,טמן , dabyr=1). The syllabary had no special character for dha. In ictil, t replaces Hebrew, as in Ethiopic. The Assyrian avoided the use of diphthongs: au is very rare; perhaps the foreign name Khauran is the only certain example of it. Ai and ya are much more common. The Gentile termination is ai, e.g. Madai, "the Medes.' Ai has a tendency to become ya or yā;1 thus ayāsi (VN) is more usually yasi; aibut, "enemies," also appears as yābut; yanu or yānu. More frequently ay or ya passed either into ah (bāhtubiyatu) or i; while in proper names an initial Hebrew was always ya (e.g. Yahua, Yahukhazi), in roots it was more generally i (e.g. imnu = D', irad=77", isibu= , isara). Even when answering to N, ai became i; e.g. inu=N. To prevent a compound vowel, hemza was largely employed, as in abu'a for abu-ya, "my father." As in modern Arabic, hemza tended more and more to be lost: in the Babylonian period it is very generally replaced by a long vowel so even utut for utuhut. 1. As in all ancient Semitic alphabets, 'a was a consonant, a soft breathing, namely, followed by the vowel a. This will explain how it is that ai is represented by 'a+'a. The second breathing here passed into y, so that we have 'aya; and hence ai. 2. H is another instance of the ambiguity arising from the employment of a foreign alphabet. It stood for h, ah, and hi. More usually the value is ah. 3. In the Babylonian inscriptions the m final very often appears as a separate character, implying that the mimmation was more strongly pronounced in Babylonia than in Assyria. The interchange of and in the cognate languages argues the weaker and later pronunciation of as v. Assyrian does not exhibit any interchange of band m. B reduplicated, however, may be changed into mb, e.g. innambu, "he is proclaimed," for innabbu, just as we find ambuba for 1, 'Iepoμßdaλ for 27, σaμBúên This is properly N; e.g. yarru=' for ; and conversely from ovupwvía. The change is an Aramaising one, and therefore exceptional in Assyrian: more frequently in (mercantile) contract tablets of late date. There is no trace of aspiration in Assyrian in the letters b, g, d, c, p, t. In Hebrew also the dagesh lene would be of late introduction, caused by Aramaic influence, as the alphabet, like the cuneiform syllabary, uses but one character for both sounds. So, too, in Arabic and Æthiopic. Equally unknown to Assyrian are the sounds elaborated by Arabic ¿, ¿, ė, b, ^, (*), and; or (as in modern Aramaic) the ƒ of Arabic and Æthiopic. The soft pronunciation of gimel, again, is not found. The accent, as in Arabic, is thrown back as much as possible. Without doubt, this was also the usage of ancient Hebrew (as is shown by the segholates) before the necessities of a rhythmic intonation of the Old Testament changed the accent. The accent is upon the antepenult, unless the penult has a long vowel or is a closed syllable. The accent is often indicated by the incorrect insertion of a long vowel or a double letter. Besides accent, Assyrian observed the laws of quantity. A long vowel was according to rule expressed, though in many cases omitted (as in the case of the double letters). In the nomina verbi a short vowel in the second syllable was generally dropped before the caseendings. The accent and the quantity seem to have coincided, as in Arabic, whenever a word possessed a long syllable not further back than the antepenult or not in the last syllable. There was a tendency to shorten vowels and words in the later period; thus the Babylonian inscriptions give us labri, for which the Assyrian is always labiru ("old"). When a word consisted of three short syllables, the second vowel was generally dropped, making the first a closed syllable long by position; thus măliců becomes malcu. The enclitic threw back the accent upon the preceding syllable, even though this had a long syllable before it; e.g. illicuniv-va (for illicūní-va), ikhdu'uninni (for ikhdhuni-ni). The doubling of a consonant was frequently disregarded even in pael1-sometimes it was replaced by a long vowel, more often by the accent merely, as in l-mu for lim-mu in contract tablets. THE PRONOUNS. The personal pronouns in the Semitic languages, as in the Aryan, are formative elements of the verb, and therefore must be considered first. SINGULAR. I, me anacu; yāti, yati, yātima. Thou, thee (masculine) = atta; (feminine) atti; cāta (cāti). We, us=[anakhni]. PLURAL. You (masculine)=attunu; (feminine) [attina]. They, them (masculine) = sunu, sun, sunutu; (feminine) sina, sin, sinatu. Attina has not been found, but analogy would lead us to this form. Anakhnu or anakhni, Dr. Oppert's conjecture, is probably right. The word is met with only in a mutilated part of the Behistun inscription (1. 3), where Sir H. Rawlinson's cast reads doubtfully a-ga-ni. As the suffix of the noun is -ni, the form anakhni is to be preferred, u being weakened to i through a false analogy of the plural termination. 1 So in Hebrew, Æthiopic, etc. Traces are Anacu is Hebrew, Phoenician N, for which in the other dialects we have only ana, ani, or eno. found in the Ethiopic tense-ending -cu, Mahri -k (Arabic and Hebrew tu and ti). The plural in all the dialects is manifestly formed from it, c becoming kh. In Coptic (and Old Egyptian) anok (and nuk)="I," anen="we"; so in Berber nekki="I," nekni="we." The relation of these subSemitic dialects to the Semitic family is very questionable. Vulgar Assyrian used anacu, in the place of the suffix pronoun, after a preposition, e.g. assu anacu, "as regards myself" (S.H.A. 190). The Arabic and Æthiopic ana, Hebrew ani, point to another form of the pronoun in ya. This has lost the final vowel in Hebrew and the initial vowel in the other two languages. It is the form that appears as the suffixed pronoun in Assyrian ya, later i and a, in Hebrew, Arabic, and Aramaic i, in Æthiopic ya. The Assyrian alone uses this without the verbal root an preceding, substituting for the latter the abstract termination tu, ti, as in sunuti by the side of sunu, or ristu, "chief," from ris, "head" (compare Ethiopic we'tu, "he," and ye'ti, "she"). Yati is often shortened to yati, just as in Arabic 'anā is used by the poets as a word of two short syllables. Yatima, "me here" (e.g. cima yātima, "like me here"), has the demonstrative ma added (as in suma), for which see below. only at the beginning of a sentence, but we find also ikbi yati, "he told me." I have not found it, except in Babylonian inscriptions, and those of the later Assyrian empire (after Sennacherib). Yati is not to be confounded with yāsi or aisi, "myself." This is yasu (), "man," (used for "self," Yati is for the most part used compare 1 Kings xx. 20, etc.), with the pronoun-suffix of the first person added (yās-i). Ya'a (S.H.A., 37, 9) is irregularly lengthened from yā, like ma'a for ma. The survival of the old word for the first personal pronoun in Assyrian is parallel to the existence of as third personal pronoun in the Phoenician-a form pre-supposed by the third person of the verb. In the second person, again, Assyrian agrees with Hebrew in assimilating the nasal to the dental, while the other dialects. have anta and ant. The Coptic ntok and Berber prefixed pronoun ewent have been compared. The interchange of guttural and dental already noticed appears in the Ethiopic tense-ending ca, ci (Mahri -k, -sh). Hebrew has in most instances shortened the feminine to att', just as Aramaic has contracted the masculine. In the plural, antumu has become in Assyrian attunu, like Aramaic antûn and the feminine plural in all the dialects. Assyrian and Æthiopic alone preserve the case-ending of the masculine, though it is found. also in Arabic poetry. Like sunuti and yāti, attunu is used in vulgar Assyrian as an accusative after the verb in place of the suffixed pronoun (e.g. altapra attunu, "I sent to you," where the preposition is ungrammatically omitted). Cāta (in the accusative) is employed for the sake of emphasis after a preceding verbal-suffix ca, which is changed into a separate pronoun by the abstract termination tu, ta (e.g. S.H.A., 180, usamkhar-ca cāta, "I cause thee, even thee, to be present," as tributary). The verbal root with which the pronouns of the first and second persons are compounded is regarded by Dr. Hincks as N, "adesse," whence the preposition ana. I should prefer און |