Scientific Explanation (the DN Model)

The Core Idea: The Covering-Law Model

According to the DN model, a scientific explanation consists of two components: the explanans, comprising statements of general laws and initial conditions, and the explanandum, a statement of the event to be explained. An explanation is a valid deductive argument in which the explanans logically entails the explanandum. For instance, the fact that water boils below 100 °C at high altitude is derived from the ideal gas laws, initial conditions of atmospheric pressure, and general thermodynamic principles. Hempel and Oppenheim contend that explanation and prediction are structurally identical: the same argument functions as an explanation when viewed retrospectively and as a prediction when viewed prospectively — a claim known as the symmetry thesis.

Key Concepts: Adequacy Conditions and the IS Variant

Hempel proposes four conditions for an adequate explanation: (1) the explanans must logically entail the explanandum; (2) the explanans must contain at least one general law; (3) the explanans must have empirical content; and (4) the sentences of the explanans must be true. The DN model covers only universal laws. For cases involving probabilistic laws, Hempel proposes the inductive-statistical (IS) model: the explanandum is inductively derived from a statistical law as a highly probable outcome. The IS model proved contentious in its own right, since the same event can receive rival IS explanations depending on the reference class invoked.

Criticisms: Asymmetry and Irrelevance Objections

The most powerful objections to the DN model concern asymmetry and irrelevance. The asymmetry problem is illustrated by Sylvain Bromberger's flagpole example: given the sun's elevation and the length of a flagpole's shadow, one can derive the pole's height in a DN-valid argument — yet the shadow does not explain the pole's height. The irrelevance objection is highlighted by Wesley Salmon's well-known example: the fact that a man who regularly takes birth-control pills does not become pregnant can be deduced from the covering law that such pills prevent pregnancy — yet the law is causally irrelevant to this individual. These objections demonstrated that derivability from a law is insufficient to capture genuine causal explanation.

Significance and Influence on Subsequent Theories of Explanation

The DN model remains the most systematic and influential attempt to analyse scientific explanation. The debates it provoked directly catalysed the emergence of the philosophy of explanation as a distinct research programme. Wesley Salmon's causal-mechanical theory grounds explanation in causal processes rather than laws. Philip Kitcher's unificationist theory identifies explanation with the capacity to systematise knowledge by reducing the number of independent argument patterns. Both alternatives were developed as direct responses to the gaps left by the DN model. In the social sciences and historiography, the DN model continues to serve as the primary reference point when debating the adequacy of functional or interpretive forms of explanation.

Key thinkers

• Carl G. Hempel (1905–1997)Principal architect of the DN model; German-American philosopher of science who formalised scientific explanation as logical deduction from covering laws.
• Paul Oppenheim (1885–1977)German-born American scientist who co-authored the landmark 1948 paper with Hempel and contributed to defining the formal adequacy conditions for scientific explanation.

Sources