Three Glossing Questions

If you're new to these arts, this is the place to ask "stupid" questions and get directions!
Post Reply
User avatar
collect_gluesticks
rupestrian
rupestrian
Posts: 15
Joined: 27 Nov 2021 00:49

Three Glossing Questions

Post by collect_gluesticks »

I am translating some example sentences in my conlang, to demonstrate how the grammar works. But I had a few questions about interlinear glosses.

1. My conlang has an enclitic, -më that marks a verb argument as an agent (as opposed to patient). This is applied to the subject in transitive sentences, and optionally in intransitive sentences (where it indicates volition). Examples:

tswáhna rómë kentsói
drink 1sg-AGENT water
"I drink water."

kyóngbo anái
lose team
"The team loses"

kyóngbo anáimë
lose team-AGENT
"The team throws the game"

What glossing abbreviation would I use for this? I thought it was AGT, but some other conlangers told me that was for agent nominalizers. Is there a term for the grammatical category of "agent marker" and is there an abbreviation for it? Would I just use ERG? (I was thinking that ERG would be inaccurate, since the marker's function is encoding volition rather than the syntactical relationship between the verb and the argument.)

2. I'm also have a head marking posession affix -nem

oikátenem ro
shelter-POSESSED 1SG
"my house"

I'm also confused as to what abbreviation to use there. I assumed POSSD (possessed case) would be the right answer, but someone told me that this isn't really a "case" since, in their words "because the possessed noun can have either agentive or non-agentive case." I'm not sure if I 100% follow, but I also suspect that POSSD is the wrong term to use here.

I have seen examples of head marking possession being abbreviated as POSS in the gloss. Wouldn't that be confusing? I've also been told it resembles a construct state (CNS, CONSTR, CNSTR)... but isn't there more to the construct state than this? What is your preferred term and abbreviation here?

3. Last question: My conlang has 3 pronouns, all singular by default, and plurality is achieved by using quantifiers ("tai rah", you two) or conjunctions ("ro na ne", I and him/her/them ). Since there are no plural pronouns, is it ok to omit SG from the gloss? So rather than glossing as 1SG, could I just gloss it as "1"? Do you think that would be confusing or not?

As you can see, I'm hung up on some little details. So I apologize for the long post, and thank you in advance. Any advice in this area is welcome!
User avatar
Creyeditor
MVP
MVP
Posts: 4603
Joined: 14 Aug 2012 19:32

Re: Three Glossing Questions

Post by Creyeditor »

First things first, glossing abbreviations can vary a lot and be shorter or longer, depending on various factors. The important part is that glossed sentences ideally come with some explanation of the relevant bits.

1. You could just use A for agent and that would be fine with me. Alternatively, you could use VOL to make sure everybody gets the volition part. Also, clitics are often marked with an equal sign (=) instead of a hyphen.

2. POSSD is the most correct gloss, IMHO, even is this is not a case and just marks a possessed noun. It is a bit long though, so I can totally see why people use POSS.

3. You do not need to mark singular here. You could mark it in square brackets if neccessary. Balinese pronouns are often glossed as 1 and 3 (second person pronouns distinguish gender). I don't think it looks to confusing.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
Salmoneus
MVP
MVP
Posts: 2674
Joined: 19 Sep 2011 19:37

Re: Three Glossing Questions

Post by Salmoneus »

First off: I wouldn't worry about it too much. It's not a bad thing to try to use standard terminology, not only because it helps make things clearer for readers but also because trying to fully understand the standard terminology often results in improving your understanding of the phenomena the terminology describes*; but at the same time, we should remember that the real objective here is only being understood, and clarity can be more important for that than technical accuracy. And sometimes you may want to change your terminology to suit the context - what you might say when describing a distinction within the language might not be what you'd say if you were trying to draw attention to a similarity, or difference, with another language, for example. Having said that...
collect_gluesticks wrote: 08 Apr 2022 02:51 I am translating some example sentences in my conlang, to demonstrate how the grammar works. But I had a few questions about interlinear glosses.

1. My conlang has an enclitic, -më that marks a verb argument as an agent (as opposed to patient). This is applied to the subject in transitive sentences, and optionally in intransitive sentences (where it indicates volition).

What glossing abbreviation would I use for this? I thought it was AGT, but some other conlangers told me that was for agent nominalizers. Is there a term for the grammatical category of "agent marker" and is there an abbreviation for it? Would I just use ERG? (I was thinking that ERG would be inaccurate, since the marker's function is encoding volition rather than the syntactical relationship between the verb and the argument.)
'AGT' is apparently right for the agentive case, so you can use that. But the others are correct that that will often be assumed to mean an agentive nominalisation. So it depends on how clear you feel you're being, and how important it is to avoid misunderstanding. [for instance, I might use 'AGT' in the grammar of the language, where it's clear in context what it means (and it's explained in a note somewhere anyway), while avoiding it for, say, a one-line translation on this forum, where you don't really want to have to explain your notation from scratch]

Conversely, 'ERG' is technically 'wrong', I guess, because an agentive isn't a true ergative (a 'true' ergative never governs monadic subjects). However, an agentive is very, very similar to an ergative (eg it's usually marked, unlike a nominative), and the distinction arguably isn't important, particularly because a given language is unlikely to have both cases. So I'd be happy just calling it 'ERG', even though it's not 100% accurate. Unless, of course, you're creating an example explicitly to discuss the active-stative alignment, in which case it too could be confusing...

----------------

But on the specific question of what 'agentive' and 'ergative' actually refer to: they're more similar than you think. Although it's convenient to think of core cases as being purely 'syntactic', they're actually also semantic: 'syntactic' functions like "this means it's the subject" or "this means it's a core argument of the verb" rely on notions like 'subject' and 'core argument' which, despite the wishes of the formalists, turn out to actually be slightly difficult to define abstractly and cross-linguistically, and their actual function in the language is often tied up in semantic distinctions.

Specifically, the key here (as it often is!) is the semantic concept of 'transitivity', as distinct from the syntactic concept of valency. As regards transitivity, the ergative and agentive are the same: both mark the subjects of transitive verbs but not the subjects of intransitive verbs (making them distinct from the nominative, which marks the subjects of all (in theory!) verbs). The difference between the two is really found elsewhere in the language: the case is termed 'agentive' (and the alignment 'active-stative') if the language recognises the possibility of bivalent intransitive verbs, and 'ergative' if it does not. [volition being one of the key factors in determining transitivity]. But this isn't really as binary a distinction as people sometimes make out, as in practice many 'ergative-absolutive' or 'nominative-accusative' languages DO systematically distinguish a class of bivalent intransitive constructions - they just set the threshold of transitivity lower than a classically 'active-stative' language does. [Germanic languages, for example, generally have inverted case assignment for verbs of perception ("I see it", even though semantically 'it' is the one acting on 'I' rather than vice versa) and in 'quirky case' constructions in which the patient can be denied the accusative, the subject can be denied the nominative, or both - both phenomena essentially mark low transitivity]

['transitivity' is a vague concept that is applied differently in different languages. But broadly (and this isn't an official definition I'm quoting so I might be forgetting something), an action is more transitive when an animate, distinct agent voluntarily, definitely, completely and successfully acts upon, and causes a change within, a distinct and specific, specified, real, inanimate, unwilling, uncolluding and unreciprocating patient. Transitive actions are often conveyed with bivalent verbs, but logically it is possible to have both univalent transitive verbs and bivalent intransitive verbs. [some people also use the terms 'monadic' and 'dyadic', instead of discussing valency, but it's not always clear whether these are syntactic or semantic terms]. Some languages explicitly mark transitivity, but most do not. However, it is common for languages to non-explicitly mark transitivity through the use of distinct constructions, at least in extreme cases, and this is often (at least part of) what split alignment and split ergativity are doing. The concept of transitivity is also tied up with the concept of animacy (potential to be the subject of a transitive verb) and of subjecthood itself]

Buuut now I'm on a tangent...
2. I'm also have a head marking posession affix -nem

oikátenem ro
shelter-POSESSED 1SG
"my house"

I'm also confused as to what abbreviation to use there. I assumed POSSD (possessed case) would be the right answer, but someone told me that this isn't really a "case" since, in their words "because the possessed noun can have either agentive or non-agentive case." I'm not sure if I 100% follow, but I also suspect that POSSD is the wrong term to use here.

I have seen examples of head marking possession being abbreviated as POSS in the gloss. Wouldn't that be confusing? I've also been told it resembles a construct state (CNS, CONSTR, CNSTR)... but isn't there more to the construct state than this? What is your preferred term and abbreviation here?
There needn't be more to the construct state than that. Although there can be. 'Construct state' is often associated also with the neutralisation of a definiteness distinction (which is why very occasionally people talk about a construct state in Celtic languages), but it needn't be.

The distinction your other correspondant is making is that possession marking 'stacks' with case, in a way that cases don't usually do. So, "I walked my friend's dog" vs "my friend's dog ate a fish". Whereas conceptually, and iirc in at least some languages, it's possible to instead have a genuine 'possessed case' that doesn't stack (so either syntactic cases must be shifted to the possessor, or such nouns are simply restricted to a certain role - so, for instance, a language might make it impossble to say "my friend's dog ate a fish" because it requires possessed nouns to be patients, and instead make you say "my friend has a dog that ate a fish" or "my friend ate a fish using his dog" or the like).

I wouldn't worry about it. I guess I'd talk about a construct state if the language resembled Afro-asiatic, or if possessed nouns were distinct in other ways, or if the same form were used more widely for other constructions as well (eg iirc some languages have the same form for possessed nouns and for topicalised subjects). Also maybe if it were zero-marked relative to a non-zero non-possessed state? Otherwise I'd just call it 'possession case' or 'possessed case'.
3. Last question: My conlang has 3 pronouns, all singular by default, and plurality is achieved by using quantifiers ("tai rah", you two) or conjunctions ("ro na ne", I and him/her/them ). Since there are no plural pronouns, is it ok to omit SG from the gloss? So rather than glossing as 1SG, could I just gloss it as "1"? Do you think that would be confusing or not?
No, that's fine. You don't need to gloss all the things that a form is assumed to mean by default (that would be an infinite list!), only the things it actively marks.
User avatar
Creyeditor
MVP
MVP
Posts: 4603
Joined: 14 Aug 2012 19:32

Re: Three Glossing Questions

Post by Creyeditor »

Just an idea I had. Why not make a compromise and call the category of possessed nouns a possessed state? This makes it clearer that the possessed noun suffix -nem can potentially be combined with the =më agentive clitic. Btw, just out of curiosity, does this work in your conlang? Something like one of the following (depending on the position of your case clitics).

Kyóngbo anái-nem ro=më.
lose team-POSSD 1-A
My team throws the game.

or

Kyóngbo anái-nem=më ro.
lose team-POSSD-A 1
My team throws the game.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
User avatar
collect_gluesticks
rupestrian
rupestrian
Posts: 15
Joined: 27 Nov 2021 00:49

Re: Three Glossing Questions

Post by collect_gluesticks »

Thank you both so much for the advice. As you could probably tell, I haven't done a lot of glossing before, and it's so helpful to get feedback from someone more experienced. I am definitely incorporating your feedback into my next revision.

Salmoneus - I appreciate your toughts on transitivity. I had gone down a similar rabbit hole just a few months ago. I used to think of alignment as just different strategies of distinguishing A from P. But it looks like, in some languages, what we call "alignment" is really a complex interaction between concepts like animacy, transitivity, and volition, which also happen to disambugate verb arguments. I wish I was studied enough to say more on this subject. But I will just say that I find it fascinating.

Creyeditor - Your two examples highlight some ambiguity with the =më clitic. Here's another example of that, involving a prepositional phrase:

otowáe tswumë́ny rah tóre rómë

sleep friend two in_front_of-AN 1-AGT

the two friends in front of me went to sleep

The issue is, I want the clitic to come at the end of a noun phrase, but it's possible for that end to be a different noun.

I'm debating whether to restrict =më further to avoid such confusion. I know that all natural languages have ambiguity like that (english: "the queen of england's crown"). I am thinking of my conlang as having an "ambiguity budget" - a little ambiguity is necessary to be natural, but too much could be confusing or impractical. Since my conlang is still a work in progress, I will decide later whether to restrict =më
Post Reply