Random ideas: Morphosyntax

A forum for all topics related to constructed languages
Cellular Automaton
rupestrian
rupestrian
Posts: 11
Joined: 08 Apr 2022 01:33

Re: Random ideas: Morphosyntax

Post by Cellular Automaton »

I've been thinking of a system where (like several Asian languages I think) every noun is preceded by a classifier noun such as "person", "animal", etc. Hopefully unlike those languages, while the noun root never declines, the classifier can be inflected for number, case, etc. You could also do a similar thing for verbs, where an adverb-like part of speech applies to every verb and takes prefixes or suffixes for tense, aspect, mood, et cetera. (Actually, in the language that I considered using this for was, verbs still conjugated for mood, but the "adverb" would have taken tense and aspect.) I'm not sure how naturalistic this is, but I feel like it's the most original system I've come up with.

I also had an idea with a word order that involved the verb appearing multiple times with different TAM. For example, "cows eat grass eat". Not sure what I could actually convey with this system, and I feel like a replacement for the second verb would develop, some kind of "proverb" or something, but I feel like if I figured it out, it could be an interesting idea.

I don't know as much about morphology or syntax as I do phonology, so I probably ran into ANADEW and/or misused some terminology somewhere just now. Oh well. Do these ideas seem interesting?
Last edited by Cellular Automaton on 04 May 2022 21:24, edited 1 time in total.
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Random ideas: Morphosyntax

Post by Creyeditor »

Interesting ideas. The first one looks a bit like articles (which often encode gender/noun class and case). There seem to be some languages with morphological case abd obligatory case marking (according to WALS) but I don't know how these systems work.

A finite adverb sounds a bit like an auxiliary verb with extra steps to be honest. What could be a reason for considering these things adverbs? Do they encode manner as their core meaning? Do they occur in the same position that you find adverbs in?

The third idea is usually called Verb Doubling, IINM. I haven't seen it used for TAM though, great idea.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
Cellular Automaton
rupestrian
rupestrian
Posts: 11
Joined: 08 Apr 2022 01:33

Re: Random ideas: Morphosyntax

Post by Cellular Automaton »

Creyeditor wrote: 04 May 2022 19:26 [...] sounds a bit like an auxiliary verb with extra steps[...]
That's exactly what I meant, just didn't know the term, so thank you!
User avatar
Sequor
sinic
sinic
Posts: 438
Joined: 30 Jun 2012 06:13

Re: Random ideas: Morphosyntax

Post by Sequor »

Cellular Automaton wrote: 04 May 2022 17:19 I've been thinking of a system where (like several Asian languages I think) every noun is preceded by a classifier noun such as "person", "animal", etc. Hopefully unlike those languages, while the noun root never declines, the classifier can be inflected for number, case, etc.
I've never heard of such article-like classifiers before, so congratulations (to the extent that my knowledge of morphosyntax typology is worth anything [xD] ).
Creyeditor wrote: 04 May 2022 19:26 Interesting ideas. The first one looks a bit like articles (which often encode gender/noun class and case). There seem to be some languages with morphological case abd obligatory case marking (according to WALS) but I don't know how these systems work.
What WALS feature(s) were you looking at? "Morphological case and obligatory case marking" sounds just like Latin or Russian... I'm sure I'm misunderstanding something. (I tried to find the relevant WALS feature but couldn't find it.)
hīc sunt linguificēs. hēr bēoþ tungemakeras.
User avatar
Omzinesý
mongolian
mongolian
Posts: 4082
Joined: 27 Aug 2010 08:17
Location: nowhere [naʊhɪɚ]

Re: Random ideas: Morphosyntax

Post by Omzinesý »

Cellular Automaton wrote: 04 May 2022 17:19 I've been thinking of a system where (like several Asian languages I think) every noun is preceded by a classifier noun such as "person", "animal", etc. Hopefully unlike those languages, while the noun root never declines, the classifier can be inflected for number, case, etc.
I'm sure some language has inflecting classifiers.
Yidiny has the category of generic nouns for example. Bantu languages code tense in noun class markers.
You could also do a similar thing for verbs, where an adverb-like part of speech applies to every word and takes prefixes or suffixes for tense, aspect, mood, et cetera. (Actually, in the language that I considered using this for was, verbs still conjugated for mood, but the "adverb" would have taken tense and aspect.) I'm not sure how naturalistic this is, but I feel like it's the most original system I've come up with.
Every word, including nouns, would be inflected for tense?
I also had an idea with a word order that involved the verb appearing multiple times with different TAM. For example, "cows eat grass eat". Not sure what I could actually convey with this system, and I feel like a replacement for the second verb would develop, some kind of "proverb" or something, but I feel like if I figured it out, it could be an interesting idea.
There would be much redundancy if you repeat the verb many times. But, maybe a lang with few verbs that can thus be short. It could also be that only some verbs are repeated and some are replaced by a shorther auxiliary. "Do" as a stem for the negation marker in English is a bit like that.
My meta-thread: viewtopic.php?f=6&t=5760
Cellular Automaton
rupestrian
rupestrian
Posts: 11
Joined: 08 Apr 2022 01:33

Re: Random ideas: Morphosyntax

Post by Cellular Automaton »

Omzinesý wrote: 04 May 2022 21:03 Every word, including nouns, would be inflected for tense?
I meant verb, not word. Silly me. That is another interesting idea, though, nouns declining for tense - maybe the subject could take tense instead of the verb? Then the object (and thus sentence) would automatically have the same tense as the subject, unless you marked it differently for some reason. There's some interesting things you could express like that: "John.REM.PST catch rabbit.REC.PST" could imply that John set up a trap which just caught a rabbit. That system would also make discussing time travel easier (especially if you kept normal verb tenses as well), if you wanted to do that a lot for some reason.
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Random ideas: Morphosyntax

Post by Creyeditor »

Sequor wrote:
Creyeditor wrote: 04 May 2022 19:26 Interesting ideas. The first one looks a bit like articles (which often encode gender/noun class and case). There seem to be some languages with morphological case abd obligatory case marking (according to WALS) but I don't know how these systems work.
What WALS feature(s) were you looking at? "Morphological case and obligatory case marking" sounds just like Latin or Russian... I'm sure I'm misunderstanding something. (I tried to find the relevant WALS feature but couldn't find it.)
Sorry, I meant to say obligatory classifiers and morphological case. I looked at features 49A and 55A.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
User avatar
Sequor
sinic
sinic
Posts: 438
Joined: 30 Jun 2012 06:13

Re: Random ideas: Morphosyntax

Post by Sequor »

Ah, I see. I'm looking at the languages in question, and I see even some famous ones are included:

Languages with cases and obligatory use of numeral classifiers:
- Burmese (8-9 cases)
- Garo (8-9 cases; Sino-Tibetan; in India)
- Semelai (3 cases; Austroasiatic; in Malaysia)
- Nivkh (8-9 cases)
- Korean (6-7 cases)
- Japanese (8-9 cases)
- Tlingit (8-9 cases)
- Yagua ("exclusively borderline case marking"; Peba-Yaguan; in Peru)

Languages with cases and optional use of numeral classifiers:
- Hungarian (10+ cases)
- Turkish (6-7 cases)
- Southern Tati (2 cases; Iranian IE)
- Persian (2 cases)
- Maybrat (2 cases; West Papuan)
- Sanuma (2 cases; Yanomam family; in Brazil/Venezuela)

Well, we can eliminate Burmese, Korean, Japanese and Persian because (from personal knowledge) their case marking doesn't mix with their numeral classifiers in morphological fusion, but I wonder about the others.
hīc sunt linguificēs. hēr bēoþ tungemakeras.
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Random ideas: Morphosyntax

Post by Creyeditor »

I just checked a Semelai grammar and it has optional numeral classifiers and a small set of case clitics with a very restricted distribution. So no luck with case marked classifiers.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
Solarius
roman
roman
Posts: 1173
Joined: 30 Aug 2010 01:23

Re: Random ideas: Morphosyntax

Post by Solarius »

Made a little system of noun class/number marking which I enjoy. Might become something but probably not.

Nouns in this language are divided into four noun classes: Feminine, Masculine, Animate, and Inanimate. Noun Class assignment is pretty firmly semantic; masculine is used for human men (plus assorted very high animacy male non-human entities, like gods), feminine is used for human women (plus, as with the masculine, various assorted high animacy female non-humans), the animate is for relatively high animacy non-humans (in practice mostly terrestrial animals) and the inanimate for everything else. Some of the wrinkles of the distinction between the latter two will make sense in the light of the way the number system works.

This language has a productive singular-plural distinction, but the nature of the inflection varies depending on noun class. This language also has a simple four tone system of low/high/rising/falling; this is indicated with no accent, macron, acute, and grave respectively.

In the feminine, the singular is unmarked and the plural is marked with -lil.

nàbí --> nàbílil
woman --> women

In the masculine, the singular is unmarked and the plural is marked with the suffix -(a)lp.

iq'e --> iq'elp
man --> men

gèrzak --> gèrzakalp
monk --> monks

In the animate, the singular is unmarked but the plural is marked, not with a suffix but with raising the tone of the final syllable. The low tone becomes a rising tone, while rising and falling tones become high. The high tone doesn't change, but very few animate nouns have high tone in this environment because it tends to get reanalyzed.

nāya --> nāyá
cat --> cats

è --> ē
cow --> cows

Finally, in the inanimate, the plural is unmarked and there is instead a singulative marker -g, which exerts a "depressor" effect on the tone of previous vowels if there's not a consonant coda sitting in the way--high tone becomes falling and rising tone becomes low, while falling and low don't change. (This is a normal phonological process in this language.)

lāmb --> lāmbg
rivers --> river

t'ùsū --> t'ùsùg
rice --> grain of rice

Historically there's been a tendency for inanimates where the singular is more salient to be filtered into the animate.

A fun feature of this system is that some nouns can be declined in multiple genders.

There are a few nouns which can vary between masculine and feminine, usually referring to social roles which can be occupied by both men and women, like hílyú "archer" or sàblyu "elder."

There is also a smaller, but rather exciting group of nouns which can be conjugated in all four noun classes. These are almost all body parts--a residue of inalienable possession marking in the protolanguage. Nouns like lílí "(front of) neck" can be masculine (to indicate a man's neck), feminine (for a woman's neck), animate (for the neck of, say, an ostrich) or inanimate (for the discussion of the concept of necks more generally, or for the narrow tapering part of an inanimate object.)

One more thing: adjectives don't exist as a separate class in this language, but rather are forms of nouns linked to their heads in an ezafe-esque construction used with genitives also. This means that adjectival nouns have their own gender, which is usually animate but not always if there are, say, strong associations to it, like how handsome in English is usually for men.
Salmoneus
MVP
MVP
Posts: 3031
Joined: 19 Sep 2011 19:37

Re: Random ideas: Morphosyntax

Post by Salmoneus »

Weirdly, I suddenly had the idea of an inanimate gender with an unmarked plural and marked singulative just a few weeks ago!
[much less interesting than your system, though]
Sevly
rupestrian
rupestrian
Posts: 20
Joined: 18 Jan 2012 21:39
Contact:

Re: Random ideas: Morphosyntax

Post by Sevly »

Very neat! Much enjoyed both the overall system and the number of little nuances
Solarius
roman
roman
Posts: 1173
Joined: 30 Aug 2010 01:23

Re: Random ideas: Morphosyntax

Post by Solarius »

Teşekkür ederim!
User avatar
Omzinesý
mongolian
mongolian
Posts: 4082
Joined: 27 Aug 2010 08:17
Location: nowhere [naʊhɪɚ]

Re: Random ideas: Morphosyntax

Post by Omzinesý »

The language
i) codes aspect with the case of the direct object, like Finnic.
ii) all verbs have an object (to step a step etc.)

That could exclude statives, like to sleep or to be tall, that cannot be perfective and aren't maybe even verbs in the language.

If words are short, it does not make sentences too long.
My meta-thread: viewtopic.php?f=6&t=5760
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Random ideas: Morphosyntax

Post by Creyeditor »

Maybe one of the object cases could be zero marked and you could drop the bare object in the context of this aspect. This would make some sentences shorter.

A possible reanalysis could be that some intransitive verbs have a bipartite structure with an infixed aspect marker whereas transitive verbs have an aspect marker on the object noun.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
Porphyrogenitos
sinic
sinic
Posts: 401
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Re: Random ideas: Morphosyntax

Post by Porphyrogenitos »

An idea I've had for a while now:

A language where basically all of the verbal morphosyntax is borne by a single, essentially opaque or suppletive particle inflected for number, person, and possibly gender or animacy of subject/agent, object/patient, and possibly oblique arguments; transitivity and other valency categories; and possibly tense and aspect. Naturally, it would have dozens and dozens of inflected forms, and would also exhibit a considerable amount of syncretism. This is inspired by languages that make heavy use of auxiliary verbs as the primary site of verbal inflection, but also the pronominal prefixes of the Iroquoian languages. Also like Iroquoian pronominal prefixes, this particle would be main way of expressing pronominal reference - independent pronouns would be rare.

Diachronically, I am thinking that they probably would have originated from a small set of auxiliary verbs, which of course historically would have originated from full, lexical verbs. Over time, as syncretism and analogy in the system progressed, and as sound change took its toll, the set of inflected auxiliary forms would become so restricted (yet also convoluted) that they would lose their identity as separate lexical items and essentially come to operate as the rather monstrous paradigm of a single preverbal particle. The actual lexical verbs would either be completely bare of inflection, or perhaps receive a single gerund-like inflection in certain contexts or constructions, like the English -ing or German ge-...-en.

I imagine there might be significant traces of lexical distinction in the particle system, e.g. through suppletion for tense or valency, or perhaps highly agentive transitive verbs would take certain alternate forms of the particle, but the speakers would still probably not conceive of the particle as "meaning" anything other than the collective expression of person, number, valency, tense, etc. The bare lexical verb following the particle would probably remain syntactically and phonetically free from the particle; e.g. adverbs could be inserted between the particle and the lexical verb, and multiple lexical verbs could be chained together to form serial verbs, and so on.

I haven't thought of any example forms for this system, and don't have much of an idea about the rest of the language... I assume it would go the head-marking route and lack case marking on nouns. Although paradoxically, I feel like it could be fun to try deriving this language from an ancient or reconstructed IE lang, considering that it would basically be taking IE synthesis to an extreme...
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Random ideas: Morphosyntax

Post by Creyeditor »

A conlang where morphological marking primarily expresses if something is a modifier or is modified by something. Nouns/adjectives/verbal nouns could have a construct state marker if they head a complex noun phrase. Nouns/adjective/verbs could have a attributive marker if the modify a noun in a compound/attributive adjective/participle kind of way. Clauses/verbs could be marked for being (in) main clauses, embedded clauses, or adjunct clauses. And adverbs could definitely be derived from adjectives or verbs in an English-y kind of way.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
Salmoneus
MVP
MVP
Posts: 3031
Joined: 19 Sep 2011 19:37

Re: Random ideas: Morphosyntax

Post by Salmoneus »

Porphyrogenitos wrote: 10 Nov 2022 01:21 An idea I've had for a while now:

A language where basically all of the verbal morphosyntax is borne by a single, essentially opaque or suppletive particle inflected for number, person, and possibly gender or animacy of subject/agent, object/patient, and possibly oblique arguments; transitivity and other valency categories; and possibly tense and aspect. Naturally, it would have dozens and dozens of inflected forms, and would also exhibit a considerable amount of syncretism. This is inspired by languages that make heavy use of auxiliary verbs as the primary site of verbal inflection, but also the pronominal prefixes of the Iroquoian languages. Also like Iroquoian pronominal prefixes, this particle would be main way of expressing pronominal reference - independent pronouns would be rare.

Diachronically, I am thinking that they probably would have originated from a small set of auxiliary verbs, which of course historically would have originated from full, lexical verbs. Over time, as syncretism and analogy in the system progressed, and as sound change took its toll, the set of inflected auxiliary forms would become so restricted (yet also convoluted) that they would lose their identity as separate lexical items and essentially come to operate as the rather monstrous paradigm of a single preverbal particle. The actual lexical verbs would either be completely bare of inflection, or perhaps receive a single gerund-like inflection in certain contexts or constructions, like the English -ing or German ge-...-en.

I imagine there might be significant traces of lexical distinction in the particle system, e.g. through suppletion for tense or valency, or perhaps highly agentive transitive verbs would take certain alternate forms of the particle, but the speakers would still probably not conceive of the particle as "meaning" anything other than the collective expression of person, number, valency, tense, etc. The bare lexical verb following the particle would probably remain syntactically and phonetically free from the particle; e.g. adverbs could be inserted between the particle and the lexical verb, and multiple lexical verbs could be chained together to form serial verbs, and so on.

I haven't thought of any example forms for this system, and don't have much of an idea about the rest of the language... I assume it would go the head-marking route and lack case marking on nouns. Although paradoxically, I feel like it could be fun to try deriving this language from an ancient or reconstructed IE lang, considering that it would basically be taking IE synthesis to an extreme...

I've also thought about this a bit.
If you're going the verbal route, one fun thing to do could be verb category indexing on the particle. It's not unusual for different verbs to take different auxiliaries - look at the lexical definition of which auxiliary is used with an abstract noun in English (take, make, do, have, etc).


The route I would like to go, however, if I ever get around to it, is more minimalist, and skip the auxiliary altogether. This can be done by having a different word order for pronouns and for full nouns, and fossilising the former with cliticisation. So "The woman ate the dog" > "Ate the woman the dog", but also "She it ate" (there has been verb-fronting, but pronouns move with the verb). Combining these two forms > "She it ate the woman the dog" (i.e. obligatory pronouns, or else pronominal clauses supplemented with postposed comments, same result). If 'it' is generally a clitic (or there is second-position cliticisation), > "Shet ate the woman the dog", with "shet" = 3sF subject, 3sInan object. [this would be needed because the language would have lost all existing role marking on both verbs and noun, but would also have fairly free movement for emphass, which would also prevent this particle from adhering to the verb as a prefix]. But I've also toyed with having auxiliary elements pop up - for instance, confusion between this construction and an equative with the copula in a cleft construction where the copula is null in the present, but then having the copula show up in other tenses, etc.
Porphyrogenitos
sinic
sinic
Posts: 401
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Re: Random ideas: Morphosyntax

Post by Porphyrogenitos »

Salmoneus wrote: 12 Nov 2022 13:44

I've also thought about this a bit.
If you're going the verbal route, one fun thing to do could be verb category indexing on the particle. It's not unusual for different verbs to take different auxiliaries - look at the lexical definition of which auxiliary is used with an abstract noun in English (take, make, do, have, etc).
That's a neat idea, thanks.
Salmoneus wrote: 12 Nov 2022 13:44 The route I would like to go, however, if I ever get around to it, is more minimalist, and skip the auxiliary altogether. This can be done by having a different word order for pronouns and for full nouns, and fossilising the former with cliticisation. So "The woman ate the dog" > "Ate the woman the dog", but also "She it ate" (there has been verb-fronting, but pronouns move with the verb). Combining these two forms > "She it ate the woman the dog" (i.e. obligatory pronouns, or else pronominal clauses supplemented with postposed comments, same result). If 'it' is generally a clitic (or there is second-position cliticisation), > "Shet ate the woman the dog", with "shet" = 3sF subject, 3sInan object. [this would be needed because the language would have lost all existing role marking on both verbs and noun, but would also have fairly free movement for emphass, which would also prevent this particle from adhering to the verb as a prefix]. But I've also toyed with having auxiliary elements pop up - for instance, confusion between this construction and an equative with the copula in a cleft construction where the copula is null in the present, but then having the copula show up in other tenses, etc.
Yeah, I suppose that's a simpler and more obvious approach, to derive them from pronominal portmanteaus.

I think I had imagined that the copula would be integrated into this system, via a stative category or something. Like I guess a monovalent "1sSubj tall" meaning "I am tall", and a divalent "1sSubj-3sMascObj tall" being "I am [a/the] tall [one]" (kind of an unnecessary distinction, I suppose?) and then "1sSubj-3sMascObj" on its own being "I am him". Hmm, but I suppose this might go against my idea that there wouldn't be any lexically distinct forms, if one functions as a copula and the others don't.

Another neat thing might be to cut down on the exuberant proliferation of inflected forms for the particle, and have it exhibit heavy syncretism, like the German definite article, whereby lots of forms are identical, but they're distributed in such a manner that ambiguities are mostly avoidable - like "1sSubj-2sObj" and "3sFemSubj-1pObj-past" could be identical, and speakers would probably be able to tell from context whether it's "I'm listening to you" or "She listened to us."
User avatar
Omzinesý
mongolian
mongolian
Posts: 4082
Joined: 27 Aug 2010 08:17
Location: nowhere [naʊhɪɚ]

Re: Random ideas: Morphosyntax

Post by Omzinesý »

This idea started from my thinkin about a future English but it could be used in any conlang.

The only agreement suffix of English verbs is the -s of sg3.
I started thinking if developed to a marker of the clause having a full-NP subject.

He do.
They do.
The child does.
The children does.

The past suffix, if we are still thinking about English the -ed, takes the same function in the past tense, while pronouns develop their own past forms.

Heed do.
Theyed do.
The child did.
The children did.

That is, past tense is coded in the pronouns if there are pronouns and in the verb if there is a verb. Similarly, verbs have a marker of having a full-NP subject in the present.
My meta-thread: viewtopic.php?f=6&t=5760
Post Reply