Early research

This chapter provides an extensive review of early and recent research on rhythm, rhythm indices, and the measurement of rhythm in relation to different varieties of world Englishes. Implications of recent research on rhythm for pronunciation teaching will be considered.

A summary of early research studies on speech rhythm has been provided in Low (2006) and in Low (2014). This section takes reference from both these works. Early research on speech rhythm tended to focus on exactly which speech unit regularly recurs such that isochrony (or equality in timing) occurs. Based on the concept of whether stresses or syllables recurred at regular intervals, Pike (1945) and Abercrombie (1967: 69) classified languages into either being stress-timed or syllable-timed. For stress-timed languages, it is the feet (comprising one stressed syllable up to but not including the next stressed syllable) that contribute to the overall perception of isochrony in timing. In the case of syllable-timed languages, it is the syllables that are believed to contribute to the perception of isochrony. However, the concept of pure or perfect isochrony became a moot point in the 1980s, with scholars proposing that isochrony should be described as a tendency rather than as an absolute. Dauer (1983) and Miller (1984), for example, suggest a continuum of rhythmic typology where languages can fall in between being stress-based at one end and syllable-based on the other (Grabe and Low 2002; Low 2006, 2014).

The earliest works classifying the rhythmic typology of the world’s languages tended to forward the strict dichotomous view where languages were considered as either being stress- or syllable-timed. Abercrombie (1965: 67) believed that it is the way that chest or stress pulses recur that helps determine the rhythmic typology of a language, and for stress-timed languages it was the stress pulses that were isochronous while for syllable-timed languages chest pulses were isochronous. A third categorization of rhythm, known as mora-timing, was proposed by another group of scholars (e.g., Bloch 1942; Han 1962; Hoequist 1983a, 1983b; Ladefoged 1975). Japanese is the only language that scholars classified as being mora-timed. Because mora-timing does not apply to English, its specific details will not be discussed further except where it is relevant in other studies detailing rhythm across languages.

While early scholars proposed stress-, syllable-, and mora-timing as a means for classifying rhythmic typology of the world’s languages, research also highlighted clear difficulties in adopting these categorical distinctions. For example, when interfoot (interstress) intervals were measured for stress-timed languages, researchers could not find evidence that their timing was roughly equal, that is, isochronous (Shen and Peterson 1962; Bolinger 1965; Faure, Hirst, and Chafcouloff 1980; Nakatani, O’Connor, and Aston 1981; Strangert 1985; Lehiste 1990). Yet others tried to find evidence that syllables were more nearly equal in timing for syllable-timed languages but failed (Delattre 1966; Pointon 1980; Borzone de Manrique and Signorini 1983).

Roach (1982) and Dauer (1983) measured the interstress intervals of different languages classified as stress- and syllable-timed. Roach’s (1982) research set out to test the claims made by Abercrombie (1967) that syllables do not vary in length for syllable-timed languages and that interstress intervals ought not be equal in timing compared to stress-timed languages. Not only did Roach not find evidence to support these two claims but he found evidence that contradicted earlier claims because there was greater variability for syllable durations for syllable-timed languages compared to stress-timed ones and interstress intervals varied more in stress-timed languages compared to their syllable-timed counterparts. Roach’s findings led him to suggest that evidence for the rhythmic categorization of languages cannot be sought by measuring timing units like syllables or interstress intervals in speech. Dauer (1983) conducted a cross-linguistic study of English, Thai, Italian, Greek, and Spanish. She found that interstress intervals were not more equal in languages classified as stress-timed, like English compared to Spanish, which has been classified to be syllable-timed. She therefore reached the same conclusion as Roach where she concluded that empirical support for rhythmic categorization cannot be found by measuring timing units found in speech. This led other scholars like Couper-Kuhlen (1990, 1993) to forward the view that isochrony is better understood as a perceptual rather than an acoustically measurable phenomenon. The experimental findings for mora-timed languages yielded mixed results. Port, Dalby, and O’Dell (1987) found some evidence that mora was nearly equal in timing in Japanese but others could not (Oyakawa 1971; Beckman 1982; Hoequist 1983a, 1983b).

Due to the experimental findings by early researchers where empirical evidence for rhythmic categorization was not related to timing units in speech, isochrony was then considered to be a tendency. This led to the terms stress-based, syllable-based, and mora-based languages in place of the earlier categorization of stress-, syllable-, and mora-timed (Dauer 1983, 1987; Laver 1994: 528–529). Grabe and Low (2002: 518) forwarded the proposal that “true isochrony is assumed to be an underlying constraint” while it is the phonetic, phonological, syntactic, and lexical characteristics of a language that are likely to affect the isochrony of speech units found in any language. These characteristics form the basis for later research attempting to hunt for acoustic validation for the rhythmic classification of the world’s languages as being stress-, syllable-, or mora-based.

Recent research

Early experimental studies on rhythm were unable to find support for isochrony by measuring timing intervals in speech. This led to the hypothesis proposed by researchers like Dauer (1983, 1987) and Dasher and Bolinger (1982) that rhythmic patterning is reliant on other linguistic properties of language such as their lexical, syntactic, phonological, and phonetic attributes. Dauer singled out three main influences on speech rhythm: syllable structure, the presence or absence of reduced vowels, and the stress patterning of different languages. She suggested that stress-based languages tend to have a more complex syllable structure make-up, and syllable-based languages also tend not to make a strong distinction between full and reduced vowels. Dasher and Bolinger (1982) also observed that syllable-based languages tended not to have phonemic vowel length distinctions, i.e., that long versus short vowels were not used as distinct phonemes, leading to long/short vowel conflations.

Nespor (1990) introduced the concept of “rhythmically mixed” or intermediate languages. For her, the strict categorical distinction was no longer tenable and languages were mainly mixed or intermediate in terms of rhythmic typology, and so-called intermediate languages exhibited shared properties characteristic of both stress- and syllable-based languages. One example of an intermediate language is Polish, which tends to be classified as being stress-based but which does not have reduced vowels, a feature that helps stress-based languages to achieve foot isochrony through compensatory shortening of syllables. Catalan is another such language, which has been classified as syllable-based but which has vowel reduction, a property that is not usually found in syllable-based languages.

Rhythm indices and the measurement of rhythm of world Englishes

The hunt for empirical acoustic validation for rhythmic classification of the world’s languages led researchers to measure the durations of some phonological properties such as vowels, syllables, or consonants. In tandem with this focus on measuring durational units in speech, several rhythm indices have been developed to capture the rhythmic patterning of different languages, as indicated by the durational properties of the different timing units in speech. A nonexhaustive summary of the main rhythmic indices developed from the late 1990s to the present will be presented here. Tan and Low (2014) also present a version of this summary of latest developments on speech rhythm using rhythm indices.

The key breakthrough in the development of rhythmic indices to measure timing intervals can be traced back to the pairwise variability index found in Low (1994, 1998) but published in Low, Grabe, and Nolan (2000). At about the same time, the rhythmic indices developed by Ramus, Nespor, and Mehler (1999), Deterding (2001), Grabe and Low (2002), and Dellwo and Wagner (2003) were also developed. The main contribution of the rhythm indices to the study of speech rhythm was to show that it was possible to find empirical evidence to classify rhythm by measuring timing intervals in speech and subjecting them to calculations made possible by the rhythmic indices. Low (2006) details the development of the earlier rhythmic indices. Ramus et al.’s (1999) index and Low et al.’s (2000) index were applied to successive consonantal and vowel intervals respectively. The earlier indices are premised on the fact that stress-based languages tend to have a greater difference durationally between stressed and unstressed syllables and have a more complex syllable structure with more consonantal clusters in the onset and coda positions. This in turn influences the overall consonantal durations, making them longer. Nolan and Asu (2009) note that one advantage of Ramus et al.’s (1999) interval measures (IM) and Low et al.’s (2000) Pairwise Variability Index (PVI) is that a researcher is able to measure the timing intervals of languages not known to them because there is no need to consider the phonological make-up of syllables. Instead, as long as one is able to segment the speech signal into vowels and consonantal intervals, it is possible to apply both these indices to their measurements with little difficulty.

To elaborate on these two indices, Ramus et al.’s (1999) IM concentrated mainly on three timing intervals that are said to vary durationally across different languages. %V measures the proportion of vocalic intervals in speech (the segment between the vowel onsets and offsets); ∆V measures the standard deviation of the vocalic intervals, while ∆C measures the standard deviation of consonantal intervals (the segment of speech between vowel offsets and onsets excluding any pauses). These three IM were applied to languages classified as stress-based (Polish, Dutch, and English), mora-based (Japanese) and syllable-based (Catalan, Spanish, Italian, and French). Their results showed that the most reliable way to classify rhythmic patterning is to use ∆V and a combination of either ∆C or %V. The problem with using either ∆C or ∆V is that standard deviations are unable to capture the successive durational patterning of successive timing intervals, be they vowels or consonants, as pointed out by Low, Grabe, and Nolan (2000).

The rhythm indices developed by Low, Grabe, and Nolan (2000) are known as the PVI. It measures the durational variation that exists between successive vowels found in an utterance. The PVI is premised on the hypothesis that the main difference between stress-based and syllable-based languages is the lack of contrast between full and reduced vowels in syllable-based languages. This hypothesis is further premised on the assumption that stress-based languages need to have compensatory shortening for feet that contain a lot of syllables so that they can approach foot iscochrony, a central property of stress-based languages. Compensatory shortening is achieved via reduced vowels in unstressed syllables. Low (1998) and Low, Grabe, and Nolan (2000) considered the claim by Taylor (1981) that it is the vowels, not the syllables, that determined the syllable-based nature of Singapore English. They compared the successive vowel durations found in British English (a stress-based language) with Singapore English (a syllable-based language). The PVI measures the mean absolute difference between successive vowels in an utterance. Absolute differences in durations between pairs of successive vowels are calculated and their means are taken (only positive values are considered by disregarding the negative sign when negative values occur). The mean difference is then calculated by dividing the difference between the successive vowel durations by the durational average of the two pairwise vowels so as to normalize for different speaking rates. To produce whole numbers, the values are multiplied by 100 and expressed as an index while the formula may be represented as

where m = number of vowel intervals in an utterance and d = duration of the kth vowel.

As the PVI measures the variation between successive vowels in an utterance, it is possible to surmise that an idealized stress-based language ought to have a high PVI while an idealized syllable-based language will have a low PVI. The highest possible PVI showing maximal variation between successive timing units is 100 while the lowest possible PVI showing no variation between successive timing units is 0. Low, Grabe, and Nolan (2000) discovered a signficant difference in PVI values between British English and Singapore English and concluded that the greater variation in the successive vowel durational units contributed to the perception of British English as stress-based and, consequently, the lack of variation in successive vowel durational units contributed to the perception of Singapore English as syllable-based. Applying Ramus, Nespor, and Mehler’s (1999) IM %V to the data, they did not find that this was useful in reflecting the rhythmic patterning of both language varieties. However, if we consider %V to be the proxy for syllable structure make-up, we can then conclude that the difference between stress- and syllable-based languages cannot be captured adequately by considering differences in syllable structure make-up. The main breakthrough in Low, Grabe, and Nolan’s (2000) work is that a measure for empirically capturing the difference between stress- and syllable-based languages could be found by measuring timing intervals in the speech signal, namely successive vowel durations.

Grabe and Low (2002) extended the investigation to 18 different languages and used both the normalized vocalic PVI values (nPVI) and the raw PVI scores for consonants (rPVI) for the investigation of prototypically stress-based languages (Dutch, German, and English), prototypically syllable-based languages (Spanish and French), and a prototypically mora-timed language (Japanese). The nPVI and rPVI were also applied to Polish and Catalan (classified by Nespor as being rhythmically mixed or intermediate) and three languages whose rhythmic patterning has never been classified (Greek, Estonian, and Romanian). Grabe and Low (2002) found further evidence to show that prototypically stress-based languages like German, Dutch, and English had higher normalized vocalic variability (nPVI) while prototypically syllable-based languages like Spanish and French tended to have a lower normalized vocalic variability (nPVI). The alternation of full and reduced vowels in stress-based languages was more prominent than in syllable-based languages. Japanese had an nPVI reading that was closer to stress-based languages but high consonantal rPVI, which resembled that found for syllable-based languages. Catalan showed traits of being rhythmically mixed, as suggested by Nespor (1990), because it has a high normalized vocalic PVI characteristic of stress-based languages but it also had a high consonantal raw PVI normally associated with syllable-based languages. The raw PVI for consonantal intervals also showed the ability to tease out further differences between different languages like Polish and Estonian, which had similar vocalic nPVI values but different rPVI consonantal values.

More recent work (Dellwo and Wagner 2003) on developing rhythmic indices has emphasized the importance of normalizing rhythm indices against speaking rates across the entire utterance. Dellow’s (2006) index, known as VarcoC, measures the standard deviation of consonantal intervals and divides the value by the mean consonantal duration in order to normalize for speech rates. VarcoC (the normalized version of ∆C) was found to be more robust than ∆C in capturing the difference between stress-based and syllable-based languages. However, Dellwo and Wagner (2003) found that normalizing for speech rate does not affect successive vocalic durations significantly and that it is therefore more important to control for speech rates when measuring consonantal intervals. White and Mattys (2007a, 2007b) devised a VarcoV in spite of Dellwo and Wagner’s suggestion that vocalic intervals need not be normalized for speech rate and found that VarcoV was able to show the influence of one’s L1 rhythm when VarcoC cannot.

More recent work on rhythm indices combines more than one index to the data. Loukina et al. (2009) found that combining two rhythm indices was more effective at classifying rhythmic differences between languages but that combining three indices did not yield better results. Studies combining different indices abound in the literature (Gibbon and Gut 2001; Gut et al. 2001; Dellwo and Wagner 2003; Asu and Nolan 2005; Lin and Wang 2005; Benton et al. 2007). Only those studies that further our understanding of the rhythmic patterning in different varieties of English spoken around the world will be highlighted.

Ferragne and Pellegrino (2004) found that the nPVI of successive vocalic intervals was a good way to automatically detect the difference in the dialects of English spoken in the British Isles but that consonantal intervals or the rPVI of consonants was not effective in detecting dialectal differences. Other studies have examined the influence of a speaker’s L1 rhythm on their L2 rhythm by combining the rhythm indices. Lin and Wang (2005) used a combination of ∆C and %V and showed that L2 speakers of Canadian English were influenced rhythmically by their L1 Mandarin Chinese. Mok and Dellow (2008) applied the following indices on their data, ∆V, ∆C, ∆S, %V, VarcoV, VarcoC, VarcoS, rPVI-C, rPVI-S, nPVI-V, and nPVI-S, where S refers to syllable durations, and found that L2 speakers of English were influenced rhythmically by their L1 Cantonese and Beijing Mandarin. Carter (2005) found that the rhythm of American English L2 Spanish bilingual speakers who had moved from Mexico to North Carolina was influenced by the L1 rhythm of Mexican Spanish. The PVI values obtained were intermediate between what one would expect for a stress-based language like English and a syllable-based language like Spanish. Whitworth’s (2002) study on English and German bilinguals showed that bilingual children in these two stress-based languages produced the same PVI values for English and German as their parents’ respective first languages. White and Mattys (2007a) used the following rhythm indices, ∆V, ∆C, %V, VarcoV, VarcoC, nPVI, and cPVI, to compare the rhythmic patterning of L1 and L2 speakers of English, Dutch, Spanish, and French. They found VarcoV to be the best discriminator between L1 and L2 speech rhythms, as a significant difference in VarcoV was found between the two groups of speakers.

Even more recent research on speech rhythm has argued for measuring other timing units such as foot and syllable durations and to consider the measurement of intensity in addition to merely timing durations (Ferragne 2008; Nolan and Asu 2009). These studies have also argued for considering the notion of rhythmic coexistence where a language can be both stress-based and syllable-based simultaneously. The proposal for considering foot and syllable durations can be challenging as it is harder than segmenting vowels and consonants in the speech signal. Furthermore, in typical stress-based languages, foot segmentation is a real issue if stressed and unstressed syllables are not significantly contrasted.

The application of rhythm indices to measuring different varieties of world Englishes has continued in recent years. Low (2010) applied the nPVI to British English, Chinese English (by a speaker of Beijing Mandarin), and Singapore English. Findings showed that while Singapore English differed significantly from British English (corroborating earlier studies), Chinese English rhythm did not differ significantly from either Singapore or British English. These findings provide support for the Kachruvian notion that Inner Circle varieties like British English provide the norms that Expanding Circle varieties like Chinese English veer towards. However, what is interesting is that at least in the rhythmic domain, Chinese English also veered towards Outer Circle norms, like Singapore English, which are supposed to be norm-developing varieties.

Mok (2011) measured the consonantal, vocalic, and syllabic intervals of Cantonese–English bilingual children and their age-matched monolingual counterparts. Results showed that at least in the syllabic domain, bilingual English speakers exhibited less variability than monolingual English speakers and this could signal a delay in the acquisition of L2 rhythm. She suggests that the lack of a strong contrast between stressed and unstressed syllables and the absence of reduced vowels in Cantonese may account for the delay. There is also evidence of syllabic simplification of Cantonese spoken by the Cantonese–English bilinguals, showing that bilingual speakers also show delay in language acquisition in both L1 and L2.

Payne et al. (2011) compared the speech of English, Spanish, and Catalan children aged 2, 4, and 6 with that of their mothers and found that they had more vocalic intervals but less durational variability. By age 6, interestingly, the children acquired similar vocalic interval patterning as their mothers but significantly different consonantal components.

Nakamura (2011) discovered that the ratio of stressed to unstressed syllables was lower for non-native compared to native speakers of English, showing that less contrast between stressed and unstressed syllables can be found in non-native English speech. Nokes and Hay (2012) applied the PVI to measure variability in the durational, intensity, and pitch of successive vowels of New Zealand English (NZE) speakers born between 1951 and 1988. The cross-generational study showed that younger speakers of NZE tended to show less of a distinction between stressed and unstressed vowels.

In recent years, Multicultural London English (MLE), spoken by different migrants in the inner city of London, has received much attention. Togersen and Szagay (2012) compared the rhythmic patterning of MLE speech compared to outer city counterparts and found that the MLE speakers had significantly lower PVI values compared to their outer London peers. The lower PVI values and more syllable-based rhythmic patterning is consistent with L2 varieties of English spoken around the world. Diez et al. (2008) found that the higher the proficiency of the L2 speaker, the more native-like their rhythmic patterning is likely to be.

Implications for pronunciation teaching

This section will discuss the relevant studies on speech rhythm that help inform pronunciation teaching and learning. What is clear from the detailed literature review of the research is that L2 rhythm is clearly influenced by L1 rhythm. Earlier research by Grabe, Post, and Watson (1999) suggested that the rhythm of a syllable-timed language like French is easier to acquire than that of a stress-timed language like English. Their evidence was found through comparing the PVI values of 4-year-old French and English children with their mothers. While French 4-year-olds had statistically similar PVI values compared with their mothers, English children clearly did not. More recent research by Payne et al. (2011) showed that by age 6, all children effectively acquired the rhythmic patterning of their mothers. The two studies taken together suggest that the syllable-timed advantage in the acquisition of rhythmic patterning levels out by the time children reach 6 years of age. This suggests that in order to capitalize on this advantage, exposure to the spoken language(s) that the child needs to learn should start from 4 years or earlier.

Another set of findings has implications for the early treatment and diagnosis of speech disorders. Peter and Stoel-Gammon (2003) looked at the rhythmic patterning of two children suspected of childhood apraxia compared to healthy controls. They found that singing a familiar song, imitating clapped rhythms, and repetitively tapping showed significant differences. This suggests that comparing the rhythmic patterning of healthy and impaired children speaking a first language can be used as a diagnostic test for childhood speech apraxia.

The rhythmic patterning of native versus non-native varieties of English also showed significant differences. PVI values for Singapore English (Low, Grabe, and Nolan 2000), Nigerian English (Gut and Milde 2002), and Hispanic English (Carter 2005) were significantly lower compared to British English speakers, showing therefore a more syllable-based tendency for non-native varieties of English. The lower PVI values are, at least in part, due to the lack of a strong contrast between full and reduced vowels. However, for teachers of pronunciation, it is important to point out that the absence of reduced vowels may in fact help rather than hinder intelligibility (Janse, Nooteboom, and Quene 2003).

In terms of the development of English as an international language (EIL), Low (2010) showed that Chinese English had similar rhythmic patterning as British English (previously described as norm-providing) and Singapore English (previously described to be norm-developing). This led me to put forward the suggestion that the Kachruvian three circles model for world Englishes requires a re-thinking, at least in the rhythmic domain. No longer is the division of world Englishes into three concentric circles relevant when, in fact, Expanding Circle varieties may display similar attributes to both Inner and Outer Circle varieties. One suggestion is the Venn diagram found in Low (2010), where there is an intersection of Expanding Circle varieties with the two other circles. In other words, the inner and outer Circles should not be contained one within the other but represent separate ends of a continuum. There is therefore the pull of the expanding circle both towards and away from inner circle norms depending on what the speakers are trying to portray or achieve with their language use. This finding has many important suggestions for reshaping the way we think about norms for pronunciation.

First of all, in the EIL classroom, there is a need to consider both local and global norms. Upholding either a local or global norm has different implications. Alsagoff (2007: 39) uses Singapore English as an example to demonstrate the difference between a globalist or localist orientation in the use of a language variety. The global or international variety is associated with “socio-cultural capital, camaraderie, informality, closeness and community membership”. In terms of the EIL pronunciation classroom and in considering instruction on speech rhythm in particular, if learners aspire towards a globalist orientation then stress-based timing should be taught. However, if learners aspire towards a localist orientation, then syllable-based timing should be the focus of the pronunciation classroom. The key here is to introduce the element of choice to the learners, allowing them to decide their identity and orientation in the EIL pronunciation classroom.

Moving to the pragmatic norms in EIL pronunciation instruction, Deterding (2012) cites Crystal’s (1995) suggestion that syllable-based timing is sometimes used by British English speakers to express irritation or sarcasm. In the EIL pronunciation classroom, instructors do need to point out the pragmatic implications when native speakers shift from stress-based to syllable-based timing so as to avoid misunderstandings in cross-cultural speech settings involving high stakes, such as in educational or business settings.

Those who argue for the importance of teaching stress-based rhythm state that it is important to achieve fluency (Cruttenden 2008) since in native varieties of English, the presence or absence of reduced vowels forms the lowest level of the prosodic hierarchy (Beckman and Edwards 1994). This is a view that is echoed by Teschner and Whitley (2004), who state that the sound system of the English language is based on the alternation of strong and weak syllables or stressed and unstressed syllables. Celce-Murcia, Brinton, and Goodwin (1996) also emphasize that stress-based rhythm helps improve the fluency of the speech of learners of English. Wong (1987: 21) considers rhythm to be one of the major “organizing structures that native speakers rely on to process speech”; thus deviation from the native-like rhythm of English might potentially lead native speakers not to fully understand the speech of non-native speakers of English who use a primarily syllable-based timing.

In the EIL classroom, there is a need to introduce both the concepts of stress-based and syllable-based timing and to point out which varieties of English exhibit stress-based or syllable-based tendencies. This is because in the EIL paradigm, it is important to note who one wishes to be understood by, and in some cases stress-based timing is important for achieving intelligibility but in other speech situations syllable-based timing might be more important.

On a final note, the fact that there are more non-native speakers than native speakers of English in the world and that China alone has about 400 million speakers of English suggests that syllable-timed rhythm of Asian varieties may well become the target model for global trade given the rising economic dominance of the region. It is therefore important to emphasize to pronunciation instructors the multirhythmic models available and the need to take student needs and local and global constraints into account when teaching rhythm.

REFERENCES

1. Abercrombie, D. 1965. Studies in Phonetics and Linguistics, London: Oxford University Press.
2. Abercrombie, D. 1967. Elements of General Phonetics, Edinburgh: Edinburgh University Press.
3. Alsagoff, L. 2007. Singlish: negotiating culture, capital and identity. In: Language, Capital, Culture: Critical Studies of Language in Education in Singapore, V. Vaish, S. Gopinathan, and Yongbing Liu (eds.), 25–46, Rotterdam: Sense Publishers.
4. Asu, E.L. and Nolan, F. 2005. Estonian rhythm and pairwise variability index. In: Proceedings of Fonetik, 2005, 29–32, Gothenburg. Accessed June 3, 2013: http://www.ling.gu.se/(en)/konferenser/fonetik2005/papers/Asu_Nolan.pdf.
5. Beckman, M.E. 1982. Segment duration and the “mora” in Japanese. Phonetica 39: 113–135.
6. Beckman, M.E. and Edwards, J. 1994. Articulatory evidence for differentiating stress categories. In: Phonological Structure and Phonetic Form: Papers in Laboratory Phonology III, P.A. Keating (ed.), 7–33, Cambridge University Press.
7. Benton, M., Dockendorf, L., Wenhua, J., Yang, L., and Edmondson, J. 2007. The continuum of speech rhythm: computational testing of speech rhythm of large corpora from natural Chinese and English speech. In: Proceedings of ICPhS, vol. 16, 1269–1272, Saarbrücken, Germany. Accessed June 10, 2013: http://.icphs2007.de/conference/Papers/1591/.
8. Bloch, B. 1942. Studies in colloquial Japanese IV: phonemics. Language 26: 86–125.
9. Bolinger, D.L. 1965. Forms of English: Accent, Morpheme, Order, Cambridge, MA: Harvard University Press.
10. Borzone de Manrique, A.M. and Signorini, A. 1983. Segmental reduction in Spanish. Journal of Phonetics 11: 117–128.
11. Carter, P.M. 2005. Quantifying rhythmic differences between Spanish, English, and Hispanic English. In: Theoretical and Experimental Approaches to Romance Linguistics: Selected Papers from the 34th Linguistic Symposium on Romance Languages (Current Issues in Linguistic Theory 272), R.S. Gess and E.J. Rubin (eds.), 63–75, Amsterdam and Philadelphia, PA: John Benjamins.
12. Celce-Murcia, M., Brinton, D.M., and Goodwin, J.M. 1996. Teaching Pronunciation: A Reference for Teachers of English to Speakers of Other Languages, Cambridge: Cambridge University Press.
13. Couper-Kuhlen, E. 1990. Discovering rhythm in conversational English: perceptual and acoustic approaches to the analysis of isochrony, KontRI Working Paper 13.
14. Couper-Kuhlen, E. 1993. English Speech Rhythm: Form and Function in Everyday Verbal Interaction, Amsterdam: John Benjamins B.V.
15. Cruttenden, A. 2008. Gimson’s Pronuncitation of English, 7th edition, London: Arnold.
16. Crystal, D. 1995. Documenting rhythmical change. In: Studies in General and English Phonetics: Essays in Honour of Professor J.D. O'Connor, J.W. Lewis (ed.), 174–179, London: Routledge.
17. Dasher, R. and Bolinger, D. 1982. On pre-accentual lengthening. Journal of the International Phonetic Association 12: 58–69.
18. Dauer, R.M. 1983. Stress-timing and syllable-timing reanalyzed. Journal of Phonetics 11: 51–69.
19. Dauer, R.M. 1987. Phonetic and phonological components of language rhythm. In: Proceedings of the 11th International Congress of Phonetic Sciences, 447–450.
20. Delattre, P. 1966. A comparison of syllable length conditioning among languages. International Review of Applied Linguistics in Language Teaching 4(3): 183–198.
21. Dellwo, V. 2006. Rhythm and speech rate: a variation coefficient for DeltaC. In: Language and Language Processing: Proceedings of the 38th Linguistic Colloquium, P. Karnowski and I. Szigeti (eds.), 231–241, Piliscsaba 2003, Frankfurt: Peter Lang.
22. Dellwo, V. and Wagner, P. 2003. Relations between language rhythm and speech rate. In: Proceedings of the 15th International Congress of the Phonetic Sciences, M.J. Solé, D. Recasens, and J. Romero (eds.), 471–474, Rundle Mall: Causal Productions.
23. Deterding, D. 1994. The intonation of Singapore English. Journal of the International Phonetic Association 24(2): 61–72.
24. Deterding, D. 2001. The measurement of rhythm: a comparison of Singapore English and British English. Journal of Phonetics 29: 217–230.
25. Deterding, D. 2012. Issues in the acoustic measurement of rhythm. In: Pragmatics and Prosody in English Language Teaching, J. Romero-Trillo (ed.), 9–24, Dordrecht: Springer.
26. Diez, F.G., Dellwo, V., Gevalda, N., and Rosen, S. 2008. The development of measurable speech rhythm during second language acquisition. Journal of the Acoustical Society of America 123: 3886.
27. Faure, G., Hirst, D., and Chafcouloff, M. 1980. Rhythm in English: isochronism, pitch and perceived stress. In: The Melody of Language, L.R. Waugh and C.H. van Schooneveld (eds.), 71–79, Baltimore, MD: University Park Press.
28. Ferragne, E. 2008. Étude Phonétique des Dialectes Modernes de L’anglais des I1es Britanniques: Vers L’identification Automatique du Dialect. PhD Dissertation, Université Lyon.
29. Ferragne, E. and Pellegrino, F. 2004. A comparative account of the suprasegmental and rhythmic features of British English dialects. In: Proceedings of “Modelisations pour I’Identification des Langues”, Paris. Accessed July 2, 2013: http://www.limsi.fr/MIDL/actes/session%20IV/Ferragne&Pellegrino_MIDL2004.pdf.
30. Gilbert, J. 1978. Gadgets: nonverbal tools for teaching pronunciation. CATESOL Occasional Papers 4: 68–78.
31. Gibbon, D. and Gut, U. 2001. Measuring speech rhythm. In: Proceedings of Eurospeech, 91–94, Aalborg.
32. Grabe, E. and Low, E.-L. 2002. Durational variability in speech and the rhythm class hypothesis. In: Laboratory Phonology 7, C. Gussenhoven and N. Warner (eds.), 515–546, Berlin: Mouton de Gruyter.
33. Grabe, E., Post, B., and Watson, I. 1999. The acquisition of rhythmic patterns in English and French. In: Proceedings of the International Congress of Phonetic Sciences, 1201–1204.
34. Gut, U. and Milde J.-T. 2002. The prosody of Nigerian English. In: Speech Prosody 2002, 367–370.
35. Gut, U., Urua, E.-A., Adouakou, S., and Gibbon, D. 2001. Rhythm in West African tone languages: a study of Ibibio, Anyi and Ega. In: Proceedings of Typology of African Prosodic Systems Workshop, Bielefeld University, Germany, May 18–20.
36. Han, M.S. 1962. The feature of duration in Japanese. Onsei no kenkyuu 10: 65–80.
37. Hoequist Jr., C. 1983a. Durational correlates of linguistic rhythm categories. Phonetica 40: 19–31.
38. Hoequist Jr., C. 1983b. Syllable duration in stress-, syllable- and mora-timed languages. Phonetica 40: 203–237.
39. Janse, E., Nooteboom, S., and Quené, H. 2003. Word-level intelligibility of time-compressed speech: prosodic and segmental factors. Speech Communication 41: 287–301. doi:10.1016/S0167-6393(02)00130-9.
40. Kachru, B.B. 1992. Teaching world Englishes. In: The Other Tongue: English across Cultures, 2nd edition, B.B. Kachru (ed.), 355–365, Urbana, IL: University of Illinois Press.
41. Ladefoged, P. 1975. A Course in Phonetics, New York: Harcourt Brace Jovanovich.
42. Laver, J. 1994. Principles of Phonetics, Cambridge: Cambridge University Press.
43. Lehiste, I. 1990. Some aspects of the phonetics of metrics. In: Nordic Prosody V, K. Wiik and I. Ramo (eds.), 206–218, Turku: University of Turku.
44. Lin, H. and Wang, Q. 2005. Vowel quality and consonant variance: a comparison between Chinese and English. In: Proceedings of between Stress and Tone, Leiden.
45. Loukina, A., Kochanski, G., Shih, C., Keane, E., and Watson, I. 2009. Rhythm measures with language-dependent segmentation. In: Proceedings of Interspeech 2009: Speech and Intelligence, 1531–1534, Brighton, UK, 6–10 September, International Speech Communications Association.
46. Low, E.-L. 1994. Intonation Patterns in Singapore English. Masters Dissertation, University of Cambridge.
47. Low, E.-L. 1998. Prosodic Prominence in Singapore English. PhD Dissertation, University of Cambridge.
48. Low, E.-L. 2006. A review of recent research on speech rhythm: some insights for language acquisition, language disorders and language teaching. In: Spoken English, TESOL and Applied Linguistics: Challenges for Theory and Practice, R. Hughes (ed.), 99–125, Houndmills, Hampshire: Palgrave Macmillan.
49. Low, E.-L. 2010. The acoustic reality of the Kachruvian circles: a rhythmic perspective. World Englishes 29(3): 394–405.
50. Low, E.-L. 2014. Pronunciation for English as an International Language: From Research to Practice, London: Routledge.
51. Low, E.-L. and Grabe, E. 1999. A contrastive study of prosody and lexical stress placement in Singapore English and British English. Language and Speech 42(1): 39–56.
52. Low, E.-L., Grabe, E., and Nolan, F. 2000. Quantitative characterizations of speech rhythm: syllable-timing in Singapore English. Language and Speech 43(4): 377–401.
53. Miller, M. 1984. On the perception of rhythm. Journal of Phonetics 12: 75–83.
54. Mok, P.P.K. 2011. The acquisition of speech rhythm by three-year-old bilingual and monolingual children: Cantonese and English. Bilingualism: Language and Cognition 14: 458–472.
55. Mok, P.P.K. and Dellwo, V. 2008. Comparing native and non-native speech rhythm using acoustic rhythmic measures: Cantonese, Beijing Mandarin and English. In: Speech Prosody 2008, 423–426, Campinas, Brazil.
56. Nakamura, S. 2011. Characteristics of contrast between the stressed and the unstressed in rhythm units observed in duration structure in English speech by Japanese learners. Pan-Pacific Association of Applied Linguistics 15(1): 177–189.
57. Nakatani, L.H., O’Connor, J.D., and Aston, C.H. 1981. Prosodic aspects of American English speech rhythm. Phonetica 38: 84–105.
58. Nespor, M. 1990. On the rhythm parameter in phonology. In: Logical Issues in Language Acquisition, I. Roca (ed.), 157–175, Dordrecht: Foris.
59. Nokes, J. and Hay, J. 2012. Acoustic correlates of rhythm in New Zealand English: a diachronic study. Language Variation and Change 24: 1–31.
60. Nolan, F. and Asu, E.L. 2009. The Pairwise Variability Index and coexisting rhythms in language. Phonetica 66: 64–77.
61. Oyakawa, T. 1971. On the directionality of segmental conditioning in Japanese. Monthly Internal Memo, 81–103, University of Berkely, Califonia.
62. Payne, E., Post, B., Astruc, L., Prieto, P., and del Mar Vanrell, M. 2011. Measuring child rhythm. Language and Speech 55(2): 203–229.
63. Peter, B. and Stoel-Gammon, C. 2003. Rhythm production during speech and music tasks in two children with suspected childhood apraxia of speech (sCAS) and two typically developing children. Poster presentation at the Child Phonology Conference, Vancouver, BC, July 2–4, 2003.
64. Pike, K.L. 1945. The Intonation of American English, Ann Arbor, MI: University of Michigan Press.
65. Pointon, G.E. 1980. Is Spanish really syllable-timed? Journal of Phonetics 8: 293–304.
66. Port, R.F., Al-Ani, S., and Maeda, S. 1980. Temporal compensation and universal phonetics. Phonetica 37: 235–252.
67. Port, R.F., Dalby, F., and O’Dell, M. 1987. Evidence for mora-timing in Japanese. Journal of the Acoustical Society of America 81: 1574–1585.
68. Ramus, F., Nespor, M., and Mehler. J. 1999. Correlates of linguistic rhythm in the speech signal. Cognition 73(3): 265–292.
69. Roach, P. 1982. On the distinction between “tress-timed” and “syllable-timed” languages. In: Linguistic Controversies, D. Crystal (ed.), 73–79, London: Edward Arnold.
70. Shen, Y. and Peterson, G.G. 1962. Isochrony in English. University of Buffalo Studies in Linsguistics Occasional Papers 9: 1–36.
71. Strangert, E. 1985. Swedish Speech Rhythm in a Cross-Linguistic Perspective, Umea: Umea Studies in Humanities 9.
72. Tan, R.S.K., and Low, E.-L. 2014. Rhythmic patterning in Malaysian and Singapore English. Language and Speech 57(2): 196–214.
73. Taylor, D.S. 1981. Non-native speakers and the rhythm of English. International Review of Applied Linguistics in Language Teaching 19: 219–226.
74. Teschner, R.V. and Whitley, M.S. 2004. Pronouncing English: A Stress-based Approach (with CD-ROM), Washington, DC: Georgetown University Press.
75. Thomas, E.R. and Carter, P.M. 2006. Prosodic rhythm and African American English. English World-Wide 27: 331–355.
76. Torgersen, E.N., and Szakay, A. 2012. An investigation of speech rhythm in London English. Lingua 122: 822–840.
77. White, L. and Mattys, S.L. 2007a. Calibrating rhythm: first and second language studies. Journal of Phonetics 35: 501–522.
78. White, L. and Mattys, S.L. 2007b. Rhythmic typology and variation in first and second languages. In: Segmental and Prosodic Issues in Romance Phonology (Current Issues in Linguistic Theory Series), P. Prieto, J. Mascaró, and M.-J. Solé (eds.), 237–257, Amsterdam, Philadelphia, PA: John Benjamins.
79. Whitworth, N. 2002. Speech rhythm production in three German–English bilingual families. In: Leeds Working Papers in Linguistics and Phonetics 9, D. Nelson (ed.), 175–205.
80. Wong, R. 1987. Teaching Pronunciation: Focus on English Rhythm and Intonation, Englewood Cliffs, NJ: Prentice-Hall Regents.