In “The Domain of Depiction,” Dominic Lopes (2005) argues against the EROS (experienced resemblance of outline shape) theory of depiction and proposes and defends a recognition theory of depiction. I would here like to focus on what I think is Lopes’ strongest objection against EROS, and to show how a defender of EROS theory might reply.
There are six cardinal truths about depiction that EROS theory attempts to explain: 1- depiction is not property-less: to depict something, you must depict it as having some property; 2- depiction is always from a point of view; 3- only visible objects can be depicted; 4- objects can only be partially misrepresented in pictures (total misrepresentation is merely non-representation); 5- understanding a picture entails knowing what the depicted object looks like; and 6- knowing how the depicted object looks is necessary and suffices for understanding a depiction of the object (2005, p. 162). In EROS theory, depiction is explained (though not defined) as an experience of resemblance of an object in outline shape. This experience is characterized as “seeing-in”, but depiction requires more than just seeing-in, it also requires the right kind of causal relation between the picture and the object, or requires that the picture is made with the intent for the object to be seen in the picture (p. 163).
Lopes offers several objections to EROS theory, and perhaps his strongest objection is his analysis of the limits of EROS. This objection takes the form of four illustrations of problematic depictions:
- A cube shown in parallel oblique perspective (EROS either says that in many cases our intuition about what is depicted is incorrect, or that EROS is so flexible that it does not have to track with objective resemblance in outline shape—either way, EROS is not necessary for seeing-in),
- Three outline shapes of parts of a cube (recognizable outline shapes depend on the presence of significant contours or boundaries in the depiction, so EROS is not sufficient for seeing-in),
- The outline of shading on a face, shown in positive, negative and outline (outline shape is only recognizable with the presence of a an illumination boundary, so again EROS is not sufficient for seeing-in), and
- R.C. James’s “Photograph of a Dalmatian” (the outline shape of the dog is not seen until the percipient sees the dog in the picture, so EROS depends upon seeing-in).
So EROS is not sufficient or necessary for seeing-in, and seems to imply circularity.
How might an EROS theorist respond? I think one avenue might be to look back at what EROS theory claims: that depiction is seeing-in plus intention/causality. The cube depicts an object as cubical, we see a cube in the picture, and we experience the outline shape in the picture as matching a cube in outline shape. How? EROS allows for multiple points of view of an object in a depiction, and this can be taken to be one such case of multiple perspectives (and EROS is still necessary for seeing-in). In the illustration of three outline shapes of parts of a cube, only the first is depicting a cube: there is no indication that the other two (the irregular shaped outlines) are intended to depict a cube. In fact, as the outlines are presented here, they are intended to not depict a cube (or they would be poor examples for Lopes’ argument). Similarly, for the outline shading on a face, the last picture is not intended to depict a face (and it does not). The “Photograph of a Dalmatian” case is a bit different. The picture intentionally leaves out significant portions of the outline shape, which are provided by the percipient (via Gestalt). The viewer never sees the outline shape of a dog, but rather creates it in their mind (infers the presence of such a shape given the parts of the picture’s surface, and the percipient’s previous experiences of outline shapes of dogs). This picture is not made with the intention of the dog being seen: the point is to not see the dog in the picture, but to infer the resemblance to the outline shape of the dog in the picture (and from there to experience the resemblance to the outline shape of the dog in the picture). So the conditions for depiction are not met (we are not intended to see the dog). In sum, these cases can be argued to be by and large not depiction (due to lack of intent) or to be cases of multiple perspectives. No circularity seems to be involved, and it seems that EROS can still be said to be sufficient and necessary for seeing-in.
Lopes, D. M. (2005). The domain of depiction. In Matthew Kieran (Ed.), Contemporary Debates in Aesthetics and Philosophy of Art, pp. 160-175. Oxford: Blackwell.