16

I'm working on an application that takes a special database of words and its word class and determines the such from a given sentence. I'm now working to see if word classes that are found in English are found in other non-European languages. If not, how would do you define word classes for Pacific, Asiatic and other languages?

So far, I can take the sentence the boy has a shirt, tag each word properly and using a Spanish database now, convert that same text to Spanish el nino tiene una camisa. In essence, I'm creating a multi-functional translation engine, but it won't be used to translate; it's more for simple human-to-machine translation.

NOTE: I wanted to post this on StackOverflow, but the question was more tied to linguistics than programming.

  • See also this question and its answers: linguistics.stackexchange.com/questions/12777/… – jk - Reinstate Monica Aug 8 '16 at 12:38
  • 1
    The class of words that start with a consonant is universal. – Greg Lee Jun 3 '18 at 8:37
  • 1
    What you probly want is an augmented tag set, which distinguishes (e.g) transitive from intransitive verbs, predicate from attributive adjective, and so on. If you're doing a lot of SEAsian languages, you'll want an overstuffed feature table, because almost all such features are critical in some area language. – jlawler Jun 3 '18 at 16:40
  • 1
    @jick I don't know, but I rolled back the version because of the introduction of a grammatical error – jk - Reinstate Monica Jun 6 '18 at 14:03
  • 2
    This may be obvious but keep in mind that translation frequently has to change word classes, even if both languages have similar classes. The way to express "I like this cake" in Japanese is roughly "This cake by me cherished is", where "cherished" here refers to the adjective suki – Eng express the state of "liking something" with a verb, J with an adjective. The German way of saying "I like travelling" is "I travel gladly", with an adverb (gern). – melissa_boiko Jun 10 '18 at 10:08
16

From Word classes and parts of speech (pdf), a 2001 paper:

Despite the theoretical problems in defining word classes in general, in practice it is often not difficult to agree on the use of these terms in a particular language. This is because nouns, verbs, and adjectives show great similarities in their behavior across languages. [...]

The general properties of nouns, verbs, and adjectives ... are sufficient to establish these classes without much doubt in a great many languages. However, again and again linguists report on languages where such a threefold subdivision does not seem appropriate. Particularly problematic are adjectives ... but languages lacking a noun- verb distinction are also claimed to exist ..., and ... adverbs ... present difficulties in all languages.

And towards the end:

Hengeveld (1992a) proposed that major word classes can either be lacking in a language (then it is called rigid) or a language may not differentiate between two word classes (then it is called flexible). Thus, `languages without adjectives' ... are either flexible in that they combine nouns and adjectives in one class (N/Adj), or rigid in that they lack adjectives completely.

Hengeveld claims that besides the English type, where all four classes (V - N - Adj - Adv) are differentiated and exist, there are only three types of rigid languages (V - N - Adj, e.g., Wambon; V - N, e.g., Hausa; and V, e.g., Tuscarora), and three types of flexible languages (V - N - Adj/Adv, e.g., German; V - N/Adj/Adv, e.g., Quechua; V/N/Adj/Adv, e.g., Samoan).

Universal language support is tricky at best, as far as I know linguists are still arguing about which aspects of language are universal and to what extent. (Universal here meaning "applying to every natural language that could conceivably be used by a human being".)

From a practical standpoint, e.g. for the purposes of making a program, there is also the question of how common the various language classes are. For example Tuscarora, mentioned as an example above, has a grand total of 52 speakers (according to Wikipedia), and it may become a business decision how far out of your way you are willing to go to support it.

IANAL (with the L here meaning Linguist), I merely read up on similar topics in a similar context a few years ago.

| improve this answer | |
  • 2
    That's very interesting. I would have expected that there would be parts of speech that WEREN'T in English, but I guess there's really just nouns and verbs, and then the "HOW" versions of each. That's all there is, and all there can be. And there are dialects of English that are at least really light on adverbs, too. – Chris B. Behrens Feb 10 '11 at 22:54
  • 4
    @Chris There are parts of speech that aren't in English, for example classifiers in Chinese, Korean and Japanese as mentioned by Benoit (and Wikipedia Parts of speech); there are probably others - this paper just focused on the major ones. Part of the difficulty of saying something universal is that there are at least 6,900 languages in the world - compare to ~200 countries - and many of them have never been properly studied, so we are biased towards the PoS used in the languages we know of. – j-g-faustus Feb 11 '11 at 0:46
  • Yeah, it did occur to me that when you did into the more granular parts of speech, pronouns, for example, that those were probably not universal, and there were bound to be linguistic curly-cues (as John McWhorter puts it) that didn't exist in English. Fun stuff. – Chris B. Behrens Feb 11 '11 at 14:52
9

There is some commonality among the word classes appropriate for analysing different languages, but they don't map completely onto each other.

(Bear in mind also that there is not one single agreed classification for any language)

So I believe all languages have verbs but in some languages (eg Japanese) it is possible to regard adjectives as a subclass of verb. Most languages have nouns, but whether adjectives belong in with them or not may vary (Lojban has no nouns, but I don't know whether there are any natural languages which haven't). I think all languages have pronouns and determiners

| improve this answer | |
  • Good point you've brought up. – Jacky Alcine Feb 11 '11 at 17:53
7

This is an attempt at a universal set of 12 parts-of-speech tags. I do not know if it is used in any online systems.

A Universal Part-of-Speech Tagset, by Slav Petrov, Dipanjan Das and Ryan McDonald.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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