Recall Fodor’s conclusion: observation is laden by theory but, unlike Hanson’s and Churchland’s accounts, not by just any theory that an observer holds. The theories that shape observation are limited to those that govern the mental module responsible for perception.
Here is an example. When we listen to speech, what we hear comprises sounds or pressure waves in the air which we manage to hear as words, sentences and even as meanings. How does that come about? Well, cognitive psychologists think that there is a mental module (a relatively isolated part of the mind/brain) responsible for ‘parsing’ the sounds and working out – from the sounds – which are the nouns, which the verbs etc and thus working out what the sentence is and hence what it means.
Now if this is so, this ‘observation’ or hearing will have to be laden with theory because what makes some sounds correspond to nouns isn’t simply there in their pitch. (It is not that all nouns are said in a squeaky voice.) So the parser mental mechanism will have to have a theory of language / grammar built into it. It hears sounds as verbs and nouns etc because it is structured by a theory of how language works.
OK, that much sounds like the theory dependence of observation again. It sounds like Hanson and Churchland. But although Fodor concedes that a parser has a theory inbuilt, and thus that its observations are theory laden, he thinks that it is not open to just any theory the person holds. So he disagrees with (Hanson and) Churchland that ANY of the theories a person holds can affect her observations. Rather, the mental mechanisms concerned with hearing and decoding speech are relatively isolated from the rest of the mind. Whilst observation is affected by theory, the theories involved are fixed and limited to the firmware of the mental mechanisms of hearing and seeing.
To make that point, he gives the example of suddenly talking about Pooh Bear and Piglet in the middle of a philosophy lecture. If Churchland were right, and all / any theory influenced perception, then our hearing would be affected by the knowledge that talking about Pooh and Piglet in a philosophy lecture is very, very rare. And so if Fodor (or I) suddenly did that, then you would not be able to ‘hear’ those sounds as about Pooh and Piglet (because your broader theoretical view is that that will not happen, and that theory itself would directly affect your hearing / observation). But that is not true. Actually, the sounds carry on sounding to be about Pooh and Piglet even if that is very odd. In other words, the theory that informs your sentence parsing mental mechanism is cut off from the broader theory about whether anyone talks about Pooh and Piglet in philosophy lectures.
So there is a limit to the theory that can affect perception / hearing.