Get access

We Are Not All ‘Self-Blind’: A Defense of a Modest Introspectionism


  • This paper originated in a talk in March 2009, as a part of a Mellon foundation series on philosophy of mind and cognitive science at Brown University. Versions of it have been presented at the Institut Jean Nicod in Sept 2009, at the Dubrovnik philosophy conference, and at the University of Oslo in 2010. I want to thank my hosts at these various institutions for the occasions—particularly Katarina Samoilova at Brown—for their useful comments and advice, as well as Mark Engelbert and Max Heiber here at Maryland for lively discussions of the topic. And I'm indebted to Jonathan Adler, Ned Block, Tyler Burge, David Bennett, Peter Carruthers, Michael Devitt, Matt Gifford, Carsten Hansen, Joe Levine, Robert Lurz, Chris Peacocke, Joelle Proust, Rachel Singpurwalla, Andreas Stokke and anonymous referees of this journal, all for calling attention to issues in drafts.

  • In the time between the finishing of a draft that Carruthers read and the publication of it here, Carruthers has now written an entire (2011a) book on the topic, inter alia, replying to that draft. I'm not going to alter the passages he cites, except to insert footnotes briefly indicating replies to his replies that I discuss at greater length in a postscript at my website (Rey, 2012). See also Carruthers, 2011b and Rey, 2011 for a further exchange between us.

Address for correspondence: Department of Philosophy, University of Maryland, Skinner Building, College Park, MD 20742, USA.



Shoemaker (1996) presented a priori arguments against the possibility of ‘self-blindness’, or the inability of someone, otherwise intelligent and possessed of mental concepts, to introspect any of her concurrent attitude states. Ironically enough, this seems to be a position that Gopnik (1993) and Carruthers (2006, 2008, 2009a,b) have proposed as not only possible, but as the actual human condition generally! According to this ‘Objectivist’ view, supposed introspection of one's attitudes is not ‘direct’, but an ‘inference’ of precisely the sort we make about the attitudes of others, an inference that has the advantage in our own case of only our own sensory data and memories, our behavior, and of the context we are in; i.e. we are all substantially self- blind. After sorting out a number of methodological and verbal issues, I argue, first, that the a priori arguments against Objectivism don't succeed, and that Gopnik and Carruthers are right to regard the issue as an empirical one. On the other hand, I think they seriously underestimate the difficulty of establishing Objectivism. It is unlikely there is an inferential procedure from the data of pure sensation, behavior and context to the relevant self-attributions that would be as spectacularly reliable as people manifestly seem to be. Moreover, there is a simpler model: the mind very likely consists of a panoply of sub-routines some of whose outputs are ‘tagged’ for their having been so processed, rather in the way that software ‘documents’ are on standard computers. Introspection plausibly consists in a person's simply attending to distinctive constellations of these tags, even though they may lack phenomenal feels. This draws attention to an important independent fact: that much of phenomenology (or ‘what it's like’ to be in a certain state) may be constituted by facts that are not phenomenal.

Get access to the full text of this article