Deborah Knight SELVES, INTERPRETERS, NARRATORS I Autobiography, understood in die standard sense of someone telling or more frequendy writing the story of her life, has been a periodic focus of philosophical attention. Partly this is because philosophers such as Augustine and Rousseau have written some of the classics of the genre. Recendy, interest in autobiography has been renewed among philosophers concerned widi narrative and especially those concerned widi die relation between interpretation, narration, and die self. In particular, there has been considerable interest in questions of narrative voice and perspective,1 as well as in questions concerning whether die self is (merely) a literary or fictional construct. One philosopher who has had much to say about selves as autobiographers is Daniel C. Dennett. He argues that the self is just an autobiographical accretion forming on or around some agent who is able to speak, to differentiate itself from others, and to call itself "I."2 The self, according to Dennett, is just the hole surrounded by the autobiographical doughnut. It is a fiction, and like any otiier fiction, what any selfamounts to will depend on die way we tell the story. But as I will argue, Dennett's view oversimplifies die complex narrational role involved in telling tales about oneself and otiiers. It also oversimplifies die relationship between the accreters of autobiography and dieir interpreters. In particular, it fails to recognize that both autobiographical accreters and interpreters are involved in a narrational activity. Perhaps it should not surprise us to discover a philosopher-novelist whose work offers a helpful counterpoint to Dennett's views about story-based selves. Rebecca Goldstein's novel, TheLate-SummerPassion of a Woman ofMind,5 contains an important insight into die narrational Philosophy and Literature, © 1994, 18: 274-286 Deborah Knight275 function presupposed by bodi interpreters and self-interpreters. Goldstein, not unlike Dennett, argues that in a fundamental sense our lives are accounted for and made sense of in storied terms. This is not an original observation. What Goldstein recognizes that Dennett does not is the complex narrational interconnection between teller and interpreter. Further, she recognizes the hopelessness of the desire for some omniscient interpreter or impersonal narrator—some third-person interpretive perspective that can simply bracket the intentions of the agent as that agent understands them. Goldstein helps us to see part of what is at stake in all discussions of storied lives, as well as what is at stake in the idea diat die self or subject of action is produced through the stories tiiat we and otiiers tell. So let me contrast Dennett's views about narrated selves with Goldstein's views. My objective is to show the limitations of Dennett's position, and die suggestiveness of Goldstein's. II Dennett's views about the self can be thought of in terms of a pair of guiding intuitions: either (Dl) the self is soul-like, the ghost in the Cartesian machine, the originator a priori ofmeanings and actions, the incorrigible source of intentions, the consciousness that preexists and directs action; or (D2) the self is discursive, which is to say that it is a product formed in language dirough die telling and exchanging of tales about ourselves. If (Dl), then the self is prelinguistic and makes each of us God-like creators and guarantors of our meanings; or, if (D2), then the selfis the fragmentary, pieced-togetiier posit that orients us through an otherwise potentially unbalancing variety of environmental and psychological inputs and outputs. Dennett treats diese two views as the only possible positions in die debate, and since his brand of anti-Cartesianism, like RyIe's before him, requires that he refute (Dl), he is left only with (D2). By refuting (Dl) and keeping (D2), Dennett allows himself to conclude that, while agents' selves are spun through the tales they tell of themselves, there is no good reason to base one's interpretation of any agent on the version of the tale told by that agent. The only perspective from which agents can be interpreted is a diird-person perspective. There can be no significant question as to first-person authority, since on Dennett's view, die first-person subjective position has...