Author: Mark Hancock

For many teachers, the very heart of pronunciation is the set of individual sounds in the target language. Yes, there is stress, intonation and connected speech too, but sounds feel more basic. However, they are also frustratingly difficult to deal with in class, and hard to integrate with other topics in the lesson. They can also leave the teacher feeling dis-empowered, especially teachers whose own accent is not like standard accents often presented in text books. It’s easy to see why this aspect of pronunciation is often neglected in the language class. In what follows, I would like to outline in a bit more detail three kinds of problems teachers have with individual phonemes, and then go on to suggest some practical solutions to these problems.

PROBLEMS

1 Accent issues

Accent issues play a part in discouraging teachers from dealing with individual sounds in class. Sometimes, pronunciation sections in books seem to assume all learners need to master RP (Received Pronunciation: a standard British accent) or GA (General American). This is particularly noticeable with vowel sounds. For example, your text book might present the ‘correct’ pronunciation for bath as having the long A sound (like art), implying that the short A sound (like cat) is wrong for this word. If the goal of the class is intelligibility, then this kind of accent-bias is totally unnecessary – the short A sound would be perfectly intelligible. Perceived accent-bias in pronunciation materials can also be very alienating for teachers whose own accent is non-standard.

2 Mixed needs

Another problem with teaching individual sounds is relevance to the learners’ needs. One student may find a particular sound difficult while the next has no problem with it at all. This problem is especially acute in classes with mixed first languages. If you teach the distinction between /r/ and /l/ for your Japanese learners, it will be irrelevant for their European classmates. Even if all your class share the same mother tongue, your coursebook may not provide material on their specific L1 issues, since such materials are often intended for a wider market than just one country. To provide targeted pronunciation activities, you will probably need to step away from the coursebook and source your material elsewhere.

3 Unconnected fragments

Sounds are difficult to fit into the storyline of a lesson. Many teachers like lessons to have a good sense of flow, some kind of narrative linking one part of the lesson to the next. But whatever your class is about, it’s unlikely to create an obvious need for a section on ship and sheep. That’s because there is no direct connection between phonemes and meaningful units of language like words and sentences. Topics rarely contain a large number of words featuring any particular sound. If you want to focus on a sound in class, it will usually have to be in a small, separate chunk of the lesson with no strong connection to the rest of it. This makes teaching segmental aspects of pronunciation feel bitty – a collection of random fragments.

SUGGESTIONS

1 Use minimal pairs

I’m going to suggest that taking a minimal pair approach to teaching pronunciation goes a long way to answering the accent issues problem. But first of all, I should make one assumption clear. I assume that for most pronunciation teachers around the world, our job is to help learners understand and be understood. In other words, the goal is intelligibility rather than training them to sound like a native. This is important, because you can be intelligible without sounding like a native.

What are the implications of this assumption? Well, it means that learners do not have to acquire sounds which are exactly the same as the standard model. For example, the vowel in the word gate does not have to be identical to the RP or GA diphthong. What’s important is that it should be different from other phonemes, for example, the vowel in get.

This is the strength of minimal pair activities: they switch the focus away from the individual sound and onto the contrast between a phoneme pair, such as gate-get. The basic message is this: it doesn’t matter what accent you have, as long as you can make the important contrasts. This message is so different from implying that learners must sound like the queen (or any other idealized native speaker). You can sound like yourself.

2 Put minimal pairs in groups

A good way to make your pronunciation lesson relevant to a mixed group of learners is to focus on a set of related pairs rather than a single pair. Take for example the set of words beat, bait, bite, bit, bet, bat. We could call this a minimal group. Within this group, there are a range of different challenges. Some learners may find it difficult to distinguish beat-bit; others may have no problem with this, but instead have a problem with bait-bet, or with bet-bat. By focusing on a minimal group, hopefully, there’s something for everybody.

For all the learners, even those who have no specific problem with any of the pairings, there is also the benefit of seeing how these vowel phonemes work together as a system, allowing them to be compared and contrasted. This connects the individual minimal pairs to the bigger picture, making the pronunciation segment of your class seem less fragmented, or bitty. This goes some way to offering a solution to problem 3 above: although it doesn’t help to integrate your pronunciation section to the topic of the rest of the lesson, it does make it more weighty chunk in its own right. You could also introduce some kind of phoneme syllabus, having a different phoneme group at regular intervals. Here is a suggestion of some groups you could cover (Hancock 2024):

Consonants

– The P group (lip consonants):  pea, bee, fee, V, we

– The T group (tongue and gum consonants): tear, deer, cheer, jeer, year

– The S group (sibilants): sip, zip, ship, vision

– The K group (palate and throat consonants): cold, gold, hold, old

– The TH group (teeth consonants): thin, tin, fin, sin, then, den, Venn, zen

– The L group (liquids): light, right, night

– The N group (nose consonants): sun, sum, sung, sunk

Vowels

– The E group (front vowels): beat, bait, bite, bit, bet, bat

– The A group (open vowels): caught, coat, cot, cat, cut

– The U group (rounded vowels): foul, foal, fool, full, gull

– The R vowel group (vowels before ‘r’): steer, stair, star, store, stir

Hancock, M. (2024), PronPack: The Minimal Pair Collection Hancock McDonald ELT

Accent is a problem in ELT, particularly in pronunciation teaching. In the real world, accents are diverse, and yet we often seem to teach as if only one or two of them are valid. Why is that, and is there any way to make pronunciation teaching more accent-friendly? In this short article, we explore those questions, and I’ll suggest that the answer may be to set a good example.

You have an accent

You sometimes hear people say things like, ‘I don’t have an accent’. On the one hand, this seems like nonsense – like saying, ‘I don’t have an appearance’. On the other hand, I guess we know what they mean. They mean that they have a way of speaking which is felt to be ‘normal’ or ‘neutral’, not marked as being strongly regional or foreign. But whatever that is, it’s an accent too. If you speak a language, you have an accent. Alene Moyer writes, ‘In any language – native or not – everyone has an accent, yet the idea of a neutral accent standard persists in our collective consciousness’ (Moyer, 2013 p.89).

Accent reduction and elocution lessons

The idea of a neutral accent is exploited by courses which offer ‘accent reduction’. If we accept Moyer’s claim, then these courses could be better described as ‘accent training’, not so much losing an accent as replacing it with another – one which is more acceptable in the community where you are living. These kinds of courses are popular nowadays with immigrants in North America, but the idea is not new. They are like ‘elocution lessons’, which were popular in the past in the UK. These were in effect accent training for native speakers, with an emphasis on social climbing: learning to speak your own language in a way which is more acceptable in more upper class circles. In this context, the model accent was ‘Received Pronunciation’ (RP), the word ‘received’ here is being used as a synonym for ‘accepted’. RP is an accent of English which is regarded as standard in the UK and elsewhere, but there is a strong evaluative element here too: the idea that this accent is not only ‘standard’ but a ‘higher standard’ than others.

‘Standard’ does not mean ‘better’

Pronunciation classes often set up a model for learners to aim at. This model is a native accent, and more specifically, a ‘standard’ native accent – RP or GA (General American). But it’s important to understand that being  ‘standard’ does not mean those accents are somehow better than other accents. John Wells points out, ‘A standard accent is regarded as a standard not because of any intrinsic qualities it may possess, but because of an arbitrary attitude adopted towards it by society’ (Wells, 1982 p34).

In the world today, English is an international language, with many more non-native than native speakers. If you want to understand and be understood by as many of those people as possible, having an RP or GA accent is no guarantee. As Wells says, those accents are not intrinsically superior – they are not, for example, more intelligible. Their usage as models is essentially random, relating more to local prejudices within the UK and US than to anything else.

Pronunciation teaching goals

Robin Walker and Gemma Archer outline two alternative pronunciation goals for learners of English – a. a native speaker accent, or b. comfortable intelligibility (Walker & Archer 2024). A native speaker accent was the goal of pronunciation teaching before the 1980s. The model chosen was overwhelmingly either RP, or GA.

The focus on comfortable intelligibility emerged when communicative approaches became more fashionable, with the idea was that we should help learners to understand and be understood. Speakers can be intelligible without necessarily speaking RP or GA.

Of course, the native speaker accent goal still exists alongside the comfortable intelligibility goal. Although the majority of learners around the world mainly need to be intelligible, some specifically want or need to sound like native speakers. To keep the two goals clear and separate, I think we could refer to them using different terms: accent training aims at a native speaker accent goal while pronunciation teaching aims at comfortable intelligibility.

Models or examples?

I’ve suggested that we keep pronunciation teaching separate from accent training, but in practice, the boundary between these two is very often blurred. A lot of teachers teach pronunciation as if they are doing accent training, correcting perfectly intelligible speech simply because it is not native-like. I think a part of the reason for this is to do with an insistence on models.

The strongest argument often given in favour of standard models (such as RP or GA) is that we need a fixed target to aim at, and why not choose one which is widely accepted? If anybody suggests abandoning a standard model, people ask, ‘But what can we replace it with?’ It’s a tough question, because no suggested replacement comes without problems. Yes, it’s a tough question, but maybe it’s not a question that needs an answer. Let me suggest that if we want to develop an accent-friendly approach to pronunciation teaching, perhaps we don’t need to attach so much importance to the question of models in the first place. Instead of models, we simply offer examples – starting with ourselves, and whatever accent we have as teachers.

Teaching by example

Let’s say standard accents such as RP, our own accents, and all other intelligible accents of English are all examples of successful English pronunciation. Any of them can serve as a model; none of them has to be the model. In talking about them this way there is no evaluative judgement going on; no assumption that one accent is superior to another.

We don’t need to protect learners from the reality of accent variation. As listeners, learners will inevitably encounter many different accents, not only the standard ones. Equally inevitably, as speakers, most learners will end up with accents different from the standards, and this is not a shameful fact we should hide away from them.

Nor do we need to create accent anxiety among teachers by implying that we should all have standard accents. If you are an intelligible speaker of English, you are a good example for your learners, whatever your accent.

Walker, R & Archer, G. (2024) Teaching English Pronunciation for a Global World Oxford: Oxford University Press

Moyer, A. (2013)  Foreign Accent: the Phenomenon of Non-Native Speech Cambridge: Cambridge University Press Wells, JC (1982) Accents of English 1: An Introduction Cambridge: Cambridge University Press

Wells, JC (1982) Accents of English 1: An Introduction Cambridge: Cambridge University Press

(First published in IATEFL Conference Selections 2023; this article is a written summary of the conference presentation)

Unlike the written word, the spoken word is different every time you hear it – think of all the different voices and accents in the world. How do listeners ever recognise these various versions as being the same word? This crucial aspect of the listening skill is known as ‘spoken word recognition’. In this presentation, we look at some of the difficulties involved in this, and some of the things we can do in the language class. The analysis is divided into four parts, which I call spelling, storing, priming and processing.

Spelling

According to John Field, the written forms of words tend to stick in the memory more strongly than the spoken forms. Unfortunately, in English, the written form is often misleading and often leads to mispronunciation. For example, many learners pronounce ‘comfortable’ like ‘come for table’.

But what about the consequences for listening? The problem is this: if the learner expects words to sound like their written form, they may not recognise them in speech. We should bear this in mind when teaching. Encourage learners to make some kind of note of words which are pronounced very differently from their spelling. If possible, give them guidance about spelling rules. For example, make sure they are aware of vowel reduction: the letter ‘a’ in ‘comfortable’ is not the same as in ‘table’!

Storing

When we hear a word, we compare it to words which are stored in our memory and look for a match. However, words don’t have only one form. John Field gives the example of ‘actually’. If you’re speaking really carefully, this may have four syllables, but said quickly it may come out as only two, like ‘ashley’. Field proposes that instead of having storing a single form of a word, the listener stores multiple versions (or ‘exemplars’) of it. To help learners build up their repertoire of stored exemplars of a word, we need to expose them to more variety. We can assemble multiple examples of the same word in different contexts, and for this purpose, the online tool called ‘Youglish’ is very useful. It’s like a search engine of video material, and you type in the word (or phrase) you want to hear, and it gives you thousands of examples in different voices, accents and speeds. In order to give learners a sense of how the spoken form of words like this vary, we can use YouGlish in class like this, or else encourage them to make regular use of it at home.

Priming

Listeners do not hear neutrally. We are primed to pay attention to features which are important in our language, while ignoring features which are not. For example, if word stress is important in your first language, you tend to notice it; if it is not, then you tend to be what Anne Cutler calls ‘stress deaf’. With our learners, we somehow need to prime them to pay attention to features which may not be common in their L1, but which are common in English. One approach is to use texts which have a high density of certain common patterns in English such as word endings. For example, I have designed this rhyme to draw attention to the ending ‘able/ible’:

They’re comfortable and durable

They’re lovable, adorable

Fashionable but sensible

To me they’re indispensable

You can make short texts with lots of examples for yourself. Try using Chat GPT: Instruct it to write a brief text containing… and then give a list of words with the suffix you want to focus on.

Processing

Listeners have to process what they’re hearing in real time. According to John Field, ‘listeners may need to form tentative matches on the basis of the available evidence and to confirm or change them as they hear more and more of the utterance’. In the examples below, after hearing the first part of a sentence, the listener understands a, but then after hearing the ending, they must change their interpretation to b:

a. It’s a fish… b. It’s official.

a. Pay a ten… b. Pay attention.

a. It’s a nun… b. It’s an onion.

Expert listeners do this all the time; learners on the other hand tend to stick with the first interpretation, no matter how bizarre. We can use dictations like the examples above in class to raise awareness of this. Read out ‘a’ first and ask learners to write what they hear. Then read out ‘b’ and ask them to correct and complete what they wrote.

References

Cutler, A. (2012). Native Listening : Language Experience and the Recognition of Spoken Words MIT Press

Field, J. (2008). Listening in the Language Classroom. Cambridge University Press

(First published in Modern English Teacher Volume 32 Issue 6)

Phrasal homophones

Look at the phrases in the table below and try saying them to yourself. Do you find that the phrases on the left sound like the phrases on the right? I mean, not just similar, but identical? It seems incredible, but for many speakers of English, they can indeed be identical. They are what I call ‘phrasal homophones’.

A	B
packs and bags plants and berries chops and potatoes cooks and meals sauce and pasta drinks and milk	pack some bags plant some berries chop some potatoes cook some meals saw some pasta drink some milk

Writing listening activities from the audio script

The existence of phrasal homophones like the examples above shows us a couple of interesting things about listening in ELT. The first is this: if you base your listening activities on the audio script (as many course book authors do), you will miss something important. If you’re working from the audio script, you tend to notice the difficult vocabulary and structures and focus on these. However, difficulties are often caused by sections of the text which appear simple in the script, and we fail to focus on these. For instance, as the phrasal homophones demonstrate, in a spoken context some may be confused with and, despite appearing completely different in print.

Scripted versus authentic listening

The phrasal homophones in the table do not always sound identical. If the speaker is articulating carefully, they may sound different. For instance, speakers may pronounce the ‘d’ in and. This raises a second issue about listening in ELT: it’s often based on scripted audio. This is created by actors in a studio reading from a script, and in these circumstances, they tend to speak more clearly than they would normally. This is why Sheila Thorn argues so forcefully that we need to integrate authentic listening into the language classroom (Thorn, 2021). Learners need to have exposure to features of spoken English which are often absent in careful speech.

Connected speech

In writing, and and some look completely different. So how is it possible for them to sound the same in the phrasal homophones above? The explanation lies in a set of features of spoken English which are often referred to by the label ‘connected speech’. Here are some of those features which explain the examples in the table:

1 The words and and some are unstressed. As often happens in such cases, the vowel is reduced to the weak, neutral vowel known as schwa. So there’s no difference between the ‘a’ in and and the ‘o’ in some.

2 The ‘d’ in and is lost. This often happens, especially when there are consonants on either side of the ‘d’. It’s so common that it’s sometimes written that way, for example fish ‘n chips. This kind of change is known as elision.

3 With the ‘d’ cut, the ‘n’ is now the last sound in and. This ‘n’ often changes to an ‘m’ if the following word begins with ‘p’, ‘b’ or ‘m’. This kind of change is known as assimilation.

4 The ‘s’ at the end of the first word links to the ‘a’ of and. This is known as linking and it often happens when one word ends in a consonant and the next begins with a vowel.

Teaching pronunciation for listening

Connected speech is an aspect of spoken English which is often covered in pronunciation materials. We often think of pronunciation as something productive, relating to the speaking skill, but pronunciation is equally important receptively, for listening. In fact, there are some aspects of pronunciation which are much more important for listening than they are for speaking, and I think connected speech is an example of this. Your learners can speak perfectly clearly without the features outlined in 1-4 above. However, as listeners, they will not be able to avoid hearing them, and it’s best for them to be prepared. I’m suggesting that we need to teach pronunciation for listening, and that includes raising awareness of connected speech. In the remainder of this article, I will suggest two ways this may be achieved using fun activities based on phrasal homophones.

Trictation

This activity is a dictation with a trick, which I’ve named ‘trictation’ (Hancock 2022). Read out a few of the phrases in the box below and give learners time to write down what they hear. It doesn’t matter if you say the phrase from Column A or Column B since they should sound the same anyway!

Now for the trick: Write up the phrases for the learners to check their answers – but only write the phrases from Column A. At this point, many in the class will look disappointed because they wrote the phrase in Column B. Finally, write up the answers in Column B too and explain that they are equally correct – that both phrases are pronounced the same.

At this stage, the learners will probably want to know how such different phrases can sound the same, and you can elicit or explain some of the features of connected speech outlined above, such as linking, vowel reduction, elision and assimilation. You don’t need to use this technical jargon, of course.

As a follow up, write some more of the phrases from Column A on the board and ask learners to guess what the sound-alike phrase in Column B could be.

• 

Dictation bloopers

Like the previous activity, this again is based on phrasal homophones and what they reveal about connected speech. In Trictation, the phrases in both the A and B columns make sense. In this activity, one of the phrases is nonsense. Learners have to work out what the intended phrase should be. The nonsense could be contextualised as what voice recognition software might write. For example, it hears cakes and biscuits, but misunderstands and writes cake some biscuits – which sounds identical, but doesn’t make sense.

1 Give out copies of the photocopiable activity Dictation bloopers or project it on a slide. Go through the example in number 1 and show how the wrong phrase and the corrected phrase sound the same.

Note that this example has all the same features of connected speech outlined above – schwa, elision, linking and assimilation, causing and and some to sound the same.

2 Ask the class to suggest what the correct versions of some of the other phrases are. They are in a grid rather than a list to suggest that learners can do them in any order. In that way, they can go straight for the ones they find easiest first.

3 After correcting each of the bloopers, encourage learners to explain how the mishearing happened if they can. This is to raise their awareness of connected speech. Features you may wish to focus on inclued vowel reduction, elision, linking and assimilation outlined above in the connected speech section. This activity also has two examples of ‘blending’, in 3 and 11. This is where the consonant at the end of one word blends together with the ‘y’ of your to make a different consonant. For example, paid your sounds like page or.

Note: Your learners may want to know why the ‘b’ of parrots’ beak sounds like  the ‘p’ parrot speak. It’s because unvoiced consonants like ‘p’ or ‘t’ sound voiced like ‘b’ or ‘d’ when they come after ‘s’ at the beginning of a word. There are other examples of this in 4 and 12.

References

Hancock, M. (2022). PronPack: Connected Speech for Listeners. Hancock McDonald ELT

Thorn, S. (2021). Integrating Authentic Listening into the Language Classroom. Pavilion Publishing

First published: IATEFL 2022 Conference Selections

Pronunciation for receptive purposes

We often think of pronunciation in terms of productive skills, but it’s equally important for receptive ones. Indeed, I would argue that some aspects of pronunciation learning are primarily for the benefit of listening – connected speech in particular. This point is made very clearly if you consider the pairs of sentences below:

            A                                             B

Give him a hug.                      Give them a hug.

Done as a favour.                   Done us a favour.      

Get a receipt.                          Get her a seat.

Gave them an aim.                  Gave her my name.

Speakers may pronounce A and B exactly the same. This is because in connected speech, features such as elision, linking and weak forms can obscure the differences. Obviously, we don’t necessarily want our learners to do this in their own speech – it’s usually better to pronounce clearly! However, as listeners, they have no choice – they’re bound to hear this kind of connected speech, and we need to prepare them for it.

Raising awareness of connected speech with micro-listening

One approach to preparing learners for real connected speech is to focus in detail on very short segments of audio – what John Field calls micro-listening. You can do this by choosing short segments of any audio text which you’re using, but an easy alternative is to use the online tool YouGlish. Type in any chunk you’re interested in and this search engine will find it for you across a whole corpus of online video material. For example, I typed in Give them a. YouGlish then searched and found the phrase in thousands of videos, and played them with a few words before and a few words after my chosen phrase. In this way, my class could hear it in many different contexts, with different speeds, voices and accents. In most of them, the class could hear how the pronoun them was reduced in connected speech, for example to ‘em, and how it was linked up to its neighbouring words.

Integrating connected speech with grammar

A focus on connected speech is important, but it can feel rather random and difficult to integrate with other aspects of a course. One idea would be to integrate it into your grammar syllabus. For instance, if you are teaching a structures such as Give them a hug (that is, ditransitive verb phrases), you can focus on object pronouns in connected speech. Most grammar structures have strings of words including function words like pronouns, articles, auxiliaries and so on – and these are exactly the kinds of words which are most affected by the features of connected speech. This is the approach I took in my book PronPack: Connected Speech for Listeners.

Saying it to hear it

Alongside micro-listening, another approach to raising awareness of connected speech involves learners actually producing it themselves. Although the procedure is productive, the objective is receptive – actually hearing yourself produce this kind of speech is one of the best ways of becoming fully familiar with how it sounds. Any kind of drill which includes examples of connected speech can be used in this approach, but one which is very easy to set up is what I call the counting drill. Here’s an example for object pronouns after ditransitive verbs. You read each line out and the class repeats:

Give ‘em a ONE, Give ‘em a TWO, Give ‘em a THREE, Give ‘em a FOUR

Send ‘er a ONE, Send ‘er a TWO, Send ‘er a THREE, Send ‘er a FOUR

Buy ‘im a ONE, Buy ‘im a TWO, Buy ‘im a THREE, Buy ‘im a FOUR

The idea is that the numbers are so predictable, the learners can focus their attention on the bits which come before and how they are connected up.

Use earworms

Another kind of drill I would recommend for a connected speech focus is a short and simple text, preferably with a bit of rhythm and rhyme. The word-play helps to make the sound of the text ‘stick in the head’ – the earworm effect. Again, you can say the text line by line getting the learners to repeat. Here’s an example, focusing on the same grammar point as the counting chant. The bold shows the stress.

Give ‘em a hug

Give ‘em your love

Send ‘er a gift

Give ‘im a lift

Make ‘er a cake

Give ‘im a break

Send us a link

Buy us a drink

Send ‘em a text

Give ‘em my best

Gimme a call

And love to you all!

A recorded version of this talk can be viewed here: https://www.youtube.com/watch?v=EwDHd2o2lhk

References

YouGlish.  https://youglish.com/

Field, J. (2008). Listening in the Language Classroom. Cambridge University Press

Hancock, M. (2022). Pronpack: Connected Speech for Listeners. Hancock McDonald ELT

TITLE: Spoken word recognition for listeners
NAME(S) OF PRESENTER(S): Mark Hancock
DAY: Tuesday 18 April 2023
TIME: 14:50-15:20
LENGTH: 30 mins
ROOM: Queen’s Suite 7 – Harrogate Convention Centre
AUDIENCE CAPACITY: 75

Knowing a word is one thing; recognising it in the continuous stream of speech is something else. How do listeners accomplish this, and how can we help our learners to achieve the same? In this presentation, we will look at research into spoken word recognition and try out some classroom activities for developing this key aspect of the listening skill.

This article first appeared in the TEIS Newsletter (TESOL Teacher Educator Interest Section Newsletter), December 2022

Know weigh!

Learners are sometimes amazed to discover that words which look completely different in written form are sometimes pronounced exactly the same. It seems almost unbelievable that know weigh sounds the same as no way! With English spelling being so unreliable, it’s no wonder that learners and teachers look for alternative ways to represent pronunciation in writing. One popular option is to write the word using the spelling conventions of your first language. For example, I once noticed White House written as guait haus in a piece of graffiti in Madrid. I often see learners using similar kinds of informal phonetics in their notebooks. I’ve done the same thing myself, representing French enfant as onfon. Seeing the pronunciation in a written form can help to understand it and fix it in the memory – ears and eyes are better than ears alone. But these kinds of informal spellings are very personal – each learner will have their own version – and they are often inaccurate. It’s useful to have something more reliable, and this is where phonemic symbols come in.

IPA symbols

In the world of English Language Teaching (ELT), it’s common to use a set of symbols to represent pronunciation, and the most widely used symbols come from the International Phonetic Alphabet (IPA). Becoming familiar with these symbols is a rite-of-passage for trainee teachers, but too often we fail to understand what they are and how they work. There’s a widespread and unhelpful belief that the symbols somehow only represent one specific accent, and I think this derives from a confusion of ‘phonemic’ and ‘phonetic’.

Phonemic versus phonetic

The first thing we should understand about the IPA that we typically use in ELT is that it’s not the full set –that is designed to cover all languages – but only that small set of symbols needed to represent English. It’s also important to note that in ELT, we normally use the symbols phonemically rather than phonetically. Let me explain this with a concrete example. There are two kinds of L – the clear L and the dark L, and there is a different phonetic symbol for each of these – [l] and [ɫ]. English does have both of these sounds, but there is no meaningful difference between them – they are simply different ‘flavours’ of the same phoneme. We use one phonemic symbol /l/ to represent this phoneme. In other words, the phoneme /l/ includes both sounds [l]and [ɫ]. Trainees should know that a phonemic symbol does not represent one exact and specific sound. By the way, make sure they notice that phonetic symbols are shown between square brackets and phonemic symbols between slash brackets!

Love them or hate them?

There seems to be a love-hate relationship between teachers and the IPA symbols. Some teachers love them, others won’t use them, or only ever use them for the observed lessons they did as trainees. So what’s the problem? The main objection which I’ve heard to the IPA goes like this: ‘My accent is not the same as the accent shown by the IPA, so I can’t use it!’ I believe this worry is based on an important misunderstanding. Phonetic symbols may represent one specific accent, but phonemic symbols don’t.

Symbols and accents

I think the phonemic symbols are best regarded as accent-neutral. Take for example the word bet in a typical English accent and a typical New Zealand accent. The vowel sounds quite different in the two accents – New Zealand pet sounds like bit to English ears. Or from the opposite point of view, English bet sounds like bat to New Zealand ears. However, we can use the same phonemic symbol /e/ for the vowel sound in both accents. This is because the symbol represents a phoneme, not a sound. If we wanted to represent a sound, we would use a phonetic symbol instead.

Phonemes are like chess pieces

The pieces in different chess sets often have slightly different shapes. For example, in one set, the knight may look like a horse’s head; in another set the knight may be a more abstract shape. But despite the differences in shape, both of these pieces play the same role in the game. Phonemes are like this. The /e/ in UK English sounds different from the /e/ in New Zealand English, but they both play the same role in the system as a whole. You could define it this way: /e/ represents the vowel sound in ‘bet’ whatever your accent. As a teacher trainer, this is the message I try to get across to trainees: phonemic symbols don’t represent only one accent; if you are an intelligible speaker of English, they can represent YOUR accent too!

Why do UK and US books often use different symbols?

If phonemic symbols are accent-neutral, then why would British and American books use different ones? I think the answer is that the differences more about academic tradition than accent. Take for example the vowel phoneme in boot, which is often given as /u:/ in UK texts but /uw/ in US ones. This difference has nothing to do with a contrast between the British and American pronunciations of boot; it is merely a different habitual use of symbols. The symbols in themselves are arbitrary – it’s the role they play in the system as a whole which matters.

A chart as a box of chocolates

Phonemic charts often look rather like a box of chocolates – a collection of intriguing symbols, each one in its own separate compartment. Naturally, our attention is drawn to the symbols, like the chocolates in the box, but what if the box itself is actually the important part? I think that’s the case with a phonemic chart – the system as a whole is more important than the individual symbols within.

A system of distinctions

So how is the box more important than the symbols? Well, it’s this: the system of phonemes in English is a system of distinctions. What matters about the vowel in bet is not so much its intrinsic quality, but more the fact that it is distinguishable from the vowels in bit, beat or bait, for example. What is important is not the precise quality of the occupant of each cell in the chart, but the fact that it is different from its neighbours. English and New Zealand speakers may pronounce those individual vowels differently, but they can still distinguish the words and that’s what counts. We have to keep the chocolates separate from one another!

What if you don’t have a distinction in your accent?

I should acknowledge a difficulty with the phonemic chart. Unfortunately, it can’t always be as accent neutral is we might want. Some accents have only one phoneme where other accents have two. It’s as if two of the chocolates in your box have melted together into one. Take for example the two vowel phonemes in full and fool. For many Scottish speakers, there’s only one phoneme here and these two words are homophones. If you are a Scottish teacher and your class asks you to explain the difference between these two symbols in the chart, you will be obliged to say something like, ‘Well, they are the same in my accent, but different in some other accents’. It’s not ideal, but nor is it a reason to reject the entire IPA. That would be like throwing the baby out with the bathwater.

The big picture

No doubt difficulties arise from time to time when we try to use the same set of phonemic symbols for a variety of accents of English, as illustrated with the full and fool example above. But I think the essential point to bear in mind in teaching and teacher training is that the IPA symbols that we use in class are phonemic and not phonetic. This means that they do not represent specific, precise sounds but rather a range of sounds, for example, /l/ represents both the clear and the dark L. It also means that they don’t represent one specific accent, but are flexible enough to accommodate a range of accents – for example, /e/ can represent the vowel phoneme in bet in both British and New Zealand accents. Dear teacher educator, the phonemic symbols can represent your trainees’ accents too; encourage them to feel that they can own them!

This article first appeared in the TEIS Newsletter (TESOL Teacher Educator Interest Section Newsletter), December 2022

“Connected Speech for Listeners”: New book is now out in electronic format on Kobo and Apple Books. Some folks prefer to hold a physical book, but the ebook has certain advantages too, such as easy click-to-hear audio for a quick idea of what this all sounds like.

Surrealism

Imagine this: You’re in the middle of the ocean, and you can see two ships. But the top part of the nearest one actually looks like a huge sheep! And above the other ship, there’s a cloud in the sky which is also the shape of a sheep. It’s a surreal scene which looks like a painting by Salvador Dali, but in fact it is part of an advertising campaign for an English language school in Brazil. ‘What on earth is the connection between this image and English?’, you may ask. But of course, you know already; ship and sheep is a minimal pair. The picture represents one of the more well-known pronunciation problems of English.