Robert Rosenberg's distinction between a spell[ing] checker and a "grammar"
checker is not an empty one.
Regrettably, however, no good grammar checker exists because no
natural-language grammar has yet been formalized. Chomsky and his followers
have made a beginning, but the two biggest problems these putative grammar
checkers confront are those of
o semantic ambiguity and
o usage consensus.
What does 'raw' mean. It may mean uncooked; as in raw carrots; it may mean
abraded and untreated, as in raw wound; it may mean off-color, as in raw joke;
it may mean untreated, as in raw sewage; etc., etc. Or again, they find
phrases like Shakespeare's 'salad days' unintelligible.
For me 'data' is the plural of 'datum'; for others it is singular. I use
'millenia' as the plural of 'millenium'; the New York Times uses 'milleniums'.
Many word processors "think" that 'nor' may occur only as the second element
of a neither...nor pair: They treat Lincoln's "The world will little note nor
long remember . . . ' as incorrect.
All of the English-, i.e., American-language grammar checkers I have
encountered are synchronic rather than diachronic. Confronted with an old
disjunctive subjunctive like that of Jack in the Bean Stalk's giant, "Be he
alive or be he dead, I'll grind his bones to make my bread", they throw up
their hands; and most do not even understand modern counterfactual-conditional
subjunctives like "If he were blind, he would be eligible for . . .".
The set of useful grammar checkers, which should be called usage checkers
instead, is currently empty.
John Gilmore Ashland, MA 01721-1817 USA
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
