Discourse marker 'like': A unified account

The English 'like' is a case of synchronic polysemy that features adjectival, nomi­nal, preposition, conjunction, and discourse marking uses. The latter constitute a mosaic of sub-functions attracting a variety of interpretations by linguists, in­cluding the functions of approximative, non-contrastive focuser, hedge, quotative, and discourse processor. While research tends to accept the quotative fun­ction, it is divided on the question of the remaining discourse functions dis­played by the marker. The examination of polysemous 'like' will begin with its diachronic evolution and its sociolinguistic distribution, then proceed to a discus­sion of previous approaches to the discourse marker function, then attempt to reconcile these within an extended grammaticalization framework, arguing that the different functions attributed to it are implicatures of the central concept of similarity/comparison, which is residual in its current use in the language