Abstract
That Kuhn is mistaken in drawing a sharp line between normal and revolutionary phases of science is shown by re-examining the role of models in extending theories to new phenomenal domains. In the light of this revision of the role of models, theory extension, which Kuhn includes in normal science, is shown to be continuous with theory replacement, which Kuhn includes in revolutionary science. Both involve language changes and the 'gestalt switches' associated with revolutionary science. These characteristics cannot be used to demarcate the two phases as Kuhn seems to think