A department of the College is a medium, not a subject. The English Department holds what is done in language; the Art Department what is done in the image; Media Studies what is done on the screen. This department holds what is done in the map — and a map is not a picture of a territory. It is an argument about what is next to what, made in a form that hides the fact it is arguing.
Every map states three things before it says anything else: what counts as a node, what counts as an edge, and what is left off. None of the three is discovered. All three are decided, by somebody, in advance — and the reader receives the decision as terrain. That is the department's object: not the places, but the deciding.
The department was working before it had a room. The Mythic Constellation has credited its charts to the Department of Cartography since it was drawn, and five other maps were built to the same grammar without anyone naming the discipline they had in common. This page is not a new project. It is the room those six were already standing in.
They share a form, and the form is a claim. Each is a force-directed graph: nodes repel, edges pull, and the layout is left to settle. Nothing is placed. Where a thing ends up is a consequence of what it is attached to, which means position is an output and not an input — the opposite of every map that puts a capital in the middle because it is the capital. A node drifts to the centre here only by being on the routes between other things.
That is also the department's standing warning about itself. A force-directed layout is legible and it is seductive, and the two are related: it produces a picture that looks like a finding before any finding has been made. The corrective is that every map here is drawn from a record that exists apart from it — the pages on disk, the commit history, the ledger — so the map can be wrong about the record and the record will say so.
A seventh map is held elsewhere on purpose. The Royalty Map belongs to Media Studies, because what it draws is a medium's genealogy rather than the site's own shape. It is worth the walk across the College anyway: it is the only map here that will refuse you a number. Choose one of its three genealogies and the four centrality models switch off, because they are arithmetics over the merged graph and a layer is a different graph — so an absent node reports as not applicable rather than as a low rank. Cartography's habit is to fill the frame. That map is the department's counter-example, kept in another room.
The other thing it demonstrates is projection. Hide a class of node and the map does not lose the connections that ran through it — the hidden node becomes a relay, its neighbours are joined directly, and the new line is drawn differently because it is a different kind of claim. A projected edge says these two are connected through something you are not looking at. Most maps in the world draw that line identically to a real one. That is the single most common lie in the medium, and it is a lie of omission committed by the renderer.