So, probably to close of this year's posts I will write one more article. Merry Christmas and a Happy new year to you all.
A while ago I had to read an article of Kortlandt, something which is always incredibly interesting, but due to his writing style also means you'll read about one page every half an hour. Nevertheless it presented some interesting things.
The article I read was Hittite hi-verbs and the indo-european perfect . Definitely worth reading, though I won't get into the details right now. Actually I'm discussing the last part of the article, which was put in as somewhat an afterthought it seems.
Kortlandt proposes that the verbal affixes *-(e)i- *-(e)m- *-(e)s- *-n- and *-t/dʰ- originally were verbal elements suffixed to roots. An interesting thought, although maybe a bit hard to prove. The affix I want to specifically talk about is the infix *-n- which Kortlandt glosses as 'to lead', which means he connects it with the verbal root *neiH-.
It takes some steps of reasoning, but might be possible. We'd have to assumed some kind of reduction of *neiH- to either *n- or *nə- I assume that just like the other suffixes this must have once been a suffix, because motivation to infix verbs inside other roots is very small.
To give a small illustration of this root, let me just give an indo-european example. We have the root *bʰeid- 'to split', which in first singular would be *bʰinéd-mi, and in first plural *bʰindmés.
Let's go back a coupla of steps, and look at an early stage of indo-european, where it would still be a suffix, you would get a construction that looks something like this *bʰaid-nə́-mə(i) (*a as a placeholder vowel) and in plural we'd see *bʰaid-nə-mə́s.
Now if we look at what these two forms would look like after syncope, you would get the following:
**bʰid-né-m(i) **bʰidnmés. I think it's obvious that a cluster *dnm wasn't popular at the time, especially since RV cluster don't seem to usually vocalise, only original VR and VRV clusters. so a non syllabic cluster *dnm. It would be only logical to expect a metathesis of the n, to cure this horrible cluster, resulting in the form we know today as *bʰindmés.
Then the singular variant, we can wonder whether *dn- is a legal clusters, but at lest *pn- is like in *pneu-, so that probably wasn't the problem. But a paradigm where a suffix *ne, infixes itself in the root when unaccented, must have felt crazy to the indo-europeans (and with reason). So they moved the whole suffix into the root. giving the perfectly accetable form: *bʰi-né-d-m(i).
So, even if this suffix wasn't from an original verbal root, I'm fairly confident that it was a suffix, rather than an infix that we find in later Indo-European.