A barometer drops; a storm follows. We can derive 'storm will occur' from the barometer reading plus physical laws. Why does citing the barometer NOT constitute a causal explanation of the storm?
ABecause the derivation uses inductive rather than deductive logic, which is insufficient for scientific explanation
BBecause the barometer reading does not produce the storm through any mechanism — both are effects of the same atmospheric pressure change, making the barometer causally inert with respect to the storm
CBecause the laws connecting barometers to storms are not universal enough to support explanation
DBecause explanation requires knowing the storm's precise location and timing, which the barometer cannot provide
This is the canonical case showing that the covering-law (D-N) model is insufficient. We can derive the storm from the barometer reading plus the correct physical laws — the D-N conditions are satisfied — yet no one thinks the barometer explains the storm. What's missing is causal relevance: the causal arrow runs from low atmospheric pressure to both the barometer reading and the storm. The barometer doesn't produce the storm; they are parallel effects of the same common cause. Genuine explanation requires tracing the real mechanism.
Question 2 Multiple Choice
Two assassins simultaneously fire lethal shots at a target. The target dies. Had Assassin A not fired, Assassin B's shot would have killed the target; had Assassin B not fired, A's would have. Under simple counterfactual causation, neither assassin caused the death. What does this illustrate?
ACounterfactual causation correctly shows that joint actions cannot have singular causes
BOverdetermination cases show that simple counterfactual dependence fails: both shots were causes, yet neither passes the 'had it not occurred' test — more refined accounts are needed
CThe counterfactual test works here: since both fired, the correct analysis is that each caused 50% of the death
DThis shows that causal explanation cannot handle intentional human action and only applies to physical events
Overdetermination is a genuine challenge for counterfactual accounts: when two independently sufficient causes both operate, neither is necessary (removing either leaves the other to produce the effect), so counterfactual dependence fails for both — yet intuitively both are causes. Resolutions include INUS conditions (each is an insufficient but necessary part of an unnecessary but sufficient condition), causal graph models (using actual vs. counterfactual causation), or probabilistic causation. These refinements preserve the core insight — explanation requires real causal mechanism — while handling complex causal structures.
Question 3 True / False
The counterfactual account of causation — 'A caused B' means 'had A not occurred, B would not have occurred' — is sensitive to whether the causal mechanism was actually operative, unlike mere correlation.
TTrue
FFalse
Answer: True
This is the central advantage of the counterfactual approach over pure correlation. The barometer correlates perfectly with the storm, but the counterfactual test reveals the asymmetry: had the barometer not dropped, the storm would still have come (because atmospheric pressure was already low). The counterfactual test correctly identifies that the barometer has no causal purchase on the storm. For the match: had it not been struck, it would not have lit — the counterfactual holds, confirming the striking as a genuine cause.
Question 4 True / False
The deductive-nomological (covering-law) model is a complete account of scientific explanation because any successful derivation of a phenomenon from universal laws and initial conditions constitutes a genuine causal explanation.
TTrue
FFalse
Answer: False
The barometer case directly refutes this: we can derive 'storm will occur' from the barometer reading plus physical laws via a valid deductive argument, but this is not a causal explanation of the storm — the barometer is causally inert with respect to the storm. The D-N model cannot distinguish genuinely explanatory laws from accidental correlations that happen to track the same underlying cause. This is the 'irrelevance' problem: D-N allows derivations using causally irrelevant factors that intuitively do not explain.
Question 5 Short Answer
Why does citing a factor that merely correlates with an effect fail to constitute a causal explanation? Use either the barometer or the match example to illustrate the difference between correlation and causal mechanism.
Think about your answer, then reveal below.
Model answer: Causal explanation requires that the cited factor actually produces the effect through a real mechanism — a physical process connecting cause to effect. The barometer correlates perfectly with the storm but doesn't produce it; both are effects of low atmospheric pressure. Tracing the mechanism for the match — friction → heat → combustion temperature exceeded → chemical reaction with oxygen — shows each step as a real physical process. A correlation that tracks the same cause without being part of the causal chain from cause to effect is explanatorily empty.
The deeper point is about causal relevance: a genuine explanation must identify factors that would change the outcome if they were changed, and do so because they are part of the mechanism — not merely because they co-vary with something in the mechanism. This is why randomized controlled trials are the gold standard in science: they establish counterfactual dependence by intervening on the putative cause, ruling out common-cause explanations. The philosophical account of causal explanation directly motivates the methodology of experimental science.