Sociality and Normativity for Robots

Abstract

The paper develops a general conceptual framework for the ontological<br/>classification of human-robot interaction. After arguing against fictionalist interpretations<br/>of human-robot interactions, I present five notions of simulation or<br/>partial realization, formally defined in terms of relationships between process systems<br/>(approximating, displaying, mimicking, imitating, and replicating). Since each<br/>of the n criterial processes for a type of two-agent interaction Á can be realized<br/>in at least six modes (full realization plus five modes of simulation), we receive<br/>a (6n n)(6n n) matrix of symmetric and asymmetric modes of realizing Á,<br/>called the ‘simulatory expansion’ of interaction type Á. Simulatory expansions of<br/>social interactions can be used to map out different kinds and degrees of sociality in<br/>human-human and human-robot interaction, relative to current notions of sociality<br/>in philosophy, anthropology, and linguistics. The classificatory framework developed<br/>(SISI) thus represents the field of possible simulated social interactions. SISI<br/>can be used to clarify which conceptual and empirical grounds we can draw on to<br/>evaluate capacities and affordances of robots for social interaction and it provides<br/>the conceptual means to build up a taxonomy of human-robot interaction.