Conlang Class

[Kuhron]

2a) I can conceive of no reason whatsoever for marking the verb with the case of any participant.
Grammatical/syntactical/functional cases, such as Nominative or Absolutive, Accusative or Ergative, Dative or Dechticaetiative, and Genitive or Construct, really only tell you what Grammatical/syntactical Relation/function the thus-cased noun has; so the morpheme marking the case of the Subject or of the primary Object on the verb could only ever have one value.

I don't understand this. It sounds like you're saying that there should be a morpheme indicating that there is a subject and one indicating that there is an object, yet not specifying further. Swahili has a set of subject prefixes for verbs and a set of object prefixes, and that's kind of what I was getting at here, except we made them obligatory even when the nouns are present.

2b) OTOH it's both naturalistic and realistic to mark the semantic role of the Subject on the verb; that's Grammatical Voice.
And marking the semantic role of the primary Object on the verb is a kind of Applicative "voice".
For instance there might be no marking to show that the Subject is the Patient (Absolutive), and likewise none to show that the primary Object is the Agent (Ergative). But if the Subject is the Agent the verb might be marked as Anti-Passive Voice. And if the primary Object is the Recipient the Verb might be marked in some particular Applicative "Voice".

We haven't gotten to voice yet, so all of this will be considered. But I still don't see how it conflicts with the marking on the prefixes. It is true that all they do is mark the syntactic role, not the semantic one, but how is this a problem? The semantic roles will likely be shown by voice.

Assuming your language is Absolutive/Ergative/Dative, then the Subject will be assumed to be the Patient or Theme, the Primary Object will be assumed to be the Agent, and the Secondary Object will be assumed to be the Recipient, unless voice-marking on the verb says different.

I thought it was common for the subject to be the agent and the primary object to be the patient. That's what this language is doing.

Possibly your language has only two Grammatical Relations, namely Subject and Object; in which case you won't have any Secondary Objects, and the case-marking of the third participant in any ditransitive clauses, will specify its Semantic Role (not its Grammatical Relation). Does your language have three-valent verbs? Ditransitive verbs? Two kinds of Object? Three Grammatical Relations?

Not yet decided, but it is likely that there will be ditransitive verbs at least, and thus another grammatical relation. Are there trivalent verbs that are not ditransitive?

Your language also needs a syntactical cases for the Subject of monovalent (that is, one-participant) intransitive clauses.
For most languages exactly one of the cases for monotransitive clauses will be re-used in the intransitive clauses.
If you have S=A that's called Nominative/Accusative alignment. If you have S=P that's called Ergative/Absolutive alignment.
(Other alignments exist; for some of them both monotransitive cases can be used in intransitive clauses, while for others there's a case used in intransitive clauses that's not used in transitive clauses.)

I said in an earlier post that it's ergative/absolutive.

I'll be sure to bring up what you have said in the next class! There are still many basic things that we haven't yet covered, but since we've focused a lot on verbs so far, this all seems like it'll be a big help.

[eldin raigmore]

I don't understand this.

Alright; just look around and find any natlang anywhere that marks the verb with the case of a participant.

Edit: See http://listserv.brown.edu/archives/cgi- ... %3BMatches for a conlang that (at least at one stage) was attempting to mark the subject's case on the verb, and how that contrasted with natlangs that come close to doing the same thing (what Klaiman calls "Information-Salience Voice Systems", as well as what some conlangers call "Trigger Systems").

It sounds like you're saying that there should be a morpheme indicating that there is a subject and one indicating that there is an object, yet not specifying further.

Definitely not what I was saying.
I was saying that that's all that could be accomplished by any system for marking the case of subjects and/or objects on the verb: and that it wouldn't be worth it, so any language that had such a system would soon lose it.

Swahili has a set of subject prefixes for verbs and a set of object prefixes, and that's kind of what I was getting at here, except we made them obligatory even when the nouns are present.

But Swahili marks the gender and number and person of the subject and object on the verb (IIRC in the order subject prefix -- tense prefix -- object prefix -- verb root). It does not mark the case of the subject nor the case of the object on the verb. I don't know of any natlang that does, and if there is one, it must be a diachronically transitory oddity.

A system like Swahili's would be cool for your conlang.

We haven't gotten to voice yet, so all of this will be considered. But I still don't see how it conflicts with the marking on the prefixes. It is true that all they do is mark the syntactic role, not the semantic one, but how is this a problem? The semantic roles will likely be shown by voice.

Then why bother marking the case on the verb?
The case of the subject will always be the same, namely the equivalent of Nominative or Absolutive or whatever in the 'lang in question. And if it has an "object" grammatical relation, the case of the object will always be the equivalent of Accusative or Ergative or whatever.
Mark the definiteness, gender, number, person, specificity, topichood or focushood, of the subject (and the object if it has that GR, and the secondary or indirect object if it has that GR) on the verb; but not the case.

I thought it was common for the subject to be the agent and the primary object to be the patient. That's what this language is doing.

That certainly doesn't mesh well with ergativity.
Maybe your conlang is mostly Accusative/Nominative with a high degree of partial ergativity?
Look up http://www.kneequickie.com/kq/Grammatical_relations.
Especially see http://www.kneequickie.com/kq/Grammatic ... Absolutive; and then also http://www.kneequickie.com/kq/Grammatic ... ns#Subject.
The term "subject", if a language has one, is almost synonymous with the term "syntactic pivot", if the language has one. One feature that most ergative languages have is that the absolutive participant -- which usually means the patient -- is the syntactic pivot. It's reasonable to say that in most unmarked transitive clauses in most ergative languages the patient is usually the subject; it's certainly usually the absolutive.

Not yet decided, but it is likely that there will be ditransitive verbs at least, and thus another grammatical relation. Are there trivalent verbs that are not ditransitive?

Such verbs can arise by causativization or applicativization (e.g. benefactive applicativization) of bivalent monotransitive verbs; and if there are any other valency-raising operations that apply to bivalent verbs, probably most of these other operations as well.
I can't think of a root verb, already trivalent as a root, that isn't ditransitive. Maybe there is one; but I never heard of one.

I said in an earlier post that it's ergative/absolutive.

Sorry ; I must have missed it.

I'll be sure to bring up what you have said in the next class! There are still many basic things that we haven't yet covered, but since we've focused a lot on verbs so far, this all seems like it'll be a big help.

I'm glad it may have helped.

You might be interested in this. Especially consider looking at M.H. Klaiman's book.

Also, go to http://listserv.brown.edu/archives/cgi-bin/wa and search for all posts containing the string "grammatical voice" since 2001.