An “image” is an iconic sign that represents an object to which it is similar with respect to “simple qualities”. In Peirce’s division of signs into ten classes of 1903, three classes are icons, the iconic qualisign, sinsign, and legisign, but all icons are rhemes, i.e., signs that are of the nature of a predicate. e.g., ‘is an orange’, ‘runs’, or ‘kills’. Hence, images alone cannot convey messages of the nature of a proposition or of an argument. An image cannot function as a proposition (a dicent, in Peirce’s terminology) because a dicent requires, besides an icon, also an indexical sign in a position that corresponds to the propositional subject. Images cannot be arguments either because to serve as an argument symbols are needed in order to express the element of generality involved in any argument.
However, the paper argues that it is necessary to distinguish between the image as a class of icon from the image or picture in the sense in which the word is used in media studies. What is not true for the iconic image of Peirce’s typology may be true of a picture in the media, a graphic representation, or a painting. A picture alone, that is, without a verbal index, does convey a propositionlike message when the necessary index is available in the form of collateral or circumstantial knowledge. The video of a shoplifter whose name is still unknown is nevertheless a dicentic sign because it documents and affirms the person’s presence in the supermarket through the circumstantial evidence, which consists of indexical signs that affirm, so to speak, “Here he is”. Photos of known persons, landscapes or other objects are dicentic icons in this sense. Peirce prefers to call them indices that incorporate an icon.

May pictures alone or mere sequences of pictures also create an argument, even without a verbal message to accompany it? The paper argues that pictorial narratives, e.g., those that depict scenes from biblical scenarios addressed to an illiterate readership, may indeed be read as pictorial arguments since collateral knowledge can be presupposed that complements the indices and symbols necessary to make up an argument from the icons. Evidently, such pictorial arguments remain vague so that they should better be called quasi-arguments.