the contempt atlas

This is the deep-dive, not the point.

A flashy contempt map is the most screenshot-able thing here — and the version that spreads is a grievance ripped from context. So before you read it: the finding is that we overestimate each other. This map is the structure underneath that.

Contempt depends on direction

The single clearest thing in the real data: who is talking to whom changes everything. Across the divide, contempt is aimed. Inside a group, the same words are used to say “they look down on us” — or worn with pride.

◆ Real Reddit sample · classifier not yet validated — suggestive, not findings

aimed attributed claimed

Across the divide

When the contempt actually crosses between groups, it's overwhelmingly aimed — not imagined.

  • Liberals & progressives Conservativesn 799
  • Conservatives Liberals & progressivesn 585
  • Conservatives Rural & working-class folksn 117
  • Liberals & progressives Rural & working-class folksn 111
  • Liberals & progressives Coastal cultural eliten 90
  • Conservatives Coastal cultural eliten 74
  • Rural & working-class folks Coastal cultural eliten 37
  • Urban professionals Rural & working-class folksn 26
  • Coastal cultural elite Rural & working-class folksn 25
  • Rural & working-class folks Urban professionalsn 24

Within a group

Talk inside a group is the opposite: people attribute the put-down to outsiders, or claim the word with pride.

  • Conservatives about themselvesn 343
  • Liberals & progressives about themselvesn 325
  • Rural & working-class folks about themselvesn 256
  • Urban professionals about themselvesn 41
  • Coastal cultural elite about themselvesn 31
  • Southerners about themselvesn 15

The lexicon network

The structural view: who aims which words at whom. Edges point from the group using a term toward the group it targets; thickness is volume. Tap a node or edge to inspect it.

Solid edges ◆ are real (unvalidated) Reddit data; dashed edges are illustrative fixtures.

Contempt vectors, ranked by volume

← Back to the perception-gap test