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

TzirTzi · Post by **TzirTzi** » Fri Oct 02, 2009 11:42 am

Radius, that does indeed look pretty awesome. The various stem forms make for fascinating looking sets - how regular will the semantic correspondences between them be?

cedh - if I was qualified to comment on climate properly, then I would. However, can I add my vote that increased axial tilt would seem an interesting option to me. ^^

Radius Solis · Post by **Radius Solis** » Sat Oct 03, 2009 11:26 am

TzirTzi wrote:Radius, that does indeed look pretty awesome. The various stem forms make for fascinating looking sets - how regular will the semantic correspondences between them be?

The J-grade has merged with the PEI i-voice, at least in that a given PEI root's J-grade form and i-voice forms end up the same. So there will be some semantic confusion between the two, and I can't decide yet what the result of that is because I don't know what the J-grade marked.

The N-grade for such verbs as preserve it will fit into the modern aspect paradigm fairly neatly, as a formative sub-class (comparable to the preservation of IE strong forms in some English verbs). As such, it will form a consistent aspectual operation.

The S-grade is lost in the fewest stems and is the most predictable to form, so I am thinking to keep it as a still-regular fossil derivation that, loosely speaking, forms resultative nouns. (Another echo from IE there - English retains only one ablauted resultative noun that I can think of, sing -> song, but I'll be keeping at least several dozen)

On to the vowels....

Perhaps a better way to illustrate the vowel alternations than the last post, is to show how they arose. The key is a stage in Eastern Isthmus where stressed vowels were lengthened before voiced consonants, unless they were already part of a diphthong in -j, or in the enviroments _CC$ or _CC+ (+ denotes a morpheme boundary other than grade infixes). Thus zero-grade roots could have either vowel length, and long roots in the S-grade and J-grade always shortened, while N-grades could go either way.

So let's look at the possibilities. Let's make up the following PI roots, all original a-voice stems, and see where their regular outcomes land. Stem classes indicated in parentheses.

Code: Select all

grade:  0        N         J        S
kat >   hat(a)   hant(a)   hat(e)   hast(a)
ket >   het(a)   hent(a)   het(e)   hest(a)
kit >   hret(e)  hrent(e)  hret(e)  hrest(e)
kot >   hot(a)   hont(a)   heot(e)  host(a)
kut >   hweot(a) hweont(a) hwit(e)  hweost(a)
kad >   head(a)  hand(a)   had(e)   hasd(a)
ked >   heod(a)  hend(a)   hed(e)   hesd(a)
kid >   hreid(e) hrend(e)  hred(e)  hresd(e)
kod >   houd(a)  hond(a)   heod(e)  hosd(a)
kud >   hwaid(a) hweond(a) hwid(e)  hweosd(a)

roots with complex codas to start with, assuming they are not from infix grades (or else grades lexicalized already in PI):

Code: Select all

        0          N          J          S
mant >  bant(a)    beart(a)   bant(e)    bast(a)
ment >  bent(a)    beort(a)   bent(e)    best(a)
mint >  brent(e)   breirt(e)  brent(e)   brest(e)
mont >  bont(a)    bourt(a)   beont(e)   bost(a)
munt >  bleont(a)  blairt(a)  bleont(e)  bleost(a)
malt >  balt(a)    beart(a)   balt(e)    bast(a)
melt >  belt(a)    beort(a)   belt(e)    best(a)
milt >  brelt(e)   breirt(e)  brelt(e)   brest(e)
molt >  bolt(a)    bourt(a)   beolt(e)   bost(a)
mult >  bleolt(a)  blairt(a)  blilt(e)   bleost(a)

I think there are also other possibilities these examples don't explore, but let's leave it here for now and see what we can predict based on the zero grade of the modern language.

1. a > a, but in the N-grade sometimes ea instead
2. e > e, but in the N-grade often eo
3. re > re, but in the N-grade often rei
4. o > o in the S-grade, sometimes ou in the N-grade and always eo in the J-grade
5. weo > weo in the S-grade, but sometimes wai in the N-grade and sometimes wi in the J-grade (also, w > l after labials)
6. ea > a, and eo and ei > e, not including weo, but all sometimes remain unchanged in the N-grade
7. ou > o in the S-grade and eo in the J-grade, and can be either in the N-grade
8. wai > weo in the S and N-grades, and wi in the J-grade
9. wi (not shown) > wi most of the time, but sometimes weo
10. i, u, and oa are not found in native words (except in wi), so verbs with those vowels do not show gradation

Much of the mess arises from trying to analyze the alternations as having reversible direction. For example a and ea often alternate in either direction, but do not always do so even when it might be expected. Similarly weo, wi, and wai alternate with each other but with poor predictability - at most, one can broadly say that wi and wai become weo in some forms and that weo can become either wi or wai.

The unfortunate part is that the alternations just aren't predictable enough - even knowing the voicing condition for lengthening, because that environment was sometimes messed up - so I am going to have to individually derive all the non-zero grades by manually sound changing them. UGH. But there's no help for it. Even in the above tables I have frankly probably made some mistakes because, despite having controlled the conditioning environments apurpose, it was still hard to keep track of all the moving parts and I don't really want to spend an extra hour right now on checking each individual form to be sure of it. The N-grade is the worst by far. I will probably end up smacking some sense into it with analogy, but not so much as to lose the interesting flavor.

A last note: In PEI, wherever an already complex coda existed in the zero grade, a copy vowel was inserted after the infix, such that the N-grade of e.g. *dasg-a- was *danasg-a-. In Kennan this ends up not mattering to the S and J-grades, but it does play with the N-grade. This is the source of -r- in the N-grade, as well as the source of so much of its variability; nasals became voiced stops except V_(C)$, and later these were lenited after the unstressed vowels collapsed. So *danasg-a- went through the following stages: *da:nasg-a-

*da:dasg-a-

*da:dsg-

dearg-.

TomHChappell · Post by **TomHChappell** » Sat Oct 03, 2009 12:35 pm

Radius Solis wrote:PEI

What's "PEI"? "Proto- Eiga-Isthmus" or something?
Maybe you meant "PIE" for "Proto-Indo-European"?

Radius Solis · Post by **Radius Solis** » Sat Oct 03, 2009 12:37 pm

Proto-Eigə-Isthmus.

There is some potential for confusion between that and Proto-Eastern-Isthmus, but I've been abbreviating that EastI or PEaI or similar.

TzirTzi · Post by **TzirTzi** » Sat Oct 03, 2009 11:23 pm

Radius -

Look's great

It'll be good to have a conlang in Akana that has something comparable to IE's oblique web of interrelated derivations and such.

So your scs include a lot of small, very specifically conditioned rules? To generate all of those "most of the time"s and "often"s ^^

You sound like you're finding applying SCs a lot of work - are you doing it by hand? Because especially if you're going to have to apply them separately to lots of derived forms, wouldn't it be easier to use IPA zounds or a similar SC applier?

Btw, I've very quickly jotted down a page with the acronyms used for languages. The two currently problematic pairs are Proto-Eige-Isthmus and Proto-Eastern-Isthmus (where I've followed your suggestion and put PEI and PEaI) and Proto-Isthmus and Proto-Isles (which are harder. I went for PI and PIl in the end - but anyone who disagrees is welcome to change them.)

Radius Solis · Post by **Radius Solis** » Sun Oct 04, 2009 2:22 am

Good idea on the abbreviations page. I made some additions, including some alternate options. For example "PNT" without hyphens gets me for being so easily confusable with NT, but at the same time I recognize it has achieved standardness, so we might as well list both.

--

IThe "often"s and the "sometimes"s result mostly from how often the sound changes interfere with each others' outcomes.

I certainly use SCAs anytime it's worth it, but for the style of SCs I tend to produce I have found that getting them encoded for a program to read is so much work that applying them manually is often faster! (Unless we're talking about many hundreds of words, which isn't the case here.) And Kennan is among my worst for that, too. The way I write sound changes, I can state a rule and know what *I* mean by it, in terms of all its contingencies, but getting a program to perform it all correctly can be a serious pain in the ass.

Or even impossible. There's actually a grammar-sensitive* rule in my changes for Kennan: many consonant clusters resulting from the Great Vowel Collapse are resolved in a manner that preferentially preserves grammatically salient consonants (such as the infixes). So with the example *danasg-a- I used before, you cannot tell just from looking at it if it's a monosyllabic root in the N-grade or a bisyllabic root in the S-grade, but the former ends up as dearg- and the latter as dasg-. And even when that's not in play, I would still be hard pressed to list the many dozens of little individual rules that add up to "clusters gravitate towards a (sibilant/approx)C(approx) structure where the C is whatever was least sonorous of all the original cluster components, around which all the others collapse and lenite as needed".

*a few such cases are attested in natlangs, despite the general rule of thumb that that doesn't happen. This instance results from analogic pressure.

the duke of nuke · Post by **the duke of nuke** » Sun Oct 18, 2009 6:45 am

I thought I'd add to this thread with a surprise contribution:
The Ndak abjad is now available as a .ttf. (link is to parent directory)
It's not really anything special, but I was bored and decided to make it; hopefully it will help anyone working on the writing of Akana. Its usage does require some explanation since the letters don't map very well to the Latin alphabet.
Latin m n p t k b d g s r l w all map to the corresponding Ndak letters.
Latin c j q v map to Ndak ts ng kw bw.
Capital letters are prenasalised, so Latin P T K Q B V D G C W map to Ndak mp nt ngk ngkw mb mbw nd ngg nts ngw

Here's an example - you can probably guess what it is

Radius Solis · Post by **Radius Solis** » Sun Oct 18, 2009 8:17 am

....wow. Well, cool, and I appreciate the effort you've gone to! The question of Ndak writing is a lot stickier than may have been apparent to you, though.

For one thing there are not one but two proposed systems on the table, mine and Legion's, for which we have both drawn up character sets; finding some way to use both systems, or a new home for one of them, would be desirable - but we just haven't gotten down and dirty in the subject long enough to work out a solution. For another, we had decided it would be unrealistic to have the full Ndak writing system be too alphabetical so early in the history - the Fáralo were supposed to be the first to invent an alphabet. In my original plan for the Ndak script, those letters were meant to be source characters for the phonetic elements of logograms (that worked on the rebus principle), rather than a standalone writing system to themselves. I'm not married to that specific plan, but at the same time, a simple abjad does bring up questions of realism: in the equivalent period on Earth, we hadn't yet moved very far beyond cuneiform and hieroglyphics.

But honestly I think it's just about time to give up and pick something, anything, so long as we're all reasonably comfortable with it. So let's see how the rest of the team weighs in. Do we want the abjad? Do we want to wait possibly many more years for me or someone else to get around to making a logography? And what about Legion's proposals from last year? I hate to just pass up work that people have already done, Legions's or Duke's either one.

Cedh · Post by **Cedh** » Sun Oct 18, 2009 9:58 am

That font is cool. I agree with Radius though that we should try to work out a good scenario for the development of Akana writing systems which includes earlier work as much as possible.

To facilitate this, I've compiled the relevant comments from earlier script discussions in a new thread at the AkanaForum. Let's discuss it in detail there.

Zhen Lin · Post by **Zhen Lin** » Sat Oct 24, 2009 4:44 am

I have moved my files over - most URLs should work if you replace www.geocities.com/low_zl/ by zhenlin.nfshost.com/akana/ but .csv.txt should be replaced by .csv. Geocities closes on Monday...

Nortaneous · Post by **Nortaneous** » Sat Oct 24, 2009 4:53 am

Radius Solis wrote:Do we want to wait possibly many more years for me or someone else to get around to making a logography?

There's no need for a full logography unless people want a logography. To develop a writing system, all you need is enough logographic signs to develop the next stage of the system.

Radius Solis · Post by **Radius Solis** » Sat Oct 24, 2009 10:26 am

Yes, we could have gone down that path. However, logographies are interesting, and the more characters you have defined originally, the greater the variety of stock to choose among in deciding what to derive later phonetic systems from.

So if you check the thread Cedh linked a couple posts back, you'll see we're already at work on one! Or at least the original ideographic system that everything else arose from. Turns out the amount of labor required isn't totally prohibitive. A little graphosyntax goes a long ways.

dunomapuka · Post by **dunomapuka** » Tue Nov 03, 2009 12:20 pm

This map has me a little lost. Which families are supposed to be which? Is Coastal yellow? Also, I didn't realize that the Western homeland was so close to the water...

the duke of nuke · Post by **the duke of nuke** » Tue Nov 03, 2009 1:13 pm

Coastal is blue, Gezoro and Tjakori are light purple, and Tmaśareʔ is either yellow or (more likely) green. The location of Çetázó is currently unresolved.

On the other hand, I imagine the map will be reworked sooner or later, so what I said can probably be disregarded.

eodrakken · Post by **eodrakken** » Tue Nov 03, 2009 3:11 pm

Dark purple is meant to be the steppe family, of which Nortaneous's lang and Iŋomœ́ should be members.

Corumayas · Post by **Corumayas** » Wed Nov 25, 2009 4:58 pm

Thedukeofnuke is working on a new daughter of Fáralo, and the subject of province/region names in Huyfárah came up on the talk page. I thought it might be good to move the discussion here where more people can see it; I'm especially hoping to get zompist's input.

As a frame of reference, Huyfárah seems to contain the following regions/provinces. I've listed them in roughly chronological order by conquest (dates mostly gleaned from this timeline), along with their major urban centers (where these exist/are known), and some suggestions for names (new names are in bold, old ones in italics).

Regions of Huyfárah

* Huyfárah proper (ancient or poetic name: Hagíbəl) may be subdivided roughly as follows:

** the highlands around/between the upper Oltu and upper Poráš (continuously Faraghin-speaking since before Tsinakan's time)
** the upper Oltu valley (conquered by the Faraghin ca. -1400; center at Barnágo)
** the lower Oltu valley and coastal plain (also conquered ca. -1400, but had to be reconquered from the Ndak ca -1170; centers at Ussor and Mæmedéi)
** Sertek (founded -762) and the lower Poráš valley (at the time this was the frontier between Faraghin and Feråjin territories)
** the upper Poráš valley (apparently absorbed after -762; center at Peimast)

The above was unified by Mentek of Ussor in -198.

Before unification, this region-- especially the Oltu valley-- was a patchwork of small baronies; maybe it was sometimes called Lu-Kənat "the barons". And maybe the parts that were Faraghin-speaking the longest could be referred to as Lu-Haran "the Faraghin". Thedukeofnuke has some other suggestions for the Oltu valley, like Šesuy-laš "crowning-land" and Oludu-laš "rights-land, inheritance-land"; these might be later names (the latter is a pun on the river name which works better in Woltu Farla).

Several ancient empires (early Rome, Achaemenid Persia) didn't incorporate their home territories into the province system, so it's quite possible that this region remained a patchwork of small territories for a long time.

Regions added to the empire after unification:

* Lu-Kpoun Dagæm (conquered -185, center at Agumosou)
* Oltumosou (settled after ca -650, brought under imperial rule -167) and Lu-Ferój ("the Feråjin", pacified beginning -167)
* Čisse (founded -142) and Lu-Doroh (the eastern frontier region, beyond the Feråjin)
* Miədu (first conquered -520 by Ussor, permanently joined to Huyfárah -133) and Lu-Rewŋlo ("the plains", between the coast and the uplands of Lu-Tal)
* Muəbaz (i.e. Momuva'e; first conquered ca -480 by Ussor, but lost and reconquered many times; only ever sporadically controlled)
* several other towns and districts of Kazəgad (e.g. Puwa, Luyoša, Ŋourlo, Nirola) with similarly complex histories
* Lu-Tal (fully brought under imperial rule under the Etúgin dynasty +230-319, though much of it was conquered earlier)
* Buruya (acquired +351)
* Wanenuteye (i.e. Fmana-hŋ-Talam, claimed +370)
* Tæm Hou (various colonies established in the ?fourth century YP)
* Lašumu (client state ca. +453-546)

A couple incidental points:

-Wanenuteye is meant to be a Fáralo-ization of /wAnEnu~tEjE~/, the name Zhen Lin originally suggested as a descendant form of Proto-Peninsular *fmana-hŋ-talam.

-I coined the name Etúgin dynasty for Etou I and his heirs-- it's a pun on the name Etou, but it also suggests the strong association of the dynasty with religious orthodoxy (so strong that political dissent took the form of religious sectarianism during their period).

This list doesn't necessarily reflect the political subdivisions of the empire, but I imagine it could. What do you all think?

dunomapuka · Post by **dunomapuka** » Fri Nov 27, 2009 5:30 pm

I find the name Wanenuteye a little unwieldy... any way we can shave it down more?

Fruithat · Post by **Fruithat** » Tue Dec 15, 2009 10:51 am

I have a proposal for Akana. I would like to place my proto-language Bwenkhir somewhere into the Akana world in its earliest stages and have it develop daughter languages.

Bwenkhir can be found here:

http://wiki.frath.net/Bwenkhir

Just tell me a bit about the geography and biology of Akana.

"Bwenkhir" means mountain people, so the proto-lang should probably be placed somewhere where mountains are.

This language started out as part of the Conproject, bot that website went down. I've worked too much on it to just abandon it.

the duke of nuke · Post by **the duke of nuke** » Tue Dec 15, 2009 5:55 pm

If you're looking for a place to put it, the mountains of west-central Tuysáfa are empty and waiting. They're much like the Urals, but a little warmer; to the west is a broad temperate coastal plain, and to the east is the great central steppe. Another possibility is the smaller mountain ranges near the south coast, which are mediterranean in climate.
Tuysáfa was settled in three waves, of which the first was palaeolithic, and the second and third neolithic.

Neon Fox, roninbodhisattva and I are currently doing work on Tuysáfa, and I'd be happy to include it if they are (and if the senior members don't have any issues).

EDIT: And it looks like another mysterious mountain civilisation has just appeared - the Wellawi, who inhabit a fertile plateau in the high mountains of western Peilaš. Ask cedh for details on them.

Cedh · Post by **Cedh** » Wed Dec 16, 2009 7:05 pm

Xeroderma Pigmentosum wrote:http://wiki.frath.net/Bwenkhir

Bwenkhir looks like it's fairly different from most languages we have in Akana so far, and I think we could find a place for it. However, I'd like to hear some opinions from the other senior participants before making a decision, and I'd also like to read your description more closely.

The latter thing is a bit difficult though: You don't use much formatting in your grammar sketch, and that makes it extremely hard to read because I have to check all the time whether the word I'm reading is Bwenkhir or English. It would be much easier to get a good picture of your language if you would, for instance, (a) use tables for the morphology paradigms, (b) write all Bwenkhir words in bold or italic, and/or possibly (c) set off glossed examples by using some kind of indentation.

abaddamn · Post by **abaddamn** » Sat Dec 19, 2009 5:28 am

boy #12 wrote:I find the name Wanenuteye a little unwieldy... any way we can shave it down more?

1. Wanenuteye
2. Wanenteye
3. Wanteye
4. Wante'a

Or
1. Wanenuteye
2. Wanenute'a
3. Wanute'a

Cedh · Post by **Cedh** » Sat Dec 19, 2009 6:50 am

In Fáralo, Wanuteye seems like a likely shortened form. Or maybe Wannuteye with geminate /n/.

dunomapuka · Post by **dunomapuka** » Sat Dec 19, 2009 3:04 pm

cedh audmanh wrote:In Fáralo, Wanuteye seems like a likely shortened form. Or maybe Wannuteye with geminate /n/.

I think it's the -eye that bothers me, it doesn't have any equivalent in Fáralo phonology. Maybe Wannutei?

Corumayas · Post by **Corumayas** » Sat Dec 19, 2009 6:40 pm

I was hoping zompist would give an opinion on this.

But in any case, Fáralo doesn't seem to have geminates; so maybe Wanutei would be better...?

Corumayas · Post by **Corumayas** » Sun Dec 20, 2009 6:21 pm

Thinking about it some more, it occurs to me that, of course, the form the name takes should depend partly on when it was borrowed and how much it was adapted to Fáralo phonotactic patterns. And that, since the island isn't extremely far away, it should maybe be borrowed earlier and adapted more (more than I was assuming, that is). (Cf. Histuəna < Siixtaguna, which is if anything more distant.)

Looking at the Fáralo sound changes, the following alternatives suggest themselves (NB: the native name is given as /wAnEnu~tEjE~/):

- The /A/ might be fronted to /æ/-- if I'm reading the sound changes right, all native /a/ does this before nasals.
- The /u~/ might become /u@/, if it was treated as stressed (as happened in Histuəna).
- Similarly, either the second or the last syllable might become /ei/ if stressed; but they seem unlikely syllables to receive stress.
- If the loan was very very early, the /j/ might become /S/; but I don't think that's likely either.

I can't find an excuse to drop the second syllable in the sound changes; it'd be an irregular contraction if it happened. But contracting the last two to /ei/ seems quite reasonable. So Wænenutei or Wænenuətei seem like good possibilities to me. (I think I like the second best, mostly because it's more different from the source...)

(Edit: on the other hand, the fronting rule doesn't seem to be applied to loans from Faraghin-- which are presumably very early; so maybe Wanenu(ə)tei would be more likely after all.)