Grammatical cases derivation

If you're new to these arts, this is the place to ask "stupid" questions and get directions!
Post Reply
User avatar
samsam
hieroglyphic
hieroglyphic
Posts: 32
Joined: 18 Jul 2018 16:40

Grammatical cases derivation

Post by samsam »

I would like to create a Satem language derived from PIE. How can I make evolve grammatical cases (and other grammatical features) ?
User avatar
eldin raigmore
korean
korean
Posts: 6352
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Grammatical cases derivation

Post by eldin raigmore »

This post won’t solve everything, but:

Case-markings can be words; when they are, they are usually called “adpositions”, or sometimes “realtorsrelators”. (Goddam autocorrect)
“Prepositions” always come just before the first word of their object noun-phrase;
“Postpositions” always come just after the last word of their object NP;
“Inpositions” sometimes come between words of their object NP, for instance,
* just after the first word, provided that’s not always the head noun, or
* just before the last word, provided that’s not always the head noun.
And there may be other kinds of adpositions.

Case-suffixes and case-prefixes usually come from postpositions and prepositions, respectively, or at least from words or clitics that might be so considered. Tagalog has case-infixes. Some 3Cons natlangs have some case-transfixes. There are languages with morphological tone that have case-tones. And there are other case-suprafixes, such as chronemes (chronofixes?) and stressemes.

Adpositions can come from nearly any part-of-speech. I think I’ve read they’re likelier to come from adverbs than from any other source ancestor-word; but I know they sometimes come from adjectives or nouns or verbal auxiliaries or even verbs.
Edit: ”Ago” (English’s only postposition) comes from a verb.
Possibly (?) you could get a case-affix directly from an adverb, or some other part-of-speech, without going through the “adposition phase”.

Some cases are “ad-verbal”, in that they tell the relationship between the cased noun and some verb (usually the main, nuclear verb of the clause); some are “ad-nominal”, in that they tell the relationship between the cased noun and some other noun.
Some cases are “syntactic”, in that they tell the syntactic role played by the cased noun in the clause or phrase, regardless of its semantic role; some are “semantic” in that they tell the semantic role played by the cased noun.

Nearly all cases mean more than one thing. Some grammarians won’t call whatever system a natlang has, a “case system”, unless at least one of its cases has both a syntactic use and a semantic use.

To mention just four cases;
“Nominative”, “Accusative”, “Dative”, and “Genitive”, are all usually used as, and/or considered as, syntactic cases.
NOM and ACC and DAT are (principally) ad-verbial; GEN is (mostly) ad-nominal.

Will that start you off?

You could check out and read Barry J. Blake’s book “Case”.
Last edited by eldin raigmore on 27 Aug 2018 17:16, edited 3 times in total.
Clio
sinic
sinic
Posts: 228
Joined: 27 Dec 2012 23:45

Re: Grammatical cases derivation

Post by Clio »

Since you're deriving a language from Proto-Indo-European, a good starting-point is probably to take a look at the grammatical cases reconstructed for PIE. My first step is always to run some nominal paradigms through the sound changes I've using to derive my language and then to see which cases remain distinct.* Formal mergers of formerly different cases might elucidate the specifics of case usage; when two cases stop being distinct, the new noun-form might take on the functions of both cases. This happened in Greek, which lost the ablative as a distinct case, so the genitive started to be used to indicate motion away from something in addition to its original functions.

Beyond what you get from applying sound changes, additional case marking could derive from any of the sources eldin mentioned above. The corollary to his post is that you can also get rid of morphologically distinct cases by having prepositions used to indicate the same meaning as the case form. The Romance languages are a great example of this phenomenon; they use words from Latin de 'from, about' to indicate possession, which used to be the purview of the genitive case.

*There's nothing wrong about working backwards a little bit, too, since this is a creative endeavor. For instance, if you know you want your language to preserve the ablative, you might consider working out certain sound changes that will keep it distinct from the other cases, or you could introduce other case markers from analogy or other sources. The Sabellian languages, for instance, imported the ablative marker *-d from the pronominal and o-stem declensions to other declensions, which resulted in ablative singulars that were distinct from the corresponding genitive singulars.
Last edited by Clio on 27 Aug 2018 17:50, edited 1 time in total.
Niûro nCora
Getic: longum Getico murmur in ore fuit
scratchpad
User avatar
samsam
hieroglyphic
hieroglyphic
Posts: 32
Joined: 18 Jul 2018 16:40

Re: Grammatical cases derivation

Post by samsam »

Thanks a lot you both!
User avatar
eldin raigmore
korean
korean
Posts: 6352
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Grammatical cases derivation

Post by eldin raigmore »

You don’t have to pick just one technique.
Lots of case languages combine case-inflection with adpositions.
For instance, Greek. (I don’t know whether I mean Classical Greek, or New Testament Greek, or Modern Greek.)
With four inflectional cases and four adpositions, I think they got about a dozen effective cases.
Or something; it’s all hearsay coming from me, I don’t really know much Greek.
Tristan Radicz
rupestrian
rupestrian
Posts: 17
Joined: 21 Sep 2018 01:56

Re: Grammatical cases derivation

Post by Tristan Radicz »

All solid advice here.

I'd like to add that if you work straight from the reconstructed PIE declension paradigms, you'll inevitably end up with some syncretism in the case endings (e. g., nominative plurals indistinguishable from genitive singulars in many of the declension classes) - it's kind of built in the system. This is not a big deal, it has been the case with all of the real world Indo-European languages after all. The exact amount of syncretism depends on the sound changes you decide to apply, and if you ever feel like the system is getting too ambiguous, you can work around the issue by employing things like syntax, accentuation, analogy and, in the end of the day, inventing a certain ending from scratch and leaving it unexplained (all of this happens in natlangs). For example, in my own Indo-European conlang the ablative and genitive singular in most paradigms have identical flexions and so do the dative and locative singular, but you can always tell which is which, because ablative and locative are used with prepositions only, while genitive and dative never take prepositions. Another solution is to replace the old case system with an innovative one, e. g. build it on the basis of some of the oblique cases - this is what Romani and Tocharian languages did.

And remember, even if rampant case syncretism is likely to induce a collapse of some of the cases with each other (or, indeed, of the whole system), this need not happen - look at what some of Czech declension paradigms look like! [:D]
User avatar
samsam
hieroglyphic
hieroglyphic
Posts: 32
Joined: 18 Jul 2018 16:40

Re: Grammatical cases derivation

Post by samsam »

Thank you a lot !
Post Reply