Page 1 of 1
Languages with optional person indexing on verbs
Posted: Sun Aug 13, 2017 6:51 pm
by con quesa
Is anyone aware of any literature on languages that have non-obligatory person marking on verbs? By this I mean languages where you can use some kind of morphology to index an argument of the verb, but that is not required if an overt pronoun exists in the clause?
Re: Languages with optional person indexing on verbs
Posted: Tue Aug 29, 2017 6:10 pm
by Pabappa
I suspect this doesnt happen or is at most highly unstable. If an inflection is optional, that would mean that the null form is meaningful on its own. This would mean that at least one form of the verb likely already uses the null form as standard. Such a setup might turn up in a language undergoing rapid sound change, such as perhaps medieval French, but I'd expect the zero-marked form of the verb to become the standard form and the pronouns to become mandatory.
Re: Languages with optional person indexing on verbs
Posted: Tue Aug 29, 2017 8:47 pm
by Sumelic
con quesa wrote:Is anyone aware of any literature on languages that have non-obligatory person marking on verbs? By this I mean languages where you can use some kind of morphology to index an argument of the verb, but that is not required if an overt pronoun exists in the clause?
Some varieties of modern French could sort of be seen like this. However, the person marking (which is what is traditionally analyzed as a "subject pronoun") is in fact obligatory in clauses with overt (disjunctive) pronouns, as far as I know; it's only "optional" (and not in a straightforward manner) with third-person verbs with an overt subject (and a fair number of verbs have unique suffixal inflections to mark third-person
plural that are always present, like 3P
savent /sav/ "know" vs. singular (1S, 2S (non-V), 3S) /sɛ/ "know").
That is,
"I like vegetables" = "J'aime les légumes"; the
j' cannot be removed even if a disjunctive pronoun "moi" is used ("moi, j'aime les légumes"; "*moi aime les légumes")
But "my brother likes vegatables" = "mon frère aime les légumes" or "mon frère, il aime les légumes".
Of course, this looks a lot like the way left-dislocation works in a language like English that is pretty much never thought of as having "person marking" on verbs. However, French subject pronouns are more closely bound to the verb than English subject pronouns; they don't normally occur before other types of words in ordinary French syntax (although the legalistic phrase "je soussigné(e)" does feature "je" before a word other than a verb).
Re: Languages with optional person indexing on verbs
Posted: Wed Aug 30, 2017 2:13 am
by Astraios
You might also look at Coptic, the subject of whose verbs can be either a bound pronominal or a disjunctive nominal phrase, which can either intervene between a TAM-marking particle and the verb itself:
ⲁϥⲕⲟⲧⲥ
a-f-kot-s
Perf-3MS-build-3FS
“he built it”
ⲁⲡϫⲟⲉⲓⲥ ⲕⲟⲧⲥ
a-p-joeis kot-s
Perf-Def.MS-lord build-3FS
“the Lord built it”
… or precede it:
ϥⲛⲁⲕⲟⲧⲥ
f-na-kot-s
3MS-Fut-build-3FS
“he will build it”
ⲡϫⲟⲉⲓⲥ ⲛⲁⲕⲟⲧⲥ
p-joeis na-kot-s
Def.MS-lord Fut-build-3FS
“the Lord will build it”
, which looks at first glance to be close to what you want. As with French, though, the bound pronouns can also optionally appear with a disjunctive nominal: ⲡϫⲟⲉⲓⲥ ⲁϥⲕⲟⲧⲥ “the Lord, he built it”, so all we seem to have is more left-dislocation with agglutinative-ish bound pronouns.
It’s not entirely clear to me, however, that the disjunctive pronouns can be used without a bound pronoun: that is, that ⲛⲁⲕⲟⲧⲥ “… will build it” can appear with a disjunctive pronoun. It seems not to be the case; that is, that only ⲛⲧⲟϥ ϥⲛⲁⲕⲟⲧⲥ is valid for “he will build it”, with both an overt pronoun and a bound affix; but I don’t have any evidence that ⲛⲧⲟϥ ⲛⲁⲕⲟⲧⲥ is invalid. So at least for disjunctive pronouns, Coptic is the same as French.
Re: Languages with optional person indexing on verbs
Posted: Thu Aug 31, 2017 10:27 pm
by Gaxa
Bantu languages have optional object marking on verbs
(here's a paper about this). As far as I know you don't find optional subject agreement though. Here's the first example from that paper from Chichewa (SM is "subject marker", OM is "object marker", and FV is the "final vowel" that appears on Bantu verbs for phonological reasons):
Code: Select all
Njuchi zinawaluma.
njuchi zi-na- wa-lum -a
bees SM-PST-OM-bite-FV
"The bees bit them"
Njuchi zinaluma alenje.
njuchi zi-na- lum -a alenje
bees SM-PST-bite-FV hunters
"The bees bit the hunters"
Njuchi zinawaluma alenje.
njuchi zi-na- wa-lum -a alenje
bees SM-PST-OM-bite-FV hunters
"The bees bit (them,) the hunters"
People have proposed various analyses of this phenomenon for various languages. There's the analysis that's been mentioned above that we have a cliticized/incorporated pronoun together with some dislocated noun phrase. The paper I linked argues that this is the case in Chichewa, making it the same as French and apparently Coptic.
However, the other analysis, which is (I think) what you want, is that you get "true" optional object agreement on the verb, as opposed to a cliticized/incorporated pronoun + dislocated noun phrase. Sambaa (another Bantu language) appears to act like this
(here's a dissertation on the topic, and the first paper I linked also talks about this). There are several arguments that we have "true" object marking, and not the French-Coptic-Chichewa type pronoun-doubling + object dislocation:
- Bantu verbs typically have two forms: a "conjoint" form and a "disjoint" form (also termed conjunct vs. disjunct or dependent vs. independent). The choice of which form to use is subject to prosodic requirements (which can vary across languages), but essentially conjoint verbs must be followed by stuff within the same prosodic phrase, whereas disjoint verbs can't be followed by stuff within the same prosodic phrase. If the verb and the object are phrased together, then we would get the conjoint form; if the verb and the object are phrased separately (likely due to the object being dislocated, as it could be for focus), then we would get the disjoint form. In Sambaa, you can get both conjoint and disjoint forms with and without doubled object marking, indicating that we don't (always) have object-dislocation together with pronoun doubling. Sambaa behaves differently from French/Coptic/Chichewa in this regard.
- There are certain phonological processes that occur between verbs and objects, like high tone spreading, that are blocked when an object is dislocated. If an object is dislocated, it is no longer in the same prosodic phrase as the verb, so phonological processes that occur in this domain are blocked. In Sambaa, these processes appear between verb and objects whether there is object agreement or not, further suggesting that we don't have dislocation of objects when object agreement shows up.
- (Optional) object agreement is possible with negative noun phrases (1sS is first singular subject, OM7 is class 7 object marker):
Code: Select all
Sichioniye kintu chochoshe.
si- chi-on -iye kintu chochoshe
NEG.1sS-OM7-see-PERF 7thing 7any
"I didn't see anything."
So it seems that there are languages, like Sambaa, that have non-obligatory person marking on verbs, at least for objects, and this marking is "true" agreement and not just pronoun doubling like in French, Coptic, or Chichewa. I'm not sure if you can get an overt pronoun object with object agreement though---I couldn't find any examples of that, but hopefully this is helpful.