How do you create a pantheon?

Portfolios are the least useful place to start

The standard approach is to enumerate domains — war, harvest, the sea, the dead — and assign a god to each. It produces a complete pantheon quickly and a dull one permanently, because a tidy division of labour gives the gods nothing to want from one another.

Mythologies people actually find memorable are messy in specific ways: overlapping claims, gods who are the same figure under two names, a domain held by someone who took it from someone else. Those irregularities are not decoration. They are the reason a pantheon can carry a plot.

Five moves that generate structure

Pick two or three. Applying all five at once produces noise.

• A contested domain Two gods both claim the sea, and their churches have separate calendars, rites and grudges as a result. Contested domains create factions on the ground automatically.
• A displaced god Someone held a portfolio before the current holder. Where the old worship survives — remote valleys, sailors, criminals — you have an instant subculture.
• A binding debt or crime One founding act the whole pantheon is still arranged around. It explains why gods who dislike each other cooperate.
• A silence A god who has not answered a prayer in three hundred years. Their clergy still exists, and what they do with that fact is a campaign.
• An overlap of names The same god worshipped as two figures in two regions, each convinced the other is a heresy. Cheap to write, enormously productive at the table.

Decide what the gods do, mechanically and narratively

Two questions determine how the pantheon behaves in play, and they are worth settling explicitly rather than by drift. Do the gods answer? And can anyone tell the difference between a god answering and a coincidence?

A setting where divine intervention is verifiable produces theocracies, tested prophecies and clergy with actual authority. One where it is ambiguous produces schisms, competing interpretations and faith as a genuine choice. Neither is more correct — but a table that has not decided will produce both by accident, and players will notice the inconsistency long before they notice the pantheon.

This is also the boundary between a pantheon and a religion. The pantheon is the roster and its entanglements; the religion is what people do about it — the rites, the institutions, the money, the schisms. They are separate design problems, and the second is usually where the play is.

A pantheon that will earn its keep

• At least two gods want something incompatible.
• One portfolio has changed hands, and someone still resents it.
• You know whether prayers are demonstrably answered.
• At least one god's worship maps onto a real group in the setting.
• You could run a session about a disagreement between two churches without inventing anything new.