Re: Falsifiability

H. M. Hubey (
22 Oct 1995 18:38:08 -0400

Eugene Thomas <> writes:

I tried replying to this but for some strange reason the
email went to the wrong guy. I still don't understand how
it happened.

>Although I have taught lessons about scientific method on a number of
>occasions, "falsifiability" is new to me. From context, I gather that
>to be acceptable to the scientific community, a model (I use the word
>"model" to avoid the apparent emotional charge of "theory", "hypothesis"
>and "speculation".) must must be accompanied by statements of tests which
>can be performed. I have tried to find a beginning science text which
>might present this idea at an introductory level. I have failed. I also
>attempted to look at histories of science.

The problems started (AFAIK) with Hempel's Raven Paradox. Until
then the idea was that an experiment could be conducted (in
physics naturally) which would verify or falsify a theory. Then
the verification was dropped and replaced by confirmation.

The raven paradox then closed that off too since logically
the confirmation amounted to something really strange. It goes
like this; The statement "Ravens are black" can be written simply

R => B

so saying that every time you see a black raven you add more
confirmation to the statement is logically equivalent to

B ' => R '

which is "Non-black things are non-Ravens." But then every time
you see a red parrot, a yellow banana or a blue car you'd also
be confirming "Ravens are black".

So only falsifiability was left. That was due I guess mostly
to Karl Popper. He has a lot of books. In short, it says that
whatever passes off as a theory must at least be falsifiable
since otherwise we can have theories like this:

* The devil is the cause of all evil.
** We cannot detect the devil.

We could never disprove that the devil doesn't cause
all evil since we cannot detect if it's around or not.


Regards, Mark