“Free will” is a black box containing our decision making algorithm.
What kind of mind would invent “free will”? The same mind that would neatly wrap up any other open ended question into a single label, be it “élan vital” or “philogeston”. Our minds are fantastic at dreaming up explanations for things, and if they are not easily empirically testable at the time, then such explanations tend to stick. Without falsifying evidence, our pet theories tend to remain, and confirmation bias slowly hardens them into what feels like brute facts.
It’s appealing because it ties up (or at least hides) loose ends. If we play taboo on free will, we might get something like the concept that people can narrow a number of possible futures into one future that is optimal. With this definition, free will would indeed exist. If, however, free will was postulated in such a say as to include some fantastical element, or another black box, Occam’s razor may strike it down. Alternatively, it may be superficially appealing enough to stick, so long as we don’t think about it too thoroughly. For example, the idea that humans are in control of their actions feels like an explanation, but contains control as a nested black box.
But what does this process actually feel like when we make such a mistake? Well, it’s based on implicit assumptions, so nothing feels amiss. You don’t realize that you are making an implicit assumption. All the loose ends look like they are tucked away, at least at a glance. but if you take a closer look, say by repeatedly asking “why?”, then you start to feel less confident. This is a sign of trouble, but you should make sure aren’t just asking “does 1 plus 1 really equal 2″ in a pretentious tone of voice. If repeated self inquiry seems to be creating a rabbit hole of nested black boxes, then you should go back to the highest level box, and try a different form of inquiry. Ask yourself if there’s anything about the nested black boxes that feel wrong. Use all your tools as a rationalist to inquire into this, and hopefully find a path of inquiry besides infinite regression. For a thorough analysis, ask yourself whether the nested boxes make observable predictions, and how those predictions might differ from reality.
With our example of playing taboo on free will to get the idea that humans are in control of their actions, we might intuit that “control” is just an inherently complex concept. Although usually applying agency during an explanation is a good way of sicking Occam’s razor on it, perhaps this is an exception. We are discussing agency itself, after all. But what does it mean to have agency? What are the observable differences in the world? When we hold these sorts of questions in our minds, and again try to play taboo, we are more likely to get something like the concept that people can narrow a number of possible futures into one future that is optimal. That’s a much different answer, because a computer program could also down-select from a number of different options, given some criteria. This answer also doesn’t leave loose ends, and doesn’t leave that nagging feeling of doubt that comes from having left something unexplained. It turns out that all our sense of confusion was contained within the mysticism implied by using the phrase “free will”. It may be easy to forget about that doubt when only taking a broad view, but as soon as you zoom in on the problem it will become detectible. We live in a messy world, and so frequently have to say say “good enough” and leave unexplained doubts due to time constraints, but when we decide that something is important and pursue the little doubting feeling to it’s conclusion, it can be incredibly satisfying. You’ve just answered one of the mysteries of the universe, after all.
So that’s what it feels like to make and then correct a mistake, but why and how did we make the mistake in the first place? Well, our minds are naturally wired around concepts of agency. This is an observable fact in others, and as in personal examples; it really does feel like the dice are out to get us, or that “the system” must be consciously malicious rather than merely incompetent. It is even more natural to endow ourselves with that same vague agency we give inanimate objects and bureaucratic systems. It’s only been in the past couple hundred years that humanity has been able to group everything under the same laws of physics. Before that, the stars obeyed their own special rules, and living things were unexplainable mysteries running on “élan vital” instead of something more akin to a combustion reaction. Unless we specifically look close enough at a belief to notice that requires the world to operate under a different set of physical principles, we will default to what is most natural to believe.
As for why the concept of free will should exist in the first place, it is because it is the most natural explanation. It is a fault in out minds that the most natural is not also the simplest, but it is also a useful feature. This form of vague, associative reasoning let’s us jump to reasonably accurate conclusions quickly. The difference between the two breakdowns of free will I gave is that one only uses known, well understood phenomena, and the other revolves around making the concept of agency unexplainable. One might also entangle the question with the concept of self and our individual sense of identity. By rolling a bunch of concepts up into one, there are less easily recognizable loose ends. The problem is that tools like associative reasoning and vague definitions aren’t enough to actually arrive at a satisfying answer. It ties up enough loose ends that we declare that we’ve solved it, and move on, ignoring the incompleteness.
Note that this in itself isn’t a completely 100% exhaustive expiation of why we naturally want to believe in free will. Where does this concept come from? How do we form it in the first place? To fully answer these, we’d have to also examine the concept if our personal sense of identity, since that is the thing that gets conflated with the “free will” concept to form the more vague and fluffy version. If someone thinks computers don’t or can’t have free will but humans can, this is likely what they mean by “free will”. Our sense of identity is a large issue, and well enough outside the scope of the question that I think it isn’t necessary for this explanation. I’ll leave that particular novel length post for someone else.
I wrote the above before reading any of the comments, but there are a couple other ideas which people touched on but I did not. I’m bringing them together here, mostly for my own future reference:
Humans have the ability to model the outside world in our own minds, including other people, but not our own minds. Because of this, it seems like our choices aren’t subject to causality. Credit, and more detail, here.
Another comment goes into more detail of why this is. In order to fully model itself, a mind would need more power that it has. Therefore, minds cannot predict their own actions with high fidelity. For minds that don’t intuitively understand concepts like recursion, this implies that their own future actions cannot be predicted, and that therefore free will exists.
“Free will” is a black box containing our decision making algorithm.
What kind of mind would invent “free will”? The same mind that would neatly wrap up any other open ended question into a single label, be it “élan vital” or “philogeston”. Our minds are fantastic at dreaming up explanations for things, and if they are not easily empirically testable at the time, then such explanations tend to stick. Without falsifying evidence, our pet theories tend to remain, and confirmation bias slowly hardens them into what feels like brute facts.
It’s appealing because it ties up (or at least hides) loose ends. If we play taboo on
free will, we might get something likethe concept that people can narrow a number of possible futures into one future that is optimal. With this definition, free will would indeed exist. If, however,free willwas postulated in such a say as to include some fantastical element, or another black box, Occam’s razor may strike it down. Alternatively, it may be superficially appealing enough to stick, so long as we don’t think about it too thoroughly. For example,the idea that humans are in control of their actionsfeels like an explanation, but containscontrolas a nested black box.But what does this process actually feel like when we make such a mistake? Well, it’s based on implicit assumptions, so nothing feels amiss. You don’t realize that you are making an implicit assumption. All the loose ends look like they are tucked away, at least at a glance. but if you take a closer look, say by repeatedly asking “why?”, then you start to feel less confident. This is a sign of trouble, but you should make sure aren’t just asking “does 1 plus 1 really equal 2″ in a pretentious tone of voice. If repeated self inquiry seems to be creating a rabbit hole of nested black boxes, then you should go back to the highest level box, and try a different form of inquiry. Ask yourself if there’s anything about the nested black boxes that feel wrong. Use all your tools as a rationalist to inquire into this, and hopefully find a path of inquiry besides infinite regression. For a thorough analysis, ask yourself whether the nested boxes make observable predictions, and how those predictions might differ from reality.
With our example of playing taboo on
free willto getthe idea that humans are in control of their actions, we might intuit that “control” is just an inherently complex concept. Although usually applying agency during an explanation is a good way of sicking Occam’s razor on it, perhaps this is an exception. We are discussing agency itself, after all. But what does it mean to have agency? What are the observable differences in the world? When we hold these sorts of questions in our minds, and again try to play taboo, we are more likely to get something likethe concept that people can narrow a number of possible futures into one future that is optimal. That’s a much different answer, because a computer program could also down-select from a number of different options, given some criteria. This answer also doesn’t leave loose ends, and doesn’t leave that nagging feeling of doubt that comes from having left something unexplained. It turns out that all our sense of confusion was contained within the mysticism implied by using the phrase “free will”. It may be easy to forget about that doubt when only taking a broad view, but as soon as you zoom in on the problem it will become detectible. We live in a messy world, and so frequently have to say say “good enough” and leave unexplained doubts due to time constraints, but when we decide that something is important and pursue the little doubting feeling to it’s conclusion, it can be incredibly satisfying. You’ve just answered one of the mysteries of the universe, after all.So that’s what it feels like to make and then correct a mistake, but why and how did we make the mistake in the first place? Well, our minds are naturally wired around concepts of agency. This is an observable fact in others, and as in personal examples; it really does feel like the dice are out to get us, or that “the system” must be consciously malicious rather than merely incompetent. It is even more natural to endow ourselves with that same vague agency we give inanimate objects and bureaucratic systems. It’s only been in the past couple hundred years that humanity has been able to group everything under the same laws of physics. Before that, the stars obeyed their own special rules, and living things were unexplainable mysteries running on “élan vital” instead of something more akin to a combustion reaction. Unless we specifically look close enough at a belief to notice that requires the world to operate under a different set of physical principles, we will default to what is most natural to believe.
As for why the concept of free will should exist in the first place, it is because it is the most natural explanation. It is a fault in out minds that the most natural is not also the simplest, but it is also a useful feature. This form of vague, associative reasoning let’s us jump to reasonably accurate conclusions quickly. The difference between the two breakdowns of
free willI gave is that one only uses known, well understood phenomena, and the other revolves around making the concept of agency unexplainable. One might also entangle the question with the concept of self and our individual sense of identity. By rolling a bunch of concepts up into one, there are less easily recognizable loose ends. The problem is that tools like associative reasoning and vague definitions aren’t enough to actually arrive at a satisfying answer. It ties up enough loose ends that we declare that we’ve solved it, and move on, ignoring the incompleteness.Note that this in itself isn’t a completely 100% exhaustive expiation of why we naturally want to believe in free will. Where does this concept come from? How do we form it in the first place? To fully answer these, we’d have to also examine the concept if our personal sense of identity, since that is the thing that gets conflated with the “free will” concept to form the more vague and fluffy version. If someone thinks computers don’t or can’t have free will but humans can, this is likely what they mean by “free will”. Our sense of identity is a large issue, and well enough outside the scope of the question that I think it isn’t necessary for this explanation. I’ll leave that particular novel length post for someone else.
I wrote the above before reading any of the comments, but there are a couple other ideas which people touched on but I did not. I’m bringing them together here, mostly for my own future reference:
Humans have the ability to model the outside world in our own minds, including other people, but not our own minds. Because of this, it seems like our choices aren’t subject to causality. Credit, and more detail, here.
Another comment goes into more detail of why this is. In order to fully model itself, a mind would need more power that it has. Therefore, minds cannot predict their own actions with high fidelity. For minds that don’t intuitively understand concepts like recursion, this implies that their own future actions cannot be predicted, and that therefore free will exists.
If we have separate neural hardware for processing human actions and for inanimate events then this might lead to the idea of free will, and then also several other odd notions.