Claim № ii.01  · Lens · Case file

“Evolutionary theory is falsifiable: it makes risky predictions that could be shown false, and so far they have held.”

Disciplines: Philosophy of science
Status: reviewed 2026-11-15
UnderExaminationTwo claims · two grades

§1 The claim, as taught

More precisely, the common descent hypothesis would have been falsified if the universal phylogenetic trees determined from the independent molecular and morphological evidence did not match with statistical significance.Theobald DL. *29+ Evidences for Macroevolution*, TalkOrigins Archive (version 2.89, 2012), Part 1. Each of its 29 predictions is paired with a “potential falsification.”

§2 The strongest evidence offered

This card is about the logic of testing. The examples matter less for what they show about evolution than for what they show about how it is tested.

Inferred

Popper's doubt, and his recantation

Karl Popper, who made falsifiability the mark of science, wrote in 1976 that Darwinism was not a testable theory but a “metaphysical research programme.” Two years later he revised his view, saying natural selection could be tested, while still criticising the tautological form of “survival of the fittest.”
Popper, Unended Quest (1976); Popper, Dialectica 32 (1978).
Inferred

Rabbits in the Precambrian

J.B.S. Haldane is reported to have said that fossil rabbits in Precambrian rock would disprove evolution. The order of fossils is a real test, though it is a test of the fossil order and of common descent, not of the mechanism.
Widely attributed to Haldane in secondary sources; no primary written source is known.
Observed

A prediction confirmed

Searching rocks of the age their model predicted, Neil Shubin and colleagues found Tiktaalik, a fish with features linking it to tetrapods. It is often cited as a successful risky prediction.
Daeschler, Shubin & Jenkins, Nature 440 (2006)
Observed

A prediction overtaken

Four years later, tetrapod trackways were reported from Poland, about 18 million years older than Tiktaalik. The prediction's timing failed. The theory was not regarded as refuted: the tracks were absorbed by placing the origin of tetrapods earlier.
Niedźwiedzki et al., Nature 463 (2010)
Inferred

Research programmes

Imre Lakatos argued that all major theories protect a “hard core” with a belt of auxiliary hypotheses, adjusted when predictions fail. A programme is healthy while its adjustments lead to new, confirmed predictions, and degenerating when they only explain away.
Lakatos, The Methodology of Scientific Research Programmes (1978)

§3 The grade: two claims hiding in one

Specific hypotheses and the theory as a whole are tested differently. Graded separately: Hover or tap a rung to see why it got its score. How the grading works →

Specific evolutionary hypotheses make predictions that can be, and sometimes are, tested

BStrongly supported
Observation
///
Specific predictions, such as Tiktaalik's location, have been made and checked.
Hypothesis
///
Specific hypotheses are clearly stated.
Prediction
//
Some predictions confirmed, some overtaken.
Experiment
//
Field searches and lab tests, not controlled experiments on history.
Replication
//
A number of successful predictions exist.
Falsifiability
//
Individual hypotheses can and do fail.
Method score 14/18

The theory as a whole could be falsified by observation

CInference
Observation
//
Major surprises, such as the Cambrian explosion and stasis, have been observed.
Hypothesis
//
The core claim is clear: descent with modification by natural causes.
Prediction
/
Failed predictions are typically met by adjusting auxiliary assumptions.
Experiment
/
No experiment could test the whole.
Replication
//
Philosophers of science disagree about how testable the core is.
Falsifiability
/
It is hard to name an observation that the core would not absorb.
Method score 9/18
ADemonstratedBStrongly supportedCInferenceDExtrapolationEAssumption

§4 What it would take

To show that the theory as a whole is falsifiable in practice, the following would be needed:

  • A named defeaterState, in advance, an observation that would count against common descent by natural causes, and that would not be met by revising the timescale or the mechanism.
  • Adjustments that predictShow that the adjustments made after surprises (punctuated equilibria, relaxed clocks, lineage sorting) led to new predictions that were then confirmed.

§5 Another reading of the same data

Falsifiability is a fair test, and it cuts both ways. A theory that could be reconciled with any possible observation would not be telling us anything about the world. So it is reasonable to ask of evolution, and of any rival, what it rules out.

Specific evolutionary hypotheses do rule things out, and the Tiktaalik search is a fine example of a prediction made in advance and met. But the story did not end there. When older tetrapod tracks turned up, the prediction's timing was wrong, and the theory adjusted without strain. That is typical. The Cambrian explosion, stasis, the conflict between gene trees and the discovery of function in “junk” DNA were all surprises, and each was absorbed. Lakatos would say that is how all research programmes behave. The question he would ask is whether the adjustments produced new successes or only protected the core.

Read through a design lens, much of the protection comes from a commitment that sits outside science: that the explanation must be natural, because the alternative has been ruled out in advance. The next card examines that commitment. The same test applies to the site's own position. A recent-creation reading has to say what it predicts and what would count against it, and the cards on this site try to state those things in their “what it would take” sections.

§6 Reasoning check

The grade weighs the evidence. This check looks at the reasoning: the step from the evidence to the conclusion. The same tests are applied to both readings. How the reasoning check works →

In the claim as taught
  • Possible

    When predictions fail, the theory's auxiliary assumptions are adjusted. That is normal science, as Lakatos noted. Whether the adjustments led to new successes, or only explained away the failure, is what decides.

  • Possible

    If “the fittest” is defined as those that survive and reproduce, “survival of the fittest” cannot be false. Defenders reply that fitness can be defined by measured traits, and when it is, the claim becomes testable.

  • Possible
    In the evidence: Rabbits in the Precambrian

    The test concerns the order of fossils, which a Flood reading also has to explain. It is offered as a test of “evolution” as a whole, including the mechanism, which the rabbit test does not touch.

In the other reading
  • Clear case

    A creation model that answers any difficulty by appeal to miracle is no more falsifiable. If the recent-creation reading is to be scientific, it must say what it predicts and what would count against it, as the cards on this site try to do.

§7 Questions to ponder

  • If tetrapod tracks older than Tiktaalik did not count against the theory, what would?
  • Is a theory that has survived many surprises well confirmed, or well protected?
  • Apply the same test to a recent-creation model. What observation would it rule out?

§8 Sources

Supporting the mainstream claim
  1. Popper K. Unended Quest: An Intellectual Autobiography. Fontana/Collins (1976), §37.C
  2. Popper K. “Natural selection and the emergence of mind.” Dialectica 32:339–355 (1978).C
  3. Peters RH. “Tautology in evolution and ecology.” American Naturalist 110:1–12 (1976).C
  4. Lakatos I. The Methodology of Scientific Research Programmes. Cambridge University Press (1978).C
  5. Kitcher P. Abusing Science: The Case Against Creationism. MIT Press (1982).C
  6. Daeschler EB, Shubin NH, Jenkins FA. “A Devonian tetrapod-like fish and the evolution of the tetrapod body plan.” Nature 440:757–763 (2006).B
  7. Niedźwiedzki G, Szrek P, Narkiewicz K, Narkiewicz M, Ahlberg PE. “Tetrapod trackways from the early Middle Devonian period of Poland.” Nature 463:43–48 (2010).B
  8. Sober E. Evidence and Evolution: The Logic Behind the Science. Cambridge University Press (2008).C
Critical & alternative perspectives
  1. Hunter CG. Darwin's God: Evolution and the Problem of Evil. Brazos (2001).C

Source grades are provisional. In production each links to a full published evaluation that readers can challenge.

§9 Your notes