Roots.—Almost all words go back to triconsonantal roots, al though the theory which refers them utimately to biconsonantal roots is not improbable ; for certain primitive words, like dm, "blood," are biconsonantal, and two consonants seem to run through many common roots, like KS through those denoting "to cut" or "strike." But the language, as now known, has passed this stage, and the parent speech, if ever there was one, is irretrievably lost. From the simple root diverse parts of speech are formed by modifying the vowels or by adding prefixes or affixes ; thus from the root qdsh are formed the Bab. qadishtu, "hierodule"; Heb. qodhesh, "holiness," qiidhosh, "holy," and miqdash, "sanctuary"; Mishn. Heb. qiddfish, "betrothal"; Syr. mqaddshfithil, "sanctifi cation," and qadhisherith, "holily"; Arab. qaddis, "priest"; Eth. qedsencl, "holiness" and so on. This facility outweighed the rarity of compound substantives; but compound propositions and ad verbs became ever more frequent.
Genders and Plurals.—The only genders are masculine and feminine, which sometimes serves for the neuter ; e.g., Heb. Whelk (fem. sing.), "good" = "goodness" and Heb. nekhohoth (fem. plur.), "honest things" "honesty." The cases took termi nations (sing. nom. -u, acc. -a, gen. -i; du. nom. acc.-gen. -ai or -e; pl. nom. acc.-gen.-I). Classical Babylonian and Arabic exhibit the later languages confused them. Thus Hebrew, Aramaic and modern Arabic use -1 (-Im or -in) and Assyrian -ii (-iini) as the universal plural ending ; Hebrew and Ethiopic retain traces of the singular endings, but Aramaic and modern Arabic have discarded them. To the singular endings Arabic added -n, when the noun was indeterminate, and Babylonian -m, of which the force is obscure; to the plural Aramaic, Arabic and sometimes Ethiopic added -n, Hebrew and Phoenician -m. The feminine singular ending was -(a)t(h), which became -ah in Hebrew and -ah in modern Arabic, and the plural ending or, in Hebrew and Phoenician, -oth; the case-endings, when retained, followed this -t. There are traces of a primitive plural, formed (as in Sumerian) by reduplication (cf. Ass.-Bab. mdmi beside; me from mit, "water"), to which the Hebrews through misunderstanding some times added the regular -oth (cf. Heb. piphiyoth beside; piyoth from peh, "mouth"). The remarkable "inner plurals," character istic of Arabic and less so of Ethiopic, are formed by modifying the vowels and sometimes adding preformative or afformative syllables; e.g., Arab. 'awldd from walad and Eth. welfid from weled, "boy." The attempt to take a collective noun like zakhiir, "male(s)" as the "inner plural" of the adjective zakhdr, "male," in Hebrew is hardly justified.
The States.—Determination and indetermination were ex pressed by the "states." Primarily the "absolute state," which took no termination, denoted the indeterminate, the "emphatic state" the determinate noun, as in Bibl.-Aram. helem (absol.), "a dream," but Pima (emph.), "the dream"; in Ass.-Bab. the abso
lute state was practically confined to the predicate, as in sinnishtum (emph.) shi iil ashshat (absol.), "that woman (is) not a wife." In time these forms too were confused or dropped (cf. classical Arab. malikun, "a king," and al-meliku, "the king," with modern Arab. malik, "a king," and al-malik, "the king"). Thirdly, the "construct state," used before the genitive case, was the lightest form of the noun, dropping the afformative -m or -n (and often the singular case-ending in Ass.-Bab.) and dispensing (if there was one) with the article (cf. Ass.-Bab. ashshat awelim, "the wife of the man"). This construction runs through the whole group, but various peri phrases with prepositions or words denoting "owner" or "property" and so on are not uncommon, especially in Assyrian, Syriac and modern Arabic. Only Hebrew, Nabataean and Arabic had a defini tive article (viz., ha-, 1- and al- respectively) ; Ass.-Bab., like late Aramaic, did not feel the need, early Aramaic used the emphatic ending 41, and the south-Arabian dialects suffixed -rt for the pur pose ; Ethiopic did not determine nouns, while Abyssinian used the suffix "his" (cf. be'si, "man," and be'sihn, "his man" = "the man"). Similarly the modern Aramaic of the Tiiru-l-tAbdin has met the difficulty by coining the forms i-, du, "the," from the demonstrative pronoun.
Numerals.—The numerals exhibit one peculiarity not yet satis factorily explained but characteristic of the whole group : mascu line nouns take feminine numerals and vice versa from three to ten.
Pronouns.—Pronouns are singularly numerous. From the pri mary element ha many, containing usually the elements d(h), z, k, 1 or m (for the remoter object) and n (for the nearer object) are built up; but other varieties occur, especially in the later languages. The personal pronouns, which are less numerous, rest on a common scheme : k or sometimes t for the first, t for the second, and h (or sh in Ass.-Bab.) for the third person. To express a possessive pronoun or the pronominal object, elements of these pronouns are suffixed to the noun or verb; e.g., Ass.-Bab. shumshu =Hebr. shemo (for shemehiC), "his name"; or Syr. qatleh=Arab. qatalahu, "he killed him." These suffixes never denote the indirect object, except in Ethiopic and very rarely in Hebrew; only early Babylonian uses a special form in -m for the dative (e.g., iddinu shu, "they gave it," but iddinushum, "they gave to him"). As a rel ative pronoun either the word used to introduce the genitive case she/ in Ass.-Bab. and zi (di) in Aramaic, or an originally demon strative pronoun, 'alladhi (cf. Heb. hallazeh "this"), in Arabic and za (cf. Eth. ze, "this") in Ethiopic, is used. Hebrew employed 'fisher, the construct state of an old noun, meaning "the place of," "where" (cf. Ass.-Bab. ashru, "place," and ashar, "where") for the purpose ; but she- was used dialectically and in late Hebrew.