Essay by Stefaan Verhulst: “In 2008, I published “Linked Geographies: Maps as Mediators of Reality” (in The Hyperlinked Society). I opened the chapter with a question that, eighteen years on, has taken on a strange new currency: “Through maps, we grasp reality. But do maps also shape our reality and our behavior? Do they determine the world we see and live in?”
My argument then was that maps should not be understood simply as mirrors of an independently existing world. They are also mediators: they contextualize and frame our perceptions of reality. And despite their frequent claims to scientific or technical neutrality, maps inevitably embody choices, purposes, and forms of subjectivity — “whether consciously or not, they contain the biases of their creators”.
I did not expect to return to that essay. But three developments in the last days and weeks have made its central claim feel less like a historical observation and more like a live governance problem. The first concerns projection: the African-led “Correct the Map” initiative, which — as of yesterday — has secured a United Nations General Assembly resolution against the Mercator default. The second concerns naming: a wave of politically driven geographical renaming, most recently the executive-order rechristening of Lake Ontario as “Lake America.” The third concerns computation: the integration of generative AI into Google Earth and comparable geospatial systems.
Each is a contest over who has the authority to determine how the world is represented and who has the leverage to change that representation. Projection, naming, and computation are simply three fronts in that contest: three layers at which the seemingly neutral map turns out to encode purpose, power, and choice.
The mediation characteristic of maps I described in 2008 has now been extended across the full stack of geographical representation—from the projection that fixes relative size to the label that assigns meaning to the model that now infers, predicts, and generates. AI is the newest and deepest of these layers, and I will argue it changes the mediation in kind rather than degree. But it is best understood as the latest front in a longer struggle over representational infrastructure, not as the whole of the story…(More)”.
