Conlang Random Thread

[Whimemsz]

.

[zompist]

Fair enough— I though the syncope would sometimes reveal unexpected vowels, but I didn't check. It looks like prefixes have that magical power, but not suffixes.

[bradrn]

I doubt there's any languages where there's literally no identifiable patterns to plural formation at all.

This is the problem - it is almost completely random. There are some vague, unhelpful patterns - for instance, words ending in geminates add a vowel to form the plural (but which vowel?), and words ending in some letters tend to lenite (but only sometimes). But for most words, the plural is random.

The problem I have may be demonstrated by these simple sound changes:

1. Vowels disappear at the end of a word. (You'd think that wouldn't be too common, but it's fairly widespread according to the Index Diachronica.)
2. Glottal stops disappear at the end of a word.

Now the ancestor forms plurals using -ʔe (for words ending in a vowel), and -e for words ending in a consonant. Now take the following words and their plurals:

Code: Select all

ansɯ ansɯʔe
ɣad͡za ɣad͡zaʔe
onɣip onɣipe
mint͡sa mint͡saʔe
leræ leræʔe
xomin xomine

Now applying the changes, you get:

Code: Select all

ans ansɯ
ɣad͡z ɣad͡za
onɣip onɣip
mint͡s mint͡sa
ler leræ
xomin xomin

See the problem? The final vowel (if there even is one) is unpredictable, since it's the last vowel of the ancestor word (if it had one). And the actual descendants (applying the full set of sound changes, not just the ones above) are even worse:

Code: Select all

anː anːɨ
jaz jaza
onːɨp onːɨf
mɨnː mɨnːa
lir lire
jomɨn jomɨn

In fact, the plurals are so badly behaved I think I may just take Akangka's suggestion and form it periphrastically. But it would be nice to have a more fusional solution...

[Xwtek]

I doubt there's any languages where there's literally no identifiable patterns to plural formation at all.

This is the problem - it is almost completely random. There are some vague, unhelpful patterns - for instance, words ending in geminates add a vowel to form the plural (but which vowel?), and words ending in some letters tend to lenite (but only sometimes). But for most words, the plural is random.

The problem I have may be demonstrated by these simple sound changes:

1. Vowels disappear at the end of a word. (You'd think that wouldn't be too common, but it's fairly widespread according to the Index Diachronica.)
2. Glottal stops disappear at the end of a word.

Now the ancestor forms plurals using -ʔe (for words ending in a vowel), and -e for words ending in a consonant. Now take the following words and their plurals:

Code: Select all

ansɯ ansɯʔe
ɣad͡za ɣad͡zaʔe
onɣip onɣipe
mint͡sa mint͡saʔe
leræ leræʔe
xomin xomine

Now applying the changes, you get:

Code: Select all

ans ansɯ
ɣad͡z ɣad͡za
onɣip onɣip
mint͡s mint͡sa
ler leræ
xomin xomin

See the problem? The final vowel (if there even is one) is unpredictable, since it's the last vowel of the ancestor word (if it had one). And the actual descendants (applying the full set of sound changes, not just the ones above) are even worse:

Code: Select all

anː anːɨ
jaz jaza
onːɨp onːɨf
mɨnː mɨnːa
lir lire
jomɨn jomɨn

In fact, the plurals are so badly behaved I think I may just take Akangka's suggestion and form it periphrastically. But it would be nice to have a more fusional solution...

But it's just simple disfixation on a singular noun. In this case, I think the final vowel on the plural will be treated as suffix. The vowel used is analogized by the semantics of the noun. For example I assign the meaning of each noun as this:

Code: Select all

anː anːɨ man
jaz jaza woman
onːɨp onːɨp house
mɨnː mɨnːa commoner
lir lire tower
jomɨn jomɨn castle

Then it will have a analogical change into

Code: Select all

anː anːa man
jaz jaza woman
onːɨp onːɨp house
mɨnː mɨnːa commoner
lir lir tower
jomɨn jomɨn castle

Now, -a is the human plural, and no change for building plural.

BTW, how can onɣip > onːɨp and onːɨf depending on plurality

[zompist]

I don't really see a problem. Unless you have, like, 20 vowels, you only have a small number of suffixes. Having to remember two forms for each word is quite naturalistic.

If it bothers you, though, why not refine that sound change that deletes all final vowels? Delete just some of them instead, e.g. all but high vowels.

[Nortaneous]

It's fine. Plurals in Seri are completely unpredictable:

	Singular	Plural
lobster	ptkamn	ptkamn
man	ktam	ktamkʷ
plant	ʔeʔe	ʔeʔet
fire	ʔamak	ʔamakx
arrow point	ʔaꭓ	ʔaꭓɬk
girl	ʃakaːm	ʃakaːmaɬk
hematite	ꭓpaːʔxʷ	ꭓpaːʔxʷɬka
estuary	ꭓtaːsi	ꭓtaːsitox
fox	ꭓeːs	ꭓeːstax
bobcat	nop	nopꭓam
roadrunner	ʃaːp	ʃaːpkox
ear	isɬitꭓ	isɬitꭓkox
road	ʔaːʔo	ʔaːʔoɬkam
black vulture	koɬkiːmet	koɬkiːmt
grasshopper	kaːtk	kaːtxk
Heermann's gull	kanokni	kanokʃox
3SG.POSS-placenta	iːꭓʷni	iːꭓʷʃax
cave	paːix	paːiɬk
round stingray	kotx	kotɬka
month	iːʃaꭓ	iːʃtoꭓ
heron	sɬeːkox	sɬeːkʷɬ
heart	imoʃ	imxʷk
knife	eːnim	eːntʃ
dog	ʔaꭓʃ	ʔaꭓaka
foot	itoaː	itoit
knee	iɸɬk	iɸɬxox
stone	ʔast	ʔasatox
person	kmiːke	komkaːk
woman	kmaːm	kmaxiːk
thing	ʃiːꭓ	ꭓiːka
Yaqui	yekim	yektʃ
shirt (loan)	kaːmiʃ	kaːmixox

[bradrn]

But it's just simple disfixation on a singular noun.

I suppose it is, but disfixation is rather rare, isn't it...

In this case, I think the final vowel on the plural will be treated as suffix. The vowel used is analogized by the semantics of the noun. For example I assign the meaning of each noun as this:

Code: Select all

anː anːɨ man
jaz jaza woman
onːɨp onːɨp house
mɨnː mɨnːa commoner
lir lire tower
jomɨn jomɨn castle

Then it will have a analogical change into

Code: Select all

anː anːa man
jaz jaza woman
onːɨp onːɨp house
mɨnː mɨnːa commoner
lir lir tower
jomɨn jomɨn castle

Now, -a is the human plural, and no change for building plural.

Gosh, that's a clever idea! I never thought of creating genders via analogy!

BTW, how can onɣip > onːɨp and onːɨf depending on plurality

The exact changes are:

Singular	Plural	(sound change)
onɣip	onɣipe
onɣip	onɣife	stop lenition between vowels
onɣɨp	onɣɨfi	vowel shift
onːɨp	onːɨfi	cluster simplification
onːɨp	onːɨf	final vowel deletion

It's fine. Plurals in Seri are completely unpredictable:

lobster	ptkamn	ptkamn
man	ktam	ktamkʷ
plant	ʔeʔe	ʔeʔet
fire	ʔamak	ʔamakx
arrow point	ʔaꭓ	ʔaꭓɬk
girl	ʃakaːm	ʃakaːmaɬk
hematite	ꭓpaːʔxʷ	ꭓpaːʔxʷɬka
estuary	ꭓtaːsi	ꭓtaːsitox
fox	ꭓeːs	ꭓeːstax
bobcat	nop	nopꭓam
roadrunner	ʃaːp	ʃaːpkox
ear	isɬitꭓ	isɬitꭓkox
road	ʔaːʔo	ʔaːʔoɬkam
black vulture	koɬkiːmet	koɬkiːmt
grasshopper	kaːtk	kaːtxk
Heermann's gull	kanokni	kanokʃox
3SG.POSS-placenta	iːꭓʷni	iːꭓʷʃax
cave	paːix	paːiɬk
round stingray	kotx	kotɬka
month	iːʃaꭓ	iːʃtoꭓ
heron	sɬeːkox	sɬeːkʷɬ
heart	imoʃ	imxʷk
knife	eːnim	eːntʃ
dog	ʔaꭓʃ	ʔaꭓaka
foot	itoaː	itoit
knee	iɸɬk	iɸɬxox
stone	ʔast	ʔasatox
person	kmiːke	komkaːk
woman	kmaːm	kmaxiːk
thing	ʃiːꭓ	ꭓiːka
Yaqui	yekim	yektʃ
shirt (loan)	kaːmiʃ	kaːmixox

Wow. Why hasn't that been analogised away by now? Now what I'd like to know is akam chinjir's query:

How do they form the plurals of newly-borrowed words?

I suppose that in comparison to that a few random vowels and suffixes is tame.

[Nortaneous]

Wow. Why hasn't that been analogised away by now?

Languages can support a lot of irregularity. And there are some patterns - I didn't type up the whole chart.

Words with suppletive plurals are common ('Seri person', 'woman', 'day', 'thing'); a lot of words ending in -k take infixed -x-; there are some common final consonant alternations (although none of them make much phonological sense: -x > -ɬ, -n > -ʃ, -m > -tʃ); and so on.

I don't know if these words have been analyzed into declensions or anything, but my guess is that it's a lot like Germanic strong verbs, only more so: there are a lot of regular patterns which are partially predictable from the citation form, and truly irregular words will tend to be common.

Now what I'd like to know is akam chinjir's query:

How do they form the plurals of newly-borrowed words?

Here are some loanwords:

ladder	kaskareːra	kaskareːrax
donkey	ʔoʔra	ʔoʔrax
shoe	paʃaːto	paʃaːtox
bedsheet	soaːno	soaːnox
goat	seaːto	seaːtox
blanket	saraːpi	saraːpitax
boat	kanoaː	kanoaːtax
soldier	santaːr	santaːrox
chicken	toːtar	toːtarox
melon	meroːn	meroːntax
lasso	ret	retkox

Spanish surnames can also take pluralization. This is completely regular: surnames ending in a sibilant take a zero marker, surnames ending in a vowel take /-x/, and surnames ending in a non-sibilant consonant take /-tax/.

[bradrn]

Thanks for all the suggestions everyone - they've really helped! Looking back through them, it would appear that I have a few options:

1. Simply preserve the irregularities (as in Seri or Welsh)
2. Preserve the irregularities, but treat it as disfixation
3. Semantics-based analogy
4. Abandon the morphological plural altogether and replace it with a periphrastic construction

(1) and (4) are very plausible. (2) is plausible, but uncommon: disfixation seems to have a very limited distribution in natlangs. I would say (3) is the best suggestion, but unfortunately it doesn't work in my case: due to the exact nature of my problem (detailed above), the plural transformation is unrelated to semantics. I think at this point I'll just go with option (1) as the best overall.

[statelessnation]

¿What on earth would you call a system like this:

Each time a noun is introduced into a sentence the last vowel is assigned a final tone (lexically all nouns have an atonic [not neutral!] final vowel, including proper nouns), either: /˧ ˥ ˩ ˥˩/ which equate to the 'identifiers' {α β γ δ}.

There 24 personal pronouns: three persons × two numbers × four identifiers.

Again the personal pronouns have six base forms, with the 'identifier' being obligatorily marked, as the four tones on the final vowel.

The first three identifiers refer to a specific person(s), & until a new noun is introduced using one of those three tones, they will continue to refer to that particular noun when those pronouns are used.

The fourth (-δ) is used to indicate a variety of nouns, and may be used again to refer to someone/something else that hasn't been introduced yet and/or is unlikely to be as regularly referenced again.

The -α -β -γ pronouns are roughly proximal topical, whilst -δ is obviate, and it is often used for generic or indefinite referrants.

These pronouns allow for a unambiguous reflexive voice: repeat the same pronoun in the two relevant positions.

I'll note that there being four identifiers for all pronouns is excessive, but I just for funsies:
• 1.Sg.α = "I", always.
• 1.Sg.β & 1.Sg.γ may be re-Narrative or Royal We.
• 3.Sg.δ is used for impersonal verbs.

1.Pl.β & 1.Pl.γ have some use, whilst "I" has an innate & irremovable high tone, the later two  can be assigned to accompaniments separately (they include the I they are (eventually) adjacent to), thus allowing for an pseudo inclusive-exclusive distinction (all would include the 1.Sg.α but not vice versa).

I imagine this may all seem crazy, but it's inspired by my limited experience with sign language - essentially I'm using a floating tone instead of signing the same noun in a different direction; bit I'm pretty sure I totally misunderstood what was being communicated anyway.

I figured such a system could be fun, but there's a lot to hammer out, not to mention the whole problem about me ultimately preferring naturalism (I love flirting with crazy systems, but they rarely hold me interest for long)

[Xwtek]

¿What on earth would you call a system like this:

Each time a noun is introduced into a sentence the last vowel is assigned a final tone (lexically all nouns have an atonic [not neutral!] final vowel, including proper nouns), either: /˧ ˥ ˩ ˥˩/ which equate to the 'identifiers' {α β γ δ}.

There 24 personal pronouns: three persons × two numbers × four identifiers.

Again the personal pronouns have six base forms, with the 'identifier' being obligatorily marked, as the four tones on the final vowel.

The first three identifiers refer to a specific person(s), & until a new noun is introduced using one of those three tones, they will continue to refer to that particular noun when those pronouns are used.

The fourth (-δ) is used to indicate a variety of nouns, and may be used again to refer to someone/something else that hasn't been introduced yet and/or is unlikely to be as regularly referenced again.

The -α -β -γ pronouns are roughly proximal topical, whilst -δ is obviate, and it is often used for generic or indefinite referrants.

These pronouns allow for a unambiguous reflexive voice: repeat the same pronoun in the two relevant positions.

I'll note that there being four identifiers for all pronouns is excessive, but I just for funsies:
• 1.Sg.α = "I", always.
• 1.Sg.β & 1.Sg.γ may be re-Narrative or Royal We.
• 3.Sg.δ is used for impersonal verbs.

1.Pl.β & 1.Pl.γ have some use, whilst "I" has an innate & irremovable high tone, the later two  can be assigned to accompaniments separately (they include the I they are (eventually) adjacent to), thus allowing for an pseudo inclusive-exclusive distinction (all would include the 1.Sg.α but not vice versa).

I imagine this may all seem crazy, but it's inspired by my limited experience with sign language - essentially I'm using a floating tone instead of signing the same noun in a different direction; bit I'm pretty sure I totally misunderstood what was being communicated anyway.

I figured such a system could be fun, but there's a lot to hammer out, not to mention the whole problem about me ultimately preferring naturalism (I love flirting with crazy systems, but they rarely hold me interest for long)

Sounds like Elkaril's variables. Or my personal marker (I has 1, 1.5, 2, 3, 4, 5, 6, and 0. 1 and 2 is like normal. 1.5 is an inclusive 1. And 3-6 is like your identifier. 0 is impersonal.

[Pabappa]

Are there any known languages that have person agreement between possessed objects and their verbs, like

Bappubo lappapo.
My knife is sharp.

Where both the noun bappub- "knife" and the verb lappap- "sharp; to cut through solid objects" are marked for 1st person? The sentence could also be parsed as "(With) my knife, I am sharp" or even "I have a knife; I am sharp". This is one of the things that I've been doing for a long time but has always in the back of mind still bothered me. There is an alternative construction:

Bappubos lappapa.
My knife is sharp.

Where the additional -s on the possessed noun promotes it to animate status and therefore turns it into a 3rd person noun. But I dont want to have to make every noun animate just to make it the agent of an intransitive verb ... which is what Poswa uses instead of adjectives.

I want to keep both setups grammatical, with the 2nd one being the lesser-used alternative, usually with the implication that the possessed object has some control over what it is and what it does. e.g. one might say in English "hmm, my knife is sharp today" as if it sometimes works well and sometimes struggles. Has anyone seen a setup like my first one, where the verb of the possessed noun must agree in person with the possessor rather than the object itself?

As for why I dont just allow *Bappubo lappapa, with disharmonious person markers, to be honest, I dont have a good answer. It just feels wrong somehow. I guess if I had to try to explain it Id say that since knives and other inanimate objects aren't persons, if they take a person marker that means they inherit their personhood from the possessor and should behave accordingly. Or maybe its because the verbs are only meaningful with an animate possessor, since e.g. the word for sharp is a verb, not just an adjective describing its shape.

[Xwtek]

1. What is difference of obviative system and topic system?
2. How does a language acquire gerund?

[Salmoneus]

1. What is difference of obviative system and topic system?
2. How does a language acquire gerund?

Topic is a pragmatic concept, referring to speech referents that are conceptually 'what the information is about', and in practical terms tend to be referred to repeatedly throughout a passage of speech. A "topic system" indicates that the information structure of sentences is grammatically encoded in such a way that the concept of topic is very important.

Proximate and obviative is a distinction between two third persons. What assigns something to proximate or obviative status is variable, and may include topicalness, but may also include focus, or deictic considerations (physical proximity, possession), or just the order in which things arise in a conversation. Even where proximity is assigned on the basis of topicality, that does not mean there needs to be any broader 'topic system' in the language. However, if there is a topic system, it may interact with proximity - for instance, it may be that only proximate third persons may be allowed to be topics.

[Qwynegold]

In the grammar of the conlang (Liu) I wrote about a couple of pages back, there is nothing written about relative pronouns. I wrote the pronoun section years ago, so I can't remember if it's because I intended the conlang to not have any relative pronouns, or if it's because I never got around to doing that. Considering that I had a ton of pronouns and had written quite a bit about them, I think it's likely that the absence of relative pronouns means that they are not supposed to exist. So then I thought that maybe I can use relative clauses instead. But the way Liu handles relative clauses is already so complicated, it just fries my brain when I try to think about how to translate "I want to work on Liu again, but I'm not sure what else to do to get the grammar more complete." Do you think it's a good idea to somehow use relative clauses instead of pronouns? Like something like "...that* which do to get the grammar more complete".

*Demonstrative pronoun.

[akam chinjir]

So then I thought that maybe I can use relative clauses instead. But the way Liu handles relative clauses is already so complicated, it just fries my brain when I try to think about how to translate "I want to work on Liu again, but I'm not sure what else to do to get the grammar more complete." Do you think it's a good idea to somehow use relative clauses instead of pronouns? Like something like "...that* which do to get the grammar more complete".

I'm not sure I understand your terminology, but it sounds as if you're asking whether to allow headless relative clauses. (Both your examples have relative clauses, and whether to allow headless ones doesn't have anything to do with whether you use relative pronouns to form them, as far as I know.)

[bradrn]

Would it be realistic to have a polypersonal agreement system where the same affixes are used for both agent and patient? If so, how would you know whether it is nominative-accusative or ergative-absolutive, since it's impossible to know whether an intransitive verb is agreeing with its experiencer using the agent or patient affixes?

[Vilike]

Would it be realistic to have a polypersonal agreement system where the same affixes are used for both agent and patient? If so, how would you know whether it is nominative-accusative or ergative-absolutive, since it's impossible to know whether an intransitive verb is agreeing with its experiencer using the agent or patient affixes?

There is also direct-inverse marking, like in Algonquian. Considering the following ad hoc system:

English	Morpheme
I, me	se-
you	ci-
he, she, it	ra-
(inverse marking)	-li
hit	kama
sleep	bara

And a hierarchy of arguments, from the most susceptible to act as an agent to the least susceptible: 1 > 2 > 3 animate > 3 inanimate.

Thus secikama can only ever mean 'I hit you', and racikama 'you hit her'. To subvert the implicit hierarchy, use the inverse morpheme: secikamali 'you hit me', racikamali 'she hits you'.
The hierarchy I gave is only an example. In Algonquian, it's actually 2 > 1 > 3, and for distinguishing between, say, two 3rd person animate arguments, one can use obviation and/or the more refined hierarchy of Navajo.

On the subject of intransitives, you get sebara, cibara, rabara, with straightforward meanings as they have but one argument.

[akam chinjir]

Would it be realistic to have a polypersonal agreement system where the same affixes are used for both agent and patient? If so, how would you know whether it is nominative-accusative or ergative-absolutive, since it's impossible to know whether an intransitive verb is agreeing with its experiencer using the agent or patient affixes?

This also made me think of direct/inverse systems.

One thing about that. There's a tendency to say that anything that isn't nom/acc must be erg/abs, but there's a bunch of ways in which languages differ from nom/acc alignment, and no particular reason to think that those languages thereby have anything substantive in common. (Some fun reading: Delancy's Blue Bird of Ergativity.) I mean, once you've settled your morphosyntactic facts, there's no additional question you need to answer of whether your language is ergative.

[bradrn]

Yes, I'm aware of direct-inverse - I quite like them, in fact - but that wasn't actually the system I'm talking about. Consider the following example morphemes:

	singular	plural
1	-xa	-xe
2	-si	-se
3	-ko	-ke

.

kode	to see
aman	to hear
yes	to exist

Then, in the system I'm describing, all actors (is that the right word?) use the personal affixes above, irrespective of whether they're agent/patient/experiencer, with the only distinguishing feature being the order they appear in the word. Then you could have the following:

kode-xa-si

I see you

ka-si-xa

You see me

aman-ke-xe

They hear us

yes-xe

We exist

As opposed to a system in which agents and patients are marked using different affixes e.g. distinguishing 1s agent/patient affix. First question: is this realistic? If it is, the issue now is if you have an intransitive verb:

yes-xe

We exist

Is this language nom-acc, since this is the experiencer having the same affix as the agent? Or is it erg-acc, since the experiencer is the same as the patient? What if the rest of the syntax doesn't help e.g. if it's SOV?