What are the various ways in which one could go about trying to demarcate science from pseudoscience?
Answer by Massimo Pigliucci
The distinction between science and pseudoscience constitutes what in philosophy of science is known as the demarcation problem, a term coined by Karl Popper in the early part of the 20th century. Popper was actually interested in solving David Hume’s famous problem of induction. Induction is the general type of reasoning by which we make inferences about things we do not know about on the basis of things we know. For instance, we have seen the sun rise many times before, therefore we reasonably infer that it will do so again tomorrow.
But Hume showed that we unfortunately do not have any solid logical foundation for such inferences. The usual justification for induction is that “it has worked so far” (call it the pragmatic response), but this amounts to say that we believe in induction on the basis of an inductive argument (it worked in the past, ergo it will work in the future), which amounts to deploying a type of circular argument – not exactly a kosher move in logic.
The problem is made more cogent by the fact that a great deal of scientific theorizing uses induction, which is why Popper was so worried. He figured that a possible solution was to move from an inductive to a deductive justification of scientific theorizing. Instead of thinking of science as making progress by inductive generalization (which doesn’t work because no matter how many times a given theory may have been confirmed thus far, it is always possible that new, contrary, data will emerge tomorrow), we should say that science makes progress by conclusively disconfirming theories that are, in fact, wrong. This is Popper’s famous criterion of falsification, which can be formulated as an instance of standard modus tollens in deductive logic: If theory T is true, Then fact F should be observed. Fact F is not observed; therefore theory T is false.
There are several reasons why Popper’s idea of falsification doesn’t actually solve Hume’s problem of induction, which we shall set aside for another time. Popper also thought that falsification could function as the demarcation criterion between science and pseudoscience: if a theory can, in principle, been falsified, then it is scientific; but if there is no way to ever reject it, regardless of what empirical evidence may become available, then it is pseudoscience. Popper thought that Einstein’s theory of relativity is a good example of the first, while Freudian psychoanalysis and Marxist theories of history exemplify the second.
The problem is that there are plenty of pseudoscientific notions that are eminently falsifiable (and have, in fact, been falsified), from astrology to homeopathy. Moreover, there is a history of scientific notions that were initially either unfalsifiable or appeared to be falsified, and yet led to advancements in science. For instance, the original version of the Copernican theory in astronomy (which posited the Sun, instead of the Earth, at the center of the solar system) didn’t do a good job at predicting the actual positions of the planets in the sky. And yet astronomers like Galileo and Kepler kept playing with it, until the latter figured out a relatively minor tweak that solved the problem: Copernicus had assumed that planetary orbits are circular, while they better approximate an ellipsis. Once Kepler introduced the change the theory worked like a charm, so that a Popper-style abandonment after initial falsification would have been unwise.
Because of the problems with the idea of falsification, Larry Laudan published a famous paper in 1983 declaring the demarcation problem dead in the water. He suggested that there is no small set of necessary and jointly sufficient conditions that define “science” or “pseudoscience”; he also claimed that if there were such a set, the only way philosophers could test their classifications of epistemic activities would if they agreed with the judgment of scientists (in which case why not jus ask the latter in the first place); and that the very term “pseudo”-science is problematic on the ground that it pretty obviously implies a negative epistemic judgment, rather than a neutral analysis.
Laudan’s paper was very influential, and did in fact manage to slow down philosophers’ work on demarcation to a trickle. This changed recently because of a collection of essays on the subject published by the University of Chicago Press (and which I co-edited with Maarten Boudry, from the University of Ghent, in Belgium). The book begins with several (belated) replies to Laudan, and then proceeds to explore a number of philosophical, historical, and sociological issues surrounding demarcation.
With respect to Laudan’s first point, several authors have pointed out that just because it is not possible to define science or pseudoscience precisely, it doesn’t follow that the two concepts aren’t meaningful and useful. Wittgenstein famously introduced the idea that a number of complex ideas share a “family resemblance,” meaning that – just like in the case of the members of a biological family – one can see that different instantiations of a concept are related to each other by various threads, even though there is no single or small set of criteria that definitely rule individual instances in or out. Wittgenstein argued that even an apparently straightforward concept like that of “game” is actually difficult to pin down, because for any group of criteria one may propose to define it (“has rules,” “it is done for fun,” “there are winners and losers”) one can easily find either games that do not fit all the criteria, or non-games that fit a number of them. So, a better way to think of science and pseudoscience is as two peaks on a continuous landscape of epistemic activities, some of which are more (or less) scientific (or pseudoscientific) than others.
As far as philosophers having to agree with whatever judgment scientists come up with (Laudan’s second point), the heck with that! Some of the most interesting criticisms of science itself have come from philosophers (e.g., about claims made by evolutionary psychology, or in the name of string theory), so it is far more constructive and interesting to see scientists and philosophers engage in a continuous dialogue about these matters, without either having to simply defer to the other.
Finally, yes, of course the term “pseudo”-science carries a negative connotation (Laudan’s third point). That is on purpose: to put a warning label on an epistemically deficient activity that only apes the trappings of science without actually being a science. And philosophers have always been in the prescriptive, not descriptive, business, so it is okay for us to deploy critical terminology, as long as the deployment is warranted by a detailed analysis.
Answer by Craig Skinner
There was lively debate about this in the 20th century. The upshot was that there is no clearcut demarcation. It’s all gone quiet now.
Suggested criteria for science were:
* falsifiability (Popper).
* puzzle-solving (Kuhn).
* progressive research programme (Lakatos).
Falsifiability: Karl Popper was impressed by how Einstein’s Theory of Gravity offered itself up for possible falsification by predicting something unexpected and testable (light bending around the sun: observations during the 1919 solar eclipse found for Einstein and against Newton) whereas Freudian psychoanalysts or Marxists could explain away any observations without giving up their theories. He suggested falsifiability as the mark of science. Scientists liked it – they were portrayed as heroic figures willing to let their beloved hypothesis be slain by a single awkward fact. But real scientists are different: they hang on to their hypotheses like grim death, blaming auxiliary hypotheses for the apparent falsification (the experimenter missed a confounding factor which affected the outcome; the blood-sugar machine was faulty: etc). Also, strictly, no hypothesis can be falsified: any observation or experiment necessarily tests several hypotheses at once, and we can always say one of the auxiliary ones is wrong, not the main one we are testing.
Puzzle-solving: Thomas Kuhn pointed out the drawbacks of falsifiability, said that scientists rarely even tested their main (paradigmatic) hypotheses, but solved puzzles within paradigm. The hallmark of science was systematic, progressive, puzzle-solving.
Progressive research programme: Imre Lakatos said that scientists (as opposed to, say, astrologers) do research, testing their views against the empirical world, and expect to make progress, discovering new things, reaching better understanding, refining and amending hypotheses.
In a famous 1981 USA court ruling that “creation science” was religion, not science (and so didn’t merit equal school classroom time with evolution), Judge Overton gave the criteria for science as:
* explanation by natural law.
* views testable against the empirical world.
* views held tentatively, not dogmatically.
* views falsifiable.
Creation science has changed its name to Intelligent Design, claims puzzle-solving and research activity, and battles on against “Darwinism”.
Ultimately demarcation depends on detailed understanding of how science works, but even here, scientists differ as to whether some views are science or not e.g. string theory, multiverse hypotheses.
Finally, if you pick up an alleged popular science book, suspect pseudoscience or a religious agenda if the blurb, review or text includes the following words, phrases or links:
* “scientific materialism”.
* “irreducibly complex”.
* “astonishingly complex molecular machines”.
* “academic freedom” (code for acceptance of creationism).
* “Darwinists/Darwinism” (real scientists usually say “biologists/evolution”).
* “blind, random, undirected” (referring to evolution).
* link between quantum physics and freewill
* link between Darwin and the Holocaust
4 thoughts on “Demarcating science from pseudoscience”
Thanks for your thoughtful responses. But Craig, “academic freedom” is code for religion or psudoscience? Really?
what about a shift from “what is science” to “what is knowledge”? that would allow many meaningful types of knowledge to be accounted for- and which can be validated by scientific communities and even the public sphere (as it was defined by Frankfurt critical theory folks)- ? – if demarcation is a difficult problem in science,- this is typically relevant in social sciences where the demarcation is almost impossible?
Thanks for answering. Very helpful summary of the issue at hand, both of you alike. Thank you.