Character relationship mapping: draw the web, then read it

You can name every character in your cast, describe their goals, quote their dialogue. Then a planning session asks a different question: what does Aria think of Zane? You pause. They have never shared a scene. Has the reader noticed that too?

Character design usually happens one character at a time, but stories happen between pairs. A relationship map is the design tool that works at that scale: a drawing of the whole cast where every named character is a node and every relationship is a line between two of them. You draw it to see the structure you actually built, not the one you believe you built.

This is the planning half of relationship work. Once the story is running, the job becomes tracking: recording each relationship's turns with chapters and causes, so nothing contradicts itself (see relationship tracking for large casts). The map comes first; you cannot track a web you never drew.

Draw every edge you mean

Put every named character on the page and draw a line for each relationship, with a label and a direction. The types that matter most:

• Allies: who backs whom, and how solid it is. Trust has degrees; "would lend money" and "would die for" are different edges.
• Rivals and enemies: what the conflict is about, and whether both sides know they are in it. A rival who does not know he is hated is a different edge from a mutual feud.
• Contrast pairs: two characters who highlight each other by differing. Criticism has a name for this, the foil, "a character who contrasts with another character" in Wikipedia's wording, and the contrast does its work whether the pair ever meets or not.
• Asymmetries: the one-way edges. Unrequited love, loyalty a retainer owes that the lord never returns, forgiveness one side has not offered. Wikipedia's love triangle article lists "unrequited love and jealousy" among the form's common themes, and they are the common themes of asymmetric edges generally.
• [Secrets between pairs](/learn/character-secrets-and-reveals): a secret shared by two characters is an edge with a third property, that everyone else is outside it. Who shares which secret with whom is relationship data, not just knowledge tracking data.

Read the map for structure

A finished map is read, not admired. Four findings carry most of the value:

• Missing connections: two characters whose goals collide, who share a history, or who both matter, with no edge between them. Either the story owes them a meeting, or one of them does not need to exist at this cast size.
• Dead pairs: characters who have an edge in your notes but never interact on the page. In planning, a dead pair is a scene you have not written and may want; a cast of nothing but dead pairs is a set of strangers who happen to share a book.
• [The hub](/learn/managing-an-ensemble-cast): a cast where every line runs through the protagonist. Everything routes through her, so no scene can exist without her, and the world feels like her entourage. One or two cross-edges between secondaries, say the rival trusting the mentor, fixes it.
• Bridges: a single edge joining two clusters. Graph theory defines a bridge as "an edge of a graph whose deletion increases the graph's number of connected components": remove it and the graph splits in two. In a cast, the bridge character is the only person both camps talk to. That is a structural role, and also a target: cut her, and the groups stop communicating, which may be exactly the story you want.

Faction-level relations are their own web layered on top of this one; see building a faction web for the institutional layer.

Turn the map into scenes

A map earns its keep when it generates work. Practical moves:

1. Give every core pair at least one scene on the page, and make the scene about the edge, not incidental to it.
2. Put two clusters in a room through their bridge, and let the bridge strain.
3. Pick one dead pair per arc and write the meeting. Fresh pairs produce fresh dialogue; it is the cheapest novelty available without adding a character.
4. Weaponize an asymmetry: let the character who does not know how the other feels learn it, at the worst moment.

Keep the map honest as things shift

Relationships move. The ally of chapter one is the betrayer of chapter forty, and a map from three arcs ago is then wrong in a way that misleads. Two habits keep it useful. Redraw at arc boundaries, the points where several relationships turned at once. And when a line changes, keep the old one visible rather than erasing it, because the sequence rival, enemy, ally is the arc. The map belongs in the story bible next to the cast it describes, and from the first redraw onward it is no longer only a design tool: it is a record with history, which is exactly when the tracking discipline takes over and keeps later chapters from contradicting earlier ones.

A relationship map is a small object: names, arrows, a few labels. Its value is that it makes the cast legible as a structure, and structure is where missing scenes, missing connections and dead pairs hide.

How Kanonara approaches this

Relationships in Kanonara are drawn as a graph of the cast in the story workspace, and each edge keeps its full history instead of a single current value, so the map you drew at planning time becomes the record you consult while writing. Changes arrive as proposals with evidence when chapters are analyzed, and nothing enters the graph without your approval. See how Kanonara works.