I am writing this because I regret my reply to Miguel Carrasquer Vidal's explanation of neuter u-stems as r/n-stems. Nevertheless, perhaps it wasn't posted -- I hope it wasn't deemed offensive, because that surely wasn't the intention. I hope I did not come across as attacking Mr. Vidal's hypothesis unfairly. I realized I am certainly not qualified to accurately critique his ideas, that sometimes I have to defer to the more learned in this group. I was merely conveying my initial impression, which was especially based on the idea of the original form of *me:nes- "month". But since writing that message, I have thought about neuter u-stems and remembered that what Mr. Vidal is saying about them does ring true for Sanskrit: I remembered that all neuter u-stems have a connecting -n- in
oblique cases before vowels, which suggests that they may have terminated in *-n at one stage, at least in these oblique cases, and hence with *-r in the nom./acc. sg. So perhaps I was too hasty to judge. But two problems arise: one is that not only the neuter u-stems have connecting -n- in Sanskrit, also the neuter i-stems have this. This suggests that these also may have been originally r/n stems -- if this was the case, wouldn't we expect -r in neuter i-stems also in Armenian? The other problem is, final -r regularly appears as visarga (-h.) in Sanskrit. If these neuter stems originally ended in *-ur or *-ir in the nom./acc. sg., wouldn't one expect to find -h. in Sanskrit, at least in some recorded forms?
Andrew Jarrette