Published online by Cambridge University Press: 01 January 2026
Perlmutter and Postal (1977 and subsequent) argued that passives cannot passivize. Three prima facie counterexamples have come to light, found in Turkish, Lithuanian, and Sanskrit. We reexamine these three cases and demonstrate that rather than counterexemplifying Perlmutter and Postal's generalization, these confirm it. The Turkish construction is an impersonal of a passive, the Lithuanian is an evidential of a passive, and the Sanskrit is an unaccusative with an instrumental case-marked theme. We provide a syntactic analysis of both the Turkish impersonal and the Lithuanian evidential. Finally, we develop an analysis of the passive that captures the generalization that passives cannot passivize.
Thank you to the Language referees and editors, whose detailed comments led to improvements throughout. Thank you to all those who discussed (subsets) of this material with us. Thank you to George Cardona for discussion of Sanskrit data (although the analysis is our own), and thank you to our consultants. We had ten primary Turkish consultants, ranging in age from mid-twenties to early forties, four from Bitlis, and one each from Adıyaman, Bursa, Denizli, Hatay, Isparta, and Mersin. We had two additional Turkish consultants, in their thirties, from Bitlis and İzmir, whose grammar is systematically different; we discuss their grammar when relevant. We had eight Lithuanian consultants, five in their late twenties, three in their late thirties to forties, six from Vilnius and Kaunas, two from Šiauliai. Glossing follows Leipzig conventions (https://www.eva.mpg.de/lingua/resources/glossing-rules.php), with these additions: act: active, aor: aorist, cm: compound marker, ger: gerund, nact: nonactive.