Philosophy, Psychiatry, & Psychology

We seem to have a special, seemingly direct, relationship to our own thoughts that we do not have to the thoughts of others; I can become aware of my thoughts in a way that I cannot become aware of yours: through introspection. Those who have delusions of thought-insertion, however, claim not only to be aware of another's thoughts, but to have another's thoughts in their own mind. These thoughts, of course, cannot actually be someone else's thoughts. However, if we take those who have this delusion at their word, it certainly seems to them that these thoughts belong to another. Although introspection is no longer generally thought to be absolutely error free, this misinterpretation seems quite a strange mistake to make. Surely, we cannot misidentify who has the thoughts we are aware of through introspection! Yet the existence of thought-insertion delusions shows that this kind of misidentification is not just possible—it actually happens. So thought-insertion delusions seem to threaten the principle of immunity to error through misidentification (IEM) of our introspective mental state attributions (Campbell, 1999a). However, in "Pathologies of Thought and First-Person Authority," Michael Young argues that thought-insertion delusions are not counterexamples to this immunity principle, once we properly formulate it.

One might gloss IEM by saying things like, "we cannot misidentify who has the thoughts that we are introspectively aware of," just as I did above. And thought-insertion delusions, if we interpret these reports straightforwardly, certainly are counterexamples to this. Yet, as Young points out, they are not counterexamples to the following principle:

Immunity Principle (IP): any agent who thinks a thought* (where thought*s are thoughts, based on introspection, attributing a mental state in the present tense) of the form 'I am F' cannot be right on the basis of the perspective grounding this thought that something is F but wrong about whether 'I am F.'

If IEM for thought*s is correctly characterized by IP, then Campbell has not provided a counterexample, for the principle was only ever meant to apply to first-personal ascriptions of mental states (which Young calls I-thought*s)—those judgments wherein you ascribe the mental state to yourself through introspection. When suffering thought-insertion delusions, patients ascribe the mental state in question to someone else. Moreover, Young argues, if we reformulate descriptions of thought-insertion delusions so that they are in the first person (so the patient's ascription denies that the mental state in question is theirs), we end up with a different sort of error—an error of mispredication. That is, we simply end up with the patients making a mistake about which mental state they are in (or are not in), rather than an identification error.

Although IP can be defended from counterexample in this way, I argue that IP does not show [End Page 161] us anything about introspection or the first person—which should make us wonder whether it really captures what is at stake in discussions of IEM. First, Young's defense of IP turns on features that are not peculiar to introspection or the first person—nor to any other kind of subject matter that is thought to have IEM. Second, although Young has given us a principle that actual-world thought-insertion delusions do not threaten, IP is not yet strong enough to capture the standard claim of IEM for introspective self-attribution. Without further supplementation, both IP and Young's defense of IP leave open the possibility that such immunity might be merely contingent. Furthermore, thought-insertion delusions provide prima facie evidence that IP is merely contingently true.

The general structure of Young's argument can be used to defend a multitude of versions of IEM principles—not just ones that involve first-personal introspective judgments. For example:

(BIP) Any agent who thinks a thought based on evidence concerning Bates College of the form 'Bates College is F' cannot be right on the basis of that evidence that something is F but wrong about whether Bates College is F.

Of course, someone could think that Bowdoin College is in Lewiston on the basis of...