Ever since Daniel Jones first devised a method of representing the cardinal English vowels, others have used similar systems to map languages.
His quadrilateral representation is the most well-known, but not necessary to the English language, unless we are to imagine a scientifically-accurate placement of every vowel. When we compare the various representations available on-line, it becomes clear, that whether to avoid copyright violations, or to attempt at being original, there are perhaps a dozen variations available, most of them minor. English could even be represented by a triangle, which is at the heart of one of the versions found in Jones’ book, An Outline of English Phonetics.
The basic idea was to show the open vowels at the bottom, the close ones at the top. On the left were the so-called front vowels, on the right, the ones pronounced near the back of the mouth. The following diagram gives some idea of the variations available:
All books dealing with pronunciation in our bibliography show tongue positions within a side view of the vocal system, helping one to determine correct tongue placement. Our book for French, by Suzanne Mercier, showed this for all the sounds of that language, but did not include a vowel chart of any type to complement its instructions on pronunciation.
A fellow member of the International Phonetic Organization, of which Jones later became president, was the Argentine D. Furman, who in 1940 published his Curso de fonética inglesa moderna, which was corrected and enlarged in the 1946 edition, taking into consideration suggestions made by Daniel Jones in a letter dated May 19, 1942, (n. p., in “Advertencia de la segunda edición“). Already in 1936, the system of the International Phonetic Organization was recommended in the official teaching programs of Argentina by a ruling (resolución) of May 4, 1936.
We will go into a bit of detail on Furman’s book, as their is nothing available about him that we could find on the Internet, except that the 1946 edition is available for sale at around 100 pesos (4 dollars), and the 1940 edition, for between 500 and over 1300 pesos. A couple of his English technical words have become quite rare.
Furman showed the most primitive representation of all, a letter on the left, a line, and a letter on the right, followed by a second and third line, but these, additionally, with sounds made in the center of the mouth, including the “u” of “sun”. He does the same for diphthongs, but only identifies one in the center [p. 11].
Admittedly, he did show some other variations. One is the vocalic triangle, which can be describe as a “V” shape inside of a box, with the letters, from top to bottom, i, e, and a on the left, and u and o on the right. These are described as the essential vowels, and are placed within categories of closed, medial, and open, “a” being the only one of the last of these categories, [p. 18]. Two pages later [p. 20], he mentions the primary and secondary cardinal vowels, together with additional items from Broad and Narrow Romic Notation; and gives a new chart, a variation of our model “H” in the image provided above, but the top is a lightly dotted line, covered by the roof of the mouth.. Except for the center sounds, all vowels are outside of the boundary.
Finally, on page 31, he gives something similar, with the roof of the mouth, but two curves, the one on the left being larger, almost half a semicircle, going over to the right, and met by a minor curve of less than half the size of the first, forming a sort of cut-off American football, pointing downward to the right. The purpose of this diagrm is to show the path which the tongue traces at what the book calls, in English diphthongs of the Old Formation. The words “Old Formation” were put into English by Furman, beside his Spanish subheading.
The idea of dividing the graph into 3 sections, as did Furman in his way, is one of two methods occasionally adopted, the other divides the quadrilateral into four: closed, semi-closed, semi-open, and open.
Another author, more prominently-listed on the search engine used, as he had the opportunity of availing himself of the Internet during his lifetime, was Francisco Sánchez Benedito. His first diagram is a variation, an upside-down equilateral triangle (type B of our illustration) superimposed upon a basic structure of 9 squares (3 x 3), labelled above and to the left. (p. 18). On the next page, he gives something like image “H”, labelled internally, with no references to rounded and unrounded. On the pages which follow, he repeats these charts – one for each vowel, but each showing the position of the Spanish sound as contrasted with the English. In our library of limited size, this was a novelty.
We begin now with a critique of the system, and with our suggestions for improvements.
The Quadrilateral as Shown by Jones 
In the 1972 edition of An Outline of English Phonetics [Cambridge University Press], there are certain images of the quadrilateral, but these are not named as such, and differ from the stylized versions found elsewhere, which take the shape of trapezoids [trapezia], with the base shorter than the parallel top, and a jutting angle prominent on the left. The images in the text, on the other hand, taper towards the left, resulting in an even more acute angle on the top left of the figure. It becomes a nest, as it were, for the simplified depiction as a triangle, usually the isosceles form. The caption, as close as it gets to saying anything concrete about the form, mentions “areas served”.
As a teaching tool for the English language, especially for the American version free of any French sounds, the triangle lends itself as the most practical tool.
It is not necessary to pretend at any attempt to rigorously map a supposed precise position in the mouth for any vowel. It is sufficient, for example, to suggest, that when one pronounces the vowel sound of “beat”, that if the tongue retains the same position, to the degree that such an action is possible, rounding of the lips will produce a different sound.
The student can experiment by varying the shape of the lips, slowly making them round, and then rounder. Some of the sounds which result will not be relevant to the language being studied, but the exercise can be useful in inculcating the learner to finding the correct tongue and lip placement.
In the same way, attempts can be made by pronouncing the open unrounded vowel, and then gradually closing the mouth. This would be especially useful when trying to teach the correct sound of the French of Spanish “e”, too often transliterated as “ay”, as in “Hey!”, or pronounced as in the “e” of the word just left behind.
The sounds to the right are less productive for English, but a learner, especially a young one, might find it amusing.
Why do we harp upon this?
Pronunciation in language teaching is on the whole is too much avoided, as far as this writer could see. An attempt to eliminate the worst sounds made by learners, has always been part of his modus operandi. This would make him lax, if we apply this criterion of Daniel Jones:
Those learning to speak a foreign language should begin their study by ear training [p. 61].
It might be useful, depending on the dictionary being used in a class situation, to teach the International Phonetic Alphabet at the same time, so that when a word is encountered for the first time in that reference book, the sound will be produced correctly from the first days of study. Exaggeration is not required, insofar as Jones has admitted that his pronunciation does not always correspond to his descriptions, for example, for the word “coarse” [p. 98]. Elsewhere, he speaks of an unstressed diphthong, such as found in “glorious”, appear to him as “rising” in tone. [p. 119, Para. 446b.]
Problems with the Representation
Once we decide that there is no strict geometrical accuracy to the diagram, especially for the English language, we can rest easy. Such may not always be the case.
Languages with a more complex vowel system, or with many diphthongs, often have these depicted inside the quadrilateral. This author has seen one version of the “u” of “fun” shown in such a central position, which was in no way in agreement with his own pronunciation, and is not borne out by other charts.
This is not to say that the letter (or any other) is pronounced in the same way in all places. This was noticed by the present writer in a 1991 Irish movie, The Commitments. The “u” had a quality similar to that of a type of “o” symbolized by the reversed “c” in the International Phonetic Alphabet.
This kind of internal mapping of sounds can be rather intimidating. In at least one of our reference books, at least the nasal sounds were shown separately. It was especially in a text for Brazilian Portuguese that the chart became cluttered up. [The book is no longer in the author’s library, and thus cannot be referenced.]
It also has seemed strange, that the sounds with the mouth almost closed somehow suggested themselves as being open (assuming that we are not paying full attention to the letter), because of their high position in the charts, and because the top line is longer than the bottom one. Such an opinion might have been formed because of the standard depiction being placed inside of the mouth of a cut-away head, to emphasize the position of the sounds.
It is to be noted that in the preceding paragraph, we have not used the idea of “close” vowels such as “i” in the linguistic sense (clarified in the paragraph preceding our solution), but in the instruction that the vowel sound of “cheese” is made by keeping the lips close together, and stretched wide, as is known to many persons from whom a smile needed to be elicited for a photograph, through the instruction, “Say ‘cheese’.”
Mercier makes it very clear, although Jones was not remiss in this regard, that this “i” is not the same in French as in someone’s native language. The reason is because of slight differences in the tongue position. We may take this as valid for any other vowel. We could also take it as a reason to begin communicating by text rather than by speech.
This is logical, insofar as it shows the correct position of the front vowels. However, an additional point of confusion enters, when we notice that the back vowels seem to be associated, by position, with the round vowels. This is so much the case, that an image which was prepared for this page is under review, to see if it has been properly labeled. It results from the vowel sound of the word “moon” being proximate to both the labels “back” and “round”.
The chart we have presented above shows that exposure to one of these may easily cause confusion when looking at a second one. We must ask ourselves if the primary vowels are shown on a separate chart from the secondary ones; or if the close vowels are represented with the round vowels on the left, or if it is the case that the primary and secondary vowels are on that side. We have found, confusingly, that Duden’s Grammatik correctly showed close and round vowels on the left, in two separate rows, but only gave one row on the right, 4 elements rounded, one unrounded. Here, only 16 vowels were represented, compared with 26 in the Duden Aussprachwörterbuch.
While Daniel Jones does not map the sounds as used in the United States, he does dedicate some pages to these. Unfortunately, he treats the American sounds as one, which is far from the truth. We do not know to which region he is referring, if any, but he must have some specific place in mind when he speaks of American English. That the American pronunciation of the letter “e” in “met” is represented by the IPA symbol “e” would make the “e” sound similar to that of French or Spanish, which is false as far as this writer is concerned. Furthermore, he sees nasalization, not only preceding the nasal consonants, but even after them. Were this true, this writer would be speaking French much better than he does.
D. Furman is also unclear as to what kind of English he refers to. His work refers to the “a study of the sounds of educated Londoners and citizens of the United States” (un estudio de los sonidos del londinense y estadounidense culto – Prefacio / [Preface]). We ask ourselves, in that Washington, Boston, New York, or where?
Most of the charts we have seen require the text to be read before using the chart. That is the only way sense could be made of them. For those unfamiliar with the system, the explanation of the vowels at the top is this: it is related to the tongue being raised as close to the roof of the mouth as possible, without touching it. On the other hand, the vowel “a” requires the tongue to be near the bottom of the mouth.
To summarize, we have the following objections to the traditional representations: intimidating mappings, confusing representation of the vowels with the mouth less open, and more confusion with the apparent proximity of the back vowels with rounded vowels. It would have been convenient to have a system that obviated the need for reading text in order to understand the charts.
The problem compounded itself even more, as an attempt was made to depict the sounds of a foreign language in a way that would be less confusing than in the texts used.
The author is not an artist, and after considering and discarding some ideas, nine more sketches were attempted, of which the tenth is shown here.
In its crude manner, it attempts to simplify the idea of front and back, rounded and unrounded. Front is what is closest to the viewer, back is – well, in the back of the image. Unrounded vowels are on the left side, which has been kept with right angles only. Rounded vowels are situated on the semicircular side. A crude attempt to show how much the lips should be parted, and in what shape, is reserved for the space between the front and back.
On the left, we show the letter “M”, because this is pronounced with the mouth shut, and is therefore a reference only, Moving towards the middle, the mouth gradually opens to its widest, until, as we jump over to the center left, we have a wide open “O”, becoming a small “o” for a pronunciation of a letter like “W”, which should not be a surprising representation, for the sound of the French “oui”, or the Semitic letters, ו [vav], and و [wāw] [Hebrew and Arabic, respectively], where both letters are used to represent the “o”, better have the “w” quality when pronounced as consonants. It will be necessary to click on the image to see the details, as the central elements have been marked in grey, to emphasize that they are internal to this case-shaped object.
While we have criticized some others for clutter, we feel we already have too much of the same, but we also believe that the learner can visualize and remember this more easily than the other systems. If this is not the case, at least a full mapping onto the correct amount of lip opening would be very useful. The “i” has been selected with a red dot, with a line leading to another dot, to the right of the “M”, to show that the lips are only slightly open. An orange dot connects an “a” to a the dot in an oval, which should have been rotated 90 degrees, but shows that the mouth is open, but not circular. The two green dots show a mouth wide open for a rounded sound, while two purple dots connect the “u” (representing the “oo” of “ooze” sound) to the rounded opening to the left of the “W”. (It has really been filled in with the purple color). Were all the vowels of a language to be learned represented on our proposed diagram, it would not be necessary to read much further.
In spite of Daniel Jones being quite scientific in some of his presentation, it is to be regretted that for the gap between the lips, the best he could do was to suggest, grosso modo, say, “narrow to medium”, or “medium to wide”. Of course, what is valid for an adult is not valid for a young child, but at least mature learners could have been given better directions. Some texts are more precise, and such information could be incorporated for a specific language.
On the other hand, we can be thankful to Jones for his treatment of other factors in the correct pronunciation, these being, amonst others, the height of the tongue (something we have referred to, mentioned in Duden’s Ausssprachwörterbuch, but in Jones, we take this as a separate element from how much the mouth is opened); location of the highest part of the tongue within the mouth; and a comment on the opening of the jaws. Perhaps these can be subsumed under the broad category of the opening of the mouth and its relative amount of rounding, or lack of it, but tongue placement is definitely important in the correct reproduction of sounds. It has just raised the question as to whether an “a” or “o” sound can be pronounced with the tongue in the position of an “l”, and how that would be represented. Apparently, this is not a concern for European languages.
The only difficulty with the chart we have presented, but not an insurmountable one, was the placement of the schwa; which had to be neither front nor back. The solution for such vowels is to have them on a level separate from the left-right M to W axis. Again, grey was used, as it is inside our “box”, but it can be quickly located by the yellow square, and clicking to enlarge.
We will now give two applications of the above system. In the first one, we do not necessarily see anything spectacular.
In the preceding diagram, one of Duden’s vowels was omitted, as it did not agree with other systems, but it was not necessary.
The diphthongs are to be understood as the sound of the letter where the arrow starts, connected with the sound of the symbol to which it points. Daniel Jones mentions approximately 21 diphthongs, not all of which are needed for practical usage. To prevent clutter, not all English diphthongs are shown. The German language manages with 3, as can be represented with the words rauh, Reihe, and reuen, represented by arrows from a to u, a to I, and ɔ to i, respectively.
Representing Turkish Vowel Harmony
The impetus for the creation of our chart was to graphically illustrate Turkish vowel harmony, in such a way that looking at the graph would facilitate memorization to a faster degree than that which could be accomplished by memorizing the information given in books. That is very well when humans will have chips built into their brains to speed up their mental processes, but for the moment, cheaper methods of information retention and retrieval are required. Here is the chart:
Using the above chart, the most important rule to remember, as it is difficult to draw, is that any vowel in a syllable can follow the same one in the syllable immediately preceding it. This has only been shown for the “ü” and “u”, by a circle, and a small arrow pointing to its boundary, but is valid for all eight vowels shown. For the remainder, one need only follow the arrows. With the back vowels, it is seen that the “a” can be followed by a “u” if the bilabials “b”, “m”, or “p” intervene, or the labiodental “v”. When applying the rule, it will be necessary to accomodate exceptions on account of foreign or compound words.
We can test some words from Turkish place names. Turkey is, in the language under discussion, Türkiye, so there is an apparent violation of the rule, in that “i” is not shown as following “ü” on our chart; but the “e” follows correctly, and all are front vowels. The capital city, Ankara, follows the rule perfectly. Republics ending in “-stan” speak Turkic languages, so we expect the rule to be applied, at least partially, in these other countries, viz. Afghanistan. Its capital, Kabul, follows the rule with back vowels – in this case, we may look at the “u” as following the “a” by the exception provided by the “b”, or through the fact that “bul” is a separate word, for which we do not care to risk any specific meaning at present. The personal name Abdul, though of Arabic origin, fits the rule, as does the verb for “answer”: cevap – a vowel change allowed because of the presence of the “v”. An outstanding example of vowel harmony is the name of the first of the named authors for our reference, Colloquial Turkish. The other author’s name may be an exception, as it is most probably of Arabic origin, if not, Greek (cf. Sinon).
This exercise should not be taken too seriously, since Turkish names may be transcribed in English, and do not respect the difference between the front dotted “i” and the one pronounced in the back, which is in Turkish an undotted “i”. These will not be found on Western-made maps.
No work similar to what has been done here has been found on the Internet, which does not mean that I have succeeded in being original. Hopefully, the strands which have been joined have allowed me this privilege.
Original versions: June 26 – July 2, 2018.
© 2018, Paul Karl Moeller.
Benedito, Francisco Sánchez, “Los Vocales”, Manual de pronunciación inglesa comparada con la española, [Madrid: Editorial Alhambra 1976], pp. 18 – 50.
Furman, D. Curso de Fonëtica Inglesa Moderna, Furman, 2a. edición [Buenos Aires: Buenos Aires (?), 1946].
Grebe, Paul, bearb., “Lautbestand”, Grammatik. 3. Auflage, [Mannheim: Bibliographisches Institut, 1973], p. 35.
Jones, Daniel. An Outline of English Phonetics, Cambridge University Press, 1972.
Mangold, Max, bearb., “Vokalviereck”, Das Aussprachwörterbuch, 2. Auflage, [Mannheim: Bibliographisches Institut, 1974], p. 27; “Lautbestand (Allophone)”, p. 31.
Mercier, Suzanne. Les sons fondamentaux du français. [Hachette, 1957]
Richard, Jack C., Platt, John, et. al., “Cardinal Vowels”, Longman Dictionary of Language Teaching and Applied Linguistics, [Longman, 1992], pp. 44-5; “Vowels”, p. 403.
Bayraktaroğlu, Arın and Bayraktaroğlu, Sinan, “Vowel Harmony”, Colloquial Turkish, [London: Routledge, 1992], pp. 13 – 15.
Lewis, Geoffrey, “Vowel Harmony”, Turkish, [Seven Oaks, Kent: Teach Yourself Books, 1989], pp. 11 – 13.
Note: The editor of Furman’s book is given as Buenos Aires on line. While such an editor may have existed, we suspect that the editor was not named, and this is merely the city of publication. Only a distributor and printer are clearly given. Perhaps, if the editor was “Buenos Aires”, it was felt that there was no need to give the place of publication.