Conlang Random Thread

[Zju]

Something about the overbar and dot (which represent featural distinctions beyond the base shape of each character) really bugs me both aesthetically and conceptually. It seems like each character ought to have a unified form without needing diacritics to write native phonological contrasts.

What do you mean, I wouldn't have thought it was a conscirpt if you didn't tell.

[mèþru]

I share your dislike of the overbars and dots, and i feel the same way even when i see them in natlangs .... at least when they cause the characters they modify to shrink.

The way I feel about it is it looks nice as long as you don't make me handwrite it, because I'm guaranteed to get the sizes wrong

[Xwtek]

It seems like each character ought to have a unified form without needing diacritics to write native phonological contrasts.

Latin uses acute accent for their long vowels.
Arabic uses a diacritic to distinguish s and sh.
Special note goes to Ancient Greek, where there is an extensive system of diacritics to write their own language.

[cedh]

What do you call the chapter about comparison, questions and negation? Is there a term that encompasses the three?

Not really afaik, but what about something like "Basic pragmatic functions" (although that might imply including imperatives and/or topicalisation in the same section too) or "Special sentence types"?

[holbuzvala]

Can we think of any reasons why a language's verbs might have both subject and direct object marking on them, but not any oblique arguments marked on the verb? I'm evolving a lang whose noun classifiers have melded onto the verb as prefixes denoting subject and object, but I'm wondering if oblique arguments would resist this bearing in mind that noun incorporation is also allowed without disturbing the number of arguments markable on the verb (the noun classifiers helped demarcate noun phrases at T₀). Here's a sample:

T₀
Pitar ki ta saka
Peter HUMAN INANIMATE SEE.pres
"Peter sees it."

T₁
Pitar kitasaka
Pitar ki-ta-saka
Peter h.subj-inan.obj-SEE.pres
"Peter sees it."

T₀
Pitar ki miki je talwa
Peter HUMAN fish ANIMATE hunt.pres
"Peter is hunting/hunts (the) fish"

Pitar ki miki talwa
Peter HUMAN fish hunt.pres
"Peter is fishing"
(N.B. the lack of a classifier following 'miki' implies that it forms a unit with 'talwa')

Pitar ki alaw je miki talwa
Peter HUMAN trout ANIMATE fish hunt.pres
"Peter is fishing for trout"

T₁
Pitar miki kijetalwa
Pitar miki ki-je-talwa
Peter fish h.subj-ani.obj-hunt.pres
"Peter is hunting/hunts (the) fish"

Pitar kimikitalwa
Pitar ki-miki-talwa
Peter h.subj-fish-hunt.pres
"Peter is fishing"

Pitar alaw kijemikitalwa
Pitar alaw ki-je-miki-talwa
Peter trout h.subj-ani.obj-fish-hunt.pres
"Peter is fishing for trout"


So, the question is: "despite the verb complex being able to absorb a noun while having two argument, is it feasible/naturalistic that the verb would resist incorporating obliques? If so, how? Perhaps by having all oblique phrases come sentence-finally? (though, what would then prevent them from binding on the back end of the verb?)

Any pointers would be most helpful.

[Pabappa]

Can we think of any reasons why a language's verbs might have both subject and direct object marking on them, but not any oblique arguments marked on the verb?

Id like to help, as this conlang has a similar setup to some of mine, but I dont think I understand the terminology. i looked up https://glossary.sil.org/term/oblique-object but i dont understand how to apply that to a language like this where the nouns are all single words. By the way, does this mean that all verbs with human agents begin with ki-? That might get tedious, and I'd consider allowing zero-marking for the most common class.

edit: are oblique arguments the same as indefinite? thats all the difference i can see in those sample sentences.

[bradrn]

Can we think of any reasons why a language's verbs might have both subject and direct object marking on them, but not any oblique arguments marked on the verb?

Isn’t this the usual situation with polypersonal marking?

[holbuzvala]

@pabappa

Apologies if my terminology was inaccurate. By ‘oblique object’, I meant any argument of a transitive verb that isn’t the subject(agent) or direct object(patient). This would include indirect objects, locative constructions, instrumentals, adverbs of manner (if applicable), and so on.

@bradrn

I’m not sure it is usual. I recall reading somewhere that Swahili can mark the indirect object in its verb complexes; and I think the pilysynthetic native American languages tend to do so as well, alongside encoding all sorts of locative/manner information.

But if it is usual, all the better

[Xwtek]

Apologies if my terminology was inaccurate. By ‘oblique object’, I meant any argument of a transitive verb that isn’t the subject(agent) or direct object(patient). This would include indirect objects, locative constructions, instrumentals, adverbs of manner (if applicable), and so on.

It's actually unrealistic to mark all of them on the verb. Most polysynthetic language only marks subject and direct object. Some also marks indirect object. But if you have four arguments, it's really weird, albeit admittedly extant. Instead, most polysynthetic language has applicative and benefactive voice.

I’m not sure it is usual. I recall reading somewhere that Swahili can mark the indirect object in its verb complexes; and I think the pilysynthetic native American languages tend to do so as well, alongside encoding all sorts of locative/manner information.

But if it is usual, all the better

But Swahili is the unusual one. I think Navajo just marks subject and indirect object. Also, I think it's more common to mark just subject and object in North America. (although it's not rare to also mark the indirect object). And the justification for the absence of marking of indirect can be pretty simple: Indirect object is governed by adposition, hence the pronoun fuses into adposition instead of into verbs.

[bradrn]

@bradrn

I’m not sure it is usual. I recall reading somewhere that Swahili can mark the indirect object in its verb complexes; and I think the pilysynthetic native American languages tend to do so as well, alongside encoding all sorts of locative/manner information.

I couldn’t find any grammars of Swahili thorough enough to answer this, but from what I could find it seems that the Bantu languages do mark indirect objects as you state. According to Wikipedia, so do Georgian and Basque. In contradiction to what Akangka says, Wikipedia does list a ‘postposition object’ prefix for Navajo; a Koasati grammar also reports indirect object prefixes. However, a grammar of West Greenlandic explicitly states that ‘In so far as one may use this term for oblique arguments in the allative case … indirect objects are not encoded in the verb’. In general, it seems that you are actually correct in saying that indirect objects are usually marked on the verb.

[holbuzvala]

@bradrn, Akangka

Seems you disagree. No matter - I'll just browse through some grammars. But I have a feeling it can be either way, which is good, as I won't have to do too much diachronic justification for whichever I choose.

[Pabappa]

I guess I cant help either. My conlangs would translate the sentences youve provided in a similar way to yours. e.g. "peter caught (a/the/some) trout" would require the "sea life" prefix, but "Peter fished" would not.

I guess what I dont get is that it looks like youre incorporating an indefinite object into the verb, and then calling it oblique. My conlangs would follow yours in not using classifiers for indefinite objects, but I wouldnt feel comfortable assuming that I should also do the same for adverbs, instrumentals, and indirect objects. To me those things dont seem to form a logical set with indefinites.

[holbuzvala]

@pabappa

I guess what I dont get is that it looks like youre incorporating an indefinite object into the verb, and then calling it oblique. My conlangs would follow yours in not using classifiers for indefinite objects, but I wouldnt feel comfortable assuming that I should also do the same for adverbs, instrumentals, and indirect objects. To me those things dont seem to form a logical set with indefinites.

Ah! I see where the misunderstanding has occurred. The sentences about fishing are not demonstrating any oblique arguments. They're merely demonstrating how 'hungry' the verbs are for incorporating elements from the sentence; and thence are illustrative to raise the question of "If the verbs are 'hungry' enough to absorb both freestanding nouns involved in the agent-patient relationship and clitic-y bits, would they be hungry enough to absorb oblique arguments too?"

[malloc]

It seems like each character ought to have a unified form without needing diacritics to write native phonological contrasts.

Latin uses acute accent for their long vowels.
Arabic uses a diacritic to distinguish s and sh.
Special note goes to Ancient Greek, where there is an extensive system of diacritics to write their own language.

True enough, although all of those developed the diacritics well after the rest of the script. Visually though, the floating pieces feel less like diacritics and more like pieces of discontinuous characters. It seems difficult to think of many scripts where so many (or even any) characters are discontinuous. The best example that comes to mind is Chinese hanzi, although that seems a poor analogy given its logographic nature. Plenty of other scripts have numerous characters without any discontinuity so it really ought to be possible. Yet making it work aesthetically while following the featural principle remains a challenge.

[Richard W]

It seems difficult to think of many scripts where so many (or even any) characters are discontinuous.

Don't English 'i' and 'j' count? Greek Ξ (most styles) Hebrew ה also spring to mind. Indic scripts are full of two part vowel symbols, some deriving from the two-part Brahmi diacritic for dependent vowel 'o'.

[bradrn]

It seems difficult to think of many scripts where so many (or even any) characters are discontinuous.

Don't English 'i' and 'j' count? Greek Ξ (most styles) Hebrew ה also spring to mind. Indic scripts are full of two part vowel symbols, some deriving from the two-part Brahmi diacritic for dependent vowel 'o'.

In Hebrew, אבגזק at least are also written discontinuously. But admittedly the cursive version of the script is geared towards avoiding these discontinuities.

It seems difficult to think of many scripts where so many (or even any) characters are discontinuous.

Maybe you could do something similar to the situation outlined above; that is, have discontinuous ‘block letters’, and then invent a less obviously featural ‘cursive’ script which is easier to write.

[Nortaneous]

Can we think of any reasons why a language's verbs might have both subject and direct object marking on them, but not any oblique arguments marked on the verb?

dunno, but Japhug only marks the subject and direct object on the verb

It seems difficult to think of many scripts where so many (or even any) characters are discontinuous.

Don't English 'i' and 'j' count? Greek Ξ (most styles) Hebrew ה also spring to mind. Indic scripts are full of two part vowel symbols, some deriving from the two-part Brahmi diacritic for dependent vowel 'o'.

the number of discontinuous English characters depends on your handwriting - for me <i j p t z> are discontinuous, in standard American cursive <i j t x> are, and in ~standard American print <f i j k t x y> are

[bradrn]

Can we think of any reasons why a language's verbs might have both subject and direct object marking on them, but not any oblique arguments marked on the verb?

dunno, but Japhug only marks the subject and direct object on the verb

I couldn’t find a grammar of Japhug in English to confirm this, but a grammar for the closely related Situ language explicitly states that ‘Transitive verbs also can express the relationship between the subject and one object in person markers which are prefixed to the verb root’ (emphasis mine). So it seems that this could be common in rGyalrongic languages.

It seems difficult to think of many scripts where so many (or even any) characters are discontinuous.

Don't English 'i' and 'j' count? Greek Ξ (most styles) Hebrew ה also spring to mind. Indic scripts are full of two part vowel symbols, some deriving from the two-part Brahmi diacritic for dependent vowel 'o'.

the number of discontinuous English characters depends on your handwriting - for me <i j p t z> are discontinuous, in standard American cursive <i j t x> are, and in ~standard American print <f i j k t x y> are

Oddly enough, my style of writing has no discontinuous characters. Possibly this could be because I use a fountain pen, which I find works best with connected lines. (This also provides a great example of the influence of writing instruments on alphabet design!)

[Raphael]

In the SCA, is there some way to get, as output, a list of the pairs of pre-change words and post-change words?

[Raphael]

Unrelated question: how do prefixes and suffixes marking things like cases and plurals and the like change over time? My impression from the few languages I know a little bit about is that after some time has passed, you might almost as well throw them out completely and introduce completely new ones, but I fully expect that people who know more about languages than I do will tell me that that's nonsense.