Sometimes, sounds are lengthened for emphasis. For example, see "とっても versus とても".

What are the rules governing this process?

  • Are there restrictions on where lengthening can be inserted?
    とっても or とてえも or とても~?
    すんごい or すご~い or すごいー?

  • Is it predictable which sound is inserted?
    すんごい rather than す~ごい or すっごい?

I've given examples, but I'm hoping to learn if any general rules exist.

  • 2
    The sound of すごい can be lengthened to not only すんごい but also すっごい and すごーい. Commented Dec 6, 2012 at 22:46
  • There are rules, but they aren't all characterized strictly as lengthening. For example すごい->すんごい isn't a lengthening of any phone. If the medial /g/ is lengthened retrogressively than it results in gemination /Q/ not an epenthetical mora nasal /N/ which is the case here. The only dominant commonality in your examples is that the words are prosodically longer. So I'm guessing that's what you meant by lengthening; prosodic lengthening via epenthesis to the the moraic skeleton?
    – taylor
    Commented Dec 7, 2012 at 19:47
  • 1
    @taylor: Isn’t it possible to interpret the change from すごい to すんごい as a result of lengthening [ŋ] in [sɯŋoi], as suggested in alexandrec’s comment? Commented Dec 8, 2012 at 11:11
  • 1
    @TsuyoshiIto Going with my textbooks it can't be a lengthening of [ŋ]. It just seems that way. [ŋ] is an allophone first of all, which means it's a fully specified segment. It doesn't make sense to lengthen an allophone, but I don't think I know enough phonology to explain exactly why other than that's just not the defined role of allophones. I'll try to do a Q on this すんごい specifically and try to work out the details if I can.
    – taylor
    Commented Dec 9, 2012 at 2:26
  • 2
    @taylor: Thanks, but I do not know why that explanation is relevant. Both [ŋ] and [g] are allophones for the same phoneme /g/ in Japanese, but I have no idea why it should mean that the change from すごい to すんごい cannot be interpreted as lengthening of [ŋ]. (Just in case, I am not saying that this change actually arose in this way. I am just saying that considering this change in this way does not contradict any facts that I know.) Commented Dec 9, 2012 at 3:48

3 Answers 3


Maybe this is an overgeneralization, but I find this often has to do with pitch accent. Example:





It seems as if inserted sounds, if they exist, all appear on low to high pitch transitions. Obviously most words do not have these sounds inserted, and often only very common interjections would use them.

As for why ん is used rather than a っ, try pronouncing すっごい. It's pretty difficult, and since intervocalic "g" is allophonic with "ng" which sounds like ん, it is natural for すっごい to turn into すんごい.

  • Are you sure about the pitch of とても? It seems incorrect to me.
    – HAL
    Commented Dec 25, 2014 at 7:49

The sungoi is most likely due to the Tokyo nasal g accent. In Tokyo, oftentimes when there is a "g-" sound in the middle of a word, they'll add in a subtle "n" right before it. They'll probably still write it as sugoi because that's what they're saying, just with an accent. This i have learned from my time on WaniKani because one of the people they use for vocab pronunciation has the tokyo nasal g thing going on and it really confused me until someone explained it.

For the lengthening, it's pretty easily compared to english. when we want to emphasize a word, we stress the syllable that already has the most stress on it. try it with "totally" or "extraordinary." In japanese, they don't really use strong and soft syllables like we do, but changes in pitch. I'd compare raising pitch to beginning a strong syllable, so modifications for emphasis go where pitch increases.

Hope this helped! :)


In both of the examples you presented, it's the first consonant of the second syllable (or mora) that is doubled. Sometimes, I also hear dekai as dekkai. I've also heard verbs like して being pronounced as しって. That too matches the pattern of CVCV(x)>CVCCV(x).

  • So すんごい doesn't exist?
    – dainichi
    Commented Oct 17, 2012 at 0:01
  • Since g and ng (as in English -ing) are interchangeable, you can have ng-g, gg and double ng. However, you will never get g-gn.
    – alexandrec
    Commented Oct 17, 2012 at 13:31
  • I think it's dangerous to generalize this to such a simple rule -- because though it is true in many cases, it doesn't hold up under closer scrutiny. One example of this is the case of すごーい, where it is the /u/ that is being lengthened (incidentally, I believe すーごい to be acceptable as well, though すごいー is not). So while consonant lengthening on the second syllable is certainly one case of sound-based emphasis, I'm not sure it fully describes the situation.
    – rintaun
    Commented Nov 26, 2012 at 17:30

You must log in to answer this question.