= 27d567bc2d48d9f40e9ccc20ea370b15)
gjm comments on Skepticism about Probability - Less Wrong Discussion
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (129)
A sufficiently skeptical position is completely immune to criticism, or to any other form of argument. I don't see what anyone could hope to do about that, beyond not bothering arguing with people who profess such extreme skepticism.
(I am reminded of a little fable I think I saw in an old OB post. Human space travellers encounter an alien planet whose inhabitants have adopted an anti-inductive principle, with the unsurprising result that pretty much everything they do is miserably unsuccessful. The humans ask them "So why do you keep on doing this?" and they say "Well, it's never worked for us before...")
Which means that anti-scepticism is a position taken on faith in the religious sense. It is, after all, the anti-sceptic who claims something can be known.
What I'm looking for is an argument that starts from no assumptions whatsoever but the self-evident, that gets to a justifiable probability theory. That would get around arguments such as the Evil Demon argument.
The sceptic also claims that something can be known: that nothing can be known.
Any conclusion can be reached by choosing the right "self-evident" assumptions; any conclusion can be denied by denying the self-evidence. It is self-evident to John C. Wright that God exists and His angels spoke to him in a vision. It had previously been self-evident to him that no such thing was possible. His rational mind pursues the implications of his latter beliefs as relentlessly as it did his previous beliefs. (His blog makes a fascinating case study in taking ideas seriously.)
What will be self-evident to you? Only you can decide that. You would be better off reading diverse books on the foundations of probability (Jaynes, Feller, Savage, de Finetti, Kolmogorov, ...) and seeing whether their arguments convince you, than asking someone to come up with an argument that you will find compelling.
I highly doubt those authors address the idea of scepticism in the sense of showing from first principles that probability is legitimate, in a way that addresses such things as Descarte's Evil Demon Argument. They are discussing how to implement probability, not whether the entire concept is legitimate.
You are denying the validity of self-evidence here, and admittedly there is a problem with properly establishing self-evident ideas. That's part of my problem, admittedly, and what I am trying to bypass somehow.
Nothing can address the Evil Demon Argument. Descartes thought it was self-evident that he thought, and therefore that he existed, but you can find various modern philosophers, and ancient ones of the Buddhist traditions, who declare that the self does not exist and claim to have none.
As a foundation for knowledge, yes, I am.
Doesn't "self-evident" mean that they don't need establishing?
"Self-evident" in the sense that they don't need any starting assumptions whatsoever. The point I am making repeatedly because others don't seem to get it is that if there is no way to justify the premise that the world exists without resort to assumptions, then we're no better than the people who believe in God on faith.
I am searching for a way to deal with the Evil Demon Argument etc. for that reason. As for said philosophers, they have a different concept of a self from Descartes and so to an extent are talking about different things.
Let us imagine two people. One believes "on faith" that (1) what their senses tell them has some correlation with how things really are, (2) their memory has some correlation with the past, and (3) their reasoning isn't completely random and broken. The other believes "on faith" those three things, and also that every statement in a certain collection of ancient documents is true, that a certain person who lived 2000 years ago was really a god in human form, that our true selves are immortal immaterial entities, and that after our deaths we will be judged and consigned to eternal bliss or eternal torment.
I'm quite happy saying that the first of those people is doing better than the second. S/he needs to assume far less; the things s/he assumes are more obviously true and more obviously unavoidable assumptions; there is less arbitrariness to them.
Both of them, indeed, fail if you judge them according to the following principle: "Everything you believe should be derived from absolutely incontestable axioms with which no one could possibly disagree". But why on earth should we do that?
You really should add a fourth- the principle of induction. You also misinterpret my premise- it is not that nobody could possibly disagree, but that the ideas could not possibly be false, even under an Evil Demon argument.
The problem is the same as the Isolation Objection to Coherentism- that if there is any actual correlation to reality, it is merely by chance rather than through actual evidence. This is because both groups have no basis for their assumptions.
Have you ever seen such a proposition? I don't think that I have. Not a single sage of recorded history has been able to come up with something whose self-evidence convinced everyone. And if someone is unconvinced, how shall you convince them, if it's "self-evident"?
What is this? If you have any unjustified belief, you are identical with someone who pays no heed to rationality at all?
And what does this have to do with probability in particular? You originally asked about probability, so I recommended works on the foundations. Even if none of them persuade you that they are a sound basis, at least you will be informed about the arguments and conceptual structures that people have created, at which point you may be able to productively search for something better.
But now you have broadened this to a requirement for a refutation of the Evil Demon/Matrix scenario. I see no possibility of any such refutation, because sufficient powers can always be attributed to the Demon/Skynet/Lizard Overlords/NSA to explain away any putative refutation. If there is a refutation, you will have to find it yourself.
I mentioned John C. Wright earlier, and there is more to say. He finds the ultimate foundation in the uncaused cause that is the Originator of all causation, the Good that needs no justification because it is the Originator of all that is good, proves their existence by the argument against infinite regress, and recognises them in the world as the Christian God, specifically as preached by the Roman Catholic Church. You could work out from that what his self-evident truths might be, for him to build these arguments on, but his actual self-evident truths are the religious visions that he had. He was never argued into any of this by the arguments that he presents (and neither am I, an atheist).
Self-evidence is a subjective property of a belief. The experience of self-evidence is the absence of experience of justification for the thing believed.
I explained my context was the refutation of philosophical scepticism in general- what I was after should have been clear.
1- You assume that the criterion of self-evidence should be based on being universally convincing. Why should this necessarily be so? Self-evidence comes when the contrary proposition simply doesn't make sense, as it were (simplistic example: free will). The question is how to deal with that with regards to demonstrating the validity of probability/induction. 2- Because the fundamental starting assumption is unjustified, we are no more justified in believing we know the truth than the people who believe in God on faith.
"Free will" is a concept, not a proposition. What is the proposition about free will that you are claiming to be self-evident, and its opposite "not making sense"?
Such an experiment doesn't exist. It's not self-evident that we don't live in a simulation in which strange things can happen.
I know it's not self-evident that we don't with certainty (and I said argument, not experiment), but I'm not trying to get to the conclusion we don't live in such a simulation- only that it is improbable.
Simple: The fact that we don't see strange things happening is bayesian evidence that we don't live in a world where that is possible.
As I already mentioned, it is probability itself which must be justified in the first place. How do you do that?
What assumptions am I granted? Can't argue anythin' from nuthin'. Even "I think" is an assumption if logic is.
This is the problem which must be dealt with. Rather than assume an assumption must be correct, you must somehow show it will work even if you start from no assumptions.
Your universal propositional calculus might not be able to generate that proposition, but my calculus can easily prove: Yours won't generate any propositions if it has no axioms.
This is precisely the problem. I was posting in the hopes of finding some clever solution to this problem- a self-proving axiom, as it were.
"Self-evident assumptions" sounds suspiciously like "axioms", yet axioms are apparently not what you seek. What exactly are you hoping to find? What would an acceptable "self-evident assumption" be?
Assumptions that can be demonstrated to be true in any possible universe.
As it has been fably noted, the assumption that cannot be verified is the induction (from past into the future). It is well-known in the area of philosophy (something with the words “induction is impossible”) that if something worked well before it won't necessarily work further (unless you assume apriori that induction is possible).
One can freely claim anything (any prediction / hypothesis) if induction is not assumed; the prediction happening to be false would not say anything about the validity of the induction and, thus, of the relevant models.
And the probabilities in particular can be justified from induction (additionally, some of the axioms are relevant to the mathematical representation of the probabilities, not to the notion itself). As well as Ockham's principle, and many other relevant ideas.
You have a point. Then how do you justify induction?
You don't (or “I don't”, if that's what you meant).
You could say something like that: “if induction is impossible then decision-making and communication are futile”.
However, that by itself does not disprove / dejustify claims that induction is possible but in other ways / with exceptions (on the lines of “the induction is possible unless applied to god^W magic”).
If you have no non-circular basis for believing in induction, surely it is irrational?
So, what are you trying to say?
"Better" isn't a function of the real world anyway- I'm appealing to it because most people here want to be rational, not because it is objectively better.
What do you mean by "rational" is not a binary?
“Better” / “preferable” / “utility” / … is necessary for “usefulness” e.g. “usefulness of this communication” (and also for decision-making).
By “not a binary” I mean the division is not into “rational” / “non-rational”, but into “more rational” / “less rational”; where “rational” is relevant to the aforementioned “better” (with regards to efficiency of optimization and also forms of communication).
… vaguely speaking.
On thought, my response is that no circular argument can possibly be rational so the question of if rationality is binary is irrelevant. You are mostly right, though for some purposes rational/irrational is better considered as a binary.
You apparently don't think it's self evident that induction works. Do you think that it's self evident that deduction works (like Descartes did)? If you do, why? If you met someone who was willing to accept the premises of a deductive argument but not the conclusion, like the tortoise in this parable, how would you convince the person they were wrong? The only way to do it would be using deductive arguments, but that's circular! So it seems that deductive arguments are just as "unjustifiable" as inductive arguments.
But if you reject deductive arguments as well, then you can't do anything. Even if you start with self-evident premises, you won't be able to conclude anything from them. Perhaps this is a hint that your standards for justification are so high that they're effectively useless.
A premise isn't self-evident because anybody whatsoever would accept it, but because it must be true in any possible universe.
Deductive arguments aren't self-evident, but for a different reason than you think- the Evil Demon Argument, which shows that even if it looks completely solid it could easily be mistaken. There may be some way to deal with it, but I can't think of any. That's why I came here for ideas.
You claim my standards of justification are too high because you want to rule skepticism out- you are implicitly appealing to the fact skepticism results as a reason for me to lower my standards. Isn't that bias against skepticism, lowering standards specifically so it does not result?
There are all kinds of things that are true in every possible universe that aren't self-evident. Look up "necessary a posteriori" for examples. So no, self-evident is not the same as necessary, at least not according to a very popular philosophical approach to possible worlds (Kripke's). More generally, "necessity" is a metaphysical property, and "self-evidence" is an epistemic property. Just because a proposition has to be true does not mean it is going to be obvious to me that it has to be true. Even Descartes makes this distinction. He doesn't regard all the truths of mathematics to be self-evident (he says he may be mistaken about their derivation), but presumably he does not disagree that they are necessarily true. (Come to think of it, he may disagree that they are necessarily true, given his extreme theological voluntarism, but that's an orthogonal debate.)
As for your question about standards: I think it is a very plausible principle that "ought" implies "can". If I (or anyone else) have an obligation to do something, then it must at least be possible for me to do it. So, in so far as I have a rational obligation to have justified beliefs, it must be possible for me to justify my beliefs. If you're using the word "justification" in a way that renders it impossible for me to justify any belief, then I cannot have any obligation to justify my beliefs in that sense. And if that's the case, then skepticism regarding that kind of justification has no bite. Sure, my beliefs aren't justified in that rigorous sense, but if I have no rational obligation to justify them, why should I care?
So either you're using "justification" in a sense that I should care about, in which case your standards for justification shouldn't be so high as to render it impossible, or you're using "justification" in Descartes's highly rigorous sense, in which case I don't see why I should be worried, since rationality cannot require that impossible standard of justification. Either way, I don't see a skeptical problem.
It seems we're using different definitions of words here. Maybe I should clarify a bit.
The definition of rationality I use (and I needed to think about this a bit) is a set of rules that must, by their nature, correlate with reality. Pragmatic considerations do not correlate with reality, no matter how pressing they may seem.
Rather than a rational obligation, it is a fact that if a person is irrational then they have no reason to believe that their beliefs correlate with the truth, as they do not. It is merely an assumption they have.