Is there a strong case for the existence of languages that lack a clear morpho-syntactic distinction between nouns and verbs?

Question

Is there a strong case for the existence of languages that lack a clear morpho-syntactic distinction between nouns and verbs? If so, what would be an example of a phrase structure for a uniclausal simple transitive sentence in such a language? For instance, what would be the phrase structure of a sentence in one such language which sentence translates as "The boy ate the burger"?

Answer 1

Every language seems to distinguish between the same basic syntactic functions. For instance, we can distinguish between predicates and arguments within a sentence, or between a head and its modifiers in a smaller phrase.

Now, in English, most words are limited as to which function they can play. Nouns like "boy" make good arguments, but they don't make good predicates on their own: you have to tack on a copula (e.g. "be a boy"). Verbs like "ate" make good predicates but they don't make good arguments on their own: you have to build a relative clause around them first (e.g. "the one who ate"). So we can say this:

[The boy]_Arg [ate]_Pred

But sentences like these, intended to mean something like "The one who ate is a boy," are ungrammatical:

[The ate]_Arg [boy]_Pred

[The ate]_Arg [boy-s]_Pred

This limitation is clear evidence for a morphosyntactic noun/verb distinction in English.

Okay. So there are claimed to be languages without this limitation: languages, that is, in which any word can take any syntactic function. The Salishan language family is the classic source for examples of this. Here's one example from St'át'imcets: in this sentence, t'ak 'go' is the predicate...

[T'ak]_Pred [ti=nk'yáp=a]_Arg

[go]_Pred [DET=coyote=DET]_Arg

"Coyote goes along"

...and in this one, nk'yap '[be a] coyote' is the predicate:

[nk'yáp]_Pred [ti=t'ak=a]_Arg

[coyote]_Pred [DET=go=DET]_Arg

"The one going along is Coyote"

(Note that the word order is backwards from English: Salishan predicates come at the start of the clause, before their subjects.)

So people have claimed, based on data like these, that there is no noun/verb distinction in Salishan languages.

The trouble is, in all the Salish languages where people have checked, it turns out that there are other morphosyntactic asymmetries between nouns and verbs. For instance, in St'át'imcets, only nouns can be the head of a relative clause. You can say this...

[ti=tupun-táli=ha]_Modifier [sqaycw]_Head

[DET=hit-3sg=DET]_Modifier [man]_Head

"The man who hit him"

but not this...

[ti=sqaycw=a]_Modifier [tupun-as]_Head

[DET=man=DET]_Modifier [hit-3sg]_Head

lit. "The him-hitter who mans"

(Once again, the St'át'imcets word order is the reverse of the English word order: the modifier comes first, before the head.)

So if you want to describe the syntax of relative clauses in St'át'imcets, you do need a noun/verb distinction after all. And this seems to be the case for the other Salishan languages as well, though the jury is still out on some of them. This handout by Seth Cable is a good place to start for information on the noun/verb distinction in Salishan (and also Wakashan) languages, and the references in it go into more detail. (The St'át'imcets examples here are all from Cable's handout.)

So okay, the Salishan languages have a noun/verb distinction. What about other language families? I'm less familiar with the literature here. Some people have made similar claims about Austronesian languages. David Gil, for instance, claims that Riau Indonesian has no noun/verb distinction. But here's a paper by Brendon Yoder arguing against Gil's conclusions, claiming that Riau Indonesian does have a noun/verb distinction after all.

Long story short: In the cases I'm familiar with where someone has said "Language X lacks a noun/verb distinction," further study has suggested that their claim was dubious our outright wrong. We still can't be dead certain that all languages have the distinction. But it would be a big surprise at this point to find a language that really and truly lacked it.

Answer 2

There seems to be no natural language with just one word class for both nouns and verbs. Although some Philippine-type Austronesian languages like Tagalog are said to have no noun/verb distinction, it seems what they have is weak distinction instead.

According to Himmelmann,

With regard to syntactic categories it may be argued that all Tagalog content words (both roots and derived words) are categorially indistinct, i.e. they may all occur in essentially the same basic syntactic positions (section 3.1)

but Kroeger disputes this saying

The distributional similarities mentioned above between nouns and verbs in Tagalog apply equally to adjectives and even PPs. Each of the four categories may appear as (i) a clausal predicate, (ii) a modifier within the NP, or (iii) a noun phrase..[but]..this difference between verbal and non-verbal clauses provides a clear distributional test for distinguishing between verbs and all other lexical categories: only a verb can function as the predicate of a verbal clause.

According to Himmelmann,

With regard to lexical categories, however, there are clear-cut categorial distinctions (section 3.2).

His conclusion:

An important aspect of the voice system in Tagalog is the fact that voice-marked words (V-words) as well as the roots from which they are derived belong to lexical categories which are very different from the lexical categories found in more familiar languages such as English. It has been repeatedly suggested that the difference pertains to the fact that Tagalog roots are precategorial and/or that there is no distinction between nouns and verbs in Tagalog. The preceding sections present a somewhat different explication of the difference between the two systems of lexical categories and its repercussions for the voice systems. Specifically, it is claimed that Tagalog roots are generally not bound and/or precategorial roots. Instead, Tagalog roots belong to different morpho-lexical classes. That is, it cannot be predicted solely on the basis of their meaning with which affixes a given root may occur. Furthermore, all kinds of roots, including roots denoting ACTIONS, allow for unaffixed uses. In their unaffixed uses, roots may denote THINGS, ANIMATE BEINGS, PROPERTIES, STATES, RESULTS OF ACTIONS, NAMES OF ACTIONS, etc., i.e. concepts which in English are generally rendered by nouns or adjectives. What roots cannot denote is the actual and specific performance of an action. Only V-words may denote the actual performance of an action. In this regard, Tagalog V-words are similar to English verbs. They differ, however, from English verbs in the following important respects:

• While Tagalog V-words clearly belong to their own morpho-lexical category, they do not belong to a special terminal syntactic category. That is, V-words have morphological and semantic properties which set them apart from all other Tagalog content words. But with regard to the positions they may occupy in a phrase structure tree they do not differ from other content words. In English, on the other hand, there is an unambiguous correlation between (morpho-)lexical and terminal syntactic categories: Membership in a given lexical (sub-)category implies membership in a specific terminal syntactic category (‘die’ is an intransitive verb, and hence a verb and not a noun).
• All Tagalog V-words are necessarily derived, while in English there are both basic and derived verbs. The derivation of Tagalog V-words is manifest both formally and semantically: Formally, the morpho-lexical category of the root is changed to the
  morpho-lexical category of V-words. Semantically, an oriented action expression is derived from an expression which denotes a THING, STATE, NAME OF AN ACTION, RESULT OF AN ACTION, etc. The notion oriented action expression conveys two things: First, oriented action expressions denote the actual performance of an action (and not the name or the result of an action). Second, they denote the actual performance of an action in such a way that at the same time they also denote one of the participants involved in the action.
• For English verbs active voice is the basic, non-derived voice. For Tagalog V-words all voices are derived in the same way. Hence, it does not make sense to consider one of the four Tagalog voices the basic voice.