Sérhes Kéttw verb questions

Museum for the best conlanging and conworldery threads. Ask mods to move threads here.
Post Reply
User avatar
Nortaneous
Sumerul
Sumerul
Posts: 4544
Joined: Mon Apr 13, 2009 1:52 am
Location: the Imperial Corridor

Sérhes Kéttw verb questions

Post by Nortaneous »

(I originally posted this in C&CQ, but right after posting, I realized it's not exactly quick anymore. Originally it was just a question about what sort of alignment this is, but I've been hitting enough walls lately with my horribly limited knowledge of syntax that I ended up having a lot more questions than that.)

I attempted Austronesian alignment in Sérhes Kéttw, but I'm not entirely sure what it is that I ended up with.

Topic marking is obligatory. The verb marks for A/P status of the topic: (ignore the weird phonological processes, metathesis, etc. in the examples)

Nián káibat rákyannenn.
né=an káibat rákyanne-t-n
1S=TOP book read-ACT.PFV-1S
I am reading a book. (although normally the pronoun and topic marker are dropped: Káibat rákyannenn.)

Káibann rákyanneran.
káibat=an rákyanne-ra-n
book=TOP read-PASS.PFV-1S
The book is being read by me.

And perfective/imperfective:

Nián káibatre rákyannengn.
né=an káibat-re rákyanne-ng-n
1S=TOP book-COLL read-ACT.IPFV-1S
I read books.

Káibatrian rákyannercan.
káibat-re=an rákyanne-cera-n
book-COLL=TOP read-PASS.IPFV-1S
I read books. (not sure how to pull this off in English)

And there's a separate set of endings in that slot for imperatives:

Céyan káibatre rákyanneyt!
céy=an káibat-re rákyanne-y-t
2S=TOP book-COLL read-ACT.IMP-2S
Read books!

Káibatrian rákyannesat!
káibat-re=an rákyanne-sa-t
book-COLL=TOP read-PASS.IMP-2S
Read books! (again, not sure how to pull this off in English; I'd translate it as "Read the books!", but that'd take the plural, not the collective)

I'm not sure how I'd handle trivalent verbs, but I'd guess it'd look something like:

Nián káibat láncakcang ḩéungenn.
né=an káibat láncak-cang ḩéung-t-n
1S=TOP book child-DAT give-ACT.PFV-1S
I am giving a book to a child.

Káibann láncakcang ḩéungran.
káibat=an láncak-cang ḩéung-ra-n
book=TOP child-DAT give-PASS.PFV-1S
I am giving a book to a child.

Láncakcangan káibat ḩéungran.
láncak-cang=an káibat ḩéung-ra-n
child-DAT=TOP book give-PASS.PFV-1S
I am giving a book to a child.

The verb template ordering so far is tense mood stem act/pass neg person:

Nián káibat eccarrákyannennan.
né=an káibat et-cal-rákyanne-t-nga-n
1S=TOP book FUT-POT-read-ACT.PFV-NEG-1S
I will probably not read the book.

So, a few questions:
1. What kind of alignment is this? Is it even realistic?
2. How could I make trivalent verbs more interesting? Would it be realistic to have another set of affixes for when the dative (or some other argument besides the agent or patient (?)) is the topic?
3. How realistic would it be to mark the object on the verb also, so even more could be left implicit? Where could that affix go?
4. Is there a better way to write the template ordering?
5. What sort of mood system should I have? (I just added the potential mood because I realized I didn't have any moods yet and I needed one for the example.) The Wikipedia article doesn't list that many moods. I might just steal everything from Finnish, but I'm trying to avoid relying too heavily on European langs.
6. What are some ways that this could handle interrogatives, especially yes-no questions? (I have a three-form system, if it matters.)
7. Are there any stupid mistakes or horribly unrealistic things in this that I've missed so far?
Siöö jandeng raiglin zåbei tandiüłåd;
nää džunnfin kukuch vklaivei sivei tåd.
Chei. Chei. Chei. Chei. Chei. Chei. Chei.

Yng
Avisaru
Avisaru
Posts: 880
Joined: Fri Jul 03, 2009 3:17 pm
Location: Llundain

Re: Sérhes Kéttw verb questions

Post by Yng »

Verb systems like these seem to be springing up all over the place recently. I tried something similar, although mine marks the topic with a null affix and any argument, including an adverbial, will be marked on the verb when topical. Apparently it's not attested in any natlang, but Tagalog has two applicative voices whose morphemes take the slot (IIRC) of the patient/agent focus morphemes, so it's really just a logical extension of that.
كان يا ما كان / يا صمت العشية / قمري هاجر في الصبح بعيدا / في العيون العسلية

tà yi póbo tsùtsùr ciivà dè!

short texts in Cuhbi

Risha Cuhbi grammar

User avatar
Nortaneous
Sumerul
Sumerul
Posts: 4544
Joined: Mon Apr 13, 2009 1:52 am
Location: the Imperial Corridor

Re: Sérhes Kéttw verb questions

Post by Nortaneous »

Yeah, it seems like a minor variation on Austronesian alignment, if I understand it correctly: nothing is marked but the direct, intransitive clauses use the agent marker, and word order is reversed.
Siöö jandeng raiglin zåbei tandiüłåd;
nää džunnfin kukuch vklaivei sivei tåd.
Chei. Chei. Chei. Chei. Chei. Chei. Chei.

User avatar
vec
Avisaru
Avisaru
Posts: 639
Joined: Tue Sep 16, 2003 10:42 am
Location: Reykjavík, Iceland
Contact:

Re: Sérhes Kéttw verb questions

Post by vec »

1. What kind of alignment is this? Is it even realistic?
I think so. I don't think it looks that weird. However, there is no overt marker on the NPs besides the topic marker, so you could interpret it differently. But notice that you are marking what seems to be a nominative subject on the verb consistently in both active and passive forms. Which makes me see this as being nom-acc. Am I missing something?
2. How could I make trivalent verbs more interesting? Would it be realistic to have another set of affixes for when the dative (or some other argument besides the agent or patient (?)) is the topic?
Well, there are several things you can do. You are already marking subject and object pretty overtly on the verb. The dative could be left unmarked, and there might be word-order related rules that apply to it. But I find the way you're handling things now kind of interesting. It is normal for the dative to be more marked, since it is more marked in abstract terms (ie. less default than nom. and acc.).
3. How realistic would it be to mark the object on the verb also, so even more could be left implicit? Where could that affix go?
It is very realistic and common. I don't know anything about austronesian alignment theories, I never understand the articles, it always seems like everyone's trying to make them out to be much more complicated than they are. So if they're what you're aiming for you have to ask for something else. But you can definitely mark the object.

See, one thing that's a bit odd that's happening here is that when you make things what you call "passive", you only delete the NP for the subject, but it seems judging by the verb that it is still the same. Passivisation usually involves bumping up the object to subject and either deleting the active subject or relegating it to an adjunct phrase, often a PP (preposition/postposition phrase). It's a form of valency reduction. Your alleged passive simply marks that the subject is not being stated in a separate NP but it is still marked on the verb. So maybe the verb isn't marking the subject anymore but the non-overt adjunct. But then, what happens if it is not 1s? What happens when it is unknown (which is often a reason for making things passive)?
4. Is there a better way to write the template ordering?
I think you did fine. You could also do a numbered list. I would call your active/passive distinction "voice" and negation "polarity".

Code: Select all

5. What sort of mood system should I have? (I just added the potential mood because I realized I didn't have any moods yet and I needed one for the example.) The Wikipedia article doesn't list that many moods. I might just steal everything from Finnish, but I'm trying to avoid relying too heavily on European langs.
Well, one of the most important modality distinction is that of possibility and necessity. Each can be either epistemic or situational. Examples (taken from the WALS article on the subject):
  1. You may go home now. [situational possibilty]
  2. You must go home now. [situational necessity]
Situational possibility is basically permission. Situational necessity is... I don't know, an obligation. Situational is sometimes called root modality.
The following are epistemic:
  1. Bob may be mistaken about the cause of the accident.
  2. Terry must be from Northumberland.
Epistemic modality is something that the speaker has deduced from the available information. As you can see, English conflates these and there is a lot of overlap. Some languages have overlap, some have clear distinctions. Some have overlap for one type of modality but not the other. A language could also have a situational and an epistemic mood which might delve into some flirtation with evidentiality.
Another important category is realis/irrealis. Whether something is real or imagined. This can be conflated with the others or made separate or not indicated at all grammatically.
Your potential case could incorporate irrealis meanings, or epistemic meanings, or possibility meanings. In any combination or none.
Modality has to do with "how" something happens or doesn't happen. Is it real? Is it possible? Is it necessary? It can also answer questions like: Is it important? Is it desirable? Is it desirable but impossible? Etc. There are infinite possibilities and I would never simply fall on copying one system or another. Make your own. And personally, I find these kinds of things most interesting when they are not entirely symmetrical, when they have gaps, irregularities, fuzzy borders and other weirdness. Play around with it. Read the WALS article and check out this list of modal adverbs to get an idea of what modality is all about.
6. What are some ways that this could handle interrogatives, especially yes-no questions? (I have a three-form system, if it matters.)
I actually don't know what a three-form system is. And this could handle them any way you want. You could mark it with polarity. You could mark it after or before it. You could not mark it at all and use intonation or word order changes. Or you could topicalise the verb. Or you could add an adverb. Or an auxiliary verb. Or a separate clause controlled by some interrogative verb. Or you could change the topic marker and make having that topic marker in interrogative mandatory? Or you could mix and match some of these. Or maybe you can come up with something even more exciting or even more boring if you want.
7. Are there any stupid mistakes or horribly unrealistic things in this that I've missed so far?
No, not that I see, but I do think that you do have to consider what's happening with your voice system a little bit. I kind of like what's going on but I'm not sure you can get away with calling your subject-dropping valency reduction passive. I'd ask TomHChappell.

Hope some of that helps or inspires.
vec

User avatar
Nortaneous
Sumerul
Sumerul
Posts: 4544
Joined: Mon Apr 13, 2009 1:52 am
Location: the Imperial Corridor

Re: Sérhes Kéttw verb questions

Post by Nortaneous »

vecfaranti wrote:But notice that you are marking what seems to be a nominative subject on the verb consistently in both active and passive forms. Which makes me see this as being nom-acc. Am I missing something?
Heh, how did I miss that? That might be a better way of thinking about it, except"
See, one thing that's a bit odd that's happening here is that when you make things what you call "passive", you only delete the NP for the subject, but it seems judging by the verb that it is still the same. Passivisation usually involves bumping up the object to subject and either deleting the active subject or relegating it to an adjunct phrase, often a PP (preposition/postposition phrase). It's a form of valency reduction. Your alleged passive simply marks that the subject is not being stated in a separate NP but it is still marked on the verb. So maybe the verb isn't marking the subject anymore but the non-overt adjunct. But then, what happens if it is not 1s? What happens when it is unknown (which is often a reason for making things passive)?
Right, it's not a standard passive. After reading the Wikipedia article on Austronesian alignment again, I think I should probably call the voices agent and patient trigger instead of active and passive. My guess is that, if the subject is unknown, the subject slot would default to 3S, which is conveniently -∅. (Or maybe I should merge 3S with 3P (currently -m) and make them both -∅. That'd probably make that work better, I think.) So:

Káibann rákyannera.
káibat=an rákyanne-ra-∅
book=TOP read-PTRG.PFV-3S
The book is being read.
I actually don't know what a three-form system is.
There are two words for "yes": one answering a positive question and one answering a negative question. Stolen from German because I'm lazy.
Or you could change the topic marker and make having that topic marker in interrogative mandatory?
I like this idea. I'll probably do that, maybe along with an interrogative slot somewhere in the verb template. (probably I could even use this to get away with not even having a class of interrogatives -- instead of having, for example, "when", I could just have time=TOP.Q)
I kind of like what's going on but I'm not sure you can get away with calling your subject-dropping valency reduction passive.
But it's not even valency reduction, unless I'm missing something. It just casts the topic in the patient role.
Siöö jandeng raiglin zåbei tandiüłåd;
nää džunnfin kukuch vklaivei sivei tåd.
Chei. Chei. Chei. Chei. Chei. Chei. Chei.

User avatar
vec
Avisaru
Avisaru
Posts: 639
Joined: Tue Sep 16, 2003 10:42 am
Location: Reykjavík, Iceland
Contact:

Re: Sérhes Kéttw verb questions

Post by vec »

Nortaneous wrote:
I actually don't know what a three-form system is.
There are two words for "yes": one answering a positive question and one answering a negative question. Stolen from German because I'm lazy.
Ah, like Icelandic too. Já, jú, nei.
Or you could change the topic marker and make having that topic marker in interrogative mandatory?
I like this idea. I'll probably do that, maybe along with an interrogative slot somewhere in the verb template. (probably I could even use this to get away with not even having a class of interrogatives -- instead of having, for example, "when", I could just have time=TOP.Q)
Cool idea. You're getting somewhere
I kind of like what's going on but I'm not sure you can get away with calling your subject-dropping valency reduction passive.
But it's not even valency reduction, unless I'm missing something. It just casts the topic in the patient role.
Yes, you're right. I think we figured out what's going on with your system but I still think you should try to get Tom to discuss this. He's an encyclopædia on this sort of stuff.
vec

User avatar
vec
Avisaru
Avisaru
Posts: 639
Joined: Tue Sep 16, 2003 10:42 am
Location: Reykjavík, Iceland
Contact:

Re: Sérhes Kéttw verb questions

Post by vec »

On second thought, based on the examples you give, it does mess with valency, at least in some way, because it seems to forbid the subject or the more-agent-like constituent from appearing in the sentence, unless that's only what's going on in your examples. {wow, that's one long sentence]
vec

User avatar
Nortaneous
Sumerul
Sumerul
Posts: 4544
Joined: Mon Apr 13, 2009 1:52 am
Location: the Imperial Corridor

Re: Sérhes Kéttw verb questions

Post by Nortaneous »

vecfaranti wrote:On second thought, based on the examples you give, it does mess with valency, at least in some way, because it seems to forbid the subject or the more-agent-like constituent from appearing in the sentence, unless that's only what's going on in your examples. {wow, that's one long sentence]
Sample error. That's just pronoun dropping.

Láncakan káibat rákyannet.
láncak=an káibat rákyanne-t-∅
child=TOP book read-ATRG.PFV-3S
The child is reading a book.

Káibann láncak rákyannet.
káibat=an láncak rákyanne-t-∅
book=TOP child read-ATRG.PFV-3S
The book is being read by a child.
Siöö jandeng raiglin zåbei tandiüłåd;
nää džunnfin kukuch vklaivei sivei tåd.
Chei. Chei. Chei. Chei. Chei. Chei. Chei.

User avatar
vec
Avisaru
Avisaru
Posts: 639
Joined: Tue Sep 16, 2003 10:42 am
Location: Reykjavík, Iceland
Contact:

Re: Sérhes Kéttw verb questions

Post by vec »

So in essence:
A verb with an agent trigger marks that whatever's marked with the topic marker is the agent.
A verb with a patient trigger marks that whatever's marked with the topic marker is the patient.
Or we could say:
The trigger assigns kind-of-constituent marking to the topic marker.
vec

User avatar
Nortaneous
Sumerul
Sumerul
Posts: 4544
Joined: Mon Apr 13, 2009 1:52 am
Location: the Imperial Corridor

Re: Sérhes Kéttw verb questions

Post by Nortaneous »

Sounds right to me. Isn't that what Austronesian alignment does? It seems to me that, at least as described here:

A verb with an agent trigger marks that whatever's marked with the direct case marker is the agent.
A verb with a patient trigger marks that whatever's marked with the direct case marker is the patient.

Also, it seems like the direct case marker is used to mark the topic (well, focus), so it seems like what I have here, unless I'm missing something, is Austronesian alignment with an unmarked indirect. I don't have the case-marking bit that Tagalog has, but that seems awesome enough that I might add it.

edit: also, questions

Kátwia saiáru?
kátw=ia sai-áru
time=TOP.Q Q-COP
What time is it? / When is it?

Káibatwia asairákyannett?
káibat-w=ia a-sai-rákyanne-t-t
book-PL=TOP.Q PAST-Q-read-ATRG.PFV-2S
Which books did you read?

Láncakcangan ḩasainákkat?
láncak-cang=an ḩa-sai-nák~k-t
child-DAT=TOP PAST-Q-death~CAUS-2S
Did you kill the child?
Siöö jandeng raiglin zåbei tandiüłåd;
nää džunnfin kukuch vklaivei sivei tåd.
Chei. Chei. Chei. Chei. Chei. Chei. Chei.

User avatar
vec
Avisaru
Avisaru
Posts: 639
Joined: Tue Sep 16, 2003 10:42 am
Location: Reykjavík, Iceland
Contact:

Re: Sérhes Kéttw verb questions

Post by vec »

Why is the murdered child dative? Is that just a peculiarity of that verb? Or does it have to do with the causative?
vec

User avatar
Nortaneous
Sumerul
Sumerul
Posts: 4544
Joined: Mon Apr 13, 2009 1:52 am
Location: the Imperial Corridor

Re: Sérhes Kéttw verb questions

Post by Nortaneous »

Causative. Although I think I'm going to drop that.
Siöö jandeng raiglin zåbei tandiüłåd;
nää džunnfin kukuch vklaivei sivei tåd.
Chei. Chei. Chei. Chei. Chei. Chei. Chei.

TomHChappell
Avisaru
Avisaru
Posts: 807
Joined: Wed Dec 28, 2005 2:58 pm

Re: Sérhes Kéttw verb questions

Post by TomHChappell »

It looks like you could use what M.H.Klaiman classified as "an Information Saliency type of grammatical-voice system". The topic noun would be marked to show it was the topic, not to show what semantic role nor what grammatical relation it occupied; the verb would have a voice-marking to show what grammatical relation and/or what semantic role the topic occupies.
So intransitive, monovalent clauses would just have the Subject case-marked as Topic.
In bivalent, monotransitive clauses,
  • if the Agent is the Topic, it will be case-marked as Topic, the Patient will be case-marked as Patient (Accusative), and the verb will be voice-marked as "Topic is Agent".
  • if the Patient is the Topic, it will be case-marked as Topic, the Agent will be case-marked as Agent (Ergative), and the verb will be voice-marked as "Topic is Patient".
In trivalent, ditransitive clauses,
  • if the Donor is the Topic, it will be case-marked as Topic, the Theme will be case-marked as Patient (Accusative), the Recipient will be case-marked Dative, and the verb will be voice-marked as "Topic is Agent".
  • if the Theme is the Topic, it will be case-marked as Topic, the Donor will be case-marked as Agent (Ergative), the Recipient will be case-marked Dative, and the verb will be voice-marked as "Topic is Patient".
  • if the Recipient is the Topic, it will be case-marked as Topic, the Donor will be case-marked as Agent (Ergative), the Theme will be case-marked as Patient (Accusative), and the verb will be voice-marked as "Topic is Recipient".
So, you see, if the Recipient is the Topic, there needn't be a participant noun that's case-marked "Dative".

Tagalog, AIUI, has an information-saliency type voice-system, but marks the focus (rather than the topic) with the definite-article-cum-nominative-preposition "ang", and then voice-marks the verb to show which semantic role the "ang"-marked noun plays.

Japanese, AIUI, also has an information-saliency type voice-system. It marks the topic with the definite-article-cum-nominative-preposition "wo", and then voice-marks the verb to show which semantic role the "wo"-marked noun plays. Or is that wrong?

I would think it would be crucial in any language in which the topic of every clause has to be explicitly marked as topic, leaving no room for any other case-marking on that topic noun, that the verb be voice-marked to show what grammatical relation or semantic role the topic-noun plays in the clause. Similarly, change "topic" to "focus" throughout that last sentence, and I think you still get a true statement.

Anyway, some languages (including, IIRC, some South or Central American languages) whose system of grammatical voice falls into Klaiman's "Information Saliency" type or class, actually mark both the topic and the focus; the verb's voice-inflection has to tell the "case" of both the topic and the focus.

Post Reply