25.	See Kaye (MSb) where I argue that the most plausible hypothesis is that some of our thought occurs in inner tokens of natural language sentences. [Harman has long maintained this view as well.]<p>