I. From pidgins to creoles

While linguistics-related scholars have come to a general agreement on the definition of such terms as “creole” and “pidgin”, it is not unusual to encounter semantic variations in the broader context of creole cultures studies. A first essential step will therefore be to establish a few clear definitions, such as accepted and used by the different studies and works discussed here.

Pidgin is the term used to designate any language or jargon, assembled as an aggregate of other languages, in order to facilitate communication between two or more groups of speakers who do not otherwise share a common language. Characteristically, pidgins are not fully developed languages in that they only possess the most limited of grammar and syntactic structures along with a simplified subset of vocabularies borrowed from their base languages (Holm 1998, p4-5). Some pidgins are born out of commercial necessity, acting as lingua franca in trade situations between distant societies (e.g. Sabir: a mix of European and Mid-eastern languages used by medieval mariners); a more significant number, however, arise during times of war or colonization, when forced labor or land occupation induce cohabitation between speakers of varying native tongues and no other ways to communicate.

With some arguing on the exact configuration required for the making of a proper pidgin (as opposed to straight assimilation of one language by another), notably the number of languages that must initially coexist (Bickerton 1995, Holm 2000), there is a general agreement on the crucial role of power struggles (McWhorter 2005): one language often dominating the other(s) and providing most of the resulting pidgin’s lexicon. As such, slavery was of course a strong catalyst, in particular due to the widespread practice among slave-owners of purposely separating speakers of a same language: ostensibly to force a quicker adoption of their own language, but also probably in order to decrease rebellious aspirations by diminishing the part of cultural identity tacked to a common language (Derrida 1996).

Pidgins, by their very definition, do not count any native speakers. When children are born into a pidgin and start interacting with each other, they do not simply acquire better fluency in what is only their parents’ second language: they invariably develop it into their own richer, full-featured language; what is generically labeled a creole by linguists (capitalized Creole usually refers to one specific language or culture). Though it should be noted that this creolization process does not always take place (there are cases where the original pidgin simply disappears in favor of another language), when it does, it has been shown to spontaneously produce new structures absent from the language spoken and taught by parents.

One of the chief debates in modern linguistics and cognitive sciences is then to define how this creolization process can be accounted for. Indeed, the potential conclusions far outreach the mere domain of creole linguistics: while some posit that the features of a creole can still be explained by the influence of its base languages (Chaudenson 2004, Holdm), others see evidence of an innate human ability to develop language structures (Bickerton 1992, Pinker 1994). The latter cognitivist position has eventually come to be held as the most probable in the scientific community at large, with proponents of the former, still bringing strong objections and pointing at important flaws in the approach.

II. Hawaian creole and pidgin

Of the hundreds of creoles known to linguists worldwide, one tend to appear recurrently in works on the topic: Hawaii Creole English (HCE). Unlike most other creoles, its pidgin-to-creole transition phase took place for the most part less than a century ago: this historical proximity makes it one of the best-studied examples of creolization in contemporary times and as such, a cornerstone of modern theories of language acquisition.

At the end of the 19^th century, a sudden boom in sugar plantations led to massive immigration of foreign workers to Hawaii. The cohabitation of over a dozen nationalities on Hawaiian plantations and the need to communicate while working, quickly resulted in the emergence of a pidgin: “lexified” by English (i.e. with English as its main contributor in terms of vocabulary), but bearing strong influences from Portuguese, Hawaiian, Cantonese, Japanese, Tagalog and Korean.

As is commonly the case in such situation, the first generation of children born to these immigrants were exposed to both their parents’ own native language and Hawaii Pidgin English (HPE), through community interactions or the supervision of adults speaking a different native language. The ensuing creolization of HPE into HCE was a unique opportunity for linguists to study the details of this transition.

In the late 1970s, linguist Derek Bickerton was able to interview both some of the original speakers of HPE (those laborers who had arrived on the island at the turn of the century) and native speakers of HCE (born shortly thereafter). Commenting on transcriptions of sample Pidgin sentences (see Figure 1), he notes: ‘there are no articles, no prepositions, no complementizers and no markers of tense or aspect. In other words, one finds only the types of grammatical item that are relatively rich in meaning, and not those whose primary function is structural.’ (B. 1992 p120). In essence, HPE is only a make-do ‘protolanguage’, missing key elements of structured speech despite its speakers’ abilities to use such structures in their own native language.

Figure 1: Samples of HPE and HCE (B. 1992 p169-170)

Hawaii Pidgin English

Hawaii Creole English

3. Ifu laik meiki, mo beta make time, mani no kaen hapai If like make, more better di time, money no can carry ‘if you want to build (a temple), you should do it just before your die – you can’t take it with you!’ 4. Aena tu macha churen, samawl churen, haus mani pei And too much children, small children, house money pay ‘And I had many children, small children, and I had to pay the rent’ 5. Luna, hu hapai? Hapai awl, hemo awl Foreman, who carry? Carry all, cut all ‘Who’ll carry it, Boss? Everyone will cut it and everyone will carry it’

6a. They wen go up there early in the morning e go plant ‘They went up there early in the morning in order to plant (crops).’ (p. 169) 6.b. I gotta go hire one carpenter e go fix the form ‘I have to hire a carpenter to fix the form.’ (p. 170) 8.a. [The guy on’ lay the vinyl] been quote me price ‘The guy who was going to lay the vinyl had quoted me a price.’ 8.b. You see [the island get coconut]? ‘Can you see the island that has coconuts on it?’

By contrast, while dissecting samples of HCE, such as spoken by the children of these very immigrants, Bickerton establishes the presence of coherent grammatical structures that were all but absent from the original Pidgin. The enhancements brought over the span of a single generation are so numerous as to make him call it a transition from proto-language to actual language. Word order (relative placement of subject, verb and complement) is a major difference: in HCE sentences, it is guided by rigid rules, where it was arbitrary in HPE, often topic-driven or dependent on the speaker’s native language. Another dramatic change is the introduction of syntactic markers, words with no meaning of their own, ostensibly helping to shape sentence structure, sometimes drawn from purely lexical words of the initial pidgin: ‘for example go in the subordinate clauses of [figure 1] 6.a conveys no sense of motion but has been downgraded to a purely formal marker of nonfinite verbs, like English to’ (Bickerton 1992 p171).

But perhaps the most notable observation about this linguistic evolutionary leap is that: ‘the grammar of the language that resulted bore the closest resemblance not to grammars of the languages of Hawaii’s immigrants; nor to that of Hawaiian, the indigenous language; but rather to the grammars of other creole languages that had come into existence in other parts of the world’. This apparent lack of direct transmission from “parent” languages to the resulting creole is one of the most crucial point in Bickerton’s support of Chomskyan nativist theories of language acquisition wherein: ‘creole languages form an unusually direct expression of a species-specific biological characteristic, a capacity to recreate language in the absence of any specific model’ (Bickerton p171).

III. A different type of creolization: Nicaraguan Sign Language

While studying the genesis of spoken creoles such as Hawaii Creole English has brought some very strong elements toward proving existence of an underlying universal grammar and a biological component to language acquisition (Bickerton’s Language Bioprogram Hypothesis), one of the major criticism leveled at these scientific proceedings is that it mostly excludes a possibly meaningful influence from either main (superstratum) or auxiliary (substratum) contributing languages to the pidgin. In all cases studied, it is impossible to rule out completely some level of interaction between children learning creole and native speakers of other languages (starting with their own parents). This interaction, some argue could be enough to provide an imitable example of grammatically structured speech, which children would then merely adapt to pidgin.

If isolated cases of “perfect” unassisted creolization exist, they are, for ethical reasons, impossible to study: when happening upon an individual who may never have been exposed to the influence of human societies (the perennial “feral children”), priority is obviously given to teaching them a language rather than studying any existing language skills they may or may not have acquired prior to being retrieved.

In such context, the genesis of modern Nicaraguan Sign Language provided as close a perfect case as seemed reasonably reachable through ethical research: a direct transition from a form of grammar-poor pidgin into a creolized, grammatically structured language, with arguably no input from any other native language.

Despite commonly held beliefs, formalized sign languages differ, travel, mix and evolve, from one country to the next, in a fashion similar to spoken languages, albeit often following wholly different geographical patterns. American Sign Language, for instance, is radically different from British Sign Language, itself related to other European sign languages in ways reminiscent of the relations tying spoken languages of a same family. In pre-1970 Nicaragua, however, no such language had been devised or adopted: deaf children would receive no particular care and would not be taught any systemized ways to communicate. Furthermore, in the absence of any targeted schooling and often any form of schooling altogether, conjugated with a very high social stigmatization, deaf individuals remained scattered throughout the country, with no contacts between themselves and therefore no way for a “signed pidgin” to spontaneously emerge.

In the late 1970s, political change in Nicaragua brought about a massive literacy campaign for all inhabitants and resulted in the opening of the first schools for the deaf. As was often the case at the time, officials recommended that lip-reading and finger-spelling be laboriously taught to children rather than any formalized sign system. While the results of this oral method were dismal, it took only a few years for deaf children thus gathered to devise, outside of class, their own way of communicating through signs. The resulting Lenguaje de Signos Nicaragüense (LSN), as it is still known and used today, shows the characteristics of a pidgin: while allowing efficient communication between “speakers”, LSN is still a protolanguage in that it lacks rigid grammar structures and formulation often varies from one person to the next. But the remarkable feat, from a linguistic standpoint, happened in the following years, when new, younger kids started attending schools for the deaf where LSN was, by then, commonly used by their older schoolmates:

‘Sign languages experts visited Nicaragua in the 1980s to study the children. They found that the older signers were actually less fluent (used less complex signs, signed more slowly or hesitantly) than the younger signers – those children who had entered the school at a later date and learned more recently. For instance, older children signed for manner and path of motion events simultaneously while younger, more fluent children signed manner and path sequentially.’ (Fasold and Linton, p226).

There again, the transition from pidgin to creole was striking enough for the resulting language to receive a name of its own: Idioma de Signos Nicaragüense (ISN). Although there may be controversy as to where exactly the switch from protolanguage to full-fledged language happened, all agree that the form of ISN in use by today’s Nicaraguan deaf children is a bona fide language, complete with complex grammar structures and syntax markers, the genesis of which seemed unaffected by either Spanish or other sign languages.

For Steven Pinker and other proponents of the hardwired language acquisition module theory, this brought irrefutable evidence that the leading factor in the creolization process comes not from the influence of a – here non-existent – substrate or superstrate language, but from the children themselves who demonstrated a seemingly innate ability to elaborate grammar rules and improve on the very incomplete language set they were handed. Pinker further developed his point by citing other cases of children whose parent were themselves not fluent in sign language, developing far better abilities, to the point of adding structures of their own, where their parents used none (Pinker 1994 p39).

IV. The impact of creolization theories

The methodical studying of the creolization process from a psycho-linguistic standpoint that was initiated in the late 1960s and its cognitivist interpretation had major consequences on two different levels:

First, they initiated a profound shift in the social and ethnological perception of creole-speaking societies. By clearly demonstrating a gap from pidgins to creoles and establishing the later as languages of their own, linguists helped shattering the image of creoles as mere “broken languages” and the pejorative, sometimes racist, pseudoscientific theories that fed on it.

A second, considerably more controversial, aspect to creolization studies is its implication on language learning theories:

In the late 1950s, verbal behaviorism (Skinner 1957) occupied the foremost position among theories of language. Building on a long tradition of analytical behaviorists, backed by the works of Wittgenstein or Quine, Skinner had successfully imposed a strictly behaviorist vision of language learning as a mere set of trial and errors: “Rules, are derived from contingencies, which specify discriminative stimuli, responses, and consequences” (Skinner 1984 p153).

In 1959, with a strongly critical review of Skinner’s book (Chomsky 1959), Noam Chomsky set in motion the shift toward a cognitive model he was to help shape over the following two decades. However, Chomsky’s earlier rebukes to Skinner’s “all behaviors, no mental states” theories rested on few empirical facts. Chiefly, the “lexical explosion” phenomenon, a phase of young children’s development during which their linguistic abilities suddenly increase in bursts, regardless of environment, suggesting the existence of relevant mental states or abilities triggering rapid learning of a language. The weakness of the “language reinforcement” hypothesis put forth by verbal behaviorism was also asserted through sheer observation: a young child does not, in fact, learn new words solely by repeating them in the presence of older individuals who repeatedly correct him.

While Chomsky’s further efforts went mostly into formalizing his hypotheses of an underlying universal grammar of language – independent of all “external languages” and therefore giving credence to the existence of a biological linguistic module – it is only with such empirical data as that brought by creolization studies, that his model could reach the level of popularity in the scientific community that it is enjoying today. By bringing evidence that entire populations were able to acquire and produce fully-structured languages from limited input and with no other “reinforcement” than their own native abilities, cognitivists were able to take to the very foundations of behaviorism.

Strong nativists like Pinker offer a diversified array of arguments against linguistic determinism, including direct attacks at the scientific validity of the research leading to it (citing Geoffrey Pullum’s rebuke of Whorf’s alleged “Eskimo vocabulary” findings: Pinker 1994 p64). Yet, while discarding the now tenuously supported Sapir-Whorf hypothesis in its stronger formulations (”Weltanschauung”) proves a rather easy task, Pinker would likely not be able to push forth his own without the help of creolization studies and the new elements they bring into the equation.

V. Alternative theories of creolization

As we mentioned in the first part: for all the momentum gained by nativist theories of creolization in the past three decades, the controversy is anything but settled. If behaviorism has never recovered from the emergence of cognitive sciences and has today lost nearly all its support among linguists, some see in connectionism a worthy successor to behaviorism. Resting on a strong neurobiological basis, yet insisting on the importance of rule-governed learning, connectionism is able to challenge nativism on its own territory: while acknowledging the existence of a particular biological structure of the mind, it has no need for, and indeed refutes, the idea of innate rules dealing with symbol manipulation.

Using only mathematical properties of neural networks, connectionists have been able to exhibit artificial patterns of learning reminiscent of creolization to the extent that they did not relate to one single input or even a simple linear combination of inputs. To effectively carry over this model to real-life cases, one needs to consider no creolization is possible without continuous influence by substrate languages, long past the pidginization stage. A moderate substratist position extends the origins of this influence to adstrate languages: those that are neither superstrate nor substrate (Holm 2000).

While substratists theories, backed by a certain amount of research in artificial intelligence and neural networks, seem to bring satisfying explanations for most spoken creoles, they still fall short on cases of pure creolization from a signed pidgin (cf. pt. III). In the case of NSL, there is no evidence that the children involved had access to any other language spoken or signed (Spanish, ASL or other) and therefore no way to account for the emergence of grammatical structures without resorting to nativist arguments (innate rules, bio-programming etc.).

Beyond the scope of this essay, yet relevant to the controversy, is the whole neurological aspect of language acquisition. To this day, an exact “geographical” mapping of language functions to specific areas of the brain is beyond reach of current neurobiology and a controversial hypothesis in itself. However, through the study of brain impairments involving speech or language acquisition, neuro-cognitivists were able to show some evidence of modularity: aphasic patients with full mental abilities and untouched non-verbal skills on one hand, mentally disabled patients with perfect, if not exceptionally developed, speech abilities, on the other. While it can be argued that modularization is not necessarily innate and can be a part of development, its mere existence implies an important biological factor to language acquisition.

VI. Beyond creole…

With the battle raging between nativist and anti-nativist proponents of language acquisition theories and the crucial arbitrating role played by emprirical data, creolization studies keep receiving acute attention from the scientific community. If today’s emerging creolization cases are fortunately triggered by very different socio-political contexts, they still do arise and will likely continue to, for as long as language diversity and the need to communicate subsist.

As for existing creoles, they generally suffer the same difficulties as many older languages sharing similarly low number of speakers: language streamlining and certain forms of explicit or implicit cultural imperialisms all concur to their slow disappearance through assimilation. This assimilation process sometimes takes the shape of “reverse creolization” or “decreolization”: a tendency of creole speakers to “correct” their language over time under the continued influence of a culturally dominating superstrate language (Sato 1994). Studying of this post-creole continuum has so far lead to renewed controversies regarding the importance of language-internal mechanisms against social and psychological factors.

References and Bibliography

Bickerton, Derek. “Creoles and the bankruptcy of current acquisition theory”. Creole Languages and Language Acquisition, edited by Herman Wekker. Berlin: Mouton de Gruyter, 1995. 33-43.

Bickerton, Derek. Language & Species. University Of Chicago Press; Reprint edition, 1992.

Chaudenson, Robert. La Créolisation : Théories, Applications, Implications. Paris: L’Harmattan, 2004.

Chomsky, Noam. “Review of Verbal Behavior, by B.F. Skinner”. Language 35, no. 1 (January-March 1959). Reprint. in The Structure of Language, edited by Fodor and Katz. New York: Prentice-Hall, 1964.

Chomsky, Noam. Syntactic Structures. The Hague: Mouton, 1957. Reprint. Berlin and New York, 1985.

Derrida, Jacques. Le monolinguisme de l’autre, Paris: Galilée, 1996.

Fasold, Ralph and Connor-Linton, Jeffrey. An Introduction to Language and Linguistics. Cambridge: CUP, 2006.

Holm, John. An Introduction to Pidgins and Creoles. Cambridge: CUP, 2000.

Holm, John. Pidgins and Creoles. 2 vols. Cambridge: CUP, 1988.

McWhorter, John H. Defining Creole. Oxford University Press, 2005.

Mufwene, Salikoko. “Pidgins and creole languages”. International Encyclopedia of the Social and Behavioral Sciences, 2002.

Pinker, Steven. How the Mind Works. W. W. Norton & Company, 1999.

Pinker, Steven. The Language Instinct. Penguin, 1994.

Roepper, Thomas. “Comment on Bickerton’s paper”, Creole Languages and Language Acquisition, edited by Herman Wekker. Berlin – New York: Mouton de Gruyter, 1995. 45-49.

Sato, Charlene J. “Language change in a creole continuum: decreolization?”. Progression and Regression in Language: Sociocultural, Neuropsychological and Linguistic, edited by Kenneth Hyltenstam and Ake Viberg. Cambridge: CUP, 1994.

Senghas, Ann, Kita, Sotaro and Özyürek, Asli. “Children Creating Core Properties of Language: Evidence from an Emerging Sign Language in Nicaragua”. Science 305, 1779 (2004).

Skinner, B. F. Abstract for “An Operant Analysis of Problem Solving”. Behavioral and Brain Sciences, 7, 583 (1984).

Skinner, B. F. Verbal Behavior. 1957. Reprint. by the B. F. Skinner Foundation, 1992.

Wittgenstein, Ludwig J. J. Philosophical Investigations. Trans. G. E. M. Anscombe. Blackwell Publishers, 2002.

Online Sources

• Wikipedia – Creolization
• Wikipedia – Creole Language
• Wikipedia – Pidgin
• Pidgin and Creole Languages – Salikoko Mufwene
• Stanford Encyclopedia of Philosophy – Behaviorism
• The Creolization of Pidgin: A Connectionist Exploration
• Wikipedia – Noam Chomsky
• Timothy Mason – Could Chomsky be Wrong
• Wikipedia – Sapir-Whord Hypothesis
• The Synthetic Modeling of Language Origins
• Wikipedia – Deep Structure