It is thought that the moraic post-vocalic consonant [N], spelt with ん, appeared in Japanese under the Chinese influence, with the influx of borrowings. Are there any kun-readings in Japanese that have the final ん? And with what processes could they have emerged?
Kun-readings are an orthographic notation for the native lexical stratum (Yamato-kotoba). So the underlying question is, are there native morphemes with closed syllables (that is, with a consonant at the end)? Originally, no; Old Japanese was a CV language (syllables started with a consonant and ended in a vowel). However, what are currently voiced consonants used to be prenasalized (as they still are in certain dialects, e.g. in Tōhoku); so that ざ was /ⁿza/, ば was /ᵐba/ etc. Modern tanabata used to be pronounced tanambada.
Around Heian times, this worked in tandem with the new Chinese layer to introduce the moraic, syllable-final N-consonant. Consider for example the onbin sound changes:
- A verb like hanasu inflected to the i-form to get the suffix -te, generating hanasite.
- A verb like hanatu, likewise, was hanatite; but elision of the vowel rendered it hanatte – a new geminated form that was only possible after CCV syllables became part of the system, along with Chinese consonant-final loans.
- A verb like asobu > asobite also underwent elision – but recall that the Old pronunciation was asombite; the pre-nasal survived elision while the voicing assimilated to the -te, generating asonde.
So Yamato morphemes got syllable-final N due to internal processes, at the same time that the Chinese wave brought another syllable-final nasal (and not just the nasal; syllable-final /p/, /k/, /t/ survived in on loans—the last two for many centuries). One way of explaining this is, multiple pressures changed the syllable structure from a simple CV to a moraic system, and once the language was moraic, syllable-final nasals could be generated even in the Yamato stratum (along with other moraic structures: geminated consonants and diphthongs).
However, the above aren't kun-readings proper, since verb inflections aren't normally written as kanji readings. Still, there are kun readings with ん in them:
- Some are the result of simple elision, like 何 nan < nani, 神 kan < kami, 懇ろ nengoro < nemogoro, or 女 onna < womina.
- Some are readings resulting from the morphophonological processes described above, like 冠 kanmuri < kauburi *[kãw̃ᵐburi], or 芳しい kambasii < かぐはしい *kangubasii.
- Some are ultimately Chinese morphemes that ended up counting as kun in the orthography, like 鞄 kaban < ?Ch. 夾板 *keap-paen; or the don of 丼 donburi (etymology unclear but probably Chinese don- + Japanese -furi; standard view traces don to 貪 *thom "covet" in 慳貪 *kʰˠɛn thom > J. kendon, "greed, gluttony", since kendon-ya is attested for donburi sellers; alternatively from 緞子 donsu, "damask").
- Frellesvig, A History of the Japanese Language.
- The Kokugo Daijiten.
- Baxter-Sargat Middle Chinese reconstruction.
- The Jōyō Kanji table and the Kanjidic file for kun-readings.