Brent VINE Att.
erôtáô: Phonology vs. Morphology
Att. erôtáô, Ion. eirôtáô 'ask' is a close derivational relative of Att. éromai, Ion. eíromai 'ask, inquire' (*erwo/e-), eré(w)ô 'id.', but the source of the -ô- in e(i)rôtáô is obscure (Chantraine: "dérivation ... inexpliquée"). Two solutions have recently been proposed, both phonological. This approach has fundamental weaknesses, leaving room for alternative explanations. Recent work by A. Nussbaum on secondary "decasuative" formations (e.g., "de-instrumental" adjectives based on instrumental case forms) suggests an attractive morphological solution to this long-standing crux of Greek word-formation.
J. Rasmussen has suggested (1986) that while eíromai, eré(w)ô require stem-forms (Pr.-Gk.) *erwo/e-, *erewo/e- < IE *h1rw-o/e-, *h1rew-o/e- (cf. Rix 1970), the -ô- of eirôtáô could be explained via two assumptions: (1) Greek inherited forms with laryngeal "enlargement" or "root extension" *-h3- (i.e., Greek has both *h1rew- and *h1rew-h3-, the latter supported by the *(H)ruH-ná-h2 of OIr. rún, OHG ru:na 'secret, rune"), and (2) a zero-grade *h1ru-h3-C would regularly yield Pr.-Gk. *erwo:C, via the "Lex Francis-Normier". Rasmussen later (1990/91) added that o-grade formations like ON rauna 'investigation' would have lost the internal laryngeal, which might explain the Greek forms that lack it.
This explanation is ingenious but problematic. The "Lex Francis-Normier" remains controversial, and no account of eirôtáô that depends on it can be compelling. It is also troubling that all the Greek material (besides eirôtáô) excludes a final laryngeal: besides eíromai, eré(w)ô, note ereunáô 'search after', Cret. ereutaí 'zêtêtaí', and Myc. e-re-u-te-re (PY), if this means 'inspector' or the like. Nor is the comparison with OIr. rún etc. ironclad.
Another solution involving a root-final laryngeal (E. Tichy ap. LIV) takes e(i)rôtáô as denominative to a *erôtá: < *h1roh1-táh2, with eréô and eíromai from *h1reh1-, since Hom. aor. ...erésthai# excludes digamma and "für eree- und eire- ist *w nicht gesichert". This derivation avoids the Lex Francis-Normier, but substitutes an improbable morphological entity. Nor is the lack of digamma in Hom. ...erésthai# problematic (Kimball 2000), and the digamma in eire- is independently supported (cf. Hdt. eirôtáô). Worse, a basis *h1reh1- separates this material from ereunáô and Cret. ereutaí.
A derived iterative like e(i)rôtáô should begin from a participial *e(i)rôtós; and this should be based on the plain *h1r(e)w- attested in eréô, eíromai, ereunáô. As Nussbaum has shown (1996, 1998), many IE formations in /-V:to-/ derive from "possessive instrumentals", i.e. substantival instrumentals in *-V-h1 (e.g., o-stem *-o-h1) which are then "adjectivized" via *-to-, as in kholôtós 'angry' (: khólos 'anger'). For e(i)rôtáô, one need only assume an o-grade action noun *h1rów-o- '(act of) inquiring', a well-developed type in Greek. According to one scenario, *h1rów-o- could have had a collective *h1rw-áh2 'inquiry'; the instr. sg. *h1rw-oh1 'by inquiry' (with zero grade from the collective) then serves as the basis for an adjectivization (Pr.-Gk.) *erwo:tós (cf. kholôtós), whence iterative e(i)rôtáô.