Page:Morris-Jones Welsh Grammar 0294.png

294 and gynnag pwy, gynnag beth are found in some lesser writings of the late period; more recently they appear in the corrupt and curiously meaning&shy;less forms gan nad pwy, gan nad beth.

As the interrogative is always predicative it is followed regularly in Ml. and Mn. W. by the relative on the analogy of affirm&shy;ative sentences; thus pwy a ŵyr ‘who [is it]  knows?’ on the analogy of Duw a ŵyr ‘[it is] God   knows’, § 162 vii (2). But this appears to be an inno&shy;vation in the case of the inter&shy;rogative, as the oldest examples omit the relative, as puy guant i (1), pa roteiste i (2), pir deuthoste i (5).

. The stems of the interrogative in Ar. were *q$u̯$o‑, *q$u̯$e‑, f. q$u̯$ā‑, also *q$u̯$i‑, *q$u̯$u-'' the last in adverbs only (Brugmann² II ii 348).—W. pwy < nom. sg. mas. *q$u̯$o‑i: Lat. quī < *q$u̯$o‑i''.—W. pa, pỿ adj. < stem *q$u̯$o- compound&shy;ed with its noun and so causing lenition; o after the labial becomes a, or remains and becomes ỿ, cf. § 65 iv (2).—W. pa, pỿ subst. < nom., ace. sg. neut. *q$u̯$o‑d, *q$u̯$i‑d: Lat. quod, quid; lenition is perhaps due to the analogy of the adj. pa, py.—W. peth < *q$u̯$id-dm̥ § 91 ii; already in Brit. the word had become indef., meaning ‘something, thing’, hence pa beth ‘what thing?’ beth is not necessari&shy;ly a shorten&shy;ing of this, as pa is not omitted in such phrases in Ml. W.; but beth is for peth (= Ml. Bret. pez ‘quid?’) which occurs in Ml. W., see i (3), with b- as in ba, by i (2), ban 55, 56.—Ml. W. pet ‘how many?’ Bret. pet < *q$u̯$e-ti § 162 vi (2).—Ml. W. pyr ‘why?’ < *q$u̯$o‑r: Goth., O.E. hwar ‘where?’ < *q$u̯$o‑r, Lat. cūr < *q$u̯$ō‑r.—W. pan < *q$u̯$an-de < *q$u̯$ām-de: cf. O. Lat. quamde, Umbr. ponne § 147 iv (4) p. 245.—Ml. W. cw, cwd, cwẟ represent different forma&shy;tions of *q$u̯$u- (q$u̯$ > k before u § 89 ii (3)) by the addition of more than one of the suffixes named in § 162 vi (2); the different forms have been confused, and can no longer be dis&shy;entangled; similar forma&shy;tions are Skr. kú-ha (h < dh), Gathav. ku-dā ‘where?’ Lat. ubi < *q$u̯$u‑dh‑, O. Bulg. kŭ-de ‘where?’

W. pam, pahám < *pa(ẟ)am < *q$u̯$od m̥bhi ‘what about?’ paẟiw or pyẟiw is obscure; no dative form seems possible; an ana&shy;logical *pod-do might give *pyẟ (as d‑d > d § 93 iii (1)) and iw may be yw ‘is’ § 77 v; so ‘to whom it is’ or ‘for what it is’.

W. bynnag, Bret. bennak, bennag, seems to be from some such form as *q$u̯$om-de ‘when’ + ac ‘and’, so that in meaning it is the literal equiv&shy;alent of Lat. cum-que, and is, like it, separable (Lat. quī cumque lit. ‘who and when’).

The demonstratives hwn ‘this’, hwnnw ‘that’ are peculiar in having a neuter form in the singular. Both are sub&shy;stantival and adjec&shy;tival. The adjec&shy;tival demon&shy;strative is placed after its noun, which is preceded by the article; thus y gŵr