methodologies {was Re: form versus meaning}

[Ellen F. Prince]

jon, i was nodding vigorously in agreement all thru your post on
methodologies -- until i did a doubletake at this line:

>And something
>else, you would probably have to go through dozens of hours of
>transcripts to come up with one example of some of the phenomena that
>fill many theoretical journals these days.  The core stuff, you know.

i did a doubletake because i first read it as an argument for NOT using
naturally-occurring data, which didn't gibe with the preceding paragraphs.
then it hit me that you might have meant something quite different...

are you perhaps saying that low frequency phenomena are any less crucial to
the whole story than high frequency ones?!? presumably you wouldn't think
much of a theory of hematology that left out type o blood (or whichever is
the least frequent)? in fact, in syntax at least, it's mainly in the rarer
forms that you see what is actually going on, structurally speaking...

in any event, my real response to your line above is: thank goodness for the
humongous online corpora we have today! and let's hope they're all fully parsed
and tagged real soon. we are really the first 'generation' of linguists that
CAN work on low frequency phenomena using naturally-occurring data (and without
having to have the perseverence and energy of a jespersen). i actually find
this the most exciting thing to have happened in my professional lifetime -- it
has made my own methodology of choice actually feasible, plus it has enabled
historical syntacticians to do some really sound work (given the total
impossibility of intuitionistic or experimental data for dead lgs).