Skip to content

Agent Spaces, Agreements, Offers, Promises, and LifeCodes

Overview

In the MAP architecture, AgentSpaces are the fundamental context for interaction, coordination, and value flow between agents. They are bounded containers in which agents interact, co-create, and build shared meaning and commitments. These AgentSpaces exist within a holarchic structure, where every agent is simultaneously a whole (in their own right) and a part of other spaces.

Each AgentSpace is enclosed by a semi-permeable membrane—defined by the promises required for membership and the expectations members can hold of one another. These membranes are codified through a structure called a LifeCode.

LifeCodes (Memetic Signatures)

A LifeCode—formerly referred to as a memetic signature—is a declaration of the values, principles, and behavioral norms that characterize an Agent or AgentSpace. It may include:

  • A statement of purpose or identity
  • Governance and decision-making models
  • Expected roles and responsibilities
  • Conflict resolution practices
  • Norms of reciprocity, transparency, and trust

For a personal agent (e.g., an individual), the LifeCode lives in their I-Space—a private AgentSpace instantiated upon joining the MAP. This is where the agent evolves their own commitments, values, and intentionality. It does not host interaction with others.

In shared or collective AgentSpaces, the LifeCode defines the culture and coordination logic of that space. It may evolve over time, fork, or be extended as the space matures.

The MAP AgentSpace

Upon joining the MAP, every agent is added to the default MAP AgentSpace—the largest public sphere for interaction. It has:

  • The lowest threshold for membership
  • The highest potential reach
  • The lowest built-in trust

It serves as a global commons for agents to discover one another, broadcast offers, or initiate higher-trust subspaces.

Offers and Promises

Agents extend offers into AgentSpaces, which include:

  • Promises they are making (e.g., services, contributions)
  • Reciprocal promises they expect in return
  • Optional capital flows (financial, social, informational)
  • Conditions, roles, or mediators

Offers are extended into chosen AgentSpaces based on trust, visibility, and purpose. High-reach spaces (e.g., the MAP sphere) offer broad visibility but weaker alignment. Smaller, high-trust AgentSpaces allow more meaningful reciprocity.

If no agents accept an offer, it may be adapted—dropping or refining promises to increase uptake. This is an area where algorithms for “region-choice” or offer optimization may apply.

Agreements and Agreement-Based AgentSpaces

When an offer is accepted, an Agreement is formed—a structured, mutual commitment. In many cases, a new Agreement-Based AgentSpace is instantiated to hold the context for the Agreement. These spaces:

  • Include all parties to the Agreement
  • Enclose the interaction and coordination logic related to the Agreement
  • Are defined exclusively by the Agreement itself

⚠️ Key Architectural Principle: Agreement-Based AgentSpaces are not agents. They do not evolve or persist as autonomous entities. Their LifeCode is strictly derived from the terms of the Agreement and does not extend beyond them.

This is especially important in contexts where the Agreement has legal force. Modeled after Creative Commons architecture, Agreements may have:

  • A human-readable summary
  • A machine-readable representation (for enforcement and interpretation by MAP)
  • A legal-form representation (for interaction with legacy institutions)

To avoid semantic drift or legal ambiguity, Agreement-Based AgentSpaces are immutable, and their LifeCodes are bounded by the Agreement itself.

Agentic AgentSpaces

In contrast, Agentic AgentSpaces are full-fledged agents in their own right. They have:

  • Evolving LifeCodes
  • Memetic identity
  • Governance structures
  • Delegation logic

These spaces can encapsulate Agreement-Based AgentSpaces as parts—treating them as modules or commitments. This enables more complex social organisms to form (e.g., DAOs, organizations, communities) without overloading the Agreement itself with cultural or governance logic.

✅ This is a clean separation of concerns: Agreements define what was promised; AgentSpaces define how agents live together.

A canonical example: A corporation may contain agreements (contracts) with customers, employees, and suppliers, but its Agentic identity is defined by its charter, governance model, and LifeCode.

Holarchy and Identity

Every Agent is a Holon—both a whole (with its own I-Space) and a part (belonging to other AgentSpaces). AgentSpaces can contain:

  • Other agents
  • Other AgentSpaces
  • Agreements

This holarchic model supports recursive coordination and composable identity. For example:

  • Two agents form a persistent relationship AgentSpace.
  • Within that, they form multiple discrete Agreement-Based AgentSpaces for specific collaborations.
  • A larger Agentic AgentSpace (e.g., a DAO or org) may contain them all as part of its operational footprint.

This structure allows:

  • Legal integrity (bounded Agreements)
  • Cultural evolution (Agentic Spaces)
  • Emergent agency (spaces becoming agents)
  • Clear delegation and accountability

Summary

The MAP architecture draws a clear ontological distinction:

Type Description Can Become Agent?
Agreement-Based AgentSpace Formed via specific Agreement. LifeCode = machine-readable terms of the Agreement. Immutable. ❌ No
Agentic AgentSpace Evolves over time. Has governance, values, memetic identity. Contains other spaces. ✅ Yes
I-Space (Personal) Private space tied to an individual agent’s LifeCode. ✅ Already is an agent

By constraining Agreement-Based AgentSpaces to the terms of their originating Agreement—and modeling broader context in enclosing Agentic Spaces—the MAP preserves both legal clarity and memetic flexibility.

This design supports trust, transparency, and coherence across a network of self-organizing agents and their shared commitments.