Conlang Random Thread

[Pabappa]

i wonder if butt=knife comes from people carrying knives by their hips? if so, it has the exact same logic that I use in Poswa for hip=cellphone=gun. but i suspect the Pomo way might have something to do with blunt force.

_______

Poswa, like all my conlangs since 2002, is SOV. The basic structure of the sentence is
SUBJ + (OBJ+INFLECTIONS) + (VERB+INFLECTIONS)
So the first word of the sentence, the subject, is uninflected. The second word (assuming there are no adjectives etc) is the object and carries the accusative case. The verb carries all the other inflections and is often the longest word in the snentence.

But there is an anomalous verb-first construvtion:

VERB + (OBJ+ACC+INFLICTIONS)

If there is a subject, it has no inflections eihter.
I could analyze trhe ""normal" verbs as being a special case of the fronted verbs, with the inflections being the same as the object's inflections in the VO construciton. This is where it gets hard to explain ..basically

popem pypopebi "eat-Ø apple-ACC-TR-1P.PAST" is the marked construction, and is rquivlent to the much more common

pypopi pobebi "apple-ACC-Ø-1P.PAST eat-TR-1P.PAST".

popem ~ pob- "to eat"
pyp "apple" (accusative pypop)
-ebi 1st person transitive past

Neither sentenceh as a subject because there are no pronouns in Poswa. THe stray -i that isnt colored in is an agreement morpheme, which is a serparate issue.

This is hard to explain but it makes me feel great because even though Poswa is supposed to imitate natlang grammar, I like it when it behaves like a loglang even so. Basically the breakthrough is: being able to analyze unmarked verbs as just a special case of marked verb clauses, where the object argument is replaced by a zero morph.. So pobebi " i ate it" is the same as popem ... -ebi "I ate the ...." which can only occur when an object appears in between the fronted verb and the inflection.

Thus, every fully inflected verb is grammatically equivalent to a clause, and every clause is grammatically equivalent to a single verb.

Sorry for my typing lately, I think its partly stress and partyly extra caffeine use due to less sleep.

[Glass Half Baked]

Sleep is for the dead.

I really love this syntax. It feels like something you would see in real life, but I can't think of any examples. You're right that it's simultaneously something clean and efficient but not artificial.

[Xwtek]

i wonder if butt=knife comes from people carrying knives by their hips? if so, it has the exact same logic that I use in Poswa for hip=cellphone=gun. but i suspect the Pomo way might have something to do with blunt force.

Probably just a homophone.

[linguistcat]

Would it be likely for a language that has two ways of forming concessive constructions to turn one into a concessive conditional? Example:

A construction that meant "Even though John smokes, he's still in good health." coming to mean "Even if John smokes, he may stay in good health."

It seems to me like it could easily change this way with time. But it would also affect the main phrase (at least going from the proto-language I'm using and its grammar), and I've thought things were likely before that others with more experience seemed to think were unlikely. So I wanted to check if this was more sound.

ETA: A little ways into this paper, I found that concessives do often become concessive conditionals. So please ignore this question.

[Xwtek]

ETA: A little ways into this paper, I found that concessives do often become concessive conditionals. So please ignore this question.

Wow, that's a cool idea.

[Kuchigakatai]

Just gonna mention that in Classical Chinese, two juxtaposed clauses without any markers of their relationship can be interpreted as temporally sequential (first A, then B), conditional (if A, then B), concessive (even though A, nevertheless B), or explanatorily consequential (because A, therefore B). There are markers to disambiguate such relationships if it's really needed though.

I think the concessive interpretation probably comes from an "A but B" interpretation morphing into "even though A, nevertheless B"... And the others can easily be semantic expansions of "A and B"...

[akam chinjir]

Just gonna mention that in Classical Chinese...

Are you counting "而" as a marker of relationship? (All of those seem fine to me with "而," without it I'm less sure.)

[Glass Half Baked]

... please ignore this question.

Never!
Regardless of how concessive constructions drift over time, there is another way to do what you want. The difference between the "even though" and "even if" examples you gave was one of realis and irrealis. Languages that have some overt marking of irrealis can use the same concessive construction for both without ambiguity.

[linguistcat]

... please ignore this question.

Never!
Regardless of how concessive constructions drift over time, there is another way to do what you want. The difference between the "even though" and "even if" examples you gave was one of realis and irrealis. Languages that have some overt marking of irrealis can use the same concessive construction for both without ambiguity.

Hmmm.... There is differentiation between realis and irrealis (among a few other things) in the proto language and the modern language this conlang is related to. Currently they both forms use the realis stem, but it wouldn't be hard to have the conspeakers reanalyze the one that becomes conditional as using the irrealis. Or maybe the stems would stop having that distinction and come to mean something else. I'm not sure what to do with that though....

[Glass Half Baked]

Breaking news! William Annis, of Conlangery Podcast fame, has published a pdf of the Kilta Reference Grammar. I especially like his use of auxiliaries and particles. https://www.dropbox.com/s/4ug1d8h4e5lf4 ... a.pdf?dl=0

[akam chinjir]

Breaking news! William Annis, of Conlangery Podcast fame, has published a pdf of the Kilta Reference Grammar. I especially like his use of auxiliaries and particles. https://www.dropbox.com/s/4ug1d8h4e5lf4 ... a.pdf?dl=0

It looks like a new version---but earlier ones have been available (and this one is now) through his conlanging page for quite some time.

[Kuchigakatai]

Just gonna mention that in Classical Chinese...

Are you counting "而" as a marker of relationship? (All of those seem fine to me with "而," without it I'm less sure.)

Yeah, I'm not counting 而 as a "marker of clause relationships" because it just barely manages to be some sort of conjunction without much meaning. I was rather thinking of 若 'if', 則 'then', 雖 'although', and the like.

Interestingly, I've come across "although X, nevertheless Y" with no conjunction or anything in poetry (I believe it was in some of the Tang poetry of Naiying Yuan et al.'s textbook series), but I bet it doesn't occur in prose.

[Glass Half Baked]

Breaking news! William Annis, of Conlangery Podcast fame, has published a pdf of the Kilta Reference Grammar. I especially like his use of auxiliaries and particles. https://www.dropbox.com/s/4ug1d8h4e5lf4 ... a.pdf?dl=0

It looks like a new version---but earlier ones have been available (and this one is now) through his conlanging page for quite some time.

Ah, yes, re-checking the date it's from January... of last year.

[Zju]

Quick poll:
Say there's a conditional sound change A > B that depends on three conditions C, D and E in the following way:



Which way to write it down would make the most sense?

a) A > B / C, D ~ E
Union is marked with a comma, intersection is marked with a tilde, union takes precedence.

b) A > B / (C, D) ~ E
Union is marked with a comma, intersection is marked with a tilde, intersection takes precedence.

c) A > B / C ~ D, E
Union is marked with a tilde, intersection is marked with a comma, union takes precedence.

d) A > B / (C ~ D), E
Union is marked with a tilde, intersection is marked with a comma, intersection takes precedence.

e) in some other way?

Bonus question: how is it written down in linguistic papers / linguistic notation?

Bonus question 2: say the same sound change A > B occurs in all cases where condition C is not present. How would it make most sense to write it down / how is it written down in linguistic papers?

a) A > B / !C
b) A > B / -C
c) A > B / ~C
d) in some other way?

[bradrn]

Quick poll:
Say there's a conditional sound change A > B that depends on three conditions C, D and E in the following way:

…

Which way to write it down would make the most sense?

I would personally just use words: ‘A → B / (C or D) and E’.

Can’t help with the bonus questions, sorry.

[Kuchigakatai]

Are sound changes like that even known to have happened in natlangs?

And yeah, these seem like situations where words would be used in a relatively ad hoc manner.

A > B / E and either C or D or both
A > B / always, except when C (...or maybe: "A > B / D or E, except when C")

I guess I can think of an example of the latter from Archaic Latin to the pre-classical Old Latin of Plautus...

o > u / _C#, except when w_

Thus, trībos > trībus, manos > manus, duenos > bonus, but servos > servos, equos > equos.

[akam chinjir]

I think when people are being careful, they don't allow disjunctive or negative conditions. This affects decisions about whether features are binary, for example. Like, if voicing is binary, then both [+voiced] and [-voiced] can condition phonological rules; but if it's privative, then only [voiced] can. And a reason for thinking that there's a [coronal] node (say) in your feature geometry is that there are phonological processes sensitive to it, and you don't want to just say (for example) [alveolar] or [dental].

What this means is that if you genuinely want a situation where *A→B/(E&(C or D)), then the careful way to put it is with two rules:

• A → B / E and C
• A → B / E and D

This makes some historical sense, in that you wouldn't especially expect the changes to have occurred together.

[Zju]

Are sound changes like that even known to have happened in natlangs?

I can't think of a sound change that happens under such combination of circumstances per se, but such description would nevertheless be useful as an intermediary stage, or as a part of a description of a difficult to capture in one line sound change.

Case is point: how to write a conditional sound change that affects consonants next to /i/, but not when they're word final? I can think of e.g.:

A > B / _i, i_ & !_#



New poll - consider the following sound change:
A word-initial consonant cluster of two stops is broken up with an epenthetic vowel that echoes the first vowel of the word, e.g.:
tka > taka
pki > piki
ktu > kutu

How would you write it down symbolically in an unambiguous way?

a) ∅ > V1 / #C_CV1
b) ∅ > V(1) / #C_CV(1)
c) ∅ > V^1 / #C_CV^1
d) ∅ > V[1] / #C_CV[1]
e) in some other way?

[Travis B.]

The problem with using words to represent conjunctions or disjunctions is that English "or" is ambiguous as to whether it means an inclusive or or an exclusive or, whereas in contexts like these one wants no ambiguity whatsoever. (In the programming world for instance "or" always means an inclusive or while "xor" or "eor" means an exclusive or, which is very important in that context.)

[TurkeySloth]

Voicing won't contrast in my setting's Elvish language's plosives but will in its fricatives. Is this attested in any natlang? Does having the language's affricates pattern like its plosives or fricatives sound more natural?