Could you check this over
Fragment of a discussion from User talk:Rua
In both cases I believe it was reanalysis.
- For -mentum, there are examples like armentum, whose earlier form (armenta) is 1st declension. This would then have been reanalyzed as the neuter plural.
- Similarly, I suspect the that -μᾰ was suppletive (*-mn̥, *-mn̥th₂) and the t-stem crept back into all the singular forms except the nominative singular.
This may just be idle rambling from having looked at these too long, but it seems plausible to me.
For me, the simplest explanation for the Greek form is just as an athematic -mn̥t-. Greek regularly loses final -t, so in the nominative the result would be -mn̥t > -mat > -ma or -mn̥t > -mn̥ > -ma (both orders are possible and give the same result). In the other forms, the -t- was not final, so it was retained. I see no reason to involve suppletion, unless I am missing something.