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

[Corumayas]

It would be nice to have Zhen Lin's input on the Lotoka migration, if he has any.

Has anyone come to any conclusions from looking at the Isles sound changes and zompist's reconstruction? Thokyunehota and Mutsipsa' appear to share one or two developments (e.g. the shift *ay *aw > e o, which I think Zele has too); could they form a subgroup of some kind?

Naxuutayi also does this.

Hm. From their sound changes, it looks like Naxuutayi and Thokyunehota are likely sisters-- they share several developments (devoicing of voiced obstruents--at least word-initial ones, palatalization before *i, denasalization/monophthongization of *VN "diphthongs") and, maybe most importantly, they have identical reflexes of not only *ay *aw but also *uy *iw (> wi yu-- the second of these seems to be shared with Affanonic, but not the first).

They don't seem very similar grammatically though.

Does anyone think that having Thokyunehota date to > 500 years before the other Isles daughters is problematic? It has more sound changes than Naxuutayi or Mutsipsa', but I don't think that's an issue in itself.

[Zhen Lin]

I believe what cedh said was what was previously agreed upon.

[Legion]

Hm. From their sound changes, it looks like Naxuutayi and Thokyunehota are likely sisters-- they share several developments (devoicing of voiced obstruents--at least word-initial ones, palatalization before *i, denasalization/monophthongization of *VN "diphthongs") and, maybe most importantly, they have identical reflexes of not only *ay *aw but also *uy *iw (> wi yu-- the second of these seems to be shared with Affanonic, but not the first).

That's because Naxuutayi was originally intended as a (deep) revision of Thokyunehota.

Does anyone think that having Thokyunehota date to > 500 years before the other Isles daughters is problematic? It has more sound changes than Naxuutayi or Mutsipsa', but I don't think that's an issue in itself.

I don't think that's a problem myself.

[Drydic]

Does anyone think that having Thokyunehota date to > 500 years before the other Isles daughters is problematic? It has more sound changes than Naxuutayi or Mutsipsa', but I don't think that's an issue in itself.

Think Hittite versus Homeric Greek/Sanskrit.

[Legion]

Tzirtzi

Some notes about the Kozado lexicon:

Verbs in Kozado have lost the habitual aspect of Adata, so the citation form is the imperfective aspect instead, which comes from Adati suffix -si, leading to -s in Kozado.

As Kozado simplified the verbal system, the 5 possible endings (according to the stem vowel) have been reduced to only two by analogy, this way:

Adata > Kozado
-esi > -es
-isi > -es
-usi > -es
-asi > -as
-osi > -as

For verbs in -asi, Adata often had irregular habitual forms, ending in a consonant (thus "zin", to live, has an imperfective form "zimasi", and leads to Kozado "zemas"), and it may be sometimes needed to refer to the Ndak Ta etymology.

So in the lexicon, you listed three verbs as leading to just "a" (to see, to turn, and to farm), but coming from Adata "ina", "en" and "a:n", from Ndak Ta "ina", "em" and "ain", they would rather lead to Kozado "enas", "amas" and "anas".

Moreover: most -as verbs in Kozado have an irregular verbal noun (because the suffix induced palatalisation), which must be listed in the lexicon. In the case of the three verbs here, these would be "iñeža", "emeža" and "eñeža".

Additionally, all verbs have two meaning, an static one, and an inchoative/dynamic one, see the original lexicon for examples.

"enas", which is part of the original lexicon, has thus the meaning "to watch" and "to start to watch".

Ultimatly, many formaly deponent -as verbs have been reanalysed as regular -es verbs, often leading to dublets.

"enas" is an example: in the lexicon, you can also find, from the same root, "enes": "to see/to glimpse"


Oh, and one more thing: for each verb in the lexicon, one must also known if they use class I or class II prefixes. This one is easy: if the first syllable of the verb had a front vowel for nucleus in Adata, they use class II prefixes. Otherwise they use class I.

Thus, "enas" and "amas" from ("ina" and "en") use class II prefixes, while "anas" (from "a:n") uses class I prefixes.


There, sorry for the crazy (there may be other diachronical informations scattered in the Kozado grammar, do not hesitate to have a closer look).

[TzirTzi]

Thanks for all the info! I thought that I'd probably have to do quite a lot more sorting out with the verbs - which is why I haven't put it into a more "official" page yet. I'm afraid I'll be away in a place with no internet (there are still such places!) for about a week and then on and off elsewhere afterwards, so it will be a while before I get much more done. But it's on my to-do list!

[Corumayas]

Legion:

I'm still confused about the tone shifts at the end of your Neo-Agaf sound changes. The changes look like this:

[middle tone] > [low tone]
V[+rhotacised][high tone] > [-rhotacised][high falling tone]
V[+rhotacised][low tone] > [-rhotacised][falling tone]
V[high tone][low tone] > [high falling tone]
V[low tone][high tone] > [high tone]
V > [+stress] / _C0#
[low tone] > [high tone] / [high tone]C0_
V > 0 / #_$

The broad question I want to ask is, what tone contrasts did you envision the language having after these changes?


In more detail:

Per your suggestion earlier, I added a rule (after the [M] > [L] change) combining [HH...] > [H] and [LL...] > [L] within a syllable. This leaves five possible tones for syllables:

high [H]
low [L]
rising [LH]
falling [HL]
rising-falling [LHL]

First of all, how are the contour tones affected by rhoticization? Surely a rhoticized contour tone shouldn't turn into a string of consecutive falling contours...

Secondly, what is the distinction between the "high falling" and "falling" tones created by your rules-- is high falling [HM] and falling [HL]? If so, shouldn't they merge in the [HL] > high falling rule that follows?

Thirdly, what outcome would you suggest for the rising-falling tone? Your rules suggest that it could merge into either the high or the high falling tone, but as written they don't quite seem to produce either outcome. (Keeping it distinct is fine, but seems slightly odd since it's somewhat uncommon and can only occur in initial syllables-- it's a variant of the rising tone, which has the same restriction but occurs much more frequently.)

Fourthly, does the shift of [L] > [H] after [H] work only with level [L] and [H]? And is it recursive? ([L] mostly occurs after the Ayasthi accent and [H] mostly before it, so there's a danger that this rule could produce a bunch of words that are all [H]...)


Lastly, a non-tone-related question: deleting word-initial vowels will delete a lot of stressed vowels. Does the stress move to the next syllable, or what?

[Legion]

Corumayas

I'm not sure what I wanted to do exactly originally; looking at what I wrote, here's what I deduced and the corrections I suggest:

The tonal change should be organised in three steps.

Step 1:

[LHL] > [L]
[LH] > [H]

Step 2:

[HL] > [H-falling]

Step 3:

[H][+rhotacized] > [H-falling]
[L][+rhotacized] > [L-falling]
[H-falling][+rhotacized] > (no change)
[+rhotacized] > [-rhotacized]


This gives us a four tones system:
High [H]
Low [L]
High-falling [HM] (or [HHL] if we don't want an [M] element)
Low-falling [ML] (or [HLL])

Fourthly, does the shift of [L] > [H] after [H] work only with level [L] and [H]? And is it recursive? ([L] mostly occurs after the Ayasthi accent and [H] mostly before it, so there's a danger that this rule could produce a bunch of words that are all [H]...)

It works only with [L] and [H], but is not recursive (this rule I had conceived because iirc otherwise the [L] was too dominant).

Lastly, a non-tone-related question: deleting word-initial vowels will delete a lot of stressed vowels. Does the stress move to the next syllable, or what?

Actually, this rule: V > [+stress] / _C0# explicitely means that stress moves in the final syllable unconditionally (it's not to be confused with the earlier rule [+voc +long] > [+stress] / _#, which had stress moved word finally only for long vowels in open syllables).

So the vowels deleted word initially afterwards are not stressed anymore (I think I did this rule to make some intervocallic realisations of consonants more phonemic).

[Corumayas]

Legion:

Those changes work perfectly, thanks. I have one question left, which I think is really the last one:

When [K] becomes voiced because of this rule:

C > [αvoiced] / _C[αvoiced]

does it become [l] or [K\]?

Also, does anything special happen with approximants, [r], and nasals, since they don't have voiceless equivalents otherwise?

(I assume the rule is recursive, so all clusters end up with homogeneous voicing.)

[Legion]

When [K] becomes voiced because of this rule:

C > [αvoiced] / _C[αvoiced]

does it become [l] or [K\]?

Also, does anything special happen with approximants, [r], and nasals, since they don't have voiceless equivalents otherwise?

(I assume the rule is recursive, so all clusters end up with homogeneous voicing.)

[K] becomes [K\]. For this consonant and the voiced one without voiceless counterpart, the result of voicing assimilation will only exist as an allophone, so nothing particular should happen to it.

And yes, the rule is recursive.

[Corumayas]

Ok. So I take it [l] > [l_0] (not [K]) also. I think it might be hard to keep those distinct, but I'll leave it for now. I should be able to put a wordlist on the wiki soon (probably after the weekend). The reflex of "Ayasthi" is [eɟ˦rɨ˦]...

[Corumayas]

I should be able to put a wordlist on the wiki soon (probably after the weekend).

It is done. I wasn't sure how to syllabify the CCC clusters, but I decided to ignore that issue and just put the lexicon up as is.

The two sets of sound changes produce completely different languages. jmcd, what do you think of them?

[Legion]

Awesome! I clearly like the result of the new changes better.

[jmcd]

I'll hopefully make more constructive comments later but straight off I see some weird symbols I don't know what mean like ˥˧, ˨ and ˧˩. Do these show tone or what?

[Zhen Lin]

Yes. Tone contours, to be precise.

[Corumayas]

Yes-- there are four tones: high [˦], high falling [˥˧], low [˨] and low falling [˧˩].

(Edit: I also see that the under-ring doesn't show up on my computer at work-- it's not very important though, it just shows that a consonant is allophonically devoiced.)

Edit again: the phoneme inventory appears to be
plosives /p b t d c J\ k g/
nasals /m n J/
fricatives /f v T K s z S Z C j\ x/
trill & approximants /r l j w/
vowels /i e a o u 1/
plus the four tones.

For comparison, the old version (after some further changes Legion suggested) has
consonants /p b t d k g m n N f s x l/
monophthongs /i y e 2 E 9 { A O o u 1/
diphthongs /EI 9I {I OU {U j{ jA/
plus vowel length.

[Corumayas]

Hey, remember how we were saying we should start our own forum? Is there any prospect of that happening? Carrying on multiple conversations in this thread is difficult, and finding previous discussion on any given topic is a major pain, even if it was relatively recent.

If a dedicated forum isn't feasible, we could make more use of the wiki talk pages. But I'm afraid not as many people check there regularly, and sometimes there isn't an appropriate page to attach a discussion to.


Among the things I'd like to discuss right now:

-Climate-- kind of a prerequisite for everthing else, I think. This depends on the map being finalized, which really should be a high priority for us.

-"Prehistory", especially as it affects the Proto-Eige-Isthmus speakers. I need to figure out where they lived and what their culture was like, and I'm not sure how that's affected by, for example, Salmoneus' Proto-Canoe scenario from a few weeks ago. The history of agriculture and technology is also super important.

-Places where there's room for more languages (and language families). This is interesting on both the large scale (e.g., I think there should more small families like Lukpanic on the west coast of Peilash-- maybe lots more), and on the small scale (I also think there could be a few more Adata daughters in the western Rathedan)... and it could be helpful to have such a list when people ask "how can I contribute?"

-The new relay, if we're still intending to do one.

-And various narrower subjects, like the Isles family reconstruction project, the early history of Sumarushuxi (including before the Isles speakers arrive), the actual languages I'm supposedly working on, etc....

[Zhen Lin]

I'm working on a Peninsular language now. Somewhat. We'll see.

[jmcd]

E, I'll admit I don't really know much about tone.

I'm sorry I took so long to answer.

Maybe we could even have two separate languages coming from it since the 2 sets are so different?

Unless we're doing that, I think it's probably otherwise best to just stick with the old changes except where the
old ones essentially go backwards.

More specific comments about what's turned out:
Homophones:
In the 1st set, maybe we should replace the word for 'sour' or 'praise' with another root since they've become the same monosyllable. Suppletion is probably kind of useful for either 'see' or 'capture'. 'to carve/'legal'(gɑt) and 'neighbour'/'gossipmonger'(geˈbɛ) have also merged in the 1st set.

And 'sweat'/'dig' and 'steal'/'play' and 'fight'/'think' 'casualty'/'owner' and 'dawn'/'grow' and 'fingernail'/'nose' were already like that. The 1st set of changes adds goat and low to the last group and the 2nd set adds big and low.

And in the 2nd on top of the homophones that already existed in Ayasthi, the words for 'to reward'/'dirty'/'to need'(ɟeʝ˧˩) and 'thick'/'ancient'(ɟe˧˩ʝɨ˨) and 'sweat'/'slut'/'skeptic'(ki˦cɨ˦) and 'ashes'/'daytime'/'river'/'to fall'/'feeble'(cɨ˦) and 'violent'/'powerful'/'urban'/'person'/foreign city'(mi˦ɟi˦) and 'blind'/'idol'(bi˦) and 'to lose'/'to fold'/'to act'(ɟen˦) and 'ready'/'loud'/'valley'(pi˦ɟi˦) and 'to walk'/'language'(ɟi˦) and 'to lack'/'to kneel'/'bark'/'to decide'/'to swim'(beʝ˧˩) and 'steep'/'to be'/'dog'(ʃi˦) and 'to mimic'/'to be depressed'(ʃi˦ɟen˦) and 'to be (past plural)'/'to meet (of councils)'/'leftward'(vi˦) and 'play music'/'try, attempt'(wos˥˧) and 'to spit'/'to sacrifice'(o˦, the way they worship the gods?) and 'heart'/'overseer'/'to count'(mi˦cɨ˦) and 'true'/'to float'(res˦) and 'fire'/'ugly'/'north'(ɬɨ˦) and 'modern'/'or'/'name'/'to journey'(mi˦) and 'city dweller'/'visible'/'doctor'/'to destroy'/'to cure'(ʝɨ˨, curing goes with doctors but not with destroying) and 'country'/'to shut'(la˦) and 'space'/'rabbit'(wi˦ɟi˦) and 'fan'/'stubborn'(wu˦bu˦ ) and 'active'/'yellow'/'upward'/'egg'(gi˦) and 'guilty/'tree'(ji˦) and 'walk'>'limp'/'cook'(e˦gji˦) and 'meet'/'trust'(mi˦go˦) and 'angry'/'Etúgə'(ju˦go˦) and 'to clean'/'sweet'(go˦) and 'mouse'/'to be (present plural)'/'to record' (jɨ˦) and 'soon'/'to promise'/'to grieve'/'chief'/'chieftan'(rɨ˦, last 2 clearly related so no problem) and 'pure'/'again'/'swamp'/'to choke'(vu˦) and 'goal'/'complex'(dos˦) and 'dark'/'to hope'/'to guess'/'to wound'(ɟo˦) and 'national'/'healthy'/'a little'/'to form'(li˦ɟi˦) and 'hit'/'remember'(zof˥˧) and 'knife'/'cloud'/'lucky'/'female'(wi˦) and 'coast'/'left'(ju˦) and 'herb'/'garden'/'uncle'(ʒi˦ɟi˦, those 1st 2 have similar meanings though) and 'salt'/'oath'/'touch'/'use'/'squeeze'/'snake'/'corral'/'hole'/'knot'/'finger'/'chilly'/'moderately'/'well'(ɟi˦) and 'laugh'/'share'(ja˦ɟi˦) and 'squirt'/'to be clumsy'/'white'/'to weary'/'bad luck'/'normally'/'male'(ɟi˦ɟi˦) and 'tool'/'to be hot (weather)'(ɟi˦ɟɨ˦) and 'right'/'skin'/'eye'/'adult, mature'/'mountain'(ɟu˦) and 'life, survival'/'be bold'/'bad'(ɟu˦bu˦) and 'deep'/'far'/'green'/'country dweller'/'grass'/'frost'/'liver'/'to bloom'/'to freeze'(bo˦, green/grass/to bloom on the one hand and frost/to freeze on the other hand though fair enough) and I'm sure there's more are now the same. We can of course have homophones but maybe that's a bit much for such a wee vocab list. English has you, ewe and yew but it doesn't top the list off with yoo, yu and ew. And even if it did, /ju:/ wouldn't have as many meanings as ɟi˦. And that's with e˦ɟrɨ˦ having a much weeer vocabulary. And that's just counting the ones that sound exactly the same, not just similar like ʃi˦ɬɨl˦ (there, yonder) and ʒi˦ɬɨl˦ (here).
And suppletion is good to do anyway since languages don't keep the roots for words constant and homophones should probably be the 1st place to look to do suppletion.

In the 1st set, are we going to have voicing assimilation for words like ˈkɛfbɑs?

In the 2nd set:
I do think that there's too much ˦ if it's going to be a tone language. Maybe there should be more processes which cause the tone to be otherwise.

I think reducing the extent of umlaut would help keep words distinct and have less /i/. Having phonemes deleted without them affecting the rest of the word at all is probably also a cause of the large amount of homophones. We definitely need more normal plosives. Normal as in not having half of them palatal. Could we be radical and just turn all nasal stops into voiced plosives and then from there into voiceless ones if they're in the right situation? /T/ could go to /t/ as well.

How can you get 2 tones on the same syllable like in oʃ˥˧ ? Or does it here mean high falling?
See in words like o˦ggɨ˦ where you have 2 consonants the same following a tone symbol, shouldn't it be better to make it either og˦gɨ˦ or o˦g:ɨ˦?

Is the 1st 'n' in vo˦bnbon˦ syllabic? If not, it seems a mighty cluster compared to the rest of the language. If yes, it seems to be the only syllabic consonant around.

wɨ˦ɟzvu˦ is both unusual and difficult. I suppose we could make e˦ɟrɨ˦ slightly Caucasian though.

ʒe˦ʝeʝ˧˩ and ɟe˥˧ʝʝɨ˨ seem especially overloaded in similar sounds. This can also be seen in the homophone examples above. Maybe there could be a dissimilation process so there's less palatals next to each other?

The j being retained in we˦ɲji˦ seems odd.

On the other hand, I like the way ɬɨl˦ comes to be a word for everywhere as well as meaning -where and -place.

[Corumayas]

E, I'll admit I don't really know much about tone.

Hm, that might be a good reason to eliminate the tonal contrasts. Or do you want to learn by doing a tonal lang?

I'm sorry I took so long to answer.

No worries; it took me forever to work through the sound changes. Anyway I'm in no great hurry here.

Maybe we could even have two separate languages coming from it since the 2 sets are so different?

That thought had occurred to me. I don't know if there's room for two languages, though, unless half the Ayasthi-speaking population suddenly migrates somewhere fairly far away.

I was hoping we'd be able to take some changes from each list and kind of combine them, but that might be difficult since they're so different.

Homophones:

Not all of those are super problematic (e.g. where they're different parts of speech); but yeah. I didn't attempt any semantic shifts or suppletion or anything, and we should obviously do that.

(I also didn't post paradigms as that would take up way too much space, but I have the various stem forms for each word. Some of the ones whose citation form merged still inflect differently.)

In the 1st set, are we going to have voicing assimilation for words like ˈkɛfbɑs?

We probably should. Legion didn't indicate how to deal with those clusters, so I left them alone.

I do think that there's too much ˦ if it's going to be a tone language.

I definitely agree. There's a rule that changes low to high after a high; we should probably eliminate that rule. Without it, a large number of the current H-H words will become H-L.

There's also a dialect of Ayasthi where all words begin with L-H (instead of contrasting L-H with H-L). Maybe we should use that dialect as our starting point to have more L in initial syllables too.

I think reducing the extent of umlaut would help keep words distinct and have less /i/.

Yeah. Another possibility is using the diphthongization from the old rules to reduce the amount of /i/ (by changing some of it to /EI/).

Having phonemes deleted without them affecting the rest of the word at all is probably also a cause of the large amount of homophones.

Agreed, there should be less of that.

We definitely need more normal plosives. Normal as in not having half of them palatal. Could we be radical and just turn all nasal stops into voiced plosives and then from there into voiceless ones if they're in the right situation?

The old rules have that change (except for the devoicing part). It seems pretty drastic, though...

I think some of the consonant changes in the new rules are just too arbitrary. A lot of consonants shift to palatals for no apparent reason (except to create more palatals).

/T/ could go to /t/ as well.

It could; on the other hand it comes from Adata /t/ so that wouldn't be very interesting.

How can you get 2 tones on the same syllable like in oʃ˥˧ ? Or does it here mean high falling?

Yes, that's high falling. (For some reason, the two tone symbols display separately on the wiki, but on the board they merge into one... the board has better unicode implementation, apparently)

See in words like o˦ggɨ˦ where you have 2 consonants the same following a tone symbol, shouldn't it be better to make it either og˦gɨ˦ or o˦g:ɨ˦?

That would make sense. I was sloppy with that b/c it would've been way more work to move all the tone symbols to match the new syllable boundaries after all the syncope.

Is the 1st 'n' in vo˦bnbon˦ syllabic? If not, it seems a mighty cluster compared to the rest of the language. If yes, it seems to be the only syllabic consonant around.

etc.

Yeah; the syncope rules delete vowels without regard to what consonants there are or what order they come in, and they produce up to three-consonant clusters. It's very messy. My preference would be to drop vowels only V(C)_(C)V, so that only two-consonant clusters can occur. (The current rule is V(C)(C)_(C)V.)

ʒe˦ʝeʝ˧˩ and ɟe˥˧ʝʝɨ˨ seem especially overloaded in similar sounds. This can also be seen in the homophone examples above. Maybe there could be a dissimilation process so there's less palatals next to each other?

The j being retained in we˦ɲji˦ seems odd.

I agree. We could add some dissimilation rules and a rule deleting [j] after a palatal consonant. (We could also try creating fewer palatals in the first place and see if that helps...)

[Zhen Lin]

[quote="Corumayas"Yes, that's high falling. (For some reason, the two tone symbols display separately on the wiki, but on the board they merge into one... the board has better unicode implementation, apparently)[/quote]

It's a font thing. Some fonts can convert sequences of level tone symbols into contour tone symbols. I believe Doulos SIL (used on the wiki) is one of them. (Well, to be precise, the font has the glyphs and various conversion tables, but it depends on the renderer what gets shown.)

[Legion]

Corumayas > As for new families of languages, I'm playing with some ideas right now, that I originally planed to put in southern Zeluzhia, but which could probably fit in any place not linguistically described yet.


Corumayas + jcmd > For tone, I suggest instead we do this: voiced obstruant (occlusives + fricatives) get voiceless, while changing a following high tone into low.

/T/ could go to /t/ or 0.

Dissimulation (either progressive or regressive) of palatals is a good idea.

We can make continuant consonants (nasals and approximants) to ease clusters.

Clusters should probably be simplified anyway.

But those are only suggestions, if you want to take it to a given direction, feel free to do so

[TomHChappell]

Dissimulation

Is it dissimulation or dissimilation?
Doesn't one mean "making things dissimilar" and the other mean "lying"?

[dunomapuka]

Among the things I'd like to discuss right now:

-Climate-- kind of a prerequisite for everthing else, I think. This depends on the map being finalized, which really should be a high priority for us.

Climate, and discussion of local flora/fauna hinging on that. I'd like to develop the Lukpanic culture some more, but I don't know anything about what kind of climate they live in and what kinds of things live and grow there.

[Legion]

TomHChappel > I keep confusing those two (dissimulation means more like, "hiding").