Adjectives

[HoskhMatriarch]

We all know that some languages basically have verbal adjectives. What's more interesting to me is that some languages seem to have nominal adjectives. Latin has a bunch of adjectives and nouns that are indistinguishable, and I thought that maybe that made the "adjective" a noun in an appositional compound rather than an adjective (at least in the English sense). I've also seen a post about how apparently Coptic just has adjectival nouns instead of adjectives. So I was wondering if I could get away with just having adjectival nouns and compounds of adjectival nouns instead of seperate adjectives in my conlang and still be only doing things natlangs do. I'd rather have fewer adjectives but adjectival verbs are worse since I can't even incorporate them into verbs like I can a series of an adjective and a noun.

[Sumelic]

Well, there is a WALS chapter about adjectives marked the same way as genitives (that is, like nominal modifiers) or as relative clauses (like verbal modifiers). Both systems are attested. In general, I think there's no problem of naturalism with not having a single clearly-defined class of "adjectives."

However, defining parts of speech is a bit tricky, and languages sometimes have sub-divisions in larger "parts of speech" categories, or there are certain words that are used much more in one way than in another way (for example, in English nouns, there are some nouns that are normally count, and some that are normally non-count.) So even if the "noun-like" words and "adjective-like" words in your language fall into the same formal class, the way they are used, their function, might still be able to be categorized as "adjectival" or "nominal", and there might be specific words that are very likely to be used one way or the other.

It's possible for a language to have things that look like nouns, act as "adjectives", but that also differ a bit in their behavior from a prototypical noun like "a person". Latin adjectives, as you've mentioned, take a lot of the same inflectional categories as nouns, and basically can be used as nouns. But unlike a regular noun, Latin adjectives agree in gender and have forms for all 3 genders, and they have comparative and superlative forms. So there are some differences.

Latin also has participles, which are sort of like "adjectival verbs," in that they share some of the properties of both adjectives and verbs. So that might be considered another "part of speech", or a subset of verbs, depending on your viewpoint.

Anyway, I guess my point is that it doesn't matter so much what you chose to call these grammatical categories; what's important is to know how the system works.

By the way, what were you thinking of when you mentioned "adjectival verbs", and why can't they be incorporated in your language?

[elemtilas]

Well, there is a WALS chapter about adjectives marked the same way as genitives (that is, like nominal modifiers) or as relative clauses (like verbal modifiers). Both systems are attested. In general, I think there's no problem of naturalism with not having a single clearly-defined class of "adjectives."

If I recall correctly, Talarian does both. What we call adjectives they might call attributive nouns. They can be possessive structures: beauty's girl = beautiful girl kind of thing. Or they can take part in a polyradical concatenation: girl-beauty-tall-among-tulips-tiptoeing would not be an impossible noun formation.

However, defining parts of speech is a bit tricky, and languages sometimes have sub-divisions in larger "parts of speech" categories, or there are certain words that are used much more in one way than in another way (for example, in English nouns, there are some nouns that are normally count, and some that are normally non-count.) So even if the "noun-like" words and "adjective-like" words in your language fall into the same formal class, the way they are used, their function, might still be able to be categorized as "adjectival" or "nominal", and there might be specific words that are very likely to be used one way or the other.

Yes. Talarian, for example, has two broad kinds of words: substantives and function words. The latter are adpositions and so forth. The former breaks down into nouns, pronouns and verbs (because they all denote the substance of meaning).

The native names are more interesting, I think, though I'd have to look up to remind myself the classification scheme. I remember it was based on the concept of a dance.

It's possible for a language to have things that look like nouns, act as "adjectives", but that also differ a bit in their behavior from a prototypical noun like "a person". Latin adjectives, as you've mentioned, take a lot of the same inflectional categories as nouns, and basically can be used as nouns. But unlike a regular noun, Latin adjectives agree in gender and have forms for all 3 genders, and they have comparative and superlative forms. So there are some differences.

Latin also has participles, which are sort of like "adjectival verbs," in that they share some of the properties of both adjectives and verbs. So that might be considered another "part of speech", or a subset of verbs, depending on your viewpoint.

I place them with nouns in Talarian. Though of course they are actually verbs, as several of them can act quasi-verbally, taking objects and exhibiting tense.

Anyway, I guess my point is that it doesn't matter so much what you chose to call these grammatical categories; what's important is to know how the system works.

Yeah. And write it all down, so you won't forget it twenty years from now!

[clawgrip]

Japanese has both adjectival verbs and adjectival nouns, and very few true adjectives (like maybe 2). I don't see any problem with making all adjectives nominal.

"Adjectives" and relative clauses are completely merged, and all clauses generally require a verbal element. Since Japanese adjectival nouns, by definition, lack the verbal element found in adjectival verbs, they thus require a form of the copula both in attributive and predicative positions.

There are two types of adjectival noun that differ only in the type of attributive particle they use, and one of these two types is, grammatically, pretty much indistinguishable from nouns. In fact, you could swap out an adjectival noun of this type for a regular noun and still have a grammatical sentence (even if it becomes nonsensical). The reverse is not always true, however, as adjectival nouns are still semantically adjectives and can't occupy all the roles of regular nouns, such as subject, object, etc.

Adjectival nouns are grammatically different from adjectival verbs, but they are nonetheless just as versatile, so I see no reason why a language couldn't have all its adjective-type words as adjectival nouns.

[gestaltist]

Japanese has both adjectival verbs and adjectival nouns, and very few true adjectives (like maybe 2). I don't see any problem with making all adjectives nominal.

"Adjectives" and relative clauses are completely merged, and all clauses generally require a verbal element. Since Japanese adjectival nouns, by definition, lack the verbal element found in adjectival verbs, they thus require a form of the copula both in attributive and predicative positions.

There are two types of adjectival noun that differ only in the type of attributive particle they use, and one of these two types is, grammatically, pretty much indistinguishable from nouns. In fact, you could swap out an adjectival noun of this type for a regular noun and still have a grammatical sentence (even if it becomes nonsensical). The reverse is not always true, however, as adjectival nouns are still semantically adjectives and can't occupy all the roles of regular nouns, such as subject, object, etc.

Adjectival nouns are grammatically different from adjectival verbs, but they are nonetheless just as versatile, so I see no reason why a language couldn't have all its adjective-type words as adjectival nouns.

What about adverbs? In a lot of European languages adjectives can double as adverbs. How does Japanese work in that regard?

[clawgrip]

All verbal elements, i.e. regular verbs, adjectival verbs, and the copula, each have two separate adverbial forms: one for in-clause modification, and one that subordinates an entire clause. There are also many adverbs that are unrelated to any sort of adjective.

[Xing]

I think it was Mark Baker's conclusion in Lexical Categories that all languages have distinct nouns, verbs and adjective as distinct parts of speech - though this distinction may be less obvious on the surface leve in some languages.

However, defining parts of speech is a bit tricky, and languages sometimes have sub-divisions in larger "parts of speech" categories, or

We could define parts of speech more or less broadly. Not all "verbs" in a given language behave in exactly the same way all the time – you might distinguish various sub-categories of verbs, with subtle differences in their syntactic behaviour. (When beginner conlangers think they can adequately describe how verbs behave in a couple of paragraphs and an inflection table, they seriously misjudge how complex real languages are.) Say that we could discern a category of "adjectival verb" or "stative verbs" – how different from other verbs must these be, in order to be considered a "separate" part of speech, rather than a subcategory of verbs? It's a rather theoretical issue – I think Baker mentioned the ability to appear as secondary predicates as one of the criteria of adjectivehood.

[Micamo]

I think it was Mark Baker's conclusion in Lexical Categories that all languages have distinct nouns, verbs and adjective as distinct parts of speech - though this distinction may be less obvious on the surface leve in some languages.

I agree with Baker's result, but with an important caveat: Baker used a purely syntactically motivated definition of the difference between nouns, verbs, adjectives, and adpositions. It doesn't apply if you define it semantically or functionally instead.

[HoskhMatriarch]

By the way, what were you thinking of when you mentioned "adjectival verbs", and why can't they be incorporated in your language?

I'm not saying that there'll be no adjectival verbs ever, since I'm pretty sure things like "towered" (=was tall) and "boomed" (=was loud) in English probably count as adjectival verbs, and in that case not having any would be unnatural. I'm saying that I don't want that to be the primary way of expressing concepts because there is all sorts of compounding and verb incorporation in my language. Even with a regular adjective, I could make something like (not with English words) "cardrive" (=drive a car) be "redcardrive" (= drive a red car). Otherwise I'd have to use multi-word constructions like "cardrive isred" that don't really seem to fit into this language very well.

If I recall correctly, Talarian does both. What we call adjectives they might call attributive nouns. They can be possessive structures: beauty's girl = beautiful girl kind of thing. Or they can take part in a polyradical concatenation: girl-beauty-tall-among-tulips-tiptoeing would not be an impossible noun formation.

...

Yeah. And write it all down, so you won't forget it twenty years from now!

"Girl-beauty-tall-among-tulips-tiptoeing" is just awesome. And the equivalent of "beauty's girl" in my conlang would probably be a compound (depending on context).

I'm definitely writing everything down. I'm making a grammar once I'm done, and a lot of recordings of the language...

[Systemzwang]

It seems to be very popular for conlangers to merge adjectives with either verbs or nouns; I guess some kinda merge them both ways (i.e. what we'd translate as nouns, some of them are verbs in the language, some are nouns). However, why doesn't anyone ever split adjectives into two new categories?

[zyma]

However, why doesn't anyone ever split adjectives into two new categories?

Anything in particular you have in mind in terms of this? If so, could you maybe give an example or two?

[gach]

As in something like for example one group of adjectives that follow similar syntax to determiners and another that's more like nouns but still has some distinct morphology from them?

You could for example say that a language has agreement on most of its adjective attributes (case etc.) and puts them after the nouns but also requires every NP to be headed by some determiner such as an article preceding the noun. However, there's a group of uninflecting core adjectives that precede the nouns and can take over the function of the mandatory determiner. You might get something like,

ART lake-LOC
"on a lake"

PROX lake-LOC
"on that lake"

large lake-LOC
"on a large lake"

PROX large lake-LOC
"on that large lake"

ART lake-LOC calm-LOC
"on a calm lake"

PROX lake-LOC calm-LOC
"on that calm lake"

I'm myself a fan of dealing away with the whole adjective class one way or another but I might actually end up incorporating something along these lines to the adjective system of Nooníí kiskn.

[masako]

We all know that some languages basically have verbal adjectives. What's more interesting to me is that some languages seem to have nominal adjectives. Latin has a bunch of adjectives and nouns that are indistinguishable, and I thought that maybe that made the "adjective" a noun in an appositional compound rather than an adjective (at least in the English sense). I've also seen a post about how apparently Coptic just has adjectival nouns instead of adjectives. So I was wondering if I could get away with just having adjectival nouns and compounds of adjectival nouns instead of separate adjectives in my conlang and still be only doing things natlangs do. I'd rather have fewer adjectives but adjectival verbs are worse since I can't even incorporate them into verbs like I can a series of an adjective and a noun.

http://en.wikipedia.org/wiki/Part_of_speech#India

Some of the oldest academic grammar completely leaves "adjectives" out. In my conlang Kala, I use descriptive verbs, i.e. "is red", "is big". I think Imralu's Ahu also does this.

[Imralu]

I see no reason why your language can't describe all adjective-like meanings using words which belong to the same noun-class as the words describing noun-like meanings. Your word for "large" can, on its own, mean "large thing" or "large one". To really make sure that the categories are actually merged, you'll need to make sure that their distributions are the same. For example, if you end up with only adjective-like meanings being able to be used in comparative constructions, as here:

• 3s.MASC COP COMP tall than 3s.FEM Grammatical.
  He is taller than her.
  
  *3s.MASC COP COMP friend than 3s.FEM Ungrammatical
  *He is more of a friend than her.

... or being treated differently in how they may be combined with other words ...

• ART man tall grammatical
  ART man REL tall grammatical
  the tall man
  
  ART man friend ungrammatical
  ART man REL friend grammatical
  the man who is a friend

Then you will have a distinction between the two word classes. If you're not careful, it's easy for a distinction to creep in somewhere.

Pretending for a second that St’át’imcets is a conlang, not a natlang, I feel like the creator was aiming for no noun-adjective distinction and let something slip and forgot for a moment. Using ti =a, there seems to be no noun-adjective distinction but using ti7 ku= there is.

ti sqaycw=a
DET man=DET
the man

ti tayt=a
DET hungry=DET
the hungry one

ti7 ku=sqaycw
DEM DET=man
that man

*ti7 ku=tayt
DEM DET=hungry
that hungry one

[HoskhMatriarch]

I see no reason why your language can't describe all adjective-like meanings using words which belong to the same noun-class as the words describing noun-like meanings. Your word for "large" can, on its own, mean "large thing" or "large one". To really make sure that the categories are actually merged, you'll need to make sure that their distributions are the same. For example, if you end up with only adjective-like meanings being able to be used in comparative constructions, as here:

• 3s.MASC COP COMP tall than 3s.FEM Grammatical.
  He is taller than her.
  
  *3s.MASC COP COMP friend than 3s.FEM Ungrammatical
  *He is more of a friend than her.

... or being treated differently in how they may be combined with other words ...

• ART man tall grammatical
  ART man REL tall grammatical
  the tall man
  
  ART man friend ungrammatical
  ART man REL friend grammatical
  the man who is a friend

Then you will have a distinction between the two word classes. If you're not careful, it's easy for a distinction to creep in somewhere.

Pretending for a second that St’át’imcets is a conlang, not a natlang, I feel like the creator was aiming for no noun-adjective distinction and let something slip and forgot for a moment. Using ti =a, there seems to be no noun-adjective distinction but using ti7 ku= there is.

ti sqaycw=a
DET man=DET
the man

ti tayt=a
DET hungry=DET
the hungry one

ti7 ku=sqaycw
DEM DET=man
that man

*ti7 ku=tayt
DEM DET=hungry
that hungry one

What's wrong with there being some distinction somewhere? As long as I can put the adjectives/nouns into compounds and incorporated verbs instead of having them have to hang out on their own because they're seperate from the noun head they describe, I'm fine. I probably should've specified that I'm trying to do things with them, that is, combine them with the noun head somehow but also have them be able to stand alone, and not just trying to be "look at me adjectives are nouns I'm weird and not IE". But the more examples of natlangs doing things where nouns and adjectives are the same, the better for my language I guess.

[Imralu]

What's wrong with there being some distinction somewhere?

Nothing at all. A distinction simply means there is a distinction. I thought you were asking about no distinction.

[HoskhMatriarch]

What's wrong with there being some distinction somewhere?

Nothing at all. A distinction simply means there is a distinction. I thought you were asking about no distinction.

Oh. No, I'm just trying to figure out what to do with adjective-type (and adverb-type for that matter) words to make them attach to noun and verb heads so I can do more with them morphologically (because something like the example I gave with redcardrive and a noun from it redcardriving doesn't seem like it would really work much, but if "redcar" were a word it would...). Not having a distinction would be cool, but I'm not sure if any language does that.

[Imralu]

Oh. No, I'm just trying to figure out what to do with adjective-type (and adverb-type for that matter) words to make them attach to noun and verb heads so I can do more with them morphologically (because something like the example I gave with redcardrive and a noun from it redcardriving doesn't seem like it would really work much, but if "redcar" were a word it would...). Not having a distinction would be cool, but I'm not sure if any language does that.

Wow. I really suck at reading comprehension then.

Random idea - maybe you could have a special form of the adjective which refers to an incorporated noun.

• red1 car driving
  
  red2 cardriving

If you were to use red1 with cardriving, it would refer to the driving, but the form red2 specifies an incorporated noun and makes it refer to the car rather than the driving.


Doesn't sound very natlangy though ...

[Znex]

Oh. No, I'm just trying to figure out what to do with adjective-type (and adverb-type for that matter) words to make them attach to noun and verb heads so I can do more with them morphologically (because something like the example I gave with redcardrive and a noun from it redcardriving doesn't seem like it would really work much, but if "redcar" were a word it would...). Not having a distinction would be cool, but I'm not sure if any language does that.

Wow. I really suck at reading comprehension then.

Random idea - maybe you could have a special form of the adjective which refers to an incorporated noun.

• red1 car driving
  
  red2 cardriving

If you were to use red1 with cardriving, it would refer to the driving, but the form red2 specifies an incorporated noun and makes it refer to the car rather than the driving.


Doesn't sound very natlangy though ...

It would probably make more sense to add an affix to the head ({cardriving}) so that {red} is undoubtedly referring to that. That's more naturalistic, I think.

eg. <car driving> vs. <cardrivingery>.

[HoskhMatriarch]

Oh. No, I'm just trying to figure out what to do with adjective-type (and adverb-type for that matter) words to make them attach to noun and verb heads so I can do more with them morphologically (because something like the example I gave with redcardrive and a noun from it redcardriving doesn't seem like it would really work much, but if "redcar" were a word it would...). Not having a distinction would be cool, but I'm not sure if any language does that.

Wow. I really suck at reading comprehension then.

Random idea - maybe you could have a special form of the adjective which refers to an incorporated noun.

• red1 car driving
  
  red2 cardriving

If you were to use red1 with cardriving, it would refer to the driving, but the form red2 specifies an incorporated noun and makes it refer to the car rather than the driving.


Doesn't sound very natlangy though ...

It would probably make more sense to add an affix to the head ({cardriving}) so that {red} is undoubtedly referring to that. That's more naturalistic, I think.

eg. <car driving> vs. <cardrivingery>.

Why would there need to be an affix on anything at all? If it's a verb, and the red were referring to that, it'd be an adverb, and if the red were referring to driving as a noun, it'd be carreddriving. Plus, if the adjectives-are-adjectival-nouns things works for a naturalistic language, it'd be an appositional compound, so it'd be really obvious, "[reddthing/car]driving", or with the other way another person mentioned "redness(gen.)cardriving".