¹ So, e.g., Brugmann-Thumb, Gr. Gr.⁴ 376.

² Confusion of form due to the change of dh to Lat. d (cf. Sommer, Handb.²178) assisted in the merger of the two verbs, but the complete obliteration of the difference would have been impossible except for the association due to meaning.

³ The question of the origin of the earliest of these forms is here ignored as being irrelevant.

⁴ If these forms were actually allowed to exist, they became ∗τ∈-ϑι and ∗-ϑι, and the obscurity of these would cause and to be welcomed as a means of escape, while ∗δο-θι would have been unaffected.

⁵ That δíωμι has in addition the first singular impf, δίδονν instead of ∗δίδω-ν, as opposed to -τί-ϑŋ-ν, has two causes. In the first place, the association between δίδωμι. and the two others was less close than between τίϑŋμι and ΐŋμι, because, on the formal side, there was no help from identity of root-vowel. Secondly, the contract form of δίδονν was closer to the root-vowel than ∗τίθονι> < ∗ would have been to ϑη-, so that there was not the same motive for avoiding it.

⁶ Cf. Boisacq, Dict, étymol. 369, and Walde, Etym. Wörterb.² 373.

⁷ Cf. Boisacq 969, and Walde 266.

⁸ It is unimportant that ∗dhēk- is found in three language groups, and ∗iēk- in only two. Since the material from Phrygian is so exceedingly scanty, the absence of a Phrygian descendant of ∗iēk- is undoubtedly accidental.

⁹ Cf., e.g., Brugmann-Thumb 377.