What is the general ablaut rule that explains examples like φρήν, πρόφρων, πρόφρονα?

Question

Φρήν (midriff, will) gives rise to the adjective πρόφρων (eager, literally motivated by will). It looks to me like the -ων comes from ablaut applied to -ην. (It doesn't look like a suffix -ων, since ν is already present in the noun. It's not a comparative -ων, and it also doesn't look like or decline like a participle. It also doesn't look like a person-who-has construction, since those form nouns, not adjectives.) The accusative is πρόφρονα. A similar example seems to be ἀπάτωρ. Summarizing:

πατήρ ἀπάτωρ ἀπάτορα

φρήν πρόφρων πρόφρονα

I'm sure there are discussions of this somewhere, but I can't seem to find anything relevant in Pharr or Smyth or by googling. It seems like there might be a rule that when a third-declension noun ends in a syllable with accented ή, and we form a third-declension adjective by prefixing it with a preposition, then we get the pattern shown above with ablaut and a receding accent, which seems to be the paradigm of κακοδαίμων (possessed).

Is this rule correct? Does it only work with η? Only with accented η? Only with adjectives formed in this way out of a noun prefixed with a preposition?

Answer 1

φρήν, stem φρεν-, reflects a Proto-Indo-European root (straightforwardly *bʰren-, though *gʷʰren- has also been posited) in the e-grade, -φρων, stem -φρον-, has the same root in the o-grade. In both cases here the vowel is lengthened due to Szemerényi's law, a PIE sound law which holds that *-VRs > *-VːR—that is, if a vowel is followed by a resonant and an -s at the end of a word, the -s is lost and the vowel is lengthened; the -s in this case is the masculine nominative singular ending. Szemerényi's law is also why πατήρ has its η, and you can tell the lengthening must be old because ε/ο become η/ω rather than ει/ου, as they would in the historical period.

φρήν is/reflects a straightforward root noun, and it having the e-grade is not unusual. The o-grade in -φρων is pretty common in adjectives, and may be conditioned by the accent moving to the preceding syllable. The details are not necessarily well-understood, but they're details belonging to Proto-Indo-European, not Greek. The ablaut here (and indeed anywhere in Greek roots) is certainly not a productive mechanism in Greek anymore, and trying to read too much into it is not likely to be helpful to you.