¹ Cf. Schwyzer-Debrunner 2.592. Their view that the oukh that occurs before rough breathing is elided from oukhí ‘not at all, certainly not’ seems unnecessary and is hard to reconcile with the difference in meaning between oukhí and ou. The chi in oukh is simply the normal Greek way of writing the phoneme sequence /kh/ that arose when a word ending in k was followed by one beginning with h.

² Cf. e.g. A. Heusler, Altisländisches Elementarbuch ³ 28.

³ Cf. the accusative Go. aiw, OIc. ey, OE ā, OHG ēo.

⁴ On alleged ∗ne- beside ∗- in compounds, cf. Puhvel, Lg. 29.24–5.

⁵ It is tempting to try to make something out of ‘pleonastic’ mḕ ou and ou mḗ, but these usages are not found before the fifth century and are comprehensible starting with a meaning ‘not’ for ou. In Homeric passages like Iliad 1.28, mḗ nú toi ou khraísmēi skêptron kaì stémma theoîo ‘lest (mḗ) the staff and garland of the god not (ou) keep thee from harm’, each negative is performing its own function.

⁶ But Odysseus' pseudonym Oûtis is conceivably a relic of accented ∗oû and not an invention of grammarians, as Schwyzer supposed, Griech. Gr. 1.377.

⁷ Cf. E. Risch, Mus. Helv. 12.70.

⁸ Students of Albanian may wish to follow up the possible connection with Alb. as, s-‘not’ pointed out by Pedersen, KZ 36.322.

⁹ Cf. my remarks in the report of the 1959 Texas Conference in Indo-European Linguistics, Evidence for Laryngeals 97.