Just when I think there’s nothing more to be said about free will—after all, we’ve hashed over most of the points here—a new piece comes along with yet another take on the issue.
The latest slant, written by Dr. Peter Tse, is either deeply misguided or, less likely, profound in a way I don’t understand. I suspect it’s the latter, an essay that, in the end, is just a “deepity.” Tse, a cognitive neuroscientist at Dartmouth College in New Hampshire, has written a 1200-word defense of free will in New Scientist called “Free will unleashed” (unfortunately, it’s behind a paywall, though I give the reference below).
Although Tse doesn’t define free will, he appears to conceive of it as the condition of human behavior when, in a given situation, with all else equal, you could have done otherwise if you reran the tape of life. I used to hold that definition, too, until I realized that if quantum indeterminacy really does play a role in our actions, then it’s possible for us to have done otherwise in a given situation—that is, with all the molecules in the universe aligned in the same way were the situation repeated—even though we’re not really affecting that decision through any kind of conscious rumination. I now prefer to define free will as Anthony Cashmore did in his article on the topic in Proc. Nat. Acad. Sci. (reference below):
I believe that free will is better defined as a belief that there is a component to biological behavior that is something more than the unavoidable consequences of the genetic and environmental history of the individual and the possible stochastic laws of nature.
That excludes any quantum effects from the notion of free will, for I believe that even the compatibilists who post here don’t find their compatibility in quantum indeterminacy.
At any rate, Tse’s insight—his defense of free will—depends on the fact that when neurons fire, causing electrical “spikes,” the firing can change the subsequent way that other neurons fire:
The missing piece is that neurons can rewire each other. Spikes don’t just trigger subsequent spikes in other neurons. Within milliseconds, they can temporarily change the degree to which synapses –; the nerve structures that pass signals to other neurons –; trigger future spikes. This reweighting of a synapse is like changing the combination on a padlock without opening it, and can happen without necessarily triggering spikes immediately. I base this claim on research from the past decade showing that rapid bursts of spikes trigger the opening of specialised synaptic receptors, altering the responsiveness of neurons to subsequent spikes.
This means that a neuron could now be driven by an input that, moments before, might have contributed nothing to its firing. For example, a nerve cell that has just responded to a touch to your forehead could now respond to someone stroking your hand.
This rapid synaptic reweighting could potentially alter the connectivity of an entire circuit, defining new neuronal paths that signals can traverse. Just as railway switches must be flipped to allow trains to pass, synaptic weights must be reset before brain signals can follow one path through a neural circuit instead other possible paths. And if information is realised in the brain at the level of circuits, not just neurons, it is no wonder that listening to spikes in single neurons has not allowed us to crack the neural code.
I’ll concede the idea of “reweighting,” which doesn’t violate my notion of how brains might work. But how does this “reweighting” give us free will? Because it means, according to Tse, that we could have done otherwise.
You might ask, though, how could that be, since the description above is still deterministic ? Reweighting is just a deterministic phenomenon: neuronal firing affects the way other neurons fire. Here he simply sneaks in quantum mechanics (my emphasis), something that he hasn’t defined as part of “neuronal reweighting”:
What does this have to do with free will? Determinists argue that because all particles follow predetermined trajectories, all events, including our lives, unfold as inevitably as a movie. Indeterminists, supported by quantum mechanics, argue the opposite –that all events are random. In either case, whether predetermined or random, there is no room for free will to make events turn out otherwise.
There is, however, a middle path to freedom between these unfree extremes. If the brain sets up criteria for future firing, and if spike timing is made random by the amplification of quantum-level events in the synapse, it is down to chance how these criteria are met. The inputs that meet criteria cannot be predicted –the outcome depends on which spikes coincidentally arrive first.
How does chance interplay with these internal criteria in real life? If I ask you to think of a politician, your brain sets the appropriate criterion in neurons involved in retrieval of information held in your memory. Perhaps Margaret Thatcher comes to mind. If it were possible to rewind the universe, you might think of Barack Obama this time, because he also meets the criterion. This process is not utterly random, because the answer had to be a politician. However, it is also not deterministic, because it could have turned out otherwise.
Here he appears to find free will in the pure stochasticity of individual quantum events. The problems are twofold: nobody wants to see free will in such indeterminacy, and we don’t even know if quantum effects have an effect on neuronal firing, much less on the behavior that results from such firing. And, absent quantum events, the system becomes purely deterministic. Since Tse thinks pure determinism rules out free will (he’s not a compatibilist), he has to do some fast-stepping to drag conscious “will” into the picture:
Factoring in rapid synaptic reweighting also gets around the argument that free will can’t exist because of the impossibility of self-causation. The argument goes as follows: we act as we do at each moment because of how our brain is physically organised at that time. So because we are not ultimately responsible for the way we are organised then, we are not responsible for the consequences of that action. It had to happen as it did, otherwise a thought could change its own neuronal basis, which is impossible. But with synaptic reweighting, mental events don’t change their present physical basis. They change the neuronal basis of possible future events.
As far as I understand this, though, “synpaptic reweighting” does not get around determinism and is not “self causation”. Absent quantum events, there’s no obvious reason to ditch determinism simply because when a neuron fires it affects how other neurons will fire. To me, this seems like semantic gobbledygook.
Tse then raises the problem that a zombie lacking consciousness could still show this behavior, but “we would not say it had free will.” In other words, to Tse, a conscious decision, one in which you can alter future events by thinking about them, is essential for free will. He then explains, using a food analogy, how consciousness really is important in giving us free will:
If consciousness plays no part in the synaptic reweighting process, there is hardly a free will worth having. (There are many definitions for consciousness, but I define it as all the information that presently is, or could be, voluntarily attended to.) [JAC: note the sneaking in of the word “voluntarily,” which is skirting dualism and, at any rate, doesn’t have anything to do with quantum indeterminacy].
Fortunately, the neural activity associated with consciousness does play a necessary role. One way to demonstrate this is using a thought experiment. Let’s say you are planning a dinner party and play out various possibilities in your mind’s eye. You imagine serving a steak, then realise that one guest is vegetarian, so set criteria “delicious; not meat” among synapses associated with memory retrieval. As described before, whatever comes to mind will meet these criteria yet could have turned out otherwise.
Let’s say spinach lasagne is the first appropriate solution that comes to mind. This solution could only have been reached through intentional manipulation of conscious thoughts, so the neural activity that gives rise to consciousness is necessary for the subsequent act of shopping for spinach. Your brain freely willed the outcome of spinach by setting up specific criteria in advance, then playing things out. Such internal deliberation is where the action is in free will, not in repetitive or automated motor acts.
I don’t get this, for it seems like pure determinism to me. The “internal manipulation of conscious thoughts” so touted by Tse is simply the working of the meat computer we call our brain. That computer running a program based on your genes and environment. So how on earth is this the same as “your brain freely willing the outcome of spinach” in a way that your decision could have turned out otherwise? How could it have turned out otherwise? How do we “manipulate” our conscious thoughts? Am I missing something?
As far as I can see, Tse is coming perilously close to saying we have a dualistic free will, but coming nowhere near justifying his claim that our own thinking can change our decisions so that we could have done otherwise. His ending makes this clear:
This way of understanding the neural code has deep implications. It means that our thoughts and actions are neither utterly random nor predetermined. This counters arguments that free will is an illusion. It shows that the conclusion derived from the dogma of determinism –that mental events, including volitional ones, cannot cause subsequent events –is wrong.
We are not mere automata or unfree characters in a deterministic movie. We can change the physical universe with our minds. For example, it was not predetermined at the big bang when and where aeroplanes would be invented. They were brought into existence by brains that could harness chance to creatively envision a different future.
This does not mean that we require a soul for free will. We don’t. My account is entirely physicalist. But our brains can set criteria, play events out internally, choose the best option, then make things happen. And it could always have turned out otherwise.
I will grant Tse one thing: I agree with him that the invention of aeroplanes wasn’t determined at the Big Bang. For between that event and the Wright brothers there were lots of events in which quantum indeterminacy could have played a role. The configuration of the universe right after the Big Bang, so Sean Carroll tells me, could have been profoundly influenced by pure physical indeterminacy. I’m also willing to grant that mutations—the raw material of evolution—could often be purely indeterminate. And if that’s so, then even the evolution of humans, or of any other species, might not have been inevitable had we, à la Gould, rolled back the earth 4.6 billion years ago, but leaving every molecule in the same place.
But that is not the same thing as saying that if I was planning a dinner party, and rolled back time a few minutes to the moment when I decided what to serve, I could have chosen meat rather than spinach.
Maybe I’m getting something wrong here, and maybe Tse is proposing a type of compatibilism with which I’m not familiar, but it seems to me that he’s simply stringing together a lot of words, mixing them with some findings in neuroscience, and coming up with a type of free will that doesn’t do what it purports to.
Cashmore, A. R. 2010. The Lucretian swerve: the biological basis of human behavior and the criminal justice system. Proc Natl Acad Sci USA 107:4499-504.
Tse, P. U. 2013. Free will unleashed. New Scientist 218:28-29.