> I'm going over the Tirelat grammar to take a break from the vocabulary
> work, and improve the documentation. The current topic is noun cases.
> It's questionable whether I should call them cases at all, since they're
> marked not on the noun but a particle attached to the start of the noun
> phrase. Case and number are fused into a single word. E.g.
>
> saj zarvi Ä‹ik
> NOM.PL bare rock
> (the) bare rocks
>
> The adjective and the noun remain uninflected. There may be an argument
> for considering nominative and accusative as the only two "real" cases,
> since there is agreement on the verb for each of those, and pronouns
> have separate nominative vs. accusative forms, but I'll continue using
> "case" for the rest of these noun phrase particles.
>
> As I've mentioned, nominative and accusative are two prominent cases.
> Most intransitive verbs have a nominative subject, and most transitive
> verbs have a subject in the nominative and an object in the accusative.
> The nominative case tends to be used for agents and experiencers, and
> the accusative case for patients (but also destinations and some other
> usages).
>
> The nominative case has three forms: su (animate singular), sy
> (inanimate singular) and saj (plural). This is the only case so far that
> distinguishes between animate and inanimate nouns, although it's
> possible that other cases may make this distinction. The accusative case
> marker is my (singular) or maj (plural).
>
> Some verbs also have an indirect object, which will be in the dative
> case (na/naj, used typically for a recipient or a beneficiary). The
> locative case (vë/vaj) represents the location of an action.
>
> Now for the more interesting ones. Some verbs, of which "veste" (to
> dance) is typical, have an extra subject (kë/kaj). I call this the
> "comitative" case, since the English translations of these verbs tend to
> use "with" where Tirelat uses this case.
>
> Another lesser used case (dy/daj) is one that I haven't had a name for,
> but I'm thinking of "ablative". Typically, it's used for non-volitional
> agents, or the agent of a passive verb, which according to Wikipedia are
> possible uses of the ablative case in Latin, and I've also considered
> using this case to represent the "source" of an action (one translation
> of English "from" or "out of").
>
> Then there are the two remaining cases, which unlike the preceding
> cases, are not associated with particular "argument slots" of verbs. The
> vocative case, used to direct a comment to a particular listener (or
> listeners) has only a single form (ee), not distinguished by singular or
> plural.
>
> The last case (u/aj) is one that I've been calling "oblique", as it
> typically is not associated with a verb; its most typical use is as the
> object of a locative noun phrase.
>
> vë lak u łivi žeġ
> LOC.SG over OBL.SG lazy dog
> over the lazy dog
>
> However, the "Swallow Song" translation on the Tirelat web page
> (original at
http://archives.conlang.info/shi/jhoezan/bibiaghein.html)
> has a couple of unusual usages of this case.
>
> jë-kóta -z u zhemi nallu, u zhemi rága
> 3s-carry-IPF OBL beautiful hour OBL beautiful year
> bringing hours of beauty, years of beauty
>
> ty rÃlh -ë-njem sha mifsi -njem zhlara-ru -ka -j u tegla
> and wheat- -bread nor legume-bread refuse-PAS-NEG-GA OBL swallow
> And the swallow does not disdain wheat-bread or pulse-porridge, either.
>
> zhúvi, zhúvi u shúru më-bala-ta -j na tegla
> open open OBL door 3s-set -DEO-GA DAT swallow
> Open, open the door for the swallow
>
> I don't know if I should try to account for these cases. Maybe it would
> be better to make a new translation from the original Czirehlat version
> from 2001 (
http://archives.conlang.info/kau/sueljei/jhorqaenghein.html),
> or directly from the original Greek.
>
> kóta -z me jémi nállu, me jémi rága
> carry-IMPF(PART) OBJ beautiful hour OBJ beautiful year
>
> te me rÃhl -e-niém sza mÃfsi -niém se tégla jelára-kai
> and OBJ wheat- -bread nor legume-bread SUBJ swallow refuse-NEG
>
> júvi, júvi se szúru bála-tai na tégla
> open open OBJ door set -DEON DAT swallow
>
> (Either the "se" or the "OBJ" in the last line must be a mistake, but
> I'm not sure whether "se" or "me" is correct in this case.)