In order to provide embodied artificial systems with rich semantics reminiscent of human language complexity, agents need mechanisms for conceptualising complex compositional semantic structure. For this we have developed a computational system called Incremental Recruitment Language (IRL). In our situated language experiments with physical robotic agents, IRL is also used to bootstrap and shape the robots’ grounded representations.
IRL is an implementation of procedural semantics, i.e. semantics is characterised as a program, the execution of a structured set of (semantic) operations. The set of semantic primitives is open-ended and adaptive, allowing the agents to dynamically adjust their semantic inventories in order to reach their goals.