Introduction
2 Pronunciation in Language Teaching
John M. Levis
- To explain why certain features should take priority
- To describe why spelling has both promise and peril for teaching pronunciation
- To explore how pronunciation varies by position and dialect
- To argue that variation is an essential part of pronunciation teaching
- To focus on important constant, vowel, and stress variations between American and British English
2.1 Priorities
Setting priorities for pronunciation teaching and learning starts with communication. Pronunciation is a servant skill to the communicative abilities of speaking and listening, and any consideration of pronunciation success must be in the context of how well it helps learners to speak so that they are understood by a wide range of listeners and to listen so that they understand a wide range of speakers. By speakers, we mean both native speakers of English and those who speak English as an additional language. This focus on communication is in line with what Suzanne Firth (1992) called ”the Zoom principle.” Pronunciation should start with a consideration of general spoken language skills (which serve broader goals of communicative competence) before zeroing in on specific features.
This view of pronunciation learning is also informed by the concepts of accentedness, intelligibility, and comprehensibility, first established as three partially related aspects of spoken language by Murray Munro and Tracey Derwing (1995). All three concepts are related to pronunciation, but not in the same way. Accentedness is most closely connected to pronunciation, and it is defined as the difference between someone’s speech and a reference accent (e.g., Standard North American English as spoken in a particular area). Native speakers of English, like native speakers of any language, are incredibly skilled at noticing differences in pronunciation that distinguish speakers who are not “from here” (wherever “here” is). Accentedness is affected by differences, however minor, in vowel and consonant sounds, stress and rhythm, and speech melody marked by pitch variations. In one study that demonstrated how attuned listeners are to differences in accentedness, Munro, Derwing, and Burgess (2010) showed that listeners were able to identify accented speech when the content was masked (listeners couldn’t hear any sounds or words clearly) and played backward. In terms of pronunciation teaching, this means that even the smallest deviation in speech (which is evident in almost any L2 speaker’s production) can be heard as an increase in accentedness.
Fortunately, the fact that accentedness is almost unhideable is not a problem for pronunciation teaching. Many aspects of pronunciation contribute to accentedness but do not affect a listener’s ability to understand or impair a speaker’s ability to communicate. Even native speakers communicating with native speakers from other dialect areas are usually able to communicate successfully despite variations in pronunciation. In other words, they find each other intelligible and comprehensible most of the time, and when there are problems in understanding, they are as likely to come from vocabulary, grammar, and unspoken cultural rules of conversation as from pronunciation.
The variations in pronunciation that are evident in L2 speech are much greater and more unexpected than those found in L1 speech, and some of these differences can lead to a lack of understanding (a loss of intelligibility) or make the listener have to work much harder to understand the L2 speaker (a loss of comprehensibility). Problems with intelligibility or comprehensibility are much more serious than increased accentedness because they are much more likely to impact communication.
Intelligibility
There is wide agreement that the appropriate goal for pronunciation teaching and learning should be intelligibility, not accentedness. Levis (2005) calls these two approaches to pronunciation the Intelligibility Principle and the Nativeness Principle. The Intelligibility Principle says that the things that should be prioritized in teaching and learning are pronunciation features that are most likely to cause someone not to be understood or to make listeners work very hard to understand a speaker. The Nativeness Principle, on the other hand, says that priorities for pronunciation teaching should be anything that marks someone as being nonnative. Because so few adult learners ever become nativelike in pronunciation, this means that almost any aspect of speech could be prioritized regardless of whether it affects understanding. It is our view that trying to become nativelike is almost never the appropriate goal for L2 learners because it is nearly impossible and unnecessary for communication. In contrast, prioritizing those features that have been shown to affect intelligibility or to promote improvement in comprehensibility is the best way to improve communication skills. Such improvement includes, at the very least, teaching with an eye to functional load and prioritizing suprasegmentals.
Functional load
Vowel and consonant mispronunciations often, but not always, cause trouble with intelligibility. A speaker who pronounces path as bath, for example, may leave listeners to work hard to understand what is being said. In other cases, mispronunciations can be impossible to understand. In one example, a graduate student, in a talk about satellites, said that it was important to have zurist power. He meant thrust, but the listeners never understood the word, even though it was repeated 10 times during the talk. How do we know when mispronunciations are likely to cause trouble?
One promising way to decide which vowels and consonants are important is through the use of functional load. Functional load is a measure of the amount of work that is done by two phonemes in distinguishing words in a language. For example, it is important to look at sounds that learners have trouble with, such as /l/-/r/ and /f/-/θ/. But these two pairs of sounds are not equally valuable for pronunciation training. If we think about how many words differ only in the sounds /l/ and /r/ in English, it is easy to come up with many minimal pairs such as lead/read, collect/correct, fell/fair, red/led, fallen/foreign, etc. (There are a lot more, actually,, and many are quite common words in English!) If we want to come up with minimal pairs for /f/ and /θ/, it is much more challenging because there are only a few possible minimal pairs (e.g., fought/thought, free/three). The /l/-/r/ minimal pair is said to have high functional load (FL) because the two sounds distinguish many words (leading to lots of opportunities for misunderstanding), while /f/ and /θ/ represent a low FL pair of sounds, with few opportunities for misunderstanding in speech.
Catford (1987), for American English, and Brown (1988), for British English, both offer lists of contrasting sounds along with their functional load calculations. Not all of these sound pairs are mistakes that learners actually make (e.g., /k/-/h/ is a high FL pair of sounds, but it is rare as a pronunciation problem), but many involve problems that are quite common for learners of various L2 backgrounds. Research has repeatedly shown that high FL sound pairs impact comprehensibility and accentedness more than low FL sound pairs. The first study, by Munro and Derwing (2006), showed that high FL errors affected comprehensibility and accentedness significantly more than low FL errors and that multiple high FL errors increased accentedness ratings but that multiple low FL errors did not. Other studies have shown similar results, with the additional finding that increased numbers of errors can also cause lower comprehensibility ratings even for low FL errors (Alnafisah et al., 2022) and that lower evaluations of free speech are correlated with higher numbers of high FL errors.
FL thus promises to help pronunciation teachers prioritize the types of mistakes that they include in classroom instruction. If a group of learners have both high and low FL problems, emphasis on high FL errors should lead to greater improvement in how L2 speakers are understood (Munro & Derwing, 2006). On the other hand, ignoring low FL errors is unlikely to cause problems with understanding. As teachers, we rarely have the time or resources to work on all the pronunciation errors that our students have, and functional load allows us to set priorities for vowel and consonant errors.
Suprasegmentals
In 1987, Maureen McNerney and David Mendelsohn wrote an article for a Canadian government publication that advocated a focus on suprasegmentals as most appropriate for the best improvement in pronunciation in the short term. Their advocacy fit well within the strong emphasis on suprasegmentals that occurred during the 1980s. Books by Rita Worg and Joanne Kenworthy, also published in 1987, provided a wide range of innovative, communicative activities for suprasegmentals. Because they affected pronunciation at the sentence level and discourse level (that is, beyond the sentence level), suprasegmentals were seen as a good fit for communicative approaches to language teaching. It was sometime after this that research demonstrated that emphasizing suprasegmentals actually did improve comprehensibility in spontaneous speech. In the first study that unambiguously showed that suprasegmentals were special, Derwing, Munro, and Wiebe (1998) examined three classes of English learners in Canada. Over the course of the semester, the first received the normal instruction for their class (with no special attention to pronunciation), the second received instruction on segmentals, and the third was taught with a global approach emphasizing suprasegmentals. In controlled tasks (e.g., reading aloud), both the segmental and suprasegmental groups improved their comprehensibility, but only the suprasegmental group showed improved comprehensibility in spontaneous speech. Since we communicate by speaking spontaneously, not reading aloud, the study shows a superiority for suprasegmental instruction. The same study was replicated by Zhang and Yuan (2020) in China, with nearly the same results. Gordon and Darcy (2016), in a related study with classes of learners in an Intensive English program in the United States, also found that instruction on suprasegmentals was superior to instruction on segmentals in terms of effects on comprehensibility. Finally, in an exploration of self-study focused on suprasegmentals, Foote and McDonough (2017) had advanced learners shadow sitcom scenes over eight weeks, paying particular attention to global features (suprasegmentals and fluency) in their shadowing. The results? In agreement with the other studies, the comprehensibility of spontaneous speech improved
Not only are suprasegmentals in general important for improved intelligibility, but individual suprasegmentals are as well. Misplaced word stress, like mispronounced vowels and consonants, can make words unintelligible. Benrabah (1997), in a study of misplaced word stress by Algerian, Indian and Nigerian speakers, found that British English-speaking listeners relied on stress patterns to interpret what was said. The word writTEN, for example, with stress on the second syllable, was heard as reTAIN. This tendency to rely on word stress to identify words was also evident in Beth Zielinski’s (2008) study of Chinese, Korean and Vietnamese speech, in which stressed syllables were the most likely locations of reduced intelligibility. Misplaced stress does not only affect native English-speaking listeners, but also nonnative listeners, according to a study by Field (2005), in which misplaced stress affected both groups equally. Fortunately, word stress seems to be quite teachable in classroom-based studies, by calling attention to stress rules (Sadat-Tehrani (2017), in relation to the development of practice strategies (Sardegna, 2022), and by practicing word stress in conjunction with physical actions (Murphy, 2004).
Prominence is another suprasegmental that is important for intelligibility and comprehensibility. Laura Hahn (2004) used a matched-guise approach to studying the connection of prominence and intelligibility. A bilingual Korean-English teaching assistant delivered the same short lecture in three guises: one with English-like prominence patterns, one with prominence patterns that contained mistakes, and one without prominence patterns but rather with Korean-like suprasegmentals. Native English-speaking undergraduate student listeners recalled the most information when they heard the lecture with English-like prominence patterns and the least when the prominence patterns contained mistakes.
Other researchers have also argued that prominence is essential for intelligibility for L2 listeners as well. Jennifer Jenkins (2000), in her study of communication between L2 speakers, established that prominence was the only suprasegmental feature that was important for international intelligibility. Like word stress, prominence is also teachable, and its effects on comprehensibility are noticeable even after a short amount of instruction. Levis and Muller Levis (2018) demonstrated this for teaching contrastive prominence for intermediate-level learners of English, who, after 150 minutes of instruction over three weeks, showed improved comprehensibility when producing spontaneous speech in which contrastive prominence was essential to meaning. In another recent study, practice with a corpus of academic English led to better perception and production of prominence (Hirschi & Kang, 2024).
Finally, intonation also affects how well listeners understand speakers. Pickering (2001) examined the English classroom speech of American English and Chinese teaching assistants who differed in the patterns of rising and falling intonation. Chinese TAs used a greater proportion of falling intonation patterns, and listeners found them harder to understand and more distant and unapproachable. Pickering suggested the unexpectedly large numbers of falling contours made listeners think the lectures continued too much information. In another study by Sereno, Lammers and Jongman (2016), when English sentences were acoustically changed to have Korean intonation, listeners found them harder to transcribe correctly (a clear indication that their intelligibility was worse for the listeners).
2.2 Orthography and Pronunciation
Language learners almost always learn pronunciation, at least to some extent, through the use of the written language. This means that they have to negotiate the way that the spelling system (i.e., the orthography) of the L2 connects written symbols to sounds. In any language, this will be a challenge because literate learners already know another orthography very well and unconsciously assume that the new orthography follows a logic similar to the one they are used to reading. This is almost never the case, of course.
In looking at the connection between orthography and pronunciation, it is helpful to see that some orthographies are harder to employ for pronunciation teaching. Orthographies fall on a continuum between transparent and opaque. Transparent orthographies are those in which each letter mostly represents a single sound (though no language is perfectly transparent). Opaque orthographies, on the other hand, have a complex relationship between spelling and sound. It should come as no surprise that English spelling is opaque, making it difficult for most language learners to use for pronunciation learning.
The peril of orthography
Research on orthography and pronunciation learning has come into its own since 2009. In general, we know that a learner’s first language orthography continues to affect how even advanced L2 speakers perceive and produce the pronunciation of the L2. Orthography can affect learners in simple ways, such as happens with the presence of silent letters in English (e.g., half, calf, salmon) or areas in which word stress affects how similar spellings are pronounced (e.g., retain, Britain; table, vegetable). Orthography can also affect learners’ pronunciation more systematically when they expect English to behave like their L1 orthography. This can be illustrated using one connection between pronunciation and orthography that has repeatedly been explored by Bene Bassetti (2023).
Her studies report on geminate consonants. Italian has a phonemic distinction between single and double consonant sounds, but English does not. Both languages, however, have single and double consonant letters. Double consonant sounds are also called geminate consonants. Italian orthography also represents double consonant sounds with double letters. So in the author’s family name, Bassetti, there is a double [s] and a double [t], both of which will be pronounced in Italian. In contrast, an English speaker pronouncing her name will only pronounce one [s] and one [t] because English spelling regularly uses double letters for single sounds. This causes a great deal of confusion for Italian learners of English, who expect double consonant letters to be pronounced with a double consonant sound (that is, a geminate). When faced with spellings like finish and Finnish, Italian learners believe that they are pronounced differently because Finnish has two <n> letters. They also pronounce them differently, with a single [n] and double [nn]. They also are more likely to hear a difference in pronunciation when they see the spelling, even though the English pronunciation of the two words is identical. Bassetti’s research also shows that this effect of orthography is quite resistant to instruction. Even after being taught that doubled letters in English spelling represent only one sound, Italian learners of English continue to make the same mistakes.
The promise of orthography
Despite the peril of orthography, connections between spelling and sound have to be addressed, if only because it is almost impossible to teach pronunciation (or indeed any aspect of language) without written representations of sounds and words. Writing out words or sentences is essential for most controlled production exercises, and even if we try to teach with minimal connections to written language, learners will find that they have to be able to connect the written language code to the pronunciation of new words or wonder how words that they know only from writing are pronounced.
English’s opaque orthography is systematic, though indirectly connected to sounds. Some influential linguists, such as Chomsky and Halle (1968), in their famous book, the Sound Pattern of English, argued that the spelling system of English was ideal because English spelling is morphologically not phonologically organized. In other words, English preserves the visual identity of words through spelling patterns of related words that may be pronounced differently, as in the <cav> of cave, cavity, excavate, where <cav> means “hole” or <electric> in electrical, electricity, electrician, where the final <c> spelling remains visually consistent but inconsistent in pronunciation.
A number of influential teachers have used spelling to help learn pronunciation. Judy Gilbert (2001) teaches vowel sounds according to whether the vowel letters say their name, what she calls “alphabet vowels,” <a, e, i, o u> sounding like [eɪ, i, aɪ, əʊ, ju]. English speakers usually are taught to think of these vowels as “long vowels.” The short vowels, on the other hand, Gilbert calls “relative vowels,” in which <a, e, i, o u> spellings represent [æ, ɛ, ɪ, ɑ, ʌ].
| Spelling | Alphabet (long) vowels | Relative (short) vowels |
|---|---|---|
| <a> | hate [eɪ] | hat [æ] |
| <e> | mete [i] | met [ɛ] |
| <i> | fine [aɪ] | fin [ɪ] |
| <o> | code [əʊ] | cod [ɑ] |
| <u> | chute [u] | shut [ʌ] |
A more complete treatment of English spelling and pronunciation can be found in the work of Wayne Dickerson, for whom spelling is a resource used to teach language learners how to predict the pronunciation of words and sentences they have never heard but perhaps only know from reading. In a 2015 chapter, Dickerson presents principles for using spelling based on a lifetime of research into spelling-pronunciation connections. Dickerson argues that pronunciation is most effectively taught using the “3 Ps” of perception, production and prediction. But his primary contribution to the field is in the use of prediction to help improve perception and production. One of his pronunciation textbooks, Speechcraft, published together with Laura Hahn (1998), uses spelling patterns to help make the prediction of sounds, word stress, and prominence accessible to teachers and learners.
2.3 Variation
Positional variation
With pronunciation teaching, we often assume that language representations are unvarying. We talk about the /p/ sound, the /t/ sound, the /r/ sound, or the /æ/ sound, but each of these phonemes has variations that are quite noticeably different. Native speakers of American English unconsciously treat the variations for each of these phonemes as representing the same sound. Still, the phoneme boundaries and the dialect variations for the sounds are often quite troubling to language learners, who do not bring native phonological intuitions to the sounds of English. Some variations in these phonemes are listed below.
| /p/ | /t/ | /r/ | /æ/ | /ɑ/ – /ɔ/ |
|---|---|---|---|---|
| pit [ph] | take [th] | right [r] | cat [æ] | horse [ɔ] |
| spit [p] | stake [t] | tar [ ] | man [eə] | caught [ɑ] or [ɔ] |
| tip [p̚] | late [t̚] | bird [ɝ] | cot [ɑ] or [ɔ] | |
| later [ɾ] | ||||
| button [ʔ] |
The /r/ also has dialect variations for the post-vocalic /r/, in which the sound is not pronounced outside of the lengthening of the preceding vowel. /æ/ may be pronounced as [ɛ] in some dialects of American English when it is in a word before /g/ or /ŋ/, such as bag or bank. Teachers need to be aware of these variations in their own speech to recognize where they differ from the prescriptive statements about sounds in teaching materials. They also need to be aware of such variations so that they do not require learners to produce distinctions that are not important in the speech of others in the area they are teaching. For example, the last column above shows a merger in our part of the US and in Canada. Some of our authors merge these two phonemes, others do not, but we all agree that teaching the distinction is not important because not even native speakers seem to care about distinguishing these two sounds.
Dialectal variation
A better-known type of variation is that which is associated with dialects. For example, the /æ/ is pronounced with a pure monophthong in Canadian English, but in most varieties of American English is often pronounced with a raised diphthong, [eə]. Both are normal for many educated speakers. The vowels in the cot-caught merger, well described in North American English varieties, are merged to a single sound [ɑ] for much of American English, but to [ɔ] for some varieties of American English and Canadian English. Moreover, few North American English speakers ever find the lack of distinction troubling in the speech of other native speakers. Both of these sounds are distinct, however, in British English varieties.
Dialects of English largely differ from each other in the realization of vowels. The monophthong vowels of cities in the Inland North (e.g., Buffalo, Cleveland, Detroit, Chicago, Milwaukee) have shifted in a chain shift from what they sounded like 70 years ago. So words like buses now sound like bosses, Dawn like Don, and block like black to speakers from other dialect areas (Labov, 2002). Another example comes from the US South, an area with nearly 70 million speakers, where tense and lax vowels have been shifting such that heat sounds like hit and vice versa (to speakers from other areas) and late sounds like let.
What does variation mean for teaching pronunciation?
Variation is important for teachers because it is important for learners. In teaching pronunciation, there are some models of English speech that are well-represented in pronunciation materials (primarily Standard British and General American), but other models are not (such as Scottish English, Irish English, Australian English, New Zealand English, or Southern US English). This lack of representation can lead to the assumption that speakers who pronounce differently are also pronouncing wrongly. English teachers who speak varieties of English that are different from the varieties represented in published materials may assume that their pronunciation is wrong and that they need to model something they don’t actually speak. It is our belief that this attitude that some language models are good and others are deficient is fundamentally wrong. Teachers can only model the variety they speak while being aware of the pronunciation of other models. In doing so, a pronunciation teacher can become a tour guide to the wonderful variations in English speech and can allow learners to make their own choices about intelligible pronunciation.
In addition, a prescriptive attitude about correct pronunciation can disadvantage nonnative teachers. We know that both native and nonnative teachers can be equally effective when teaching pronunciation (Levis et al., 2016) and that training is a more important qualification than nativeness, especially in a world with a ubiquity of Englishes. Part of this training should include knowledge of variation and which variations are important to insist on (e.g., positional variation) and which are not (e.g., dialect variation). Levis (1999, p. 16) provides some guidelines for addressing variations in pronunciation.
- Teachers should be aware of the most common systematic variations of pronunciation in common educated dialects of English.
- Teachers should be encouraged to model their own dialect but should allow students to use any acceptable variation in pronunciation.
- Students should have access to more pronunciation models than they can get from the teacher or the textbook alone.
2.4 American and British English variations
The two major varieties modeled in published pronunciation materials are British and American English. British English pronunciation is often referred to as Received Pronunciation (RP) or Standard Southern British English (SSBE). This is a non-regional variety, and even though as few as 3% of the population of the UK actually speaks this variety (Upton, 2015), it dominates how British English is viewed around the world. The UK has many other varieties with their own pronunciation systems, and these varieties represent the great majority of speakers in the UK. The rise of Estuary English, a hybrid of RP and Cockney (Rosewarne, 1984, 1994), continues to be associated with the greater London area and Southeastern England and is perhaps what many outsiders now identify with British English pronunciation.
American English pronunciation, or what is often referred to as General American (GA), is also a non-regional variety, even though it is associated mostly with the pronunciation of the American midwest and west. About half of the speakers in the US (and Canada) speak some version of GA. Other well-known varieties in the US are Southern US English (actually a wide variety of pronunciation systems that share some distinctive features), New York English, and Boston English, among others. These varieties are identifiable because they have pronunciation variations that are culturally stereotyped (such as the monophthong [a] for [aɪ] in words like ride and night, or the non-rhotic pronunciation used by many New Yorkers or Bostonians).
There are more differences between British English and GA pronunciation than can be addressed in even a full book, but some are important enough for teachers to be aware of. These include vowel differences, consonant differences, and word stress differences. More differences can be found in Celce-Murcia et al. (2010).
Consonants
/r/
The pronunciation of /r/ is one of the most important varietal differences between American and British English. AmE is largely a rhotic variety, whereas BrE is largely non-rhotic. This means that in postvocalic position (e.g., car, form, care), BrE speakers do not say the /r/, but AmE speakers do. In BrE, however, post-vocalic /r/ will be pronounced if the following word starts with a vowel (e.g., care about), whereas AmE /r/ is always pronounced. (Note: Not all British English varieties are non-rhotic, nor are all AmE varieties rhotic, so these generalizations refer primarily to the varieties found in most published materials for pronunciation teaching.)
BrE also has an intrusive /r/ that appears when a word ends with schwa, as in vodka or Cuba. Thus vodka and tonic would be pronounced [vɑdkɚntɑnɪk], with a rhotic quality to the schwa. Some AmE non-rhotic varieties do the same thing. Former President John F. Kennedy, a non-rhotic speaker, pronounced Cuba as [kjubɚ] and idea as [aɪdiɚ].
A common variation of /r/ in BrE is to use a labio-dental approximant [ʋ], which is not quite a [w] or [r]. The late Oliver Sacks, a prominent neurologist, used this pronunciation. This link illustrates Sacks’s pronunciation: Oliver Sacks Ted Talk. Listen for the words with <r> and you will notice the [ʋ] pronunciation.
/l/
Another difference between BrE and AmE pronunciation is found in the pronunciation of dark /l/, as in words like fall, whole, build. Dark (or velarized) /l/ or [ɫ], is a co-articulated allophone of the /l/ phoneme. That means it has both an alveolar lateral articulation using the front of the tongue, like regular or bright /l/ (as in laugh, lot, lift) and a velar approximant articulation, in which the back of the tongue is raised to the velar area (like the co-articulation of [w]). It is very common in BrE that the front articulation does not occur while the velar articulation does. As a result, fall may be pronounced as [fɒʷ], whole as [hoʷ], and build as [bɪʷd]. This pronunciation is sometimes referred to as /l/-vocalization. /l/-vocalization is also the normal pronunciation in Southern US English, but is not common in other areas of the US or Canada. There are very few arguments for the importance of dark /l/ for pronunciation teaching. Jenkins (2000), in her study of English as a Lingua Franca pronunciation, says flatly that this allophone should not be taught because it does not affect intelligibility. Most pronunciation teachers likewise give this sound very low priority, seeing it as affecting accentedness only.
Vowels
There are also notable differences in the inventory of vowels in BrE and AmE. GA usually is listed as having around 15 vowel phonemes, while BrE has around 20. The difference in number has to do with two areas on the vowel chart: low back vowels and vowels before /r/. There are also many dialect-specific differences that reflect different choices of vowels, such as the first vowel sound in process. In BrE, the vowel is [o] while AmE prefers [ɑ].
Low (back) vowels
In BrE, there are three low-back vowel phonemes: /ɑ/, /ɒ/ and /ɔ/. /ɑ/ is the vowel in father, /ɒ/ in cot, and /ɔ/ is found in caught. /ɒ/ and /ɔ/ are both rounded vowels, with /ɒ/ being lower than /ɔ/ (that is, /ɒ/ is pronounced with a more open jaw). In AmE, /ɔ/ does not really exist except as an allophone before /r/, as in words like horse and four. AmE materials, however, use /ɔ/ as the symbol for the rounded low-back vowel, even though most speakers with this low-rounded vowel actually use /ɒ/ (a rounded low-back vowel).
Many GA speakers actually only have one low-back vowel, /ɑ/, rather than the three found in BrE. A key difference between GA and Canadian English is that Canadian English pronunciation uses only /ɒ/. So both GA and Canadian English pronunciations have a single low-back vowel, just not the same vowel.
Another difference between BrE and AmE is found with low vowels. In many borrowed words, such as drama, pasta, macho, and mafia, the vowel in BrE is more front in the mouth, closer to [æ] or [a], while AmE speakers prefer [ɑ]. Canadian English speakers pattern with the BrE pronunciations rather than with AmE. Finally, there are other words such as bath and dance in which AmE speakers prefer [æ] while BrE use [ɑ]. These are well-known variations between the varieties.
| Word Example | BrE | AmE | CnE |
|---|---|---|---|
| father | [ɛə] | [ɑ] | [ɒ] |
| cot | [ɒ] | [ɑ] | [ɒ] |
| caught | [ɔ] | [ɑ] | [ɒ] |
| pasta | [æ] / [a] | [ɑ] | [æ] / [a] |
| bath | [ɑ] | [æ] | [æ] |
r-vowels
Another area in which BrE descriptions have vowels that AmE does not are with vowels before post-vocalic /r/, as in square, near, force, cure. Recall that BrE is a non-rhotic variety, so tense vowels before /r/ are pronounced as lax diphthongs with a centralized (schwa) offglide or simply pronounced as a lengthened vowel. In AmE, a rhotic variety, the distinction between tense and lax vowels is also neutralized but the /r/ is pronounced. Most importantly, BrE descriptions describe these vowels as separate phonemes, but American English does not.
| Word Example | BrE | AmE |
|---|---|---|
| Square | [ɛə] | [ɛr] |
| Near | [ɪə] | [ɪr] |
| Force | [ɔː] | [ɔr] |
| Cure | [ʊə] | [ʊr] |
Word stress
An often-noticed area of difference between British and American English is in word stress. British English speakers say BALlet while Americans prefer balLET, and American English prefers FRUStrate to the British stress pattern frusTRATE. Americans say CORollary, a difference from the British coROLlary, and the British adVERtisement differs from the American adverTISEment. Although these differences in stress seem extremely common, Berg (1999) found that such differences only make up a little more than 1% of words the two varieties share. In addition, many of the differences reflect clear patterns. Recent French borrowings, for example, are usually pronounced with initial stress in British English (e.g., GArage) but with final stress in American English (gaRAGE). In contrast, two-syllable -ate words like frustrate and donate have initial stress in American English but final stress in British English. In another example, words with a final -ary have initial stress in American English but are stressed before the -ary suffix in British English. Berg’s article includes an appendix of all stress differences he found, and it is striking how many of these are place names (e.g., Baghdad) and compounds (e.g., halftime), likely with no changes in vowel quality and thus the likelihood that intelligibility will not be affected.
2.5 Conclusion
For successful pronunciation teaching, teachers need to understand the content they are teaching (the phonetics and phonology of English), but they also need to understand how to set priorities. It is rare for even enthusiastic teachers and learners to have unlimited time to work on pronunciation, and in most instructional contexts, there is very little time afforded for pronunciation teaching. This means that teachers need to know what pronunciation features are likely to lead to communicative improvement, and then target their teaching toward these features. This is what we mean by an intelligibility-based approach. Pronunciation is an essential skill in oral communication, but not all pronunciation errors are equally important in promoting oral communication. It makes little sense in most circumstances to work on errors that will not lead to more successful L2 communication. It makes a lot of sense to work on errors that will.
2.6 References
Alnafisah, M., Goodale, E., Rehman, I., Levis, J., & Kochem, T. (2022). The impact of functional load and cumulative errors on listeners’ judgments of comprehensibility and accentedness. System, 110, 102906. https://doi.org/10.1016/j.system.2022.102906
Bassetti, B. (2023). Effects of orthography on second language phonology: Learning, awareness, perception and production. Taylor & Francis. https://doi.org/10.4324/9780429343117
Benrabah, M. (1997). Word-stress–a source of unintelligibility in English. International Review of Applied Linguistics (IRAL), 35(3), 157-165. https://doi.org/10.1515/iral.1997.35.3.157
Berg, T. (1999). Stress variation in British and American English. World Englishes, 18(2), 123-143. https://doi.org/10.1111/1467-971X.00128
Brown, A. (1988). Functional load and the teaching of pronunciation. TESOL Quarterly, 22(4), 593-606. https://doi.org/10.2307/3587258
Catford, J. C. (1987). Phonetics and the teaching of pronunciation: A systemic description of English phonology. In J. Morley (Ed.), Current perspectives on pronunciation: Practices anchored in theory (pp. 87-100). TESOL Press.
Celce-Murcia, M., Brinton, D., Goodwin, J. & Griner, B. (2010). Teaching pronunciation: A reference for teachers of English. Cambridge University Press.
Chomsky, N., & Halle, M. (1968). The sound pattern of English. MIT Press.
Derwing, T. M., Munro, M. J., & Wiebe, G. (1998). Evidence in favor of a broad framework for pronunciation instruction. Language Learning, 48(3), 393-410. https://doi.org/10.1111/0023-8333.00047
Dickerson, W. B. (2015). Using orthography to teach pronunciation. In M. Reed & J. Levis (Eds.), The handbook of English pronunciation (pp. 488-504). Wiley Blackwell. https://doi.org/10.1002/9781118346952.ch27
Firth, S. (1992). Pronunciation syllabus design: A question of focus. In S. Avery & P. Ehrlich, Teaching American English pronunciation (pp. 173-183). Oxford University Press.
Foote, J. A., & McDonough, K. (2017). Using shadowing with mobile technology to improve L2 pronunciation. Journal of Second Language Pronunciation, 3(1), 34-56. https://doi.org/10.1075/jslp.3.1.02foo
Gilbert, J. (2001). Six pronunciation priorities for the beginning student. The CATESOL Journal, 13(1), 173-182. https://www.catesoljournal.org/wp-content/uploads/2014/07/CJ13_gilbert.pdf
Gordon, J., & Darcy, I. (2016). The development of comprehensible speech in L2 learners: A classroom study on the effects of short-term pronunciation instruction. Journal of Second Language Pronunciation, 2, 56-92. https://doi.org/10.1075/jslp.2.1.03gor
Hahn, L. D. (2004). Primary stress and intelligibility: Research to motivate the teaching of suprasegmentals. TESOL Quarterly, 38(2), 201-223. https://doi.org/10.2307/3588378
Hahn, L. D., & Dickerson, W. B. (1998). Speechcraft: Discourse pronunciation for academic English. University of Michigan Press ELT. https://doi.org/10.3998/mpub.8399
Hirschi, K., & Kang, O. (2024). Data‐driven learning for pronunciation: Perception and production of lexical stress and prominence in academic English. TESOL Quarterly. https://doi.org/10.1002/tesq.3302
Jenkins, J. The phonology of English as an international language. Oxford University Press.
Kenworthy, J. (1987). Teaching English pronunciation. Longman.
Labov, W. (2002). Driving forces in linguistic change. In Proceedings of the 2002 International Conference on Korean linguistics (pp. 1-24). Retrieved from http://www.danielezrajohnson.com/labov_dflc.pdf
Levis, J. M. (1999). Variations in pronunciation and ESL teacher training. TESOL Matters, 9(3), 16.
Levis, J. M. (2005). Changing contexts and shifting paradigms in pronunciation teaching. TESOL Quarterly, 39(3), 369-377. https://doi.org/10.2307/3588485
Levis, J. & Muller Levis, G. (2018). Teaching high-value pronunciation features: Contrastive stress for intermediate learners. CATESOL Journal, 30(1), 139-160.
Levis, J. M., Sonsaat, S., Link, S., & Barriuso, T. A. (2016). Native and nonnative teachers of L2 pronunciation: Effects on learner performance. TESOL Quarterly, 50(4), 894-931. https://doi.org/10.1002/tesq.272
McNerney, M., & Mendelsohn, D. (1987). Putting suprasegmentals in their place. TESL Talk, 17(1), 132-40.
Munro, M. J., & Derwing, T. M. (1995). Foreign accent, comprehensibility, and intelligibility in the speech of second language learners. Language Learning, 45(1), 73-97. https://psycnet.apa.org/doi/10.1111/j.1467-1770.1995.tb00963.x
Munro, M. & Derwing, T. (2006). The functional load principle in ESL pronunciation instruction: An exploratory study. System, 34, 520-531. https://doi.org/10.1016/j.system.2006.09.004
Munro, M. J., Derwing, T. M., & Burgess, C. S. (2010). Detection of nonnative speaker status from content-masked speech. Speech Communication, 52(7-8), 626-637. https://doi.org/10.1016/j.specom.2010.02.013
Murphy, J. (2004). Attending to word-stress while learning new vocabulary. English for Specific Purposes, 23(1), 67-83. https://doi.org/10.1016/S0889-4906(03)00019-X
Pickering, L. (2001). The role of tone choice in improving ITA communication in the classroom. TESOL Quarterly, 35(2), 233-255. https://doi.org/10.2307/3587647
Rosewarne, D. (1984). Estuary English: Rosewarne describes a newly observed variety of English pronunciation. Times Educational Supplement, 29-30.
Rosewarne, D. (1994). Estuary English: tomorrow’s RP? English Today, 10(1), 3-8. https://doi.org/10.1017/S0266078400000808
Sadat‐Tehrani, N. (2017). Teaching English stress: A case study. TESOL Journal, 8(4), 943-968. https://doi.org/10.1002/tesj.332
Sardegna, V. G. (2022). Evidence in favor of a strategy-based model for English pronunciation instruction. Language Teaching, 55(3), 363-378. https://www.doi.org/10.1017/S0261444821000380
Sereno, J., Lammers, L., & Jongman, A. (2016). The relative contribution of segments and intonation to the perception of foreign-accented speech. Applied Psycholinguistics, 37(2), 303-322. http://dx.doi.org/10.1017/S0142716414000575
Upton, C. (2015). British English. In M. Reed & J. Levis (Eds.), The Handbook of English Pronunciation (pp. 251-268). Wiley Blackwell. https://doi.org/10.1002/9781118346952.ch14
Wong, R. (1987). Teaching pronunciation: Focus on English rhythm and intonation. Language in Education: Theory and Practice, No. 68. ERIC.
Zhang, R., & Yuan, Z. M. (2020). Examining the effects of explicit pronunciation instruction on the development of L2 pronunciation. Studies in Second Language Acquisition, 42(4), 905-918. https://doi.org/10.1017/S0272263120000121
Zielinski, B. (2008). The listener: No longer the silent partner in reduced intelligibility. System 36(1), 69-84. https://doi.org/10.1016/j.system.2007.11.004