I am learning Korean pronunciation, and find it reported that the distinguishing feature of Korean consonants such as orthographic ㅂ (b) and ㅍ (p) is aspiration. However, to my ears both (b) and (p) (and by extension, the pairs (d,t), (g,k), (j,ch)) sound aspirated, when spoken by my Korean friends. (I am Bengali, and am quite confident at detecting aspiration.) Now, the Wikipedia article on Korean phonology does mention that for many young speakers both are aspirated, and distinguished primarily by pitch patterns. My query, however, is that while I do acknowledge (and hear) the tonal differences, I also feel that (p) is aspirated more strongly than (b) --- similar to the difference between English (p) and Japanese (p), but definitely distinct from (p) in Bengali. Is there any study (references are welcome) on the degree of aspiration in Korean? And, if my observation is correct, would Korean be the only language to contrast degree of aspiration, that is to say, are there other languages that use such a parameter at a phonemic level?

[P.S.: I don't know how to put angle brackets to mark orthography, hence the parentheses surrounding the p's and b's.]

  • 2
    To clarify, even if both ㅂ and ㅍ are aspirated in Korean to different degrees (as you say you hear), this doesn't mean that Korean contrasts degrees of aspiration phonologically. The degrees of aspiration could be treated as a detail of the phonetic implementation of some other underlying contrast. However, it is an interesting question whether there are any analyses that make use of "degrees of aspiration" (of Korean or of any other language). Apr 30, 2017 at 5:55
  • 2
    This may interest you: jbe-platform.com/content/journals/10.1075/kl.13.01djs . May 2, 2017 at 9:52
  • 1
    @sumelic, that was exactly my point: If there are analyses that treat the Korean case as phonemic contrast between degrees of aspiration, especially considering (as user6726 mentions in an answer below) Korean has a third set of stops that are clearly unaspirated (albeit with "tenseness"). May 2, 2017 at 22:16

2 Answers 2


The notion "degree of X" really requires a three-way distinction to be valid, as in degrees of length (Estonian, Saami, Dinka), nasalization (Palantla Chinantec) or breathiness (Bor Dinka). If there are only two systematic values, we say "it is" or "it isn't", and we don't have to say "how much" (degree). In the present case, it is likely that you are using one language (Bengali) to set a standard of comparison for another language.

Korean has been subject to a half century of acoustic analysis, and it is clear that there has been recent language change where /p/ et al. in Seoul Korean have become "more aspirated", over time and compared to other dialects. It is now to the point that the unaspirated stops of Seoul Korean are comparable in voice onset time to the contrastively aspirated stops of languages like Apache and Khonoma Angami (Cho & Ladefoged). Some languages have extreme degrees of aspiration, for example Navaho [k] at 45 msc vs [kʰ] at 154 msc; 28 and 128 msc respectively for the same pair in Tlingit. From that perspective, p in Korean is "somewhat aspirated" and ph is "more aspirated, compared to Bengali or other Indic languages where the phonologically unaspirated stops don't have a very long VOT at all.

What makes Korean so special is that it has a third kind of consonant, transcribed typically as pp, which is termed 'fortis; tense; glottal; geminate' – all of these seem to be appropriate descriptions of the facts. These 'tense' consonants have an even lower VOT, which is the element missing for claiming that there are "degrees of aspiration:. However, the 'tense' consonants also have some kind of glottalization associated with them, at least for older speakers. It is possible that some time in the future, the p, pp, ph contrast will develop into a three-way difference in aspiration alone, but at present, there are enough other differences associated with these consonants that we can legitimately treat the system as having two orthogonal two-way differences (aspiration versus glottalization).

In one way of looking at it, English could provide another example of three-way aspiration differences (though not contrasts). It is well-known that aspiration of voiceless stops is governed by a rule where they are aspirated in foot-initial position and not aspirated elsewhere. This is often treated as a phonological (categorial) allophonic process. However, it has also been observed that for many speakers, the contrast between "g" and "k" does not involve voicing, it involves aspiration ("g" is really unaspirated [k] and "k" is aspirated [kʰ]). Putting these two facts together, one can say that phonetically we have three aspirations states: unaspirated ("g"), lightly aspirated ("k" not in foot-initial position), and more-aspirated ("k" foot-initially as in "cap")

There are plenty of ways to reduce the significance of these aspiration differences. One can maintain that the "k" / "g" contrast is indeed one of voicing and the failure of vocal fold vibration in phonetic outputs is the result of something about phonetic implementation; or one can say that that contrast is indeed phonologically one of aspiration, but phonetic implementation amplifies the degree of aspiration of foot-initial aspirated stops.

The situation with English is different from that of Korean in that English only has two kinds of underlying consonants as opposed to three in Korean, and "what is underlying" is one definition of "phonemic". But the other definition of "phonemic" refers only to phonetic outputs and the ability to reduce two phonetic categories to one by rule, referring only to available phonetic properties and not abstract phonological ones. Because of resyllabifications (an abstract phonological process), you can actually get all three stop types in comparable contexts: "a Benny" (unaspirated), "up any" (lightly aspirated), "a penny" (most aspirated). This is typically dealt with by referring to juncture / word boundary, but junctures or boundaries are not phonetic events, they are abstract phonological units. Since I don't subscribe to the strict surface-oriented view of the concept "phoneme", I don't know whether advocates of that theory would then hold that English also has "degrees of aspiration". This handout and paper by Vaux gives an overview of relevant facts and theoretical interpretations of the matter.

  • 1
    Thanks for the great answer. It all makes more sense now. Just to make it clear, I wasn't using Bengali to set the standard for anything. My point was that in other languages that are said to contrast aspiration that I have heard myself, including Mandarin, Icelandic, other Indo-Aryan languages (Hindi-Urdu, Kashmiri), and even ones that do so allophonically, like English and German, my ears always pick up what I interpret as a clear lack of aspiration contrasted with its presence. So, the case in Korean struck me as somewhat "odd." May 2, 2017 at 22:13

I am bilingual in Korean and English. I have always held the view that the difference between 'ㅂ' and 'ㅍ' is not aspiration -- they are both aspirated as you have observed. However, the uniqueness of the 'ㅂ' is that it is a voiced, aspirated, bilabial plosive as oppsed to the unvoiced 'ㅍ', which is 100% identical in sound to the English 'p'. However, no such equivalent sound exists in English for the Korean 'ㅂ', and neither is there a Korean equivalent for 'b'.

This idea of voiced and aspirated consonants also occurs in the Korean 'ㄱ','ㅈ','ㄷ', and 'ㅅ'. If you listen carefully, you will notice that they do not sound like the English 'g', 'j', 'd', and 's' consonants. Here is my classification of various Korean consonants:

ㄱ - voiced, aspirated, velar, plosive

ㄷ - voiced, aspirated, alveolar, plosive

ㅂ - voiced, aspirated, bilabial, plosive

ㅅ - unvoiced, aspirated, alveolar, fricative

ㅈ - voiced, aspirated, palatal, approximant

If it is difficult to accept that aspiration can occur simultaneously with voicing, then I would propose that these consonants are unvoiced and aspirated with the difference being that the voicing comes much sooner than their normally unvoiced counterparts (the English 'k', 'd', 'b', 'j' sounds).


  • 1
    Korean ㄱ,ㄷ,ㅂ are articulated unvoiced at word initial position, so voice cannot be the defining feature of these consonants.
    – Ignatius
    Feb 15, 2019 at 23:33
  • @Taegyung: I wonder whether Justin is referring to the pitch pattern of the following vowel. I've heard that the Korean consonants ㄱ,ㄷ,ㅂ are supposed to affect the pitch of a following vowel in a similar way to voiced consonants. Feb 16, 2019 at 1:48

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.