With regard to self-concept, my claim in the Rorty book is certainly not that, as sociologists of ideas, we should somehow let intellectuals tell their own stories. As I’ve noted above, the accounts intellectuals give of their own lives are often highly problematic from the standpoint of sociological realism. That does not mean, however, that their self-narratives are irrelevant for explaining their actions. Beyond the point about grievances, the specific way in which I think they are relevant is that such narratives tend to be built around categories of intellectual selfhood into which thinkers understand themselves to fall. My argument is that thinkers labor under social-psychological pressure to do work they see as consistent with these categorical self-concepts. As to the methodological question of how one goes about reconstructing self-concepts, the answer I proposed in the book is to look for as many instances of self-talk as one can find – in interviews, autobiographies, correspondence, diaries, speeches, and so on – and attempt to discern from these the most salient and least evanescent categories of personhood a thinker places himself in at any given juncture. This reconstruction will have been carried out with maximum objectivity if most observers of that self-talk – including, where possible, the intellectual himself – will agree that those in fact seem to have been the most salient identity categories.
– Neil Gross