Re: theory (was: Re: Greenberg's Word Order Universals)

John Cowan wrote:

> On Fri, 15 Sep 2000, J Matthew Pearson wrote:
>
> > Linguistics is the only field I know where theoretical work--and even the idea that
> > there should *be* theories--is routinely dismissed.  Nobody would seriously suggest
> > that physicists or chemists or psychologists or economists should confine themselves
> > to the collection of data and avoid positing theories to explain that data.
>
> Trust me, in physics the experimentalists take exactly the same attitude.
> And in geology, too, as you can see from McPhee's _Basin and Range_.
> The prejudice may be summed up thus:  Theoreticians are people who will
> cheerfully sacrifice inconvenient facts to beautiful theories, and may
> even develop meta-theories about the unimportance of counterexamples.
Well, I can understand how an experimentalist might get frustrated if s/he felt that
his/her results were not being taken seriously by theorists.  What I don't understand is
how a prejudice against *bad* theories gets amplified into a prejudice against theory in
general.  What the hell do the experimentalists think they're doing, if not testing
theories?

My impression (no doubt a biased one) is that much of the rancor on the part of
experimentalists stems from a misunderstanding of how progress is made in theory-building,
which is to start with sweeping generalizations and then gradually move to more nuanced
treatments of the facts.

No theory of language can explain everything, and so any theory will necessarily have some
'inconvenient facts' left over.  But that's no reason to abandon theoretical inquiry.  On
the contrary, the only way to proceed is to come up with the most beautiful theory
possible to explain the facts you have, and then look to see what that theory fails to
explain, and build from there.


> The contrary prejudice, of course, emanates from theorists: that
> experimentalist (or fieldworkers, as the case may be) are
> anti-intellectual stamp-collecting baboons, who wouldn't know
> a decent generalization if it fell out of a tree on their heads.
Maybe I'm naive, but I don't recall ever hearing one of my fellow theoristical linguists
characterize fieldworkers this way--except in knee-jerk defense against the accusation on
the part of said fieldworkers that linguistic theory is a load of crap.

Matt.