THE LINGUA FRANCA CORE Introduction In Chapter 1 we saw how English is now used predominantly as a lingua franca, We explored a number of problems that this new role can produce in terms of pronunciation and, in response to these problems, we examined three different teaching approaches for ELF. I suggested that the best of the three was based on the Lingua Franca Core (LFC). In this chapter, I will briefly describe the research that led to the notion of the Lingua Franca Core, which I will then describe in detail. I will further go on to show how the LEC is open to ‘fine-tuning’ by the inclusion of new empiri- cal data. Finally, I will explain how the LEC isa starting-point, rather than an end-point, in terms of pronunciation teaching and learner goals. Searching for priorities In Chapter 1 I explained how until now intelligibility in spoken English has been described almost exclusively in terms of native-speaker listeners, with English operating as a foreign language. Asa result, any attempts to prioritize pronunciation work for learners have been strongly influenced by contem- porary beliefs about how native speakers make themselves understood to other native speakers. Thus, prior to the 1980s, the priority in pronunciation in ELT was competence in the individual sounds. Vowels and consonants were the object of most classroom activities, and it was generally accepted that until learners had attained sufficient competence in the perception and production of these sounds, it would not be advisable to progress to higher levels, Experts fele that stress, rhythm, intonation, and certain characteristics of connected speech were only appropriate for advanced learners. “The arrival of Communicative Language Teaching in the early 1980s brought about an almost complete reversal of these priorities. The argument put forward to justify this change was that ‘in the absence of complete mutilation26 The Lingua Franca Core of the phonemes by the non-native speaker, the suprasegmentals will carry the day because they bear the meaning of the message’ (Stevens 1989: 183). Analysis of how native speakers used stress, rhythm, and intonation to con- struct their spoken messages led ro the conclusion that learners needed to focus their attention on these features, rather than on individual sounds. Since the dramatic swing to suprasegmentals in the 1980s, work on individ- ual sounds has slowly made its way back into ELT coursebooks at all levels. However, what has not changed is the focus on the native speaker as the interlocutor. As result, pronunciation work in almostall currently available materials assumes that learners are preparing themselves for interactions with native speakers. In the light of what I said in Chapter 1, it should be clear that this assumption does not correspond to the principal use of English today. A number of ELT experts and applied linguists were aware of this mismatch, including Jennifer Jenkins. She realized that there was virtually no empirical information as to how English operates at a phonological level when itis used asa lingua franca. In order to remedy this, she gathered empirical data about intelligibility in spoken English between non-native speakers. To get her data, Jenkins recorded interactions between learners working in pairs and groups in communicative tasks in the classroom. She also collected field data over a period of three years in both classroom and social settings (Jenkins 2000). The subjects of all of her studies were learning English in the UK, and were ‘of upper-intermediate to low advanced level as recognized by the University of Cambridge Local Examinations Syndicate (UCLES)’ (2000: 87). One aim of Jenkins’ research was to determine to what extent breakdowns in communication in ELF settings were due to problems at a phonological level, and to what extent they were due to problems in vocabulary, grammar, general knowledge, and so on. The analysis of her data produced some sur- prising results. For example, in one study Jenkins noted a total of 4o instances of communication breakdown, of which 27 were attributable to pronuncia- tion, eight to lexis, two to grammar, and three to other causes. ‘Two things stand out from this data. ‘The first is the almost negligible role of incorrect grammar asa cause of miscommunication in ELE. This is in marked. contrast with the importance most coursebooks give to grammar, and the amount of classroom time most of us as teachers dedicate to this area. The second point is of direct relevance to this book, and it is the fact that pro- nunciation was found to be the most important cause of breakdowns in ELF communication. Jenkins’ work also led her to question current understanding of the relative importance of individual sounds as opposed to the suprasegmental features of English pronunciation. As we have just seen, Communicative Language Teaching placed attention heavily on suprasegmentals at the cost of work on‘The Lingua Franca Core individual sounds. However, Jenkins’ research into NNS-NNS interactions does not support this emphasis. In her study, all of the breakdowns in com- munication that could be attributed to pronunciation ‘were caused by the transfer of Lt sounds.’ (Jenkins 2000: 88). m2 @3 a4 Key: 1 = pronunciation; 2 = vocabulary; 3 = grammar; 4 = other causes Figure 2.1 The causes of communication breakdown in ELF (Jenkins 2000) “These results raised the question as to why individual sounds were involved in breakdowns to such an extent. The answer, in Jenkins’ opinion, lay in the dif- ference between what happens in native-speaker interactions, and what was happening in interactions between the non-bilingual, non-native speakers that she was studying. Native speakers, she suggested, together with non- native speakers with very high levels of proficiency, use top-down processing to access meaning, and to re-interpret what they think they have heard when the perceived meaning does not make sense. That is to say, they use informa- tion that may not actually be in the words they hear in order to understand the intended message. In contrast, speakers at lower levels of proficiency employ essentially bottom-up processing when listening. This means that they are heavily dependent on the acoustic signal - the actual sounds that they hear. This dependence can be so great that they become completely thrown by deviations in individual sounds, even when there are clear linguis- tic or extra-linguistic clues that contradict the sense of what they think they heard. In one example from Jenkins’ data, a Japanese student describes a picture to her Swiss-German partner. Both students can see a set of six pictures. ‘The task was for the Swiss-German student to listen and identify which picture the Japanese speaker was describing. He had problems completing 27cil The Lingua Franca Core the task, however, because the Japanese student talked about ‘three [led] cars’. ‘This made him think about cars for hire (car rental). In other words he had understood ‘let’ cars. There was nothing in any of the pictures, however, that suggested car hire. More importantly, only one of the six pictures contained cars at all, and these were all red. ‘The confusion was only resolved when the Japanese speaker corrected her pronunciation of the ‘r’in ‘ted’. A I don’t understand the ‘let’ cars. B Let [let] cars? Three [red] cars (very slowly) A Ah, red B Red. A Now [ understand. I understood car to hire, to let. Ah, red, yeah, I see. (Jenkins 2000: 81) Ina top-down approach, listeners with a high level of proficiency would have been able to use the contextual clues. They would have immediately focused on the picture with the cars, and would have used the fact that they were all red to re-interpret the acoustic signal. In the study, the Swiss-German listener relied too heavily on what he had heard, and using a bottom-up process, was unable to make use of the contextual clues. In general, Jenkins concluded that ‘given speakers’ frequent inability to “say what (they) mean” pronunci- ation-wise, which is compounded by listeners’ seemingly ubiquitous use of bottom-up strategies, pronunciation is possibly the greatest single barrier to successful communication in ELF communication’ (ibid. 2000: 83). “The analysis of her empirical data enabled Jenkins to identify the features of English pronunciation that repeatedly caused breakdowns in ELF commu- nication. This in turn allowed her to establish the Lingua Franca Core, a list of pronunciation items central to maintaining mutual intelligibility of ELE. The Lingua Franca Core “The LFC identifies four areas where it is thought to be essential to eliminate error from a speaker's pronunciation if he or she wants to be intelligible in ELF communication: 1 individual consonant sounds 2 groups of consonants (clusters) 3 vowels 4 nuclear stress placement.The Lingua Franca Core Consonant sounds In general, when dealing with differences between the sound system of their own language and that of English, speakers employ a substitution strategy. “That is to say, when a consonant of English does not occur in their first language, they subconsciously replace it with something from their first language that they believe is the same. ‘This substitution strategy can cause serious confusion for both native speaker and non-native speaker listeners. ‘The replacement of /f/ with /p/, for example, a substitution typical of Korean, Malay, Tagalog, and Thai and speakers of English, means that ‘coffee’ will be understood as ‘copy’. Similarly, the consonant /tf/ is not present in most accents of Portuguese, and is replaced by the sound /f/. This results in ‘chair’ sounding like ‘share’. Because of the impact the substitution of consonants has on ELF commu- nication, the LEC requires speakers to be competent, both receptively and productively, in all but two of the consonant phonemes of English. The two exceptions are the voiceless and voiced ‘th’ sounds, /0/ and /8/, as in the words think and then, respectively. The LFC also gives additional guidance as to the optimum pronunciation of five more consonants: /p/, /t/, /k/, /I/, n/. Descriptions and teaching ideas for the consonants of English are widely available in existing pronunciation manuals, and Chapter 5 offers further guidance for learners from ten specific L1 backgrounds, so I will limit my comments here to these seven sounds. /0/ and /o/ ‘These two consonants are classed as dental fricatives: with the tip of the tongue protruding between the top and bottom teeth, air from the lungs is expelled from the mouth. Although they are strongly identified with the pronunciation of English in the minds of ELT practitioners like ourselves, /0/ and /3/ are notably absent from many languages in the world, including some native-speaker varieties of English, such as Irish, Jamaican, or New York. This absence suggests a certain inherent difficulty in their pronun- ciation, a difficulty that is confirmed by the fact that they are often the last consonant sounds of English that native-speaker children learn to pronounce correctly. Some children, in fact, never learn to pronounce them; they sub- stitute /6/ with /f/, to produce ‘fink’ instead of ‘think’, and replace /8/ with a ‘d’-like sound that makes ‘then’ sound like ‘den’. Travelling through London one day, I overheard a young girl playing with the words ‘that thing’, which she variously pronounced ‘that thing’, ‘dat thing’ and ‘dat fing’. At one point in her game she asked her younger brother what he did, listened to his reply, and then declared ‘I can say “that thing”, but “dat fing” is easier’. Wealso now know that /0/ and /8/ are especially resistant to classroom teach- ing techniques (Menyuk 1968; Eckman 1977; Pennington 1996). Despite 2930 The Lingua Franca Core teachers’ persistent and well intentioned efforts, many learners, just like native speakers, consistently substitute these two sounds with similar sounds that they find ‘easier’. Examples of these substitutions can be found on many tracks on the CD, and notably on Tracks 3, 4, 51.7: 9: 10; EI, 13, 19, 20, 22, 25, 28, and 30. These substitutions characterize the pronunciation of non- native speakers even when they achieve very high levels of competence in all other aspects of the language. The difficulty in pronouncing /6/ and /3/ is so significant and so widespread, in fact, that their use is forbidden in com- munication between pilots and air-traffic controllers for certain key aviation terms, Non-native and native speakers alike are required to pronounce ‘three’ as ‘tree’, and ‘thousand’ as ‘tousand’, for example, in international ‘airspeal’. ‘Absent from many of the world’s languages, as well as from numerous native- speaker varieties of English, inherently difficult to pronounce, notoriously resistant to classroom teaching, and unnecessary for ELF intelligibility: it is not difficult to understand why the LEC does not include the dental frica- tives of English. The LFC does not, however, stipulate that speakers should be discouraged from using them if they are part of the speaker's mother tongue pronunciation. Rebecca Dauer, for example, wrongly interpreted the LFC in this way: ‘...from my 30 years of teaching experience, I do not think having students replace /6/ and /8/ with /ff and /v/ is very helpful’ (Dauer 2005: 546). The LEC does not suggest that /6/ and /3/ be ‘replaced’. It simply insists that they are not necessary for intelligibility in ELF. Ipl, ltl, and [ki Six consonants of English, /p/, /b/, /t/, /d/, Iki; /g/, ate collectively known as plosives (or stops) because of the way they are produced. They involve momentarily holding (stopping) the air in the mouth, and then suddenly (‘explosively’) releasing it, Three of these plosive (/p/, /t/ /k/), are frequently classed as voiceless, and three as voiced (/b/, /d/, /g/). However, in certain positions the key difference between voiced and voiceless plosives is aspira- tion, the release of a small puff of air immediately after a sound is made. The voiceless plosives /p/, /t/,/k/ are aspirated when they are on their own at the beginning of a stressed syllable as in ‘pen’, ‘tip’, ‘attend? or ‘because’; their voiced counterparts /b/, /d/, and /g/ are never aspirated. If speakers fail to aspirate /p/, /t/, and /k/ adequately, they will make ‘pear’ sound like ‘bear’ ‘tin’ sounds like ‘din’, ‘coat like ‘goat’, and so on. Correct aspiration of /p/, /el, [kl is therefore essential for ELF intelligibility. hl In English the sound /t/ is made by placing the tip of the tongue against the alveolar ridge. However, when ‘t comes between vowels, as in words like ‘water’, ‘butter’, or ‘matter, itis pronounced slightly differently in AmericanThe Lingua Franca Core English, which has a single, rapid tap of the tongue against the alveolar ridge. ‘This variation is closer phonetically to /d/ than to /t/, and this can make ‘matter’ sound like ‘madder’. The LEC proposes that learners use the British English /t/ in preference to the American English alternative. My In native-speaker English, the pronunciation of in words like ‘lip’ or ‘below’ is not the same as in words like ‘doll’ or ‘mill’. In ‘lip’ and ‘below’, the comes before a stressed vowel, and is pronounced in a way that is known as ‘clear’ /I/, with the tip of the tongue raised to the alveolar ridge in the front of the mouth. In contrast, in ‘doll’ and ‘milk’ the comes after a stressed vowel, and a variant known as ‘dark’ /I/ is used. Here, the back of the tongue is also slightly raised. Dark /I/ is typical of certain native speaker accents such as RP and GA, but it is not necessary for intelligibility in ELF communication. For speakers of some L1 backgrounds, it easier to substitute dark /l/ with an [u}-like vowel rather than use clear /I/. This makes ‘milk’ sound like ‘miuk’ [miok] and ‘doll’ like [dou]. The LFC includes (u} as a valid alternative to dark /I/. An example of this can be heard on Track 15 on the CD where Speaker B pronounces ‘schoo!’ as [sku:4], and again on Track 11, where ‘film’ is pronounced as [frum]. Interestingly, this same substitution was stigma- tized in the past among RP speakers, probably because of its association with the working-class London Cockney accent. Today, however, the use of [0] is increasingly common among young, educated speakers in London and South East England, and may become the standard pronunciation for /I/ in this position for this group. hf ‘The situation with /r/ is similar to that of the dental fricatives /6/ and /8/, both in terms of inherent difficulty and in terms of variation, Some native speakers never learn to pronounce /t/ in the way that is considered standard for the English they speak. In addition, the type of /1/ used changes from one group of native speakers to another. The /r/ of Scottish English is quite different to that of RP or GA, for example. The LEC does not describe exactly how /+/ should be pronounced, but as with /6/ and /3/, the learner's mother tongue will be a deciding factor. The most common variant of /1/ in native-speaker Englishes is [1], a sound that is made by holding the tip of the tongue just behind the alveolar ridge, but without any contact being made. This variant, which is known as a post-alveolar approximant, is notoriously difficult for learners of English who have a different variant in their mother tongue. Another common variant of /r/ is the trill [r], where the tip of the tongue repeatedly touches the alveolar ridge. This variant is a feature of some 3132 The Lingua Franca Core native-speaker Englishes such as Scottish English, and is typical of the pro- nunciation of many other languages including Arabic, Polish, northern dialects of German, Russian, and Spanish. The Spanish speaker on Track 14 uses a trilled /+/ with the words ‘other’ and ‘your’, for example. Spanish also usesa third variant of /r/, a /d/-like flap, [r]. We met this variant in the section on /t/. A fourth variant, the uvular [1], is the made in the throat and is char- acteristic of French and southern German accents of English. This variant is not as widespread as any of the others, so is not likely to be as intelligible in ELF communication. Finally, the /r/ of the German speaker in Tracks 1, 5, 15, and 23 is mainly the result of the speaker rounding her lips, such that in some places it sounds more like /w/. In addition to differences in articulation, native-speaker varieties of English can be put into one of two groups with respect to when they pronounce an ‘ in the spelling of a word. In one group ‘r is always pronounced regard- less of whether it comes before or after vowels, These varieties are known as thotic, and include GA, Canadian English, Scottish English, Irish English, and many Caribbean Englishes. In contrast, non-rhotic varieties such as RP, ‘Australian English, and most varieties of New Zealand English, do not pro- ounce an ‘t’ when it comes after a vowel or before a consonant as in words like ‘car’ or ‘card’. For the LEC, Jenkins opted for the GA rhotic variety of /+/. She chose this option ‘mainly because’ is indicated orthographically in situ- ations (i.e, post-vocalic) where it does not feature in RP pronunciation when a word is spoken in isolation (for example, four) or is followed by another consonant (for example, ‘four books’) rather than a vowel (for example, ‘four eggs’) (2000: 139). That is to say that because rhotic varieties reflect the ‘r’in spelling more closely, they should be easier for ELF users, and should increase ELF intelligibility. Consonant clusters Clusters are groups of two or more consonants. These are found at the begin- ning of individual words, as in ‘cluster’, or ‘group’, in the middle of words, as in ‘cluster’ or ‘consonants’, and at the end of words, as in ‘consonants’. English allows up to three consonants in any position in a word. The word ‘strengths’, for example, has initial and final three-consonant clusters. In word-final position, four-consonant clusters are possible, such as in ‘texts’ [teksts]. Clusters can also occur across word boundaries. A cluster of four consonants is formed, for example, across the word boundary in ‘consonant cluster’. Clusters are not common to all languages, and even when they are, not all languages allow the same combinations of consonants. The outcome of this is that many learners find the pronunciation of English consonant clusters difficult. Faced with this difficulty, they employ two broad strategies:The Lingua Franca Core * deletion of one of the consonants in the cluster * insertion of a short [1]- or [e]-like vowel between two of the consonants. Of these two strategies, deletion affects intelligibility considerably. ‘The dele- tion of the ‘sin ‘strain’, for example, produces ‘train’, and taking the ‘train’ is not the same as taking the ‘strain’! If deletions occur in several places in an utterance, intelligibility can be seriously threatened. Many listeners, for example, would have problems understanding [ju wo jo ‘bade 'peze] (‘You want your birthday present’). On Track 5 of the CD Speaker G pronounces ‘English’ as [ingi/], deleting the /I/. On Track 22 the same speaker deletes the /l/ on ‘blog’, whilst the Malay speaker on Track 26 deletes the /r/ in ‘problem’. Deletions like these can threaten intelligibility. ‘The second strategy learners use, addition of a vowel, is found to cause fewer problems in ELF. ‘The insertion of a short /e/, for example, is typical of Spanish Lr accented English whenever a word begins ‘s+consonant’. Thus, ‘Spain’, with an initial cluster that is not found in Spanish, is pronounced [e'spein]. If the speaker deletes cither of these consonants, the outcome will be either ‘pain’ or ‘sane’. To say you are ‘in pain’ or ‘insane’ is certainly not the same as being ‘in Spain’. Jenkins gives a graphic example of the impacts of addition and deletion when she explains how ‘a Japanese learner’ rendering (in context) of the word ‘product’ as [po'rodakuto] was perfectly intelligible to her ... receiv- ers, whereas the Taiwanese learner’s version [‘podak] was not’ (Jenkins 2000: 142). The speakers on Tracks 6, 19, 21, and 22 all employ this strategy. As a general rule, then, simplification of consonant clusters by addition of a short vowel is an acceptable strategy and will help to retain intelligibility in ELF communication. Deletion, on the other hand, is likely to have the opposite effect, and should be avoided. Although deletion is not an acceptable strategy in word- initial clusters, it can be used with certain clusters that occur in the middle or at the end of words in English. Native speakers commonly pronounce words like ‘postman’, ‘aspects’, or ‘next week’ as ['pausman], ["zspeks]. and [neks'wisk]. In all three examples the /t/ in the middle of the respective cluster has been deleted. This deletion of sounds is known as elision, and is characteristic of native-speaker English in all contexts except those of slow, careful speech. The clision of It] or [d/ when either is the central consonant of three in a cluster does not seem to affect intelligibility in ELE. In addition, eliminating one of the three consonants in a cluster has clear benefits for many learners because it makes the resulting two-consonant cluster easier to pronounce. Almost without exception the speakers in Tracks 21-30 elide the /t/ or /d/ in clusters. There are many other examples of the correct elision of /t/ or /d/ on the other tracks, particularly Tracks 3 and 4. 33The Lingua Franca Core Vowel sounds “There is far more variation in the vowels of English than in the consonants. “The vowel in the word ‘bat’, for example, is pronounced in more or less the same way in American, Australian, and British English, but very differently in New Zealand English, where it sounds more like ‘bet’; standard North American English uses the same vowel in ‘hot’ and ‘heart’, but these are dif- ferent vowels in British English; Australians pronounce ‘day’ in a way that makes it sound like ‘die’ to non-Australian ears. Differences in vowel qualities are also marked as you travel around Inner Circle countries, especially the UK. In England, for example, the word ‘bus’ is pronounced /bas/ in the south, but /bus/ in many parts of the north. Similarly, words like ‘grass’ or ‘bath’ are pronounced /gres/ and /ba6/ in some areas, and /gra:s/ and /ba:0/ in others. In an attempt to make sense of the huge variation in native-speaker accents the phonetician Bryan Jenner examined what they all have in common (1989). The outcome of his research was a graded list of pronunciation pri- otities called the ‘Common Core’ (1989). This differentiated between two characteristics of English vowels, their quantity (length) and their quality, and placed much greater importance on length: Native accents show such enormous differences in vowel quality that it cannot be claimed thae these are vital for mutual intelligibility. Nor can it be claimed that there is a minimum set of vowel shapes that must be acquired by the foreign learner, since different native varieties do not make all the oppositi found in SBE [Standard British English] or any other ‘standard’ variety. (Jenner 1989: 3) “The LFC takes a similar position and requires teachers and learners to focus their attention on the long-short differences between vowels’ rather than achieving any exact, native-speaker quality. In addition, the empirical evi- dence that the LEC is based on revealed a significant number of occasions where communication broke down when the long vowel of ‘bird’ or ‘her’ (2:1) was mispronounced. Because of this, the LFC includes the need for a good approximation to the native-speaker quality of the central vowel Jax/, On'Track 7, for example, the Japanese Lr speaker pronounces the word ‘term as [tam], which proves almost impossible to understand. “The LFC focus on vowel length applies to diphthongs as well as pure vowels. English diphthongs have approximately the same length as the long English pure vowels. ELF users whose first language has no diphthongs will tend to shorten English diphthongs in a way that makes them sound like pure vowels. Words like ‘note’ and ‘not’ will sound the same, for example. The answer to this problem for ELF intelligibility lies in achieving sufficient length in the diphthong rather than in trying to attain an exact, native-speaker quality.The Lingua Franca Core Problems with the length of English vowels are marked not only in speakers’ attempts to distinguish between minimal pairs that constitute long-short pairs such as ‘bear’ and ‘bit’, which have different vowels, but also between pairs that share the same vowel. With pairs like ‘mate-made’ or ‘ice-cyes’, where both words have the same vowel quality, the difference that listen- ets perceive is purely one of length. This happens because in English the length of a vowel varies according to its context. Broadly speaking, a vowel in English can be found in one of three environments: in an open syllable, as in words like ‘bee’, ‘200’, or ‘throw’ in a syllable closed by a voiced (lenis) consonant, as in words like ‘made’, ‘eyes’, bag’, ‘leave’, ‘peas’, etc. in a syllable closed by a voiceless (fortis) consonant, as in ‘mate’, ‘ice’, ‘back’, ‘leaf’, ‘peace’, etc. Vowels in open syllables have been shown to have approximately the same length as vowels in syllables closed by a voiced (lenis) consonant. However, when an accented syllable is closed by a voiceless (fortis) consonant (i.e. /p/, Iel, Ik/, Itfl, Ifl, 101, /s/, [f/)y this has the effect of shortening the preceding vowel. The word ‘mate’, for example, is perceived by listeners as being differ- ent from ‘made’ not because of the voiceless/voiced nature of the final /t/ or /d/, but because the vowel in ‘mate’ is shorter than the same vowel in ‘made’. Vowel length also explains the difference between ice/eyes, back/bag, leaf! leave, or peace/peas, with the vowels in ‘ice’, back’, ‘leaf’, and ‘peace’ being shorter. ‘Two general points are worth making before finishing this section on the LEC and vowels: 1 Ofthe languages covered in Chapter 5, for example, only Arabic, German, and Japanese have vowel systems where differences in vowel length can create differences in meaning. Learners from languages where vowel length is more or less constant, such as Chinese, Greek, Malay, Russian, or Spanish, usually have important problems in this area. ‘The LEC focus on length is not to say that learners are free to use any vowel in any place, Ifa speaker pronounces ‘beat like ‘Bart’ one time, for example, like ‘bird’ the next, and like ‘board’ the third time, communica- tion will most probably break down. There is a great deal of tolerance of variation in vowel quality in ELF, but the qualities of individual speak- ers must be reasonably consistent. Someone who is used to listening to American English, for example, might initially be confused by the way a British English speaker pronounced words like ‘caught’ or ‘thought’. But if the speaker in question always does this, the listener will quickly ‘tune in’ to this different quality. Vowel qualities for ELF pronunciation need to be consistent in this same way, and in Chapter 4 we will look at how this consistency can be achieved. x 35OEE _ Er ee 36 ‘The Lingua Franca Core Nuclear stress placement “The last of the four major pronunciation features that make up the LFC is the area of nuclear stress. Analysis of spoken English shows that speakers divide what they say into groups of words. ‘These often coincide with grammatical structures such as clauses or simple sentences. However, they can be as short as one word, and on average tend to be around four words. Various names have been given to these naturally occurring groups of words, including tone units, thought groups, and word groups. Iwill use the term ‘word group’ in this book. “The first benefit that comes from the use of word groups is that they break the speech flow up into manageable, meaningful blocks of information. This makes the task of the listener easier for two reasons. Firstly, it provides the information in packages that bring out the meaning more clearly than if the speaker simply pauses at random. Secondly, the small pauses between each word group allow the listener time to process what he or she has heard. Breaking speech up into word groups in this way is especially valuable in monologues, as opposed to dialogues or conversation. Typical ‘monologues’ include lectures at institutions of higher education, presentations in the world of business, instructions in public places such as stations or airports, or commentaries by tour guides at monuments and other tourist attractions. In these situations the listener cannot easily interrupt the speaker and request clarification, and the onus is on the speaker to make life as easy as possible for the listener. Using word groups is one way of doing this. Not pausing adequately, in contrast, can havea serious impact on intelligibil- ity. Firstly, it can lead to breaks in the speech flow in unexpected places, and this will reduce intelligibility. A weather forecaster saying: 1 ‘We can expect heavy rain in the southwest from Monday onwards. The situation will improve slowly.” is telling us something quite different from: 2 ‘We can expect heavy rain in the southwest. From Monday onwards the situation will improve slowly.” Secondly, poor pausing, or more specifically, too little pausing, reduces plan- ning time for the speaker. This can easily lead to errors of pronunciation and/ or vocabulary, which will in turn complicate the job of the listener. In spoken English, one syllable in each word group is prominent. That is to say, it is made to stand out from the others. To do this, the chosen syllable is made louder and longer than the others. In addition, the most significant pitch movement (tone) in the word group coincides with this syllable, which is called the nucleus (or tonic). The correct placement of nuclear stress is important for intelligibility in ELF:The Lingua Franca Core 37 Nuclear stress, whether unmarked (on the last content word in the word group) or contrastive (somewhere else) is the most important key to the speaker's intended meaning. It highlights the most salient part of the message, indicating where the listener should pay particular attention, And contrastive stress is especially important in English as it does not have the morphological and syntactic resources that many other languages have to highlight contrasts: English has few inflections and the word order is relatively inflexible. (Jenkins 1997: 18) Ina conversation between friends, the word group ‘T've rented a flat’ could be said in three different ways: 1 T've rented a FLAT. 2 T've RENTED a flat. 3 I'VE rented a flat. Int, the nuclearstress falls on the last lexical item. Thisis known as unmarked stress, and is, we could say, a ‘neutral’ way of saying this phrase. It tells the listener what type of accommodation the speaker has taken. In both 2 and 3, the nucleus is earlier in the word group. Both are examples of contrastive stress, and the placement of the nucleus on a word other than the last lexical item gives these two word groups different meanings. By choosing to make ‘rented’ the nucleus, for example, the speaker draws attention to the idea of not having bought a flat. In 3, the focus has been put on the fact that it was the speaker who rented the fat, and not some other person mentioned in the conversation. Both speakers on Track 5 make effective use of contrastive as does Speaker B on Track 15. stres: Using nuclear stress inappropriately can seriously confuse listeners by drawing their attention to the wrong part of the message. In the example below (Jenkins 1997: 19), a Swiss speaker explained to her Taiwanese listener how many cigarettes she smoked a day. ‘The Taiwanese partner responded by saying ‘You smoke more than I DO.’ There was no attempt at contras- tive stress, which correctly done would have led to ‘You smoke more than / do’, with the nuclear stress on ‘T’ in order to make the contrast between the speaker's smoking habits and his partner's. Instead, the nucleus fell on the last word in the word group, and, as a result, the speaker had to repeat the utter- ance several times before he was understood. In another example from her field data (ibid.: 18-19), Jenkins reports on a Hungarian speaker asking his Brazilian and Swiss-French partners for a col- oured pen. To do this he asks ‘Do you havea blue VUN?,, placing the nuclear stress on the last word. This incorrect placement was further compounded byan error in the sound /w/, which was pronounced as [v], probably because of Lr transfer. The outcome was that neither partner understood the request and simply repeated ‘Vun?’ ‘Vun?’, the word that the misplaced nuclear stress had drawn their attention to. Had the speaker said ‘Do you have a BLUE oc Cl lh LL rh SLE ELE rl lc lc rl LrThe Lingua Franca Core vun?’, his partners would probably have understood him despite the /w/—/v/ substitution. Non-core features A number of features of English pronunciation that are widely considered to be central to intelligibility for English as a Foreign Language (EFL), where the listener is a native speaker, are noticeably absent from the LFC. These non-core features include: * /0/, /8/, and dark /I/ * exact vowel quality + pitch movement (tone) * word stress + stress-timing + vowel reduction, schwa, and weak forms + certain features of connected speech linking, assimilation, coalescence [explained the reasons behind the omission of /0/, /8/, dark //, and vowel quality from the Lingua Franca Core in the previous section of this chapter. In this section I will concentrate on the three remaining non-core items. Before doing this, two points need clarifying: 1 The LEC allows for variation in the non-core items. That is to say that individual speakers ate free to pronounce these items differently in the knowledge that variation here will not result in any lack of ELF intelligibility. 2 In traditional approaches to pronunciation, variation in the non-core fea- tures of the LEC is perceived, at best, as a producing a foreign’ accent. All too often, however, variation is equated to ‘error’. With ELF this is not the case. For example, if two speakers pronounce ‘sold’ with different vowel qualities, the difference is not a question of correctness. It is simply an instance of L2 regional variation. We do not question this sort of variation between speakers from the Inner Circle, and should not question itamong speakers from the Expanding Circle in ELF communication. Non-core items with no impact on ELF intelligibility Pitch movement (tone) ‘As teachers, most of us will have had our own experience of students who scem perfectly capable of imitating a tone that has been modelled for them in class, but who are unable to use the same tone appropriately when participat- ing in communication tasks later. This inability to transfer tones from drills and other imitation activities to acts of spontaneous speech is to do withThe Lingua Franca Core the way tone is selected. The choice of a particular tone is made at the very beginning of the sequence of neurolinguistic processes involved in human speech. These processes do not allow speakers to consciously select tone in the same way that they can consciously articulate a particular consonant. Instead, speakers subconsciously use the same tone that would be appropri- ate for the same message in their mother tongue. Inaddition to this neurolinguistic obstacle of subconscious L1 phonological transfer, some pronunciation experts now feel that the complex ways that tone is used in English are ‘unteachable’. By this they mean that it cannot easily be generalized into rules that are simple enough to be taught effectively in the classroom. For pedagogical purposes, it might in fact be helpful to think about the various aspects of pronunciation along a teachability-learnability scale. Some things, say the distinction between fortis [voiceless] and lenis [voiced] consonants, are fairly easy to describe and generalise — they are teachable. Other aspects, notably the attitudinal function of intonation, are extremely dependent on individual circumstances, and therefore nearly impossible to isolate out for direct teaching. (Dalton and Seidlhofer 1994: 72-73). A third problem with the teaching of tone is that pronunciation experts seem unable to agree as to the actual meaning of individual tones. Richard Cauldwell, for example, spent many years working at Birmingham University’s English Language Research Unit, where he further developed the discourse intonation theory originally laid down by David Brazil (1997). Despite this invaluable experience, in a posting to an e-list discussion among international pronunciation experts about the meaning of different tones, Cauldwell declared that: after working for nearly twenty years with Discourse Intonation on examples of spontaneous speech I no longer feel that tones ‘mean’ anything. My view is that they are one of the ways of making your speech interestingly variable, Particular meanings already contextually present can be cued (drawn attention to) by particular realisations ofa tone ... but, in my view, it is a misrepresentation of reality to say that rising tones have a dominant referring meaning, Or indeed, any meaning, (Cauldwell 2006) Neurolinguistically inaccessible, pedagogically unteachable, possibly mean- ingless — even though they are not damaging to intelligibility in ELE, the teaching of tones is not a good investment of classroom time. Word stress Word stress is another non-core feature where variation is unlikely to have a negative impact on ELF communication. ‘This was clearly the case on Track 2, where Speaker C made numerous non-standard pronunciations in terms 340 ‘The Lingua Franca Core of words stress. Jenkins (2000: 150) regards word stress as ‘something of a grey area’, and points out that although icis reasonably important for native- speaker listeners, it rarely causes intelligibility problems among non-native speakers, except when an error in words stress occurs in combination with an error in a particular sound. In terms of classroom practice, the full set of rules that govern word stress in English is probably unteachable because of its complexity, just as with the rules governing tone. Word stress, for example, depends on the number of syllables in a word, the origin of any suffixes, the position of a word in a phrase (‘thirteen’ can be stressed on either syllable depending on where it comes in an utterance), the grammatical category of a word, and so on. However, even though our goal is ELF intelligibility, it may be worth paying some attention to word stress for two reasons: 1 The exact role of word stress in ELF is not yet fully understood, and some studies suggest that incorrect word stress could have a negative impact for both native and non-native speaker listeners (Field 20053 Rajadurai 2006). 2 Work on word stress provides us with preparatory exercises for work on nuclear stress placement. As we saw in the previous section, this is essential to ELE and the mechanisms for perceiving and placing stress at the level of whole utterances is the same as the mechanism for perceiving and placing stress at word level. Stress-timing Stress-timing refers to the way that native speakers of English appear to ‘squash’ syllables together so that stressed syllables in the speech flow can come at more or less regular intervals. Stress-timing is a key feature of the rhychm of native-speaker English, and has come to occupy a significant place in the teaching of pronunciation. Today, in fact, it enjoys priority status in many teaching programmes, despite the fact that the division of languages into stress-timed or syllable-timed is widely questioned. Jonathan Marks, for example, points out that ‘{allthough the notion of stress-timing is often j referred to in pedagogic models of English pronunciation, and forms the basis for some classroom materials for pronunciation development, there appears to be no hard evidence that it really exists’ (Marks 1999: 191). Peter Roach makes the same point: “Thete are many laboratory techniques for measuring time in speech, and measurement of the time intervals between stressed syllables in connected English speech has not shown the expected regularity; moreover, using the same measuring techniques on different languages, it has not been possible to show a‘The Lingua Franca Core real difference between ‘stress-timed’ and ‘syllable-timed’ languages. (Roach 1991: 123) Whether it exists or not, work on stress-timing would not be a good use of classroom time for students whose target is ELF, particularly as it is not necessary for intelligibility in ELF. More importantly, the features of native- speaker pronunciation that are claimed to create the effect of stress-timing can all have a negative impact on ELF intelligibility, as we will now see in detail. Non-core features with a negative impact on ELF intelligibility Vowel reduction, schwa, and weak forms In order to make stressed syllables more prominent in the speech flow, native speakers ‘weaken’ unstressed syllables. To do this, they replace the vowels in unstressed syllables, usually with weak vowel schwa (/a/), although also with the short vowels /a/ ot /u/. This process of vowel reduction is also applied to unstressed syllables within words, and this is reflected in dictionaries. The word ‘chocolate’, for example, can be found transcribed as both ['tfok.*lot] and ['tfpk hit]. The transcription shows that the leccer ‘a’ in the written word is pronounced either as /a/ ot /1/. Similarly, the second letter ‘o’ in the written form is pronounced as /a/ or, as the use of superscript indicates, is simply not pronounced at all. That is to say, it is so strongly reduced that it disappears altogether, leaving the spoken word with only two syllables. Vowel reduction is also apparent in the unstressed pronunciations of gram- matical categories such as pronouns, auxiliary verbs, articles, and prepositions. All of these words have two pronunciations in English, with a ‘strong’ form where the vowel is given its full value, and a ‘weak? form, where the vowel is reduced. ‘The strong form is used for citing the word, or for when itis delib- erately made the nucleus of a word group for reasons of contrastive stress. Compare the vowels in the auxiliary verb ‘can’, in: ‘Ican DO it.’ (unmarked stress; ‘can’ = weak form pronunciation = [kan]) ‘| CAN do it.’ (contrastive stress; ‘can’ = strong form pronunciation = [ken]) Weak forms (and vowel reduction) are given considerable attention in EFL coursebooks and in pronunciation programmes. However, despite this, attention the vast majority of learners, including many who become fluent bilinguals, use few weak forms other than ‘a! and ‘the’. In this sense, despite the fact that ic is easy to formulate clear rules about weak form use, they are unteachable. (Jenkins 2000: 147). 41