Zompist Bboard Again

rotting bones wrote: ↑Thu Aug 20, 2026 2:22 pm

Richard W wrote: ↑Thu Aug 20, 2026 2:11 pm Isn't Coptic gender agreement in true adjectives restricted to Greek loans?

This is a tendency, but isn't sabe/sabh an exception? https://shs.hal.science/halshs-03092432/document

Ouch! <h> as a vowel?

Table 4 gives the word as a pair of gender-marked nouns, 'wise man' / 'wise woman', but we may be seeing how a bit of Greek morphology could insert itself. Native Coptic sex-marking looks messy, and I suspect it ceased to be productive.

Richard W wrote: ↑Fri Aug 21, 2026 3:34 am Ouch! <h> as a vowel?

Interesting orthographic choice. The letter <h> does derive from the Greek letter eta which represented the vowel /ɛː/ and that's presumably the logic here.

Richard W wrote: ↑Fri Aug 21, 2026 3:34 am Ouch! <h> as a vowel?

malloc wrote: ↑Fri Aug 21, 2026 8:04 am Interesting orthographic choice. The letter <h> does derive from the Greek letter eta which represented the vowel /ɛː/ and that's presumably the logic here.

<h> transcribes hata in the Coptic fonts I'm used to and also in the linked pdf: https://en.wikipedia.org/wiki/Coptic_script Hata (ⲏ) is pronounced /e/: saβɛ/saβe in the Sahidic dialect of Upper Egypt. Late Coptic and Greco-Bohairic sound very different.

Richard W wrote: ↑Fri Aug 21, 2026 3:34 am Table 4 gives the word as a pair of gender-marked nouns, 'wise man' / 'wise woman', but we may be seeing how a bit of Greek morphology could insert itself. Native Coptic sex-marking looks messy, and I suspect it ceased to be productive.

Unlike Laulai, Coptic is not considered to be a mixed language. Any further Hellenization, and it might have become a mixed language if it didn't die out. If it had become a mixed language, the morphology might have been affected in the manner you describe, and possibly in other ways Greek differs from Coptic.

A Greek-Egyptian mixed language could have been split-ergative. Coptic marks TAM and the subject using verbal prefixes. Although the subject is marked with a prefix, pronominal suffixes are attached to the end of a transitive verb. Ancient Greek marks subjects and TAM at the end of the verb stem. Greek is, of course, strictly accusative. If imperfective actions preserve the Coptic verbal prefixes while perfective actions borrow the Greek aorist past tense suffix, speakers might reinterpret the Greek suffixes as marking the patient. The transitive agent might use a different oblique marker. This would make the past ergative.

This is actually a more scientific plan for creating a mixed language than what I currently have for Laulai. I didn't set out to make Laulai a mixed language. That's apologetics I invented to justify the way I used it. My excuse for Laulai is that it was a tiny language spoken in one valley. All its relatives were absorbed into Greek, so what survived is very likely to be mixed.

There are reasons for Coptic to have been more resistant to mixing than Laulai. Egyptian had a large population of speakers and an established native literary tradition. Given that, it's kind of shocking how Hellenized it is. I've read ideology played a role. Egyptian tradition was associated with idolatry once the bible was accepted, given the things it says about Egypt.

quinterbeck wrote: ↑Thu Aug 20, 2026 10:05 pm

Leilis wrote: ↑Thu Aug 20, 2026 12:42 pm ... the distinction between the verbs tag and elf. Here are the lexicon entries (deleting a word I'm not sure works):

tag {?PH *s-takʰa} v be arranged, be set in order; signify, mean, symbolise; be fated; inf arrangement; system; meaning; fate; act administrator. tegien f order. teigoi m rubric, chart. tegéid f part of a system, class; method; declension or conjugation; musical mode, key; cobitegéid {cobi ‘belly’} f digestive system, cusentegéid {cusen ‘heart’} f circulatory system. tanga m sign, symbol, token; curitanga {curi ‘market’} f logograph, pictographic label. | antag v (of a fyrd or army) be arranged for battle; inf battle-lines; battle; act warlord’s commander; ivantegir act commander over an army from more than one fyrdland. antegéin f battlefield. artag v be done out of order or against a system; (music) be sung out of key. intag v be rightly ordered. brenatag {bren ‘hand’ < bar} v be gestured at with the hand, be pointed at.

elfa {?PH *ælpʰa; cf. ilva} f kind, sort, class, family; kinship, relationship. inelfa f ratio; teiginelfa f the ratio of a circle’s circumference to its radius (two pi or tau). aiyelfa {aia} f language, tongue. incireilfa {incir ‘be inflected’ < cir} f a particular declension or conjugation. teimbeteilfa {teimbot ‘dance’} f a metrical segment of the form long-short-long-short. | elf v be related (to com. person or thing); be sorted, classified (with com. persons or things). corelf v (grammatical) be modified, be limited, be determined, be defined; corelfoi m pronoun or determiner. inelf v have some numerical relationship. teigelf v be neatened, be tidied.

It seems like concepts under tag generally have more internal structure than those under elf (aielfa 'language' being the main exception)

I think this is starting to get at it, yes. It seems that tag is 'arranged' slightly more in the sense 'prepared (for use)', with an internal structure or structure determined by the end goal, and elf is 'arranged' slightly more in the sense 'classified, organised', with the structure having to do with grouping things (not surprising from elfa 'kind'!). It does seem to break down with some of the derivatives, like why is 'language' aiyelfa 'class of speech' instead of *aiyetaga 'system of speech'? Maybe because it sounds better the sense 'family/kinship' comes in here, it's a 'kind of speech' but also 'a kinship of speech', i.e., people of the same ethnicity often have the same language? Anyway I think I'm getting a clearer idea of the distinction.

rotting bones wrote: ↑Fri Aug 21, 2026 9:14 am There are reasons for Coptic to have been more resistant to mixing than Laulai. Egyptian had a large population of speakers and an established native literary tradition. Given that, it's kind of shocking how Hellenized it is. I've read ideology played a role. Egyptian tradition was associated with idolatry once the bible was accepted, given the things it says about Egypt.

I've read that the use of Coptic for legal documents received a massive boost from the brief period of Persian rule (619 to 628) before the Eastern Roman empire regained control. It's quite conceivable that language replacement (Coptic to Greek) was already underway. One could compare Coptic to the language of blazons in English - that's far more Romanised (specifically, Frenchified) than bookish English.

Richard W wrote: ↑Fri Aug 21, 2026 5:12 pm I've read that the use of Coptic for legal documents received a massive boost from the brief period of Persian rule (619 to 628) before the Eastern Roman empire regained control. It's quite conceivable that language replacement (Coptic to Greek) was already underway.

This guy says Iranians have never preserved and are currently not preserving their practices abroad: https://youtu.be/3pL6uERL85w This guy is a liberal Islamist type, so it could just be culture war bellyaching.

Richard W wrote: ↑Fri Aug 21, 2026 5:12 pm One could compare Coptic to the language of blazons in English - that's far more Romanised (specifically, Frenchified) than bookish English.

What are the conditions under which a language becomes mixed as opposed to dying out? I assumed it has something to do with institutional practices. If speakers are immersed in institutions that use a different language, they will eventually leave their old language behind. This is why having an established literati is likely to make the old language endure. On the other hand, if society is developed like in Egypt, that would facilitate a shift to a new language en masse from global pressure.

Perhaps the fact that the Laulai were an isolated tribe of goatherds in Arcadia helped preserve their language for long enough. There was a scholarly class, but they belonged to a (by then) local cult who had discovered what they believed were all kinds of mystical properties in their language. Even so, Laulai might have only survived as a holy language as long as the cult endured if they had remained in the region.

Are there any languages that start off without grammatical gender/class but have nominative/oblique case markings (they're simple vowels) and then create animate/inanimate gender classes by suffixing the animate/inanimate markers AFTER the case-marking suffixes?

Stage one:
-i (nom) / -a (obl)

Stage two:
-i-m (animate nom) / -a-m (animate obl)
-i-s (inanimate nom) / -a-s (inanimate obl)

Ahzoh wrote: ↑Mon Aug 24, 2026 9:54 pm Are there any languages that start off without grammatical gender/class but have nominative/oblique case markings (they're simple vowels) and then create animate/inanimate gender classes by suffixing the animate/inanimate markers AFTER the case-marking suffixes?

Stage one:
-i (nom) / -a (obl)

Stage two:
-i-m (animate nom) / -a-m (animate obl)
-i-s (inanimate nom) / -a-s (inanimate obl)

Speaking generally:

I would say this depends on whether one of the case-marking suffixes is considered a 'null' suffix or if the case-marking suffixes are still seen in a way as cliticized postpositions rather than true suffixes. If either of those are true, then I'd say that the gender markers would likely go before the case-markers; otherwise, I'd say that the gender markers would likely go after the case markers. If some of the case-marking suffixes are seen as true suffixes while others are seen as being postpositions of sorts, then you could have a mix, where the gender markers go after the true suffixes but before the postpositions.

In your case, though, I'd say they'd come after if your case markers behave like suffixes but before if your case markers behave like clitics.

Travis B. wrote: ↑Tue Aug 25, 2026 12:43 pm

Ahzoh wrote: ↑Mon Aug 24, 2026 9:54 pm Are there any languages that start off without grammatical gender/class but have nominative/oblique case markings (they're simple vowels) and then create animate/inanimate gender classes by suffixing the animate/inanimate markers AFTER the case-marking suffixes?

Stage one:
-i (nom) / -a (obl)

Stage two:
-i-m (animate nom) / -a-m (animate obl)
-i-s (inanimate nom) / -a-s (inanimate obl)

Speaking generally:

I would say this depends on whether one of the case-marking suffixes is considered a 'null' suffix or if the case-marking suffixes are still seen in a way as cliticized postpositions rather than true suffixes. If either of those are true, then I'd say that the gender markers would likely go before the case-markers; otherwise, I'd say that the gender markers would likely go after the case markers. If some of the case-marking suffixes are seen as true suffixes while others are seen as being postpositions of sorts, then you could have a mix, where the gender markers go after the true suffixes but before the postpositions.

In your case, though, I'd say they'd come after if your case markers behave like suffixes but before if your case markers behave like clitics.

The end result is supposed to look like this The language doesn't really have postpositions, but maybe they are fossilized determiners

Unfortunately the biquinary principle is quite a double-edged sword for my minimalist project. It reduces the number of roots required but it also results in longer and more repetitive words for numbers. Something tells me that Dolly Parton would not have gotten too far if she had to sing "five-four to ten-five-two" (as my project uses 24 hour time) instead of "nine to five". Perhaps the binary approach of "eight-one to hex-one" would go down easier.

Moose-tache wrote: ↑Wed Aug 26, 2026 11:54 pm Syncope would like to know your location

Syncope would know your location in certain environments.

Moose-tache wrote: ↑Wed Aug 26, 2026 11:54 pm Syncope would like to know your location

Is that referring to the use of 24 hour time? It seems like a useful way to avoid the need for specifying AM or PM.

malloc wrote: ↑Thu Aug 27, 2026 4:33 pm

Moose-tache wrote: ↑Wed Aug 26, 2026 11:54 pm Syncope would like to know your location

Is that referring to the use of 24 hour time? It seems like a useful way to avoid the need for specifying AM or PM.

I read it as a suggestion that a sound change of syncope could help delete some syllables in those pesky long numbers.

Aside from the biquinary principle, I must also consider how to build larger numbers. I quite like successive powers of a myriad (10000) since it seems a nice square number. That said, thousands would probably work better since the metric system assumes powers of a thousand if nothing else. The construction "twenty-three hundred forty-five" sounds cooler than "two thousand three hundred forty-five" but also more formally complicated.

malloc wrote: ↑Fri Aug 28, 2026 10:16 am The construction "twenty-three hundred forty-five" sounds cooler than "two thousand three hundred forty-five" but also more formally complicated.

And of course, English uses the snappier "twenty-three forty-five" for the year 2345. (The pattern took a while to re-establish itself for dates like 2003.) For round numbers, although nominally using the plodding method of using the unit names for non-zero units, tens, hundred, thousands, tens of thousands and hundreds of thousands in a large number, for round numbers Thai colloquially says the equivalent of 'two thousand three' for 'twenty three hundred'. It didn't take me long to comprehend when I first encountered it. As befits language efficiency, actually conveying 'two thousand and three' is more complicated.

I've been working on Heláin numbers and just decided earlier this week that orders of magnitude could be compounded in pairs: as if we said in English, 'one-twentythousand, three-fourhundred, fifty-six' for 123,456. (It's base-6, though, and also a few pairs can't be compounded because they're ambiguous.)

Leilis wrote: ↑Fri Aug 28, 2026 4:39 pmI've been working on Heláin numbers and just decided earlier this week that orders of magnitude could be compounded in pairs: as if we said in English, 'one-twentythousand, three-fourhundred, fifty-six' for 123,456. (It's base-6, though, and also a few pairs can't be compounded because they're ambiguous.)

Sounds interesting. I experimented with base-6 but quickly realized it would need numerous words for powers of six to build larger numbers. This seems like an interesting way to mitigate that problem.

Currently my proposed system has words for one through five and ten and hundred, a suitably minimalist set of morphemes. That just leaves the question of what comes next.

malloc wrote: ↑Fri Aug 28, 2026 9:05 pm

Leilis wrote: ↑Fri Aug 28, 2026 4:39 pmI've been working on Heláin numbers and just decided earlier this week that orders of magnitude could be compounded in pairs: as if we said in English, 'one-twentythousand, three-fourhundred, fifty-six' for 123,456. (It's base-6, though, and also a few pairs can't be compounded because they're ambiguous.)

Sounds interesting. I experimented with base-6 but quickly realized it would need numerous words for powers of six to build larger numbers. This seems like an interesting way to mitigate that problem.

Currently my proposed system has words for one through five and ten and hundred, a suitably minimalist set of morphemes. That just leaves the question of what comes next.

I like in my languages to use biquinary for one through nine, have a number for ten, add biquinary numbers to ten for 11-19, have a number for 20, use multiples of 20 plus bi-quinary numbers for 20-29, 40-49, 60-69, and 80-89, use multiples of 20 plus ten plus biquinary numbers for 30-39, 50-59, 70-79, and 90-99, and have numbers for 100, 1000, and 10000 which are multiplied by biquinary numbers and added to the preceding numbers.

I am completely lost on how tone should manifest in my language, mostly regarding interactions with codas

T = /p t k/ = [p~pʰ t~tʰ k~kʰ]
D = /b d g/
S = /s x/
N = /m n l w j/

At least for open syllables, it's either gonna be TV = high, DV = low, and SV/NV = mid or TV/SV = high and DV/NV = low

I want to have some contouring but none of the codas elide and the only thing lost is voicing distinction. There might be some sandhi such as homoorganic T/D > S before homoroganic S