What Is This Thing Called Science?

1. Science progresses through conjectures and refutations, not proof.

We learn from our mistakes; and that in finding that our conjecture was false we shall have learnt much about the truth, and shall have got nearer to the truth.

Falsification over verification. Karl Popper's falsificationism posits that science advances by proposing bold conjectures and then rigorously attempting to refute them. Unlike inductivism, which seeks to prove theories, falsificationism emphasizes the importance of identifying and eliminating false theories. This approach acknowledges the inherent limitations of induction and the impossibility of definitively proving scientific laws.

Trial and error. Scientific progress is a process of trial and error, where theories are constantly tested and refined. Theories that withstand repeated attempts at falsification are considered stronger, but never proven true. This ongoing cycle of conjecture and refutation drives the advancement of scientific knowledge.

Embracing bold conjectures. Falsificationism encourages scientists to propose daring and speculative theories, even if they seem unlikely or counterintuitive. The more falsifiable a theory is, the more informative it is, as it makes definite claims about the world that can be tested. This emphasis on boldness contrasts with the inductivist's cautious approach, which favors theories that are already well-supported by evidence.

2. Observation is active, public, and fallible, not passive and private.

Care in making the observations cannot make up for want of space.

Active intervention. Observation in science is not a passive reception of sensory data, but an active process involving practical intervention and manipulation. Scientists actively design experiments, control variables, and use instruments to explore and test their hypotheses. This active approach allows them to go beyond mere observation of naturally occurring phenomena.

Public scrutiny. Scientific observations are not private, subjective experiences, but public and objective claims that can be checked, criticized, and added to by others. Scientists communicate their findings through publications, presentations, and collaborations, allowing for peer review and independent verification. This public nature of scientific knowledge ensures its reliability and objectivity.

Fallibility of observation. Observation statements are not infallible, but are subject to error and revision. Judgments about the adequacy of observation statements depend on presupposed knowledge and can be influenced by the observer's background and expectations. This fallibility highlights the need for rigorous testing and critical evaluation of observational claims.

3. Experimental results, not just facts, form the basis of science.

To acquire facts relevant for the identification and specification of the various processes at work in nature it is, in general, necessary to practically intervene to try to isolate the process under investigation and eliminate the effects of others.

Relevance matters. Science requires relevant facts, not just any facts. The vast majority of observable facts are irrelevant to scientific inquiry. Relevant facts are those that can help answer specific questions or solve particular problems within a scientific discipline.

Intervention is key. To understand the processes at work in nature, it is often necessary to intervene practically to isolate the process under investigation and eliminate the effects of other factors. This is the essence of experimentation. Experimental results, rather than mere observations, provide the factual basis for science.

Experimental results are not straightforward. Experimental results are not simply given via the senses, but must be worked for. Establishing experimental results involves considerable know-how, practical trial and error, and the exploitation of available technology. Experimental results are also fallible and can be updated or replaced as technology and understanding advance.

4. Inductive arguments can't logically derive scientific laws.

Logic can simply reveal what follows from, or what in a sense is already contained in, the statements we already have to hand.

Limitations of deduction. Logic is concerned with deduction, the derivation of statements from other statements. While logic is truth-preserving, it cannot establish the truth of factual statements. Logic can only reveal what follows from, or is already contained in, the statements we already have.

Induction vs. deduction. Scientific laws are general statements, while observation statements are specific claims about particular events. Inductive arguments proceed from a finite number of specific facts to a general conclusion. Such arguments are not logically valid, as the conclusion goes beyond what is contained in the premises.

The problem of induction. Scientific knowledge cannot be derived from the facts if "derive" is interpreted as "logically deduce." Inductive arguments, which proceed from a finite number of specific facts to a general conclusion, are not logically valid. This poses a fundamental problem for the idea that science is based on the facts.

5. Falsifiability is a key criterion for scientific theories.

If a theory is to have informative content, it must run the risk of being falsified.

The essence of falsifiability. A scientific hypothesis must be falsifiable, meaning that there exists a logically possible observation statement or set of observation statements that are inconsistent with it. This criterion distinguishes scientific theories from non-scientific ones.

Informative content. Falsifiable hypotheses are informative because they make definite claims about the world, ruling out ways in which it could possibly behave but in fact does not. Unfalsifiable statements, on the other hand, tell us nothing about the world.

Examples of falsifiable and unfalsifiable statements:

• Falsifiable: "All swans are white" (can be falsified by observing a black swan)
• Unfalsifiable: "Either it is raining or it is not raining" (true regardless of the weather)

6. Sophisticated falsificationism emphasizes novel predictions and theory comparison.

The more a theory claims, the more potential opportunities there will be for showing that the world does not in fact behave in the way laid down by the theory.

Relative falsifiability. Sophisticated falsificationism shifts the focus from the merits of a single theory to the relative merits of competing theories. A newly proposed theory is acceptable if it is more falsifiable than its rival, especially if it predicts a new kind of phenomenon not touched on by its rival.

Ad hoc modifications. Modifications to a theory that are designed merely to protect it from falsification are rejected. Acceptable modifications are those that increase the theory's falsifiability and lead to new testable consequences.

Confirmation of novel predictions. Confirmations of novel predictions are crucial in the falsificationist account of the growth of science. A newly proposed theory must make some novel predictions that are confirmed before it can be regarded as an adequate replacement for a falsified theory.

7. Kuhn's paradigms highlight the structure of scientific revolutions.

Science starts with problems, problems associated with the explanation of the behaviour of some aspects of the world or universe.

Paradigms and normal science. Thomas Kuhn's account of science emphasizes the role of paradigms, which are the general theoretical assumptions, laws, and techniques adopted by a scientific community. Normal science involves working within a paradigm to articulate and develop it, solving puzzles and accommodating the behavior of the real world.

Crisis and revolution. When anomalies become serious and undermine confidence in the paradigm, a crisis state develops. This crisis is resolved when an entirely new paradigm emerges and attracts the allegiance of more and more scientists, leading to a scientific revolution.

Incommensurability. Kuhn argues that proponents of rival paradigms "live in different worlds" and subscribe to different sets of standards and metaphysical principles. This makes it difficult to compare paradigms logically, as there is no neutral ground for evaluation.

8. Lakatos's research programs balance core stability with belt flexibility.

The positive heuristic consists of a partially articulated set of suggestions or hints on how to change, develop, the 'refutable variants' of the research program, how to modify, sophisticate, the 'refutable' protective belt.

Hard core and protective belt. Imre Lakatos's methodology of scientific research programs (MSRP) proposes that a science consists of a hard core of fundamental principles and a protective belt of supplementary assumptions. The hard core is considered unfalsifiable, while the protective belt is modified to accommodate new evidence.

Heuristics. Research programs are guided by a negative heuristic, which advises against tampering with the hard core, and a positive heuristic, which provides guidance on how to develop the protective belt. The positive heuristic maps out a program of research, including the development of mathematical and experimental techniques.

Progressive vs. degenerating programs. A progressive research program is one that leads to novel predictions that are confirmed, while a degenerating program is one that fails to do so. The replacement of a degenerating program by a progressive one constitutes Lakatos's version of a scientific revolution.

9. Feyerabend's anarchism challenges the notion of a fixed scientific method.

What remains are aesthetic judgments, judgments of taste, metaphysical prejudices, religious desires, in short, what remains are our subjective wishes: science at its most advanced and general returns to the individual a freedom he seems to lose in its more pedestrian parts.

Against method. Paul Feyerabend argues that there is no fixed scientific method and that science does not possess features that render it necessarily superior to other forms of knowledge. His anarchistic theory of knowledge proposes that "anything goes" in science.

Galileo as an example. Feyerabend uses Galileo's innovations in physics and astronomy to illustrate his point. He argues that Galileo did not follow any fixed scientific method, but instead used propaganda and trickery to persuade his contemporaries to accept his views.

Advocacy of freedom. Feyerabend's anarchistic account of science is situated in an ethical framework that places a high value on individual freedom. He argues that the institutionalization of science in our society is inconsistent with this freedom.

10. Methodical changes in method are essential for scientific progress.

The idea that science can, and should, be run according to fixed and universal rules is both unrealistic and pernicious.

Against universal method. There is no universal, ahistorical method of science that contains standards that all sciences should live up to. The idea of a universal and ahistoric method is implausible and even absurd.

Piecemeal change. There are methods and standards in science, but they can vary from science to science and can be changed, and changed for the better, within a science. At any stage in its development, a science will consist of some specific aims, methods, standards, theories, and observational facts, each of which is subject to revision.

Galileo's change in standards. Galileo's acceptance of telescopic data over naked-eye data represents a change in the standards of science. By making a rational case for this change, Galileo was able to convince his opponents and bring about a change in the way science was done.

11. Bayesianism offers a probabilistic approach to scientific inference.

I have always stressed the need for some dogmatism: the dogmatic scientist has an important role to play.

Bayes' theorem. Bayes' theorem is a theorem in probability theory that prescribes how probabilities are to be changed in the light of new evidence. It is used to calculate the probability of a hypothesis in the light of evidence, taking into account prior probabilities and the likelihood of the evidence given the hypothesis.

Subjective Bayesianism. Subjective Bayesians interpret probabilities as subjective degrees of belief. They argue that this interpretation can do full justice to science and that the beliefs of individual scientists can be made to converge given the appropriate input of evidence.

Applications of the Bayesian formula. The Bayesian approach can be used to capture and sanction typical modes of reasoning in science, such as the law of diminishing returns when testing a theory against experiment and the rationale for retaining a hard core in the face of apparent falsifications.

12. The new experimentalism emphasizes the independent life of experiment.

I can therefore gladly admit that falsificationists like myself much prefer an attempt to solve an interesting problem by a bold conjecture, even (and especially) if it soon turns out to be false, to any recital of a sequence of irrelevant truisms.

Experiment with a life of its own. The new experimentalism emphasizes the independent life of experiment, arguing that experimental results can be substantiated and experimental effects produced by an array of strategies involving practical interventions, cross-checking, and error control and elimination in a way that can be independent of high-level theory.

Learning from error. The new experimentalists focus on the positive role played by error detection in science. Experiments that serve to detect an error in some previously accepted assertion serve a positive as well as a negative function, positively identifying an effect not previously known.

Limitations of the new experimentalism. While the new experimentalism has brought philosophy of science down to earth in a valuable way, it would be a mistake to regard it as the complete answer to our question about the character of science. Experiment is not so independent of theory as the emphasis of the new experimentalism might suggest.

Last updated: February 14, 2025