Nick to Owen: ... > Yet, just as you could never get the world to > agree that emotionality was just the number of fecal boluses left by a rat > in an open field maze, you will never get the world to agree that entropy is > just the output of a mathematical formula. They might say, "that is a > useful measure of entropy, but that is not what it IS." To put the matter > more technically, no matter how much reliability a definition buys you, it > still does not necessarily buy you validity. The same point might be made > about f=ma. (I fear being flamed by Bruce, at this point, but let it go.) > Non fingo hypotheses and all that. One could, like a good positivist, > simply assert that a thing IS that which most reliably measures it, but few > people outside your field will be comfortable with that, and everybody, even > including your closest colleagues, will continue to use the word in some > other sense at cocktail parties. It was my position that the lab bench > meaning and the cocktail meaning have some common core that we have some > responsibility to try to find.
This is a very pure example of the semantic drift that drives me crazy, in that "the lab bench meaning" was the *first* meaning: the word DID NOT EXIST before it was coined (in its adjectival form, in German, by composing badly-understood-by-its-coiner morphemes from Greek, by the physicist Clausius) in 1865. Tait (an early knot theorist and somewhat of a religious nut, as well as a thermodynamic theorist) brought it into English three years later, but changed its sign (more or less). By 1875 Maxwell had changed it back to what it now is. During this period of time the concept expressed by "entropy" became clearer, as did the whole field of thermodynamics, and eventually a good mathematical formalism for it developed--"good" in the sense that it "made sense" of results from "the lab bench" by reducing downwards (if I have your phrase right? I dunno, maybe upwards, or both ways?) so as to (1) define "entropy" of a macroscopic system in terms of the statistical behavior of the ensemble of microscopic entities participating in that system, and (2) facilitate calculations (some exact, some asymptotic) of the "entropy" (and similar thermodynamic quantities) which (3) often agreed with "lab bench" observations. When Shannon came along to study signals and noise in communication channels, he had the insight to see that *the same mathematical formalism* could be applied. He did *not* have the insight (or dumb luck) of Clausius, so he overloaded the already-existing Common English word "information" with a new, technical, mathematical meaning. *That* rather quickly allowed visionaries, hucksters, and cocktail partiers to talk about "information theory" without understanding much or any of its technicalities. It also (I suspect; but here I am arguing ahead of what data I happen to have around, so this may merely be my default Enraged Bloviator talking) encouraged the same gangs of semantic vandals to appropriate the word "entropy" to their various malign uses. (For what it's worth, the OED doesn't have citations of non-specialist uses of thermodynamic "entropy" until the mid 1930s--by Freud [as translated by a pair of Stracheys, not by Jones] and a Christian apologist; non-specialist uses of information-theoretic "entropy" appear to hold off until the mid 1960s.) So, to whatever extent the vernacular ("cocktail party") meaning(s) of entropy has or have a common core with the technical ("lab bench") meaning(s), it is because that core REMAINS FROM THE TECHNICAL MEANING after the semantic shift, and not because (as I *think* you mean to imply in your sentence containing the word "positivist") there is some ("common core") concept which BOTH the technical AND the vernacular meanings are INDEPENDENT ATTEMPTS to "reliably measure". If "we have a responsibility to try to find" anything, I think it is to try to find *why* some people insist on (1) glomming onto bits of jargon with very well-defined in-domain meanings, (2) ignoring much or all of those meanings while re-applying the jargon (often without ANY definition to speak of) in a new domain, while (3) refusing to let go of some (or all) of the Impressive Consequences derived in the original domain by derivations that (4) depend on the jettisoned definitions (and the rest of the technical apparatus of the original domain). Grrrh. ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com