Systemic approach to meaningful representations.
An existing systemic approach to meaning generation (usable for animals, humans and robots) has recently been extended to the notion of meaningful representation.
The starting point is a system submitted to a constraint (stay alive, avoid obstacle, ..) that receives information from its environment and compares it with its constraint.
The generated meaning is the connection existing between the received information and the constraint. It triggers an action aimed at satisfying the constraint (avoid predator, turn right/left, ..).
A Meaning Generator System (MGS) has been introduced. It can be used for any agent submitted to constraints.
The notion of meaningful representation of an entity for an agent is defined as being the networks of meanings related to the entity, with the action scenarios.
See book chapter and IACAP 2011 proposed abstract
Such meaningful representations are used in the approach on artificial consciousness based on self-consciousness (http://robots.ne t/article/2244.html).
More is to come on this.
Christophe Menant