I have read several articles that claim that phonemically, /j/ and /i/ are the same and distinguished from each other by being syllabic or not. What confuses me is that I can hear the difference between the English, "ye" /jiː/, and the "ee" sound /iː/. Or similarly, why we can distinguish a "w" sound from "womb" /wuːm/.

Are they always pronounced this way in other languages or can they truly be identical phonetically, and just be reliant on syllable structure? Are these pronunciations flipped around in any language?

  • They're not necessarily phonetically identical. For me, [j] is more closed than [i], so my tongue lowers slightly as I say [ji]. Same for [w] and [u].
    – Draconis
    Commented Aug 13, 2017 at 16:15
  • "/j/ and /i/ are phonemically the same" is categorically wrong. "[j] and [i] are allophones of the same phoneme in language X" could be correct, depending on the language.
    – curiousdannii
    Commented Aug 13, 2017 at 16:23
  • I think the question is a) can you reliably distinguish [i] from [j] from sound quality alone, absent syllabic context and timing information; and b) if so, does that hold for all languages with a /i/ /j/ distinction (= are there languages where the only difference between /i/ and /j/ are syllabic context and duration?) Commented Aug 13, 2017 at 19:25
  • @leoboiko Yeah, that is what I am asking. Also whether or not that holds up throughout languages (by which I mean some languages regard those two sounds as one sound) on account of sounds being sort of an approximate spectrum.
    – N.D.H.
    Commented Aug 13, 2017 at 19:35
  • 2
    also: this is orthogonal to actual phonetics, but as it regards speaker's perception of the phonemes, I think one has to consider the influence of the writing system. My hunch is that literate Portuguese speakers will tend to think of /j/ as just a "quick /i/" in part because both are written as ‹i› & learned as ‹i›s in childhood; whereas in English, which has orthographically distinct ‹y› and ‹w›, speakers are more likely to think of them as "different sounds". In Japanese it took a lot of debate to even establish the existence of diphthongs like /oj/, undistinguished from /oi/ in writing. Commented Aug 13, 2017 at 20:01

2 Answers 2


An article that claims that /j/ and /i/ are phonemically the same and distinguished from each other by being syllabic or not is either very confused about the concept of "phoneme", or doesn't know the difference between "phonemically" and "segmentally".

The standard late-SPE analysis of the distinction is that [j,w] are [-syllabic] and [i,u] are [+syllabic]: they have a difference in a feature. With the advent of CV phonology in Clements & Keyser 1983, the feature syllabic was exiled in favor of some kind of suprasegmental distinction, such as "is dominated by C vs V", or "Is dominated by 'N' (nucleus)", and then things got more complex in moraic theory. In fact, in moraic theory, "consonantal" was redefined to essentially mean "non-syllabic" (j, w, h are [+consonantal]), which is okay because as argues by Hume & Odden, there is no good evidence for the feature consonantal (as defined in SPE).

Under the CV and later but not moraic accounts, the difference is not in terms of a segmental feature, it is in terms of a prosodic property. If one incorrectly things of "phonemes" as being defined solely in terms of segmental features, it would be true that [j,i] and [w,u] are featurally identical. (Indeed, moraic theorists do not uniformly embrace the "let's call them [+consonantal])" account of glides a.k.a semivowels.

The question of being "phonemic" is a question about phonological contrast, and not about phonetic properties. There is great phonetic similarity between [i,u] and [j,w], but that doesn't mean that [j] and [i] phonemically contrast. It is an open question whether there is a real, hard-core contrast between glides and vowels (and much ink spilled on the topic).

Although I don't accept this analysis as correct, the examples "year" versus "ear" can be phonemicized as /iir/ vs. /ir/, so there is no phonemic contrast. The rule is (as a first approximation) that high vowels become glides before vowels. Those who deny that there is a phonemic contrast between vowels and glides must analyze all glides as coming from corresponding vowels, by some rule. (Logically, they could also hold that all high vowels come from glides, but nobody does that).

Levin in her dissertation "A metrical theory of syllabicity", and subsequent post-CV phonology theories have sort of eliminated the question of vowel / glide contrast from serious consideration, because the distinction is purely predictable from prosodic position: being the head of the nucleus means you are a vowel, otherwise you are a glide. The discussion then has to move to the issue of nuclear vs. non-nuclear status.

As for the phonetics of glides vs. high vowels, there's sort of a mixed bag out there. The two main tendencies are that a glide is shorter than the corresponding high vowel, and glides are more constricted. Sometimes you get other differences, such as that in Logoori, /j/ is (typically produced as, though optionally) a very fronted tongue-tip down approximant, unlike /i/.

  • So could there be a language where the [j] phoneme less constricted and longer than [i]? If so, how do we distinguish them from each other?
    – N.D.H.
    Commented Aug 13, 2017 at 20:54
  • @N.D.H. Suppose the language has syllable-level phenomena (for example, a pitch accent that attaches at syllabic boundaries, or all glides have coalesced in a certain dialect, or it counts moræ for something and onsets don't count, etc.) If we have language-specific criteria to delimit syllables, we can then distinguish [j] from [i] based on whether it occurs in the nucleus or onset/coda. If, on the other hand, the language has absolutely no clear criteria to distinguish syllables, then I think the distinction between "vowels" and "semivowels" doesn't seem very useful phonemically. Commented Aug 13, 2017 at 21:49
  • 1
    @N.D.H., if you have phonological evidence that tells you, that's what you rely on. If all you have is the phonetics, then you simply mis-transcribed everything and all of you "j"s are really [i] and vice versa.
    – user6726
    Commented Aug 13, 2017 at 22:35

Saussure proposed a phonetic theory of syllables according to which syllable onset sounds have increasing oral aperture for their duration (called explosive), syllable offset sounds have decreasing oral aperture for their duration (called implosive), and (Saussure doesn't say this, but it seems implied) syllabic nuclei, which have constant oral aperture. This is part of the discussion of the syntagmatic nature of speech in the Cours de linguistique general.

The charm of this theory is that it does not require a separate level or dimension of syllable position or metrical structure, since it makes sense on a purely articulatory level. That is not to say there is no relevant notion of structure other the stream of articulations in speech, since Saussure also describes a syntagmatic structure for syllables (but, if I understand him, that is not a separate dimension, but is rather another aspect of langue).

Personally, I do not understand how Saussure can fit syllables like [staps] into his theory, since aperture sometimes decreases in a syllable onset or increases in an offset. Accordingingly, I have proposed in other postings here a version of this theory with +/- stress replacing Saussure's explosive/implosive.

Theories with separate dimensions for articulation and prosodic structure must assume ad hoc principles for relating the two.

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.