Free will redux: What is true free will?

[Piggy]

Suppose I am the subject of an experiment where I have to freely choose to press either a red button or a green button every time I hear a buzzer. The scientists observing me have a machine that can somehow predict the state of every atom in my body several seconds in advance. If the world is deterministic at the lowest level then this is logically possible (though it almost certainly will never be a practical proposition, of course).

This means that just before they sound the buzzer the scientists know which button I will press. They know what my body will do in advance because my body is just atoms and they have predicted the position of every one of them. Is there some "higher" level in which I don't press the button that was predicted for me, the button that my fingers actually press? Only if this higher level is actually some sort of delusion.

It is not the case that a knowledge of all atomic states -- with no reference to higher-level information -- will produce an understanding of what you are thinking. That assertion is completely unfounded.

 

[hodgy]

I'm sorry, but I don't follow this line of thought at all. I just don't understand what you're saying. Can you rephrase?

Sorry - I was a bit opaque. My point is that even if everything we observe has an apparent cause, and so is apparently not random, there is a larger random factor which is that the universe exists at all (or exists with a starting state of x rather than y). Events at the micro level may apparently have a cause but at the macro level existence itself is random (a random selection of existence over non-existence or of existence like x over existence like y).

So far, my argument implies that every event is (at least in a macro context) random, even in a
n apparently causal universe - so you might ask me how this suports predetermination. There are several approaches I could take here but I'll try the one that I think is the simplest.

In a universe in which both random and caused events are possible, the caused events must be determined - a choice event is the result of preceeding state and events. Whilst an agent may be able to predict (given all relevant data) an event, they cannot prevent it since it is by definition the result of all causal variables.

Does the possibility of random events break determinism? No. A random event is by defintion unpredictable by any agent. If the agent cannot predict the event they cannot prevent it, therefore it is no different in practice from a caused event.

The only real difference between a random and caused event in such a system is that the caused event occurs within a pattern sequence of at least 2 whilst the random event occurs within a pattern sequence of only 1.

As far as I understand the term, if the elaboration of a process involves sub-processes which must be described in probabilistic terms (in other words, we would say there is an element of randomness invovled) then one could, in theory, rewind the process to an earlier state, let it run again, and the result would be a different end state to some degree at least from the state it was in at the point where the rewinding began.

But you have introduced 'rewinding' - this is the what gives opportunity for difference in your model. The universe prior to the rewind is determined to be x and the universe x rewound and re-run is determined y.

 

[Piggy]

If "free will" is what we normally think of it--that is, a conscious free will, in which we actively make decisions and determine our own actions--it certainly is shy about itself. The more we know about our behavior, the less we see room for the sort of free will that makes a difference. And so, discussions of free will dissolve into "randomness", which is not free will at all, and "unconscious choice", which tends toward the oxymoronic, and lots and lots of "it feels free to me", which is utterly irrelevant.

Very well put. I think this is an excellent summary of the perils currently beseting the notion of free will.

The only thing I see which could possibly save the idea is a model of the emergent "entity" of the "conscious self" (that is, the emergent phenomenon of consciousness) which could explain how this emergent phenomenon is able to add information to the loop, or influence the neural activity of the brain.

We don't have any such model. But the field of inquiry is so young, I don't believe we can say that the possibility has yet been eliminated.

In any case, before we can officially recognize a thing called "free will" or write out its death certificate, our models of consciousness (and of emergent phenomena) are going to have to be much more complete than they are.

And that's why I say that it ain't dead.

 

[Piggy]

In a universe in which both random and caused events are possible, the caused events must be determined

Ok, here's where you lose me.

Random does not mean uncaused.

All events have causes. Deterministic events have only one possible outcome for a given set of causes (including environmental variables). Random events do not have only one possible outcome, requiring that the range of outcomes be expressed in terms of probabilities.

 

[Piggy]

Regarding the non-sufficiency of aggregation

To illustrate my point that macro-level entities and phenomena are not explicable merely by reference to the actions of lower-level entities, let's consider vortices.

The development of hurricanes and tornadoes, for example, is only explicable with reference to macro-level entities with macro-level properties, such as pressure zones, temperature zones, and winds.

It is not the case (as everyone here will readily admit) that these things happen because all the atoms happen to be going in the right direction.

Why do these atoms form the patterns they do in these events?

We cannot find the answer exclusively within the atomic-level actions and interactions. If we try to do this, we find that each atom in the vortex is acting the same as atoms that are not in the vortex.

So why have they formed a vortex?

The vortex forms because macro-level "objects" (to use the term loosely) are interacting with one another at that level of granularity. The qualities of these objects are properties of these objects at that level, and only of objects at that level. They are not even present at the micro level (what's the pressure and viscosity of an atom?).

By the same token, essential properties of the lower-level constituents are effectively invisible at the macro level. We can speak meaningfully of the behavior of vortices whether they are composed of gases such as air, liquids such as milk and coffee, or stars -- all of which have very different lower-level properties.

Clearly then, the notion of aggregate explanations is false.

 

[hodgy]

Ok, here's where you lose me.
Random does not mean uncaused.

I agree, a random event can be said to have caused itself.

All events have causes.

In which case determinism wins

Deterministic events have only one possible outcome for a given set of causes (including environmental variables). Random events do not have only one possible outcome, requiring that the range of outcomes be expressed in terms of probabilities.

I contend that what you are proposing here is a determined event with a random variable - that's fine but I do not see how a random element supports free will.

 

[hodgy]

Also, 'probabilities' are a construct based on observation, not logic. I am reminded of IQ and similar tests in which we are asked to complete a sequence:

Given 2,4,6,x - what is x? Of course the answer is 8 but there is no logical reason why it should be so. The answer is 47 if the sequence runs 2,4,6,47,2,4,6,47....

'Probablity' means one of 2 things:

1 - there are factors we do not understand and that offers a range of results.
2 - there is a random element to the outcome.

Either way (as I have previously explained) determinism wins,

 

[Piggy]

I agree, a random event can be said to have caused itself.

That's not true either.

Random actions have causes other than themselves. Cause is not the issue at all.

It's a matter of outcomes, not causes.

If cause A within context B will always yeild result X, then the action thus described is entirely deterministic.

If cause A within context B will yeild result X in 60% of trials, result Y in 30% of trials, and result Z in 10% of trials, then the action thus described is to some extent random.

Chemical reactions are deterministic because we know what we're going to get when the reagents and environments are strictly controlled.

Quantum behavior is to some extent random because no matter how strictly we control the initial conditions and environment, we can only describe the outcomes in terms of probabilities.

Rolling dice at a craps table is random because the results can only be described in terms of probabilities. However, this case is different from controlled quantum experiments (or appears to be) because the randomness arises from variations in B (the environmental variables) at each trial. If B were held perfectly constant (e.g. identical motion of the throwing hand, identical starting position of the die, identical wind currents, etc.), it is assumed that the results would be identical every time.

To define random as "having no cause or being its own cause" is a highly unorthodox, even idiosyncratic, definition of the term. Seems to me that events which have no cause or are their own cause would be some type of singularity.

 

[Piggy]

I do not see how a random element supports free will.

Me, neither.

That is not my contention.

My argument is based on the locus of change within the brain/mind system.*

If the conscious "i" is only "along for the ride" -- that is, if non-conscious modules are doing 100% of the work, and it only seems to us like we consider and decide, for example, that we want paper instead of plastic -- then there's no free will.

On the other hand, if the "i" is not merely passive, but may be meaningfully said to be an active element in the feedback loop that underlies the activity of the mind and gives rise to the emergent phenomenon of consciousness itself, then free will is possible.


*"Mind" being defined here as a particular subset of brain activity, specifically that which we identify with what may be commonly called "the conscious self".

 

[Jimbo07]

The important element is whether our conscious selves (these phenomena which we all inherently feel ourselves to "be") are directing our own actions, or whether the unconscious mechanistic bits of our brains are actually running the show, and it only feels like we're directing anything at all.

Oh, if he's anything like predisposed to how I feel, then this is a political smoke screen for, "We have no free will, just the illusion of it." Personally, I've been sliding more toward the camp that consciousness itself is an illusion.

Ignoring the difference between free will and its mere appearance is just silly. Concluding that free will can exist if its appearance exists is like saying that something is a real unicorn because it looks like one.

No... remember walks like a duck, talks like a duck... I think he's getting at, 'we can't tell.' It feels like free will to us... even if it's not, in some sort of fluffy metaphysical sense.

This is the same error. If we're discussing "true free will" here, as the stated thread topic stipulates, then it's insufficient to be satisfied with the trivial and obvious truth that, boy, sure seems like we got it.

Oh. Well, whether one is satisfied or not... sure seems like we got it.

I'm a fan of the whole, 'illusion of free will' thing, I guess...

 

[Piggy]

Also, 'probabilities' are a construct based on observation, not logic.

Ok. But that's not really important. What's important is that, for random events, we must resort to this "construct based on observation" because identical trials don't yield identical results.

We can't just do it once and then know what's going to happen every time, so we're forced to run multiple trials, observe the results, and calculate the probabilities.

Now, the craps game I mentioned above, this would be an example of random results arising from a "random element" (variation in environmental variables and initial conditions from trial to trial) within a deterministic system.

Quantum experiments, however, are another thing altogether. I haven't read any QM papers in many years, but when I was studying the subject, the mainstream interpretation was that identical initial conditions and identical environments will result in different outcomes from trial to trial.

In fact, this is exactly what made QM so shocking, and sparked Einstein's famous retort that God does not play dice.

And yet, as predictions were made and further tests constructed and carried out, it seemed that there was indeed true randomness at the quantum level: identical initial conditions, identical environment, identical stimulus, variable results. True randomness. And, if that were true, the end of the model of the clockwork universe which had stood for centuries.

1 - there are factors we do not understand and that offers a range of results.
2 - there is a random element to the outcome.

Either way (as I have previously explained) determinism wins.

It has been proposed that quantum randomness may be a situation like #1 above -- the result of some hidden variable that our current technology can't detect. But I don't believe that's the mainstream interpretation. Perhaps some of the physics folks can chime in here.

But now we get into the issue at hand -- human behavior.

If we were to rewind the day to several hours ago, and somehow quell all random factors in the environment, so that nothing in the universe would happen differently unless I did something different... would I end up sitting here again typing these exact words? Would "I" be capable of "doing something different"?

Truth is, we don't know.

Quantum mechanics has shown that we can be surprised, and discover what is apparently true randomness where we never expected it.

Can the conscious mind introduce true randomness into the world on an entirely different level of magnification?

I don't think we understand enough about consciousness, and about the qualities and behaviors of emergent systems in general, to know for certain at this point one way or the other.

 

[Piggy]

No... remember walks like a duck, talks like a duck... I think he's getting at, 'we can't tell.' It feels like free will to us... even if it's not, in some sort of fluffy metaphysical sense.

Talk about political smoke screens.

I can't believe you'd come onto a board like this and actually put forth the argument "walks like a duck, talks like a duck" and expect to be taken seriously.

What you're proposing is that the illusion of X is the same as X itself. Or that "Sure seems that way to me" is a valid way to decide the truth of a proposition. That's patent nonsense.

Now, maybe you don't care whether true free will exists, but since that's the topic of this forum, saying that it doesn't matter to you is not particularly productive.

And I'm not talking fluffy metaphysics. I'm talking about the brain.

Go back to the story of my mother's birthday. If, at each stage, my conscious self has no active part in what's going on, if it's something of a passive witness to the activity of non-conscious modules of my brain, then "I" (this conscious entity which I consider myself to be) didn't choose anything at all. And if that's the case, it makes no sense to say I had any free will in the matter.

You might find the distinction not worth looking into, but if so, why are you posting here?

 

[Piggy]

I'm a fan of the whole, 'illusion of free will' thing, I guess...

Oops. Sorry. Should have read that before I barked at you.

<Emily Litella voice> Nevermind. </Emily Litella>

 

[chriswl]

This is simply not true.

Yes it is.

The actions of macro-level phenomena cannot be described by simply aggregating the actions of atoms.

Yes they can. This is such a basic idea in physics that it is people seldom ever bother to express it but believe me (or talk to a scientist) it is absolutely the truth. I mean that, take your statement and change the "cannot" to "can" and it becomes perfectly correct. Same with your comments about vortices - where did you get that stuff from? Not only is it not true its quite impressive how precisely opposite to the truth it is. If you doubt this you are rejecting physics and you are out there with Lifegazer in the distant lands of woo. I'm sorry, I don't mean to be rude but that's just the way it is.

Macro-level objects and phenomena interact based on properties which pertain on a macro-level, which do not even exist on an atomic level.

Not really. We can describe them as doing so but this is just shorthand for our convenience. We could predict their behaviour just as accurately, in fact actually more accurately, if we calculated the trajectory of each atom individually but this would of course be hopelessly impractical.

 

[President Bush]

Does Behaviorism's process of "manipulate(ing) the situation, and find(ing) that organisms (including, but not limited to, people) lawfully respond to these manipulations, altering their choices accordingly" involve any decision-making?

Is "decision-making" the same thing as "choice"? If so, then yes, it does. It involves choices, or decisions, that are every bit as constrained by environmental variables as are the ones they are studying. Skinner's book "Cumulative Record", reprinting his earlier experimental work, makes this point explicitly (thus the title).

What are you talking about?

A behaviorist "manipulate(ing) the situation, and find(ing) that organisms (including, but not limited to, people) lawfully respond to these manipulations, altering their choices accordingly" is not every bit as constrained by environmental variables as the subjects they study.

Subjects being manipulated, responding lawfully to these manipulations, and altering their choices accordingly are in the process of having their environment constrained by the behaviorist her/himself. To the point, at least, of the behaviorist being in charge of her/his controlled experiment.

 

[chriswl]

It is not the case that a knowledge of all atomic states -- with no reference to higher-level information -- will produce an understanding of what you are thinking. That assertion is completely unfounded.

I was talking about physical actions, not thoughts. Are you now claiming we may have freedom of thought but something prevents us acting on it? Like I could freely choose to press the red button but find myself inexplicably forced to press the green one, as predicted by the scientists? That's not much of a free will.

In any case, thoughts are surely the result of with activity in the brain and this is physical. And so it is determined.

 

[Piggy]

This is such a basic idea in physics that it is people seldom ever bother to express it but believe me (or talk to a scientist) it is absolutely the truth.

Well, educate me, then. Telling me that this is just the way it is doesn't help me much. This issue is important to me. Can you provide some resources that would help a non-scientist get the skinny on the topic in a reasonable time?

I wish I could cite my sources, but I had to sell almost all my books when I moved 3 years ago, and all the science went.

 

[Piggy]

It is not the case that a knowledge of all atomic states -- with no reference to higher-level information -- will produce an understanding of what you are thinking. That assertion is completely unfounded.

I was talking about physical actions, not thoughts. Are you now claiming we may have freedom of thought but something prevents us acting on it? Like I could freely choose to press the red button but find myself inexplicably forced to press the green one, as predicted by the scientists? That's not much of a free will.

In any case, thoughts are surely the result of with activity in the brain and this is physical. And so it is determined.

Sorry, my mistake. I imposed the notion of thought, but if our thoughts have no causitive role in our actions -- if, as suggested in the "mother's birthday" example, they are an after-effect -- then they don't matter here. According to your model, the scientists would also be able to predict the thoughts that arose in your mind as you heard the buzzer and pressed the button. {ETA: Assuming, for the sake of the example, that our knowledge of brain processes was advanced enough for us to completely understand the correlation between physical states of the brain and conscious thoughts.}

But for me to accept that model, I'm going to have to have an understanding of why it must be true that all information is present at the lowest level of granularity. Looking forward to getting more info on this.

 

[Piggy]

chriswl, I think you might want to hold off before you slap a woo label on the notion that aggregation is insufficient to fully explain the behavior of macro-level systems.

It seems like some sane folks are interested in the theory.

In the search for a "theory of everything," scientists scrutinize ever-smaller components of the universe.... Stanford physics professor [Robert] Laughlin, awarded the 1998 Nobel Prize for Physics, argues that smaller is not necessarily better. He proposes [that] if we try to analyze things too closely, we risk not understanding how they work on a macro level. In many cases, the whole exhibits properties that can't be explained by the behavior of its parts.

"Emergence" is the notion that the whole is more than the sum of its parts. John Holland, a MacArthur Fellow known as the "father of genetic algorithms," says this seemingly simple notion will be at the heart of the development of machines that can think for hemselves.

Woo?

Some other interesting links:

EmerNet: Emergent neural computational architectures based on neuroscience

Self-Organized Biological Dynamics and Nonlinear Control: Toward Understanding Complexity, Chaos and Emergent Function in Living Systems

The Dynamical Systems Approach to Cognition: Concepts and Empirical Paradigms Based on Self-Organization, Embodiment, and Coordination Dynamics

 

[Piggy]

A couple more resources on emergence:

The Emergent Systems Meeting page and Emergence Reading List

And much as I hate to cite an Encyclopedia of Philosophy, here's the Emergent Properties page from Stanford's.

Reductionism, Emergence, and Effective Field Theories - full text available in PDF format here.

Still want to call it woo?