Loading field map
How do scientists decide between competing theories? Is there a universal rational method that guarantees the best choice, or is theory choice shaped by historical context, social pressures, and personal judgment? This question has driven a rich debate in philosophy of science, producing a series of frameworks that each offer a distinctive picture of how science changes and how theories are evaluated.
Inductivism (1600–1950) held that science progresses by accumulating observations and generalizing from them. Theory choice was a matter of comparing the amount and variety of confirming evidence: a theory was better if it was supported by more inductive inferences. However, inductivism struggled with underdetermination—when multiple theories fit the same data, induction alone could not decide between them. This limitation set the stage for a more critical approach.
Critical Rationalism (1930–Present) responded to inductivism's difficulties. Karl Popper argued that science does not proceed by confirmation but by falsification. A theory is scientific if it makes risky predictions that could be falsified. Theory choice, for Popper, is about selecting the theory that has survived the most severe tests and has greater empirical content. This replaced the inductive logic of confirmation with a deductive logic of falsification. Yet critics noted that actual scientific practice rarely follows strict falsification—scientists often protect theories from refutation by adding auxiliary hypotheses. Popper's framework offered a clear demarcation criterion but left open how to choose between theories when both have survived tests.
Kuhnian Paradigm Theory (1962–Present) transformed the debate. Thomas Kuhn argued that science alternates between periods of normal science, guided by a shared paradigm, and revolutionary periods when paradigms compete. During revolutions, theory choice is not fully rational in the Popperian sense because paradigms are incommensurable—they differ in standards, concepts, and worldviews. Kuhn's account challenged the idea that a single method governs all theory choice. Instead, persuasion, community consensus, and historical contingency play central roles. This provoked a crisis for rationalist accounts: if paradigms are incommensurable, how can we ever say one theory is objectively better?
Epistemological Anarchism (1975–Present) radicalized Kuhn's insights. Paul Feyerabend argued that there is no universal method of science—"anything goes." He rejected both inductivism and falsificationism as unrealistic constraints on scientific creativity. For Feyerabend, theory choice is a matter of local, pragmatic, and even political considerations. His anarchism coexists with Kuhn's paradigm theory as a more extreme version of the same critique: where Kuhn saw revolutions as rare and structured, Feyerabend saw permanent methodological pluralism.
Methodology of Scientific Research Programmes (1970–Present) attempted to reconcile Popper's rationalism with Kuhn's historical evidence. Imre Lakatos proposed that science progresses through research programmes, each with a hard core of central assumptions and a protective belt of auxiliary hypotheses. A programme is progressive if it predicts novel facts and degenerating if it merely accommodates anomalies. Theory choice becomes a matter of comparing the progressive or degenerating status of competing programmes over time. This preserved a rational criterion while acknowledging that theories are not abandoned at the first falsification. Lakatos's framework coexists with Kuhn's as a more formal, rationalist alternative that still respects historical complexity.
Research Traditions (1977–Present) offered another reconciliation. Larry Laudan argued that science is best understood as a set of research traditions—broad frameworks that guide problem-solving. A tradition is evaluated by its effectiveness in solving empirical and conceptual problems. Laudan's approach differed from Lakatos by emphasizing problem-solving success rather than novel prediction, and by allowing for more gradual change. It also addressed Kuhn's incommensurability by suggesting that traditions can be compared in terms of their problem-solving track record. Research Traditions thus provided a more moderate rationalist alternative that absorbed some of the historical turn's insights.
Sociology of Scientific Knowledge (1970–Present) shifted attention from the logic of theory choice to the social processes that produce scientific knowledge. The Strong Programme, led by David Bloor and Barry Barnes, argued that all knowledge, including scientific, is socially constructed—shaped by interests, negotiations, and institutional contexts. This framework challenged both rationalist and historical accounts by denying that theory choice can be explained solely by evidence or method. It coexists with earlier frameworks as a radical alternative that treats scientific change as a sociological phenomenon.
Social Constructivism (1980–Present) extended the claims of SSK, arguing that scientific facts are constructed through social interactions. Laboratory studies (e.g., Latour and Woolgar's Laboratory Life) showed how scientific consensus emerges from negotiations rather than from objective evidence. Social constructivism remains in tension with rationalist frameworks: it denies that theory choice is primarily about truth or empirical adequacy, while rationalists insist that evidence plays a decisive role.
Feminist Philosophy of Science (1980–Present) expanded the debate by examining how gender biases and values influence scientific practice and theory choice. Feminist philosophers argued that traditional frameworks—both rationalist and social—overlook the ways that science can be shaped by patriarchal assumptions. Feminist approaches do not reject rationality but call for a more inclusive, value-aware rationality. They coexist with SSK and social constructivism as part of the social turn, but they also critique those frameworks for sometimes ignoring the role of epistemic values. Feminist philosophy of science thus adds a normative dimension: theory choice should be sensitive to social justice and diverse perspectives.
Today, no single framework dominates the subfield. The debate has become pluralistic, with different approaches addressing different aspects of scientific change. Critical rationalism remains influential in discussions of demarcation and testability. Kuhnian concepts like paradigm and incommensurability are widely used in history and sociology of science. Lakatos's research programmes and Laudan's research traditions offer tools for appraising scientific progress. SSK and social constructivism have shaped science and technology studies (STS). Feminist philosophy of science continues to develop accounts of values in science.
What do these frameworks agree on? Most now accept that theory choice is not a purely algorithmic process; historical and social factors matter. They also agree that scientific change is not simply cumulative—revolutions and shifts in perspective occur. Where they disagree is on the extent to which rationality can be preserved. Rationalists (Popper, Lakatos, Laudan) argue that despite social influences, there are objective criteria for comparing theories. Social and feminist approaches argue that values and interests are inescapable and that rationality must be reconceived. The legacy of each framework persists: Kuhn's work remains a touchstone for any discussion of scientific revolutions; Lakatos and Laudan provide frameworks for evaluating research programmes; SSK and feminism have permanently expanded the scope of philosophy of science to include social and political dimensions.
In recent years, Bayesian confirmation theory (from the sibling subfield of confirmation and evidence) has offered a formal tool for modeling theory choice under uncertainty, but it has not resolved the deeper disagreements about the role of values and social context. The field remains vibrant, with ongoing debates about the nature of scientific objectivity, the role of values, and the best way to understand scientific change.