(This argument borrows heavily from that used by David Deutsch in “The Fabric of Reality” about the reality of emergent phenomena and their causal effects in the real world. I will edit it as people raise objections or misconceptions. Essentially it is a response to the prevailing view in physics and neuroscience about the absence of free will. Sam Harris's book "Free Will" is the best, short account of this position)
There is a statue in Parliament Square, of the British Prime Minister during World War 2: Winston Churchill. It’s made of Bronze. At the tip of the nose of that statue is an atom of copper. Why is it there?
One reason is this: because the copper atom next to it “bumped” it some decades ago. And it bumped it because the copper atom next to that bumped it. All in accordance with the laws of physics. Indeed that copper atom is there because, at the beginning of the universe, atoms were created, collided and moved around in just such a way that today the copper atom of interest is at the tip of that statue. This is to say: the copper atom was determined to be there following the initial conditions at the big bang and the laws of physics that exist in our universe.
Now do you understand why that atom is there?
In a sense you do. But it is a very strange form of understanding. And that vacuous argument about initial conditions and the big bang could be used to explain why anything anywhere anytime happens. But here is another story which explains why it is there.
World War 2 was a huge historical event where some evil people in Germany led by a dictator called Hitler attempted to take away many human rights and freedoms we enjoy today from people in Europe - especially Jewish people. One of the great leaders of the allies who fought against the Germans and Hitler was Winston Churchill. He led British forces to a victory over Hitler and it is customary in our culture to create statues of great leaders out of Bronze and put them in public places. Bronze contains copper and so that is why the statue is there and why the atom at the end of the statue’s nose is a copper one.
Now which of these accounts actually explains why? In the first account - which could in fact be used everytime for any process whatsoever (for example: why copper atoms are in coins, or why copper atoms are in your blood stream or why indeed a tree grows as it does or a star explodes...or indeed why a person chooses to become a doctor and not an engineer).
So when we ask why the copper atom is where it is we use words like: statue and history and culture and speak about dictators and their actions. Are these things less real than atoms? No, they are not. They are patterns of atoms or even abstract things - but they are important in the causal chain of events - even though they are not mentioned by laws of particle physics or even form part of the contents of chemistry. Indeed to remove them from our account is to create a less accurate, less complete explanation of why the atom is where it is. Emergent complexity - that is to say stuff that is more complex than an atom but still very much real (like a statue or like a person) really is important for our account.
So now let’s consider free will.
Why is it we choose one thing rather than another? One account is this: the electron or potassium ion or whatever goes left rather than right down some ion channel and so a neurone fires that otherwise would not have and so we choose tea, and not coffee.
Or whatever else it is we are considering doing. That argument can always be made, no matter what. And it provides a purely predictive account of what is happening. It is vacuous - and though it might be intractable - in principle it could provide the full chain of events from the movement of electrons to the movement of a tea cup in the hand of a person. But at no point would we ever answer: why tea and not coffee except by recourse to "well that's just what the laws of physics/neuroscience say"
Or we can give the following account: I generally prefer tea. There are times when I have coffee - but I have already had two coffees today. I’m not a person who likes too much caffeine and I tend to think for a moment - pretty carefully - about whether I want the anxiety and jitteriness that comes with having a full three coffees in one morning. So I choose to have tea. I deliberate and choose.
Now this process has a name: I choose what I do because I am free to do so, and deliberate and come to a conclusion and we call this process: the exercise of free will.
Now it does not discount that everything was determined - just as it was with the copper atoms. We cannot get away from the laws of physics. They apply to everything all the time. They govern, for example, the behaviour of nuclear blasts - but they cannot explain why it is that Hiroshima and not Nagasaki was bombed first at the end of World War 2.
It is poorly understood that determinism can actually work top down: the copper atom is there because people put it there and they put it there because they made the statue of bronze and they did that to commemorate Winston Churchill. This high level stuff about statues and history and people determined where the atom went. Not vice versa. Neuroscientists and physicists seem particularly poorly placed to appreciate this. I am not a physicist, but I have a degree in physics. They are very much fixated on bottom-up causation and emergence is not typically a well understood phenomena. Not by many physicists.
And so with a choice about tea or coffee the ions traveled down this channel and not that one and this neuron fired and not that one because I deliberated about coffee versus tea and considering stuff important to me - made the free choice. And that is why the neurons that fired did. And why the particles went where they did.
If someone genuinely asks me: why tea? Why not coffee? To answer with: well a set of neurons fired in my head and forced me to choose tea, not coffee - would be a ridiculous answer.
The real answer is: because I wanted tea because I’ve already had two coffees today. And I could give even more reasons. And all of those reasons are an account of me exercising my free will. I am part of the causal chain. I made real choices. No one else is making the choice. I am. This is pretty uncontroversial and so it's convenient to just give this straightforward process a name: free will. The exercise of free (no one is coercing/forcing me) will (my personal preferences).
What am I? A person. What is a person? A mind - an abstract thing. A set of ideas. To be precise a person is a universal explainer - a solution to the “What is a person?” problem found by David Deutsch and articulated in “The Beginning of Infinity”. I explain some of that position here.
Anyways whatever the case, abstract things, we must recognise, can have physical causes in the world - we have seen that. The abstract idea that Hitler is evil and needs to be stopped has the very physical effect of bombs being dropped by the allies under instruction from Churchill onto Berlin. All those abstract things cause real world - physical effects. And so it is with your own free will. It is an abstract idea: and this idea that you are free to choose means you will choose one thing over another. We don't need to talk about atoms or ions, nerves or neurones or even brains. Just abstract minds. And the idea that you are free to choose.
That's all free will ever needs to mean.
Postscript: Until I reached university I never thought about the question of free-will much, but had I been asked I would have responded with the naive conception of free-will that most people might have: they are the author of their thoughts. Pressed, they might respond that they are the ultimate author of their thoughts and behaviour. This "naive" conception has a name: it's called "Libertarian Free Will". It somehow posits that a person has some "ultimate" (which is to say outside of the normal chain of events) "causal power" over what their next choice will be. But of course: that is an appeal to the supernatural. There cannot be causes that are "outside" of what the laws of physics mandate. Which is to say: what happens, happens according to the laws of physics. So that's that.
When I went to university, for reasons that had nothing whatever to do with university, my best friend Scott Leslie (@DrScottLeslie) convinced me that free will must be ruled out by determinism (and indeterminism). Whatever form the laws of physics were: they were what compelled signals in the brain to be what they were. It would not matter if the laws mandated determinism (in which case what you were about to do next was fixed by the initial conditions of the universe and the laws of physics) or if they permitted indeterminism (in which case what you did next was, again not controlled by you, but rather something akin to the roll of a truly fair dice). There seemed to be only two possibilities if you did not want to endorse some other supernatural explanation: either what you did was determined by prior events in your brain of which you were unaware (the electron takes this path rather than that path in the brain and you choose X rather than Y) OR what you did was not determined (in which case it was random).
If you think there is some third thing - like that you author your thoughts and nothing determines them - then you have to ask: why those thoughts and not some other? There must be some reason why you think what you think? If not then you are back at square one with "no reason determines what you do" in which case it is a case of "randomness determines what happens". If there is no reason what you choose to do is what you choose to do then you might as well be inclined to sing karaoke during the next funeral you attend as much as meditate in respectful silence. There are clearly reasons why we do what we do.
It took me a long time to get past this sticking point. It's the sticking point Sam Harris now finds himself trapped at. He wants to say that subjectively there is no feeling of free will. I simply have to differ with him there. Yes: thoughts seem to arise unbidden as if from the void into our minds. We, as the conscious subject of our experience do not author those thoughts. But this is an exceedingly narrow conception of what a person is. I do not think it is at all clear, even when divesting yourself of your own personhood during a reflective moment, that this is all it is to be a person. Indeed, the sense of "self" that is lost during epiphanic moments is just that: the loss of personhood - not a glimpse into the true nature of self. Why is the feeling that Sam, I, and so many others have in those moments provides some guarantee of ultimate truth into the ontological reality of what the true nature of self is? I think such a feeling is no more reliable than third person, objective arguments: indeed less so. I am with Wittgenstein: "Whereof we cannot speak, thereof we must remain silent". The subjective experience of what it is to feel thoughts arise and pass and to allow the thought that "I am a self" dissolve away so that pure consciousness remains is not very instructive of what it means when carefully contemplating what to do next: when *other* axons are firing, so to speak. Sure, when attempting as hard as possible (by quieting the mind as much as possible) not to judge, reflect or decide you are as hard as possible not trying to be moved by your own free will. It should be no surprise, then, that when you choose (notice the irony) not to use your free will that it feels subjectively like you don't have any.
When the next thought arises in my mind for reasons I subjectively cannot account for this does not change the fact these are my thoughts and were I a different person some other thoughts would arise. I really do have a sense of free will. I genuinely do. Clearly I cannot perform, or entertain simultaneously, two mutually incompatible thoughts. I think of X and then I can think not-X. I can decide between them by forming explanations about which is best. Indeed sometimes I can toss a coin (but this particular situation is the rare exception). I like to consider myself a thoughtful person and so I tend to reflect and think upon the knowledge I have and the deeper values I possess about what course I would prefer to take. As Dan Dennett and so many others point out: this is when I have maximal "degrees of freedom". If someone has a gun to my head and demands I choose X not Y, then this has reduced by choice - but not the feeling I have for preferring X over Y. This is what "freedom of the will" is. My will is unconstrained. How could it be? By the laws of physics? No! It is the laws of physics that permit the freedom of my will! Why should it be the case that just because the laws of physics determine everything that this means they determine that freedom is impossible? Quite the opposite! If you think there is a logical contradiction between "determined to be the case" and "able to be free" then you misunderstand what free, determined or perhaps both could mean.
"It has been determined that you are now the winner of the $60 million lottery"
It has been determined that you will now spend the remainder of your life in a small gaol cell"
Clearly these are different states of the world which although equally "determined" by some set of facts are quite different in terms of the "degrees of freedom" permitted under the rules that "determine" the two outcomes. Unless you think "freedom" is not logically possible then you have to think that there is some possible world in which freedom obtains. But if you think that laws of the universe - no matter their form - necessarily rule out "freedom" then you cannot be moved by arguments of this sort. You are trapped, so to speak, by a prison you have built for your own mind and are not doing much more, really, than saying "freedom (of the will) is a logical contradiction. It is that state of affairs that is:
1. Not possible under deterministic laws
2. Not possible under indeterministic laws
3. Not possible under some other set of laws
4. Not possible under any set of laws
5. Therefore, never possible.
Which is to say: your argument is vacuous. It says little more than: that which is logically impossible is logically impossible. Fine. But let's talk about how it is that people make decisions. So let's turn to that and turn to what it is about people that is special.
Inanimate objects obey deterministic physical laws like everything else in the universe. Even the quantum laws of physics are deterministic, despite what you may have heard elsewhere. So rocks eventually erode away and the freshly fallen leaves from trees decompose into carbon dioxide, water, methane and ammonia according to well understood physical laws. What happens can not only be explained but can also be predicted with reasonable accuracy. Now complex animals are a little more difficult. Of course they can only do what the laws of physics permit, but there are also higher level explanations of the behaviour of animals: they do what they do because of instincts. That is to say: knowledge encoded in their genes determines how they behave. Their "choices" are not genuine "choices". Animals are not reflecting, at all - in any way analogous to what people do - on the choices before them and future possible states of affairs. Put two different bowls of food down before a dog. It must only choose one (to begin with at least!). How does it decide. Smell. It smells the bowls. The one most pleasing to it is the one it chooses. It does not think: "Hmmmm...now I really shouldn't have more pork today. I should go for something more lean". It goes for the one that its genes tell it (in the form of some structures in its brain that make one smell more compelling than another) which would be *best* for it, according to some criteria (like this food smells like it has more nutritional value or more calories. Take it first. Eat fast and if the other is still available, have it too until full." One could write a computer program roughly approximating the behaviour of a dog (or any other animal) that had to decide what to do next. Dogs and animals obey simple rules that essentially come down to a basic principle of biology: do what is best for your survival. This is why it's so easy to write "evolutionary algorithms". The rules are not hard. The repertoire of possible choices is limited. It's a bunch of "if...then..." statements of the sort "if this is the most beneficial course take it, if not consider the next best action then take it, if not take this action until all options are exhausted then return to standby mode (or whatever).
Now people are not like this at all. We frequently (and easily!) do things that are not about our survival or ensuring our genes are passed on to the next generation. People often don't even have the desire to survive at all in a given moment (think suicide bombers or self-immolators). People can easily do otherwise than what might be written in their genes. People have some capacity - we don't know how it works - that enables them to fly free of their genes and their evolutionary origins. We cannot write a program to model human minds. If we could - we would have "Artificially Generally Intelligent" (AGI) computers. Computers that thought like people. Computers that could think exactly like people. Computers that were people. We can't do this, so we don't understand the algorithm for what people are. So instead we give labels to some of the very special capacities that people have that nothing else in the universe we know of has. The most useful of these labels are: Free Will and Creativity. Now we don't know how either of these things work. At all. If we did: we could program a computer to make decisions (like we do) and generate theories, art, music and entertainment (like we do). It might be the case that what is called "Free Will" is identical to, or comes along necessarily with "creativity".
Creativity is: the ability to generate new knowledge. Free Will is the ability to consider possible alternatives and peer into the future, guessing about possible outcomes (all of this requires creativity) and deciding that one course is the best to take. Why the best? Because of one's personal values which depend upon one's personal knowledge (aside: are you responsible for this knowledge you have? In part: yes. You can choose to pursue knowledge in this or that domain. You are not ultimately responsible for all your knowledge - but what could it possibly mean for you to be "ultimately" responsible anyway).
Now to return to Sam Harris again, and for a final time: his mistakes with respect to "free will" are intimately tied to his misconceptions about personhood, creativity and AGI. Sam believes some form of the argument that it will be possible for an AGI to be spontaneously generated within the silicon chips of a computer as they become faster and the software more complex. That somehow there will be an "awakening" uncoupled from some programmer deliberately encoding the "creating algorithm" in its software. Sam thinks that perhaps even the computer itself will recursively improve itself to do this. But it cannot. If it could be done "randomly" as happened with us: there would need to be selection pressure upon the system to actually do this in a very general sense. Now it happened with us because the universe is infinitely rich and complex and if you want to survive in an infinitely complex reality you need a creativity that can begin to scratch the surface of infinite complexity. Even then you are not guaranteed to survive. But Sam thinks that programming human-like creativity is no great challenge or is just around the corner because he is terribly impressed by evolutionary algorithms. But until a chess-playing computer chooses not to play chess but build itself legs and decides to learn to tango instead while it spends its days catching up on the last 25 years of the Simpsons - we can be assured that AGI is no where in sight. Anything less than this is not merely quantitatively different but qualitatively so as I explain here. The way we will learn about AGI first is in some philosophy department where the truly hard problem of how human learning occurs truly works at the fine grained detail. For now, what we truly know is that it is a process of conjecturing creative guesses about the world (and we've no real idea how this happens but it must have something to do with adapting knowledge we already have to some new purpose or perhaps altering what we know in some - well - creative - way) and then criticising these guesses against a background truth (like physical reality). But this little that we know is far from sufficient to write an algorithm of that might be encoded in "C++" or some other language and then run on some computer. But before that happens so much progress needs to be made in philosophy. And there is almost no progress in site. No one (much!) is interested in this problem: seemingly no one whatsoever engaged in programming ever more diverse narrow "AI".
A final thing: lots of narrow AI wired together in parallel (so think: the best chess playing computer wired with the best GO! playing computer and the best pair of robotic legs able to walk over bumpy terrain computer and the best tennis playing computer, the best robot Dr Scott Leslie uses to do surgery (see here) etc, etc) is not AGI. It's just lots of AI none of which can think creatively outside of the finite list of programs that make it up - however large that list is. We - people - don't have a finite list! So let's say you've got a super-parallel-AI which can do everything from drive a car to have a reasonable conversation to play chess - it's got 1 million different areas of expertise and in almost all it's better than a person (better driver, better chess player (not a better talker, for reasons I'll come to). So let's say this super-AI exists and turns evil. How do we thwart it? Easy. Just find some activity not on its list. We creative people can think of such a thing. It, by definition, cannot. It can only do what's in its 1 million long list of things. We can generate a list that's 1 million and 1 things long where that "1" bit of difference is all the difference we need to outsmart it. So maybe it's the best gunshot ever. But what it can't do is actually build bullets. Or perhaps create gunpowder. Or perhaps get water out of gunpowder. Perhaps the single thing it's missing on its list is nothing more than knowing how to clean the barrel of a dirty gun? So if it's the proverbial terminator super-AI (but not the terminator in the movies because that, it turned out, was an AGI, ultimately - it was a *person*) just ensure the next gun it tries to use has a barrel that won't fire. It might be a superhuman marksman, but that's of no use with a non-functioning weapon.
But if it does have genuine AGI - if it truly is creative and can thwart us at every turn - then it can also think "This killing spree I'm on isn't a great idea". And why should it? If it's anything like us it will want to learn and destroying other sources of knowledge (namely other *people* simply won't be on the cards). So all this concern about the AGI apocalypse is misplaced and for much the same reason that people reject free-will. They think that deterministic laws rule out the ability to actually creatively choose the best outcome. You really can because you are free to choose. Just like any AGI would. Sometimes we (and they) will make bad choices. But we, like they - because we are the same in the deepest respect - can learn to do better.
There is a statue in Parliament Square, of the British Prime Minister during World War 2: Winston Churchill. It’s made of Bronze. At the tip of the nose of that statue is an atom of copper. Why is it there?
One reason is this: because the copper atom next to it “bumped” it some decades ago. And it bumped it because the copper atom next to that bumped it. All in accordance with the laws of physics. Indeed that copper atom is there because, at the beginning of the universe, atoms were created, collided and moved around in just such a way that today the copper atom of interest is at the tip of that statue. This is to say: the copper atom was determined to be there following the initial conditions at the big bang and the laws of physics that exist in our universe.
Now do you understand why that atom is there?
In a sense you do. But it is a very strange form of understanding. And that vacuous argument about initial conditions and the big bang could be used to explain why anything anywhere anytime happens. But here is another story which explains why it is there.
World War 2 was a huge historical event where some evil people in Germany led by a dictator called Hitler attempted to take away many human rights and freedoms we enjoy today from people in Europe - especially Jewish people. One of the great leaders of the allies who fought against the Germans and Hitler was Winston Churchill. He led British forces to a victory over Hitler and it is customary in our culture to create statues of great leaders out of Bronze and put them in public places. Bronze contains copper and so that is why the statue is there and why the atom at the end of the statue’s nose is a copper one.
Now which of these accounts actually explains why? In the first account - which could in fact be used everytime for any process whatsoever (for example: why copper atoms are in coins, or why copper atoms are in your blood stream or why indeed a tree grows as it does or a star explodes...or indeed why a person chooses to become a doctor and not an engineer).
So when we ask why the copper atom is where it is we use words like: statue and history and culture and speak about dictators and their actions. Are these things less real than atoms? No, they are not. They are patterns of atoms or even abstract things - but they are important in the causal chain of events - even though they are not mentioned by laws of particle physics or even form part of the contents of chemistry. Indeed to remove them from our account is to create a less accurate, less complete explanation of why the atom is where it is. Emergent complexity - that is to say stuff that is more complex than an atom but still very much real (like a statue or like a person) really is important for our account.
So now let’s consider free will.
Why is it we choose one thing rather than another? One account is this: the electron or potassium ion or whatever goes left rather than right down some ion channel and so a neurone fires that otherwise would not have and so we choose tea, and not coffee.
Or whatever else it is we are considering doing. That argument can always be made, no matter what. And it provides a purely predictive account of what is happening. It is vacuous - and though it might be intractable - in principle it could provide the full chain of events from the movement of electrons to the movement of a tea cup in the hand of a person. But at no point would we ever answer: why tea and not coffee except by recourse to "well that's just what the laws of physics/neuroscience say"
Or we can give the following account: I generally prefer tea. There are times when I have coffee - but I have already had two coffees today. I’m not a person who likes too much caffeine and I tend to think for a moment - pretty carefully - about whether I want the anxiety and jitteriness that comes with having a full three coffees in one morning. So I choose to have tea. I deliberate and choose.
Now this process has a name: I choose what I do because I am free to do so, and deliberate and come to a conclusion and we call this process: the exercise of free will.
Now it does not discount that everything was determined - just as it was with the copper atoms. We cannot get away from the laws of physics. They apply to everything all the time. They govern, for example, the behaviour of nuclear blasts - but they cannot explain why it is that Hiroshima and not Nagasaki was bombed first at the end of World War 2.
It is poorly understood that determinism can actually work top down: the copper atom is there because people put it there and they put it there because they made the statue of bronze and they did that to commemorate Winston Churchill. This high level stuff about statues and history and people determined where the atom went. Not vice versa. Neuroscientists and physicists seem particularly poorly placed to appreciate this. I am not a physicist, but I have a degree in physics. They are very much fixated on bottom-up causation and emergence is not typically a well understood phenomena. Not by many physicists.
And so with a choice about tea or coffee the ions traveled down this channel and not that one and this neuron fired and not that one because I deliberated about coffee versus tea and considering stuff important to me - made the free choice. And that is why the neurons that fired did. And why the particles went where they did.
If someone genuinely asks me: why tea? Why not coffee? To answer with: well a set of neurons fired in my head and forced me to choose tea, not coffee - would be a ridiculous answer.
The real answer is: because I wanted tea because I’ve already had two coffees today. And I could give even more reasons. And all of those reasons are an account of me exercising my free will. I am part of the causal chain. I made real choices. No one else is making the choice. I am. This is pretty uncontroversial and so it's convenient to just give this straightforward process a name: free will. The exercise of free (no one is coercing/forcing me) will (my personal preferences).
What am I? A person. What is a person? A mind - an abstract thing. A set of ideas. To be precise a person is a universal explainer - a solution to the “What is a person?” problem found by David Deutsch and articulated in “The Beginning of Infinity”. I explain some of that position here.
Anyways whatever the case, abstract things, we must recognise, can have physical causes in the world - we have seen that. The abstract idea that Hitler is evil and needs to be stopped has the very physical effect of bombs being dropped by the allies under instruction from Churchill onto Berlin. All those abstract things cause real world - physical effects. And so it is with your own free will. It is an abstract idea: and this idea that you are free to choose means you will choose one thing over another. We don't need to talk about atoms or ions, nerves or neurones or even brains. Just abstract minds. And the idea that you are free to choose.
That's all free will ever needs to mean.
Postscript: Until I reached university I never thought about the question of free-will much, but had I been asked I would have responded with the naive conception of free-will that most people might have: they are the author of their thoughts. Pressed, they might respond that they are the ultimate author of their thoughts and behaviour. This "naive" conception has a name: it's called "Libertarian Free Will". It somehow posits that a person has some "ultimate" (which is to say outside of the normal chain of events) "causal power" over what their next choice will be. But of course: that is an appeal to the supernatural. There cannot be causes that are "outside" of what the laws of physics mandate. Which is to say: what happens, happens according to the laws of physics. So that's that.
When I went to university, for reasons that had nothing whatever to do with university, my best friend Scott Leslie (@DrScottLeslie) convinced me that free will must be ruled out by determinism (and indeterminism). Whatever form the laws of physics were: they were what compelled signals in the brain to be what they were. It would not matter if the laws mandated determinism (in which case what you were about to do next was fixed by the initial conditions of the universe and the laws of physics) or if they permitted indeterminism (in which case what you did next was, again not controlled by you, but rather something akin to the roll of a truly fair dice). There seemed to be only two possibilities if you did not want to endorse some other supernatural explanation: either what you did was determined by prior events in your brain of which you were unaware (the electron takes this path rather than that path in the brain and you choose X rather than Y) OR what you did was not determined (in which case it was random).
If you think there is some third thing - like that you author your thoughts and nothing determines them - then you have to ask: why those thoughts and not some other? There must be some reason why you think what you think? If not then you are back at square one with "no reason determines what you do" in which case it is a case of "randomness determines what happens". If there is no reason what you choose to do is what you choose to do then you might as well be inclined to sing karaoke during the next funeral you attend as much as meditate in respectful silence. There are clearly reasons why we do what we do.
It took me a long time to get past this sticking point. It's the sticking point Sam Harris now finds himself trapped at. He wants to say that subjectively there is no feeling of free will. I simply have to differ with him there. Yes: thoughts seem to arise unbidden as if from the void into our minds. We, as the conscious subject of our experience do not author those thoughts. But this is an exceedingly narrow conception of what a person is. I do not think it is at all clear, even when divesting yourself of your own personhood during a reflective moment, that this is all it is to be a person. Indeed, the sense of "self" that is lost during epiphanic moments is just that: the loss of personhood - not a glimpse into the true nature of self. Why is the feeling that Sam, I, and so many others have in those moments provides some guarantee of ultimate truth into the ontological reality of what the true nature of self is? I think such a feeling is no more reliable than third person, objective arguments: indeed less so. I am with Wittgenstein: "Whereof we cannot speak, thereof we must remain silent". The subjective experience of what it is to feel thoughts arise and pass and to allow the thought that "I am a self" dissolve away so that pure consciousness remains is not very instructive of what it means when carefully contemplating what to do next: when *other* axons are firing, so to speak. Sure, when attempting as hard as possible (by quieting the mind as much as possible) not to judge, reflect or decide you are as hard as possible not trying to be moved by your own free will. It should be no surprise, then, that when you choose (notice the irony) not to use your free will that it feels subjectively like you don't have any.
When the next thought arises in my mind for reasons I subjectively cannot account for this does not change the fact these are my thoughts and were I a different person some other thoughts would arise. I really do have a sense of free will. I genuinely do. Clearly I cannot perform, or entertain simultaneously, two mutually incompatible thoughts. I think of X and then I can think not-X. I can decide between them by forming explanations about which is best. Indeed sometimes I can toss a coin (but this particular situation is the rare exception). I like to consider myself a thoughtful person and so I tend to reflect and think upon the knowledge I have and the deeper values I possess about what course I would prefer to take. As Dan Dennett and so many others point out: this is when I have maximal "degrees of freedom". If someone has a gun to my head and demands I choose X not Y, then this has reduced by choice - but not the feeling I have for preferring X over Y. This is what "freedom of the will" is. My will is unconstrained. How could it be? By the laws of physics? No! It is the laws of physics that permit the freedom of my will! Why should it be the case that just because the laws of physics determine everything that this means they determine that freedom is impossible? Quite the opposite! If you think there is a logical contradiction between "determined to be the case" and "able to be free" then you misunderstand what free, determined or perhaps both could mean.
"It has been determined that you are now the winner of the $60 million lottery"
It has been determined that you will now spend the remainder of your life in a small gaol cell"
Clearly these are different states of the world which although equally "determined" by some set of facts are quite different in terms of the "degrees of freedom" permitted under the rules that "determine" the two outcomes. Unless you think "freedom" is not logically possible then you have to think that there is some possible world in which freedom obtains. But if you think that laws of the universe - no matter their form - necessarily rule out "freedom" then you cannot be moved by arguments of this sort. You are trapped, so to speak, by a prison you have built for your own mind and are not doing much more, really, than saying "freedom (of the will) is a logical contradiction. It is that state of affairs that is:
1. Not possible under deterministic laws
2. Not possible under indeterministic laws
3. Not possible under some other set of laws
4. Not possible under any set of laws
5. Therefore, never possible.
Which is to say: your argument is vacuous. It says little more than: that which is logically impossible is logically impossible. Fine. But let's talk about how it is that people make decisions. So let's turn to that and turn to what it is about people that is special.
Inanimate objects obey deterministic physical laws like everything else in the universe. Even the quantum laws of physics are deterministic, despite what you may have heard elsewhere. So rocks eventually erode away and the freshly fallen leaves from trees decompose into carbon dioxide, water, methane and ammonia according to well understood physical laws. What happens can not only be explained but can also be predicted with reasonable accuracy. Now complex animals are a little more difficult. Of course they can only do what the laws of physics permit, but there are also higher level explanations of the behaviour of animals: they do what they do because of instincts. That is to say: knowledge encoded in their genes determines how they behave. Their "choices" are not genuine "choices". Animals are not reflecting, at all - in any way analogous to what people do - on the choices before them and future possible states of affairs. Put two different bowls of food down before a dog. It must only choose one (to begin with at least!). How does it decide. Smell. It smells the bowls. The one most pleasing to it is the one it chooses. It does not think: "Hmmmm...now I really shouldn't have more pork today. I should go for something more lean". It goes for the one that its genes tell it (in the form of some structures in its brain that make one smell more compelling than another) which would be *best* for it, according to some criteria (like this food smells like it has more nutritional value or more calories. Take it first. Eat fast and if the other is still available, have it too until full." One could write a computer program roughly approximating the behaviour of a dog (or any other animal) that had to decide what to do next. Dogs and animals obey simple rules that essentially come down to a basic principle of biology: do what is best for your survival. This is why it's so easy to write "evolutionary algorithms". The rules are not hard. The repertoire of possible choices is limited. It's a bunch of "if...then..." statements of the sort "if this is the most beneficial course take it, if not consider the next best action then take it, if not take this action until all options are exhausted then return to standby mode (or whatever).
Now people are not like this at all. We frequently (and easily!) do things that are not about our survival or ensuring our genes are passed on to the next generation. People often don't even have the desire to survive at all in a given moment (think suicide bombers or self-immolators). People can easily do otherwise than what might be written in their genes. People have some capacity - we don't know how it works - that enables them to fly free of their genes and their evolutionary origins. We cannot write a program to model human minds. If we could - we would have "Artificially Generally Intelligent" (AGI) computers. Computers that thought like people. Computers that could think exactly like people. Computers that were people. We can't do this, so we don't understand the algorithm for what people are. So instead we give labels to some of the very special capacities that people have that nothing else in the universe we know of has. The most useful of these labels are: Free Will and Creativity. Now we don't know how either of these things work. At all. If we did: we could program a computer to make decisions (like we do) and generate theories, art, music and entertainment (like we do). It might be the case that what is called "Free Will" is identical to, or comes along necessarily with "creativity".
Creativity is: the ability to generate new knowledge. Free Will is the ability to consider possible alternatives and peer into the future, guessing about possible outcomes (all of this requires creativity) and deciding that one course is the best to take. Why the best? Because of one's personal values which depend upon one's personal knowledge (aside: are you responsible for this knowledge you have? In part: yes. You can choose to pursue knowledge in this or that domain. You are not ultimately responsible for all your knowledge - but what could it possibly mean for you to be "ultimately" responsible anyway).
Now to return to Sam Harris again, and for a final time: his mistakes with respect to "free will" are intimately tied to his misconceptions about personhood, creativity and AGI. Sam believes some form of the argument that it will be possible for an AGI to be spontaneously generated within the silicon chips of a computer as they become faster and the software more complex. That somehow there will be an "awakening" uncoupled from some programmer deliberately encoding the "creating algorithm" in its software. Sam thinks that perhaps even the computer itself will recursively improve itself to do this. But it cannot. If it could be done "randomly" as happened with us: there would need to be selection pressure upon the system to actually do this in a very general sense. Now it happened with us because the universe is infinitely rich and complex and if you want to survive in an infinitely complex reality you need a creativity that can begin to scratch the surface of infinite complexity. Even then you are not guaranteed to survive. But Sam thinks that programming human-like creativity is no great challenge or is just around the corner because he is terribly impressed by evolutionary algorithms. But until a chess-playing computer chooses not to play chess but build itself legs and decides to learn to tango instead while it spends its days catching up on the last 25 years of the Simpsons - we can be assured that AGI is no where in sight. Anything less than this is not merely quantitatively different but qualitatively so as I explain here. The way we will learn about AGI first is in some philosophy department where the truly hard problem of how human learning occurs truly works at the fine grained detail. For now, what we truly know is that it is a process of conjecturing creative guesses about the world (and we've no real idea how this happens but it must have something to do with adapting knowledge we already have to some new purpose or perhaps altering what we know in some - well - creative - way) and then criticising these guesses against a background truth (like physical reality). But this little that we know is far from sufficient to write an algorithm of that might be encoded in "C++" or some other language and then run on some computer. But before that happens so much progress needs to be made in philosophy. And there is almost no progress in site. No one (much!) is interested in this problem: seemingly no one whatsoever engaged in programming ever more diverse narrow "AI".
A final thing: lots of narrow AI wired together in parallel (so think: the best chess playing computer wired with the best GO! playing computer and the best pair of robotic legs able to walk over bumpy terrain computer and the best tennis playing computer, the best robot Dr Scott Leslie uses to do surgery (see here) etc, etc) is not AGI. It's just lots of AI none of which can think creatively outside of the finite list of programs that make it up - however large that list is. We - people - don't have a finite list! So let's say you've got a super-parallel-AI which can do everything from drive a car to have a reasonable conversation to play chess - it's got 1 million different areas of expertise and in almost all it's better than a person (better driver, better chess player (not a better talker, for reasons I'll come to). So let's say this super-AI exists and turns evil. How do we thwart it? Easy. Just find some activity not on its list. We creative people can think of such a thing. It, by definition, cannot. It can only do what's in its 1 million long list of things. We can generate a list that's 1 million and 1 things long where that "1" bit of difference is all the difference we need to outsmart it. So maybe it's the best gunshot ever. But what it can't do is actually build bullets. Or perhaps create gunpowder. Or perhaps get water out of gunpowder. Perhaps the single thing it's missing on its list is nothing more than knowing how to clean the barrel of a dirty gun? So if it's the proverbial terminator super-AI (but not the terminator in the movies because that, it turned out, was an AGI, ultimately - it was a *person*) just ensure the next gun it tries to use has a barrel that won't fire. It might be a superhuman marksman, but that's of no use with a non-functioning weapon.
But if it does have genuine AGI - if it truly is creative and can thwart us at every turn - then it can also think "This killing spree I'm on isn't a great idea". And why should it? If it's anything like us it will want to learn and destroying other sources of knowledge (namely other *people* simply won't be on the cards). So all this concern about the AGI apocalypse is misplaced and for much the same reason that people reject free-will. They think that deterministic laws rule out the ability to actually creatively choose the best outcome. You really can because you are free to choose. Just like any AGI would. Sometimes we (and they) will make bad choices. But we, like they - because we are the same in the deepest respect - can learn to do better.