Liverpool English

Liverpool English (LE) is the variety of English spoken in Liverpool and much of the surrounding county of Merseyside, in the north-west of England. After London, the north-west of England is the most densely populated of all regions in England and Wales, with the population of Liverpool standing at around 450,000. LE itself is said to have developed in the middle of the 19th century, after rapid immigration from Ireland during the Irish potato famines of 1845–1847 (see Knowles 1973). Arguably as a result of this immigration, as we will see, there are some similarities between LE's phonological system and those of Irish Englishes. Of course, as we might expect, the phonological system of LE maintains its connection with other northern Englishes, too.

There is a greater amount of previous work on LE than on many other accents of British English, particularly other varieties of the north-west of the country. The earliest systematic study of LE, Knowles (1973), remains the seminal work and is where the widest range of phonological features is considered. Later work, including De Lyon (1981), Honeybone (2001), Sangster (2001) and Watson (2006, 2007), has tended to restrict its focus to a smaller number of variables which are amongst the variety's most characteristic features. This article bases most of its descriptive detail on data gathered during fieldwork carried out by the author (see Watson 2007), but at times information is gleaned from elsewhere (most notably Knowles’ early work) to provide comparison.

The transcription of the reading passage is based on the speech of a 21-year-old working-class female speaker who was born in the district of Netherton, in the north of Liverpool, and has always lived there. She self-identifies as having a ‘broad’ Liverpool accent, although the difficulty in remaining objective about such labels should be acknowledged. There is considerable phonetic variation in LE according to age, gender and socioeconomic class, although this is an area where modern research is lacking. In what follows, where the speaker produces some phonetic feature which is known to be atypical of LE or where a feature varies according to some sociolinguistic parameter, this will be pointed out.

Consonants

Whilst the consonant system of LE is phonologically identical to most other varieties of English English, there is much allophonic variability. Of course, this is to be expected in every variety, but as we will see, the realisational potential of certain LE phonemes is much greater than elsewhere.

• 

Dental fricatives /ɵ, ð/ are often realised as dental stops [t ̪, d̪] both word-initially, medially and finally, although dental fricatives are also found. This is arguably a feature which has been innovated into LE from varieties of Irish (Honeybone 2004). It has recently been suggested that TH-fronting, the process by which /ɵ, ð/ are realised as labiodental fricatives [f, v] and which is frequent in many other British varieties (e.g. Milton Keynes, Reading, Hull (Williams & Kerswill 1999), Newcastle (Watt & Milroy 1999)), is infrequent in LE (Watson 2005). The speaker can be heard using both standard fricative variants ([ɵ] in north, [ð] in other) and the more localised dental stop ([d̪] in then). She does not use a labiodental variant.

Post-vocalic /r/ is absent in LE, so that words like car, farm, park are r-less. In pre-vocalic and intervocalic positions, /r/ is typically realised as [ɹ] or [ɾ]. The tap is common in intervocalic position (e.g. mi[ɾ]or, ve[ɾ]y) but can also occur when /r/ follows an onset obstruent (e.g. st[ɾ]ip, b[ɾ]eath, f[ɾ]ee). The speaker's variety is somewhat atypical of basilectal Liverpool English in this respect, as she uses the standard variant [ɹ] in all positions (e.g. ag[ɹ]eed, a[ɹ]ound, st[ɹ]onger). The tapped realisation of /r/ is not categorically absent from her repertoire, however, as she uses [ɾ] as the linking /r/ in stronger[ɾ] of the two. The labiodental variant, [ʋ], is not a feature of LE, despite it spreading in other accents of British English (Foulkes & Docherty 2000).

LE is similar to other accents in the north of England in that the /ɡ/ in 〈ng〉 clusters is maintained. For example, the speaker realises along as [əlɒŋɡ]. In the 〈-ing〉 morpheme, forms with the velar nasal and plosive are found, as in singing [sɪŋɡɪŋɡ], but a realisation of [ən] is also likely (e.g. sing[ən], walk[ən]). The speaker uses [ən] for the 〈-ing〉 of making. Another similarity between LE and elsewhere is the dropping of /^h/, most often in high frequency grammatical words (e.g. the speaker's realisation of him, who, his). H-dropping is not categorical, however, as the speaker's maintenance of [^h] in more he blew testifies. /^h/ is frequently maintained in lexical words (e.g. as it is in the speaker's hard).

It is in the system of plosives where the widest range of phonetic variation is attested. As in other varieties of English, voiceless plosives /p, t, k/ are aspirated in word-initial position, except when following syllable initial /s/. /p, t, k/ can also be aspirated in word-final and utterance-final positions (Knowles 1973, Watson 2007). The voiceless stops are frequently realised with noticeable preaspiration utterance-finally, which might manifest itself either as a period of glottal noise or as oral frication which is produced homorganically to the stop. As is the case in Newcastle (see e.g. Docherty & Foulkes 1999), preaspiration in Liverpool English is primarily the domain of female speakers (Watson 2007).

As well as these aspirated and preaspirated variants, there is an additional range of plosive realisations which are more or less unique to Liverpool. Most of these realisations can be described as processes of lenition – a term frequently used to group together a series of phonological weakenings which turn underlying plosives into affricates and fricatives (see e.g. Lass 1984; Harris 1990, 1994; Honeybone 2002). Indeed, plosive lenition is arguably one of the most characteristic features of Liverpool English, and one which forms a major part of the variety's stereotype.

For /t/, affrication is common word-initially, whilst spirantisation is common in intervocalic and word-final positions. The speaker uses an affricated /t/ in word-initial position in two (see figure 1). She spirantises /t/ as well, but there are a limited number of potential candidates in the reading passage due to phonological environment restrictions. The presence of sibilant fricatives for /t/ does not lead to the loss of phonological contrast with /s/. Honeybone (2001) transcribes lenited /t/ as [] which, following Pandeli et al. (1997), implies a fricative with a flat cross-sectional tongue shape (signaled by [ɵ]) at a precisely alveolar place of articulation (signaled by the double-underscore diacritic which is adapted from the ‘extended IPA’ used in the transcription of disordered speech). There are durational differences, too, with a longer phonetic fricative for /s/ than for /t/ (Sangster 2001). Recent work has suggested that there is a wide range of possibilities of ‘stopless /t/’, which can all be described as fricatives, but which are all articulated with varying degrees of approximation (Watson 2007). Rather than these realisations being the result of articulatory undershoot, there is evidence that fine-grained phonetic differences provide indexical information in LE, as they have been found to do elsewhere (see e.g. Docherty & Foulkes 1999 for Newcastle, and Jones & Llamas 2003 for Middlesbrough). Thus, these realisations represent the learned articulatory behaviour of the speakers (Docherty & Foulkes 2000, Foulkes & Docherty 2006). For example, LE /t/ for male speakers is often realised as the [] described above, but for female /t/ the relationship between the oral and glottal gesture is more variable, so that realisations such as [^hs^h] are common. An example of a preaspirated, postaspirated fricative /t/ is provided in figure 2. Space restrictions inhibit detailed discussion of these realisations in this illustration, but see Watson (forthcoming) for a more detailed examination.

Figure 1 An affricated /t/ (circled) in word-initial position in the speaker's production of the word ‘two’.

Figure 2 A preaspirated, postaspirated sibilant fricative (circled) for /t/ in pre-pausal out. The speaker was a 15-year-old working-class female (Watson 2007).

As well as realising /t/ as an oral fricative, it can also be debuccalised to [^h]. In older speakers, this occurs only in pre-pausal position in a small set of high frequency monosyllabic (pseudo)function words with short vowels (e.g. it [ɪ^h], what [wɒ^h], not [nɒ^h], that [d̪´^h], lot [lɒ^h]). For younger speakers, the realisation of /t/ as [^h] can also occur in polysyllabic words which end in an unstressed syllable (e.g. market [mːxɪ^h], maggot [mɡɪ^h], aggregate [´ɡɾɪɡɪ^h]). As the debuccalisation of /t/ does not occur in any other variety of north-western English, it has been suggested that the extension of the process to polysyllabic words is an innovation which represents phonological divergence from supralocal norms (Watson 2006, 2007). Because pre-pausality is a conditioning environment for the realisation of /t/ as [^h], there are no potential candidates in the reading passage. Connected to the realisation of /t/ as [^h], in terms of the environments in which it occurs, is the realisation of /t/ as a rhotic (typically [ɹ] but also [ɾ]). /t/ → [ɹ] and /t/ → [^h] can occur in a similar sets of monosyllabic words, but whereas pre-pausality is necessary for the realisation as [^h], the conditioning factor for [ɹ] is the presence of a following vowel (e.g. get off [ɡɛɹɒf], that apple [d̪´ɹ´pəl], lot of [lɒɹəv]). The realisation of /t/ as [ɹ] is not unique to Liverpool English, but can be found in a range of northern English varieties (see e.g. Wells 1982). Like for /t/ → [^h], the tightly constrained phonological environment in which /t/ → [ɹ] occurs means it is not produced by the speaker in the passage. The final realisation of /t/ is one which is very frequent in just about all varieties of British English; the realisation of /t/ as [ʔ]. Early accounts of LE noted that the glottal stop was rare (Knowles 1973), and more recent work has corroborated this (Watson 2005, 2006, 2007). Knowles (1973) does comment, however, that [ʔ] is possible preceding /l/ or other syllabic consonants. The speaker can be heard using [ʔ] in disputing [dɪspʲuːʔnn] and immediately [ɪmiːdiəʔliː].

It is also common for speakers to realise /k/ as an affricate or a fricative, too, as the speaker's various tokens of cloak testify. The exact place of articulation of the fricative is mostly conditioned by assimilation to the preceding vowel. That is, palatal fricatives can be found following the close front monophthong [iː] and closing diphthongs [ēi,´i] (e.g. week [wiːç], like [ĺiç]), and more dorsal fricatives are attested following low and back vowels (e.g. back [bx], dock [dɒχ]). These dorsal fricatives can be velar or uvular. Fricativisation of /p/ also occurs, typically to [ɸ], but this is much less frequent than that of /t/ or /k/.

Phonetic fricatives are also found for the voiced plosives /b, d, ɡ/, although of these the lenition of /d/ is by far the most common. Because final-devoicing is common in LE, as it is in other varieties of English, the fricative realisations of /b, d, ɡ/ are rarely voiced in final position. As with the lenited variant of /t/ described above, that of /d/ does not result in neutralisation with /z/. Instead, the fricative is articulated with a flat cross-sectional tongue shape which is not unlike the fricative variant of /t/ described above. The speaker does not affricate or fricativise /b/ or /ɡ/, but /d/ is realised as a fricative in could (see figure 3). In the transcription that follows, taking Pandeli et al.’s (1997) lead, I transcribe the fricative variant of /d/ as [], with the caveat that it is frequently devoiced.

Figure 3 Realisation of /d/ as [] (circled) in the speaker's production of the pre-pausal word could.

Vowels

• 

Figure 4 Vowel trapeziums of monophthongs and diphthongs in Liverpool English.

The similarities between the vowels of Liverpool English and those of other northern English varieties are numerous. The distinction between [ʌ] and [ʊ] that is found in accents of southern England (and elsewhere, e.g. in varieties of English outside the British Isles) is not found in LE. Thus, words such as foot, put, butcher, bus and putt all have [ʊ]. Similarly like other accents in the north, LE has the short [a] in words such as bath, dance, and grass, where southern English varieties have the longer [ɑː]. As might be expected, there is complex sociolinguistic patterning here. According to Knowles (1973, 1978), the lack of distinction between /ʌ, ʊ/ and the use of the short [a] in bath words is most robust in working-class speakers, with some middle-class speakers modifying towards higher prestige RP-like norms. Knowles (1978: 86) suggests that this differentiation often leads to hypercorrection, with speakers sometimes producing utterances such as good luck [ɡʌd lʊk] and black castle [blɑːk kæsl]. Some middle-class speakers may have a [ʌ]~[ʊ] type contrast by using a more centralised, schwa-like variant for one of the pair of vowels (e.g. good luck [ɡʊd lək]). Although the long [ɑː] is not used for bath words, it is used in words such as start and palm, as it is elsewhere England. In Liverpool, however, for some speakers the quality of the vowel is often more front, resembling [aː] (e.g. start [staː], shark [ʃ´ːx]). Recent work has suggested that this fronter variant is used mostly by male speakers, with female speakers preferring the back variant [ɑː] (Watson 2005). The speaker follows this pattern, using [ɑː] in hard.

For many speakers of LE, there is a lack of contrast between the vowel in words like square, hair, and that in words like nurse, her. Typically, both sets have a front variant such as [ēː] or [ɛː], or even [ɪː]. De Lyon (1981) establishes no less than 18 different realisations for the vowel in square words, ranging from [ɛə] to [ɜ̟] to [ɜː], and lists 15 different forms of the vowel for nurse words, including [ɜ], [əɛ] and [œ]. It is not clear from de Lyon (1981), however, whether these variants pattern in sociolinguistically structured ways, as their distribution is not successfully quantified. It could be, for example, that the variation is more a result of phonological environment than any sociolinguistic patterning, but more research is required to investigate this further. According to Knowles (1973), middle class speakers are more likely to have an RP-like distinction between nurse and square words, using [ɜː] and something like [ɛː] respectively. In the passage, the speaker uses a close front variant [ɪː] in first.

As with other accents in the north of England, words such as book, cook and look typically have a long vowel in Liverpool rather than the short [ʊ] found elsewhere. The long vowel is typically produced in an advanced position, most frequently as a central [ʉ] but also as a fully front [ý]. This results in minimal pairs such as look [lʉːk]/luck [lʊk], and book [bʉːk]/buck [bʊk]. The use of [ʉː] in look words occurs more often in the working class than the middle class for Knowles’ (1973) speakers, but recent work has suggested that this feature is recessive, occurring less frequently in younger people (Watson 2005). The vowel in e.g. goose and hoop, is also typically articulated in an advanced position. The speaker produces an advanced variant in blew and two.

Perhaps the biggest difference between Liverpool English vowels and those of many other northern English varieties is that Liverpool English has diphthongs in words like face [fēɪs] and goat [ɡɛʉt]. Diphthongs are also used in choice [tʃɔɪs], price [pɹ´ɪs] and mouth [mʊɵ], as they are elsewhere. For some speakers the vowel in price can monophthongise before certain consonants (e.g. time [tːm], five [fːv]) but this has not yet been systematically investigated. There is also variation in the vowel of goat, which most typically has a fronted onset and offset (e.g. ɛʉ) but can also be realised as [ēʊ, ɛʊ, əʉ, ēʉ] or the more standard [əʊ].

Stress and intonation

Previous work on LE intonation is minimal. Only Knowles (1973) has provided a systematic auditory investigation of prosodic issues, although more recent acoustic work is underway (Grabe 2004, Grabe, Kochanski & Coleman 2005). Knowles (1973) argues that LE pitch range is narrower than other varieties of English, and that this makes the distinction of tones difficult to detect. However, LE is known to have intonational similarities to other northern English accents, and also to share features with varieties of Irish English. For example, Knowles (1973: 188) refers to a tone which he calls the step, in which a high level tone follows the initial rise. This is comparable to the rise-plateau which is one of the most common tones in Belfast English (Grabe et al. 2005). More systematic investigation is required if we are to understand the relationship between the prosodic system of LE and that of other English varieties.

Transcription of recorded passage

Two transcriptions are provided. The first uses the phonemic symbols outlined above, whilst the second is a narrower phonetic transcription that focuses specifically on the pronunciation of the speaker and demonstrates some of the variation that has been encountered throughout this illustration.

Phonemic transcription

ðə nɔːɵ wɪnd ən ðə sʊn wə dɪspʲʉːtən wɪtʃ wəz ðə strɒŋɡə wɛn ə tŕvələ kēɪm əlɒŋɡ ŕpt ɪn ə wɔːm klɛʉk ðēɪ əɡriːd ð´t ðə wɒn ʉː fēːst sʊksiːdəd ɪn mēɪkən ðə tŕvələ tēɪk ɪz klɛʉk ɒf ʃʊd biː kənsɪdəd strɒŋɡə ð´n ðiː ʊðə ən ðɛn ðə nɔːɵ wɪnd blʉː əz ^hɑːd əz iː kʊd bʊt ðə mɔː ^hiː blʉː ðə mɔː klɛʉsli dɪd ðə tŕvələ fɛʉld ɪz klɛʉk əŕʊnd ɪm əndt ĺst ðə nɔːɵ wɪnd ɡēɪv ʊp ðiː ətɛmpt ðɛn ðə sʊn ʃ´ɪndʊt wɔːmli ən ɪmiːdɪətli ðə tŕvələ tʊk ɪz klɛʉk ɒf ən sɛʉ ðə nɔːɵ wɪnd wəz əbĺɪdʒd tə kənfɛs ð´t ðə sʊn wəz ðə strɒŋɡə əv ðə tʉː

Phonetic transcription

ðə nɔːɵ wɪnd nn ə̝ sʊm wə dɪspʲʉːʔm wɪtʃ wəz̥ ð st↑ɒ͂ŋɡəwɛn ə tʃ↑´vlə k^hēɪm əlɒ͂ŋɡ ↑´p〉t ɪn ə wɔːm k^hlɛʉxd̪ēɪ ɡriːd̪ ð´t〉 ðə wɒn ʉː fɪːs ʊksiːd^z ɪm mēɪçən̪ ə tʃ↑´vlə t^sēɪç ɪz klɛʉx ɒf ʃʊb〉 bi k^hnsɪdəd stʃ↑ɒ͂ŋə ðnðiː ʊðə ən d̪ɛn̪ ə nɔːɵ wɪm bluː əz ^hɑːð[] əz iː k^hʊð bʊt ̪ d̪ə mɔː ^hiː bluː ðə mɔː k^hl ̥ əʉsliː d^zɪd̪ ðə tʃ↑´vlə fəulð ː k^hlɛʉx əŕʊnd^z m əńʔ ĺs ðə nɔːɵ wɪŋ ɡɪv ʊp ðiː ət^sɛmp〉t^s d̪ɛn̪ ə sʊn ʃ´ːɪndʊp〉 wɔːmliː´n əmiːð əʔḭː ðə tʃ↑ ̥´vlə t^sʊx əz k^hlɛʉx ɒf n sɛʉ ðə nɔːɵ wɪm wəz əbĺɪʒ ə k^hnfɛs ðəʔ ðə sʊm əzː sʃtrɒŋɡəɾ əv̥ ðə t^sʉː