Testing Free Will with Single-Neuron Measurements of Pre-Conscious Activity
technologyreview.com
technologyreview.com
Luckily Kreiman's work isn't about free will, but neural mechanisms of decision making. And it's quite interesting in that context. He (and Bok and Boyden) say as much in the article, but for some reason scientific journalism has to create equivocal contexts for everything because they're convinced whatever they're reporting on isn't interesting enough for what it is. And apparently it works, because it ends up here.
I am thrown into existence, then I absorb information about how the environment works, then I make decisions based on that information. How else could it be? If my locus of control exists outside my physical brain, how useful could it possibly be since the data I use to make ALL decisions is stored in my physical brain? If we propose that information is also stored in some type of cosmic non-corporeal spirit-brain, nothing really changes, the same questions of agency simply move from the physical brain to the realm of the inexplicable spirit-brain.
Great question. You've more or less just articulated the philosophical problem of free will.
Why do we care what the term 'free will' means? Mainly because the concept features heavily in the way we think about (and judge) human behaviour.
If I fail to save a drowning person because I can't swim or am tied to a tree, most people would not judge me as morally responsible for them drowning. If I could have saved them, but didn't, most people would judge me as morally responsible. A theory that explains what "could have saved them" means in that context is a theory of free will.
Accepting that we have free will means that "could have done otherwise" explanations have a legitimate place in our moral thought. Rejecting it means that we need to find alternative explanations for the moral difference between the two situations I described (or just deny that there's a relevant difference). Which is certainly possible, but counterintuitive to many (/most?) people.
I don't think it makes a difference. "Free will" or not, the validity of "could have done otherwise" explanations remain unchanged. As long as physics permits, indeed, they "could have done otherwise", but ultimately, the stream of bytes that steer our machinations (be they predetermined or divinely inspired) result in some action being taken based on input from the environment. I also disagree that human morality is capable of responding to the absolute nature of our choices, regardless of their fundamental constituents (see my answer to sapient)
A theory that explains what "could have saved them" means in that context is a theory of free will.
What could have saved them is a different set of input or a different algorithm processing that input. If you can even conceive of a situation in which you would not save them, but then another one in which you would, you've implicitly acknowledged a state change in the private instance variables that underpin your decision making process, even if it's a change as subtle as "I just felt like it".
You misunderstand. I am defining "free will" as the property that makes people moral agents --- ie. able to do otherwise in such a way that features in (some peoples') moral judgements. I am not claiming that we in fact possess this property (I think that's an open question).
It isn't a substantial account of what free will is. It's a bit more meta: an account of what a substantial account of free will is. Which is important, because as you said yourself: what the hell does 'free will' even mean?
Imagine you're trying to understand a codebase with as much legacy baggage as human thought. You do so by identifying modules, finding out what depends on each one and what it depends on, testing and later maybe improving on it, or discarding it. Free will is one of those modules and before we try to 'fix' it, we need to know what it does.
My point is that all conceivable systems of human morality are equally incapable of distinguishing between humans that possess "the property that makes them moral agents" and those that do not. If the universe is deterministic, then human morality is just another link in the chain. Morality cannot address the fundamental nature of our choices because it exists in a system where the only available input are the consequences of and the perceived motivations for human behaviors. Hitler didn't accidentally slaughter European Jews, his hateful feelings and aspirations for power drove him to towards those actions, even if he didn't choose to be hateful and power hungry. The state of determinism does not modify the criteria for moral judgement.
Imagine you're trying to understand a codebase with as much legacy baggage as human thought. You do so by identifying modules, finding out what depends on each one and what it depends on, testing and later maybe improving on it, or discarding it. Free will is one of those modules and before we try to 'fix' it, we need to know what it does.
In that analogy, I don't think you can describe free will as a module, contemplating free will is more like asking "is the x86 instruction set etched into the processor or is there a homunculus living in the chip that reads incoming instructions and executes them appropriately". The answer to that question has no bearing on whether or not we judge the "human thought program" to be malware (e.g. for corrupting the memory space of another program), what we do is first decide "what was the intention of this program?" and then assess if the memory corruption is a bug or the program's intentionally destructive payload. The absolute nature of the instruction execution mechanism doesn't really matter with regard to how we form judgments about which programs we can all agree are damaging to benign programs and thus a threat to the stability of the system as a whole.
The point is that many do, so if there are no agents (as defined above), it means that we should reject moral theories that we might otherwise accept. That's a significant result whether or all kinds moral judgement require agency (and as you rightly point out, they don't. As did I...).
As far as I can see, it's an open question whether there are any agents. It depends partly on whether the universe is deterministic (which is basically all that these 'scientist solves the free will problem' articles concern) and partly on whether it's possible to give an account of agency that is compatible with determinism.
Basically, if given identical situations we always chose the same response then we don't have free will. The only other option is randomness, basically given identical physical and spiritual situations we somewhat randomly pick an outcome.
Further DNA is tied to identity making the classic argument if you had lived his life you could have made the same choices not be true.
To my mind, people care about free will because one or both of:
1) They want there to be space for a soul to be really guiding actions (otherwise God is punishing souls for no reason, and that breaks some theologies). Other dualist notions can also have troubles here.
2) They confuse physical determinism with classical determinism. Think Greek mythology; if something was fated, it would happen "whatever you do" - which motivates despair, "so why bother?", &c (for the Greeks, the answer was Virtue Ethics). In physical determinism, your actions are a part of what's "fated" - including your reflecting on determinism, including your making whatever decision you make - but if you were a different you, and made a different decision, or didn't try as hard, or tried harder, that would lead to a different outcome, so while you are uncertain about what you're deciding, you're also uncertain about the outcome, and you need very much to try.
It boils down to this - what is a choice? I would contend that you do have a real choice and that the choice is made by you - again, insofar as there is a "you". It just happens to be the case (assuming determinism) that one could know what that choice was going to be, if they possessed more knowledge than anyone actually has.
I don't see it that way. The you that is follows from the (you, stuff that isn't you) at the previous moment. Your actions and the results are both a consequence of the previous (you, stuff that isn't you).
In other words, I'm having trouble differentiating you from the stuff that isn't you. I just imagine the world is a giant mechanical machine, churning along according to some physical laws, where the you is just a conscious entity that identifies itself with small subset of atoms in the machine and believes it can affect the movement of those atoms.
I don't think that question is more interesting or important because I don't think it makes a tangible inquiry. A universe with free will looks exactly the same as one without. What do we care about when we ask "are we truly responsible for our actions?" What is the distinction between true responsibility and deterministic responsibility? Whether you freely decided to reply to my comment or were determined to do so based on an unbroken chain of physical events since the beginning of the universe, the logic behind your locomotion remains the same, a decision was calculated based on data stored in your physical/non-corporeal brain; if you had a choice, you would have done the same thing, if you didn't have a choice, you would have done the same thing.
If we have no free will, then a human killer is as much to 'blame' for killing as the crocodile is.
To me, the idea that moral culpability might hinge on the true nature of free will seems flawed since morality is a construct defined within the scope of physical consequences that arise from human behavior.
We ascribe moral culpability to the human killer not because of their choice to kill, but because of our own perception that the killer understands what it means to take another life. You might say "more precisely, because of their choice to kill within the context of that understanding", but human morality is not equipped to consider the absolute nature of our choices. Be it a truly free decision or a determined one, the human killer's programming matches our description of morally wrong i.e. "understands what it means to take an innocent life, but does so anyway".
For instance, clearly, in such a world, punishment only has value if we model that the outcome of punishment steers the offender away from the undesirable behaviour in future -- since there is no "free will" (whatever that is), if punishment doesn't work, we can stop wasting time on it and expend our energy on more constructive responses to undesirable behaviour (e.g. rehabiliation, and if that is also deemed impossible, at least agree to extract useful output from the transgressor rather than waste energy punishing them, which is a net loss for society).
I think this is the crux of my disagreement.
I don't see how the validity of determinism has any bearing on the efficacy of punishment. Determined or not, people react to punishment in measurable and generally predictable ways, if we discover that all human actions are determined, nothing has changed, all the same motivations for punishment remain in place; it (ostensibly) continues to discourage undesirable behavior, communicates society's disapproval of that behavior, and provides victims with a sense of justice. If anything, determinism only underscores the notion that individuals will react to stimuli, a fact that is equally true if those reactions were "freely" chosen by the individual.
You assume we can choose which direction to go towards. With no free will, we have as much choice in expending our energy as offenders had in committing offenses. It's absurd to state 'we can' or 'we should' while assuming a fully deterministic world.
Choice is most definitely a factor determining moral culpability. Consider a driver who drives into a pedestrian killing him, because at an unfortunate moment, the driver was paralyzed by an unexpected stroke. The driver witnessed the entire event with full understanding of what it means to take a life. Is he morally culpable? I would say no, because he was not in control. You can also imagine similar scenarios involving coercion or accidents.
The key point is whether we are in control of our actions, or merely witnesses with the illusion of control. I don't see how this can be separated from moral culpability.
Right, but in that example, the question of choice doesn't come into play. The paralyzed driver knows what it means to take a life, but they cannot act, regardless of determinism's validity.
The key point is whether we are in control of our actions, or merely witnesses with the illusion of control. I don't see how this can be separated from moral culpability.
In my view, it doesn't matter. From a moral perspective, there is no distinction between the illusion of control and actual control, the variables used to calculate a moral judgement (behavior and the intent driving that behavior) remain unchanged, even if the actor didn't have a choice regarding their intent.
Yes, but doesn't determinism mean that all of us 'cannot act'?
> From a moral perspective, there is no distinction between the illusion of control and actual control
Are you saying a fully capable driver would be culpable if he was under the illusion of full control while he drove over a pedestrian? Consider a vehicle malfunction, which goes unnoticed by the driver, leading him to believe he didn't brake or steer fast enough. Is the driver still at fault or is it the car? Doesn't actual control matter?
Let's say, yes. Now what? If all of us 'cannot act' then from a moral perspective there is nothing left to consider. Instead, we ought to shift our concerns to a system that we at least have the illusion of control over (especially since that illusion also encompasses everything we think we really care about in life), which in our reality brings us right back to where we started. If determinism is true, then our morality is part of the deterministic chain, just because we cannot control it doesn't mean that the rules don't have consequences for our reality, in fact, that is the only context in which they have any meaningful consequences at all.
Everyone will continue to behave exactly as if they could act, and we can make predictions about their behavior exactly as if they had the power to make choices, and our system of morals will continue to handle illusory intentions and behaviors just as well as it handles real ones, because at the very moment that a choice bubbles up into our reality, it's already absorbed into the apparent perception that we are making choices. That apparent reality is the only reality as far as we can tell, and certainly the only one where we can actually think about taking actions. Even if its just an illusion, the illusion is the reality we experience.
Are you saying a fully capable driver would be culpable if he was under the illusion of full control while he drove over a pedestrian? Consider a vehicle malfunction, which goes unnoticed by the driver, leading him to believe he didn't brake or steer fast enough. Is the driver still at fault or is it the car? Doesn't actual control matter?
That thought experiment misses the point because the unnoticed malfunction is literally the physical explanation for the accident. We wouldn't consider the driver culpable because he wouldn't have crashed had the malfunction not occurred. On the other hand, a driver that deliberately runs down a person because they enjoy murdering pedestrians is the physical cause of the event, even if they didn't choose to enjoy murdering pedestrians, they in fact do enjoy it, and will continue to do so unmitigated as long as our moral constructs do not react to it. This scenario plays out the same whether the driver truly makes the choice to run over pedestrians or if he just thinks he made the choice. In this case, determinism didn't make the driver a murderer against his will, determinism defined what that will would be, and the physical consequences are put into motion based on that will.
Here's an interesting paper he wrote on the subject: http://www.pnas.org/content/107/10/4499.full
For some reason people just can't accept determinism; possibly because it often seems to lead to a belief in nihilism and that is judged (perhaps rightfully so) to be destructive/dysfunctional.
On the other hand the only way free will makes sense if it's a property of all the universe and thus essentially is just the phenomenological experience of the passage of time.
Since we conflate free will and consciousness that would mean people would have to accept that there's nothing special about human consciousness and that your dog is also conscious and that even your table must then in some be 'at least a little-bit' conscious.
And that is something we also categorically seem to refuse to accept.
Either way I'd say determinism and universal consciousness are just different sides of the same coin and our opposition to them says more about our adherence to religious concepts and insistence about humanity's 'special status' in the universe than anything else.
On top of that I have all sorts of biological activities in my own body that are not subject to my will. I can't simply choose to stop my heart, or to stop breathing, or to not heal a wound in my skin.
And the counter side is true as well: if you damage my brain, I can't just freely choose to continue thinking the same way as before the injury.
On top of that there are the emotional constraints. Even though I physically could kill my family, I won't. But do I freely choose not to do so, or am I prevented from doing so by emotional reactions that I did not choose to feel? Why don't I choose to just starve to death? Why don't I step off of high places, even though I feel a compulsion to do so? Are those examples of free will, or examples of constraints on my free will?
All that said, there are still quite a lot of decisions and actions by an individual person that we cannot predict with any precision or reliability. From a purely practical perspective, it is useful to assume the existence of free will because otherwise, why can't we predict these things?
Why? I've never heard a coherent non-religious explanation of what substance or system that free will is even supposed to be. I don't see how it could help at all from a practical perspective. Would it somehow help us understand next weeks weather (that we cannot predict with any precision or reliability) better if we assume that wind and water have free will?
Would a person still be pleased and surprised to get a nice birthday card if they believed it was the inevitable result of the clockwork universe?
That's a rationalization of society's reaction to people's behavior. If you need to make up things that you don't absolutely believe exist in order to justify actions that you're taking in the real world, that's bad.
>Would a person still be pleased and surprised to get a nice birthday card if they believed it was the inevitable result of the clockwork universe?
Not believing in a soul or spirit doesn't give you precognitive powers, other than in the fact that your reasoning about the future doesn't involve magical effects, so is generally going to be more accurate.
Nobody knows if this is true or not, or even whether we could know this to be true or not.
For more info: http://plato.stanford.edu/entries/determinism-causal/
For example, I chose to ride a boat today, but it was raining, so I read a book instead.
That was what I was questioning, as the poster above noted, this isn't something you can assume. It's not really an argument for or against free will, just that particular explanation of why we don't have free will.
The point is that certain things are logically axiomatic, I could give you many more examples. So the logical axiom that we exist is used in modern society. To suggest that is is untrue and that science cannot proceed while accepting the existence of the self will only backfire and scientific endeavors in that direction will prove increasingly futile.
Isn't science about doing one thing, but doing it well? I thought if it can't be tested and disproved (as the dream possibility), then it's ‘out of scope’.
> the logical axiom that we exist is used in modern society.
By the above, this axiom would not in any way involve science—another framework would've to deal with it.
(IANAS, though.)
EDIT: I think it cannot for the simple reason that the so called individual is trying to figure out if there is scientific evidence of his own existence.