This book provides the first empirical study of the history and spread of mediopassive constructions. It investigates the productivity of the pattern, the spread of the construction in Modern English, and looks into text type-specific preferences for the construction. On a more abstract level, it combines the corpus-based description of mediopassive constructions with cognitive linguistic models, drawing largely on notions such as ‘prototype’, ‘family resemblances’, ‘patch’ and ‘construction’. The theoretical modelling is largely based on data from real texts. These come from publicly available machine-readable corpora, text-databases and a single-register ‘corpus’ (American mail-order catalogues). The study combines the corpus-based approach with cognitive theories and is therefore of interest to both empirical and theoretical linguists.