Multi-variate coding for possession: methodology and preliminary results name of the journal/book: De Gruyter Mouton
Natalia Chousou-Polydouri, David Inman, Thomas Huber and Balthasar Bickel publish the first paper presenting a new methodology to code nominal possession for a typological database. Nominal possession is a feature that on multiple levels shows great diversity among the languages of the world, making it an awkward hurdle in cross-linguistic comparisons in the past. Both within and across languages one finds varying strategies to mark how one noun is possessed by another. For instance, some noun classes (often body parts and kinship terms) might be impossible to un-possess or require a special construction to denote non-possession, and other noun classes (often animals and natural elements) might be impossible to possess or require special constructions in some languages.
The methodology presented in this paper brings a high-resolution representation of differential possession, allowing different dimensions of possession to be studied: Semantics (what semantic classes tend to group together in possession constructions), valence (what classes can/must have a possessor or not), and construction shapes are all brought together in a multi-variate coding scheme.