Book review: Defining Creole

[Harold F. Schiffman]

Forwarded from Linguist-List, Announced at
http://linguistlist.org/issues/16/16-309.html

AUTHOR: McWhorter, John H.
TITLE: Defining Creole
PUBLISHER: Oxford University Press
YEAR: 2005


Reviewed by Elizabeth Grace Winkler, Department of Linguistics, University
of Arizona

In Part 1, "Is there Such a Thing as a Creole" McWhorter explains the
motivation for putting together this series of conference presentations
into book form and the philosophical reasoning behind why there should be
a distinct category of languages called 'creoles'. For McWhorter, the
creole categorization is not based solely on the socio-historical
development of these languages as is traditionally defined by creolists.
He asserts that there is a synchronic or typological rationale as well.

One difficulty McWhorter faces in establishing this claim is convincing
creolists to give up a long held, almost knee-jerk defense of creoles as
"equal" to other languages. This derives from past and ongoing
disparagement of these varieties as incomplete and unsystematic varieties.
As could be noted from verbal responses at Society for Pidgin and Creole
Linguistics meetings and writings elsewhere, proposing that creoles are in
some way different in their structural makeup makes some distinctly
uncomfortable. However, if there is nothing structurally unique about
creoles, then their study, McWhorter points out, contributes little more
to our understanding of language development and change than the study of
any other language - a statement that he contends few creolists would be
comfortable with regardless of their positions on his theories.
Nevertheless, there has been considerable resistance to the proposal that
there is a set of synchronic features that differentiate creoles from
other languages.

For some time McWhorter has been asserting that there are features, or
more accurately a systematic LACK of features, that provides a synchronic
description of creoles. He posits three features that are typologically
absent from creoles specifically because they are newly formed languages.
Accordingly, these features may develop later during the lifecycle of a
language. They are:

1. the use of grammatical inflection via affixing;
2. the development of productive, nontransparent derivational affixes; and
3. the use of tone to either mark lexical differences or as grammatical
markers.

These are language features that McWhorter claims would not have been
found in pidgin languages because they are unnecessary to basic
communication (which will be elaborated on later in this review). Because
McWhorter asserts that a pidgin stage is crucial to creolization, it is
useful to take a look at the languages that contributed to pidginization.
McWhorter illustrates how some of the contributing languages make
considerable use of grammatical inflection, tone and nontransparent
derivational morphology yet these features are not present in the pidgins
nor in the creoles that develop from them. Accordingly, "the birth of many
creoles as pidgins leads us to the hypothesis that the natural languages
of the world (which do not include pidgins) displaying the three
particular traits above will be creole languages, and that conversely, no
older languages will display them" (p. 11).

He does not claim that all creoles are perfect examples of the Creole
Prototype, rather that because gradience is a normal feature of all
languages "creoles confirm to the hypothesis in varying degrees" (p. 19).
For example, Sango has a limited number of inflectional affixes and makes
use of tone for grammatical marking. Another reason a limited use of these
features may be present in some creoles is due to continued use of the
substrate languages for a considerable period of time during creolization
as exemplified by the presence of Ijo inflectional affixes of Berbice
Dutch Creole. Furthermore, intense contact with the superstrate during
creolization can cause the acquisition of these features as with
Runionnais French Creole. In addition, intense contact after creolization
provides an opportunity for these features to be acquired as in the case
of Louisiana Creole, which has existed in close contact with Cajun French
since plantation days.

In Chapter 2, "The World's Simplest Grammars are Creole Grammars"
McWhorter challenges the truism that grammars of all languages are equally
complex, though that complexity may be manifest in different aspects of
the system. His challenge rests on two unanswered questions: 1) how do you
realistically measure supposed complexity for comparison, and 2) by what
mechanism is this "equality" forced upon languages? The mechanism for how
they somehow "calibrate themselves according to comparison with one
another" (p. 44) is unexplained.

His claim is not that creoles are the MOST simplistic system possible
under universal grammar, but that they are fundamentally different in the
quality of their simplicity. Often creoles are described in comparison to
superstrate, in terms of what features they "lack". McWhorter takes a
different tactic. He prefers the view that non-creole languages are over-
specified and have accumulated what McWhorter terms "ornamental
elaboration" (like inflection). Ornamental features are those that appear
only in a subset of the world's languages and do not contribute anything
necessary to communication. He also calls this tendency of older languages
"grammaticalized overkill" (p. 77).

Another concern expressed by some of his critics is that McWhorter is
arguing that creole languages are in some way simplistic. Again, this is a
misunderstanding based on his approach of looking at over- versus non-
redundant-specification. He is solely arguing that older languages are
over-specified with superfluous marking that serves no communicative
purpose and that it is the quality of creole simplicity that is of
interest, since it is fundamentally different than simplicity of other
languages.

In the remaining chapters of Part One, McWhorter takes to task some of the
competing theories of creole development. In "The Rest of the Story"
McWhorter contends that explanations for creole grammar, like superstrate,
substrate and syntax-internal explanations, do not take into account a
pidginization stage and thus lack explanatory adequacy for the vast data
that he has collected.

McWhorter then points out a number of items that the syntax-internal
approaches (best elaborated by Michel DeGraff) fail to explain, for
example, the lack of ergativity in creoles, which is a marked construction
across the world's languages. He discusses the case of Korlai Portuguese
Creole (not ergative) which is derived from Marathi, an ergative language.
Because no pidgins are ergative, McWhorter posits that ergativity is a
feature of older languages. McWhorter also elaborates on half a dozen
other features distributed in Founder Community languages that appear in
neither pidgin nor creole languages.

McWhorter claims that the syntax-internal explanations only explicate part
of the picture and also make predictions that have not been borne out by
data. McWhorter asserts that any grammatical feature from his "marked"
list will not be acquired during pidginization, and thus, will not be
available for transfer during creolization. McWhorter does not completely
reject DeGraff's theory and details how the two theories complement each
other.

In the next chapter McWhorter takes to task specific claims of LeFebvre,
DeGraff and Arends through a detailed comparison of Haitian with Fongbe
and Gbe (substrates for Haitian Creole) along with an analysis of
Saramaccan. First he shows that Saramaccan does not pass its progenitor in
complexity in any consistent way. It is just beginning the stage of
"ornamentalization" of its systems. Then he shows how Lefebvre's
relexification theory fails to explain how Haitian Creole differs from
Fongbe. Finally McWhorter asserts that Arends' claim that not enough data
is available on a wide enough variety of pidgin and creole languages is no
longer fully accurate and not necessarily pertinent anyway. In the study
of other natural languages, there are often gaps, but the research claims
made are still considered valid.

Chapter 5 focuses on Mufwene's Founder Principle which taken in the
extreme maintains that superstrate is the primary contributor to creole
development; therefore, creoles are solely dialects of the superstrate.
Furthermore, because these varieties began on homesteads that predated the
creation of huge plantations, Africans had more access to the superstrate,
and thus, acquired adult second language versions of the language. Later,
when the homesteads morphed into plantations, it was the speech of these
slaves that was the model for the hundreds of newly arrived Africans who
were to have little direct contact with the superstrate as spoken by
native speakers. It was here that additional reduction occurred. Mufwene
claims that creoles are just contact varieties for which the contact was
quite significant.

McWhorter refutes the main claims: first, through an analysis of the
history of Suriname and Martinique, he shows that although Mufwene is
correct in asserting that homesteading was significant, he is incorrect in
his assessment of how much and what language was acquired there. McWhorter
provides many speech samples from early documents that clearly show
features in the speech of homestead slaves that make sense only if they
acquired a pidgin at some point.

Next McWhorter takes on the claim that creoles are just varieties of the
superstrate because many features of these varieties can be found in non-
standard French varieties of the time. McWhorter counters by pointing out
the numerous important features that can only be explained via substrate
influence. He also points out that it is not enough to show that certain
features occurred in older nonstandard varieties of a language; the
variety must also be substantially available to slaves as they were
acquiring the new language.

In the introduction to Part Two, McWhorter details some of the concerns
about what he terms the "culture cult" which has pressured, to some
degree, the direction of recent creole work, a pressure not commonly
experienced by researchers of, for example, Scandinavian languages. He
says that much of this grew out of a visceral reaction to Bickerton's
early assertions that substrate influence was insignificant. McWhorter
maintains that singular focus on substrate contributions, to the exclusion
of other explanations that may have better explanatory adequacy, does no
service to furthering our objective understanding of creole development.

The chapters of Part II look at a number of grammatical constructions to
evaluate how they came to be in creole languages. Chapter 6, focusing on
zero-copula, is a rich data-laden presentation about how this feature is
common not just to pidgins and creoles, but to second language
interlanguage varieties. McWhorter argues against this being a substrate
feature because he believes that because it would not have been in a
pidgin, and thus, it could not have transferred to any creole. He believes
that zero-copula is better explained by normal second language acquisition
processes.

In Chapter 7 McWhorter focuses on the emergence of predicate negation in
Saramaccan. He has chosen Saramaccan specifically because it has had very
limited European language influence since it developed in ex-slave
communities located far in the interior. This linguistic isolation has
caused some scholars to speculate that it is a closer reflection of
universal grammar (UG) than other creoles with more contact influence.

In the following chapter "A case for Genetic Relationship", McWhorter make
a fairly convincing argument that the English-based creoles of the
Atlantic are all descendants of a stabilized West African Pidgin. He
builds his case not by looking at shared lexical items (since these
creoles share the same lexifier, it is not a useful argument) but by
focusing on features that he asserts are not traceable to either the
superstrate or substrate groups. He details six such features and provides
a wealth of examples. These are compared with examples from non-creole
languages to show how the processes match historical processes that
happened to non-creole languages.

He claims that it is unlikely that these remarkably similar forms would
appear independently in so many creoles found in distinct places nor can
they be explained as a product of diffusion. Only shared inheritance
effectively accounts for their similarities. He convincingly argues
against a substrate explanation because although these grammatical
morphemes may have the same function in the substrate languages, they are
phonetically dissimilar. It is unlikely that early creole speakers would
have all chosen the same or very similar forms if they had not been
previously available to speakers via a locally used pidgin.

In "Creole Transplantation" McWhorter looks at how 'limited access' to the
target language appears to be the common denominator of theories of creole
development. He shows how the outcome can be the same in the different
situations in which creoles emerged and compares places 1) where the
demographics are the same, and 2) where the quality of the contact with
the superstrate was significant. Again genetic shared history can only
account for similarities found across diverse situations.

Then, in a discussion on the lack of Spanish-based creoles, McWhorter
shows that the same conditions existed in the colonies of Spain as
elsewhere and yet almost no creoles developed there. Spanish colonies,
like Mexico, had the same proportions of slave-to-European, and the same
lack of rich language contact with the superstrate. Nor did Spanish slaves
acquire Spanish because the Spanish were "kinder" to them as sometimes
claimed. The same brutality existed despite laws requiring better
treatment. McWhorter discounts Schwegler's (and other's) contention that
there were creoles in these areas that have since disappeared. He asks why
creoles would only disappear from the Spanish areas and not elsewhere. The
answer lies elsewhere. He attributes the lack of Spanish-based creoles to
the fact that the Spanish did not maintain slave ports in Africa;
therefore, a Spanish pidgin was not created and was not available as a
target to slaves and others in the new world. Still the fact that Spanish
slaves did manage to learn Spanish must be explained. McWhorter posits
that although the recently arrived adults in the community only managed to
acquire second language approximations of Spanish, their children, having
access to language acquisition abilities of their youth, were able to pick
up Spanish. This of course begs the question "Why didn't this happen to
other creoles?"

McWhorter claims that the English creoles originated as pidgins in Africa
and became creoles precisely because the pidgin was available to New
World- born children as a target. The children could have learned English,
like the children in Hawaii could have learned Standard English, yet
created Hawaiian Creole from their community's pidgin, despite the fact
that they were taught standard English in school.

Therefore, McWhorter posits that creoles developed in situations in which
there was a pre-existing pidgin available to the speakers. The need to
learn the lexifier language was circumvented because the pidgin was the
target. Having the expanded pidgin as the target neatly deals with
something else that has always troubled me as well. When I read that a
creole "lacks" a feature, usually an ornamental one at that, or has "null
specification" (like zero-copula), it implies that acquiring such a
feature is an expected step on the road to language "development". It is
telling about our general theoretical bent that we have generally focused
on the lack of over-specification rather than the elegance of non-
redundancy.

Part III begins with a lengthy section on how and why English became a
minimally specified language. Purpose of this section is to show that not
all languages are over specified equally contrary to the oft-held
assumption that they are. McWhorter does not claim that English is a
creole, rather that the extent of its restructuring can shed some light on
contact phenomena because this is another type of incomplete second
language acquisition by adults.

In the final chapter "Where does Black English come from?" McWhorter
argues for the new dialectologist position (DP) rather than the creolist
hypothesis for creole genesis. The DP, though not particularly politically
correct, has certainly been well substantiated by statistical comparisons
of Black English features with nonstandard varieties of English to which
slaves would have been exposed. The majority of the chapter is an analysis
of "The English History of African American English", edited by S.
Poplack. McWhorter's conclusion is that although the contributors to this
volume and others have established that Black English is essentially a
dialect of English, the proponents of DP have not addressed the existence
of creole features whose presence is better explained by the creole
hypothesis.

The main thrust of his argument for why Black English is a dialect of
English is that diaspora varieties of Black English are quite varied and
do not appear to be transplanted Gullah as has been asserted. McWhorter
believes that Gullah was basically spoken only in the area in which it is
now found.

He takes to task traditional views of the creole continuum and posits
(like others before him) that mesolectal and basilectal varieties could
have coexisted from early on considering that slaves on homesteads would
have been in a better position to acquire a second language variety of
English than slaves on large plantations. As homesteads turned into
plantations or as homestead slaves were sold to plantations, they may have
provided a language target. Thus, features that appear in Black English
that had disappeared from white varieties of English before the dominance
of large plantations, may have been acquired through these homestead
slaves.

McWhorter also asserts that it is necessary to look at both what creole
features are both present and absent to understand the puzzle of Black
English. Black English is similar to creoles in the features that it is
MISSING (zero copula, negative concord, limited inflectional morphology)
which he claims can be accounted for by normal Second Language Acquisition
processes like ellipsis or overgeneralization. What make Black English
different is the LACK of creole features that appear in substrate
varieties and pidgins, like serial verb constructions, features that are
found in creoles throughout the Atlantic area. Put simply, Black English
looks more like a West Germanic language than a West African one and
always has.

CONCLUSION

Some general comments first about the structure of the book. Although the
text was primarily taken from past conference presentations and shorter
publications, it is not simply a recitation of these works. It has
benefited both from McWhorter's addressing many of the critiques of and
commentaries on aspects of those earlier works as well as from his oft-
heard requests for data from lesser-known languages to support or refute
his claims. The breadth of the language samples offered is one of the
strong aspects of this work.

Although it is certainly not what the author intended, the sections of the
book are so different that they will appeal individually to a variety of
specialties and interests and may be read independently of the rest of the
text -- especially the chapters of the final section of the book.

Lastly, in terms of general comments, I would have like to have had a
concluding chapter in which McWhorter tied together or summarized his
arguments.

The fact that the features of the Creole Prototype are missing from early
creoles and only develop over time seems plausible from the extensive
evidence that McWhorter provides. But if this is accurate, we are back to
creoles being no different from other languages in form. They are just new
languages, which have not as yet accumulated the flotsam and jetsam of
older languages. McWhorter only briefly addresses that point -- "to even
elaborate on the fact that creoles will eventually be synchronically
indistinguishable from older languages is a rather vacuous detour. The
discussion has an air about it of implying that this transformation will
constitute 'progress'" (p 100). I have to disagree because this smacks of
a cultural justification rather than a linguistic one. When creoles
become "older" languages, the presence or absence of the "male nipples" of
language, as McWhorter terms these over-specifications, will be judged the
same way we judge languages like English today that lack a good deal of
ornamentation that is found in many other world languages.

As convincing as are many of McWhorter's arguments on the surface (and
well substantiated by examples from a broad range of languages, both
creole and non-creole) what I am going to have to do, as I expect others
will as well, is to spend a considerable amount of time to both reanalyze
his arguments and his assessment of competing explanations. As long as
McWhorter has been presenting these ideas at conferences and in journals,
he has been asking for data from scholars who work with languages with
which he is less familiar. The publication of this book, presenting these
papers in one place, should broaden the response to his queries and help
to either support or disprove the arguments he makes. I certainly expect
that this book will evoke responses from scholars of opposing viewpoints.

Finally, McWhorter is very good at teasing out both the small, and major,
inconsistencies in our long-held assumptions about creoles and language
change and whether we agree with him or not, he often shake us out of our
comfortable complacency over unchallenged issues. He's the 800 pound
gorilla in the room. You may not want to hear him, but he's hard to
ignore.

ABOUT THE REVIEWER

Elizabeth Grace Winkler is a visiting professor in linguistics at the
University of Arizona, USA. Her research publications have concentrated on
African substrate influence on the English-lexifier language Limonese
Creole and codeswitching between Spanish and Limonese Creole in Costa
Rica. She has also authored a dictionary of Kpelle, a Mande language of
Liberia. She is currently writing on differing uses of tag questions by
gender among the Afro-Limonese: a community in which the connections
between language and power differ from many of the previous groups
studied.