Conlang Random Thread

[Ahzoh]

I recall something said about ejectives being unstable if more than one exist in a word, but I wonder if words could still have multiple ejectives or if a word like /tʼɑn.ˈdɑsʼ.sʼim/ is unstable and that one of the ejectives dissimilates.

[bradrn]

I recall something said about ejectives being unstable if more than one exist in a word

I don’t recall anything like that. Mayan and Caucasian languages are living counterexamples.

[Ahzoh]

I recall something said about ejectives being unstable if more than one exist in a word

I don’t recall anything like that. Mayan and Caucasian languages are living counterexamples.

Wait, no, it was said here: https://cbbforum.com/viewtopic.php?f=8&t=7516

[Ahzoh]

I don't really know how to analyze my nouns and how they relate to their word-class and cases. Though I suppose it could be analyzed as:

Code: Select all

NounRoot-Class-Plural-Case > Word (sg / pl)
rēb     -i    -:     -m    > rēbim  / rēbīm
surğ    -u    -:     -m    > surğum / surğūm
ezu     -i    -:     -m    > ezīm
mazi    -u    -:     -m    > mazūm 

So I'm basically wondering if the class is closer to the root and thus forms a "stem" or if it's closer the case and plural. To add complexity to the matter, in the construct state of nouns, the "class" ending changes to -a for all nouns (although both the class and case suffixes are removed for the nominative singular, so it basically reduces the noun to the bare root).

This question of the closeness of the class suffixes would determine how susceptible they are to sound changes.

[Creyeditor]

Supposing that the class marker is some kind of noun class/gender/theme vowel that depends on the root in question, I think it should be closer to the root. But theme vowels in particular are known to show up in unexpected places. I wouldn't worry to much about the neutralization in the construct state. That looks like syncretism to me.

[Ahzoh]

Supposing that the class marker is some kind of noun class/gender/theme vowel that depends on the root in question, I think it should be closer to the root. But theme vowels in particular are known to show up in unexpected places. I wouldn't worry to much about the neutralization in the construct state. That looks like syncretism to me.

It's basically like how the -o and -a gender suffixes work in Spanish. It's primarily feminine versus masculine, though depending on the semantic value of the root it can also be domestic versus wild, natural versus manmade, or concrete versus abstract.

I'm not sure the -a class suffix is "syncretism" since it only exists to mark the construct state and probably existed independently of the other affixes rather than as a fusion of them..

---

I like my case system though I feel the way I have set it up limits me to what I can do with cases, in terms of ablaut and stress shifting.

[Man in Space]

Éukaĝ n sl. Hïḫkl. Uám ar? Aĝłíl hé łn? Áge n łn hé, tehar, áín uarák—mmmmmm—árus láus mêḫ łn!
Lamp oil. Rope. Bombs? You want it? It's yours, my friend, provided you have enough—mmm—rupees!
– Morshu, Link and the Faces of Evil

éukaĝ

oil

n

GEN

sl

lamp

hïḫkl

rope

uám

bomb

ar

PL

aĝłíl

want

hé

3SG

łn

2SG

áge

COP

n

GEN

łn

2SG

hé

3SG

tehar

friend

áín

if

uarák

have.with

mmmmm

(vocalized pause)

árus láus

ráus.láus/PL

mêḫ

sufficient

łn

2SG

[Ahzoh]

I came up with a thing where noun cases would be divided between "basic cases" (nominative, vocative, accusative, and instrumental) and "stacking cases" (equative and genitive) where the latter are able to modify a head noun (like adjectives) and engage in suffixaufnahme as well as causing the head noun to be placed in a construct state. These case morphemes are also quirky in that their vowels are in vowel harmony with the gender-marking stem-vowel.

ilig rēbili "a manlike fish" [nom, voc] (other cases: ilgas rēbilis [acc], ilgak rēbilik [ins])
ilig rēbiñi "a man's fish" [nom, voc] (other cases: ilgas rēbiñis [acc], ilgak rēbiñik [ins])
Compare with:
ilgum sağğalum "an old fish" [nom] (other cases: ilgū sağğalū [voc], ilgus sağğalus [acc], ilguk sağğaluk [ins])

Rēbā mīyala tallīntis natlaltin
Rēb-ā mīy-a-la tallīn-ti-s na-tlal-tin
man-CONS.PL.NOM 2ms-0-EQU society\CONS-1cp.POSS-ACC NFUT-destroy\ACT-3mp
"men like you would poison/destroy our society."

Now, I'm not sure if adjectives should take these forms to agree with their nouns.

[bradrn]

I came up with a thing where noun cases would be divided between "basic cases" (nominative, vocative, accusative, and instrumental) and "stacking cases" (equative and genitive) where the latter are able to modify a head noun (like adjectives) and engage in suffixaufnahme as well as causing the head noun to be placed in a construct state. These case morphemes are also quirky in that their vowels are in vowel harmony with the gender-marking stem-vowel.

This is pretty normal. e.g. Dixon lists just this property to justify why he doesn’t consider the ‘genitive’ a noun case: because it tends to have very different properties, such as suffixaufnahme, compared to more prototypical cases like accusative.

[Ahzoh]

And I'm not sure if the stacking cases should also take the nominative suffix when modifying a noun in the nominative case. The lack of suffixal -m is rather characteristic of the nominative case in the construct state, a state which adjectives cannot be placed in and why they also end in suffixal -m.

That is, ilig rēbilim "manlike fish" or ilig rēbili? Compared to ilgum sağğalum "old fish"

Alternatively, basically looking like this:
Ezu Nardiñim ḫallalim "the mighty Boar of Narad (nominative)" or Ezu Nardiñi ḫallalim
Ezu Nardiñī ḫallalī "the mighty Boar of Narad (vocative)" or Ezu Nardiñi ḫallalī
Ezâs Nardiñis ḫallalis "the mighty Boar of Narad (accusative)"
Ezâk Nardiñik ḫallalik "the mighty Boar of Narad (instrumental)"

I dunno, not sure I much like the idea of modifying nouns agreeing in case with head nouns, it often ends of making certain noun phrases (like proper names) inelegant.

[Ahzoh]

I decided against the language having a genitive case and stuck with the old system of the possessor simply taking the same case as its head. Though that did leave me with a question as to how a possessing noun agrees with a head noun in the equative. I am loathe to have the possessor use the equative affix to agree with its head in the equative case. This is also a matter that must be solved for adjectives.

My aversion to having the dependents agree with the equative is twofold: I want to avoid hap(lo)logy with adjectives and determiners (and sound change strategies have not produced appealing results.) and I want to avoid abiguity for when a noun in equative case (behaving in an adjectivelike capacity) modifies a head noun that is also in the equative case.

Dependents agreeing in case with their heads:
surğas alādis
surğ-a-s alād-i-s
chicken-C.SG-ACC warrior-M.SG-ACC
"a warrior's chicken"
surğās alādis
surğ-ā-s alād-i-s
chicken-C.PL-ACC warrior-M.SG-ACC
"a warrior's chickens"

If dependents agree using the equative suffix:
surğala alādili
surğ-a-la alād-i-li
chicken-C.SG-EQU warrior-M.SG-EQU
"like a warrior's chicken"
surğāla alādili
surğ-ā-la alād-i-li
chicken-C.PL-EQU warrior-M.SG-EQU
"like a warrior's chickens"

Sure ok, but then you get messes like this:
surğus alādilis ḫallalilis
surğ-u-s alād-i-li-s ḫallal-i-li-s
chicken-F.SG-ACC warrior-M.SG-EQU-ACC strong-M.SG-EQU-ACC
"a chicken that is like a strong warrior"
surğūs alādilis ḫallalilis
surğ-ū-s alād-i-li-s ḫallal-i-li-s
chicken-F.PL-ACC warrior-M.SG-EQU-ACC strong-M.SG-EQU-ACC
"chickens that are like a strong warrior"

Again, if dependents agree using the equative suffix:
surğulu alādilili ḫallalilili
surğ-u-lu alād-i-li-li ḫallal-i-li-li
chicken-F.SG-EQU warrior-M.SG-EQU-EQU strong-M.SG-EQU-EQU
"like a chicken that is like a strong warrior"
surğūlu alādilili ḫallalilili
surğ-ū-lu alād-i-li-li ḫallal-i-li-li
chicken-F.PL-EQU warrior-M.SG-EQU-EQU strong-M.SG-EQU-EQU
"like chickens that are like a strong warrior"

I mean I suppose i like the suffixaufnahme, but ḫallalilili [ˈxɑl.lɑ.li.li.li] is disgusting and, as I said, attempts to fix it using sound changes ends up being unpleasant or it changes the stress of words that cause the stress pattern to be asymmetrical with the heads.

[jal]

How do you handle noun compounding? Because juxtasposing two nouns seems like the primary way of handling noun compounding.


JAL

[Ahzoh]

How do you handle noun compounding? Because juxtasposing two nouns seems like the primary way of handling noun compounding.

Is that a question directed at me or to the thread in general?

[jal]

Is that a question directed at me or to the thread in general?

Oh, sorry, yes, directed at you.


JAL

[Ahzoh]

Is that a question directed at me or to the thread in general?

Oh, sorry, yes, directed at you.

I have three ways of compounding:

• Genitive construction (e.g. īz ḫuzūm "horse god / god of horses") is used for endocentric compounds (A+B denotes a special kind of B)
• the basic closed compound (e.g. ṭandaṣṣim "oxhorn, a type of polearm") is used for exocentric compounds (A+B denotes a special kind of an unexpressed semantic head)
• A subset of closed compound called the "ē-compound" (e.g. yatēḫūnum "land-and-sky" > "the world, everything that exists") is used for copulative compounds (A+B denotes 'the sum' of what A and B denote).

The ē-compound is more commonly found in Literary Vrkhazhian whereas Colloquial Vrkhazhian disprefers its use. There is certainly aesthetic considerations involved in the formation of these compounds, e.g. wadēlaḫḫim ("shield-and-weapon") is preferred over laḫēwaddim ("weapon-and-shield") even though they both have the same semantic meaning of "equipment". Likewise yatēḫūnum is preferred over ḫūnēyatum.

[quinterbeck]

I just wrote the following:

Occasionally, pronominal clitics (including demonstratives and quantifiers) have no host to attach to. This usually occurs when the pronoun is itself the object of a predicate light verb, or when it is part of a coordinated noun phrase. In these cases, one of two semantically light base morphemes is supplied to host the clitics, depending on whether the referent is personal or impersonal. These are bound morphemes, and only occur to host clitics.

Base	Referent type	Gloss
an	impersonal	IPSL
or	personal	PSL

Does anyone have suggestions to improve the terminology or glossing? It's a dummy morpheme, but can I get more specific?

Here are some examples, including normal behaviour of pronominal clitics

Object of a predicate light verb:
Eunwea orig.

eun=we=a

COM=3s=1s

or=ig

PSL=2p

We are with you.

Gen emde anha?

gen

Q

em=de

ID=this

an=ha

IPSL=what

What is this?

Part of a coordinated noun phrase (the base reduces when the coordinating prefix ends in a vowel):
Na yeam gei moulgiha?

Na

POL

yeam

FUT.ACT.2s

gei

do

mou-l-gi-ha

or-PSL=other=some

Will you do it, or someone else?

Common constructions for 'only', 'nearly', 'just over'
(1) coordinated
Dugwe homi mehhegildor naineu.

Du-g=we

PST:ACT-E=3s

homi

buy

meh-hegildor

six-egg

nai-n=eu

and.not-IPSL=more

He only bought six eggs.

(2) appositioned
Dugwe homi mehhegildor aneureo.

Du-g=we

PST:ACT-E=3s

homi

buy

meh-hegildor

six-egg

an-eu=reo

IPSL-more-no

He only bought six eggs.

Some speakers also use these bases as a strategy to extract pronominals when they are further modified
Nu ada yamdi anman anona demainado

Nu

GNO

ad=a

PROP=1s

yamdi

find

an=man

IPSL=all

anon=a

REL.EXP=1s

demain=ado

need=abst.pl

I can find everything I need

[bradrn]

I just wrote the following:

Occasionally, pronominal clitics (including demonstratives and quantifiers) have no host to attach to. This usually occurs when the pronoun is itself the object of a predicate light verb, or when it is part of a coordinated noun phrase. In these cases, one of two semantically light base morphemes is supplied to host the clitics, depending on whether the referent is personal or impersonal. These are bound morphemes, and only occur to host clitics.

Base	Referent type	Gloss
an	impersonal	IPSL
or	personal	PSL

Does anyone have suggestions to improve the terminology or glossing? It's a dummy morpheme, but can I get more specific?

I think this is fine as is. I occasionally see people gloss similar things as ‘epenthesis’ EP, but I’m not sure this is totally applicable here. If you really want a special term, perhaps you could gloss this as ‘clitic host’ or something.

The whole thing feels somewhat odd, though… people often take insensitivity to host word class as one of the defining features of clitics, and I’d expect them to simply cliticise to the next word no matter what it is. Requiring a specific host feels more like an affix than a clitic to me, though it’s by no means impossible.

(To a first approximation I tend to think of clitics as simply normal words which phonologically happen to be grouped with an adjacent word. To a second approximation it’s more involved than that, because you also get clitics which don’t correspond to normal words, and these tend to have special syntax.)

[quinterbeck]

The whole thing feels somewhat odd, though… people often take insensitivity to host word class as one of the defining features of clitics, and I’d expect them to simply cliticise to the next word no matter what it is. Requiring a specific host feels more like an affix than a clitic to me, though it’s by no means impossible.

(To a first approximation I tend to think of clitics as simply normal words which phonologically happen to be grouped with an adjacent word. To a second approximation it’s more involved than that, because you also get clitics which don’t correspond to normal words, and these tend to have special syntax.)

That is true... I did refer to this group (pronouns, demonstratives, and quantifiers) as affixes for a long time. I believe I switched to calling them clitics because for the most part they have the same syntactic behaviour as substantives.

Basic syntactic rules

The word after a noun modifies it:

ragdi

dog

hoed

black

'black dog'

ragdi=de

dog=this

'this dog'
(Well, clitics determine or quantify. So pure pronouns can't take this role.)

The word after a verb is its object:

dum

eat

oream

fish

'eating'

dum=man

eat=all

'eating everything'
(Demonstratives and quantifiers are pronouns in argument slots)

Same goes for light verbs when they modify (SVO order)

ragdi

dog

in

GEN

eoneg

woman

'the woman's dog'

ragdi

dog

in=a

GEN=1s

'my dog'

When a light verb heads a clause, the following word is its subject (VSO order)

ad

PROP

ragdi

dog

deana

good

'the dog is good'

ad=id

PROP=anim.sg

deana

good

'it is good'

The functions can stack, so verb+(argument+modifier) can attach together:

ad

PROP

ragdi

dog

hoed

black

di

RESV*

deana

good

'the black dog is good'
argideman deana

ad=gi=de=man

PROP=other=this=all

deana

good

'all these others are good'
*resumptive particle

Not-so-basic stuff

Personal pronouns have special stacking behaviour indicating mixed-person groups (no 1p pronoun)

rug-no-a

LOC=3p=1s

place=this

hir=de

'we're here'
Partly because of potential ambiguities with the pronoun stacking, I don't allow verb+subject+object to coalesce into one phonological unit. It's just a basic constraint I've had since the beginning.

Then again, these morphemes always attach directly to the head of their phrase, because I don't allow modifiers to scope over more than one preceding word. That's very un-clitic-like...

ragdi-man

dog-all

hoed

black

'all black dogs'
??

ragdi

dog

hoed-man

black-all

'??a dog of all blackness' ('a completely black dog'?)

Plus there is some sensitivity to word class - these morphemes don't attach to particles, e.g. negative:

em-mu

ID-that

reo

NEG

or-a

PSL-1s

'that's not me'

[jal]

ragdi

dog

hoed

black

'black dog'

ragdi=de

dog=this

'this dog'

If "this black dog" is "ragdide hoed" instead of "ragdi hoedde" I'd consider "de" not a clitic, but what the right term is for a particle that attaches phonologically to its head, but not like a clitic to an entire phrase, I don't know.


JAL

[bradrn]

… what the right term is for a particle that attaches phonologically to its head, but not like a clitic to an entire phrase, I don't know.

Those are clitics too. At the most fundamental level, a clitic is something which grammatically behaves like a separate word, but is phonologically dependent. (Though as usual linguists (mis)overapply it to various extents so it ends up as something like a catch-all category.) I haven’t yet gotten enough time to properly think about quinterbeck’s post, but the question to ask here is: can the ‘clitics’ be used in all the same places as any other grammatical word? Or are they more restricted in their host and/or positioning? If the former, they’re clitics; if the latter, I’d call them affixes.