Occam's Proof

[Keneke]

I shall state my theory of Occam's Razor as a proof. I make no claims as of yet, as it all is simply ruminations from a severly damaged mind, namely mine.

Let's say I theorize that all other things being equal, parsimony rules. Not just as a suggested course of action, but as a hard and fast rule. Of course, this brings in such unquantifiable variables such as the equality (and indeed knowledge) of "all other things".

SIZE=1]As an aside, let me say that as our knowledge increases, we forever approach the concept in equality of all other things.

lim (knowledge) = "all other things"
kn->inf.
[/SIZE]

The leap here (and this is complete conjecture) is that perhaps that parsimony not only suggests rightness, but proves rightness. It doesn't do that, as far as we know, but perhaps it could be proven or disproven? Is there even one proof against it? How easy it would be if it were, for this line of maddening reason could be struck down at once. It is even worth attempting? Perhaps it could be mathematically shown that

lim (Parsimony) = correctness
kn->inf.

And Occam's Razor could become Occam's proof, thereby disproving God, apple pie, and the perfect marriage. ;)

Of course this is mere conjecture and folly, but it still makes me wonder why a philosophical effort like this has never been attempted. Is it ludicrous from the start?

[Sundog]

Originally posted by Keneke
Is it ludicrous from the start?

Yes.

You're from 1930. I introduce you to Eliza. Is it more parsimonious to believe that there are millions of tiny electronic decisions taking place in a microscopically small space, incomprehensibly quickly, creating the illusion that there is someone there, or is it more parsimonious to believe that there's simply a person on the other end fooling us?

[csense]

Originally posted by Keneke


"And Occam's Razor could become Occam's proof, thereby disproving God..."



You might be interested to know that Ockham was a nominalist. He did not believe in universals, except of course one...God.

[Keneke]

Originally posted by Sundog


Yes.

You're from 1930. I introduce you to Eliza. Is it more parsimonious to believe that there are millions of tiny electronic decisions taking place in a microscopically small space, incomprehensibly quickly, creating the illusion that there is someone there, or is it more parsimonious to believe that there's simply a person on the other end fooling us?

First, let's straighten up your argument? Instead of 1930, let's just say "no knowledge of the existence of computers". Does that work just as well? Now then, I hypothesized that the limit (as knowledge approaches infinity) of parsimony is correctness. That is (if I remember my college Calculus correctly) the more I know, the better choice I make. Having no knowledge of computers takes us so far from the limit we want to go towards that we could hardly deduce a thing from it.

An example of that would be the limit of 1/n as n->inf. if n=1, (one thing we learn, such as introducing me to Eliza), then the answer is 1, but the limit approaches zero as we learn more and more.

I think the disconnect here is that I did not want to postulate that Occam's Razor be fully and blindly used as a proof, and if I stated such in the beginning, that was misleading and I apologize. What I am trying to find out is if parsimony can be used as proof in a special situation.

It's like that "God of the Gaps" discussion. How long will we have evidence of God before we have proof he does not exist? (Or does, for that matter...) Is it more likely we'll simply forget about gods in 10,000 years with the matter still unsolved? IS there a point along the line, or at the end of an infinte line (impossible, I know) where choosing A over B is concretely, logically possible?

It still hasn't been disproven. I do have the intuitive, simple thought that it's not right at all, but still it nags me. Are we forever to stare at the gap in wonder, no matter how infintesimally small? Or can we bridge the gap logically adn with proof?

[Pyrian]

In biological studies, 3 standard deviations or somewhat greater than 98% accuracy IIRC is generally considered "proof"; or at least proof enough to release a drug!