Reconciling Universalists and Substratists

Table of Contents

Introduction

X. Setting up the Priorities - Some Miscellanea

1. Bickerton's Bioprogram – a Theory and its Opponents
1.1 The LBH vs Substrate Theory
1.2. Salikoko Mufwene on LBH

2. Substrate Theory Reconsidered
2.1 Mufwene and Arguments in Favour
2.2 Mufwene on the Shortcomings of the Substratist Framework
2.3 Creolization from an SLA Perspective

3. Mufwene and Alliances: a Universal-Substrate Synthesis
3.1. Defining Creole Genesis
3.2 The Role of Universals
3.3 The Nature of Universals

4. Concluding Remarks

Bibliography

Introduction

By way of an introduction to the following paper, I would like to draw here on a quote taken from one of Salikoko Mufwene's essays: "...creolists generally agree on the nature of the sociohistorical contexts which have produced these languages, but they disagree essentially on the natures of the linguistic processes which resulted in them." (1986:129).

This sentence quite neatly captures what the general pidgin/creole-debate is all about. The various approaches to pidginization and creolization and on how, i.e. by which underlying processes, the respective language systems supposedly came into being have this one thing in common: they all entail, respectively proceed from the assumption in the first place, that they have something decisive to say about the nature of language in general. Therefore the different positions are often defended most decidedly, trying, or so it seems, to lay claim to a final definition of language in one or the other light. As such, I like to describe this phenomenon as some kind of linguistic-philosophical debate. And this is what the subject of the following paper shall be about: What are the various approaches, how convincing are they, i.e., who has the best arguments or is able to disprove opposing views best? In this sense, the following will be a theoretical rather than practical, case-study paper. The discussion can be roughly described in terms of two major opposing viewpoints: the universalist one and a more cognitive-oriented, functional-pragmatic. The latter is called substratist for the most. The two camps tend to put either more weight on the structural or the sociohistorical aspect respectively. It is especially the nativization phase, known as creolization, which interests me most in this paper.

Providing the major part of the material to be discussed, Salikoko Mufwene's work will serve as a guiding line through this paper; this being for two reasons: first, he quite simply seems to have to say something about every aspect of creolization and how they work together. Second, he does so by, quite convincingly I should say, drawing from a well-structured minimal set of principles on which he can act in various discussion points. I hope to be able to make the meaning of this more obvious as the paper proceeds. For this purpose, selected papers will now be looked at in detail, for one part stemming from the work of Derek Bickerton, on his part being the most influential representative from the universalist of the two opposing sides. On the other hand the very Salikoko Mufwene will be regarded and, to a lesser degree, Jacques Arends, those taking the part of the critical voices. I am going to try, though, to take into account comments made by other parties, as well. To this effect also other essays will, if only in parts, be considered. Structure-wise, the main body will be threefold: one part focusses on the specifics of the Bickertonian universalist model, another one is dedicated to the substratist line of argumentation and a third will give a general overview of Mufwene's theoretical contributions. The Bickerton-related part will be dealt with in more detail than the substratist one. This is for the mere and simple reason that Bickerton's work and theorizing has traditionally effected many arguments than any individual representative of the substratist side. This I wish to account for by including one specific contribution by Jaques Arends, selected for its exemplary character with regard to criticism of Bickerton's work in general and preceded by a brief recapitulation of the specifics of the same. Apart from that, both sides, universalist and substratist, will be discussed in the light of this paper's title – i.e. from a Mufwenian viewpoint – be it in methodological and conceptual terms or simply good or bad theorizing. The aim here is to lay out the relative position Mufwene is taking towards substratist and universalist theory respectively. For the sake of completeness I am also going to insert some considerations regarding Second Language Acquisition (SLA) studies. The focus here will be on the contributions made by S. E. Carroll, who seems to draw on principles very similar to those of Mufwene's. Having thus established a number of facts we will look at the conceptual framework that Mufwene proposes as a whole.

The title of this paper most obviously excludes approaches other than the ones specifically named, but this is for a reason. This reason, then, I will briefly explain in a first, or rather preposed, part which has the mere function of forming a basis from which to start. This is hopefully going to justify the focus I will be taking concerning the two major camps. For this purpose, I chose a structure that in a way violates the usual one by putting "minor" thoughts at the very beginning. It will be along the lines of Mufwene's argumentation, as well, that I am going to do so.

X. Setting up the Priorities - Some Miscellanea

The purpose of this chapter is in principle to define those theories to be excluded from the discussion in this paper: the simplification hypothesis and Hancock’s (1986) "African geo-genetic hypothesis"^[1]. As the case lies, we can be quite straightforward in listing the reasons for doing so. In fact, we can limit ourselves to a single one: either party, advocates to the simplification approach or Hancock, finally has to resign to one or the other higher framework, universal or substratist in character. In Mufwene 1986, the author points out that Hancock himself believes that either substrate influence or universals had a role to play in the formation of one African language, called Guinea Coast Creole English (132), which he believes to be the source of the English-based Caribbean languages. As for the simplification hypothesis (as advocated by well-established linguists^[2] ) Mufwene again, and righteously so, asks for an explanation as to "what principles guided and constrained the putative simplifications" (132) which were to result in structures showing undeniable similarities among creole languages world-wide. Although Schuchardt and Silverstein apparently go for a universalist tendency, in general there has been no attempt at a wholesome explanation so far, neither is the nature of the proposed universals at all made clear. Again, it all boils down to the revelation that there must be either some inter-language forces at work, universals as that, of the languages participating in a given contact situation – as proposed by Bickerton. Or, alternatively, substrate grammatical features must be looked at with regard to a possible explanation as claimed substratists such as Jaques Arends. As it were, either of the two will have a role to play in the chapter to follow.

1. Bickerton's Bioprogram – a Theory and its Opponents

Again, a Mufwenian quote will serve for establishing a starting point: "...the fact that the same basic distinctions obtain word-wide among various creoles is certainly significant" (1986:139).

Obviously, these similarities are the main reason why universalist ideas developed in the first place. This in mind, I would like to first sum up the prototype of all universalist theory, i.e., its pivotal ideas. One central concept developed by Bickerton with regard to his Language Bioprogram Hypothesis (LBH) is probably the one of unigenerationality, meaning in essence that, according to "the innate bioprogram that determines the form of human universals" (Bickerton 1981: 134), a nativized creole is to unfold within the 1stgeneration of plantation-born children, slaves for the most, or not at all. Furthermore he believes pidginization to be an act of SLA whereby gradual relexification and restructuring of the native language takes place (1977, 1981a). Creolization then he sees as the realization of the language bioprogram, only fully operative during first language acquisition (FLA) and producing so called "rogue grammars" in absence of an input (mother tongue) language. The latter he believes not to play any role in creole formation. This he argues on the grounds that the pidgin input which first generation children encounter is said to be "degenerate" (1996:34) and highly diffuse, thus forcing the speaker-to-be to resort to this inherent mechanism. He also assumes a specific scale of “creoleness” whose core terms are "semi-creolization" and the one of "true creoles" which he makes conditional on certain linguistic prerequisites. The basis for his theoretical framework is essentially made up of his own studies, conducted on Hawaiian Pidgin and Creole.

In defence of his bioprogram against substratist views, he argues that the similarities between creoles and certain African languages, as pointed out by substratists can all but be called "superficial[...]" (1986: 25). He says that "producing [a] superficially similar surface structure" cannot give sufficient proof; to this end only a grammar can be acceptable. He especially stresses the notion of "surface structure", thus already implying an inherently U(niversal)G(rammar)-based approach and way of argumentation (25). Rules, he goes on to argue, that make for differences between languages on the surface structure do not exist at deep structure. While this may be perceived as a bit of a killer argument which can do away with a number of objections, his theory still offers some points which lend themselves to heated discussions.

1.1 The LBH vs. Substrate Theory

In this subchapter I will first turn to a general discussion initiated by Jaques Arends (1992) which is essentially of a methodological nature, criticising Bickerton, as will be seen, for drawing wrong conclusions from the wrong kind of data. I am also going to include some thoughts of my own, in part triggered by Arend’s critique, finally followed by a focus on Mufwene's wordings. But first we shall have a look at the details of Arends' comments on the well-known approach of Bickerton.

One of the many points of dispute is the ad hoc – gradualness-distinction implying the first-generation creolization idea (Bickerton) on the one hand and a step by step approach on the other, this being suggested by Arends (1992), among others. Arends provides the grounds for, respectively derives an explanation out of sociohistorical and demographic facts which is in line with Mufwene's theory. Therefore, we can fit in this part within a Mufwene-based approach^[3]. His first critique of Bickerton's ab ovo approach bears on the latter's claim that the creoles as we know them are not related to each other whatsoever; a claim so central to his theory that it would be rendered absolutely worthless as soon as it was to be refuted convincingly. After all, in this case one single substrate source could be assumed to have existed which simply spread its features among the many creole languages, just to mention one possible alternative. As opposed to Bickerton, he takes a "transgenerational" view, as he calls it (374), meaning essentially that it took several generations for a Creole to come about as a fully working system. He draws his conviction from the assumption that Bickerton's data could not possibly reflect the real creolization processes involved in the respective formations.

He states the faults as follows: In view of the fact that Bickerton draws on current creolization data one cannot assume those to reflect exactly this first stable stage that a language at some point "culminates" into (371). The 20thcentury data may only work within a synchronic approach which, it follows, can impossibly explain a diachronic phenomenon such as language formation, here pidginization/creolization, in the course of time. Bickerton, he goes on, seems to be justifying this by just another unproven assumption, namely just the one concerning unigenerationality: If a creole develops within one generation during which unalterable universalist structural elements come into action, then there is "no need for Bickerton to take historical records [...] into account even if they had been available to him" (372). As he makes clear, there are historical records available for evaluation (372) which must not simply be ignored if one really wanted to get at the truth. Furthermore, he quotes sources from which it seems to be obvious for mere demographical reasons that nativization by children cannot have occurred in most cases: there had been too few of them around to have had a significant influence. So he believes that the continuous creolization process was not primarily carried out by locally-born children acquiring a creole as their 1stlanguage, simply because there had been only one or two children present on the Hawaiian plantations during a longer period of time which roughly coincided with the “critical phase” (he states the years between 1670 and 1690, p. 375). There must have been, then, a longer stage in which SLA was the foremost language formation process. From data drawn from Price (1976) he derives a mere 10% of creole speakers for a significant period of time who, in turn, were surrounded by "salt water slaves" (375) that were posing a counterweight to a possible nativization process. So the demographic factors seem to be working against Bickerton's account. Arends also argues that restructurings of Sranan between 1750 and 1850 can only have been effected by the successive cohorts of adult African slaves learning English as their second or third language. Creolization, it must finally follow, is a gradual and differential process to which also Hawaiian makes no exception. Yet another conclusion to be drawn from all this is that the scaling of different creoles according to degrees of "creoleness" cannot be considered to be based on real-world fact, thus making groundless notions such as the one of “semi-creoles”, as advocated by Bickerton (see above).

Apart from these demographic considerations, I feel that other objections as to Bickerton’s methodology and manner of arguing have to be raised, as already hinted at by Arends in the preceding paragraph. Therefore, I would now like to include some thoughts of my own in relation to Bickerton's line of argumentation with respect to a paper called "Creoles and West African languages: a Case of Mistaken Identity?" (1986)^[4]. First of all, I believe him to be superficial at pints – as we have seen this is what he himself accuses substratists of – one example for this would be his explanation concerning a common focus marker that can be found in Yoriba and Vata, two West African languages: Solely based on the fact that the two are quite unrelated he claims this particular feature to be very common and therefore widespread (28). Therefore, he speculate, from a substratist viewpoint it should ideally be included in many creoles, thus jumping to conclusions which lack empirical profoundness. This particular feature not behaving, as substratists allegedly predict, he takes as sufficient proof of his own point of view. Also the mere fact that it is only one property that he enlists followed by such a far reaching conclusion should be regarded sceptically. Some vague claim is made about the possibility of there having been a stage in Saramaccan when the creole patterned in one particular feature like Haitian ("...some historical evidence [...] which suggests...", 29), on this basis, though, he concludes that it did so in general (35), thus implying creole-resemblances between Haitian and Saramaccan on a very unsolid basis. He himself also disregards a possible superstrate influence, claiming a universalist explanation where the substratist putatively fail to come up with a satisfactory account. The differences between West African languages and creoles pointed out by him, though, might be explained by superstrate influence (36). As we will see later on, though, the superstrate role has been neglected badly in many other approaches, as well, so one should perhaps not hold this too much against him individually. Either way his reaction to the African genesis theory seems slightly opportunistic in my eyes when he writes that: "If such rules happen to be present in the input in certain cases [...], the creole will acquire such rules, not because they are in the input [...], but because such a rule is required by the structure of the emerging language" (1977: 51). While, on the one hand, it serves as a defence of his blueprint concept, this statement also hints at another point of issue important within the Bickertonian debate, namely the one of the role of the input. We will come back to this shortly when we look at what Mufwene has to comment on exactly this.

Finally, there is of course always the question of why, if a bioprogram can be assumed to exist, languages differ at all. Bickerton explains this in a rather woolly manner by historical changes (1981) bringing about the deviation of natural languages from UG, an explanation which Arends at some point called "obscure" (1992: 371).

[...]

^[1] a term coined by Mufwene (1986:131)

^[2] such as Schuchardt (1914), Jespersen (1922), Bloomfield (1933), Hall (1966) and Silverstein (1972)

^[3] Mufwene actually draws on Arends in his own argumentation (see e.g. 2000: 69)

^[4] A title, by the way, which mocks one of his critics, John Holm, who used the same phrase in one of his papers (1986: 272)