Doris Payne dlpayne at OREGON.UOREGON.EDU
Sat Oct 28 17:37:05 UTC 1995

I'm interested by W. Stoke's musings over whether "grammaticalization" is
something that exists/a process that occurs more in the cognition of
linguists, rather than general language users per se.  I have mused over
the same thing with regard to the time-honored Inflection-Derivation
distinction.  Perhaps inflection vs. derivation are categories that exist
only in the language systems of linguists. I would be less inclined to
take such a heritical position with regard to at least the extremes of a
lexical/non-automization vs. grammatical/automization distinction. This
leads me to the next point...

i.e., on whether the development of (a) one type of grammatical morpheme
(e.g. a pronoun) into another type of grammatical morpheme (e.g. a tense
marker) [cf. Spike Gildea's comments],  (b) one discourse pattern into a
syntactic structure [cf. Susanna Cumming's comments], and (c) a lexical
item (e.g. a noun or verb root) into a grammatical morpheme (e.g. an
adposition) [cf. Spike's comments] should all equally be termed

It seems to me that the "correct" answer to this is whether the type of
human cognitive processing involved in all three changes is the same, or
different.  If the type of change cognitively is the same, then call them
all by the same term.  If the type of change is somehow substantively
different, call them by different terms.  Perhaps I'm naive here, but it
does seem to me that all three involve (re)routinization or
(re)automization of language behavior.

For a case study of the potential development of highly automated syntactic
structure out of less-automated discourse patterns (cf. Susanna Cumming's
comments), consider the following article:

        Payne, Doris.  1994.  "OVSu versus VSuO in Panare (Cariban):
        Do syntax and discourse match?" _Text_ 14.581-609.

In this article I refered to the process in question as "syntacticization",
but only because I did not want to get derailed into arguments over
whether it was, or wasn't "grammticalization" according to one person or
another's tradition. Since I was not dealing with lexical vs. grammatical
morphemes in this particular paper, it was a safe cop-out.

        -- Doris Payne

More information about the Funknet mailing list