How do you like your vowels: soft or hard?

Learning to speak your first language is an automatic process that in the absence of hearing or language pathology leads to guaranteed and absolute success. An infant born into a Swahili-speaking family cannot help but learn it. A child whose mother speaks to them in Finnish and whose father speaks to them in Romanian will become a fluent speaker of those two languages. Learning your second language, on the other hand, is a difficult process and whilst we may achieve grammatical success, pronunciation will represent a significant and most likely permanent challenge for us. Linguists working in language acquisition have shown that our ability to discriminate linguistic sounds rapidly diminishes as we home in on the sounds of our native language. As adults we can’t help but hear foreign language sounds through the perceptual sieve of our first language. We might not be able to discern the difference between two sounds which our language lacks; we may hear the sounds represented by the German letters ö and ü as both sounding like the exclamation eurgh! We might also erroneously hear a new sound as being the same as a sound in our native language. For example, Hindi has different types of d sound which are as distinct for native speakers of that language as the difference between z and k in English. But because English doesn’t use different kinds of d sound for differentiating words, native English speakers learning Hindi often fail to hear the difference between words like डाल (branch) and दाल (dahl) which both end up sounding like dahl.   

If we can’t hear a difference, it’s unlikely that we’ll be able to produce it. In 1910, the Danish linguist and language teacher Otto Jespersen described a scene which will be instantly familiar to anyone who has ever tried to teach someone else to say something in an unfamiliar language: 

‘The pupil reads some word in some miserably erroneous way, the teacher stops him and pronounces the words in, let us assume, the correct way. The pupil tries to imitate that pronunciation, but fails, and thus we have an endless repetition of the same word by the teacher, followed very often on the part of the pupil by an equally endless repetition of nearly the same bad pronunciation as before’

Such drills are often accompanied by resources and teacher instructions that label the sounds with which the learners are grappling. Irish consonants are ‘broad’ or ‘slender’. Swedish vowels are ‘soft’ or ‘hard’. Th can be ‘thick’ or ‘thin’ in English. What do these terms mean and do they actually help language learners break the stressful and boring learning pronunciation drill scenario Jespersen describes? How can we instruct someone to produce the difference between ö and ü and the Hindi words for ‘branch’ and ‘dahl’? If I direct you to make a ‘soft’ vowel instead of a ‘hard’ one, what do you physically need to do?

Phonetics is the branch of Linguistics concerned with using empirical methods to examine the sounds of human language. It’s this area of Linguistics that Jespersen was calling upon his fellow language teachers to employ in their classes at the start of the previous century. He wanted teachers to use phonetic terminology and instruction in classes to enable learners to appreciate the mechanism of sounds and improve their pronunciation.  

Before following in Jespersen’s lead, some caveats need to be applied which Jespersen most likely wouldn’t have applied himself. Firstly, pronunciation in second language acquisition is extremely variable. Our emerging second language pronunciation is controlled by sounds transferring across from our first language as well as lots of social and cognitive factors often beyond our control. This means it’s highly unlikely we’re going to achieve native-like pronunciation in our second language. Secondly, we don’t want to eradicate foreign accents- not just because it’s probably impossible but also because my slightly shaky pronunciation in Swedish is part of my identity as a learner. I’m not going to pronounce things like a native speaker every time and this is perfectly fine- especially as native speaker pronunciation can vary enormously depending on the social context or the rate at which someone speaks. David Abercrombie, the first Professor of Phonetics at the University of Edinburgh, referred in 1949 to ‘comfortably intelligible’ pronunciation which he defined as ‘pronunciation which can be understood with little or no conscious effort on the part of the listener’. This, and not native-like pronunciation, is the sensible and attainable goal of second language learning and teaching. But how can I be comfortably intelligible in Irish and make sure I say ‘lunch’ and not ‘lion’ if my teacher tells me that the l in the word lón is ‘broad’ whilst this l in leon is ‘slender’?

Irish’s distinction between ‘slender’ and ‘broad’ consonants is a prime example of how Phonetics can help learners. An internet search of these terms quickly illustrates how frustratingly difficult these sounds are for beginner level learners of Irish. There are numerous videos of people reading words aloud and language learner forums full of calls for help.

‘Broad’ consonants are in fact velarized. This means that they are produced with involvement of the velum, the anatomical term for the soft palate at the roof of the mouth. If you curl your tongue tip backwards you can feel a hard ridge at the roof of your mouth. Move your tongue back even further and you can feel a seem where the hard palate becomes fleshy and soft. This is the velum. ‘Broad’ consonants in Irish are produced with a velar glide, meaning when you make a consonant you also move your tongue back towards the velum at the same time. This happens for lots of English speakers when they say the word tool; the l is velarized and can sometimes sound like a w. So the l in lón is produced with the tip of the tongue in the position for an l and the back of the tongue approaching the velum.

Contrasting with velarized consonants in Irish are ‘slender’ consonants, or more phonetically speaking palatalized consonants. This means they are produced with the tongue near or touching the hard part of the roof of the mouth. Spanish Ñ is a palatalized sound with which many readers may be familiar; the difference between ñ and n in Spanish involves the presence or absence of palatalization. To make a palatalized d, put your tongue in position for an English d sound, but roll the front part of your tongue (not the tip!) tip upward when you release the d.

Palatalized sounds are those which have a y-like sound overlaid on the consonant by involving the front part of the tongue and the hard palate. So instead of referring to ‘broad’ and ‘slender’ sounds, terms which are as instructionally transparent as talking about ‘sad’ and ‘happy’ consonants, we can give our learners some phonetic instruction. Bó (cow) is pronounced differently than beo (alive) because the former is velarized and the latter is palatalized; although, we don‘t necessarily need to use these terms at all. We could simply say “The b in bó should be pronounced with the back of the tongue bunched up” and “The b in beo should have a y-sound in it so push the front of your tongue up as you release the sound”. 

Giving phonetically informed instruction to learners requires some extra knowledge on the part of the instructor. Language teachers don’t need to understand or use complicated phonetic terminology. In order to give meaningful pronunciation instruction they do, however, need to have an insight into how their own mouths differentiate the sounds of the language they are teaching. The Irish example may strike a chord with any Irish language teachers or learners, however, let’s now turn to consider two broad phonetic parameters which are of direct and practical use to teachers and learners of all languages.

Voiced versus voiceless

Voiced sounds are those which are produced with vibrating vocal cords and voiceless sounds describe sounds which don’t have any vocal cord vibration. An English example would be the distinction between pat and bat. These two words are produced in exactly the same way: b and p have the lips in the same place and are differentiated only by the presence or absence of voicing. Put simply, you can feel a buzz in your throat in voiced sounds like b and z that you can’t feel in voiceless sounds like p and s. Languages don’t always have separate letters for voiced and voiceless sounds; think of the words this and thanks, which begin with voiced and voiceless consonants respectively.

Once we understand the difference between voiced and voiceless sounds we can start to produce distinctions in new sounds by switching our voicing on and off. Welsh has a series of nasal consonants which can switch to voiceless ones under certain grammatical conditions. Books and teachers often call the voiceless equivalents ‘thin’ or ‘quiet’, however, we can use our new-found phonetic insight to give a more detailed and ultimately useful description of these sounds. If we start saying n and draw out the sound into a long nnnnnnnnnnn before switching off our voicing we’re left with a sound that is ultimately just letting air escape through our nose- this is the voiceless nasal sound written nh in Welsh. We can switch off our voicing to produce voiceless equivalents of all consonants. In English, we do it with b, d, g to make p, t, c, but languages like Icelandic and Welsh take things a step further by producing voiceless nasals like Ngh as well as voiceless r and l sounds. Once learners appreciate the difference between voiced and voiceless sounds in their own language, they can then turn this new perceptual tool on their second language and understand how a ‘thick’ sound is different than a ‘thin’ sound.

Vowels

Vowels can differ along a number of dimensions, but here we will look at just three: the positioning of the lips, tongue and mouth. Firstly, vowels can be rounded or unrounded, for example the word tea is pronounced with spread and narrow lips whereas two is pronounced with rounded, pouting lips. Secondly, we can move our tongues to change the quality of a vowel sound. The word tea contains a front vowel: the tongue is far forward. If we say tea and then two, not only do we have to round our lips, but we can also feel our tongue retracting slightly. Phoneticians describe the vowel in two as being a back vowel. If we start with the word tea and round our lips without moving our tongue backwards, then we have a front rounded vowel, more specifically we have the French word tu (you).

Another parameter along which vowels can vary is height, or rather how open our mouth is. If we start with two and round our lips without retracting our tongue as before we get the French word tu. Our mouths are very narrowly opened when producing these sounds- a mouth position phoneticians describe as close. If we then open our mouth a little wider, we will be producing a close-mid rounded vowel, like in the Swedish word tö. Try moving between these by starting with tea and then drawing out the vowel before adjusting the roundedness of your lips and the height of your mouth to create different vowel sounds.

Figure 1: Cardinal Vowel Quad

We can employ all these vowel parameters to give language learners instructions on new sounds. For example, to teach a learner how to say the vowel in the French word tu we could listen to their attempt and then give some pointers. “You need to round your lips” or “It’s a front vowel, so move your tongue forward” or “It’s a close vowel, so don’t open your mouth so widely”. Visual aids can help. The grid above is a sort of metaphor for the mouth. The phonetician Daniel Jones used this vowel quad to describe reference vowels against which a language’s vowels could be compared. The quad uses symbols from the International Phonetic Alphabet. Vowels appearing in pairs have the rounded version on the right and the unrounded version on the left. The vowel in see would be right at the top left (around the I symbol), the vowel in two would be at the top right (around the u symbol).  Quads like these take centre-stage in any introductory course in Phonetics, however, they can also prove an invaluable pronunciation resource for language learners when simplified and repurposed. We don’t need the phonetic symbols or annotations used on the quad above and we don’t need to place sounds in exactly the correct phonetically informed position as Jones did. Instead we can just use actual letters or particular words as examples and encourage learners to think about how some are back or front by pointing to their places on the quad. Below is an example vowel quad for a selection of long vowels in Swedish allowing us to see that ö is a rounded version of e. We can also see that both e and ö are front vowels and produced in opposition to the back vowel å.

Figure 2: Some Swedish vowels

There is a lingering danger that those of us pressing for phonetically informed pronunciation resources may be calling for one set of meaningless terms like ‘slender’ or ‘broad’ to be replaced by a set of technical but equally jargonistic set of labels. Of course, what Otto Jespersen wouldn’t have wished to see is language teachers adding to the stressful pronunciation scenario he described by yelling phonetic commands to learners who simply don’t understand them. In order, for Phonetics to be helpful in classroom settings learners need to be familiar with any phonetic terms or resources that are used.  Students have come to the language class to learn Welsh or German- they haven’t come for a graduate course in phonetic science. But to get our students speaking in a comfortably intelligible way and understanding the sounds of the language they are learning they need and deserve more than endless repetition drills and meaningless descriptions like ‘soft’ or ‘hard’. Language learners are intelligent people possessed of the human faculty of language; language is something they can do. Phonetics can help learners tap into this innate potential and appeal to what Jespersen called their ‘practical instincts’. So we don’t want our vowels ‘hard’ or ‘soft’ because this doesn’t mean anything. We want them front or back, rounded or unrounded, close or open.   

Abercrombie, D., 1949. Teaching pronunciation. ELT Journal, 3(5), pp.113-122.

Jespersen, O., 1910. What is the use of phonetics. Educational review.

A version of this text first appeared as an article in Languages Magazine:
Pulman-Slater, J. 2020. How do you like your vowels: soft or hard? Language Magazine, February issue, pp. 29-31 [click to go to issue]