06-13-2010 , 04:13 PM
Why shouldn't it work the other way around?

If one could (if one were a suitable epistemic agent to be able to) predict the future based on the present states of the universe and the laws of nature, why couldn't one retrodict the past based on that same information?
06-13-2010 , 04:30 PM
Quote:

And we know what that means - two propositions are contradictory when (from one definition) a statement of the form "a and not a" can be inferred from those propositions.

Free will implies it's possible to choose the toast.

Determinism implies it's not possible to choose the toast.

Looks like an "a and not a" to me.

PairTheBoard
06-13-2010 , 06:07 PM
Quote:
Originally Posted by Aaron W.
Maybe you don't realize what you're saying about Turing machines. You've got issues with the question of finite/infinite machines here. Are you assuming that there are only a finite number of states of the universe? Or will it be necessary to have some sort of "infinite" precision in order to successfully compute the state at some other time?

Turing machines are only allowed to have a finite program that works with a finite alphabet. Furthermore, while you have a (semi-)infinite tape, the only initial tapes that are allowed must be finite (if you ignore the blanks).
I'm assuming that each state is finite, and that the process by which one state can be predicted from any other state is also finite. Or I should say, these are the consequences of the assumption that a Turing machine can successfully compute one state given any other state. There's no implication that the total set of states is finite or even countable.

(Incidentally, the assumption of a finite universe is itself another reason why this is a "narrow" form of determinism.)

Quote:
This doesn't seem to be a requirement of fatalism. Are you sure you're not making up stuff as you go again?
I was fairly sure. But now that I think of it, it might be possible to have a case where we can alter the parameters of a state of the universe and the Turing machine will give the same result. Is it possible for this to be true for all "choice" parameters? I don't know. If yes, then fatalism is possible. If no, then it isn't.

My intuition says "no," but I don't actually have any basis for that, so maybe I'm wrong.

Quote:
I have no idea how to interpret this with your current definitions.
Any s associated with a time in the future can be used to derive any s associated with a time in the past.

Being able to predict the past given the future isn't a feature of all forms of determinism. In fact, some forms of determinism don't feature any form of predictability.

Quote:
I don't even know what this means.
Some people don't believe that everything in the universe can be described as a single vector. Much less a finite vector. Or that the universe can even be (perfectly) described mathematically at all.
06-13-2010 , 06:08 PM
Quote:
Why shouldn't it work the other way around?

If one could (if one were a suitable epistemic agent to be able to) predict the future based on the present states of the universe and the laws of nature, why couldn't one retrodict the past based on that same information?
That's why I think my view of time makes the most sense.

But clearly some people view the future as something "special," the present as something "even more special," and the past as something "not so special."
06-13-2010 , 06:24 PM
Quote:
Originally Posted by PairTheBoard
Free will implies it's possible to choose the toast.

Determinism implies it's not possible to choose the toast.

Looks like an "a and not a" to me.
Changing "free will" to "choice," this would do the trick. Except that I don't think you can find any definition of possibility based on which both of those statements are true. Certainly, nobody has actually demonstrated that either follows.

At a bare minimum, you are making massive assumptions about possibility.

And let's talk about the type of possibility you're referencing. Choice does imply some forms of possibility, and determinism does negate some forms of possibility. However, there is no single form of possibility that choice implies and determinism negates. What you have here is, at best, an argument from equivocation of the term "possible."
06-13-2010 , 06:27 PM
Wait...what?

If you're a Dennettian compatibilist then you think that you can perfectly predict the future and retrodict the past based on full knowledge of the states (and microstates) of the universe and the laws of nature.

This implies that next week, although you may think that you have a choice at breakfast between toast and a bagel, you will only be able to pick the bagel (since your action next week is determined for you to eat the bagel); yet you explicitly denied this: you claimed that you really DO have the choice between toast and bagel.

Also, PTB basically restated my position concerning the incompatibility thesis implying an internal inconsistency (what I've been referring to as a contradiction) for compatibilism...and what is your response?
06-13-2010 , 06:34 PM
Quote:
Changing "free will" to "choice," this would do the trick. Except that I don't think you can find any definition of possibility based on which both of those statements are true. Certainly, nobody has actually demonstrated that either follows.

At a bare minimum, you are making massive assumptions about possibility.

And let's talk about the type of possibility you're referencing. Choice does imply some forms of possibility, and determinism does negate some forms of possibility. However, there is no single form of possibility that choice implies and determinism negates. What you have here is, at best, an argument from equivocation of the term "possible."
HAH, I have a publication in that journal.

From my quick survey of that paper, his arguments do not apply. No one is denying counterfactual possibility: had things been different, then things could have been different. Even the hard determinist can accept that (Dennett's "I Could Have Done Otherwise...So What?" article is all about whether someone could or could not have done otherwise is not sufficient to determine whether they're responsible). So, that paper that you're referencing really has nothing to do with our current discussion!

I have suggested that determinism rules out all sorts of metaphysical possibility which we consider to be a necessary condition for choice. Basically, as a simplification, there is NO sense in which it is possible for you to choose the toast except as a counterfactual: had the conditions of the universe or the laws of nature been different, then in THOSE possible worlds, you could choose the toast. But, we're not in those worlds and in the actual world, it is impossible for you, in a week, to do anything other than eat the bagel.

It's a very weak move if your account of responsibility avows the determinist thesis and is forced to awknowledge that in the actual world you couldn't have done otherwise -- and that we can perfectly predict and retrodict future and past actions, respectively, based on full knowledge of the states (and microstates) of the universe and the laws of nature -- and that the only sense that you can have 'choice' is that, well, in other possible worlds, had things been different, then I could 'choose' the toast.

What say you, sir?
06-13-2010 , 06:56 PM
How exactly are you defining "choice"? If you defined it somewhere else in the thread could you please point it out? Thanks
06-13-2010 , 07:02 PM
Quote:
Originally Posted by Matt R.
How exactly are you defining "choice"? If you defined it somewhere else in the thread could you please point it out? Thanks
Sure, the cognitive process of selection from a range of options.
06-13-2010 , 07:12 PM
Quote:
Sure, the cognitive process of selection from a range of options.
Okay, using that definition of choice, respond to my post(s). Specifically, the bagel/toast "choice." Based on there being only one thing that you will do next week at breakfast (toast vs bagel: you'll eat the bagel) in the actual world, how is it that you're "choosing" from a "range of options"?
06-13-2010 , 07:24 PM
Quote:
Sure, the cognitive process of selection from a range of options.
OK, so when you say "cognitive process", what do you mean exactly? Is a cognitive process purely a function of the workings of the central nervous system or is something else going on? (And when I ask you to explain what you mean by cognitive process, I mean in terms of determinism)
06-13-2010 , 07:26 PM
Quote:
Originally Posted by Matt R.
OK, so when you say "cognitive process", what do you mean exactly? Is a cognitive process purely a function of the workings of the central nervous system or is something else going on? (And when I ask you to explain what you mean by cognitive process, I mean in terms of determinism)
Just accept some sort of physicalist answer that it's the brain doing something with some input and producing some output. No fancy supervenience or what-not. Some functionalist answer is probably the way to go for the determinist.
06-13-2010 , 07:32 PM
Quote:
Just accept some sort of physicalist answer that it's the brain doing something with some input and producing some output. No fancy supervenience or what-not. Some functionalist answer is probably the way to go for the determinist.

That's what I think the answer pretty much has to be for a determinist, which is why I don't understand the difficulty with the domino analogy. That's why I'm wondering what the heck madnak is talking about with respect to "choice" and "cognitive process"... I'd like him to be more precise in defining what he means by "cognitive process" because I really have no clue what he's saying at the moment.
06-13-2010 , 07:41 PM
I don't think 'cognitive process' really matters here...it's more important to be clear on what is meant by 'choice' and having a set of options from which to select.

To me, "selecting from a set of options" directly implies that I could in the actual world choose A or B.

But, in a determinist system, agent S will be determined to select A which means they were never able to select B. This is the toast/bagel example. Based on full knowledge of the current state (and microstates) of the universe and the laws of nature, one could predict what madnak will have for breakfast next week. Let us suppose that madnak either has a bagel or toast for breakfast. He thinks that he's 'choosing' on each morning but we can already know what he'll do next week: select the bagel. Now, in what sense is the toast an "option" if he can't possibly select it?

madnak needs to answer this. Either he needs to deny that determinism implies the predictability and determination of future events, or he has to accept that there's only one 'option' even though there may be the illusion of multiple options. If the former, then I don't know what sort of determinism he ascribes too (it's certainly not Dennettian); if the latter, then he needs to explain how it makes sense to call this a 'choice' when he defines choice to be selecting from a (real) range of options: if he accepts the right disjunct then he appears to contradict his definition of choice.
06-13-2010 , 07:48 PM
Quote:
HAH, I have a publication in that journal.

From my quick survey of that paper, his arguments do not apply. No one is denying counterfactual possibility: had things been different, then things could have been different. Even the hard determinist can accept that (Dennett's "I Could Have Done Otherwise...So What?" article is all about whether someone could or could not have done otherwise is not sufficient to determine whether they're responsible). So, that paper that you're referencing really has nothing to do with our current discussion!
No one is denying counterfactual possibility? So in that case, it is possible. I could have chosen differently.

You've been saying it's "not possible" for me to take the toast, which without qualification I take to mean "it is a necessary falsehood that I will eat the toast," "there are no possible worlds in which I will eat the toast," and so on.

Now you're saying that there is some sense in which it is possible that I will eat the toast. According to the definitions we've been working with, that's enough to establish a "choice."

Quote:
I have suggested that determinism rules out all sorts of metaphysical possibility which we consider to be a necessary condition for choice.
Wait, we consider metaphysical possibility to be a necessary condition for choice? I never got the memo. But even so...

Quote:
Basically, as a simplification, there is NO sense in which it is possible for you to choose the toast except as a counterfactual: had the conditions of the universe or the laws of nature been different, then in THOSE possible worlds, you could choose the toast. But, we're not in those worlds and in the actual world, it is impossible for you, in a week, to do anything other than eat the bagel.
Well, my footing here isn't what it could be, but SEoP paints a more nuanced picture. And it describes metaphysical possibility under exactly the heading of possible worlds ("Φ is metaphysically possible if and only if Φ is true in some metaphysically possible world"). Wiki agrees, and describes the type of possibility you're talking about as "temporal possibility." In fact, none of the accounts on either page seem to agree with what you're saying here - maybe I'm missing something. And some of the accounts listed under the SEoP entry clearly contradict your position.

Regardless, it's not at all clear that free will changes things. If we're not talking about a multiverse, then there will only be one actual future. Either I will choose the toast or I will choose the bagel. There is only one reality. Ultimately, there is only one "real" choice, only one choice I'll end up making, if there's only one actual reality. Whether that choice is realistic or not doesn't change the fact that I will make it. So temporal possibility deals with an alternate "possible world" (in which I make the choice that I didn't actually make) just as much as any counterfactual possibility, putting it firmly in that category.

So the only basic difference I can see in possibility between determinism and indeterminism is in epistemic possibility, in that an indeterministic event is by definition impossible to predict (and ergo an indeterministic choice always involves epistemic possibility), however in determinism it is possible in theory to actually predict the future perfectly, and so we can't 100% rule out the epistemic impossibility of my choosing toast. On the other hand, this is somewhat weak as even though it's possible in theory to predict whether I'll eat toast, that doesn't imply that I (or anyone else) actually know what I'll have for breakfast until the moment the decision is made. That means that it can be epistemically possible for me to choose the toast in a deterministic universe.

Quote:
It's a very weak move if your account of responsibility avows the determinist thesis and is forced to awknowledge that in the actual world you couldn't have done otherwise -- and that we can perfectly predict and retrodict future and past actions, respectively, based on full knowledge of the states (and microstates) of the universe and the laws of nature -- and that the only sense that you can have 'choice' is that, well, in other possible worlds, had things been different, then I could 'choose' the toast.
I define "couldn't have done otherwise" logically as "did otherwise in some possible world." Therefore, according to how I use the terms, "couldn't have done otherwise in the actual world" is nonsensical. I view "could" and "couldn't" as references to possible worlds. To my knowledge, this is a fairly standard approach to modality. Metaphysical possibility is even defined as such in that link. Are you telling me otherwise?

Quote:
What say you, sir?
I say that, at a minimum, your position depends on your particular interpretation of modalities.

I suppose if you claim that it is necessarily false that I will eat the bagel in a deterministic universe, then you are justified in claiming that free will and determinism are contradictory. Otherwise, I would argue that while you may, based on your own assumptions about possibility, rule out the possibility of compatibilistic free will in your own beliefs, you do not have a basis for concluding that compatibilistic free will is self-contradictory.

Either way, there is no reason to believe that someone adhering to another of the many differing approaches to reasoning about possibility will accept your conclusions. Also according to some definitions of "could choose otherwise" and "is possible to choose otherwise" (namely those referencing counterfactual possibility, if nothing else), determinism clearly is compatible with choice (and the definitions of choice and possibility that I use fall under that category - which may give us a purely semantic disconnect or a more substantive disagreement, it's not obvious to me right at the moment which it is).
06-13-2010 , 07:51 PM
Quote:
Okay, using that definition of choice, respond to my post(s). Specifically, the bagel/toast "choice." Based on there being only one thing that you will do next week at breakfast (toast vs bagel: you'll eat the bagel) in the actual world, how is it that you're "choosing" from a "range of options"?
I'm going through a process of consideration of various actions (I believe they can be classed as possible actions, however let's call them "legal" actions to help an analogy I'm about to make), and eventually taking one of those actions.

This is similar to how a computer selects a move in chess; it enumerates every legal move (every "legal" action), it weighs the different moves against one another, and then it selects one move. The only reason the computer is not choosing (by how I view choice) is because the computer is incapable of cognition.

Computers are clearly deterministic, so if computers are capable of selecting from a range of options, then I assume it's clear that deterministic humans are also capable of doing so.
06-13-2010 , 07:55 PM
Quote:
Originally Posted by Matt R.
OK, so when you say "cognitive process", what do you mean exactly? Is a cognitive process purely a function of the workings of the central nervous system or is something else going on? (And when I ask you to explain what you mean by cognitive process, I mean in terms of determinism)
I think it's "cognitive" if it involves awareness and deliberation. As to whether awareness and deliberation can be described purely as a function of the nervous system, I expect that they can in theory but sometimes it seems like they can't.

At any rate, I'm basically an agnostic on that - while I don't think awareness is "special" in exactly the way that some people do, and while I don't include any "specialness" in my view of the world, and even make statements that may appear very strong in this regard, there are subtleties. I'm a solipsist for instance, so my view of consciousness isn't quite what people may assume it is (in particular I see a huuuuuge distinction between my consciousness and that of another person, and I think it's possible that nobody other than myself makes choices at all).
06-13-2010 , 07:57 PM
The point is that your brand of "possibility" is empty. It's not sufficient for responsibility: that's the very position that's up for debate.

Turning to possible worlds will not save you here: it's the wrong move. A Libertarian holds that in the actual world it's an open question whether you'll choose the bagel or the toast. But, for a compatibilist, soft determinist, or hard determinist, it is not: the answer is that you will not choose the toast. We're not talking about possibility in terms of counterfactuals and possible worlds.

You need to understand that "metaphysical possibility" =/= "counterfactual possibility." That's only one interpretation and thanks to David Lewis, the debate has moved on PAST that description of possibility. This sort of appeal to authority of (LOL) wiki and (less LOL) SEP indicates that you don't know the debate well enough to know that this has happened: interpreting possibility in terms of possible worlds or counterfactuals has been rejected.

Furthermore, it is not interpreted in terms of epistemic possibility. Given full knowledge of the states (and microstates) of the universe and the laws of nature, no one denies that one would know the future given determinism. Epistemic possibility requires that the agent not have full knowledge of the states and/or laws and/or the computational power to make the prediction. But, that is not what is at issue here: it's largely irrelevant.

This is precisely the point: for determinism, it is necessarily false in the actual world that you will eat the toast. Furthermore, it is necessarily true in the actual world that you will eat the bagel next week.

Do you agree or disagree?

Answer that and we'll go from there...
06-13-2010 , 07:58 PM
Quote:
Just accept some sort of physicalist answer that it's the brain doing something with some input and producing some output. No fancy supervenience or what-not. Some functionalist answer is probably the way to go for the determinist.
Well, that's probably the easiest way to do things, so we may as well assume that. But, I don't know. Assuming that could get me into trouble down the line, as some of my unexamined assumptions (unexamined for the purpose of this thread) may be related to the fact that I'm not commited to a physicalist view of the subject (and I'm actually opposed to a strict physicalist view).
06-13-2010 , 07:58 PM
Quote:
I'm going through a process of consideration of various actions (I believe they can be classed as possible actions, however let's call them "legal" actions to help an analogy I'm about to make), and eventually taking one of those actions.

This is similar to how a computer selects a move in chess; it enumerates every legal move (every "legal" action), it weighs the different moves against one another, and then it selects one move. The only reason the computer is not choosing (by how I view choice) is because the computer is incapable of cognition.

Computers are clearly deterministic, so if computers are capable of selecting from a range of options, then I assume it's clear that deterministic humans are also capable of doing so.
"Computers are clearly deterministic" may beg the question but let's put that aside.

Your analogy is not tight. Here, we need the computer to reduce the "legal" actions to some set where all options are equally utility maximizing: what will it choose and how?

If a computer must choose between 2 equally weighted options, how will it do it?

(Hint: I'm being socratic)

edit: But please respond to the other post...this is not important right now.
06-13-2010 , 07:59 PM
Quote:
Well, that's probably the easiest way to do things, so we may as well assume that. But, I don't know. Assuming that could get me into trouble down the line, as some of my unexamined assumptions (unexamined for the purpose of this thread) may be related to the fact that I'm not commited to a physicalist view of the subject (and I'm actually opposed to a strict physicalist view).
Then use a powerful philosopher's trick: assume for the sake of argument. If you realize that you need to re-examine the assumption because it led to something unpalatable, then cross that bridge when you get there.
06-13-2010 , 08:05 PM
Quote:
To me, "selecting from a set of options" directly implies that I could in the actual world choose A or B.
Again, you'll have to explain to me exactly what you mean by "could in the actual world choose A or B." To me, "could" means "did in some possible world." Clearly that's not what you mean by "could," so what do you mean?

Furthermore, why is it relevant? Why are possible worlds somehow "insufficient?" Apparently because of your view of responsibility.

Quote:
But, in a determinist system, agent S will be determined to select A which means they were never able to select B. This is the toast/bagel example. Based on full knowledge of the current state (and microstates) of the universe and the laws of nature, one could predict what madnak will have for breakfast next week. Let us suppose that madnak either has a bagel or toast for breakfast. He thinks that he's 'choosing' on each morning but we can already know what he'll do next week: select the bagel. Now, in what sense is the toast an "option" if he can't possibly select it?
In two senses: First, he selects it in some possible world. Second, it is one of the courses of action that his decision process takes into account and evaluates.

Quote:
madnak needs to answer this. Either he needs to deny that determinism implies the predictability and determination of future events, or he has to accept that there's only one 'option' even though there may be the illusion of multiple options. If the former, then I don't know what sort of determinism he ascribes too (it's certainly not Dennettian); if the latter, then he needs to explain how it makes sense to call this a 'choice' when he defines choice to be selecting from a (real) range of options: if he accepts the right disjunct then he appears to contradict his definition of choice.
The determinism I'm defending in this thread is very different from the "determinism" I actually accept. And I think you are going overboard in your evaluation of Dennett's compatibilism. He is extremely careful about the claims he actually makes (even if it may seem like he's making a certain claim, often if you examine exactly what he says, he's not saying what he seems to be saying).
06-13-2010 , 08:06 PM
Quote:
Then use a powerful philosopher's trick: assume for the sake of argument. If you realize that you need to re-examine the assumption because it led to something unpalatable, then cross that bridge when you get there.
Well, that's what I did with determinism, but I think it turned into a huge mess in the middle of this thread (probably partly because I spent a couple of weeks outside the forum). So I'm a bit gun-shy. But yeah, we can assume physicalism. If anything it seems to make my position easier.
06-13-2010 , 08:08 PM
I'm using Dennett against you. That's why it's relevant. Dennett doesn't think that it's really informative to know whether an agent could, or could not have done otherwise in some other possible world. (There are further requirements to make it informative.)

But, the reason why it's important is that whether or not an agent could have done otherwise in some other possible world does not distinguish between the libertarian and the compatibilist! Therefore, your distinction does not matter: it's irrelevant.

Instead, the difference is that the libertarian believes that an agent could have done (or could do) otherwise in the actual world.

Do you understand that? If yes, we'll move on...if not, discuss.
06-13-2010 , 08:12 PM
Quote: