Grammaticalization Quickie Thread

[StrangerCoug]

Thought I'd revive the old thread on the basis of the Sound Change Quickie Thread and my earlier Contechnology Quickie Thread.

I've got a pair of sister agglutinative languages I'm working the kinks out of so I can make a proper grammar out of them. In both, adjectives follow the nouns they modify, and there's a productive genitive suffix /ɗe/ in both of them, one use of which is to form compound words (in which modifiers follow the head, so they often end in this genitive suffix). I also use it to derive language names, usually in what would be the format "language of (place)" in English. The problem is that my place names so far are all compound words, and I find it unaesthetic to simply apply the /ɗe/ suffix twice consecutively for different functions. Do I have any other options besides just possessed case?

[bradrn]

[...] adjectives follow the nouns they modify, and there's a productive genitive suffix /ɗe/ in both of them, one use of which is to form compound words (in which modifiers follow the head, so they often end in this genitive suffix).

I don't quite understand. Do you mean that the genitive can be applied to both nouns and adjectives?

[Kuchigakatai]

If you have place names like:

grove chicken-de
valley fern-de

presumably, you'd make the phrase genitive to modify the word "language":

tongue grove-de chicken-de
tongue valley-de fern-de

Alternatively, you could use a derivational affix like -mi, meaning '-ese':

grove-chicken-de-mi
valley-fern-de-mi

Or a prefix or such. If the language is commonly talked about, reducing the length of the term would be a good idea, cf. Latin "Portus Callus" + -ensem > portucalensem > "Portuguese/português" (syncope eliminated the syllable -al-).


Notice how both strategies were used in Latin: to say 'in [language]', it used the suffix -e, but otherwise it used the noun "lingua" + a derived adjective. Latine 'in Latin' (adverb), lingua Latina 'Latin, the Latin language' (noun phrase).

Latine scriptum est. 'It is/was written in Latin.'

E lingua Graeca in (linguam) Latinam vertentur. 'They will be translated from Greek into Latin.'

[StrangerCoug]

[...] adjectives follow the nouns they modify, and there's a productive genitive suffix /ɗe/ in both of them, one use of which is to form compound words (in which modifiers follow the head, so they often end in this genitive suffix).

I don't quite understand. Do you mean that the genitive can be applied to both nouns and adjectives?

They can, but I think I can see where the confusion is coming from. Let me try to clear it up.

The two languages I have in question are named after places that are themselves formed by compounding words like this (glossing into English): bay-south-GEN (for what I call South Bay in English) and bridge-stone-GEN (for what I call Stone Bridge in English), with "bay" and "bridge" implicitly in the unmarked case, which will likely be the nominative. However, I don't want the full language names to gloss to "speech bay-south-GEN-GEN" and "speech bridge-stone-GEN-GEN" and I'm looking for options that I may not be aware of.

Edited to add:

If you have place names like:

grove chicken-de
valley fern-de

presumably, you'd make the phrase genitive to modify the word "language":

tongue grove-de chicken-de
tongue valley-de fern-de

I like this idea

[vergil]

So I've been having some trouble with the evolution of one of my conlangs' morphology, and this seemed like the place to ask the questions I had.

This regards the Dative case, /-t/ (and /-tɛ/ and /-te/) in Pretistelen, particularly its development in one of the daughter languages.
One of the phonological changes is {t, d} > ʔ / V_V, V_#, which in the case of the Dative would change the marker to /-ʔ/, which is rather closer to the unmarked nominative case than I like (especially since I plan on also having ʔ > Ø later).
My intial thought was to just have the vowel before the Case Marker lengthen (which I might do anyway for some nouns) but others are already long. Changing the standard case ending to match one of the other declensions runs into similar problems: /ɛ:ɛ/ just feels wrong.

Complicating matters is that the Dative has 3 tasks:
1. Dative/indirect object (This one got offloaded to a separate case, adding <ros> "toward" to the case ending for <-tros>, so this task isn't really a consideration).
2. Instrumental.
3. Object of stative postpositions.

I see a few strategies for dealing with the issue.
1. Offload all the tasks to other cases.
2. Combine or switch the case marker with a postposition (as for Indirect Object task).
3. Some combination of (1) and (2).
4. Substitute a case marker from another language.

I'm partial to (3), offloading the postpositional task to the Accusative/Genitive and adding a postposition to the Dative to create an Instrumental, but I'm not sure what postposition I'd use (I know Comitative "with" is an option, but I've been keeping that pretty distinct so I'm not sure how much an option that really is ). The other problem is a number of the postpositions I'd be using (<ea> "at", <aifa> "near, with") run into the same sort of vowel chain problem as changing the case marker.

Thoughts? Erm, I'm also looking for any resources on historical morphology development.

[Pabappa]

Lithuanian has some cases that are derived from other cases plus a suffix. Hungarian might do this too. I like option 3.

[cedh]

For the object of stative postpositions, you could simply let everything go its regular way so that -t > -ʔ > -Ø, effectively merging with the nominative case in this usage. Alternatively, maybe the -t gets treated like a word-internal coda consonant before stative postpositions, if this triggers interesting cluster simplifications (for example, -t C- might become -Ø =CC- with an initial geminate on the postposition).

For the instrumental function, extending the old dative case with a postposition seems a good thing to do. For suitable postposition semantics, you might want to take a look at page 29 of the Conlanger's Thesaurus.

Another option, of course, would be to analogically extend the -tɛ ~ -te variant of the case suffix (before or at roughly the same time as the sound change t > Ø / _#), so that the original dative case stays distinctive.

[vergil]

Update: I'm using <tes> "through" + the Dative for the Instrumental.

I'm also using <tes> as an intensifier / telicity marker for verbs, which might be complicating things somewhat, but I figure this is a decent additional application.

So I'll have -ttɛs, -tɛʔɛs, and -teʔɛs. This actually looks pretty managable.

[Qwynegold]

For a simple solution I would've just suggested this:

My intial thought was to just have the vowel before the Case Marker lengthen (which I might do anyway for some nouns) but others are already long.

...and let those nouns that already had a long vowel at the end be unmarked. Unless long vowel is the most common way nouns end.

[StrangerCoug]

What's the difference between the first-person jussive and the cohortative? I'm trying to decide on what moods to include in the conlang I'm creating, and I'd like two of them to be the imperative, which semantically will cover the jussive as well, and one other deontic mood for non-commands. The hortative was one of the ones I was thinking, but in my scratchpad I currently have a general volitive in its spot.

Additionally, in the event I don't use a specific morphological hortative, what are some other ways to maintain as clean an imperative-hortative split as possible?

[StrangerCoug]

What's the difference between the first-person jussive and the cohortative? I'm trying to decide on what moods to include in the conlang I'm creating, and I'd like two of them to be the imperative, which semantically will cover the jussive as well, and one other deontic mood for non-commands. The hortative was one of the ones I was thinking, but in my scratchpad I currently have a general volitive in its spot.

Additionally, in the event I don't use a specific morphological hortative, what are some other ways to maintain as clean an imperative-hortative split as possible?

Does anyone know? Are they similar enough that I can just say there never developed a lexical distinction in the first person for what I'm talking about?

[Kuchigakatai]

Does anyone know?

Both "imperative" and "jussive" come from Latin synonym verbs of command, imperō (imperātum) and iubeō (iussum), both meaning 'to command/order somebody to do something'. "Jussive" is basically a word of convenience in some grammatical traditions for something that is imperative-ish at least some of the time and yet is morphologically distinct from the imperative proper.

For example, Latin had a 3rd person imperative: active sg. amātō pl. amantō, passive sg. amātor pl. amantor; but these forms were rarely used as in Classical times they were mostly replaced by the present subjunctive (amet, ament, amētur, amentur). The present subjunctive is sometimes used for 2nd person commands as well, though rarely by Classical times. The present subjunctive was then said to have "jussive" force when used this way, and so was called the coniunctīvus iussīvus (that is, the present subjunctive is basically something like an imperative when used this way, but true formal 3rd and 2nd person imperative forms do exist so let's not get them confused by using the same word).

For another example, Standard Arabic has morphologically distinct imperatives. There is also a morphologically distinct paradigm of inflections that formally looks like some type of indicative mood thing but is used for command-related things half of the time (the other use is to negate verbs with past semantics), so in the Western tradition of Arabic grammar the word "jussive" was found to be useful for this imperative-ish paradigm (while keeping it distinct from the imperative proper).

So when you ask what the difference is between a 1st person jussive and a 1st person cohortative, I read it as asking what the difference between an imperative that is a "secondary" imperative in some sense and a cohortative is.

As for the second question in your first post, if you must clearly distinguish imperatives and cohoratives in some sense, you can perfectly do it with syntax... Use an auxiliary verb, an adverb, some kind of construction with a subordinate finite verb, a modal particle...

[StrangerCoug]

Does anyone know?

Both "imperative" and "jussive" come from Latin synonym verbs of command, imperō (imperātum) and iubeō (iussum), both meaning 'to command/order somebody to do something'. "Jussive" is basically a word of convenience in some grammatical traditions for something that is imperative-ish at least some of the time and yet is morphologically distinct from the imperative proper.

Knew I was missing something xD

As for the second question in your first post, if you must clearly distinguish imperatives and cohoratives in some sense, you can perfectly do it with syntax... Use an auxiliary verb, an adverb, some kind of construction with a subordinate finite verb, a modal particle...

It probably makes best sense to just go ahead and do that.

[dɮ the phoneme]

I'm trying to get a nominative-accusative system out of a split-ergative one, where the split is based on an animacy hierarchy. How might that occur?

[Xwtek]

I'm trying to get a nominative-accusative system out of a split-ergative one, where the split is based on an animacy hierarchy. How might that occur?

Can you give the current declension system?

[Xwtek]

I have abandoned my Asent'o, but I'm still curious how do you grammaticalize relative pronoun. I couldn't use relative pronoun in my new conlang due to a different environment. (And also because of more extensive voice system).

[Ælfwine]

I can't answer the above question, which actually I would like answered as well, as I have a similar issue in my Crimean Gothic conlang which has criticized -thata to verbs (c.f. varthata, malthata). However, whether it replaces marking for person, as it seems to have done, or simply add onto it, I am unsure.

Edit: perhaps it introduces relative clauses somehow?

[Pabappa]

Are you guys saying you want to make a verb form that means "(that) which X"? Like a participle, or is it something different? I wouldnt know what ancestral form would lead to the /-thata/ being found at the end of the verb. I guess youre asking about person marking because you could have a three-way contrast like

I am the one who drew-1P the picture.
You are the one who drew-2P the picture.
They are the ones who drew-3P the picture.

Where the ancestral form of the verb would be different for each case (and if this is SOV, it would be at the end). It sounds like a participle to me, but Im not sure i'm seeing it right .... all that comes to mind is that Slavic languages did it the other way, where the person markers are stuck after the participial form of the verb and I think might derive from the conjugated forms of "to be".

This could be an interesting way to develop gender marking on verbs if you want to.

[Ælfwine]

Didn't some Romance dialects develop gender marking on their verbs?

That's an interesting idea, Pabappa. Though in my case, all the verb forms are the same, I'm just unsure how person is marked for (assuming it is.)

The three phrases I am going off of in question (nb not my conlang, rather from the corpus I am working with:)

ich malthata "i say that"
tzo vvarthata "you made that"
ies varthata "he made that"

As far as I can tell, there's no person marking here (the difference between <vv> and <v> seems to be one of transcription.) So my question is what happened to it? Did the clitic obscure it somehow? Perhaps the pronouns used here are resumptive pronouns that buffer the verb due to the loss of person in this case.

My idea now is that this form originates from relative pronouns in relative clauses becoming cliticized to the end of the verb. How this fully works I've not figured out yet.

[Pabappa]

Maybe it just comes from having the 3rd person be the only form available to use in such a position in the first place. I typed the example in English because I dont actually know a language that uses person marking inside of a relative clause. My understanding is that in e.g. Spanish, all such clauses are in the 3rd person regardless of who the agent of the verb is. Even if this is an areal feature, if we assume it goes back to PIE it would be that way in Germanic too.