v. (1) The negative relative is nom. acc. ni, nid, Ml. W. ny, nyt; this form is also used in the gen., in the loc. after lle, and in cases governed by prepositions; but the adverbial form generally (e.g. after pryd, modd, fel, megis, paham, pa fodd, etc., and adverbs like braidd, odid, etc.) is na, nad, Ml. W. na, nat. In Late W. there is a tendency to use the a form everywhere.
Nom.: Nyt oes yndi neb ny’th adnapo r.m. 3 ‘there is in it no one who will not know thee’. Gwyn ei fyd y gŵr ni rodia Ps. i 1.—Acc.: yr hynn ny welsynt Ỻ.A. 12 ‘that which they had not seen’; cenedl nid adweini Es. lv 5; also with a redundant ‑s: llyna beth ny‑s gwrthodaf-i c.m. 42 ‘that is a thing which I will not refuse (it)’.—Gen.: y drws ny ẟylywn ny ẏ agori b.m. 41 ‘the door which we ought not to open’, lit. ‘whose its opening we ought not’.—Loc.: lle ny wyper Ỻ.A. 26 ‘[in] the place where it is not known’. After a prep.: ny roẟei hi…iẟaw r.m. 33 ‘to whom she did not give’. Adv.: pryt na Ỻ.A. 26, w.m. 183, r.m. 85, pryd na Jer. xxiii 7, D.G. 29, g. 297; mal na c.m. 20; braidd na D.G. 50.
(2) The perfective particle ry may introduce a rel. clause; see § 219 v.
vi. (1) The relative pron. a probably comes from the Ar. relative *i̯os, *i̯ā, *i̯od: Skr. yá‑ḥ, yā́, yád, Gk. ὅς, ἥ, ὅ. It was a proclitic in Brit., and pretonic *i̯o might become *i̯a § 65 vi (2); this was metathesized to ai the oldest attested form, as in hai-oid b.s.ch. 2 ‘which was’, ai torro hac ay dimanuo y bryeint hunn l.l. 121 ‘who breaks and who dishonours this privilege’, hai bid cp. ‘which will be’; and ai was reduced to a, a trace of ae occurring in Ml. W., see i.—To explain the soft mutation after it we have to assume that in Kelt. the nom. sg. m. was *i̯o like that of *so, *sā, *tod: Gk. ὁ, ἡ, τό (forms without ‑s are older, and *i̯o might be a survival).—The verb syẟ, yssyẟ represents regularly *estíi̯o = *estí i̯o; it differs from yssit ‘there is’, which sometimes precedes it, as yssit rin yssyẟ vwy b.t. 28 ‘there is a secret which is greater’, § 189 iii (3). The acc. a (< *i̯om) prob. had a radical initial after it at first, cf. ae gulich i above, and a gulich…‘which…moistens’ four times in b.b. 46.
(2) In Ar. adverbs were formed from pronominal and other stems by adding various suffixes, many of which began with a dental: thus, denoting place, *‑dhi (Gk. πό-θι ‘where?’ ὅ-θι ‘where’), *‑dhe, *‑dha (Skr. i-há ‘here’, Gk. ἰθα-γενής), *‑ta (Gk. κατά, W. gan < *km̥-ta); whither, *‑te (Gk. πό-σε? < ‑τε, Goth, hvaþ ‘whither?’); whence, *‑dhem (Gk. ‑θεν), *‑tos (Skr. yá-taḥ ‘whence’, Lat. in-tus, W. hwn‑t ‘hence’); manner, *‑ti (Skr. í-ti ‘thus’, Lat. iti-dem), *‑thā (Skr. ka-thā́ ‘how’, yá-thā ‘as’, Lat. ita < *i-tā); time, *‑dā (Skr. ya-dā ‘when’), *‑te (Gk. ὅ-τε ‘when’); Brugmann² II ii 728–734. To these may be added the adj. of number formed with *‑ti (Skr. ká-ti ‘how many?’ W. pe‑t id., Lat. quo‑t, Skr. yá-ti ‘as many’).
