¹ ZCP 14.9, 16.301–3, 6.558.

² MSL 17.337.

³ CCG §239.

⁴ Ériu 3.18.

⁵ VKG §319.

⁶ VKG §550 (2) and (3).

⁷ Hdb. §§536, 554–63.

⁸ VKG §602; CCG §448.

⁹ On the Syntax of the Verb in Old Irish, Ériu 12.197.

¹⁰ Ibid. 214.

¹¹ Ériu 3.20.

¹² Thurneysen pointed out that conid ‘that he is’, nand (nant) ‘that he is not’, which presumably ended in *est, do not lenite a consonant, and yet they fail to prefix h to vowels; see ZCP 1.4. In this article, published in 1897, he had proposed an infixed subject pronoun -d- in such forms as conid, nidan (nitan) ‘we are not’, and suggested that it was originally a common 3 sg. and pl. form which had spread analogically. But he later abandoned this explanation. The d is a particle of obscure origin, perhaps IE ḽod, sometimes with relative force, which appears also in Welsh, and has no importance for the present discussion; see Hdb. §§507, 781; VKG §638 (2); CCG §478 (2).

¹³ Little is known of the personal endings in Gaulish (see G. Dottin, La langue gauloise 122), but there is the evidence of to-mez-eclai and dugiiontiio cited above.

¹⁴ F. Henry, Irish Art in the Early Christian Period 46, 142–3, and passim (London, 1940).

¹⁵ C.Donahue, The Testament of Mary = Fordham University Studies, Language Series, No. 1 (New York, 1942).

¹⁶ On the date of this glossary, which is largely Old Irish, see E. MacNeill, Eriu 11.112 (1932). Another example, perhaps earlier, is: ní fetar teacht a n-agaid chainte an chumachtaigh sin ‘it is not possible to oppose the word of that mighty one’, Togail na Tebe, 1076.

¹⁷ Danske Vidensk. Selsk. Hist.-fil. Medd. 11.5 (1926).