Questions: The Falsifiability Criterion and Its Problems

Question 1

When Uranus showed orbital anomalies inconsistent with Newtonian predictions, Le Verrier and Adams responded by hypothesizing the existence of Neptune rather than abandoning Newton's laws. This is an example of:

• A. Falsifying Newtonian mechanics, since the core prediction had empirically failed
• B. An illegitimate ad hoc modification, because introducing unobserved entities is always epistemically dishonest
• C. The Duhem-Quine thesis in action — an auxiliary hypothesis (the complete inventory of massive bodies) was revised rather than the core theory
• D. A failure to apply Popper's falsifiability criterion, since Neptune was unobservable at the time

Answer: (C)

This is the Duhem-Quine thesis in its most instructive form. The tested bundle included Newton's laws plus the auxiliary assumption that all relevant massive bodies were accounted for. When the prediction failed, logic alone could not determine whether Newton's laws were wrong or the auxiliary was. Le Verrier and Adams correctly diagnosed the auxiliary — and Neptune's discovery vindicated them. But the logical form of the move is always available: any anomaly can, in principle, be absorbed by revising an auxiliary hypothesis. This does not automatically make such moves illegitimate — it shows that falsification is never logically compelled.

Question 2

A single experimental result clearly contradicts a well-established theory's prediction. What does the Duhem-Quine thesis imply about what this result logically establishes?

• A. The core theory is false, since it generated the failed prediction
• B. Logic alone cannot determine which element of the tested bundle — core theory, auxiliary hypotheses, or measurement assumptions — is responsible for the failure
• C. The measurement instruments must have malfunctioned, since well-established theories are presumed correct pending multiple replications
• D. Both the core theory and all auxiliary hypotheses are equally falsified by the anomalous result

Answer: (B)

The Duhem-Quine thesis says that experiments always test conjunctions of hypotheses, never single claims in isolation. When a prediction fails, the logic (modus tollens) tells you that at least one member of the bundle is false, but not which one. Scientists must use judgment, background knowledge, and further investigation to diagnose the fault — logic alone is silent. This is why Popper's simple falsificationism — one failed prediction, one falsified theory — does not accurately describe either the logic or the practice of science.

Question 3

The Duhem-Quine thesis implies that it is always epistemically possible to protect any core theory from falsification by adjusting auxiliary hypotheses, but this does not mean every such protection is legitimate.

• T. True
• F. False

Answer: True

True. The logical structure always permits rescuing the core theory by blaming an auxiliary. But this permissibility is logical, not epistemological. The question of whether a given auxiliary revision is legitimate depends on whether it generates novel testable predictions, coheres with independent evidence, or simply absorbs anomalies without explanatory gain. The Neptune hypothesis was legitimate because it generated the testable prediction 'look here' that was subsequently confirmed. Freudian post-hoc reinterpretation of any patient behavior is less legitimate because it makes no new predictions.

Question 4

A single clearly anomalous experimental result is logically sufficient to falsify a scientific theory, as Popper's falsificationism requires.

• T. True
• F. False

Answer: False

False. This is exactly what the Duhem-Quine thesis refutes. The anomalous result falsifies the entire bundle (theory + auxiliaries + measurement assumptions), not the theory alone. Logic cannot identify which component failed. In practice, a single anomalous result typically triggers investigation into whether the result is reproducible, whether the instruments were working correctly, and whether auxiliary hypotheses might be the culprit — not immediate abandonment of the theory. Popper's idealized picture does not match the logical structure of how scientific testing works.

Question 5

Why does the Duhem-Quine thesis pose a fundamental challenge to Popper's falsifiability criterion as a demarcation between science and non-science, and how does it motivate Lakatos's notion of research programs?

Model answer: Popper's criterion requires that a scientific claim be falsifiable — that there exist possible observations that could refute it. The Duhem-Quine thesis shows that no single claim is ever tested in isolation: anomalies always underdetermine which part of the tested bundle is false, and any core theory can be protected by revising auxiliaries. This means 'falsifiability' is a property of theories in bundles, not individual claims, and scientists can always maintain any core theory in the face of evidence. Lakatos responds by shifting from theories to research programs: the 'hard core' is protected by a 'protective belt' of auxiliary hypotheses that absorbs anomalies, and the program is progressive (legitimate) if it generates novel confirmed predictions, or degenerative (pseudoscientific) if it only adds epicycles to explain past anomalies.

The key insight is that the unit of scientific appraisal cannot be the individual theory — it must be something larger. Lakatos captures what Popper missed: that protecting a core theory from falsification is standard scientific practice, and the question is whether that protection is generative (leading to new discoveries) or sterile.