Australian English phonology

Australian English (AuE) is a non-rhotic variety of English spoken by most native-born Australians. Phonologically, it is one of the most regionally homogeneous language varieties in the world. Australian English is notable for vowel length contrasts which are absent from most English dialects.

The Australian English vowels, , and  are noticeably closer (pronounced with a higher tongue position) than their contemporary Received Pronunciation equivalents. However, a recent short-front vowel chain shift has resulted in younger generations having lower positions than this for the former three vowels.

Vowels


The vowels of Australian English can be divided according to length. The long vowels, which include monophthongs and diphthongs, mostly correspond to the tense vowels used in analyses of Received Pronunciation (RP) as well as its centring diphthongs. The short vowels, consisting only of monophthongs, correspond to the RP lax vowels. There exist pairs of long and short vowels with overlapping vowel quality giving Australian English phonemic length distinction.

There are two families of phonemic transcriptions of Australian English: revised ones, which attempt to more accurately represent the phonetic sounds of Australian English; and the Mitchell-Delbridge system, which is minimally distinct from Jones' original transcription of RP. This page uses a revised transcription based on Durie and Hajek (1994) and Harrington, Cox and Evans (1997) but also shows the Mitchell-Delbridge equivalents as this system is commonly used for example in the Macquarie Dictionary and much literature, even recent.


 * As with General American, the weak vowel merger is nearly complete in Australian English: unstressed is merged with  (schwa) except before a following velar. New Zealand English takes it a step further and merges all instances of  with  (even in stressed syllables), which is why the New Zealand pronunciation of the dish name fish and chips as  sounds like 'fush and chups' to Australians. In Australian English,  is restricted to unstressed syllables, as in most dialects.
 * The trap-bath split is a regional variable in Australia, with the vowel  being more common in South Australia than elsewhere. This is due to the fact that that state was settled later than the rest of Australia, when the lengthened pronunciation was already a feature of London speech. Research done by  shows that the word graph is pronounced with the  vowel  by 86% speakers from Adelaide, whereas 100% speakers from Hobart use the  vowel in this word: . There are words in which the  vowel is much less common; for instance, Crystal reports that both the word grasp and the verb to contrast are most commonly pronounced with the  vowel:, . This also affects the pronunciation of some placenames; Castlemaine is locally , but speakers from outside of Victoria often pronounce that name  by analogy to the noun castle in their local accent.

Monophthongs

 * The target for is closer to cardinal  than in other dialects. The aforementioned phrase fish and chips as pronounced by an Australian ( in narrow transcription) can sound a lot like feesh and cheeps to speakers of New Zealand English and other dialects, whereas words such as bit and sit may sound like beat and seat, respectively.
 * The sound is usually pronounced as a diphthong (or disyllabically, like ) only in open syllables. In closed syllables, it is distinguished from  primarily by length  and from  by the significant onset in the latter.
 * tends to be higher than the corresponding vowel in General American or RP. The typical realization is close-mid, although for some speakers it may be even closer (according to John Wells, this pronunciation can occur only in Broad varieties). A recent change is the lowering of  to the  region.
 * For some Victorian speakers, has merged with  in pre-lateral environments, and thus the words celery and salary are homophonous as . See salary-celery merger.
 * The sound is traditionally transcribed and analysed the same as the short, but minimal pairs exist in at least some Australians' speech. It is found in the adjectives bad, mad, glad and sad, before the  sound (for example, hag, rag, bag) and also in content words before  and  in the same syllable (for example, ham, tan, plant). In South Australia, plant is usually pronounced with the vowel sound , as in rather and father. In some speakers, especially those with the broad accent,  and  will be shifted toward  and , respectively.
 * There is æ-tensing before a nasal consonant. The nasal sounds create changes in preceding vowels because air can flow into the nose during the vowel. Nasal consonants can also affect the articulation of a vowel. Thus, for many speakers, the vowel in words like jam, man, dam and hand is shifted towards . This is also present in General American and Cockney English. Length has become the main difference between words like 'ban' and 'Ben', with 'ban' pronounced  and 'Ben' pronounced.
 * is pronounced as open front by many younger speakers.
 * As with New Zealand English, the / vowel in words like park, calm and farm  is central (in the past even front) in terms of tongue position and non-rhotic. This is the same vowel sound used by speakers of the Boston accent of North Eastern New England in the United States. Thus the phrase park the car is said identically by a New Zealander, Australian or Bostonian. This vowel is only distinguished from the  vowel by length, thus: park  versus puck.
 * The phoneme is pronounced at least as high as , and has a lowered F3 that might indicate that it is rounded . The ⟨ɜ⟩ glyph is used — rather than ⟨ɘ⟩ or ⟨ɵ⟩ — as most revisions of the phonemic orthography for Australian English predate the 1993 modifications to the International Phonetic Alphabet. At the time, ⟨ɜ⟩ was suitable for any mid central vowel, rounded or unrounded.
 * The schwa is a highly variable sound. For this reason, it is not shown on the vowel charts to the right. The word-final schwa in comma and letter is often lowered to  so that it strongly resembles the  vowel : . As the latter is a checked vowel (meaning that it cannot occur in a final stressed position) and the lowering of  is not categorical (meaning that those words can be also pronounced  and, whereas strut is never pronounced ), this sound is considered to belong to the  phoneme. The word-initial schwa (as in enduring ) is typically mid : . In the word-internal position (as in bottom ),  is raised to : , as in American English roses . Thus, the difference between the  of paddock and the  of panic lies in the backness of the vowels, rather than their height: . In the rest of the article, those allophones of  are all transcribed with the broad symbol ⟨ə⟩:  etc.  is also broadly transcribed with ⟨ɪ⟩: , which does not capture its closeness.

Diphthongs

 * The vowel has an onset, except before laterals. The onset is often lowered to , so that beat is  for some speakers.
 * As in American English and modern RP, the final vowel in words like happy and city is pronounced as (happee, citee), not as  (happy-tensing).
 * In some parts of Australia, a fully backed allophone of, transcribed , is common before . As a result, the pairs full/fool and pull/pool differ phonetically only in vowel length for those speakers. The usual allophone is further forward in New South Wales than Victoria. It is moving further forwards, however, in both regions at a similar rate.
 * The second elements of and  on the one hand and  on the other are somewhat different. The first two approach the  vowel, whereas the ending point of  is more similar to the  vowel , which is why it tends to be written with ⟨ɑe⟩ in modern sources. John Wells writes this phoneme , with the same ending point as  and  (which he writes with ⟨ʌɪ⟩ and ⟨ɔɪ⟩). However, the second element of  is not nearly as different from that of the other fronting-closing diphthongs as the ending point of  is from that of , which is the reason why ⟨ɑɪ⟩ is used in this article.
 * The first element of may be raised and rounded in broad accents.
 * The first element of is significantly lower  than in many other dialects of English.
 * There is significant allophonic variation in, including a backed allophone before a word-final or preconsonantal . The first part of this allophone is in the same position as , but  differs from it in that it possesses an additional closing glide, which also makes it longer than.
 * is shifted to among some speakers. This realisation has its roots in South Australia but is becoming more common among younger speakers across the country.
 * The phoneme is rare and almost extinct. Most speakers consistently use  or  (before ) instead. Many cases of RP  are pronounced instead with the  phoneme in Australian English. "pour" and "poor", "more" and "moor" and "shore" and "sure" are homophones, but "tore" and "tour" remain distinct.

Examples of vowels

 * One needs to be very careful of the symbol, which represents different vowels: the vowel in the Harrington, Cox and Evans (1997) system (transcribed  in the other system), but the  vowel in the Mitchell-Delbridge system (transcribed  in the other system).
 * The fourth column is the OED transcription, taken from the OED website.

It differs somewhat from the ad hoc Wikipedia transcription used in this article. In a few instances the OED example word differs from the others given in this table; these are appended at the end of the second column following a semicolon.

Consonants
Australian English consonants are similar to those of other non-rhotic varieties of English. A table containing the consonant phonemes is given below.


 * Non-rhoticity
 * Australian English is non-rhotic; in other words, the sound does not appear at the end of a syllable or immediately before a consonant. So the words butter, here  and park  will not contain the  sound.


 * Linking and intrusive
 * The sound can occur when a word that has a final $⟨r⟩$ in the spelling comes before another word that starts with a vowel. For example, in car alarm the sound  can occur in car because here it comes before another word beginning with a vowel. The words far, far more and farm do not contain an  but far out will contain the linking  sound because the next word starts with a vowel sound.
 * An intrusive may be inserted before a vowel in words that do not have $⟨r⟩$ in the spelling. For example, drawing will sound like draw-ring, saw it will sound like sore it, the tuner is and the tuna is will both be . This  occurs between,  and  and the following vowel regardless of the historical presence or absence of . Between ,  and  (and  whenever it stems from the earlier ) and the following vowel, the -ful pronunciation is the historical one.
 * Flapping
 * Intervocalic (and for some speakers ) undergo voicing and flapping to the alveolar tap  after the stressed syllable and before unstressed vowels (as in butter, party) and syllabic  or  (bottle, button ), as well as at the end of a word or morpheme before any vowel (what else , whatever ). For those speakers where  also undergoes the change, there will be homophony, for example, metal and medal or petal and pedal will sound the same ( and , respectively). In formal speech  is retained.  in the cluster  can elide. As a result, in quick speech, words like winner and winter can become homophonous (as ). This is a quality that Australian English shares most notably with North American English.


 * T-glottalisation
 * Some speakers use a glottal stop as an allophone of  in final position, for example trait, habit; or in medial position, such as a  followed by a syllabic  is often realized as a glottal stop, for example button or fatten. Alveolar pronunciations nevertheless predominate.


 * Pronunciation of
 * The alveolar lateral approximant is velarised  in pre-pausal and preconsonantal positions and often also in morpheme-final positions before a vowel. There have been some suggestions that onset  is also velarised, although that needs to be further researched. Some speakers vocalise preconsonantal, syllable-final and syllabic instances of  to a close back vowel similar to, so that milk can be pronounced  and noodle . This is more common in South Australia than elsewhere.


 * Yod-dropping and coalescence
 * Standard Australian English usually coalesces and  into  and  respectively. Because of this palatalisation, dune is pronounced as, exactly like June, and the first syllable of Tuesday  is pronounced like choose . That said, there is stylistic and social variation in this feature.  and  in the clusters  and  are similarly affricated.
 * Word initial and  have merged with  and  respectively. Other cases of  and  are often pronounced respectively  and, as in assume  and resume  (ashume and rezhume).
 * Similarly, has merged with  word initially. Remaining cases of  are often pronounced simply as  in colloquial speech.
 * and other common sequences of consonant plus, are retained.
 * For some speakers, (or "sh") may be uttered instead of /s/ before the stressed /tj/ sound in words like student, history, eschew, street and Australia – As a result, in quick speech, eschew will sound like esh-chew. According to author Wayne P. Lawrence, "this phonemic change seems to be neither dialectal nor regional", as it can also be found among some American, Canadian, British and New Zealand English speakers as well.

Other features

 * Between voiced sounds, the glottal fricative may be realised as voiced, so that e.g. behind may be pronounced as either  or.
 * The sequence is realised as a voiceless palatal fricative, so that e.g. huge is pronounced.
 * The word foyer is usually pronounced, as in NZ and American English, rather than as in British English.
 * The word data is commonly pronounced, with being the second most common, and  being very rare.
 * The trans- prefix is pronounced, even in South Australia, where the trap–bath split is significantly more advanced than in other states.
 * In English, upward inflexion (a rise in the pitch of the voice at the end of an utterance) typically signals a question. Some Australian English speakers commonly use a form of upward inflexion in their speech that is not associated with asking questions. Some speakers use upward inflexion as a way of including their conversational partner in the dialogue. This is also common in Californian English.

Relationship to other varieties
Australian English pronunciation is most similar to that of New Zealand English; many people from other parts of the world often cannot distinguish them but there are differences. New Zealand English has centralised and the other short front vowels are higher. New Zealand English more strongly maintains the diphthongal quality of the NEAR and SQUARE vowels and they can be merged as something around. New Zealand English does not have the bad-lad split, but like Victoria has merged with  in pre-lateral environments.

Both New Zealand English and Australian English are also similar to South African English, so they have even been grouped together under the common label "southern hemisphere Englishes". Like the other two varieties in that group, Australian English pronunciation bears some similarities to dialects from the South-East of Britain;  Thus, it is non-rhotic and has the trap-bath split although, as indicated above, this split was not completed in Australia as it was in England, so many words that have the  vowel in Southeastern England retain the  vowel in Australia.

Historically, the Australian English speaking manuals endorsed the lengthening of before unvoiced fricatives however this has since been reversed. Australian English lacks some innovations in Cockney since the settling of Australia, such as the use of a glottal stop in many places where a would be found, th-fronting, and h-dropping. Flapping, which Australian English shares with New Zealand English and North American English, is also found in Cockney, where it occurs as a common alternative to the glottal stop in the intervocalic position. The word butter as pronounced by an Australian or New Zealander can be homophonous with the Cockney pronunciation (which could also be  instead).

AusTalk
AusTalk is a database of Australian speech from all regions of the country. Initially, 1000 adult voices were planned to be recorded in the period between June 2011 and June 2016. By the end of it, voices of 861 speakers with ages ranging from 18 to 83 were recorded into the database, each lasting approximately an hour. The database is expected to be expanded in future, to include children's voices and more variations. As well as providing a resource for cultural studies, the database is expected to help improve speech-based technology, such as speech recognition systems and hearing aids.

The AusTalk database was collected as part of the Big Australian Speech Corpus (Big ASC) project, a collaboration between Australian universities and the speech technology experts.