Key Takeaways: Unfalsifiable by Design

The Big Idea

Unfalsifiable ideas — claims structured so that no possible evidence could disprove them — are the second major entry mechanism for wrong answers in knowledge production. They resist correction not through authority (Chapter 2) but through their internal architecture: they are designed, intentionally or accidentally, so that they can accommodate any observation.

Core Concepts

Popper's Insight

The value of a theory is measured not by how much it can explain, but by how much it forbids. A theory that explains everything explains nothing.

Four Mechanisms of Unfalsifiability

  1. Post-hoc rationalization — Explains any outcome after the fact without predicting outcomes in advance (Freud, market commentary)
  2. Ad hoc auxiliary hypotheses (epicycles) — New assumptions added specifically to save the theory (Ptolemaic astronomy, DSGE models)
  3. Moving goalposts — Success criteria redefined in response to failure (tech predictions, policy claims)
  4. Definitional immunity — Key terms defined so broadly any outcome fits ("culture," "authenticity")

The Falsifiability Spectrum

Level Description Example
1 Strictly falsifiable Einstein's light-bending prediction
2 Practically falsifiable Dark matter detection experiments
3 Indirectly falsifiable Individual evolutionary claims
4 Not yet falsifiable String theory (debated)
5 Unfalsifiable in principle "Everything happens for a reason"

Lakatos's Refinement

  • Progressive programme: Generates novel predictions, occasionally confirmed
  • Degenerating programme: Only accommodates existing evidence, adds epicycles

The Five-Question Diagnostic

  1. What would disprove this?
  2. Has anyone tried to disprove it?
  3. When evidence challenges it, what happens?
  4. Has complexity increased without predictive power increasing?
  5. Could proponents and skeptics agree on a test?

The Epicycle Test

Count qualifiers. Were they predicted or reactive? Do they only accumulate? Has predictive power increased with complexity?

Epistemic Audit — Chapter 3 Addition

After this chapter, your audit should include: falsifiability assessment of core claims, identification of any epicycles, and spectrum placement for key frameworks.

What's Coming Next

Chapter 4: The Streetlight Effect — why fields study what's measurable instead of what matters. Introduces Goodhart's Law and the McNamara Fallacy.

Quick Reference:

A theory that explains EVERYTHING → explains NOTHING
A theory that predicts NOTHING → teaches NOTHING
The best theories → the ones that RISK being wrong