Jamuaet, 1875.] SKETCH OF SAB.EAN GRA MM AR. 26 I. Votive inscriptions, usually engraved on bronze tablets or stone slabs, occurring in the interior or in the immediate vicinity of temples. A multitude of national and local deities are mentioned in them, and these prove the Satjffiau pantheon to have been prodigiously rich. It. Votive tablets, called by Halevy 7 V - cynimes, belonging to strangers who completed some act of devotion in the sanctuary and there inscribed their names and descent. These in- scriptions are engraved on slabs suitable for the purpose, and the formula? vary but Blightly. The chief interest of these inscriptions cent res in the large number of territories, towns, and tribes mentioned in them, affording materials for a restoration of the ancient geography and ethnography of Southern Arabia. HI. Architectural inscriptions, traced on the walls of temples and other public edifices, in order to commemorate the name of tin.- building, or of the persons who had contributed to it. In this latter case, care is taken to indicate the exact dimensions of the portion constructed by each man, and if a stranger was among them his country and tribe are mentioned. Inscrip- tions of this kind constitute the majority in Halevy's collection. IV. Historical inscriptions, intended to an- nounce a victory gained over a foe, or to i memorate an event wherein the author plays a part. Under this head fall the tex Mareb and of S i r w a h , the inscription on Hisn G'ura'b, and especially the inscription OH the monolith of S ir w ii h, the copy of which was stolen from Halevy by the Arabs. V. Police orders, engraved on pillars at the entrance of temples or other public localities, in to warn the people against the commission of damages antler pain of fines. These tests are very interesting, because they show great otacm in the civil organization, as well • existence of a penal code among the Sabasans. VI Funerary inscriptions, not many of which have as yet been discovered, but which prove that the Salwans were in the habit of carrying away their dead from inhabited places into isolated valleys, and up mountains, where they erected small houses for them. ^^^^^^^ Sal/wan Alphabet. Tn the following sketch of Sabsan grammar I shall give only what has been fixed by valuable authorities, not the least of whom is Halevy, whom I intend to follow closely. I shall only mark by signs of interrogation words not yet fully authorized, and shall designate the inscrip- tions of Fresnel, Osiander, and JJuK'vy respec- mvly by Fr. Us. Hal. or H. Although the alphabet is at present well fixed, 1 append jI Jiarmonic Bctbaati, Ui'h I Arabic ! adduce some peculiarities ■rs ; mention a few, the occur- rence of which is rare, and whose value was I first very well fixed ; enumerate those letters which are apt to give rise to confusion ; and, lastly, I shall montion snch letters as may, by their too great proximity to each other in certain inscriptions* sometimes be mistaken for one letter. The n generally remains after the preforma- tive letters of the Imperfect, thus : — rerr, tw. There are, however, a few exceptions j ytfr (U- 152, 14-*), itt {Os. viu. lit). In the Miuamn dialect the n is sometimes considered as a vowel : thns we very often find pa (ccsttj H. 188, 5), tto-m (IT. 199, 1), nrons (JT. Ill, 5) for p, to, rtfte. This takes place even in the divine name vr-srs ill. 222, 1 : 229, -I). whu-h is certainly derived from the radical vl-i (y^j = nsi). The' gutturals, n, fr. r, ta nt no peculiarity. The letters i ami like the first radical of verbs, are often elided by the servile letters : — per, np* ! * «wv» a nrr | when the i forms the third radical it generally remains unchanged, thus :— <n) -oc. •----, ww ; but also ■&>, nih occur ((.' I, '.')• The servile n usually becomes o in the Mhunan and in the Hadramaut dialect ; thns irr, n, the sufhx of the third person, appears in these dialects in the form ir D, for instance icji or c», ceres* instead of im (rci)* crcca* (■rareta*, cc*=h). In the same manner the fourth form of the verb, which is in ordinary Sabamn brcn, becomes ttto in these dialects : thus the Sabsean "3pn, ninn in the in and Hadramaut dialect will be *%v* rnrc- Tliis is another approach to the Assyrian, and in general to the languages of East Africa. This form answers to the Aramaean tare and to the Ethiopic tencH- The change of the Bervile n into d is much
- Hal&vy'* inscriptions here referrwl to will be found in
1 Odunuer'* are given is No. 7. Oct. 1873, Ac the Journal Asiatique (1873 For— Kui).
