Conlang relay [relocated] (aka "The Cursed Relay")

[TzirTzi]

I've started work on Siixtaguna culture - beginnings of a draft can be found on the wiki here. I've gone with the earlier dates as they work better for what I want to do and seem to fit Isles and Isthmus culture movements as well...



On the issue of N-T, I'm sure I can make sure I always use a hyphen . Another possible solution would be renaming this group of languages the Coastal Siixtaguna family (CS).



Legion - I've also been having a look at expanding the Kozado lexicon, as it seemed to be the single smallest (and thus most straight-forwardly completable) project needing work. It turns out, of course, to be more complicated. I've got a few questions about the sound changes, if you have time to go over them?

The Adata page suggests that irregular stress is marked with an underline, but as far as I can see on the lexicon page, no irregular stresses are marked. However, you have e.g. epiáp, suggesting that you can see irregular stresses in the Adata lexicon. Is this a display fault with my computer or is this info from elsewhere?

You have a sound change /j/>/Z/ which doesn't seem to apply at all regularly. Intervocallicaly it sometimes occurs e.g. /moje:/>/maz\e/, but at other times doesn't. Initially it usually doesn't, e.g. /ja:ti:/>/ieTi/, /ja:ra/>/iar/, etc. It also doesn't seem to happen in C_V position, e.g. /itjan/>/iTjo/.

You have a sound change deleting final unstressed vowels which also seems unreliable. Sometimes it occurs e.g. /abize/>/abiz/, and sometimes it doesn't e.g. /dado/>/dado/.

The vowel shifts don't always seem to match what I understand from the sound changes. For example, short /a/ should change /a/>/A/>/o/ and long /a:/ should change /a:/>/a/. However in /dado/ the short /a/ remains unchanged, and in /do/ (from /da:n/) a long /a:/ becomes /o/.

I also have some words whose etymologies I can't work out, but I'll leave those until I'm more sure of the sound changes...

Thanks!

[Cedh]

The Siixtaguna culture looks interesting, and now that you've replaced "insular" with "continental" the distribution of the groups makes sense too. It was a bit hard to follow in the first version.
Incidentally, you now have this:

shared cultural traits across the isles and pencontinental make regular communication across large distances likely

I like this word... does it refer to the whole Siixtaguna subcontinent? Or is it an accident of the search&replace function?

<hr>

The Adata page suggests that irregular stress is marked with an underline, but as far as I can see on the lexicon page, no irregular stresses are marked. However, you have e.g. epiáp, suggesting that you can see irregular stresses in the Adata lexicon. Is this a display fault with my computer or is this info from elsewhere?

Irregular stress used to be marked in the original Adāta lexicon, but Dewrad recoded his website in late 2007, and the new version of the lexicon does not mark stress.

However, Zhen Lin once posted a two-part .csv lexicon in this thread:

• Adāta lexicon
• Fáralo loanwords in Imperial Adāta

Stress is marked, and usually corresponds to the original lexicon. You should be aware of two issues though: First, this lexicon already includes some of the semantic drift between Adāta and Ayāsthi (and for some words, Mavakhalan). Second, sometimes it has stressed non-initial short vowels, which shouldn't normally be possible with the phonotactics of Adāta. But then, stress placement or prosodical lengthening/shortening is one of the things one would most expect to vary between different dialects of a language.

<hr>

I am currently revisiting the Hitatc family. Expect some material on the wiki tomorrow.

Introductory texts and phonology sketches for Proto-Hitatc, Proto-Wan-Mlir and Ktacwa are now up.

[Legion]

The Adata page suggests that irregular stress is marked with an underline, but as far as I can see on the lexicon page, no irregular stresses are marked. However, you have e.g. epiáp, suggesting that you can see irregular stresses in the Adata lexicon. Is this a display fault with my computer or is this info from elsewhere?

What Cedh said, those aren't on display in the new page (which is one of the reason I stoped working on Kozado).

You have a sound change /j/>/Z/ which doesn't seem to apply at all regularly. Intervocallicaly it sometimes occurs e.g. /moje:/>/maz\e/, but at other times doesn't. Initially it usually doesn't, e.g. /ja:ti:/>/ieTi/, /ja:ra/>/iar/, etc. It also doesn't seem to happen in C_V position, e.g. /itjan/>/iTjo/.

My understanding of Adata phonology was that "i" designated /j/ only in post-vocalic position, and /i/ everywhere else. So yes, the change is to be taken as being applied only in post-vocalic position.

You have a sound change deleting final unstressed vowels which also seems unreliable. Sometimes it occurs e.g. /abize/>/abiz/, and sometimes it doesn't e.g. /dado/>/dado/.

That's most likely analogy: dado is an adjective and thus takes the adjectival suffix -o (from Adata -ax).

The vowel shifts don't always seem to match what I understand from the sound changes. For example, short /a/ should change /a/>/A/>/o/ and long /a:/ should change /a:/>/a/. However in /dado/ the short /a/ remains unchanged, and in /do/ (from /da:n/) a long /a:/ becomes /o/.

That, however, are probably mistakes on my part. I made the error of making the sound changes go [a a:] > [A a], which is counter intuitive, you generally expect the opposite to happen. So there's probably other words with those two vowels that shifted the wrong way. That's not too problematic though, those king of irregular sound changes are not that are (as an example, Vulgar Latin had many verbs spontaneously passing from 2nd to 3rd conjugation, and vice versa). Feel free to keep this irregularity with new words or to avoid it.

I also have some words whose etymologies I can't work out, but I'll leave those until I'm more sure of the sound changes...

I can't always remember myself what the etymologie are, but do ask.

[TzirTzi]

The Siixtaguna culture looks interesting, and now that you've replaced "insular" with "continental" the distribution of the groups makes sense too. It was a bit hard to follow in the first version.
Incidentally, you now have this:

shared cultural traits across the isles and pencontinental make regular communication across large distances likely

I like this word... does it refer to the whole Siixtaguna subcontinent? Or is it an accident of the search&replace function?

Yeah... for some reason I originally had it in my head that insular meant "of the mainland". When I discovered that it meant "of islands" I changed it to continental by find&replace.

However, Zhen Lin once posted a two-part .csv lexicon in this thread:

• Adāta lexicon
• Fáralo loanwords in Imperial Adāta

Stress is marked, and usually corresponds to the original lexicon. You should be aware of two issues though: First, this lexicon already includes some of the semantic drift between Adāta and Ayāsthi (and for some words, Mavakhalan). Second, sometimes it has stressed non-initial short vowels, which shouldn't normally be possible with the phonotactics of Adāta. But then, stress placement or prosodical lengthening/shortening is one of the things one would most expect to vary between different dialects of a language.

Thanks, I should be able to use that.

My understanding of Adata phonology was that "i" designated /j/ only in post-vocalic position, and /i/ everywhere else. So yes, the change is to be taken as being applied only in post-vocalic position.

[...]

That's most likely analogy: dado is an adjective and thus takes the adjectival suffix -o (from Adata -ax).

Ah, ok, thanks.

That, however, are probably mistakes on my part. I made the error of making the sound changes go [a a:] > [A a], which is counter intuitive, you generally expect the opposite to happen. So there's probably other words with those two vowels that shifted the wrong way. That's not too problematic though, those king of irregular sound changes are not that are (as an example, Vulgar Latin had many verbs spontaneously passing from 2nd to 3rd conjugation, and vice versa). Feel free to keep this irregularity with new words or to avoid it.

Cool, I will indeed keep all of the original vocabulary and call anything that doesn't quite fit an irregularity.

I can't always remember myself what the etymologie are, but do ask.

Well, with the above attitude it probably doesn't matter too much anyway... Once I've finished the rest, I'll come back and ask you about anything I still can't work out.

[Dewrad]

My understanding of Adata phonology was that "i" designated /j/ only in post-vocalic position, and /i/ everywhere else.

I do recall pointing out that you were wrong there

[Ran]

Does anyone know where I can find the sound changes for the Isles languages? I know that Basilius designed Affanonic (have you put up the sound changes anywhere?) and has found correspondences for Pparwak, and Cedh has done an etymological dictionary for Mutsipsa', but has anyone done the other ones yet? I'd be interested in helping out if I can.

[Zhen Lin]

Second, sometimes it has stressed non-initial short vowels, which shouldn't normally be possible with the phonotactics of Adāta. But then, stress placement or prosodical lengthening/shortening is one of the things one would most expect to vary between different dialects of a language.

I presume you are referring to words such as iába. In these cases I interpreted the initial i as /j/, going by the etymology. For words such as edák... well, those are loanwords.

[Cedh]

Ran, welcome back!

Second, sometimes it has stressed non-initial short vowels, which shouldn't normally be possible with the phonotactics of Adāta. But then, stress placement or prosodical lengthening/shortening is one of the things one would most expect to vary between different dialects of a language.

I presume you are referring to words such as iába. In these cases I interpreted the initial i as /j/, going by the etymology. For words such as edák... well, those are loanwords.

No, I'm referring to verb stems like banê, banáza- or dēmê, dēmáza-, i.e. verbs which ended in a diphthong in Ndak Ta. IIRC the regular Adāta sound changes would result in a stress shift to the initial syllable in the long stem. But of course this may simply have been leveled by analogy.

[EDIT]
Going through the SC just now, it seems that these verbs shouldn't normally even have that /z/. It results from the insertion of an epenthetic schwa after the diphthong, which shouldn't have been necessary based on NT phonotactics. Which is not really an issue though, because Adāta has a Gezoro substratum, and Gezoro did not allow as many clusters as Ndak Ta.

Here's what I get for the paradigm btw:

Code: Select all

NT                   ->   Adāta (SC)           (Adāta traditional)

banai      banaindi       banê    banâthi      banê      banázathi
banain     banaibe        banân   banêbe       banázan   banázabe
banaisti   banais         banâsi  banâ         banázasi  banázā

For comparison, with epenthetic schwa:

banai      banai@ndi      banê      bánazathi
banai@n    banai@be       bánazan   bánazabe
banai@sti  banai@s        bánazasi  bánazā

[Dewrad]

No, I'm referring to verb stems like banê, banáza- or dēmê, dēmáza-, i.e. verbs which ended in a diphthong in Ndak Ta. IIRC the regular Adāta sound changes would result in a stress shift to the initial syllable in the long stem. But of course this may simply have been leveled by analogy.

[EDIT]
Going through the SC just now, it seems that these verbs shouldn't normally even have that /z/. It results from the insertion of an epenthetic schwa after the diphthong, which shouldn't have been necessary based on NT phonotactics. Which is not really an issue though, because Adāta has a Gezoro substratum, and Gezoro did not allow as many clusters as Ndak Ta.

A couple of things to point out:

First and foremost, those sound changes are not authoritative- that list was not written by me, and I'm afraid the "attested" Adata words here are to be treated as canonical even when they are seemingly "contradicted" by that soundchange list. To borrow a term from Ill Bethisad, quod scripsi scripsi.

Secondly, development of stress from Ndak Ta to Adata is complicated- I remember gleefully making it needlessly complex and introducing irregularities at whim (my grammar-writing four years ago didn't reach the faintly obsessive levels it has done since either- this is also probably why I've not actually released anything substantial since Proto-Western). So honestly, it might just be safest to assume that the rule given in the grammar is a tendency, not an absolute. Appeals to substrata, dialects etc can be made.

[Radius Solis]

Does anyone know where I can find the sound changes for the Isles languages? I know that Basilius designed Affanonic (have you put up the sound changes anywhere?) and has found correspondences for Pparwak, and Cedh has done an etymological dictionary for Mutsipsa', but has anyone done the other ones yet? I'd be interested in helping out if I can.

Welcome back to the fray. As far as I know, you've just listed the full extent of work that's been done-- for the most part, what you see on the wiki is what you get. It would be great to have you helping.

Of course there were the original attempts at reconstruction way back in 2005, but most of those came to nothing, except for Zomp's. And I don't know where his reconstruction is, or if it's even still online.

[Ran]

Thanks for the welcomes. =) I've been looking over the conworld and while I'm still really unfamiliar with it, I'm impressed with Basilius' analysis of the Isles language family and I think that we should go with it.

So how about something like this? I'm not sure if this would work with the geography we have but it might make the most sense.

Instead of the Isles family, let's have a Tuysafa-Ttiruku language family. Proto-TT is the ultimate PI that Basilius has been talking about; my PI is now Proto-Tuysafa, while the current Isles languages are reclassified as Ttiruku languages.

This means that we would need to design Proto-Ttiruku (which can be based on zompist's reconstruction!! which would be pretty cool), then make Proto-TT from Proto-Tuy and Proto-Tti. (Proto-Tuy and Proto-Tti would probably end up as mutually intelligible dialects of PTT.) Once we have that, we would have both a more "naturalistic" classification for the I languages, as well as a springboard for future development in Tuysafa.

[zompist]

My reconstruction is still up, here:

http://www.zompist.com/proto-ran.html

[Ran]

Awesome! Thanks, I've been looking for that for ages.

[dunomapuka]

I have just uploaded a big wordlist for Proto-Fisherman.

[jmcd]

I'm fine with it sticking to the old changes except that I prefer the treatment of the affricates in the new changes. And what I said already about the treatment of h\. This is because with the old changes the voicing system and the points of articulation available to plosives goes straight back to the Adata one.

I'm done running the lexicon through the old changes, and I'll put up the results on the wiki tomorrow.

I agree that the shift of approximants to voiced stops is awkward, it pretty much just reverses the changes from Adata to Ayasthi.

That is also true but what I was referring to is the fact that the plosive inventory of both Adata and old Agaf consists of p, b, ph, t, d, th, k, g and kh. I think it would be best to avoid this. Avoiding it can be done by taking the changes which apply to Ayasthi affricates from the new changes list instead of the old one. That way, we'd have palatal plosive added but not aspirates.

[TzirTzi]

My understanding of Adata phonology was that "i" designated /j/ only in post-vocalic position, and /i/ everywhere else.

I do recall pointing out that you were wrong there

Legion - given the above, would you prefer I continued with the idea that Adata <i> was only /j/ postvocalically (which would mean having a sound change [j]> C_, #_) or shall I derive new words with /j/ in other positions?

[Legion]

Tzirtzi > I'd prefer you continue with this idea, yeah (another possibility is to take Adata phonology as it, but to make the [j] fortition change conditional, this way: j > Z /V_), if only to keep consistent with the existing vocabulary.

[Zhen Lin]

No, I'm referring to verb stems like banê, banáza- or dēmê, dēmáza-, i.e. verbs which ended in a diphthong in Ndak Ta. IIRC the regular Adāta sound changes would result in a stress shift to the initial syllable in the long stem. But of course this may simply have been leveled by analogy.

Ah yes, those. I wasn't really sure what to do with them (verbs with final long vowels) really. I suppose I could revise Ayasthi again to reinstate this conjugation... but otherwise it could be interpreted as analogy, yes.

[Basilius]

zompist > Wowwww!

So, we still have a wordlist for Thokyunèhotà :)

BTW, it seems that the wordlists you used may include items not attested otherwise. Do you have the wordlists themselves?

Ran, Dewrad > welcome back, you've been missed! :)

Legion > I still suggest that you make Thokyunèhotà public - we are already warned that you consider it "noobish" :)

And, if Thokyunèhotà is not completely abandoned, what if we still use the old name for the archipelago, Thumapahìthì? It's been used nobody knows how many times in discussions...

As for 'do' -> imperfective marker, wouldn't it be better to use the imperfective of 'do', níni?

boy #12 > the wordlist looks promising :) waiting for the grammar...

Ran > no, I haven't published my SC's for Affanonic yet, which is indeed a big fault of mine. There are tricky points with them, so reconstructing them would be really difficult without a few cues...

My correspondences for Ppãrwak are based on very scarce material and leave a lot of questions unanswered. Some of them may be difficult (e. g. the attested SC's if applied without any corrections will lead to rather unnatural phoneme frequencies).

Corumayas and Cedh tried to reconstruct the SC's for Mûtsipsa' and Máotatšàlì, and I attempted at Máotatšàlì too; it seems that everybody got stuck with the Máotatšàlì tones... :(

As for possible regroupings - my current impression is that the two "Northern" langs are most deviant, but their potential archaisms are different; that is, it remains unclear if they form a natural subgroup.

Zele, Ppãrwak, and especially Naxuutayi are much more "conservative" in the sense that their ancestral dialects may need only minor adjustments differentiating them from your description of PI.

Affanonic was started when I still hoped to deduce everything from your version of PI; accordingly, it only implies a different quality of /h/ and a special treatment of prosodics in "true" reduplicates ("true" reduplicative syllables are divided from the stem by "=" in the lexicon).

BTW, I kinda like the term used by Zompist: "Ranic" ;)

[EDIT]

In other words, right now the probable subgroupings look approximately like the following:

(Ppãrwak, (Zele, (("Original PI", Affanonic), Naxuutayi))), Mûtsipsa', Máotatšàlì.

[Ran]

Basilius: > Do you have an explanation for the way the Isles languages came out the way they did? For example, why would Affanonic and Naxuutayi be closer to "my PI" than Pparwak and Zele?

[Basilius]

Basilius: > Do you have an explanation for the way the Isles languages came out the way they did? For example, why would Affanonic and Naxuutayi be closer to "my PI" than Pparwak and Zele?

Do you mean an extralinguistic explanation? I think Salmoneus' idea (cf. also subsequent discussions somewhere around this post) may work: the initial spread was very rapid, so the geographical distribution of the daughters does not match their genetic relationships.

On the other hand, the areas where Affanonic and Naxuutayi are spoken are geographically close to each other, while the most deviant northern langs move furthest... the latter in particular may mean that the respective deviant dialects of PI were spoken in a peripheral area (still in Tuysáfa) which was affected by the invasion earliest?

Or do you mean the grounds for grouping the languages like I did? It is all based on a tentative weighing of the potential archaisms. For example, I am inclined to think that the agreement rules in Mûtsipsa' are indeed better interpreted as a rudiment of nominal classes (although Rory hardly meant that), and that the use of tones in Máotatšàlì morphology (derivational in particular) is based on an archaism of similar depth.

An important deviation from "your PI" in Ppãrwak and Zele (not found in Affanonic or Naxuutayi) is postpositive placement of attributes available as an option. In addition to that, Ppãrwak has its strange nominative marker -(i)p (would be a regular development from some **[-iʔ]), but particularly interesting are its verbal forms: some are difficult to map onto PI tenses/moods, while others seem to reflect a form of non-sensory present with last consonant of the stem left intact.

Oh, BTW, a question I wanted to ask months ago: what was your idea about the imperative in PI? None of the descriptions of the daughter langs mentions the imperative - guess why :)

[Ran]

Ah. I clearly forgot about it. Does this mean I can make something up?


-a -> -a
-ay, -i, -uy -> -ay
-aw, -iw, -u -> -aw
-am, -im -> -am
-an, -in, -un -> -an
/ _(s,h,q)#

No aspect distinction

EDIT:

I realize that this can be seen as inconsistent with the nonsensory present, where we have -u > -i instead of -u > -iw.

[Basilius]

Ran > Let's see...

In Affanonic that would imply the imperative always formed with an auxiliary. This would agree with the division of verbs into lexical perfectives and imperfectives, since e. g. imperfective imperative can be perceived as expressing the demand that a continuous/repetitive state should *begin*. Also, Affanonic levels out all the original PI tense distinctions by generalizing the sensory epistemic mood, and the auxiliary will help to leave this scenario as is. On the other hand, an imperative based on alternations will produce not too cumbersome forms if the auxiliaries are monosyllabic... that's fine.

However, all the other langs seem to preserve the tense distinctions with various rearrangements; without an investigation it is difficult to decide if additional forms based on stem alternations will fit well in the attested paradigms...

I'd suggest leaving an escape hatch: besides the form you've just devised, the language could have e. g. a more polite form based on a tense/mood that seems more appropriate to you plus a discourse particle. This would work worse for Affanonic but probably better for all the other langs.

I think u -> iw is OK: the non-sensory present form is probably irrelevant here, since **iwms is impossible.

[Ran]

I'm sorry, I meant the sensory, not the nonsensory present.

EDIT:

The sensory paradigm would suggest that and are analyzed as /i/ and /u/ but the proposed imperative would suggest that and are analyzed as /ij/ and /uw/.

[Ran]

How about we try to figure out the sound changes for the Isles languages, then see if the paradigm works? If we're going to design around the Isles languages we should probably figure out what we're working with.