Free Will

[Aradia]

I don't think so. In any idea of a future that hasn't happened (to us) but has happened for any third party observer, the future is as much a fact as the past is. You can't change the past.

That relies on time being linear relative to a plane beyond itself. Would it necessarily be?

 

[Aradia]

No, that is not the point, which you again have missed. You need to look at this from a philosophical standpoint, not a technical one.

Actually, this is a very technical discussion. The thread is about what free will _is_, and how it _works_. Not just abstract conceptual babbling.

But philosophically, we can consider system A to be separate from system B. And in such a case, there is no way to determine what the next number in the pseudorandom sequence will be (assuming we've put the RNG in a separate process). Therefore, from a philosophical POV, we can point at the system and say that it is partly deterministic, partly random.

"Philosophically" is not a general allowance to ignore reality. You claim in a previous post that things are either: (1) Random, and not determined by past events, or (2) Deterministic, and determined by past events, or (3) a hybrid.

Yet you still have not explained how, in REALITY, a hybrid system could work. Now you're somehow claiming that the randomness is actually outside of our universe(?!). Philosophy is meaningless if it can't be practically applied.

Obviously, the above definition does not fit our real-world implementations of RNG's if we take our whole universe (or even the entire computer) into consideration, but it does fit if we only take a particular algorithm into consideration. It's called building a model. I'm sure you're familiar with the technique.

Alas, when dealing with the concept of free will, we do need to take our whole universe into consideration. So where does that leave your argument?

Therefore, when considering the model I originally described, the description of being partly random, partly deterministic fits. I know it's not possible to implement in real life, but that is inconsequential to the discussion at hand. We are, after all, not dealing with implementations here, but abstracts.

No, we're dealing with implementations. Our implementation. How it has been implemented in OUR universe, for OUR lives. Study quantum mechanics.

 

[Asimov]

That relies on time being linear relative to a plane beyond itself. Would it necessarily be?

It was stated in the OP that relative to us, God knows both Past, Present and Future.

In order to know anything, it would have to be true. Thus, any knowledge of the future would have to be true.

If the future is known, it's already happened, relative to God. Even though to us it hasn't happened yet.

 

[Aradia]

It was stated in the OP that relative to us, God knows both Past, Present and Future.

In order to know anything, it would have to be true. Thus, any knowledge of the future would have to be true.

If the future is known, it's already happened, relative to God. Even though to us it hasn't happened yet.

Which future? Is there a reason why our future would necessarily be linear to god, even though it's linear to us?

I apologise, btw, for any late replies. Got a new job yesterday, so I've not had much time to check the forums.

 

[Asimov]

Which future? Is there a reason why our future would necessarily be linear to god, even though it's linear to us?

I didn't say it was linear to God. In fact I'm pretty sure I indicated that it couldn't be linear to God because he is aware of all aspects of time that are relative to us.

I apologise, btw, for any late replies. Got a new job yesterday, so I've not had much time to check the forums.

Congrats on your new job.

 

[smog]

Oh goodie. Another free will discussion.

Ultimately, the realization of whether free will exists or not does not come from arguments that we are machines or God's omnipotence. It comes from the realization that the very concept of free will is incoherent. Free will does not exist for the same reason that four-sided triangles don't - the definition is a contradiction of terms.

Let me demonstrate. There are two types of systems that we know can theoretically exist - random systems and deterministic systems. There may be more - I'll get to that later.

In a completely random system, effects do not depend on any prior cause. This makes things that happen in a random system "free" in the sense that it's impossible to tell what will happen in the future, but all basis for order and rationality are obviously missing in such a system. Obviously, neither our universe or our minds are completely random for this very reason.

In a completely deterministic system, all effects depend fully on prior causes. This means that, given perfect knowledge of the system, one can perfectly predict the behaviour of that system. Given partial knowledge of the system, one can approximate the future behaviour of that system. The approximation will improve with the available information. This fits rather well with both our universe and our selves, although scientists tell us that there are some things that seem to not be deterministic. In any event, we can approximate the behaviour of those around us, and we get better at it the better we know the people around us. Obviously, a deterministic system can be rational, but it is certainly not free.

There is a third option. A system that is partially deterministic and partially random. Computer based decision making systems are typically constructed according to this model (although the randomness is not true randomness, but that has no relevance here). Such a system can be very similar to a deterministic system, but even with perfect knowledge of the system it is not possible to fully predict its behaviour. However, when the system does not behave according to a "perfect" prediction, it does not do so out of rationality but out of chance. The behaviour in that instance will be chaotic. Therefore, these occasions are signs of free actions that do not fit the description of will, because they cancel out the effects of premeditation. This system therefore does not fit the idea of free will.

So my question to free will proponents is this: if none of the 3 types of systems above fit free will, then just what the heck IS free will? I've asked that question many times, and NOONE has been able to give me an answer.

So how can you argue for something that isn't even possible to define? How do you propose that free will can exist when the very concept of free will is so nebulous that noone knows what it is? I've heard people declare that it is none of the three options above (which is why I brought them up). But a definition must include a positive description - saying what free will isn't is not enough. We must know what it IS if you are to assert its existence.

Technically, one could come up with a reasonable definition of free will under the wholly deterministic label.

Let U be the universe and {E, O} be a partition of U, respectively the entity which we want to determine if it has free will or not and the outside world. For instance, I could partition the universe into my brain and everything else.

Now let's look at the following two probability distributions: P(O(t+1)|O(t)) and P(O(t+1)|O(t), E(t)). The former represents what the outside world will be like in a moment, given knowledge of what it is now. The latter represents what the outside world will be in a moment, given knowledge of what it is now, PLUS knowledge about the entity E.

From there, we could simply define free will to be a measure of the difference between those two distributions (could be something like Kullback–Leibler divergence). Indeed, the greater the difference, the more the outside world depends on E. Basically, instead of splitting the universe into a deterministic system and a random system as you did earlier, I split it into an entity (the one we want to know if it has free will or not) and the rest and then I look at the amount of correlation between the two.

Of course, there are drawbacks to this definition. For example, somebody could make a program, erase all source files, commit suicide and then, technically, the free will of the program would be great. Worse yet, a nuclear bomb would have low a priori probability but great impact, hence great free will. Unfortunately I don't think a formal definition of free will can avoid those cases. Well, I guess we could throw in a measure of complexity of E or some sort of normalization, but that's probably putting too much effort into it.


Regardless, the trap pretty much everybody falls in when discussing free will is that they never even bother to think about the entity they're talking about. When I say X has free will, what is X? From what I can see of debates, X is inexplicably taken in such a way that it cannot be anything but the empty set. I know this sounds weird, but when people talk about the free will of X, they are all really talking about the free will of nothing at all, because they never bother to define X and their discourse is inconsistent with X being nonempty.

What happens with you fill X? Obviously, X will contain a certain amount of information. Furthermore, evidently, X cannot "choose" what information it will contain, because if X was anything else, it would not actually be X, it would be something else. It's nonsensical for an entity to choose their own nature. Therefore, knowledge of X should not bear any consequence on its free will, and in fact it is required in order to identify X in the first place. From there, we could directly conclude that free will is nonsense because it's not even logically possible for an entity to have options: there's a bijection between actions and people, you do Y because you are X and you are X because you do Y. On the other hand, we could say that free will should be a measure of how the behavior of X as a physical entity is correlated to its information content. For example, a rock would have lower free will than a human because its behavior is much less dependant on its contents than the behavior of a human is on his or her brain. (Un?)surprisingly, it doesn't actually have anything to do with determinism or randomness.

Now, as I said, the gates are wide open for many things to have free will that "we might not want to", but you know, there is no denying that the concept of free will is much more emotional than it is rational.


So the point I'm raising is that sure you can say that determinism implies that we can trace back our actions to physical causes, but if those causes are actually part of how we are defined in the first place, well... I hope you can see the loophole there.

Another way to view things is to consider that a person is defined by her behavior and that any different behavior would actually correspond to another person. In that situation, people do choose their behavior, because if they did not, that would imply that several behaviors could be attributed to the same person, which is contradictory with the previous hypothesis. Counterintuitive, I know, but we humans are typically pretty bad at inferring in the correct direction

Consider the computer program to be system A. Our universe as a whole is system B. Obviously, system A is part of system B. An agent in system B can look at the random number generator and conclude that it is deterministic. Since A is part of B, this is still true in A from a technical POV.

But philosophically, we can consider system A to be separate from system B. And in such a case, there is no way to determine what the next number in the pseudorandom sequence will be (assuming we've put the RNG in a separate process). Therefore, from a philosophical POV, we can point at the system and say that it is partly deterministic, partly random.

Technically and philosophically, each invokation of the RNG leaks a certain quantity of information from A to B (a bit, for example). Assuming the RNG has a finite description, this means that as time tends to infinity, enough information would leak into B for A to be deterministic from B's perspective.

 

[Hnefi]

Actually, this is a very technical discussion. The thread is about what free will _is_, and how it _works_. Not just abstract conceptual babbling.

"Philosophically" is not a general allowance to ignore reality. You claim in a previous post that things are either: (1) Random, and not determined by past events, or (2) Deterministic, and determined by past events, or (3) a hybrid.

I think you need to go back and read my first post again, because you have obviously completely misunderstood it. My point is that there are three kinds of system (relevant to this discussion) that are conceptually well defined. Given a system and complete information about it, we can determine which kind of the three alternatives it fits into for that very reason.

I argue that this is NOT possible with free will, because free will is NOT well defined (or defined at all). Regardless of what physical laws we operate in, there is no concievable system that does not fit into any of the three previous categories while free will proponents often claim the contrary. Disproving my claim is easy: provide a positive definition of free will that excludes a free will process from the other three mentioned above. Noone ever has.

Yet you still have not explained how, in REALITY, a hybrid system could work. Now you're somehow claiming that the randomness is actually outside of our universe(?!). Philosophy is meaningless if it can't be practically applied.

While one can argue about the usefulness of philosophy, that does not change the fact that this is a purely philosophical discussion. Free will belongs firmly in that category and my entire argument is about what is logically, not physically, possible.

Alas, when dealing with the concept of free will, we do need to take our whole universe into consideration. So where does that leave your argument?

In the realm of philosophy, obviously.

No, we're dealing with implementations. Our implementation. How it has been implemented in OUR universe, for OUR lives. Study quantum mechanics.

Why? This has absolutely zilch to do with QM, because my argument is in no way dependent on whether there actually IS any randomness in the universe or. not. As I said previously, I'll leave that to the actual physicists. My argument is simply that free will is a logical contradiction and unless that position turns out to be false, it means that free will cannot exist, regardless of true randomness.

 

[Hnefi]

Smog, your definition of free will is one that most of us naturalists would be happy to agree upon. It is trivially obvious that such free will exists and that is a meaningful concept. One flaw, of course, is that according to that definition a rock or a cog also has a degree of free will.

But there is a reason why I chose to classify processes into my previous three categories. The reason is that when you ask a free will proponent what it actually is, it turns out that it is a process that is not deterministic ("we are more than our brains, genes and history"), not random ("randomness is not rational") and not a mix between the two. I have proposed different definitions of free will in various discussion throughout my life and they were all rejected because they fit into one of these categories.

Usually, these people maintain that free will stems from the "soul" or something equivalent and we can't show that nondeterministic, nonrandom free will doesn't exist because we don't know enough about souls. My argument therefore targets the logical contradiction that free will implies, which makes it useless to invoke "souls" in support for the concept.

 

[Aradia]

I didn't say it was linear to God. In fact I'm pretty sure I indicated that it couldn't be linear to God because he is aware of all aspects of time that are relative to us.

Linear in a slightly different sense. If there are an infinite number of futures that are possible, and god knows them all, then he can know our past/present/future while allowing us free will. Think of a choose-your-own-adventure book, if you've ever seen one. =)

 

[Aradia]

I argue that this is NOT possible with free will, because free will is NOT well defined (or defined at all). Regardless of what physical laws we operate in, there is no concievable system that does not fit into any of the three previous categories while free will proponents often claim the contrary. Disproving my claim is easy: provide a positive definition of free will that excludes a free will process from the other three mentioned above. Noone ever has.

Why must free will necessarily be defined outside of those three systems? Free will as a hybrid makes the most sense. It's a shame that you have no idea how such a system would work.

While one can argue about the usefulness of philosophy, that does not change the fact that this is a purely philosophical discussion. Free will belongs firmly in that category and my entire argument is about what is logically, not physically, possible.

In that case, this conversation is pointless. Perhaps you should talk about the IPU instead.

Why? This has absolutely zilch to do with QM, because my argument is in no way dependent on whether there actually IS any randomness in the universe or. not. As I said previously, I'll leave that to the actual physicists. My argument is simply that free will is a logical contradiction and unless that position turns out to be false, it means that free will cannot exist, regardless of true randomness.

Of course not. You can use words like "random" and "deterministic" and "hybrid", but heaven forbid you actually learn how they work in order to properly support your argument.

 

[Hnefi]

Why must free will necessarily be defined outside of those three systems? Free will as a hybrid makes the most sense. It's a shame that you have no idea how such a system would work.

It doesn't. I originally wouldn't. But as I said in the post before this one in reply to Smog, free will proponents often claim that free will is something else than those three types of systems. I claim that it's impossible and I've been trying to tell you this for half a thread now.

In that case, this conversation is pointless. Perhaps you should talk about the IPU instead.

Many discussions are without practical utility. That doesn't mean they are pointless. I quite enjoy discussions about free will - at least when my opponent actually listens to what I have to say.

Of course not. You can use words like "random" and "deterministic" and "hybrid", but heaven forbid you actually learn how they work in order to properly support your argument.

And heaven forbid you actually respond to my arguments instead of building up strawmen based on analogies I've made.

ETA: If you consider philosophy to be useless, why are you posting in the philosophy forum?

 

[Aradia]

It doesn't. I originally wouldn't. But as I said in the post before this one in reply to Smog, free will proponents often claim that free will is something else than those three types of systems. I claim that it's impossible and I've been trying to tell you this for half a thread now.

You've done a poor job of it, though. You complain that free will advocates cannot define free will or how it's supposed to work, and yet you refuse to logically define a random-deterministic hybrid or how it's supposed to work. Slightly hypocritical, I must say.

Many discussions are without practical utility. That doesn't mean they are pointless. I quite enjoy discussions about free will - at least when my opponent actually listens to what I have to say.
And heaven forbid you actually respond to my arguments instead of building up strawmen based on analogies I've made.

Your analogy is directly relevant. It is your only definition for a random-deterministic hybrid, and it falls flat.

ETA: If you consider philosophy to be useless, why are you posting in the philosophy forum?

Actually, I said that philosophy without practicality is meaningless. Perhaps you should listen to what I actually say.

Let us define "random" as a future event not based on any past events, per your definition in post #32. Let us define "deterministic" as a future event based entirely on past events, also per your definition in post #32. Free will could theoretically be defined as "self-determining". This type of system could effect a future event through a self-contained thought process that is both separate from past events, and at the same time non-random.

 

[Asimov]

Linear in a slightly different sense. If there are an infinite number of futures that are possible, and god knows them all, then he can know our past/present/future while allowing us free will. Think of a choose-your-own-adventure book, if you've ever seen one. =)

Regardless of there being an "infinite number" of possible futures, God still knows the one that will come about.

If you're equating God to the reader of a Choose your own adventure, then you're still indicating that we don't have free will.

 

[Hnefi]

You've done a poor job of it, though. You complain that free will advocates cannot define free will or how it's supposed to work, and yet you refuse to logically define a random-deterministic hybrid or how it's supposed to work. Slightly hypocritical, I must say.

This is just dishonest. I gave you two adequate examples of hypothetical systems that are essentially hybrid in this sense. That any practical implementation of these systems would depend on pseudorandomness rather than "true" randomness is inconsequential in a philosophical discussion because the point was to illustrate how such a system might use both deterministic and random mechanics to work. Given that there IS a random factor and deterministic laws (which may or may not ultimately be the case in this universe), my examples hold.

In fact, even from a technical POV my examples still hold, because the functionality of the systems described would not be altered by "true" randomness because of the very way the systems are designed. They don't depend on the causality of the RNG precisely because that would defeat its purpose.

Your analogy is directly relevant. It is your only definition for a random-deterministic hybrid, and it falls flat.

No, it is not my only definition. It is not a definition at all. You never asked for a definition until now; you asked for how such a system might work. I felt a rough example would be more appropriate than a definition because you gave the impression of being computer illiterate and the only examples I know of are computer related.

And until now you hadn't responded to my actual argument; instead you keep nitpicking and switching the subject away from the logical absurdities introduced by free will and instead focus on the practical issues of implementing a random system - which is a secondary issue at best.

Actually, I said that philosophy without practicality is meaningless. Perhaps you should listen to what I actually say.

You also said that this discussion was meaningless in reply to my point that this discussion is about the logical possibility of free will, not the practical problems of implementation. Such discussion can still have practical implications because if it is shown that something is logically impossible, it also means it is practically impossible.

Let us define "random" as a future event not based on any past events, per your definition in post #32. Let us define "deterministic" as a future event based entirely on past events, also per your definition in post #32. Free will could theoretically be defined as "self-determining". This type of system could effect a future event through a self-contained thought process that is both separate from past events, and at the same time non-random.

Finally, an attempt to respond to what I'm actually arguing. Good work. Maybe this discussion is still salvageable.

What you say is true. A system that is internally deterministic but is not affected by past outside events would be free from outside influence, thus fitting the "free" qualifier. Such a system could also be rational when making decisions that it has knowledge about, which is a necessary condition to be able to talk about "will" in a meaningful sense. So yes, one could call this free will.

The problem is that per the definition you give, such a system would not be able to take its environment into account. It would not be able to make any meaningful choices in a given setting, because it would be unaware of the alternatives it has to choose from. Indeed, it would be unaware that there even is a choice to be made to begin with.

Because of this, the system will be incapable of making rational decisions about the world around it. Since it is meaningless to talk about will if there is no rationality, the will qualifier disappears whenever decisions have to be made that are in any way connected to outside conditions - and I would argue that those are the only kinds of decisions we can observe.

One way to solve this, which I have considered and rejected before for obvious reasons, is to violate the time requirement. If the system can make decisions based on what it will experience in the future it can still make rational decisions about its surroundings without being affected by past events. But I find that solution absurd.

 

[Aradia]

Regardless of there being an "infinite number" of possible futures, God still knows the one that will come about.

If you're equating God to the reader of a Choose your own adventure, then you're still indicating that we don't have free will.

Only as a reader, though. Not as the person making the choice. It would be like the character in the book telling you which option to take, and you trying to tell him not to look behind the shower door, except he can't hear you, and then he gets cut into little bits.

Or something. Anyway. Kinda like that. =)

 

[Aradia]

This is just dishonest. I gave you two adequate examples of hypothetical systems that are essentially hybrid in this sense.

Hardly.

In fact, even from a technical POV my examples still hold, because the functionality of the systems described would not be altered by "true" randomness because of the very way the systems are designed. They don't depend on the causality of the RNG precisely because that would defeat its purpose.

Except you never actually described in any detail how the system would be designed. At any rate, it doesn't matter, because computers aren't hybrid, and computer simulations aren't hybrid, and they're a poor analogy.

(snip)

Finally, an attempt to respond to what I'm actually arguing. Good work. Maybe this discussion is still salvageable.

What you say is true. A system that is internally deterministic but is not affected by past outside events would be free from outside influence, thus fitting the "free" qualifier. Such a system could also be rational when making decisions that it has knowledge about, which is a necessary condition to be able to talk about "will" in a meaningful sense. So yes, one could call this free will.

*sigh*

The problem is that per the definition you give, such a system would not be able to take its environment into account. It would not be able to make any meaningful choices in a given setting, because it would be unaware of the alternatives it has to choose from. Indeed, it would be unaware that there even is a choice to be made to begin with.

Nope. The mechanism of determination is self-contained, but can take input from the outside environment. As opposed to the hybrid system, a computer _is_ a good analogy for this. The CPU is the self-contained thought process. It can take input from keyboards and mice and such, but the actually decision making is separate from the input data.

Because of this, the system will be incapable of making rational decisions about the world around it. Since it is meaningless to talk about will if there is no rationality, the will qualifier disappears whenever decisions have to be made that are in any way connected to outside conditions - and I would argue that those are the only kinds of decisions we can observe.

See above.

 

[Hnefi]

Nope. The mechanism of determination is self-contained, but can take input from the outside environment. As opposed to the hybrid system, a computer _is_ a good analogy for this. The CPU is the self-contained thought process.

But such a system violates the definition you gave before. If the system receives input before making a decision, that input is the cause of the processes that eventually lead to the decision. Then the "thought process" you talked about in your definition is not separate from past events.

Obviously, the internal structure of the system will also play a big role, but ultimately it was the input that led to a decision being made and also the nature of that decision. Basic stimulus-response relationship.

It can take input from keyboards and mice and such, but the actually decision making is separate from the input data.

No, this is never ever the case in a computer. If a program makes decisions based on input, that input data must be taken into account. In doing so, the decision is obviously not separate from said data.

As opposed to the hybrid system, a computer _is_ a good analogy for this.

Are you trying to argue that computers have free will?

 

[elman]

Regardless of there being an "infinite number" of possible futures, God still knows the one that will come about.

If you're equating God to the reader of a Choose your own adventure, then you're still indicating that we don't have free will.

You keep assuming that if God knows what you are going to chose it is not a choice. That is not a good assumption. If I know how someone is going to chose, it does not mean I controled the choice.

 

[elman]

I don't think so. In any idea of a future that hasn't happened (to us) but has happened for any third party observer, the future is as much a fact as the past is. You can't change the past.

But it is not past to us and we can still change it. If we do, that will be the past the third party observed and not some other one. The past the third party observes is the one created by us.

 

[DailyBlessings]

You keep assuming that if God knows what you are going to chose it is not a choice. That is not a good assumption. If I know how someone is going to chose, it does not mean I controled the choice.

That is my problem with this whole argument. It sounds completely illogical to me that knowledge of a choice would somehow cause the choice to disappear. Yet when asked for clarification or support for their idea, proponents just repeat what they have already said and insist that it is logically necessary. I'm not sure how they got this notion but I think in part they are commiting the fallacy of confusing cause and effect, i.e. God always knows in advance what choice A will be, therefore God caused choice A to occur. This similar to claiming that mold was the cause of the bread going bad- correlation =/= causation. Either that or they are inappropriately assuming that the only way God could know the future is if he used science to make a prediction based on a deterministic causal chain. Either way, it is not as logical as apparently assumed.