The only languages for which I have found a book (not even a webpage) is for Hebrew and Arabic. Are there books or webpages that contain all the noun declensions and verb "conjugations" (or noun and verb forms generally) for different languages? In particular I am looking for:

  • Tibetan (Found a book on Tibetan verbs, but not the conjugations)
  • Sanskrit (Found verb roots on a Wiki, but not sure how thorough, and no conjugations)
  • Cree
  • Ancient Egyptian
  • Coptic
  • Gothic
  • Chinese
  • Korean
  • Japanese
  • Vietnamese
  • Thai
  • Khmer
  • Georgian (the only halfway book is 1200 verb list, but it doesn't seem to have conjugations at all)
  • Armenian
  • Sumerian
  • Akkadian
  • Phoenician
  • Navajo
  • Nahuatl
  • Zapotec
  • Quechuan
  • Ancient Greek
  • Classical Latin

Basically I am looking for the conjugation/declension/"form" tables for these languages, essentially the data so that I know that I am morphing a verb or noun correctly. Because, like in Gothic, while you may find an "a-stem noun" (or even a list), I suspect that there are slight variations in the form of the modified noun in a case-specific manner. So having a list of all the forms of each verb and noun would confirm whether or not applying a general pattern to a new verb or noun is correct.

I have looked for verb and noun books for all of the above languages and didn't find anything worthwhile. Wondering if this just doesn't exist, or the types of places I should be looking.

It doesn't have to be exactly compiled into tables. It can be in any form but basically I would like to find all the forms of the words mapped out in some way, so I can know if a "generic pattern" can be accurately applied to new forms following the same pattern (like -er verbs), or if it is an outlier/irregular pattern like tener and tengo in Spanish.

For some reason, the best site so far has been Wiktionary, like for Ancient Greek Verbs. However, they just list the possible verbs, no conjugations or anything. I have heard Ancient Greek verbs are highly irregular, and many students have a hard time learning the edge cases. So I would (in this case) like to find a list of all Greek verbs in conjugated form, so I could see the conjugations. But same applies to any language (cross-linguistics).

This is the best I've found for Ancient Greek, which isn't saying much.

If no such reliable, respected (i.e. not spam), aggregated resource exists, what do I have to do to put together a list myself? Scan each book for the few pieces they contain? Or what would be effective?

  • This sounds like an impossible request for some languages. For Korean (with very rich verbal inflection), just explaining all the possible suffixes will easily take up a third of a decent grammar book. – jick Oct 2 '19 at 2:03
  • Chinese noun declensions and verb conjugations? I mean even in Old Chinese, that's a stretch. – Michaelyus Oct 2 '19 at 10:17

Such a thing generally doesn't exist, since it wouldn't be useful.

For languages with complicated verbal morphology, such a list would take up several volumes without really communicating much. In Lingála, for example, a back-of-the-envelope calculation says there are over 72,000 morphologically distinct finite forms for each verb! (*)

For comparison, this is almost the length of the first Harry Potter book—for a single verb stem. A quick search through a dictionary shows 200-some verbs, which is a bit on the low end (it's a small dictionary). So listing every form for every verb would be over fourteen million words. That's twenty-some copies of War and Peace filled with nothing but lists of verb forms. And that's with low estimates for every factor, and not even getting to nouns…

However, these fourteen million verb forms are created using a relatively small set of relatively simple rules. Experiments have shown that humans learn a language by internalizing these rules, not by memorizing every form they see (once again, the wug test is the classic example). And it takes significantly less space to list out these rules than to show all fourteen million forms. This is what any decent grammar will do, listing out the morphological rules, along with any relevant exceptions that have to be memorized specifically.

To put it differently: those fourteen million forms have a very low Kolmogorov complexity, and can be very easily compressed down into a couple of pages of code. Since there's no real reason to deal with the uncompressed version, nobody ever wastes the time and energy and resources to print it all out.

What you can find, if you really want them, are automatic paradigm generators, such as the one you linked for Sanskrit. This is basically what Wiktionary does to fill in its inflection charts automatically. If you search for [language name] morphological analyzer you can sometimes find very good open-source tools for this sort of thing.

(*) The actual number of forms per verb is significantly higher. For this estimate I left off infinitives, imperatives, all possible homophones (where the meaning is different but the surface form is the same), and didn't allow stacking the same extension more than once. In theory, extension stacking means a verb can grow infinitely long, though in practice this would never happen. So the number of possible forms, in theory, is unbounded.

  • Oh wow that's a good example, thanks! So then since there are cases where there are tens of thousands / millions of forms of a single item, then I imagine you are bound to have edge cases that don't fit the pattern. Likely too, since we're not recording every form, your "rules" will get some things wrong. So the grammar must be a living document and can't be predefined perfectly like a programming language. Is that what you're saying? – Lance Pollard Oct 2 '19 at 7:42
  • That is, your rules are going to be inaccurate in some edge cases that you didn't encounter. If you speak based on the rules in the grammar, you will get that wrong, and have to be corrected live, then you internalize that, etc. But often these cases won't get included in the grammar simply because there's (probably) too many edge cases to keep track of. – Lance Pollard Oct 2 '19 at 7:44
  • And also, then why do they have books on it in Hebrew and Arabic? – Lance Pollard Oct 2 '19 at 7:45
  • 3
    @LancePollard There are absolutely books on it—but they record the rules, not just a list of forms. Basically, our brains are ruthlessly efficient at compression. Mentally, we don't store these millions of forms—we only store the rules, and the exceptions. This is one way that things get regularized over time, since removing the exceptions makes it more compressible. – Draconis Oct 2 '19 at 15:16
  • 1
    @LancePollard For example, English has been pruning away its strong verbs over time (strong verbs are ones which change their vowels in the past tense), replacing them with weak verbs—since for us, weak verbs are regular and strong verbs are exceptions. Old forms like "holpen" have been replaced with new, regularized forms like "helped". – Draconis Oct 2 '19 at 15:17

For the Saami languages, you can get paradigms here. E.g. select North Saami, All tools, Paradigm generation then (presumably) full paradigm. You will then fill in a word, optionally a part of speech, then it gives you the full paradigm. Though the paradigm generator is in English and other languages, the dictionary resources aren't in English, but you can get lexical items clicking around here. Their resources are open if you're up for it, you can download the databases and code and build a machine to generate massive paradigm lists. As far as I can tell, the North Saami material is absolutely correct (w.r.t. the written language: there are individual phonetic details not covered). You might consider putting Saami on your list.

The concept of "verb form" is a bit anomalous for Chinese, Khmer and Vietnamese. Goldstein & Nornang Modern spoken Tibetan has good paradigmatic information. For Navaho, you could look here (a meta-guide to using Young & Morgan).

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