Kouwenberg, N.J.C. Gemination in the Akkadian Verb. Studia Semitica Neerlandica 33. Assen: Van Gorcum, 1997.
Kouwenberg challenges the consensus introduced by Goetze’s influential article, namely that the D-stem is primarily a denominative stem (it is used to create new verbs from nominals). He allows that Goetze was correct to be suspect of the idea that somehow doubling of the middle radical denotes “strengthening” of the basic meaning of the verb, since such arguments were unsophisticated in his day; however, modern linguistic study seems to have confirmed that the intuition of the classical grammarians was actually correct.
As Goetze outlined, the D-stem seems to have three basic functions: intensive-plurality, causative-transitivization, and denominative. The assumption of 19th and early 20th century grammarians was that the D-stem was a derivation from the G-stem and that intensification was the primary function. This led to the causative meaning since intensity could imply that the agent made someone else perform the action (see Brockelmann’s 1908 grammar, p. 509, as well as GKC et al). Note that Poebel argued similarly, but put plurality as the primary function of the stem, suggesting that the gemination is a remnant of reduplication of the whole root and in many languages reduplication is used to express plurality.
Kouwenberg argues that there is a clear relationship between gemination and plurality (be it habituality, durativity, or plurality of the subject or object) which he labels iconic. The Semitic languages are iconic in general as meaning is built very predictably from the combination of roots and patterns. The simpler patterns tend to correspond to simpler meanings, while more complex patterns correspond to more complex meaning. Roots are extended either (1) by means of affixes or (2) by reduplication (in whole or part), vowel lengthening, or gemination.
Of the second class, reduplication obviously corresponds to plurality while vowel lengthening doesn’t seem to have any inherent meaning but is used generally to mark a contrast between two forms. Gemination seems to be somewhere in between, but in analyzing relevant Akkadian forms with gemination, Kouwenberg concludes that they overwhelmingly reflect plurality. Thus, gemination seems to be closely related to reduplication.
For instance, the function of the parras nominal form seems to be largely to serve as a plural for simple adjectives denoting dimensions, and the same holds for many purrus forms. Gemination also marks plurality in many verbal forms such as the Gtn used for iteration, frequentativity, and plurality. The D-stem of transitive verbs are also used frequently to underline plurality of the object. Lastly, the present-future iparras seems to have grammaticalized plurality into a tense form.
Goetze was correct to note the relationship between the stative and the D-stem, however Kouwenberg argues that it has nothing to do with Goetze’s three categories (durative, perfect, and passive). These are merely descriptions of the semantics of particular verbs. Rather, the D-stative purrus should be seen as an adjective used originally to refer to inherent qualities and to underline plurality.
In the Semitic languages, especially seen in Neo-Aramaic, it is common to derive new verbal forms from ajectives and infinitives. Cross-linguistically, verbs deriving from adjectives with the meaning “to be X”, “to become X”, “to make X”, or “to cause X” are the most common category. Even in English, where there is no real “causative” marker, there are examples such as “to lengthen”, “to cool”, etc.
In Akkadian, almost all common simple adjectives have a derived purrus form, and most also have a D-stem with a factitive function. The case of transitive verbs can also be related to the adjective. Interestingly, highly transitive verbs tend to have a corresponding verbal adjective much more frequently than verbs with low transitivity. The frequency of use of a D-stem with a given verb correlates with the frequency of the occurrence of its verbal adjective. Thus in both cases, the D-stem began as a denominative form derived from the adjective. As this form spread, however, the D-stem clearly became derived from the G-stem.