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

Basilius · Post by **Basilius** » Sat Jun 13, 2009 10:00 am

1) I wikified a few paragraphs on Affalinnei, the state of the Affanons.

As usual, they must be full of flaws, inconsistencies and confused wordings - please don't hesitate to correct.

2) Also, I uploaded the two attested Zele texts with tentative Proto-Isles glosses.

There seem to occur a few curious constructions in the texts; thus, '(you aren't calling) for my benefit' is construed simply as (qúquq-piwtams) háy qí, i. e. "('not'-'call'-PRES.NON-SENS.) 'I' 'good' ".

Also, predicative adjectives look a bit bizarre, and there are examples of compound verbs looking like bare noun stem + ni.

Basilius · Post by **Basilius** » Sat Jun 13, 2009 10:43 am

Rory wrote:Maybe I'm misunderstanding, but I see both options as involving equal amounts of work - in both cases, you have to explain how X became Y.

Only, one of the options allows one to *choose* a suitable X, while the other does not look feasible at all, at first glance :)

Rory wrote:I also don't see the option of explaining how Mutsipsa' arose from "Ran's PI" as being particularly complicated. I regard much of language change as arising from probabilistic linguistic variation, partly due to social and geographic stratification.

To me, this sounds like answering "By natural selection!" to a question like "How did it happen that people have exactly five fingers on each hand?" :)

Imagine that a variety of English starts to use a construction like "horse cart and" (with "and" etymologically identical to the usual "and"). Honestly, I don't see how to explain this without speakers' conspiracy ("Let's speak thusly to mock those linguistic nerds") and can't figure out what the intermediary stages might look like. And with PI it's a bit more difficult because of the interwoven morphology (declinable conjunctions).

Rory · Post by **Rory** » Tue Jun 16, 2009 11:06 am

Basilius wrote:
Rory wrote:Maybe I'm misunderstanding, but I see both options as involving equal amounts of work - in both cases, you have to explain how X became Y.
Only, one of the options allows one to *choose* a suitable X, while the other does not look feasible at all, at first glance

But in both cases we need to explain the relatedness of "Ran's PI" and Mutsipsa'.

Imagine that a variety of English starts to use a construction like "horse cart and" (with "and" etymologically identical to the usual "and"). Honestly, I don't see how to explain this without speakers' conspiracy ("Let's speak thusly to mock those linguistic nerds") and can't figure out what the intermediary stages might look like. And with PI it's a bit more difficult because of the interwoven morphology (declinable conjunctions).

But it's not a case of a group of people suddenly switching from 100% "horse and cart" to 100% "horse cart and" - it's a gradual process, involving levels of free variation. The inflections on Mutsipsa' adjectives arose as general word order flexibility rose; this increase in flexibility allows constructions like "horse cart and" just as well as "horse and cart"; then, for whatever social reason, "horse cart and" won the competition and became the standard construction.

Thanks for raising these issues and questions, I wouldn't have thought about them otherwise

Corumayas · Post by **Corumayas** » Tue Jun 16, 2009 1:27 pm

Ok, so I did get busy with other things, but I expect to finish running Legion's Neo-Agaf changes in the next day or so. Once those results are wikified I'll probably take a break from that project.

Legion: Are intervocallic Cj clusters produced by this rule:

you wrote: > [j w] / _V

affected by this rule:

you wrote:C > [+long] / {V[+long],[ɹ]}_V

-- for example, would [ɨː.ʒjɨ] become [ɨʒ.ʒjɨ] > [ɨ.ɟjɨ], or stay as [ɨ.ʒjɨ]?

EDIT: Also, should [V.ɹjV] arising from the first rule become [Vj.jV]?

EDIT AGAIN: As long as you haven't replied yet, I might as well ask one more. What exactly is happening in this rule:

[+voc +long] > [+stress] / _#

-- does the pitch contour of the word change to move the pitch accent to the newly stressed syllable? Or does the pitch contour stay the same, and the new stress is realized in some other way?

EDIT 3: And does [a] act as a front vowel?
Hell,

Legion · Post by **Legion** » Thu Jun 18, 2009 10:35 pm

Basilius > I'd rather let you wikify anything about Thyokynehota. Let's just say that the language was extinct as a native language by -1000 YP, but was subsequently used as a lingua franca untill -500 YP, being gradually replaced by Naxuutayi.

Corumayas >

Are intervocallic Cj clusters produced by this rule:
> [j w] / _V

affected by this rule:
C > [+long] / {V[+long],[ɹ]}_V

-- for example, would [ɨː.ʒjɨ] become [ɨʒ.ʒjɨ] > [ɨ.ɟjɨ], or stay as [ɨ.ʒjɨ]?

No. /j/ doesn't count as a vowel in clusters.

does the pitch contour of the word change to move the pitch accent to the newly stressed syllable? Or does the pitch contour stay the same, and the new stress is realized in some other way?

No, the pitch countour isn't affected by the stress positional change (this was done precisely in order to make pitch countour more phonemic). The new stress is realised in terms of intensity and duration (but is not very distinct nor strong).

And does [a] act as a front vowel?

No.

Corumayas · Post by **Corumayas** » Fri Jun 19, 2009 2:43 am

Cool, thanks very much Legion!

(Please disregard that "Hell" at the bottom of my last post, btw... it's a remnant of something else I copy-pasted and missed deleting, not a response to your sound changes!)

Another question though (sorry), about this rule:

V[-low -stressed] > 0 / VC0_C0V

I was going to ask if the rule was iterative, but then I noticed it's not C_C but C0_C0. And this C0 appears in a couple other rules too. What's it mean?

Post by **zompist** » Fri Jun 19, 2009 4:39 am

Basilius wrote:Imagine that a variety of English starts to use a construction like "horse cart and" (with "and" etymologically identical to the usual "and"). Honestly, I don't see how to explain this without speakers' conspiracy

How does any syntactic change start? Why did we switch from "I think not" to "I don't think"? Why did Latin change from SOV to SVO?

But if you want a route, perhaps "horse and cart" goes to "horse and cart and", and later on the first "and" is deleted.

Legion · Post by **Legion** » Fri Jun 19, 2009 5:30 am

Corumayas > subcript and superscript numbers after a symbol have the respective meaning of "a minimum of X" and "a maximum of X".

The the C[subscript]0 means "a minimum of 0 consonant". The rule simply means that interior non low unstressed vowel get deleted, regardless of the number of consonants surrounding them (including if there are none). Given the intensity of the rule, however, it should probably be tone downed, by also posing a maximum of consonants around which this can happen, like, 2 before and 1 after (so it could only happen in open syllables, preceded by a maximum of 2 consonants). In these conditions, recursivity shouldn't be too much of a problem.

Corumayas · Post by **Corumayas** » Fri Jun 19, 2009 12:51 pm

Ah, ok. Where do I start deleting vowels-- front or back of the word? For instance, in a [VCVCVCV] word, should it become [VCCVCV] > [VCCCV], or [VCVCCV] (and stay there)?

Also where should a deleted vowel's tone go-- the previous syllable maybe?

pentalarc · Post by **pentalarc** » Fri Jun 19, 2009 12:59 pm

Okay, awfully newbie question, but this wasn't here last time I was here.

Is this where we are supposed to post updates to our conlang, rather than in the thread where we introduced our conlang?

(Or, more accurately, what exactly is this thread for?)

Pentalarc

Corumayas · Post by **Corumayas** » Fri Jun 19, 2009 1:11 pm

This thread is for a specific conworld that a bunch of us are working on together. New participants are always welcome. If you want to know more, the FAQ on our wiki is probably a good place to start.

Legion · Post by **Legion** » Sat Jun 20, 2009 12:07 am

Corumayas wrote:Ah, ok. Where do I start deleting vowels-- front or back of the word? For instance, in a [VCVCVCV] word, should it become [VCCVCV] > [VCCCV], or [VCVCCV] (and stay there)?

Also where should a deleted vowel's tone go-- the previous syllable maybe?

Deletion of vowels should start from the left (thus in this case you'd get [VCCCV].

The tone will simply be deleted, cause here it's not just a resyllabification, but a loss of syllable.

Basilius · Post by **Basilius** » Sat Jun 20, 2009 5:37 pm

I am sorry for being so slow with my replies...

Legion wrote:Basilius > I'd rather let you wikify anything about Thyokynehota. Let's just say that the language was extinct as a native language by -1000 YP, but was subsequently used as a lingua franca untill -500 YP, being gradually replaced by Naxuutayi.

Extinct by -1000 YP? This would mean that the latest version of the language was spoken only about 500 years after the start of the Isles migrations... 500 years being about the distance between today's French and the language of Rabelais... That is, enough time for a language to become different, not enough to become a different language :)

Any ideas?

zompist wrote:How does any syntactic change start? Why did we switch from "I think not" to "I don't think"?

The question I'd like to know how to answer is not "why?" (which hardly can be answered in principle). It was about intermediary stages.

The construction "I do not think" was availble and grammatical for centures before it became the default with "indefinite" tenses. Its positive counterpart is still available, and I wouldn't be too exacting about a "future English" using I do think as the default instead of I think (actually, I seem to recall reading somewhere about English dialects with this feature).

zompist also wrote:Why did Latin change from SOV to SVO?

There was no such development. It's just a good illustration of the trap with the notion of "default WO" which tends to be perceived as implying a more rigid WO than actually observed. Latin already had SVO as an option (not even especially emphatic).

So, what happened in both of your examples was marginalization of the option that used to be the default. But there is no difficulty in pointing to the source of the new default option - which type of difficulty is the main point behind each of my "issues".

and on top of all that zompist wrote:But if you want a route, perhaps "horse and cart" goes to "horse and cart and", and later on the first "and" is deleted.

So you do think this would be a natural-looking scenario with English. People starting to use "horse and cart and". Teens-and grannies-and, amidst casual-and natural-and conversation. You do. That's really puzzling :)

Rory, Zompist >

OK, I imagined myself trying to substantiate my point... all those ramified chains of arguments and counter-arguments... yet probably perceived primarily as attempts to explain how artificial and doubtful the features of other people's projects and the associated scenarios are... No, that'll look too depressing :)

My actual point was that it would be fun e. g. to find class agreement in "Ultimate PI" etc. And to discover some detail about Isles migrations in passing.

Post by **zompist** » Sat Jun 20, 2009 6:01 pm

My point was that any syntactic change sounds odd and arbitrary if you look at the endpoints in isolation. Your reference to "teens and grannies" is just yokelish. Is there supposed to be some linguistic reason why teenagers can't accept a syntactic change? If they do, would you be astonished that some of them grow up to be grannies?

My examples were supposed to make you think, not to make you kneejerk. Patterns for very basic syntactic operations change dramatically, for all kinds of reasons. Look at some actual historical examples— I've lost any interest in doing your work for you.

(BTW, look more carefully at "I do not think"; it's not an "option", it was originally precisely the old pattern: verb + not. I'd say the change is a combination of do-support becoming mandatory, plus reanalysis. Similarly, of course Latin was not exclusively SOV, but it's just crankishness to pretend that no syntactic change has taken place.)

Basilius · Post by **Basilius** » Sat Jun 20, 2009 6:39 pm

Zompist >

I am terribly sorry for having irritated you. Also, thank you for pointing to my language in this discussion being inadequate. I'll know for the future that jokose wordings and emoticons can do the opposite of what they were supposed to do.

On the other hand, I am terribly happy that my non-native perception of English frees me from perceiving immediately all the obertones behind expressions like "yokelish" or "kneejerk". This leaves some space for hope, sort of.

I believe that I can adduce enough historical examples myself, and I assure you that I did think of yours.

Please take my apologies seriously - I do feel wrong about how it's turned.

Legion · Post by **Legion** » Sun Jun 21, 2009 12:45 am

Basilius wrote:I am sorry for being so slow with my replies...

Legion wrote:Basilius > I'd rather let you wikify anything about Thyokynehota. Let's just say that the language was extinct as a native language by -1000 YP, but was subsequently used as a lingua franca untill -500 YP, being gradually replaced by Naxuutayi.
Extinct by -1000 YP? This would mean that the latest version of the language was spoken only about 500 years after the start of the Isles migrations... 500 years being about the distance between today's French and the language of Rabelais... That is, enough time for a language to become different, not enough to become a different language

Any ideas?

Uuuh… the wiki says PI was spoken by -2000 YP, I had assumed the divergences started before the migrations, and, according to what you say about reconstruction, even before PI itself…

Basilius · Post by **Basilius** » Sun Jun 21, 2009 3:51 pm

Legion wrote:Uuuh… the wiki says PI was spoken by -2000 YP, I had assumed the divergences started before the migrations, and, according to what you say about reconstruction, even before PI itself…

Yes, there seem to be discrepancies... maybe, the datings stem from some older discussions that I missed.

In principle, the dates could be revisioned - after all, they are supposed to be coherent with the rest, including the history of Sumarušuxi.

The main problem is the migrations of the Peninsular peoples. They settled down in the Lotoka area, in all probability having migrated there via Sumarušuxi. This probably means that their migrations must pre-date those of the Isles speakers. But this map, marked "valid", shows that by -1900 there were no Peninsular speakers on either Sumarušuxi or the Lotoka coast.

I don't know how to put these fragments together.

Corumayas · Post by **Corumayas** » Sun Jun 21, 2009 6:14 pm

I'm not sure what you see as discrepencies. As far as I can tell, the timeline Legion is suggesting is approximately like this:

ca -2000 to -1500: Proto-Isles dialects begin diverging in Tuysáfa
ca -1500 to -1000: Isles speakers migrate west to Sumarušuxi etc.
ca -1000 to -500: Thokyunehota is spoken as a lingua franca in Sumarušuxi
ca -500 to +100: Naxuutayi replaces Thok. as the dominant language of Sumarušuxi

The first part of this is nothing new.

(I'm not sure about having Thok. become a dead language by -1000; that does seem too early to me. But that doesn't look like the most important part of Legion's proposal, since he still has it as a lingua franca until -500. I'd rather simply say it became extinct, both as a native language and as a lingua franca, around -500.)

I don't see a conflict with the Peninsular migrations either; on the one hand, they could reach Lotoka via Dagæm and Affalinnei instead (in fact this route requires shorter sea crossings); on the other, there's enough time between -1900 and -1400 or so for them to get there before the Isles speakers show up, if we decide that's really necessary.

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?

Basilius · Post by **Basilius** » Sun Jun 21, 2009 6:48 pm

OK. Everybody is alive and writing comments in acceptable language. So.

Zompist >

It was indeed very stupid of me to respond in a mocking manner to your proposed solution of the "horse cart-and" problem. Please do consider my apologies - offending you wasn't what I had on my mind.

"Teens and grannies" were my clumsy attempt to address your native intuition about English: I hoped you'd admit that "I do think" (as a future-English equivalent of the development "think not" -> "don't think") is different from "horse-and cart-and". That is, that the latter does indeed require some unusual setting to emerge and spread. And I am still interested to hear your opinion on that.

After a pause, I must admit that "kneejerking" (especially if inderstood as "an act of reflectory kicking") was not too bad a description of my response to your Latin example, in particular. As I tried to explain in a pm, it's a holywar issue for me, so it's indeed a problem: it can blinden me as for the ways I can hurt people. I don't think it would be prudent to enter the theoretical detail in this thread: it doesn't look like the right place for holiwars and flame, and I do have strong opinions.

That said, I believe the "horse cart-and" problem remains unsolved, and your examples are in fact very helpful in formulating why. While you're right that syntactic changes may look odd at first glance, a closer examination seems to show that the visible oddity is associated solely with the marginalization of a once prevalent syntactic pattern; on the other hand, the new prevalent pattern usually has an easily identifiable source in the grammatical patterns available on the ancestral stage.

With "horse cart-and", we are lacking such a source in "Ran's Proto-Isles", and I don't see - perhaps, it's just me - any easy and natural-looking way to obtain such a source via some moderately complex multistep scenario. Which is why I still believe that treating it as reflecting an archaic trait of "Ultimate PI" might be an easier solution.

* * *

Corumayas wrote: I'd rather simply say it became extinct, both as a native language and as a lingua franca, around -500.

This looks like the right solution. Legion?

Corumayas wrote: I don't see a conflict with the Peninsular migrations either; on the one hand, they could reach Lotoka via Dagæm and Affalinnei instead (in fact this route requires shorter sea crossings); on the other, there's enough time between -1900 and -1400 or so for them to get there before the Isles speakers show up, if we decide that's really necessary.

So basically we need to chose between the two options - their implications are a bit different :)

Corumayas wrote: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?

To me, these look more like either areal or independent developments (I'll try to add some explanations later).

Post by **zompist** » Sun Jun 21, 2009 7:39 pm

Basilius wrote:That said, I believe the "horse cart-and" problem remains unsolved, and your examples are in fact very helpful in formulating why. While you're right that syntactic changes may look odd at first glance, a closer examination seems to show that the visible oddity is associated solely with the marginalization of a once prevalent syntactic pattern; on the other hand, the new prevalent pattern usually has an easily identifiable source in the grammatical patterns available on the ancestral stage.

Well, sometimes. Certainly there can be variation in the parent and just one of the variations survives. But that's not the only possibility.

Another is reanalysis. The "I do not think" example superficially matches the old rule (verb + not), but that rule isn't part of the spoken language any more.

Or, borrowing. The formula "X conjunction Y" is quite alien to Quechua, but it's been borrowed from Spanish.

A strange reordering might be the endpoints of a complicated progression-- e.g. development of the negative in French: non volo

je ne veux

je ne veux pas

je veux pas.

Or consider the slang negation "Not", as in "It's a beautiful world. Not." Imagine this becoming the standard-- it would seem quite inexplicable!

And there are other possibilities. I don't think we can say it's impossible to go from "X and Y" to "X Y and". It'd be more elegant if the "and" morpheme wasn't identical (as in the French negation example). Or of course if there were other similar changes in the language so that "-and" isn't the only example of cliticization.

Radius Solis · Post by **Radius Solis** » Sun Jun 21, 2009 7:58 pm

Basilius - I think everything will be okay. You sometimes come across a bit strongly, so maybe it would be a good idea to watch out for that when you're posting about something that matters a lot to you. In return, the rest of us should keep in mind that your non-native command of English can cause more issues than just vocab usage, but also understandable errors in judging tone (how people come across to each other).

For example, in your last post, you described previous comments as having "acceptable language" - I'd call that a minor tone mistake, since you clearly wanted to sound conciliatory. I'm guessing you probably didn't know that "acceptable language" sorta implies you are judging everyone's language to make sure it's good enough for you - as though we're kids and you're a schoolteacher threatening to punish anyone who acts out of line.

I'm sure nobody took offense, because it's obvious you didn't intend to sound that way. I only mention it as an example.

Mbwa · Post by **Mbwa** » Sun Jun 21, 2009 8:11 pm

Wow. I must say that this project is very impressive. I've sorta kinda looked at it before and it is very cool, but I didn't know new members could join. This may be something I'd like to be a part of at some point, if nobody minds.

I'll start reading up on the conhistory. One thing I'd really like to do, if the situation fits, is design a sort of trade pidgin out of two languages. That would be awesome.

Radius Solis · Post by **Radius Solis** » Sun Jun 21, 2009 8:16 pm

zompist wrote: Or consider the slang negation "Not", as in "It's a beautiful world. Not." Imagine this becoming the standard-- it would seem quite inexplicable!

And there are other possibilities. I don't think we can say it's impossible to go from "X and Y" to "X Y and". It'd be more elegant if the "and" morpheme wasn't identical (as in the French negation example). Or of course if there were other similar changes in the language so that "-and" isn't the only example of cliticization.

The slang example could work in a case like this too. If, for a humorous example, the original pattern was "X and Y" for two but "X, Y and Z" for three (just like English), but then slang arose where you always list a third thing after any two NPs. Like the kitchen sink. So at first, the usual pattern stays "the husband and wife" while the slang pattern distorts this to "the husband, the wife, and the kitchen sink", perhaps to indicate an exhaustive list. Later on the sink could drop out, since the original joke has gotten stale, and you're left with "the husband, the wife, and..." as your way to form any exhaustive list... and after that it's only a short step for the exhaustive formula to replace the unmarked formula.

Legion · Post by **Legion** » Mon Jun 22, 2009 1:15 am

Corumayas wrote: 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.

Basilius wrote:
Corumayas wrote: I'd rather simply say it became extinct, both as a native language and as a lingua franca, around -500.
This looks like the right solution. Legion?

Let's go with that.

Cedh · Post by **Cedh** » Mon Jun 22, 2009 1:58 am

Basilius wrote:
Corumayas wrote: I don't see a conflict with the Peninsular migrations either; on the one hand, they could reach Lotoka via Dagæm and Affalinnei instead (in fact this route requires shorter sea crossings); on the other, there's enough time between -1900 and -1400 or so for them to get there before the Isles speakers show up, if we decide that's really necessary.
So basically we need to chose between the two options - their implications are a bit different

The most likely route for the Lotoka people is via Fmana-hŋ-Talam, where we also have a Peninsular language later on. I don't know whether they'll pass through Dagæm as well; geographically this seems likely, but they don't need to if it doesn't fit in with Legion's plans for Komeyech. For the step onto the Siixtaguna subcontinent, I'd imagine that the Lotokans first settle in Affalinnei and on the coasts of the bay/fjord east of that, bypassing Sumarušuxi. The Affanons would then migrate to the mainland (replacing the Lotokans in the western part of their lands) maybe two or three centuries later, bypassing the eastern half of the Lotoka sphere. As for timing, I'd suggest that all this happens sometime between -1200 and -800.