So here is my second guess, based only on the diachrometric method and not on regular sound correspondences, subgroupings and sound changes like the last one (I admit, I assumed a lot of fortition)

Here's my idea about homologous parts:

Code: Select all

```
ABCDEFG H
1. fro i n
2. ŋ ɯx ʌ ʒ
3. ŋ ɵð ɤɤʁ
4. æθr uː
5. frig ʌ n
6. x eθ ɵ n
7. h attro ŋ
8. h əttrʌ n
9. yʑ ɯ̃
10. ð oːɲ
11. ɛz ĩ
12. gʒɛh y n
```

A: I am really not sure where the /f/ came from. /x/,/h/,/ŋ/ and /g/ seems to hint at *g IMO.

B: If we look at 12. this might have been the original position of *r, since it yielded /r/ and /ʒ/.

C: The first vowel. We've got nearly everything here, but since I assume sesquisyllabicity, this has to be either *a or *i. My feeling somehow tells me *a since there are only two high vowels in the descendants.

D: There seem to be two groups involved here. /tt/,/z/ and the other coronals can be traced back to *tt, but /h/, /g/ and /x/ look different from the rest. Maybe a cluster *ttk?

E-F: I will just assume that the /r/s here come from some kind of weird metathesis. I mean why not?

G: Again, this is rather hard. The descendants have front and back, rounded and unrounded, high and mid, and tense and lax vowels. There are no low vowels and there are no front mid vowels. Based on the following statistics I will assume *ɤ here.

- 3/12 front

8/12 back

1/12 central

7/12 unrounded

5/12 rounded

5/12 high

7/12 mid

H: This again looks like *g because it yields /ŋ/,/ʁ/ and /ʒ/.

So, all in all my second guess is *grattkɤg. I don't really feel it's better than my first guess though. This crazy initial /f/s really bug me.

**Edit: **Maybe rather *graktɤg, the gemination could just be assimilation. Also here is a new and detailed grouping.