1 90 Bekher''s Digammated Text of Homer. 



of loiiia. ; neither of which, as we have seen, is much to be relied on. 

 As to the derivation of a relative fo? from JfFos by omission of x, Cur- 

 tius remarks that " the only phonetic analogy which could be called 

 in to support it, is that of the High German wer (for hwer) z= Gothic 

 hvos, cf. Eng. tohat. But the loss of the feeble h j^roves little for 

 that of k ; and how improbable that of the two consonants the Greek 

 would give up the perfectly familiar x in favor of the unstable digam- 

 ma, wavering even from the earliest time." [It may be observed in 

 passing, that Curtius' own derivation of eoixa from dik, dyik, yiJc, is 

 liable to the same objection: it makes the Greek give up the familiar 

 d in favor of the unstable and perishing y.] " Still less," he con- 

 tinues, " can it be proved that Sk. yas has come from kyas, and that 

 ka with the secondary kva, ki/a, is the common root of all these widely 

 ramified pronouns. Finally, the demonstrative meaning of the Greek 

 bg in xal cig sq)i] speaks against this derivation, and recommends the 

 assumption that the originally demonstrative stem i, with the second- 

 ary form ya, lies at the basis of the Greek relative." As the demon- 

 strative use in xaJ o? e(pt] is confined to the nominative, while in the 

 accusative we have nctl t6v^ it seems to me quite possible that the og is 

 for 6, by confusion of the two forms bg and o, so much alike in ap- 

 pearance, though so diverse in origin. Curtius then adds, as Schlei- 

 cher also does, that if the Greek relative did really begin with f, it 

 would be preferable to explain it from a stem sva, which appears 

 with relative force in Gothic sve, as, whence the German so in its 

 relative use. This relative stem sva was long ago recognized in 

 Greek by Curtius himself (Kuhn's Zeitschrift, iii, 75, 76), though 

 only in the merest relic, the adverb qo^j as, which the Alexandrine 

 critic Zenodotus read in two passages of the Iliad {B, 144, Z, 499). 

 Lottner afterwards, in Kuhn's Zeitschrift, ix, 320, proposed to derive 

 all the forms of the Greek relative from this same stem. But the 

 traces of a digamma in the Greek relative are much less frequent and 

 decided than we should naturally expect to find them if this were its 

 real derivation — much less so than in the forms of the possessive bg, 

 ri, bv, his, her, its, which come from a stem of the same sound sva, 

 though of widely diflferent import. Possibly the fact which we have 

 noticed, that the adverb cb? difiers so much from the other forms of 

 the relative in the indications which it shows of a consonant initial, 

 may warrant the conjecture that they are of different origin, that in 

 fact <^ came from the digammate stem sva, while bg, ohg, baog, and 

 the rest, are akin to the Sanskrit yas, yd, yat, and came from a stem 

 with initial y-sound. 



