Tangut-ish isolating lang: a scratchpad

Substantial postings about constructed languages and constructed worlds in general. Good place to mention your own or evaluate someone else's. Put quick questions in C&C Quickies instead.
Post Reply
Porphyrogenitos
Lebom
Lebom
Posts: 168
Joined: Thu Dec 22, 2011 1:13 pm
Location: Ohio

Tangut-ish isolating lang: a scratchpad

Post by Porphyrogenitos »

I'm creating this thread to avoid spamming up the conlang tag on Tumblr, where I've been posting this material up until now. This is all just grammatical stuff right now, as I haven't figured out much of anything in the way of phonology or vocabulary. But it will come eventually. I guess I'm kind of taking the opposite route of a lot of conlangers.

This language isn't meant to be set in any fictional universe, and has some loglang-like tendencies. It's primarily just a personal language for my own aesthetic and artistic enjoyment.

Overview

Firstly, it's totally isolating, or will probably be as isolating as I can make it. Zero marking will feature prominently. It is strictly VSO. And yeah, as far as I can figure it's just gonna be nominative-accusative.

Subject and object are indicated simply by word order. Other grammatical roles are indicated by prepositional particles:
see 3.s man LOC store
"He sees [the] man in [the] store."
There will be a "zero person" pronoun/placeholder that can fill in for the subject, and I suppose any other grammatical role. (I'm going to use "0" for it, unless someone can suggest a better gloss.)
see 0 3.s LOC store
"He was/is seen at the store."

give 3.s 0 DAT 1.s
"He gives [something] to me."
Relative clauses

I keep encountering difficulties with the details of relative clauses, so I won't go into a large discussion yet. But I can say a few things.

Relative clauses are zero-marked. They occur when a verb directly follows a noun, which never happens in any other circumstance. I'm pretty sure I'm going to be doing something with a resumptive particle, or a resumptive pronoun, like this:
woman see 1.s 3.s.RES
"The woman who I saw [her]"
Verbs and adjectives

Verbs are, of course, not marked for anything. The role of adjectives will be taken by stative verbs and a small, closed class of adjectives consisting of numbers and other determiner-ish words, as well as perhaps a few basic descriptive concepts. These "true adjectives" will be placed before the noun, as with the prepositional particles. (They will also come after the prepositional particles.)
give 1.s food DAT two bird blue
"I give/gave [the] food to the two birds which are blue."
"I gave the two blue birds food."
And we thus have this distinction:
blue bird
"The bird is blue."

bird blue
"The bird that is blue"
Nominal and dependent clauses

A clause can be made into a nominal/dependent clause with a subordinative particle. The clause as a whole is positioned in the sentence as if it were a noun and can be modified by prepositional particles.
please SUBR see 1.s woman tall 1.s
"That I saw the woman who is tall pleases me."
"It pleases me to see the tall woman."
This construction can be used to fulfill the role of infinitives in English and other languages:
want 1.s SUBR go 1.s LOC store
"I want that I go to the store."
"I want to go to the store."
A different particle, a deverbalizer, can be used to convert the clause into what is essentially a deverbal gerund in English.
like 1.s DEV sing 1.s
"I like my singing."
As opposed to:
like 1.s SUBR sing 1.s
"I like that I sing."
"I like to sing."
Or, this example:
like 1.s DEV sing
"I like [the act of] singing [in general]."
As opposed to:
like 1.s SUBR sing
"I like that someone is singing."
"I like that there is singing."
Deverbal clauses not only take the position of a noun in a sentence, but can also be modified by adjectives and relative clauses, unlike regular subordinate clauses.
award 3.p 3.s.f CAUS two DEV capture 3.s.f bear
"They awarded her because of her two capturings/captures of the bear."
(I still need to figure out how exactly relative clauses modifying deverbal phrases will work without creating syntactical ambiguity.)
Last edited by Porphyrogenitos on Fri Nov 28, 2014 6:13 pm, edited 1 time in total.

Porphyrogenitos
Lebom
Lebom
Posts: 168
Joined: Thu Dec 22, 2011 1:13 pm
Location: Ohio

Re: Isolating, zero-marking lang: a grammatical scratchpad

Post by Porphyrogenitos »

Phonological considerations

I have a tendency to try to make my conlangs have self-segregating morphologies, so that's where a lot of the loglang-ish aspects come in.

With this language, I am going to accomplish that by having initial stress on all words. Stressed syllables will undergo vowel breaking. The beginnings of words will thus be distinguishable via allophone.

It'll probably be something like this:

/e/ > /ai̯/
/o/ > /au̯/
/i/ > /əi̯/ or /eu̯/
/u/ > /əu̯/ or /oi̯/
/a/ will remain unchanged in stressed syllables, but will become /ə/ in unstressed ones

Even with this, though, this language just kind of a grammatical skeleton I can throw any sort of phonology and lexicon over. Which is a shame, as I enjoy creating phonologies, and I'd like to have a phonology and lexicon that's more closely tied to the grammar of the language.

I know I'm thinking of a pretty simple syllable structure, perhaps with mostly mono- and disyllables, so I've considered creating homophonous pairs that are distinguished by grammatical gender. However, I don't want to require every word to have a gender-bearing determiner, but if I don't, there'll be a large amount of semantic ambiguity. Does anyone have any other ideas to spruce up the grammar of this language in regards to its phonology and vocabulary?

User avatar
Benturi
Sanci
Sanci
Posts: 63
Joined: Sun Jul 04, 2010 5:50 pm

Re: Isolating, zero-marking lang: a grammatical scratchpad

Post by Benturi »

Nice isolang so far. I'm not sure if this is the kind of advice you're asking for, but with your five vowels plus 20 consonants, you'd have 2000 possible CVC root forms and 10000 possible CVCV root forms (assuming no restrictions), which should be enough for a naturalistic conlang. (I'd include CVC roots as a special form of CVCV roots, like say, final -u gets dropped after sonorant or something, but that's just a personal taste.)

Porphyrogenitos
Lebom
Lebom
Posts: 168
Joined: Thu Dec 22, 2011 1:13 pm
Location: Ohio

Re: Isolating, zero-marking lang: a grammatical scratchpad

Post by Porphyrogenitos »

Benturi wrote:Nice isolang so far. I'm not sure if this is the kind of advice you're asking for, but with your five vowels plus 20 consonants, you'd have 2000 possible CVC root forms and 10000 possible CVCV root forms (assuming no restrictions), which should be enough for a naturalistic conlang. (I'd include CVC roots as a special form of CVCV roots, like say, final -u gets dropped after sonorant or something, but that's just a personal taste.)
Thanks! I keep ending up constraining myself in various ways when it comes to the phonology - I don't want consonant clusters, but on the other hand I don't want a lot of vowels. I want a limited amount of coda consonants, but on the other hand I don't really want tone. So one or more of those things will have to give at least a little, haha.

Null subjects and the zero pronoun

As illustrated in a couple of my previous examples, verbs can be used without any object or subject. (I know there's a number of names for this, so I'll just call it a null-subject verb). But I also discussed the presence of a zero pronoun. I've decided I'm going to keep both of those constructions - or, at least, the null-subject will be kept provisionally.

Null-subject verbs will primarily be used for verbs that naturally have no agent, such as "to rain". That is, if I end up choosing to express that and other concepts in such a manner. Their other major use will be for deverbal constructions referring to the abstract concept or general notion of a particular action, rather than any particular act carried out or experienced by any individual or group. Some examples would be:
DEV see
seeing, sight

DEV build
building, construction
The zero pronoun will, however, be used in constructions approximating the passive voice, among other situations. Its meaning will sometimes be closer to the English words "someone" or "something" than to a true nullity. It can take the place of any noun, including those modified by prepositional particles. For example:
give 0.NOM 3.s DAT 0.OBL
It was given to someone.
As you can see, the zero pronoun actually distinguishes between the nominative and oblique cases. This is the only case distinction in the language. I came up with it to solve a little problem I will describe in the section on relative pronouns.

A null subject cannot be used when the verb takes an object, as this would result in the object being parsed as a subject (since word order is all that distinguishes them). The zero pronoun must be used in this situation.

The relative clause

As stated earlier, relatives clauses occur when a noun is immediately followed by a verb, which never happens in any other circumstance.

A resumptive particle is used before the pronoun in the relative clause that refers to the modified noun. However, a series of rules exist for the eliding of both the resumptive particle and the referent pronoun.

First rule of elision: If the referent pronoun is the first pronoun in the relative clause, then the resumptive particle may be elided.

Thus, the following underlying phrase:
woman see RES 3.s 1.s
the woman who saw me
Becomes on the surface:
woman see 3.s 1.s
Second rule of elision: If the referent pronoun is the only non-zero pronoun in the relative clause, it may be elided.

This works in conjunction with the first rule, and allows for a variety of handy constructions that can be used derivationally, such as:

Underlying:
person teach RES 3.s
the person who teaches
Surface:
person teach
the person who teaches, teacher
However, these two rules result in an ambiguity. The following underlying phrases:
woman see RES 3.s child
the woman who saw the child

woman see child RES 3.s
the woman who the child saw
Both reduce to the same surface form:
woman see child
In such a situation, it will always be assumed that the referent pronoun in the relative clause is the subject. If the referent pronoun is the object, then a special case-marked resumptive pronoun is used to disambiguate:
woman see child RES.obj
the woman who the child saw
Gosh, I hope all this makes sense.

User avatar
kanejam
Avisaru
Avisaru
Posts: 257
Joined: Sat Mar 30, 2013 1:16 pm
Location: New Zealand

Re: Isolating, zero-marking lang: a grammatical scratchpad

Post by kanejam »

Porphyrogenitos wrote:First rule of elision: If the referent pronoun is the first pronoun in the relative clause, then the resumptive particle may be elided.
Why not just change this rule to state that the 'if the referent pronoun is the first noun phrase' which would make more sense and prevent the ambiguity without relying on a weird resumptive pronoun.
If you cannot change your mind, are you sure you have one?

Here's a thread on Oscan.

Porphyrogenitos
Lebom
Lebom
Posts: 168
Joined: Thu Dec 22, 2011 1:13 pm
Location: Ohio

Re: Isolating, zero-marking lang: a grammatical scratchpad

Post by Porphyrogenitos »

kanejam wrote:
Porphyrogenitos wrote:First rule of elision: If the referent pronoun is the first pronoun in the relative clause, then the resumptive particle may be elided.
Why not just change this rule to state that the 'if the referent pronoun is the first noun phrase' which would make more sense and prevent the ambiguity without relying on a weird resumptive pronoun.
After revisiting this little language and giving it some thought, I think I've just decided to remove the zero pronoun entirely, and instead add a passivizing particle that will be placed immediately before the verb. I think this will take care of a great deal of ambiguity and add some extra functionality to the deverbalizing particle. I'm also going to keep the resumptive particle, but revise the rule regarding elision so that the referent pronoun and its resumptive particle may both be elided together if and only if they are the first noun clause of the relative clause, kind of like what you suggested.

So we have this relative clause here:
woman love man RES 3.s
"the woman who the man loved [her]"

But the pronoun cannot be elided, as it is not the first noun clause of the relative clause. To enable its elision, the clause must be rephrased as a passive:
woman PAS love man
"the woman who was loved by the man"

Essentially, the passivizer switches the positions of the subject and the object.

Some solid ideas about phonology coming...sometime.

Porphyrogenitos
Lebom
Lebom
Posts: 168
Joined: Thu Dec 22, 2011 1:13 pm
Location: Ohio

Re: Isolating, zero-marking lang: a grammatical scratchpad

Post by Porphyrogenitos »

I just realized that "switching the place of the subject and object" doesn't actually create a true passive, because it doesn't demote the subject and promote the object, but I'm just going to call the particle that.

So today I was fiddling around and decided to just start writing out syllables for some sample sentences, without any definite phonology in mind. And what I ended up going with was something rather Tangut-inspired. Ultimately, in the battle between monosyllabicity and simple syllable structure, monosyllabicity won out.

Tangut looks something like this:
dja2-tji1-thjowr2-mju2 gjii2 bju1

dja1-mji1-khwej2-mo2
(From here.)

So after writing out some loose ideas, I formalized them into the following syllable structure:

Image

I suppose it's also somewhat Cantonese-inspired?

For polysyllabic words - which will only feature in loanwords, mainly proper nouns - I think I will permit word-medially any consonant or consonant cluster that is permitted word-initially. I may also allow word-medial coda consonants to be followed by a single word-initial consonant, as well. Stress will always fall upon the first syllable in the case of polysyllabic words.

Now I'm trying to think of/think up an attractive script of some sort to write this language in.

User avatar
Nortaneous
Sumerul
Sumerul
Posts: 4544
Joined: Mon Apr 13, 2009 1:52 am
Location: the Imperial Corridor

Re: Tangut-ish isolating lang: a scratchpad

Post by Nortaneous »

what Tangut looks like depends heavily on the reconstruction

go read Amritas, a conlang based on that would be *interesting*
Siöö jandeng raiglin zåbei tandiüłåd;
nää džunnfin kukuch vklaivei sivei tåd.
Chei. Chei. Chei. Chei. Chei. Chei. Chei.

Post Reply