than bantering. Nevertheless, the statement shows that Stevens has learned
something in the course of the narrative, has moved closer to Ishiguro’s ethi-
cal beliefs about the importance of affect in human relationships than dur-
ing his first unenthusiastic responses to Mr. Faraday’s bantering. As Stevens
moves in this direction, rhetorical readers move toward him both ethically
and affectively.
Again, as the emphasis on consequences suggests, this distinction between
estranging and bonding unreliability is based on the rhetorical effect of the
given unreliability on the authorial audience. The taxonomy of six types
of unreliability I mention above (misreporting, misreading, misregarding,
underreporting, underreading, underregarding) arises not from an analysis
of effects but from an analysis of two main variables of the communicative
exchange among implied author, narrator, and authorial audience: (a) the axis
of communication along which the unreliability occurs and (b) whether the
particular communication indicates that the authorial audience needs to reject
the narrator’s perspective or supplement it. The distinction between estrang-
ing and bonding unreliability cuts across the taxonomy of six types. More
simply, any one of the six types can function as estranging unreliability or as
bonding unreliability.
Applying Phelan’s Shaver (see the preface) yields the results that the most
important theoretical interventions so far are (a) the distinction between
bonding and estranging unreliability and (b) the claim that the effects of unre-
liability vary across a wide spectrum. However, since most previous work has
been on what I call estranging unreliability, I want to focus now on the diverse
ways authors deploy bonding unreliability by identifying six of its subtypes. As
I do, I shall consider whether each subtype is more likely to occur along one
of the three axes of communication than the other two, and, thus, the kind
of distance that each type is likely to reduce. This catalog of subtypes is illus-
trative rather than exhaustive, and, again, its purpose is to further clarify the
concept and to show why it matters. More generally, I hope the catalog adds
a significant dimension to our understanding of how authors and audiences
interact through the resource of unreliable narration.
SIX SUBTYPES OF BONDING UNRELIABILITY:
LOCAL AND GLOBAL EFFECTS
The first subtype takes advantage of the single text with two tellers, two audi-
ences, and two purposes by rendering the narrator’s communication literally
unreliable but metaphorically reliable. It most typically occurs along the axes of
ESTRANgINg UNRELIABILITY, BONDINg UNRELIABILITY • 101