Iconic graphs: An exercise in topological phenomenology

Abstract

John Sowa has often pointed out that his proposal of conceptual graphs develops and updates intuitions by Peirce, as well as by other thinkers in the areas of psychology and linguistics like Selz and Tesni~re. To the ideas of the latter refers the Fields medal Ren6 Thorn, who developed catastrophe theory and made several innovative proposals for the application of catastrophist formalism to natural language. Thorn's ideas have been developed in specifically linguistic terms by Wildgen, while Petitot has paid particularly close attention to the relationship between language and visual perception (Cf. in particular [Wildgen 1982] and [1994], [Petitot 1985] and [1995]). The contributions of Thorn, Wildgen and Petitot can be better understood as aspects of a broader research programme which we may call 'cognitive semantics.' In its turn, cognitive semantics is a component of an even broader programme which undertakes the transition from the paradigm of computational systems theory to the paradigm of dynamic systems theory. The first embedding that of the modern developments of proposals by Tesni~re and other structuralists working in cognitive semantics can be characterized by setting them against Chomskian theory. Indeed, cognitive semantics can also be viewed as a reaction against the cultural and academic hegemony of generativism, since cognitive linguistics regards language as a product of cognitive processes. For cognitive semanticists, "the basic principles of the cognitive organization of the world (perception of reality, construction of mental models, reference to encyclopedic knowledge, conceptualization, mental images and schemata, the grounding of meaning in bodily and social experience, and so on) can and should be accommodated into linguistic analysis ... On the other hand, the idea that the components of grammar should be analyzed separately is also rejected" ([Serra Borneto 1993], 445-6). Cognitive semantics, moreover, rejects the thesis of the overall independence of syntax from semantics. Generally speaking, this theory articulates into the following two branches: