"Geda", part 3

« previous post | next post »

Earlier this week (11/12/18), under the rubric "Of knots, pimples, and Sinitic reconstructions", we discussed the origins and meaning of the fascinating Sinitic word "geda" ("pimple; knot; lump").  That, in turn, was prompted by our initial acquaintance with "geda" in "Too hard to translate soup" a couple of months before (9/2/18).  After considering a possible source in Indo-European, Turkic, Tungusic, and Mongolic, there seemed to be a bit of momentum in favor of the last named family.

Since "geda" first appeared in a significantly large number of citations in written Sinitic during the Mongol Yuan Dynasty (1271-1368) about a thousand years ago, it was thought advisable to look at an earlier stage of Mongolic rather than simply referring to modern Mongolian forms.  So I thought of asking Daniel Kane, a rare specialist in Khitan, which is generally considered to be a Para-Mongolic language, whether he had any thoughts on the matter.

Here is Daniel's reply:

The word *geda "looks like" it is non-Chinese, and there are very many such words in northern dialects. Nie Hongyin noticed ages ago that some Kitan names in Chinese transcription reveal an underlying monosyllable  *Ce-la < Cra, but no one has been able to incorporate this into the broader picture.

In the larger Chinese dictionaries we can find *geda written in various ways, and you are right, it must mean something like lump. This may be Kitan, but there is no evidence. The Kitan texts are about 40 funerary inscriptions and a few odds and ends in other texts. It would be strange to find a word like lump. To find out what is in an inscription, one has to transcribe and translate – so far only four have been partially deciphered. Transcribing these takes up an enormous amount of time. Even the parts of the text that can be understood are mainly Chinese names, titles, place names, etc., which prompted Nie Hongyin to suggest that no (or very few) Kitan words (real Kitan words) are known, an observation which made Liu Fengzhu apoplectic.

Of the Kitan words per se, the majority opinion is that they can be compared with Mongol, which everyone seems to think is self evident, but I am in a minority of one in expecting more evidence. If the Mongol theory has anything in it, Mongol would be the place to look for geda 'lump' .

All that having been said, I went through the Kitan texts at my disposal, looking for some word which could be transcribed as *geda", but I could not see it in this very meagre amount of Kitan texts. So I did not get anywhere. I started work on transcribing another text late last year, but some health problems interfered with that, and I have not been able to continue. Sooner or later someone is going to have to learn how to transcribe these texts, and I might have to wait until the next generation. The students at Yale* were very bright and reasonably interested, but I think Andrew Shimunek is probably more likely to continue the great task, at least in the Western world.

[*YouTube videos of the Yale Kitan workshop may be accessed here.]

So much for "geda" through a Khitan lens, at least for the present time.



  1. Victor Mair said,

    November 15, 2018 @ 12:45 pm

    From Alexander Vovin:

    As far as I can tell, there is nothing similar in Middle Mongolian (MM). There is Eastern MM ɣada 'cliff', but it is a stretch: I might live with 'hill', but not with a 'cliff'. Also note that geda violates vowel harmony (we would expect either ɣada or gede) , so chances that it is Mongolic are not too great. Maybe Juha knows something similar from modern Mongolic languages. A very quick and quite superficial search in Yeniseic did not reveal anything either.

  2. Benjamin E. Orsatti said,

    November 15, 2018 @ 1:26 pm

    Nepali, perhaps?

    Entry Grammar Meaning Nepali
    dɔ n. nut, seed, kernel geḍā.


  3. Benjamin E. Orsatti said,

    November 15, 2018 @ 1:28 pm


    In Modern Nepali, there is also "गुदी" (Gudī), which means "inside, core, interior".

    Very likely a spurious etymology, but fun to think about, just the same.

  4. Eidolon said,

    November 15, 2018 @ 9:00 pm

    The Beijing variant is gāda, and I have also heard of it pronounced gēde, so the violation of vowel harmony is probably not a big deal; it could've been borrowed as ɣada or gede, before experiencing phonetic change in Sinitic where vowel harmony is not a requirement. The lack of a root in Middle Mongolian is a bigger problem, and may indicate that the term was not ultimately of Mongolic provenance, even though the vector of transmission may have been through Mongolic.

  5. Chris Button said,

    November 16, 2018 @ 10:24 am

    The Beijing variant is gāda, and I have also heard of it pronounced gēde, so the violation of vowel harmony is probably not a big deal; it could've been borrowed as ɣada or gede, before experiencing phonetic change in Sinitic where vowel harmony is not a requirement.

    The time period for the borrowing is at the tail end of Middle Chinese before we move into what is termed "Early Mandarin". I would have expected some trace of the ru-sheng (stopped codas) in both syllables to still have been present at that stage before disappearing in Early Mandarin although they were most likely lenited in some way. Maspero's "Dialecte de Tch'ang-ngan" (1920:41-44) has an interesting discussion in this regard regarding possible frication of the final stops of which my proposal of Mongolian хавдах "swell" is somewhat supportive (although quite possibly entirely misled). Prof. Mair's observation of the vacillation of -p and -t at the end of the final syllable in terms of its reconstructed Middle Chinese form is particularly telling in terms of the state of flux of the final codas during that period.

    Sooner or later someone is going to have to learn how to transcribe these texts, and I might have to wait until the next generation. The students at Yale* were very bright and reasonably interested, but I think Andrew Shimunek is probably more likely to continue the great task, at least in the Western world.

    Presumably that is due to lack of funding rather than interest? Personally, I think that sounds like a wonderfully rewarding way to spend one's time!

  6. Eidolon said,

    November 16, 2018 @ 9:57 pm

    Indeed, the examples should have been presented with -p or -t final codas. In fact, ɣət dɑt is quite close to ɣa da if we simply assume that ɣa da was borrowed into late Middle Chinese with -t codas, after which there was a shift from a > ə > e, i in most mainstream varieties of Sinitic for the first syllable, and a loss of -t in Early Mandarin. But the semantics of "cliff" are problematic, as mentioned above, though I'd like to consider the göbdürigüü class of words brought up by Timothy May in the previous thread:

    göbdürigüü: elevation, wart, blister, pimple.

    göbdürigüde: to have pimples.

    gübegegen: hillock, knoll, mound

    güber-e: a bruise or welt caused by lashing

    A cursory search reveals the following:


    There is no Middle Mongolian root listed, but there is Middle Turkic: qabar-, qaba, with semantics "to swell, form blisters; thick, swollen; hill, mound". Could this have been related to Middle Mongolian ɣa da and Mandarin geda?

  7. Chris Button said,

    November 17, 2018 @ 6:53 am

    On the other hand, perhaps focusing too much on traces of final codas (in whatever form) is barking up the wrong tree. The following is from Pulleyblank (1965) "The Chinese Names for the Turks":

    Middle Chinese had consonantal clusters neither at the beginning nor at the end of a syllable, and the only way in which such clusters in a foreign word could be represented was to make use of an additional syllable in Chinese, only the initial consonant of which
    counted so far as the transcription was concerned. We find syllables of more than one type used for this purpose but one of the commonest types is certainly the ju-sheng syllable, that is, a syllable closed by a final stop. This may seem surprising since one might suppose that to have to discount not only a following vowel but also a final consonant
    would be more troublesome than to have to discount only the following vowel. The explanation is undoubtedly that the ju-sheng was the
    shortest in duration of all the Chinese tones and therefore best suited to represent absence of duration, in spite of the difficulty of the unwanted

    … Examples of final clusters are in the nature of things much less common, since they were not often found in the languages which the Chinese were called upon to transcribe. We do however find cases where a foreign word ended in a final consonant which could only be adequately represented by the initial of a Chinese syllable. In such
    cases the Chinese syllable chosen was likely to be a ju-sheng, for the same reasons that led to the choice of such syllables in initial or medial position.

    In that case, perhaps we should be looking for a monosyllabic source word?

  8. Victor Mair said,

    November 19, 2018 @ 8:27 am

    From Gerd Carling:

    These types of words (knot, knob, Sw knut, knottra, knyta, knyte etc.) are very complex, since they are affected by sound symbolism in language history. In particular is the growth of this type of clusters of sound symbolic terms frequent in Germanic (I have written about that, cf. links below and attached pdf). It is relatively easy to track etymologies back into a proto-root (here, it is PIE basis *gen- ‘press, knead’ IEW:370) but it is very hard to establish valid cognates in the daughter languages. I don’t think that even nôdus is a valid cognate to knot.

    Etymology and the European Lexicon; Etymology and iconicity in onomatopoeia and sound symbolism: A Germanic case study

    G Carling, N Johansson, "Motivated language change: processes involved in the growth and conventionalization of onomatopoeia and sound symbolism," Acta linguistica hafniensia 46.2 (2014), 199-217.

    pdf available here:




    [VHM: Also available from me if you have a hard time getting it from one of these sources, though that shouldn't be a problem.]

  9. Ben Orsatti said,

    November 19, 2018 @ 10:36 am

    Chris Button said,
    "In that case, perhaps we should be looking for a monosyllabic source word?"

    How about: 括 (*kʷāt (~gʷ-))?


    WARNING: I'm neither a Sinologist, nor a linguist, but even a broken clock…

  10. Chris Button said,

    November 19, 2018 @ 9:12 pm

    @ Ben Orsatti

    I was actually referring to the possibility that a monosyllabic source word in Mongolian, Tungusic, Turkic etc. might have been treated as bisyllabic in Chinese.

  11. Victor Mair said,

    November 20, 2018 @ 6:55 am

    @Chris Button

    At first I felt the same way as you about Ben Orsatti's proposal to consider 括. I thought that it didn't even have the right galaxy of meanings ("draw together; include; embrace; enclose; contract", though Starostin gives "bind; tie up; bring together") for "geda" ("pimple; lump; clump; knot; node"). But then I scrolled down to Long-range etymologies, Austric etymology, and Austro-Asiatic etymology in Starostin's StarLing cited by Ben Orsatti, and I saw that Starostin does cite a number of words meaning "knot" that have the right configuration of sounds. So it's something you might want to think about a bit more.

  12. Victor Mair said,

    November 20, 2018 @ 5:16 pm

    From David Dettmann:

    I found a Japanese language study of Daur vocabulary by Hitoshi Kuribayashi through WorldCat with an entry you might be interested in. That dictionary file may be downloaded from this site:


    If you scroll to page 92 (according to the pagination at the bottom of each page), you'll find a Daur entry for xild, written under the Mongolian script word hilidü(n). The Chinese gloss given (or maybe it is a Japanese-influenced gloss) as xuǎn 癬, which I think was intended by the author to mean "ringworm". There are supposed to be Khitan influences on Daur, right? I wonder if it could be that xild became geda?

    I also located words for "cabbage" and scanned for the Chinese word yuán 圆 ("circular; round") (since round cabbage is called gēda bái 疙瘩白 ("geda white") in the NE but couldn't find anything like geda. If you are interested in seeing those entries for vegetable names, they're on page 153.

    In any case, Daur might be a language to explore for that geda connection. There are other dictionaries out there, including one published in Inner Mongolia in 2017. No nearby lending libraries have that one yet.

  13. Penglin Wang said,

    November 20, 2018 @ 7:39 pm

    I’d like to concur with Daniel. Written Mongolian has two forms: (h)ilidü(n) and (h)ilde (Lessing 1995 Mongolian-English Dictionary pp. 403, 408), with which Dagur xild (hild would be a better transcription) is cognate. I estimate that (h)ilde was a colloquial form pronounced in speech. If this colloquial form was there in Kitan, it would be possible phonetically for Chinese geda to have a connection with Mongolic (h)ilde (either Mongolian or Kitan), because the postvocalic liquid l might be omitted in Chinese pronunciation or transcription. This implies that the form (h)ilidü(n) with its syllable li could hardly be fed into Chinese geda. By the way, my experience living in the Dagur region tells me that this skin trouble was not rare among the Dagurs, and thus the word should have existed since early times.

    David mentions the Kitan influences on Dagur. Yes, we have some pieces of evidence to show the unique connection of Dagur with Kitan. Liaoshi (878, 389, 930) records the Kitan numeral for ‘five’ as 討 (tao) and the Kitan word for ‘iron’ as 曷朮 (heshu), which correspond to Dagur tāw ‘five’ and kaso ‘iron’, respectively. Moreover, I have identified Kitan 沓温司 coindexed with Chinese 五院司 (wuyuan si) and henceforth connected Kitan 沓温 (tawen) used as an adjective with Dagur tāwǝn ‘five’ used as an adjective as well. Thus, Dagur tāw and tāwǝn must be reflections of Kitan numerals tāw and tawen (For more detail see Penglin Wang 2015 Number Conception and Application (p.211). Daniel, in his book The Kitan Language and Script (2009:43) considers the Kitan word ki.ên to be corresponding to the Chinese xian ‘county.’ Here Chinese x as a palatal fricative was represented as a velar stop k in Kitan. Dagur has such a phonetic representation (Penglin Wang 2018: Linguistic Mysteries of Ethnonyms in Inner Asia (p. 117).

  14. Victor Mair said,

    November 20, 2018 @ 8:33 pm

    From Pamela Kyle Crossley:

    This is a very interesting discussion, partly because there doesn’t seem to be much to grab onto before Cantonese. That would lead me to think that the place to look might be Tibetan, Burmese, Thai, Vietnamese. I also think there is some possibility of medical vocabulary from Tibetan.

    The links from Kitan to Da’ur are clear enough (genetically at least), but that only reinforces the emerging argument (from Juhanen,, Kane, Shimunek, etc) that Kitan might have been more Tungusic than Mongolic. Overall on the basis of what everybody has written so far it sounds like you should be looking west and south, not north.

  15. Chris Button said,

    November 20, 2018 @ 9:26 pm

    @ Penglin Wang

    If this colloquial form was there in Kitan, it would be possible phonetically for Chinese geda to have a connection with Mongolic (h)ilde (either Mongolian or Kitan), because the postvocalic liquid l might be omitted in Chinese pronunciation or transcription.

    Actually the "l" can be accounted for by Pulleyblank's (1978 – Nature of MC Tones, p.176-177) development of Maspero's proposal (cited above) regarding the lenition of the Middle Chinese stop codas in the development of Early Mandarin where he associates a rhotic articulation with a -t coda supported by the fact that final -t was often transcribed in foreign scripts as liquid -r at that time (and as -l in Sino-Korean).

    As a result, I think Daur "xild" / Mongolian "(h)ilde" is an extremely promising proposal.

    @ Victor Mair

    So internal splitting rather than an external loanword? Wouldn't that either require some kind of euphonic association between the two syllables, or an original liquid onset in the second syllable?

  16. Victor Mair said,

    November 20, 2018 @ 10:40 pm

    From Stefan Georg:

    There is no shortage of Dagur dictionaries, I uploaded some of them for you. I remain skeptical, though, whether you will find there a source for geda – if the word is there in some shape, it will be most probably just the Chinese word. I also cannot imagine how hildün/xild could be geda. Regarding Dagur and Khitan – Dagur is sometimes thought to be the continuation of Khitan, but this, I think, rests on shaky ground. Most features of Dagur can be accounted for by viewing it as a somewhat archaic member of Common Mongolic

  17. Victor Mair said,

    November 21, 2018 @ 8:18 am

    From Juha Janhunen:

    Just a note on Khitan and Dagur: There is a tradition of scholarly folklore that the Dagur are somehow connected with the Khitan, or even descendants of the Khitan. Some Dagur scholars themselves believe that this is so. Linguistically this is impossible. The Dagur language is a direct heir of Proto-Mongolic, which is more or less the same as "Middle Mongol", the language spoken by Chinggis Khan and his contemporaries and documented in a large number of sources in different scripts. The Khitan language, as much as we know of it on the basis of the Khitan Small Script (and some Chinese sources), was a Para-Mongolic language, fairly far from Proto-Mongolic, and with features that are not present in either Middle Mongol or Dagur.

    Even so, it is true that the Dagur language was probably the first to branch off Proto-Mongolic, This apparently happened around the time of Mongol state building, or immediately before. For instance, the Dagur were never comprised by the literary traditions of Middle Mongol, or by the Mongol script. It seems that the linguistic ancestors of the Dagur moved along the Argun river upwards to the Middle Amur-Zeya basin, where they survived until the 17th century, when they were moved by the Manchus to the Nonni (Nenjiang) basin in Central Manchuria. In the Middle Amur region, the ancestors of the Dagur were subjects to the Jurchen, who were originally the principal adversaries of the historical Mongols. So, it seems that the Dagur were Mongolic speakers who chose to remain on the Jurchen side against the historical Mongols – until the Mongols destroyed the Jurchen empire in 1234. Since the Jurchen (Jin) were the political heirs of the Khitan (Liao) in Manchuria there is, thus, an indirect connection between the Dagur and the Khitan as well. There are also a few lexical items, like the word for "iron", which is supposedly shared by Dagur and Khitan. These could be either archaisms inherited from Pre-Proto-Mongolic, or loanwords from Khitan to Dagur.

  18. Victor Mair said,

    November 21, 2018 @ 8:36 am

    From Zeyao Wu, who hails from the Northeast:

    It is interesting about the idea that "geda" may come from Kitan. I think I have heard someone mentioned that before, but he did not give a specific evidence.

    However, for 疙瘩白, I have never heard of it. I asked my parents and grandparents. They also had no idea about it, and my father said that we usually say 蘿蔔疙瘩, 芥菜疙瘩(refers to "雪裏蕻 xue li hong", a kind of vegetable) and 白菜幫子. I am thinking that maybe 疙瘩白 is used in the other two provinces in Dongbei. Among the three provinces (Heilongjiang, Jilin and Liaoning) in Dongbei, their topolects do have some subtle differences. I really think in dongbei topolect, the meaning of 疙瘩 is complicated:

    1. the youngest child: 他是家裏的老疙瘩. (He is the youngest one in the family.)
    2. indicate a specific place: 你放那疙瘩. (You can put in that place.)
    3. knot: 你係個疙瘩在這兒. (You make a knot here.)
    4. bun (hairstyle): 我要梳個疙瘩啾. (I want to put my hair in a bun.)
    5. 三七疙瘩話: a fixed phrase. Bad words. Complaints.

RSS feed for comments on this post