Bob, Herb, et al.:

Like Herb (and for most of the same reasons Herb has already discussed),
I have trouble accepting "sentenceness" as something that pre-exists our
definitions as a kind of fundamental category, at least in the way we've
traditionally defined sentences. Cultural categories *always* seem
natural to those accustomed to them, so the fact that sentences seem an
"intuitively obvious" category doesn't mean that they are necessarily
givens (in other words, saying sentences are a basic unit of human
language might be a bit like saying that desserts are a basic unit of
human cuisine). "Sentence" isn't really the same as "enough language to
get the job done," and it's the latter that has the greater chance (I
think) of being universal. 

There are two problems in particular that I've seen occurring repeatedly
in linguistic discussions of grammar; they aren't really problems if one
takes "sentence" as a label for something that's useful to work with as
a methodological unit, but they are real issues if claims are made about
the "reality" of what the label refers to. 
 

The first of these is the way that analyses relying on ellipsis are
motivated by ideas of what constitutes a sentence, but then are
reconceptualized as *supporting* the ideas that motivated them in the
first place (apologies for returning to ellipsis, since I've posted on
it too many times before, but it's relevant here. Honest. No, really).
The "less-than-sentence-but-still-good" examples that Herb, Craig, and
others have pointed out can usually be analyzed as full sentences with
elided parts -- but only if one starts with the assumption that since
those utterances *sound* okay in context, they must therefore be
sentences in the way we've chosen to define sentences, and hence adding
invisible parts is warranted. There's nothing wrong with a carefully
constrained ellipsis argument, but there IS something wrong with using
it to support one's claims of what a full sentence is -- it's circular
reasoning. 

The other is the effect of using a single pair of opposing terms like
"competence vs. performance." Setting up a binary opposition encourages
people to think that anything that is not A is by necessity B, with some
of the consequences that Herb has already mentioned. There's one sense
of "performance" in which the term encompasses speaker "goofs": slips of
the tongue, false starts, the occasional fit of coughing interrupting an
utterance and the like. But once we make a definition of sentence,
anything that diverges from that definition then becomes a "goof" in a
way that it wasn't before. Saying that slips of the tongue don't
necessarily fall within the purview of your theory is a very different
matter than saying that "ya eat yet?" doesn't, but having only the one
pair of terms leads us to group both as "performance" (or to rescue the
second via an ellipsis argument). A construct of the theory
("sentencehood") becomes the basis on which data is judged *relevant* to
arguments about whether the construct itself is valid. Once that
happens, you don't have a theory anymore; you have a faith (and it's a
really boring, nerdy one, without even a decent holiday attached).  

Bill Spruiell
Dept. of English
Central Michigan University

To join or leave this LISTSERV list, please visit the list's web interface at:
     http://listserv.muohio.edu/archives/ateg.html
and select "Join or leave the list"

Visit ATEG's web site at http://ateg.org/