[[maraid]] and absolute vs. conjunct issues

Apparently the mainstream theory of Old Irish absolute vs. conjunct etymology in Celtic circles was that Celtic had innovated a apocope of -i in the primary IE verbal endings that was blocked by a conveniently transphonologized enclitic. The encliticized verb forms became absolute forms and the apocopated forms became conjunct forms. This theory is so mainstream that virtually everyone mentioning the subject today follows it (this does not mean that I myself support it), the only exceptions seeming to be Matasović and Kortlandt.

This theory causes headaches in places like maraid, since your form *marati would end up truncated to *marat, but we see mair as the Old Irish conjunct with the palatalization appearing out of thin air.

A few outsiders of Celtic circles (Kortlandt and Matasović) are completely baffled by this theory. There's obvious problem of *ɸeruti (“last year”) leading to Old Irish uraid which obviously conflicts with the desire to get a sound law that apocopates the 3sg ending *-ti down to *-t and *-nti to *-nt. In an attempt to eliminate it, theorists note the accusative definite article being always used before uraid in Old Irish and extrapolate an accusative *(ɸ)erutam into Proto-Celtic (I personally believe the use of the accusative article is analogical). Kortlandt notes that such an accusative form of that adverb has no IE parallel and *péruti shows blatant fossilization in other IE branches already.

In short, the business of reconstructing Celtic verbs is more complicated than we expect.