Internal Reduplication among the t-Derived Verbal Stems of Akkadian.

• Author: Testen, David

Alongside the familiar mechanisms of which Akkadian morphology avails itself to form verbal stems--affixation, ablaut, and gemination--a small but solidly documented set of verbs employs what has been described as reduplication of medial syllables. The overall effect for such verbs is that we encounter a recapitulation of the second radical, with an interposed vowel -a- (or -e- under the appropriate conditions) and gemination of the second instance (-...[C.sub.2]a[C.sub.2][C.sub.2]...-). The most securely documented of these verbs constitute what has been called the "Dtr" type, forms that would be parsed as Dt-Stem verbs were it not for the appearance of this ancillary syllable. (1) The verb utlellu 'to exceed in number' (also 'to raise oneself; to praise') happens to possess an exceptionally broad range of attested forms, allowing us to compile a fairly comprehensive set of finite and non-finite forms (see Table l). (2)

In some but not all of these verbs, researchers have identified special semantic nuances--reciprocality, iterativity, or a systematic escalation in the value or intensity of the event. (3) Alongside such Dtr verbs it is not uncommon to find straightforward Dt verbs performing similar functions--cf, e.g., from Old Babylonian Mari, Dtr uk-ta-sa-sa-ra-am '(the army) gathers' (ARM 26/1,287 no. 121:12) vs. in the preterite Dt uktassiram (ARM 6 58:17)--and in large part the various values that have been seen in the Dtr constitute specialized functional extrapolations from familiar semantic corollaries of the mediopassive marker -t-. (4) At the same time, there are also formally comparable verbs that show no obvious signs of the semantic values in question but nonetheless display syllabic reduplication--cf, e.g., gana ut-le-li-ma ina eldtu tisbi (TCL 6 51:37f.) 'come, raise yourself and sit on high' (Table 2). (5)

A comparatively well-documented instance is the morphologically convoluted geminate-root verb meaning 'to pray' (*[??]h-n-n), which displays forms that alternate between the standard Dt and the reduplicated Dtr models. As Kouwenberg observes, attested Old Babylonian shapes of this verb are consistent with either the Dtr model *y-u-htanannVn- (cf. pres./impf. li-te-ne-en-ne-en [St. Reiner p. 188:2], precative lu-te-ne-ni-in [AbB 10 111 r. 5]) or the Dt model *y-u-htannVn- (pres./impf. u-te-en-ne-en [KUB 37 72 r. 8], also Standard Babylonian prec. li-te-nin [Hem. 94: 41]); the noun utnen-u, utmn-u, utninn-u 'prayer'is presumably a lexicalized nomen actionis associated with this verb. This verb presumably constitutes the East Semitic cognate to a well-attested set of West Semitic Dt-Stem verbs (cf. Hebrew hithannen 'sought favor', and Syriac 'ethannan 'sought favor, made supplication'; cf. Arabic tahannana 'showed compassion'). Although the West Semitic forms show no sign of the reduplication seen in Akkadian, the existence of these cognates suggests that the East Semitic lexeme for "pray" is to be traced back to Common Semitic, in which case its morphological eccentricities may well be worth taking seriously. (6)

The West Semitic languages offer no convincing counterpart to the Akkadian Dtr formation, and we are probably justified in regarding the formal properties of these verbs as the product of developments specific to East Semitic. (7) The initial impression that these data give is that the "Dtr" class originated as a subcomponent within the Dt-Stem verbal lexicon and gained formal and semantic autonomy by (arbitrarily?) creating a supplementary syllable and garnering specific lexical functions. (8)

From a historical linguistic perspective, however, we should also consider the possibility that the reduplicated shapes represent, not sportive deviations from the canonical Dt-Stem, but fossils that happen to preserve details that offer us insight into the prehistory of the Dt. The synchronic morphology of the canonical Akkadian Dt-Stem as we find it described in the grammars is eminently simple: it is formed through the infixation of a -t(a)-element into the corresponding D-Stem (Table 3).

Most frequently, pervasive systematicity of this sort reflects the outcome of a relatively recent restructuring. As will be seen below, comparative data from West Semitic suggest rather strongly that Akkadian has fundamentally reconfigured its Dt-Stem morphology. There is thus a good possibility that at least certain deviations from the historical Akkadian norm that we observe here might represent earlier phases of the morphological machinery of the Dt-Stem.

The core of the Akkadian conjugational system lies in the relationship between the present/imperfective stem (-parras-, etc., incl. Dt -ptarras-) and the preterite stem (-prus-, etc., inch Dt -ptarris-); of these two, the latter typically corresponds to the West Semitic imper-fective/jussive stem. There are two overarching mechanisms in Akkadian through which the fundamental temporal/aspectual bifurcation is realized, the choice between the two being dependent on the class membership of the verb in question. The first verbal class (called here the "- u" type) consists of verbs belonging to the Basic stem (G-Stem), the N-Stems (tri- and quadriradical), the Gt-Stem, and the iterative derived stems built from these (Gtn, Ntn, NQtn), while the verbs displaying the second type (the "+ u" type) are composed of the D-Stem, the various Causative-Stems (S, SQ, SD), and their mediopassive and iterative derivatives (Dt, St, Dtn, Stn). (9) Each of these two classes possesses its own characteristic expression of the present/imperfective, viz. syllabic expansion vs. vocalic ablaut. As it happens, this bifurcation in the tense/aspect-marking mechanisms maps neatly onto a distinction in the quality of the preradical vowel: verbs of the first type above (- u) employ the preradical -a- (after the first-person singular *'- and the second person /-) or -/'- (for the first-person plural n- and third person *y-), while verbs of the second type (+ u) employ -u- (Table 4).

As we have seen above, both the Dt- and Dtr-Stems behave in conformity with the morphological rules for the + u regimen: the tense/aspect opposition for these verbs formally resides in the stem vowel (Dt pres. u-ptarras- vs. pret. u-ptarris-), displaying a systematic ablaut-relation between the present/imperfective stem-vowel -a- (or -e~) and preterite The contrast in the stem-vowel quality is often not visible in final-weak verbs like *tumammu or utlellu, but it is clear in the examples of the Dtr below, which are drawn from CAD dananu mng. 4 (Table 5).

The rules governing the formation of the present/imperfective are specific to Akkadian, but the overarching morphological split of the verbal lexicon into "- u" and "+ u" types must surely reach back to the Common Semitic stage. Given the close match in the bifurcating preradical systems of Akkadian and Literary Arabic--cf. Akk. G a-prus vs. D u-parris alongside Arabic G'-a-f'al vs. D '-u-fa"ii--a fundamental distinction between verbal classes that employed the preradical vowel *-u- and those that did not must have been part of the grammar of the common ancestor of East and West Semitic. (10)

While we may be quite confident in reconstructing the existence...