Approaching philosophy as science is not new. It has had a few spectacular successes, such as the wholesale transfer of cosmology from science to philosophy, and a lot of failures, judging by the long list of unanswered philosophical questions (about 200, according to wikipedia). It also has the special pitfall of philosophically uninformed scientists answering the wrong question:-
How to determine causality from observational data;
What causality is is the correct question/.
where the perception that humans have free will comes from;
Whether humans have the power of free will is the correct question.
where human moral intuitions come from.
Whether human moral intuitions are correct is the correct question.
Oh, if you count that one as a question, then let’s call that one solved too.
Whether humans have the power of free will is the correct question.
Disagree; I think this is what it looks like to get the question of where the perception comes from wrong.
Whether human moral intuitions are correct is the correct question.
Disagree for roughly the same reason; the question of where the word “correct” comes from in this statement seems like the actual query, and is part of the broader question of where human moral intuitions come from.
Oh, if you count that one as a question, then let’s call that one solved too.
Solved where?
Whether humans have the power of free will is the correct question.
Disagree; I think this is what it looks like to get the question of where the perception comes from wrong.
How can philosophers be systematically wrong about the nature of their questions? And what makes you right?
Of course, inasmuch as you agree with Y., you are going to agree that the only question to be answered is where the perception comes for, but this is about truth, not opinion: the important point is that he never demonstrated that.
Whether human moral intuitions are correct is the correct question.
Disagree for roughly the same reason; the question of where the word “correct” comes from in this statement seems like the actual query, and is part of the broader question of where human moral intuitions come from.
if moral intuitions come from God, that might underpin correctness, but things are much less straightforward in naturalistic explanations.
On one level, by the study of dynamical systems and the invention of differential equations.
On a level closer to what you meant when you asked the question, most of the confusing things about ‘causality’ are actually confusing things about the way our high-level models of the world interact with the world itself.
The problem of free will is a useful example of this. People draw this picture that looks like [universe] → [me] → [my future actions], and get confused, because it looks like either determinism (the idea that [universe] → [my future actions] ) isn’t correct or the intuitive sense that I can meaningfully choose my future actions (the idea that [me] → [my future actions] ) isn’t correct.
But the actual picture is something like [universe: [me] → [my future actions] ]. That is, I am a higher-level concept in the universe, and my future actions are a higher-level concept in the universe, and the relationship between the two of them is also a higher-level concept in the universe. Both determinism and the intuitive sense that I can meaningfully choose my future actions are correct, and there isn’t a real conflict between them. (The intuitive sense mostly comes from the fact that the higher level concept is a lossy compression mechanism; if I had perfect self-knowledge, I wouldn’t have any uncertainty about my future actions, but I don’t have perfect self-knowledge. It also comes from the relative importance of decision-making as a ‘natural concept’ in the whole ‘being a human’ business.)
And so when philosophers ask questions like “When the cue ball knocks the nine ball into the corner pocket, what are the terms of this causal relation?” (from SEP), it seems to me like what they’re mostly doing is getting confused about the various levels of their models, and mistaking properties of their models for properties of the territory.
That is, in the territory, the wavefunction of the universe updates according to dynamical equations, and that’s that. It’s only by going to higher level models that things like ‘cause’ and ‘effect’ start to become meaningful, and different modeling choices lead to different forms of cause and effect.
Now, there’s an underlying question of how my map came to believe the statement about the territory that begins the previous paragraph, and that is indeed an interesting question with a long answer. There are also lots of subtle points, about stuff like that it’s interesting that we don’t really need an idea of counterfactuals to describe the universe and the dynamical equations but we do need an idea of counterfactuals to describe higher-level models of the universe that involve causality. But as far as I can tell, you don’t get the main point right by talking about causal relata and you don’t get much out of talking about the subtle points until you get the main point right.
To elaborate a bit on that, hopefully in a way that makes it somewhat clearer why I find it aggravating or difficult to talk about why my approach on philosophy is better, typically I see a crisp and correct model that, if accepted, obsoletes other claims almost accidentally. If you accept the [universe: [me] → [my future actions] ] model of free will, for example, then nearly everything written about why determinism is correct / incorrect or free will exists / doesn’t exists is just missing the point and is implicitly addressed by getting the point right, and explicitly addressing it looks like repeating the point over and over again.
This is also where the sense that they’re wrong about questions is coming from; compare to Babbage being surprised when a MP asked if his calculator would give the right output if given the wrong inputs. If they’re asking X, then something else is going wrong upstream, and fixing that seems better than answering that question.
Oh, if you count that one as a question, then let’s call that one solved too.
Solved where?
On one level, by the study of dynamical systems and the invention of differential equations.
Nope. On most of the detailed questions a philosopher might want to ask about causality , physics comes down firmly on both sides. Physics is not monolothic.
Does causality imply determinism?
(In)determinism is an open question in physics. Note that “differential equations” are used in both classical (deterministic by most accounts) and quantum (indeterminstic by most accounts) physics.
Must causes precede effects?
Perhaps not, if timeless physics, or the theory of closed timelike curves, is correct.
Is causality fundamental?
It is in causal dynamic triangulation, and a few other things. otherwise not.
Both determinism and the intuitive sense that I can meaningfully choose my future actions are correct, and there isn’t a real conflict between them.
Which may be true or false depending on whatever “meaningfully” means. If “meaningful” means choosing between more than one possible future, as required by libertarian free will, then determinism definitely excludes meaningful choice, since it excludes the existence of more than one possible future.
The main problem here is vagueness: you didn’t define “free will” or “meaningful”. Philosophers have known for a long time that people who think free will is compatible with determinism are defining it one way, and people who
think it is not are defining it another way. If you had shown that the libertarian version of free will is compatible
with determinism, you would have shown something momentous , but you actually haven’t shown anything because you haven’t defined “free will” or “meaningful”.
Incidentally, you have also smuggled in the idea that the universe actually is, categorically, deterministic. (Compatibilism is usually phrased hypothetically). As noted, that is actually an open question.
The intuitive sense mostly comes from the fact that the higher level concept is a lossy compression mechanism;
Explaining the feeling of having free will, is a third definition, something different yet again. You don’t see much about in mainstream philosophical literature because the compatibility between a false impression of X and the non-existence of X is too obvious to be worth pointing out—not because it is some great insight that philosophers have never had because they are too dumb.
Having a false impression of X is the least meaningful version of X, surely!
That is, in the territory, the wavefunction of the universe updates according to dynamical equations, and that’s that. It’s only by going to higher level models that things like ‘cause’ and ‘effect’ start to become meaningful, and different modeling choices lead to different forms of cause and effect.
So is causality entirely high level or does it have a fundamental basis?
To elaborate a bit on that, hopefully in a way that makes it somewhat clearer why I find it aggravating or difficult to talk about why my approach on philosophy
I find it aggravating to keep pointing out to people that they haven’t in any way noticed the real problem. It seems to you that you have solved the problem of free will just because you are using concepts in such a vague way that you can;t get a handle on the real problem.
(In)determinism is an open question in physics. Note that “differential equations” are used in both classical (deterministic by most accounts) and quantum (indeterminstic by most accounts) physics.
For the human level, it is irrelevant whether quantum physics is lawfully deterministic or lawfully following a quantum random number generator. It is still atoms boucing according to equations, except that in one case those equations include a computation of a random number. If every atom is secretly holding a coin that it flips whenever it bounces off another atom, from the human level it makes no difference.
People are often mesmerized by the word “indeterministic”, because they interpret it as “that means magic is possible, and my thoughts actually could be changing the physical events directly”. But that absolutely doesn’t follow. It the atoms flips a coin whenever it bounces off another atom, that is still completely unrelated to the content of my thoughts.
Quantum experiments that show how particles follow some statistical patterns when moving through two slits, still don’t show any connection between the movement of the particle and the human thought. So this is all a huge red herring.
If you don’t understand how whether the atom is flipping a truly random coin when bouncing off another atom, or whether it only follows a computation that doesn’t include a random coin is completely irrelevant for debating human “free will”, then you are simply confused about the topic.
Maybe this will help:
Imagine that a master has two slaves. The first slave receives a command “today, you will pick cotton the whole day”. The second slave receives a command “today in the morning, your foreman will flip a coin—if it lands head, you will pick cotton the whole day; if it lands tails, you will clean the stables the whole day”. Is the second slave any more “free” than the first one? (Just because until the foreman flips the coin he is unable to predict what he will be doing today? How is that relevant to freedom? If the foreman instead of a coin uses a quantum device and sends an electron through two slits, does that make the difference?)
People are often mesmerized by the word “indeterministic”, because they interpret it as “that means magic is possible, and my thoughts actually could be changing the physical events directly”.
Perhaps laypeople are that confused, but what we are talking about is Yudkwosky versus professional philosophy.
Philosophers have come up with a class of theory called “naturalistic libertarian free will”, which is based on appealing to physical indeterminism to provide a basis for free will, without appeals to magic. (eg Robert Kane’s).
But that absolutely doesn’t follow. It the atoms flips a coin whenever it bounces off another atom, that is still completely unrelated to the content of my thoughts.
You speak as though your thoughts are distinct from the physical behaviour of your brain...but you don’t actually believe. Plugging in your actual belief that thoughts are just a high-level description of fine-grained neural processing, then the question of Fw becomes the following:
“How can a physical information-processing system behave in a way that is, seen from the outside indeterminstic (unpredictable in principle) and also, within reasonable limits, rational, intelligent and agentive.
(ie from the outside we might want to preserve the validity of “X did Y because they thought it was a good idea” but only as a high-level descritption, and without thoughts appearing in the fundamental ontology).
That is the problem that naturalistic FW addresses.
If you don’t understand how whether the atom is flipping a truly random coin when bouncing off another atom, or whether it only follows a computation that doesn’t include a random coin is completely irrelevant for debating human “free will”, then you are simply confused about the topic.
Do the reading I’ve done before calling me confused. You guys would sound a lot more rational f you could get into the habit of saying “I know of no good argument for Y” instead of “Y is wrong and anyone who believes it is an idiot”.
Imagine that a master has two slaves. The first slave receives a command “today, you will pick cotton the whole day”. The second slave receives a command “today in the morning, your foreman will flip a coin—if it lands head, you will pick cotton the whole day; if it lands tails, you will clean the stables the whole day”. Is the second slave any more “free” than the first one? (Just because until the foreman flips the coin he is unable to predict what he will be doing today? How is that relevant to freedom? If the foreman instead of a coin uses a quantum device and sends an electron through two slits, does that make the difference?)
The usual fallacy: you are assuming that the coin flip is in the driving seat, but actually no part of brain has to act on any particular indeterminstic impulse. If an algorithm contains indeterminsitc function calls embedded in determinstic code, you can’t strip out the deterministic code and still be able to predict what it does.
You speak as though your thoughts are distinct from the physical behaviour of your brain...but you don’t actually believe.
More like: my thoughts are implemented by the interaction of the atoms in my brain, but there is no meaningful relation between the content of my thoughts, and how the atoms in my brain flipped their coins.
[...] is it a reasonable stipulation to say that flipping the switch does not affect you in any [in-principle experimentally detectable] way? All the particles in the switch are interacting with the particles composing your body and brain. There are gravitational effects—tiny, but real and [in-principle experimentally detectable]. The gravitational pull from a one-gram switch ten meters away is around 6 * 10-16 m/s2. That’s around half a neutron diameter per second per second, far below thermal noise, but way above the Planck level.
My point is that technically there is an interaction between the content of my thoughts and how the individual atoms in my brain flip their coins (because the “concent of my thoughts” is implemented by positions and movements of various atoms in my brain), but there is still no meaningful correlation. It’s not like thinking “I want to eat the chocolate cake now” systematically shifts the related atoms in my brain to the left side, and thinking “I want to keep the chocolate cake for tomorrow” systematically shifts the related atoms in my brain to the right side.
If the atoms in my brains would receive different results from flipping their coins, could it change the content of my thoughts? Sure. Some thought impulses carried by those atoms could arrive a few nanoseconds sooner, some of them a few nanoseconds later, some of them could be microscopically stronger or microscopically weaker. According to chaos theory, at some moment later, an imaginary butterfly in my mind could flap its wings differently, and it could make the difference between whether my desire to eat the cake wins over the plan to put it in the fridge, if the desires are sufficiently balanced. On the other hand, the greater imbalance between these two desires (and the shorter time interval for changes to chaotically propagate through the system), the smaller chance of the imaginary butterfly to change the outcome.
But my point is, again, that there is no meaningful correlation between the coin flips and the resulting thoughts and actions. Suppose you have two magical buttons: if you press one of them, you can make all my cake-decision-related atoms receive a head on their coins, if you press the other, you can make them all receive tails. You wouldn’t even know which one to press. Maybe neither would produce the desired butterfly.
The conclusion is that while technically how the atoms flip their coins has some relation with the content of my thoughts, the relation is meaningless. Expecting it to somehow explain the “free will” means searching for the answer in the wrong place, simply because that’s where the magical quantum streetlight is.
“How can a physical information-processing system behave in a way that is, seen from the outside indeterminstic (unpredictable in principle) and also, within reasonable limits, rational, intelligent and agentive.
The aspects that are “unpredictable in principle” are irrelevant to whether it seems rational and agentive.
A stone rolling down the hill is technically speaking “unpredictable in principle”, because there is the “Heisenberg’s uncertainty” about the exact position and momentum of its particles, and yet it doesn’t seem rational nor agentive. If this argument does not give “free will” to stones, it shouldn’t be used as an explanation of “free will” in humans, because it is not valid in general.
More like: my thoughts are implemented by the interaction of the atoms in my brain, but there is no meaningful relation between the content of my thoughts, and how the atoms in my brain flipped their coins.
There is a relationship between your brain state and your thoughts, which is that your thoughts are entirely
constituted by, and predictable from, your brain state. Moreover, the temporal sequence of your thoughts is constituted by and predictable from you the evolution of your brain state, whether it is determinsitic or indeterministc.
I see no grounds for saying that your thoughts lack a “meaningful” connection to your brain states in the indeterministic case only, … but then I don’t know that you mean by “meaningful”. Care to taboo it for me?
My point is that technically there is an interaction between the content of my thoughts and how the individual atoms in my brain flip their coins (because the “concent of my thoughts” is implemented by positions and movements of various atoms in my brain), but there is still no meaningful correlation. It’s not like thinking “I want to eat the chocolate cake now” systematically shifts the related atoms in my brain to the left side, and thinking “I want to keep the chocolate cake for tomorrow” systematically shifts the related atoms in my brain to the right side.
No. Its more like identity. You seem, to be saying that your thoughts aren’t non -physical things are causing physical brain states. That’s something. Specifically, it is a refutation of interactionist dualism...but, as such it doesn’t have that much to do with free will, as usually defined. If all libertarian theories were a subset of interactionist theories, you would be on to something,, but they are not.
The conclusion is that while technically how the atoms flip their coins has some relation with the content of my thoughts, the relation is meaningless.
Taboo meaningless, please.
Expecting it to somehow explain the “free will” means searching for the answer in the wrong place, simply because that’s where the magical quantum streetlight is.
Saying it is the wrong answer because it is the wrong answer is pointless. You need to find out what naturalistic libertarianism actually says, and then refute. It.
The aspects that are “unpredictable in principle” are irrelevant to whether it seems rational and agentive.
So much the better for naturalistic libertarianism , then. One of the standard counterargument to it is that the more
free you are , the less rational you would be.
A stone rolling down the hill is technically speaking “unpredictable in principle”, because there is the “Heisenberg’s uncertainty” about the exact position and momentum of its particles, and yet it doesn’t seem rational nor agentive.
Which would refute the claim that indeteminism alone is a sufficient condition for rationality and agency. But that claim is not made naturalistic libertarianism. Would it kill you to do some homework?
If this argument does not give “free will” to stones, it shouldn’t be used as an explanation of “free will” in humans, because it is not valid in general.
This is like saying that if physics does not result in consciousness in stones, we shouldn’t admit that it results in consciousness in humans.
I have no particular reason to think that we have libertarian free will. But we do make choices, and if those choices are indeterminate, then we have libertarian free will. If those choices are indeterminate, it will in fact be because of the indeterminacy of the underlying matter.
If your argument is correct, something more is needed for libertarian free will besides choices which are indeterminate. What is that extra component that you are positing as necessary for free will?
This is like saying that if physics does not result in consciousness in stones, we shouldn’t admit that it results in consciousness in humans.
My point exactly. If physics does not result in consciousness in stones, then “physics” is not an explanation of consciousness in humans.
And neither is “quantum physics” an explanation of free will in humans (as long as we use any definition of “free will” which does not also apply to stones).
What is that extra component that you are positing as necessary for free will?
Well, the philosophers are supposed to have some superior insights, so I am waiting for someone to communicate them clearly. Preferably without invoking quantum physics in the explanation.
My guess is that “free will” belongs to the realm of psychology. We can talk about when we mean when we feel that other people (or animals, or hypothetical machines) have “free will”, and what we mean when we feel that we have “free will”. That’s all there is about “free will”. Start with the experiences that caused us to create the expression “free will” in the first place, and follow the chain of causality backwards (what in the world caused us to have these experiences? how exactly does that work?). Don’t have a bottom line of “X, in principle” first.
So… what would make me feel that someone or something has a free will? I guess “not completely predictable”, “not completely random”, “seems to follow some goals” and “can somewhat adapt to changes in its environment” are among the key components, but maybe I forgot something just as important.
But whether something seems predictable or unpredictable to me, that is a fact about my ability to predict, not about the observed thing. I mean, if something is “unpredictable in principle”, that would of course explain my inability to predict it. But there are also other reasonable explanations for my inability to predict—some of them so obvious that they are probably low-status to mention—such as me not having enough information, or not having enough computing power. I don’t see the atoms in other people’s brains, I couldn’t compute their movements fast enough anyway, so I can’t predict other people’s thoughts or actions precisely enough. Thus, other people are “not completely predictable” to me.
I see no need to posit that this unpredictability exists “in principle”, in the territory. That assumption is not necessary for explaining my inability to predict. If there is no reason why something should exist in the territory, we should avoid talking about it like it necessarily exists there. The quantum physics is a red herring here. My inability to predict systems reaches far beyond what the Heisenberg’s uncertainty would make me concede. The vast majority of my inability to predict complex systems such as human brains—and therefore the vast majority of my perception of “free will”—is completely unrelated to quantum physics. (Saying that the quantum noise is the only thing that prevents me from reading the contents of your brain and simulating them in real time would be completely delusional. Probably no respected philosopher holds this position explicitly, but all that hand-waving about “quantum physics” is pointing suggestively in this direction. I am saying it’s a wrong direction.)
And how I believe in my own “free will”? Similarly, I can’t sufficiently observe and predict the working of my own brain either. (Again, the quantum noise is the least of my problems here.)
Adding to my previous comment, to explain the point about stones more fully:
I understand libertarian free will to mean, “the ability to make choices, in such a way that those choices are not completely deterministic in advance.”
We know from experience that people have the ability to make choices. We do not know from experience if they are deterministic in advance or not. And personally I do not know or care.
Your objection about the second part seems to be, “if the second part of the definition is satisfied, but only by reason of something which also exists in stones, that says nothing special about people.”
I agree, it says nothing special about people. That does not prevent the definition from being satisfied. And it is not satisfied by stones, since stones do not have the first part, whether or not they have the second.
My point exactly. If physics does not result in consciousness in stones, then “physics” is not an explanation of consciousness in humans.
Generic physics doesn’t even even account for toasters. You need to plug in structure.
And neither is “quantum physics” an explanation of free will in humans (as long as we use any definition of “free will” which does not also apply to stones).
An explanation all in itself. or a potential part of an explanation , including other things, such as structure.
My guess is that “free will” belongs to the realm of psychology. We can talk about when we mean when we feel that other people (or animals, or hypothetical machines) have “free will”, and what we mean when we feel that we have “free will”. That’s all there is about “free will”. Start with the experiences that caused us to create the expression “free will” in the first place, and follow the chain of causality backwards (what in the world caused us to have these experiences? how exactly does that work?). Don’t have a bottom line of “X, in principle” first
Tracing the feeling back might result in a mechanism that produces a false impression of freedom, or a mechanism
that results in freedom. What you are suggesting leaves the question open.
I see no need to posit that this unpredictability exists “in principle”, in the territory.
Who do yo think is doing that? The claim is hypothetical..that if indeterminism exists in the territory, then it could provide the basis for non-illusory FW. And if we investigate that, we can resolve the question you left open above.
This is all fine, for how you understand the idea of free will. And I personally agree that it does not matter whether the world is unpredictable in principle or not. I am just saying that people who talk about libertarian free will, define it as being able to make choices, without those choices being deterministic. And that definition would be satisfied in a situation where people make choices, as they actually do, and their choices are not deterministic because of quantum mechanics (which may or may not be the case—as I said, I do not care.) And notice that this definition of free will would not be satisfied by stones, even if they are not deterministic, because they do not have the choice part.
In the previous comment, you seemed to be denying that this would satisfy the definition, which would mean that you would have to define libertarian free will in an idiosyncratic sense.
Yes. Viliam is assuming that if you actions correspond to an non-deterministic physics, it is “randomness” rather than you who are responsible for your actions. But what would the world look like if you were responsible for your actions? Just because they are indeterminate (in this view) does not mean that there cannot be statistics about them. If you ask someone whether he wants chocolate or vanilla ice cream enough times, you will be able to say what percentage of the time they want vanilla.
Which is just the way it is if the world results from non-deterministic physics as well. In other worlds the world looks exactly the same. That is because it is the same thing. So there is no reason for Viliam’s conclusion that it is not really you doing it; unless you were already planning to draw that conclusion no matter the facts turned out to be.
I find it aggravating to keep pointing out to people that they haven’t in any way noticed the real problem. It seems to you that you have solved the problem of free will just because you are using concepts in such a vague way that you can;t get a handle on the real problem.
What process do you use to determine which problem is more ‘real’? That seems like our core disagreement, and we can probably discuss that more fruitfully.
The more you diverge from discussing the problem in the literature, the less you are really solving the age old problem of X, Y or Z, as opposed to a substitute of your own invention.
Of course there is also a sense in which some age old problem could be a pseudo problem—but the above reasoning still applies. To really show that a problem is a pseudo problem, you need to show that about the problem as stated and not, again, your own proxy.
To really show that a problem is a pseudo problem, you need to show that about the problem as stated and not, again, your own proxy.
I see, but it seems to me that people are interested in age old problems for three main reasons: 1) they have some conflicting beliefs, concepts, or intuitions, 2) they want to accomplish some goal that this problem is a part of, or 3) they want to contribute to the age old tradition of wrestling with problems.
My main claim is that I don’t care much about the third reason, but do care about the first two. And so if we have an answer for where an intuition comes from, this can often satisfy the first reason. If we have the ability to code up something that works, this can satisfy the second reason.
To give perhaps a cleaner example, consider Epistemology and the Psychology of Human Judgment, in which a philosopher and a psychologist say, basically, “for some weird reason epistemology as a field of philosophy is mostly ignoring modern developments in psychology, and so is focusing its attention on the definition of ‘justified’ and ‘true’ instead of trying to actually improve human decision-making or knowledge acquisition. This is what it would look like to focus on the latter.”
No, it does not. If you do not care about that age-old problem, you don’t have an obligation to show anything about it. You can just ignore the pseudo problem and deal with the actual problem you’re interested in.
Vaniver was saying that causality is entirely high level.
That cannot be the case, though, because it means that causality itself is caused by the low level, which is a contradiction.
The true meaning of cause is just “what has something else coming from it, namely when it can help to explain the thing that comes from it.” This cannot be reduced to something else, because the thing it was supposedly reduced to would be what causality is from, and would help to explain it, leading to a contradiction.
That cannot be the case, though, because it means that causality itself is caused by the low level, which is a contradiction.
Disagreed, because this looks like a type error to me. Molecular chemistry describes the interactions of atoms, but the interactions of atoms are not themselves made of atoms. (That is, a covalent bond is a different kind of thing than an atom is.)
Causality is what it looks like when you consider running a dynamical system forward from various starting points, and noting how the future behavior of the system is different from different points. This is deeply similar to the concept of ‘running a dynamical system’ in the first place, and so you might not want to draw a distinction between the two of them.
My point is that our human view of causality typically involves human-sized objects in it, whereas the update rules of the universe operate on a level much smaller than human-sized, and so the connection between the two is mostly opaque to us.
I’m not sure I understand what you are saying, and I am very sure that you either did not understand what I was saying, or else you misinterpreted it.
I was using “cause” in a very general sense, where it is almost, but not quite, equivalent to anything that can be helpful in explaining something. The one extra element that is needed is that, in some way, the effect comes “from” the cause. In the situation you are calling causality, it is true that you can say “the future behavior comes from the present situation and is somehow explained by it,” so there is a kind of causality there. But that is only one kind of causality, and there are plenty of other kinds. For example “is made out of” is a way of being an effect: if something is made out of something else, the thing that is made is “from” the stuff it is made out of, and the stuff helps to explain the existence of the thing.
My point is that if you use this general sense of cause, which I do because I consider it the most useful way to use the word, then you cannot completely reduce causality to something else, but it is in some respect irreducible. This is because “reducing” a thing is finding a kind of cause.
It looks to me like you’re saying something along the lines of ‘wait, reverse reductionism is a core part of causation because the properties of the higher level model are caused by the properties of the lower level model.’ I think it makes sense to differentiate between reductionism (and doing it in reverse) and temporal causation, though they are linked.
I agree with the point that if someone is trying to figure out the word “because” you haven’t fully explained it until you’ve unpacked each of its meanings into something crisp, and that saying “because means temporal causation” is a mistake because it obscures those other meanings. But I also think it’s a mistake to not carve out temporal causation and discuss that independent of the other sorts of causation.
Vaniver was saying that causality is entirely high level.
Maybe. But Yudkowsky sometimes writes as though it is fundamental.
That cannot be the case, though, because it means that causality itself is caused by the low level, which is a contradiction.
It would mean causality is constituted by the low level. Nowadays, causation means efficient causation, not material causation.
This cannot be reduced to something else, because the thing it was supposedly reduced to would be what causality is from, and would help to explain it, leading to a contradiction.
As before …efficient causation is narrower than anything that can explain anything.
I agree, it would not be a contradiction to think that you could explain efficient causality using material causality (although you still might be wrong.) But you could not explain material causality in the same way.
Approaching philosophy as science is not new. It has had a few spectacular successes, such as the wholesale transfer of cosmology from science to philosophy, and a lot of failures, judging by the long list of unanswered philosophical questions (about 200, according to wikipedia). It also has the special pitfall of philosophically uninformed scientists answering the wrong question:-
What causality is is the correct question/.
Whether humans have the power of free will is the correct question.
Whether human moral intuitions are correct is the correct question.
Oh, if you count that one as a question, then let’s call that one solved too.
Disagree; I think this is what it looks like to get the question of where the perception comes from wrong.
Disagree for roughly the same reason; the question of where the word “correct” comes from in this statement seems like the actual query, and is part of the broader question of where human moral intuitions come from.
Solved where?
How can philosophers be systematically wrong about the nature of their questions? And what makes you right?
Of course, inasmuch as you agree with Y., you are going to agree that the only question to be answered is where the perception comes for, but this is about truth, not opinion: the important point is that he never demonstrated that.
if moral intuitions come from God, that might underpin correctness, but things are much less straightforward in naturalistic explanations.
On one level, by the study of dynamical systems and the invention of differential equations.
On a level closer to what you meant when you asked the question, most of the confusing things about ‘causality’ are actually confusing things about the way our high-level models of the world interact with the world itself.
The problem of free will is a useful example of this. People draw this picture that looks like [universe] → [me] → [my future actions], and get confused, because it looks like either determinism (the idea that [universe] → [my future actions] ) isn’t correct or the intuitive sense that I can meaningfully choose my future actions (the idea that [me] → [my future actions] ) isn’t correct.
But the actual picture is something like [universe: [me] → [my future actions] ]. That is, I am a higher-level concept in the universe, and my future actions are a higher-level concept in the universe, and the relationship between the two of them is also a higher-level concept in the universe. Both determinism and the intuitive sense that I can meaningfully choose my future actions are correct, and there isn’t a real conflict between them. (The intuitive sense mostly comes from the fact that the higher level concept is a lossy compression mechanism; if I had perfect self-knowledge, I wouldn’t have any uncertainty about my future actions, but I don’t have perfect self-knowledge. It also comes from the relative importance of decision-making as a ‘natural concept’ in the whole ‘being a human’ business.)
And so when philosophers ask questions like “When the cue ball knocks the nine ball into the corner pocket, what are the terms of this causal relation?” (from SEP), it seems to me like what they’re mostly doing is getting confused about the various levels of their models, and mistaking properties of their models for properties of the territory.
That is, in the territory, the wavefunction of the universe updates according to dynamical equations, and that’s that. It’s only by going to higher level models that things like ‘cause’ and ‘effect’ start to become meaningful, and different modeling choices lead to different forms of cause and effect.
Now, there’s an underlying question of how my map came to believe the statement about the territory that begins the previous paragraph, and that is indeed an interesting question with a long answer. There are also lots of subtle points, about stuff like that it’s interesting that we don’t really need an idea of counterfactuals to describe the universe and the dynamical equations but we do need an idea of counterfactuals to describe higher-level models of the universe that involve causality. But as far as I can tell, you don’t get the main point right by talking about causal relata and you don’t get much out of talking about the subtle points until you get the main point right.
To elaborate a bit on that, hopefully in a way that makes it somewhat clearer why I find it aggravating or difficult to talk about why my approach on philosophy is better, typically I see a crisp and correct model that, if accepted, obsoletes other claims almost accidentally. If you accept the [universe: [me] → [my future actions] ] model of free will, for example, then nearly everything written about why determinism is correct / incorrect or free will exists / doesn’t exists is just missing the point and is implicitly addressed by getting the point right, and explicitly addressing it looks like repeating the point over and over again.
This is also where the sense that they’re wrong about questions is coming from; compare to Babbage being surprised when a MP asked if his calculator would give the right output if given the wrong inputs. If they’re asking X, then something else is going wrong upstream, and fixing that seems better than answering that question.
Nope. On most of the detailed questions a philosopher might want to ask about causality , physics comes down firmly on both sides. Physics is not monolothic.
Does causality imply determinism? (In)determinism is an open question in physics. Note that “differential equations” are used in both classical (deterministic by most accounts) and quantum (indeterminstic by most accounts) physics.
Must causes precede effects? Perhaps not, if timeless physics, or the theory of closed timelike curves, is correct.
Is causality fundamental? It is in causal dynamic triangulation, and a few other things. otherwise not.
Which may be true or false depending on whatever “meaningfully” means. If “meaningful” means choosing between more than one possible future, as required by libertarian free will, then determinism definitely excludes meaningful choice, since it excludes the existence of more than one possible future.
The main problem here is vagueness: you didn’t define “free will” or “meaningful”. Philosophers have known for a long time that people who think free will is compatible with determinism are defining it one way, and people who think it is not are defining it another way. If you had shown that the libertarian version of free will is compatible with determinism, you would have shown something momentous , but you actually haven’t shown anything because you haven’t defined “free will” or “meaningful”.
Incidentally, you have also smuggled in the idea that the universe actually is, categorically, deterministic. (Compatibilism is usually phrased hypothetically). As noted, that is actually an open question.
Explaining the feeling of having free will, is a third definition, something different yet again. You don’t see much about in mainstream philosophical literature because the compatibility between a false impression of X and the non-existence of X is too obvious to be worth pointing out—not because it is some great insight that philosophers have never had because they are too dumb.
Having a false impression of X is the least meaningful version of X, surely!
So is causality entirely high level or does it have a fundamental basis?
I find it aggravating to keep pointing out to people that they haven’t in any way noticed the real problem. It seems to you that you have solved the problem of free will just because you are using concepts in such a vague way that you can;t get a handle on the real problem.
For the human level, it is irrelevant whether quantum physics is lawfully deterministic or lawfully following a quantum random number generator. It is still atoms boucing according to equations, except that in one case those equations include a computation of a random number. If every atom is secretly holding a coin that it flips whenever it bounces off another atom, from the human level it makes no difference.
People are often mesmerized by the word “indeterministic”, because they interpret it as “that means magic is possible, and my thoughts actually could be changing the physical events directly”. But that absolutely doesn’t follow. It the atoms flips a coin whenever it bounces off another atom, that is still completely unrelated to the content of my thoughts.
Quantum experiments that show how particles follow some statistical patterns when moving through two slits, still don’t show any connection between the movement of the particle and the human thought. So this is all a huge red herring.
If you don’t understand how whether the atom is flipping a truly random coin when bouncing off another atom, or whether it only follows a computation that doesn’t include a random coin is completely irrelevant for debating human “free will”, then you are simply confused about the topic.
Maybe this will help:
Imagine that a master has two slaves. The first slave receives a command “today, you will pick cotton the whole day”. The second slave receives a command “today in the morning, your foreman will flip a coin—if it lands head, you will pick cotton the whole day; if it lands tails, you will clean the stables the whole day”. Is the second slave any more “free” than the first one? (Just because until the foreman flips the coin he is unable to predict what he will be doing today? How is that relevant to freedom? If the foreman instead of a coin uses a quantum device and sends an electron through two slits, does that make the difference?)
Perhaps laypeople are that confused, but what we are talking about is Yudkwosky versus professional philosophy.
Philosophers have come up with a class of theory called “naturalistic libertarian free will”, which is based on appealing to physical indeterminism to provide a basis for free will, without appeals to magic. (eg Robert Kane’s).
You speak as though your thoughts are distinct from the physical behaviour of your brain...but you don’t actually believe. Plugging in your actual belief that thoughts are just a high-level description of fine-grained neural processing, then the question of Fw becomes the following:
“How can a physical information-processing system behave in a way that is, seen from the outside indeterminstic (unpredictable in principle) and also, within reasonable limits, rational, intelligent and agentive.
(ie from the outside we might want to preserve the validity of “X did Y because they thought it was a good idea” but only as a high-level descritption, and without thoughts appearing in the fundamental ontology).
That is the problem that naturalistic FW addresses.
Do the reading I’ve done before calling me confused. You guys would sound a lot more rational f you could get into the habit of saying “I know of no good argument for Y” instead of “Y is wrong and anyone who believes it is an idiot”.
The usual fallacy: you are assuming that the coin flip is in the driving seat, but actually no part of brain has to act on any particular indeterminstic impulse. If an algorithm contains indeterminsitc function calls embedded in determinstic code, you can’t strip out the deterministic code and still be able to predict what it does.
More like: my thoughts are implemented by the interaction of the atoms in my brain, but there is no meaningful relation between the content of my thoughts, and how the atoms in my brain flipped their coins.
Somewhat related to this part in “The Generalized Anti-Zombie Principle”:
My point is that technically there is an interaction between the content of my thoughts and how the individual atoms in my brain flip their coins (because the “concent of my thoughts” is implemented by positions and movements of various atoms in my brain), but there is still no meaningful correlation. It’s not like thinking “I want to eat the chocolate cake now” systematically shifts the related atoms in my brain to the left side, and thinking “I want to keep the chocolate cake for tomorrow” systematically shifts the related atoms in my brain to the right side.
If the atoms in my brains would receive different results from flipping their coins, could it change the content of my thoughts? Sure. Some thought impulses carried by those atoms could arrive a few nanoseconds sooner, some of them a few nanoseconds later, some of them could be microscopically stronger or microscopically weaker. According to chaos theory, at some moment later, an imaginary butterfly in my mind could flap its wings differently, and it could make the difference between whether my desire to eat the cake wins over the plan to put it in the fridge, if the desires are sufficiently balanced. On the other hand, the greater imbalance between these two desires (and the shorter time interval for changes to chaotically propagate through the system), the smaller chance of the imaginary butterfly to change the outcome.
But my point is, again, that there is no meaningful correlation between the coin flips and the resulting thoughts and actions. Suppose you have two magical buttons: if you press one of them, you can make all my cake-decision-related atoms receive a head on their coins, if you press the other, you can make them all receive tails. You wouldn’t even know which one to press. Maybe neither would produce the desired butterfly.
The conclusion is that while technically how the atoms flip their coins has some relation with the content of my thoughts, the relation is meaningless. Expecting it to somehow explain the “free will” means searching for the answer in the wrong place, simply because that’s where the magical quantum streetlight is.
The aspects that are “unpredictable in principle” are irrelevant to whether it seems rational and agentive.
A stone rolling down the hill is technically speaking “unpredictable in principle”, because there is the “Heisenberg’s uncertainty” about the exact position and momentum of its particles, and yet it doesn’t seem rational nor agentive. If this argument does not give “free will” to stones, it shouldn’t be used as an explanation of “free will” in humans, because it is not valid in general.
There is a relationship between your brain state and your thoughts, which is that your thoughts are entirely constituted by, and predictable from, your brain state. Moreover, the temporal sequence of your thoughts is constituted by and predictable from you the evolution of your brain state, whether it is determinsitic or indeterministc.
I see no grounds for saying that your thoughts lack a “meaningful” connection to your brain states in the indeterministic case only, … but then I don’t know that you mean by “meaningful”. Care to taboo it for me?
No. Its more like identity. You seem, to be saying that your thoughts aren’t non -physical things are causing physical brain states. That’s something. Specifically, it is a refutation of interactionist dualism...but, as such it doesn’t have that much to do with free will, as usually defined. If all libertarian theories were a subset of interactionist theories, you would be on to something,, but they are not.
Taboo meaningless, please.
Saying it is the wrong answer because it is the wrong answer is pointless. You need to find out what naturalistic libertarianism actually says, and then refute. It.
So much the better for naturalistic libertarianism , then. One of the standard counterargument to it is that the more free you are , the less rational you would be.
Which would refute the claim that indeteminism alone is a sufficient condition for rationality and agency. But that claim is not made naturalistic libertarianism. Would it kill you to do some homework?
This is like saying that if physics does not result in consciousness in stones, we shouldn’t admit that it results in consciousness in humans.
I have no particular reason to think that we have libertarian free will. But we do make choices, and if those choices are indeterminate, then we have libertarian free will. If those choices are indeterminate, it will in fact be because of the indeterminacy of the underlying matter.
If your argument is correct, something more is needed for libertarian free will besides choices which are indeterminate. What is that extra component that you are positing as necessary for free will?
My point exactly. If physics does not result in consciousness in stones, then “physics” is not an explanation of consciousness in humans.
And neither is “quantum physics” an explanation of free will in humans (as long as we use any definition of “free will” which does not also apply to stones).
Well, the philosophers are supposed to have some superior insights, so I am waiting for someone to communicate them clearly. Preferably without invoking quantum physics in the explanation.
My guess is that “free will” belongs to the realm of psychology. We can talk about when we mean when we feel that other people (or animals, or hypothetical machines) have “free will”, and what we mean when we feel that we have “free will”. That’s all there is about “free will”. Start with the experiences that caused us to create the expression “free will” in the first place, and follow the chain of causality backwards (what in the world caused us to have these experiences? how exactly does that work?). Don’t have a bottom line of “X, in principle” first.
So… what would make me feel that someone or something has a free will? I guess “not completely predictable”, “not completely random”, “seems to follow some goals” and “can somewhat adapt to changes in its environment” are among the key components, but maybe I forgot something just as important.
But whether something seems predictable or unpredictable to me, that is a fact about my ability to predict, not about the observed thing. I mean, if something is “unpredictable in principle”, that would of course explain my inability to predict it. But there are also other reasonable explanations for my inability to predict—some of them so obvious that they are probably low-status to mention—such as me not having enough information, or not having enough computing power. I don’t see the atoms in other people’s brains, I couldn’t compute their movements fast enough anyway, so I can’t predict other people’s thoughts or actions precisely enough. Thus, other people are “not completely predictable” to me.
I see no need to posit that this unpredictability exists “in principle”, in the territory. That assumption is not necessary for explaining my inability to predict. If there is no reason why something should exist in the territory, we should avoid talking about it like it necessarily exists there. The quantum physics is a red herring here. My inability to predict systems reaches far beyond what the Heisenberg’s uncertainty would make me concede. The vast majority of my inability to predict complex systems such as human brains—and therefore the vast majority of my perception of “free will”—is completely unrelated to quantum physics. (Saying that the quantum noise is the only thing that prevents me from reading the contents of your brain and simulating them in real time would be completely delusional. Probably no respected philosopher holds this position explicitly, but all that hand-waving about “quantum physics” is pointing suggestively in this direction. I am saying it’s a wrong direction.)
And how I believe in my own “free will”? Similarly, I can’t sufficiently observe and predict the working of my own brain either. (Again, the quantum noise is the least of my problems here.)
Adding to my previous comment, to explain the point about stones more fully:
I understand libertarian free will to mean, “the ability to make choices, in such a way that those choices are not completely deterministic in advance.”
We know from experience that people have the ability to make choices. We do not know from experience if they are deterministic in advance or not. And personally I do not know or care.
Your objection about the second part seems to be, “if the second part of the definition is satisfied, but only by reason of something which also exists in stones, that says nothing special about people.”
I agree, it says nothing special about people. That does not prevent the definition from being satisfied. And it is not satisfied by stones, since stones do not have the first part, whether or not they have the second.
Generic physics doesn’t even even account for toasters. You need to plug in structure.
An explanation all in itself. or a potential part of an explanation , including other things, such as structure.
Tracing the feeling back might result in a mechanism that produces a false impression of freedom, or a mechanism that results in freedom. What you are suggesting leaves the question open.
Who do yo think is doing that? The claim is hypothetical..that if indeterminism exists in the territory, then it could provide the basis for non-illusory FW. And if we investigate that, we can resolve the question you left open above.
This is all fine, for how you understand the idea of free will. And I personally agree that it does not matter whether the world is unpredictable in principle or not. I am just saying that people who talk about libertarian free will, define it as being able to make choices, without those choices being deterministic. And that definition would be satisfied in a situation where people make choices, as they actually do, and their choices are not deterministic because of quantum mechanics (which may or may not be the case—as I said, I do not care.) And notice that this definition of free will would not be satisfied by stones, even if they are not deterministic, because they do not have the choice part.
In the previous comment, you seemed to be denying that this would satisfy the definition, which would mean that you would have to define libertarian free will in an idiosyncratic sense.
Yes. Viliam is assuming that if you actions correspond to an non-deterministic physics, it is “randomness” rather than you who are responsible for your actions. But what would the world look like if you were responsible for your actions? Just because they are indeterminate (in this view) does not mean that there cannot be statistics about them. If you ask someone whether he wants chocolate or vanilla ice cream enough times, you will be able to say what percentage of the time they want vanilla.
Which is just the way it is if the world results from non-deterministic physics as well. In other worlds the world looks exactly the same. That is because it is the same thing. So there is no reason for Viliam’s conclusion that it is not really you doing it; unless you were already planning to draw that conclusion no matter the facts turned out to be.
What process do you use to determine which problem is more ‘real’? That seems like our core disagreement, and we can probably discuss that more fruitfully.
The real problem is the problem as discussed in the literature.
So, implicitly, “the more professional philosophers care about a problem, the more real it is”?
The more you diverge from discussing the problem in the literature, the less you are really solving the age old problem of X, Y or Z, as opposed to a substitute of your own invention.
Of course there is also a sense in which some age old problem could be a pseudo problem—but the above reasoning still applies. To really show that a problem is a pseudo problem, you need to show that about the problem as stated and not, again, your own proxy.
I see, but it seems to me that people are interested in age old problems for three main reasons: 1) they have some conflicting beliefs, concepts, or intuitions, 2) they want to accomplish some goal that this problem is a part of, or 3) they want to contribute to the age old tradition of wrestling with problems.
My main claim is that I don’t care much about the third reason, but do care about the first two. And so if we have an answer for where an intuition comes from, this can often satisfy the first reason. If we have the ability to code up something that works, this can satisfy the second reason.
To give perhaps a cleaner example, consider Epistemology and the Psychology of Human Judgment, in which a philosopher and a psychologist say, basically, “for some weird reason epistemology as a field of philosophy is mostly ignoring modern developments in psychology, and so is focusing its attention on the definition of ‘justified’ and ‘true’ instead of trying to actually improve human decision-making or knowledge acquisition. This is what it would look like to focus on the latter.”
No, it does not. If you do not care about that age-old problem, you don’t have an obligation to show anything about it. You can just ignore the pseudo problem and deal with the actual problem you’re interested in.
All this is posited on having made a claim to have solved problem an existing problem. Read back.
Vaniver was saying that causality is entirely high level.
That cannot be the case, though, because it means that causality itself is caused by the low level, which is a contradiction.
The true meaning of cause is just “what has something else coming from it, namely when it can help to explain the thing that comes from it.” This cannot be reduced to something else, because the thing it was supposedly reduced to would be what causality is from, and would help to explain it, leading to a contradiction.
Disagreed, because this looks like a type error to me. Molecular chemistry describes the interactions of atoms, but the interactions of atoms are not themselves made of atoms. (That is, a covalent bond is a different kind of thing than an atom is.)
Causality is what it looks like when you consider running a dynamical system forward from various starting points, and noting how the future behavior of the system is different from different points. This is deeply similar to the concept of ‘running a dynamical system’ in the first place, and so you might not want to draw a distinction between the two of them.
My point is that our human view of causality typically involves human-sized objects in it, whereas the update rules of the universe operate on a level much smaller than human-sized, and so the connection between the two is mostly opaque to us.
I’m not sure I understand what you are saying, and I am very sure that you either did not understand what I was saying, or else you misinterpreted it.
I was using “cause” in a very general sense, where it is almost, but not quite, equivalent to anything that can be helpful in explaining something. The one extra element that is needed is that, in some way, the effect comes “from” the cause. In the situation you are calling causality, it is true that you can say “the future behavior comes from the present situation and is somehow explained by it,” so there is a kind of causality there. But that is only one kind of causality, and there are plenty of other kinds. For example “is made out of” is a way of being an effect: if something is made out of something else, the thing that is made is “from” the stuff it is made out of, and the stuff helps to explain the existence of the thing.
My point is that if you use this general sense of cause, which I do because I consider it the most useful way to use the word, then you cannot completely reduce causality to something else, but it is in some respect irreducible. This is because “reducing” a thing is finding a kind of cause.
It looks to me like you’re saying something along the lines of ‘wait, reverse reductionism is a core part of causation because the properties of the higher level model are caused by the properties of the lower level model.’ I think it makes sense to differentiate between reductionism (and doing it in reverse) and temporal causation, though they are linked.
I agree with the point that if someone is trying to figure out the word “because” you haven’t fully explained it until you’ve unpacked each of its meanings into something crisp, and that saying “because means temporal causation” is a mistake because it obscures those other meanings. But I also think it’s a mistake to not carve out temporal causation and discuss that independent of the other sorts of causation.
Maybe. But Yudkowsky sometimes writes as though it is fundamental.
It would mean causality is constituted by the low level. Nowadays, causation means efficient causation, not material causation.
As before …efficient causation is narrower than anything that can explain anything.
I agree, it would not be a contradiction to think that you could explain efficient causality using material causality (although you still might be wrong.) But you could not explain material causality in the same way.