New Zealand English - the physics perspective

You may have read in the papers, seen on the web or heard on the radio about our research into the New Zealand English Language dialect. The mathematical modelling work was undertaken in collaboration with Gareth Baxter (a physicist at the University of Wellington, New Zealand), Bill Croft (a linguist at the University of New Mexico, USA) and Alan McKane (a physicist at the University of Manchester, UK).

A point unfortunately missed by the media is that none of this would have been possible without a comprehensive set of data having previously been published by a (separate) team of researchers based at the University of Canterbury, New Zealand. The outcome of these many years of hard work is published in the book by Elizabeth Gordon, Lyle Campbell, Jennifer Hay, Margaret Maclagan, Andrea Sudbury and Peter Trudgill “New Zealand English: Its origins and evolution” (Cambridge University Press, 2004). More details can be found on the Origins of New Zealand English (ONZE) project homepage. A theory for new dialect formation, based on these data, has been advanced in a companion volume by Peter Trudgill “New-dialect formation: The inevitability of colonial Englishes” (Edinburgh University Press, 2004). Both books are a fascinating read.

What’s physics got to do with it?

A large part of physics involves the study of how things move - planets around a sun, water down a pipe or electrons in a solid, for example. Here we’re looking out how linguistic utterances flow around a population of speakers. Particularly relevant to the formation of New Zealand English is the fact that initially many different variants of, for example, vowel sounds, were in use by different speakers, but ultimately, very similar variants were used by almost the whole population. According to the empirical study mentioned above, this happened very quickly - over 50 years, or roughly two generations of speakers.

Our contribution has been to take specific models of language use and see if they can explain this rapid pace of change in the population of 100,000 to 1,000,000 speakers living on the islands between 1850 and 1900. The mathematical methods used have much in common with the physics of gases, liquids and solid where you have large numbers of particles and try and predict the behaviour of the system as a whole given the (statistical) tendencies of its constituents.

What kind of changes are you interested in?

The empirical study describes changes in the pronunciation of various sounds, mostly vowels (such as that in ‘dress’, found also in words like ‘step’, ‘bread’ and ‘seven’). By looking at the geographical origins of the early immigrants to New Zealand (mostly different regions of Britain and Ireland) and painstakingly counting utterances of such words in a set of sound recordings made in the 1940s of speakers born in the second half of the 19th century, the raising of the ‘dress’ vowel (and other changes) could be charted over time. Any model we come up with must be capable of reproducing these kinds of changes.

The data also further suggest that the dialect is formed through a procedure of elimination of different variants present within the range of British and Irish regional dialects of the speakers who first settled in New Zealand. A theory for how these changes took place is presented in the book by Trudgill. The central questions from a physicist’s point of view are: which variants end up surviving; and how long does this process of elimination take?

What are your conclusions?

This is to some extent a work in progress, and we haven’t completely unravelled the question of how the elimination process was so rapid in such a large population of individuals who could not possibly have met each other. Mass media can't be invoked as an explanation: quite apart from there being no evidence of it facilitating language acquisition by children, it hadn't even been invented at the time that New Zealand English was formed. However, our modelling suggests that social differences between speakers were relevant to the process of new dialect formation in the isolated community, contrary to (our interpretation of) the theory proposed by Trudgill. Whilst one finds that the dialect that is formed does reflect the proportions of variants used by the immigrant populations in a similar way to that seen empirically, it appears that in a socially neutral model, the rate of change is simply too slow for the process to be complete after fifty years.

The reason for this can be understood through a tennis-match analogy. If two players are very evenly-matched, they must play for a long time to determine the winner: many games go to a deuce, and the sets to tie-breaks before the rare event of one player breaking the other’s serve occurs. Variant vowel sounds must similarly fight it out to the bitter end, unless speakers have a preference towards one over the other. Whilst we believe these preferences are social in origin, this remains a hypothesis yet to be confirmed, and their precise nature remains unclear.

So, contrary to what you may have heard, the riddle of the rapid formation of the stable, relatively homogeneous present-day New Zealand English dialect is not entirely solved.

The details of our model and analysis can be found in this online preprint.