Metaphor, Metonymy, and Personification in the Language of Robotics

Abstract

This paper presents an approach to meaning that relates the (pseudo-)problems raised by the language of robotics to the fundamental distinction between semantic and representational truth. As shown by Plato's Cratylus or the controversy between Descartes and Arnauld on the "falsity" of ideas, "objective" theories of truth generally neglect this distinction, which cannot be captured by the Fregean dichotomy between "reference" and "sense", and involves two notions of the "world"-the objective one of the "word-as-it-is" and the subjective one of the "world-as-experienced". Relying on a typology of metaphors that emphasizes the difference between "pre-wired" and "non-pre-wired" metaphors, as well as the specificity of those metaphors that are not processed "on line", one can provide arguments that support the thesis that the language of robotics, which bears on artifacts, uses not so much semantic as representational personification produced by either metaphor or metonymy. While metonymy is grounded on rational assignments of intentions, metaphor ensures compactness of thought and expression.