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

[Cedh]

Looks like a good start!

However, I find it a bit strange that some of the symbols selected for creating the alphabet represent verbs or adjectives rather than nouns. Of course the logographic Ngauro script would have contained glyphs for such concepts, but I suspect that the Ndak would use an all-noun list when choosing glyphs for their alphabet. Also, including both "woman" and "female" seems unlikely to me.

In addition I'd suggest that the script should contain separate glyphs for lone nasals on the one hand, and nasal+stop clusters on the other hand, because the latter are such a distinctive part of NT phonology (and do appear word-initially). This would also enable Adāta and Ndok Aisô to have separate glyphs for their aspirate series because these developed mostly from such clusters.

[Radius Solis]

Hmm. The glyphs themselves are definitely a keeper, and thank you! As for the alphabet system I'm not sure, because at some point a couple years back we decided that it wasn't a good idea for the Ndak to have had a true alphabet - we were going to leave that to the Fáralo to invent. My original script plan for NT was for it to have a hurky system cobbled together out of both Ngauro logographs and native runes. I came up with the runes as far back as 2005 but the rest stalled for lack of inspiration on the Ngauro side, and exactly how to combine the two systems for Ndak Ta.

Got any ideas for what to do?

Another minor issue is that it seems likely to me, if the Ndak did have an alphabet, that they would perceive the presence of an actual nasal coda consonant and not quite recognize the true nature of the nasal vowels. So the nasal-diacritic might be more realistic to write full-size as a separate character.

But please don't take this post as harsh criticism. I'm really glad to see you're working on the script!

[Legion]

Cedh > well, the selection of symbols is not really supposed to make sense – the way I see it, the Ndak first try to adopt the Ngauro script as it was, but quickly found it impractical and narrowed down the number of symbols, keeping the most common ones and starting to use them for their sound, ultimatly leading to the canonical selection, which is really a random choice resulting of chance, and not an educted construction, so this explain why verbs and adjectives ("female" is the adjective) ended up as source for glyphs.

The real life justification is that it was sometimes difficult to find Ndak Ta words starting with the desired sound and being common words susceptible to end up in a selection of symbols associated with common words – sometimes the nouns failed me so I had to ressort to adjectives and verbs; but if you have suggestions reguarding this, I'm all ear!.

I don't know about having glyph for nasal + stop clusters, maybe the daughter scripts could just make them by turning the nasal into diacritics (also note that Adata is supposed to use an adaptation of the Tjakori script, unrelated to this one).


Rad > I did some try with none alphabetic system, but I didn't find it very practical nor pretty; and abjad seems unlikely with all those vowel sound to note, while an alpha syllabary makes more sense, but if they have symbol for vowels word initially, why not using them elsewhere as well (if we *really* want an abugida though, we can just use the vowel symbols as diacritics when they follow a consonant).

For writing the nasal vowel as simple vowel + ng, I'm not against that idea, though I guess, as the nasal vowel contrast with oral vowel + nasal consonant, the nasalisation consonant would eventually end up as a diacritic, I guess (especially if we go for an abugida and represent the vowels as diacritics, then a diacritic for nasalisation, à la Devanagari, would be the logical outcome).


I'll play a bit more with the abudiga idea, then...


And I don't have a problem with criticism, we don't want a loosy script, do we ?

[Legion]

Okay, here's some try at an abugida/alpha-syllabary thing (it's Tsinakan text, obviously):



(I'm still experimenting a bit with some letter forms)

[Cedh]

sometimes the nouns failed me so I had to ressort to adjectives and verbs; but if you have suggestions reguarding this, I'm all ear!

What about the following:

- ntâug "forest" and ngkai "egg" for /nt Nk/ instead of /n N/, and mpaiswa "fish" for /mp/ (instead of mpepm "breast").
- mbontai "seed", ndai "tree", nggol "foot" for /mb nd Ng/ (if you want to use the existing "leg" glyph for the latter, baus "ox" would be a good alternative for /b/).
- mabm "mouth", nenga "leaf", and ngane "neck, anything narrow connecting two things" for /m n N/.
- saungwe "water" for /s/.
- wedn "herb, vegetable" or wampum "hollow" > "kettle" (this meaning is attested in Buruya Nzaysa) for /w/. Alternatively, use the same glyph for consonantal /w/ and vocalic /u/ (this works best in a true alphabet of course).

/mbw Nkw nts/ would then be the only consonants without a glyph based on a nominal concept. I don't know what to do with these currently.

For vowels, I would not distinguish nasality in the oldest version of the script, so we need only /i e a o u ai au/.

- itwam "arm" is a good choice for /i/; another good one would be imbi "moon".
- empi "grass" and okwa "ear" are also good.
- For /a/, you could either stay with asa "woman", choosing any of the three glyphs you currently have for "woman", "female", and "breast", or choose ali "fire". I also like the glyph for âka "air"; you could use the same glyph for /ai/ via aimwa "snake" though, or for /u/ via ulwo "rope", which is the only noun starting with /u/ in the entire lexicon.
- Another good word for /ai/ might be ailàu "flower".
- For /au/ I think I would go with aunti "river", but your glyph for âukwai "sculpture" is good too.


Your longer example above looks good, but a bit squiggly overall. Some letterforms should certainly be simplified further.

What kind of material did the Ndak write on? Stone, at least sometimes (e.g. the original Tsinakan stele). What else? Clay tablets? Parchment? Papyrus? This would all have a notable effect on the shape of the glyphs...

[Legion]

I like your vowel suggestions.

For the preconsonantic nasal however, I don't think it's a good idea to come up with a distinct glyph for each possible prenasalised consonant, as it will considerably augment the number of glyph while it's not really necessary:

In the above example I've used a system where prenasalisation is shown with a "ng" diacritic (small circle) on the left of the consonant. And I pushed the consistency of the system further, so that actually, any nasal that isn't before a vowel is treated as a secondary feature and rendered by a diacritic, just like nasality of vowels.

I don't know on what Ndak Ta was supposed to be writen though, I guess Radius knows better!

[Radius Solis]

My original intention was for the Ndak to have two "modes" of writing which would each require their own letterforms... 1. the inscription mode, which would be used for carving into stone and clay tablets, and would have to be fairly angular. And 2. the brushed mode, for easy daily use on papyrus/pounded bark/whatever, painted on with a brush like they did in China and with therefore fairly rounded letterforms. The inscriptions by which NT is attested would be in mode 1 but all the daughter scripts would come from mode 2.


Here are the runes I originally developed for Ndak Ta. The first column indicates the pre-Ndak source runes. The second column is Imperial Ndak mode 1 and the third column is mode 2. There are no vowels because I didn't want the runes to be an alphabet, or even a complete writing system - I wanted them supplemented with a lot of Ngauro glyphs. You can see that I arrived at a number of characters for clusters by ligatures of more basic forms:



(Don't worry, I will not be heartbroken if we do not use these runes for NT - I can always adapt them to another language. But I do like them, they're one of only two conscripts I've ever done that were really worth keeping.)


--------------

I do still have reservations about the Ndak using an alphabet. It took 4000+ years and a chain of adaptations between half a dozen languages before we arrived at a true alphabet on Earth from the original Cuneiform - and the Ndak are 1500-2000 years after the first invention of writing on Akana and only one adaptation away from it.

So if we really are going to do an alphabet, I am hoping we can make it a really inefficient one - like by having dozens of "unnecessary" characters that are used for clusters, or having multiple characters based on their position in a word, or even just having multiple characters for each sound that are chosen at the whim of the writer because the phonemic principle just hasn't occurred to them yet. Or even including one or two hundred Ngauro logograms for particular words, that get interspersed with letters throughout a text, like Kanji in Japanese. That's my favorite option, but it's probably too much work.

Any of these options would make a much greater wealth of source material for daughter-scripts to draw on, too. They could have all kinds of quirks arising from the clumsy NT system, and they wouldn't all have to have similar letterforms - derivers could pick and choose which forms to keep.

[Legion]

Rad > well, as it is, I'm really going for an abugida à la Devanagari (the large example I provided is already of this kind). And a few letters will have several possible forms, so that the daughter languages can mess with that.

I like your idea of having two distinct modes according to the kind of support, I'll work on that!


Your script is not bad either, we should use it for something (the brushed mode looks like Tengw-- er, like the Klingon script).



(let's bet on how many times we can say "script" before Serali pop ups in the thread)

[Legion]

Okay, here's my last try at the thing, I've used several of the newly suggested letter forms, and I've tried to simplify the shapes here and there.



Here's what it gives in use:

[Legion]

Yes, no, anyone?

[Radius Solis]

LOOKS LIEK TENGW-
O WAIT

More seriously, hopefully people will start chiming in soon. And it's definitely very attractive!

[Corumayas]

Wow! I've been busy with Christmas travel and stuff (won't be home until the middle of next week) and haven't had much time to spend online... but this is something I have to comment on.

I agree with Radius that an alphabet is not what we want. Nor, in my opinion, would an abugida be particularly likely. Abjads, alphabets, and abugidas are, I think, entirely a product of the uniquely quirky way the Egyptian script works-- the original Western Semitic abjad being inspired by the Egyptian one, while (as I understand it) both the Indic abugidas and the Greek alphabet developed from its descendent abjads, and each of those types of script was really only invented once. Most independently invented scripts seem to have been logosyllabic, and most of their descendants are either the same, or simplified to pure syllabaries.

Personally I'd love to have a complex, messy logosyllabic script, like Akkadian cuneiform or Kanji. What I was hoping for was for someone to make a logosyllabic script for Ngauro once I had at least a basic phonology and lexicon for it, and then start the relay off by adapting that script to Ndak Ta. If the script was borrowed from the Ngauro, as Sumerian cuneiform was adopted to write Akkadian (and Chinese characters for Japanese), I think it'd make the most sense if the phonetic readings of the characters were based on Ngauro words.

However, I recognize that 1. designing something like that would obviously be a lot of work, and 2. my progress on EI and Ngauro has been very slow, and it'll probably be months before Ngauro is ready to have a script designed for it. Besides which, I like the way both Legion's script and Radius' one look (though they're rather different in style). So hopefully we can come up with a satisfactory way to use them both. Maybe Legion's characters could be part of the larger logosyllabic system...?

[Legion]

Well, I have oppositions to having a logographic script for Ndak Ta, but they are mostly of purely technical nature:

Designing a logographic script, for whatever language, will take a long time, and the result will be that once this script is ready, it is unlikely anybody will be willing to do anything but just cary it over as it is on other languages; I mean, I'm willing to make manually evolve, say, the shapes of 30~40 characters, by writing them over and over again, to get a new script. I'm not willing to do that with 100 times more characters.

By extension, this will seriously hinder any possibility of interesting evolution with the script. With a small number of symbols, it's easy to simplify them, to developp things like cursive, interesting ligatures, different moods and stuff. Of couse, all those things also exist with logograms, but developping them from our perspective is a herculean task I do not want to get into.



Now, to more general opposition, first, the way I saw the Ndak Ta developping a script was also paralel to a real world case: how the egyptian glyphs was strongly symplified to write semitic languages, and indeed, with sound values for letters based on the sound the aforementioned semitic languages had for the corresponding word (eg, the glyph for "palm (of the hand)" represented the sound /d/ in Egyptian, but was used to represent the sound /k/ in its adaptation, according to the semitic word "kapp").

Now, they invented the system as an abjad mostly because:
1)Egyptian's phonemic representation sub-system was already adjabic.
2)Semitic languages weren't bother much with such a system, what with the triconsonantism, not allowing vowel onset, and not having many vowels to begin with.


My theory is that the Ndak would have:

First, adopted the Ngauro script, using the glyphs for their meaning (probably still during the Proto-Talo-Edastean period). Then, they started simplifying the system, using glyphs for their sound value in Ndak Ta (because after all, Ndak Ta was what they were trying to write).

But now, I have good reasons to believe that, if we definetly exclude the possibility of a true alphabet, an abugida is the most likely way Ndak Ta would have been writen because:
-An adjab would be impractical, as Ndak Ta is quite vowel happy.
-But a syllabary would be not only impractical, but also a nightmare to design, because of Ndak Ta high number of possible different syllables, and because many of these are seldom attested in initial position, making the finding of a glyph difficult for said thing. And I mean, even if we find ways to work around and stuff, a syllabary for Ndak Ta would still require at the very least 107 symbols (and that's really the scrict minimal number, provided we use diacritics and other means for anything that is more complex than CVC (where V is only an oral monophthong!).

If we were actually to come up with a more strictly phonemic syllabary, we could quickly end up with more 300 glyphs, some of which would only serve rarely...

My Abugida system, on the other hand, conveniantly goes with 24 glyphs and a few diacritics derived from those glyphs (and all the forms are actually derived from hypothetical Ngauro glyph, with an original design clearly evocating of their meaning). With this system, evolution by other conlangers will be easy and quick to implement, interesting orthographic or calligraphic quirck will be implemented, and so on.



Also, after checking, it turns out Abugida were in fact invented several time independently. While the Brahmi script and its descendants are the most famous ones, Ethiopia gives us Ge'ez, attested since 5th century BC, and probably derived from Sabean (and abjad).

This is also true for alphabets: it seems that the Okhron and later Uyghur alphabets were also created out of another abjad, Sogdian, without any influence from western alphabets.

(And even cuneiforms were adopted into highly simplified form – the old Persian cuneiforms were mostly a syllabary, which only retained five logograms of very common words).


(ps: all this should be considered material for discussion and debate, not a strict refusal to change my ways)

[Radius Solis]

I think a good compromise might be if we were to have the Ndak first borrow Ngauro logograms and then adapt some into the above abugida, while at the same time preserving a hundred or two of the most common word-glyphs from Ngauro. (We need not wait for anything about Ngauro to make a script for it if it was purely logographic, nor do we need to waste a bunch of time creating thousands of glyphs for Ngauro.) I'm picturing Japanese Kanji surrounded by kana, but with an abugida system for the kana. Except not so many logograms as official Kanji has, not by a long way.

The abugida portion of the script would serve well as the basis for future scripts, but having the remaining logograms would 1. be more realistic and 2. give script derivers much more material to work with. This way, also, Legion's existing work won't need any drastic revisions, just the addition of the preserved logograms. And I can definitely help come up with logograms, if you want.

That said, I will note that whatever the Ndak system ends up being, it must be capable of writing the full language with inflections intact. Because that's how we get the Tsinakan text - and indeed, attestation of its grammatical system at all.



Another question we should ask ourselves: what languages will be adopting scripts from Ndak Ta's? Definitely not Adāta, which uses the Tjakori syllabary (and is far more suited to a syllabary than any other Edastean language). I think the answer is:

1. The Fáralo writing system, which will probably be the basis of much of the latter-day world's writing systems just as the Latin script is the basis for so many of ours. It should probably be alphabetic or defectively alphabetic. I can easily see the Faraghin trying to write their language with Ndak symbols and quickly dumping the logograms, as well as failing to get the hang of abugidas and dropping the vowel diacritics and "inherent vowels" in favor of fully expressed vowels every time.

2. Everything else. Other systems will no doubt arise from Ndak Ta, many of which may remain in use indefinitely (aside from evolving), but which never become more than locally dominant. Ndok Aisô may have such a script. Naidda definitely will. Komejech too, I assume. My intention for Naidda is for it to have a quite baroque writing system, pretty but nasty to learn, and that this will have been replaced by Fáralo script for all common uses and survive only in holy writings and the like.

[Legion]

Well, I like this compromise idea, as long as we do not have too many logograms for Ndak Ta, I think it can stay manageable. 100~200 glyphs, maybe we should so the swadesh list, numbers, and some termes dimed importants (like "king", "country", and names like "Kasadgad", "Tsinakan", etc).

I also like the idea of evolution for Faralo.

For Komejech, I'd plan to pretty much drop the logogram (mostly due to the nature of morphology, with productive initial consonant mutations, hard to represent otherwise), but keep the abugida system, going for a more cursive feel (with a rather irregular spelling – which means that when the island is conquered by Huyfara, the native script will come in competition with the imported system).

[Zhen Lin]

It looks very nice... on paper (or papyrus). I have my doubts about this being a monumental script though, but I suppose it just means that Ndak Ta builders will have to spend more time chiselling out all the curves. Either that, or they would have to paint it on...

[Legion]

Well, this is the running mood, designed for writing on papyrus and stuff – there will be a different mood for carving (probably even following suggesting Radius' suggestion that only the later survived as archeological evidence, but modern scripts will be derived from the former).

[Corumayas]

I can support the compromise too. One idea I had is that maybe, while early Ndak writing used a larger and more complex system, most or all of the descendent systems are based on a small subset of characters... so all we'd need to start the relay is that subset.


If the Ndak didn't use Ngauro as a prestige/literary language at all, it might be more likely that they'd devise a new script based on the Ngauro one but using Ndak Ta words for the phonetic values, rather than just adopting characters with their Ngauro phonetics. There could also have been some of both at first, but the Ndak-based system proved easier to use and won out.

I do think it'd be very cool if at least some of the logographic characters had both Ngauro and Ndak Ta readings (as happened with cuneiform and is still true of kanji). But if the Ngauro readings are only used in special situations, then we could just add the alternate readings to those characters later.


Also, based on Legion's research, it sounds like an abugida is a more plausible invention than I thought (especially if we keep the pre-Ndak consonantal runes). I'd say that any other abugidas and alphabets in Akana should be derived from or inspired by the Ndak script though. I just remembered that the Takuña script, and the Mûtsipsa' one adapted from it, was said to be an abugida, so probably that should be ultimately derived from the Ndak script as well.


So, in conclusion, yay, let's move forward with the script!



Edit: Although, for what it's worth, many syllabaries only have characters for (C)V syllables, which would reduce the number of characters required to a more reasonable amount.

Edit 2: Something else just occurred to me: if there are no phonetic readings shared between the Ngauro script and the Ndak one, it will be much harder for later scholars to ever decipher Ngauro. Also, in thinking about Ngauro's phonology I've been keeping an eye on Ndak Ta's with the idea that a Ngauro syllabary should be adaptable to NT. Maybe I should stop worrying about that.

[Legion]

Corumayas: actually, I counted – even with only glyphs for (C)V syllables, even with "V" standing for oral monophtongue, and everything else handled by diacritics or digraphs, a syllabary for Ndak Ta would still need at least: 18 (17 consonants + null onset) * 6 (5 monophthongs + null nucleus) - 1 (null onset + null nucleus > no glyoh needed) = 107 symbols.

Add the two diphthongs in the equation and we reach 143. Add the nasal vowels, and we're already with 233 symbols, and still representing only (C)V syllables!



On an unrelated note: do we have somewhere, in readable form, the sound changes from Ndak Ta to Fáralo? I'm trying to reverse-engineer them, but I must say, that's not easy!

[Corumayas]

Hmm. Only nasals can occur with null nucleus, so you don't actually need quite as many as 107... more like 18 consonants * 5 vowels + 3 sesquisyllabic nasals = 93. You could even get away with just one sesquisyllabic nasal since it's always next to a homorganic stop, which would give 91. (Not that that's that much less than 107, really.)


For what it's worth, a tentative phoneme inventory for late Ngauro is something like this:

/p t k b d g m n N f v s z l j/
/a e i o u/ plus some diphthongs, at least some of which are realized as monophthongs in closed syllables (e.g. /ai~E/). There may be a simple tone and/or vowel phonation system too.

In the course of its history, Ngauro has at least one chain shift, where /j w/ > /z v/, followed by /r/ > /j/. The diphthongs may move around a bit too, but the monophthongs pretty much have to stay put for the correspondences to work.


Ndak Ta to Fáralo changes were posted here:
http://www.spinnoff.com/zbb/viewtopic.p ... 882#566882

[Legion]

Oops, I meant not syllabic consonants, but any consonant not followed by a vowel (so coda consonants as well), which we obviously need some way to represent.

Thank you for the Fáralo sound changes!

[Legion]

So, Komejech grammar is still progressing, and to prove this, here's my new year's present: Tsinakan text in Komejech, enjoy!

(š ž ch c = /S Z x ts/, the rest is as in IPA)

Pi zef Zinega, ju lu djege teidi, ju lu djege ech lu las ech Kezakas, ju lu mechas ech Cof o Ibei:

"Ža feibi i ne lu jepelo ech lu meibi eje, adeipe zekou že eze lo ras fajetu. Pipei zef lo ras fajetu roumach adeš ach lo chimeš ejech: "Adeš o djege teidi lu meibi egai. Zofižape opa ras fajetu. O zetan ju o negai. Taf adei o depei lu žus roume fei ai zežou ne lu jepelo ech lu meibi egai."

Lere feibi i, ju lu mechas ech Cof o Ibei, ne lu jepelo ech lu meibi eje, ža non ža lo žas fajetos roumach adeš ach zekou že, non ža lo bjeti ech Pezei. Peizofon ach o ulan ža lu joumu teizile lu dou eje. Pin zef: "Nerei eje, ju lu ban ech lo bi! Irefen i so ras roumach adeti ach lo chimeš ejech o žoumach pebon sach ju o depei. Nete zibaman zo kejan lo charas ech lu las jeižu etou, Nerei eje! Žerei lo zunegane!"

Peižason Pezei so reizi za lu jam eje. Ulan i o zapen lu igai ža lu jesai eje. Zofižan si poka rou žjes kjech roumach irainabe ach ezi. Zofižan. Eižulan jasiš, pes o kjere, o pilen ža lu ras ech Kezakas."


Waiting for a full gloss, here's some general grammatical considerations:
-It's strongly verb initial.
-There's an elaborate system of initial consonantic mutations.
-There's an intricate system of cases (only three cases, but their use is complex)
-Personnal pronouns have strong, independant forms, and weak, cliticized forms.
-There's no syntactical pivot, so Komejech is quite free with droping out of the sentence already mentionned verbal arguments.
-There are inflected prepositions.

That is all for now.

[Corumayas]

This thread got me thinking about the history of cities and city size in Akana. Is the progression Salmoneus sketched there likely to be universal, or could it have happened differently?

If the former is true, then presumably the Ngauro civilization roughly corresponds to Uruk, while the next explosion in city size is probably associated with our Classical period (though it may have started earlier with Ngahexoldod in the first millenium BP, not to mention Xshalad or elsewhere). It seems like only the capitals of large states would exceed 100k in population (until the industrial revolution); certainly Ussor after about -200 is in this category, and maybe Athale at the height of its empire. I don't know if any Kascan cities would reach this size before the rise of Wi'makwa.

The role of trade/goods acquisition in the growth of cities and states is key, I think. The Ngauro cities grow, and their political influence expands, largely on the basis of trade bringing raw materials (chiefly stone, copper, and tin I think) to the lower Aiwa in exchange for high-quality manufactures produced by specialists. The need to secure the trade routes and sources is what leads them to colonize Lasomo. (This is apparently what happened in Mesopotamia during the Uruk period; the main exports of Uruk cities seem to have been fine wool cloth and maybe wine).

Anyone else have thoughts about this?

[zompist]

I don't agree with a few of Salmoneus's statements. Namely:

1. I think the idea of 'colonial cities' is wrong, at least if he's taking Greece as the model. An example: the state (polis) of Megara had about 24,000 inhabitants. Of these, no more than 3000 lived in Megara town (asty). The vast majority of Greeks were rural. Only imperial Athens was much of a city in the modern (or even medieval) sense.

2. Imperial Rome very likely had no more than 250,000 residents.

3. The Persians were not a very urban culture— Persepolis was barely a settlement, and even Babylon and Ctesiphon were tiny compared to imperial Rome.

But he's quite right to emphasize the smallness of ancient cities, and the possibility of going backwards, as in Dark Ages Europe.

[con quesa]

So, it's been ages since I've properly conlanged. I've been stuck with minute details of my own conlangs for a long time, and I've been too busy/lazy to try to work through them. I figure I need a fresh project to work on, and I'd like to start participating in Akana.

I created Zhaj (damn. About two years ago, back when I was applying for college), but I haven't kept up with what the project has become since then. So I'll have to spend some significant time reading the wiki and through this thread. I'd like to rework Zhaj, since I kind of half-assed it at the time to meet the deadline (in my defense I was doing UC applications at the same time), but it looks like people have created daughter languages from it, so that might not be cool. Bar that, I'll have to find another corner of the world to work in.

It looks (from my cursory reading of the wiki; I'll go through everything in more detail probably tomorrow) like most of the work thus far has happened on one continant in the northern hemisphere. If reworking Zhaj would be problematic, maybe I could work on something on another continant. Once I actually read through what is here so far, I'll have better ideas I'm sure; I'm mostly just posting this to commit myself to doing some sort of Akana-related stuff, and announce to everyone that I'm interested.