I was reading some stuff on how complexity affects causal inferences and came upon this interesting paradox. From what other scholars have to say, the issue is still debated and there is no conclusion one way or the other.
The paradox was originally penned by a German scholar, Hempel. His hypothesis was simple. His statement was that "all ravens are black". From a statistical standpoint, if this null is hypothesized as being credible the alternative would be "all non black things are not ravens". Therefore if we observe a red cricket ball, this lends further support to our credible null that all ravens are black. The question is intuitively what is the relationship between the color of a cricket ball and the color of a raven? I cant think of any!
Now this clearly indicates that using the inductive approach fails intuition. A fundamentally simple argument which can put a lot of empirical research under tremendous pressure!
PS: http://en.wikipedia.org/wiki/Raven_paradox
Thursday, May 31, 2007
Hempels Paradox
Posted by Mad Max at 12:29 PM
Subscribe to:
Post Comments (Atom)
Posts
5 comments:
Can you please elaborate on this? I can google and understand what you mean but it would be easy this way :)
@mad-max: quite interesting...
@brainwaves: My understanding...How do you say that some statement or observation to be true? Is it because it is true or that you dont find a counter example to prove that it is not true? As we all know, many theories in physics are true because we dont have a counter example to prove otherwise...If "all ravens are black" is the statement and if all we see in our day to day life are cricket balls, which is the converse (all non black things are not ravens), then the statement implication is true though there is no relation between a cricket ball and a raven whatsoever....
very interesting. But I dont' see where the paradox is... without a sufficient data set it is impossible to make a generalisation isn't it?
@ Brainwaves: Mindframes has done what I should have done in the main blog
@ Mindframes: Thanks for the succinct explanation.
@ Mano: hmm..am not sure I understand your comment Mano...the idea is you can have infinite observations, but the fact that whatever you observe (like the fact that you order chinese food) adds to the probability that all ravens are black...therefore data is not an issue...the idea behind the paradox is constructing clear hypotheses...to be more precise, it has to do with Necessary and Sufficient conditions...well goes way back to our high school calculus but it plays such an important role when it comes to thinking about issues
@mad max: My bad. I get it now (had to do some reading- getting a bit rusty).
Post a Comment