Introduction and background

English is used as a lingua franca in most parts of the world (Ozaki, Reference Ozaki2011). However, problems and issues related to learning English are country specific (Nagamine, Reference Nagamine2011), because most of the difficulties in foreign language learning arise from L1 interference (Flege, Reference Flege and Strange1995). Since this study focuses on acoustic analysis of a phonological feature of Pakistan English (PakE), we outline the historical background of the issue very briefly. Pakistan is a linguistically rich country. More than 70 languages are spoken in Pakistan (Rahman, Reference Rahman1996). Saraiki, Balochi, Sindhi, Punjabi and Pashto are the major indigenous languages of the country. More than 90% of the total population speaks these languages. Pakistan came into being in 1947. It inherited English as a language of education, law, the judiciary and media from the British colonial masters. The British rulers also used the English language in India for official correspondence. Therefore, English became a very effective tool and symbol of power in the subcontinent. As a result, people of the subcontinent feel pride in learning English. Although the colonial period has ended and the English rulers have departed to their homeland, English still remains the language of ruling elite in Pakistan and India.

However, after gaining independence from the British colonial rulers, the people of Pakistan did not have any live native speaker model to follow. Consequently, a new variety of English developed in Pakistan (Ali, Reference Ali and Baumgardener1993) which has different features from those of British or American English. Since the current study focuses on PakE, we focus on the relevant features of PakE only.

PakE has some grammatical features which are different from native English (Baumgardener, Reference Baumgardener1993; Rahman, Reference Rahman1991, Reference Rahman, Bolton, Botha and Kirkpatrick2020). Nowadays, these features of PakE are not considered grammatical errors or deviations from standard norms. They are accepted features of a variety of English. Similarly to African English (Jenkins, Reference Jenkins2009) or Indian English (Costa, Reference Costa2019), PakE has established itself as a separate independent variety of English (Gargesh, Reference Gargesh, Nelson, Proshina and Davis2019). One of the prominent features of PakE is that Pakistani speakers produce English alveolar /t/ as a retroflex [ʈ] (Mahboob & Ahmar, Reference Mahboob, Ahmar and Schneider2004; Rahman, Reference Rahman1991). The current study provides a discussion of implications of this feature of PakE with reference to voice onset time (VOT).

VOT was first discussed in detail by Lisker and Abramson (Reference Lisker and Abramson1964). VOT is the time interval between the burst of a stop and onset of voicing for the following vowel (Docherty, Reference Docherty1992). It is an acoustic correlate of aspiration and voicing (Foulkes, Docherty & Jones, Reference Foulkes, Docherty, Jones, Foulkes, Docherty and Jones2010). As already known in the literature, the place of articulation of plosives has a strong effect on VOT ranges of stops (Cho & Ladefoged, Reference Cho and Ladefoged1999; Lisker & Abramson, Reference Lisker and Abramson1967; Nasukawa, Reference Nasukawa2010). In this study, we reproduce the results of previous experiments (Syed, Reference Syed2013a, Reference Syed2013b, Reference Syed2015; Syed & Saleem, Reference Syed and Saleem2019) and compare those with the findings of two current experiments to determine the VOT ranges of PakE plosives in two different contexts. We also anticipate the implications of the findings for future changes expected to occur in PakE.

Previous studies

The VOT of PakE plosives produced by speakers of various social and linguistic backgrounds has been studied by the lead author and colleagues over the previous decade. In an experiment, Syed (Reference Syed2013a) recorded a list of English words produced by 34 males studying MA English in Pakistan. The word list also included teeth and steal. The L1 of all participants was central Saraiki of southern Punjab, which is considered the standard dialect of Saraiki (Bashir & Conners, Reference Bashir and Conners2019). In this experiment, each of the stimuli was produced six times, i.e. three times as isolated words and three times in a carrier sentence. The carrier sentence was ‘I say . . . again’. The difference in VOT of the alveolar stop produced in the carrier sentences and that in isolated words was not statistically significant. Therefore, a cumulative mean VOT was taken in all six repetitions. The mean VOT of /t/ in the word teeth was 14.73 ms and that in the word steal was 20.50 ms. Therefore, there was a statistically significant difference of almost 6 ms in the VOT in English coronal stops in two different contexts.

Syed (Reference Syed2015) repeated the same experiment with a group of ten male teachers of English who had been teaching in the Southern Punjab where Standard Saraiki is spoken as a common language of the people (Kula & Syed, Reference Kula and Syed2020). The same research methods as adopted in the previous experiment, i.e. three repetitions of the target sound in isolated words and three in carrier sentences, were adopted in this study with the same stimuli (teeth, steal included in the list of stimuli). Again, no significant difference was found in the mean VOTs of stops obtained in single words and those produced in the carrier sentences. Therefore, the means of six repetitions by each participant were averaged. The mean VOT of the coronal stop in the word teeth was 19.34 ms and that in the word steal was 26.63 ms. Again, there was a difference of 7 ms in the VOT in the two variants of English coronal stop produced by English language teachers in Pakistan. The difference was also statistically significant.

In a third experiment of this series, Syed and Saleem (Reference Syed and Saleem2019) recorded a list of words of English in the speech of two groups of PakE speakers. The participants of one group comprised 29 male Pakistanis who had learnt PakE in Pakistan and afterwards had moved to England for higher education. The second group comprised 30 male educated PakE speakers in Pakistan. The L1 of both groups was central Saraiki. The list of stimuli in this study also included the words teach and steal. The participants produced each word three times from a list in a word-reading task. Acoustic analyses revealed that the UK-based PakE speakers produced the English coronal stop in the words teach and steal with a mean VOT of 25.60 and 30.07 ms, respectively. The participants of the second group, i.e. Pakistan-based PakE speakers, produced a coronal stop in the word teach with an average VOT of 19.31 ms but that in the word steal was produced with a mean VOT of 25.02 ms. A significant difference of approximately 6 ms VOT was found in the production of the context-specific variants of the English coronal stop in their speech.

Further acoustic analyses confirm that place and manner of articulation of /t/ in PakE is the main reason behind the fact that PakE speakers produce /t/ on word-initial position e.g. in words like ‘teach’ with shorter VOT than in clusters in words like steal. Previous research has already very well established that in PakE, the English coronal stop /t/ is produced as retroflex [ʈ] (Mahboob & Ahmar, Reference Mahboob, Ahmar and Schneider2004; Rahman, Reference Rahman1991) as is the case in Indian English (Gargesh, Reference Gargesh and Kortmann2004; Sailaja, Reference Sailaja2009; Wiltshire, Reference Wiltshire2005). But /t/ is produced as an alveolar stop in British or American English. In terms of feature geometry (Clements & Hume, Reference Clements, Hume and Goldsmith1995), retroflex sounds being post-alveolar are posterior ([-anterior]) and alveolar stops are anterior by place of articulation. English /s/ is also anterior by articulation. Therefore, it is difficult for Pakistanis to produce a retroflex [ʈ] immediately after another anterior /s/ because it is humanly unmanageable to articulate two opposite gestures simultaneously or consecutively without a vowel between them, particularly when the speaker is producing a non-native cluster which is illegitimate in his/her L1. As a result, many PakE speakers produce English coronal /t/ as alveolar after /s/ in words such as steal, but as retroflex [ʈ] in other contexts. In order to confirm this view, Syed and Saleem (Reference Syed and Saleem2019) obtained F3 values of the vowels in the words steal and teach in productions of 59 participants. The results revealed that the F3 of these Pakistanis was significantly lowered in vowels in teach than in steal. The F3 value of a vowel adjacent to a retroflex consonant is normally lowered (Steriade, Reference Steriade, Hume and Johnson2001). Thus, Syed and Saleem (Reference Syed and Saleem2019) claim that PakE speakers whose L1 is Saraiki produce English coronal stops as retroflex or posterior in all other contexts but in st-clusters in word-initial position, they produce it without a retroflex gesture, i.e. as an alveolar stop as used in British or American English. This results in a shorter VOT duration for coronal stops in word-initial position but a slightly longer VOT duration for the same in st-clusters.

The experiment by Syed (Reference Syed2013a) was repeated in another study (Syed, Reference Syed2013b) with 12 native speakers of Pashto who were studying at the University of Essex in England. They had learnt PakE in Pakistan and had come to England in pursuit of postgraduate degrees. They were asked to produce a list of English words which also comprised teeth, steal, tall, stall, tool, stool, among other words. Each of the stimuli was produced six times by each participant, i.e. three times in isolated words and three times in the carrier sentences. The list of stimuli in this study also contained words with a coronal stop immediately followed by three quantum vowels. One of the objectives of this study was to determine the impact of an immediate vowel on the VOT of the preceding stop. Therefore, a non-parametric test was applied on the VOTs obtained from coronal stops in various vocalic contexts. The impact of the adjacent vowels was found to be non-significant on the VOT of coronal stops. Therefore, the mean values of VOTs obtained in different words in two contexts were averaged. The results show that Pashto speakers produced English /t/ with a mean VOT of 35.64 ms and in words starting with st-clusters the mean VOT was 33.17 ms. The mean difference was found to be non-significant (p=.535). The results of this study show that unlike L1-Saraiki speakers, L1-Pashto speakers of PakE do not produce coronal stops in two different contexts with significantly different VOT ranges. The findings of the study (Syed, Reference Syed2013b) are different from those with L1-Saraiki speakers of PakE.

The above-quoted results show that Saraiki speakers of PakE produce English coronal stops in word-initial st-clusters with a VOT which is approximately 6-7 ms longer than that in other contexts. However, the same trend is not seen in the PakE speech of those Pakistanis who speak Pashto as their mother tongue (Syed, Reference Syed2013b). The current study aims to acoustically analyze English coronal stops in the speech of those Pakistanis who speak Balochi and Sindhi as mother tongues to further confirm this trend in PakE.

Research methodology

The purpose of this study is to determine the VOT ranges of English coronal plosives in two contexts, i.e. word-initial position in monosyllabic words and in words starting with st-clusters produced by PakE speakers who speak Sindhi and Balochi as their mother tongues. Two experiments are conducted for this purpose, the details of which are given below.

Analysis and discussion

The study of VOTs of coronal stops in the speech of Balochi and Sindhi speakers of PakE reflects a consistent trend, which is also already observed (as summarized above) in the speech of PakE speakers whose L1 is Saraiki. It means that L1-Balochi, Saraiki and Sindhi speakers of PakE produce coronal plosives in st-clusters with approximately 2–7 ms longer VOTs than in that in word-initial position. On the other hand, in the speech of L1-Pashto speakers of PakE, this trend is not confirmed. On the basis of all these studies, we can generalize that in the speech of those PakE speakers who speak Saraiki, Sindhi and Balochi, the word-initial coronal stop is produced with slightly longer VOT than in st-clusters. The reason for this lies in the articulation of this sound in PakE. Since PakE speakers substitute British English alveolar stop /t/ with retroflex [ʈ] in all other contexts, they have developed a category of this sound in the phonemic inventory of PakE which is basically retroflex. However, it becomes articulatorily difficult for them to produce a retroflex soon after /s/ because a retroflex sound is geometrically [-anterior] while /s/ is [+anterior]. The speakers cannot produce a sequence of two consecutive sounds with two opposite gestures, particularly in a cluster which is not permitted in the phonology of their mother tongues (e.g. Saraiki, Sindhi, Balochi). As a result, they produce the coronal stop in word-initial st-clusters with an apical gesture which increases the VOT of the stop whereas the retroflex gesture reduces the VOT of consonants (Steriade, Reference Steriade, Hume and Johnson2001).

A possible question regarding this conclusion is why this pattern is not found in the speech of those PakE speakers whose L1 is Pashto. An apparent reason is that Pashto is very rich in consonant clusters (Khan & Bukhari, Reference Khan and Bukhari2014). It very frequently uses st-clusters. Therefore, the articulatory difficulty which other Pakistanis experience in the production of st-clusters in English is resolved by Pashto speakers because of positive transfer from the L1. Scores of words of Pashto start with st-clusters (David, Reference David2014). Since Pashto speakers do not experience any difficulty in the production of a posterior sound (retroflex [ʈ]) immediately after an anterior consonant (alveolar /s/) because of facilitative interference of the L1, they do not need to change the place of articulation of PakE from post-alveolar retroflex to alveolar. To our knowledge, no other language of Pakistan allows st-clusters in word-initial position. Therefore, Pakistanis other than those who speak Pashto face difficulty in accurate pronunciation of the British English st-clusters and have resolved this difficulty by producing /t/ as an alveolar when it occurs after /s/ word-initially but the same is produced as a retroflex in all other contexts because the actual phonetic category developed for British English /t/ in PakE is a retroflex [ʈ] as demonstrated in the previous studies (Mahboob & Ahmar, Reference Mahboob, Ahmar and Schneider2004; Rahman, Reference Rahman1991; Syed, Ansari & Gopang, Reference Syed, Ansari and Gopang2017).

Finally, different results obtained in the speech of PakE speakers of different L1s indicate that PakE develops under the influence of speakers’ mother tongues. Keeping in view a large number of languages spoken in Pakistan, we anticipate that Pakistanis may also have different types of Englishes in their country. Researchers have already pointed out similar trends in India (Wiltshire, Reference Wiltshire2005; Wiltshire & Harnsberger, Reference Wiltshire and Harnsberger2006). As is apparent from the data discussed in this study, L1-Saraiki speakers have established two different context-specific allophones of English coronal stops. L1-Balochi and Sindhi speakers also seem to follow the same track. But L1-Pashto speakers have different pronunciation for the PakE target sound because of the grammatical structure of their L1 (Pashto). This leads us to conclude that, first, a new feature is emerging in PakE which has not yet been properly recognized in academic circles; second, since the indigenous languages of Pakistan are structurally different from each other, PakE versions spoken by speakers of different L1s are also different. In this scenario, different Englishes are expected to emerge in Pakistan under the cover term of Pakistan English.