The Knowledge Graph: Seeing the Connections Between Reasoning Errors

Most people, when they first open the Knowledge Graph, do what you'd expect: they zoom in, click a node, read the description, move on. That's a fine way to explore. But the real value of the graph isn't in the individual nodes — it's in the edges between them.

What the nodes are

Each node in the graph represents one aspect from the TellDear taxonomy — one of 535 named, defined patterns of reasoning error or manipulation. The nodes are color-coded by dimension. Size and position reflect the aspect's connectivity — how many other aspects it relates to.

Highly connected nodes — aspects like emotional appeal, framing, or cherry-picking — sit closer to the center and appear larger. These are the workhorses of manipulation: versatile, frequently combined, hard to isolate.

What the edges mean

An edge between two aspects means they are conceptually related — they co-occur frequently, one enables the other, or they are the inverse of each other. Confirmation bias and cherry-picking are linked because the cognitive tendency to seek confirming information is exactly what makes selective evidence presentation effective as a manipulation technique. One is the psychological vulnerability; the other is the deliberate exploit of it.

Some edges connect aspects within the same dimension. Many cross dimensions. The cross-dimension edges are often the most revealing: they show where a cognitive bias becomes the mechanism for a logical fallacy, or where a discourse mechanic creates the conditions for propaganda to land.

How to use it in practice

The most useful way to approach the graph is with a specific text in mind. After analyzing a piece of rhetoric, you'll have identified one or two aspects you're fairly confident about. Find those nodes in the graph and follow their edges.

You'll typically find that the aspects you identified cluster with others you hadn't noticed. A speech that uses appeal to fear will often also use scapegoating and false dilemma — and the graph shows you why: they're structurally related. Once you know one is present, the others are worth looking for.

What the graph reveals about manipulation strategies

Sophisticated manipulation rarely relies on a single technique. The graph reveals the natural "constellations" — clusters of aspects that tend to appear together because they're mutually reinforcing. Populist rhetoric, for instance, typically combines us vs. them framing, appeal to the people, scapegoating, loaded language, and oversimplification. These aren't chosen randomly — each one prepares the ground for the others.

When you can see a constellation in a text, you've moved from identifying individual errors to understanding a strategy. That's a qualitatively different kind of critical insight.

Navigating the graph

A few practical tips: use the search to jump directly to a specific aspect. Filter by dimension to focus on one lens at a time. Click any node to see its full description, related aspects, and verification steps. The zoom controls let you move between the overview (where clusters are visible) and the detail level (where individual connections become legible).

The graph rewards slow exploration more than quick scanning. Spend twenty minutes following edges from a starting point you find interesting — you'll understand the taxonomy's architecture better than any amount of list-reading would provide.