贏 doanh

Why does the Vietnamese pronunciation of this character start with <d>? From what I have read, this letter represented a /ð/ in Middle Vietnamese. Reconstructions claim that the Middle Chinese pronunciation was /jiajŋ/ (as reconstructed by Edwin Pulleyblank) or /jĭɛŋ/ (as reconstructed by Wang Li). How do you explain the borrowing of the /j/ sound as /ð/?

  • d in Vietnamese represents /j/ in Southern and Central dialect and /z/ in Northern dialect Aug 23 '15 at 5:07

I found this answer satisfying: /j//dʲ//z///j/. Edwin G. Pulleybank wrote in Some Notes on Chinese Historical Phonology:

Vietnamese d and v are also partly derived from earlier *j and *w respectively as we can see from other comparisons with Mương (Thompson 1976). This accounts for the fact that d corresponds regularly in Sino-Vietnamese with LMC j (initial yu 喻 Grade IV). Presumably Vietnamese *j, including Sino-Vietnamese j, had become in the seventeenth century when the missionary orthography was invented. The phonetic change is comparable to that by which Latin Jacobus has given Diego in Spanish. After this change *j merged with from softened t. Some modern dialects preserve this pronunciation but in Hanoi the pronunciation is z and in Saigon it is j whatever the origin. It is likely that this is the result of a secondary loss of occlusion since the seventeenth century rather than a continuation of earlier *j.

I thought the comparison to Latin JACOBVS giving rise to Spanish DIEGO was rather brilliant, (everything that follows is an edit/addendum…) although as noted in one of the comments below, Diego actually comes from Sant+yego.

That isn’t to say that it didn’t happen that way. Scottish Gaelic underwent a fortitioning process:

  • /v/ → /p/ Scots vervane, werwane ‘vervain’ → bearbhain /pɛɾavɛɲ/
  • /ʍ/ → /kʰ/ Scots quhel ‘wheel’ → cuidheall /kʰujəl̪ˠ/
  • /w/ → /p/ Middle English wall → balla /pal̪ˠə/
  • /f/ → /p/ Latin fundus → bonn /pɔun̪ˠ/ (foundation)
  • /θ/ → /t̪ʰ/ Norse þrǣll → tràill /t̪ʰɾaːʎ/ (slave)
  • /h/ → /t̪ʰ/ Scots hogsheid ‘hogshead’ → tocasaid /t̪ʰɔʰkəs̪ətʲ/
  • /j/ → /kʲ/ English yawl → geòla /kʲɔːl̪ˠə/

The connection between /d/ and /j/ can also be found in the Japanese languages. In the book Is Japanese Related to Korean, Tungusic, Mongolic and Turkic?, Martine Irma Robbeets writes:

…the assumption that proto-Japanese had voiced obstruents is stronger in the case of pJ *b and pJ *d due to some modern southern Ryukyu dialects that have b- and d- corresponding to OJ w- and OJ y-. As for the reconstruction of pJ *d, the dialectal evidence rests entirely on a single dialect Yonaguni, which reflects initial OJ y- as d-. The dialectal evidence is rather weak because the Yonaguni d- is restricted to the initial position. Moreover, d- also appears in borrowings from Chinese such as Yo. dasai ‘vegetables’ which reflects a fortition y- > d-. Therefore, I doubt the antiquity of Yonaguni d- and I am reluctant to reconstruct pJ *d- as an ancestor for OJ y-.

Later on, Robbeets writes:

An argument contra the reconstruction of pJ *y is that from the cross-linguistic viewpoint a fortition of a palatal glide to a dental stop is an unnatural sound change, but Miyake mentions a similar development from pre-Middle Vietnamese *y to Middle Vietnamese *d.

It is far more likely, then, that the fortition of /j/ to a dental happened via some intermediate fricative or affricate. A similar fortitioning process changed the pronunciation of some words borrowed from Chinese from a sibilant to dental stop:

  • (/sam/) → <tam>/<tâm>
  • (/sim/) → <tâm>/<tim>
  • (/ɕiaj/) → <thế>/<thá>/<thé>/<thể>
  • (/ɕiaw/) → <thiểu>/<thẹo>/<thểu>/<thỉu>/<thiếu>/<thẻo>
  • 使 (/ʂɨ/) → <sứ>/<sử>/<sửa>/<thửa>
  • (/siajŋ/) → <tính>/<dính>/<tánh>

If /ð/ was indeed one of the former pronunciations of <d> as commonly stated, then the sound change might have gone the route: /j//ðʲ//dʲ//z///j/. The fortition was at least strong enough for the missionaries to assign the phoneme to <d>. Of course, none of this is for certain and anything without citation is just my pure conjecture.

  • 1
    Actually, I'm suspicious about the "Diego" example (I might ask a separate question about it!) but the part about Vietnamese seems pretty convincing. Jun 15 '17 at 19:31
  • To elaborate, I'm not aware of any other example of Latin /j/ > Spanish [dj] word-initially. Normally it became /ʝ/; in some cases (especially before back vowels) it became /x/ (from devoicing and backing of earlier /ʒ/). Wiktionary connects "Diego" to the form "Santiago," and while that seems to have problems of its own (I don't think the voicing of the "t" is regular) it seems more plausible to me. Jun 15 '17 at 23:14
  • However, the parallel change is regular in Latin -> Italian (Giacomo), or am I missing something?
    – Colin Fine
    Jun 16 '17 at 16:01
  • @ColinFine: Latin /j/ > Italian /dʒ/ word-initially is regular, as far as I know, but I wouldn't say that's exactly parallel to the supposed Spanish change. Italian "gi" /d͡ʒ/ is standardly analyzed as a single phoneme, while Spanish "di" /dj/ is standardly analyzed as a sequence of two phonemes. Jun 16 '17 at 19:24
  • Nice edit! Another thing this reminds me of is the change of various palatalized consonants in Proto-Greek to Classical Greek /tt/ in certain dialects. The voiced equivalent is thought to have been pronounced /zd/, which didn't become /d/ in any variety of Greek as far as I know, but it seems like it wouldn't be too big a change. Jun 22 '17 at 14:01

It may be instructive to note that Sino-Vietnamese readings are not the only Sinitic-based Vietnamese lexemes loaned into Vietnamese: there exist earlier substrata of Chinese loans (compare 呉音 go-on and 漢音 kan-on for Japanese kanji, where 漢音 is the "canonical" Tang China loan). Additionally, there are indications that the Sino-Vietnamese are not strictly limited to the Middle Chinese of the Tang court, but perhaps later borrowings.

However, going from Middle Chinese /j/ to Middle Hanoi Vietnamese /ð/ to Modern Hanoi Vietnamese /z/ is well-attested. Middle Chinese 喻母 /j/ with 四等 (i.e. a particular distinction in chongniu) generally went down this route. Compare 遙 diêu, 引 dẫn, 鹽 diêm, 異 dị.

Fortition of /j/ is not exactly uncommon, often via [ʝ] to [ʒ]. Spanish has been going through this change at the moment, and it is (currently) most salient in the Rioplatense region.

The real outlier though, is that 贏 was division III in Middle Chinese, so it "should" have become a h- or v- in Vietnamese! How it ended up reallocated under a different division is an interesting question, yet to be explored.

If you think 赢 > doanh is weird, look up what happened to 民!

  • a lot of syllables start with m in Chinese has become d in Vietnamese: 妙 diệu, 名 danh, di 彌 (in Amitābha 阿彌陀佛) Aug 29 '15 at 17:03

Lưu Vĩnh Phúc has explained correctly how the phoneme written as "d" is realised in modern Vietnamese dialects. But in the context of this question we also need to ask why the Jesuit missionaries who invented Vietnamese Latin script in the 17th century decided to represent this phoneme as “d” rather than as something else. Modern scholarship seems to favour the view that in 17th-century Vietnamese (sometimes called “Middle Vietnamese”) "d" did represent some sort of voiced dental stop. Of course, we do not know exactly how it was pronounced any more than how exactly Middle Chinese was pronounced. Comparative linguistics is mainly about establishing regular sound correspondences rather than about discovering the physiological rationale of these correspondences.

To return to the original question: it is not, I think, so much about how to “explain the borrowing of the /y/ sound as /ð/” as about an internal development in Vietnamese which led to the evolution of borrowed [j] first to a dental stop and then to [z] in some dialects and [j] in others. The big question is whether the [j] in Southern dialects is really a retention of an original [j] – specifically in Chinese loan words – or whether it represents a secondary reversion to an older pronunciation.


Abstracting away from the specific loanword relationship, there are two kinds of "ð", a fricative and an approximant. The lenited "δ" (oft transcribed as ð) of Danish is an example, and "ð" in some Kurdish dialects is quite similar. There are also a few Bantu language which have a "non-palatal" [j], that is, a [j] (IPA for y) which doesn't have the same tongue-raising usually associated with [j], which sounds very much like the dental approximant of Danish and Kurdish. If Vietnamese had a non-palatal /j/ at that point in its history, that would explain the pattern of borrowing.

  • d still represents /j/ in Vietnamese nowadays Aug 23 '15 at 5:40

The sound /ð/ doesn't exist in Vietnamese. And d in Vietnamese doesn't represent any of the D-like sound as you expected. The d sound in Vietnamese is /ɗ/ which is a voiced alveolar implosive and written by the character Đ/đ

In Northern Vietnamese the letter d represents the phoneme /z/ which is no way similar to /j/1. However in the remaining dialects of Vietnamese d still represents the phoneme /j/, so there's no difference from Chinese.

1Note that /y/ is the sound of u in French or German ü, not the sound of y

  • well, z and j seem somewhat similar to me, and they must have descended from some common ancestor. Both are voiced, and neither are labial. Aug 23 '15 at 6:18
  • IMHO z and j may be somewhat similar, but /z/ and /j/ are not Aug 23 '15 at 6:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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