Pierre Duhem’s book La théorie physique: son objet, sa structure set out a logical analysis of how physical theories connect to experiments. It argued that in physics, a test checks a whole network of assumptions (about instruments, auxiliary hypotheses, and theory), not a single claim in isolation. This became a key starting point for later debates about whether evidence can uniquely determine a theory.
Karl Popper’s Logik der Forschung (later translated as The Logic of Scientific Discovery) argued that scientific method should emphasize falsifiability—whether a theory can, in principle, be refuted by observation. This sharpened attention to what it means for evidence to “count against” a theory. Duhem’s point about tests involving background assumptions later became a challenge for simple versions of Popper’s falsification story.
Hans Reichenbach articulated (and refined) the distinction between how scientists come up with ideas (discovery) and how they support them with reasons and evidence (justification). This helped focus philosophy of science on rules of testing and confirmation rather than on creativity or psychology alone. Later underdetermination debates largely operated within the “justification” side: what evidence can logically support.
W.V.O. Quine’s paper “Two Dogmas of Empiricism” rejected a sharp analytic–synthetic divide (truths “by meaning” vs. “by fact”) and also criticized reductionism (the idea that each statement can be tied to a specific set of experiences). Quine argued instead for a holistic view: statements face experience as a connected system. This expanded Duhem’s point from physics-focused testing to a broader claim about how beliefs are confirmed or disconfirmed.
Quine collected key essays (including “Two Dogmas”) into From a Logical Point of View, giving his arguments a stable, book-length platform. The collection made confirmational holism—testing and confirmation applying to bodies of belief rather than single sentences—central to postwar analytic philosophy. This consolidation helped turn underdetermination into a mainstream methodological issue, not only a point about physics experiments.
Duhem’s book became much more accessible to English-language philosophers and scientists through Philip P. Wiener’s translation, published by Princeton University Press. This helped integrate Duhem’s views on theory testing into mid-20th-century Anglophone philosophy of science. The later “Duhem–Quine” pairing was easier to make once both authors were widely read in the same circles.
In Word and Object, Quine introduced “radical translation” thought experiments and argued for the indeterminacy of translation: even with all possible behavioral evidence, more than one incompatible translation manual can fit. This gave underdetermination a new form tied to language and meaning, not only to physics experiments. It reinforced the idea that evidence may fail to uniquely fix theory, interpretation, or meaning.
Thomas Kuhn’s The Structure of Scientific Revolutions argued that mature sciences often develop through “normal science” under shared paradigms, interrupted by crises and revolutions. Competing paradigms can be difficult to compare using neutral standards, a challenge related to underdetermination: the same data may be understood differently under different frameworks. Kuhn’s historically grounded approach broadened method debates beyond purely logical models of confirmation.
Quine’s essay “Epistemology Naturalized” argued that traditional foundational epistemology had not succeeded and proposed studying knowledge as part of empirical science (especially psychology). This “naturalist” turn changed how philosophers framed method questions: instead of seeking a special, a priori justification for science, they increasingly asked how scientific inquiry actually works. Underdetermination concerns remained, but were often discussed alongside human cognition and scientific practice.
Imre Lakatos’s “Falsification and the Methodology of Scientific Research Programmes” argued that scientists typically test sequences of theories organized as “programmes,” not single hypotheses. This approach aimed to keep Popper’s emphasis on criticism while acknowledging Kuhn-like historical patterns. It also provided tools for discussing underdetermination: when faced with conflicting evidence, a programme can be adjusted in different ways, and method must judge which adjustments are better.
By the 1970s, underdetermination—especially in Quine’s stronger form—was widely treated as a central issue about scientific rationality: whether any hypothesis can be preserved by altering auxiliary assumptions, and what counts as a legitimate response to failed predictions. Major reference works summarize this as a move from Duhem’s more limited claim (focused on theoretical physics and scientific “good sense”) to Quine’s broader thesis about theory choice and evidence. The debate increasingly turned from whether underdetermination exists to how science manages it in practice (simplicity, coherence, predictive success, and other standards).
In later 20th-century philosophy of science, Duhem and Quine were increasingly discussed together under a single heading: the “Duhem–Quine” problem or thesis. The label captured a common lesson: tests confront webs of assumptions, so evidence alone may not uniquely determine which part of a theory network to revise. As a result, underdetermination became a lasting backdrop for debates about scientific method, theory choice, and what scientific evidence can logically guarantee.
Pierre Duhem and W.V.O. Quine on Underdetermination (1906–1970s)