As we have seen, “vowel” is a manner of articulation. Accordingly, there are no manner distinctions in vowels. As for places of articulation, vowels may be labelled palatal or velar. Actually, front and back, respectively, are more common terms used for these two categories in vowels. Labiality in vowels is referred to as rounded. However, the most important property of vowels is tongue height or the openness of the mouth.
Vowels may be distinguished by how open the mouth is during their pronunciation. The a, eg in start, is very open, the i, eg in geese, and the u, eg in goose, are quite close. The wider opening of the mouth is accompanied by the lowering of the tongue, therefore a is an open or low vowel, while i and u are close or high vowels, these two pairs are synonymous, the first of each refers to the mouth, the second to the tongue.
The close/high vowels i and u are different in two respects. In the pronunciation of i the highest point of the tongue is below the palate, this is the frontest possible location for the highest tongue position. The vowel i is thus front or palatal. On the other hand, u is back or velar (again synonymous terms). The shape of the lips is also different in these two vowels: for i they are spread or unrounded,If you say cheese, you spread your lips and thus look as if you were smiling. for u they are rounded. Lip rounding usually accompanies back vowels, while front vowels are pronounced with spread lips. This is a universal markedness tendency: any human language is expected to have liprounding for front vowels only if that language also has unrounded front vowels, and vice versa, unrounded back vowels are only expected to occur in languages that have rounded back vowels too.
So some languages have only front unrounded vowels, others have these and front rounded vowels. The rounded pair of i is spelt ü in eg German or Hungarian, the IPA symbol for it is y. Similarly, in some languages the back rounded u has an unrounded counterpart, symbolized as ɯ by the IPA. In Turkish script, for example, this sound is represented by the letter ı (dotless i).
Vowels may not only be either high or low, some vowels are between these two edges. Languages differ in whether they have one or two (some even three) vowels between the highest and the lowest vowel. For a start, the IPA provides two symbols between i and a: e and ɛ. That is, i is high/close, e is high-mid/half-close, ɛ is low-mid/half-open, and a is low/open. These four vowels are called cardinal vowels #1 to #4, respectively. They are shown in the diagram to the left with the tongue positions associated with them.
The same four-step scale is posited for back vowels, giving us four more cardinal vowels, #5: ɑ, #6: ɔ, #7: o, #8: u. One thing to note is that while all of the front vowels above are unrounded, three of the back vowels, u, o, ɔ, are rounded. This is not surprising: as we have already seen, front vowels are unrounded in the default case, back vowels rounded. The low back ɑ, however, is not rounded. Rounding is not really significant for low vowels, languages rarely distinguish two low vowels that only differ in liprounding. In spite of this, as we are going to see soon, the IPA provides symbols for rounded low vowels too. The other notable thing is the difference between a and ɑ. These two symbols are needed if we distinguish between front and back low vowels.Not all phonologists agree with this though, some would argue that the low vowel is not sensitive to the front vs back difference.
The positions in these diagrams were taken from x-ray photos. The first such photos from 1917 are shown here. The black dots are the image of a lead chain the person in the photos has half swallowed for the sake of the experiment. This shows clearly the shape of their tongue. The red dots indicate the highest point of their tongue.
The eight cardinal vowels are shown in this chart together with their conventional numbers. The arrangement you see here is called Jones vowel chart after Daniel Jones, who had x-ray photos made of his head while pronouncing different vowels. (Yes, you have just seen his skull and jaw!)
We have seen that it is not only along the high–low and the front–back scales that vowels vary, but liprounding may also be distictive in them. The rounded counterpart of i is y, the unrounded counterpart of u is ɯ. As you may already suspect, each cardinal vowel has a counterpart which is rounded for the front vowels and ɑ, and unrounded for the rest of the back vowels. This set of eight vowels are called secondary cardinal vowels. They are shown here along with their primary counterparts, which are set in boldface. The first of each pair is unrounded, the second rounded.
You may recall that we said some languages distinguish as many as five vowel heights. To be able to transcribe so many vowels, the IPA provides a further symbol between ɛ and a: æ. Some languages also distinguish between two kinds of i’s: the “normal” i and another one, which is more centralized, ie more towards the centre of the vowel chart. The latter type of i is labelled lax, its symbol is ɪThis symbol not only lacks the dot, it is also a capital letter of a smaller size, technically a small capital., while the “normal” i is tense. The same dicothomy is also available for the back high u: tense u contrasts with lax ʊ (sometimes visualized as ɷ). Furthermore, similarly to the nonhigh nonlow vowels, there are vowels that are neither front, nor back. These are called central. The best-known central vowel is in the middle of the vowel chart, and is called schwa: ə.
Finally here is the Jones vowel chart with all the vowels we have talked about so far, and some more central vowels.
Since it leads to a somewhat different result, we will briefly describe an acoustic analysis of the vowel space too.
In acoustic terms vowels are described by their formants. Without going into too much detail here, we may say that a complex sound — like a vowel — is made up of a series of resonances at various frequencies. These frequencies can be shown on a spectrogram, like the one below. There are frequency ranges where resonances gather in different vowels, these are their formants. These are the darker grey areas, labelled as “F1” and “F2” in red in the spectrogram below.
The lowest frequency range is called the first formant (F1), the second lowest the second formant (F2), and so forth. The first formant of high vowels is lower than the first formant of low vowels. The second formant of front vowels is higher than the second formant of back vowels, or at least the first two formants are further apart in front vowels than in back vowels.
low F1 = high vowel, high F1 = low vowel
low F2 = back vowel, high F2 = front vowel
Lip rounding modifies the relationship of F2 and F3, r-colouring (see below) is a result of lowered F3, but this much will suffice for us.
An alternative vowel chart
The Jones vowel chart shown above is organized by the articulatory properties of vowels. However, based on their formants of the same vowels, another, quite similar chart can be produced, this is shown below.
In this chart there is a single low vowel, a, other low vowels appear higher (ie their F1 is lower). We can also see that both front and back rounded vowels have a lower F2 than their unrounded counterparts. This means that a front rounded vowel is acoustically closer to a back unrounded vowel than to a back rounded vowel, or a back unrounded vowel to a front unrounded vowel. Being acoutically closer means they are more confusable.Hungarians render Russian ɯ (spelled ы) as y. This is weird in articulatory terms: these two vowels are very different from that point of view. We see that acoustically they are much closer. This is why front vowels are typically unrounded and back vowels rounded: in this way vowels are more difficult to be misheard.
Other vowel contrasts
We have seen that vowels may be high or low (or somewhere in between), they may be front or back (or in between), and they may be rounded or unrounded (which strongly correlates with their being front or back). There are some further categories in which two vowels can be different
During the default pronunciation of a vowel, the velum is raised and air does not escape through the nose. If the velum is lowered and air does go through the nose too, the result is a nasalized vowel. The IPA marks this by a tilde diacritic above the vowel symbol, eg ã is nasalized a. In many languages vowels are nasalized before a nasal consonant (eg m, n), especially if this nasal consonant is not followed by a vowel (eg ant is pronounced as ãnt). It also occurs that it is only the nasalization of a vowel that distinguishes two words, eg French faitfɛ ‘fact’ vs finfɛ̃ ‘end’, beaubɔ ‘nice’ vs bonbɔ̃ ‘good’, etc.Some varieties of Hungarian also have such contrasts: úszoluːsol ‘swim-2sg’ vs unszolũːsol ‘urge’, elcseszɛlʧɛs ‘screw up’ vs elcsenszɛlʧɛ̃s ‘steal-2sg’.
Another possible modification to vowels is retroflexivization, ie curling the tongue back, during the pronunciation of the vowel. The vowel so pronounced is said to be r-coloured or retroflex or rhotic. This is not a very common way of pronouncing vowels, we mention it though because some accents of English, the best-known of which is General American,GA for short, this is the most widespread accent in the US and parts of Canada. have such vowels in eg nursenɚs, startstɑ˞t, northnɔ˞θ. Pairs where only rhoticity is the difference are not many: budbəd vs birdbɚd, papɑ vs parpɑ˞.
Vowels may also be distinct in their length relative to each other. Vowel length is often accompanied by other differences: English hathat and hearthɑːt have vowels that are different in quality, as well as quantity (= length).The same situation can be found in Hungarian hathɑt ‘six’ vs háthaːt ‘back’, though in this case the short vowel is back and the long one is front. There are pairs that are very similar, practically the same in quality, thus only differ in quantity, eg terrortɛrə vs tearertɛːrə, Siriussɪrɪjəs vs serioussɪːrɪjəs.Zugzug ‘nook’ vs zúgzuːg ‘buzz’ in Hungarian.
A diphthong is a vowel that changes its quality from beginning to end. Diphthongs are usually transcribed by two vowel symbols, the first representing the starting point of the diphthong, the second its ending point. For example, the word eye could be transcribed ai, while owe could be əu.
Depending on the relationship of the two halves of a diphthong, we can distinguish closing and opening diphthongs. In a closing diphthong the end half is a close vowel, in an opening diphthong it is an open vowel. There also are centring diphthongs, these end in the central vowel, ə. Current British English, as well as General American, have closing diphthongs — like ai in eye or fly and au in now or cloud — and opening diphthongs, like ia in yank or ua in wax.Note that opening diphthongs are usually transcribed as ja or wa. Also note that old-fashioned transcriptions of BrE contain also centring diphthongs in words like nearnɪə or curekjʊə.
Either the beginning or the end of a diphthong is syllabic: it constitutes the centre of the syllable. If the first part is syllabic, we talk about a falling diphthong — like ai or au — if the second part, it is a rising diphthong — like ia or ua. That is, the end of the word say is a falling diphthong, the beginning of the word yes is a rising diphthong.
The nonsyllabic part of a diphthong is called an offglide (in falling diphthongs) or an onglide (in rising diphthongs). It is very difficult to tell if an offglide or an onglide is a vowel or a consonant (a glide, recall, we labelled j and w glides earlier).Some experts call glides semivowels or semiconsonants. The following table shows three different ways of transcribing these sequences.
The British tradition is to transcribe offglides as vowels, and onglides as consonants (kau, sei, but waks, jes). American transcribers often use the consonantal symbol for both onglides and offglides (kaw, waks, sej, jes).In the American tradition the symbol y is often used for the glide j. Practically nobody uses the nonsyllabic vowel symbols of the middle column, probably because they are technically too difficult to produce. We are going to transcribe offglides by the consonant symbols.