What's a "creole"?

[Michel DeGraff]

Sally makes lots of interesting and constructive points --- as usual.  Let
me try and add to these.

For one recent critique of the family-tree model, see Dixon's _The Rise and
Fall of Languages_.  Dixon's main point is that such a model is in many
cases inapplicable.  Dixon also analyzes various (alleged?) empirical and
theoretical flaws in the comparative method.  Unfortunately, he too falls
in (what I'd call) the ideological trap of brushing creoles aside as freak
cases of language creation.

Re Sally's comments on Haitian Creole vis-a-vis genetic relationships.
Here it may be worthwhile to invoke the Latin-to-Romance example again.  In
comparing Latin to its Romance `descendants', we have to face the facts
that many aspects of, say, French have no counterparts in Latin.  Consider
for example case morphology, (canonical) word order, scrambling, articles.
At the same time, certain aspects of French are due to `substrate'
influence; that is, to the influence of the Gaul and Frankish speakers
acquiring some variety of Latin in the context of language contact (due to
imperialist conquests --- sounds familiar?). Given the typological distance
between Latin and French, the grammar of French certainly cannot be argued
to come "primarily" from Latin.  Similar remarks apply to, say, the history
of English from Germanic (Meillet has some discussion of these facts; for
example, he observed that the use of word order in French and English in
order to express relations between phrases is a *creation* of these
languages.  Meillet's conclusion was that such innovations didn't have any
model in Latin or Old Germanic.)

As it turns out, it could be argued that along *certain* parameters (such
as case morphology, word order, scrambling, articles) French and HC are
more similar to each other than Latin and French.   Ditto regarding the
typological distance between, say, English and Jamaican Creole, on one
hand, and that between Old Germanic and English, on the other hand.

Question: What then is the *structural* litmus for deciding where "language
change" ends and where "creolization" begins?

At this point, it looks like we're talking about (quite "fuzzy")
differences in degree and rate rather than about fundamental differences in
quality.  It seems to me that *all* cases of language acquisition entails
"innovations"/"creation" as constrainted by UG and as influenced by
contingent socio-historical factors such as degrees of language contact.
There is no "language transmission" per se: the learner *always*
(re-)creates the `target' grammar based on primary linguistic data that are
the output of pair-wise distinct idiolects/dialects/languages in contact.

Here Meillet had the right insight when he considered that:

  "Each child must on his own acquire the capacity to understand the
  speech of people in his community. ...   Language is not given to
  him en bloc, all in one piece. ...  Thus, for each individual,
  language  is a total re-creation, carried out under the
  influence of the surrounding environment.  There could not exist a more
  absolute discontinuity."

Thus, "language transmission" (whether "normal" or "abnormal"!) is nothing
but an epiphonemanal illusion, and the `Creole' cases of language creation
only make obvious what in other cases may be too subtle to observe.  (I
discuss such observations in the MIT Press volume _Language Creation and
Language Change_; see p14ff.)

As for why there should be a "Society for Pidgin & Creole Linguistics" ---
or why I call myself a "creol-ist": Note that there is also an "Annual
Conference on African Languages" --- which I attended this past March at
Boston University.  There are also many university-level programs for the
study of "African languages".  Yet no African-ist that I know would argue
that there exists a (fuzzy or categorical) set of structural criteria that
"identify a reasonably coherent set of [African] languages".

There's a lot of very important concerns (socio-linguistic, historical,
political, ideological, etc.)  that Creole languages and their congeners
(e.g. Black English) share in common --- including the fact that these
languages are still among the most stigmatized in the world.  I do think
that the term "creole" is still one of the most sociologically- and
politically-loaded terms in contemporary linguistics.  And I seriously
think that `creolists' can learn a lot from one another independently of an
(elusive?)  structural/cognitive litmus test for creolization. This is
actually why I subscribe to the Chinook list.

                                 -michel.
___________________________________________________________________________
MIT Linguistics & Philosophy, 77 Massachusetts Ave, Cambridge MA 02139-4307
degraff at MIT.EDU        http://web.mit.edu/linguistics/www/degraff.home.html
___________________________________________________________________________