Emergence of Autonomy in Contemporary Artificial Agents: Any Novelty?

The article by Arnellos, Spyrou and Darzentas discusses the question of whether an artificial system, such as a robot or artificial life, can be made genuinely autonomous. The negative answer that they present is sound, but seems to us to be a simple consequence of the definitions adopted. We argue that the theoretical considerations presented, based on the notions of functional emergence and meaning, need further explanation in order to give support for their argument. As it is presented so far, not very much novelty is added to the already existent research on autonomous and cognitive robotics and artificial life. Also, we would like to challenge the defenders of the “theory of self-referential systems” to set up conditions in which the theory would be considered falsified. The problem of whether life and autonomy could in principle be sustained with components other than carbonbased organic molecules seems to involve still unknown empirical aspects, and cannot be answered simply with theoretical considerations.