Deriving an agglutinating language

If you're new to these arts, this is the place to ask "stupid" questions and get directions!
Ælfwine
roman
roman
Posts: 940
Joined: 21 Sep 2015 01:28
Location: New Jersey

Re: Deriving an agglutinating language

Post by Ælfwine »

Isfendil wrote:
Ælfwine wrote:
Creyeditor wrote:Some people distinguish between core (or morphosyntactical) cases like nominative, accusative, genitive, dative cases and adverbial cases (or semantic cases) like locative, superessive and translative case.
Oh, I see...

Say, any conlangs that people have done that went through the process of agglutination that I may study? Or better yet, any good free online articles/books that do not require a bachelor's degree in linguistics I can read? I just want to study more examples.
Maybe we can demonstrate this to you with our own conlangs? What exactly do you want to see?
I'm not sure. Maybe a step-by-step process-from a proto-lang to a modern language- on how perhaps an isolating or fusional lang got to be agglutinating.

I'm toying around with the two languages that I know - Old Norse and Latin - and seeing what can be regularized. In Old Norse I've noted that in some cases, the first person singular pronoun ek was sometimes attached to the end of the verb. I've extended this tendency to most other singular pronouns.

ex.
hafða + ek = hafðak “I have”
kann + ek = kannk “I can”
heyr + tú = heyrtu “You hear”
fórt + sú = fórsu “She went” (though regularized, this is probably not in the past tense.)

This is a good start, but I am still a bit lost on where to go from here.
My Blog

A-posteriori, alternative history nerd
clawgrip
MVP
MVP
Posts: 2257
Joined: 24 Jun 2012 07:33
Location: Tokyo

Re: Deriving an agglutinating language

Post by clawgrip »

In deriving Nandut from Naduta, I went from two cases, nominative and objective, to four pseudo-cases: nominative-direct and nominative-oblique, and objective-direct and objective-oblique. This was caused entirely from sound changes, specifically:
1. voicing of unvoiced stops
2. loss of final vowels
3. devoicing of final voiced stops

Adding a postpositional clitic eliminates rules 2 and 3, causing two distinct forms to appear:
with the example suffix day (nday in Nandut):

Code: Select all

Naduta
                      singular   plural
        nominative    phuta      phurta
        nominative    phuta day  phurta day
        objective     phuy       phurey
        objective     phuy day   phurey day

No difference with or without the postposition in Naduta. But:

Nandut
                      singular   plural
nominative direct     put        pōt
           oblique    puda.ndai  pōda.ndai
objective  direct     pī         purē
           oblique    pī.ndai    purē.ndai
In the animate gender, objective direct and oblique are identical, but in other genders they are different.

So my point is, clitics that alter the application of sound changes in a word can become dependent morphemes, creating new cases.
Ælfwine
roman
roman
Posts: 940
Joined: 21 Sep 2015 01:28
Location: New Jersey

Re: Deriving an agglutinating language

Post by Ælfwine »

clawgrip wrote:In deriving Nandut from Naduta, I went from two cases, nominative and objective, to four pseudo-cases: nominative-direct and nominative-oblique, and objective-direct and objective-oblique. This was caused entirely from sound changes, specifically:
1. voicing of unvoiced stops
2. loss of final vowels
3. devoicing of final voiced stops

Adding a postpositional clitic eliminates rules 2 and 3, causing two distinct forms to appear:
with the example suffix day (nday in Nandut):

Code: Select all

Naduta
                      singular   plural
        nominative    phuta      phurta
        nominative    phuta day  phurta day
        objective     phuy       phurey
        objective     phuy day   phurey day

No difference with or without the postposition in Naduta. But:

Nandut
                      singular   plural
nominative direct     put        pōt
           oblique    puda.ndai  pōda.ndai
objective  direct     pī         purē
           oblique    pī.ndai    purē.ndai
In the animate gender, objective direct and oblique are identical, but in other genders they are different.

So my point is, clitics that alter the application of sound changes in a word can become dependent morphemes, creating new cases.
This bump is a bit more than a year late but more relevant than ever.

Through a discussion with Isfendil, I've decided to bring back agglutination in my romlang. In order to do this, I want to create a new accusative and dative case from Latin. I've so far come up with a dative that is from the Vulgar Latin preposition, a(d). I've been thinking of ways to make the agglutination fairly naturalistic.

Note that the plural is -i, which like in Romanian disappears and palatalizes before consonants. Also note the sound change t -> z / V_V. The rules are otherwise similar to your Nandut's.

Code: Select all

not "note"
                 singular   plural
nominative       not        nöc (nøts)
dative           noza       nözai
I don't know if you are still around, but if you are: look good to you Clawgrip?
My Blog

A-posteriori, alternative history nerd
clawgrip
MVP
MVP
Posts: 2257
Joined: 24 Jun 2012 07:33
Location: Tokyo

Re: Deriving an agglutinating language

Post by clawgrip »

That seems fine to me. I like how the plural works as well, though I wonder if the plural might appear in front of the case ending, if it was already part of the noun.

not - nöc
noza - nöza

Possibly the i does not disappear when followed by ad, so you could have noza - nöze (ia → e) or something along those lines.
Ælfwine
roman
roman
Posts: 940
Joined: 21 Sep 2015 01:28
Location: New Jersey

Re: Deriving an agglutinating language

Post by Ælfwine »

Spoiler:
clawgrip wrote: 03 Sep 2016 13:25 The development of the Romance future tense to replace the Latin future tense is a good example of how to develop new suffixes. If the personal suffixes are regularized before the auxiliary becomes an enclitic, then it can preserve its form quite regularly.

"I walk" etc.
ambulo
ambulas
ambulat

"I will walk" etc.
ambularbo
ambularbas
ambularbat

from ambulare hab- we get our future suffix -arb

If we allow stacking of infinitives, we can get more suffixes.

if possum "can" is regularized to posso:

"I can walk"
ambularpso

rp may be reduced to pp, which may again reduce to p, so:

ambulare posso "I can walk " etc.:
ambulapso
ambulapsas
ambulapsat

ambulare posse habeo "I will be able to walk" etc.:

ambulapsarbo
ambulapsarbas
ambulapsarbat

with plurals:
ambulapsarbons
ambulapsarbasans
ambulapsarbatans

ambul-aps-arb-o-ns
walk-POT-FUT-1-PL

(I want to reduce ambul- to ambl-, but that's a different matter)
Clawgrip,

I’m looking back at this post as a reference to how I can structure my verbs, but I have a few questions about your thinking process behind it.

First of all, what marks the infinitive mood? I’m guessing the root is the bare infinitive? It’s worth noting that in Hungarian (and Portuguese) the person suffix is added onto the infinitive suffix. You basically merged the infinitive with the future, though. Could there be a compromise between these two ideas, perhaps keep the future as <b> and the infinitive as <ar> separate? Or should I simply innovative a new infinitive — or leave as is?

Second of all, what made you decide to go with ar as opposed to er or ir for “arb?” And what happened to the -er and -ir conjugations, were they just replaced/eliminated entirely? Outside of vowel harmony, could I see the form -erb or -irb for the future?

Likewise, why does the plural suffix have a nasal consonant? I’m guessing the <a> in it is epithetic.

Finally, I’m thinking a subjunctive mood would be formed from VL *ɛsserɛ and the imperative from *deβerɛ. Does this sound logical?
My Blog

A-posteriori, alternative history nerd
clawgrip
MVP
MVP
Posts: 2257
Joined: 24 Jun 2012 07:33
Location: Tokyo

Re: Deriving an agglutinating language

Post by clawgrip »

It's been quite a while since I wrote that, so I had to look back a bit on what I did.

First, I need to point out that this was not some deeply thought-out conlang or anything; it was just an example I came up with for that post.

Aside from the innovations regarding person marking and tenses, it is essentially just Latin.

To address your questions:
First of all, what marks the infinitive mood?
I guess it would be -are as in Latin, though if I were doing this for real, I would shorten it to -ar at least.
I’m guessing the root is the bare infinitive?
The root is ambul-
You basically merged the infinitive with the future
I didn't merge the infinitive with the future, I merged the infinitive with habeo to get a new future tense, which is exactly what happened to create the French and Spanish future tenses. I just retained the /b/, which they did not, and I retained my regularized personal endings, which are something else in French and Spanish.

Code: Select all

Latin        French   Spanish  This
amo          aime     amo      amo
amas         aimes    amas     amas
amat         aime     ama      amat

amare habeo  aimerai  amaré    amarbo
amare habes  aimeras  amarás   amarbas
amare habet  aimera   amará    amarbat
Essentially, in French and Spanish, the future tense is just the infinitive ending plus a personal ending that may or may not be different from the standard personal ending. In this agglutinative language, the personal endings are all regularized and the /b/ of habeo is retained.
Second of all, what made you decide to go with ar as opposed to er or ir for “arb?”
The infinitive ending for ambul- is -are. This is the same as Spanish.
And what happened to the -er and -ir conjugations, were they just replaced/eliminated entirely?
I would probably have them as different conjugation classes, though you could eliminate them if you wanted. It would still be agglutinative, but the exact ending added depends on the class of the stem.

Code: Select all

Latin          French    Spanish   This
dormio         dors      duermo    dormo
dormis         dors      duermes   dormis
dormit         dort      duerme    dormit

Latin          French    Spanish   This
dormire habeo  dormirai  dormiré   dormirbo
dormire habes  dormiras  dormirás  dormirbas
dormire habet  dormira   dormirá   dormirbat
It seems like what I have come up with now, the verb class determines which epenthetic vowel should be added, e.g. the personal endings are just -o, -s, -t, and the class determines if it should be a, i, or e. So maybe the future should be -rb(e) rather than -rb(a)?
why does the plural suffix have a nasal consonant?
I think the a was epenthetic. The plural suffix was just an on-the-fly regularization of -m(u)s, -t(i)s, and -(u)nt. Two nasals and two ending in s, so I just merged them to -ns.
Finally, I’m thinking a subjunctive mood would be formed from VL *ɛsserɛ and the imperative from *deβerɛ. Does this sound logical?
Let me get back to you on this in a bit. I'm not an expert in Latin and it will take me a bit of looking to answer this!
Ælfwine
roman
roman
Posts: 940
Joined: 21 Sep 2015 01:28
Location: New Jersey

Re: Deriving an agglutinating language

Post by Ælfwine »

clawgrip wrote: 20 Dec 2017 13:06 It's been quite a while since I wrote that, so I had to look back a bit on what I did.

First, I need to point out that this was not some deeply thought-out conlang or anything; it was just an example I came up with for that post.
That's fine. I should have realized that you probably couldn't remember a post from a year ago. :p
I guess it would be -are as in Latin, though if I were doing this for real, I would shorten it to -ar at least.
Yet in your original examples, you only included a potential mood, otherwise leaving it unmarked. That's why I thought the unmarked root marked the infinitive, as in English, as -arb only marked the future tense from my PoV. Or did -arb mark both?

My thinking is that the class vowel -a (or -e, or -i) would also mark the infinitive, -rb would mark the future, and -o, -s, -t mark the person.
I didn't merge the infinitive with the future, I merged the infinitive with habeo to get a new future tense, which is exactly what happened to create the French and Spanish future tenses. I just retained the /b/, which they did not, and I retained my regularized personal endings, which are something else in French and Spanish.
I meant that, of course, and not the Latin future.
Essentially, in French and Spanish, the future tense is just the infinitive ending plus a personal ending that may or may not be different from the standard personal ending. In this agglutinative language, the personal endings are all regularized and the /b/ of habeo is retained.
It's worth noting that some Spanish dialects have retained /b/, even intervocally, as in Aragonese.
The infinitive ending for ambul- is -are. This is the same as Spanish.

...

I would probably have them as different conjugation classes, though you could eliminate them if you wanted. It would still be agglutinative, but the exact ending added depends on the class of the stem.

...

It seems like what I have come up with now, the verb class determines which epenthetic vowel should be added, e.g. the personal endings are just -o, -s, -t, and the class determines if it should be a, i, or e. So maybe the future should be -rb(e) rather than -rb(a)?
Vowel harmony may definitely screw with the conjugation classes, though I am not sure in what ways. So it could be both.

https://en.wikipedia.org/wiki/Vowel_harmony#Hungarian
I think the a was epenthetic. The plural suffix was just an on-the-fly regularization of -m(u)s, -t(i)s, and -(u)nt. Two nasals and two ending in s, so I just merged them to -ns.
Okay. I thought something similar, but wasn't sure.
Let me get back to you on this in a bit. I'm not an expert in Latin and it will take me a bit of looking to answer this!
I always appreciate the help, Claw. I asked some other people and they said *deβerɛ was fine for an imperative, yet I can't think of a Romance language that grammaticalizes it. *esserɛ could easily become a passive construction, though if I follow Hungarian's path I may not want that. (Hungarian is a topic prominent language, for example.)
My Blog

A-posteriori, alternative history nerd
Ælfwine
roman
roman
Posts: 940
Joined: 21 Sep 2015 01:28
Location: New Jersey

Re: Deriving an agglutinating language

Post by Ælfwine »

I'm having a little bit harder time trying to regularize Eastern Romance. Your original example works well for a Western Romance language, but less so for an Eastern Romance once. While I considered moving the location, I still wanted to do a Romance language based in Hungary and agglutination is unavoidable there with the strong Avar and Magyar substrates.

So going back to verbs, this is what I found:

*Eastern Romance seems to prefer -i in the indicative present 2nd person singular and lacks -s entirely to form the plural. (Italian ami to Spanish amas).
*Except Italian, Eastern Romance doesn't seem to use habeo to construct future tenses, which is unfortunate. Neither does Ladin and Romansch, which are nearby. It is used in the perfect in Romanian, however.
*I've noticed that some Northeastern Italian dialects have a tendency to use clitic subject pronouns to rebuild person marking. Maybe that's an idea?

Of course I can handwave some of these as exceptions to the La Spienza line are common, but as I already have my nouns based around Eastern Romance forms, this seems unlikely. One good thing is that Dalmatian and Istriot at least became quite quickly analytic, so pushing the hands of the clock forward more shouldn't be too hard, the question is mostly "how."

Once again any help would be appreciated.
My Blog

A-posteriori, alternative history nerd
Post Reply