I’m doing a daily close reading of Murakami with my sweetheart and this came up:


She asks: Why Kyakkanteki Jijitsu, instead of Kyakkanteki na Jijitsu?

To recap: Japanese has two kinds of noun-like words, nouns proper and na-adjectives. These adjectives behave morphosyntactically pretty much like nouns, the main distinction being that na-adjectives modify nouns with na, whereas nouns use no:


no Jijitsu “humankind’s truth”

na Jijitsu “simple truth”

(Another major distinction is that na-adjectives don’t take “case particles”; that is, unlike nouns, adjectives can’t work in main argument positions, as subjects (-ga) or objects (-wo).)

Na-adjectives interest me because they have a number of borderline cases:

  • There are noun and na-adj pairs that sound identical but mean different things: Heiwa na kuni “Peaceful country” vs. Heiwa no kuni “The country of peace”.
  • In some other cases, a word can work both as a noun or as na-adj, with no clear distinction in meaning; Makkura no Kutsu = Makkura na kutsu = “pitch-black shoe”; Tokubetsu na Hito = Tokubetsu no Hito = “special person”. Some words prefer one class, to a greater or lesser degree; others are pretty much 50/50. (I once wrote a script to count this on ja.wikipedia; any of these days I’ll post it here.)
  • A few words that we’d expect to be adjectives are nouns: Byōki no hito “sick person” (person of sickness?), Haiiro no kutsu “ashen shoe” (shoe of ashen color?). The weirdness of these is likely an artifact of translation – of coming to Japanese with Indo-European expectations. Formally they’re just nouns.

Let’s get back to Kyakkanteki. -teki is a suffix that transforms nouns into na-adjectives, similar to English “-ive” in “objective”:

Kyakkan no Jijitsu Truth of object
Kyakkanteki na Jijitsu Objective truth

The Digital Daijirin dictionary agrees with this description of -teki, and even lists -teki words (including kyakkanteki) as regular na-adjectives. Which they are.

Except for one detail:

  1. Kyakkanteki na Jijitsu
  2. Kyakkanteki Jijitsu
  3. Inshōteki na Jijitsu
  4. Inshōteki Jijitsu
  5. Kantan na Jijitsu
  6. Kantan Jijitsu
  7. Akiraka na Jijitsu
  8. Akiraka Jijitsu
  9. Kyakkanteki no Jijitsu

The asterisk means that form doesn’t normally occur. My Japanese intuitions are wonky, but a cursory browsing of Google results, and asking a handful of natives, supported the analysis above. If this is correct, it follows that:

  • 9 shows that teki-adjectives aren’t dual-class words, like Tokubetsu, nor noun/adjective pairs, like Heiwa; they’re pure na-adjectives.
  • But 1–4 show that teki-adjectives may optionally drop the na (cf.). This counts as particle-dropping “headline style” (a native speaker tells me that “it sounds cool”).
  • And 5–8 shows that other na-adjectives cannot drop the na, even in headline style.

Therefore: teki-adjectives are a subclass of na-adjectives, distinguished by the fact that they (and only they) can modify nouns directly?…

Could this na-less usage be a direct echo of Chinese grammar, via writing? 的, Jap. -teki, in Mandarin is de, the genitive particle; it has the same function as Japanese no and na (plus a few others). 的 de isn’t the Classical Chinese genitive (zhī 之), but Schuessler says its attributive usage dates as early as Song, so certainly it could have influenced Japanese. One Japanese teacher I know jokingly calls headline-style Japanese “Chinese” (kanji piles with nary a kana particle in between); perhaps she’s onto something…

8 thoughts on “Teki-adjectives

  1. The Nihon Kokugo Daijiten says that, yes, it was a borrowing from Chinese (and quite a recent one at that — early Meiji):

    〔二〕名詞、特に抽象的な意味を表わす漢語の名詞や体言的な語および句について、体言、または形容動詞語幹をつくる。中国語の助辞の用法にならって、明治初期の翻訳文のなかで、英語のtic などの形容詞的な語の訳語として二字の漢語につけて用いられ出してから、学術的な文章などに多く用いられる。

    The relevant sub-definition:


    It seems that earlier in its history it was also used as an all-purpose nominalizer, which seems unnatural today but of course fits with how 的 is used in Chinese:



    It’s interesting because by the Meiji period you don’t really think of Chinese having much influence on Japanese any more — more the other way around, 漢語 translating new concepts from the west are coined in Japan and imported by China — but clearly the literati were still looking to Chinese as spoken in China, not just pre-imported morphemes, when looking for solutions to new translation problems.

  2. Question solved, thanks! I should save some monies for a copy of the Kokugo Daijiten…

    (2) above seems to be the source of the Wikipedia definition I linked to. I’d never have guessed that this usage would be as new as Meiji! I like how 「学術的な文章などに多く用いられる」 is self-demonstrating.

    I find a good number of results for “客観的を” and “圧倒的を” etc. – they’re a bit hard to Google because of quotative use (“「圧倒的」を”), but many appear to be positive matches. Is this an still-active nominal use of -的? Examples:

    • 誰かと一緒にいる方が圧倒的を楽しいわ
    • コレで今年も圧倒的を目指す。
    • 来年が、楽しみ。圧倒的を目指して、どれもどの舞台も。
    • 好印象的を持たれやすい人ですね。

    体言、または形容動詞語幹」 is looking better than Daijirin’s 「形容動詞の語幹をつくる」… I’m even finding more valid examples of “的の”, now that I look more closely…

  3. I had a quick look in Martin’s Reference Grammar of Japanese (most relevant pages are pp.762-3); he states the optionality of ‘na’ for -teki derived adjectives.

    Interestingly, he also says we may occasionally find -teki NO, rather than -teki NA used to modify. On a very quick check, it was quite difficult to get a sense of whether this is genuinely widespread nowadays — the only construction that seemed genuine is -teki no you da/dearu, which admittedly is not what would be expected for a standard na-adjective.

    Finally, if I can make a comment about your background assumption that things which modify nouns using -no must be nouns nouns while those using -na are adjectives. Sorry for my laziness in not citing my sources, but I do think there’s a case for saying that some adjectives (either obligatorily or optionally) use -no rather than -na to modify. If, for instance, compatibility with adverbs like totemo is a test of adjective-hood, then although nouns should fail it, I think -no-using adjectives would still pass: how about ??totemo okoko no hito / totemo byooki no hito ; totemo kiree na hito (maybe my judgements are off though…).

  4. Thanks for the comment, jo!

    I found more -teki nos than I was comfortable with on Twitter :)

    totemo is an interesting test because it groups “nouny” na-adjectives with “verby” i-adjectives (which is why Baker loves this test, since he’s trying to establish a clean, universal distinction between nouns, verbs, and adjectives in all languages (though he missed the fact that totemo is used with certain “stateful” verbs, e.g. totemo omowanai, which makes me think the criteria may be semantic, not syntactic)).

    I agree, it seems like “no-adjectives” take totemo and that’s definitely interesting. However, prototypically, na-adjectives to not take subject and direct-object position, while our mysterious “no-adjectives” generally accept -wo and subject-ga (otoko wo naosu; byōki wo naosu; *shizuka wo sagasu).

    The best I can do is to think that it depends on the tests you choose, and that, as Satoshi Uehara has argued, the categories are gradual (there are roots that ~prefer~ noun-like or na-adj-like morphosyntactic behavior, but accept both; others accept both equally, without a clear preference; and still others always choose one or the other). So “no-adjectives” are syntactic nouns with a hint of na-adj (mainly because their semantics lead to attributive use); makkuro-words live in the middle of the way, and can work as either noun or na-adj; kirei is a little bit more nouny than shizuka; and chiisai is an i-adj with a little bit of na-adjness in it (since it accepts chiisa-na but not *chiisa da). I’m preparing a longer post about this… eventually :)

  5. As a preview, here’s something I wrote in Portuguese after counting stuff, with the Japanese Wikipedia as a corpus: https://namakajiri.net/letras/2014/morfo/nakeiyosi.pdf

    The table in 3.4.2 is a count of words preceded by totemo and not marked as adjectives in EDICT. 3.4.3 are words not marked as nouns in EDICT, but found followed by -wo (though all were marked as meishi in the Daijirin); and the last table on page 22 compares some words followed by -no and -na in different proportions. This was a dumb count, without trying to properly analyze the whole sentences, but there’s still interesting results, I think.

  6. Thanks for the replies — very interesting. I was familiar with Miyagawa (1987) which is partly what I had in mind when commenting earlier, but much of the other stuff is new to me, and the data is very interesting.

    In principle I’m certainly sympathetic to this idea of categories as gradual. Tangentially related (and cross-linguistically rather than language-internally), I was always more convinced by the ‘spectrum of pronominality’, where Japanese pronouns are more noun-like than, say, English ones are, than by claims Japanese doesn’t have pronouns at all. I don’t mean to open up another thorny subject, though.

  7. @”categories as gradual”—i think Japanese pronouns are really a great and accessible example for “spectrums” instead of “hard” categories. After all, what is used in Japanese where an IE speaker would expect a pronoun can be surprising: often, there’s nothing; often, a name or a job designation is used; commonly, the same speaker will use a range of different words to identify themselves and others according to situation, cf the many words used for ‘I` (ore, boku, watashi, atashi, watakushi, honjin, …). Some of these words (like chichi used by a father when talking to his kids) only *function* as pronouns. It’s sure possible, even common to hear “daddy must go to work now” in an English-speaking family, but in Japanese, chichi certainly ‘blends’ in better with all the diverse ores and bokus of that language.

  8. “Totemo” is an interesting example because its contemporary usage as (basically) an adverb meaning “very” is itself a case of syntactic boundary-blurring. If you go back a few decades you can still find peevers moaning about the kids today, using “totemo” other than in conjunction with negated verb, etc.

    Naturally the story isn’t quite that simple, but it seems fair to say that even usages like “totemo kirei”, that are completely unremarkable today, would have at least sounded odd or slangy a few centuries ago.

    (“Zenzen” is currently undergoing a similar evolution, with the yoof of today getting hassled for saying things like “zenzen ikemasu yo” [or more likely “zenzen ikessu yo”, since the entire -masu form is on its way to being entirely replaced by -(Q)su as a non-conjugating politeness particle].)

Leave a Reply to Matt Cancel reply

Your email address will not be published. Required fields are marked *