(Conlangs) Q&A Thread - Quick questions go here

[Creyeditor]

My Celtic/Greenlandic language (before any contact with Greenlandic) would have a very complicated pluralization system (as does modern Irish). After all, there are like 12 declension patterns in Old Irish, and some of them don't even have a consistent pluralizer anyway!

However, the definite article always causes predictable mutations in the plural on most nouns:

• The plural of guedhe is guedhi; with the definite article, these are an chuedhe and nna gguedhe.
• The plural of cbhand is cbhanda; with the definite article, these are an chbhand and nna ccbhanda.
• The plural of ríochain is ríognoi; with the definite article, these are end ríochain and nne rríognoi.

Would it be plausible, then, for the plurals to regularize ─ or become nonexistent ─ and have the inflections of the article and the mutations they cause be the sole way of indicating number?

That looks naturalistic to me.

Okay, I have another question, this time about pitch accent.

Can a pitch accent language exhibit tone sandhi and assimilation like tone plateaus, or Meeusen's rule?

Yes

[Arayaz]

My Celtic/Greenlandic language (before any contact with Greenlandic) would have a very complicated pluralization system (as does modern Irish). After all, there are like 12 declension patterns in Old Irish, and some of them don't even have a consistent pluralizer anyway!

However, the definite article always causes predictable mutations in the plural on most nouns:

• The plural of guedhe is guedhi; with the definite article, these are an chuedhe and nna gguedhe.
• The plural of cbhand is cbhanda; with the definite article, these are an chbhand and nna ccbhanda.
• The plural of ríochain is ríognoi; with the definite article, these are end ríochain and nne rríognoi.

Would it be plausible, then, for the plurals to regularize ─ or become nonexistent ─ and have the inflections of the article and the mutations they cause be the sole way of indicating number?

That looks naturalistic to me.

All right, thank you! Would you expect the plural inflections to remain at all? Perhaps just on common words?

[Creyeditor]

I might expect the mutations to become analogized to all plural forms, such that mutation becomes the proper plural marking.

[Ahzoh]

Would it make sense for a subject/agent noun to take the nominative case if the verb is stative (transitive or intransitive) and the ergative case if the verb is dynamic (transitive or intransitive)? And would it make sense for inanimates to possess nom-acc syncretism but maintain a separate ergative case? My rationale is that the ergative case is also possess instrumental/secondary object and ablative functions.

[WeepingElf]

Would it make sense for a subject/agent noun to take the nominative case if the verb is stative (transitive or intransitive) and the ergative case if the verb is dynamic (transitive or intransitive)? And would it make sense for inanimates to possess nom-acc syncretism but maintain a separate ergative case? My rationale is that the ergative case is also possess instrumental/secondary object and ablative functions.

As I said on the other board, this makes enough sense to me at least to pursue in a conlang, which is what you are asking about. No, it is the wrong way round - I had somehow failed to notice.

[Salmoneus]

Would it make sense for a subject/agent noun to take the nominative case if the verb is stative (transitive or intransitive) and the ergative case if the verb is dynamic (transitive or intransitive)?

I have a problem here, because by definition statives can't be transitive. [wikipedia does use the example '[be able to] play the piano', but I think that's something different, and it isn't covered by wikipedia's own lists of categories]. They can be bivalent in some languages (like "I hear the clock" - it's not transitive because the clock is not actually affected in any way, in in many languages you'd have to say the equivalent of "I hear [from/of] the clock").

It's certainly not unusual to mark transitivity through differential case marking. Usually there'd be the option to do this with intransitive forms of dynamics as well - so "I have eaten" would take different marking from "I have eaten [it]". But maybe your language just doesn't allow detransitivisation without conversion into a stative (i.e. there are no dynamic intransitives in it).

Differential marking using two different forms of non-absolutive (rather than just erg-abs) seems stranger, but I suppose plausible. After all, quirky subject is a thing.

I'm not convinced you have an ergative, though. If, with dynamic verbs, the "ergative" is used for the subject of intransitive verbs and the agent of transitive verbs... that's the nominative case, surely?

And would it make sense for inanimates to possess nom-acc syncretism but maintain a separate ergative case? My rationale is that the ergative case is also possess instrumental/secondary object and ablative functions.

So let me get this straight with examples:

1. the water is cold: water-CASE1 colds

2. the man understands: man-CASE2 understands

3. the water surrounds the cat: water-CASE1 surrounds cat-CASE1

4. the man likes the dog: man-CASE2 likes dog-CASE1

5. the rock falls: rock-CASE3 falls

6. the dog barks: dog-CASE3 barks

7. the rock crushes the cabbage: rock-CASE3 crushes cabbage-CASE1

8. the cat bites the dog: cat-CASE3 bits dog-CASE1


This doesn't seem inconceivable to me. But I do think you've mislabelled your cases!

CASE1 here is obviously an accusative, and CASE3 is clearly a nominative. The only wrinkle there is that inanimate subjects take the accusative in low-transitivity (stative) contexts, which seems very believable. [both low-animacy subjects and low-transitivity verbs will tend to 'de-subjectivise' the subject - make it less like a prototypical subject]

The odd bit is CASE2. This case is only used with animate subjects of stative verbs. That maybe makes sense, specifically because it's paradoxical: animate subjects are high-transitivity but stative verbs are low-transitivity, so it's believable that a language would find the juxtaposition odd and not be willing to use the normal nominative in these instances.

In isolation, without looking at the wider case system, I would call CASE2 the "dative", because its semantics are fairly dative and the dative case is often used in exactly this context in other languages (eg German, Spanish).

A follow-up question would be whether dative subjects in this language are always true syntactic subjects (as is usually argued for Icelandic), or whether they are actually indirect objects treated peculiarly with implied subjects (as is traditionally argued for Spanish, "me gusta", etc), which would depend on a wider range of behaviours.

[Ahzoh]

I have a problem here, because by definition statives can't be transitive. [wikipedia does use the example '[be able to] play the piano', but I think that's something different, and it isn't covered by wikipedia's own lists of categories]. They can be bivalent in some languages (like "I hear the clock" - it's not transitive because the clock is not actually affected in any way, in in many languages you'd have to say the equivalent of "I hear [from/of] the clock").

I've heard that verbs like "understand", and "know" are stative transitive verbs. At least in English they are unable to take the inchoative aspect (You cannot say "I am knowing the answer").

But forcing all "transitive" stative verbs to be intransitive would certainly make it easy for me to just decide that the majority of verbs in Vrkhazhian are monotransitive.

[Arayaz]

I have a problem here, because by definition statives can't be transitive. [wikipedia does use the example '[be able to] play the piano', but I think that's something different, and it isn't covered by wikipedia's own lists of categories]. They can be bivalent in some languages (like "I hear the clock" - it's not transitive because the clock is not actually affected in any way, in in many languages you'd have to say the equivalent of "I hear [from/of] the clock").

I've heard that verbs like "understand", and "know" are stative transitive verbs. At least in English they are unable to take the inchoative aspect (You cannot say "I am knowing the answer".)

That's not inchoative, that's continuous. Inchoative would be "I begin to know the answer." Weird, but not ungrammatical, especially if you say "I am beginning to know the answer."

[Ahzoh]

Right, oh well, I was just thinking of this and mixed it up with the section above.
https://en.wikipedia.org/wiki/Stative_v ... s_analysis

[Salmoneus]

I have a problem here, because by definition statives can't be transitive. [wikipedia does use the example '[be able to] play the piano', but I think that's something different, and it isn't covered by wikipedia's own lists of categories]. They can be bivalent in some languages (like "I hear the clock" - it's not transitive because the clock is not actually affected in any way, in in many languages you'd have to say the equivalent of "I hear [from/of] the clock").

I've heard that verbs like "understand", and "know" are stative transitive verbs. At least in English they are unable to take the inchoative aspect (You cannot say "I am knowing the answer").

They're stative verbs, but they're not transitive - the "agent" is not instigating the "action", there is no action, and the "patient" is not affected by the "action". If anything, the opposite is true: if you know the answer, it's really the answer acting upon you and making you different, while the answer itself is unaffected by your knowledge of it.

Unless you're just using "transitive" as a synonym of "bivalent" or "dyadic". But that would be a wasted opportunity, since the subject you're really asking about here - with dynamic vs stative, animate vs inanimate, etc - is transitivity.

[Ahzoh]

A follow-up question would be whether dative subjects in this language are always true syntactic subjects (as is usually argued for Icelandic), or whether they are actually indirect objects treated peculiarly with implied subjects (as is traditionally argued for Spanish, "me gusta", etc), which would depend on a wider range of behaviours.

Well, since Vrkhazhian is a secundative language, Case3 would also mark the secondary object so that man-Case3 gives dog-Case1 a bone-Case3.

Unless you're just using "transitive" as a synonym of "bivalent" or "dyadic"

Yes. Only understood transitivity as "can it take objects/non-subject arguments?", which I guess is more in the realm of valency and not "agentiveness"

[Omzinesý]

Transitivity is a messy term.

Semantic transitivity: an agent somehow changes the patient.
Syntactic transitivity: the verb has a direct object.

The two are of course linked, because the syntactic transitive construction is defined as the one coding a prototypical transitive semantic action, basically killing.

I understood very well that Ahzoh meant syntactic transitivity.

[LinguoFranco]

So, on the page for Polysynthesis, Wikipedia lists two types of polysynthetic languages: affixal and compositional.

How accurate is this? Is there any overlap between the two in natlangs?

[Omzinesý]

So, on the page for Polysynthesis, Wikipedia lists two types of polysynthetic languages: affixal and compositional.

How accurate is this? Is there any overlap between the two in natlangs?

Read "A structural typology of polysynthesis' Johanna Mattissen, available on internet. I found it illuminating.

Typology is always typology. It's not right or wrong. It's just theory. But I think the distinction is useful. And of course languages can be both.

[VaptuantaDoi]

So, on the page for Polysynthesis, Wikipedia lists two types of polysynthetic languages: affixal and compositional.

How accurate is this? Is there any overlap between the two in natlangs?

If you want a good conlang-specific intro to polysynthesis, there's this guide from the old ZBB. As for Wikipedia's two types, they definitely exist, although I don't know if it's a useful distinction to make. I guess all poly(synthetic)langs are somewhat affixal – they all have some bound morphemes (I guess it would be possible to make a polylang without bound morphemes, but it would be weird).

[Solarius]

I'm working on a diachronic conlang with a noun morphology based on this sketch, and am struggling with an issue.

The ancestor language, Deligic, marks the singulative form with -g, even when doing so creates fairly intense consonant clusters. This isn't an issue as Deligic allows these. However, the descendant language, Wikay simplifies clusters, so that -g is deleted after stops and ejectives, while preserved elsewhere.

Deligic: /déld/ "ice" --> /déldg/ "piece of ice"
Wikay: /d͡ʒǽld/ "ice" --> /d͡ʒǽld/ "piece of ice"

The trouble is, I really like -g, especially how weird and cludgy it is after stops or consonant clusters. So I'm thinking of reintroducing it in these environments via analogy. Which option sounds best?
1. Have it be reintroduced by analogy, regardless of the new consonant cluster rules.
2. Have it be reintroduced by analogy, with an epenthetic vowel [ə]. However, /ə/ is a pretty marginal phoneme elsewhere in the language, only appearing in loanwords and as an underspecified offglide in diphthongs (and not for all speakers), so this might be implausible.
3. Let it stay deleted.

There are similar issues for most other affixes in the nominal paradigm, but I don't have the same fondness for them, and this rule doesn't lead to a loss in grammatical distinctions. Thus, I'm leaning towards letting them stay deleted, but should I try to bring them in line with whatever I do with the singulative?

[Omzinesý]

The trouble is, I really like -g, especially how weird and cludgy it is after stops or consonant clusters. So I'm thinking of reintroducing it in these environments via analogy. Which option sounds best?
1. Have it be reintroduced by analogy, regardless of the new consonant cluster rules.
2. Have it be reintroduced by analogy, with an epenthetic vowel [ə]. However, /ə/ is a pretty marginal phoneme elsewhere in the language, only appearing in loanwords and as an underspecified offglide in diphthongs (and not for all speakers), so this might be implausible.
3. Let it stay deleted.

Of course the answer is whatever you like the most. I understood that you like to preserve the marker. I also understood that you don't like to have those complex clusters. So, it's probably 2. but does the epenthetic vowel have to be /ə/ if it's not a native vowel? What about having, say, -ag instead?

[Dormouse559]

The trouble is, I really like -g, especially how weird and cludgy it is after stops or consonant clusters. So I'm thinking of reintroducing it in these environments via analogy. Which option sounds best?
1. Have it be reintroduced by analogy, regardless of the new consonant cluster rules.
2. Have it be reintroduced by analogy, with an epenthetic vowel [ə]. However, /ə/ is a pretty marginal phoneme elsewhere in the language, only appearing in loanwords and as an underspecified offglide in diphthongs (and not for all speakers), so this might be implausible.
3. Let it stay deleted.

A couple more possibilities:
4. Reanalyze the suffix based on vowel-final nouns. Maybe a common singulative target is elda; it notably contains a cluster that would drop /g/ word-finally. Its singulative elda-g gets reinterpreted as eld-ag, with the ending generalizing to nouns that have dropped /g/. That won’t preserve the consonant clusters, but it’s no worse than 3 in that respect.
5. Simplify singulative clusters differently. This is similar to 1, except that in an additional step, the cluster with restored /g/ is simplified again while preserving /g/ (e.g. /déldg/ > /d͡ʒǽld/ > /d͡ʒǽldg/ > /d͡ʒǽlg/). You could also reduce the number of steps by just saying the singulative suffix resists elision and forces whatever other sound changes will preserve the phonotactics. That seems to create irregularity, depending on how you feel about that.

I like the idea of picking one major strategy that most nouns use, and then having a few other nouns use others, just to keep L2 speakers and children on their toes

[Creyeditor]

I really like Dormouse's idea about exceptional cluster resolution.

Another (synchronic) possibility: in some languages certain affixes can violate the cluster restrictions that apply in other contexts. In English for example the past tense suffix <-ed> can result in a coda cluster of voiced plosives, e.g. <dubbed>, something that is not found in simple forms (or any other AFAIK), except in the word <smaragd. Similarly, German allows coda clusters as complex as <Herbsts> and <Arzts> only in genitive forms of nouns. No simple noun or other word could end in such a consonant cluster.
So you might actually say, complex clusters ending in /g/ are not allowed, except if the clusters includes the singulative -g.

[Salmoneus]

Are these really example of certain forms being exempt from phonotactic rules, or are they just examples of new clusters being formed through sound change after others were lost?

In English at least, the problematic suffixes - -ed, -es - are formed by elision of schwa, subsequent to whatever process prevented/eliminated such clusters being formed in the roots they were attached to.