Foucault’s indictment of resemblance is based in an assumed (though carefully historically constituted) progression of the role of language. Language in the classical period, according to Foucault, was centered on naming—the proliferation of words used to resemble things:

The art of language was a way of “making a sign”—of simultaneously signifying something and arranging signs around that thing; an art of naming, therefore, and then by means of a reduplication both demonstrative and decorative, of capturing that name and enclosing it and concealing it, of designating it in turn by names that were the deferred presence of the first name, its secondary sign, its figuration, its rhetorical panoply. (O.T. 43)

This process provoked the growth of commentary to expand the complex interplay of naming. Foucault’s relatively cryptic comments on classical painting and image making presuppose an analogous process of the proliferation of resemblance outside linguistic convention:

Separation between linguistic signs and plastic elements; equivalence of resemblance and affirmation. These two principles constituted the tension in classical painting, because the second reintroduced discourse (affirmation exists only where there is speech) into an art from which the linguistic element was rigorously excluded. Hence the fact that classical painting spoke—and spoke constantly—while constituting itself completely outside language; hence the fact that it rested silently in a discursive space; hence the fact that it provided, beneath itself, a kind of common ground where it could restore the bonds of sign and image. (This is Not a Pipe 53)

Stated in another way: “There must be, in the things represented, the insistent murmur of resemblance; there must be the perpetual possibility of imaginative recall” (O.T. 69). Classical painting “speaks” by recalling, through resemblance, the thing that it represents. It affirms in silence. Classical writing constitutes representation by a figuration, a metonymic connection deferred from the original thing. Central to this project is the imagination; only imagination can close the distance between linguistic symbol, image, and thing.

Foucault and Baudrillard champion an alternative to this slavery to resemblance through the play of similitude. In essence, this also relies on imagination but it is an imagination freed from all concrete referents. The basis of this “new” form of knowledge is thought itself—as Magritte echoes in a letter to Foucault:

The words Resemblance and Similitude permit you forcefully to suggest the presence—uttterly foreign—of the world and ourselves. Yet, I believe these two words are scarcely every differentiated, dictionaries are hardly enlightening to what distinguishes them.

It seems to me that, for example, green peas have between them relationships of similitude, at once visible (their color, form, size) and invisible (their nature, taste, weight). It is the same for the false and the real, etc. Things do not have resemblances, they do or do not have similitudes.

Only thought resembles. It resembles by being what it sees, hears, or knows; it becomes what the world offers it.

Foucault champions Magritte’s usage of linguistic elements to subvert the project of resemblance, to render realism impotent. Magritte has broken down the divide between linguistic and pictorial elements by assaulting the problem of reproduction on both fronts by painting thought. The progression of epistemé promoted in The Order of Things is problematic on a number of fronts due to totalizing assumptions:

  1. The classical mode of thought assumed that the world was finite and could be named

  2. Seventeenth century thought imposed grammar and classification upon those names

  3. Eighteenth century thought assumed that these classifications constituted knowledge.

  4. Nineteenth century thought worked to organize that thought (based in naming, affirmation and resemblance) into systems.

As Foucault himself explored, such blanket models for modes of thought seldom hold up under scrutiny. The presence of deeply evolved grammars and modes of classification in the classical age, the argument for sensual knowledge outside the finite in the theories of the sublime in the eighteenth century, and the rebellion against systems by Blake and others in the nineteenth century are chinks in the armor of this totalizing mode of historical thought. However, Foucault amply highlights the tension between linguistic modes of knowledge and imagistic ones.

Foucault’s isolation of the problem of representation into resemblance versus similitude can also be restated as a unique privilege granted to metonymy in the classical age, and a displacement of that privilege into metaphor in the modern and postmodern ages. It is yet another variation on the imagistic versus linguistic mode. Language is rooted in metaphor; images are rooted in metonymy.

The flight from representation in modern art can be characterized as a linguistic turn. Absent of a referent, these works demand description not as an accessory, but as a necessity for their existence. Doing this, they do not affirm the presence of an object, but instead generate discourse which simulates the effect they are meant to provoke.