zompist bboard

In another thread, dhokarena56 wrote:OK, here's the evidence I have for relating OS and Kebreni. Admittedly I don't have time to do the comparative method right now, so this will smack of Uralo-Altaic rather than Indo-European; it's based on typological similarities.

What looks like by far the most likely connection to me is Wede:i / Skourene. Let's start with the first five numerals:
Aside from "three", which may be non-cognate, these are very suggestively similar. Let's move on to the pronominal affixes now (and note that both Skourene and Wede:i lack real full personal pronouns--the Skourene ones are transparently derived from verbs+pronominal morphology). Removing the suffixes that evidently indicate gender in Skourene (and ignoring the 3rd person forms, which don't look cognate), we're left with:
Just in comparing these two sets of possible cognates, from very stable parts of the vocabulary, we can begin to see a few possible sound correspondences, although obviously nothing conclusive yet without examining a lot more. But, for instance, we have OS g : Wd k in "two" and the 2sg affix; and we have OS d : Wd t in "four" and the 1pl affix.

Now I'm going to try using the Swadesh list on these puppies.

Interesting, and plausible! But the morphology is going to be hell to reconstruct.

Yeahhhhh...I might leave that part for someone else!

Anyway, Swadesh list is done, but it lacks some terms that I should check and see if they're cognate too (like directions and more colors and stuff). A question mark means I can't find a native term with that definition in the lexicon; a double-dash indicates that the meaning in question is expressed by a derivation from some other root with a different meaning, and so is unlikely to be cognate. In one case I have "[NIL]", because I found evidence for the morpheme in compounds and stuff, but it's not listed in the lexicon. Also this table may end up looking like shit because there's no way I'm going through and making sure everything lines up.

Alright, enough of that. Let's get this party started!
Now, on to what all this means. We can remove all the ones for which one or both of the languages lacks a word corresponding to that definition, as well as ones where one or both of the languages uses a derivation from some other root to express that meaning. Ignoring terms that show no evidence of being related, and focusing instead on those which look promising, I think we're left with the following:

(1) "who/what" = OS -beg- / Wd -bogu-
(2) "other" = OS -gog / Wd jok
(3) "long" = OS -dim / Wd śim
(4) ??"small" = OS -its / Wd śiya
(5) "woman/nurture" = OS maşḍ / Wd ma:r
(6) "fish" = OS ţag / Wd sak(ana)
(7) "forest" = OS geŋ / Wd jen
(8) "hair" = OS bol- / Wd po:
(9) "head" = OS ?um- / Wd yuma
(10) "nose/smell" = OS sig- / Wd źuko
(11) "foot" = OS gan / Wd keŋ
(12) ?"eat" = OS ŋels- / Wd go:źi
(13) "bite" = OS gim- / Wd kimu
(14) "laugh" = OS sand- / Wd źe:ti
(15) ??"see" = OS ḍişn- / Wd liu
(16) ?"die/kill" = OS keşg- / Wd koku
(17) ?"hunt" = OS nesd- / Wd ŋo:du
(18) ?"hit" = OS serḍ- / Wd śoŋu
(19) "lie down" = OS ŋeḍ- / Wd neru
(20) ?"sit" = OS barḍ- / Wd waŋu
(21) "turn" = OS miḍ- / Wd muri
(22) ?"fall" = OS targ- / Wd źeku
(23) "count" = OS merg- / Wd boyok-
(24) "say/speak" = OS neln- / Wd yonu
(25) ?"sing" = OS riŋ- / Wd ŋinu
(26) "star" = OS maḍaŋ / Wd ma:in
(27) ??"water" = OS les- / Wd lu:
(28) "river" = OS ḍel / Wd ye
(29) "lake" = OS dit / Wd śi
(30) ??"wind" = OS ḍegl- / Wd do:
(31) ?"burn/fire" = OS rel- / Wd ŋol-
(32) "night" = OS selg- / Wd seki
(33) ?"year" = OS past- / Wd wata
(34) "rotten/bad" = OS ḍer- / Wd yur

Some of these are pretty tenuous, but others look very convincing. There's also a number of regular correspondences here. For instance OS ḍ- regularly corresponds to Wd y-, as in sets 28 and 34, as well as the number "two" from the previous post. On the other hand, word-internally (and sometimes word-finally?) the correspondence is OS -ḍ- to Wd -r-: see sets 19, 21, and possibly 5; however, 26 would seem to not follow this rule. OS voiced stops seem to mostly corresponded to Wd voiceless stops: see sets 2, 6, 8, 10, 11, 13, 14, ?16, ?22, ?23, and 32, as well as my previous post. Also notable is that in many cases consonant clusters in OS correspond to long vowel+consonant sequences in Wd (presumably a case of compensatory lengthening): see sets 5, ?12, 14, ?17, and ??30; there seem to be several exceptions here, though. And these are just the correspondences that immediately jump out at me! I'm sure there's more, but I've been working on this for a while now so I'm not going to bother with that immediately.

EDIT: A few more correspondences. OS ţ : Wd s, for which see 6 as well as the numeral "four" (in my previous post). OS m- seems to correspond to Wd b- in a few cases (set ?23 and the numeral "one") but not in others (sets 5, 21, and 26). Wd also apparently palatalized some instances of **g (or **k?), since there are a few cases where OS g corresponds to Wd j: see sets 2 (the first OS "g") and 7. More often, though, it corresponds to Wd k, as in the numeral "two", the 2sg pronominal affix, and sets 2 (the second OS "g"), 6, 10, 11, 13, ?16, ?22, ?23, and 32. In set 1 it appears to be g in Wd as well.

[EDIT 5/29/2012: At least two of these sets, based on the given etymology of the Wede:i forms, can't be cognate, so I've eliminated them]

By the way Zompist I better get, like, mad extra bonus points for this.

Yes, you've got it, congrats. I'm happy the mystery has lasted for five years, though now I guess I'll have to make something more subtle yet...

There's actually about 150 cognates; a large fraction have semantic changes as well.

zompist wrote:Yes, you've got it, congrats. I'm happy the mystery has lasted for five years, though now I guess I'll have to make something more subtle yet...

There's actually about 150 cognates; a large fraction have semantic changes as well.

Did you actually derive the language from a proto-lang you built?

Also: excellent work Whimemsz! What do you think we should call this family?

Atom wrote:

zompist wrote:Yes, you've got it, congrats. I'm happy the mystery has lasted for five years, though now I guess I'll have to make something more subtle yet...

There's actually about 150 cognates; a large fraction have semantic changes as well.

Did you actually derive the language from a proto-lang you built?

Also: excellent work Whimemsz! What do you think we should call this family?

Taësic, from "ours", in the same vein as Nostratic.

It has a name, but as a bit of recognition for Whim, if he can suggest a reasonable name I'll use that.

(I'm not sure about putting it on Almeopedia; I kind of like having some things less obvious.)

Impressive work Whimemsz, congratulations!

That's some hard work there. Way to go, Whims!

There should be a special ZBB Award for unlocking Almean mysteries.

Yay! Congratulations!

Now I have no excuse but to start reconstructing Proto-Western...

I shall try to figure out some morphology tonight. However, the typology of the two is so fricking different, it'll take a while.

zompist wrote:It has a name, but as a bit of recognition for Whim, if he can suggest a reasonable name I'll use that.

It's actually much harder than it sounds to come up with a good name. At first I thought about going the "Penutian" route, but that would end up as something like ḍogyokian or Yokḍogian, which I think we can all agree are pretty hideous. How about a cash prize in lieu of that honor?

Bomotian, however, isn't too bad, or Bomotic. Yonunelnic?

Yokḍogian isn't that bad.

No, it's pretty terrible, actually.

"Bomotic" isn't half bad. It looks a lot like Omotic, though. And Omotian sounds a bit sillier to me.


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

A few more likely or possible cognate sets:

(35) "south" = OS ţar / Wd saŋ
(36) ?"north" = OS ḍan / Wd lun
(37) "east" = OS kom / Wd ju
(38) "west" = OS boḍ / Wd wor
(39) "brother" = OS ḍed- / Wd roda
(40) "sister" = OS meld- / Wd beda
(41) "hemp" = OS naşgu / Wd naka
(42) "rule/inherit" = OS baţş- / Wd paźiwa "king"
(43) "branch/arm" = OS daļg- / Wd te:
(44) "deer" = OS mulg- / Wd buka
(45) "beard" = OS ŋarḍ / Wd ge:ŋ
(46) ??"inside" = OS ḍadŋ- / Wd ra
(47) ?"cat" = OS kuļiŋ "mountain lion" / Wd jo:na
(48) "here/this" = OS luļ / Wd lil
(49) "before" = OS mur- / Wd bu
(50) "upward/above" = OS dem- / Wd śen
(51) "below/under" = OS men- / Wd ben
(52) "middle" = OS saln- / Wd sai

A number of these display the same correspondences as I laid out in my earlier posts. For instance, the ubiquitous OS vd.stop : Wd vls.stop, in sets 41, 42, 43, 44, and possibly ?50 (if Wd ś reflects palatalization of an earlier *t). OS ţ : Wd s is illustrated in 35. The OS m- : Wd b- correspondence I mentioned in my last post as a possible one seems to be exemplified in these sets several times: see 40, 44, 49, and 51. The Wd correspondences to OS ḍ seem even more complex now. There are no examples here of OS ḍ- : Wd y-, but there are several cases where initial OS ḍ- corresponds to Wd r- (the same correspondence I took to be regular when word-medial and possibly -final; see also set 38): sets 39 and ??46; possibly set ?36 is to be included here as well, with Wd l a later development? (this would then perhaps be parallel to set ??15) It also looks like there is a regular (though odd) correspondence of OS rḍ and Wd ŋ, exemplified here by set 45, but in my previous post as well, in sets ?18 and ?20 [the question marks can probably be removed: these look like pretty secure cognates to me now].

I'll also add that it now looks like the numeral "three" is probably cognate between the two families as well. There are a number of sets in which OS d or t correspond to Wd ś or ź before a front vowel: 3, ?22, 29, and 50.


EDIT: I added two more cognate sets. Also I'll point out that the marker for Y/N questions in OS, -si may be related to either the Wd marker for negatives -ze or the Wd marker for Y/N questions, -źu. The OS prefix ŋok- "finish" may be related to the Wd suffix -ok "past tense", as well?

rotting ham wrote:There should be a special ZBB Award for unlocking Almean mysteries.

Maybe this can get the "Best Thread" award this year, then?







(It should)

(53) "each/every" = OS -doļ / Wd to:l (not sure how I missed this one...)

Okay, a few more possible connections between morphological affixes (see the edits to my previous post for a few others).

OS me- "pretense/doubt" is possibly connected to Wd -mi "reportative"
OS ru- "quickly/urgently" can probably be connected to Wd -ro "imperative"
OS ge- "case reset" is maaaaybe connected with Wd -ka "connective"?
OS nen- "comparative" may be connected to Wd -ne "contrastive"

Whimemsz wrote:

zompist wrote:It has a name, but as a bit of recognition for Whim, if he can suggest a reasonable name I'll use that.

It's actually much harder than it sounds to come up with a good name. At first I thought about going the "Penutian" route, but that would end up as something like ḍogyokian or Yokḍogian, which I think we can all agree are pretty hideous. How about a cash prize in lieu of that honor?

I think a character/nation on Almea is the highest honour available m'self.

PS. This is going to have extra interesting implications when Tzhuro is released - I wondered how much of a proto-Lenani language had been developed or whether Mark would be "working backwards" on that one - now I think we know the answer!

Whimemsz wrote:It also looks like there is a regular (though odd) correspondence of OS rḍ and Wd ŋ, exemplified here by set 45, but in my previous post as well, in sets ?18 and ?20 [the question marks can probably be removed: these look like pretty secure cognates to me now].

Per sets 31 and 35 this could be just r ~ ŋ actually.

MrKrov wrote:No, it's pretty terrible, actually.

I don't get what's so bad about it. It's just a name, it doesn't need to be pretty or anything.

Azulene wrote:

Whimemsz wrote:It also looks like there is a regular (though odd) correspondence of OS rḍ and Wd ŋ, exemplified here by set 45, but in my previous post as well, in sets ?18 and ?20 [the question marks can probably be removed: these look like pretty secure cognates to me now].

Per sets 31 and 35 this could be just r ~ ŋ actually.

Good point!

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

Alright, now that I have a better handle on some of the various correspondences, I can spot a few more cognates (or potential cognates) in the Swadesh list that I missed on the first pass:

(54) "big" = OS riḍ- / Wd ŋor
(55) "seed" = OS ḍad / Wd rada

Also, for ease of future reference, might as well assign the other words from various posts correspondence set numbers:

(56) "one" = OS moţ / Wd bo
(57) "two" = OS ḍog / Wd yok
(58) "three" = OS ded / Wd śir
(59) "four" = OS darţ / Wd tause
(60) "five" = OS bim / Wd pina
(61) "1sg" = OS -ŋ(o)- / Wd -no
(62) "1pl" = OS -dor- / Wd -tu
(63) "2sg" = OS -g(o)- / Wd -ku
(64) "2pl" = OS -beş- / Wd -we

I hesitate to include the possible connections of morphological affixes, since all involve semantic or pragmatic shifts, and all are very short and thus have a high probability of being similar due to chance. Dhokarena, what happened to your grammatical analysis? >:|


Also, this would be a lot friggin' easier if we knew stuff about Tžuro, Mei, Fei, etc.

Whimemsz wrote:Also, this would be a lot friggin' easier if we knew stuff about Tžuro, Mei, Fei, etc.

Well unless Zompist has been hiding something all these years...

Anyways, just a little something for me to help visualize. I recolored the language map to show only the highest level organization for language families (Eastern, Wedei-Littoral (or whatever you decide to call, and Ur-Kagot).