A second objection of Walton’s proves more telling, and it begins to point
the way toward combining and integrating elements of Goodman’s conven-
tionalism with classical resemblance theory. Walton notes that Goodman has
difficulty explaining the greater realism of, for example, a painting by Verm-
eer compared with one by Braque.^30 In order to explain this difference,
Walton argues, we must distinguish betweendepictionanddescription(refer-
ring by means of the use of language).ContraGoodman, depiction is not just a
differentkindof language for referring; it works differently. To depict an
objectoin a workwis toprescribethat an audience, in looking atwis to
imaginethat it is looking ato.“A work depicts a particular actual object if in
authorized games [of imagining or making-believe] it is fictional [i.e. part of
the game] that that object is what the viewer sees.”^31
This account of depiction explains the varieties of successful visual repre-
sentation and their connections with varying historical habits and conven-
tions. Many quite different marked surfaces and three-dimensional objects
are such that we can successfully imagine that in looking at them (literally)
we are looking at a represented object. Suitably instructed, we can imagine
that in looking at a bicycle seat and handlebars we are looking at a bull’s
head,^32 and we can imagine that in looking at 4-by-6 black and white photo-
graph we are seeing a multihued circus carousel or a 6-foot tall man. Styles of
visual representation together with instructions for seeing represented
objects“in”them do change over time and place.
But Walton’s analysis also explains both our sense of the visual immediacy
of the represented object and the comparative realism of some representa-
tions. When we imagine seeing a represented objectoin a workw, then this
imagining suffuses our perceptual experience ofo.“Suitably internalized, the
principles of make-believe guide the imaginings that inform one’s perceptual
experience.”^33 In thus pretending, we really do seem to ourselves to seeo
itself. A visual representation is then comparatively realisticnotwhen it
“directly resembles”what it represents: any black-and-white snapshot of a
middle-sized object at a medium distance is much more like any other such
snapshot than it is like the thing represented. Rather, a visual representation
(^30) Walton,Mimesis as Make-Believe, p. 299. (^31) Ibid., p. 297.
(^32) Picasso’sBull’s Head(1943) appears on the cover of the paperback edition of Walton’s
book.
(^33) Ibid., p. 302.
Representation, imitation, and resemblance 35