Along with cats and cowboy boots a long-running theme of Jerry Coyne’s website has been Jerry’s arguments against any form of “free will”. This usually leads to long comment-thread arguments between the incompatibilists (or “hard determinists”) and the compatibilists amongst Jerry’s readers.
I get the impression that sometimes the incompatibilists don’t properly understand a compatibilist view. They often accuse compatibilists of disliking determinism, of hankering after dualism, hoping that something will turn up that will overturn current science, or of just equivocating. Here I want to explain compatibilism to those determinists who take an incompatibilist stance (“hard determinism”). It is not aimed at libertarian dualists!
First, let’s be clear on the two stances. Compatibilism asks whether, given a deterministic universe, one can arrive at sensible and coherent meanings of terms such as “choice”, “freedom” and indeed “free will”. The compatibilist says yes; the incompatibilist says no, regarding such terms as too tainted by the dualistic idea that humans have a non-material “soul” that can make “choices” that are independent of the physical state of the brain and which thus violate the laws of physics.
Second, we should also be clear that the compatibilist is not disagreeing with the incompatibilist over any aspect of science. The compatibilist is only disagreeing over the meaning of concepts such as “choice” and “freedom”. Thus:
Step 1: Embrace determinism
A “choice” or “decision” that a brain makes is determined by the physical state of the brain at the time of the decision; and that physical state is determined by the physical state shortly prior to that; et cetera in a deterministic chain.
Having said that, we immediately need an aside. What about quantum indeterminacy? What about quantum effects coupled with deterministic chaos? Yes, in the long run, the system is most likely indeterministic. But, it is almost certain that our brain’s decisions are sufficiently deterministic that we can neglect quantum indeterminacy for current purposes. That’s because brains are hugely expensive in evolutionary terms (requiring huge amounts of energy to run, needing hugely extended childhood nurturing, and forcing big compromises in female anatomy to enable the swollen brain to pass through the birth canal). Thus they could only evolve if strongly selected for, and that can only happen if the genes that program for brains have a strong influence over the decisions that the brain ends up making. You don’t need a hugely extended brain of 1014 neural connections in order to do quantum dice throwing, and thus such brains would not have evolved if all they did was produce effectively random outputs.
Secondly, even if quantum indeterminacy were involved to some degree, that is irrelevant to anything we could regard as a “will”, since random dice throwing does not give you a “will”. I will thus presume that the system is sufficiently deterministic that, for current purposes, we can regard it as entirely deterministic.
Both compatibilists and incompatibilists make exactly the same Step 1 and have the same attitude to determinism.
Step 2: What is a “choice”?
Let us consider the following conversation, between a father and a child (and I’m willing to bet that even incompatibilists have conversations like this one in their everyday life):
Dad: Would you like an ice cream? You can choose which flavour.
Kid: Strawberry please!
Dad: Here you are.
Kid: on tasting the ice cream: But I don’t like strawberry. 😦
Dad: Well you chose it!
What is a “choice”? What is the kid doing when “choosing”? We (both compatibilists and incompatibilists) are agreed that the child will make a “choice” that is determined by the physical state of the system. That exact child in that exact situation and environment would always make that same choice and ask for the same flavour.
But, humans are so complicated and our environments so complicated that most of the time one human would not be able to guess which option another human being will select. Thus the father is simply asking the child to report their preference. Nothing in the father’s request that the child choose implies anything dualistic or contrary to the laws of physics.
While humans generally might not know another human’s preference, they may have a fairly good idea of the range of preferences humans tend to have and thus the range of choices they tend to make in the range of environments that humans tend to find themselves in.
Thus, we can define “choosing” as the deterministic selection of one option, from among the range of options that would be opted for by a typical range of human beings in a typical range of situations.
Thus the compatibilist agrees with the incompatibilist that a given human in a given situation will always make the same choice. But, what usually matters more is that a range of humans (all with somewhat different natures) in a typical range of situations will make a range of choices. After all, when interacting with another human, you treat them as one member of that range, since you don’t know enough about them to predict their exact choice.
Thus, if you’re tasked with ensuring that a whole class of children can choose their ice cream flavour, or that a whole plane-load of passengers can have their preference of meal option, what matters to you is the range of typical human choices. Nothing about stocking up with the right ratio of chicken versus beef meals has anything to do with dualism or with causation contrary to the laws of physics. That’s simply not what is relevant.
In making the choice the human brain is processing information and selecting an option with the aim of fulfilling a goal (e.g. selecting a flavour with the goal of enjoying eating the ice cream). This concept of choice (being deterministic) applies just as well to chess-playing computers and aircraft autopilots. The chess computer has a goal (winning the game) and processes information to arrive at the option that best furthers that aim. The aircraft autopilot, likewise, processes information and selects the option that best fulfills the aim of level flight, or whatever it is programmed to do.
Thus the term “choice” is one that we use about complex information-processing systems that make goal-oriented selections among options. (And, yes, both the goal and the information-processing towards that goal are determined by the prior physical state of the system.)
We use the concept “choice” where that information processing is sufficiently complex that the outcome would not be obvious to a third party. It is, in essence, a concept that accepts our ignorance of the full details of the low-level information processing that is going on. Thus we don’t use the term “choice” about simpler behaviours, such as a house brick falling under gravity, where we do understand the reasons for the outcome.
In the above conversation the father is exasperated that the child asked for strawberry but then didn’t like it. This would make no sense in a quantum-dice-throwing scenario, since then there would be no reason to expect the randomly-arrived-at choice to be one that the child actually likes. It does make sense under determinism, since one would expect information from past tastings to have influenced the decision. A young, inexperienced and partially-trained neural network will, however, make this sort of error. The parent is thus exasperated in the same way they might be with a clumsy child.
How would the incompatibilist react to the above account? They might react by declaring that such a conception of “choice” is invalid, and that the word should only be used for dualistic un-caused choices that are unrelated to the physical system, and so might that want to retire the word “choice” from the language.
Why would the father be exasperated? Afterall, to the incompatibilist the child had no “choice” and was only doing what the laws of physics dictated that the child do, so why get annoyed? Of course the father might retort that he also had no choice other than to get exasperated!
My suggestion is that, after the incompatibilist has purged the language of all the words tainted with dualism, they would then have to invent a new set of concepts, complete with a new set of words, along the lines of the above compatibilist account. Afterall, those concepts are unavoidable in social interactions.
Step 3: Understanding “freedom”
To an incompatibilist the word “free” in the term “free will” can only mean that the will is divorced from physical reality, and that what the will wants is not the product of the physical system.
Yet, that is not how the word is used in every other context in the English language. In most contexts, the term “free” does not mean that the desire arises independently of physics, it denotes an absence of external constraints that prevent someone acting on their desires. Thus, for example, “free speech” is not about dualism, it’s about social control by others.
Here are some common usages: Free speech; free press; freedom of religion; free style; free load; free radical; freed from jail; free lunch; free fall; free agent; free to leave; freed from slavery; free man; set the birds free; kick your legs free; free form.
Oxford Dictionaries defines “free” as meaning: “Able to act or be done as one wishes; not under the control of another”. It is only in the specific case of the phrase “free will”, out of line with most other usages in English, that “freedom” is taken to be about dualism and the non-operation of the laws of physics. Indeed, if I describe an object as “in free fall”, the whole point is that it is obeying the laws of physics.
Thus, the concept “freedom” has a sensible and coherent meaning in a deterministic world. That meaning is used commonly in everyday life, and is the one adopted by compatibilism.
Step 4: Understanding “moral” responsibility
If a car breaks down we can say that a faulty spark plug is “responsible” for the failure. But obviously the spark plug is not “morally responsible”. To an incompatibilist, the term “morality” is ceded to the religious. It is taken to refer to acts by an immaterial soul that are judged against a standard of absolute rightness or wrongness.
The compatibilist agrees with the incompatibilist in rejecting the religious and dualistic conception of morals. However, as with the above terms “choice” and “freedom”, the compatibilist regards “morals” as about the social interactions of humans. At root, moral judgements are feelings that we have about how humans treat each other. Such feelings help determine how we ourselves act, and by having such feelings we can also influence how others act.
The difference between a spark plug and a fellow human is that no amount of disapproval will affect the spark plug, whereas expressing disapproval will affect another human. Thus “moral” responsibility means susceptible-to-social-opprobrium responsibility. By establishing expectations, rules and punishments we encourage behaviour we like and deter behaviour we deplore.
The incompatibilist will object that, if someone commits a crime despite the deterrence, then they had no “choice”, since the behaviour was determined. True, replies the compatibilist, but given a range of different humans in a range of similar situations, many of them would have been deterred. We hold someone “morally” responsible if many other humans in similar situations would have been deterred and thus would not have committed the crime.
We would, though, not hold someone “morally” responsible if the bad acts were the result of a brain tumour, since social opprobrium would have no effect on the brain tumour. Thus morality is a pragmatic concept about social interactions, and about attempts to influence each others’ behaviour.
The incompatibilist will object that this is a very different concept from the religious and dualistic one, to which the compatibilist agrees entirely, but states that this compatibilist conception of “moral responsibility” is coherent and sensible, and is a necessary part of human society.
Step 5: Interpreting “free will”
As the compatibilist sees it, notions such as “choice”, “freedom” and “morality” are all about human social interactions, and have coherent and sensible meanings in a deterministic world. Even if we end up using different words for these notions, the concepts themselves are required to understand and talk about human society.
Through history, though, most humans have been religious and dualist. When interpreting society they have tended to produce a commentary in terms of dualistic notions such as a “soul” and absolute conceptions of morality. Thus the term “free will” is most commonly interpreted as the action of a soul that acts independently of material. The incompatibilist rightly rejects such notions, and also regards the term “free will” itself as so tainted by dualistic notions that it is best rejected along with dualism. The incompatibilist may also regard words such as “choice” and “moral” as similarly tainted, and might then want to embark on a wholesale re-writing of the language.
The compatibilist is just as clear in rejecting dualism, but only rejects the dualistic commentary about terms such as “choice” and “moral”. Rather than re-writing the language, the compatibilist prefers to re-interpret the relevant words in a manner compatible with a deterministic world. An analogy is with the word “life”, which was not abandoned when science moved from a vitalistic to a materialist conception of life.
While the dualistic interpretation of “free will” is traditional, the compatibilist interpretation also has a long history. Further, that meaning is common in every day life: “Did you sign this contract of your own free will, or were you coerced?”. That question is not about dualism, it is about pressure from other humans, in the same way that “free speech” is not about dualism, it’s about social restrictions.
Einstein paraphrased Schopenhauer’s compatibilism as “Man can do what he wills but he cannot will what he wills”, and compatibilist interpretations have been common back to Hume and before.
Jerry Coyne’s main complaint about compatibilism is that retaining the term “free will” can give people the impression that dualism is tenable after all. Perhaps, though I’ve not seen any evidence saying it is easier to convert a dualist to incompatibilism rather than compatibilism.
In my view, incompatibilists would anyhow be more-or-less forced to adopt compatibilist conceptions of human social interactions, in order to operate in the everyday world, even if they ultimately choose to use different language. While I’m not wedded to the particular phrase “free will”, the terms “choice” and “freedom” and “moral” are needed in everyday life, and replacing them with different words would be way too much trouble.
There is little substantive difference between the compatibilists and the incompatibilists, and certainly there is no difference over the science or over determinism. In many ways incompatibilism is simply a version of compatibilism that is so busy still fighting the dualists (a laudable enterprise) that it hasn’t yet worked out what it would do when it wins. If it weren’t for the spectre of dualism, incompatibilists would realise that they are actually compatibilists, all bar the shouting and some semantics.