Yoshikawa introduced the basic classification system of the Sumerian verb according to their hamtu/maru stem in a 1968 paper. There he proposed a so-called affixation class, the maru class of whose verbs was derived simply by suffixing a particle -e to the hamtu base. This became known as the maru -e. Yoshikawa saw this -e embedded in the so-called '-ed' morpheme, arguing that it was actually a compound /-e/ + /-d/. As a result, analyzed the pronomial maru suffixes (-en, -e, -enden, -enzen, and -ene) into the maru -e plus reduced forms (i.e. -e + -n, -0, -nden, -nzen -ne).

Edzard, on the other hand, argued for a different analysis, in which the maru -e was actually part of the pronomial suffixes used in the maru conjugation. In this view, Yoshikawa's affixation class actually consisted of verbs whose stem did not change between hamtu and maru (unveranderung klasse).

Thompson argued along the lines of Edzard. In objecting to Yoshikawa's analysis on phonological grounds, she stated that the 1. pl and 2. pl suffixes (-nden, -nzen) have a consonant combination uncharacteristic of Sumerian phonology.

Attinger (1993) cited examples of intransitive verbs like ku4 which exhibit both ku4-ku4 and ku4-re forms denoting the maru stem. He also observes the absolutive pronomial suffixes for intransitive verbs all begin with an -e (-en, -enden, -enzen, etc). Such an example, however, may simply be an instance where there are two ways of marking maru in the same verb (Rubio).

The analysis is made more difficult by the fact that the actual form of the suffixes is somewhat obscured by the writing system. As an answer to Thompson's objection, Rubio has suggested that the maru -e could have been reinterpreted as part of the pronomial suffixes (see Rubio). Moreover Kienast has argued that the suffixes should actually be (-nge and -ze) to better reflect symmetry in the independent and possessive pronomial suffixes.

Rubio also makes an important point from the standpoint of linguistic typology which supports Yoshikawa's interpretation. Namely, it would be odd that some of the Sumerian verbs (class II, III, and IV) exhibit different hamtu and maru stems while class I verbs do not, instead relying on agreement patterns. Also, there is a hamtu/maru opposition exhibited in the non-finite forms with the suffixes -a and -ed. In non-finite forms agreement patterns are not relevant, so some stem distinguishing mechanism must be present for class I verbs (rewrite!).

Attinger, Pascal 1993 Elements de linguistique sumerienne:….

Edzard, 1971 Hamtu, Maru, und freie Reduplikation….

Thompson, Sumerian grammar.

Yoshikawa, 1968 Maru and Hamtu aspects

Yoshikawa, 1974 The Maru Conjugation in the Sumerian Verbal System

Rubio, Gonzalo. 2006 Sumerian Morphology