I am studying speech recognition by Lawrence Rabiner's book. I am unable to find a proper and easy to understand answer for the following question :

Difference between production of vowels, diphthongs and semi-vowels

The author has just stated a single statement about their difference like
Vowels : Vowels are produced by exciting fixed vocal tract with quasi-periodic pulses of air caused by vibration of vocal cord
Diphthongs : Produced by varying the vocal tract smoothly between vowel configuration appropriate to diphthongs
Semi-Vowels : Characterized by a gliding transition in vocal tract area function between adjacent phonemes.

Can I get some more points ?

  • You want to understand the terminology, the acoustics, or the biological apparatus? Sep 7, 2012 at 8:15
  • Basically i am preparing for exams so it would be nice to have difference in both acoustics as well as the biological way so that i can write maximum points. Sep 7, 2012 at 8:28
  • Maybe expand your question a little. I'll edit my answer to say that I'm only really tackling the terminology aspect. Some other people here will bring some great acoustics stuff I bet. Sep 7, 2012 at 8:35
  • 2
    a related question
    – user483
    Sep 7, 2012 at 12:28

3 Answers 3


Assuming you know what vowels are, diphthongs are phones which may be described phonetically as a sequence of two vowels (or vowel targets) but which function (phonologically) as a single vowel phoneme in a given language. Semivowels are phones which are phonetically vowels (or vowel-like) but which function as a consonant phoneme in a given language. The distinction between phonetics and phonology is crucial as phonetically identical phones may function as a vowel phoneme in one language, and as a consonant in another. As well, a phonetically identical sequence of two vowels may function as a diphthong (ie single phoneme) in one language and as a sequence of two vowel phonemes in another.


I'm just going to tackle the "terminology" aspect of your question:

  • Diphthongs (and triphthongs) are a kind of vowel. You can contrast them with "pure vowels".

  • Pure vowels have a single "target" where diphthongs (etc) have two (or more) targets. So for the English vowel sound in "pie" the first target is like the "a" in "car" and the second target is like the "ee" in "fee". This seems to be unintuitive for many people not acquainted with linguistics and could be compounded in English where the spelling of diphthongs often doesn't reflect its components.

  • Semivowels can function as either vowels or consonants, depending on their context (what comes before and after them).

  • @hippietrail----great man. How do you understand all these things so nicely ? I never thought of this example. Could you suggest some good books? Sep 7, 2012 at 8:31
  • @hippietrial-- can i see speech recognition as a career, i am interested in it but find it very difficult? I cannot up-vote your answer since i don't have 15 repution. Sep 7, 2012 at 8:33
  • I dunno. I became fascinated when I was about 10 years old I think. In Australia we followed the British system of IPA in dictionaries where each phoneme had one symbol and each diphthong had a pair of symbols. It seemed odd so I started analysing. That's nerdy kids for you (-: Sep 7, 2012 at 8:34

In an addition to hippietrail's great answer, here's the visual representation of English diphthongs as it appears here:

Diphthongs chart


/eɪ/ as day, pay, say, lay
/aɪ/ as sky, buy, cry, tie
/ɔɪ/ as toy, boy
/ɪə/ as beer, hear
/eə/ as bear, pair, hair
/ʊə/ as tour, poor
/əʊ/ as phone, no, go
/aʊ/ as how, cow

As a side note, an automated speech recognizer should be able to determine at which point of the chart the speaker's voice stays at a given moment. Having this in place, detecting diphthongs is no more difficult than finding a gesture on a 2D map, and there are plenty of effective algorithms for it.

  • @ bytebuster----- But the book from which i am studying it says that there are 6 diphthongs in American English, they are Sep 8, 2012 at 11:57
  • (1) /a^y/ ( it is actually 'a' raised to 'y' without the symbol '^', but i don't know how to write that here) (as in buy) Sep 8, 2012 at 11:58
  • 1
    @R122 The book you are reading uses semi-formal notation for the sounds. The map above uses IPA, a phonetic alphabet. Although it may look hard to learn for the first time, it is very unambiguous in terms of indicating specific sounds. Sep 8, 2012 at 12:10
  • 2
    @R122 The vowel diagram is universal as it indicates human vocal tract; most people can't produce anything beyond this map. The points indicate specific vowels for English. The arrows are also specific to English as they indicate diphthongs that are bound to English vowels. The symbols, as I mentioned before, belong to IPA phonetic alphabet. Sep 8, 2012 at 12:13
  • 1
    @R122 The problem with semi-formals is that they are different everywhere, you can't re-use your knowledge. It may be difficult to realize that "boy" can be [bo^i], [bo:y], [boʲ], [boi], or anything else. Sep 8, 2012 at 12:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.