What Makes a Thinking Machine? Computational Semiotics And Semiotic Computation

Abstract

Semiotics is considered here as a relational and ontogenetic approach to describing cognition and communication in signifying systems. Implementing a semiotic approach to computing thus would require a computable and scalable signifying space where signs can be arbitrarily created, related, interpreted and deliberated. The author argues that although signs are representations, a signifying space cannot be realized under the current representational paradigm of recording and processing static data in a hierarchical data space. A ‘semiotic machine’ instead must implement a genetic epistemology of cognition based on assimilation and pure relations. The ‘Pile’ system introduced in this essay is supposed to meet these requirements and is described as a ‘semiotic computation’ system structurally enabling deductive as well as inductive and abductive processes of self-reflection, deliberation and interpretation, commonly associated with ‘thinking’.