Defining interaction protocols using a commitment-based agent communication language

Source: DOI Authors: Nicoletta Fornara, Marco Colombetti Date: 2003

Core idea

Multiagent interaction protocols are more meaningful when message semantics are grounded in social commitments rather than only in allowed message sequences.

Key claims

• Protocol state should be understandable in terms of who is committed to what.
• Messages matter because they create, modify, discharge, or cancel commitments.
• Open systems need semantic interaction contracts, not just message choreography.

Harness takeaway

If a harness wants a durable log of observed goals, the primary unit should be commitment state transitions created by interaction, not a scalar estimate of trustworthiness.