This paper assumes that in order to explain rather than describe language change, historical linguists should not only consider what happens diachronically at the language output level but also, crucially, what speaker-listeners do at the processing level. The reason for this is that the structure of the language is shaped by the properties of the neurolinguistic mechanism underlying both language use and language learning. It will be argued that analogy as an important principle in grammar formation is the main mechanism in grammaticalization and in change in general when looked at from a processing point of view. The paper discusses the workings of analogy in a number of cases in the history of English which have traditionally been interp...