25 English influence on the pronunciation of re-awakened Aboriginal languages

This chapter explores the influence of literacy and teaching, by first language speakers of English, on the pronunciation of Aboriginal languages in the context of language re-awakening in New South Wales (NSW). Wherever languages are learned in the absence of a generation of first language speakers we find that the learners’ first language will have a major impact – the linguistic resources that you have to build on play a strong role in shaping the new language that you acquire. This paper canvasses some pronunciation changes currently taking place in NSW in the context of learning revitalised languages. It raises the need for open discussion about the authenticity of re-created languages and argues that, for re-created languages, phonemic orthographies might not be the best choice. While this paper focuses on New South Wales its arguments may be relevant to other parts of the country where re-creation-type programs are underway. What is being learned in revitalisation programs Language re-awakening work undertaken in NSW typically involves learners whose first language is Australian English (from standard to Aboriginal English varieties) engaged in the learning of Aboriginal languages. The input that learners receive is generally either written language in the form of wordlists, learner guides or other pedagogical materials, or spoken language samples modeled by someone else who also learned pronunciation from written sources. In some lucky cases there are still Elders with enough speaking knowledge to record words as pronunciation guides, however the usual scenario involves careful decision-making about how words should be pronounced and sentences constructed, under two serious restrictions: the absence 1 School of Behavioural, Cognitive & Social Science, University of New England. 294 Re-awakening languages of any community of first language speakers of the target language, and the paucity of the materials available. Such learning is fundamentally different from normal second language learning. When you learn a second language you can access information to answer any questions that arise, and you have the option of immersion among first language speakers. In NSW these options are not available. The paucity of materials available for even the best documented languages (probably Awabakal, Bundjalung, Gumbaynggirr, Paakantji and Wiradjuri), provide us with basic grammatical descriptions, but tell us little about such simple things as how to have a conversation. Learning a language under these restrictions inevitably induces changes in that language. Some changes, such as creating new vocabulary, result from deliberate language engineering. Others, such as changes to pronunciation and grammar, are likely to be less deliberate and may largely result from the inherent difficulties of learning a language in the absence of native speaker models. For these reasons, although the goals of revitalisation programs are often worded in terms of ‘getting our old language back’, the outcomes of many are likely to be quite different from the traditional languages that they are based on. This is no criticism, just a statement of the inevitable. No language has ever ceased to be spoken and then later revived in a way that is the same as the earlier form. Even the muchcited example of Israeli (Zuckermann 2005) turns out to now be, although healthy, a Germanic/Hebrew hybrid language, vastly different from Hebrew as it was last spoken. We understand now that, because any language reflects the communicative needs and social world of its speakers, the same language cannot do that for two groups of people displaced in time, society and culture. With respect to pronunciation in particular, wherever a generation of learners revitalises a language in the absence of first language speakers, the learners’ first language will have a major impact on the sound system of the target language. Details of the changes taking place Here we consider some of the ways that changes are taking place in NSW languages in the context of revitalisation learning. We can find examples of induced change in all areas of language. Sometimes we find that verb suffixes become simplified, so that a single form of a verb is used in a non-inflecting way for all tense categories, for example in Paakantji the use of the present participle ending {-ana} on all remembered verb forms regardless of their actual tense (Thieberger 2002, p. 322). In other cases we find case marking on nouns either simplified or avoided, and even case suffixes detached and used like prepositions. Syntactically we can hear the development of simplified or fixed word order, often based on English. We also find many changes taking place in sound systems, and here we’ll focus on just four types of pronunciation change. Literacy and oracy 295 Neutralisation of rhotic contrasts Most NSW languages have traditional phonemic contrasts among more than one r sound – usually a flap or trill written as rr, contrasting with a continuant (more like the English r) often written as r. Some of the northern NSW coastal languages appear to have a third r sound. Many early written sources failed to distinguish among these sounds so, in many cases, it is difficult to know which pronunciation is right. In the context of language revitalisation programs many learners have circumvented the question by adopting various simplification strategies. Some pronounce the continuant r in all cases, a quite natural conflation for anyone whose first language is a variety of English. A few learners go the other way (what’s known as hyper-correction) and pronounce the trill rr in all cases. Other people might adopt the strategy of using only one sound mostly, but being careful to distinguish between them for just those important minimal pairs, for example being careful to pronounce wirri and wiri differently, but otherwise just using a single r sound where it doesn’t really affect the meaning. Loss of variation in stop Most NSW languages have just a single series of phonemic stops (sounds that block off airflow completely). Orthographies for these languages usually use a single series of symbols, either b, dh, d, dj, g or alternatively p, th, t, tj, k. In NSW the voiced symbols happen to have predominated, though there are some exceptions like Paakantji. Being phonemes means that these stops function as contrasting sounds in the minds of their first language speakers. But choosing to write them with either b, dh, d, dj, g or with p, th, t, tj, k tells us nothing about how they would have been traditionally pronounced. In any given language it was likely that both voiced and unvoiced stop sounds could be heard, depending on what part of the word they appeared in and what other sounds surrounded them. To use a made-up example, a word [pabap] with unvoiced stops initially and finally but voiced stops medially, could be written phonemically as babab in one language but as papap in another, even though it is pronounced identically. In NSW revitalisation programs, phonemic orthographies have been widely adopted under considered input from linguists who tend to promote them as being the best linguistic practice. They are best practice for first language and second language literacy, however phonemic orthographies tempt Aboriginal people trying to reawaken a language in the absence of first language speakers, falling back on their knowledge of English orthography, to pronounce such words ‘as they are spelled’. So babab tends to be pronounced as [babab], and papap tends to be pronounced as [papap]. This is happening quite widely in NSW, so we tend to now hear that Paakantji begins with a [p], and Gamilaraay with a [g], regardless of how they might have once been pronounced. Where previously in each language the phonetic realisation of stops depended on word position and preceding or following sounds, now that pattern is 296 Re-awakening languages being replaced by one where stop sounds, at all places in a word, are more likely to be either all voiced or all voiceless. Because the voiced symbols have predominated in NSW, we are currently hearing an escalation of voiced stop pronunciations; that is, orthography is driving change in pronunciation. Affricated realisation of palatal stops While the finer articulatory details of the realisation of palatal stops can vary considerably across Aboriginal languages (see Butcher 1995), it is likely that when most NSW languages were spoken as first languages their palatal stops were unaffricated stops made with tongue tip down and blade raised. Now it is increasingly common to hear palatal stops (International Phonetic Alphabet [IPA] symbol [ɟ] and [c]) realised as palato-alveolar affricates (the j of English jam, IPA symbol [dʒ], or the ch of English chew, IPA symbol [tʃ]), so putative word badjanu is pronounced [bʌdʒʌnu] rather than [bʌɟʌnu]. This is phonetically a fairly natural shift, so a link to English is not necessary. However the influence of English is the likely explanation here. A contributing factor is the many, well-intended, learner pronunciation guides (see Reid 2008, p. 5 for an example) that casually describe palatal stops as being ‘like ch in English ‘chew’’. Neutralisation of unstressed vowels Vowel inventories differ in only small ways across NSW with typically three vowel places and often also a short/long vowel contrast, yielding systems of six phonemic vowels, typically written as a, aa, i, ii, u, uu. As is fairly typical of small vowel systems (Butcher 1994) in NSW languages we find the traditional pattern of some minor allophony, but generally vowel phonemes are quite discrete. There is little evidence of any vowel sound being an allophone of more than one phoneme. Nor is there widespread evidence of the centralisation of unstressed vowels. This can be contrasted with English where schwa [ə] is an allophone of most vowel phonemes, and the common realisation of vowels in u