What's a "creole"?

[Sally Thomason]

Alas, definitions of "pidgin" and "contact language" are just as
controversial as definitions of "creole": the field of pidgin/creole
studies is absolutely stuffed with controversy, starting with
definitions of what the field is about.

The idea of defining pidgins & creoles historically, as Michel
suggests (unless I'm misinterpreting Michel's comments) and as Dave
now suggests, is to my mind the only way to go: i don't believe
that there's any possibility of identifying creoles, much less
pidgins, by any "laundry list" of features, as Bickerton and others
have proposed (for creoles only).  Terry Kaufman & I, in a 1988
book, argued for a historical definition; I've spelled out the
argument in more detail in a 1997 paper ("A typology of contact
languages", in a book edited by Donald Winford and Arthur Spears).

The notion of genetic relationships in linguistics is deeply
embedded in a whole complex of theories & methods of historical
linguistics, and these have been developed and tested and have
enjoyed considerable success for well over a hundred years now.
They've been borrowed by anthropologists ("cladistics") and
adapted for some other historical sciences (e.g. by Darwin, in
The Descent of Man).  But basically, it's the same notion of
descent with modification that's familiar from evolutionary
biology.  The reason Spanish is considered a daughter language
of Vulgar Latin is that it can be shown to have descended by
successive modifications from Latin lexicon & structure; there
are of course aspects of Spanish that have foreign sources, but
they aren't sufficient in quantity or quality to disrupt the
straight line of genetic descent.

The reason Haitian Creole (for instance) isn't descended from
French (or Latin) is that, although the bulk of its lexicon is
certainly derived from French, its grammar can't be shown to
have come primarily from French or any other language (though
this last assertion is controversial; there's a major theory that
claims that its grammar is basically that of a single West
African language, Fongbe).  So it doesn't fit into
the historical linguist's picture of a language that belongs
in a family tree.  Likewise for most other languages that
have traditionally been called creoles, and for most languages
that have traditionally been called pidgins.   (Probably I
shouldn't use Haitian Creole as an example, because I'm
definitely no expert on the language; if Michel tells me that
it actually does have a lot of French grammar, I'll believe
him!  But if Haitian Creole doesn't fit my description, there
are a lot of creoles that do.)

But not all of them.  Hiri Motu (New Guinea) does have a
fair amount of Motu grammar as well as Motu lexicon; Reunionese
(Indian Ocean) has quite a bit of French grammar as well as
French lexicon.  Bajan (Barbados) has quite a bit of English
grammar.  Etc.  So there are borderline cases: it is not always
going to be possible to decide whether a language is genetically
related to its lexifier language or not.  Any putative pidgin
or creole obviously can't be related in the historical linguistic
sense to any of its *non*-lexifier languages, because the lexicon
won't fit.

This notion of fuzzy boundaries & undecidable cases is familiar,
of course, from many areas of historical linguistics -- not a
surprise, since historical linguistics is all about things that
can change into other things.  Consider the language vs. dialect
distinction: a nice clear distinction except in the borderline
period (which can be permanent in some cases, I guess) when
the degree of divergence between two diverging daughters of a
parent language isn't quite close enough to call them clear
dialects of a single language, and not quite far enough to call
them clearly separate languages.

None of this takes into account the undoubted issues of politics,
especially racism, that Michel raises.  I don't mean to suggest
at all that those are trivial matters.  But I do think that the
notions "pidgin" and "creole" are useful --- I think they identify
a reasonably coherent set of languages.  So I wouldn't be eager
to dissolve the Society for Pidgin & Creole Languages/Linguistics
on the ground that there's no there there.  (I can easily imagine
the usefulness of finding a less inflammatory term than "creole",
though.)

  -- Sally Thomason