Talk:Hypothetico-deductive method
|
|
This statement in the current article is incorrect:
- "This problem is related to the problem of induction, and arrises because it is logically invalid to infer a general case – a hypothesis – from any series of specific observations."
Induction cannot be "logically invalid" because induction is non-logical it has nothing to do with logic which is deductive by definition. B 23:10, Dec 11, 2003 (UTC)
Garr. I've gone on a wild goose chase, around and around, and back to the start. Indeed, Bayes does solve this "paradox" which is no paradox.
We solve it by saying "certainly if you show me a non-white thing, and I see it is not a swan, it actually does increase the validity of my hypothesis, just that a swan being white inncreases the validity a lot, and a non-white thing being a non-swan increases the validity of my hypothesis an almost-insignificant amount."
Read the article on the Raven paradox to see how easily it's solved using modern logical and statistical method.
- Did you take a look at the talk page for raven paradox? Banno 11:34, Jun 11, 2004 (UTC)
I would submit that this "paradox" is poorly based on the Method in the first place- at least for the scientifically minded. Philosophers may decide to debate other aspects of the method that are decidely less used. For scientific usage, a hypothesis is: hy·poth·e·sis Audio pronunciation of "hypothesis" ( P ) Pronunciation Key (h-pth-ss) n. pl. hy·poth·e·ses (-sz)
1. A tentative explanation for an observation, phenomenon, or scientific problem that can be tested by further investigation.
(The American Heritage® Dictionary of the English Language, Fourth Edition)
Thus, "All Swans are white" is not a hypotheis. It is a statment of generality of observations. Statements or obbservations alone do not provide a structure for prediction and experimentation - i.e. an Method. With out a method, we are aimless. No?
The statement "all swans are white" is only logically equivalent to "all non-white things are not swans" if it is a double implication, i.e.
Swan <-> White is equivalent to NOT(Swan) <-> NOT(White).
So if your hypothesis is
Swan -> White (which I think is more common), this is actually not equivalent to Not(Swan) -> Not(White).
An observation of a non-white thing will not (dis)prove anything if I'm correct.