lengthy deconstruction of Harris’s arguments? Is this a big waste of time?”.
The issue is not whether or not I can get as many readers as Harris. Nonetheless, when you’re trying to show people that they’ve been seduced by some bad ideas, you want to be able to get through to those people. I don’t know what percentage of Harris’s original audience will read Dennett’s critique (or any other detailed, philosophically informed critique), but I’m confident it’s a small subset of the millions have bought and read his books. Perhaps it’d be just a few people interested in the topics Harris covers, and with a sufficient interest in philosophical issues to work through some difficult conceptual terrain — and I don’t imagine that this characterises the majority of Harris’s readers. (That said, I’d rather be talking to that subset. Who wants to be to philosophy what Justin Bieber is to music?) The fact that Dennett took the time to highlight the flaws in Harris’s arguments about free will makes me believe it’s worth me doing the same with the rest of his work.
1. At the Beyond Belief conference in 2006, physicist Lawrence Krause challenged Harris on his views on reincarnation. Harris replied:
Reincarnation … who knows? It may, err, I mean, I have no … well who knows in the sense that there are no … I mean there are these spooky stories where a kid… [Someone in the audience, Krauss I think, shouts “Come on!”] … OK, I’m not … reincarnation … you are on firm ground being sceptical of reincarnation, let me say that … I hear there’s all this data, someone like Dean Radin writes a book about it, Brian Josephson, a Nobel Laureate in Physics, blurbs it. I don’t have the time to do the meta-analysis or the statistical expertise, so, so I’m awaiting the evidence. I don’t want to talk about reincarnation.
I don’t know whether Harris still harbours a soft spot for reincarnation, but it certainly isn’t part of the best scientific picture of the world we have. The worst thing about this reply, however, is the way it justifies entertaining belief in reincarnation. Spooky stories? A book blurb by an eccentric Nobel laureate? Is this really what Harris thinks empiricism adds up to? No. If someone tried to suggest that belief in God was reasonable because there was were some spooky anecdotes and [Person With Qualification X] says God exists, Harris would laugh them out of the room. For empiricists like Harris, anecdotes and arguments from authority are supposed to carry little evidentiary weight. I guess different standards apply when it comes to Harris’s own pet beliefs.