Is Free Will Possible in a Deterministic Universe?

[mihaiv]

Are the causal relations that we see around us complete and going back to the birth of universe? Is there any real randomness? Is there anything else other than causal relations and randomness in order to be able to have free will?
I know these are open question in philosophy and science. Please come with arguments.

 

[johng23]

Yes, the universe is deterministic. I know it would be more helpful if I elaborated, but unfortunately that's out of my control.

 

[mihaiv]

Yes, the universe is deterministic. I know it would be more helpful if I elaborated, but unfortunately that's out of my control.

So, you say complete causality, no randomness, no free will...

 

[johng23]

To even entertain the notion of determinism seems ludicrous to me. The ultimate reality is whatever humans perceive. All of our reasoning, all of our logic, is ultimately a product of this perception. To then use the faculty of reason to suggest a higher reality than perception seems pointlessly contrary. It's not testable, it's not tenable. What purpose would you have for denying the most self evident aspect of life?

You might as well jump in front of a bus if you think you have no free will. I mean, it's not like you could stop yourself.

 

[imiyakawa]

So, you say complete causality[I think you mean determinism]

Nobody can say this. QM could be random (most interpretations).

no free will...

Oh boy. If materialism: All I'll say is, if the brain follows physical law, then the brain follows physical law. No matter whether you want to find the explanatory variables by reducing or thinking in terms of strong emergence. Laws that govern higher order interactions are still laws. Randomness doesn't help the picture, as all it introduces are fully determined probability distributions of events. You need a third type of causality, one beyond anyone's wildest imagination. Even if you wish to propose that laws somehow break down in extremely complex systems, this doesn't give you free will. The antecedent of that breaking down is determinism/randomness which forms an unknown level of constraint on that 'breakage'.

If someone proposes a coherent model of how such a causality could arise, they will be the most unprecedented genius ... ever. But it probably won't be done.

But even though the brain follows the constraint of physics - random or otherwise - FAPP (for all practical purposes) to yourself, you have free will.

You might as well jump in front of a bus if you think you have no free will.

I'm going to assume you're a materialist, if you're not the debate is moot because there isn't physical causal closure. I'm also going to assume you hold a definition of free will that people on the street think they have. So, do you wish to propose a mechanism outside of randomness and determinism? I'm all ears, really. But unfortunately we have no reason to suspect the brain doesn't follow laws, which means no free will under the definition I provided that I assumed you think is correct. Self-evidence is never a good argument. Ever.

Even if consciousness is 'X' processes, and there are specific laws that arise at that level of interaction that can't be reduced, that is no argument whatsoever for the existence of free will. They are still laws. Brain is still constrained under those laws.

 

[Pythagorean]

Our universe appears to be pretty damn deterministic. If causality is a coincidence of many superimposing random events, than that would be pretty miraculous in itself, especially with how often something like, say, gravity passes the existence test (100% in all my experiments... I have a record of scars outlining the detailed impact of each free-body fall).

When it comes to organisms and free-will, you're talking about very complicated, difficult to predict, determinism (but determinism none the less): Nonlinear systems, chaos, extreme sensitivity to initial conditions and particular kinds of perturbations. This is why determinism has no affect on individuality (because each person, in addition to having a slightly different set of genes, experiences different aspects of the world).

Of course, for organisms to persist in a stable manner through time (through reproduction) they have to all conform to certain laws, somehow. They have to be able to predict dangers and avoid death until they reproduce. Of course, this isn't the whole story. If a biological system is stressed to the point where it cannot function properly, it may just give up and die.

This is why it's difficult (impossible, from a deterministic perspective) for your average, non-stressed organism, to willfully step in front of a bus that has a good chance of killing or maiming him (or her). Which is why you're suggestion is silly. It's very unlikely that anybody here will be able to step in front of it because of determinism's influence on freewill.

 

[arunma]

To even entertain the notion of determinism seems ludicrous to me. The ultimate reality is whatever humans perceive. All of our reasoning, all of our logic, is ultimately a product of this perception. To then use the faculty of reason to suggest a higher reality than perception seems pointlessly contrary. It's not testable, it's not tenable. What purpose would you have for denying the most self evident aspect of life?

You might as well jump in front of a bus if you think you have no free will. I mean, it's not like you could stop yourself.

But really, is any metaphysical view of time completely testable? Sure, some models can be ruled out. For example, the fact that physical laws depend on the time derivatives certain physical quantities means that presentism (i.e. the view that only the present exists) is probably not correct. But even quantum indeteminacy doesn't rule out the possibility that on a metaphysical level, the future exists now and has thus already been determined. We can certainly make epistemological claims about whether or not we can predict the future. And quantum indeterminacy would seem to suggest that we can't. I guess this would rule out Laplacian determinism. But I don't know if it would rule out all forms of determinism.

 

[imiyakawa]

This is why determinism has no affect on individuality (because each person, in addition to having a slightly different set of genes, experiences different aspects of the world)...It's very unlikely that anybody here will be able to step in front of it because of determinism's influence on freewill.

Could you clarify your wording here. It seems you've snuck in "free will" without providing a reason to suspect it is there. It also seems you think determinism runs the brain but somehow the brain is not constrained by determinism.

 

[Pythagorean]

Bad wording; I believe free will to be a human construct. I never implied the brain was not constrained by determinism: that is actually my point.

 

[nismaratwork]

Our universe appears to be pretty damn deterministic. If causality is a coincidence of many superimposing random events, than that would be pretty miraculous in itself, especially with how often something like, say, gravity passes the existence test (100% in all my experiments... I have a record of scars outlining the detailed impact of each free-body fall).

When it comes to organisms and free-will, you're talking about very complicated, difficult to predict, determinism (but determinism none the less): Nonlinear systems, chaos, extreme sensitivity to initial conditions and particular kinds of perturbations. This is why determinism has no affect on individuality (because each person, in addition to having a slightly different set of genes, experiences different aspects of the world).

Of course, for organisms to persist in a stable manner through time (through reproduction) they have to all conform to certain laws, somehow. They have to be able to predict dangers and avoid death until they reproduce. Of course, this isn't the whole story. If a biological system is stressed to the point where it cannot function properly, it may just give up and die.

This is why it's difficult (impossible, from a deterministic perspective) for your average, non-stressed organism, to willfully step in front of a bus that has a good chance of killing or maiming him (or her). Which is why you're suggestion is silly. It's very unlikely that anybody here will be able to step in front of it because of determinism's influence on freewill.

Let me ask this directly: Given the same initial conditions you believe the same scenario will always play out identically? This is, I think, contrary to basic principles of QM, so you do not believe that element of it, or do you simply believe in a "guiding hand" beyond all of that?

 

[apeiron]

Let me ask this directly: Given the same initial conditions you believe the same scenario will always play out identically? This is, I think, contrary to basic principles of QM, so you do not believe that element of it, or do you simply believe in a "guiding hand" beyond all of that?

A good question. And QM uncertainty would seem to say we can't even have "the same initial conditions" except as a thought experiment. There is an inherent graininess or vagueness in the real world (once you get down to small size/high energies).

But say you could imagine a twin pair of wavefunctions, then your point here is that the collapse would be probabilistic. And my rejoinder to that is this is true only because QM does not model the collapse aspect - the boundary constraints that "determine" (well, constrain) the QM outcomes. Constraints can be weak in some cases (making the outcomes freer, or more unpredictable). Or stronger in others (making the outcomes quite predictable in practice).

Furthermore, if we are bringing in the insights of "deterministic chaos", then we can turn things right around. A system released from any initial conditions will track towards its attractor. Does not matter how you start, you end up in the same general place. And this is indeed a constraints based view. It is the global that dominates the local (rather than the local adding up to create the global story).

This is why the whole determined vs random, chance vs necessity, debate is so confused. It is not a case of either/or, but about the dynamic nature of an interaction.

So if we want to talk about determinism in relation to complex systems, rather than the physically most simple, then a better intuition primer is to ask how a seed becomes a tree.

Is the final outcome determined, or random, or a mix of both?

Perhaps finally the fact that there is an interaction becomes obvious. A seed is a vague potential. Depending on where the seed falls, the nature of the soil and light, what grows up around it, the eventual tree can have all kinds shape. It can be tall, bent, twisted. Even in "identical conditions" no two trees will grow exactly the same as the tiniest deviations at the earliest stages of growth will be magnified in non-linear fashion.

But equally, a tree always grows to look like a tree. There is enough information by way of boundary constraints (both in the seed's genes and the regularities of the environment - a sun overhead, etc) to determine the outcome.

 

[nismaratwork]

A good question. And QM uncertainty would seem to say we can't even have "the same initial conditions" except as a thought experiment. There is an inherent graininess or vagueness in the real world (once you get down to small size/high energies).

But say you could imagine a twin pair of wavefunctions, then your point here is that the collapse would be probabilistic. And my rejoinder to that is this is true only because QM does not model the collapse aspect - the boundary constraints that "determine" (well, constrain) the QM outcomes. Constraints can be weak in some cases (making the outcomes freer, or more unpredictable). Or stronger in others (making the outcomes quite predictable in practice).

Furthermore, if we are bringing in the insights of "deterministic chaos", then we can turn things right around. A system released from any initial conditions will track towards its attractor. Does not matter how you start, you end up in the same general place. And this is indeed a constraints based view. It is the global that dominates the local (rather than the local adding up to create the global story).

This is why the whole determined vs random, chance vs necessity, debate is so confused. It is not a case of either/or, but about the dynamic nature of an interaction.

So if we want to talk about determinism in relation to complex systems, rather than the physically most simple, then a better intuition primer is to ask how a seed becomes a tree.

Is the final outcome determined, or random, or a mix of both?

Perhaps finally the fact that there is an interaction becomes obvious. A seed is a vague potential. Depending on where the seed falls, the nature of the soil and light, what grows up around it, the eventual tree can have all kinds shape. It can be tall, bent, twisted. Even in "identical conditions" no two trees will grow exactly the same as the tiniest deviations at the earliest stages of growth will be magnified in non-linear fashion.

But equally, a tree always grows to look like a tree. There is enough information by way of boundary constraints (both in the seed's genes and the regularities of the environment - a sun overhead, etc) to determine the outcome.

Nice way of working foliation in there! :) Yes, I see your point, but it can't be said that the fate of any given seed is to become a tree. The seed could fail, or a rogue insect or fungus could alter or end its growth. Even at this level of complexity, the outcomes may be limited (dead, alive, healthy, ill, shape of branches) to tree or not-tree, but is that deterministic in the way that has been discussed? Will any two or three BB events produce the same CMB? Uncertainty at a level well above the HUP can mean that the seed is ingested by a bird, pooped into river, where it eventually freezes for millenia. The fate of a seed is not deterministic, unless you're controlling the behavior of all other factors. It CAN only ever be a tree or a seed or a sapling, but it can have many fates that are unpredictable.

 

[apeiron]

Nice way of working foliation in there! :) Yes, I see your point, but it can't be said that the fate of any given seed is to become a tree. The seed could fail, or a rogue insect or fungus could alter or end its growth. Even at this level of complexity, the outcomes may be limited (dead, alive, healthy, ill, shape of branches) to tree or not-tree, but is that deterministic in the way that has been discussed? Will any two or three BB events produce the same CMB? Uncertainty at a level well above the HUP can mean that the seed is ingested by a bird, pooped into river, where it eventually freezes for millenia. The fate of a seed is not deterministic, unless you're controlling the behavior of all other factors. It CAN only ever be a tree or a seed or a sapling, but it can have many fates that are unpredictable.

This is my point. Is the model we have of simple systems going to be the same we have of complex ones?

The universe at the physical level can be best modeled in certain terms (where random and determined may make some kind of pragmatic sense). Then life and other forms of complexity may need to be measure across different dimensions (say spontaneity and autonomy).

A really good model of things would be able to span both dimensions of description (simplexity and complicity). But continuing to focus solely on the language of one of these dimensions (is it all determined, is it all random?) is not the way to see this larger model.

And yes, all sorts of fates can befall a seed. But you are betraying the standard prejudices in saying "The fate of a seed is not deterministic, unless you're controlling the behavior of all other factors."

Constraint is not control (which would be strong determinism - by localised agents). Constraint is exactly what it says - a global limiting that restricts the space of the possible and so increases the chances of something locally actual. You could call it a weak determinism (though it can be pretty powerful). Or better yet, just give up on the notion of determinism as an ontological category.

 

[brainstorm]

Free will may be no more complex than the ability to perceive multiple possibilities of action and to exercise conscious judgment in selecting one action over others. If rationality was 100% convergent, there might not be free will. If rationality could not be transcended in favor of irrational choices, there might also be no free will.

I would guess if you could program a computer to recognize multiple choices and some algorhythm for logically weighing them against each other, and a random function that made it possible for the computer to apply multiple reasoning methods and methods for combining and overweighing conflicting results from those different methods, it would still need a random-choice generator to escape endless conflicts as to which choice was optimum.

I don't think it would ever exercise the free-will to just go with its best guess at some point.

 

[Pythagorean]

Let me ask this directly: Given the same initial conditions you believe the same scenario will always play out identically? This is, I think, contrary to basic principles of QM, so you do not believe that element of it, or do you simply believe in a "guiding hand" beyond all of that?

where do you get this from about QM? QM is deterministic!

 

[apeiron]

Free will may be no more complex than the ability to perceive multiple possibilities of action and to exercise conscious judgment in selecting one action over others. If rationality was 100% convergent, there might not be free will. If rationality could not be transcended in favor of irrational choices, there might also be no free will.

I would guess if you could program a computer to recognize multiple choices and some algorhythm for logically weighing them against each other, and a random function that made it possible for the computer to apply multiple reasoning methods and methods for combining and overweighing conflicting results from those different methods, it would still need a random-choice generator to escape endless conflicts as to which choice was optimum.

I don't think it would ever exercise the free-will to just go with its best guess at some point.

There has been a lot of work to try to build artificial intelligence. And the best approaches are based on machines (neural nets) generating non-random states of prediction, then learning from these "choices".

So it is a top-down, constraints-based, interactionist approach. You do make a best guess about what is about to happen next. Then the world may surprise you. And you then reshape your guessing so as to make a better guess next time.

Freewill is then just about having a goal and an expectation about how the goal will be achieved. The goal becomes the constraint which shapes the best guess.

 

[JoeDawg]

Are the causal relations that we see around us complete and going back to the birth of universe?

Causal relations are apparent in our small part, and in our understanding, of the universe. The birth of the universe is another question entirely. Even seemingly random things can appear to contain patterns.

I'd say determinism is a prejudice we have, which isn't to say its bad or not useful.

Is there any real randomness?

Randomness certainly exists in terms of perspective, objective randomness might be a myth, or it might just be rare... but then there is no real pattern too randomness. So our universe may just be in a state of being randomly consistent.

Is there anything else other than causal relations and randomness in order to be able to have free will?

Consciousness... and the nature of consciousness is central, ie whether it is computational, emergent, or some magical soul-thing.

 

[imiyakawa]

Is there anything else other than causal relations and randomness in order to be able to have free will?

You would need a new type of causality in order to have actual free will. I seriously doubt free will can exist under materialism. The illusion of it is no problem.

Causal relations are apparent in our small part, and in our understanding, of the universe. The birth of the universe is another question entirely

What do you mean by "another question entirely."

Free will may be no more complex than the ability to perceive multiple possibilities of action and to exercise conscious judgment in selecting one action over others.

Bad definition, I think. I prefer the definition of the ability to act outside of the total constraint of physical law.

 

[GeorgCantor]

Causal relations are apparent in our small part, and in our understanding, of the universe. The birth of the universe is another question entirely. Even seemingly random things can appear to contain patterns.

I'd say determinism is a prejudice we have, which isn't to say its bad or not useful.

Randomness certainly exists in terms of perspective, objective randomness might be a myth, or it might just be rare... but then there is no real pattern too randomness. So our universe may just be in a state of being randomly consistent.

Consciousness... and the nature of consciousness is central, ie whether it is computational, emergent, or some magical soul-thing.

Well put.

I've been thinking about this soul-business for quite some time but it seems impossible to find the right framework(or even any framework at all) to try to make a case on it.

Alfred Whitehead's views of 'blobs of perception' being fundamental do not address the seeming free will issue(they are consistent with relativity and probability though).

 

[apeiron]

Even seemingly random things can appear to contain patterns.

Randomness IS pattern. It is precisely the patterns that maximise entropy within a set of informational constraints.

There really is no better recent paper on this than...
http://arxiv.org/PS_cache/arxiv/pdf/0906/0906.3507v1.pdf

The neutral patterns share
a special characteristic: they describe the patterns of nature that follow from
simple constraints on information. For example, any aggregation of processes
that preserves information only about the mean and variance attracts to the
Gaussian pattern; any aggregation that preserves information only about the
mean attracts to the exponential pattern; any aggregation that preserves
information only about the geometric mean attracts to the power law pattern.
I present a simple and consistent informational framework of the common
patterns of nature based on the method of maximum entropy. This framework
shows that each neutral generative model is a special case that helps to
discover a particular set of informational constraints; those informational
constraints define a much wider domain of non-neutral generative processes
that attract to the same neutral pattern.

 

[brainstorm]

There has been a lot of work to try to build artificial intelligence. And the best approaches are based on machines (neural nets) generating non-random states of prediction, then learning from these "choices".

So it is a top-down, constraints-based, interactionist approach. You do make a best guess about what is about to happen next. Then the world may surprise you. And you then reshape your guessing so as to make a better guess next time.

Freewill is then just about having a goal and an expectation about how the goal will be achieved. The goal becomes the constraint which shapes the best guess.

My point was that if an AI system had the ability to endlessly reason about which choice to make, how would it decide when to undertake action despite its continuing reasoning process? If it randomly selects a moment, that is not free will, is it? It is a response to a command to generate a random moment in time and act at that moment. It does not have the free will to CHOOSE between continuing to reason or to act. It needs a command-protocol to base its "choice" on.

 

[skeptic2]

My point was that if an AI system had the ability to endlessly reason about which choice to make, how would it decide when to undertake action despite its continuing reasoning process? If it randomly selects a moment, that is not free will, is it? It is a response to a command to generate a random moment in time and act at that moment. It does not have the free will to CHOOSE between continuing to reason or to act. It needs a command-protocol to base its "choice" on.

One might well assume that an AI system that has the ability, not only to reason about which choice to make, but also the ability to make a choice, must have a criterion or criteria for making the choice programmed into it. Once the criteria are satisfied, it would choose. In this case the choice would be determined by the input data and the given criteria. It's certainly not random, but is it freewill? The problem isn't so much whether making a choice determined by external data and a predefined set of criteria amounts to freewill but perhaps how the criteria were determined.

 

[brainstorm]

One might well assume that an AI system that has the ability, not only to reason about which choice to make, but also the ability to make a choice, must have a criterion or criteria for making the choice programmed into it. Once the criteria are satisfied, it would choose. In this case the choice would be determined by the input data and the given criteria. It's certainly not random, but is it freewill? The problem isn't so much whether making a choice determined by external data and a predefined set of criteria amounts to freewill but perhaps how the criteria were determined.

It would still have to satisfy its criteria. If it couldn't, could it make the choice to go ahead and make a choice on the basis of what it had figured out so far?

 

[apeiron]

My point was that if an AI system had the ability to endlessly reason about which choice to make, how would it decide when to undertake action despite its continuing reasoning process? If it randomly selects a moment, that is not free will, is it? It is a response to a command to generate a random moment in time and act at that moment. It does not have the free will to CHOOSE between continuing to reason or to act. It needs a command-protocol to base its "choice" on.

And what I was saying is that people actually doing AI would not think of building a machine with that approach - one based on a need for a separate randomness generating process to break out some kind of deterministic loop.

Turing machines are completely deterministic devices and would have such a problem. But brains are nothing like Turing machines.

 

[brainstorm]

And what I was saying is that people actually doing AI would not think of building a machine with that approach - one based on a need for a separate randomness generating process to break out some kind of deterministic loop.

Turing machines are completely deterministic devices and would have such a problem. But brains are nothing like Turing machines.

That was my point, i.e. that brains have free-will. They are only deterministic to the extent they freely choose to follow command-protocols and algorithms. A brain can emulate AI, but AI cannot emulate a brain, or at least not the part capable of free will.

 

[apeiron]

They are only deterministic to the extent they freely choose to follow command-protocols and algorithms.

I'd like to see the neuroscience behind that statement.

 

[brainstorm]

I'd like to see the neuroscience behind that statement.

What neuroscience? Have you been following this thread? The issue was that a computer could follow various command protocols, including an algorithm that would allow it to randomly choose an action in the event of non-convergent reasoning, but in no case could it simply decide for itself to make a choice based on what it had come up with so far. A human brain can do this, so what can explain this ability to short-circuit the reasoning process to implement a tentative decision except free will?

 

[apeiron]

What neuroscience? Have you been following this thread?

Hah, I have been following THIS thread but not THAT thread - https://www.physicsforums.com/showthread.php?t=411323

But I would still like to see evidence that the brain operates by "command protocols and algorithms" if that is your belief.

 

[brainstorm]

Hah, I have been following THIS thread but not THAT thread - https://www.physicsforums.com/showthread.php?t=411323

But I would still like to see evidence that the brain operates by "command protocols and algorithms" if that is your belief.

(sigh) if you would have read the thread, you would have seen that I posted that the brain ONLY operates by "command protocols and algorithms" by choice and that free will actually governs the choice to apply such deterministic structures, when, and how. If anything is deterministic about the brain, it is subconscious or semi-conscious associations and involuntary thoughts and emotions. Decision-making itself can be influenced by deterministic processes but one's response to them is the result of free-will, I believe. I don't see how neuroscience, specifically, is the best approach to this topic.

 

[apeiron]

A human brain can do this, so what can explain this ability to short-circuit the reasoning process to implement a tentative decision except free will?

My basic point here is again that thinking about complexity in terms of simplicity is the source of most modern philosophical errors.

To see how "freewill" (anticipation, autonomy, etc) can arise in complex adaptive systems, it is better to start from simple examples of complexity (ie: simple forms of life), rather than simple examples of simplicity (ie: Newtonian deterministic mechanics).

One really cute little example is the way a flagella-driven bacterium like E. coli makes intelligent choices.

E. coli swims along, driven by the anti-clockwise cork-screw paddling of its flagella, seeking food. While spinning anti-clockwise, the flagella are all tangled together and the bacterium swims straight.

Covering its surface are receptors that can scent sugars and amino acids. While the receptors are picking up traces of food, the instruction to the flagella is to keep rotating anti-clockwise. But if the concentrations start to fall off, then the flagella are told to reverse. This then untangles the flagella and sends E. coli into a random tumbling spin until the scent begins to pick up again, at which point straight-line swimming can resume towards the source of the food.

So we have here a beautifully simple feedback mechanism. And no philosophical problem at all about how a bacterium switches from a determined, directed, action to an undetermined, randomly exploratory, undirected one.

Extract the principles, scale them up to more complex animals like cats and humans, and we have naturalistic explanations of "freewill".

You won't find anything sensible to latch on to while you stay stuck down at the level of atoms and wavefunctions (or binary computer circuits and information). But a training in biology just makes freewill a non-issue philosophically. There is still the complexity of brains to explain, but no big deal about animals making intelligent choices as a result of fundamental asymmetries (dichotomies) embedded in their design.

Brains work one way when they are smoothly anticipating the world, then switch to a different approach when they encounter errors or suprises. There does not have to be a little homunculus compartment in the brain that does the chosing. The world is always out there to drive things along.