Realizational differences

In the second type of accent difference, part of the system of phonemes may be the same for two or more accents, but the realizations of that phoneme or set of phonemes will vary. For instance, in SSBE, SSE and GA, /l/ has two main allophones, being clear, or alveolar [l] before a stressed vowel, as in light, clear, but dark, velarized after a stressed vowel, as in dull, hill. This distribution of allophones is not the only possibility in English, however. In some accents, /l/ is always realized as clear; this is true, for instance, of Tyneside English (or ‘Geordie’), Welsh English, and some South African varieties. On the other hand, in Australia and New Zealand, /l/ is consistently pronounced dark; and indeed, realizations may be pharyngeal rather than velar, or in other words, pronounced with a restriction even further back in the vocal tract. In London English, there is a further allophone of /l/, namely a vocalized (or vowel-like) realization finally or before a consonant: in sell, tall, people, help, /l/ is typically realized as a high or high mid back vowel like [υ] or [o]. For younger speakers, vocalization is also taking hold in medial position, in words like million; and the process is also spreading beyond London, as part of the shift towards so-called ‘Estuary English’, a mixture of SSBE and London English which is arguably becoming a new standard for young people, especially in urban centres in the south of England.

The other English liquid consonant, /r/, also provides plenty of scope for realizational differences. /r/ is typically an alveolar or slightly retroflex approximant for SSBE and GA, but at least in medial position, is frequently realized as an alveolar tap in SSE (the tap is also a common realization in South African English). In some parts of the north of England, notably in Northumberland and County Durham, a voiced The other English liquid consonant, /r/, also provides plenty of scope for realizational differences. /r/ is typically an alveolar or slightly retroflex approximant for SSBE and GA, but at least in medial position, is frequently realized as an alveolar tap in SSE (the tap is also a common realization in South African English). In some parts of the north of England, notably in Northumberland and County Durham, a voiced uvular fricative [ʁ] is quite commonly found, although this may be receding gradually.

In other areas of northern England, this time notably Yorkshire, Tyneside and Liverpool, [ɹ] appears as an allophone of /t/, typically between vowels and across a word-boundary, as in not on [nɒɹɒn], lot of laughs [lɒɹə …],get a job [ εɹə …]. In Merseyside, voiceless stops are very generally realised as fricatives or affricates in word-final position, so that cake, luck, bike will be [keIx], [lυx], [baIx]: whereas in Scots and SSE the appearance of [x] in loch constitutes a systemic difference, as there are minimal pairs establishing an opposition of /x/ and /k/, in Liverpool the velar fricative is clearly an allophone of /k/, so that the accent difference between, say, SSBE and Merseyside English in this respect is realizational, but not systemic.

Turning to vowels, one particularly salient example involves the FACE and GOAT vowels, which in SSBE, NZE and Australian English are pronounced consistently as diphthongs. In GA, the FACE vowel is diphthongal, while the GOAT vowel may be a monophthong; and in SSE and SgE, both are monophthongal, with the predominant allophones being high-mid [e] and [o] in both accents. The NURSE vowel in SSBE is mid central; the same phoneme in NZE is very generally rounded, while in SgE it is typically raised to high-mid back unrounded, or high back unrounded [ɯ] (as we might expect, Hokkien has, Malay has bothand [ɯ], but both lack).

Sometimes, although these realizational differences have no direct impact on the phoneme system, they do lead to neutralizations of otherwise consistent contrasts. For instance, we saw in the last section that SgE speakers raise /ε/ to [e] before plosives and affricates; the monophthongal pronunciation of /e/ as [e] in FACE words, and the lack of any systematic vowel-length distinction in SgE means that the contrast of /ε/ and /e/ is suspended in this context, leading to identical pronunciations of bread and braid, or wreck and rake. It is also possible for realizational differences in vowels to lead to allophonic differences in consonants. For instance, right at the beginning of this book, we identified an allophonic difference between velar [k] and palatal [c], with the latter appearing adjacent to a front vowel. In SSBE, SSE and GA, this will mean that velar realizations will be produced in cupboard and car, palatals in kitchen and keys. However, the distribution differs in other varieties of English, depending on their typical realizations of the FLEECE and KIT vowels. In NZE, FLEECE has a high front diphthong, so that keys will still have [c]; but no fronting will take place in kitchen, since the KIT set in NZE has central [ə]. On the other hand, in Australian English, KIT has a rather high, front [i] vowel so that kitchen will certainly attract a palatal [c]; but in some varieties at least, the diphthong in keys is central [əI], which will therefore favor a velar allophone of /k/.