14

Among conlangers, AllNoun is a notable syntax because it only makes use one part of speech / word class, which is analagous to nouns. A natural language I've heard of (but I can't remember or find a reference to it anywhere, I think it was an Australian aboriginal language) only has three verbs. I've heard that there are some North American languages that don't exactly have nouns, but I'm not actually sure if this is true or not.

What other languages don't have parts of speech or word classes? Which do they lack, and how does their syntax work in a way so that they are not needed?

hippietrail
  • 14,687
  • 7
  • 61
  • 146
Peter Olson
  • 1,412
  • 1
  • 13
  • 23
  • Aboriginal? For where? Australia? Asia? N America? S America? Etc? – Mitch Sep 19 '11 at 00:02
  • @Mitch Australia – Peter Olson Sep 19 '11 at 01:04
  • I had not heard of AllNoun before this, but from the description, it looks like a spoken version of Discourse Representation Structures – prash Sep 19 '11 at 06:02
  • Do you mean Jingulu? Pensalfini (2004) deals with it in Natural Language and Linguistic Theory 22. – Arne Sep 19 '11 at 19:30
  • While I don't have an answer to your question, I think it would be fruitful to explore language isolates in the hope that these languages have not developed well for one reason or another (trade, wars, etc.). – prash Sep 19 '11 at 06:02
  • @prash How is a language that has no surviving relatives, or that doesn't have a clear distinction between two parts of speech, less "well-developed"? – Alek Storm Sep 19 '11 at 06:58
  • Ah! Well-developed! I need to write a large article on this some day. No one seems to share my view that some languages are objectively better (or better developed) than others. I'll leave aside the poverty of vocabulary (which can always be invented, especially for open-class words) and come straight down to grammar. Are you a programmer? Have you observed how some programming languages offer an easier syntax to express oneself, and how another makes a task tedious, or how one language is suited for one domain, while another is for another? I hold spoken languages up for similar scrutiny. – prash Sep 19 '11 at 08:10
  • (cont.: space constraints here, as before. I'll mention one example.) Consider http://en.wikipedia.org/wiki/Pirah%C3%A3_language leave aside the people who speak it, their culture, etc. Read the section on Embedding. The language does not have words for colours of for (any!) numbers, but for this exercise, let's say it does. Assume a vocabulary as large as that of English. Imagine how you would express the idea "The boy's green ball was bigger than the girl's red ball, but smaller than her green ball." – prash Sep 19 '11 at 08:27
  • I would be interested to see an objective metric you could come up with to measure language expressiveness. You asked me to disregard the culture of the Pirahã people and consider their language on its own; this is impossible. Language and culture are inextricably linked. It may simply be that in the Pirahã's native environment, colors are identifiable from context - if they have no way to change the color of a particular object (say, by painting it), then most objects have a usual color, no? Water is blue, plants are green, etc. – Alek Storm Sep 19 '11 at 17:05
  • I am indeed a programmer, and I have personally seen that one person's preference for a language depends more on aesthetics than any "objective" suitability for a task, as exemplified by the eternal "Language Wars" between the top four or five most popular languages of the day. If you've developed an objective metric for this domain of languages as well, I'm sure these people would love to see their debate resolved. – Alek Storm Sep 19 '11 at 17:08
  • @Alek Storm: w.r.t. your first comment: I agree with all you say here. That is why I originally asked you to assume a large vocabulary. What I was trying to focus on was that the grammar of the language makes expressing my example sentence difficult. W.r.t. your second comment: I do not have an objective metric. I do not think I ever will, because I am not a linguist and this is not what primarily interests me. continued... – prash Sep 19 '11 at 17:19
  • Irrespective of objective metrics for both spoken and programming languages, can we at least qualitatively say that COBOL and BASIC are bad languages for designing complex systems? (It's been done, to be sure.) All I'm trying to get at, for now, is that not all languages are equally well-suited at expressing various kinds of ideas. (disregard vocabulary!) I can make a better argument if you agree with this. If not, I withdraw. – prash Sep 19 '11 at 17:24
  • let us continue this discussion in chat – Alek Storm Sep 19 '11 at 17:26
  • Are you asking about the traditional concept of *part of speech* or the modern concept of *word class*? – hippietrail Sep 20 '11 at 11:02
  • 2
    @hippietrail We are on a linguistics site, so what the 8th grade English teacher or Aristotle thinks is only interesting as a question of the history of science. So I think we should read the question as if it says "word class" – MatthewMartin Sep 20 '11 at 16:27
  • @Matthew: OK I will go ahead and edit the question to clarify both the aboriginal issue and the POS vs word class issue. I don't think it's safe to assume everybody with a linguistics question is familiar with all the current changing terminology though. – hippietrail Sep 20 '11 at 16:30
  • 1
    @prash et al. A quick mental experiment is in order. Would a piraha speaker be struck dumb (speechless) when confronted with a boy with a large green ball, a girl with a smaller red ball? Its unlikely. Language without simple words for color typically use other strategies, like "leaf colored", "blood colored". Being able (or not) to use a given strategy (lexicalization) doesn't exclude the option to use other strategies, including discourse (speaking at length to express a scenario). – MatthewMartin Sep 20 '11 at 16:31
  • @hippietrail I like to use links to wikipedia to get me permission to use my favorite $50 words. Wikipedia's basic linguistics articles are excellent. – MatthewMartin Sep 20 '11 at 16:32
  • @Matthew: I was going to do that earlier to clarify the two terms but currently Wikipedia covers both in one article. I will link to it in a broad way though. I think linking to either Wikipedia or specialist Linguistics sites for technical terminology is always good and edits to add them should be encouraged. – hippietrail Sep 20 '11 at 16:35
  • @MatthewMartin: I'll responded here: http://chat.stackexchange.com/rooms/1376/discussion-between-alek-storm-and-prash – prash Sep 20 '11 at 16:40
  • @PeterOlson: Were you aware that by some analyses some languages completely lack lexical word classes? By some accounts in Chinese and Polynesian languages any word can function as any part of speech. That means part of speech is for those language is down to the syntax rather than the lexicon. Dictionaries conservatively take the traditional approach however. – hippietrail Sep 20 '11 at 21:11
  • @hippietrail Could you point me to these analyses? I think it's pretty much accepted across linguistics that every language has at least two POS (part of speech) categories and most have more. Why do you distinguish between morphology and syntax in this? English, having zero-derivation, needs syntax to define it's POS categories. – Gaston Ümlaut Oct 11 '11 at 11:57
  • @Gaston: No I don't have any references. I'm not a real linguist, I just read a lot about languages but I have read on multiple occasions of accounts that either Polynesian or Chinese languages don't really have word classes. I can't say whether linguists were writing these ideas. I'll certainly be keeping my eyes open for references now though that I have a site to use them on (-: – hippietrail Oct 12 '11 at 08:20

5 Answers5

11

For many languages it is claimed that they lack adjectives (for instance Thai), and that either nouns or verbs are used instead. Others claim that all languages have adjectives, if you look closely enough. Japanese has two types of adjectives: noun-like and verb-like. However it can also be analysed as having no adjectives, if the verb-like adjectives were considered just to be odd verbs and the noun-like adjectives were considered just to be odd nouns. In some languages adjectives are a closed class of words - are they still then adjectives? Not if being an open class is part of the definition of adjective :) (An open class may gain and lose words while a closed class is of a much more stable size.)

Ditto for languages that are claimed to lack nouns or verbs, it's a matter of how "noun" and "verb" is defined. AFAIK there is not yet consensus on any language lacking nouns or verbs.

Lacking articles (the, a) is not so uncommon, but I don't think there are languages that lack determiners as a whole (articles are a subset of determiners).

kaleissin
  • 4,163
  • 24
  • 36
  • Under the traditional concept of *part of speech* article is not one of the parts but is a kind of adjective. In modern approaches articles are part of the *word class* of determiners. – hippietrail Sep 20 '11 at 11:06
  • 1
    How part of speech is defined depends on language and also varies over time. "Part of speech" itself is outmoded, isn't it supposed to be called "word class" now? – kaleissin Sep 20 '11 at 14:22
  • Indeed, and I made a comment about that on the question itself. Either way it seems that articles are not regarded as parts of speech or a word class, but a subcategory. – hippietrail Sep 20 '11 at 14:41
  • 4
    'Part of speech' outmoded? In linguistics each of part of speech (POS), word class, and lexical category (and various combinations) are used, interchangeably. – Gaston Ümlaut Oct 11 '11 at 12:01
8

Since the statement "X has no Y" can be disproved with just one word, you might find more researchers willing to say "X has very few Y". There are several languages in Papua New Guinea that have remarkably few verbs, sometimes as few as twenty.

There is an aboriginal language that some people claim has only three verbs, but when I looked at the reference grammar, the verb turns out to be incredibly complex, so I suppose one could just as easily argue that the language had lots of verbs and a required morpheme that came in three versions.

Here is another interesting paper about languages with very few verbs.

Ref. The Papuan languages of New Guinea

MatthewMartin
  • 2,914
  • 19
  • 33
3

I suspect one could make the claim that most languages either lack counters (or measure words) or that they've become (mostly) extinct. An example in English would be cattle. One does not say "I have 3 cattle", instead one says "I have 3 head of cattle". The book I suspect you are thinking of is Women, Fire and Dangerous Things. In it, in Chapter 6, there is also a short discussion of a single Japanese counter, hon (which is used to count long thing things such as sticks, pencils, rivers, phone calls, ball game hits & pitches and dried snakes) and how these categorize how these are in some respects metaphors for how one thinks about the world.

Tangurena
  • 755
  • 4
  • 13
3

Re AllNoun: as this is a conlang, and an artificial 'language', I'd suggest that it can't be considered a true language until it has native speakers, and that if it did have native speakers they would expand the language to suit the full range of requirements of human beings, such that it would have at least two POS categories, one that includes 'verb'-type words and one with 'noun'-type words. (Schachter 1985). This is what happens when speakers turn a pidgin into a creole.

Yes, some languages have a small, closed verb category. But in these languages the verbs are highly generic and combine with other elements (which may include coverbs, adjectives, nouns, other verbs, etc) to create verb complexes (or complex predicates) that have more specific meanings. In this way they are able to express everything that a human being needs to. (Schultze-Berndt 2000)

I doubt there are any languages without nouns. The smallest number of POS categories that I have heard about in any language is two, a 'verbs' category and a 'nouns' category. But each of these categories will contain numerous words that can be used to fill the functions that in other languages will fall into different POS categories; ie, the full range of functions is there it's just that, morphosyntactically, they are compressed into two POS categories. Thus Australian languages are usually analysed as having no distinct category of adjective. Instead, words that have adjectival meanings will be members of the same category that includes nouns, in a category often called 'nominal' in grammars of Australian languages. (Goddard 1985)

Finally, many languages will lack some part of speech that's present in some other languages, eg English doesn't have relators or coverbs. It's impossible to give a brief answer to how languages deal with this and in fact, this is what linguistic analysis is largely about, finding out how each language uses its particular set of grammatical and lexical elements to achieve the full range of functions that humans need. I've already described how languages with a small, closed verb class are able to express the full range of meanings required by speakers. Another example could be the question, how does English get along without politeness markers? It does so by using special vocabulary and morphosyntactic constructions, and I doubt many native speakers feel a lack. (Schachter 1985)

Gaston Ümlaut
  • 6,697
  • 30
  • 46
0

Petrov, Slav, Dipanjan Das, and Ryan McDonald. "A universal part-of-speech tagset." arXiv preprint arXiv:1104.2086 (2011). contains an interesting comparison of the number of tags between different annotated data sets in different languages:

enter image description here

The number of tags depends on how much fine-grained the annotations are intended to be, as well as the language itself.

The appendix contains a proposed mappings from language-specific part-of-speech tags to our universal part-of-speech tags.

enter image description here

enter image description here

More information can be found on the Universal Dependencies website.

Franck Dernoncourt
  • 1,588
  • 2
  • 12
  • 35