For words like 帰ります is there a specific reason they're Godan instead of Ichidan in any case? Things like a change from an historical pronunciation of the word, grammar relating to the radicals of the character or is it literally random with each of these decisions genuinely being made arbitrarily? For 帰ります just for the sake of example, what was the original reason it's Godan and how many other いる/える Godan verbs have similar stories?


1 Answer 1


I think you are approaching this the wrong way. You seem to assume that all verbs that end in -eru or -iru should be ichidan. But there's no reason why that should be true. In linguistics there are no "shoulds". It's just the way that the verbs are.

Let's try looking at this from another angle. Consider the inflections for the ichidan verb kaeru = "to change" (変える). As you know, inflections are the part of the word that changes; in languages like English and Japanese, that happens at the end of the word. I will separate the part that doesn't change (the "base", "radical" or "stem") with an hyphen:

  • kae-nai
  • kae-masu
  • kae-ru
  • kae-re-ba
  • kae-yo

Now compare the stem for the godan verb kaeru, "to return" (帰る):

  • kaer-a-nai
  • kaer-i-masu
  • kaer-u
  • kaer-e-ba
  • kaer

Notice how, for the ichidan verb 変える, the base or stem ends in a vowel, -e. For the godan verb 帰る, the stem (=the part that never changes) ends in a consonant, -r. Because of this, linguists call the traditional "ichidan" verbs vowel-stem verbs, and "godan" verbs consonant-stem verbs.

So the fact that a verb has conclusive inflection (shūshikei 終止形, dictionary form) ending in -eru does not, by itself, tell us how it inflects. We need to know whether the stem ends in a consonant or in a vowel. "To return" has the stem kaer-, and "to change" is kae-. It's just that the conclusive form of vowel-stem verbs (ichidan verbs) adds -ru, so that, by coincidence, they end up sounding just like consonant-stem verbs whose final consonant happens to -r.

Why are some verbs vowel-stem, and some verbs consonant-stem?

We don't know. The oldest records we have of Japanese are from the Nara period, and at this stage the distinction was already present. In ancient Japanese consonant verbs used four extra vowels, not five (so they were 四段 yodan, four-grade, in Japanese terminology). And there were more kinds of vowel-stem verbs, with some rules which made the vowel change, and more irregular verbs, too. But the basic distinction between vowel-stem and consonant-stem verbs, those like kaer-u vs. kae-ru, goes back as far as we have evidence. For example, the verb kaer-u "to return" has been reconstructed as *kapyer- (かへる)¹, while "to change", kae-ru, had the ancient stem *kape- (かへ-).²

Some linguists have proposed that all verbs were originally consonant-stem, and vowel-stem verbs were derived from them. Others have argued for the opposite. As far as I can tell, there's no definitive answer yet.

Frellesvig, in A History of the Japanese Language (pp. 119–), notes two phenomena that happen with ancient vowel-stem verbs. The first is that, in many cases, the vowel-stem verb appears to be derived from an adjective, with a change in vowel (aka-ki "red" → ake-ru "redden, lighten"; sabu-si "lonely" → sabwi-ru "grow desolate"). In other cases, the vowel-stem verb appears to be derived from a consonant-stem, changing the transitivity (tuk-u "stick to" → tuke-ru "attach something"). These alternations (both the change in vowel quality, and the addition of an extra vowel) have been theorized (by Ohno, Unger and others) to come from some prehistoric suffix *⁠-y, which formed diphthongs like akay-, later changing into ake- etc.

If this theory is correct, the reason why 変える kae-ru is a vowel verb is that, a very long time ago, it was formed with this hypothetical *⁠-y suffix, possibly from *kapay- (compare 変わる kawar-u < kapar-u). And the reason why 帰る kaer-u is not a vowel verb is that it never had such a diphthong. But all of this is hypothetical; and even if it's correct, there are still many modern vowel-stem verbs which would not trace to these *⁠-y-verbs. The proposed *⁠-y formant would have created most of the vowel-stem verbs, but not all³ (e.g. mi-ru "to see" and ki-ru "to put on" appear to have been vowel-stems from the beginning).


[1] There used to be two kinds of e, called in Japanese linguistics the "A-type" (甲) and the "B-type" (乙). I'm following Frellesvig in interpreting A-type as ye, and B as a plain e. Also the はへひほふ sounds used to be pa,pe,pi,po,pu; in the middle of words they later became wa,we,wi,wo,u, then finally a,e,i,o,u (so kapyeru > kaweru > kaeru). The orthography changed, accordingly, from かへる to かえる. These ancient pronunciations aren't relevant for the discussion, so you can gloss over all that.

[2] One of the verb inflection rules was that, when two vowels met, you picked the second. So kapyer- plus the conclusive suffix -u resulted in kapyeru かへる; but kape- + -u > kape+u > kapu, かふ. The unchanged stem was visible when you added a consonant-initial suffix, like kape-te かへて, which didn't change the original vowel. This rule accounts for the "bigrade" (ニ段) verbs of traditional grammar.

[3] The verbs thought to be made with the *⁠-y formant are called "thematic". For those familiar with traditional grammar, the thematic class corresponds to lower bigrades 下ニ段, upper bigrades 上ニ段, and possibly the s-irregular サ変.

  • Could this also explain why vowel-stem verbs only end in -e or -i (which is this question)?
    – Earthliŋ
    Jul 9, 2018 at 9:57
  • 1
    @Earthliŋ Maybe! If I'm reading this correctly, all vowel-stem verbs descend from either a) thematic verbs once built from the formative *-y, or b) a handful of upper monograde verbs with a monosyllabic stem (mi-, ki-…). This brings to mind two questions: i) why the *-y element always resulted in -i/-e, and ii) whence the monosyllabic -i verbs, and how are they related to the thematic ones. I'd have to dig a bit deeper on the literature to answer that properly, though… Jul 10, 2018 at 7:05
  • @Earthliŋ, Boiko: Another possibility is that the 二段・一段 verb forms are defective developments from underlying 四段 full-paradigm verbs. We see this in the modern language primarily with 下二段 forms in verb clusters like つく・つける, with similar related developments also producing 五段 forms paired with 下二段 forms as in かかる・かける, both arising from earlier かく. Some of the oldest 上一段 verbs might also be noun derivations. Jul 11, 2018 at 16:28
  • @Earthliŋ, Boiko: For research on developments within the scope of historical Japanese, this paper presents a compelling case for one mechanism by which 下二段 verbs would develop -- essentially as an outgrowth of the potential / 自発 sense of an otherwise transitive verb, used to describe the agent performing the action. The 終止形 stem for 変える was かふ, suggesting overlap / cognacy with other かふ verbs (like 買う, 交う, etc.). The putative prehistoric //-y// mentioned above might be cognate with OJP / classical ゆ, the 自発・受け身 suffix. Jul 11, 2018 at 16:47

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .