¹ The ideas presented here have been brewing for some time. I collected the data while compiling a Korean-English dictionary with my colleagues Sung-Un Chang and the late Yang Ha Lee. The etymologies have benefited from discussions with a number of scholars in Korea, notably Lee Sung Nyong, who helped me gain an understanding of the structure of Middle Korean. I have taken into account the native Japanese etymologies suggested by Ohtsuki Fumihiko's etymological dictionary Daigenkai (Tokyo, 1932), but explicit reference is made to only a few of these. Earlier versions of this work were presented before linguistic groups at Yale, Harvard, and the Universities of Michigan, Washington, and California (Berkeley) ; the present version has benefited from suggestions received on those occasions.

² I have omitted from these charts the symbol *Y (see Note on Correspondence 25), the cover symbols *n (for homorganic nasals in Etymologies 142, 293, 302) and *V (for unspecified vowel), and the three symbols *e (Etymology 299), *a (Etymology 307), and *r (Etymology 305) that represent rejected or doubtful correspondences.

³ Another example: arrow. II. phyëng-cak ‘medium-sized arrow’, cak-sal ‘harpoon’, cak-sali ‘cross-sticks’, cak-täki ‘stick with forked head’ : yá ‘arrow’. 10–17 + -k:#. *ja(k).

⁴ Note that the Japanese developments in green and hemp call for open juncture before the consonant.

⁵ Korean pitulki, MK pit°/_uli < MK pitɔlki ‘dove’ may contain this etymon and be directly or indirectly cognate with J páto ‘dove’; other possible examples of the correspondence sequence pi : pa are pis : páuki ‘broom’ (and J paké 'brush'?) and pisïs : pasu 'aslant'. Various problems have led me to set these etymologies aside.

⁶ Ohtsuki's etymology ‘roads between [ricefields]‘ is ingenious but unconvincing.

⁷ Korean also has culluki ‘a kind of crane’, apparently an aberrant development with a different suffix.

⁸ The derived adverb sukósi-ku indicates an underlying infinitive from (?*)sukosí-i (listed by Ohtsuki) ‘is little’. Also related is sukobúru ‘extremely, muchly’ < OJ ‘a little bit’.

⁹ I have elsewhere suggested that “kom might be from kolh-ïm ‘going stale’. But another possibility is a shortening from ”komph-angi (with the common diminutive suffix). The reconstruction would then be *kwɔnbxyi, with the intrusive homorganic nasal of first of all, jump, and pan; note also the loanword K tampä ‘tobacco’.

¹⁰ Cf. the final s in K tas(ës) ‘five’, yës(ës) ‘six’, and yël-ës ‘(ten or so >) a fair number’ ; see remarks under two.

¹¹ Ohtsuki treats íne as a contraction of ipi-ne, a suffix-extended form of OJ ipi ‘cooked rice’; if the opposite is true (and ipi is actually a contraction from ipine), my first suggestion may be worth a second thought. (Both Ohtsuki's etymology and mine would assume the variant OJ yóne to be a secondary development from íne.)

¹² But nïkk-ëw- ‘be felt’ indicates that K -i- is a suffix and the final *y should not be reconstructed.

¹³ Still another candidate is J turá ‘a line, a row’, whence tura-nár- ‘be lined up’ and turá-n(e)- ‘line them up’; but see accompany and two for a better berth for these words.

¹⁴ This assumes that the Korean form is not contracted from hëpi- ‘dig out’. There is a Korean dialect form hachi- ‘row’ that is probably related.

¹⁵ Since Japanese also displays vowel alternation in these etyma (ko/ka as in ka-ku > ka-u > koo ‘thus’, so/sa), can we reconstruct the alternation *o/a with Correspondence 17b?

¹⁶ Ohtsuki considers the Japanese noun a contraction from te-arap-i ‘hand-washing’. In any event, my etymologies must be rejected because of an error in one of the correspondences; since this error is of a type all too easy to commit, it may be instructive for the reader to discover it for himself. (I have eliminated references to washtub in the List of Correspondences.)

¹⁷ But the -k may be internal to Korean; probably related are ōl ‘strand’, olmu ‘noose’, and perhaps oli ‘strip’ and ola ‘red rope for binding criminals’ (unless the latter is from *polk- = pulk- ‘be red‘).

¹⁸ Ohtsuki's two suggestions (< pa-siwam-i ‘leaf-shriveling’ and < pari-siba-mi ‘pin-brushwood-fruit‘) seem farfetched.

¹⁹ A better shape fit for the Japanese word is K mat < MK mot/mos ‘eldest’ (?< 'mos 'extreme'), 2–21d–6b [if we assume an original -t] -19c, *mádu. (Note that the symbol á in the Japanese citation merely represents modern accent, but *á = ɔ:a.) The two etymologies are not necessarily in competition with each other, since K mëncë may contain an abbreviation of cëk 'time', with the preceding nasal either intrusive or a relic of the pseudocopula adnominal form (i)n; cf. ince ‘now’ < i ‘this’ + cëk ‘time’. For another possible place for K mat, see bottom.

²⁰ The Korean -i may be a verbalizing suffix; cf. the derivatives tëm-pëng(/tampang) ‘bustling’, tëm-ppëk(/tam-ppak) ‘recklessly, greedily’.

²¹ And, in fact, erroneous, since there is no -m in the Korean form. If the words are cognate we must assume secondary -h in Korean and -ma in Japanese, *k(x)al (with 12b).

²² To answer my own question, no. The likely etymology is tol.. / “tïl(..), MK tï'lï(h) 'wild; field', noph-tïli ‘unproductive ricefield in high [noph-] place with little water’ (?< *tɔl) : tá ‘ricefield’, 6–21d–12b. *tál.

²³ There are many more etymologies to be uncovered. Here is one that is of interest because it provides an attractive alternant to an unacceptable etymology of Ramstedt's : face. III. thal ‘mask’ : turá ‘face’. 6–5c–21a–11–23. *txɔ⇛.—Only in page proof did I discover that I have inadvertently included sour twice (Etymologies 211 and 316). And it has been brought to my attention that Etymology 37 (cage) must be reconsidered in the light of the Japanese historical spelling wori, which I had overlooked. Two additional modern Korean expressions that contain the word choli (tail, Etymology 236) are “cepi choli ‘swallowtail shape’ and hoy choli ‘a switch’.