Bingol, IbrahimAltinkilic, Umran2026-05-152026-05-1520261657-463X2389-9638https://hdl.handle.net/20.500.12514/10949Traditional descriptions of voice in Kurmanji and Zazaki often foreground an active-passive opposition, and some Zazaki-oriented descriptive traditions additionally claim that Zazaki has two passive types: a canonical passive formed with the auxiliary ameyene 'come' plus an infinitive, and a second suffixal/added passive. Building on a structural diagnostic- the optional licensing of an overt agent phrase-this article argues that the alleged second passive is not passive at all but an event-centered middle construction: it can be semantically passive-like yet systematically resists agent-phrase insertion. We show that true passives are parallel in Kurmanji and Zazaki (hatin/ameyene + infinitive) and that the extra passive arises when dialectally constrained datasets (especially Southern/Cermug Za zaki, where ameyene-passives are reported to be rare) conflate passives with middles. In contrast, Northern Var to Zazaki patterns closely with Kurmanji, favoring the auxiliary-passive strategy. Finally, we argue that middle alternations are constrained by verb class: only a subset of ambitransitives behaves as labile/ergative predicates that readily form middles, whereas highly general transitives such as kerdene 'do' resist the middle alternation but passivize straightforwardly. Using parallel Kurmanji-Zazaki paradigms, the paper provides a compact toolkit for separating passive from middle and offers a dialect-sensitive reanalysis of Todd [1] and Pamukçu [2].eninfo:eu-repo/semantics/closedAccessMiddle VoiceZazakiKurmanjiPassive VoiceErgativityAgent PhraseArticle