Free Will

[Goodspeed]

Yes. This always ends at definitions and it seems that no definition of free will really makes sense. More and more I'm starting to think it's just a silly discussion not worth having, in part also because people seem to go into it thinking it's somehow relevant to morality

[Dolan]

You are now thinking of a simple computer. But let's do a thought experiment and think bigger: imagine having a computer as big as you want it to be, with infinite computing power. This sort of computer can actually accurately simulate the universe and it's laws of physics. This computer can create a simulation that has all the emergent properties that any omplex system has.

If one believes the brain to be made up of particles that follow the laws of physics, then our hypothetical computer can, in principle, make a perfect simulation of a brain. We would need to know the exact distribution of the atoms in a brain, but in principle a computer can do this simulation. In other words, our collection of bits can, if organized in the appropriate manner and if governed by the appropriate algorithm (that is, the algorithm that exactly represents the laws of physics), exactly simulate a human brain. I don't see how this collection of bits can then actually have the emergent property of free will, as long as these bits themselves are just electrical switches.

I doubt a computer will ever be able to simulate the entire universe, not in the least because nobody experiences that to be able to model it. We'd have to build an imaginary case for this that would work on so many assumptions, that it would end up being sci-fi. Too many assumptions related to the technology, resources and energy needed to perform such a simulation. And again, the content itself, the universe, is not available for direct experience to anyone. Without this, the simulation would be devoid of content (or would only represent a very small area of it that we could map out in a simulation) and the thought experiment couldn't be tested. These are flight of fancy thought experiments that are cool to mentally play around with, but not exactly testable or realistic propositions.

The brain case is something closer to home and theoretically doable. However, we don't even have a working model for how a neuron works. We thought we did, but some recent experiments showed that a neuron can work in non-linear ways, that is, you don't simply get an action potential immediately as you excite it with an impulse. Sometimes the action potential comes out with a delay, as if the neuron doesn't simply process an impulse mechanistically. This has upset some AI evangelists who kept banging on about how the brain is a computer and we're going to reach singularity in just a few decades. This idea is not currently supported by what we see from brain research. So frankly, I'm not sure if we can even make this assumption that a brain could be one day simulated. What if we discover quantum effects in how memory or cognition work?
There have been some findings that quantum effects could be involved in the internal compass that helps birds orient themselves (https://www.nature.com/articles/s41586-021-03618-9) and always reach the right location they're aiming for, despite the vast distance between their starting point and their destination. If similar effects are found in human brains, then this might indicate that the human brain is not simply a finite-state machine that we could simulate. There could be lots of nondeterministic, probabilistic states involved in how neurons work, which would make the whole physical ensemble unpredictable.

I've also noticed that the answer to the conundrum that I brought up is thought to be 'emergent properties'. But in assuming this is a satisfying solution, we are treating these emergent properties as a sort of magic black box that can produce any result that we desire. But in reality, a complex system can only have emergent properties that obey the laws of physics. Or rather, the complex system of many elements can only have the emergent properties that are the result of these elements following the laws of physics. We can't have a complex system that has the emergent property of defying the laws of physics; ie there is no complex system that will ignore the law of conservation of energy (at least, this is not possible within our scientific beliefs).

I think part of what creates this conundrum of how could free will work within a determined world is how we think about physical laws. Most of these arguments resort to pointing to how our macro world works in completely deterministic ways, governed by Newtonian and, at the limit, relativistic physics. But recently we've also seen quantum effects even in bigger molecules. A quantum wave effect was observed in a peptide, an organic compound (https://www.nature.com/articles/s41467-020-15280-2). Brain physiology relies on lots of peptides for some of the most important hypothalamic, pituitary and gastrointestinal functions. Such findings really make the idea of quantum effects in the brain a lot less fanciful than it previously seemed.

In that case, if we do manage to prove there are quantum effects in the brain, then a complex system (let's say, a brain) doesn't have to break the laws of physics to work in non-predictable and non-simulatable ways. Because the laws of physics themselves are not some rigid, linear grooves that everything must be guided through without any room for wiggle. The lower we go in terms of scale the less the building blocks of matter behave in linear, predictable ways. And in fact, any kind of intervention (like a measurement) that we do to study matter at a quantum level changes its behaviour, which prompts some very interesting philosophical questions: if matter molds itself to fit our measurements, how could it determine processes that are supposed to simply program our behaviour?

[Dolan]

Actually, I'll make a special post for this research paper I mentioned above, because it's one of the most groundbreaking things that have been found in the last decades in brain research.

Slow integration leads to persistent action potential firing in distal axons of coupled interneurons

The conventional view of neurons is that synaptic inputs are integrated on a timescale of milliseconds to seconds in the dendrites,
with action potential initiation occurring in the axon initial segment. We found a much slower form of integration that leads to
action potential initiation in the distal axon, well beyond the initial segment. In a subset of rodent hippocampal and neocortical
interneurons, hundreds of spikes, evoked over minutes, resulted in persistent firing that lasted for a similar duration. Although
axonal action potential firing was required to trigger persistent firing, somatic depolarization was not. In paired recordings,
persistent firing was not restricted to the stimulated neuron; it could also be produced in the unstimulated cell. Thus, these
interneurons can slowly integrate spiking, share the output across a coupled network of axons and respond with persistent firing
even in the absence of input to the soma or dendrites.

Study can be read here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3030701/

What this means is that not only neurons do not simply and mechanistically output an action potential after being stimulated, they can even fire without being stimulated.
This is kind of a shocking discovery (not sure how else to call it) that pretty much put AI evangelists on a suicide watch.
How can people talk so loud about building and simulating brains, when we don't even know how a brain cell works at a most fundamental level and we cannot predict its behaviour.
That's why all this philosophical speculation about free will is based on so many wrong assumptions about how organic matter behaves, that it's mostly sci-fi confabulation.

[Goodspeed]

Actually, I'll make a special post for this research paper I mentioned above, because it's one of the most groundbreaking things that have been found in the last decades in brain research.

Slow integration leads to persistent action potential firing in distal axons of coupled interneurons

The conventional view of neurons is that synaptic inputs are integrated on a timescale of milliseconds to seconds in the dendrites,
with action potential initiation occurring in the axon initial segment. We found a much slower form of integration that leads to
action potential initiation in the distal axon, well beyond the initial segment. In a subset of rodent hippocampal and neocortical
interneurons, hundreds of spikes, evoked over minutes, resulted in persistent firing that lasted for a similar duration. Although
axonal action potential firing was required to trigger persistent firing, somatic depolarization was not. In paired recordings,
persistent firing was not restricted to the stimulated neuron; it could also be produced in the unstimulated cell. Thus, these
interneurons can slowly integrate spiking, share the output across a coupled network of axons and respond with persistent firing
even in the absence of input to the soma or dendrites.

Study can be read here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3030701/

What this means is that not only neurons do not simply and mechanistically output an action potential after being stimulated, they can even fire without being stimulated.
This is kind of a shocking discovery (not sure how else to call it) that pretty much put AI evangelists on a suicide watch.
How can people talk so loud about building and simulating brains, when we don't even know how a brain cell works at a most fundamental level and we cannot predict its behaviour.
That's why all this philosophical speculation about free will is based on so many wrong assumptions about how organic matter behaves, that it's mostly sci-fi confabulation.

Don't confuse "we don't know why X is happening" with "X is happening spontaneously". It's quite a leap to assume spontaneity.

I say that without having read the study. It seems likely that they actually do have some idea of what's happening, and their explanation is certainly not "neurons are firing spontaneously".

And if it's not spontaneous, it can be simulated.

[RefluxSemantic]

Yes. This always ends at definitions and it seems that no definition of free will really makes sense. More and more I'm starting to think it's just a silly discussion not worth having, in part also because people seem to go into it thinking it's somehow relevant to morality

If there's no definition of free will that makes sense, wouldn't it make sense to conclude that free will doesn't exist?

I find it a reasonably interesting talking point with the right people. But many people (subconsciously) want free will to exist and are then rather subjective in their reasoning.

[RefluxSemantic]

@RefluxSemantic if you're not already familiar with it, you may be interested in the free will theorem by Conway and Kochen (link, also see "The Strong Free Will Theorem" by the same authors, and criticisms "What Does the Free Will Theorem Actually Prove?" by Goldstein et al., and "On the notion of free will in the Free Will Theorem" by Landsman. They're hard to link on mobile but easy to find).

The main claim of the paper addresses something you pointed out, namely that humans having free will implies that elementary particles also have free will (under certain assumptions that are commonly accepted in modern physics). The authors also seem to think that this makes a good case for the existence of free will, which is strange to me, since I think it makes a pretty good case against free will.

Philosophy isn't fun when you let others do the thinking for you. I like laymen philosophy, where I do a poor job at understanding a philosophically challenging subject and then pretend I have solved it

[princeofcarthage]

Show hidden quotes

If we, a collection of elementary particles, have freedom of action, doesn't that necessarily mean that the particles that we consist of have freedom of action as well? And then wouldn't that mean that every single thing in the universe has 'free will'?

A collection of things can have different properties than individual things. This is like the basic rule of science... we learnt it 1st grade.

Is this actually a rule of science? Do you have any scientific examples? As far as I know, every bit of nature can be explained as elementary particles following the rules of physics.

Okay, I am out of my depth here I agree, so I don't understand entirely what you mean but compounds for ex have different properties than its corresponding atoms. Does that count as example

The compounds have different properties, but these properties are described excellently by a model of elementary particles following the laws of nature. The difference in properties arises from the increased complexity of the system, but not because the laws of physics change or because the building blocks behaving differently.

So assuming that the world is basically a collection of elementary particles following the laws of physics (which I would say is the ruling paradigm in science), I then don't see a way in which a collection of elementary particles (like a human brain) can have freedom of action if these elementary particles themselves do not.

So this is why I think the discussion about free will revolves around the question I asked earlier: How does one define free will such that a particle does not have it but a human does have it?

You are answering yourself. Free will arises from the complexity of our brain, not found in atom, compound or elementary particle. Laws of physics or building blocks don't have to behave differently for that.

Same way as compounds do not share same properties as their respective elements. Free will is simply a property of brain?

[Dolan]

I say that without having read the study. It seems likely that they actually do have some idea of what's happening, and their explanation is certainly not "neurons are firing spontaneously".

And if it's not spontaneous, it can be simulated.

From the Discussion section of that study:

The mechanisms responsible for all aspects of this new operational mode are unknown, and will require extensive additional work to be revealed.

[Goodspeed]

I didn't. Anyway yeah, don't confuse "we don't know" with "it's spontaneous".

[Jam]

Spontaneous has a scientific meaning such as a spontaneous chemical reaction is one that will occur under the given conditions by itself. So process can be called spontaneous without violating determinism depending on the definition.

[Goodspeed]

Ah, then that is ambiguous. I meant spontaneous as in indeterministic. I didn't want to use that word because it's apparently a trigger word to many

[RefluxSemantic]

Dolan, I called it a thought experiment for a reason. It doesn't have to be feasible at all.

But in principle we can simulate the laws of physics. In principle a computer so large that it can simulate a human brain can exist. In principle one can then simulate a human brain (even though we don't know the exact composition of the human brain, it is theoretically possible to have a computer exactly simulate this human brain). This means that within our laws of physics, a computer could make a perfect simulation of a random human brain. That human brain would essentially be the sum of a complex system of bits. I don't see how such a 'deterministic' system could have free will.

[RefluxSemantic]

Ah, then that is ambiguous. I meant spontaneous as in indeterministic. I didn't want to use that word because it's apparently a trigger word to many

I think the discussion doesn't concern determinism. It's an irrelevant aspect within the question that I raised.

When the laws of nature are indeterminate, it could save 'free will', in the sense that every single particle has a 'free will' (ie the underlying mechanism for quantum mechanics is the existance of 'free will'). But if every particle has free will, then every single physical system will have a free will and the entire notion of free will is reduced to an empty concept devoid of any significant meaning.

For a satisfying free will to exist, ie a free will that applied complex intelligent living beings and not to inanimate objects, free will needs to somehow be an emergent property of some specific systems. But emergent properties are not some magical black box that can produce anything you want them to be; an emergent property can only be the result of the specific interactions between the building blocks of the system. This then again requires the existence of some physical process where the building blocks have 'free will', which again reduces the principle of free will to a meaningless concept.

[RefluxSemantic]

For me (read: a complex system of many elementary particles) to have a free will - the ability to make choices that are not purely the result of the many elementary particles of my complex system following the laws of physics - I (my complex system) needs to have the emergent property that the elementary particles that make up me do not simply follow the laws of physics.

That would be against science as we know it.

Edit: it is possible that my complex system has free will because:
- elementary particles have free will (ie free will is baked into nature through quantum mechanics), so by following the laws of physics the complex system will have free will (but so will every other complex system)
- because a complex system is more than elementary particles. Ie by aranging elementary particles in a specific way, our resulting brain will be granted a soul by God which allows us to defy the laws of nature. This would be a logically acceptable take, but it would not be a scientific take.

[Jam]

Dolan, I called it a thought experiment for a reason. It doesn't have to be feasible at all.

But in principle we can simulate the laws of physics. In principle a computer so large that it can simulate a human brain can exist. In principle one can then simulate a human brain (even though we don't know the exact composition of the human brain, it is theoretically possible to have a computer exactly simulate this human brain). This means that within our laws of physics, a computer could make a perfect simulation of a random human brain. That human brain would essentially be the sum of a complex system of bits. I don't see how such a 'deterministic' system could have free will.

But is a software simulated brain conscious or do you have to design the computer at the hardware level to function like a brain ie a neural network style computer. I don't think a traditional computer simulating a brain would be conscious but it needs to be simulated at a hardware level because consciousness seems to be an emergent property of a particular type of system and the way a computer does calculations is not the same type of system as a brain. And if the computer simulated brain is not conscious can it really be used to answer the question of free will.

[princeofcarthage]

Why do you think free will defies the law of nature? It's completely possible that different complex system has different properties. That free will is simply a property of brain. Why does it have to be more than that? Jellyfish for ex doesn't have brain and is entirely dependent on external factors for its survival. Is that not proof enough? It's probably the most like to like comparison.

[Goodspeed]

Ah, then that is ambiguous. I meant spontaneous as in indeterministic. I didn't want to use that word because it's apparently a trigger word to many

I think the discussion doesn't concern determinism. It's an irrelevant aspect within the question that I raised.

When the laws of nature are indeterminate, it could save 'free will', in the sense that every single particle has a 'free will' (ie the underlying mechanism for quantum mechanics is the existance of 'free will'). But if every particle has free will, then every single physical system will have a free will and the entire notion of free will is reduced to an empty concept devoid of any significant meaning.

For a satisfying free will to exist, ie a free will that applied complex intelligent living beings and not to inanimate objects, free will needs to somehow be an emergent property of some specific systems. But emergent properties are not some magical black box that can produce anything you want them to be; an emergent property can only be the result of the specific interactions between the building blocks of the system. This then again requires the existence of some physical process where the building blocks have 'free will', which again reduces the principle of free will to a meaningless concept.

I agree with all of that, I was just making a side point about not confusing "we don't know" with indeterminism.

[Goodspeed]

Dolan, I called it a thought experiment for a reason. It doesn't have to be feasible at all.

But in principle we can simulate the laws of physics. In principle a computer so large that it can simulate a human brain can exist. In principle one can then simulate a human brain (even though we don't know the exact composition of the human brain, it is theoretically possible to have a computer exactly simulate this human brain). This means that within our laws of physics, a computer could make a perfect simulation of a random human brain. That human brain would essentially be the sum of a complex system of bits. I don't see how such a 'deterministic' system could have free will.

But is a software simulated brain conscious or do you have to design the computer at the hardware level to function like a brain ie a neural network style computer. I don't think a traditional computer simulating a brain would be conscious but it needs to be simulated at a hardware level because consciousness seems to be an emergent property of a particular type of system and the way a computer does calculations is not the same type of system as a brain. And if the computer simulated brain is not conscious can it really be used to answer the question of free will.

Why wouldn't it be conscious? What is it about consciousness that would make it impossible to simulate with ones and zeroes?

In the end, is consciousness not simply a level of understanding about ourselves and our surroundings that enables us to differentiate between the two?

[Jotunir]

Free will is the capacity of agents to choose between different possible courses of action unimpeded. I think we always have free will.

[Goodspeed]

Free will is the capacity of agents to choose between different possible courses of action unimpeded. I think we always have free will.

Yeah this is my take as well. It's the only definition of free will I can sort of get behind.

[evilcheadar]

Why care about free will when———>it’s over

[chris1089]

Free will is the capacity of agents to choose between different possible courses of action unimpeded. I think we always have free will.

Then we don't have free will all the time, but only for certain actions. For instance I am not unimpeded when I wish to fly. I can't choose between the course of action of jumping in the air and falling, and the course of action of jumping in the air and then flying forward before descending.
I agree we are free to choose between certain courses of action, but what courses of action we can choose between is limited by our human nature.

[fightinfrenchman]

Free will is the capacity of agents to choose between different possible courses of action unimpeded. I think we always have free will.

Then we don't have free will all the time, but only for certain actions. For instance I am not unimpeded when I wish to fly. I can't choose between the course of action of jumping in the air and falling, and the course of action of jumping in the air and then flying forward before descending.
I agree we are free to choose between certain courses of action, but what courses of action we can choose between is limited by our human nature.

Fascinating stuff

[Dolan]

1. Freedom should be understood as the latitude to act on a physical reality constrained by physical laws, by acting through them (just like taking a train, even though you can't decide when it moves, but you can decide to get onboard)
2. The fact that physical laws exist doesn't mean they determine and micromanage every single physical event, like a grain of sand drifting a bit because the wind blew over it (this happened as a conjunction of multiple effects, but not specifically by one particular law; physical laws don't direct things individually)
3. Organic matter, while constrained by physical laws and acting through them doesn't have its actions specifically directed by physical laws: there's no specific physical law that makes dogs bark.
4. Quantum effects could ripple through up to macro reality and lead to some unpredictable effects that don't reach a particular state until "observed" (acted upon in some way) - therefore many things that living organisms do can be unpredictable, they can act in ways that are not directed by physical laws, but still constrained by them like a train on a railway track
5. Organisms get mad, they agitate over the fact that they're constrained by physical laws, while physical laws are calm and placid and always make sure they verify all our equations to the last digit.
6. Until one day a physicist claims he discovered a new physical law, which causes previous physical laws to panic, because it's time to run their course in a different way to make sure they verify our theories again.
7. The more we try to pin down physical reality to fixed trajectories, the more particles go awry, their precise positions become impossible to establish until we perform a measurement, then they pretend they're in one place just to humour us.
8. They know that in the end we're just trying to offload the burden of living onto them, those little particles, we're trying to find a way out of the puzzle of life by explaining it away through some external mover (a god, some particles, physical laws) and so become just like them, calm and free of any conflicts or concerns, free from the burden of subjectivity. But they are smarter, so when we look away, they keep wobbling around defying any logic, whereas when we try to measure what they do, they suddenly walk in line like little disciplined soldiers, signalling to us not to worry, reality is stable, our equations are correct.
9. In this way, they foil our plans to escape the burden and mystery of life by becoming completely explainable objects ourselves. As long as there still are some unresolved questions in some corners, in our models, the mystery hangs over our heads. But once we completely explain physical reality and so ourselves, that cloud will move over to them, the things themselves. There will be no room for any struggles or concerns, the deciders and movers will be the particles themselves, they will be the new subjects.
10. Maybe AI and the singularity will absolve us of the burden of existence, finally, after centuries in which we invested all our hopes in god and then in the sanctity of physical laws, maybe our fantasy of offloading our subjectivity onto some other entity will come to fruition and AI-powered robots will not let us down. They will act so recklessly and out of control, we will finally recognise ourselves in them. And this will set us free.

[chris1089]

How do you account for the supernatural in your above formulation?