3,417 followers
@haspelmath @MaxPapillon1 But to return to linguistic categories, if I understand your position better now, why not just use something like the Canonical or Multivariate approaches to typology which are already out there? https://t.co/yakvquPbx0