AbstractThis article analyses the relation between corpus linguistics and grammaticalisation. It shows how corpora and the use of corpus-linguistic methods enrich the empirical basis of grammaticalisation theory and, in some important instances, even help it to develop its theoretical foundations. The article discusses an incipient and hither to unrecorded instance of grammaticalisation in present-day English to test the potential of the corpus-linguistic working environment for the investigation of such difficult empirical cases. It also explains the three reasons for using corpora in research on grammaticalisation.