Slezak, P. P. (2002). Thinking about thinking: language, thought and introspection. Language & Communication, 22, 353–373. doi:10.1016/S0271-5309(02)00012-5 http://journalpsyche.org/articles/0xc0a7.pdf
Abstract:
I do not think that the world or the sciences would ever have suggested to me any philosophical problems. What has suggested philosophical problems to me is things which other philosophers have said about the world or the sciences. (G.E. Moore, 1942, p. 14) Peter Carruthers has made a vigorous attempt to defend the admittedly unfashionable doctrine that we think ‘in’ language, despite its displacement by something like Fodor’s ‘language of thought’. The idea that we think in language has considerable intuitive persuasiveness, but I suggest that this is not the force of good argument and evidence, but a familiar kind of introspective illusion. In this regard, the question of language and thought derives a more general interest, since the illusion is independently familiar from other notorious disputes in cognitive science such as the ‘imagery debate’.
[Citing Place (1956)]  
Citing Place (1956) in context (citations start with an asterisk *):
Section 15. Conclusion: the phenomenological fallacy
Carruthers writes:
Remember that according to the cognitive conception, when a speaker utters a sentence their utterance will express a thought by constituting it, not by encoding or signaling it. (1996, p. 122)
How sentences of a natural language could conceivably ‘‘constitute’’ thought is utterly unclear because such sentences require interpretation. It is an illusion that the sentences themselves might function in cognition because it is we as theorist who are surreptitiously doing the work by using our own intelligence to compensate for the deficiencies of the theory. Pictures also appear to serve imagery by somehow ‘‘constituting it’’, but again only because we as theorists can understand or perceive them. Thus, Kosslyn (1994, p. 14) claims that pictorialism is vindicated by the discovery of retinotopic maps of distal stimuli on the brain of experimental animals. Of course, these are pictures, but not for the animal, only the theorist.
I have suggested that the very naturalness or indeed compellingness of the view we have been considering is, paradoxically, the strongest evidence against it. This irony is understood when it is seen that the compellingness of the view does not derive from its systematic scientific merits but from the seductive persuasiveness of a certain misleading picture which foists itself upon us as a consequence of certain inherent cognitive arrangements. The very possibility of introspective self-reflection encourages the mistake of thinking that the ‘‘objects’’ of our internal experiences are similar to their external vehicles. This ad hominem charge gains considerable force from the surprising ubiquity of the mistake across seemingly independent domains and times. Thus, to take just one example, it is remarkable that in a well-known article by Place (1956), first to articulate a materialist theory of mind, he writes under a heading ‘The Physiological Explanation of Introspection and the Phenomenological Fallacy’:
... for all its emotional appeal [Sherrington’s argument for dualism] depends on a fairly simple logical mistake, which is unfortunately all too frequently made by psychologists and physiologists and not infrequently in the past by the philosophers themselves. This logical mistake, which I shall refer to as the ‘phenomenological fallacy’, is the mistake of supposing that when the subject describes his experience, when he describes how things look, sound, smell, taste or feel to him, he is describing the literal properties of objects and events on a peculiar sort of internal cinema or television screen ... (1956, p. 37)
Significantly, Place observes in a footnote that his case ‘‘often seemed a lost cause’’ and it seems likely that he is reflecting on the persistent difficulty of arguing against the strong contrary intuitions derived from introspection.
I have argued that if imagery cannot plausibly be in pictures, then thoughts cannot be in language either, and for the same reasons.