In the languages I'm familiar with I can't think of any cases of semivowels other than the "w" and "y" sounds /w/ and /j/. So are there any others and if so, which are most common beyond these two?

  • I think the nasals, liquids, and fricatives are also sometimes considered semi-vowels.
    – Cerberus
    Oct 30, 2011 at 1:19
  • 10
    @Cerberus No, those are not semi-vowels although they can be syllabic nuclei. In fact, semi-vowels are the opposite, they're vowel (or vowel-like) sounds used on the edge of a syllable, so not acting as syllabic nuclei. Oct 30, 2011 at 2:39

4 Answers 4


You listed:

  • [j] which is a palatal approximant, seen in words like you [ju:] (English);
  • [w] which is a labio-velar approximant, seen in words like "weep" [wi:p] (English). Following a comment request: It's a labio-velar because it's articulated with the back part of the tongue raised toward the soft palate (called also "velum"velar) while rounding the lips.

There seem to be two others:

  • [ɥ] which is a labio-palatal approximant, seen in words like "月 yuè", [ɥœ˥˩] in Mandarin or, choosing a "closer" example, in "bonne nuit", [bɔnnɥi] (French);
  • [ɰ] which is a velar approximant, seen in the Korean word for doctor, 의사 [ɰi.sä]
  • @Otavio Macedo: Thanks, your correction was actually needed in the other items too, so I added it. :)
    – Alenanno
    Oct 30, 2011 at 10:07
  • 1
    The mixing of [ɰ] and [ɣ] is very confusing for me. I'm familiar with both Spanish and Korean and just don't know how to comprehend the Wikipedia articles. Oct 31, 2011 at 7:40
  • 5
    @hippietrail It is by no means Alenanno's fault; the Wikipedia article is flat-out wrong. Whether or not there is an alternation in Spanish, [ɣ] is not a semivowel. I suppose this is why wikis are editable; I'll go correct it now.
    – Alek Storm
    Oct 31, 2011 at 17:52
  • 3
    @hippietrail Even linguists get confused sometimes :). The IPA is updated periodically with new symbols for sounds that have become "common enough" to warrant a special designation - the labiodental flap [ⱱ] is the most recent example. Before this happens, however, researchers describing the language will often reappropriate a different symbol to stand for the new sound, as long as there is no confusion within the paper. Often, their transcriptions are cited in other works without describing the notation, leading to a great deal of confusion.
    – Alek Storm
    Oct 31, 2011 at 19:21
  • 2
    @NikhilBellarykar Added a short explanation. :) If it's still unclear I can expand it.
    – Alenanno
    Feb 6, 2012 at 9:54

The behavior with semi-vowels has to do with their level of sonority. Typically, a vowel is the peak of sonority in a syllable. Semi-vowels are vowels which are not the peak of sonority because they precede and/or are followed by a more sonorous (open) vowel.

In principle, any vowel could become a semi-vowel if it preceded or was followed by something more sonorous (think less constriction) than it was, perhaps with the exception of low vowels (because they are the most sonorous sounds). In reality, I am not sure about the existence of mid semi-vowels, outside of the off-glide in falling diphthongs as in /aʊ/. Alek Storm listed the semi-vowels for the high vowels. Any other types of semi-vowels would probably not have distinctive symbols in the IPA because of their rarity.

It is also worth noting that it is not mandatory that a vowel become a glide if it is precedes or is followed by something more sonorous. A language might instead treat each vowel as the nucleus of its own syllable. It depends on the language's syllabification rules.

As a side note, vowels need not be the nucleus of a syllable: anything which is more sonorant than the surrounding segments can be. Languages vary on what types of segments they will allow to be syllabic nuclei, but Berber famously allows fricatives and stops: tf.tkt 'you sprained'.


This ties into how one marks up diphthongs and polyphthongs. Strictly speaking you can add a non-syllabic marker (like the bow /̯̯̯/, or the more-open diacritic to closed vowels or the more-closed diacritic to an open vowel) to any vowel to produce a semi-vowel suitable for a diphthong, if how you define diphthong is a semivowel preceding or following a vowel.

As usual for IPA, frequent/common semivowels get their own symbols, which is why we have /j w ɥ ɰ/.

(Pet peeve: IPA isn't very good at showing diphthongs at a phonetic level, or rather, there seems to have been insuffcient interest in typological comparisons of diphthongs such that there is no agreed upon way of marking the difference between for instance English /aj/ in "hi" and Norwegian /ɑˑi̯/ "hai" or what have you.)

  • Hm, fascinating how different browsers show the ipa-symbols included above. YMMV
    – kaleissin
    Oct 30, 2011 at 16:43
  • 2
    How does making a close vowel more open, or an open vowel more closed, produce a semivowel? And what is '...a semi-vowel suitable for a diphthong'? I don't think anyone defines a diphthong as a combination of semivowel and vowel. A diphthong is a monosyllabic vowel with two phonetic vowel targets. A semivowel is a vowel articulation used as a non-syllabic segment. Oct 31, 2011 at 3:38
  • 2
    @kaleissin A more open [i] would be [ɩ] or [e]. A more open version of a vowel is another vowel. It is sometimes said that the [j] approximant has a narrower aperture than its equivalent vowel (a smaller aperture means less vowel-like). Do you have a ref for that definition of 'diphthong'? I think in linguistics it's pretty much agreed that diphthongs are a movement from one vowel target to another, occurring as a single syllabic nucleus. Or maybe the other way around is better: this kind of phone is given the name 'diphthong' in linguistics. Oct 31, 2011 at 9:19
  • 1
    [ɩ] is no longer proper IPA. Btw, how would you mark something that is inbetween /i/ and /e/? The point is, not all people use the symbols the same way. There are several traditions in phonetic transcription, they use partly overlapping symbols and they use them differently. You can't say that X is always so-and-so without knowing what tradition/branch it comes from, that goes for plenty other sciences too.
    – kaleissin
    Oct 31, 2011 at 12:08
  • 1
    @kaleissin Yes you're right, sorry, I didn't even notice myself that I had used lower-case [ɩ] instead of upper-case [ɪ], my mistake. Standard IPA (ie when not working within one of the traditions with its own transcription practices) is very strict about how the symbols are used. To mark something between [i] and [e] I would use a lowering or raising mark on one of those symbols or maybe [ɪ], depending on the acoustic analysis. But that's got nothing to do with indicating a semivowel. Oct 31, 2011 at 20:49

Wikipedia: Semivowel has a useful list, which adds some more details to Alennano's answer. In addition to j, ɥ, ɰ, w, it says that some authors list ɹ and ɻ as semivowels.

Your Answer

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

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