Between 1946 and 1948, Radio New Zealand established a Mobile Unit that visited many towns throughout the country, seeking out people who were longtime residents of small towns in order to record their oral histories. In 1986 Elizabeth Gordon was told of these archived recordings. The uniqueness of this data corpus is that the speakers, born between 1851 and 1904, were all participants in the formation of a new dialect, New Zealand English (NZE). It is unlikely that other data sets will be found in which tape recording technology and a first generation of speakers come together. English had arrived in 1840 with the original colonizers, who were mainly English, Scottish, and Irish. The story gets complicated by the arrival of many English-speaking immigrants from Australia, descendants from a penal colony founded in 1788, which had formed its own new dialect earlier with input from English, Scottish, and Irish settlers. Moreover, settlers often spent time first in Australia and then moved to New Zealand, and there was considerable contact between Australia and New Zealand from the start. It is within this historical setting and with this database that the Origins of New Zealand English (ONZE) project researchers set out to describe early NZE in order to examine its origins.
Although there is no detailed description of the interviews – for instance, no discussion of technical or stylistic quality, length, or sample transcriptions – it is clear that these are ordinary people. The map of the geographical distribution of the Mobile Unit interviews indicates that the speakers came mostly from small towns, with concentrations in the southern part of the South Island and the northern part of the North Island. It is clear that the radio broadcasters were not interested in the urban settlements of Auckland and Christchurch, although Wellington was included. Since the analysis shows that towns of mixed ethnic makeup are most advanced in the changes that will become NZE, one can perhaps speculate that these two major cities were the most advanced.
The project researchers represent an impressive group with a number of specialties among them: historical linguistics, dialect geography, sociolinguistics, and phonetics, all perspectives that inform the data analysis and interpretation. That the researchers have different interests in the interpretation of the data is clear though understated, and it is probably this characteristic that makes the book so compelling. This is exciting science with much at stake.
The book opens with three chapters devoted to background information about the ONZE project, New Zealand English, and the settlement history of New Zealand. This is followed by a thorough account of the plethora of commentaries, myths, speculations, and theories about the origin of New Zealand English in particular and about new dialect formation in general.
A methodological chapter presents the three procedures used in the data analysis. The first was a broad auditory perception study of 95 speakers; it involved listening to interviews to obtain an overall evaluation of the speakers' phonetics and phonology and to listen for specific features of lexical phonology. Narrow IPA transcriptions were made where relevant, and some relative frequencies were noted. The variables included all of the vowel phonemes of NZE, including the unstressed vowels, consonants such as /r/ and /l/, and other items of particular interest. Although no statistical analysis was made of these data, the information was computerized and could be interrogated for specific linguistic features. Second, an acoustic analysis was made of the entire vowel systems of five women and five men; the normalized vowel plots are given in an appendix. The third approach to the data involved the detailed quantitative analysis of eight linguistic variables and four social variables for 59 speakers. The linguistic variables included the presence of post vocalic /r/; the presence of word-initial /h/; the /h/ ∼ /hw/ contrast; the centralization of the KIT and DRESS vowels; the raising of the TRAP vowel; the fronting of the START vowel; and the merger of the unstressed vowels. An index was derived for each speaker for each linguistic variable, and this became the input to the statistical analysis. The social variables were speaker's date of birth, sex, parents' birthplace; and town type (Scottish, English, Mixed, etc.). Statistical analysis revealed a break between the linguistic patterns of speakers born before and after 1875, and almost all of the quantified variables are reported with this age variable.
The researchers use three statistical techniques in the quantitative analysis of the data. They used linear regression to model the speakers; a linguistic index was calculated for each speaker for each linguistic variable, and the social characteristics (sex, date of birth, parents' birthplace and type of town the speaker lived in) retained by the model are said to play a statistically significant role in predicting the patterns of variation. Logistic regression (similar to VARBRUL) was also used to model the contextual structure of the linguistic variables, including preceding and following environments, as well as the social variables. Raw data are displayed on the graphs, and the authors decided not to provide the correlation coefficients for ease of interpretation by readers unfamiliar with the techniques. It is surprising, however, that the correlation coefficients are not presented in an appendix for those readers who do want to examine them. Further data exploration involved the use of CART regression trees; this statistic produces a series of binary splits in the data in an attempt to predict, for instance, the speakers' index of rhoticity. It was the repeated split of the variables based on speakers' date of birth that led to the discovery of a significant age division between speakers born before and after 1875.
Chapter 6, “The Variables of Early NZ English,” is the empirical heart and takes up almost a third of the book. Nineteen phonological variables are discussed; the eight mentioned above are given the full quantitative treatment. In all cases, the account of the variables begins with a detailed phonetic description of the variants of the variable and is followed by an often lengthy discussion of the potential variants that were brought to New Zealand by the English and Scottish migrants in particular, but also by Irish and Australians in some cases. Many of the variants being studied had already begun to change in England and Scotland. Of particular interest to the researchers are Scottish features (centralization of the KIT vowel or [r] pronunciation of non-prevocalic /r/, for instance), because many scholars and other commentators attribute a number of features of NZE to Scottish input. Following this introductory material, the data are analyzed using the techniques described above. Although it is not easy or perhaps even wise to generalize across so many variables, one is left with the impression that almost all of the social and linguistic variables are statistically significant, as is the time factor (born before or after 1875). For a number of variants, the results challenge the received wisdom; for instance, it is speakers whose parents were born in New Zealand who centralize the KIT vowel, and not those with Scottish-born parents. The authors recommend caution in claiming a Scottish origin for this feature.
Chapter 7 returns to the subject of ideas about the origin of NZE that were canvassed earlier. Most are dismissed quite easily, including the Cockney and Scottish connections. The “Cockney” label was used to disparage NZE rather than accurately to describe it, and Scottish is shown to have influenced primarily the rhoticity of a small region in New Zealand. However, even these speakers showed low frequencies for rhoticity, and it has almost disappeared in modern NZE. The two conflicting origin stories considered at length are that NZE has been influenced by Australian English (AusE), and that Trudgill's theory of new dialect formation can account for the facts without considering any AusE influence. This theoretical juxtaposition builds up throughout the book, and readers may be disappointed by the final outcome – especially knowing that it would be hard to find a better data set to demonstrate that one case is more persuasive than the other. For both perspectives, both supportive and ambiguous evidence is brought to bear, leading to the conclusion that neither theory can be fully supported nor can either theory be dismissed. Trudgill 2004 has published his own book arguing his interpretation in support of new dialect formation.
I would like to add a comment about the use of the vowel variant in words of the DANCE class; much is made of the variable use of /æ/ and /a:/. In arguing for the process of reallocation (where, in the process of new dialect formation, such a variable acquires a sociolinguistic function rather than being a distinguishing characteristic of a separate dialect), the authors inexplicably illustrate their point by using an AusE example rather than one of the NZE variables. They claim that the [æ ∼a:] for the DANCE class is a sociolinguistic variable; Horvath & Horvath's (2001) study of six cities in Australia found this variable to be a strong regional variant but not a sociolinguistic variant in any city.
This book would make an ideal centerpiece for a graduate seminar: the data set is truly unique, the multiplicity of analytical approaches is instructive, and the interpretation of the findings make lively discussion likely.
