Here several languages have each independently lost something which the tongues most akin to them have preserved. Similarly, the strengthening of the pronoun of the third person by -t or 4/4 was customary only in Ethiopic, Sabaean and Phoenician, though perhaps found also in certain Arabic particles. Aramaic alone had no reflexive with prefixed n- and Hebrew no causative theme with prefixed sha-. Several languages originally had a passive formed by internal vowel-change ; Arabic alone exhibits the complete sys tem; there are some traces of it in Hebrew and Aramaic, very few in the modern Arabic, and virtually none in Ass.-Bab., or, when it became literary, in Ethiopic. Again, phonetic resemblances occasionally indicated no kinship but were due to later growth. For instance, the masculine plural ending was in Hebrew -Im but in Aramaic, as in modern Arabic, -in; yet Aramaic originally had -m, whereas classical Arabic had a after the -n (in -Cola and -Ina), where the change of m into n between two vowels is very im probable. These endings, therefore, may originally have been dis tinct. Great caution, too, must be exercised in drawing conclusions from points of agreement between the vocabularies of the various languages. For instance, Assyrian, Hebrew and Ethiopic had the same word for many objects which the other Semites called by different names; and the same might be said of Hebrew and Sabaean. Finally Moabite, which was otherwise indistinguishable from Hebrew (although, if the vocalization were known, its pro nunciation might be found to have differed considerably), shared one striking peculiarity with Ass.-Bab. and the southern group, viz., a derived verbal theme formed by infixing t (cf. Moab. qthm, "fought," and Arab. iltaljam, "was hotly contested," with Hebr. nillzam, "fought"). But to build any theory on such facts is haz ardous, since the words or forms so cited are either found, though with a changed nuance, in at least one cognate language or actually occur with the same sense exceptionally or in poetical or archaic texts.
Certainly Arabic (with Sabaean, Mahri and Socotri) and Ethi opic were closely related and formed a distinct group over against the Hebraeo-Phoenician, Aramaic and Ass.-Bab. groups. Only in these southern languages are found the peculiar "inner plurals," consisting in the employment of certain forms denoting abstracts for the plural. They agree in inserting a between the first two consonants to enable a verb to take a direct instead of an indirect object (e.g., Arab. kataba, "corresponded with," and takdtaba, "corresponded together," from kataba, "wrote"), and in using a between the last two consonants of the perfect in every active theme (cf. Eth.'asdaqa with Hebr. hisdiq,"justified"). Yet Hebrew has isolated examples of both peculiarities (e.g., ashen, "slan dered," with '6 for a, and qiddash; "sanctified," if the vocalization is correct). This is not disproved by such exceptions nor violated by the fact that the aspirated dentals th, dh, z in Arabic are replaced in Ethiopic, as in Hebrew and Ass.-Bab., by the pure sibilants s (or sh in the latter language), z, s, but in Aramaic by the simple dentals t, d, t, which seem closer to the Arabic sounds. Even after the separation of the northern and southern groups all must have possessed these sounds, like Arabic, but have afterwards simpli fied them in one direction or another; hence accidental similarities resulted. Ethiopic kept 4, the most peculiar of them, distinct from s, whereas Aramaic confused it with the guttural (`ayin) and Ass.-Bab. and Hebrew with s; all therefore once possessed this
as a distinct sound. The sibilant s (sin) occurred only in Hebrew, Phoenician and old Aramaic ; it must have closely resembled sh (shin) with which Ass.-Bab. confused it, since it was represented by the same sign, and in later times was changed into an ordinary 5. The division of the Semitic languages, therefore, into northern and southern is justifiable. It must, however, be remembered that such a division was not an instantaneous occurrence but that even after it, intercourse was maintained between tribes speaking kin dred dialects and that intermediate dialects between the groups once existed. Further, Hebraeo-Phoenician and indeed Aramaic stand apart from Ass.-Bab. in more things than in the script; e.g., in the preservation of the gutturals and the absence of verbal themes with -tan- infixed. The simplest classification is perhaps the following; A. Ass.-Bab. . B. (i.) Hebraeo-Phoenician and Aramaic ; (ii.) Arabic and Ethiopic.
Their connection with the Hamitic languages is disputed. They clearly, however, influenced Egyptian, which has pronouns (cf.
Egt. 'nky with Hebr. "I") and pronominal affixes (cf. Egt. -ty with Hebr. -td, "thou," and Egt. -ty with Ass.-Bab. -at, "she") resembling the Semitic forms, a pseudo-participle like the Ac cadian "permansive," certain verbal stems analogous to the Semitic themes, and many similar words (cf. Egt. myw and Ass.-Bab.
"water"). The Berber languages show similar but less frequent resemblances. Similarities, however, between Semitic and Indo European words are either superficial (cf. Skt. shash and Hebr. shesh, "six," whose root proves them unconnected) or due to borrowing (cf. Gk. xpvaos, which was borrowed from the Phoen.
"gold").
The script in all except Ass.-Bab. is derived from Phoenician and the writing runs from right to left in all except Ass.-Bab. and Ethiopic ; only in a few Sabaean and early Ethiopic inscriptions is it alternate. Accadian alone had a script, borrowed from Sumerian, which primitively represented vowels, while Ethiopic very soon added them to an originally consonantal script ; in both the signs represent syllables and not letters. In the others vowel signs, which were never regularly used, were added above and below the line, first by the Syrians in imitation of the Greek vow els and afterwards by the Hebrews and Arabs. Special numerical signs were invented or borrowed (for the so-called "Arabic numerals" are Indian) by all except the later Jews, who used letters of the alphabet for the purpose. No syllable might begin with a vowel, a breathing ('aleph or hamzah), written ', being used to support an initial vowel; there were five gutturals of varying hardness, viz., h, h,t(cayin) and k (gayin). Arabic alone had all six gutturals; Mandaean confused all and Ass.-Bab. merged all except It, and sometimes and Ii in vowels. There were also 12 sibilants and dentals as well as the ordinary consonants, which interchanged in the various languages according to definite rules; there were usually only two labials, b and p or f , although Hebrew had an aspirated b and p and Ethiopic two forms of p. The stress lay on the consonants, perhaps most of all in Syriac, and the vowels (originally only a, i and u) might be freely modified, especially through proximity to gutturals (cf. Ass.-Bab. sibi and Heb. shebhat, "seven").