In greek, the indicative mediopassive in the 3rd person singular and plural are -εται and -ονται, however in PIE, the alpha was originally an o. Additionally, in the imperfect, the endings also have o in them. Why did this shift to alpha, since it seems like an omicron and iota seems phonologically legal in greek, such as in the masculine plural article οἱ.

  • 1
    I'll consult Willi's Origins of the Greek Verb later when I get a chance. If you can get your hands on it it's an excellent read on the PIE verbal system through the lens of Greek
    – Tristan
    Nov 29, 2023 at 13:24
  • Thanks for the suggestion, I'll try to look at it.
    – Blubber
    Nov 29, 2023 at 13:40
  • 1
    Greek takes a capital g, We should make an effort...
    – Lambie
    Nov 29, 2023 at 16:42

1 Answer 1


Willi's Origins of the Greek Verb argues that it's by analogy from the 1sg, 2sg (from PIE *(m)h₂ei̯ & *soi̯/-sai̯), noting Arcado-Cypriot and Myceanaen forms in -τοι:

In Greek *-toi̯ > -τοι is attested in Arcado-Cyprian and Mycenaean; -ται is analogical after the 1sg./2sg.

Note that Willi does not assume the 1sg & 2sg directly continue the PIE middles, but instead consist of the h₂e-conjugation (stative) endings appended to the active athematic ones (-mi conjugation), and then followed by the primary -i in the present tense.

These are badsed on a combination of endings from both series/conjugations. ... the ending *-o acquired a mediopassive value and could be added to the 3sg. ending *-t of the m-conjugation form new 'oppositional' middles...

... Since Vedic 'primary' 2sg. med. -se can continue either *-sai or *-soi, a late-PIE 'primary' *-s-t(h₂)a-i̯ may have been 'regularised' after active *-si into *-sai̯ (> Ved. -se), while the vocalism of only 'secondary' *-s-th2e > *-s-t(h₂)a > 'regularised' *-sa was aligned with that of 3sg. *-to. This would imply that Proto-Grreek possessed a 'primary' middle series 1sg. *-mai̯ (cf below), 2sg. *-sai̯, 3sg. *-toi̯ vs 'secondary' *-mā(m) (cf.below), *-so, *-to, and it would thus account for the general a-vocalism of the 'primary', but not 'secondary', middle endings of most historical dialects (Att. -μαι, -σαι, -ται vs. -μην, -σο, -το).

[footnotes ommitted]

This does require two analogies (one for 2sg. *-st- > *-s-, and another for 3sg. *-oi̯ > *-ai̯).

The ablaut of the PIE middle is not consistently reconstructed (Sihler & Beekes reconstruct o-grade, whilst Fortson & Ringe reconstruct e-grade). An o-grade explains the lack of a-vocalism in the secondary middle (i.e. imperfect & aorist middles) but makes the a-grade in the primary middle (i.e. present middle) problematic, whilst the e-grade is exactly reversed (explaining the a-vocalism of the primary, but not the lack of it in the secondary middle). Willi's explanation using concatenation of the stative and active (athematic) endings explains both.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.