Ok, so first, I guess I'll summarize the current understanding that y'all have produced. I'll critique it "blind", I will only be giving tips as a person with no expertise in this language family (but expertise in diachronics) might see. You have to be careful of what qualifies as valid criticism of your model. I might not be giving "correct" criticism, but rather a red herring. It depends what you throw at me. (Also, it might seem rather harsh, depending on how you take it, so put up iron walls
(I think this will be extremely helpful; it's not a good idea to not read this post)
First, just recapping from the OP, and the Protolang post:
The phonology is
/m m̊ n n̊ ŋ ŋ̊/
/p t̪ t k b d̪ d g/
/s ś h j w ɰ ɾ/
/a ɛ e i u o ɤ/
If you got other stuff in your answer, I would highly suggest to correct that.
Secondly, the language is sesquisyllabic, meaning polysyllabic words have at least 1 minor syllable. Minor syllables can only contain /a i ∅/ [ə ɨ ∅].
Next is the groupings proposed by Creyeditor and Sangi39
Creyeditor's grouping goes something like this
Code: Select all
| | |--11
First comment on this structure: I have no idea why you wouldn't place 4, 9, 10, 11 in one group given your reconstruction. Isn't /ɛθrỹ/ and /æθroːɲ/ pretty similar? If you're unwilling to group these together, then I'd assert you either see that
1) the group you've selected is extremely close to the basal group
2) the group you've selected is paraphyletic
, and not cladal, which is a major problem in your reconstruction
Onto Sangi's divisions (This graph was hell to make)
Code: Select all
| | |------3
| | |
| | |------9
| | | |--10
| | |
| | |----11
Since this is all pulled from your sound changes, I'll just comment on those. There's no valid reason as to why 1 shouldn't be in the əθroɲ group. It's pretty easy to imagine ətroɲ > əθroɲ > θroɲ, and in fact might even be simpler than the root you have provide. If you do this, there's also good reason to add 5, and increase the parsimony
. A similar argument could be made as to adding 6 and making the initial split based on the fricatization of the medial t.
Another reason to cast doubt upon the internal structure is the extremely strong similarities between 7 and 8, which you order as having split early in the history. The difference between 4 and 7 is much larger than the difference between 7 and 8. This either suggests that 7 and 8 are so highly conservative as to retain extremely similar features in a central innovative location, yet 4 was in a more separated from 7 as to diverge quickly. While this is possible, it seems a little handwavy.
Now onto critiquing protowords
Ashtâr Balînestyâr *ŋ̊VC(t̪/d̪)ɾBFŋś
There really isn't much to say here, since a lot is missing. What makes you think that there has to be a back and front vowel together? If you're basing it off of just 1, you seem to end up ignoring the large amount of long back vowels in a lot more languages. Also, how do explain the final /ʁ/ in 3? Is it from ŋś? That's not too unreasonable if you believe in something like ŋś >ɣʰ > ʁ.
There's no minor syllable. As a trisyllabic word in the protolanguage, the word must
contain a minor syllable. The only allowed vowels in a minor syllable are *a and *i. This can be extremely important in determining the fate of a word. Secondly, how do you explain the final nasals in a large number of languages? Are you saying that the final approximant became a nasal? That seems like too many steps in contrast to just hypothesizing a final nasal. Lastly, why do you think the vowel in the middle should be an *o? I'm not sure I see your reasoning Well, I do, but let's pretend I don't
Why are there two r's? How would the initial and medial r be lost? Also, is there a good reason as to why we see the change ttr>g? Also, the reverse happening in ç > ð is pretty eccentric.
sangi39 *hə'troɲ or hə'troin
You didn't look at the protolang phonology. Shame. There's enormous realism problems in your sound changes that make them hard to believe. They make me feel like you started playing the Pogostick game
rather than actually reconstruct.
As the map creator I have several harsh comments for everybody
I'm pretty interested to know why everybody thinks literally only fortition happened in the history of this language.
I also am wondering why very few of you guys have decided to group up families: it's actually a relatively decent technique for avoiding contamination from large language families, although to be honest, it seems like it would be less successful for this cognate set than just the phoneme.
It's also probably pretty reasonable to use my "diachronemic" method. Identify different homologous
parts and place them in context in the word. The hardest part is deciding which parts are homologous.