Thermocline: how activist ideas reach the middle

Which ideas are crossing from activists to the mainstream, what carries them across, and what stops them.

ACTIVIST LAYER warm, small, fast-moving THE THERMOCLINE MAINSTREAM cold, large, slow to move crossed bounced off

In the ocean A thermocline is the thin layer where warm surface water meets the cold, deep water below. Temperature drops sharply across it, so things move freely within each layer but rarely cross between them.

In politics Ideas behave the same way. The activist layer is small, hot and fast: moral, us-versus-them language spreads quickly there, mostly among people who already agree. Most ideas never reach the mainstream below.

What gets through A few ideas do cross: when their language turns plain, when they find an institutional home (a training, a style guide, a grant criterion), or when a study gives them cover.

Where each idea stands

Speed and reach inside the movement, and outside it. The gap is the thermocline.

The thirteen tracked ideas

Each idea has its own file: its origin, foundations and spread, the roles it assigns, its use against institutions, its funding and the pushback it has met.

Crossing rising into the mainstream nowCrossed reached the mainstream earlier; now familiar, though it may be contestedReceding falling in both tiersRolled back pushed back by policyContained strong inside the movement only

The extreme ideas behind the ones that crossed

Most ideas that crossed into the mainstream had a harder sister inside the movement. The sister supplies energy and the moral case; the softer version supplies reach. Knowing the sister tells you what the softer word is for.

What crossedThe sister idea that hatched itStatus
Privilege“White skin privilege” (1967), from Marxist organizers who wanted white workers to renounce it; later “abolish whiteness” and the journal Race Traitor.Crossed
Safe space“No platform”: the 1970s student-union policy of denying speakers a stage, now deplatforming.Crossed
Lived experience“Objectivity is white supremacy culture”: the claim that evidence and neutral standards are tools of oppression.Crossing
Settler colonialism“Decolonization is not a metaphor”: the demand for literal return of land, and “globalize the intifada.”Crossing
Democratic socialism“Abolish capitalism,” and the openly Marxist caucuses inside the Democratic Socialists of America.Crossing
Billionaire tax“Abolish billionaires” and “every billionaire is a policy failure.”Crossing
Police reform (illustrative, not a tracked idea)“ACAB” and “abolish the police.”Crossing
Equity“The only remedy to past discrimination is present discrimination” (Ibram X. Kendi, 2019).Rolled back
IntersectionalityThe “progressive stack”: ranking who may speak by identity.Rolled back
Systemic racism“The system isn’t broken, it was built this way”: every institution as white supremacy.Receding
Microaggression“Words are violence”: speech treated as physical harm.Receding
Land acknowledgment“Land back”: actual return of land, not a spoken ritual.Receding
Late capitalismRevolutionary Marxism, and “eat the rich.”Receding

Our coding of each concept file. Decolonize is left out because it has not crossed.

Few ideas follow the textbook life cycle

IncubatingGaining traction in academia and activist circles.

Document the lineage and the sister idea behind it.

BreakoutReaching the mainstream for the first time.

Explain it to the middle first: say plainly what it means in practice before its backers’ softened version sets the meaning.

InstitutionalizedWritten into trainings, rules and budgets.

Pull policy and budget levers.

ContestedLosing ground with the public.

Press the advantage; don’t overreach.

Dashed line: the textbook path. × marks where each ran into trouble. Our coding of each concept file’s data.

Why most ideas never reach the middle

Moral-emotional language ("evil," "disgust," "betray") spreads faster, but mostly inside the group that already agrees. What wins the in-group keeps an idea from crossing.

+12%

More sharing per moral-emotional word (meta-analysis of 4.8 million posts). The original 2017 study found 20%, mostly within like-minded networks.

+67%

Higher odds of sharing for each word attacking the other side (2.7 million posts).

Sources: Brady et al. (2017), PNAS · Brady et al. (2025), replication and meta-analysis · Rathje et al. (2021), PNAS

What keeps it credible

Show the losses

Where grievance politics is winning, we say so.

Grade the sources

Every claim labeled by evidence quality and linked.

Openly focused on grievance and identity politics. Accurate, not neutral.