Talk:Japanese grammar/adjectives

This page is for the discussion of 形容詞 and 形容動詞 as they pertain to the Japanese grammar article. The material was moved here because the old talk page was messed up. — Kaustuv 16:28, 2005 Jan 19 (UTC)

And I messed it up. My apologies again, and thanks for cleaning things up. Squidley

Adjectives?

Kaustuv and I have a good-natured discussion going on about what the proper terminology for keiyooshi and keiyoodooshi should be. He suggested taking it to sci.lang.japan, but I abandoned that group years ago because of its high volume of irrelevant postings. (Even so, there are intelligent and knowledgeable people contributing to the group, like Bart Mathias.)

Here's the issue. Keiyooshi are commonly called "adjectives," and keiyoodooshi are commonly called "na-adjectives" or "adjectival nouns." However, the term adjective refers strictly to those words that modify nouns and do nothing else. These words in Japanese not only modify nouns; they can also predicate sentences (that is, act as the verb). By definition, these words are not adjectives.

I, as a trained linguist and specialist in Japanese, think that it is important to use the correct terminology, and so "adjective" should be abandoned for something more accurate. Kaustuv, on the other hand, makes some good points with comments like this:

Are we so constrained by technical definitions that we cannot translate keiyoushi as "adjective"? Why can we not understand "adjective" in the context of a Japanese grammar as separate from some narrow definition of "adjective" preferred by some linguists?

He also notes that everything he's ever read about Japanese grammar uses the term adjective for these words.

Here's my response.

First, and I'm sure you know this, the number of people who believe a proposition has no impact on its validity (this is the logical fallacy called argumentum ad populum, and no, I'm not saying that Kaustuv has fallen for it). Even if every book written on Japanese grammar uses the term adjective, it does not make those words adjectives. However, it does put a strong burden on me to show that this usage is mistaken.

I use adjective in its technical sense (defined above), which is not necessarily the one generally understood by people at large. So why should non-specialists be forced to deal with unknown or unfamiliar terminology? Well, as I noted elsewhere on this page, using a convenient but wrong term can be harmful for students who become confused by unexpected behavior on the part of the words they thought were (for example) adjectives but really aren't.

Adjectives exhibit a certain range of behavior. If a word doesn't meet the minimum requirement--modifying nouns--it's not an adjective. (Verbs also modify nouns, but they do a lot of other things, too.) Adjectives may do some other things, like agreeing with the modified noun for gender. This is found in many Germanic languages and (to the best of my knowledge) all the Romance languages, but is not a necessary feature of the class. However, adjectives may not predicate sentences; that task is left to verbs. Keiyooshi and keiyoodooshi may predicate sentences, and therefore do not meet the definition of the term adjective.

So why not broaden the definition of "adjective"? Well, then you have a problem. You have some adjectives that predicate sentences, and some that don't. You have two classes of words that predicate sentences, rather than just one. It makes things messier.

But what about the other noun modifers, the verbs? Well, they can do a lot more, like inflect for agreement with the subject (in some languages); inflect for tense, aspect, mood, voice, and so forth; and predicate sentences. In fact, a sentence without a verb isn't even a sentence, but adjectives are not part of the minimum requirement for sentences. So the grammatical differences between verbs and adjectives are greater than their similarities. A little overlap can be OK, but it should be avoided when possible--and it's definitely possible here.

Not all languages have all word classes or grammatical structures. The Chinese languages lack adjectives entirely. Chinese, Japanese, and Korean all lack articles, as well as grammatical gender. Japanese and Korean lack pronouns. English has no postpositions. I don't see how it can be helpful to use incorrect terminology, even if it is familiar.

One of the important lessons that students of a language need to learn is that their target language is not just like their native language. They need to learn that it's not just the words and pronunciation that are different, but that languages differ in grammar and other important areas.

The term I (and at least some other linguists) use for keiyooshi is stative verb. Stative verbs describe a state, rather than an action, and have certain grammatical characteristics, like not inflecting for the continuative or imperative. There are two basic conjugational classes of stative verbs in Japanese. Some are keiyooshi, and others are regular verbs, like aru "to be (inanimate)." The difference between them is strictly in terms of how they conjugate; the difference between stative and dynamic verbs (the "other" kind of verb) is grammatical, not conjugational. This means that not all verb of a certain type need to conjugate in the same way; all they need to do is appear in the same syntactic frame.

So what? What's the deal? Well, as I see it, it boils down to accuracy versus popularity. As sociable as I am, I would rather be (professionally) accurate than popular. Squidley 00:46, 19 Jan 2005 (UTC)

I certainly cannot complain about accuracy; therefore, I shall ask you to explain one more thing before I pipe down. You have raised a point that "adjective" is a poor description of 形容詞. I should note that Martin, at least, shares your viewpoint to a certain degree: his "A" class in primarily a generative schema for "nuclear" or "predicative" sentences (pp 32--33). Shibatani on the other hand prefers the traditional view of adjectives a proper lexical class: "In addition to the major lexical categories of nouns, verbs and adjectives, ..." (p 215). Suppose I grant your point that "adjective" is inaccurate in an article on Japanese grammar. You still need to make a good case for "stative verb". I ask you to address the following points:

If "adjective" gives readers a false familiarity, then surely "verb" has the same fault! No 形容詞 has a causative or passive conjugation, and only a few of them have volitional forms. You can't take a 形容詞 and construct respectful forms like with verbs: お(V-連用)に{する|なる}. No 形容詞 has a desiderative conjugation like with verbs: (V-連用)たい.

It conflates 形容詞 with the very real class of 動詞 that are "stative" because they don't permit all aspects to be marked. Explain to me how translating both 状態動詞 and 形容詞 as "stative verb" clarifies anything. Do you suggest subordinating 形容詞 under the banner of 動詞? Do you simply deny that 動詞/形容詞 is a valid lexical distinction? Either view is radical enough that I ask you again to furnish citations. Note: asserting your qualifications (which, by the way, the reader has no way of verifying) isn't a citation.

Not all 形容詞 are stative. Take any dynamic verb and negate it, for instance: 呼ばない (in the sense of "doesn't call", not "won't call").

If you want to persist with banishing "adjective" from the article, then I think you should pick some other term for them than "stative verb". This will be my last message on this topic; if you continue to use "stative verb", I shall wash my hands off this article. — Kaustuv 16:28, 2005 Jan 19 (UTC)

First, thanks for cleaning up my mess. I really appreciate it, and I'm sorry for messing up the page in the first place.

Second, I wouldn't dream of saying something like "because I'm an expert and therefore I'm right." (For those just joining us, I have a Ph.D. in Linguistics and am a professor of Japanese.) It's the old "appeal to authority" fallacy (argumentum ad verecundiam), and while I am an authority in some sense, it is an invalid argument because there is disagreement among the the experts.

As for citations, I have to admit that grammar is not my forte; I'm more of an interested outsider. (My specialities are phonology and historical linguistics, but I was trained in grammar as well.) As I (unfortunately) have limited access to the appropriate specialist literature at my current institution, I'd have an awfully hard time coming up with citations.

I'll try to address your other concerns later. I recognize that you're an intelligent, open-minded person, and I'll do my best to show you that the view I'm putting forth is valid, but I'm not sure I can do much more than I already have--but I'll give it a shot. Squidley 22:54, 19 Jan 2005 (UTC)

Here, I’ll respond to Kaustuv’s concerns point-by-point. I’ll preface his comments with a K; my comments follow immediately below.

K: If "adjective" gives readers a false familiarity, then surely "verb" has the same fault!
Actually, using "verb" has precisely the right effect. It alerts the learner that these words are not adjectives, but verbs—which is exactly the point.

It has an effect certainly, but one of begging the question. Kaustuv 23:13, 2005 Jan 21 (UTC)

It points out that these things inflect and predicate sentences—the very functions that verbs perform.

YES! Call 形容詞 "adjectival predicates" or something like that! I'll happily cheer you on if you mean to actually talk about what predicates, "restrictives" (as Martin calls them), etc. are. That would be an actual improvement in clarity in the article. Kaustuv 23:13, 2005 Jan 21 (UTC)

We agree in substance but not in terminology. I think I can make changes that we can both live with. Squidley 23:17, 24 Jan 2005 (UTC)

I would argue that the best way to translate individual keiyooshi is by adding "to be" before the English adjective, like this:

akai "to be red"

Doing this gives a better understanding of Japanese sentences, e.g.,

sora ga akai

sky SUBJECT be-red

"The sky is red."

sora ga akakatta

sky SUBJECT be-red-PERFECT

"The sky was red."

There is no single word in the Japanese sentences that corresponds to English “is” or “was,” yet these are well-formed Japanese sentences. Like all well-formed sentences, they have predicates. Those predicates are none other than the keiyooshi. Add ‘’da’’ to these sentences, and they become grammatically unacceptable. In contrast, the English sentences without “be” verbs are unacceptable. This shows that keiyooshi are verbs.

K: No 形容詞 has a causative or passive conjugation, and only a few of them have volitional forms.
While it is true that the keiyooshi do not conjugate for causative or passive, causation and passivization can be achieved periphrastically.

Yes, certainly, but the conjugations are not parallel. No one is arguing that you can't coax 形容詞 into these roles with auxiliaries like する, なる, おもう, etc. Kaustuv 23:13, 2005 Jan 21 (UTC)

The fact that keiyooshi inflect 'at all is the point. True adjectives may agree with nouns, but they do not, in and of themselves, have any sort of past/perfect/future/continuative/conditional/whatever. Different paradigms for inflection is not a basis for saying they aren't verbs. Squidley 23:17, 24 Jan 2005 (UTC)

This means that instead of there being a special inflected form for something, you get the meaning by adding words. For example, English verbs have only two forms for tense: past and non-past. We express future periphrastically: I will go, I am going to go. In Japanese, causative and passive for keiyooshi are also periphrastic, and are accomplished thus:

Causative: akaku suru

e.g., Taiyoo ga ringo o akaku shita.

“The sun has reddened the apples.”

Passive: akaku naru

e.g., Kono ki no ringo ga akaku natta.

“The apples on this tree have reddened.”

When you say “volitional," I assume you mean things like yokaroo. I don’t think that volitional is the right term for this, because even though this form is similar to that for dooshi, e.g., ikoo, the meaning is different. That’s because the semantics are different. But if you’re trying to say that “only a few” keiyooshi have such forms, that, I’m afraid, is incorrect: all keiyooshi have this form.

What I meant is that only a few 形容詞 are semantically volitional. 良かろう is close to the only one I think of. Kaustuv 23:13, 2005 Jan 21 (UTC)

K: You can't take a 形容詞 and construct respectful forms like with verbs.
Keiyooshi do, in fact, have polite/respectful forms. However, with the exception of ohayoo gozaimasu (literally, “to-be-early-POLITE”) and omedetoo gozaimasu (literally, “to-be-auspicious-POLITE”), they are exceedingly rare in the modern language. Their formulation isn’t too tough: some have an optional o at the beginning, they have gozaimasu at the end, and the keiyooshi itself is in what we might call the “Kansai adverb” form, which is the adverb in -ku, with the k dropped and the vowels assimilated, ombin-like, to each other. More examples:

ookii : ookyuu gozaimasu
takai : (o)takoo gozaimasu
muzukashii : (o)muzukashuu gozaimasu
ooi : oo gozaimasu (an exceptional form)

Again, this doesn't answer my point. The respectful constructions are not parallel to 動詞. Kaustuv 23:13, 2005 Jan 21 (UTC)

And again I say that what matters is not the paradigm in which the word inflects, but that it inflects at all. Squidley 23:17, 24 Jan 2005 (UTC)

K: No 形容詞 has a desiderative conjugation like with verbs.
Desiderative forms are also made periphrastically. Let’s just remember that desideratives necessarily have animate subjects.

”Regular” verb

migaku : migakitai "to polish" : "[I] want to polish"

Keiyooshi

kuroi : kuroku shitai "to be black" : "[I] want to blacken"

So what does this show? It shows that the inflectional criteria you outlined are, in fact, met by keiyooshi. That they are met in different ways only shows that they belong to different inflectional paradigms, and does not show that they are not verbs. In fact, given all the verbal inflecting they do, I don’t see the basis for denying that they are verbs.

K: It conflates 形容詞 with the very real class of 動詞 that are "stative" because they don't permit all aspects to be marked. Explain to me how translating both 状態動詞 and 形容詞 as "stative verb" clarifies anything.
Conflating keiyooshi and that class of regular dooshi that are also stative shows that they have similar limitations on their inflections. The main criterion for determining whether or not a verb is stative is syntactic: does it have a continuative/progressive form? Can it appear in the imperative? If not, it’s stative. It doesn’t matter if the inflectional paradigms differ. (Just so we’re all on the same page, “dynamic” is the opposite of “stative.”) Let’s look at a couple of English verbs:

dynamic to carry:

PROGRESSIVE I am carrying it

IMPERATIVE Carry it!

PAST I carried it

stative to know:

PROGRESSIVE *I am knowing it (* means it’s unacceptable)

IMPERATIVE *Know it!

PAST I knew it

Note that the past tenses are formed differently; this is because they are part of different inflectional paradigms.

Now compare those with Japanese verbs:

dynamic kiku “to hear”:

CONTINUATIVE kiite iru “[I] am listening.”

IMPERATIVE kike “Listen!”

PERFECT kiita “[I] heard.” (What's called "past tense" in Japanese is, in my opinion, perfect aspect, but that's another story)

stative (dooshi-type) aru “to be INANIMATE”:

CONTINUATIVE *atte iru

IMPERATIVE ?are (shows up in some set phrases but is non-productive and does not seem to be imperative in meaning); (? means it’s of questionable acceptability)

PERFECT atta “[It] was.”

stative (keiyooshi-type) akai “to be red’’:

CONTINUATIVE (no form)

IMPERATIVE (no form)

PERFECT akakatta “[It] was red.”

What it clarifies is that keiyooshi inflect and predicate, just like other verbs. It shows that their grammatical behavior is not like that of adjectives. I think that’s a pretty important thing to know--don’t you?

Important certainly, but again calling them "verbs" is not as important as highlighting the differences between "adjectives in English" and "adjectives in Japanese". Kaustuv 23:13, 2005 Jan 21 (UTC)

Which is the crux of our disagreement. Keep reading! Squidley 23:17, 24 Jan 2005 (UTC)

K: Do you simply deny that 動詞/形容詞 is a valid lexical distinction?
Dooshi versus keiyooshi is, to my mind, not a lexical distinction but an inflectional one.

A point that is not corroborated by anything in the literature. Really now, can it be so hard for you to find a single citation that agrees with you here? If your point is so valid, why haven't you published a paper on it? I would think that far more important than editing a Wikipedia article. Kaustuv 23:13, 2005 Jan 21 (UTC)

An excellent point. I think that Hisami Springer's 1993 University of Hawai'i Ph.D. dissertation, "PERSPECTIVE-SHIFTING CONSTRUCTIONS IN JAPANESE: A LEXICASE DEPENDENCY ANALYSIS", probably calls them verbs, because her advisor is the person who convinced me that they're verbs and not adjectives.

Thanks! My university's library doesn't appear to carry it. I will check in U. Pitt's East-Asian library tomorrow. Their collection is many times larger than CMU's. I have modified the article to use "i-type adjective" and "na-type adjective", following the recomendation of Bart Mathias to categorize them based on their inflections, but this is a place-holder. I would agree to calling 形容詞 "predicate adjectives". I think "verb" to too many people denotes "action"; for a while I was thinking of suggesting "verbal adjective" as a compromise, but then we are in a quandary about 形容動詞 which can't be "verbal (ad)nouns" as there are a class of nouns that take "-suru" that deserve the name "verbal noun" a lot more. Anyway, from my reading of Crystal's definition, he says that the four criteria are for English, and different criteria might apply to other languages. This is a good out for my side of the argument: 形容詞 are just different. — 00:56, 2005 Jan 25 (UTC) (That was by me, Kaustuv; for some reason Wikipedia is messing up my signature line.)

No luck, sadly. I'll write to Hisami Springer. — Kaustuv 04:03, 2005 Jan 26 (UTC)

As for publishing an article myself, my specialties are phonology and historical linguistics, but I implicitly and explicity call them verbs in my papers.

One set inflects in one way, the other, another. But both are still verbs, because both inflect and predicate sentences--two of the primary identifiers of verbhood. (Another criterion is selection of arguments, but I don't see how it really helps here.)

K: Either view is radical enough that I ask you again to furnish citations.
As for citations, as noted above, sadly, I do not have access to the right kind of materials at my current institution’s library, and since grammar isn’t my specialty, I don’t have appropriate references in my personal library. I’m simply doing this off the top of my head, based on the principles I learned and my training. If my arguments are bad, please point out their weaknesses. You might see something that I’m missing.

Wikipedia is not the place for radical departures from the zeitgeist. Please understand this.

This is the most convincing argument you've made. I concede.

I don't think asking for citations is unreasonable. At least use your real name and institution so I can check that you are indeed speaking the truth about your qualifications, etc. You can't wear your academic hat and simultaneously use the "don't have access" excuse. — Kaustuv 23:13, 2005 Jan 21 (UTC)

One of the ways I practice "safe Internet" is by remaning anonymous. A friend of mine has a blog that came up and bit him in the @$$. The reason? The political opinions on his blog are not the ones most popular with most people in academia. Although he's a brilliant scholar and a gifted teacher, one who doesn't bring his politics into the classroom, his association with his opinions has cost him at least one job. Since I don't have tenure either, I prefer to stay anonymous. Academic freedom isn't free until you have tenure. (Bart, on the other hand, is retired, so he can say whatever he wants.)

Also, as I'm sure you know, not all institutes of higher education are Class I Research Institutes, granting Ph.D.s in dozens of fields. Restricted library collections are the norm, not the exception, in academe. Squidley 23:17, 24 Jan 2005 (UTC)

K: Note: asserting your qualifications (which, by the way, the reader has no way of verifying) isn't a citation.
As I’ve said before, I don’t believe that asserting my qualifications is an argument. I hope that I haven’t said anything even remotely resembling “because I said so” here. But I am going by definitions. Adjectives don’t predicate sentences. This is one of the basic characteristics of this word class. This is not a definition I created; it’s the one used by linguists around the world. Maybe we need to change this definition. If that’s the case, please make your arguments for why this definition is inadequate or inappropriate. I’ll listen.

K: Not all 形容詞 are stative. Take any dynamic verb and negate it, for instance: 呼ばない (in the sense of "doesn't call", not "won't call").
This, I can’t agree with. The plain negatives are stative verbs. Do they have continuative forms? No.

Taking a leaf out of your own book: 呼んでいない.

Do they have imperative forms? No.

呼ばないでくれ

Good points. Time to think... Squidley 23:17, 24 Jan 2005 (UTC)

Do they inflect like other keiyooshi? Yes. Do they meet the definition of stative verbs? Yes. What grammatical reasons are there for not calling these stative verbs? I can’t think of any--can you? All the plain negative forms of all verbs are stative! It was a shocking realization for me, but in time, I learned to accept it.

Now, I understand that virtually everything you have read calls keiyooshi adjectives.

Not just virtually. Everything. Including plenty of Google searches. Kaustuv 23:13, 2005 Jan 21 (UTC)

I’ve read most of the same things, and in none of them do the authors ever justify this label.

Martin justifies them in excruciatingly great detail. That is why I wish you would follow his strategy of talking about predicate formation in Japanese instead of this silly game of calling 形容詞 verbs. Kaustuv 23:13, 2005 Jan 21 (UTC)

Sometimes, the status quo is wrong. I’ve given the standard, accepted definition of both adjective and verb, and I believe I have shown that keiyooshi fit the latter, not the former.

K: If you want to persist with banishing "adjective" from the article, then I think you should pick some other term for them than "stative verb."
Why? Based on what principle(s)?

Based on the principle that Wikipedia is not a place for grinding axes. So far you have done nothing to dispel this image. Kaustuv 23:13, 2005 Jan 21 (UTC)

As noted above, you're right. I don't think I'm grinding an axe on this point, but maybe I am. Regardless, my heterodox (yet correct! ;-) views are perhaps not best aired on Wikipedia. Squidley 23:17, 24 Jan 2005 (UTC)

If you can demonstrate why they are not stative, I’ll be the first to admit my error and change my terminology.

My objection all along has been with "verb", not "stative". Call them "stative predicates", and I won't have an objection left. Kaustuv 23:13, 2005 Jan 21 (UTC)

To paraphrase Galileo, and yet they are verbs. Squidley 23:17, 24 Jan 2005 (UTC)

K: This will be my last message on this topic; if you continue to use "stative verb", I shall wash my hands off this article.
Well, that’s one approach, and I can’t say it’s the wrong one, but it seems uncharacteristic of you. I’m very impressed by the way that you were able to acknowledge that Japanese doesn’t have pronouns. I would think that you would be equally open-minded about other issues as well.

Having an open mind and taking an anonymous voice on the Internet seriously, and especially when all other sources disagree, are different things. I conceded "pronoun" because Martin and Shibatani both essentially agreed with you. Not the case with 形容詞. Kaustuv 23:13, 2005 Jan 21 (UTC)

Well, I will await your comments. I will also start my revision of the Adjective section, but your reaction shows clearly that I will have to be careful to make the reasoning clear. I think I’ll start by calling them adnominals, even though this label isn’t 100% accurate.

As an aside, I will note that Japanese does have some true adjectives, like chiisana and onaji, yet I find it interesting that most authors ignore them, and they don’t even seem to have a proper name of their own--just rentai. Squidley 19:26, 21 Jan 2005 (UTC)

I just remembered something regarding that last point on my way home. Martin does talk about 大きな etc. Page 747 in the discussion of "restricted nominals; adnouns" says words like 大きな, 小さな, おかしな, etc.

[...]are stylistic options in place of the regularly adnominalized ookii [...] this option is available only if the epitheme is extruded: ookina hako 'a big box' but not *ooki-na hazu da 'ought to be big'. [emphasis mine]

Though, in characteristic form, he also has the following confounding passage in page 754:

NOTE: The following adjectival nouns also form adjectives, as shown: koma-ka, komakai 'fine'; atata-ka, atatakai, 'warm'; [...] okasi(-)na, okasii 'comical'; tiisa(-)na, tiisai, 'small'; ooki(-)na, ookii, 'large'—the last three adjectival nouns are highly defective. [emphasis mine]

To understand what he means by 'defective', you have to see page 787:

Adjectival nouns which occur only in the adnominal (... na) and the infinitive (... ni) but lack the predicative form, i.e., a nuclear sentence (* ... da) from which they can be derived.

So there you have it. Your point that these are "adjectives" in the sense that they cannot predicate sentences is most certainly known to Martin. Kaustuv 01:10, 2005 Jan 22 (UTC)

On re-reading: I sound more angry than I should be (and am). This is perhaps a good indication that I ought to get out, as I should have done a few rounds ago. Have fun with this article. (I mean that sincerely.) Kaustuv 23:31, 2005 Jan 21 (UTC)

I think I, too will find other ways to pass the time. Squidley 23:17, 24 Jan 2005 (UTC)

As a matter of curiosity, I have asked this question to sci.lang.japan, trying to phrase the issue neutrally. Feel free to join if you can tolerate the banter for a while. — Kaustuv 00:04, 2005 Jan 23 (UTC)

In particular, read this response. Kaustuv 18:37, 2005 Jan 23 (UTC)

And this one. I think it is pretty clear now that you have a very weak case. — 20:56, 2005 Jan 23 (UTC)

OK, thanks for the links. Thanks for challenging me to rethink my views, and to argue for them! It has honestly been a lot of fun, as well as intellectually stimulating. Until we meet again... Squidley 23:17, 24 Jan 2005 (UTC)

I checked out the thread on sci.lang.japan; it was of much higher quality than I expected. Who knows--I may even join the fun! Anyway, here's my rejoinder:

1. Definition of "adjective" from Crystal: "they can occur in a post-verbal of *predicative* position, e.g., /the man is *big*/"

Let's ignore word order and concentrate on the important fact that adjectives in predicative position are accompanied by verbs. Keiyooshi predicate sentences by themselves. In comparison:

sora ga akai.
*sora ga akai da.
*The sky red.
The sky is red.

QED. (I hope I don't sound upset--I'm not!)

(As an aside, I see that you doubt whether be-verbs are truly verbs or just "auxilliaries." I'd say they're verbs--in fact, main verbs, since they're the ones that inflect, agree, etc. when present in a sentence, and their subordinate arguments are limited in their inflection. Notice their full designation: "auxilliary verbs." But this is another point.)

2. Further definition of "adjective" by Crystal: "However, not all adjectives satisfy all these criteria (e.g., /major/, as in /a major question/, does not occur predicatively - cf. /*The question is major/)"

In other words, some adjectives have defective distribution, but all verbs may predicate sentences. All keiyooshi may predicate sentences. QED.

As for sentences like "She fat," that's another language--not English. It has different rules. In Ebonics/Jive/Black Vernacular English/whatever you want to call it, the words that are adjectives in English are--are you ready?--stative verbs! Russian, mentioned in the thread, has--get ready--predicate nouns! Hmm. Maybe I'm coming around to "predicate adjectives"--but just maybe!

Someone in the thread got their knickers twisted about imposing Indo-European language categories on non-IE languages. I agree. In fact, I'm getting my panties in a knot over that point exactly. I've been saying that not all languages have all word categories, translation isn't classification, etc. To a certain degree, it's about terminology. As far as I can tell, we agree that keiyooshi aren't adjectives; we just fail to agree on what they are. Oh well.

As for Martin's reference grammar, my copy of it is in the middle of the Pacific, in a storage locker in Honolulu, and I am on the West Coast, so I'm not able to check it right now. However, your reading of Martin seems to be excellent. I especially liked how you mentioned "in characteristic form, he also has the following confounding passage." That's great! You are correct about Martin knowing that there's a set of what I call "true adjectives," but he's one of very few who acknowledges them. Squidley 00:29, 25 Jan 2005 (UTC)

summary/comment

Hi. Interesting comments. Thanks.

My summary, comment, & questions:

Why should keiyoshi be called adjectives? It seems that keiyoshi & keiyodoshi are called adjective due to their semantics. Am I right? If so, this is generally not a good way to define word categories (i.e. parts of speech) since semantics is so unreliable. Is there any morpho-syntactic criteria to support this grouping? If there is nothing syntactic or morphology, then how can the adjective category be valid? Perhaps the word choice of adjective is a pedagogical issue?
The grouping together of verbs & keiyoshi into a single category seems to be based on primarily morphological criteria, namely the possibility of inflection. This category is then subdivided by using various morpho-syntactic criteria, resulting in things like Verb_{stative-adjectival}, Verb_{stative-nonadjectival}, Verb_non-stative, etc. Is my understanding correct?
There is a question of whether the difference in inflectional paradigm between keiyoshi and other inflected words is great enough to warrant two separate categories. Am I correct here?

A further comment on textbooks: Many grammar books of English are rather unsophisticated in their combination of both semantic and syntactic criteria to define word classes. So, even if most Japanese textbooks do the same, this is hardly convincing to a linguist. Hopefully, the linguistic reference grammars will not do this.

Verbs as adjectives: As is probably already known by some, not all languages have adjectives (i.e. a separate & distinct word class that functions adjectivally). So just because a given word must be translated into English as an adjective doesnt mean that that the given word is an adjective or that there must be an adjective word class. I just want to re-emphasize this.

In short, after reading the above, I dont see why the keiyoshi & keiyodoshi should both be called adjectives, and I dont see that considering keiyoshi to be certain type of verb would be such a bad proposal.

Anyway, thanks for the discussion. Peace. — ishwar (SPEAK) 19:31, 2005 Apr 10 (UTC)

Again, if you want to call keiyoushi verbs, then you should furnish citations. Until then, the standard followed by every book I've ever encountered is to call keiyoushi adjectives, and I don't see why Wikipedia should follow any non-standard classification. (Incidentally, please be careful when editing: you had inadvertently stripped a large chunk of initial text.) — Kaustuv 22:02, 2005 Apr 10 (UTC)

Hi. So, are you only opposed to using the term verb to encompass the inflected word classes? Why not call all of these predicates? These names are really just labels of word classes after all (some structuralist linguists simply used terms like Class 1, Class 2, etc.). Then, you can have something like:

Predicates
1. Verb (i.e. predicate type 1)
2. Adjectival predicate (type 2) = keiyoshi
3. Copula (type 3)
Nouns
1. Noun (type 1)
2. Adjectival noun (type 2) = keiyodoshi
etc., etc....

What I see is that user Squidley is defining the term adjective based on morpho-syntactic criteria and this is dispreferred (by Squidley) since some readers/learners may mistakenly equate what are termed adjectives in Japanese as the same as adjectives in other languages. I think that many other authors are using the term adjective that is probably based mostly on a semantic definition. However, you are objecting to new terminology since it would be new and/or uncommon.

I think that you two are coming to an agreement in terms of the dissimilarity in morpho-syntactic behaviour between the keiyoshi & keiyodoshi. Right? I make the suggestion that you both could use the term predicate instead of verb to encompass the entire inflected-able-to-be category. You could use the word adjectival in adjectival predicate and adjectival noun to connect this to the popular terminologies of i-adjective and na-adjective (to be consistent you may need to call the other non-keiyoshi stative verbs adjectival verbs or stative-adjective verbs)

Opinions? — ishwar (SPEAK) 01:34, 2005 Apr 11 (UTC)

I would be fine with, and would in fact encourage, calling both verbs and adjectives predicates. In fact, your framework is almost exactly how tango would be classified in any kokugo textbook. The difference is for keiyoudoushi, which are also predicates in traditional grammars. The traditional treatment is to lay out words in two axes: "things that can predicate" (PRED) and "things that inflect" (INFL). Keiyoushi and doushi are (+PRED,+INFL); keiyoudoushi are (+PRED,-INFL); meishi, fukushi and joshi are (-PRED,-INFL). Of the (+PRED,+INFL) class, the difference between keiyoushi and doushi is entirely based on how they inflect and what inflections are possible. (Note the oddity that keiyou suru meishi are keiyoudoushi.)

I said almost at the beginning of this discussion that a proper treatment of predicates (and restrictives) would be a worthy addition to this article. Sorry to be too repetitive, but my opposition to calling keiyoushi "verbs" begins and ends with the observation that no major text on Japanese linguistics in the English language—not Shibatani's The Languages of Japan, not Kuno's The Structure of the Japanese Language, nor even Martin's A Reference Grammar of Japanese—calls keiyoushi "verbs". Squidley's definition of adjective is, as far as I can tell, Crystal's definition, which specifically lays out the criteria for English. Obviously what works for English won't work for Japanese. Now, there is a fairly important point to be made here about the coarseness of the popular verb/adjective lexical distinction, but calling keiyoushi verbs is, in my opinion, not the right approach. — Kaustuv 02:51, 2005 Apr 11 (UTC)

Negatives

First, to anchor my comment to the pertinent text:

---
K: Not all 形容詞 are stative. Take any dynamic verb and negate it, for instance: 呼ばない (in the sense of "doesn't call", not "won't call").
This, I can’t agree with. The plain negatives are stative verbs. Do they have continuative forms? No.

Taking a leaf out of your own book: 呼んでいない.

Do they have imperative forms? No.

呼ばないでくれ

Good points. Time to think... Squidley 23:17, 24 Jan 2005 (UTC)

---

One important point to note with the "continuative" example noted above is that the continuation is not of the negative ending nai, described as a jodōshi in the copy of Shogakukan's Kokugo Daijiten that I have on hand, but of the preceding verb stem itself, as yonde iru – the nai remains just that, ない, and gets stuck onto the end of the verb iru.

Likewise, the "imperative" example above does not exhibit an imperative form of nai, as the imperative is wholly contained within the ending verb kureru. The de in the middle isn't even a conjugative de, such as the -te in nakute, but is rather the simple copula de. I've never heard a good explanation for why negative requests use de instead of the -te form, but perhaps it has something to do with the stative nature of nai?

Just my ¥2 here --- Eirikr 04:50, 4 Apr 2005 (UTC)

Thanks for your comment. Those "continuative" and "imperative" forms were not serious suggestions, but rather to demonstrate to Squidley that if one should allow periphrastic marking of aspects as a form of 'inflection', then there is no reason to claim that keiyoushi have no 'continuative form', etc. Kaustuv 07:18, 2005 Apr 9 (UTC)

My point all along has been that keiyooshi conjugate, predicate, and do other verby things. It actually matters little in what way they are defective as verbs; what matters is that they meet the primary criteria. In my estimation (and that of at least some other linguists) keiyooshi act like verbs, so that's what we call them.

Kaustuv's point (or at least one of them) is that this is not the status quo, and that therefore the practice of calling keiyooshi "verbs" is inappropriate for Wikipedia. On this point, we agree, though I am happy to see the term sneaked in to the keiyooshi section.

It appears that since everything Kaustuv has read does not use the term verb when describing/defining keiyooshi, he does not care for that term. I have tried, in vain, to convince him that in spite of this, they really are verbs. I know he is willing to change his mind when given a good argument--he came around to the position that Japanese has no pronouns--but for whatever reason(s), I just can't seem to convince him on this point. Oh well. The process has been fun anyway. Squidley 22:20, 18 Apr 2005 (UTC)

I have repeatedly said that I am willing to be convinced by citations that argue that keiyoushi should be called "verbs". I fully concede that my understanding of the issues is shaky and incomplete, and I wouldn't be surprised to hear that Martin's and Shibatani's taxonomies, to pick two prominent examples, are superceded by more recent treatments. Not being a linguist myself, I've spent a lot of time searching for such a citation on Google, and failing that, sparked a discussion on the newsgroup sci.lang.japan. This second tack seemed to give a convincing argument for sticking with the traditional taxonomy. (I never received a response from Hisami Springer to my initial query.) My impression is that this usage is not popular in the literature, so it is too controversial for a Wikipedia article.

Unfortunately, the process has long stopped being fun for me. I think I have said the same things a dozen times by now. I'll be content with whatever compromise is cooked up. — Kaustuv 22:57, 2005 Apr 19 (UTC)

Did you notice that I found a mainstream reference that calls keiyooshi verbs? McClain's 1981 Handbook of Modern Japanese Grammar, page 87. Still, it's just a drop in the bucket.

I agree with Kaustuv that calling keiyooshi "verbs" is inappropriate for Wikipedia, but I remain happy to discuss why I think that term is not correct. Squidley 23:05, 19 Apr 2005 (UTC)

Thanks for that reference. I tracked down McClain's book, and here is the relevant extract from p.87:

ADJECTIVES (VERBAL ADJECTIVES)

As attributive adjectives, Japanese adjectives function like their English counterparts—that is, by preceding the nouns they modify. But as predicate adjectives they differ from English in that they function as verbs. For example, akai in the sentence Kono ringo wa akai [...] "This flower is red," means not just "red" but "is red." In the sentence, Kono hana wa akai desu, desu does not have a copulative verb function, but rather is used to make the ending level more polite [...]. Because Japanese adjectives are thus a special class of verbs, they conjugate. [...]

I must say that I remain unconvinced. Firstly, McClain's argument that keiyoushi are verbs appears to beg the question. Secondly, she says that keiyoushi are verbs therefore they conjugate, when I think the more correct statement (which I believe is also Squidley's argument) is that they are verbs because they conjugate. Thirdly, I think she means that Japanese adjectives are a special class of Japanese verbs, because verbs in a language like English don't occur in the attributive position (or do they?). Lastly, I think her "red"/"is red" distinction echoes Squidley's point that the predicative akai is stative, but I am not sure if she considers "is" to be the verb in "this flower is red". From my brief skim through her book, her grammar is fairly non-traditional in many other ways—for instance, counting keiyoudoushi as meishi, or treating the volitional conjugation as a separate stem form—but she claims that her presentation is better for the student. She might well be right. I will agree that something needs to be added to the main article to explain this reconstruction of keiyoushi. —Kaustuv 19:33, 2005 Apr 20 (UTC)

True adjectives

This showed up as a question over on the Talk page for Japanese language – what are "true" adjectives? The discussion above gives some possible examples, but I note that chiisa-na is basically a na adjective, with the alternate and somewhat more common i adjective form of chiisai, whereas onaji is to some extent a corruption from an i-type, as evidenced by its adverbial form of onajiku. Which would seem to suggest that all "adjectives" (as generally discussed in teaching materials) fall into two categories, the i-adjectives and the na-adjectives. Is there something I'm missing here?

Cheers, --- Eirikr 04:50, 4 Apr 2005 (UTC)

Martin (and I) disagree with the assessment that chiisa-na is "basically a na adjective". Martin describes it as a "highly defective" precopular (ad)nominal (IIRC). I believe there is a citation of that in the mess of text above. Basically, chiisa-na does not allow a whole slew of markings that you would get with a sensible na-type adjective such as shizuka: contrast *chiisa-ni vs. shizuka-ni, or *chiisa-na hazu vs. shizuka-na hazu, etc. I would claim that chiisa-na is a (restricted) stylistic variant of chiisa-i, which is the "true" adjective. — Kaustuv 07:26, 2005 Apr 9 (UTC)

Good point -- now that I think it through fully, I realize I've never heard the phrase chiisa-de either.

However, I'm still puzzled as to what the so-called "true adjectives" would be. Given your point above, it would seem we still only have na and i adjectives, with some irregulars (ooki-na, chiisa-na, and the truncated onaji). What words would fall under this third category of "true" adjectives?

Thanks, --- Eiríkr Útlendi 02:59, 10 Apr 2005 (UTC)

I have no idea what this category of "true" adjective is; it's not a class I have advocated for, nor particularly care to define. As far as classification is concerned, we have to ask ourselves whether we want to err on the side of detail or simplicity. Assuming we prefer detail, here's how Martin chunks up the keiyoudoushi class (i.e., na and no-type adjectives):

adjectival noun

Eg. shizuka-na, hade-na

adjectival noun, defective

marks only adverbial (-ni) and adnominal (-na). Eg. setsu-na

adjectival noun, highly defective

marks only adverbial (-ni) or only adnominal (-na). Eg. tadachi-ni, chiisa-na

precopular noun

Eg. tama-no, muji-no, ippan-no, tokubetsu-no, betsu-betsu-no, naname-no, hasu(kai)-no

precopular noun, defective

lacking adverbial (-ni). Eg. utte-tsuke-no, kansetsu-no, hon'nen-no, kinshou-no, akushitsu-no

precopular noun, highly defective

marks only the adnominal (-no). Eg. zekkou-no, zantei-no

adnoun

Eg. kono, iron'na, hon-no, aru

(Paraphrase of p.180-2) Now if you think a bit more about these classes, you'll see that the keiyoudoushi are anything but regular. There are examples like isasaka, sugu, hitasura, etc. that are impossible to place in any single adjectival, precopular or adverbial class, depending solely on their use. One has to wonder at this point if there really is any sense to these classifications.

Note that these are all basically nouns. The "adjective" class, keiyoushi, are not nouns, though many are formed from nouns: atataka-i from atataka, yawaraka-i from yawaraka, etc. This class is called the i-type class for obvious reasons, but they are "pure" in the following additional sense: they are adjectives in bungo (the literary/classical language). Bungo in fact has exactly two styles of "adjectives"—the ku-type and the shiku-type—which differ only in how to calculate their stem from their predicative form. For the ku-type, you have to delete the -shi (i.e., takashi → taka-, nashi → na-), whereas for the shiku-type the stem equals the predicative form (i.e., suzushi → suzushi-). Otherwise, they have identical endings. In Modern Japanese the distinction between these two classes is removed.

While I'm quoting Martin, here's what he says about onaji (p.743):

onaji 'the same'--also onnaji and onnashi. In the literary language this is a special kind of adjective with the forms onaji-ki for the attributive and onaji//-shi//^* for predicative, following a rule of haplology that obligatorily reduces -shi-shi and -ji-shi in adjective predicative forms. If you run across onaji-ku, it is a somewhat formal or semi-literary usage fror onaji (you)-ni; and onaji-i, from onaji-[k]i, is a pseudo-literary form for onaji (+N). Although onaji you ni is more common, onaji ni also occurs. ...

^* this notation means that -shi is disallowed

The point of all this is that lexical classification in Japanese is a mess. Most textbooks simply (and rightly) ignore the complexity. — Kaustuv 04:39, 2005 Apr 10 (UTC)

Hmmm. It seems I need to read some Martin. Would the text in question be A Reference Grammar of Japanese by any chance? I am currently slogging through Part One of Shibatani's The Languages of Japan and am quite interested at the fun overlap between Ainu and Japanese, but even more surprised and intrigued by the seeming overlap with Indo-European (sine, tu for one, two, similar to Korean hana, dwu, and random words like wakka for water and hosi for hose, i.e. a garment you put on your legs, and the hone (J) - pone (A) - bone (E) similarity). At any rate, do let me know what I should be reading next. :)

Incidentally, I quite agree with your notion that the language books need not delve into the complexity of lexical categorization. While splitter tendencies of being as specific as possible can be helpful in certain situations, in others, they only serve to make discussion more confusing instead of clearer. For example, I'm not entirely certain what Martin means by classifying word stems like chiisa- and shizuka as nouns. While shizuka can form a predicate when followed by desu, chiisa- never appears on its own...?

Thank you for your time in this discussion. I very much appreciate being able to talk about these ideas in an informed and interested context. Cheers, --- Eiríkr Útlendi 15:25, 10 Apr 2005 (UTC)

True adjectives are non-predicating words that modify nouns. This means that verbs and copular forms are not adjectives, even though they may be used to modify nouns.

Who cares about classification? Well, that may be a deeper question than we can get into here. The basic issue, as far as linguistics is concerned, is figuring out a set of word categories that can be used for all languages. We have a problem in that the only truly universal word classes are noun and verb. Beyond that, languages differ in whether or not they have other categories. So we want to have enough categories to fit everyone, but not more than are necessary.

We then get into the issue of categorization: do we create new word classes, subcategorize, or create a dichotomy? Creating new word classes willy-nilly doesn't help, so we want to avoid that when possible. Subcategorizing words within existing categories is good, though you may find that what you thought was one class of words in multiple subcategories may turn out to be more than one class. Dichotomies tell you that you're probably making a mistake somewhere.

Let's move on to some concrete examples. In accusative languages (a category we don't need to worry about here), verbs can be intransitive or transitive. The difference is in the number and kind of arguments they take: intransitive verbs take one argument, a subject, e.g., Bubba ran, and transitive verbs take two, a subject and an object, e.g., Maisie slapped Bubba. Now, what about eat? Buford ate at noon (the time phrase is added for naturalness and has no bearing on the argument here) and Buford ate 4 greaseburgers at noon are both OK. Do we have a class of verbs that are sometimes transitive and sometimes intransitive? Hmm, dichotomy... Let's say instead that a word is defined by three qualities: form, meaning, and distribution. Form means its pronunciation; meaning is obvious; and distribution means where it shows up in a sentence and what shows up around it. Since we already have the classes of transitive and intransitive verbs, we can just say that there are two words eat: eat₁ is transitive, and eat₂ is intransitive.

This solution has a potential problem: have we created more words than necessary? No, because we find many cases of otherwise-identical words differing only in distribution (i.e., word class), like blue the noun and blue the adjective, and book the noun and book the verb.

So, the relevance is: does Japanese have adjectives, and if so, what are they? I maintain that the the words that are traditionally called "adjectives" in Japanese are verbs, because that is the syntactic (and morphological) behavior they exhibit. On the other hand, there are some "true" adjectives in Japanese, words whose only syntactic function is to modify nouns.

So there you have it. I hope this helps. Squidley 17:04, 19 Apr 2005 (UTC)

Keiyou doushi / meishi

Hey there Kaustuv, By what metric are the keiyou doushi presumed to be nouns? Is this Martin's categorization? Only I'm thinking about the shizuka example, and fail to see how shizuka could be a noun, unless it's simply in that it must be followed by the copula to form a complete sentence or predicate. Thanks, --- Eiríkr Útlendi 14:34, 12 Apr 2005 (UTC)

Sorry, I am not sure what you're referencing, and I might have mischaracterized Martin's classification. Shizuka is certainly not a meishi—not in traditional grammars, and not in Martin's grammar; rather, it is an "adjectival noun". The name of this class is traditional, but misleading: ANs are not a subcategory of noun, but a sort of sibling under some larger nominal class. ("Sort of" because there is a fair bit of overlap between keiyoudoushi and meishi.) Shizuka does have the most important feature of all nominals—that of needing -da for its inflections—so calling it a noun is not a huge mistake. The mistake would be essentially that shizuka cannot be a subject in the modern language, but if you look at the development of -na and -no, they both come from the classical verb nari (which itself might have come from no ari). —Kaustuv 08:19, 2005 Apr 16 (UTC)

It would seem the only confusion is mine, so thank you for clearing that up. I have not yet had the chance to read Martin (I'm currently digging through Shibatani's Languages of Japan), and so my misunderstanding is mostly due to lack of familiarity with the particulars of this discussion. However, your post here explains the gap in my comprehension quite well. :) Might it be acceptable to also call these "nominal adjectives", in that they are more adjectives than nouns? Cheers, --- Eiríkr Útlendi 18:05, 16 Apr 2005 (UTC)

A common term for keiyoo dooshi is copular noun. I'm not certain why the term noun is used, but I do understand the couplar part: they require a copula, both when modifying (e.g., shizukana basho) and when in predicate position (e.g., koko wa shizuka da).

They aren't adjectives in that they appear with copulas when modifying; notice that English adjectives, for example, don't need anything else when modifying. They are similar to adjectives in that they cannot predicate on their own, and need copulas to do that for them. Rather than "nominal adjectives," McClain (1981) "explained" them with the phrase adjectival nouns, which is also a fairly common term.

FYI, nari, the "parent" of na, is widely thought to have come from ni ari, while da and desu, and their dialectal relatives ja, ya, and dasu are from de arimasu. no has a different origin, as I recall. Incidentally, the ja in ja nai is from dewa, and was once written jaa (read Akutagawa and you'll see it like that).

One more FYI: In her Handbook of Modern Japanese Grammar, McClain used the term verb to describe keiyooshi. Look on page 87. Squidley 17:14, 18 Apr 2005 (UTC)