Relative who and the actuation problem

This paper presents an empirical analysis of the beginning of a linguistic change in the relativiser system in one variety of English and attempts an answer to the hitherto unsolved actuation problem: why does a change begin in a language at a particular time and place but not in the same language at other times or in other places? We show that although relative who occurs with the same frequency and the same grammatical constraints in two related varieties of London English, a topic marking feature for who has developed in only one of these varieties. We offer an explanation that combines sociolinguistic and formal approaches to the analysis of variation, arguing that this combination allows a better understanding of the mechanism for the actuation of a grammatical change.