Is a language without nouns possible? And another one without verbs?
And other ones without adjectives or adverbs?
Is there some real examples? (In preference: non-constructed languages, because everything is possible with constructed languages, but you can mention a few, it's also interesting)
(Credit for the question: TseDanylo)
It is not possible for there to be a human language that does not have a way of referring to entities, or to predicate states and actions of an entity. If that is what you mean by "noun" and "verb", then all languages have nouns and verbs. However, noun, verb, adverb, adjective are typically treated by linguists as "word classes", defined in terms of how they are structured or function with other word classes (and not in terms of what they mean), and it is (at this point in linguistic theory) not entirely clear that the distinction is vital. For example, Salishan languages are flexible in noun / verb relations, so a root may be more noun-like with a certain prefix, and more verb-like with a different affix (suffix, in Lushootseed). Nouns and verbs both have past and future tenses. One can typically discern, on a language specific basis, some formal basis for calling certain words "nouns" versus "verbs", such as "adding s- makes this a noun", but it may not yield any significant benefit to making such a distinction. The formal properties of English nouns, verbs and adjectives do not define an "architype" for such words in most languages.
One way of understanding the underlying issue is that there are two different questions you can ask about word classes. One is whether there is always compelling evidence in all languages for distinguishing nouns and verbs (or verbs and adjectives), and the answer to that is "No"; but, you can also ask if it is always possible to define such a distinction in a language, on the premise that there might be mandatory but not particularly useful categories of analysis, and the answer to that is "Yes" (but IMO this begs the question of whether there are mandatory categories in grammar).
There is an artificial language called Kēlen which lacks verbs (or supposedly). It instead uses four 'relationals' to show how the noun phrases interact with each other. It was designed to be alien, but people have supposedly translated texts written by the author back into a natlang. And she has I believe translated texts herself into the language.
To me it looks more like a language with a very minimal verb system. But perhaps that's what you're looking for?
I have also heard of some odd philosophical languages that treat all 'actions' as a state of being, or treat nouns as being transient phases rather than objects (this mostly just boils down to nouns having tense).
As for natlangs that do? No, there are none. Its one the few linguistic universals: all languages differentiate nouns from verbs. Of course parts of speech aren't universal. Chinese has 'coverbs' that act as both prepositions and verbs. And eastern languages in general tend to treat adjectives as just a sub-class of verbs (they double as verbs meaning 'to be X', and can even take tense, even if used adjectivally). But there is no language that puts nouns and verbs into the same category. Well, confrontational languages kind of do, but often times the word has different meanings depending on whether its used as a noun or verb (a fly vs to fly, produce vs to produce). Honestly its more like a lazy form of derivation.
I agree and disagree with user6726--humans are naturally concerned with entity versus action distinctions (exactly the same as user6726 puts it), but that does not necessarily mean that such distinctions are directly encoded into a language lexically. The question, I think, is less whether or not a language has verbs or does not, but rather to what extent do conversational participants have to work in order to disambiguate information in an utterance.
Example: Riau Indonesian is notorious for being the language sans noun and verb classes. Speakers of the language, though, clearly use contextual cues in conversation in order to discern whether a given utterance is to be construed as an action or an entity. As such, Riau Indonesian categorically defies attempts at cleanly splitting "words" into one of the before mentioned categories or the other. Language developes with respect to the environment that a culture blossoms within, and as such researchers need to take that into account. Even location can influence the context of an utterance and whether or not a given segment is noun-ish, or verb-y.
Not to forget, we have this way in English too, using contextual cues to decipher whether the word acts as a verb or as a noun.
Examples:
The designers of Lojban found a way to combine noun, verbs, and adjectives into single-word predicates called brivla that become noun, verbs, or adjectives according to some extremely fluid process that nobody understands except for a handful of Vulcans and Houyhnhnms. The creators of Lojban did something I presumed impossible: they created a programming language harder to understand than C++!
Because everything in the universe is either an object (noun) or a process (verb), no, it is not possible for a general purpose human language to omit these two categories. Adjectives are the end result of processes (state).
Therefore because everything in the universe is an object (noun), process (verb), or the end result of a process (adjective), it follows that every human language has the lexical categories of nouns, verbs, and adjectives, even if professional linguists may not categorize them exactly like a Latin grammarian would.
