Talk:두꺼비

, I don't think Vovin's speculations have sufficient validity here:


 * The variation between Middle Korean and Modern  >  demonstrates that the historical form must have been, since  would produce Modern.
 * The variation between MK and MdK  >  is identical to that between MK  and MdK  >, so the internal derivation from this semantically easily related adjective is more likely and in fact what has usually been considered by Korean scholars (e.g. 우리말어원사전, 김민수). There is also a Middle Korean pitch accent match between "toad" and "to be thick", as the latter is an irregular rising-pitch stem which is synchronically from  and the loss of the final minimal vowel is expected.
 * There are some manuscripts of the Jilin leishi in which is written as, so  may actually have been a later copyist's error of the variant form , in which case the whole argument falls apart. Note that the preceding glosses are , , , , and.

I've amended the etymology accordingly.--Karaeng Matoaya (talk) 01:43, 9 December 2020 (UTC)

, I think you might be right, but there are a few observations to be made:
 * Yes, *twukVtep would produce Modern Korean *twuktep. However Proto-Korean intervocalic *k lenites to h, so a (diachronically) closer form would be
 * twuhVtep, which can unproblematically lead to twuthep. Nonetheless, there is no internal evidence that points to this form instead of *twutVhep<*twutukep. in which case the form tutkep might be an early dialectal word predating intervocalic lenition.
 * I think I've seen references to Kim's dictionary before (Kim 1997, isn't it?), so the appeal to authority here works, but the main thing is definitely the pitch. if it is indeed rising, then the argument is very strong.
 * I don't doubt this, but I'm really curious where you took it from.

Edited accordingly.


 * , on your first point, I think it's exactly that: at some point and  must have coexisted (the latter applied the vowel-dropping first, which made the lenition of  unnecessary), since the variation between the Middle and Modern Korean forms aren't explainable otherwise. This points to an ancestral.
 * On the final point, I was going off this article, but it seems like the media misrepresented this as they are prone to do. I consulted Jin's academic work, and what it turns out is that the two earliest manuscripts of the Jilin leishi actually have and, both of which obviously contain typographic errors. The "mainstream" Qing version of the Jilin leishi was derived from the same tradition as one of these two early manuscripts rather than representing an independent transmission from Sun Mu, which suggests that a later scribe or printer noticed that  did not fit the context and corrected it to the visually similar . Jin's point is that since there is no firm manuscript evidence that  actually was the correct form, we might as well look at the context, in which  is a better fit among the lice and ants and flies. He connects this to , apparently a dialectal word for "bedbug" besides its primary meaning of "whore", since Jilin leishi final  corresponds to both Korean  and .--Karaeng Matoaya (talk) 00:49, 12 December 2020 (UTC)

If is, as he claims, a correction for, then it would be more difficult to reconcile with  as the graphic similarity becomes moot, unless this scribe intended to write , but rendered it as  , while the scribe responsible for the first manuscript somehow turned  into. Since you have access to Korean-language works, could you check out 안병호 1985 계림유사 와 고려 시기 조선어, 강신항 1980 계림유사 고려 방언 연구 or 남풍현 2009 고대 한국어 연구 to see if they've got anything to say about it? Besides that, does Jin cite any source for this particular meaning of ? The article you linked appears to claim it's a dialectal word but it doesn't give any references.


 * , 강신항 1980 cites previous Korean scholars who think the Jilin leishi manuscripts simply omitted the character for the first syllable, so that it was actually writing the 거비 part of 둗거비 (he does not advance his own hypothesis). Incidentally, I find even this more convincing than Vovin for several reasons:
 * Again, the difficulty of Vovin's hypothesis to explain modern in virtually all dialects as opposed to
 * Middle Korean 둗거비 is actually attested in 1517, so the hypothesis fails to explain what is dialectal divergence in Middle Korean itself
 * the Middle Korean word for "to be thick" varied between 두텁다 and 둗겁다 in even the earliest Hangul sources, and if we accept the notion that this adjective is the source for the "toad" word, the latter word would also have varied between 두텁 and 둗겁, which is what we see in Middle Korean
 * Vovin needs to be an irregular metathesis posited to produce 두텁:
 * is not a minimal vowel and is less likely to be subject to vowel loss; it's almost always retained in the causative suffixes, for example, rather than becoming coda
 * The prefix is not identified, and I can't think of what Vovin might have in mind—my best guess is that he might mistakenly think that  has such a prefix, when it's actually the verb stem  "to dig through"
 * Unfortunately I do not currently have access to the newer works :(
 * Jin cites 이희승's 1961 국어대사전, which I do not have access to, which he says lists 갈보 as a vulgar word for bedbug. He also says he heard it himself as a child, so take that as you will.--Karaeng Matoaya (talk) 01:29, 15 December 2020 (UTC)

I do not think Vovin tried to identify the prefix, but rather simply accepted it since it supports the external comparison. This is somewhat surprising given his strong dislike for "unaccounted segments", but then again, the paper was written in 2000. I also doubt there is any problem with the syncope of *i since the loss of non-minimal vowels, while not as common, is sufficiently attested: LMK psol 'rice'< PK pAsAr, LMK alph 'front' < Paekche aripisi LMK pto- 'pick, pluck, gather' < Manchu fata- 'id.', LMK ptho- ‘to play on a stringed instrument, to beat cotton out’ < Manchu fithe- 'id.'. Therefore, we are left with three hypotheses, none of which is completely satisfactory:

1. Kang's hypothesis: twutkep 'toad' < dialectal *twutkepu- 'thick' < *twutVkepu- 'id.' and twuthep 'toad' < twuthepu- 'thick' < *twutVkepu- 'id'. Jilin Leishi *kitpe accounts for the second part only: *-kitpe > *-kep. The loss of so many sounds is problematic and it is pure speculation that an outsider writing down a local word would remove part of it.

2. Vovin's hypothesis: twuthep 'toad' < *twukVtep < *twu- 'unknown prefix' + *kitpe < Manchu-Tungus *kitibe. There is an unaccounted segment, a metathesis which is not mandated by phonotactics and fails to explain the forms twutkep 'toad' and twuthep- 'thick'.

3. Jin's hypothesis: derivation as in 1. Jilin Leishi *kitpe reflects Contemporary Korean dialectal word 갈보 based on the assumption that syllable coda l goes back to earlier t. It fails to account for the vowel discrepancies and the textual argument is problematic.

It is unlikely that Vovin did not know about these forms, which makes me wonder whether he does have some sort of alternative explanation. I will try to contact him via email, even though there is little chance he will respond and even less chance that he actually has any.


 * , I'm not quite sure I agree about : the Baekje forms with or the Tungusic loans discussed by Vovin would be quite ancient ones (and Vovin seems to imply this, e.g. a potential loan from Parhay times in the late first millennium is mentioned as a "comparatively recent loan"), whereas the Jilin leishi represents a late stage of Old Korean only three centuries and a half removed from Middle Korean. My feeling is that there would simply have been less time for something as drastic as what Vovin proposes, compared to attestations from the Nihon shoki or prehistoric borrowings.
 * Generally speaking, a major shift like >  would stand out for me given that most Jilin leishi ancestors of Middle Korean forms are not very removed from the Middle Korean forms (in the other items Vovin discusses in Vovin 2000, we're just one sound shift removed: lenition of  for "head", final consonant dropping for "tongue", final syllable dropping for "rain", minimal vowel syncope for "big").
 * On the phonetics of, I have just noticed another weakness with Vovin's argument. Further down the Jilin leishi we have : "for 'reading books', they say 乞鋪". This form is agreed to be ancestral to Middle Korean "to read books" (cf. pages 64 and 76 in The Languages of Japan and Korea edited by Nicolas Tranter), so Jilin leishi  seems to actually be transcribing Korean.
 * This being the case, even if Vovin was otherwise correct, we would have, not . This would not lead to Middle Korean twuthep even with metathesis.
 * On Jin's hypothesis, Jilin leishi corresponds to all Old Korean dentals (Vovin also mentions this in Vovin 2000, p. 145), so even  could have been written as 虼鋪. There is only an issue with the first vowel, so Jin's hypothesis is certainly the most persuasive if we only take the phonology into account. Nonetheless, now that I know there is a lack of actual textual support for "bedbug", I'm reluctant to support this hypothesis.
 * I think it's a good idea to ask Vovin, especially since his views in general have evolved since. Vovin was still a follower of Altaic in 2000, wasn't he?--Karaeng Matoaya (talk) 10:43, 15 December 2020 (UTC)

Thinking about it, I wonder if it could be that twuthep ~ twutkep 'toad' derives from the verb 'to be thick', while the Jilin Leishi actually reflects an unrelated loan from Manchu. The last point you made is still valid, but this might be the best explanation so far.

By the way, check out this entry on Vovin's website. It offers an etymology for twuti-cwuy.
 * I didn't know about the site, I'm really quite excited!
 * On twuti, I'm again rather skeptical that an internal etymology isn't the more plausible. The conventional etymology for "mole", given by the National Institute of the Korean Language here, is twuti- "to dig through" + cwuy "rat". twuti- is an attested Middle Korean verb, which is a point in its favor.
 * On twuten, twutulk, I'm again quite skeptical. The conventional etymology is an internal connection to 돋다 "to bulge; to protrude". "Bank; levee" aren't really "earth" in general but earth that bulges or swells. Compare Contemporary Korean, , , , and the ideophones , , . So the "earth" connection does not really work in light of the broader use of the twut- morpheme in Modern Korean, which all point to a primary meaning of "bulging".
 * I've checked Whitman and Vovin on this, but they only cite Middle Korean twuten and twutulk and their modern reflexes, which makes me wonder if they were unaware of the wider use of the morpheme in MdK. I might actually inquire Vovin about this via emial.--Karaeng Matoaya (talk) 02:13, 16 December 2020 (UTC)