Passives, their PIMPs, and implicit control

This talk presents a new generalisation about implicit control in the context of passives, building on Pitteroff & Schäfer (2019): iff a type of passive can be construed as an impersonal passive with unergative verbs, then it also allows implicit control. The revision of Pitteroff & Schäfer’s original generalisation is motivated by the distribution of implicit control in languages such as Croatian, which have more than one passive, and by novel data showing a contrast between different types of passive with respect to the availability of Wh-extraction in implicit control configurations. I propose that an important role in deriving the two-way split (between passives that do and those that do not allow the aforementioned configurations) is played by the featural makeup of the passive implicit argument (PIMP), which may vary both among and within languages. The proposal thus sides with approaches assuming that PIMPs have a place in the syntactic component (Landau 2010, Bhatt & Pancheva 2017, Michelioudakis 2021) and lends support to the view that passives may come in different guises (e.g., Legate 2014, Alexiadou et al. 2015, Legate et al. 2020).