An attempt to show that of the three claims that follow from W.D. Marslen-Wilson's and L.K. Tyler's (1980: 'The temporal structure of spoken language understanding') view on language processing, viz. (i) that language interpretation is on-line interactive, (ii) leads to only one uniform and (partially) non- linguistic representation, that (iii) results from the application of both linguistic and non-linguistic knowledge, (iii) is correct, (i) needs clarification and modification, and (ii) is wrong.