12 Stratton L. Ladewig In fact, Fred W. Householder summarizes this passage by saying “this is hardly a satisfying account of the middle”30. The morphology of these ‘middle’ verbs’ can indicate either activity or passivity. He illustrates his point with ἐλουσάμην, ἐποιησάμην, and ἐτριψάμην. Each of these is parsed according to its morphology as aorist, middle, indicative, first person, singular. Nevertheless, Apollonius Dyscolus informs us that these verbs can function as actives or passives31. This is evident in his comparison and contrast in the verbal sample used. He says that ἔτριψα and ἔλουσα are different in meaning from ἐτριψάμην and ἐλουσάμην, which seems to indicate that he sees these middle forms as related to the passive. He stops short of saying that they were passive because these words do not actually carry a passive function. In contrast, ἐποιησάμην (and προηκάμην32) are described as being similar to ἐποίησα (and προῆκα), which seems to indicate that these middle verbal forms connote an active function33. The remainder of the passage is especially difficult. He explains that those who consider the interchange of voice functions as valid are ignorant because such an exchange would be unnatural and ungram- matical. Householder explains that there are only two voice possibilities for Apollonius Dyscolus: active and passive. The middle forms do not carry a distinct middle meaning34. His examples seem to illustrate that middle forms for which a passive sense would be expected are found with a transitive active35. In conclusion, although Apollonius Dyscolus’s first description of the middle voice is cryptic, one thing is clear: there is room within his understanding of voice for a discrepancy to exist between the form and function of the verb with reference to voice. It should also be observed that this understanding of voice corresponds closely to the definition of diathesis by Dionysius Thrax. However, the situation in Dionysius Thrax is the opposite of the one found here in Apollonius Dyscolus. In Dionysius Thrax, the verbs have active or passive morphology but middle function. In Apollonius Dyscolus, the verbs have middle morphology but 30 Householder, Syntax of Apollonius Dyscolus, 166. 31 More specifically, Apollonius Dyscolus may have in mind the concepts of transitive and intransitive, respectively (Householder, Syntax of Apollonius Dyscolus, 165). 32 It seems odd that Apollonius Dyscolus would introduce another example into the discussion at this point. This further illustrates the cryptic nature of his description of diathesis here. 33 Signes-Codoñer, “Definitions of Middle Voice”, 20. 34 Householder, Syntax of Apollonius Dyscolus, 165. 35 Householder, Syntax of Apollonius Dyscolus, 166.