Instrumentalism is the philosophical position that scientific theories are not descriptions of an underlying reality but tools-instruments-for predicting observable phenomena, organizing experience, and guiding practical intervention in the world. Holders of this view contend that the question of whether a theory is “true” in some metaphysically robust sense is either unanswerable or meaningless; what matters is whether the theory works. Instrumentalism is held by a significant strand of philosophers of science, practicing scientists, and pragmatist philosophers, and has historically served as a methodological stance among scientists who wish to bracket metaphysical commitments while pursuing empirical research.
Theories are calculational devices, not pictures of reality. Instrumentalists argue that a scientific theory is best understood as a formal apparatus-a set of equations, rules, and procedures-that takes observational inputs and yields observational outputs. Pierre Duhem, one of the earliest systematic defenders of this view, held that physical theory “is not an explanation” of natural phenomena but “a system of mathematical propositions, deduced from a small number of principles, which aim to represent as simply, as completely, and as exactly as possible a set of experimental laws.”1) On this account, the internal machinery of a theory-atoms, fields, forces, curved spacetime-need not correspond to anything real; it is scaffolding that holds the empirical predictions together.
The underdetermination of theory by evidence. Instrumentalists point to the logical fact that any finite body of observational evidence is compatible with multiple, mutually inconsistent theoretical accounts. If evidence cannot uniquely determine which theory is true, they argue, the demand for theories to be literally true is unreasonable. The appropriate response is to choose the theory that is most useful-most predictively powerful, most parsimonious, most fruitful for further research-without pretending the choice settles metaphysical questions.
The pessimistic meta-induction cuts against realism. Instrumentalists frequently invoke the history of science as evidence against scientific realism. Theories once regarded as literally true-caloric fluid, phlogiston, the luminiferous ether, Newtonian absolute space-were eventually abandoned. If past successful theories have been radically wrong about the entities they postulated, instrumentalists ask, why should confidence in the literal truth of present theories be warranted? Better to regard success as a mark of instrumental adequacy, not verisimilitude.
Empirical adequacy is the proper goal of science. Bas van Fraassen's constructive empiricism, a close relative of instrumentalism, holds that a theory need only be “empirically adequate”-it need only “save the phenomena,” correctly representing the observable parts of the world-to fulfill science's legitimate aims.2) Instrumentalists broadly share this emphasis: science aims to control and predict experience, not to achieve a God's-eye view of a theory-independent world.
Pragmatic virtues are the real criteria of theory choice. Instrumentalists hold that the virtues scientists actually employ-simplicity, predictive accuracy, scope, fruitfulness, coherence with other accepted theories-are pragmatic virtues, not truth-tracking ones. A theory is chosen because it is a better instrument, not because it more closely mirrors the deep structure of nature.
Instrumentalist impulses appear as early as ancient astronomy, where the Ptolemaic system was sometimes defended not as a literal description of the heavens but as a mathematical device for calculating planetary positions. The prefatory remarks (whose authorship was long disputed) to Copernicus's De revolutionibus (1543), likely written by Andreas Osiander, famously urged readers to treat the heliocentric model as a calculating convenience rather than a physical claim-an early and influential statement of the instrumentalist posture in the context of the Copernican revolution.3)
In the early modern period, instrumentalism became entangled with debates over Newtonian mechanics. Newton's own famous declaration hypotheses non fingo-“I feign no hypotheses”-regarding the cause of gravity was read by many as an instrumentalist concession: the mathematics of universal gravitation worked magnificently, but Newton declined to assert what gravity was at a deeper level. Subsequent interpreters of Newtonian gravity as instrumentalism elaborated this reading into a broader methodological stance.4)
The late nineteenth and early twentieth centuries saw instrumentalism take systematic philosophical form. Ernst Mach argued that science is economical description of sensations, and that entities such as atoms-which he long refused to accept as real-were at best hypothetical constructs useful for calculation.5) Duhem developed a sophisticated philosophy of science centered on the view that physical theories aim at mathematical representation of experimental laws, not causal explanation of hidden realities. John Dewey and the American pragmatist tradition gave instrumentalism a broader philosophical grounding, tying the value of ideas to their consequences in experience and action.
In the mid-twentieth century, instrumentalism influenced the logical empiricists of the Vienna Circle, though thinkers like Rudolf Carnap preferred to dissolve metaphysical questions about theoretical entities rather than simply bracket them. The position became a default working philosophy for many physicists grappling with quantum mechanics, where the famous Copenhagen interpretation-associated with Niels Bohr and Werner Heisenberg-was widely read as counseling physicists not to ask what quantum states “really are” between measurements but to use the formalism to predict outcomes. The slogan “shut up and calculate,” though of disputed and later origin, captured the spirit.6)
Van Fraassen's The Scientific Image (1980) gave instrumentalism renewed philosophical rigor under the label constructive empiricism, distinguishing carefully between the observable and unobservable, and arguing that belief in the empirical adequacy of a theory is sufficient for rational acceptance.
Pierre Duhem (1861-1916): French physicist and historian of science whose The Aim and Structure of Physical Theory remains the most systematic early defense of the view that physical theory aims at mathematical representation, not causal explanation. Duhem also argued-in what became known as the Duhem-Quine thesis-that individual hypotheses cannot be tested in isolation, since any test requires auxiliary assumptions.
Ernst Mach (1838-1916): Austrian physicist and philosopher who insisted science is economical description of sensations. His skepticism about unobservable entities such as atoms influenced a generation of physicists and philosophers, including the young Einstein (who later broke with Mach's anti-atomism).
John Dewey (1859-1952): American pragmatist philosopher who developed a broad instrumentalism about ideas generally, arguing that concepts and theories are tools for solving problems arising in experience, evaluated by their practical consequences.
Niels Bohr (1885-1962): Danish physicist whose Copenhagen interpretation of quantum mechanics is widely regarded as instrumentalist in spirit. Bohr held that quantum mechanics describes the outcomes of experimental interactions, not an observer-independent microscopic reality.
Bas van Fraassen (b. 1941): Dutch-American philosopher of science whose constructive empiricism defends a rigorous version of the view that science aims at empirical adequacy rather than truth about unobservables. His The Scientific Image is the most cited contemporary statement of the anti-realist position in philosophy of science.
Thomas Kuhn (1922-1996): Though not a strict instrumentalist, Kuhn's account of paradigm shifts-in which incommensurable frameworks succeed one another without obvious convergence on truth-is frequently marshaled by instrumentalists as evidence against the realist picture of cumulative progress toward truth.7)
Instrumentalism is not monolithic, and its proponents disagree on several important questions.
How to draw the observable/unobservable line. Van Fraassen's constructive empiricism depends on a distinction between what is observable by the unaided senses and what is not. Critics within the broadly anti-realist camp argue that this line is drawn too arbitrarily-that a cell seen through a microscope is not categorically different from one seen with the naked eye-and that the position requires either a principled account of observability or a retreat toward more thoroughgoing instrumentalism that abandons the distinction.
Whether instrumentalism collapses into pragmatism or idealism. Some instrumentalists hold that theories represent nothing beyond patterns in experience; others argue that this collapses the position into a form of idealism that denies the independence of the world. Pragmatist instrumentalists, following Dewey, are comfortable treating inquiry as interaction between organism and environment without positing mind-independent realities; more deflationary instrumentalists prefer to remain silent on such questions.
The status of mathematics. Some instrumentalists are troubled by the role of abstract mathematical structures in theories. If the equations of quantum field theory are mere instruments, what explains the unreasonable effectiveness of mathematics in natural science? Some instrumentalists respond that mathematical effectiveness is itself a pragmatic achievement; others find this answer thin.
Convergent realism vs. instrumentalism about mature science. Some proponents of instrumentalism are willing to grant that highly mature, stable theories-classical mechanics within its domain, thermodynamics-may describe genuine regularities, while insisting that the deeper theoretical posits remain mere instruments. Others hold the instrumentalist stance uniformly across all theory levels.
<footnotes />