Is gradable vs absolute a universal distinction?

Question

Inspired by multiple questions on ELU and in particular this recent question about 'correct', I wonder whether French has the similar concept of gradable vs absolute adjectives.

The idea is that some qualities are either the case or not. For example, if something is 'unique', only one thing can be that (otherwise, logically, it isn't unique). This is called 'absolute'.

And then other qualities can afford degrees; this is called 'gradable'. One thing can be red, and another in comparison can have a higher value of red.

And what this means for grammar is that a gradable adjective allows the syntactic formation of comparative and superlative (redder/reddest), but absolutes do not (*'more unique').

The question here is whether this concept is universal across languages. Do all languages have some adjectives (or modifiers) that are syntactically not allowed with comparative constructions? Are there other ways of managing comparison syntactically (partially gradable? absolute but with a middle 'not either' syntactic marker?) etc.

I feel like Indo-European languages would disallow comparatives with semantically absolute qualities, but maybe Chinese, with bǐ and zuí comparatives may. I can't tell.

Or is this distinction purely a semantics or pragmatics issue? The Chinese comparative construction is more predicate based rather than around a modification of an adjective.

Frankly, in English absolutes are used all the time in comparatives, but almost always in a jocular (or thoughtless) manner.

Answer 1

I think there is a distinction here, and it's cross-linguistic—but your example falls on the wrong side.

Most English adjectives (that aren't already in comparative/superlative form—"*more best") are comparable. But a handful are not, such as "only", "galore", "additional", and maybe "daily".¹ And "unique" is not in that small category. Compare:

• John is more annoying/unique/*only/*galore/*additional/*daily than Mary.

That "*more only" is blatantly ungrammatical, and "more unique" is clearly different. If you want to correct someone's usage of "more unique", you have to get them to think consciously about the logic. Which is a good sign that this isn't a rule of grammar—unless you want to stretch the grammar to rule out, e.g., "John is taller than both Mary and Susan, but shorter than both Bill and Susan".

So, why are we taught that "more unique" is ungrammatical? Because English pedagogy isn't based on linguistics, it's based on pedantry. If "only" and "unique" both translate Latin "unicus", they must be in the same class, so "more unique" must be just as wrong as "more only". Also, if "most unique" is logically incoherent, it must be grammatically incoherent, if only kids these days would learn to think grammatically (or, at least, if only kids in the past hadn't broken English so it's not logical like Latin). As a matter of style, you shouldn't say "more unique" unless you're making an ironic point or jokingly being illogical or whatever, but that doesn't mean it's not grammatical English.

Presumably logic is universal. So, in any language, using a comparative construction with an adjectives whose meaning includes being singular, like "unique", will lead to an illogical sentence. Which implies that any language that has a community of prescriptive grammarians like English's will probably have a similar prescription. But how universal such communities are, that's a sociological question, way outside of linguistics.

---

If you're just interested in "more unique" and not "more only", you can probably stop here.

---

So, what makes actually uncomparable² adjectives uncomparable? I think it's that they're not predicable:

• John is tall/annoying/unique/*only/*daily.
• John is taller/more annoying/more unique/*onlier/*more only/*dailier/*more daily than Mary.

As a side note,³ beyond not being predicable, most of the uncomparable adjectives are weird in at least one other way—and in different ways from each other. "Only" acts almost as if it were already a superlative: "John is the *tall/*taller/tallest/only doctor in town." And "daily" seems to be restricted to nouns formed from verbs (not just gerunds; nouns formed by semi-regular morphology work too): "John is a daily drinker/competitor/?doctor/*man." And "galore" and "aplenty" are mandatorily postnominal.

---

My hypothesis is that cross-linguistically, any predicable adjective is grammatically comparable.

First, "unique" is a comparable adjective in the Romance languages, just as in English,⁴ with the same "if you think about it consciously, it's illogical" result. And they do have truly uncomparable adjectives:

• John es más alto/molesto/único/*sólo/*diario que Mary. (John is more tall/annoying/unique/*only/*daily than Mary.)

… which also aren't predicable:

• John es alto/molesto/único/*sólo/*diario.

---

But those languages are still cousins of English. What about farther afield? I think the same holds true. In fact, I think there are many languages where "anything predicative is comparable" is a grammatical rule, or at least tautologically follows from the rules.

For example, in Japanese, while I am far from a fluent speaker (and invite correction here), you can turn any predicative sentence into a comparative sentence without needing to do anything to the adjective: just add a comparand with the "than" particle,⁵ or add a sentence-adverb for a bare comparative:

• Jon wa segatakai. (John topic-particle tall-as-verb: John is tall.)
• Jon wa Mari yori segatakai. (John topic-particle Mary than-particle tall-as-verb: John is taller than Mary.)
• Jon wa motto segatakai. (John topic-particle sentential-more tall-as-verb: John is taller.)

The same works with desu-copula sentences:

• Jon wa aka desu. (John topic-particle red-as-noun is: John is red.)
• Jon wa motto aka desu. (John topic-particle more red-as-noun is: John is redder.)
• Jon wa Mari yori aka desu. (John topic-particle Mary than-particle red-as-noun is: John is redder than Mary.)

… even when the predicate is a noun, which couldn't be used attributively:

• Jon wa shoonen desu. (John topic-particle boy is: John is a boy.)
• Jon wa motto shoonen desu. (John topic-particle more boy is: John is more of a boy.)
• Jon wa Ken yori shoonen desu. (John topic-particle Ken than-particle boy is: John is more of a boy than Ken.)
• Jon wa yuniiku desu. (John topic-particle unique is: John is unique.)
• Jon wa motto yuniiku desu. (John topic-particle more unique is: John is more unique.)
• Jon wa Mari yori yuniiku desu. (John topic-particle Mary than-particle unique is: John is more unique than Mary. Notice that "yuniiku" is a recent borrowing from English "unique", but it's a noun.)

---

And let's take that insight back to English:

• John is more man than Ken will ever be.

Anything that can be a predicate can be compared, not just adjectives, even in English?⁶

That's a little odd to at least some speakers (see the comments), and may even invite an account based on either deleting "of a" or reducing the adjective "manly" or something similar. But look at examples where the contrast is between two predicates over one entity, rather than two entities over one predicate:

• She's more woman than child.
• He's more machine now than man.
• Tyrell Corporation: more human than human.

---

One tricky test:

• ?Jon wa Mari yori mainichi desu. (John topic-particle Mary than-particle every day is: ?John is more every-day than Mary)

That's semantically very odd in English, but in the right context I think still acceptable: "Do John and Mary come here every day? No, but John is more every-day than Mary." And I think the same is true in Japanese. So, this is still a different class of "wrong" from using "more only" in English.⁷

I don't think there is a "more only" class in Japanese, or many other languages. Uncomparable adjectives are words that act attributively like adjectives (and, in inflected languages, agree like adjectives when doing so), but can't be predicated; English and Spanish have a few, but Japanese has none.

---

But for a real test, you'd want to raise a child as a native Lojban speaker and see whether they accept the direct translation of "John is more unique than Mary" as a grammatical sentence.⁸ I'll bet they will.

---

So, in conclusion:

• A few English and Spanish adjectives, like "additional", are not predicates, and they're not comparable.
• A few English adjectives, like "galore", have weird syntactic features that get in the way of the usual comparison constructions, and they're not comparable.
• There seem to be no English adjectives that are predicable and syntactically regular and can't be compared.
• Japanese has no non-predicable adjectives and no syntactically-irregular adjectives, and all of its adjectives as comparable.
• It looks like even non-adjectival predicates can be compared, in both English and Japanese.

So, "uncomparable" reduces to "either not predicable, or syntactically irregular in a way that blocks using them in a comparison construction".

---

_{1. Of course most of these are polysemous. For example, "only" is an adverb, "John is only dancing", and an adjective of restricted distribution, "John is an only child/*guy/*tall guy", and another adjective of free distribution, "John is the only doctor/guy/tall guy in town". Let's ignore the first two senses here.}

_{2. Oddly, many prescriptivists use "incomparable" to describe adjectives like "unique". Someone should tell them that "incomparable" means "beyond compare", not "unable to be compared", so they're really saying that these are superlatively excellent words…}

_{3. I'm not sure it really is a side note. If, say, these words are also stored—and primarily accessed—as partial constructions, like [N aplenty] as an N' meaning "having sufficient N", that would explain why they're resistant to being thrown into normal A constructions.}

_{4. I think it's a modern (not just as opposed to Norman, but after early modern English) borrowing from French; if so, this isn't surprising.}

_{5. I'm ignoring the fact that "yori" can also be used more like "from" than "than", which should pull all the "yori" sentences toward a second interpretation, because I think the comparative interpretation is much stronger here, enough to avoid the ambiguity. But I could be wrong, especially with the last one—"??John is every day from Mary" isn't much odder than "?John is more every-day than Mary" in English, after all.}

_{6. This implies that the usual gloss of "more" as an adverb is just wrong. But then we've already got "The more I read, the more I like it", and various other constructions, that are even harder to explain (outside of a construction grammar or similar framework).}

_{7. Notice that if I translate "mainichi" in its more usual gloss as "daily", it's much more unacceptable in English. Maybe because it's now taking the form of one of English's small class of non-predicable adjectives, so it's not just a matter of fixing up the semantics with discourse/pragmatic context, but overriding a grammatical judgment?}

_{8. There's no Lojban word meaning "to be unique", but that's fine; Lojban is designed to let you productively invent new compounds. So, "la djan zmadu la maris cizra pamei" ("proper-name John is-more-than proper-name Mary (in-the-property-of) to-be-some-contextually-determined-compound-having-to-do-with-weirdly/distinctly/especially-singular") is a perfectly legitimate sentence in Lojban grammar. And, if your previous discourse (or a native Lojban community…) has established "cizra pamei" as here meaning "to be unique", what it legitimately means is "John is more unique than Mary".}