that (a) Twyla and Roberta are either black or white, not black/white (duck or
rabbit, not duck/rabbit); (b) they know each other’s race and therefore so does
she; and (c) she shapes the story so that her audience cannot know which one
is which.^5
As Zunshine explains, metarepresentation is a term referring to our cogni-
tive ability to keep track of the sources of information. Thus, an act of metarep-
resentation has two parts, one concerned with the source of the information
and one concerned with the content of the information. The first sentence of
this paragraph provides a handy example: it provides a content (the definition)
and a source for that content (Zunshine). Your understanding of my sentence
adds an additional source: “Phelan writes that Zunshine says that metarepre-
sentation is . . .” When we monitor sources as well as content, we remain open
to the possibility that the content is unreliable or otherwise limited, as the
expression “consider the source” indicates. When we lose track of sources, we
are susceptible to mistaking lies, distortions, and other unreliable messages for
accurate reports, interpretations, or evaluations. Zunshine impressively shows
how a focus on source-tracking can illuminate our experience of unreliable
narrators (her main examples are Richardson’s Robert Lovelace and Nabokov’s
Humbert Humbert) and of detective stories.
As a rhetorician, however, I am struck by Zunshine’s relatively weak com-
mitment to the (implied) author as the ultimate source of fictional communi-
cation. By a weak commitment, I mean that Zunshine recognizes this point in
theory but does not give much weight to it in her interpretive practice. In her
discussions of Clarissa and Lolita, Zunshine focuses not on the choices made
by Richardson and Nabokov but rather on the character narrators’ own fail-
ures in monitoring themselves as sources and their consequences for readerly
decisions about unreliability. In this way, Zunshine effectively treats unreliable
narration as if it works primarily along a single communicative track running
- One might take the duck/rabbit figure as a closer analogue to “Recitatif ” by contend-
ing that Morrison simultaneously constructs Twyla and Roberta as white and as black. In this
view, rather than creating a single storyworld and impeding her audience’s access to one of
its crucial features, Morrison simultaneously creates two distinct storyworlds and thus two
distinct narratives: in one Twyla is white and Roberta black and in the other their racial iden-
tities are reversed. One could make the analogy even tighter by contending that Morrison
makes it impossible for us to determine whether there is one storyworld or two. I find some
merit in both hypotheses but ultimately find them less compelling than the view that Morri-
son constructs a single storyworld and denies our access to part of it. This view corresponds
more closely to most readers’ experience. In addition, the hypothesis about two storyworlds
foregrounds the synthetic component of Morrison’s characters and her storyworld and thus
emphasizes their fictionality, when every other feature of the narrative points to Morrison’s
interest in minimizing the difference between her storyworld and the actual world—and the
realities of its racial politics.
TONI MORRISON’S DETERMINATE AMBIgUITY • 155