FreeWordOrder SerialVerb Concordial VerbClass

A forum for discussing linguistics or just languages in general.
Post Reply
User avatar
eldin raigmore
korean
korean
Posts: 6389
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

FreeWordOrder SerialVerb Concordial VerbClass

Post by eldin raigmore » 25 Jul 2019 22:38

@Lao Kou, you may remember something relevant on CONLANG-L from Sep and Oct 2007.

Does anyone know of any natlang(s) that have or had both free word-order and serial verb constructions?
If so, what were these FWO SVC languages?
And what strategies do or did they use to say which participant-nouns go with which verbs?

Or do all real-life SVC natlangs use word-order a lot?

—————

Relatedly;

Does anyone know of any natural language that has concordial verb-classes?
If so, what are they and where can someone find out how they work?

—————

ObConLang:

There seem to be three or four conlangs, like Chris Bates’s ngwaalq’ and Lao Kou’s Gearthnuns and Jesse Bangs’s Yivrian and something by John Vertical, with concordial verb-classes.
Are there more?

And, which and whose conlangs are FWO SVC?

User avatar
Ser
cuneiform
cuneiform
Posts: 115
Joined: 30 Jun 2012 06:13
Location: Vancouver, British Columbia / Colombie Britannique, Canada

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by Ser » 26 Jul 2019 00:35

What do you mean by "concordial verb classes"? I had quick look Lao Kou's Gearthnuns thread and I couldn't identify anything that could be called that. (That thread is very long though, and it's likely it is there somewhere.)

User avatar
eldin raigmore
korean
korean
Posts: 6389
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by eldin raigmore » 26 Jul 2019 04:54

Ser wrote:
26 Jul 2019 00:35
What do you mean by "concordial verb classes"? I had quick look Lao Kou's Gearthnuns thread and I couldn't identify anything that could be called that. (That thread is very long though, and it's likely it is there somewhere.)
See https://listserv.brown.edu/cgi-bin/wa?A ... om&P=11582

Also see https://listserv.brown.edu/cgi-bin/wa?A ... om&P=26520

And https://listserv.brown.edu/cgi-bin/wa?A ... com&P=5396

And look at https://wals.info/chapter/30, https://wals.info/chapter/31, https://wals.info/chapter/32, and https://wals.info/chapter/44.

A concordial verb-class would be a lexically-inherent obligatory (ie every verb has one) feature of the verb, not necessarily marked on the verb itself, with which some other words would be required sometimes to agree, or concord.
The words which would have to be marked to agree with the verb’s class, might include its auxiliaries, and/or its adverbs, and/or pro-verbs that co-refer with it, and/or some or all of its arguments.
The verb-class might be semantically based, like ingestive verbs or venitive verbs or andative verbs.
Class might be morphologically based, like first conjugation or third conjugation or whatever.
Class might be phonologically based, like first or last consonant or vowel of the verb-root.
It might be syntactically based, like intransitive vs monotransitive vs ditransitive, or, can’t have a clause as an argument vs must have a clause as an argument vs may or may not have a clause as an argument.
Or some classes might be pretty arbitrary. Like normal verbs vs differential-object-marked “quirky-cased object” verbs vs “quirky-cased subject” verbs.

—————

Does any of that help?

User avatar
Ser
cuneiform
cuneiform
Posts: 115
Joined: 30 Jun 2012 06:13
Location: Vancouver, British Columbia / Colombie Britannique, Canada

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by Ser » 26 Jul 2019 06:48

VERB CLASSES!
BECAUSE VERBS CAN BE MASCULINE OR FEMININE TOO!

Thank you eldin, yes, that was helpful.

My first thought is that this is kind of attested in Arabic. Verbal nouns (akin to infinitives in European languages) take a gender, quite unpredictably masculine or feminine among basic verbs but more predictably among derived verbs, and when they appear in the "cognate accusative" construction they can be modified by an adjective that agrees in gender.

A couple examples taken from some websites for Arabic learners:

darasa-hu daraasatan Tawiilatan
study.PST.3SG.MASC-3SG.MASC study.VN.ACC.SG.[FEM] long.ACC.SG.FEM
'He studied it for a long time.' (adapted from here)
(VN = "verbal noun". The verbal noun of darasa 'to study', daraasatun, is feminine. The adjective Tawiilun 'long' modifies it agreeing with it, which semantically acts like an adverb: "for a long time".)

iSbir Sabran jamiilan
be.patient.IMPER.2SG.MASC be.patient.VN.ACC.SG.[MASC] beautiful.ACC.SG.MASC
'Be beautifully patient.' (taken from here)
(The verbal noun of Sabara 'to be patient', Sabrun, is masculine. The adjective jamiilun 'beautiful' modifies it agreeing with it, which semantically acts like an adverb: "beautifully".)

User avatar
eldin raigmore
korean
korean
Posts: 6389
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by eldin raigmore » 26 Jul 2019 13:18

Ser wrote:
26 Jul 2019 06:48
VERB CLASSES! BECAUSE VERBS CAN BE MASCULINE OR FEMININE TOO!
Thank you eldin, yes, that was helpful.
My first thought is that this is kind of attested in Arabic. Verbal nouns (akin to infinitives in European languages) take a gender, quite unpredictably masculine or feminine among basic verbs but more predictably among derived verbs, and when they appear in the "cognate accusative" construction they can be modified by an adjective that agrees in gender.
A couple examples taken from some websites for Arabic learners:
....
Thanks, Ser! That’s very helpful, too!

User avatar
Ser
cuneiform
cuneiform
Posts: 115
Joined: 30 Jun 2012 06:13
Location: Vancouver, British Columbia / Colombie Britannique, Canada

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by Ser » 26 Jul 2019 16:23

eldin raigmore wrote:
26 Jul 2019 13:18
Thanks, Ser! That’s very helpful, too!
Although I should add I failed to mention that a verb can have more than one verbal noun, and it is possible that one is feminine and the other one or two are masculine, say.

User avatar
eldin raigmore
korean
korean
Posts: 6389
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by eldin raigmore » 26 Jul 2019 20:43

Ser wrote:
26 Jul 2019 16:23
Although I should add I failed to mention that a verb can have more than one verbal noun, and it is possible that one is feminine and the other one or two are masculine, say.
That might, or might not, be an “oops”.
Isn’t it generally true for nouns that their gender doesn’t change as they decline?
(Well, Bantu noun-classes are paired, a singular one with a plural one; and two different singular classes can be paired with the same plural class; so they might be an exception to that last sentence.)

I think we might want concordial verb-class to be the same for any verb as it conjugates.
I also think we might want the finite versions of the verbs to have class that some of their participants must agree with.

But any correct answer to the opening post will be welcome. I just ask that repliers explicitly mention any caveats or exceptions, as you just did in the post I quoted.

Thanks!

User avatar
Ser
cuneiform
cuneiform
Posts: 115
Joined: 30 Jun 2012 06:13
Location: Vancouver, British Columbia / Colombie Britannique, Canada

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by Ser » 26 Jul 2019 22:12

eldin raigmore wrote:
26 Jul 2019 20:43
Isn’t it generally true for nouns that their gender doesn’t change as they decline?
(Well, Bantu noun-classes are paired, a singular one with a plural one; and two different singular classes can be paired with the same plural class; so they might be an exception to that last sentence.)
Yes, in Arabic the gender of nouns is (kind of) retained through inflection, but the thing is, for Arabic basic verbs, the verbal noun is essentially derived in an unpredictable way. The verbal noun "transfix" (or triconsonantal template) used in any given basic verb is unpredictable, and by extension so is the gender (masculine/feminine) of the result. There is no pattern that explains why wajada 'to find' has wujuudun (masculine) while kataba 'to write' has kitaabatun (feminine). However, for most derived verbs, the verbal noun is a predictable inflection that ends up with masculine gender.

I would like to point out that Romanian has a very large subcategory of nouns that are masculine in the singular but feminine in the plural, the so called "neuter" gender nouns, which contrast with whole-masculine and whole-feminine nouns. French and Spanish have a few similar nouns (Sp. arte 'art' is masculine in the singular but feminine in the plural*), and Italian has some nouns that may or may not change gender in the plural depending on the meaning (braccio becomes feminine in the plural when it means '(body) arm; nautical fathom', but stays masculine when it means 'candleholder arm, mechanical/robotic arm').

* For those who know some Spanish, no, this is not me getting confused by whole-feminine nouns like el águila ~ las águilas. The case of arte is different.

User avatar
qwed117
mongolian
mongolian
Posts: 4481
Joined: 20 Nov 2014 02:27

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by qwed117 » 27 Jul 2019 05:12

Ser wrote:
26 Jul 2019 22:12
eldin raigmore wrote:
26 Jul 2019 20:43
Isn’t it generally true for nouns that their gender doesn’t change as they decline?
(Well, Bantu noun-classes are paired, a singular one with a plural one; and two different singular classes can be paired with the same plural class; so they might be an exception to that last sentence.)
Yes, in Arabic the gender of nouns is (kind of) retained through inflection, but the thing is, for Arabic basic verbs, the verbal noun is essentially derived in an unpredictable way. The verbal noun "transfix" (or triconsonantal template) used in any given basic verb is unpredictable, and by extension so is the gender (masculine/feminine) of the result. There is no pattern that explains why wajada 'to find' has wujuudun (masculine) while kataba 'to write' has kitaabatun (feminine). However, for most derived verbs, the verbal noun is a predictable inflection that ends up with masculine gender.

I would like to point out that Romanian has a very large subcategory of nouns that are masculine in the singular but feminine in the plural, the so called "neuter" gender nouns, which contrast with whole-masculine and whole-feminine nouns. French and Spanish have a few similar nouns (Sp. arte 'art' is masculine in the singular but feminine in the plural*), and Italian has some nouns that may or may not change gender in the plural depending on the meaning (braccio becomes feminine in the plural when it means '(body) arm; nautical fathom', but stays masculine when it means 'candleholder arm, mechanical/robotic arm').

* For those who know some Spanish, no, this is not me getting confused by whole-feminine nouns like el águila ~ las águilas. The case of arte is different.
re: arte
arte poética and arte abstracto are both singular, yet use different genders. Romanian's 'neuter gender' comes from -um/a 2nd declension neuters in Latin, like ovum/ova > ou/ ouă, which can mean both eggs and testes according to wikipedia
Spoiler:
My minicity is Zyphrazia and Novland
What is made of man will crumble away.

User avatar
eldin raigmore
korean
korean
Posts: 6389
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by eldin raigmore » 27 Jul 2019 06:57

akam chinjir on the ZBB recommended we google the search string “nonconfigurational serial verb”. Doing so they (and I) found references to a language called Wambaya.
I haven’t read about it yet but it looks pretty relevant.

User avatar
Ser
cuneiform
cuneiform
Posts: 115
Joined: 30 Jun 2012 06:13
Location: Vancouver, British Columbia / Colombie Britannique, Canada

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by Ser » 29 Jul 2019 07:07

qwed117 wrote:
27 Jul 2019 05:12
re: arte
arte poética and arte abstracto are both singular, yet use different genders.
Oh, right. I would say there are exceptional lexicalized phrases where arte is feminine, the three main ones being arte poética and the poetic (metrical) terms arte mayor and arte menor, but generally (as in 98% of the time) singular arte is masculine while the plural is feminine. It is the only one of its kind, I'd say.

I suppose you can compare it with mar. Technically either masculine or feminine, but I estimate it's masculine 98% of the time (in both singular and plural).
Romanian's 'neuter gender' comes from -um/a 2nd declension neuters in Latin, like ovum/ova > ou/ ouă, which can mean both eggs and testes according to wikipedia
Ou/ouă is nice but exceptional. Most inherited neuter words actually just get the new *-or-ī > -uri ending (basically the ending of corpus/corpora after grabbing the plural ending -ī/-ēs > -i, an ending that also crept into the 1st declension: lūna lūnae > lună luni) or otherwise the feminine -e plural (which I think comes from nominative -ae).

corpus corpora > corp corpuri 'body'
pectus pectora > piept piepturi 'chest'
latus latera > lat laturi 'side'
faenum faena > fân fânuri 'hay'
cor corda > cord corduri 'heart'

vīsum vīsa 'seen thing' > vis visuri/vise 'dream'

peccātum peccāta > păcat păcate 'sin'
bracchium bracchia > braț brațe 'arm'
caput capita > cap capete 'head'
vās vāsa 'bowl, dish; tool' > vas vase 'bowl, dish; boat'

User avatar
Lao Kou
korean
korean
Posts: 5665
Joined: 25 Nov 2012 10:39
Location: 蘇州/苏州

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by Lao Kou » 30 Jul 2019 18:09

eldin raigmore wrote:
26 Jul 2019 04:54
Ser wrote:
26 Jul 2019 00:35
What do you mean by "concordial verb classes"? I had quick look Lao Kou's Géarthnuns thread and I couldn't identify anything that could be called that. (That thread is very long though, and it's likely it is there somewhere.)
See https://listserv.brown.edu/cgi-bin/wa?A ... om&P=11582

Also see https://listserv.brown.edu/cgi-bin/wa?A ... om&P=26520

And https://listserv.brown.edu/cgi-bin/wa?A ... com&P=5396

And look at https://wals.info/chapter/30, https://wals.info/chapter/31, https://wals.info/chapter/32, and https://wals.info/chapter/44.

A concordial verb-class would be a lexically-inherent obligatory (ie every verb has one) feature of the verb, not necessarily marked on the verb itself, with which some other words would be required sometimes to agree, or concord.
The words which would have to be marked to agree with the verb’s class, might include its auxiliaries, and/or its adverbs, and/or pro-verbs that co-refer with it, and/or some or all of its arguments.
The verb-class might be semantically based, like ingestive verbs or venitive verbs or andative verbs.
Class might be morphologically based, like first conjugation or third conjugation or whatever.
Class might be phonologically based, like first or last consonant or vowel of the verb-root.
It might be syntactically based, like intransitive vs monotransitive vs ditransitive, or, can’t have a clause as an argument vs must have a clause as an argument vs may or may not have a clause as an argument.
Or some classes might be pretty arbitrary. Like normal verbs vs differential-object-marked “quirky-cased object” verbs vs “quirky-cased subject” verbs.
Does any of that help?
Well, I'm still not sure how "concordial" necessarily relates to Géarthnuns. Yes, there are a handful (fistful?) of adverbs that "agree" with the auxiliary in tense only, not in voice; I think this speaks more to the nature of such adverbs and not to verb agreement, per se. Verb class has overlapped with noun class back in the hoarfrost of the early language; i.e. early *basic* verbs in "-l" deal with speech acts and sensory verbs; some motion verbs end in "-z". But that's semantic, I guess, and hardly invokes any type of agreement on the verb side. How "quirky" that makes it in terms of agreement (with verbs; nouns are a standard declension system we are all familiar with).
道可道,非常道
名可名,非常名

User avatar
eldin raigmore
korean
korean
Posts: 6389
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by eldin raigmore » 31 Jul 2019 02:10

Lao Kou wrote:
30 Jul 2019 18:09
....
Well, I'm still not sure how "concordial" necessarily relates to Géarthnuns. Yes, there are a handful (fistful?) of adverbs that "agree" with the auxiliary in tense only, not in voice; I think this speaks more to the nature of such adverbs and not to verb agreement, per se. Verb class has overlapped with noun class back in the hoarfrost of the early language; i.e. early *basic* verbs in "-l" deal with speech acts and sensory verbs; some motion verbs end in "-z". But that's semantic, I guess, and hardly invokes any type of agreement on the verb side. How "quirky" that makes it in terms of agreement (with verbs; nouns are a standard declension system we are all familiar with).
(That’s not a very good objection to “concordial”, but it’s an excellent objection to “verb-class”!)

Was I wrong in thinking that you were the one who contributed the post about Géarthnuns to that CONLANG LISTSERV thread “THEORY: Gender in Verbs” eleven years and nine or ten months ago?
If you were the submitter*, you might re-read it, and might recall to mind what you were thinking then. It’s more than possible I misunderstood it, but if so, I’d need you to explain it to me, if you won’t mind.
OTOH, if you weren’t the submitter, it’s likely that whoever was*, misunderstood your conlang (or your description thereof) as it was at that time.
Or, maybe, they saw something you haven’t seen yet. Maybe because it’s not there; maybe because you just weren’t thinking along those lines.

*Douglas Koller

—————

In any case; I think you are the world’s expert on Géarthnuns!

User avatar
eldin raigmore
korean
korean
Posts: 6389
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: FreeWordOrder SerialVerb Concordial VerbClass

Post by eldin raigmore » 31 Jul 2019 20:18

https://www.researchgate.net/publicatio ... neighbours seems to indicate that in Waanyi:
* verbs in serial-verb constructions can come in any order and
* the verb-phrases within an SVC can be discontinuous.

—————
In https://pdfs.semanticscholar.org/3d61/f ... 14407e.pdf I see:
Spoiler:
2.4.6 The verbal classification approach (McGregor 2002)
Many works on Australian bipartite verbal structures assume some kind of verbal classification more or less implicitly, e.g., Capell (1979), McGregor (1990), Reid (1990, 2011), Nicolas (2000), Schultze-Berndt (2000), Bowern (2004). But by far the most explicit treatment of bipartite verbal con- structions as verbal classification constructions in Australian languages is presented by McGregor (2002). In his approach, the inflecting verbs classify coverbs into classes, similarly to nominal classification. That is, the coverbs form classes based on their compatibility with inflecting verbs. Three dif- ferent factors may play a role in the classification, i.e., valency, Aktionsart and vectorial configuration. (emphasis mine — eldin raigmore)
As he points out, this classification differs from verb classes as proposed by e.g., Fillmore (1968) or Levin (1993) who divide verbs based on tran- sitivity, felicity etc. While these classifications are based on the inherent properties of the verbs, i.e., the classification is covert, the verbal classifica- tion in McGregor’s (2002) sense is overt. He further distinguishes his type of verbal classification with verbal classification in Amerindianist linguistics in which the verb serves to classify the object.
What is important to the discourse of Australian bipartite verbal con- structions is the distinction between the verbal classification approach pro- posed by McGregor (2002) and the event classification approach proposed by others, most prominently Schultze-Berndt (2000), but also Bowern (2004). As laid out above, in the verbal classification approach the inflecting verb serves to classify the coverbs. In contrast, in the event classification ap- proach, the inflecting verb serves to classify the event described by the coverb and inflecting verb together. I agree with Barone-Nugent (2008) in that both approaches are not mutually exclusive, as the event classifi- cation approaches usually encompass the verbal classification in discussing event classification based on the classes of coverbs that are determined by the compatibility of the coverbs with inflecting verbs.
McGregor (2002) further argues that Australian bipartite constructions should not be considered complex predicates and that consequently, fusion or union analyses are not applicable. His main reason for the rejection of a complex predicate analysis lies in his reservations towards the notion of a ‘head’. In his view, the definition of a syntactic head is not clear and even if it were, the light verbs and coverbs in Australian bipartite verbal complexes could not be heads at the same time. This, then, excludes the construction from a complex predicate analysis.

Complex predicates crosslinguistically 67
However, if a verbal head is defined as an element which determines the argument structure and semantics of an event, both inflecting verbs and coverbs are indeed heads and form a complex predicate together as co-heads. Bowern (2004) and Barone-Nugent (2008) share the view that a classification approach is not mutually exclusive with a complex predicate analysis.
So, what is “vectorial configuration”?

User avatar
eldin raigmore
korean
korean
Posts: 6389
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Concordial VerbClass

Post by eldin raigmore » 07 Aug 2019 02:16

Can anyone think of a good use for concordial verb-class, other than telling which arguments go with which verbs in serial-verb-constructions, in languages with flexible or pragmatic or “free” word-order?

For instance; maybe a clause has two or more verbs; and some of the clause’s other constituents, “go with” one or some of the verbs but don’t “go with” one or some of the other verbs. Especially if there are other constituents that “go with” a different proper subset of the clause’s verbs.

The constituents in question needn’t be arguments of the verb(s) in question. (They might be auxiliaries or adverbs or something.)
And the verbs needn’t form a serial-verb construction. They might be co-ordinately conjoined; or subordinately conjoined; or something.

—————

Possibly related; maybe a verb’s dependent words are marked to agree with the verb’s modality-or-mode-or-mood. And maybe subordinated verbs (ie verbs of subordinate clauses) are always m/m/m -marked (eg as “subjunctive”) to show that they’re subordinate. Then the words dependent on the main clause’s verb would probably have a different mood-mark than the words dependent on the subordinate clause’s verb.

Something reminiscent of, if not like, this, happens in the prototypical uses of logophoric pronouns.
Logophoric pronouns are prototypically used to co-refer to certain nominals in clauses that contain an indirectly-quoted complement clause. (Such a clause is embedded, but ordinarily not dependent for part of its meaning on the main clause, and so ordinarily not technically a “subordinate” clause.)
A “logophoric first person” pronoun is a third-person pronoun that co-refers with the speaker of the (indirectly?)-quoted clause.
A “logophoric second person” pronoun is a third-person pronoun that co-refers with the addressee of the (indirectly?)-quoted clause.

Possibly if a clause contains an embedded clause, other constituents might be marked to show whether they depend on the verb of the main clause, or instead depend on the verb of the embedded clause.

That might be simpler if the two verbs were of different concordial verb-classes. Or not.

——————

Suppose some NP or other constituent, were an argument both of the main clause and of the embedded clause.
Should it be case-marked to show whether and how it participates in the main clause, and also to show whether and how it participates in the embedded clause?
For instance, some nominal or pronominal thing might be the subject or direct object or indirect object or oblique argument or even an adjunct of the main clause, while also being the subject of a relative subordinate clause which modifies it (as an adjective would).
How should that be shown?
Obviously, the nominal or pronominal modified by a relative clause, must be the topic of the RC; but in some languages it could be the direct or indirect object, or some other argument, of the RC, instead of the subject. Does it need to be marked to show that?
Would concordial verb-class help? At least sometimes?

——————

In conditional statements, how should constituents of the protasis that depend on the protasis’s verb, be marked to show that, in contrast to constituents of the apodisis that depend on the apodosis’s verb? And vice-versa; how should things that go with the apodosis be marked to show they don’t go with the protasis, if they don’t? And how about things that go with both? Especially if they relate to the protasis’s verb different than the apodosis’s verb?

Would concordial verb-classes help?

==============================

OK, I’ve talked about a lot of two-or-more-verb situations that aren’t serial-verb-constructions.

How about dependents of verbs that aren’t its arguments?
Like auxiliaries and adverbs?
I would think even in SVC languages there are likely to be two-or-more-verb clauses in which some auxiliary or adverb goes with just one of the verbs, or doesn’t go with one of the verbs.
How is that shown?
Would concordial verb-class help?

Post Reply