I am exceedingly frustrated by >Daniel Dennett’s response to Sam Harris’ recent book on free will.
It serves as the most elaborate, learned, and desperate hand-waving I’ve ever witnessed. It was such a weak argument that it looked more like an example that a brilliant philosophy professor, like Daniel Dennett, might use to highlight poor arguments to his students. Sadly it wasn’t a strawman used for instruction—it was his real position.
Here’s what he basically said:
- It seems like we make choices, so we do.
- It’s useful to hold people responsible for their actions, so moral responsibility is real.
I just saved you ~30 minutes of exasperation. But let’s take it piece by piece.
Harris and others need to do their homework if they want to engage with the best thought on the topic.
This dismissive, scholarly tone is found throughout the essay. I’d find it less annoying if his argument wasn’t silly.
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.
Um, well, back at you:
When you encounter a growing number of your top-end scientific friends who reject free will, you should consider that perhaps you have missed something. Nowhere in your argument do I see you giving the opening that you’re asking others to give.
However, the ‘free will’ that compatibilists defend is not the free will that most people feel they have. (p16)
First of all, he doesn’t know this. This is a guess, and suitably expressed questionnaires might well prove him wrong.
Seriously? Do you really think that, in a country where only half of the population believes in evolution, any significant percentage of people are going to have an advanced belief in free will?
No. They aren’t. Most believe that people make choices independently of causes, to a significant degree, and therefore deserve reward and punishment. This is the basis of the American justice system and of much of our culture. This highly nuanced dance that Dennett is doing isn’t on the radar because they don’t even have radar.
Ask someone why a murderer deserves to die. Ask 1,000 people in a scientifically valid poll. You’ll find that most people believe the following: The murderer had a choice. That means that despite their bad upbringing, despite their drugged out mom, despite whatever hardships, they had the concrete, tangible, and available OPTION to not commit that murder.
So they are 100% guilty. Period.
That’s the resolution that most people have in this country when it comes to considering free will. Not everyone, but most. Go ahead. Try the poll. I’m happy to be proven wrong.
But let’s continue:
Maybe many people, maybe most, think that they have a kind of freedom that they don’t and can’t have. But that settles nothing. There may be other, better kinds of freedom that people also think they have, and that are worth wanting. (Dennett, 1984)
A couple of things are interesting about this quote. First, like I’ve been saying in my various free will arguments over the last several years, Dennett does not believe in Absolute Free Will. He also says quite clearly, and I think in opposition to his quote above, that many (and maybe most) people believe in a type of free will that they don’t and can’t have.
Agreed. But isn’t that exactly what Harris said?
The randomness gambit
Next topic, and this one is unbelievable:
And to the extent that the law of cause and effect is subject to indeterminism—quantum or otherwise—we can take no credit for what happens.
Dennett, supposedly giving a counterexample:
You must correctly answer three questions to save the world from a space pirate, who provides you with a special answering gadget. It has two buttons marked YES and NO and two foot pedals marked YES and NO. A sign on the gadget lights up after every question “Use the buttons” or “Use the pedals.” You are asked “is Chicago the capital of Illinois?”, the sign says “Use the buttons” and you press the No button with your finger. Then you are asked “Are Dugongs mammals?”, the sign says “Use the buttons” and you press the Yes button with your finger. Finally you are asked “Are proteins made of amino acids?” and the sign says “Use the pedals” so you reach out with your foot and press the Yes pedal. A roar of gratitude goes up from the crowd. You’ve saved the world, thanks to your knowledge and responsible action! But all three actions were unpredictable by Laplace’s demon because whether the light said “Button” or “Pedals” was caused by a quantum random event. In a less obvious way, random perturbations could infect (without negating) your every deed. The tone of your voice when you give your evidence could be tweaked up or done, the pressure of your trigger finger as you pull the trigger could be tweaked greater or lesser, and so forth, without robbing you of responsibility. Brains are, in all likelihood, designed by natural selection to absorb random fluctuations without being seriously diverted by them—just as computers are. But that means that randomness need not destroy the rationality, the well-governedness, the sense-making integrity of your control system. Your brain may even exploit randomness in a variety of ways to enhance its heuristic search for good solutions to problems.
Um, what? How is Daniel Dennett—one of the best philosophers of our time, capable of making such an obvious mistake?
Let me walk you through this:
- Harris says that randomness doesn’t grant you freedom
- Dennett says he has a counterexample, and then shows an example of randomness, to which a human responds in a predictable way (given the input), and then claims a free choice was made
What? This is logically blasphemous. If you’re trying to prove that the human made a free choice you can’t use a response to random stimuli as your evidence. You can program a “moral” robot to make the proper choice if the light flashes a certain way as well, so if this happens did it too make a “responsible action”?
And if you say that this is different then you must describe how it’s different, which means all your work is still in front of you. The randomness of the light was a distraction, not a solution.
Next, Dennett attacks the concept of the replay of the past as a thought experiment for the lack of free will:
To say that they were free not to rape and murder is to say that they could have resisted the impulse to do so (or could have avoided feeling such an impulse altogether)—with the universe, including their brains, in precisely the same state it was in at the moment they committed their crimes. (p17)
Just not true. If we are interested in whether somebody has free will, it is some kind of ability that we want to assess, and you can’t assess any ability by “replaying the tape.” (See my extended argument to this effect in Freedom Evolves, 2003) The point was made long ago by A. M. Honoré in his classic paper “Can and Can’t,” in Mind, 1964, and more recently deeply grounded in Judea Pearl’s Causality: Models, Reasoning and Inference, [CUP] 2000. This is as true of the abilities of automobiles as of people. Suppose I am driving along at 60 MPH and am asked if my car can also go 80 MPH. Yes, I reply, but not in precisely the same conditions; I have to press harder on the accelerator. In fact, I add, it can also go 40 MPH, but not with conditions precisely as they are. Replay the tape till eternity, and it will never go 40MPH in just these conditions. So if you want to know whether some rapist/murderer was “free not to rape and murder,” don’t distract yourself with fantasies about determinism and rewinding the tape; rely on the sorts of observations and tests that everyday folk use to confirm and disconfirm their verdicts about who could have done otherwise and who couldn’t.
Rely on observations and tests of everyday folks? Is that what he really said?
Everyday folks think Bigfoot is real. Everyday folks are waiting for Jesus to return. Everyday folks don’t believe in evolution. Everyday folks can’t find Australia on a map.
Now, let me show you what’s wrong with his rather clever car example. Here he’s talking about capabilities of the car, and how at any given moment you’re seeing what it is doing, and not what it can do.
But one problem: Cars don’t press their own gas pedals.
So if you’re arguing that humans are just like cars, then I agree! That’s a problem for you, not for me (or Harris). Yes, you can go 40 MPH, or 60 MPH, or swerve to avoid the pedestrian, or swerve to hit him—but either way there is a cause for that outcome.
For the car it’s the foot on the gas pedal, and for the human it’s a combination of chemicals coming together in his brain based on the laws of physics. Either way, they aren’t making free choices despite having different settings, or “capabilities”.
What Dennett is missing here is that most people wouldn’t make the argument he’s making. What they would say is that at any given point, as they’re making a decision, they get complete freedom. So if you replayed the tape you’d get another free choice, which would result in a different outcome.
Cars don’t get that.
What he’s doing here (and elsewhere) is alternating between embracing the common man’s view of free will and deviating from it based on convenience.
Harris can’t take credit for the luck of his birth, his having had a normal moral education—that’s just luck—but those born thus lucky are informed that they have a duty or obligation to preserve their competence, and grow it, and educate themselves, and Harris has responded admirably to those incentives.
He keeps making the same error. Let’s try it with Robots:
Robots can’t take credit for the luck of their birth and their normal moral education—that’s just luck—but those born thus lucky are informed that they have a duty or obligation to preserve their competence, and grow it, and educate themselves, and the robots has responded admirably to those incentives.
Taking input and formulating a complex response is not freedom. Plants do it. Animals do it. Computers do it. That is not the standard, and if it were then most living things would qualify; it’d just be a question of degree.
Transparency is key
If you don’t know what your soul is going to do, you are not in control.
Really? When you drive a car, are you not in control? You know “your soul” is going to do the right thing, whatever in the instant it turns out to be, and that suffices to demonstrate to you, and the rest of us, that you are in control. Control doesn’t get any more real than that.
I am stunned that he calls this an argument.
Let’s try this with Feeling God.
Really? When you pray, are you not feeling God? You know “your soul” is going to do the right thing, whatever in the instant it turns out to be, and that suffices to demonstrate to you, and the rest of us, that you are feeling God. feeling God doesn’t get any more real than that.
Actually, it does get more real than that. How about something being true whether or not someone is experiencing it? People see ghosts. People hear voices. People feel like they’re making decisions. Some of these experiences correspond to what we call reality, and others don’t. And we have rules for differentiating them.
So just pointing to the fact that something is being experienced is most definitely not enough. Not even close.
Making an author from the middle of a chain
Harris ignores the reflexive, repetitive nature of thinking. My choice at time t can influence my choice at time t’ which can influence my choice at time t”. How? My choice at t can have among its effects the biasing of settings in my brain (which I cannot directly inspect) that determine (I use the term deliberately) my choice at t’. I can influence my choice at t’. I influenced it at time t (without “inspecting” it).
Brilliant. He just proved that computers have free will.
No free will at all
Where is the freedom in being perfectly satisfied with your thoughts, intentions, and subsequent actions when they are the product of prior events that you had absolutely no hand in creating?
Once you stop thinking of free will as a magical metaphysical endowment and start thinking of it as an explicable achievement that individual human beings normally accomplish (very much aided by the societies in which they live), much as they learn to speak and read and write, this rhetorical question falls flat.
Let’s unpack that.
- An explicable achievement: what exactly is that? I guess he means trivial. Simple. Normal. Non-special.
- That human beings normally accomplish: ok, same thing
- Aided by the societies in which they live: ok, that’s just another set of inputs coming into the mix. I think I agree…
- Much as they learn to speak and write: ok, but chimps learn to speak and write. And computers will likely be learning very shortly here—if you don’t agree they already are
None of these things get us to freedom unless you’re describing it in the loosest and most useless way possible. As I said in the opening, he’s basically saying that if it feels free then it is, and that’s the best you’re going to get.
Well, I don’t have to accept the best illusion I’m going to get and call it freedom. I’d rather, for the sake of human dignity and the respect for reason, acknowledge that it’s an illusion and work within that framework. It’s more true, which I think is generally more healthy.
Only the choices available
Am I free to do that which does not occur to me to do? Of course not.
Again, really? You’re playing bridge and trying to decide whether or not to win the trick in front of you. You decide to play your ace, winning the trick. Were you free to play a low card instead? It didn’t occur to you (it should have, but you acted rather thoughtlessly, as your partner soon informs you). Were you free to play your six instead? In some sense. We wouldn’t play games if there weren’t opportunities in them to make one choice or another. But, comes the familiar rejoinder, if determinism is true and we rewound the tape of time and put you in exactly the same physical state, you’d ignore the six of clubs again. True, but so what? It does not show that you are not the agent you think you are. Contrast your competence at this moment with the “competence” of a robotic bridge-playing doll that always plays its highest card in the suit, no matter what the circumstances. It wasn’t free to choose the six, because it would play the ace whatever the circumstances were whereas if it occurred to you to play the six, you could do it, depending on the circumstances. Freedom involves the ability to have one’s choices influenced by changes in the world that matter under the circumstances.
Again, he just described many different types of decision-maker. Computers do this. Apes do this. Dogs do this. Fish do this.
The more a word means the less it means. He has far too much bolted onto “free will”. It’s so diluted at this point that it reduces to feeling like you make choices and/or adjusting your decisions based on inputs.
Way. Too. Broad.
Here he’s talking about card game options:
If determinism is true, are these real opportunities? Yes, as real as an opportunity could be: thanks to your perceptual apparatus, your memory, and the well-lit environment, you are caused/determined to evaluate the situation as one that calls for playing the six, and you play the six.
He’s failed to even address Harris’ point. If you don’t know the rules, you don’t have the option to use one of those rules to win. If you happen to have a lack of concentration while playing a card, you may play the wrong one.
These things are not in your control. They happen to you. They trim your options down to the limited set that present themselves to you from, well…you don’t know where. That was Harris’ entire point. Whatever options get presented for you to choose from are labeled as your freedom, and you’re not even thinking about all the others that weren’t presented.
What’s free about that?
Oh, right, it’s as free as it can be, and we should be happy with that.
Sure, and a shackled slave is free to run from the plantation in lots of different ways:
- In a straight line
- In a zig-zag pattern
- Towards the river
Picking from available options is illusory freedom because it ignores the fact that you were only presented a few choices, and you weren’t the one who chose them. This is true without even mentioning the uncomfortable bit that you’re also not the one picking afterwards.
We use the same tools to influence our own desires as we use to influence other people’s desires. I doubt that he denying that we ever influence other people’s desires.
Yes, and domino # 47 influences domino # 48. Now let’s group dominos 45-55 together and call them “me”.
The bait and switch
As I have said, I think compatibilists like Dennett change the subject: They trade a psychological fact—the subjective experience of being a conscious agent—for a conceptual understanding of ourselves as persons. This is a bait and switch. The psychological truth is that people feel identical to a certain channel of information in their conscious minds. Dennett is simply asserting that we are more than this—we are coterminous with everything that goes on inside our bodies, whether we are conscious of it or not. This is like saying we are made of stardust—which we are. But we don’t feel like stardust. And the knowledge that we are stardust is not driving our moral intuitions or our system of criminal justice.
Dennett admits to being confused by this. The key is in the last sentences:
Dennett is simply asserting that we are more than this—we are coterminous with everything that goes on inside our bodies, whether we are conscious of it or not. This is like saying we are made of stardust—which we are. But we don’t feel like stardust. And the knowledge that we are stardust is not driving our moral intuitions or our system of criminal justice.
In short, Dennett is saying that because the uncontrolled falling dominos of our actions exist within our bodies, they are ours, and therefore we have freedom.
That’s how he’s getting there.
Harris and I say, “Well, your decisions are just coming to you from your brain. You don’t even know why you’re doing that.”
Dennett says (not a real quote), “So what? It’s happening in my brain, and my brain is mine, therefore I did it.”
That’s playground stuff. Unacceptable. Are you taking responsibility for the bacteria in your body as well? What if the mutant bacteria that destroys the world comes from YOUR intestines? Was that part of you as well? Can we hold you responsible? Of course not.
Thoughts like “What should I get my daughter for her birthday? I know—I’ll take her to a pet store and have her pick out some tropical fish” convey the apparent reality of choices, freely made. But from a deeper perspective (speaking both objectively and subjectively) thoughts simply arise unauthored and yet author our actions. (p53)
What would an authored thought look like, pray tell? And how can unauthored thoughts author our actions?
The same way a previous domino falling in a line “authors” the next one falling.
Compatibilism amounts to nothing more than an assertion of the following creed: A puppet is free as long as he loves his strings.
This is in no way supported by anything in his discussion of compatibilism.
Maybe not, but it’s all over the entire essay that Dennett just wrote. He doesn’t care that we can’t control the options we have before us, nor the faculties that control how we chose among them. He ignores these limitations; he doesn’t even see them as important.
That’s an embrace of the puppet strings, plain and simple. He’s saying that if a puppet feels like it’s a real boy, then it’s pedantic not to call it one. This is an unacceptable position for a thinking person in the 21st century—philosopher, scientist, or otherwise.
Embracing the illusion
Getting behind our conscious thoughts and feelings can allow us to steer a more intelligent course through our lives (while knowing, of course, that we are ultimately being steered).
So unlike the grumpy child (or moody bear), we intelligent human adults can “grab hold of one of our strings”. But then if our bodies are the puppets and we are the puppeteers, we can control our bodies, and thereby our choices, and hence can be held responsible—really but not Ultimately responsible—for our actions and our characters. We are not immaterial souls but embodied rational agents, determined (in two senses) to do what is right, most of the time, and ready to be held responsible for our deeds.
No. No. No.
Harris put “grab hold of one of our strings” in quotes for a reason (because it’s not real). Then he explicitly says that “we are ultimately being steered.” He’s talking about the difference between experiencing choice and actually having it. This point seems hard for many to grasp.
Dennett wants to say that since we can “take control of our lives”, or “write a book”, then we can be held responsible for our actions.
All three are illusory. We are saying that some embrace of this illusion is functional, and practical, and thus we do it all the while knowing it’s not real.
That may seem strange to some, but it’s nowhere near as strange as calling it true freedom. Freedom means something. Use another word.
Back to the top
As I said at the start: Dennett’s argument reduces to this:
- We have free will because we feel like we do
- We have moral responsibility because it’s practical to behave as if we do
This is a reckless assault on truth in the name of wishful thinking. And it would make much more sense if he approached free will like he approaches consciousness. Here is one of his favorite quotes from Lee Seigel:
I’m writing a book on magic, I explain, and I’m asked, Real magic? By real magic people mean miracles, thaumaturgical acts, and supernatural powers. No, I answer: Conjuring tricks, not real magic. Real magic, in other words, refers to the magic that is not real, while the magic that is real, that can actually be done, is not real magic.
To anyone who follows Dennett’s work, it is obvious from his use of quotes like this (and many others) that he is a pure materialist. He believes consciousness is a bag of tricks (and has said exactly that on numerous occasions). He has spent years of his life deconstructing it from its semi-magical state.
He should do the same with free will, and that starts with clarity of language.
Using his favorite analogy above, he doesn’t have to be against magic to dissuade others from thinking it’s real. Magic is nice, and so can be the experience of making choices.
What I believe he should do is acknowledge the difference between absolute and practical free will, and then clearly state that he supports the practical version and not the absolute. This is the clarity that’s lacking in his arguments. He uses purposely opaque language, which, combined with his authoritative tone, casts his entire argument in the black light of an appeal to authority.
He should speak plainly. Perhaps he would say that Absolute free will is false, but we can work with Practical free will in a way that grants us personal responsibility. I don’t currently accept that, but it’s an argument I’d be eager to hear from him. But instead of that we get obfuscation, condescension, and a curious obsession with a much less useful part of the debate—the definition of a term.
An outright denial of Absolute free will by Dennett would span much of the distance between these camps, as most don’t realize that he outright denies that Absolute free will is possible (note that compatibilists ARE determinists—it’s in the definition).
At that point we could begin a far more interesting discussion: How to (and whether to) build a societal structure of personal responsibility, reward, and punishment in a world where we experience choice but don’t actually have any. That’s the discussion we should be having.
I wait eagerly for Harris’ response.
[ 30.01.14: Added some additional comments to the closing. ]
[ 01.02.14: Here’s the argument I think Dennett should be making for modern compatibilism. ]
- Absolute vs. Practical Free Will (the source of Dennett’s confusion)
- The Two-Lever Argument Against Free Will
- I was writing out of frustration here, and purposely kept it raw, but I have tremendous respect for Daniel Dennett and all he has done for philosophy. I’ve read many of his books and will read anything he puts out. It is only on this one topic that he infuriates me.
- I dislike Dennett’s repeated use of “little book” to describe Harris’ essay. Over time it becomes more obviously pejorative.
- Aaron Swartz attacking Dennett’s deductive argument for compatibilism.
- A previous post of mine on Dennett and free will.
- My recent post on Why Free Will Matters