ENESTERNISMENOI (Was: De-Inflection Software?)
Carl W. Conrad
cwconrad at artsci.wustl.edu
Sat Sep 21 20:16:51 EDT 2002
At 4:23 PM -0700 9/21/02, Richard A. Stauch wrote:
>
>Finally, Smyth writes, "As far back as we can follow the history of the
>Indo-European languages we find only complete words; hence their
>analysis into component morphological elements is merely a scientific
>device for purposes of arrangement and classification" [193 N].
>Nevertheless, if we took the rules and tables listed in Smyth Parts II
>and III, boiled them down and added tables of roots and such, a computer
>could use them much more quickly than a human can. And when we consider
>that a computer program is nothing but an extension of human
>capabilities, it isn't really cheating, after all. I believe this is a
>project that can be done, and I am very surprised that, so far, it has
>not.
One final comment which may sound sarcastic but really is deadly serious:
that's a neat citation from Smyth. I really think such a project is
eminently do-able, BUT I think that in order to create it you would have to
learn thoroughly what all the interacting factors in the morphology of all
inflected forms actually are and HOW they interact (e.g. for verbs, how
consonantal stems alter before different consonantal endings such as GRAF
in GEGRAMMAI, GEGRAYAI and GEGRAPTAI as well as the infinitive GEGRAFQAI;
how augment replaces reduplication in certain double-consoant verbs like
ENESTERNISMENOI, etc., etc.). And ... once you have mastered all these
interacting factors that enter into the multiplicitous inflectional system
of Greek--especially the verb--you'll have a splendid program--but YOU
WON'T NEED IT ANY MORE, because you'll be able instantly to guess
intelligently at the lemma and recognize precisely what form you have
before you without using the newly-formed program at all. Or to say that in
different words: you would have to learn all that you really need to know
about the interaction of elements in Greek inflection in order to create
the program, and then you wouldn't need the program.
--
Carl W. Conrad
Department of Classics, Washington University (Emeritus)
Most months:: 1647 Grindstaff Road/Burnsville, NC 28714/(828) 675-4243
cwconrad at artsci.wustl.edu OR cwconrad at ioa.com
WWW: http://www.artsci.wustl.edu/~cwconrad/
More information about the B-Greek
mailing list