TL;DR: Pseudoscience is a doctrine or practice that presents itself as science but lacks the core epistemic qualities of science, especially the mechanism of self-correction in the face of contrary evidence. Drawing the line between science and pseudoscience is difficult because both the concept of science and the concept of pseudoscience are conventional and value-laden, and no single criterion is accepted by all philosophers. Even Popper's falsifiability is not enough: Kuhn (1970) and Thagard (1978) point out that astrology did make predictions and did fail, so its weakness is not that it cannot be refuted, but that its failures did not generate research puzzles for a community of interpreters and did not lead to theoretical progress. Hansson (2009) proposes a narrow definition at the level of individual claims, while Laudan (1983) argues that all demarcation criteria fail and should be replaced by the question of whether a belief has good evidence. The multicriteria approach is currently seen as the most pragmatic, though it does not give a definitive answer for every borderline case.
What is pseudoscience, and why is drawing the line between science and pseudoscience far harder than it looks
Why is this question more important than it appears?
Distinguishing science from pseudoscience is not just an intellectual game for philosophers. It is a question with direct practical consequences, and in many fields those consequences are a matter of life and death.
According to the Stanford Encyclopedia of Philosophy's entry on science and pseudoscience, there are at least six areas where this distinction becomes a prerequisite for making the right decision:
- Climate policy: the scientific consensus on human-caused climate change is clear, yet science denial has delayed climate action for decades.
- Health care: medical pseudoscience produces interventions that do not work and are sometimes dangerous, and it draws patients away from evidence-based medicine.
- Environmental policy: we need to distinguish precautionary measures based on valid but incomplete evidence from measures responding to a threat for which there is no evidence at all.
- Expert testimony in court: legal systems need to determine the reliability of different kinds of evidence, and sometimes one side wants to present a non-scientific view as solid science.
- Science education: advocates of certain forms of pseudoscience, creationism in particular, seek to get their teachings into the school curriculum.
- Journalism: we need to distinguish legitimate scientific debate from an effort to sell a pseudoscientific view as if it were a valid scientific one.
What these areas have in common is this: without criteria for recognising pseudoscience, we will treat it like any other mistaken belief, and that is methodologically wrong. Pseudoscience is not merely wrong; it also comes equipped with a mechanism that makes it immune to refutation.
How does pseudoscience differ from bad science and from non-science?
The short answer: pseudoscience is non-science that presents itself as science. It is not the same as bad science, nor the same as neutral non-science.
SEP distinguishes three categories that are often confused. Non-science is made up of fields that lie outside the scope of science but do not pretend to be science, literature or music for example. Bad science is science practised poorly, but it still operates by scientific norms and is capable of correcting itself. Pseudoscience is a doctrine or practice that is both not science and presents itself as science.
The key difference lies in the mechanism of defence. Compare a palaeontologist who misidentifies a dinosaur femur with a creationist who believes humans once lived alongside Tyrannosaurus rex. The palaeontologist who is wrong will listen to colleagues, discuss the matter in depth, and if the argument is strong enough, he will be convinced. A typical creationist, by contrast, will sidestep the crucial counterarguments and fall back on the escape routes and evasive strategies already built into his belief system. It is precisely this self-immunising mechanism that makes pseudoscience impossible to refute through ordinary scientific argument.
Why is the word "science" itself so hard to define?
The first difficulty in the demarcation problem is that the concept of "science" is not nearly as clear as we assume. The everyday use of the word is both descriptive and normative.
When an activity is recognised as science, that recognition usually implies that it plays a positive role in the effort to seek knowledge. But the concept of science was also formed through a historical process, with many contingent factors affecting what we call science and what we do not. Whether we call a claim or a doctrine "scientific" depends both on its subject area and on its epistemic quality. The first part is highly conventional, the second is highly normative and tightly bound up with foundational questions of epistemology and metaphysics.
There is one notable linguistic difference. The English word "science" is used mainly for the natural sciences and fields considered similar to them. Political economy and sociology count as sciences, while literary studies and history usually do not. The German word "Wissenschaft", meanwhile, has a much broader meaning, covering every academic discipline, the humanities included. SEP observes that the German term draws the line better around the kind of systematic knowledge at issue in the dispute between science and pseudoscience, because the historical distortions presented by Holocaust deniers are very similar in nature to the scientific distortions presented by creationists and homeopaths.
What does Popper's criterion of falsifiability actually solve?
Karl Popper proposed falsifiability as the demarcation criterion, and it remains the most widely discussed one. He set it out in Logik der Forschung (1934, printed 1935), with an expanded English edition in 1959 titled The Logic of Scientific Discovery.
The core idea is the logical asymmetry between confirmation and refutation. To confirm the claim "all swans are white", we would have to observe every swan, which is practically impossible. But observing a single black swan is enough to refute the claim. The logical structure here is modus tollens: from law L we derive Q, but we observe not-Q, therefore L is false.
According to Thornton, S., Karl Popper, Stanford Encyclopedia of Philosophy (revised 31 July 2026), Popper also stressed that falsifiability is a logical criterion, not a methodological one. He distinguished it clearly from falsificationism, the methodological approach in which the scientist actively seeks evidence that refutes a theory. The purpose of falsifiability is to make a theory predictive, testable and useful in practice.
According to Popper, K. (1963), Conjectures and Refutations: The Growth of Scientific Knowledge, London: Routledge & Kegan Paul, chapter 1 "Science: Conjectures and Refutations" (a lecture at Peterhouse, Cambridge, 1953), Popper's personal motivation came from his experience with psychoanalysis in the 1910s, especially with Alfred Adler. Whatever observation was put forward, psychoanalysis could explain it. The reason it could explain everything was that it ruled nothing out. Popper held that this was a failure, because such a criterion cannot lead to any prediction.
Why has the falsifiability criterion been criticised?
Popper's criterion is logically neat but runs into serious problems when applied to real science.
According to Stanford, K., Underdetermination of Scientific Theory, Stanford Encyclopedia of Philosophy, the Duhem, Quine thesis holds that decisive experimental refutation is impossible, and that no scientific hypothesis on its own can yield a prediction, because testing a hypothesis requires background assumptions. In other words, when a prediction fails to match observation, we cannot be sure whether the fault lies with the central hypothesis or with one of the many auxiliary assumptions. According to Thornton, S., Karl Popper, Stanford Encyclopedia of Philosophy (revised 31 July 2026), Popper replied that falsifiability is a logical criterion, and that empirical research has the Duhem problem and others besides, but that this does not weaken the logical criterion.
Another criticism targets Popper's own dismissal of Marxism as non-science on the grounds that it resists negative evidence. This is a methodological stance, yet the problems with that stance are presented as limitations of falsifiability.
Finally, many critics argue that falsifiability offers no comparable proposal for justifying theories or research programmes, and supports only a methodology based on critical discussion. The efforts of Russell, the Vienna Circle, Lakatos and others to establish a rigorous justification for scientific theory are something the author has not been able to verify against primary sources, and so they are noted here only as a remark left open.
What do Kuhn and Thagard add to the falsifiability criterion?
Both authors take astrology as their central example, and both hold that astrology's weakness does not lie in being irrefutable.
According to Thomas S. Kuhn (1970), Logic of Discovery or Psychology of Research?, in I. Lakatos and A. Musgrave (eds.), Criticism and the Growth of Knowledge, Cambridge University Press, pp. 1-24, astrology did make predictions and did get them wrong, so falsifiability alone cannot rule it out of science. What Kuhn stresses is that astrology's weakness is not that it cannot be refuted, but that its failures generated no research puzzles for a practising community to go on solving. Kuhn also notes that ancient astrologers were not frauds: they believed in the basis of their craft and worked within the framework of a rule-governed tradition.
According to Paul R. Thagard (1978), Why Astrology is a Pseudoscience, PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association 1978(1): 223-234, a field is a pseudoscience when it is less progressive than alternative theories, accumulates anomalies, and its practitioners make no attempt to develop the theory or to compare it with other theories. Thagard holds that astrology meets these conditions. This criterion does not require astrology to be irrefutable; it requires it to answer for its own lack of progress.
What do the various criteria have in common and where do they differ?
The demarcation criteria proposed over the history of the philosophy of science can be summarised as follows:
| Criterion | Core content | Main proponents | Main limitation |
|---|---|---|---|
| Verifiability | A claim must be capable of confirmation by observation | Logical positivists | Also rules out scientific claims that cannot yet be tested |
| Falsifiability | A theory must rule out some observations | Karl Popper | Runs into the Duhem, Quine thesis |
| Puzzle-solving | A doctrine must be able to solve its internal problems | Thomas Kuhn | Hard to apply to non-empirical fields |
| Scientific progress | A field must show cumulative progress | Various authors, among them Paul R. Thagard | Progress is hard to measure and easily abused |
| Epistemic norms | Conformity with the epistemic norms of the scientific community | Various authors | Norms change over time and across fields |
| Reliability of the claim | A claim in a scientific field that is so unreliable it cannot reasonably be believed, yet is presented as scientific knowledge | Sven Ove Hansson | The line between "unreliable" and "not yet sufficiently evidenced" can blur |
SEP observes that there is far more agreement on particular cases than on the general criteria those judgements are supposed to rest on. This is a sign that important philosophical work remains to be done on the relationship between science and pseudoscience.
How many demarcation criteria have been proposed?
According to Angelo Fasce (2017), What do we mean when we speak of pseudoscience? The development of a demarcation criterion based on the analysis of twenty-one previous attempts, Disputatio. Philosophical Research Bulletin 6(7): 459-488, twenty-one criteria have been proposed in the philosophical literature. Fasce reviews each of them, and the comparison table in the section above is a condensed reading of that list. The fact that the number reaches twenty-one is part of the answer to why the demarcation problem is hard: not because criteria are lacking, but because there are too many of them and they are not fully compatible with one another.
Do all philosophers think a demarcation criterion is needed?
No. This is a point many writings on pseudoscience pass over: there is a tradition that opposes the demarcation project itself.
According to Larry Laudan (1983), The Demise of the Demarcation Problem, in R. S. Cohen and L. Laudan (eds.), Physics, Philosophy and Psychoanalysis, Boston Studies in the Philosophy of Science 76, Springer, pp. 111-127, every demarcation criterion ever proposed has failed. Laudan argues that we should replace the question "is this science?" with a more pragmatic one: is this belief well evidenced, and is it worthy of belief. For Laudan, "pseudoscience" is not a useful epistemological category but merely a term of abuse.
Laudan's position stands in direct opposition to that of Sven Ove Hansson (2009), Cutting the Gordian Knot of Demarcation, International Studies in the Philosophy of Science 23(3): 237-243. For Hansson, a claim is pseudoscientific when it belongs to a scientific field, is so unreliable that it cannot reasonably be believed, and is nevertheless presented as scientific knowledge. This is a narrow definition: it does not require demarcating an entire field, only assessing individual claims. This framing sidesteps some of Laudan's criticisms, because it does not try to find a criterion that demarcates fields.
The opposition between Laudan and Hansson shows that the demarcation problem is not merely a technical matter but a philosophical question still under debate.
How does the multi-criteria approach work?
Since no single criterion holds up, many philosophers turn to a multi-criteria approach: draw up a list of indicators and assess a doctrine by how many of them it violates.
A typical list usually includes indicators such as: claims that cannot be tested or refuted, the absence of a self-correcting mechanism, reliance on personal testimony rather than public evidence, the use of vague or invented terminology, resistance to scrutiny by the community, claims isolated from the rest of scientific knowledge, and the use of strategies that immunise the doctrine against refutation.
The advantage of this approach is that it reflects the reality that pseudoscientific cases are highly diverse. The disadvantage is that it gives no decisive answer: how many indicators are enough to call a doctrine pseudoscientific? The boundary becomes blurred, and that is precisely why the demarcation problem is harder than it looks.
Why can the concept of pseudoscience not be value-neutral?
One important point SEP stresses: the term "pseudoscience" has carried a clearly pejorative meaning throughout its history. It would be as strange for someone to proudly describe their activity as pseudoscience as to boast that it is bad science.
The Latin word "pseudoscientia" was already in use in the first half of the 17th century in discussions about the relationship between religion and empirical inquiry. The earliest known use of the English word "pseudoscience" dates to 1796, when the historian James Pettit Andrew called alchemy a "fabulous pseudoscience". The word has been in frequent use since the 1880s.
Because the pejorative connotation is an essential feature of the word "pseudoscience", an attempt to extract a value-neutral definition from it would be meaningless. A term that is evaluative by nature must be defined in evaluative terms. This is often difficult, because identifying that evaluative component tends to be controversial.
Are all cases clear-cut?
No. This is a point many accounts of pseudoscience pass over: there are borderline cases that are genuinely hard to resolve.
One example is emerging fields that lack a solid theoretical foundation but have not been refuted either. Another is doctrines once regarded as mainstream science but later abandoned, such as the ether theory or spontaneous generation. If we applied present-day criteria naively, we would have to call them pseudoscientific, but that would be historically unfair.
SEP notes that the distinction is also time-bound: a doctrine can move from being regarded as science to being regarded as pseudoscience, and vice versa. This means that a judgement about pseudoscience is not a permanent judgement but one tied to the state of knowledge at a particular moment.
What neighbouring concepts are easily confused with pseudoscience?
SEP distinguishes pseudoscience from a number of neighbouring concepts that are often used interchangeably:
- Scepticism: an attitude that demands evidence proportionate to the claim. This is a legitimate scientific attitude, not pseudoscience.
- Resistance to facts: a tendency to refuse to accept evidence for psychological or ideological reasons. It can appear among both followers of science and followers of pseudoscience.
- Conspiracy theory: explaining events through the plots of secret forces. It often accompanies pseudoscience but is not the same as it.
- Bullshit: speech that is indifferent to the truth. It differs from pseudoscience in that pseudoscience genuinely cares about being taken as science.
- Epistemic relativism: the view that there are no universal epistemic norms. This is a philosophical challenge to the entire demarcation project.
Distinguishing these concepts matters because each calls for a different response. One cannot treat a legitimate scientific sceptic the same way as an advocate of pseudoscience equipped with a self-immunising mechanism.
So what is the pragmatic answer to the demarcation problem?
No single criterion solves the problem, and that is not a failure of philosophy but a feature of the problem itself. The most pragmatic approach available today is to combine several indicators, to stay clearly aware of the conventional and evaluative character of the concepts involved, and to accept that some borderline cases cannot be decisively resolved.
The most important point SEP stresses is this: pseudoscience differs from ordinary mistaken beliefs in that it comes with a mechanism that immunises it against counter-evidence and defends it against critical scrutiny. It is this feature that makes it require different treatment, and it is this feature that makes criteria for recognising it necessary. The demarcation problem is hard because it requires us to define both science and pseudoscience, while both concepts are historical, conventional and evaluative. But hard does not mean hopeless: the strong agreement on particular cases shows that we have reliable intuitions, even when we have not yet managed to express those intuitions as formal criteria.
