The Mono-Thinking Surprise Test
Use predictability and surprise to detect borrowed ideological packages
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 94%
The Mono-Thinking Surprise Test diagnoses ideological package thinking through prediction. Take one known opinion and forecast the person's positions across unrelated issues. Then sample their actual work and note whether any conclusion genuinely surprises you. Near-perfect predictability suggests that group identity may be supplying answers before issue-specific reasoning occurs. Surprise alone does not prove quality, so the second check is whether the unexpected view rests on coherent evidence and logic. Apply the same test inward by listing your own positions and asking which ones you could defend without naming your political camp. Williamson contrasts highly predictable commentators with figures such as Bill Maher and Sam Harris, whose conclusions sometimes violate audience expectations. The goal is not contrarianism; it is evidence that each question was allowed to remain open long enough to be considered on its own.
Origin
Williamson describes mono-thinking as wearing a preassembled ideological onesie. His practical diagnostic is to ask when a favored commentator last surprised you with a position you could not predict from their tribe.
Core principles
- 01One opinion should not mechanically determine every other opinion
- 02Serious thinkers reason issue by issue
- 03Occasional surprise is evidence against total tribal capture
- 04Dissent can look disloyal to allies and insufficient to opponents
How to run it
- 1
Choose a varied sample
Collect positions across several unrelated topics rather than several versions of the same debate. Include at least one issue that divides the person's usual allies.
Pro tip Use full arguments rather than isolated social posts.
Watch out A narrow specialist may be predictable because the same evidence applies repeatedly.
- 2
Predict before looking
Based on one or two known views, write what you expect the person to say on the remaining issues. Do this before checking their actual positions.
Pro tip Record confidence as well as the prediction.
Watch out Retrospective claims of predictability are distorted by hindsight.
- 3
Measure genuine surprise
Compare the forecast with the person's real conclusions. Flag positions that meaningfully diverge from the expected tribal package.
Watch out Random inconsistency is not independent thought.
- 4
Inspect the reasoning
Ask whether both predictable and surprising positions follow issue-specific evidence and logic. Distinguish independent judgment from performative contrarianism.
Pro tip Look for the strongest opposing argument the person acknowledges.
- 5
Run the test on yourself
Predict your own views from your group identity and identify beliefs you cannot defend without referring to the group. Reopen those questions individually.
Pro tip Start with a position where dissent would cost social approval.
Watch out Changing a view only to appear independent is another form of audience capture.
In the wild
Williamson names Bill Maher and Sam Harris as people whose takes he cannot reliably forecast from one political cue. He has some predictive power, but they regularly reach conclusions he did not expect, suggesting that individual issues can override team alignment.
→ Surprise becomes a prompt to inspect the reasoning rather than assign a simple tribal label.
Williamson argues that if one opinion lets him accurately predict everything else a person believes, the person has probably adopted a complete ideological outfit rather than assembled positions issue by issue.
→ Predictability exposes where a package may be substituting for thought.
Common mistakes
Rewarding surprise for its own sake
An unpredictable take can be incoherent or attention-seeking; the reasoning still has to survive examination.
Testing only one topic cluster
Consistent answers inside one domain may reflect a coherent model rather than ideological capture.
Is it for you?
Best for
Media consumers and creators evaluating whether a commentator or their own thinking is guided by reasons or a preassembled political package.
Not ideal for
Technical domains where many conclusions legitimately follow from one shared model or body of evidence.
From the transcript
“if I know one of your opinions and from it I can accurately predict everything else that you believe You're probably not a serious thinker.”
“you've taken a onesie and put it on and called it an outfit.”
“When was the last time that a content creator that you like or a person that you read some news contributor, when was the last…”
From the episode
J.K. Rowling & The Cost of Speaking Freely - Warren Smith - #1005
Warren Smith