Like most people, it seemed obvious to me that we have free will and up until about three years ago, I had not thought about it much or questioned it. Then I read Sam Harris’ small book Free Will (2012, Free Press) which made the case that we do not have free will. In my previous posts on consciousness, I noted that Annaka Harris, David Chalmers and Anil Seth all had some discussion of free will in their books. So I re-read Sam Harris’s book, now with a much deeper understanding of its relationship to consciousness, and came away largely convinced that he is right in seeing free will as an illusion.
Free Will is a very short book, 66 pages of main text amounting to around 15,000 words. It is very well written and jargon-free and makes a very strong case that free will is an illusion. Thoughts and intentions emerge from background causes of which we are unaware and over which we exert no conscious control. Well worth reading, whatever your views on free will.
For most people, consciousness not only involves a sense of self but also the feeling of being in control of your actions, of being the author of your thoughts. In other words, you have free will, the ability to consciously choose among potential decisions or actions. I refer to this as free will in the ordinary sense. This idea of free will emerges from a felt experience, and most people do not question it. That included me until relatively recently.
Three main positions on free will
Harris notes that in the philosophical literature we find three main approaches to the problem: determinism, compatibilism and libertarianism. Determinists believe that our thoughts and actions are fully determined by internal and external background causes and free will is an illusion. Compatibilists accept determinism and redefine “free will” as being free from any outer or inner coercions that would prevent the person from acting on his actual desires and intentions.
Libertarians (no relation to the political philosophy) believe that free will occurs outside of physical causation, whether as the causal action of consciousness on the physical brain or perhaps via metaphysical entities such as a soul. While both determinists and libertarians believe determinism and free will are incompatible, I will refer to libertarians in this review as incompatibilists to avoid confusion with the extreme right-wing political libertarians. Harris notes that our modern understanding of brain and behaviour strongly supports a determinist view and that “today, the only philosophically respectable way to endorse free will is to be a compatibilist”.
David Bourget and David Chalmers surveyed 931 philosophy faculty members at 99 institutions in many countries and found that 12% of philosophers were determinists, 59% compatibilists, 14% incompatibilists, and 15% had other views. Leaving other views aside, this means that 71% of philosophers do not believe in free will in the ordinary sense and only 14% do. This means that almost three quarters of philosophers agree with Sam Harris’ case that (ordinary) free will is an illusion, except that some are quite angry with him for not accepting their redefinition of free will as the freedom from coercion (more of this later).
Objective evidence
Harris’s case rests on objective, scientific observations on the one hand, and subjective evidence from introspection and meditation on the other hand. The scientific evidence includes:
- Observations that brain activity occurs a significant time interval before conscious awareness of the intention to do something.
- Given the right experimental manipulations, people can be led to believe that they consciously intended an action, when they neither chose it nor had control over their movements.
- Hypnotized people who are asked why they have done things that were suggested by the hypnotist will confabulate reasons for their actions that have nothing to do with the actual reason.
To this I would add the experiments done with split brain patients, where an action is requested to one side only of the split brain. When the patient is asked via the other side of the split brain why they did that action, the other side will invent a plausible reason that it believes.
Subjective evidence
Sam Harris devotes more of his book to discussion of our subjective experience, arguing that free will is not only an illusion, but it doesn’t even correspond to any subjective fact about us. Introspection soon shows us that thoughts and intentions emerge from background causes of which we are unaware and over which we exert no conscious control. Seeming acts of volition merely arise spontaneously and cannot be traced to a point of origin in our conscious minds. As an experienced meditator, I am well aware that my thoughts appear spontaneously in my mind, and I can no more decide what I will next think or intend until a thought or intention arises in my mind. To directly observe this is to understand that we are not the authors of our thoughts and actions in the way that people generally suppose. You might respond that you can think connected chains of thought, but meditators know that these arise and cease for reasons outside our conscious control.
Part of the felt experience of free will is to feel that you could have chosen to do something other than what you did, to think the thought “I could have done otherwise” after doing whatever you did do. But it is not possible to go back and make a different decision. This is an untestable belief. The perception that you could have, is actually an understanding that you could make a different decision if similar circumstance arose in the future. And that is likely true, particularly if what you did do last time had undesirable consequences. Read Harris for a fuller discussion of this.
What do the authors of the other books on consciousness that I have been reading think about free will? Anneka Harris agrees with Sam Harris and summarizes the same case, probably just as well since she is married to him. I briefly discuss the views of the other authors below.
Daniel Dennett
Daniel Dennet is a compatibilist who agrees with Sam that we do not have free will and our choices are determined, not free. But he redefines free will as the freedom to do what is determined and is very aggressive about attacking Harris for claiming there is no free will (even though Harris is talking about the type of free will he agrees is an illusion). Einstein expressed this compatibilist view very clearly as follows: In a 1929 interview in The Saturday Evening Post, he said: ‘I do not believe in free will. I believe with Schopenhauer: we can do what we wish, but we can only wish what we must.’
Dennett has written a book Freedom Evolves (2003) in which he defends his compatibilist view that that the concept of free will should be redefined so that it no longer involves a free choice among alternatives but rather refers to our freedom to do that one thing that we must wish to do, in other words our freedom from coercion. I think even this definition is incoherent, since if there is external coercion, it is just one of the external determining factors in the determination of our choice (either to change it because we are coerced, or to not change it and suffer the consequence of the coercion). As Eyal Moses’ 2010 review of his book stated:
“For Dennett, the significance of free will is that it is the basis of morality and moral responsibility, of engaging in moral judgment and holding people responsible for their actions. His thesis is that while free will in the ordinary sense is an illusion, these consequences of free will are real and compatible with his deterministic model of the universe, so free will should be redefined to refer to these consequences. Dennett suggests that calling an action “freely chosen” should not mean that the person had some other possible alternative action (which Dennett claims is never true), but rather should mean that we are justified in holding the person morally responsible for that action.”
Dennett has written a long, condescending and incoherent review of Sam Harris’ book. Harris has responded to this here, and Daniel Miessner has written a devastating critique of Dennett’s arguments and examples in his review of Harris’s book. Miessner summarizes Dennett’s position as (1) We have free will because we feel like we do and (2) It’s useful to hold people responsible for their actions, so we must tell people that free will is real.
If you want to dig a little deeper into the issues around understanding the true nature of free will, I can recommend reading Dennett’s review, Sam Harris’ response and Miessner’s detailed critique.
David Chalmers
Chalmers made a fairly strong case in his book that consciousness is an epiphenomenon of mental functioning and itself has no causal role. But he resisted drawing this conclusion, saying he is agnostic and hopes to find some subtle role for consciousness in causing behaviour. Annaka Harris identifies such a potential causal role for the behaviour of someone who is reporting on their experience of meditating on their empty consciousness (formless awareness). Its hard to imagine how a zombie could exhibit such behaviour.
Chalmers examines a number of potential strategies for avoiding epiphenomenalism but concludes that none of them justify rejecting the conclusion that consciousness is mostly epiphenomenal, perhaps with some subtle exceptions. This clearly is fundamental to the issue of whether or not we have free will, and I find it odd that Chalmers does not discuss free will directly. He is on record as saying he does not have strong feelings about free will. For some reason, he appears to act coy around expressing a view on free will, though he is clearly sceptical and tending towards compatibilism, saying in the Scientific American interview that “If it just means you can do what you want to do, then, well, that seems pretty straightforward. If it’s the ability to do something completely non-deterministic, well, I don’t know if we have that.”
Anil Seth
Anil Seth also comes down very clearly on the side of no “spooky free will” (free will where consciousness causally intervenes in the flow of physical events). He discusses in some detail how intentions are formed in the brain before we become aware of making a decision and the very strong feeling that our “self” has made the decision and is causing the action. He is also a compatibilist, also quoting Schopenhauer in slightly different words “Man can do what he wills, but he cannot will what he wills.”
But Seth, somewhat like Dennett, is clearly unwilling to state that free will is an illusion. Indeed, he says that the conscious experience of volition is as real as any other conscious perception, such as a visual experience of colour. I find it somewhat difficult to take him seriously after he has just reviewed at length all the neurological experiments showing brain activity a significant time before a conscious volition is experienced and that people will experience volition if they are fooled into thinking they are controlling a series of events or the real cause is hidden from them.
What Seth actually means by saying the experience of volition is “real” is that it is indeed an experience we have, even though it does not reflect a reality in which our “self” makes choices from a range of options, any of which the self could have chosen. Seth argues that the reason we have this experience of volition is that it is indispensable to our survival and assists us to realize that we can learn from our previous “voluntary” actions, to possibly make a different choice next time. He may well be right.
In conclusion
The objective evidence, well discussed in some detail by Anil Seth, and the subjective evidence, well laid out by Sam Harris, and which I have also examined in some depth, together make a strong case that free will in the ordinary sense, is an illusion, or as Sam Harris puts it more strongly, even the illusion of free will is an illusion. It seems to me very likely that not only are free will and the sense of a conscious self illusions, but that consciousness is largely along for the ride, and plays at most only a limited causal role in behaviour and thought.
I have not touched on the implications of this for moral responsibility. Sam Harris spends quite a bit of time discussing this in his book, and Daniel Dennett sees moral responsibility as necessary and that it justifies redefining free will so people can be told they have it. There is quite a body of empirical evidence around this issue and I will examine it, as well as evidence on the free will beliefs of the general population and specific groups, in Part 2 of this post on free will.
Mountains and rivers 
I view free will not as a feeling, but rather as an event. Consider the waiter in the restaurant. He watches the customer browse the menu for a while, and then the customer calls him over and gives the waiter a dinner order, such as “I will have the Chef Salad, please”.
The waiter is not a mind reader, and has no idea what went on in the customer’s head. What neurological activity was below the customer’s conscious awareness versus what activity happened consciously does not matter to the waiter. The waiter has certain knowledge that the customer looked at a menu of possible dinners, and then selected one dinner he would order.
Based on this empirical evidence, the waiter takes the order to the chef, and later brings the dinner, and the bill, to the customer who ordered it.
To me, this is “ordinary free will”. Whether the customer was having any illusions about what was going on or not, the waiter certainly knows what happened. He knows that the menu contains a list of the many things that every customer can order. He knows that the customer considered those options in some fashion and then spoke his intention to the waiter, “I will have the Chef Salad, please”.
The customer chose, from the many things that he “could” order, the single thing that he “would” order.
We call this event “choosing”. Choosing inputs two or more options, applies some criteria of comparative evaluation, and outputs a single choice. The choice is usually in the form of an “I will X” where X is the thing we have decided we will do. This chosen intent then motivates our subsequent thoughts and actions.
It is commonly understood, that when people are free to make this choice for themselves, that it is a “freely chosen will”, commonly referred to simply as “free will” (see the first definition in any general purpose dictionary).
Free will is a deterministic event, reliably caused by our own goals and reasons, and is fully compatible with a deterministic view of the universe.
From the waiter’s perspective, the customer is responsible for paying for the dinner he ordered. This perspective is also available to the customer. The customers observe themselves considering the options on the menu, and narrowing their options until they settle upon what they will order.
Neither the waiter nor the customer can attribute this choice to any other entity. Regardless of the prior causes of the customer’s goals and reasons, they are now the customer’s own goals and reasons. And they are the reliable cause of the customer’s choice.
The neuroscience does not change this. It enlightens us as to the role of unconscious processes in decision making. But it can never assert that decisions are not being made.
I’ve only just noticed that you had commented on my post. Sorry for overlooking this and thank you for a stimulating response.
I’ve reviewed studies which have asked ordinary people what they understand free will to be and it is indeed the general understanding that it means at the point they were given the choice of the two icecreams they could have chosen either at that point in time. But they chose only one, and the choice was the product of a whole set of causal factors, genetics, personality, lived experience, preference for flavours, or who knows, the decision to make the decision based on the toss of a coin, or to ask another person to make the decision for them. The one thing no-one can do is actually go back and make a different decision. The belief that you could have freely made the other choice is an untestable belief. The perception that you could have, is actually an understanding that you could make a different decision if similar circumstance arose in the future. And that is likely true, particularly if what you did do last time had undesirable consequences. Read Sam Harris for a fuller discussion of this.
In my view, it is an illusion to think we could have done otherwise (at the original decision point) UNLESS we completely redefine free will to refer not to free choice of options but to having the unconstrained ability to do the thing we had chosen to do (compatibilism). I think its basically dishonest to redefine free will to be something other than what most people understand by it, and then tell people they have free will. Which is what most philosophers are now doing.
Here’s a Reddit post where I explained the issue of “can” versus “will” in more detail:
Causal Determinism: A World of Possibilities
From the Same Studio that Brought You “Cause and Effect”…
It is sometimes suggested that a deterministic world, limited to one actual future, eliminates all other possibilities. But this is short-sighted. The same evolved intelligence that produced the notions of cause and effect, from which determinism derives its “causal necessity”, also produced the notion of possibilities.
What are Possibilities?
Possibilities exist solely within the imagination. We cannot walk across a possible bridge. We can only walk across an actual bridge. But this does not mean that possibilities are useless figments of our imagination. Possibilities are very important, because we can never build an actual bridge without first imagining a possible bridge.
In the safe sandbox of the imagination, we can run through many bridge design choices, estimate the likely outcomes of each, and choose the one we think is best. In the imagination we can lay out a plan of action, test it in our minds before we test it in the field, to see what steps must come in what order to successfully construct our bridge. Only then are we prepared to build a real bridge, strong and durable.
Uncertainty Necessitates Possibilities
If we were omnisicent, and already knew every detail of what would happen in the future, then we would have no need for the notion of possibilities. We would never use words like “can”, “might”, or “may”, because we would always know exactly what “will” happen.
But, of course, we are not all-knowing. Quite often, we only have clues as to what will happen, clues that only tell us with certainty what “can” happen, but not what “will” happen. Special words, like “can”, “might”, or “may”, shift us from the context of actuality to the context of possibilities. And whenever we do not know for certain what “will” happen, we imagine what “can” happen, to better prepare for what does happen.
From Many to One
Whenever we must make a choice, there will be two or more options, and we must select one. Each option is a possible future. Some choices are small things, that affect our immediate future. Will we wear the white shirt or the blue shirt today? Will we have cereal or pancakes for breakfast? Other choices are major things that determine the course of our lives. Which college will we attend? What career will we pursue? Will we buy a house now or later?
Each choice selects a single actual future from among the possible futures available to us. From among the many things that we can do, it is up to us to select the single thing that we will do.
Within the domain of our choices, the single inevitable future will be chosen by us from among the many possible futures we will imagine.
There is a many-to-one relationship between what can happen and what will happen, and between what we can choose and what we will choose. This many-to-one relationship continues to exist when we reflect upon our past choices. There are many things that we could have done, but only one thing that we would have done.
This many-to-one relationship, between can and will, between possibility and actuality, is a matter of logical necessity, and thus cannot be altered by causal determinism.
Causal determinism may safely assert that we “would not have done otherwise”, but it cannot logically assert that we “could not have done otherwise”.
Everyone Makes Mistakes…
Hey, what?! But we’ve always heard that causal determinism implies that we “could not have done otherwise”!
Sorry, but we cannot conflate what “can” happen with what “will” happen, without destroying the logical mechanism we evolved to deal with matters of uncertainty.
Conflating “can” with “will” creates a paradox, because it breaks the many-to-one relationship between what can happen versus what will happen, and between the many things that we can choose versus the single thing that we will choose.
Using “could not” instead of “would not” creates cognitive dissonance. For example, a father buys two ice cream cones. He brings them to his daughter and tells her, “I wasn’t sure whether you liked strawberry or chocolate best, so I bought both. You can choose either one and I’ll take the other”. His daughter says, “I will have the strawberry”. So the father takes the chocolate.
The father then tells his daughter, “Do you know that you could not have chosen the chocolate?” His daughter responds, “You just told me a moment ago that I could choose the chocolate. And now you’re telling me that I couldn’t. Are you lying now or were you lying then?”. That’s cognitive dissonance. And she’s right, of course.
But suppose the father tells his daughter, “Do you know that you would not have chosen the chocolate?” His daughter responds, “Of course I would not have chosen the chocolate. I like strawberry best!”. No cognitive dissonance.
And it is this same cognitive dissonance that people experience when the hard determinist tries to convince them that they “could not have done otherwise”. The cognitive dissonance occurs because it makes no sense to claim they “could not” do something when they know with absolute logical certainty that they could. But the claim that they “would not have done otherwise” is consistent with both determinism and common sense.
Causal determinism can safely assert that we would not have done otherwise, but it cannot logically assert that we could not have done otherwise. If “I can do x” is true at any point in time, then “I could have done x” will be forever true when referencing back to that same point in time. It is a simple matter of present tense and past tense. It is the logic built into the language.
Literal versus Figurative
One might ask, “How did we come to make this error in the first place?”. It comes from using figurative language.
Causal Determinism tells us that every event is both an effect of prior events and the cause of new events. Thus, every event is said to be “causally necessary”, in that it must happen where and when it happens, exactly as it does happen. But, what else would anyone expect?
We’re all used to the notion of cause and effect, and we take it for granted in everything that happens and in everything that we do. Causal necessity weaves these simple instances of cause and effect into a chain of events. One thing leads to the next, and so on, as far back in time, or as far forward, as anyone can imagine.
What are we to make of this? Well, nothing really. It is simply the way things happen. We open the restaurant menu and encounter a list of possibilities, the many things we can order for dinner. We consider these options in terms of our own desires, our own dietary goals. Our own reasoning causally determines what we will order for dinner.
It was always going to happen exactly as it did happen, with us in control of what we would have for dinner.
But some people look at the causal chain and suggest to us that, if our choice was causally necessary, from any prior point in time, then “it is AS IF we never had a choice at all.” That’s a “figurative” statement. We often use metaphors, similes, personification, hyperbole and other figures of speech in our communication. But figurative statements share one serious problem: Every figurative statement is literally false.
Take the statement “it is as if we never had a choice at all”. It suggests that, because our choice was inevitable, we were not really making a choice. But we literally (actually, objecively, empirically) did make a choice. In fact, had we not made a choice, the waiter would have never brought us our dinner.
So, figurative statements may be colorful and rhetorical, but they cannot be taken literally, without distorting the truth.
Thus, causal necessity, through figurative usage, acquired many implications that are simply false. When we remove these many false suggestions, causal determinism once again becomes simple cause and effect, and not some monstrosity trying to rob us of our freedom and control.
Whenever a choosing event appears in a causal chain, there will always be at least two things that you CAN choose even though there will be only one thing that you WILL choose. The notion of “possibility” evolved to cope with our lack of omniscience. If we were omniscient, we would never use the word “can”, but only the word “will”. It is only when we do not know what will happen (or what we will choose) that we invoke the context of possibility, and imagine what can happen (or what we can choose).
There is a logical blunder in the notion that determinism means we “could not have done otherwise”. Unfortunately, this blunder is entrenched in the traditional notion of determinism. While what can happen constrains what will happen (if it cannot happen then it will not happen), the reverse is not true. What will happen does not constrain what can happen or what could have happened.
For example, you buy a couple of ice cream cones, one chocolate and the other vanilla. You say to your daughter, “I have two ice cream cones. You can choose the chocolate or the vanilla, and I will have the other.” Then she chooses the chocolate.
Now what will she say when you say to her, “You COULD NOT have chosen the vanilla”? If she’s bright she will say, “You must be lying. Just a moment ago you said I can choose the vanilla, and now you’re saying I could not have chosen it. Either you were lying then or you are lying now!”
And, she’s right of course. But if you told her “You WOULD NOT have chosen the vanilla”, she will easily agree. She knows she like chocolate best, and would not have chosen the vanilla.
The statement, “I chose the chocolate even though I could have chosen the vanilla” is considered true in both its parts.
And this is the situation with every choosing operation. There will always be more than one thing that we CAN choose, even though there is only one thing that we WILL choose. And because “I can choose vanilla” was true at the beginning, “I could have chosen vanilla” will be forever true of this choice at the end.
People who object to the claim that they “could not have done otherwise” are not making a metaphysical assertion. They are simply using the English language correctly.
Determinism cannot validly assert that we “could not have done otherwise”. It may only truthfully assert that we “would not have done otherwise”.
The cause of this error is figurative thinking. If it is the case that only one thing would happen, then it seems to us AS IF only one thing could happen. But, like every other figurative statement, it is literally false.