Pseudoscience
Pseudoscience is a body of claims that keeps the appearance of science while dropping the procedure that lets science find out it was wrong. The prefix comes from the Greek for false, but the word is not a synonym for false. A pseudoscientific claim can turn out to be true by accident, and a scientific claim is wrong most of the time on the way to being right. The label describes how the claim is handled, not whether it happens to be correct.
That distinction is the reason the word is worth using carefully. Applied to a conclusion you dislike, it is name calling. Applied to a method, it identifies something specific: a practice arranged so that no possible observation can count against it.
What pseudoscience has in common with science
Pseudoscience has a great deal in common with science, which is precisely why it works. The resemblance is not accidental and it is not usually a fraud. A quasi-scientific practice typically has all of the following, and each one is a real feature of real science too.
- Technical vocabulary, often borrowed from physics or biology and used with adjacent meanings such as energy, frequency, resonance and quantum.
- Data. Charts, measurements, case records and numbers, sometimes collected diligently over decades.
- Practitioners with credentials, conferences, journals and internal disagreements about details.
- A causal story that explains a wide range of observations in a satisfying way.
- Anomalies that mainstream science genuinely has not explained, correctly identified.
What is missing is never the trappings. It is the mechanism by which the field could be told it is wrong and would then change. Confirmation bias operates identically in both cases, in every human being; the difference is that a scientific practice builds machinery to counteract it, such as control groups and blinding, preregistration and independent replication, and a pseudoscientific practice does not.
Six warning signs, all of them about method
Six warning signs recur across otherwise unrelated pseudosciences, and each one is a defect in procedure rather than in content. No single sign is conclusive, and legitimate young sciences show one or two while they find their footing. Three or four together is the pattern.
- Unfalsifiable claims. The central assertion forbids no observation. Ask what would have to happen for it to be false and there is no answer, or the answer is another version of the claim.
- Immunity to disconfirming results. A failed test produces a new condition rather than a revision: the conditions were not right, the observer was too skeptical, the effect is shy. Each rescue is plausible on its own and the sequence of them is the tell.
- Reliance on testimonial. The evidence base is a collection of individual accounts, which have no control group and therefore cannot distinguish an effect from natural change, regression to the mean or the placebo effect.
- A claimed conspiracy of suppression. The absence of supporting research is explained by the motives of those who would have to publish it. This converts the strongest kind of counterevidence, sustained failure to replicate, into further confirmation.
- No self-correction. The field cannot name a claim it has abandoned. Real disciplines maintain a visible graveyard: phlogiston, the luminiferous ether, caloric. A practice with no discarded claims has never tested one.
- Stagnation. The core doctrine is essentially unchanged after a century and generates no new predictions, only new applications of the old ones.
The demarcation problem: why philosophers have no single clean test
The demarcation problem is the name philosophers of science give to the search for a criterion that cleanly separates science from everything else, and the honest report is that no proposed criterion has survived. Karl Popper offered the best known candidate in 1934: a theory is scientific if it is falsifiable, if it forbids something observable. Popper used astrology as his worked example of a system whose predictions were elastic enough to accommodate any outcome.
Falsifiability is indispensable and insufficient. Thomas Kuhn argued in The Structure of Scientific Revolutions in 1962 that working scientists do not abandon a theory at the first failed prediction, and would get nowhere if they did, since every theory is born surrounded by anomalies. Imre Lakatos proposed a repair in the 1970s: judge a research programme over time rather than a claim at a moment, and ask whether it is progressive, generating new predictions that turn out right, or degenerating, spending its energy on excuses for the last failure. Larry Laudan went further in a 1983 paper titled The Demise of the Demarcation Problem, arguing that no set of necessary and sufficient conditions has ever worked and that the term is doing rhetorical rather than analytic labor.
Where does that leave the reader? With a family resemblance rather than a definition. Pseudoscience is recognized the way a counterfeit is recognized, by a cluster of features none of which is individually decisive, and that limitation should be stated openly rather than papered over. Anyone who tells you the line is sharp is selling something.
Pseudoscience versus science, compared on the things that differ
Pseudoscience and science differ most clearly in what each does with a result it did not want. Comparing them on conclusions is useless, because the conclusions of a real science change every decade. Compare them on conduct instead.
| Situation | A scientific practice | A pseudoscientific practice |
|---|---|---|
| A prediction fails | Revises or abandons the hypothesis and publishes the failure | Adds a condition explaining why the test did not apply |
| An outsider tries to replicate | Supplies the protocol and the raw data | Explains that the outsider lacked the necessary sensitivity or intent |
| Evidence is requested | Points to controlled comparisons with stated methods | Points to individual accounts and length of tradition |
| The field is criticized | Treats criticism as free labor | Treats criticism as evidence of a hostile establishment |
| Twenty years pass | The content has changed and old claims are gone | The content is the same and the vocabulary is more modern |
Is a wrong theory pseudoscience?
No. A wrong theory that was tested, defended honestly and eventually surrendered is one of the ordinary products of science. Phlogiston was a serious eighteenth century chemical theory, and the ether was a serious nineteenth century physical one that Albert Michelson and Edward Morley failed to detect in 1887, after which physics gave it up. Neither was pseudoscience. Both were wrong.
The reverse case matters just as much. Alfred Wegener proposed continental drift in 1912 and was rejected by most geologists for decades, largely because he had no mechanism. He was right, and the rejection was not a conspiracy but the normal, sometimes slow, operation of a field that demanded a mechanism and eventually found one in seafloor spreading. Being unpopular is not evidence of being suppressed, and being suppressed would not be evidence of being right. A claim that leans on the Wegener story as its argument has told you which of the six warning signs it is exhibiting.
Can pseudoscience become science?
Yes, and it has, though the crossing always requires giving something up. Alchemy accumulated real laboratory technique alongside doctrine that could not be tested, and chemistry emerged from it by keeping the procedures and discarding the parts that no experiment could touch, a separation Robert Boyle was already pressing in The Sceptical Chymist in 1661. The eighteenth century practice of animal magnetism produced, after its central claim failed a controlled test in 1784, a residue of genuine questions about suggestion that later became legitimate research topics.
The condition is always the same: the field has to state a claim that could come out false, submit it to a test it might not survive, and then accept the answer. That is the scientific method compressed into one sentence, and it is the only entry fee. Astrology has been the standing example in the demarcation literature for a century, and the methodological reason is not that its subject matter is absurd but that its outputs are typically framed in terms broad enough to fit any life, which is the same mechanism that makes cold reading work on strangers who have supplied nothing.
The test to run before you use the word
The word pseudoscience earns its place only when you can answer three questions about the practice, not about its conclusion. Ask: What observation would the people who hold this claim accept as showing it is false? What claim has this field abandoned in the last twenty years? And when a test of it failed, what changed, the theory or the conditions of the test? If the first question has no answer and the second returns nothing, the third has already told you what you will find. If all three have real answers, you may be looking at a science you simply think is mistaken, and the honest word for that is mistaken.