zompist bboard

Warning: this post may be a bit hard to digest, especially if you aren't familiar with alignment. It's also quite late, so I apologise on beforehand for any errors or weird sentence constructions.

Morphosyntactic alignment is about how the different core arguments of a verb are grouped into cases (wether these are explicitly marked or not is irrelevant. Well... most of the time, anyways*). Let's take a look at the different kinds of core arguments:

S: Subject (sometimes Experiencer) of an intransitive sentence
A: Agent of a transitive sentence
P: Patient of a monotransitive sentence
R: Receiver of a ditransitive sentence
T: Theme of a ditransitive sentence

Intransitive: He (S) sleeps
Monotransitive: He (A) hits her (P)
Ditransitive: He (A) gives her (R) a book (T)

Let's look at the grouping of S, A, and P first.

• Accusative languages group S and A into the nominative. P becomes the accusative. f.e. English
• Ergative languages group S and P into the absolutive. A becomes the ergative. f.e. Basque
• Tripartite languages assign a different case to each: S is the intransitive, A the ergative and P the accusative. f.e. Wangkumara
• Active languages (sometimes called split-ergative) group S with P or with A, depending on an external factor. This can be wether S is in control or not, wether the verb represents a state or an action, ... or it can be completely arbitrary. The two resulting cases are commonly called agentive and patientive, but also ergative and accusative respectively. f.e. Georgian, sometimes
• Some other languages exhibit even weirder alignments, like an animacy hierarchy (aka direct/inverse system). I'm not sure about this, but I think I even read about a language which groups A with P and marks S differently. Which is pretty damn crazy.

Now let's look at the grouping of P, R and T. This part of alignment usually receives less attention... a pity, really.

• Dative languages (or direct/indirect-object-languages) group P with T. R becomes the dative. f.e. English
• Dechticaetiative languages (or primary/secondary-object-languages) group P with R. T becomes the dechticaetiative. f.e. Yoruba
• There are also languages that group P, R and T in a single case, and rely completely on context in ditransitive sentences to make out which is which.

Some examples:
English is accusative-dative.


Basque is ergative-dative.


Yoruba is accusative-dechticaetiative.






My own language, Lembrin is (semantically and syntactically) active-dechticaetiative.

It makes a control-based distinction in intransitive sentences between agentive and patientive. It is also Fluid-S which means that you aren't forced to use one of them, you can use both and get subtle differences in meaning:
cura lis
run-VRB I-AGT
I run (and I am in control)

li cura
I-PAT run-VRB
I run (and I am not in control)
One could say this when being chased, or something.

The opposite is Split-S which means that you have to use the agentive with some verbs, and the patientive with others.

Lembrin is dechticaetiative:
oi dauma lis stymo
you-PAT give-VRB I-AGT chair-DCH
I give you the chair

On top of that, Lembrin has a few verbs which are semantically monotransitive, but syntactically behave like ditransitive sentences with either the agent or the patient missing. An example is cør, to agree with. The subject is in the agentive or the patientive, depending on control, whereas the object is in the dechticaetiative:
cøra lis réasso
agree-VRB I-AGT document--DCH
I agree with the document (and I am in control)

li cøra réasso
I-PAT agree-VRB document--DCH
I agree with the document (and I am not in control)
One could say this when one is forced to sign it, for example.





So, what's your language's morphosyntactic alignment?** You are encouraged to draw a diagram and provide examples





*: Some languages show different alignments morphologically and syntactically.
**: I realise that this question is perhaps a bit oversimplified: some languages show different alignments depending on other factors like verb tense, 1st, 2nd or 3rd person, ... in that case, feel free to elaborate

Well I'm not going to argue with you!

(serriously, I'll have a look over it in the moning)

Interesting:

Well, Berag groups A and P together, with S marked differently. Actually this has evolved from an archaic finite-infinite distinction- nouns aren't marked.

As for the "secondary" alignment... That's harder to get my head around I think Berag would actually mark them differently... In the case of "He gives her a book" then "her" would be benefactive and "book" would be instrumental.

I've confused myself now....

Sander wrote:[*]Dechticaetiative languages (or primary/secondary-object-languages) group P with R. T becomes the dechticaetiative. f.e. Yoruba
[*]There are also languages that group P, R and T in a single case, and rely completely on context in ditransitive sentences to make out which is which.[/list]

I'm not very familiar with RT groupings, but surely it's (theoretically) possible for a language to group P, R, and T all separately? Tripartite style?

uh.. What do the colors in the diagram mean? I think im on the cusp of understanding this...

The red is for the S A P grouping, and the black is for the P T R group. (I think.)

In which case, my language is ergative-dative. The Subject, Patient, and Theme are all in the absolutive case, with the Agent marked in the ergative case. The Receiver is in the ablative case and requires a postposition (which could be one of a few).
(This is assuming I understand this concept.)

So, as examples, we have (lazily glossed):
Intransitive:
Ge's tiktik-'kap.
Building-ABS fall.apart-has
The building has fallen apart.

Monotransitive:
Thei 'sirwo-iong 'ke.
he-ERG kill-is tree-ABS
He's killing the tree.

Ditransitive:
Sai pam-'dyc pem-sme ngom.
man-ERG give-is father-ABL-BENEFICIARY water-ABS
The man's giving his father water.

Worth noting, though, is that in a ditransitive sentence, the theme is placed at the end of the sentence, whereas the recipient directly follows the verb, where the patient would be in a transitive sentence.


Remember, though, you can't spell dechticaetiative without dative.

woah, i completely didn't consider alightment when i made Extmws, maybe i can sneak it in. as far as i've considered, most Extmws verbs change from transitive/intransitive on the usage. i think it might be tripartite.

with an A, R, T sentence it looks like this:
He carried the food to her.
È-Ba-vy wè-syz atlac-yab Ba.

subject recieves no mark, but if it is an agent it can recieves the particle -twx (by way of/because of), but the theme recieves particles like -syz used when the theme is able to be moved by hands, i.e. "smaller than a breadbox", -vex used with abstracts or intangibles, -çyç used when the theme is tangible, but can't be moved by hands like liquids), -sewz is for tangibles that are movable by heavy machinery or magick i.e. "bigger than a breadbox", and finally -gáàn for tangibles that are completely unmovable. this disction is all a part of the implicit gender of an object, and don't just go with verbs that express movement. the verb jañenec means "to desire/to want" but uses receptors just like the verb atlac "to carry" a different receptor, though, would be used if what was desire was food (-syz) or a house (-sewz). the reciever gets the usual prepositional and preposition-like particles, such as -vy, -fy, -jy, -ly, and -tby.

there's a lot of grammar rules i haven't explored yet and a couple of rules that need adjusting.

Sander wrote:I'm not sure about this, but I think I even read about a language which groups A with P and marks S differently. Which is pretty damn crazy.

See my thread with identical title in L&L

Im not sure what my conlang would be classified as...

There are markers indicating what is agent and patient in the verb. In the nouns themself, they show wheather they are the subject or object or whatever.

My current project is both active-stative (with split triggered by some voices and aspects) and direct-inverse - as for the second thingie, it is dative.

Lauzheu is accusative and dechticaetiative, for the most part. The situation is muddled by the fact that nouns are not declined at all; what you may have as case I applied to word order, which in my language is very strict. Some examples will help.

Most verbs that can take three arguments (such as lau "say", toi "teach", and fu "give") follow the dechticaetiative pattern:

-- So s-fu-s sha u vi.
me IMP.give.ACT child DAT money
I give money to the child.
-- So s-lau-s sha u t vi n-mo-n.
me IMP.say.ACT child DAT that money PERF.exist.PAS
I say to the child that there is money.

(Don't mind the abbreviations; I use "DAT" for because I couldn't think of anything else to translate the particle "u".)

Also, though I say the language is accusative, it borders the line on ergativity, in that what are normally considered "intransitive verbs" are actually transitive verbs in the passive voice, so the "subject" is technically the patient of a verb with an unspoken agent. Example:

-- So s-goi-s siu.
me IMP.yellow.ACT lizard
I yellow-ify the lizard.
-- Siu s-goi-n so.
lizard IMP.yellow.PASS me
The lizard is yellow-ified by me.
-- Siu n-goi-n so.
lizard PERF.yellow.PASS me
The lizard has been yellow-ified by me.
-- Siu n-goi-n.
lizard PERF.yellow.PASS
The lizard has been yellow-ified. = The lizard is yellow.

A select few verbs (mostly motion verbs such as "move") can have up to seven arguments, although more than four are not at all common.
-- So s-beu-s siu u keDoisk u keYespank.
me IMP.move.ACT lizard DAT germany DAT spain.
I move the lizard from Spain to Germany. [i.e. You're traveling Europe with your lizard]
-- So s-beu-s siu u keDoisk u keYespank u keFalansk u kePolutuk u kePolantk.
me IMP.move.ACT lizard DAT germany DAT spain DAT france DAT portugal DAT poland.
I move the lizard from Spain to Germany by way of France towards (but not reaching) Portugal and away from (but not starting in) Poland. [A rather awkward sentence.]
-- So s-beu-s siu u u u u u kePolantk.
me IMP.move.ACT lizard DAT DAT DAT DAT DAT poland.
I move the lizard away from (but not starting in) Poland. [A natural, if slightly comical-sounding sentence.]

Réĝledh is an active (mostly split-S, based on control), dative language.

For natlangs--because I can--Ojibwe is a direct/inverse, dechticaetiative (how on earth is that pronounced?) language.

weldingfish wrote:See my thread with identical title in L&L

Now that I'm thinking about it, I believe Payne (in DMS) says there's no known language that uses such a system--it would certainly be a very surprising thing for a natural language to have. Maybe such a language has been described since DMS was first published, though.

Janaharian: Accusative, dative
Proto-Qalic (ancestor of Janaharian, as well as several less well developed langs): Split ergative, dative
Ndasti: A mix of accusative and semifluid-S (don't ask), dative
Qamet Oryä: Direct/inverse (although, rather like a fluid-S lang, it can express a "control" distinction with verbs), dechticaetiative

I think Faranit is accusitive dative, but it depends on the specific situation (both the pro/noun and the verb).

Whimemsz wrote:dechticaetiative (how on earth is that pronounced?)

/dEk.t@."keI.S@.tIv/, I think.

Rory wrote:I'm not very familiar with RT groupings, but surely it's (theoretically) possible for a language to group P, R, and T all separately? Tripartite style?

Sure I didn't include it because I haven't come across a natlang that does it.

The situation is muddled by the fact that nouns are not declined at all; what you may have as case I applied to word order, which in my language is very strict.

As I said, rigid word order is case marking just as much as morphological changes are. Lembrin indicates the dechticaetiative with -i, -o or -u, depending on noun class. The agentive/patientive distinction is only marked on (some) pronouns and particles. The agentive -s has been lost on nouns because the patientive always precedes and the agentive always follows the verb.

Just a quick note to say thank you for this. I've been reading up these issues this week on Wikipedia and the info is all over the place, whereas yours is clear and concise!

To be honest I didn't know that dechticaetiative languages existed before this thread, thank you for mentioning it! I had never thought about there being other ways to do it than dative before.
I wonder though, why did they decide to call it dechticaetiative? Why? Why not something easier, like un-dative or something?

---

My language (unnamed so far) is active fluid-S and dative. It marks the subject of both transitive and intransitive clauses as either agentive or patientive depending on volition. So:

I-ag fall
"I fall (voluntarily, not by accident)"

I-pt fall
"I fall (by accident, I might have tripped on a stone or something)"

I-ag see you-pt
"I see you (voluntarily)"

I-pt see you-pt
"I see you (involuntarily)"

The thing is that the fluid-S-ness does not only concern subjects, but extends to include also direct objects. So you get:

I-ag see you-ag
"We see each other, I see you and you see me" (both are agents, but the focus is on the subject)

I-pt see you-ag
"You make me see you" (the subject is the patient, and the object is the agent here)

But when it comes to indirect objects there's nothing too exciting, it basically works like English.

Everything you ever wanted to know about Gevey verb transitivity and valency. If anyone would care to explain to me how this should be described in proper linguistic terms I'd be eternally grateful!

Sander wrote:

Rory wrote:I'm not very familiar with RT groupings, but surely it's (theoretically) possible for a language to group P, R, and T all separately? Tripartite style?

Sure I didn't include it because I haven't come across a natlang that does it.

In my experience, if something is theoretically possible, then there's probably a language somewhere that does it. Let's look in the Caucaus, and Papua, we're bound to find it in one of those places

I would add one more category to this:

O: Object of a transitive sentence; used in an active language when the Agent does not fulfill the active condition(s). (In my own conlangs, when the Agent doesn't fulfill the active conditions, he's marked as a Subject.)

vohpenonomae wrote:I would add one more category to this:

O: Object of a transitive sentence; used in an active language when the Agent does not fulfill the active condition(s). (In my own conlangs, when the Agent doesn't fulfill the active conditions, he's marked as a Subject.)

Thus that allows you to have the active-stative split even on transitive verbs ? Neat !

Legion wrote:

vohpenonomae wrote:I would add one more category to this:

O: Object of a transitive sentence; used in an active language when the Agent does not fulfill the active condition(s). (In my own conlangs, when the Agent doesn't fulfill the active conditions, he's marked as a Subject.)

Thus that allows you to have the active-stative split even on transitive verbs ? Neat !

Yep. I'm not sure there's a natlang that does it this way, but the idea was inspired by Mohawk, which is active. I've been calling sentences with a subject and object "Subject-Object Intransitive," since I've been defining "transitive" sentences as those that fulfill the active conditions. But I may rewrite that part of the Noyahtowa grammar to be in line with this system of terminology, since it'll then be easier for others to understand. I've been doing this "morphosyntactic alignment" stuff for years, but I didn't know the exact name for it.

It's not that much unlike how the average nom-acc erg-abs split lang works after all : You have three cases, but only two at a time are used, according to the situation (these language typically have a common nominative/absolutive case but distinct ergative and accusative cases).


Edit : even really close - your language :


1 - NOM verb ABS
2 - ABS verb OBJ

Average split language :

1 - ERG verb NOM/ABS
2 - NOM/ABS verb ACC

Hallo!

Old Albic is active-dative. Another conlang I am working on (working title "Razaric") will be ergative-dechticaetiative.