Unusual verbal person-marking systems

[Chagen]

I've come to realize that there's not much information on there comparing sets of verbal person marking. WALS, for instance, has chapters on things like which argument is marked, or if the 3S is zero-marked, but I can't find anything on what kinds of persons are attested cross-linguistically.

From my experience, it appears that the most prototypical marking is 1, 2, and 3, often with plurality also distinguished (so a system of 1S/2S/3S/1P/2P/3P). Many languages also have an exclusive/inclusive distinction in the 1P. Then a few also have a "fourth-person", which is usually impersonal. Oddly enough, languages which have noun gender/class often don't distinguish it in the verbal markers (or maybe this is just another weird quirk of Indo-European languages) But what else is attested? I'm especially interested in things which veer completely off this familiar pattern. For instance, are there languages which have interrogative markers (i.e languages where you can say, e.g, do-PST-INTER for "who did (this)?"). Or ones with a negative suffix. Or other kinds of unusual "persons".

I'm interested in this mainly because I'm thinking of giving my Sunbyaku reboot a dense system with eleven distinct persons (1S/2S/3S.ANIM/3S.INANIM/4S/1P/2P/3P/4P/INTER/NEG), with possibly even more (such as a suffix which specifically means "which one of you/them"), and I'm wondering if this amount of complexity is attested in real life.

[Frislander]

For instance, are there languages which have interrogative markers (i.e languages where you can say, e.g, do-PST-INTER for "who did (this)?"). Or ones with a negative suffix. Or other kinds of unusual "persons".

The first two aren't really "person": interrogative marking in its basic form like the one you give here would mean something like "was it done/did someone do it" i.e. a polar interrogative not a WH one, while the negative is definitely not a "person" thing, though both of these could be fused with the marking for person/number.

As for "unusual persons", there doesn't seem to be anything radical anywhere: 1st-2nd-3rd is a pretty strong distinction to make, and languages without it have yet to be found anywhere.

[Chagen]

For instance, are there languages which have interrogative markers (i.e languages where you can say, e.g, do-PST-INTER for "who did (this)?"). Or ones with a negative suffix. Or other kinds of unusual "persons".

The first two aren't really "person": interrogative marking in its basic form like the one you give here would mean something like "was it done/did someone do it" i.e. a polar interrogative not a WH one, while the negative is definitely not a "person" thing, though both of these could be fused with the marking for person/number.

Is it truly inconceivable that a language would have a verbal person marker corresponding to "who/what", as it does "I", "you", and "he/she/it"? Your sentence is really weird as you're prescribing for no justifiable reason that my example can only mean one thing no matter what the language. As for the negative marker, in Sunbyaku at least, I envision it working like so (sorry for having only glosses):

last.night=TOP road=LOC move-GER=LOC see-PASS-PST-NEG
No cars were seen driving down the road last night

DEM=TOP all be.happy-NEG
Not a single one of them is happy about this

[Frislander]

For instance, are there languages which have interrogative markers (i.e languages where you can say, e.g, do-PST-INTER for "who did (this)?"). Or ones with a negative suffix. Or other kinds of unusual "persons".

The first two aren't really "person": interrogative marking in its basic form like the one you give here would mean something like "was it done/did someone do it" i.e. a polar interrogative not a WH one, while the negative is definitely not a "person" thing, though both of these could be fused with the marking for person/number.

Is it truly inconceivable that a language would have a verbal person marker corresponding to "who/what", as it does "I", "you", and "he/she/it"? Your sentence is really weird as you're prescribing for no justifiable reason that my example can only mean one thing no matter what the language. As for the negative marker, in Sunbyaku at least, I envision it working like so (sorry for having only glosses):

last.night=TOP road=LOC move-GER=LOC see-PASS-PST-NEG
No cars were seen driving down the road last night

DEM=TOP all be.happy-NEG
Not a single one of them is happy about this

Yes I pretty much am saying that, because that's how it works. WH-questions are more "wordy" than polar-questions: if you're going to mark polar questions on the verb like that, you're best strategy is to have a separate wh-interrogative mood and use an affix representing and indefinite subject/object to represent the wh-word. A structure like "do-PST-INT" will always be a polar-interrogative.

[cromulant]

It seems like it would be pretty easy for wh- words to become verbal affixes. Especially if the lang is verb-initial and wh-fronting. Another possibility would be a "wh-mood" used with a Phillipine voice system.

[hwhatting]

I don't know any language that marks questions or negation under the category person. The languages I know that mark negation on the verb mark it as a separate suffix from person. That makes a certain sense, as you can negate action in any person ("I didn't do it, you didn't do it", etc.) The languages I know that mark questions on the verb actually mark polar questions, and here the same applies as for negation - polar questions need to be combinable with all persons. But that is not necessary for wh-questions when the subject is not known, so it is at least possible that a language would have a special personal ending for such questions. But I don't know any such language.

[Chagen]

I don't know any language that marks questions or negation under the category person. The languages I know that mark negation on the verb mark it as a separate suffix from person. That makes a certain sense, as you can negate action in any person ("I didn't do it, you didn't do it", etc.) The languages I know that mark questions on the verb actually mark polar questions, and here the same applies as for negation - polar questions need to be combinable with all persons. But that is not necessary for wh-questions when the subject is not known, so it is at least possible that a language would have a special personal ending for such questions. But I don't know any such language.

Sunbyaku's negative person has a somewhat different meaning than the base negative:

shirabemasu
be.happy-NEG-3P.STAT
They are not happy

shiraberoku
be.happy-NEG-NEG.PER.STAT
None of them are happy

The first is a simple negation, the second uses the negative person and emphasizes that none of the referred to people are happy in the slightest. Is this really not attested anywhere on Earth?

[Sumelic]

For instance, are there languages which have interrogative markers (i.e languages where you can say, e.g, do-PST-INTER for "who did (this)?"). Or ones with a negative suffix. Or other kinds of unusual "persons".

The first two aren't really "person": interrogative marking in its basic form like the one you give here would mean something like "was it done/did someone do it" i.e. a polar interrogative not a WH one, while the negative is definitely not a "person" thing, though both of these could be fused with the marking for person/number.

Is it truly inconceivable that a language would have a verbal person marker corresponding to "who/what", as it does "I", "you", and "he/she/it"? Your sentence is really weird as you're prescribing for no justifiable reason that my example can only mean one thing no matter what the language. As for the negative marker, in Sunbyaku at least, I envision it working like so (sorry for having only glosses):

last.night=TOP road=LOC move-GER=LOC see-PASS-PST-NEG
No cars were seen driving down the road last night

DEM=TOP all be.happy-NEG
Not a single one of them is happy about this

Yes I pretty much am saying that, because that's how it works. WH-questions are more "wordy" than polar-questions: if you're going to mark polar questions on the verb like that, you're best strategy is to have a separate wh-interrogative mood and use an affix representing and indefinite subject/object to represent the wh-word. A structure like "do-PST-INT" will always be a polar-interrogative.

Chagen never said he would "mark polar questions on the verb like that". You're misinterpreting the first example; the context makes it clear that Chagen was asking about using an interrogative affix on the verb for non-polar questions to ask about the identity of the subject.

That certainly seems possible to me. Polar questions might use another verb affix, or there might be a separate negative particle that needs to be used with any type of question (so e.g. "INT do-PST-3" is "did he do (this)?" while "INT do-PST-INT" is "who did (this)"?").

It doesn't seem likely, but I don't see why it couldn't exist as a rare feature.

As for "unusual persons", there doesn't seem to be anything radical anywhere: 1st-2nd-3rd is a pretty strong distinction to make, and languages without it have yet to be found anywhere.

The question was specifically about verbal person marking. There are certainly many languages that don't have any verbal inflection for person, or that only distinguish incompletely between different persons in verb inflection (English is one of them).

I'm interested in this mainly because I'm thinking of giving my Sunbyaku reboot a dense system with eleven distinct persons (1S/2S/3S.ANIM/3S.INANIM/4S/1P/2P/3P/4P/INTER/NEG)

What's an example where you would make a number distinction in the 4th person? From what I understand, this is usually used to refer to affixes meaning something like "one" or "someone," so it's often underspecified for person.

I don't know any language that marks questions or negation under the category person. The languages I know that mark negation on the verb mark it as a separate suffix from person. That makes a certain sense, as you can negate action in any person ("I didn't do it, you didn't do it", etc.) The languages I know that mark questions on the verb actually mark polar questions, and here the same applies as for negation - polar questions need to be combinable with all persons. But that is not necessary for wh-questions when the subject is not known, so it is at least possible that a language would have a special personal ending for such questions. But I don't know any such language.

Sunbyaku's negative person has a somewhat different meaning than the base negative:

shirabemasu
be.happy-NEG-3P.STAT
They are not happy

shiraberoku
be.happy-NEG-NEG.PER.STAT
None of them are happy

The first is a simple negation, the second uses the negative person and emphasizes that none of the referred to people are happy in the slightest. Is this really not attested anywhere on Earth?

I don't know. I've read that negative clauses with a separate negative word and no verbal negation, like "None of them are happy," are typologically unusual. It's more common in general to use the normal verbal negation plus some other words; in this context, maybe a construction like "All of them are not happy", "Each of them is not happy", or "There isn't one of them that is happy" would be more probable (you can read some about negation strategies here: http://www.academia.edu/18532901/On_the ... ve_concord)

[Frislander]

I don't know any language that marks questions or negation under the category person. The languages I know that mark negation on the verb mark it as a separate suffix from person. That makes a certain sense, as you can negate action in any person ("I didn't do it, you didn't do it", etc.) The languages I know that mark questions on the verb actually mark polar questions, and here the same applies as for negation - polar questions need to be combinable with all persons. But that is not necessary for wh-questions when the subject is not known, so it is at least possible that a language would have a special personal ending for such questions. But I don't know any such language.

Sunbyaku's negative person has a somewhat different meaning than the base negative:

shirabemasu
be.happy-NEG-3P.STAT
They are not happy

shiraberoku
be.happy-NEG-NEG.PER.STAT
None of them are happy

The first is a simple negation, the second uses the negative person and emphasizes that none of the referred to people are happy in the slightest. Is this really not attested anywhere on Earth?

That's not a 'negative person', that incorporating meanings normally expressed by separate quantifiers into the person-marking system, and there doesn't appear to be any language that does this.

You could again use an idefinite-subject/object marker and couple it with the standard negative to form this meaning, like what I suggested with the interrogatives above.

[Sumelic]

I don't know any language that marks questions or negation under the category person. The languages I know that mark negation on the verb mark it as a separate suffix from person. That makes a certain sense, as you can negate action in any person ("I didn't do it, you didn't do it", etc.) The languages I know that mark questions on the verb actually mark polar questions, and here the same applies as for negation - polar questions need to be combinable with all persons. But that is not necessary for wh-questions when the subject is not known, so it is at least possible that a language would have a special personal ending for such questions. But I don't know any such language.

Sunbyaku's negative person has a somewhat different meaning than the base negative:

shirabemasu
be.happy-NEG-3P.STAT
They are not happy

shiraberoku
be.happy-NEG-NEG.PER.STAT
None of them are happy

The first is a simple negation, the second uses the negative person and emphasizes that none of the referred to people are happy in the slightest. Is this really not attested anywhere on Earth?

That's not a 'negative person', that incorporating meanings normally expressed by separate quantifiers into the person-marking system, and there doesn't appear to be any language that does this.

You could again use an idefinite-subject/object marker and couple it with the standard negative to form this meaning, like what I suggested with the interrogatives above.

Why are you so hung up on terminology? You understand what Chagen means, right? It's not like linguistic terminology is always used consistently in natlangs anyway. If I want to call a construction in a conlang "negative person," why can't I?

[Frislander]

For instance, are there languages which have interrogative markers (i.e languages where you can say, e.g, do-PST-INTER for "who did (this)?"). Or ones with a negative suffix. Or other kinds of unusual "persons".

The first two aren't really "person": interrogative marking in its basic form like the one you give here would mean something like "was it done/did someone do it" i.e. a polar interrogative not a WH one, while the negative is definitely not a "person" thing, though both of these could be fused with the marking for person/number.

Is it truly inconceivable that a language would have a verbal person marker corresponding to "who/what", as it does "I", "you", and "he/she/it"? Your sentence is really weird as you're prescribing for no justifiable reason that my example can only mean one thing no matter what the language. As for the negative marker, in Sunbyaku at least, I envision it working like so (sorry for having only glosses):

last.night=TOP road=LOC move-GER=LOC see-PASS-PST-NEG
No cars were seen driving down the road last night

DEM=TOP all be.happy-NEG
Not a single one of them is happy about this

Yes I pretty much am saying that, because that's how it works. WH-questions are more "wordy" than polar-questions: if you're going to mark polar questions on the verb like that, you're best strategy is to have a separate wh-interrogative mood and use an affix representing and indefinite subject/object to represent the wh-word. A structure like "do-PST-INT" will always be a polar-interrogative.

Chagen never said he would "mark polar questions on the verb like that". You're misinterpreting the first example; the context makes it clear that Chagen was asking about using an interrogative affix on the verb for non-polar questions to ask about the identity of the subject.

That certainly seems possible to me. Polar questions might use another verb affix, or there might be a separate negative particle that needs to be used with any type of question (so e.g. "INT do-PST-3" is "did he do (this)?" while "INT do-PST-INT" is "who did (this)"?").

It doesn't seem likely, but I don't see why it couldn't exist as a rare feature.

Well then gloss it as something more distinct than just INT to avoid confusion with the polar version, like what I did there.

However, as this is the L&L forum, give me some natlang examples which handle things this way.

As for "unusual persons", there doesn't seem to be anything radical anywhere: 1st-2nd-3rd is a pretty strong distinction to make, and languages without it have yet to be found anywhere.

The question was specifically about verbal person marking. There are certainly many languages that don't have any verbal inflection for person, or that only distinguish incompletely between different persons in verb inflection (English is one of them).

I rather got the impression that Chagen was asking about persons as a concept:

I can't find anything on what kinds of persons are attested cross-linguistically.

But clearly I misinterpreted.

I'm interested in this mainly because I'm thinking of giving my Sunbyaku reboot a dense system with eleven distinct persons (1S/2S/3S.ANIM/3S.INANIM/4S/1P/2P/3P/4P/INTER/NEG)

What's an example where you would make a number distinction in the 4th person? From what I understand, this is usually used to refer to affixes meaning something like "one" or "someone," so it's often underspecified for person.

Well '4th person' can mean several different things, some of which are OK with taking plural marking, others less so.

I don't know any language that marks questions or negation under the category person. The languages I know that mark negation on the verb mark it as a separate suffix from person. That makes a certain sense, as you can negate action in any person ("I didn't do it, you didn't do it", etc.) The languages I know that mark questions on the verb actually mark polar questions, and here the same applies as for negation - polar questions need to be combinable with all persons. But that is not necessary for wh-questions when the subject is not known, so it is at least possible that a language would have a special personal ending for such questions. But I don't know any such language.

Sunbyaku's negative person has a somewhat different meaning than the base negative:

shirabemasu
be.happy-NEG-3P.STAT
They are not happy

shiraberoku
be.happy-NEG-NEG.PER.STAT
None of them are happy

The first is a simple negation, the second uses the negative person and emphasizes that none of the referred to people are happy in the slightest. Is this really not attested anywhere on Earth?

That's not a 'negative person', that incorporating meanings normally expressed by separate quantifiers into the person-marking system, and there doesn't appear to be any language that does this.

You could again use an idefinite-subject/object marker and couple it with the standard negative to form this meaning, like what I suggested with the interrogatives above.

Why are you so hung up on terminology? You understand what Chagen means, right? It's not like linguistic terminology is always used consistently in natlangs anyway. If I want to call a construction in a conlang "negative person," why can't I?

Again, let's have some natlang examples.

[Sumelic]

Again, let's have some natlang examples.

Yeah, I'm looking for those right now! Are you an expert on the typology of interrogative marking in languages? Otherwise, I don't see how you can confidently state "natlangs don't do things this way" just because you can't think of a natlang that does things this way.

So far, the closest to a hint I have found that some natlangs might have this is in the following passage from "Content Questions in American Sign Language" by by Kathleen-Renee Binns-Dray (PDF download: https://www.google.com/url?sa=t&rct=j&q ... 4068,d.amc)

More languages than not use particles in polar questions.
Among those languages that do employ a particle for polar questions, it can occur
in initial, second, or final position, or in two of the three; in a few languages, the
particle occurs in some other position. Because content questions are generally
identified by the use of a question word in most spoken languages, most do not
also employ particles to signal content questions. To classify languages
typologically, Dryer proposes five types of languages with regard to question
particles.

This says content questions are "generally" identified by the use of a question word in "most" spoken languages; this may just be hedging, or it might indicate that there are some known natlangs where not all types of content questions involve a separate question word.

[Sumelic]

Here is something that is maybe somewhat similar to what Chagen mentioned:

Non-interrogative signs as question words.

Some content questions may be formed by using a non-interrogative sign with a facial expression for content questions (to be described below), without a question word (see also example 16).
(8) time? ‘What time is it?’
money? ‘How much does it cost?’
age index2? ‘How old are you?’
health index2? ‘How are you?’

(Question and Negation in Israeli Sign Language, Irit Meir)

However, that seems more a matter of intonation.

[Sumelic]

According to the grammar by Eung-Do Cook that I found, it seems that some types of content questions in Tsilhqút’ín can be indicated by the verbal prefix "hú-" without any separate question word or indefinite word (540). The meaning is broad, but in fact it cannot be used to question the identity of the subject, as in Chagen's example (a separate question word/indefinite word does need to be used for that). Cook says the prefix can be used to ask "how," "where" and "what" questions.

[So Haleza Grise]

Gurr-goni has a very intricate system of personal marking that's hard to summarise. Basically, the number distinction can be seen as distinguishing between "minimal", "unit-augmented" and "augmented", where "minimal" is the base number, and then you add +1 for unit-augmented, and more than that for augmented. For third person, this is the same as calling it singular/dual/plural, but there is a separate category for "first + second person" - ie. "you and I". In that case, "minimal" is two people and "unit augmented" is three people, not two.

Another factor is gender marking. Gender is occasionally marked on verbs, specifically in the "unit-augmented" number. There are (to simplify somewhat) four nominal genders but only one distinction is marked on verbs: "feminine" versus "non-feminine". The rule of thumb in the third person is that if at least one of the NPs is female use feminine agreement, otherwise use masculine agreement. In the case of the first and second persons, it's the gender of the "and one other" of the unit augment that determines the agreement: if this is female use feminine, if it is male use masculine.

Person agreement exists for subject and object but the way it operates is somewhat complicated: there are fewer distinction forms for every distinct combination of person, number and gender of both subject and object. The free pronouns do not exactly line up with the pronoun marking on verbs; in particular "1+2" is not distinguished in non-minimal free pronouns.

[Nortaneous]

in Modo, 1SG and 2 are marked with a high toneme on the verb, and 1PL and 3 are marked with a low toneme on the verb

[Nachtuil]

Does any language conjugate verbs for subject noun class?

[Sumelic]

Does any language conjugate verbs for subject noun class?

That's a relatively common thing to mark on verbs, I think. Within Indo-European, there's Russian past tense (historically derived from a participle, so it inflects for gender but not person). This occurs in other Slavic languages too, IIRC, but I don't remember which. Verbs in Semitic languages also generally conjugate for gender of 3rd-person and sometimes 2nd-person subjects.

Bantu languages also have noun class subject prefixes on verbs, but apparently in some languages these behave more like separate words or like clitics than like inflectional agreement.

[Frislander]

In Northeast Caucasian Gender is practically the only thing verbs agree with their subject in for many languages, and I think the Northwest paradigms also find space for a masculine-feminine distinction as well.

[Nachtuil]

Thanks guys! That is really good to know. It is so wild that the past tense of some Russian verbs have a different inflectional pattern like that from the present tense. I want to inflect verbs by gender but wasn't sure if it is done a lot so I figured I would ask.

[cromulant]

Classical Ainu has the following system:

Third person arguments are always zero-marked.

Number is only distinguished in the second person. The second person affixes only indicate the presence of the second person in the clause, with no indication as to grammatical role. Intransitive subjects, transitive agents, and transitive patients are all (non-zero) marked identically.

Grammatical role is only distinguished in the first person, and the agreement follows tripartite alignment. But there is no number distinction.


However, Colloquial Ainu adds inclusive and exclusive first person plurals to the paradigm. The inclusive forms are the same as the Classical Ainu number-indifferent forms; the singular and exclusive forms are innovations. Again, all three first person paradigms have tripartite alignment.

[vokzhen]

The Ainu system reminded me of Ayutla Mixe, which isn't what Chagen is looking for in the OP but I still think is worth bringing up.

In Ayutla Mixe, there are two conjugation patterns, traditionally called dependent and independent. These don't really have to do with subordination, but rather whether or not their are non-arguments before the verb (including things like tense particles), which ends up with word order correlations: dependent marking is overwhelmingly SOV in order (with OSV and SVO as other possibilities), while independent marking freely varies between all possible orders (though the most common for transitives is OV with no explicit subject). Dependency is marked by different sets of person prefixes and aspect-mood suffixes (neutral [traditionally incompletive], a rare and dying completive, and irrrealis). In addition, agreement is hierarchical, agreeing only with the highest-animacy argument (1>2>3>3').

For dependent inflection, 1st person is accusative, with n- marking subject and agent, and ʂ- marking patient. 3rd person is ergative, 3rd person agent t- and subject/patient j-. 2nd person is also ergative, with subject/patient 3 m-, but it has the same agent marker ʂ- as the 1st person patient, making ʂ- ambiguous between 2nd acting on 3rd, 2nd acting on 1st, and 3rd acting on 1st.

For independent inflection, 1st person is tripartite, with subject ∅-, agent n-, and patient ʂ-. 2nd person is unaligned, they all take m-. 3rd person is ergative, with agent j-, but it uses ∅- for both subject and patient, making zero-marking ambiguous between a 1st person subject, 3rd person subject, and 3' acting on 3rd. However, occasionally or 3rd person agents, t- from the dependent conjugation shows up on independent transitives with no NP arguments, which then takes the independent intransitive neutral aspect marker -p.

Plurality is marked with a suffix -tɨ. It can mark any core argument plural, generally but not obligatorily the one referenced by the person marker. In independent conjugation it contracts with the neutral aspect marker, -tɨ-yp for transitives merges with -tɨ-p for intransitives. Likewise it contracts in the neutral dependent, -t-y yields -t (dependent has no transitive-intransitive distinction). Imperatives have -t(ɨ), and elicited examples of completives have -t.

There is also a 1st plural inclusive with a dedicated suffix that suppresses normal aspect-mood marking: -ɨˀn is obligatory for irrealis, except after the perfect or a long vowel where it is -(ˀ)n, while neutral aspect commonly has -jɨˀm ~ -jɨˀn but allows the former as well. The 1INC can trigger changes in laryngeal state of the preceding vowel, similarly but not always identically to the aspect-mood suffixes: -men- "come" independent inflections /memp/ neutral /meˀmp/ irrealis /me:n/ completive /meˀenjɨˀm/ 1INC.

There is an inverse suffix -ɨ, but it is never present for 1st persons, in the completive aspect, or for singular agents in dependent conjugation.

[So Haleza Grise]

Does any language conjugate verbs for subject noun class?

Gurr-goni does this (and Russian does in the past tense as well)

[2+3 clusivity]

Take a look at Na-Dene' langauges: Some of them not only cut cross-referncing for the non-1/2 person(s) up into proximate and obviate, but have seperate areal, impersonal, etc. marking. To me, it is especially interesting to see that some of these destinctions are marked in both the "subject" and "object" slots, as opposed to the reflexives and reciprocals which are marked exclusively in the "object" slot on the verb. Take a look especially at Hupa, Navajo, and Tlingit. Eyak looks like it may do something similar skimmig though wikipedia but it's less clear.

[Frislander]

Take a look at Na-Dene' langauges: Some of them not only cut cross-referncing for the non-1/2 person(s) up into proximate and obviate, but have seperate areal, impersonal, etc. marking. To me, it is especially interesting to see that some of these destinctions are marked in both the "subject" and "object" slots, as opposed to the reflexives and reciprocals which are marked exclusively in the "object" slot on the verb. Take a look especially at Hupa, Navajo, and Tlingit. Eyak looks like it may do something similar skimmig though wikipedia but it's less clear.

In this grammar of Saanich it gives a prefix xʷ- which it calls a "locative" but is actually a verbal prefix which appears to be doing the same thing as the Na-Dené examples you give.