Similarities and differences in settlement

Australia and New Zealand were both discovered for Anglophones by James Cook. Australia was first settled as a penal colony in 1788 and New Zealand was officially settled following the Treaty of Waitangi (signed between the Crown and the New Zealand tribes) in 1840, though by that time there was already a great deal of contact with Maori people and a fair amount of de facto settlement. There was also considerable trade by 1840 between Australia and New Zealand (Bauer 1994a: 382). With dates of settlement so close together, and the close links that have, post-British settlement, always existed between the two countries, Australia and New Zealand are often seen from a European perspective as forming a larger coherent area (the antipodes, or Australasia), losing track of the notion that the distance between Sydney and Wellington is about the same as the distance between London and Labrador.

The biggest difference between the settlements, though, is the difference in the nature of the earliest settlers. Various Australian sites were settled as places of transportation for convicts and the population was made up of those convicts and the people who were sent to be in charge of them. New Zealand was largely settled by those who desired to own land but either could not get land in Britain or had been dispossessed of the land they had held.

At later periods, there was also a big difference in the patterns of immigration, with Australia taking large numbers of settlers from Italy and Greece, and later from Vietnam, areas from which the flow of immigrants to New Zealand was relatively limited.

Assumed sources

Hammarström (1980) argues, on the basis of pronunciation alone, that Australian English is derived from vernacular London English. More recent scholarship sees this as unlikely. Rather, nineteenth century London English and Australian English are mixed dialects with approximately the same inputs (Trudgill 1986; Turner 1994). The same is true for New Zealand English (Gordon et al. 2004; Hay, Maclagan, and Gordon 2008: ch. 5). The inputs are not exactly the same for Australian and New Zealand English (see above), but they are similar enough to have led to similar-sounding outputs. Where New Zealand English is concerned, the similarities are intensified by the fact that there was a great deal of contact between New South Wales and New Zealand in the second half of the nineteenth century, despite the distances involved. That high level of contact continues to this day.

The picture from pronunciation alone is not necessarily particularly clear, but when we look at other factors, the similarities between early Australian English and early New Zealand English are striking. The number of new vocabulary items they share and the number of expressions borrowed from a wide range of British dialects that they share (Bauer 1994a, 2000) can only be explained by close contact between the two.

Variation: regional, social, historical

Both Australia and New Zealand have long been said to be homogeneous linguistic areas (see Bauer 2008for some discussion). This homogeneity is not social or ethnic, but regional. Only one regional dialect is readily recognized in New Zealand, that of southern Otago and Southland (Bartlett 1992). It is differentiated from the English spoken in the rest of New Zealand by a relatively high level of rhoticity and by a number of vocabulary items and expressions that are clearly Scots in origin. There are other regional dialects in New Zealand (Ainsworth 2004; Bauer and Bauer 2005; Kennedy 2006) but they are not part of the lay perception of dialects in New Zealand. Australia is now developing regional dialects (Bryant 1989; Bradley 1989), but they are quite new.

Social variation is readily recognized. This was addressed by Mitchell and Delbridge (1965) by assigning varieties of Australian English to one of three layers: broad, general, and cultivated. The three-way split reflects the split paraphrased by Kurath (1972: 164) in another context as one between “the folk, the middle class and the cultured”. It is not clear that this three-way division ever held for Australia, given that there is considerable leakage between the levels (Bernard 1989); in New Zealand the labels were, to a large extent, adopted uncritically from the Australian experience, without any experimentation or testing. It remains a useful set of labels for dividing up the spectrum of accents in the two countries in an unsystematic way, but it cannot be given scientific content today.

There is also gender-based and ethnic variation in the two countries. I shall have little to say about these in this contribution. There is a great deal of evidence of women leading the way in phonetic change in New Zealand as they do elsewhere in the English-speaking world (see, for example, Holmes 1997and Maclagan 1998, 2000), and forms that appear nearer the “cultivated” end of any social spectrum are often more common in women’s speech. Ethnic variation is very different in the two countries, but in New Zealand so-called “Maori English” often reflects phonetic features from the “broad” end of the social spectrum (though see Warren and Bauer 2004for more detail and commentary on exceptions).

There is also a great deal of historical change in Australian and New Zealand English, so that recordings from the 1940s sound strange on both sides of the Tasman Sea that separates the two countries. Comments will be made on historical developments, particularly recent ones, in the course of this contribution.

Australasia as a single linguistic area with variation within it

On the basis of the discussion above, there is a sense in which we can see Australia–New Zealand as a single linguistic area with some regional diversification within it. Not only is there the evidence of lexis mentioned above, there is also a certain amount of (controversial) evidence that New Zealanders cannot recognize Australian accents as infallibly as they think they can and vice versa (Bayard 1995; Weatherall, Gallois, and Pittam 1998). Certainly, people external to the two countries have difficulty in distinguishing the two. Accordingly, in this contribution, Australia and New Zealand will be treated together. This should not be interpreted as meaning that Australian and New Zealand Englishes are “the same dialect” or “the same variety”; they are not. However, treating them together allows a relatively economical way of looking at the phonetics of the varieties.

The author’s point of view

The author, though an Englishman, is resident in New Zealand and is more familiar with the New Zealand situation than with the Australian one. Accordingly, in this presentation, the New Zealand versions will tend to be taken as the default, while Australian versions are treated as variation on a New Zealand theme. This may do something to make up for the occasions where New Zealand English has been treated as a variant of Australian English. Forms that are specifically Australian or New Zealand will be marked as “AuE” or “NZE” respectively. When speaking about both together, forms will be marked as “ANZE”.

Acoustics

In Table 15.1 figures for formant 1 and formant 2 of the vowels of New Zealand and Australian English are provided. The figures are derived by averaging the seven values given for New Zealand male speakers in Easton and Bauer (2000) and the two values given for Australian male speakers in the same publication. Since the earliest of these speakers were recorded 30 years before the most recent, the figures are likely to be rather conservative values.

Note that when a vowel sound is produced, two major resonators in the vocal tract, one between the lips and the point at which the tongue most obstructs the vocal tract and one between the most posterior part of the tongue and the vocal folds, each produce sound at a specific resonant frequency. These show up on a sound spectrogram or other analysis of the sound wave as bands of acoustic energy, termed formants. The relationship between formant frequency and vowel position is not always linear, but indicates relative articulatory position, with greater values for Formant 1 showing lower vowels, greater values for Formant 2 showing fronter vowels. Even in these figures, some of the shibboleths that distinguish AuE vowels from NZE vowels can be seen: the more open (lower) and retracted (backer) KIT vowel in NZ, the more open DRESS and TRAP in AuE.

Articulation, the stressed vowels

A general comment on the description of the vowels in Australian and New Zealand Englishes is probably in place. Lip rounding in these varieties is not to be equated with a pouting gesture (as it would be, for instance, in French). Rather the lips may appear quite spread and tense, but the air flow is directed through a narrow channel between the lips. The lips are held in a relatively neutral position at virtually all times. A “rounded” vowel in these varieties is thus not articulatorily the same as a rounded vowel in RP, and this may have an effect (yet to be determined) on the acoustic characterization of the vowels. For more details on vowels in general see Bauer and Warren (2004).

DRESS

The DRESS vowel is more open in AuE (where it is typically transcribed as [ɛ], although that makes it look rather more open than it actually is) than in NZE (where it is typically transcribed as [e], which also makes it look more open than it actually is). Innovative NZE pronunciation in the last few years has found some tokens of this vowel overlapping with tokens of FLEECE (Bell 1997), from which it is distinguished in terms of vowel length. Despite this, DRESS is quite a long vowel in both AuE and NZ, which helps explain the neutralization with SQUARE in NZE before /r/ (see the next section below on neutralization).

Neutralization

The patterns of neutralization of stressed vowels before /r/ and /l/ are not the same in AuE and NZE. NZE has a fuller set of these neutralizations, especially before /l/, while the patterns of neutralization in AuE appear to be still developing. The difficulty in describing these neutralizations is that they are sociolinguistically variable, with the result that neutralization is not always a transitive relationship: that is, if A is neutralized with B and B is neutralized with C in the same environment, it does not follow for any given speaker that A will be neutralized with C. Table 15.2 sets out the most developed cases of neutralization, with examples and comments.

Articulation, the unstressed vowels

There are no COMMA-KIT minimal pairs in unstressed syllables: villages and villagers, chatted and chattered are either homophonous or are consciously distinguished by the use of the NURSE vowel in villagers and chattered.

This means that there are basically two unstressed vowels to be considered, with a third one arising from the vocalization of /l/. The happY vowel is phonemically associated with the FLEECE vowel, the two vowels of seedy being perceived as “the same” while the two of city are clearly different in NZE. This is supported by the fact that the happY vowel can be diphthongized, as the FLEECE vowel may be, though whether the range of diphthongization is the same in the two cases has not been investigated. The happY vowel is also often rather longer than might be expected, but this will be considered in the next section below on general comments on length.

The other vowel, which we can term the commA vowel, has a range of realizations that may overlap auditorily with STRUT, especially in the word-final position or with NZE KIT. In the phrase-final position this vowel is often lengthened considerably (see also below in the next section).

At least in NZ, this vowel is typically used in contexts where RP would have syllabic consonants, providing a vocalic nucleus for these weak syllables. This vowel is also used to distinguish a few pairs such as groan and grown, where the past participle marker has a full syllable with the comma vowel.

General comments on length

As in the south-eastern varieties of British English from which ANZE fundamentally derives (neither AuE nor NZE retain phonological traces of distinctively northern or south-western British features), there is a distinction between long and short vowels. Despite the distinction being fundamentally the same as that found in Britain, the actual vowels that can be paired in terms of length are not the ones that are paired by the Jones’ systems of transcription, nor those paired by the orthographic system. Phonetic pairings are those shown in Table 15.3. It should also be noted that the STRUT vowel is not strictly a checked vowel, as it is in RP. The common phrase see ya! (a farewell) regularly ends in a stressed STRUT vowel.

As well as this phonological vowel length, there are two types of phonetic vowel length. The first is the type also found in other varieties of English, whereby vowels are lengthened in syllables without codas or in syllables where the coda is a voiced obstruent. Such lengthening has been referred to above as occurring “in lengthening environments”. Nothing further will be said about this. There is also important prosodic lengthening, particularly at phrase boundaries or for emphasis. In most places, this does not disturb phonological structure, but a phrase-final STRUT may be perceived by speakers of these varieties as being the START vowel. It should be noted that this is true even when the STRUT vowel in question is the commA vowel, which has become more open in the phrase-final position. Thus a phrase like Look at the koala can be perceived as having final START. This same phenomenon may account for the final happY vowel being perceived as being the same as the FLEECE vowel.

Plosives

The voiced plosives /b/, /d/, and /ɡ/ are weakly voiced, as in other varieties of English. Where they are in coda-position, the length of the preceding sonorant is often the main clue available as to the phonological voicing of the plosive, sonorants being longer before phonologically voiced obstruents than before voiceless ones.

The grave (i.e., noncoronal) voiceless plosives /p/ and /k/ behave differently from the coronal (more narrowly, alveolar) /t/. Initially in a stressed syllable, all of these plosives are aspirated/affricated. The plosive /p/ is usually just aspirated, the other two are usually affricated to a greater or lesser degree, the quality of the friction occurring after /t/ suggesting a tongue-tip articulation for the plosive. Intervocalically before an unstressed or weakly stressed syllable, /p/ and /k/ are aspirated (/k/ is probably affricated), while /t/ is voiced (Silby 2008), usually with a quick enough articulation for a transcription as [ɾ] to be reasonable, but sometimes with an articulation that is not easily distinguishable from a [d]. In the pre-consonantal or pre-pausal position, the voiceless plosives may be unreleased, weakly released, aspirated (affricated), or glottalized (as set out in Table 15.4). Research on the distribution (phonological or sociological) or these variants (Holmes 1995a, 1995b) is now outdated, and incomplete in that it covers only /t/. Where glottalization is employed, there may be either glottal reinforcement or glottal replacement (see Wells 1982for the terminology). Where glottal replacement is found (as in the utterance overheard between NZ students a few years ago of [ʃɐːʔ ɐːːʔ] “shut up”), the glottal stop may take on phonemic status in its own right. The glottalized variants are far more current in NZE than in AuE (see Tollfree 2001on the situation in Australian English), and have arisen in the last fifteen years or so. This material is summarized in Table 15.4.

Fricatives

As in other varieties of English, the voiced fricatives are only weakly voiced, as also noted above for the voiced plosives. In NZE this sometimes results in some apparent movement between voiced and voiceless categories: for instance, president and precedent may become homophonous, the first fricative in positive may be variably voiced or voiceless, as may the first fricative in pessimistic. Equation has /ʃ/ rather than /ʒ/, which makes it morphophonemically regular. While thither usually has a voiceless initial fricative, this is presumably due to a voiceless model from Scottish English, rather than a local devoicing. In Maori English, /z/ (particularly in the final position, e.g., in freeze) is said to be devoiced even more than in mainstream New Zealand English (Bell 2000; Warren and Bauer 2004).

The fricatives /θ/ and/ð/ are increasingly replaced by /f/ and/v/ in the speech of younger speakers. Some words, such as with, are particularly strongly affected (Campbell and Gordon 1996; Wood 2003). This is not perceived as standard at this stage, however. In Maori English these fricatives may be affricated as [tθ] and [dð] (see Bell 2000).

The fricative /s/ is replaced by /ʃ/ (a) when followed by a phonemic /j/, where the /j/ may or may not be assimilated, so that consume can be [kənʃʉːm] or [kənʃjʉːm]; (b) optionally before a /ʧ/, as in student /ʃʧʉːdənt/; and (c) increasingly before /tr/, as in strong [ʃtɹɒŋ]. These forms are not proscribed and those in (a) may be considered cultivated. The voiced equivalent /z/ is palatalized only in environments corresponding to (a), so that presume can be [pɹɘʒʉːm] or [pɹɘʒjʉːm].

/l/

The phoneme /l/ is typically pronounced with a darker quality than in British RP, though for many speakers there is nonetheless a difference in quality between pre-vocalic and pre-consonantal allophones. However, there is variable vocalization of /l/ pre-consonantally or finally. In parallel with the historically earlier vocalization of /r/, this leads to instances of linking /l/, so that there may be no phonetic [l] in feel bad ([fiːɯ bæd]) but a phonetic [l] in feel it (fiːlʲ ɪt]).

The vocalization of /l/ leads to a range of phonetic outputs. In the simplest cases there is a back vowel, typically unrounded, of variable vowel height, perhaps [ɯ] or [ɤ]. In some cases the /l/ appears not to gain any realization at all (see just below) and in yet other cases, the /l/ is realized as a lengthening of the previous vowel (in some cases accompanied by a distinctive quality of that vowel). The clearest example of this last phenomenon is restricted to New Zealand and happens after the GOOSE vowel (possibly neutralized with the FOOT vowel before a historical /l/ anyway). In isolation, the GOOSE vowel is very front (transcribed here as [ʉ], but occasionally further forward than that suggests, possibly [Y]); where there is a following underlying /l/, this vowel is realized as [u]. In a word like school or pool, this [u] may be lengthened, but remain a monophthong: [skuː]. Similarly, in a word like milk, with the KIT vowel followed by /l/ and an obstruent, the /l/ may not be present, but the preceding vowel may be realized as [ɤ], [mɤk] (NZE), although this is only one of a number of potential realizations of this word, others including [miɤk] or [miɯk] (with a close onset to the diphthong). Following long back vowels as in walls there may be only minimal or no diphthongization of the vowel, and so no marker of the erstwhile /l/ at all, with pronunciations like [woːz] or [wo^əz].

Before /l/ there is considerable neutralization of vowel contrasts, which was discussed above in the section on neutralization.

/ɹ/

Both Australian and New Zealand Englishes are typically described as nonrhotic, but this hides a multitude of variable usages. Not only is there the Southland-Otago accent of New Zealand, which is typically characterized as rhotic, even though it tends to be rhotic only in the context of an immediately preceding NURSE vowel, as in word [wɵːrd] with some variable rhoticity following the LETTER vowel (letter may be [leɾəɹ]) and considerably less, if any, rhoticity following FORCE and START (in words like warm and farm), but there is also increasing rhoticity on both sides of the Tasman, particularly after NURSE and LETTER, but also occasionally elsewhere, in a way that at the moment appears to be essentially random. In New Zealand, this increasing rhoticity appears to be partly dialectal and partly ethnic (greater among Maori and Pasifika speakers – i.e., speakers with Pacific Island ethnicities), but is heard sporadically from speakers who do not belong to these groups (Kennedy 2006: Marsden 2013). Surprisingly, linking rhoticity in NZE has spread to following the MOUTH vowel, so that how/r/ever and now/r/ and again are frequently heard (Hay and Warren 2002).

The /r/ in all these instances is an apical alveolar-to-post-alveolar approximant [ɹ]. Devoicing and frication of the /r/ are found as in other varieties of English, following voiceless plosives and alveolar plosives respectively. An allophone [ɾ] is variably heard following /θ/ in words like through, although [ɾ] in the intervocalic position would be perceived as realizing an alveolar plosive.

The semi-vowels [j] and [w]

As in other varieties of English, [j] and [w] are devoiced and may be fricated following an initial voiceless plosive in a stressed syllable, so that pewter and cute may be [pçʉːtɐ], [pj˚ʉːtɐ], [k˖çʉːt], [k˖j˚ʉːt]. Where an alveolar plosive and [j] arise in a cluster, the output is generally an affricate, [ʧ] or [ʤ], so that dune and June become homophones.

Where one of FLEECE, GOOSE, GOAT, FACE, PRICE, CHOICE vowels (but not NZE MOUTH, for which see above) forms a sequence with another vowel, a glide arises between the two vowels, agreeing in backness and rounding with the first vowel, to prevent the hiatus. This occurs in sequences like see[j] it, be[j] in, do[w] it, lie[j] in, de[j]ontic, go[w] on, pro[w]active, and so on. These intrusive elements are distinct from the full phonemes /j/ and /w/: say ‘S’ is not homophonous with say ‘yes’, nor is know it homophonous with no wit. The intrusive elements are shorter, less firmly articulated and (where [w] is concerned) often less rounded; they are nevertheless auditorily distinguishable from the vowels that surround them.

Yod-dropping (Wells 1982), the loss of /j/ immediately following a coronal consonant, is variable. It has vanished completely following /s/ (as in superintendant) in the last 20 years, is still variable following /θ/ (as in enthusiasm), and is apparently partly lexically determined following /n/ (as in new; in New Zealand the loss is particularly noticeable in the item New Zealand). Following /t/ or /d/ the result is an affricate rather than /j/-loss, as described above.

In conservative varieties (and to a certain extent, in southern accents of New Zealand), /hw/ in which is still distinct from /w/ in witch, but the distinction is dying fast.

Nasals

There is little to say about the nasal consonants themselves, which behave as in other major varieties of English, but it should be noted that the nasality from nasal consonants easily spreads to adjacent segments. This is dealt with further below in the section on voice quality.

Stress

Word stress functions basically as in other standard “inner circle” varieties of English, and while there are minor lexical differences in occurrence, these do not disturb the fundamental system. However, even speakers who use this system of stress seem unsure about it. I was in the interesting position recently of having a class of first-year undergraduates tell me that revenue is stressed on the third syllable, even though they were pronouncing it with first-syllable stress. Moreover, in broadcasting, stress is more variable than might be expected, perhaps particularly so in noun–noun constructions, where the position of the main stress is notoriously difficult to predict.

Rhythm

While the fundamental underlying stress-timing inherited from British English is still present, it appears to be weakening, probably more so in NZE than in AuE. In the first place, this is due to the use of full vowels (the phonetic nature of which is determined by spelling pronunciation) where unstressed vowels are normal in RP or other British and American varieties. Pairs such as effect and affect, Johnston and Johnstone, are often distinguished by the use of different full vowels (recall that there is no COMMA-KIT distinction in unstressed syllables; speakers who cannot distinguish villagers from villages equally cannot use these means to distinguish effect and affect; see Bauer 1994b). Even grammatical words may be heard with full vowels in contexts where they are not stressed. This has the effect of leveling out the difference between stressed and unstressed syllables. In New Zealand, the trend away from stress-timing may be exaggerated by the effect of Maori English, where the rhythm is based on the original mora-timing of Maori, now being lost as vowel length is eroded in Maori (Maclagan et al. 2004).

Intonation

Discussion of the intonation of Australian and New Zealand Englishes has tended to focus on the High Rise Terminal first noted in print by Benton (1965) and discussed in detail for Australian English in Guy et al. (1986). This rise occurs on statements, and is used as a pragmatic device to check comprehension or to draw attention to critical parts of a narrative (Warren and Britain 2000; Warren 2005). It is perceived by outsiders as a questioning intonation, but is phonetically distinct from the intonation used to ask real questions (Warren and Daly 2005).

Otherwise the intonation patterns of these varieties are, in everyday usage, rather flat, and not as varied as RP is reported to be. Ainsworth (2004) reports on more varied intonation patterns in one area of New Zealand.

Voice quality

It seems likely that one of the distinctive features of English from this part of the world is voice quality, and it may also be that voice quality helps distinguish ethnic varieties in New Zealand and social varieties everywhere. However, this area has not been investigated in any depth or in any phonetic detail. Features that seem to be relevant include a generally relaxed articulation, including lack of great articulatory precision and for some varieties a rather slow delivery (though see Robb, Maclagan, and Chen 2004on the speed of NZE), an overall back resonance, and variable nasalization. The nasalization varies from the effect of adjacent nasal consonants to widespread nasal air flow, so that the vowels in had and ham are not auditorily distinct in their nasal quality.