Agency in the Physical Sciences is a workshop at ECAL 2017, to be held September 4, 2017 in Lyon, France.  

The study of agents in physical or physics-like systems is at the foundation of artificial life. The concept of an agent can be seen as an abstraction of living organisms that focuses more on action, perception, and goal-directedness (normativity if you like) than replication and growth. Many systems studied in artificial life research e.g. cellular automata and reaction-diffusion systems can be seen as artificial physics that exhibit primitive “agent-like” structures. “Agent-like” currently remains an intuitive classification; this workshop aims to provide a platform for research directed towards making such a classification in a formal or quantitative way.  

More precisely, this workshop addresses questions like:

  • What sorts of physical structures can be said to be agents?
  • What can be gained by understanding them as agents?
  • What sorts of agents are they?
  • And by virtue of what physical properties are they those sorts of agents?

This workshop aims to explore questions relating to agency in its most primitive forms, both in simulation and in physical reality. To this end, we invite both theoretical and empirical contributions on primitive agency.

  • The emphasis for theoretical contributions is on clear definitions and quantification: we are most interested in concepts that lend themselves to a rigorous analysis.
  • Empirical contributions should describe apparently purposeful, or otherwise apparently agentive, behaviour in systems simpler than those previously known to display such behaviour.