I read that "Keri" is a verb suffix used to denote the realization of something, and is also poetic- I was wondering if anyone knew where it came from, kanji etc. For more context it was used with ari- like otoko arikeri. I think it was meant to mean "there was a man."

  • In my dictionary(Obun-sha Zen-yaku Ko-go-ji-ten) "けke りri" is derived from "来ki 有りari". Some sample sentences in the dictionary are from "Man You Syu" the oldest document, so the derivation of "ke ri" has happened in very ancient times, I think.
    – user4688
    Mar 24 '14 at 10:25

Just doing a quick survey of the kanji spellings used for けり in the first five books of the Man'yōshū, after excluding false positives (matches for けり belonging to -ku verbs), here's the breakdown for spellings by frequency and 甲類【こうるい】 (ke1) vs. 乙類【おつるい】 (ke2):

  • 来: 21 -- N/A, non-phonetic use
  • 家里: 11 -- ke1ri
  • 家利: 5 -- ke1ri
  • 家理: 2 -- ke1ri
  • 鶏里: 2 -- ke1ri
  • 計理: 1 -- ke1ri
  • 有: 1 -- N/A, non-phonetic use

Though the kanji used as man'yōgana vary, the readings are consistently of the 甲類 variety. As noted in Shibatani's The Languages of Japan starting from around page 134, some scholars view the 甲類 variant e1 as arising from i + a, consistent with the previously mentioned theories that keri derives from ki + ari.

Shogakukan's Kokugo Dai Jiten Dictionary (their title in the copyright notice, though the "dictionary" is redundant) explains the origins of keri as:

A fusion of past recollective auxiliary verb き and 有り, or of 来 (ki) and 有り

As @DariusJahandarie noted, the past recollective き has no 連用形{れんようけい}, so き + [verb] is unlikely. I've poked around in a few dictionaries now, and nowhere can I find any mention that き might have been irregular in this way (i.e. that it might have allowed for the 終止形{しゅうしけい} of き to attach to another following verb). Likewise, I can find no mention that 有り could be preceded in verb compounds with anything other than the 連用形, which apparently rules out the possibility of [verb in 終止形] + 有り.

Given also the prevalence of spelling けり with 来 in what appears to be a conceptual as opposed to phonetic way, it seems most likely that けり derives from 来{き} + 有{あ}り.

That said, human language is nothing if not wonderfully exceptional, so past recollective き + 有り remains a potentially viable hypothesis despite the unlikely grammatical construction, and in some ways this seems to be the better semantic match.


Since this is a very old construction, I don't think there is an absolutely clear origin, but my understanding is that the popular theory is

      k-u + ar-i → k-i-ar-i → ker-i

where k-u is the カ変動詞 "to come" and ar-i is the ラ変動詞 "to be".

However, there is also a minority theory of

      ki + ar-i → ki-ar-i → ker-i

where ki is the 体験回想 (recollective) 助動詞 and ar-i is the ラ変動詞 "to be". This theory doesn't make much sense to me, since to my knowledge the 助動詞 ki is never syntactically in this position otherwise (i.e., it doesn't have a 連用形).

Your Answer

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