Using formalist grammars

William C. Spruiell 3lfyuji at CMICH.EDU
Sun Feb 27 20:30:12 UTC 2000


    There's a crucial difference between arguing that formalist grammars
aren't useful and arguing that *particular* formalist grammars aren't.
The first assertion is, as has been pointed out, fatally flawed;
probably most of us have gotten valuable information from at least one
grammar written within a formalist framework. The second assertion is,
of course true: there are relatively useless formalist grammars
(begging, for the moment, how to define "useless").  The question at
this point is why those particular grammars appear relatively useless,
and whether their condition results from something specifically
formalist or whether it results from a more general "grammar-writing
sin" that everyone, formalists and functionalists alike, are prone to. I
would argue that there are at least three problems with problematical
formalist grammars, and there are at plenty of other grammars that
suffer from at least two of these:

(1) Data bias. We all evaluate and elicit data based on large sets of
hypotheses that we operate on either consciously or unconsciously; these
sets are influence by our theoretical background. So, many formalist
grammars exclude data that would have bearing on things functionalists
are interested in. I'm sure many functionalist grammars exclude data
bearing on things formalists are interested in. One of the marks of a
good grammar is that the author tries to anticipate what the greatest
number of linguists might want to know about, and includes enough
"ground level data" (transcriptions of actual utterances with at least
some description of apparent intended meaning or context) for those with
different interests to evaluate.

(2) Jargon inflation. Both formalists and functionalists use
field-internal jargon that can render a grammar opaque to those
"outside" (especially if this is conjoined with severe data bias). In my
opinion, formalists are worse about this, partly, I think, because of
misguided attempt to rhetorically position the field more towards
computer science, where faculty get paid more and no one cares if your
teaching evals are poor. But functionalist grammars can be just as
jargon-intensive; I don't think anyone would accuse either Tagmemicists
or Systemic-Functionalists of being formalists, but there's a sizable
jargon hurdle to overcome (for those from other backgrounds) before
using their grammars.

(3) Accepting theory-internal constructs as empirical data. We all do
this to some extent of course -- every grammar I've ever read presented
data at least in some spots as "words," even though "words" are products
of analysis. I've encountered some formalist grammars, however, that
seem to feel much freer about including "zero" elements as actual
empirical data than I felt comfortable with.  Lefebvre and Muysken's
1988 examination of Quechua nominalizations, for example, bases part of
an argument on the observation that two sentences (their examples 22 and
23 on p. 174) are different -- even though the sentences are identical
if you remove their null elements. This particular problem, however, is
easily solved if the grammar simply provides an "unanalyzed" sentence
before adding whatever theory-internal constructs are needed on the next
line, and many formalist grammars do this.

In conclusion, I have also found a number of formalist grammars
relatively useless, but part of the reason for that is that I'm not
interested in what they're focussing on and I don't have time to learn
large amounts of jargon internal to a theory which, for other reasons, I
might think is misguided. I can't fault those authors very much for
that, though, since they'd react the same way to any grammar I'd write.
I think we *can,* however, propose sets of principles that any writer of
grammars should follow to make her/his work more accessible to everyone.

_______________________________
William C. Spruiell
Dept. of English Language and Literature
Central Michigan University



More information about the Funknet mailing list