HI, 1S75.] SKETCH OF SAMAK GRi 27 ninre strictly observed in the Hadramaut than in the Minooan dialect", where sometimes the nsaal form occurs. The first of these dialects, how- ever, displays another peculiarity, namely, if an u~s to be added to a word terminating with n it is not amxed, but the n is changed to n, tSnua fate, rhR, hch nn (^- H. 1), for vrehs, Dim, cnr. But this n, instead of representing an absorp- tion of d, may perhaps only represt! aspirated pronunciation of the feminine n, as is usual in Barbary and in some districts of Yemen. Permutation between rt and n occurs seldom, as in tmrn (Fr. CV. 1). rra (N. H. 1), and ranm* (H. G. I, 8) for wsm, m*, pinN, but this may perhaps be an error of the copy. A much more important permutation is that according to which the sound 5 (>f) is rendered by a simple : (j/)like sa {0&. xvii. 8-9) is dyy{ibid. .win, 10). In the same way the root nr33 (H. 169, 2) must assimilate with the Arab °*^t and the root hidden in the form pnr {Mai. 639, 3) does not differ from that occurring in tkv (Ob. iv. 10-11- This permutation proves that the Sabooans always pronounced a £ hard, as some tribes still do, and not like;', as is most usual with modern Arabs. In the repetition of two consonants the full uncontracted form is preferred in Sabeeon. This peculiarity appeals to be confined to the liquids S "3, 3i "i ; for instance : — Wvp {Os. s. 7), pesrr (D- M.G.xix. 1), pa (0s. xv. 2), mr (H. 191, 1), instead of jVct pr.-r, r- r > ** • The proper noun
- sm must certainly be pronounced Wadd>
(«£j) "friend," as the form of the participle (n)TTO (#. 187, 2) proves, The roots of the Sabamn language are mostly triliteral, and present all the forms occurring in the other Semitic languages -.— I. Perfect roots : — -qd, pot, tcd, rn, mi; TI. vt, **e mpii fen, OCT* -rn. tti, ^cn : "vr, OB', rr, tr ; III. i*r, ^ and s-r: ibi, tt, to ; ere, ytf, 2?. rrr; £o, an, p ; IV. v">, V? and »^ : tht, tOT. «* ; 'a. t», ia, «=P, «:, *TQi nhq, wn. V. Mixed roots :— <ci, fr, »o», Hit tti. As already observed, qnadrilitoral roots arc scarce, and seera to occur only in some pr nouns, such as cisrn or mtf*t, TQfti and in the name of the divinity info?, which latter is abri to nr« when it designates a man's name, nnrrn and rrtspm are contracted from -innrswi, -vftarepvu The noun tim is contracted from rttVvfi I .-o). The word t&VD " ram" (E. lift, «•. &c ) appears to be formed from -armi, "he who outers ' the house :" which epithet may refer to the un- concernedness of this animal. The degradation of the sibilants may be re- presented by the following scale : — i-t*-t t> — fc - □•-.=!*- n** — n These transitions do not occur with regeds and in a logical manner in the Semitic languages; they are possible without being necessary. T In- forms accompanied by an asterisk are common both to the Arabic and to the Sabwan ; the latter, however, may degrade the original o a degree further by transforming it into fi. TheAramasan a last stage of degrad.it ion In the other consonants the Sabuean generally agrees with the phonetics of the Arabic languor sometimes, however, it deviates, and follows a manner peculiar to itself. Among cases of this kind the fact is to be pointed out that the Arabic words t ** pom and ^* who are both rendered by p, so that it becomes difficult to distin£Tiish them from the substantive a ton, but iu some cases this change does not take- place, and the word occurs exactly as in Arabic. The exuet determination of the letters equi- valent to p and ijo is due to Osiander, hut he was not so successful in determining the equivalent of £■ There is also another character the value of which was debated, but is now believed by Halevy to represent an intermediate sound be- tween the Arabic letters £» and o. (See p. 26. B.) The Sabsans rivalled the Egyptians and the Assyrians in the extreme care with which they produced their graphic texts : hence the inscrip- tions of Yemen am numbered among the most beautiful of antiquity. They are traced on stone or metal, and present a monumental character which Beems to have l>een immutably fixed in very remote times, else it could not have sub- sisted with such uniformity from the banks of the Euphrates to Aden, Somo details observed in certain letters are not confined to a fixed region, but arise oulv from the sculptor s manner. But, in spite of the general neatness of the characters. it is impossible for copyists not to confound with other certain letters, especially when they have to deal with a text obliterated, or seen
