Until we have a definition of consciousness that allows us to determine whether something is conscious or not I don’t really see the point of debating the compatibility of determinism and consciousness (whatever that word is supposed to mean).
Not just our own consciousness, but the consciousness of everything in the universe. If such a crackpot idea is compatible with hard determinism, then that’s a blow for the latter.
>If such a crackpot idea is compatible with hard determinism, then that’s a blow for the latter.
Is it really? I don't think compatibility with a crackpot idea/thing brings a blow to an orthogonal idea.
For example, I could believe that Mars is an alien base. This is crackpot, but it doesn't give any blow to countless other ideas it's compatible with, e.g. economic or social theories, gravity, explanations of historical eras, etc.
Why would computational explanations of consciousness be more ridiculous than the non-explanation that is asserting everything is conscious in some way (which is basically what panpsychism means)?
At what degree of computation does consciousness emerge? There must be some level of computation at which there’s zero consciousness and then one rung above that there’s a non-zero degree of consciousness.
Where could that line possibly be?
I personally cannot reasonably even guess at such a boundary, so panpsychism is a much cleaner explanation.
These two ideas aren’t really mutually exclusive anyway. I.e. you can believe that consciousness is what it feels like to compute, but then “what is compute” except a particularly structured form of a thing that’s happening everywhere and all the time (particles and energy bouncing around) in more or less structured forms. And if you say “well not all energy/particle interactions are computation, even of an abysmally low degree,” then return to the first question: what types of interactions or what degrees of structure are non-computational where one rung above it they are computational?
The computational explanation would hold that there is some specific computation that can be identified as what we call consciousness. This would be equivalent to how we say that a specific computation is what we call "sorting".
It's not a gradualist notion. A computation either is or isn't consciousness. If it is, then an agent which is currently performing that computation is conscious. If an agent is performing a computation that is not consciousness is not conscious.
Similarly, certain computations represent a sort. Agents which are performing that computation are sorting. The clearest case is a computer running a sort algorithm, but there are also purely physical processes that perform the same computations, and animals or people that do it - so "agent" here is a very broad term. Agents which are performing a computation that is not a sort are not sorting.
Of course, determining which computation is actually consciousness is a very difficult problem, well beyond our current knowledge. As such, it is also obvious that it may turn out that this is not a correct model. But I don't see why this would be considered a ridiculous model of what consciousness means. It is also a model which can, in principle, be falsified, and it could actually help resolve certain difficult real-world problems. For example, if we did have a satisfying definition of which computations represent consciousness, we could then use it to determine at which point a child becomes conscious, or whether a particular animal is conscious or not (assuming we consider that conscious === should not be killed arbitrarily, which is a completely different debate).
By contrast, panpsychism doesn't really resolve anything and can't be used to determine the answer to this type of moral question even in principle (if any system is somewhat conscious, then even swiveling my chair means that I am destroying a conscious system; but also creating another one).
A “computation” is a concept that we have in our heads. Computations only actually occur as an interaction among some substrate of the physical universe. So you’re still saying that if you arrange atoms and energy in the right configuration then we go from zero consciousness to some consciousness. If you can find that configuration I’m certainly open to being convinced.
Even your “non-gradualist” explanation does in fact run into gradualist problems, as you mention one of when a child “becomes conscious.”
You assert that this theory is falsifiable: how so? You can’t even falsify the notion that there’s only one consciousness at all in the entire universe (yours, the one you’re “doing”).
Regarding moral implications, inconvenient moral conclusions are not the same as “no help in arriving at moral conclusions.” You just named one such conclusion that you’d be compelled to arrive at (I wouldn’t, personally), so it’s not really true it doesn’t help determine the answer.
> A “computation” is a concept that we have in our heads. Computations only actually occur as an interaction among some substrate of the physical universe. So you’re still saying that if you arrange atoms and energy in the right configuration then we go from zero consciousness to some consciousness. If you can find that configuration I’m certainly open to being convinced.
A computation is a physical process. A pendulum swinging is a computation of the path that the pendulum will take (and of many other things). Whether there is some conscious being that observes and understands that computation or not is irrelevant to whether the computation is happening (that is, it won't affect the result in any way).
And yes, I am saying that one possible theory of consciousness is that some particular arrangement of matter and energy evolving in a certain way over time is consciousness. That is what I mean by a computational theory of consciousness. I am not in any way claiming I could define today which particular arrangement and evolution that is - I explicitly said I believe the specific model is far beyond our current understanding of the subject.
> Even your “non-gradualist” explanation does in fact run into gradualist problems, as you mention one of when a child “becomes conscious.”
I wouldn't call a theory that tells you that a system is either conscious or unconscious "gradualist". Yes, the system is not conscious at time A and becomes conscious at time A+1. This happens exactly the same way as any other computation: my processor is currently not performing a sort, and when it reaches the sort() call, it becomes sorting all of a sudden; when the call to sort() finishes, it is no longer sorting. Similarly, we can imagine a procedure consciousness(); when a system starts executing this procedure, by definition it becomes conscious; when it finishes executing it, it becomes not conscious.
> You assert that this theory is falsifiable: how so? You can’t even falsify the notion that there’s more than one consciousness at all in the entire universe (yours, the one you’re “doing”).
Well, it is falsifiable in two ways. First, there is the same extent as any mathematical model is falsifiable: we can check whether it accurately accounts for our intuitions. It's the same way we can check whether the formal definition of a continuous function is correct.
Now, once we accept that the proposed model matches what we intuitively mean by consciousness, it is further just as falsifiable as any other physical theory - it makes it so that the question "is system A conscious" (within the formal definition of consciousness I defined above) becomes a measurable claim about whether a particular system is performing a particular computation.
> Regarding moral implications, inconvenient moral conclusions are not the same as “no help in arriving at moral conclusions.” You just named one such conclusion that you’d be compelled to arrive at (I wouldn’t, personally), so it’s not really true it doesn’t help determine the answer.
If the conclusion is "any act you make is just as morally evil, since it leads to destroying a conscious system", that's not just inconvenient, it is by definition not helpful in making moral choices. It is as useless for arriving at moral conclusions as a very convenient model which says "no entity is conscious, so any choice you make is morally good".
Alternatively, if you believe the consciousness of a subject is irrelevant to moral choices, then any model of consciousness is also not helpful in arriving at moral conclusions.
And, if you believe that the amount of consciousness is the relevant bit for moral choices about a system, than still panpsychism is no help, since it only asserts that all things are conscious, it doesn't promise a model for determining which are more conscious and which are less conscious (you would need an additional model on top of panpsychism, p...
I'd recommend going back to spontaneous generation first. A very respectable, old idea. Or perhaps the existence of a soul? That's even older, and is still heavily defended until this day. If lineage is so important, belief in a soul trumps pan-psychism.
Consciousness is the thing and the only thing that you have direct experience of. Everything else could be an illusion. You could be a brain in a jar and all of the reality you perceive be a simulation, but you would still be conscious.
I know that I am conscious and I cannot doubt it, but at the same time I cannot prove it to you. Furthermore, I strongly suspect and assume that other human beings are conscious, but I cannot verify this scientifically (i.e. measure it somehow).
This means that consciousness is a phenomenon that cannot be doubted (in the first person) but possibly eludes the scientific method. This really grates people who have a strong emotional attachment to the dogma that science must be able to account for all of reality.
People confuse the metaphysical commitment of materialism with scientific fact, and thus conclude that consciousness must be an emergent phenomenon of neural complexity, and that it is silly to believe that anything else could be conscious. This feels like an hyper-rational, scientific stance, but really it is not. Science works perfectly fine within other metaphysical commitments such as idealism, panpsychism, etc.
This is not to say that materialism is not correct, I do not know if it is or not. My point is that many contemporary "scientific types" tend to believe that their certainty about materialism being the correct framework is grounded on science, when it is not. It is often an unexamined and quasi-religious belief.
I understand the argument but I disagree and would phrase it the other way around.
We don't have evidence of conscious experience because we don't have evidence for a division between conscious experience and "existence/physical stuff" in the first place. This division is created by our brains and the model we have of the world but in reality the only evidence we have is that something exists.
Therefore, what we call subjective experience is evidence for the existence of what we experience, but until we can empirically show that there is a difference between what we experience and what reality is, the only thing we can say is that it exists. Also, that's the problem with the hard problem of consciousness - it assumes a division between the two, and we have no evidence of such a division. And lastly if it's not clear i am not a panpsychist either because that assumes a division on some level again. We don't know what existence itself is and how it functions really, at least not the brains function within existence.
> Therefore, what we call subjective experience is evidence for the existence of what we experience, but until we can empirically show that there is a difference between what we experience and what reality is, the only thing we can say is that it exists.
The division is between our third-person descriptions of reality in science, and our first-person "experience". You can't logically derive a first-person perspective only from third-person facts. That's what's at the heart of the hard problem. The eliminativist simply says that the first-person facts are a cognitive illusion, and so we really only have third-person facts, and the panpsychist says all third-person facts have an accompanying first-person component most of which we can't access.
> Furthermore, I strongly suspect and assume that other human beings are conscious, but I cannot verify this scientifically (i.e. measure it somehow).
I don't think this is fully true, if you really examine this belief. You do have certain criteria that you use to determine, at least quasi-scientifically, if you should believe that some other entity is conscious or not. You say you believe "humans beings are conscious", but I would bet you don't believe that dead human beings are conscious, or that fertilized eggs are conscious. You may or may not believe that they are still conscious while under general anaesthesia, or while they are in clinical death. You may very much believe that certain things that are not human beings are also conscious - maybe dogs, or even all mammals, or even all vertebrates.
Your criteria probably start from the initial premise: I am conscious. Then, you apply a simple argument: I perform certain actions and/or display certain behaviors, and I have this conscious motivation for them. This other being performs these behaviors very similarly, so they probably have a similar consciousness motivating them to do so. You then test this assumption, and come up with more refined observations (e.g. not every similarity is equally related to consciousness - for example, you move around, and shadows move around, but it seems unlikely that shadows have consciousness so perhaps "move around" is not a good test for consciousness).
If something has (a) a robust temporal-spatial sense, a "reference frame" in time and space, (b) physical boundedness, (c) a robust and general predictive function which is capable of abstraction, (d) independent volition, and (e) behavior that is non-deterministic in the weak sense that its actions cannot easily be foreseen by any arbitrarily powerful computer within the universe that it resides in, then: That thing is conscious.
It's possible to imagine entities that don't necessarily satisfy all conditions and are nevertheless somehow conscious -- i.e., a sufficiently powerful AI that has no clear physical bounds and a poor (but not entirely absent) spatial reference frame, but which possesses independent volition and can independently simulate concepts -- but if all conditions are clearly met, then an entity is unambiguously conscious.
This is not novel in any way, but its style and the way you can ask questions can definitely help shape your ideas and steer you towards topics to research further. If you are not overly familiar with these concepts I can see this being helpful and interesting.
Personally I tried to tease novel techniques for compression out of it. It got me curious and at least it forced me to learn more about the domain to explain why its ideas were wrong.
Maybe that's a thing: it's so confidently (sometimes subtly) incorrect that it may take effort to learn how it's wrong. This itself can be an interesting exercise that wasn't really possible before.
Are those ideas not essentially orthogonal and therefore never were in conflict? Determinism is about what can or will happen given the current state of the universe, panpsychism is trying to explain where your experience of the current state comes from. There might be some conflict, for example if you conclude from your experience of having free will that you actually have free will, but those do not seem central enough to necessitate conflict.
1. Moral Responsibility: The concept of moral responsibility would need to be re-evaluated. [...]
You are not going to reevaluate anything in a deterministic universe.
"I have noticed even people who claim everything is predestined and that we can do nothing to change it look before they cross the road." (Stephen Hawking)
People get so hopelessly confused about determinism and minds...
Because he makes an easy mistake in the quote? I would say the real confusion will start showing up once you try making sense of the words »noticed« and »claim« in that sentence.
I think that determinism is not a useful theory, and I think that is what Hawking was getting at in his wry comment.
Even if the universe is deterministic, it doesn’t feel deterministic, and determinism doesn’t help me with ethics or morality. It reminds me of The Stranger.
[...] determinism doesn’t help me with ethics or morality.
It does. When we look at criminals we nowadays take into account medical and psychological conditions, we are aware that people are not always in full control of what they are doing. This is obviously a weaker form than full blown determinism, but we certainly take into account that people are not completely free in their actions at all times. Similarly the assumption of a deterministic universe will influence ethics and morality, even though it will of course not be the only thing under consideration. [1]
[1] Ignoring the big complication that talking about considerations and things like that in a deterministic universe is itself something that needs a lot of care.
> It does. When we look at criminals we nowadays take into account medical and psychological conditions, we are aware that people are not always in full control of what they are doing.
But that doesn't change anything one way or the other. The efficacy of policies can be determined by measuring the outcomes. Policies which reduce crime / recidivism rates are good; policies that don't are bad. Whether or not anybody has free will makes no difference, the outcomes are what they are whether or not people have free will.
If raising the DUI fine 10x causes a drop in DUI rates, then that's a good policy even if people don't have free will. And if it has no effect on DUI rates, then it's an ineffectual policy even if people have free will. Free will is irrelevant, measurable outcomes are what matters.
Policies which reduce crime / recidivism rates are good; policies that don't are bad.
Killing all people will reduce the crime and recidivism rates to zero. At the very least you want some more nuance.
The efficacy of policies can be determined by measuring the outcomes.
Killing a healthy person and taking two of its organs to save two terminally ill people waiting for an organ transplant will increase the number of years lived. Is outcome really all that matters?
Breaking the window of a car because you like breaking stuff and breaking the window of a car to unsuccessfully save a dog left behind in the car on a hot day, are they really morally the same because the outcome is the same? Yes, you talked about policies and not individual actions, but with some more thinking I could certainly find a good example for policies.
Meet me half way here, of course I'm simplifying the actual process of determining which laws are worth having and which aren't because the nitty gritty Nth order effects of laws aren't what we're discussing here. If killing one person to harvest their organs saves two lives, but also makes people cry in anguished despair at the immoral and unjust state of their society, then that's a bad law. But that's got nothing to do with free will or lack thereof.
My assertion, the point I'm making, is that at no point does the existence or nonexistence of free will factor into the process of determining which laws are worth having. The way people feel about laws does factor into determining what laws are good or bad, but that is true whether or not people's feelings are deterministic or derived from free will.
I think we actually agree on this, we just have to separate two pieces. When it comes to determining which kind of behavior we want, what we want to encourage and discourage, then determinism or not is probably not relevant. I would not completely exclude the possibility that it might sneak in somewhere, but ad hoc I can not think of anything.
The other piece is more the implementation and enforcement of the rules, how do we deal with people violating the rules. Here I think determinism or not has to influence the way we set things up. With determinism you can not really justify putting people into small prison cells, they should instead be allowed to live a normal live except from being prevented to come into situation in which there bad behavior would kick in again.
So depending on what exactly we call the law - whether only the desired or undesired behavior or also including policies around implementation and enforcement - I could agree to both, determinism plays [almost] no role or it does.
I think he does not make a mistake at all. It is more than just a snarky joke, it actually gets at the heart of the issue, which is that determinism of physical laws happens at a different level of description than considerations about human minds such as making a decision, or any thought process at all. And you can see in the long sibling thread once again what kind of confusion ensues when people don't take that distinction seriously.
Minds are implemented within physics, but human concepts w.r.t. minds (e.g. responsibility, choice, freedom..) are only valid on the level of the minds, without any reference to the underlying physics. Any time people use physics to make judgments about morality they are just hopelessly confusing themselves by switching between "standard" physics and dualist metaphysics without noticing.
Minds are implemented within physics, but human concepts w.r.t. minds (e.g. responsibility, choice, freedom..) are only valid on the level of the minds, without any reference to the underlying physics.
I agree that the concepts we use to deal with human minds are higher level concepts. I am somewhat hesitant to call them emergent because that does not seem to quite fit the common usage of the term, but if hard pressed, I would probably call them emergent at the end of the day, just in a more complex way than the color of things or the wetness of water emerges. I would also agree that some of the concepts are in one way or another substrate independent, i.e. they are not strongly tied to the underlying physics, i.e. you could build the same concept on top of other physics or maybe oven on top of something completely different, say mathematical set theory.
But...I don't think it valid to think of physics and minds as completely separable. If physics is deterministic, you do not get to introduce a fundamentally nondeterministic decision process at the level of minds. You can have one that seems nondeterministic to the mind because it lacks knowledge of all the positions and momenta of all the particles, but it would not be truly nondeterministic. You can also talk about a fictitious nondeterministic decision process that is just not realizable in the deterministic universe.
So I think there are at least two classes of human level concepts, those that are independent of the underlying physics and those that are not. And some times there will be overlap, prime example free will. Even in a deterministic universe you can have free will in one of the compatibilistic senses, say doing what you want to do, but you will neither get to determine what you want do, nor what you do or whether the two are the same.
We of course mean the same thing. Reevaluation has, at least to me and in the context it was used in here, this quality of deciding to have another look. Well, in some sense you will even in a deterministic universe decide to have another look, this really comes down to how you determine and count the available options.
I never understood hard determinists who also claim they're not fatalists and spend their lives discussing morality (eg Sam Harris). I guess they have no choice whether they do...
Even if everything is deterministic, human society can only function successfully with the illusion of free will.
If we admit nobody can be blamed for their actions, we can't punish them and thus the incentives not to do bad things disappear and the whole shebang collapses. Just cause and effect.
That we exist as we do is a result of evolution selecting for perception of free will.
The question should not be whether or not people have "free will" over their actions, since free will is impossible to emperically measure. Rather, the question should be whether deterrence and/or rehabilitation have measurable impacts on how people behave.
Does Policy X reduce Crime Y? If yes, then it doesn't matter if people have free will or not, Policy X is effective. If no, then it doesn't matter if people have free will or not, Policy X is not effective. Either way the cookie crumbles, free will doesn't matter. So the question of free will is quite literally irrelevant.
If we admit nobody can be blamed for their actions, we can't punish them and thus the incentives not to do bad things disappear and the whole shebang collapses.
That's not true. You can put people behind bars, say for example murderers, without blaming them for what did but in order to protect others. Even today we will, for example, lock people with dangerous mental illnesses up in order to protect others from them, even if we know that what they did, they did because of some condition outside of their control.
This still implicitly asserts that the murderers are to blame for something. "Protecting others" carries an implicit assertion that failing to abide by this principle means you are broken, rather than that the principle itself is mistaken, as would be true if you really believed blame didn't exist. And then you still have to account for whether the transgressor was coerced rather than was acting from their beliefs and desires (aka made a free choice). You can't really escape free will.
> "Protecting others" carries an implicit assertion that failing to abide by this principle means you are broken, rather than that the principle itself is mistaken, as would be true if you really believed blame didn't exist.
Not at all. People seek to stop fires in order to protect others, but few believe fire carries blame for burning people. Of course, you need to believe that there is determinism (the fire is causing the burns, so stopping the fire will stop the burns), but that's it.
> And then you still have to account for whether the transgressor was coerced rather than was acting from their beliefs and desires (aka made a free choice).
This is still just a question about causality. The problem you are trying to solve is that people are getting hurt. You need to determine the ultimate cause for that in order to effectively solve it. So, if A is killing people that you want to live, you will want to first stop A from doing so. Next, you will have to determine what was causing A to kill people - if A was doing so because B paid for it, then you probably need to stop B as well: otherwise B may pay C to kill people even if A is in jail. If this has been determined, A may or may not need to be jailed at all, based on beliefs about whether A may still continue to kill people even if B has been stopped.
None of this requires some metaphysical notion of blame, it only requires a belief in determinism and a model of causality of behavior. A computer could be built that would take such decisions in a purely deterministic way, and using only information about purely physical facts (A performed act x, B performed act y, other people that share characteristics d, e, f with A performed act x, etc).
> but few believe fire carries blame for burning people
Because fire doesn't have intentions, beliefs and mental states that respond to feedback about right/wrong, and that govern its future behaviour.
> he problem you are trying to solve is that people are getting hurt. You need to determine the ultimate cause for that in order to effectively solve it. So, if A is killing people that you want to live, you will want to first stop A from doing so. Next, you will have to determine what was causing A to kill people - if A was doing so because B paid for it, then you probably need to stop B as well: otherwise B may pay C to kill people even if A is in jail. If this has been determined, A may or may not need to be jailed at all, based on beliefs about whether A may still continue to kill people even if B has been stopped. None of this requires some metaphysical notion of blame
It's frankly amazing that you can perform an in depth analysis of how to ascertain who is to blame, and yet conclude that blame is not required. You're performing this elaborate dance trying to avoid the word "blame" despite the fact that this analysis is exactly what everybody means by "blame". There's nothing metaphysical about it.
You can call it blame, of course, but then we also have to say that the fire is to blame for hurting these people. The point was to show that the same analysis for why we should stop a fire (which is not considered a moral agent) is or can be carried out for why we should jail someone. Since we don't need to believe that the fire has free will in order to put it out, it stands to reason we also don't need to believe that a killer has free will in order to justify jailing them.
> The point was to show that the same analysis for why we should stop a fire (which is not considered a moral agent) is or can be carried out for why we should jail someone. Since we don't need to believe that the fire has free will in order to put it out, it stands to reason we also don't need to believe that a killer has free will in order to justify jailing them.
The key difference is that fire is not considered a moral agent and a person is. That's why the justification is required.
The difference between moral agents and amoral agents is only relevant if you care about non-physical effects of morality. The legal system wouldn't need to change if we just believed that no one is a moral agent - this is what I'm trying to say. So, to the extent that free will is required for being a moral agent, doing away with belief in free will would not require any changes in the legal system.
I don't see how you reach that conclusion. The law is specifically targeted at liability and responsibility of moral agents. That's why it has its own definition of free will of moral agents that is agnostic to determinism. That's why you can't sue a fire but you can sue a person. In the future, you will almost certainly be able to sue certain types of AI agents as well.
I also assert, with evidence [1], that this legal understanding of free will is how most people think about it. I can't make any sense of your argument unless you're employing a different definition of free will.
There are many laws or regulations aimed at the protection of humans from things that are not other humans - regulations on what food or medicine is allowed, regulations for how states have to protect people from fire, regulations for putting down animals that are dangerous to humans etc.
There are of course separate laws for protecting humans from other humans, and that is typically where questions of liability and intent etc arise. But I would still argue that these still have nothing to do with beliefs about (strong/non-compatibilist) free will (though of course they historically are related to free-will-based religious conceptions). For one thing, these laws are explicitly targeted at human persons. There is no aspect in the law that would make it apply to a chimp, for example, even if the chimp were determined to have free will just as much as a human. You can't sue a chimp, even if you believe that it is a moral agent: you can only sue a human person. You also can't sue any type of AI agent, and would not be able to without a change in the law.
So I would say that existing regulations which apply to non-human beings are also proof that we are perfectly capable of creating laws that don't rely on beliefs about the free will of the entities which these laws govern. The fact that current laws for humans make use of free will terminology is thus incidental, not fundamental. All current legal principles could be re-stated without requiring free will as well - this is what I was trying to claim (though I did claim more than that, for which I was wrong).
I will also note that this thread was started by someone who was claiming that you couldn't use the current legal system if you believed the universe is deterministic. The paper you cite works against this argument: it claims that the common notion of free will that most people apply is in fact compatible with a deterministic universe. If you also agree with this paper, then we are probably mostly in violent agreement overall, and in disagreement with sanitycheck's original claim.
This still implicitly asserts that the murderers are to blame for something.
Well, this will quickly turn into the nuances of definitions of words. Blame can have several subtilely different meanings, so call it blame if you want. I will avoid that problem by saying that we go from asserting that a person did something wrong and could and should have done otherwise to asserting that a person did something wrong.
"Protecting others" carries an implicit assertion that failing to abide by this principle means you are broken [...]
Yes, if we have the principle that one person should not harm another and you do not follow that principle, then in a deterministic universe we would say that you are not behaving as expected, that you are broken if you want. Well, there is probably some nuance required here, are you really broken and will keep harming others or was this more like an exceptional situation making you misbehave once but unlikely to do so again?
"[...] rather than that the principle itself is mistaken, as would be true if you really believed blame didn't exist.
Take the example of some manufacturing robots, if one of them breaks and keeps crashing into others, then you will remove and replace the broken one, you are not going to blame it or assign any responsibility to it, because one of the wires connecting a sensor got lose. I do not see why enforcing a policy, categorizing behavior into acceptable and unacceptable behavior, requires blame in any form. We just categorize, we do not even have to look why something ended up in one category or the other. A related but different question is what our policy is, where the line between acceptable and unacceptable behavior lies.
And then you still have to account for whether the transgressor was coerced rather than was acting from their beliefs and desires (aka made a free choice).
That falls under the point mentioned above, more nuance, whether the behavior was situation specific or is likely inherent. But that is again independent of free will - whether I am determined to do bad things or freely decide to do bad things, one should try to prevent me from doing bad things.
And it matters even less in case of coercion, or maybe it is even more complicated in the case of free will. There is not really anything to be said about deterministic coercion to do a bad thing, but if you have free will, maybe you should have resisted more? This will not always apply, but at least sometimes you will have a choice to resist the coercion and probably experience some negative consequences in return, or to go along a cause negative consequences for others.
> a person did something wrong and could and should have done otherwise to asserting that a person did something wrong.
These statements seem equivalent. To say someone did something wrong is to say they should have done something else.
> then you will remove and replace the broken one, you are not going to blame it or assign any responsibility to it
Yes, you are literally saying this robot is to blame for the problem. The only difference with intelligent beings is that they can understand what they did wrong if they are given feedback, and thus change their future behaviour, and that's why holding them responsible requires a different kind of action than holding a dumb robot responsible.
When AI advances to the point that the robot too can understand this, then it would make sense to also hold it responsible in a similar way to humans.
Note that responsibility and blame don't necessarily entail punishment. This is a common mistake. In many prior discussions, I've found people jumping to hard determinism rather that Compatibilism simply because they don't like punishment, but nothing in Compatibilism necessarily requires punishment.
To say someone did something wrong is to say they should have done something else.
They important word is could not should.
Yes, you are literally saying this robot is to blame for the problem.
I am saying the robot caused the issue - for which you can use the word blame if you want to - not that the robot is responsible or could have done otherwise - for which I would reserve the word blame most of the time - which he could not as the issue was caused by a lose cable.
[...] than holding a dumb robot responsible.
I don't think you can hold a robot responsible without stretching the definition of holding responsible quite far.
I disagree emphatically with the inclusion of "could" in the sense that you are using it. When a mother scolds her child for misbehaving, she's saying the child should have acted otherwise, not that the child could have chosen differently if you rewound time and presented the child with the same choice again.
When laypeople agree that someone "could" have done otherwise, they're not agreeing with nondeterminism, they mean that any mentally competent person with common knowledge would have chosen differently given the same circumstances, ie. that there was no mental, health-related or coercive element that was forcing them to choose wrongly.
And this should be obvious if you think about it: we have centuries of legal precedent based on exactly this reasoning, and most people find it perfectly reasonable. That's why free will in law will never be overturned, despite what some hard determinists think. Whether people are deterministic is simply irrelevant to such judgments.
> not that the robot is responsible or could have done otherwise
Frankfurt debunked the principle of alternate possibilities' relevance to responsibility more than 40 years ago. I don't know why people still think it's relevant.
> I don't think you can hold a robot responsible without stretching the definition of holding responsible quite far.
This sort of exchange is not at all uncommon:
Frank: what was responsible for the server downtime?
Alice: an update messed up the routing tables on our routers.
The intended meaning is very clear: use "who" for blameworthy agents, and "what" for blameworthy non-agents. You can of course say this sort of informal dialogue is imprecise and not in line with formal definitions, but nobody would find it a stretch or confusing.
We are stuck fighting over the definitions of words. And I kind of lost sight of the initial issue.
Even if everything is deterministic, human society can only function successfully with the illusion of free will. If we admit nobody can be blamed for their actions, we can't punish them and thus the incentives not to do bad things disappear and the whole shebang collapses.
If you kill people, we will lock you up, because we think killing people is bad and we want to prevent you from doing it again. You could just implement and enforce [1] that rule, period. Doesn't even sound unreasonable to me, not even without further elaboration. So when you are saying
If we admit nobody can be blamed for their actions, we can't punish them [...]
then this can at most mean that we can not punish them with a reasonable justification, we obviously can just punish them. One thing another comment reminded me of is that we are no longer punishing people for what they did, at least not in the sense of some kind of revenge. Anyway, if free will or blame or whatever plays any role, then only for building a reasonable justification for our rules.
[1] Ignoring for the moment that you could not actually decide to implement or enforce a rule like that in a deterministic universe - it would either appear out of the previous state at some time or not and enforcement would happen or not.
In a deterministic mindset, your choice is still specific to you, even though it is determined by the state of the universe at the time it was made. So, your choice to commit a certain act is very much still determined by the existence of laws that would punish you were you to do that act (among many other things, of course). A simplistic model is training a reinforcement learning (RL) agent: you design a reward/punishment system to determine the behavior of the agent. Typically people don't consider that the agent is morally evil if it commits certain acts, but they still punish the agent if it does them, so that it learns not to. And of course, we know by construction that a RL agent is living in a fully deterministic world and has no free will.
The difference is that humans are not RL agents, they're living breathing beings with feelings. We can't (under our moral frameworks which are inescapable) just delete people who are sub-par, we need to be convinced that they bear moral responsibility for their actions first. To kill or imprison someone who had no choice about their actions would be unjust.
> To kill or imprison someone who had no choice about their actions would be unjust.
That is not the problem under discussion. The question actually sits in the definition of "choice". You seem to believe that something constitutes a (moral) choice only if the agent could do it or not without regard to its entire life of experiences up until that point. The more common notion of (moral) choice is that it constitutes any act which an agent takes without being coerced completely into it by some other agent. If my whole life up to this moment caused me to shoot you, that was a moral choice I made. If a dog bites my hand and pushes my finger into the trigger causing me to shoot you, then I didn't make a moral choice to shoot you (however, if someone is holding a knife to my throat and telling me they will kill me unless I shoot you, I'm still making a moral choice if I do so).
I should also note that laws don't typically make any mention about whether you had a choice in your actions. A related but extremely different criterion, used in insanity defenses, is whether you understood the nature of your actions at the time of committing the act. This is obviously fully compatible with a deterministic universe: I can understand the consequences of my actions even if I am fully compelled by circumstance to do them. Conversely, a person who is suffering from severe hallucinations, even in a world with strong free will, might shoot you while believing, say, that they are choosing to shoot a water gun at a painting. They made an explicit choice by any definition, but it's not a morally evil act since they didn't understand what they were actually doing.
You say "choice", I say there is never a choice. There is no option, no decision to make, nothing you do is of your own volition, or your responsibility. You are indeed compelled to do everything you will ever do whether that's eating a bacon sandwich or drowning puppies in a bathtub. Understanding is irrelevant and perhaps much of "consciousness" is post-hoc justification.
I've brought it back to justice because it's a central pillar of human civilisation, and just looking at the legal system (though my thought is much broader) there are such concepts as "the age of criminal responsibility". What if we say that nobody is responsible for anything? Would we have got where we are now? It's a license to act in an entirely selfish (in the same sense as genes) way.
A decision is understood as any point where an agent takes a course of action based on: its perception of circumstances, its evaluation of available options, its evaluation of possible consequences of each option, and its given logic/values/morals. Nowhere in this kind of definition is there some expectation of a "free" choice (one that doesn't depend on the whole history of the agent so far).
You can see this definition in contexts like "decision trees" (which no one believes have consciousness or free will), and in common language about algorithms/programs (where people say the algorithm decides or chooses what to show you, for example - again with no illusions that the algorithm has free will).
To kill or imprison someone who had no choice about their actions would be unjust.
To simplify things, assume a serial killer that would kill again, or a thief that would continue stealing. Why is imprisoning this person worse than more people falling victim to this person. And I really want to keep the focus on someone repeatedly breaking the law, one off crimes are probably much more tricky to deal with. The choice here is not between imprisoning and not imprisoning a person for doing something they could not avoid, in case of not imprisoning that person you have to take into account future harm. Same here, who knows what people will do in the future, even in a deterministic universe, but let us pretend we know, we can make things more complicated with probabilities and whatnot later.
I never understood why "nobody could be blamed for their actions" is some kind of argument against determinism. Neither the perpetrator of the act nor the punisher can choose. Why does the punisher exist outside the deterministic system but the perpetrator does not? The perpetrator's logic is the same: if I do this, the punisher has no choice but to punish me. Everyone is along for the ride. The very desire to control the actions of others within some "moral" framework also exists within the system.
As danbruc said elsewhere in this thread:
Even in a deterministic universe you still have to deal with criminals in one way or another, and you have to make the rules and implement them
A moral framework that this happens (deterministically) under doesn't exist to undermine the deterministic nature of the criminals, but in direct deterministic consequence of the existence of criminals.
The state machine of the universe has led to my desire for a moral framework to punish criminals, the existence of criminals, the criminal acts, and the application of a moral framework to recognize and punish criminals. Under determinism, these are all intertwined.
I see it as free will exists (if it does by some definition) at the time a decision is made based on the inputs (the state of the universe) at that point. "Doing something different in the same situation" (having choice) is impossible because "the same situation" is impossible because the universe continues to run. Given the same input (the universe), why would you "choose differently"? You have no reason to. If output is not based on input, the output is arbitrary and random, which also undermines the existence of free will.
I'm not just talking about criminals, I'm talking about normal human interaction. Game theory. How would societies form if not for cooperation? How is cooperation ensured without feedback mechanisms to enforce it? This can all emerge from deterministic processes, same as many other things in nature.
In your final paragraph you seem to be saying (meaningful) free will doesn't exist? If so, I agree. The thing I really wonder about is why it seems that it does.
Oh, sorry, I'm not trying to contradict what you said, rather support and offer a (perhaps different) angle or detail. Dealing with criminals is just an example. The same applies for the desire to form or not form a society.
As for "why it it seems that [free will exists]", I want to say that a combination of the Incompleteness Theorem, the fact that we don't know enough to grok the entire state of the universe, that we don't know what the future holds, and that we have the ability to reflect upon the past that gives the "illusion" of choice. But even asking that question seems to assume that we are discussing the illusion of free will from a perspective _outside of the universe_. Just like we must assume the universe exists, we must also assume that free will _seems_ like it exists.
Dig into compatibilism to learn how these two concepts can coexist.
What you'll find is what's at the heart of most philosophical disagreement - a lack of rigorous definitions. In this case, you probably don't have a standalone definition of free will. If your definition is effectively a derivative of determinism, such as "free will = not determinism," then this is what binds you.
Disagree. I think compatibilism's definition of "free will" is something considerably less than what I consider to be free will. I think compatibilism weakens the definition of free will in order to make it compatible, but it's not really arguing what it sounds like it is.
There are many angles to this. The first one would of course be what you said, there would be no choice otherwise. I think in general a deterministic universe would not be too different from one with true free will, at least from the perspective of our current knowledge and capabilities. People would still make decisions - do one thing and not another - have ideas, talk about them, be influenced by others, just the ultimate root why things go one way and not another would be different.
Even in a deterministic universe you still have to deal with criminals in one way or another, and you have to make the rules and implement them, and for making the rules you have to think about the nature of the universe and what that means for morality. The real mystery for me would here be, why would a deterministic universe bring about a system of morality that takes the determinism of the universe into account?
Or more generally, why can you do anything useful in a deterministic universe? Why are your perception, your thoughts, and your words connected in a meaningful way? Why is it not the case that the universe is in state A, you perceive B, think C, and say D? There is one apple, you see two, think of three oranges, and say half a fork.
Back to the initial question, you could also consider that you are wrong, determinism might be false and you have free will after all. It would then be a bit of a waste to not use your capacity and instead throw your hands up because you - incorrectly - think that everything is set in stone anyway.
Moral culpability would seem to go out the window in a deterministic universe, no? Not in the sense that people shouldn't be held responsible for actions, but in the sense that retributive punishment becomes a much harder sell.
A murderous Roomba may warrant separation from other Roombas, or some sort of intervention which may incline it less towards murder, but we don't put it in a box or pry off its plastic for the mere sake of 'deserving it'.
At least in some countries we have not been locking people up for punishment for quite some time. You go to prison to protect the public, change your behavior, and deter others from committing crimes, not for punishment. The public perception, I would guess, has not quite kept up with this development, I think that prison is for punishment is still a widespread idea.
> we have not been locking people up for punishment for quite some time
Whether or not this is true seems kind of beside the point.
All I'm saying as that, for an arbitrary reader, a consequence of determinism is that the morals of "desert" are meaningfully impacted—whether that convinces them against retribution, or buttresses their existing notions against it.
There’s no reason to think that they’re incompatible in the first place. May as well ask it to synthesize a theory in which Linux and relational databases are compatible.
Determinism is a dead end that renders the human mind useless. It has its roots in the mechanistic cause-effect nature of Newtonian physics. Rationalist philosophers like Rene Descartes and Baruch Spinoza create rigid, math-like world-views believing in this. Later, David Hume would prove them wrong by demonstrating that causality is not a simple problem at all.
I don't even understand how flipping a coin to make a decision isn't the death of hard determinism. Or is it that random events can be inputs to determinism?
How is flipping a coin incompatible with a deterministic universe? Seems to me that it only seems random because our muscle control isn’t good enough to flip the side we want every time.
First of all, in a Newtonian model, flipping a coin is not a random event, the result of the coin flip is determined by the exact forces applied, which are determined by the muscle movements of the flipper, which are determined by chemical reactions in their brain etc.
Secondly, even true randomness doesn't significantly alter the world-view of Determinism, at least not when applied to the problem of consciousness (as in the current context). Regardless of whether the future state of the world is fully determined by the past state, or whether the future state of the world is fully determined by the past state + random events, consciousness and conscious decisions play no role.
That is, the feeling that you decide whether to go to work or not, that today you could have decided otherwise if you had wanted to, is still an illusion in either model. In pure determinism, this was determined the moment the universe came into being. In determinism+true randomness, whether you went or not was unknowable ahead of time, but you still didn't have any input: say, some cosmic ray flipping a switch in your brain or not determined whether you did go.
> First of all, in a Newtonian model, flipping a coin is not a random event, the result of the coin flip is determined by the exact forces applied, which are determined by the muscle movements of the flipper, which are determined by chemical reactions in their brain etc.
Nope.
In simple newtonian classical mechanics it is entirely possible to have a phase space in which for all epsilon greater than zero any initial conditions that lead to { heads } have a neighbouring set of intial conditions within epsilon radius that lead to { tails }.
ie. It is possible to have coin flipping game in which no amount of precision in setting intial conditions can guarantee the outcome.
This is the implication of Lorenz's butterfly and Smale's horseshoe map.
True, though I don't believe there is any reason to think such systems would actually be physically realizable. To me this seems similar to the dome shape where a ball sitting at rest on the top would randomly move down at any time in any direction - it is part of the mathematics of Newtonian mechanics, but it is not necessarily part of the physics of it (since that dome shape requires infinite smoothness, which can't exist in the physical model).
Newtonianism doesn't seem relevant to my question. It is a random event if you're not trying to make it not-random. I could just say the word "heads" and continue with my day based on the "heads" decision if I wanted to affect the outcome.
I'm saying: is randomness such as this an input to a deterministic model? This brain decides to flip a coin and base its next action on that coin's outcome: is that a future past which determinism cannot see?
> Newtonianism doesn't seem relevant to my question. It is a random event if you're not trying to make it not-random. I could just say the word "heads" and continue with my day based on the "heads" decision if I wanted to affect the outcome.
If we're discussing whether the future is fully determined by the past, this point is rather important. In a fully classical (newtonian) world, the future is fully determined by the past - so the result of that coin flip was already determined not just at the moment the coin was thrown, but in fact the moment the universe began.
Now, with QM, whether this remains true or not is much more debatable. In the many-worlds interpretation for example, it does remain true: in that interpretation, the whole universe evolves according to the Schrodinger equation, with all possible outcomes of a quantum measurement actually happening, just not in the same place in phase space. In the more common interpretations, there is true randomness, with only one outcome happening from a random event and no way to predict which it will be.
So, in the most common interpretations of QM, while coinflips may not be very random, you could instead measure the spin of an electron and base your decisions on that, and no one could predict the future if you do so.
Still, note that you are not choosing anything, in either interpretation. Your next action is as unpredictable to you as it is to anyone else, if there is true randomness. Your thoughts, beliefs, experiences, do not influence your actions in any way if you base your actions on a random event. So, if you believe that a deterministic universe is not compatible with true free will, then you should also believe that a random universe is just as incompatible with true free will.
Agree about coin flipping - I was thinking about the same thing after I posted. Better source of randomness required.
On your final paragraph - is what you're saying only true if you base all your decisions on the quantum mechanical equivalent of a coin flip? Or are you referring to something else?
First of all, I believe the majority of philosophers of mind today are compatibilists - they hold that free will is in fact compatible with determinism.
Secondly, there is very clearly no room at all for strong free will in any physical theory we have. By "strong free will" I mean the idea that conscious beings are able to decide their next action at some moment T+1 without either (a) a sufficiently intelligent agent being able to predict with 100% accuracy their decision from enough knowledge about the state of the world at moment T, or (b) the decision being a purely random event. Given that human beings are made entirely out of elementary particles, there is simply no way for some part of the human to influence the elementary interactions that make up our beings.
Of course, if you believe in dualism of some kind, where there is some form of a transcendent soul that doesn't obey physical theories but can intervene in them, then you can add this to any theory.
The way free will can be made compatible with determinism is to accept that our choices are not fundamentally different from the choices that a search algorithm makes in its operation. They are fully determined by our previous experience and environment (up to some randomness), but different people will still make different choices, more or less appropriate. Just like different implementations of a search algorithm can work better or worse, or just differently, so two can different people make better or worse choices, or simply different ones. But, just like the algorithm, we are still each entirely the product of our entire life so far, and our nature as human beings.
I think that pansychism and determinism are interesting theories, but it is not clear to me what I will learn about them by reading ChatGPT potentially hallucinate about them?
Arguing about determinism is boring and pointless. If determinism is true, then logic and reason has no bearing on anybody’s conclusion. It is only if determinism is false that logic and reason are in any way meaningful, and thus anything is worth arguing about.
Another way:
If determinism is true, then there is no point in arguing about it.
So:
If it is worth arguing about, then according to the laws of logic, it is false.
If it is not with arguing about, then why are you arguing about it.
Even if hard determinism is true then the act of 'communicating a logical argument' is still part of the deterministic chain of events and could potentially influence the beliefs and actions of others.
> If determinism is true, then logic and reason has no bearing on anybody’s conclusion.
That is obviously false. In a deterministic model, your beliefs are determined by all of the inputs you ever received. It is in fact stronger than a model of the world that includes strong free will: there, an agent can hear a perfectly valid argument that explains in all ways that the agent understands this that 1+1=2, but the agent can decide not to believe it for no reason whatsoever (since its decision would be independent from anything that they had ever heard or experienced or thought about before).
It is in fact pure free will that would make any kind of logic or reasoning useless, since, again, your decisions would not be in any way determined by your life up to the point of the decision.
Heh, was once able to get chatgpt to give an articulate explanation about how the single electron model of the universe would explain both 1. consciousness and 2. why elephant breasts look like human breasts.
"emergence" is a copout though. The answer was more along the lines of describing the electron as the primary interface between the instantaneous "now" and the future, and the breast being the interface between generations.
I did notice that it referred Spinoza as if his views were in opposition to Panpsychism.
Spinoza actually proposed that God(consciousness) exists within all things, and that studying physics and the nature of the universe is to study the psychology of God.
Like he's literally mentioned on the SEP page for Panpsychism [1]
I see where chatGPT would get confused. Spinoza is identified as a Rationalist, which is generally more aligned towards determinist thought.
Looking at it, I wonder if perhaps Spinoza's views are already in alignment with both Deterministic and Panpsychist thoughts (which, as others have mentioned, are not inherently mutually exclusive anyway).
What's written there seems not just correct but pretty obvious to me. I never insist I'm correct though, in this kind of matters expecially as there hardly is a way to check what can be only concluded from speculation and feeling.
If this is true, then universe is local and real and all of our observations are correlated with previous events and we will never be able to find out the truth of it through observation.
Although, if everything is unified/interconnected under the same consciousness and this manifests as non-local effects in the universe, then there's room for the universe not being deterministic.
Similarly, maybe determinism is a wrong way of looking at it. Maybe there is no true relationship between one moment and the next, and it is only your observation that sees the ordering of events. Maybe all of the events exist everywhere all at once and their ordering, progress, unfolding is just a flaw of the observation. How would an object like a photon perceive the universe and can it perceive given that photon never meets time?
This sounds clever but is content-free. The "new theory" that chatGPT synthesizes simply restates its assumptions.
Here's what it calls the "fundamental principles" of hard determinism and panpsychism:
* Hard Determinism: This is the belief that every event in the universe is determined by antecedent causes and conditions. There is no genuine free will; everything that occurs is the inevitable result of prior circumstances.
* Panpsychism: The idea that consciousness, or at least some basic form of it, is pervasive throughout the universe. Every entity, no matter how small, has some element of consciousness or experiential quality.
And here's the first two "foundational principles" of its new theory:
* Universality of Experience: The fabric of the universe is interwoven with rudimentary consciousness or experiential quality. This means everything from subatomic particles to galaxies possesses some level of conscious experience. In this view, consciousness isn't just a by-product of complex systems like the human brain but is instead a fundamental property of reality.
* Chain of Determined Consciousness: Every event, interaction, or change that takes place in the universe is pre-determined by prior conditions. Each of these events involves conscious entities (be it elementary particles, cells, animals, or galaxies) experiencing the unfolding of the deterministic chain. Each particle, while following a pre-determined path, experiences its journey in its own unique way.
But these are just the first two points in different language! The only new thing is that chatGPT says the particles are conscious while they pursue their predetermined path.
This is like saying "my new theory of light explains how it can be both a wave and a particle. Principle 1: light is composed of waves. Principle 2: light is simultaneously composed of particles! Where's my Nobel?"
It doesn't really matter whether the universe is determinstic or not, because even if it were, it would be impossible to build anything to exploit that determinism on any useful scale, and so the future remains unknowable. You can think of the universe as a giant computer calculating its own future one picosecond at a time. There is no way to 'jump ahead' and find out the future. Hence it doesn't matter whether it's determinstic or not. And probably never be possible to tell anyway. And anyway if it's not determinstic, what is it instead, Random? How is that any better or any different?
How does free will approximate in this metaphysics? Imagine your skull as a boundary between whats inside it and whats outside it - what you do next is mostly to do with whats happening inside your skull rather than outside it. Apart from certain scenarios like freefall, skydiving or something. And yes the workings of your brain have been determined by things outside your skull over time - your experiences and memories and genetic inheritance and so on. But the proximate cause of what you do next is mostly whats inside your skull. So lets just call that agency or free will and be done with it.
93 comments
[ 3.0 ms ] story [ 193 ms ] threadIs it really? I don't think compatibility with a crackpot idea/thing brings a blow to an orthogonal idea.
For example, I could believe that Mars is an alien base. This is crackpot, but it doesn't give any blow to countless other ideas it's compatible with, e.g. economic or social theories, gravity, explanations of historical eras, etc.
Where could that line possibly be?
I personally cannot reasonably even guess at such a boundary, so panpsychism is a much cleaner explanation.
These two ideas aren’t really mutually exclusive anyway. I.e. you can believe that consciousness is what it feels like to compute, but then “what is compute” except a particularly structured form of a thing that’s happening everywhere and all the time (particles and energy bouncing around) in more or less structured forms. And if you say “well not all energy/particle interactions are computation, even of an abysmally low degree,” then return to the first question: what types of interactions or what degrees of structure are non-computational where one rung above it they are computational?
It's not a gradualist notion. A computation either is or isn't consciousness. If it is, then an agent which is currently performing that computation is conscious. If an agent is performing a computation that is not consciousness is not conscious.
Similarly, certain computations represent a sort. Agents which are performing that computation are sorting. The clearest case is a computer running a sort algorithm, but there are also purely physical processes that perform the same computations, and animals or people that do it - so "agent" here is a very broad term. Agents which are performing a computation that is not a sort are not sorting.
Of course, determining which computation is actually consciousness is a very difficult problem, well beyond our current knowledge. As such, it is also obvious that it may turn out that this is not a correct model. But I don't see why this would be considered a ridiculous model of what consciousness means. It is also a model which can, in principle, be falsified, and it could actually help resolve certain difficult real-world problems. For example, if we did have a satisfying definition of which computations represent consciousness, we could then use it to determine at which point a child becomes conscious, or whether a particular animal is conscious or not (assuming we consider that conscious === should not be killed arbitrarily, which is a completely different debate).
By contrast, panpsychism doesn't really resolve anything and can't be used to determine the answer to this type of moral question even in principle (if any system is somewhat conscious, then even swiveling my chair means that I am destroying a conscious system; but also creating another one).
Even your “non-gradualist” explanation does in fact run into gradualist problems, as you mention one of when a child “becomes conscious.”
You assert that this theory is falsifiable: how so? You can’t even falsify the notion that there’s only one consciousness at all in the entire universe (yours, the one you’re “doing”).
Regarding moral implications, inconvenient moral conclusions are not the same as “no help in arriving at moral conclusions.” You just named one such conclusion that you’d be compelled to arrive at (I wouldn’t, personally), so it’s not really true it doesn’t help determine the answer.
A computation is a physical process. A pendulum swinging is a computation of the path that the pendulum will take (and of many other things). Whether there is some conscious being that observes and understands that computation or not is irrelevant to whether the computation is happening (that is, it won't affect the result in any way).
And yes, I am saying that one possible theory of consciousness is that some particular arrangement of matter and energy evolving in a certain way over time is consciousness. That is what I mean by a computational theory of consciousness. I am not in any way claiming I could define today which particular arrangement and evolution that is - I explicitly said I believe the specific model is far beyond our current understanding of the subject.
> Even your “non-gradualist” explanation does in fact run into gradualist problems, as you mention one of when a child “becomes conscious.”
I wouldn't call a theory that tells you that a system is either conscious or unconscious "gradualist". Yes, the system is not conscious at time A and becomes conscious at time A+1. This happens exactly the same way as any other computation: my processor is currently not performing a sort, and when it reaches the sort() call, it becomes sorting all of a sudden; when the call to sort() finishes, it is no longer sorting. Similarly, we can imagine a procedure consciousness(); when a system starts executing this procedure, by definition it becomes conscious; when it finishes executing it, it becomes not conscious.
> You assert that this theory is falsifiable: how so? You can’t even falsify the notion that there’s more than one consciousness at all in the entire universe (yours, the one you’re “doing”).
Well, it is falsifiable in two ways. First, there is the same extent as any mathematical model is falsifiable: we can check whether it accurately accounts for our intuitions. It's the same way we can check whether the formal definition of a continuous function is correct.
Now, once we accept that the proposed model matches what we intuitively mean by consciousness, it is further just as falsifiable as any other physical theory - it makes it so that the question "is system A conscious" (within the formal definition of consciousness I defined above) becomes a measurable claim about whether a particular system is performing a particular computation.
> Regarding moral implications, inconvenient moral conclusions are not the same as “no help in arriving at moral conclusions.” You just named one such conclusion that you’d be compelled to arrive at (I wouldn’t, personally), so it’s not really true it doesn’t help determine the answer.
If the conclusion is "any act you make is just as morally evil, since it leads to destroying a conscious system", that's not just inconvenient, it is by definition not helpful in making moral choices. It is as useless for arriving at moral conclusions as a very convenient model which says "no entity is conscious, so any choice you make is morally good".
Alternatively, if you believe the consciousness of a subject is irrelevant to moral choices, then any model of consciousness is also not helpful in arriving at moral conclusions.
And, if you believe that the amount of consciousness is the relevant bit for moral choices about a system, than still panpsychism is no help, since it only asserts that all things are conscious, it doesn't promise a model for determining which are more conscious and which are less conscious (you would need an additional model on top of panpsychism, p...
A respectable philosophical idea with thousands of years behind it, and new impetus the past couple of decades.
Always best to have at least a basic understanding of a subject before using such dismissive terms.
You can read about it here:
https://plato.stanford.edu/entries/panpsychism/
but that is essentially The Hard Problem.
I know that I am conscious and I cannot doubt it, but at the same time I cannot prove it to you. Furthermore, I strongly suspect and assume that other human beings are conscious, but I cannot verify this scientifically (i.e. measure it somehow).
This means that consciousness is a phenomenon that cannot be doubted (in the first person) but possibly eludes the scientific method. This really grates people who have a strong emotional attachment to the dogma that science must be able to account for all of reality.
People confuse the metaphysical commitment of materialism with scientific fact, and thus conclude that consciousness must be an emergent phenomenon of neural complexity, and that it is silly to believe that anything else could be conscious. This feels like an hyper-rational, scientific stance, but really it is not. Science works perfectly fine within other metaphysical commitments such as idealism, panpsychism, etc.
This is not to say that materialism is not correct, I do not know if it is or not. My point is that many contemporary "scientific types" tend to believe that their certainty about materialism being the correct framework is grounded on science, when it is not. It is often an unexamined and quasi-religious belief.
Therefore, what we call subjective experience is evidence for the existence of what we experience, but until we can empirically show that there is a difference between what we experience and what reality is, the only thing we can say is that it exists. Also, that's the problem with the hard problem of consciousness - it assumes a division between the two, and we have no evidence of such a division. And lastly if it's not clear i am not a panpsychist either because that assumes a division on some level again. We don't know what existence itself is and how it functions really, at least not the brains function within existence.
The division is between our third-person descriptions of reality in science, and our first-person "experience". You can't logically derive a first-person perspective only from third-person facts. That's what's at the heart of the hard problem. The eliminativist simply says that the first-person facts are a cognitive illusion, and so we really only have third-person facts, and the panpsychist says all third-person facts have an accompanying first-person component most of which we can't access.
I don't think this is fully true, if you really examine this belief. You do have certain criteria that you use to determine, at least quasi-scientifically, if you should believe that some other entity is conscious or not. You say you believe "humans beings are conscious", but I would bet you don't believe that dead human beings are conscious, or that fertilized eggs are conscious. You may or may not believe that they are still conscious while under general anaesthesia, or while they are in clinical death. You may very much believe that certain things that are not human beings are also conscious - maybe dogs, or even all mammals, or even all vertebrates.
Your criteria probably start from the initial premise: I am conscious. Then, you apply a simple argument: I perform certain actions and/or display certain behaviors, and I have this conscious motivation for them. This other being performs these behaviors very similarly, so they probably have a similar consciousness motivating them to do so. You then test this assumption, and come up with more refined observations (e.g. not every similarity is equally related to consciousness - for example, you move around, and shadows move around, but it seems unlikely that shadows have consciousness so perhaps "move around" is not a good test for consciousness).
If something has (a) a robust temporal-spatial sense, a "reference frame" in time and space, (b) physical boundedness, (c) a robust and general predictive function which is capable of abstraction, (d) independent volition, and (e) behavior that is non-deterministic in the weak sense that its actions cannot easily be foreseen by any arbitrarily powerful computer within the universe that it resides in, then: That thing is conscious.
It's possible to imagine entities that don't necessarily satisfy all conditions and are nevertheless somehow conscious -- i.e., a sufficiently powerful AI that has no clear physical bounds and a poor (but not entirely absent) spatial reference frame, but which possesses independent volition and can independently simulate concepts -- but if all conditions are clearly met, then an entity is unambiguously conscious.
Personally I tried to tease novel techniques for compression out of it. It got me curious and at least it forced me to learn more about the domain to explain why its ideas were wrong.
Maybe that's a thing: it's so confidently (sometimes subtly) incorrect that it may take effort to learn how it's wrong. This itself can be an interesting exercise that wasn't really possible before.
1. Moral Responsibility: The concept of moral responsibility would need to be re-evaluated. [...]
You are not going to reevaluate anything in a deterministic universe.
People get so hopelessly confused about determinism and minds...
Even if the universe is deterministic, it doesn’t feel deterministic, and determinism doesn’t help me with ethics or morality. It reminds me of The Stranger.
It does. When we look at criminals we nowadays take into account medical and psychological conditions, we are aware that people are not always in full control of what they are doing. This is obviously a weaker form than full blown determinism, but we certainly take into account that people are not completely free in their actions at all times. Similarly the assumption of a deterministic universe will influence ethics and morality, even though it will of course not be the only thing under consideration. [1]
[1] Ignoring the big complication that talking about considerations and things like that in a deterministic universe is itself something that needs a lot of care.
But that doesn't change anything one way or the other. The efficacy of policies can be determined by measuring the outcomes. Policies which reduce crime / recidivism rates are good; policies that don't are bad. Whether or not anybody has free will makes no difference, the outcomes are what they are whether or not people have free will.
If raising the DUI fine 10x causes a drop in DUI rates, then that's a good policy even if people don't have free will. And if it has no effect on DUI rates, then it's an ineffectual policy even if people have free will. Free will is irrelevant, measurable outcomes are what matters.
Killing all people will reduce the crime and recidivism rates to zero. At the very least you want some more nuance.
The efficacy of policies can be determined by measuring the outcomes.
Killing a healthy person and taking two of its organs to save two terminally ill people waiting for an organ transplant will increase the number of years lived. Is outcome really all that matters?
Breaking the window of a car because you like breaking stuff and breaking the window of a car to unsuccessfully save a dog left behind in the car on a hot day, are they really morally the same because the outcome is the same? Yes, you talked about policies and not individual actions, but with some more thinking I could certainly find a good example for policies.
My assertion, the point I'm making, is that at no point does the existence or nonexistence of free will factor into the process of determining which laws are worth having. The way people feel about laws does factor into determining what laws are good or bad, but that is true whether or not people's feelings are deterministic or derived from free will.
The other piece is more the implementation and enforcement of the rules, how do we deal with people violating the rules. Here I think determinism or not has to influence the way we set things up. With determinism you can not really justify putting people into small prison cells, they should instead be allowed to live a normal live except from being prevented to come into situation in which there bad behavior would kick in again.
So depending on what exactly we call the law - whether only the desired or undesired behavior or also including policies around implementation and enforcement - I could agree to both, determinism plays [almost] no role or it does.
Minds are implemented within physics, but human concepts w.r.t. minds (e.g. responsibility, choice, freedom..) are only valid on the level of the minds, without any reference to the underlying physics. Any time people use physics to make judgments about morality they are just hopelessly confusing themselves by switching between "standard" physics and dualist metaphysics without noticing.
I agree that the concepts we use to deal with human minds are higher level concepts. I am somewhat hesitant to call them emergent because that does not seem to quite fit the common usage of the term, but if hard pressed, I would probably call them emergent at the end of the day, just in a more complex way than the color of things or the wetness of water emerges. I would also agree that some of the concepts are in one way or another substrate independent, i.e. they are not strongly tied to the underlying physics, i.e. you could build the same concept on top of other physics or maybe oven on top of something completely different, say mathematical set theory.
But...I don't think it valid to think of physics and minds as completely separable. If physics is deterministic, you do not get to introduce a fundamentally nondeterministic decision process at the level of minds. You can have one that seems nondeterministic to the mind because it lacks knowledge of all the positions and momenta of all the particles, but it would not be truly nondeterministic. You can also talk about a fictitious nondeterministic decision process that is just not realizable in the deterministic universe.
So I think there are at least two classes of human level concepts, those that are independent of the underlying physics and those that are not. And some times there will be overlap, prime example free will. Even in a deterministic universe you can have free will in one of the compatibilistic senses, say doing what you want to do, but you will neither get to determine what you want do, nor what you do or whether the two are the same.
If we admit nobody can be blamed for their actions, we can't punish them and thus the incentives not to do bad things disappear and the whole shebang collapses. Just cause and effect.
That we exist as we do is a result of evolution selecting for perception of free will.
That's my theory, anyway.
(Make of my username what you will.)
Does Policy X reduce Crime Y? If yes, then it doesn't matter if people have free will or not, Policy X is effective. If no, then it doesn't matter if people have free will or not, Policy X is not effective. Either way the cookie crumbles, free will doesn't matter. So the question of free will is quite literally irrelevant.
That's not true. You can put people behind bars, say for example murderers, without blaming them for what did but in order to protect others. Even today we will, for example, lock people with dangerous mental illnesses up in order to protect others from them, even if we know that what they did, they did because of some condition outside of their control.
Not at all. People seek to stop fires in order to protect others, but few believe fire carries blame for burning people. Of course, you need to believe that there is determinism (the fire is causing the burns, so stopping the fire will stop the burns), but that's it.
> And then you still have to account for whether the transgressor was coerced rather than was acting from their beliefs and desires (aka made a free choice).
This is still just a question about causality. The problem you are trying to solve is that people are getting hurt. You need to determine the ultimate cause for that in order to effectively solve it. So, if A is killing people that you want to live, you will want to first stop A from doing so. Next, you will have to determine what was causing A to kill people - if A was doing so because B paid for it, then you probably need to stop B as well: otherwise B may pay C to kill people even if A is in jail. If this has been determined, A may or may not need to be jailed at all, based on beliefs about whether A may still continue to kill people even if B has been stopped.
None of this requires some metaphysical notion of blame, it only requires a belief in determinism and a model of causality of behavior. A computer could be built that would take such decisions in a purely deterministic way, and using only information about purely physical facts (A performed act x, B performed act y, other people that share characteristics d, e, f with A performed act x, etc).
Because fire doesn't have intentions, beliefs and mental states that respond to feedback about right/wrong, and that govern its future behaviour.
> he problem you are trying to solve is that people are getting hurt. You need to determine the ultimate cause for that in order to effectively solve it. So, if A is killing people that you want to live, you will want to first stop A from doing so. Next, you will have to determine what was causing A to kill people - if A was doing so because B paid for it, then you probably need to stop B as well: otherwise B may pay C to kill people even if A is in jail. If this has been determined, A may or may not need to be jailed at all, based on beliefs about whether A may still continue to kill people even if B has been stopped. None of this requires some metaphysical notion of blame
It's frankly amazing that you can perform an in depth analysis of how to ascertain who is to blame, and yet conclude that blame is not required. You're performing this elaborate dance trying to avoid the word "blame" despite the fact that this analysis is exactly what everybody means by "blame". There's nothing metaphysical about it.
The key difference is that fire is not considered a moral agent and a person is. That's why the justification is required.
I also assert, with evidence [1], that this legal understanding of free will is how most people think about it. I can't make any sense of your argument unless you're employing a different definition of free will.
[1] https://philarchive.org/rec/ANDWCI-3
There are of course separate laws for protecting humans from other humans, and that is typically where questions of liability and intent etc arise. But I would still argue that these still have nothing to do with beliefs about (strong/non-compatibilist) free will (though of course they historically are related to free-will-based religious conceptions). For one thing, these laws are explicitly targeted at human persons. There is no aspect in the law that would make it apply to a chimp, for example, even if the chimp were determined to have free will just as much as a human. You can't sue a chimp, even if you believe that it is a moral agent: you can only sue a human person. You also can't sue any type of AI agent, and would not be able to without a change in the law.
So I would say that existing regulations which apply to non-human beings are also proof that we are perfectly capable of creating laws that don't rely on beliefs about the free will of the entities which these laws govern. The fact that current laws for humans make use of free will terminology is thus incidental, not fundamental. All current legal principles could be re-stated without requiring free will as well - this is what I was trying to claim (though I did claim more than that, for which I was wrong).
I will also note that this thread was started by someone who was claiming that you couldn't use the current legal system if you believed the universe is deterministic. The paper you cite works against this argument: it claims that the common notion of free will that most people apply is in fact compatible with a deterministic universe. If you also agree with this paper, then we are probably mostly in violent agreement overall, and in disagreement with sanitycheck's original claim.
Well, this will quickly turn into the nuances of definitions of words. Blame can have several subtilely different meanings, so call it blame if you want. I will avoid that problem by saying that we go from asserting that a person did something wrong and could and should have done otherwise to asserting that a person did something wrong.
"Protecting others" carries an implicit assertion that failing to abide by this principle means you are broken [...]
Yes, if we have the principle that one person should not harm another and you do not follow that principle, then in a deterministic universe we would say that you are not behaving as expected, that you are broken if you want. Well, there is probably some nuance required here, are you really broken and will keep harming others or was this more like an exceptional situation making you misbehave once but unlikely to do so again?
"[...] rather than that the principle itself is mistaken, as would be true if you really believed blame didn't exist.
Take the example of some manufacturing robots, if one of them breaks and keeps crashing into others, then you will remove and replace the broken one, you are not going to blame it or assign any responsibility to it, because one of the wires connecting a sensor got lose. I do not see why enforcing a policy, categorizing behavior into acceptable and unacceptable behavior, requires blame in any form. We just categorize, we do not even have to look why something ended up in one category or the other. A related but different question is what our policy is, where the line between acceptable and unacceptable behavior lies.
And then you still have to account for whether the transgressor was coerced rather than was acting from their beliefs and desires (aka made a free choice).
That falls under the point mentioned above, more nuance, whether the behavior was situation specific or is likely inherent. But that is again independent of free will - whether I am determined to do bad things or freely decide to do bad things, one should try to prevent me from doing bad things.
And it matters even less in case of coercion, or maybe it is even more complicated in the case of free will. There is not really anything to be said about deterministic coercion to do a bad thing, but if you have free will, maybe you should have resisted more? This will not always apply, but at least sometimes you will have a choice to resist the coercion and probably experience some negative consequences in return, or to go along a cause negative consequences for others.
These statements seem equivalent. To say someone did something wrong is to say they should have done something else.
> then you will remove and replace the broken one, you are not going to blame it or assign any responsibility to it
Yes, you are literally saying this robot is to blame for the problem. The only difference with intelligent beings is that they can understand what they did wrong if they are given feedback, and thus change their future behaviour, and that's why holding them responsible requires a different kind of action than holding a dumb robot responsible.
When AI advances to the point that the robot too can understand this, then it would make sense to also hold it responsible in a similar way to humans.
Note that responsibility and blame don't necessarily entail punishment. This is a common mistake. In many prior discussions, I've found people jumping to hard determinism rather that Compatibilism simply because they don't like punishment, but nothing in Compatibilism necessarily requires punishment.
They important word is could not should.
Yes, you are literally saying this robot is to blame for the problem.
I am saying the robot caused the issue - for which you can use the word blame if you want to - not that the robot is responsible or could have done otherwise - for which I would reserve the word blame most of the time - which he could not as the issue was caused by a lose cable.
[...] than holding a dumb robot responsible.
I don't think you can hold a robot responsible without stretching the definition of holding responsible quite far.
I disagree emphatically with the inclusion of "could" in the sense that you are using it. When a mother scolds her child for misbehaving, she's saying the child should have acted otherwise, not that the child could have chosen differently if you rewound time and presented the child with the same choice again.
When laypeople agree that someone "could" have done otherwise, they're not agreeing with nondeterminism, they mean that any mentally competent person with common knowledge would have chosen differently given the same circumstances, ie. that there was no mental, health-related or coercive element that was forcing them to choose wrongly.
And this should be obvious if you think about it: we have centuries of legal precedent based on exactly this reasoning, and most people find it perfectly reasonable. That's why free will in law will never be overturned, despite what some hard determinists think. Whether people are deterministic is simply irrelevant to such judgments.
> not that the robot is responsible or could have done otherwise
Frankfurt debunked the principle of alternate possibilities' relevance to responsibility more than 40 years ago. I don't know why people still think it's relevant.
> I don't think you can hold a robot responsible without stretching the definition of holding responsible quite far.
This sort of exchange is not at all uncommon:
Frank: what was responsible for the server downtime?
Alice: an update messed up the routing tables on our routers.
The intended meaning is very clear: use "who" for blameworthy agents, and "what" for blameworthy non-agents. You can of course say this sort of informal dialogue is imprecise and not in line with formal definitions, but nobody would find it a stretch or confusing.
Even if everything is deterministic, human society can only function successfully with the illusion of free will. If we admit nobody can be blamed for their actions, we can't punish them and thus the incentives not to do bad things disappear and the whole shebang collapses.
If you kill people, we will lock you up, because we think killing people is bad and we want to prevent you from doing it again. You could just implement and enforce [1] that rule, period. Doesn't even sound unreasonable to me, not even without further elaboration. So when you are saying
If we admit nobody can be blamed for their actions, we can't punish them [...]
then this can at most mean that we can not punish them with a reasonable justification, we obviously can just punish them. One thing another comment reminded me of is that we are no longer punishing people for what they did, at least not in the sense of some kind of revenge. Anyway, if free will or blame or whatever plays any role, then only for building a reasonable justification for our rules.
[1] Ignoring for the moment that you could not actually decide to implement or enforce a rule like that in a deterministic universe - it would either appear out of the previous state at some time or not and enforcement would happen or not.
That is not the problem under discussion. The question actually sits in the definition of "choice". You seem to believe that something constitutes a (moral) choice only if the agent could do it or not without regard to its entire life of experiences up until that point. The more common notion of (moral) choice is that it constitutes any act which an agent takes without being coerced completely into it by some other agent. If my whole life up to this moment caused me to shoot you, that was a moral choice I made. If a dog bites my hand and pushes my finger into the trigger causing me to shoot you, then I didn't make a moral choice to shoot you (however, if someone is holding a knife to my throat and telling me they will kill me unless I shoot you, I'm still making a moral choice if I do so).
I should also note that laws don't typically make any mention about whether you had a choice in your actions. A related but extremely different criterion, used in insanity defenses, is whether you understood the nature of your actions at the time of committing the act. This is obviously fully compatible with a deterministic universe: I can understand the consequences of my actions even if I am fully compelled by circumstance to do them. Conversely, a person who is suffering from severe hallucinations, even in a world with strong free will, might shoot you while believing, say, that they are choosing to shoot a water gun at a painting. They made an explicit choice by any definition, but it's not a morally evil act since they didn't understand what they were actually doing.
I've brought it back to justice because it's a central pillar of human civilisation, and just looking at the legal system (though my thought is much broader) there are such concepts as "the age of criminal responsibility". What if we say that nobody is responsible for anything? Would we have got where we are now? It's a license to act in an entirely selfish (in the same sense as genes) way.
You can see this definition in contexts like "decision trees" (which no one believes have consciousness or free will), and in common language about algorithms/programs (where people say the algorithm decides or chooses what to show you, for example - again with no illusions that the algorithm has free will).
To simplify things, assume a serial killer that would kill again, or a thief that would continue stealing. Why is imprisoning this person worse than more people falling victim to this person. And I really want to keep the focus on someone repeatedly breaking the law, one off crimes are probably much more tricky to deal with. The choice here is not between imprisoning and not imprisoning a person for doing something they could not avoid, in case of not imprisoning that person you have to take into account future harm. Same here, who knows what people will do in the future, even in a deterministic universe, but let us pretend we know, we can make things more complicated with probabilities and whatnot later.
As danbruc said elsewhere in this thread:
Even in a deterministic universe you still have to deal with criminals in one way or another, and you have to make the rules and implement them
A moral framework that this happens (deterministically) under doesn't exist to undermine the deterministic nature of the criminals, but in direct deterministic consequence of the existence of criminals.
The state machine of the universe has led to my desire for a moral framework to punish criminals, the existence of criminals, the criminal acts, and the application of a moral framework to recognize and punish criminals. Under determinism, these are all intertwined.
I see it as free will exists (if it does by some definition) at the time a decision is made based on the inputs (the state of the universe) at that point. "Doing something different in the same situation" (having choice) is impossible because "the same situation" is impossible because the universe continues to run. Given the same input (the universe), why would you "choose differently"? You have no reason to. If output is not based on input, the output is arbitrary and random, which also undermines the existence of free will.
In your final paragraph you seem to be saying (meaningful) free will doesn't exist? If so, I agree. The thing I really wonder about is why it seems that it does.
As for "why it it seems that [free will exists]", I want to say that a combination of the Incompleteness Theorem, the fact that we don't know enough to grok the entire state of the universe, that we don't know what the future holds, and that we have the ability to reflect upon the past that gives the "illusion" of choice. But even asking that question seems to assume that we are discussing the illusion of free will from a perspective _outside of the universe_. Just like we must assume the universe exists, we must also assume that free will _seems_ like it exists.
What you'll find is what's at the heart of most philosophical disagreement - a lack of rigorous definitions. In this case, you probably don't have a standalone definition of free will. If your definition is effectively a derivative of determinism, such as "free will = not determinism," then this is what binds you.
https://en.m.wikipedia.org/wiki/Compatibilism
Even in a deterministic universe you still have to deal with criminals in one way or another, and you have to make the rules and implement them, and for making the rules you have to think about the nature of the universe and what that means for morality. The real mystery for me would here be, why would a deterministic universe bring about a system of morality that takes the determinism of the universe into account?
Or more generally, why can you do anything useful in a deterministic universe? Why are your perception, your thoughts, and your words connected in a meaningful way? Why is it not the case that the universe is in state A, you perceive B, think C, and say D? There is one apple, you see two, think of three oranges, and say half a fork.
Back to the initial question, you could also consider that you are wrong, determinism might be false and you have free will after all. It would then be a bit of a waste to not use your capacity and instead throw your hands up because you - incorrectly - think that everything is set in stone anyway.
A murderous Roomba may warrant separation from other Roombas, or some sort of intervention which may incline it less towards murder, but we don't put it in a box or pry off its plastic for the mere sake of 'deserving it'.
Whether or not this is true seems kind of beside the point.
All I'm saying as that, for an arbitrary reader, a consequence of determinism is that the morals of "desert" are meaningfully impacted—whether that convinces them against retribution, or buttresses their existing notions against it.
Secondly, even true randomness doesn't significantly alter the world-view of Determinism, at least not when applied to the problem of consciousness (as in the current context). Regardless of whether the future state of the world is fully determined by the past state, or whether the future state of the world is fully determined by the past state + random events, consciousness and conscious decisions play no role.
That is, the feeling that you decide whether to go to work or not, that today you could have decided otherwise if you had wanted to, is still an illusion in either model. In pure determinism, this was determined the moment the universe came into being. In determinism+true randomness, whether you went or not was unknowable ahead of time, but you still didn't have any input: say, some cosmic ray flipping a switch in your brain or not determined whether you did go.
Nope.
In simple newtonian classical mechanics it is entirely possible to have a phase space in which for all epsilon greater than zero any initial conditions that lead to { heads } have a neighbouring set of intial conditions within epsilon radius that lead to { tails }.
ie. It is possible to have coin flipping game in which no amount of precision in setting intial conditions can guarantee the outcome.
This is the implication of Lorenz's butterfly and Smale's horseshoe map.
https://en.wikipedia.org/wiki/Lorenz_system
https://en.wikipedia.org/wiki/Horseshoe_map
https://en.wikipedia.org/wiki/Phase_space
I'm saying: is randomness such as this an input to a deterministic model? This brain decides to flip a coin and base its next action on that coin's outcome: is that a future past which determinism cannot see?
If we're discussing whether the future is fully determined by the past, this point is rather important. In a fully classical (newtonian) world, the future is fully determined by the past - so the result of that coin flip was already determined not just at the moment the coin was thrown, but in fact the moment the universe began.
Now, with QM, whether this remains true or not is much more debatable. In the many-worlds interpretation for example, it does remain true: in that interpretation, the whole universe evolves according to the Schrodinger equation, with all possible outcomes of a quantum measurement actually happening, just not in the same place in phase space. In the more common interpretations, there is true randomness, with only one outcome happening from a random event and no way to predict which it will be.
So, in the most common interpretations of QM, while coinflips may not be very random, you could instead measure the spin of an electron and base your decisions on that, and no one could predict the future if you do so.
Still, note that you are not choosing anything, in either interpretation. Your next action is as unpredictable to you as it is to anyone else, if there is true randomness. Your thoughts, beliefs, experiences, do not influence your actions in any way if you base your actions on a random event. So, if you believe that a deterministic universe is not compatible with true free will, then you should also believe that a random universe is just as incompatible with true free will.
On your final paragraph - is what you're saying only true if you base all your decisions on the quantum mechanical equivalent of a coin flip? Or are you referring to something else?
Secondly, there is very clearly no room at all for strong free will in any physical theory we have. By "strong free will" I mean the idea that conscious beings are able to decide their next action at some moment T+1 without either (a) a sufficiently intelligent agent being able to predict with 100% accuracy their decision from enough knowledge about the state of the world at moment T, or (b) the decision being a purely random event. Given that human beings are made entirely out of elementary particles, there is simply no way for some part of the human to influence the elementary interactions that make up our beings.
Of course, if you believe in dualism of some kind, where there is some form of a transcendent soul that doesn't obey physical theories but can intervene in them, then you can add this to any theory.
The way free will can be made compatible with determinism is to accept that our choices are not fundamentally different from the choices that a search algorithm makes in its operation. They are fully determined by our previous experience and environment (up to some randomness), but different people will still make different choices, more or less appropriate. Just like different implementations of a search algorithm can work better or worse, or just differently, so two can different people make better or worse choices, or simply different ones. But, just like the algorithm, we are still each entirely the product of our entire life so far, and our nature as human beings.
Another way:
If determinism is true, then there is no point in arguing about it.
So:
If it is worth arguing about, then according to the laws of logic, it is false.
If it is not with arguing about, then why are you arguing about it.
That is obviously false. In a deterministic model, your beliefs are determined by all of the inputs you ever received. It is in fact stronger than a model of the world that includes strong free will: there, an agent can hear a perfectly valid argument that explains in all ways that the agent understands this that 1+1=2, but the agent can decide not to believe it for no reason whatsoever (since its decision would be independent from anything that they had ever heard or experienced or thought about before).
It is in fact pure free will that would make any kind of logic or reasoning useless, since, again, your decisions would not be in any way determined by your life up to the point of the decision.
It's such a good bullshitter.
Seems pretty reasonable to me
I did notice that it referred Spinoza as if his views were in opposition to Panpsychism.
Spinoza actually proposed that God(consciousness) exists within all things, and that studying physics and the nature of the universe is to study the psychology of God.
Like he's literally mentioned on the SEP page for Panpsychism [1]
I see where chatGPT would get confused. Spinoza is identified as a Rationalist, which is generally more aligned towards determinist thought.
Looking at it, I wonder if perhaps Spinoza's views are already in alignment with both Deterministic and Panpsychist thoughts (which, as others have mentioned, are not inherently mutually exclusive anyway).
[1]https://plato.stanford.edu/entries/panpsychism/
Although, if everything is unified/interconnected under the same consciousness and this manifests as non-local effects in the universe, then there's room for the universe not being deterministic.
Similarly, maybe determinism is a wrong way of looking at it. Maybe there is no true relationship between one moment and the next, and it is only your observation that sees the ordering of events. Maybe all of the events exist everywhere all at once and their ordering, progress, unfolding is just a flaw of the observation. How would an object like a photon perceive the universe and can it perceive given that photon never meets time?
Here's what it calls the "fundamental principles" of hard determinism and panpsychism:
* Hard Determinism: This is the belief that every event in the universe is determined by antecedent causes and conditions. There is no genuine free will; everything that occurs is the inevitable result of prior circumstances.
* Panpsychism: The idea that consciousness, or at least some basic form of it, is pervasive throughout the universe. Every entity, no matter how small, has some element of consciousness or experiential quality.
And here's the first two "foundational principles" of its new theory:
* Universality of Experience: The fabric of the universe is interwoven with rudimentary consciousness or experiential quality. This means everything from subatomic particles to galaxies possesses some level of conscious experience. In this view, consciousness isn't just a by-product of complex systems like the human brain but is instead a fundamental property of reality.
* Chain of Determined Consciousness: Every event, interaction, or change that takes place in the universe is pre-determined by prior conditions. Each of these events involves conscious entities (be it elementary particles, cells, animals, or galaxies) experiencing the unfolding of the deterministic chain. Each particle, while following a pre-determined path, experiences its journey in its own unique way.
But these are just the first two points in different language! The only new thing is that chatGPT says the particles are conscious while they pursue their predetermined path.
This is like saying "my new theory of light explains how it can be both a wave and a particle. Principle 1: light is composed of waves. Principle 2: light is simultaneously composed of particles! Where's my Nobel?"
How does free will approximate in this metaphysics? Imagine your skull as a boundary between whats inside it and whats outside it - what you do next is mostly to do with whats happening inside your skull rather than outside it. Apart from certain scenarios like freefall, skydiving or something. And yes the workings of your brain have been determined by things outside your skull over time - your experiences and memories and genetic inheritance and so on. But the proximate cause of what you do next is mostly whats inside your skull. So lets just call that agency or free will and be done with it.