It's a fact that the grammar core of most European languages (not only IE ones) can be analysed in a relatively precise common framework. Of course I do not know much of these languages, but the basic concepts of nouns, verbs, adjectives, verb tenses, pronouns etc. seem to be relevant not only for Indo-European languages, but also for some of their neighbours (Finno-Ugric, Turkish, Semitic, Dravidian...) A perusal of a Finnish Grammar book, for example, will give to the IE reader an occasional surprise (vowel harmony, say), but won't transport him to another planet.

I'm not interested in a list of linguistic curiosities, but in cases where a serious grammatical analysis has been made that required a brand new look on the language. For example, a language grammar lacking a verb vs. noun contrast would certainly qualify.

Of course, while Lojban grammar perfectly illustrates the kind of exotic I'm after, I'm only interested in natural examples.

This is something like the whimsical companion of a more precise question I asked earlier.

9 Answers 9


This is a failed example of what you are looking for, but perhaps it will be interesting for you nonetheless.

At a certain point (in the late 70s, if my memory serves – but I could be off by a bit), linguists working in the generative tradition started describing certain languages as "non-configurational," meaning that they belonged to a class of languages which lack word order (by contrast with most languages familiar to linguists, which were of course acknowledged to have word order).* Various explanations were pursued as to how non-configurational languages could have a grammar (making sure that morphological agreement happens, that verbs have the right number of arguments, etc.) without traditional syntactic notions that are often dependent on order. The paragon example of such a language was held to be Warlpiri (Pama-Nyungan, Australia), but it has now been shown to be analyzable in terms of familiar syntactic theories.

This 1983 article by Hale in NLLT is an example of a non-configurationality analysis of Warlpiri. This 2002 MIT dissertation by Legate** argues for an analysis of the language in line with standard syntactic theories, and by extension that "non-configurational languages" do not form a separate class.

*This is of course soemething of a simplification of the literature on non-configurationality.

**Full disclosure: a professor of mine.


I think role and reference grammar might be the kind of thing you're after. From the PDF overview at that site:

The motivating questions for RRG were, ‘what would a linguistic theory look like if it were based on the analysis of languages with diverse structures, such as Lakhota, Tagalog, Dyirbal and Barai (Papua New Guinea), rather than on the analysis of English?', and ‘how can the interaction of syntax, semantics and pragmatics in different grammatical systems best be captured and explained?’

I'll refrain from much metacommentary, but just keep in mind that its origins were in a fairly different era than the present -- since then there's a ton of (mainstream) linguistic work in non-IE languages as well, and there is much less anti-semantics/pragmatics sentiment embedded in mainstream syntax than there used to be.


Diversity with respect to lexicalization of different word classes is an interesting issue, and you will find some interesting examples in the RRG literature (see @kgr's answer) of cases where the noun-verb distinction is more difficult to motivate. One point made by Croft in his book "Radical Construction Grammar," however, is that it is in general possible to analyze a language so that it has whatever number of lexical categories you like, so at a close level of scrutiny it is hard to motivate such claims as "Language X does not distinguish nouns and verbs."

Outside of lexicalization patterns, there are some decidedly non-IE patterns which are documented in high quality descriptions of various languages. Some examples:
Languages where the concept of syllable is called into question: Bella Coola, Gokana
Languages where almost any category can predicate: Classical Nahuatl, Mwotlap
Languages where the concept of subject is called into question: Acehnese
Languages where cardinal directions are used in place of "left" and "right": Guugu Yimithirr
Languages with up to half a dozen consecutive verbs in the same clause: Isu

  • What do you mean by "any category can predicate"?
    – Joe
    Commented Nov 16, 2011 at 16:39
  • 1
    Let's say there are words in a language which has this property, translatable as "eagle", "two", "no", and "black". If all of these words could predicate, then they might be used in such a way that they were translatable into English as "to be/become an eagle"; "to be/become two in number"; "to be missing or non-existent"; "to be/become black."
    – user483
    Commented Nov 16, 2011 at 16:49
  • @jlovegren But then, couldn't we just point to any word that predicates and call it a verb?
    – Qwertie
    Commented Nov 19, 2011 at 1:44
  • You can do that, but you will usually have reasons for wanting to distinguish word classes on other grounds besides predication. Imagine that one group of words, call it group A, inflects for person and number, but another group, group B, doesn't. And only group B words can control agreement on the inflectional marking that group A words undergo. But both group A and B can predicate, so they are both verbs. You will have difficulty describing how the inflectional morphology works.
    – user483
    Commented Nov 19, 2011 at 3:44
  • For syllables, you might add Miyako. Commented Aug 22, 2012 at 22:58

There are many good descriptions of languages from all over the world, many very different to IE languages. You ask about languages with no contrast between noun and verb: Samoan (and some other Polynesian languages) are the usual examples here, but this issue is still disputed. It's important to remember that the range of variation of languages is not without limits.

Regarding Warlpiri and free word order, its worth remembering that Latin has very free word order too, and in fact it seems that languages with case-marking often have pretty free word order.


I'd start with looking for descriptions that don't pertain to follow a specific framework like Minimalism, LFG or any of the many others but try to use and if necessary invent structures and terms that make sense for a prticular language. It may make comparisons harder but it may also may lead to a eureka moment. Schachter and Otanes' reference grammar for Tagalog, from 1972 is supposedly a good example.


The simple answer is that there have been analyses attempted; however, sociocultural factors make it very difficult to have true analyses of non-IE languages. Most of the people working on grammatical frameworks do it from an IE standpoint, and thus the tools simply never get developed to work in a true cross linguistic context.

  • I would disagree. There are many excellent grammars of non-IE languages, from language families all around the world. The field I'm most familiar with (Australia and Papua New Guinea) has many examples. While it is true that some of the major theories of language have in the past been based largely in IE languages (with a few favourite others), there is much grammatical description that is not locked into particular theories, but are eclectic and functional, drawing on whatever linguistic concepts and tools make sense for the particular language. Commented Oct 8, 2011 at 12:41
  • Here's an example, a sketch grammar of Teop, an Oceanic language (Austronesian family) from Bougainville Island. I'm not sure if it's 'exotic' enough but it's quite different from Germanic and Romance languages. Commented Oct 8, 2011 at 13:05

There are a fairly large number of works on theoretical syntax applied to North American indigenous languages, for example on Stʼatʼimcets (Lillooet), Halkomelem, Plains Cree, Blackfoot, Navajo, and Slave, where the linguists have worked firmly within the Government & Binding → Minimalism framework. The work on the syntax of Salishan languages gives serious problems for syntax in terms of part-of-speech categories (noun vs. verb vs. adjective), and also in word order (they are mostly VSO); see work by Lisa Matthewson, Henry Davis, Martina Wiltschko, and others. Athabaskan, Algonquian, and Eskimo-Aleut languages challenge the idea that sentences need to have subjects and objects, since all of these families frequently show sentences consisting solely of verbs and perhaps adverbs; see work by e.g. Rose-Marie Déchaine, Meredith Johnson, Clare Cook, and Jeff Meuhlbauer. The Athabaskan (and Na-Dene) languages strongly challenge the concepts of morphology, since in these languages the verbs are constructed in a non-concatenative way that looks much more like syntax than it does conventional morphology; see work by Keren Rice, Leslie Saxon, Eloise Jelinek, and Ken Hale, among several others.


Some Native American languages obligatorily mark sentence perspective. Aymara is one of them and the following paper provides many examples: Pragmatic Structures in Aymara


I think you either underestimate the uniqueness of Finnish, or have read a wrong book on it. Among its other interesting features, Finnish has verb declention, marked evidentiality, active paradigm (like Aimara or Kechua) and near-to-obligatory (in some cases) possessive markers. None of this categories is expressed in mainstream Finnish grammars.

Oh, and did I mentioned the five types of 'infinitives' in Finnish?

The grammar of Finnish in its traditional form does have a lot in common with Germanic grammars (e.g. so-called 'weak stems' vs 'strong stems'), but a comparative analysis of other Finno-Ugric languages demonstrates the categories absent in IU languages (evidentiality, noun comparison, and a certain non-discrimination between parts of speech).

To answer the part of your question about languages without parts of speech, Chinese, Yoruba or any other 'isolating' language would serve as a perfect example of a language without parts of speech.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.