In response to:
Where Does Language Come From? from the June 24, 1982 issue
To the Editors:
Linguist colleagues of Derek Bickerton (your review of his Roots of Language [NYR, June 24]) may find his scholarship less superb than do psychologists Bruner and Feldman. Bickerton has always told a good tale, whether as journalist or novelist, but such skill can sometimes serve mainly to dazzle. This time, there are substantial reasons to think that his exciting discovery of a “bioprogram,” an innate predisposition to learn particular categories in human language, is in fact illfounded.
Universals of language may well exist; it would be odd if a species which depends so much on talk did not have some predisposition to talk in an ordered fashion. The question is whether any real evidence for a “bioprogram” can be found in creole languages. Bickerton argues that many of their syntactic peculiarities (i.e. those not found in the European languages providing most of, their vocabulary) resulted from the same linguistic universals which, he claims, play an important role in children’s acquisition of their first language and were crucial in the origin and development of language itself.
This is an extreme but not altogether original hypothesis; in 1880 Adolfo Coelho speculated that creole languages “owe their origin to the operation of psychological or physiological laws that are everywhere the same, and not to the influence of the former languages of the peoples among whom these dialects are found” (“Os dialectos românicos ou neo-latinos na Africa, Asia e América,” Boletim da Sociedade de Geographis de Lisboa).
There are many parallel syntactic structures in a wide range of creoles from Jamaican English to Haitian French or the creole Portuguese of Guine-Bissau. Many of these structures seem to be word-for-word translations of structures found in the West African languages spoken by many of the ancestors of those who today speak creole languages.
If history can explain these parallels, then Bickerton’s case for a “bioprogram” is not convincing. Bickerton attempts to prove that the creoles’ similarities resulted from language universals rather than from the influence of earlier “substrate” languages by citing the case of Hawaiian Creole English and stating flatly (p. 72) that it “shares none of the substratum languages of the other creoles.” The trouble is, it does.
Early migrant laborers to Hawaii’s plantations at a crucial stage in its language development included not only Portuguese-speakers from the Azores and Madeira Islands, but also the “Black Portuguese” from the Cape Verde Islands off the African coast. Their creole language had itself been influenced by African languages. Moreover, there is considerable evidence that restructured varieties of Portuguese and other languages, born in the slaving areas of West Africa, brought African syntactic structures not only to the New World but also to Asia and the Pacific, including areas from which there was later substantial migration to Hawaii. The simplest explanation of the linguistic structures of these creoles still seems to lie in the unusual but amply documented social history of their speakers.
Bickerton’s credibility is further marred by his manipulation of facts that run counter to his argument and by his frequent factual errors. To cite a single example, he tries to prove that the word order of Hawaiian Creole English could not be derived from its substrate languages (Chinese, Spanish, or Portuguese) by asserting that “Portuguese and Spanish are freer in their ordering, tolerating certain types of verb-first sentences, but the equivalents of VOS [Verb-Object-Subject] sentences like /27/ [no laik plei futbawl, dis gais ‘These guys don’t like to play football’]…would be ungrammatical in these languages” (p. 21). As a matter of fact, Portuguese does have such word order: Nao gostam (V) de jogar futebol (O), estes garotos (S) ‘(They) don’t like to play football, these guys’—which is perfectly acceptable spoken, informal Portuguese in Brazil.
There may well be languages universals, but Bickerton has simply not fashioned an argument rigorous enough or accurate enough to prove this.
Hunter College, CUNY
New York, New York
Jerome Bruner and Carol Fleisher Feldman replies:
John Holm may have his particular facts right, and we would be in no position to argue with them. The creolists, plainly, will have to settle the details.
But if the best he can do to support his argument for the theory of substrate influence on Creoles is to cite the influence of a Black Creole Portuguese on Hawaiian Creole English, then he must be a wildly enthusiastic diffusionist. The likelihood of Portuguese blacks having significantly influenced creolization in Hawaii is so remote as to theoretically desperate. It makes Bickerton’s alternative less outrageous by comparison, even if Bickerton’s claim is unimaginative, as we argued.
There remains, however, the very general question raised by Bickerton’s work. There appear to be four distinctions that recur in simple syntactic form in virtually all Creoles, and they also appear in the “errors” of children’s speech where simple forms are not available in their cultural language—Specific-Nonspecific, State-Process, Punctual-Nonpunctual, and Causative-Noncausative. Where they are available, they tend to be acquired errorlessly. Diffusion could not possibly account for their presence in children’s language. We doubted in our review whether a “bioprogram” alone produced them in various Creoles. Our opinion was that they could more readily arise in response to universal cultural requirements. And if they could come about by diffusion, indeed, why just this set of distinctions?
Linguist Holm, by the way, will find that not all his linguistic colleagues share his waspish sentiments toward Dr. Bickerton. Let him speak for himself.
September 23, 1982