A while ago I wrote a little argument claiming that no version of a privileged access theory could account for disagreement about foundational questions in mereology. Practically no one believed it, but despite that I didn’t decide to write it up as anything further. I did, however, have one other thought on the matter that seems worth at least putting here.
The motivation for the idea that externalism about content and privileged access are incompatible is fairly easy to see. Intuitively, it is only plausible that we have privileged access to the intrinsic features of our occurrent mental states. Even that kind of privileged access claim may be too strong, but it seems that is the strongest we could have. It is generally implausible that I have privileged access to the extrinsic features of my occurrent states. To argue by example (i.e. fallaciously) it is plausible that I have privileged access to the fact that I am experiencing a particularly vivid shade of red now, it is implausible that I have privleged access to the fact that I am experiencing a more vivid shade of red than anyone else in my building is currently experiencing.
It is a matter of no little delicacy how to turn this consideration into an argument for the incompatibility of externalism and privileged access. As John Heil pointed out back in 1988, in the right kind of context, privileged access to the intrinsic features of my occurrent thoughts might suffice for knowledge of their wide contents. His point was simply that an externalist about content will think that the content of a higher-order thought about my own occurrent thought will have its content fixed by the environment, just as the first-order occurrent thought will have. If the first-order and higher-order thought are in the same environment (which, barring some sci-fi type fantasy, they are) and the higher-order thought is sensitive to the intrinsic features of the first-order thought, my having that higher-order thought might constitute knowledge of the wide content of the first-order thought. None of this is to say that no argument from the general inaccessibility of extrinsic features of the thought to the inaccessibility of their content will go through, just that, well, it is delicate how to get one to work.
The point that doesn’t seem to have been brought up much in the literature on this subject is that the same kind of consideration tells against privileged access on most internalist theories of content. Most people deny that we have privileged access to our unconscious beliefs and desires. And most people think that we have at least some unconscious beliefs and desires, though Freud probably overstated how many, and how dramatic, they are. And most internalists about content are holists, who think that the content of a particular thought is determined, inter alia, by its relations to other thoughts.
I think that if you hold all three of those views, and they are fairly plausible, there is at least a prima facie case that you’ll have to give up privileged access about content. For now you too think that the content of an occurrent thought is an extrinsic feature of that thought, although it may well be an intrinsic feature of your mind.
This is only a prima facie case because it isn’t obvious how to turn it into a full-blown argument. I made a few attempts to do so over the weekend and all of them ran aground on Heil-style considerations. The agents I considered didn’t know all the relations that obtained between their conscious beliefs and their unconscious beliefs that determined the content of their conscious beliefs, but still their higher-order beliefs about their conscious beliefs counted as knowledge of content, because those higher-order beliefs were also related to their subconscious beliefs in just the right way. So that’s why this is a blog entry rather than a new little draft paper, because I can’t yet see how to pull off the most important step in the argument. Anyone who has any ideas about how to do so is more than welcome to try. Do let me know how the results turn out.