The idea the “correlation does not imply causation” is common and prevalent. (And, among the populace in general, probably ought to be more prevalent.) But I wish to split hairs, as I often do.
Correlation in some instances does not imply causation. However, correlation that persists in the face of all our efforts to part the two is, in fact, synonymous with what we call “causation”. Hume prefered to call causation “constant conjuction” for this very reason. That is, when we see that one event always “causes” another, what we are really seeing is that one event has always been “constantly conjoined” to the other.1
Entries (RSS)
July 2nd, 2007 at 7:34 am
I’ve usually seen this as “correlation does not equal causation”, which is probably a more useful way of phrasing it. It’s perhaps the biggest problem of experimental biology; systems are so complex that /proving/ A causes B is very, very difficult indeed.