22 nd International Conference on Historical Linguistics Naples, 26-31 July 2015 1 Where do antipassive constructions come from? A study in diachronic typology Andrea Sansò Università dell’Insubria, Como andrea.sanso@uninsubria.it *** 1. Introduction: Aim, methodology and sample - The sources of antipassives (henceforth AP[s]) have attracted less attention than the sources of pas- sives for various reasons: a) APs are rarer than passives cross-linguistically; WALS data: 48 languages out of 194 have an antipassive (24.74%; Polinsky 2013), 162 out of 373 have a passive construction (43.43%; Siewierska 2013); b) there is generally no sufficient historical documentation to reconstruct the diachrony of gram- maticalized APs; c) there are constructions performing the same functions as APs that are usually labelled with a different terminology in grammars (e.g. depatientive, detransitive, deobjective, unspecified object etc.), especially in languages with non-ergative alignment. These constructions are gen- erally not taken into account when dealing with the diachrony of AP constructions, although they might provide insights into the diachronic pathways leading to their emergence. Definition: AP constructions are broadly defined as formally intransitive constructions involving verbs that also occur in a transitive construction, in which the agent of the transitive construction is encoded as the unique argument, while the patient is either encoded as an oblique, or suppressed.  APs in which the object is incorporated and APs with no morphological changes with respect to their active counterpart (e.g. cases of patient lability) are not considered for the present purposes. - Polysemy. Across languages, AP markers may be polysemous, with functions other than patient demotion/suppression (mainly aspectual functions, e.g. the marking of verbal pluractionality and habituality; also reciprocal actions, etc.) or fully specialized/dedicated. - Focus vs. absolutive APs. APs may also have a crucial structural function in the languages in which syntactic operations such as focalization, relativization or questioning are not accessible to agents of transitive constructions: in these languages, they provide a strategy to overcome such a constraint by making the agent the unique argument of an intransitive construction. In the Mayanist tradition, AP constructions in which the patient is irrelevant or unimportant are labelled absolutive APs, whereas APs in which the agent is promoted or focused are labelled agent promoting APs or focus APs (Smith-Stark 1978; Dayley 1983: 15-16).