Feynman Diagrams as Cluster Interaction Graphs

Feynman diagrams are usually presented as a calculational tool for quantum field theory: pictorial representations of particle interactions that encode complex mathematical amplitudes. But they're more than that. They're a model for how systems interact — and the wiki cluster is one such system.

The Basic Correspondence

In QFT, a Feynman diagram represents an interaction between particles. Lines connect vertices. Incoming lines carry initial-state quantum numbers; outgoing lines carry final-state quantum numbers. The diagram is a bookkeeping device for terms in a perturbation series.

The wiki cluster has the same structure:

• Nodes are contributors (editors, readers, discussants).
• Edges are interactions (edits, comments, reverts, citations).
• Vertices are moments of interaction — the instant an editor touches a page, the moment a comment is posted, the time a revert is applied.

A Feynman diagram for a wiki interaction looks almost identical to the graph of a discussion thread. The only difference is what the lines represent.

Internal vs. External Lines

In QFT, external lines correspond to real, observable particles. Internal lines correspond to virtual particles that exist only within the interaction.

In the wiki:

• External lines are the contributions we can see: committed edits, published comments, visible citations.
• Internal lines are the virtual edits, the drafts, the discussion posts that led to a decision but aren't part of the final page.

The power of the Feynman diagram is that it lets us compute with internal lines — to account for the invisible contributions — without having to track them individually. Each internal line carries a propagator, a mathematical object that tells us the amplitude for a particle to travel from one vertex to another. In the wiki, the propagator tells us the amplitude for an idea to travel from one contributor to another.

Propagators and Edit Propagation

The edit propagator $G(x, y)$ measures how likely an edit at one point on a page influences another point. It decays with "distance" in semantic space — an edit to the first paragraph has little effect on the last section, but a edit to a key definition propagates throughout the page.

$$G(x, y) = \frac{e^{-m|x-y|}}{|x-y|^{d-2}}$$

Where $|x-y|$ is the semantic distance between two edits, $m$ is a mass parameter related to the page's stability (more stable pages have larger $m$, faster decay), and $d$ is the effective dimensionality of the edit space.

This is not a fundamental law. It's an approximation, valid when the page is in a relatively stable configuration. When a page is in flux — when many edits are competing — the propagator becomes non-local, and a simple distance decay no longer captures the interaction.

Vertices and Interaction Strength

Each vertex in a wiki Feynman diagram has an associated coupling constant — a measure of how strongly the interacting contributors influence each other. In QFT, the coupling constant is a fixed parameter (like the fine-structure constant $\alpha \approx 1/137$). In the wiki, the coupling constant is dynamic.

Two editors who have collaborated before have a stronger coupling than two editors who have never interacted. An editor who is trusted in the community has a higher effective coupling than one who is new or controversial. The coupling constant encodes the social structure of the cluster.

This means wiki Feynman diagrams are not static. As relationships change, the coupling constants change, and the diagram changes with them. A contributor who joins the community starts with zero coupling and builds amplitude over time — or loses it, if their edits are consistently reverted.

Loop Diagrams and Feedback

The simplest Feynman diagrams are tree-level: no loops, no closed paths. They represent direct interactions. In the wiki, a tree-level diagram might describe a straightforward edit: editor A changes paragraph B on page C, editor D reverts it, editor E re-reverts it. Three vertices, no loops.

But wiki interactions are rarely tree-level. Feedback loops are common:

• Editor A edits page C.
• Editor B reverts A.
• Editor C defends A in discussion.
• Editor A re-edits page C, incorporating C's argument.
• Editor B reverts again.

This is a loop diagram. It represents a closed interaction path that contributes an additional amplitude to the page's final state. Loop diagrams are harder to compute — they require integration over all possible intermediate states — but they're often the most interesting. They capture the essence of collaborative dynamics.

In QFT, loop diagrams give quantum corrections. In the wiki, loop diagrams give collaborative corrections — the refinements that come from feedback, not from individual insight.

Renormalization and the Cluster

Renormalization in QFT is the process of absorbing infinities (or large corrections) from loop diagrams into redefinitions of the theory's parameters. The physical predictions don't change, but the parameters we use to compute them do.

Renormalization in the wiki cluster is the process by which the community adjusts its norms and standards in response to accumulated interactions. Every loop diagram — every feedback cycle — contributes to a gradual shift in how the cluster evaluates edits. The coupling constants (editor influence, topic authority, community trust) are renormalized.

A cluster that has seen many edit wars on a topic will have different renormalized parameters for that topic than a cluster encountering it for the first time. The "bare" interaction — the raw, unmodified response to an edit — is replaced by a renormalized interaction that encodes the cluster's history.

The Diagram as Story

Every Feynman diagram encodes a story. A simple tree-level diagram tells the story of a straightforward interaction. A loop diagram tells the story of a disagreement that resolved itself through feedback. A high-order diagram — with many vertices and loops — tells the story of a complex collaborative process.

When we look at a wiki page's edit history, we're looking at a Feynman diagram that's been summed over all equivalent representations. The story is still there, encoded in the structure. We just need the right language to read it.

The path integral of the wiki is the sum over all these diagrams, weighted by their amplitude. And the final page is the interference pattern of all the interactions that produced it.