Distributed Knowledge in Simplicial Models

The usual semantics of multi-agent epistemic logic is based on Kripke models, defined in terms of binary relations on a set of possible worlds. Recently, there has been a growing interest in using simplicial complexes rather than graphs, as models for multi-agent epistemic logic. This approach uses agents’ views as the fundamental object instead of worlds. A set of views by different agents about a world forms a simplex, and a set of simplexes defines a simplicial complex, that can serve as a model for multi-agent epistemic logic. This new approach reveals topological information that is implicit in Kripke models, because the binary indistinguishability relations are more clearly seen as n-ary relations in the simplicial complex. This paper, written for an economics audience, introduces simplicial models to non-experts and connects distributed computing, epistemic logic and topology. Our focus is on distributed knowledge and its fixed point, common distributed knowledge. These concepts arise when considering the knowledge that a group of agents would acquire, if they could communicate their local knowledge perfectly. While common knowledge has been shown to be related to consensus, we illustrate how distributed knowledge is related to a task weaker to consensus, called majority consensus. We describe three models of communication, some well-known (immediate snapshot), and others less studied (related to broadcast and test-and-set). When majority consensus is solvable, we describe the distributed knowledge that is used to solve it. When it is not solvable, we present a logical obstruction, a formula that should always be known according to the task specification, but which the players cannot know.

💡 Research Summary

The paper introduces simplicial complexes as an alternative semantic foundation for multi‑agent epistemic logic, replacing the traditional Kripke‑style possible‑worlds models. In a simplicial model each vertex represents an agent’s local state (its “view” of the world) and a maximal simplex (an (n‑1)‑dimensional cell) corresponds to a global state that is compatible with the local views of all n agents. This geometric representation makes the indistinguishability relations of Kripke frames explicit as higher‑order connections: two worlds are indistinguishable to a single agent when they share that agent’s vertex, they are indistinguishable to a pair of agents when they share an edge, and so on. The paper emphasizes the duality between Kripke frames and chromatic simplicial complexes, showing that moving from a global to a local perspective yields a richer, n‑ary relational structure.

The central logical notions studied are distributed knowledge (DK) and its fixed‑point common distributed knowledge (CDK). DK of a group A means that the conjunction of all agents’ local information would be sufficient to infer a formula φ; in the simplicial setting this is captured by φ holding throughout the (|A|‑1)‑connected component that contains the actual simplex. CDK iterates this requirement over all groups in a set of groups, yielding a condition that φ is true everywhere in the relevant higher‑dimensional connectivity component. This topological interpretation provides a clear visual analogue of the otherwise intricate fixed‑point definition.

To illustrate the power of this framework, the authors analyze the majority‑consensus task, a weaker variant of classic consensus where the output must agree with the majority of initial inputs. They consider three communication models:

1. Immediate Snapshot (IS) – every round all agents write their local state and instantly observe all writes. The simplicial complex evolves by repeatedly “filling in” higher‑dimensional simplices, preserving full (n‑1)‑connectivity. In this model DK accumulates enough to solve majority consensus.

2. Broadcast – a single agent broadcasts its state to all others. If the broadcast creates a local articulation point (a vertex whose removal disconnects the complex), the complex splits into separate components. Some agents then lack DK about the broadcasted value, and CDK fails globally. The authors formalize the resulting impossibility with a logical obstruction of the form ¬K_A φ, showing that the required knowledge cannot be attained.

3. Test‑and‑Set (TAS) – agents compete to perform a “set” operation; the first succeeds and later “test” operations cannot see that choice. This competitive pattern also introduces a topological cut, breaking the necessary connectivity for DK, and again yields a logical obstruction to majority consensus.

The paper’s key technical insight is that k‑connectivity of the simplicial complex is the precise condition for solvability of the majority‑consensus task. When the complex remains k‑connected, every group can aggregate its local information, achieving DK and consequently CDK, which enables a correct protocol. When connectivity is broken (by articulation points or competitive writes), the required distributed knowledge cannot be formed, and no protocol can satisfy the task specification.

To make the ideas concrete, the authors revisit the classic “muddy children” puzzle. In the simplicial representation each child’s perspective is a colored vertex, and each possible world is a triangle. The teacher’s announcements and successive questions correspond to removing simplices that are inconsistent with the observed knowledge. The evolution of the complex visually demonstrates how common knowledge emerges after three rounds, something that is more opaque in the traditional Kripke diagram.

Overall, the paper bridges epistemic logic, distributed computing, and algebraic topology, offering a unified language for reasoning about what groups of agents can know after communication. It shows that distributed knowledge is not merely a logical construct but has a concrete topological signature that determines the feasibility of coordination tasks such as majority consensus. The work is presented for an economics audience, avoiding heavy technical proofs while still delivering new results on broadcast and test‑and‑set models, and it opens avenues for further research on larger agent sets, fault‑tolerant settings, and richer epistemic operators.