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

[HolyHandGrenade!]

I think the most important ones to decide are degree of synthesis, any kind of mandatory inflection (ie there is no neutral form), head/dependent marking, and which word classes your language has a strong tendency towards.

[HolyHandGrenade!]

My protolang had a 3-way distinction between dual, paucal, & plural. In the modern language, this shifted: dual>paucal, paucal>plural, & plural>superplural. I thought it would be cool if the plural pronouns somehow gained a 4-way distinction: dual, paucal, plural & superplural, but I’m not sure exactly how that would happen. Any ideas? I do have a suffix for associative plural so that could be something.

[Khemehekis]

Is superplural a numerus that means something like "more", as in "more groceries"?

[Arayaz]

Is superplural a numerus that means something like "more", as in "more groceries"?

No. A superplural means, roughly, "a lot of." "More" is a comparative.

Also, that was your 4,444th post!

[Davush]

I have a question about participles. Recently I've been wanting to make a language where participles 'take over' and become a major part of the verbal system, but this leads to the question: where do participles come from?

The obvious answer is adjectivising affixes of some kind, but more specifically, if I want an active (or agent-oriented) participle and a passive (or patient-oriented) participle, which adjectivisers would lead to which? The languages that have participles like this seem to have had them forever so there is no clear answer there.

For example, PIE -nt- seems to mostly make active participles in its descendants, while -to makes passive participles – what was it about -nt- that lent it an 'active' reading rather than passive (and vice versa for -to)? Does the distinction have something to do with the semantics of the suffixes (perhaps at an earlier stage), or did they just 'specialise' arbitrarily since both also seem to broadly be adjectivisers/nominalisers? This is more of a general question not a debate on PIE in particular: if participles are formed by de-verbal adjectivising suffixes, how could this end up in an active/passive distinction (or can it be/is it likely that different affixes just specialise arbitrarily)?

Any thoughts welcome!

[Creyeditor]

Participles can also come from nominalized forms. In English you can have forms that are relatively neutral. 'I was in a beating.' can mean I was beaten, I beat s.o. or both.
Your participles could also extend their function from an attributive (relative clause-like) function to being combined with verbs or verb-like words. So 'the eating man' could lead to 'I am eating'. And 'the (being-)eaten man' could lead to 'I am (being) eaten'. Hope this helps.

[Davush]

Participles can also come from nominalized forms. In English you can have forms that are relatively neutral. 'I was in a beating.' can mean I was beaten, I beat s.o. or both.
Your participles could also extend their function from an attributive (relative clause-like) function to being combined with verbs or verb-like words. So 'the eating man' could lead to 'I am eating'. And 'the (being-)eaten man' could lead to 'I am (being) eaten'. Hope this helps.

Thank you, Creyeditor. I suppose my question was more about the specific morphemes that are used to form participles - where do they come from? How does one morpheme become associated with active and the other passive participles (e.g., PIE -nt- Vs -to).

The pathway from participle > 'fully finite' verb is clearer as this has observably happened a lot. It's the actual formation of the participle that is more obscure and what I'm asking about. I understand that nominalisation/adjectivisation is one way but my question is more fundamental, I suppose.

[Creyeditor]

Hmm, I am not sure that I get your question. Is it about orientation? Okay, let's take Finnish -ttu as an example. According to Wiktionary it derives from a combination of the passive suffix -tta and a category changing vocalic suffix in Proto-Finnic, probably either the nominalizer -u or the adjectivizer -ut. Does that help? Finnish also has an active participle suffix -va which seems to derive from a more general adjectivizer -pa in Proto-Finnic.
Then there is the Arabic prefix mu-, which according to Wiktionary derives from a relative pronoun, such that a mudarris 'teacher' is a contraction of man yudarris '(he) who teaches'. This also means that similar prefixes occur in differently oriented participles.
I am still not sure if this answers your question any better.

[Davush]

Hmm, I am not sure that I get your question. Is it about orientation? Okay, let's take Finnish -ttu as an example. According to Wiktionary it derives from a combination of the passive suffix -tta and a category changing vocalic suffix in Proto-Finnic, probably either the nominalizer -u or the adjectivizer -ut. Does that help? Finnish also has an active participle suffix -va which seems to derive from a more general adjectivizer -pa in Proto-Finnic.
Then there is the Arabic prefix mu-, which according to Wiktionary derives from a relative pronoun, such that a mudarris 'teacher' is a contraction of man yudarris '(he) who teaches'. This also means that similar prefixes occur in differently oriented participles.
I am still not sure if this answers your question any better.

Thanks and apologies if my question was confusing. Put more simply: if language A has no participles, how might language B (the descendant of language A) acquire participles (language internally, not via borrowing)? Specifically, how might it acquire an active/passive participial distinction?

As has been mentioned, adjectivisers/nominalisers seem to be common pathway to participles *but* it is not clear what morphemes tend to become associated with 'active' participles, and what become associated with 'passive'.

In the Finnish example, -ttu contains an already explicit passive morpheme (-tt-) so that's pretty clear cut as to how it ends up in the passive participle (although it begs the question of what -pa originally meant and how it differed from -ut). But what about IE *-nt- vs. *-to, where could these come from? Obviously it's unknowable, so I'm basically asking for ideas/speculation here. Similarly, the semitic active participles in CaaCiC differ very much from the passive (ma)C(a)Cuu/iiC. Both CaaCiC and C(a)Cuu/iiC are likely originally 'adjectival', but we can't go much further back than that - nonetheless, they must have originated somewhere.

I can imagine adjectivising morphemes might come from words such as 'like, having, containing', genitive formations, and probably lots of other things, but to me, none of these immediately suggest an orientation so my further question is are there any clues as to *which types* of adjectivising/nominalising morphemes end up being used in active vs. passive participles? Or basically, what could the 'etymology' of active vs. passive participle markers (via adjectivisers) be, going back to the earliest stages?

The Finnish example is that it contains a (maybe fossilised?) passive marker, but the Semitic and IE forms do not. Different adjectival formations have clearly 'specialised' towards an orientation when being used in participles.

To use a made-up example:

Let's say proto-lang has adjectival morphemes -ma and -ka. These are 'generic' adjectivisers at this stage in the language. The question is: why does -ma end up forming active participles, but -ka passive participles? Is it something to do with the original (now bleached) semantics of these morphemes, perhaps before they were grammaticalised as affixes? Or could it really just be (more or less) arbitrary in that speakers simply 'decided' -ma was active participles, and -ka was for passive participles (despite both suffixes just being abstractly 'adjectivisers)?


Or an even more abstract question: is it common for morphemes to more or less abstractly 'specialise' into different grammatical functions (such as having multiple generic adjectivisers > active/passive participle markers) with no particular reference to their original semantics or usage?

(Sorry if this has made my question even more confusing...!)

[Creyeditor]

I think the question might actually be close to impossible to answer. It seems that participles with a fixed orientation (i.e. active vs. passive) are mostly found in language family where they are old enought to be difficult to trace back. I look at the Shagal book on participles and it seems that not coding orientation is more frequent than expected. 42 languages in her sample contrast for orientation and 52 languages do not contrast orientation.

But let's speculate, I mean that's why we are conlangers. Imagine we start with a language with relative clauses and no participles. Relative pronouns are marked for case. Relative clauses can occur without any noun that they modify. Pronominal subjects can also be dropped. This gives us sentences like the following.

tulpa re-nom porto
plug REL-NOM carry
'the plug that carries it'

tulpa re-acc porto
plug REL-ACC carry
'the plug that it carries'

Now, imagine the relative pronouns get reduced and attach to the verb forms, et voila, we have a participle system with orientation.

tulpa röm-porto
plug PTCP.ACT-carry
'the carrying plug'

tulpa räc-porto
plug PTCP.PASS-carry
'the carried plug '

If we think of adjectivizers, we would probably have to think about their origin, too. I mean, what is the difference between a deverbal adjective and a participle anyway.

[Omzinesý]

Haspelmath has a classical article that sais that the present subject-oriented and perfect object-oriented participles are the most common. Perfect emphasizes results and results appear in the object. Present emphasizes action and it's done by the subject.

(Citation to Haspelmath ends and interpretationstarts.) So, if participles get oriented, say because of areal influence, the perfect participle acquires the object-orientation and the present participle acquires the subject-orientation.

Finnish system (simplified)

Subject Object
Present -va -tta-va
Past -nut -ttu

You see that the Finnish present object-oriented participle is a combo of two morphemes, the present (subject-oriented) participle and the pasive -tt.

[Davush]

If we think of adjectivizers, we would probably have to think about their origin, too. I mean, what is the difference between a deverbal adjective and a participle anyway.

Yes indeed - this was essentially my question: if there are multiple ways of forming deverbal adjectives (which aren't initially specified for orientation), can/would speakers simply 'arbitrarily' start specialising one form for active, and one form for passive (despite them initially not having orientation).

You see that the Finnish present object-oriented participle is a combo of two morphemes, the present (subject-oriented) participle and the pasive -tt.

I think the Finnish example is more clear cut in that the passive morpheme is retained in the passive (object-oriented) forms, my question was more about cases such as IE and Semitic, where there is no clear passive morpheme, but it is interesting nonetheless. Perhaps adjectival formations that are more about a resultant 'state' might tends towards becoming a passive participle as you mention perfect(ive) emphasises results (and by extension, the object).

So we could imagine something like:

An adjectival morpheme used to mean 'having/in the state of' > passive participle. While a morpheme used for something more like 'the one who XYZ' > active participle.

---

Anyway, a semi-related question. I had a fun idea, but I'm not sure if this is attested/at all plausible:

Let's say the language has a relative morpheme *tV-.

tira ti-haki 'the man who speaks'

The t-, however, is reinterpreted as an 'agreement marker' matching the initial consonant of the modified noun so that eventually the prefix is CV- where C = the first consonant of the noun:

tira ti-haki
kuta ki-haki
sakra si-haki
etc.

This would then spread to adjectives for a kind of bizarre agreement system which is marked by the initial consonant of the noun appearing on adjectives.

[DesEsseintes]

Greetings, O Master of Jugs upon Platforms!

I like a lot of what’s going on in the thread on your latest lang. I also deeply sympathise with the deep resentment you must be harbouring towards your word forms, which, albeit lovely, will inevitably stab at the heart of their creator.

Anyway, a semi-related question. I had a fun idea, but I'm not sure if this is attested/at all plausible:

Let's say the language has a relative morpheme *tV-.

tira ti-haki 'the man who speaks'

The t-, however, is reinterpreted as an 'agreement marker' matching the initial consonant of the modified noun so that eventually the prefix is CV- where C = the first consonant of the noun:

tira ti-haki
kuta ki-haki
sakra si-haki
etc.

This would then spread to adjectives for a kind of bizarre agreement system which is marked by the initial consonant of the noun appearing on adjectives.

According to Corbett, agreement systems based purely on a formal component (phonological or prosodic) are not attested. All known natlang agreement systems have a semantic component.

[Davush]

Greetings, O Master of Jugs upon Platforms!

I like a lot of what’s going on in the thread on your latest lang. I also deeply sympathise with the deep resentment you must be harbouring towards your word forms, which, albeit lovely, will inevitably stab at the heart of their creator.

According to Corbett, agreement systems based purely on a formal component (phonological or prosodic) are not attested. All known natlang agreement systems have a semantic component.

Thanks! I think this was floating around in my head somewhere hence the uncertainty. At least I can throw that idea out now.

[Khemehekis]

Also, that was your 4,444th post!

Coolness! I didn't even notice!

[Nel Fie]

Anyway, a semi-related question. I had a fun idea, but I'm not sure if this is attested/at all plausible:

Let's say the language has a relative morpheme *tV-.

tira ti-haki 'the man who speaks'

The t-, however, is reinterpreted as an 'agreement marker' matching the initial consonant of the modified noun so that eventually the prefix is CV- where C = the first consonant of the noun:

tira ti-haki
kuta ki-haki
sakra si-haki
etc.

This would then spread to adjectives for a kind of bizarre agreement system which is marked by the initial consonant of the noun appearing on adjectives.

Unless I'm mismatching two unrelated ideas, and at risk of throwing in a wrench after you already replied to DesEsseintes, something like this seems to actually exist in natlangs, although rare. It's called "alliterative agreement", and attested to various degrees in different families, though Bantu languages seem to feature it most prominently.

More information here:

https://en.wikipedia.org/wiki/Agreement ... _agreement
http://glottopedia.org/index.php/Alliterative_agreement

And as the links mention, Corbett has a book on agreement with a chapter about it.

According to Corbett, agreement systems based purely on a formal component (phonological or prosodic) are not attested. All known natlang agreement systems have a semantic component.

I'm not sure what to make of it, but Corbett's own book, the Cambridge Textbook in Linguistics "Agreement", has an entire chapter on alliterative agreement. Either there's an inconsistency in his statement, or broad generalisation which intentionally omits the idea as an irrelevant outlier, or I'm mismatching the idea presented by Davush with something else.

[Davush]

I'm not sure what to make of it, but Corbett's own book, the Cambridge Textbook in Linguistics "Agreement", has an entire chapter on alliterative agreement. Either there's an inconsistency in his statement, or broad generalisation which intentionally omits the idea as an irrelevant outlier, or I'm mismatching the idea presented by Davush with something else.

Hello Nel Fie and thank you for this. Yes - alliterative agreement is precisely what this would be! I might be misremembering Corbett but was the idea that if a language has something like alliterative agreement, it will also have a more 'usual' form of agreement (i.e., purely formal agreement won't be the only form of agreement)...? I don't have the text to hand to check though.

[Nel Fie]

Hello Nel Fie and thank you for this. Yes - alliterative agreement is precisely what this would be! I might be misremembering Corbett but was the idea that if a language has something like alliterative agreement, it will also have a more 'usual' form of agreement (i.e., purely formal agreement won't be the only form of agreement)...? I don't have the text to hand to check though.

As you suggest, he points out that all instances of alliterative agreement either co-exist with non-alliterative agreement (e.g. an alliterative singular with a non-alliterative plural), or that other explanations than straightforward phonological copying exist - at least in his opinion. I haven't read anyone else's writing on the matter, and he doesn't go into much detail about his own criticisms. But he also states that he basically doesn't have a definite conclusion to offer because he finds that the data and analysis's are insufficient to be definite, or at least were back when the chapter was written - which was at least eighteen years ago.

So, while that somewhat answers what Corbett's view was at the time of writing the book, that might not reflect the facts on the question of alliterative agreement. I'd love to take a dive for more information, but sadly I don't have the time to spare.

[HolyHandGrenade!]

Anyway, a semi-related question. I had a fun idea, but I'm not sure if this is attested/at all plausible:

Let's say the language has a relative morpheme *tV-.

tira ti-haki 'the man who speaks'

The t-, however, is reinterpreted as an 'agreement marker' matching the initial consonant of the modified noun so that eventually the prefix is CV- where C = the first consonant of the noun:

tira ti-haki
kuta ki-haki
sakra si-haki
etc.

This would then spread to adjectives for a kind of bizarre agreement system which is marked by the initial consonant of the noun appearing on adjectives.

Unless I'm mismatching two unrelated ideas, and at risk of throwing in a wrench after you already replied to DesEsseintes, something like this seems to actually exist in natlangs, although rare. It's called "alliterative agreement", and attested to various degrees in different families, though Bantu languages seem to feature it most prominently.

More information here:

https://en.wikipedia.org/wiki/Agreement ... _agreement
http://glottopedia.org/index.php/Alliterative_agreement

And as the links mention, Corbett has a book on agreement with a chapter about it.

According to Corbett, agreement systems based purely on a formal component (phonological or prosodic) are not attested. All known natlang agreement systems have a semantic component.

I'm not sure what to make of it, but Corbett's own book, the Cambridge Textbook in Linguistics "Agreement", has an entire chapter on alliterative agreement. Either there's an inconsistency in his statement, or broad generalisation which intentionally omits the idea as an irrelevant outlier, or I'm mismatching the idea presented by Davush with something else.

Even so, it seems like it would take a lot more and a lot stronger coincidences for it to be reanalyzed so drastically like this.

[Nel Fie]

Even so, it seems like it would take a lot more and a lot stronger coincidences for it to be reanalyzed so drastically like this.

Probably true. But since the final idea is in itself not implausible, and if Davush wanted it to occur, then it's less a question of "is this plausible synchronically?" and more a question of "what needs to happen for it to be plausible diachronically?".