increases its value to the philologian. In history and geography, the style and manner are simple, often indeed careless; and the whole wears rather the character of compilation.

For our purpose, it is particularly important to become acquainted with the lexical works upon the Arabic language, compiled by native grammarians; and with the manner in which our modern lexicons have been derived from them. Here, two classical lexicographers meet us at the outset. The first is Abû Nasr Ismael ebn Hammad al Djauhari, (i. e. the jeweller,) commonly called Djauhari or Jauhari, who died in 398 of the Hegira, or A. D. 1107, a Turk by birth; who, after many journeys among the tribes most celebrated for the purity of their language, compiled a dictionary to which he gave the title of the pure; in which he has collected about forty thousand words, but with the exclusion of provincialisms and phrases not entirely pure. It is arranged, like most of the Arabic lexicons, according to the final letters; and the significations are supported by numerous citations, mostly from poets and grammarians now lost.65 The second is Medjeddin Mohammed ben Yakûb el Firuzabadi, known most commonly by the latter name, of Firuzabad in Persia, who died in 817 of the Hegira, or A. D. 1415. He was the author of a yet more complete dictionary under the title of the Kamoos, or the Ocean;66 in which all words, even the most infrequent, are inserted; but with the omission of the citations of authorities, of which at least not very many occur. The number of words amounts to about sixty thousand; and the

65 A Turkish translation of this work was printed at Constantinople in 1728, in 2 vols. folio. The translator's name is Wan-Kuli, i. e. servus Wanensis, properly Mohammed ben Mustapha of Wan in Armenia. A Specimen of Djauhari was published in Arabic and Latin by E. Scheid, 1774. 4to. Complete manuscripts are rare.

66 After this work had lain for centuries inaccessible to oriental philologists, except a few,-of whom the Hollanders, as A. Schultens, Scheid, and N. W. Schroeder, have most frequently made use of itit has at length appeared complete in print: The Kamoos, or the Ocean, an Arabic Dictionary, etc. Calcutta 1817, 2 vols. folio. See the Leipz. Lit. Z. 1818, No. 200. Allg. L. Z. 1820, No. 121 sq. In the Arabic preface the author speaks of another monstrous work of sixty volumes, called the Lamé, which he had previously composed, and of which the Kamoos is only an abstract; but the passage leayes it doubtful, whether he had actually executed this work, or only begun it. See De Sacy in the Journal des Savans, Juin 1819.

compiler states, that he gives in this work the fruits of his perusal of two thousand of the most applauded Arabic writers.

From these two original lexicons, have been chiefly derived all the modern Arabic lexicons; and that in the following order. The earliest, by Antonio Giggeius,* contains a translation of the Kamoos, the words only being arranged in the occidental manner according to the initial letters. But not only is the general arrangement of the work exceedingly inconvenient, inasmuch as the significations and derivatives of each root stand confusedly mingled together; but the translation also, as we are now in a situation to perceive, is not to be confided in, and is full of errors; arising no doubt partly from the false readings of the probably incorrect manuscript, and partly also from a false apprehension of the meaning of the original, in consequence of the no small difficulties which attend the perusal of it. The Latin, moreover, is barbarous, and often leaves the reader in doubt as to the sense which the translator means to express. A better work, beyond all comparison, is that of J. Golius,† which is as yet the most useful Arabic lexicon. This author, who had opportunity to perfect his knowledge of the language in the East, took Djauhari as his basis, supplied from the Kamoos what was deficient, and made use besides of many other lexicographers and writers with much learning and judgment. But he wholly neglected the proper arrangement of the significations; so that the primitive signification sometimes stands last. Something more was done in this respect by Castell in the Arabic part of his Heptaglotton; he also supported the different significations by quotations from the biblical versions and from some writers on medicine and natural history, e. g. Avicenna, and supplied also many words and forms from the Kamoos. But just these supplements are not at all to be depended on; because he has mostly, if not exclusively, made use only of Giggeius; and the scholar has cause here to be very much upon his guard. The very useful dictionary for other purposes, compiled by Francis a Mesgnien Meninski, and republished by Bernhard von Je

* Thesaurus Linguae Arabiae, Mediol. 1632. 4 vols. fol. Lexicon Arabico-Latinum, L. B. 1653. fol.

See his preface.-Golius was for a time with the Dutch ambassador in Morocco; travelled in 1625 and the following years, in the Levant, to Aleppo, Arabia, and Mesopotamia; and died in 1687 as Professor of Arabic at Leyden.

nisch, 1780-1802, in 4 vols. folio, is rendered less useful for etymological investigation, from the very circumstance of its strictly alphabetical arrangement; and besides, the Turkish and Persian are the principal objects in it.

While now these writers, and especially the Kamoos, are most invaluable and authentic sources for the knowledge of the existing idiom, yet the scholar must not seek with the same confidence in them for etymological research, for remarks upon the primitive significations of words, and the like. For these objects the Arabic Scholiasts on difficult writers, as on Hariri and the Moallakat, are much more fruitful sources; and A. Schultens especially deserves great praise for his labour in working up and developing this rough mass of facts, although he may have here and there gone too far in his etymological conjectures. Many etymologies of Schultens' school, lexically treated, and with a cautious separation of what is conjecture and what is fact, are found in Willmet's Arabic Lexicon; in which at the same time special reference is had to certain Arabic writers.* The object to be aimed at in a new lexicon, such as that of Professor Freytag of Bonn, which is now in press at Halle, is partly a renewed use of the native lexicographers and a certain number of the best authors,67 and partly the above mentioned more correct arrangement and deduction of the significations.

The personal and continued perusal of Arabic writers will be indispensable to the truly learned interpreter of the Old Testament; and will always be to him a rich source of parallels and comparisons for language in the broadest sense of the word, as also for ideas, poetical figures, etc. But still, the lexical helps which are extant, will suffice for him who knows how to employ them; and will enable him to determine almost every where the proper usus loquendi, and to detect the errors which have

* The Koran, a portion of Hariri, and Ebn Arabshah. The title of the work is: Lexicon Ling. Arabicae in Coranum, Haririum, et Vitam Timuri, L. B. 1784.

67 The late Professor Berg of Duisburg had made very important collections for this purpose; his copy of Golius, written full with copious citations, is now in the university library at Bonn. Good service may also be rendered by the Arabic Index to Hariri, which De Sacy has subjoined to his splendid edition of the Mekamath (Paris 1822. fol.) in reference to the Arabic commentary. [The first volume of the lexicon of Freytag was published in 1830; for a notice of it, see the Bibl. Repos. vol. I. p. 197, 198.-Ed.

found their way into many philological helps, partly through a careless use and misunderstanding of the lexicons, and partly through the mistakes which have crept into the most usual of the lexicons themselves.68

In a language so copious and so widely extended as the Arabic, it could not well be expected that there should not be a great variety of dialectical differences; and we find in fact that many a dialectical idiom of this sort corresponds more nearly to the Hebrew, than the common Arabic language of books. This is particularly the case with the so called Vulgar Arabic; which indeed much more resembles the Hebrew, both in a grammatical respect by its fewer and shorter forms, and also in its far less degree of copiousness. Even the still remoter dialects of the Moors and of the natives of Morocco and Malta,69 contain many peculiarities, which connect themselves back with the most ancient idioms; just as it is often the case, that the same phenomena of language often repeat themselves unexpectedly, in the remotest periods and districts of the same. people.*

68 The best grammars of the Arabic language are the following: Erpenii Grammatica Arabica, ed. A Schultens, L. B. 1767. 4to. Jahn Arabische Sprachlehre, Vienna 1796. De Sacy Grammaire Arabe, 2 tom. Paris 1810; reprinted with large additions, 1832. Rosenmüller Institutiones ad fundam. linguae Arabicae, etc. Lips. 1818. T. C. Tychsen Grammatik der Arabischen Schriftsprache für den ersten Unterricht, Gött. 1823; a very useful compendium, but full of typographical errors. Ewald Grammatica critica linguae Arabicae, etc. Vol. I. Gött. 1831.-ED.

69 "Ktyb yl Klym mâlti 'mfysser byl-latin u byt-taljân, seu Liber dictionum Melitensium, h. e. Mich. Ant. Vassalli Lexicon MelitenseLatino-Italum," Romae 1796. 4to.-However much a certain pride of ancestry may dispose the Maltese themselves to refer their language back to the ancient Punic, still all its peculiarities may be most naturally explained, as the best helps sufficiently shew, from the modern vulgar Arabic, without the necessity of going back to so early a

source.

* In the dialect of the Tayitic Arabs, 9, like the Hebrew 77,

stands for the relative; in that of Yemen,

i.q., myr

2, to sit. In Maltese we find e. g. ghad, fut. x,) to say, relate, declare, i. q. 779, 77977.

In the Ethiopic language*, we possess a very remarkable relic of the Hamyaric, or dialect of southern Arabia, which, since the time of Mohammed, has been wholly supplanted by the present written Arabic; the Ethiopic people having been, as is well known, a colony from southern Arabia, which wandered across the Red sea." 70 Its peculiar written character also seems to be of Arabian origin, and to have been primitively the same with that of the ancient Hamyaric. The language accords in general, indeed, most nearly with the Arabic, but is less rich and cultivated; and has quite a number of words in common with the Hebrew and Aramaean, which are not found in the Arabic.72 The literature is exclusively of an historical and ecclesiastical nature; but of the not unimportant historical writings, nothing has as yet appeared in print. The date of these writings falls between the introduction of Christianity into Ethiopia under Constantine the Great, and the fourteenth century; when this language, by a revolution in the government, was supplanted by the Amharic, which is still spoken in Abyssinia, while the Geez dialect is employed only as a written language for every species of written works or documents. The Ethiopic, both in a lexical and grammatical respect, has been handled with uncommon skill and learning by J. Ludolf; whose Grammar and Lexicon, in the second editions of them,73 are among the best philological

Called at present in Abyssinia the Geez language, and used exclusively in the church and as the language of writing and books; while the Amharic is the spoken language of the court and people.-ED.

70 See the Syllabus Vocum harmonicarum, subjoined to Ludolf's Lexicon Aethiopicum, Ed. 2.

71 See on the Shemitish origin of this alphabet, the author's article 'Amharische Sprache', in Ersch and Gruber's Encyclopaedia; in which deduction Kopp, some years later, though independently, coincides; Bilder u. Schriften der Vorzeit, II. p. 344.

72 See Ludolf's Comment. ad Hist. Aethiopicam, p. 57; and also generally the author's article 'Aethiopische Sprache und Literatur' in Ersch and Gruber's Encyclopaedia, II. p. 110 sq.

73 Jobi Ludolfi Grammatica Aethiopica, Ed. II. Francf. ad. M. 1702, fol. Lexicon Aethiopicum, Ed. II. ibid. 1699, fol. The first edition of these works (by Wansleben, Lond. 1661, 4to.) is very imperfect; and has been mostly incorporated into Castell's Heptaglotton. Among the later philologians who have devoted themselves to the Shemitish languages, only a few have exhibited satisfactory proofs of any fundaVOL. III. No. 9.

5