Pattern of variation What to study

34

Chapter 4 The Data

This chapter narrows the focus of our study to the appropriate segment of HCE and describes the actual sources of data used.

4.1 Variation

Variation is probably the single most difficult challenge faced by this study. The problem with studying HCE is that HCE is not any one thing. It is a collection of registers, some of which are fairly close to Standard English SE, while others are only barely intelligible to the monolingual SE speaker.

4.1.1 Pattern of variation

How to best describe the pattern of variation has been a matter of discussion. In Tsuzaki 1971:330, HCE is described as having four coexisting and overlapping systems: An English-based pidgin, an English-based creole, a standard and a nonstandard dialect of English. According to Day 1972:146 however, “we are not dealing with variation of a number of different forms between systems in the Hawaiian Post-Creole Continuum, but a single pattern of variation.” Perlman 1973 concurs with Day, adding however that HCE does not fit into a strictly LINEAR continuum model. Finally, Peet 1978:11 argues unsuccessfully, I think that both the coexisting systems model and the continuum model could be correct simultaneously. 1 It seems to me that neither the coexistent systems hypothesis nor the continuum hypothesis make any obvious predictions. In lieu of any compelling evidence, in this study I simply adopt the terminology of the more popular of the two views, the continuum hypothesis. Although the exact nature of the variation in HCE is debatable, the situation is clearer on some more relevant issues. It is generally agreed that decreolization has had a major disruptive effect on the HCE continuum; however, among those closest to the HCE situation Labov, Grimes, and Forman p.c., it is generally agreed that the language has, for the most part, stabilized.

4.1.2 What to study

If HCE is not any one thing, what then is the proper object of study for the purposes of this investigation? I believe the inescapable conclusion is that the basilect is the register of choice for our study. I want to emphasize that I do mean to describe the actual speech of individuals, rather than create 1 According to Peet 1978:11, this is because Tsuzaki, a native speaker is perhaps giving the ‘emic’ insider’s viewpoint while Day also Perlman, Odo incidentally, a native speaker, Bickerton, etc. are adopting the ‘etic’ outsider’s viewpoint. It is not clear exactly what point Peet is trying to make. He seems to be saying that the insider Tsuzaki sees the language as having four distinct systems, while the outside investigator Day sees a continuum with no non-arbitrary divisions. That is, Day and the others have provided us with the external objective view of HCE while Tsuzaki provided us with the internal subjective view of native HCE speakers. It seems strange to evoke the dichotomy of ‘etic’ and ‘emic’ to contrast these two investigations. In my understanding, the ‘etic’ viewpoint is what the investigator brings with him to his investigation and the ‘emic’ viewpoint is what the investigator attempts to obtain during the course of his investigation. No researcher attempts to attain to the ‘etic’ view, since he already has it. It is doubtful that Day would have characterized his results in those terms. And if Day’s proposal was, in fact, taken from the ‘etic’ point of view, then it is most certainly wrong since if he had any particular viewpoint in mind, it would been the ‘emic’ viewpoint. 35 an artificial, idealized, and exaggerated system containing a conglomeration of the forms most deviant from SE, as some linguists have done. Labov 1990:25 warns us against such descriptions characterizing them as “a kind of grammaire des fautes, a pedagogical caricature of the language rather than a portrait of it.” Such a system exists only in the minds of linguists rather than in the minds of ordinary speakers. Labov ibid. continues: Whatever value such constructions have for teacher or learners, they cannot contribute any- thing to our study of the adequacy of languages. If such “pure” creoles are not used, they are not used for some good reason. We can only argue from grammars that are used for the serious purpose of communication in everyday life. The selection of the basilect as the object of our study helps us deal with some of the problems associated with investigating a language closely related to SE. First of all, the basilect tends to be the most stable part of the continuum. Second, it should be obvious that the part of HCE which provides us with the most useful data is the part which differs most from SE. A study searching for the NSM primitives in SE would be a meaningless and redundant exercise since the NSM formulation is specified in SE. Third, as Labov 1990:6 observes, closely related systems tend to interfere with each other more than distinct ones. In our study, a major concern is how having such closely related systems might affect speaker’s intuitions about sentences. As previously mentioned, decreolization is widespread in the continuum. Speakers are often competent in a range of registers. So, for example, it might be easy to mistake an SE sentence for an HCE sentence if it is pronounced with HCE phonology. HCE can be very tolerant of SE loan words, depending on the lexical domain. Furthermore, NSM clauses are simple, and NSM discourse structure bears a remarkable resemblance to HCE discourse structure. A question naturally arises, “How does one know when one has good basilectal data?” I believe I should state outright that I am unaware of any objective criteria for determining what good basilectal data is. Currently, “you’ll know it when you hear it” is the only guide I have.

4.2 Data Sources