Philosophy

Matthew Babb offers a strikingly elegant argument for, and explanation of, the essential indexicality of intentional argument. His two key thoughts are that intentional action always involves intentions, and intentions are essentially indexical. In particular, every intention is indexically about the agent whose intention it is, i.e. de se. In this paper, I set out two models on which at least some intentions are not de se – they are impersonal – and I show that these models are compatible with the data Babb points to. I also set out some more data that an account of essential indexical cases ought to be responsive to. Its span suggests that the claim that all intentions are de se, even if true, cannot be what explains what is going on in essential indexical cases involving intention.

The notion of immunity to error through misidentification (IEM) has played a central role in discussions of first-person thought. It seems to be a way of making precise the idea of thinking about oneself 'as subject'. Asking whether bodily first-person judgments (e.g. 'My legs are crossed') can be IEM is a way of asking whether one can think about oneself simultaneously as a subject and as a bodily thing. The majority view is that one cannot. I rebut that view, arguing that on all the notions of IEM that have so far been successfully defined, bodily first-person judgments can be IEM.

Are non-indexical action rationalizations necessarily incomplete, because of a missing indexical component? Bermúdez ('Yes, indexicals really are essential', Analysis 2017) argues that they are. Two things make the argument unpersuasive. First, it assumes that all action rationalizations involve attitudes that are about the agent. Second, it assumes that the attitudes expressible using 'I' are themselves indexical. Each is an assumption that believers in complete but non-indexical action rationalizations can and do reject. Surprisingly though, a more effective argument can be obtained by switching focus from indexical attitudes about agents to indexical attitudes about times. The debate about the de se and the de nunc have much less in common than is standardly assumed. It is the de nunc not the de se which provides the knock-down case for essential indexicality.

John Perry offers an unusually substantive, and initially plausible, account of the conceptual role of first-person thought. This paper critiques Perry's account, particularly in what it says about action explanation, and offers a partial alternative. It also identifies three high-level assumptions about what accounts of conceptual roles should look like that plausibly explain why Perry's account goes off track in the ways that it does – this is the top-down distortion of the title. Identifying and arguing against the three assumptions helps point in the direction of a better account.

Why are de se mental states essential? What exactly is their de se-ness needed to do? I argue that it is needed to fend off accidentalness. If certain beliefs – e.g. nociceptive or proprioceptive or introspective beliefs – were not de se, then any truth they achieved would be too accidental for the subject to count as knowing. If certain intentions – intentions that are in play whenever we intentionally do anything – were not de se, then any satisfaction they achieved would be too accidental for the subject to count as intentionally acting. How states hook onto their referent is relevant in a systematic but underexplored way to whether they non-accidentally achieve their aim – truth in the case of beliefs, satisfaction in the case of intentions. In the relevant cases, the way of hooking on to a referent needed to avoid being accidental is the way a de se state hooks on to its referent.