Has the demarcation problem been solved?

Dec 02 2006 Published by under Philosophy, Science and pseudo-science

Revere stirs the pot (of chicken soup) to ask why alternative therapies are presumptively regarded as pseudo-science. The reflexive response of the quackbusters has been that alternative therapies fall on the wrong side of some bright line that divides what is scientific from what is not -- the line of demarcation that (scientists seem to assume) Karl Popper pointed out years ago, and that keeps the borders of science secure.
While I think a fair amount of non-science is so far from the presumptive border that we are well within our rights to just point at it and laugh, as a philosopher of science I need to go on the record as saying that right at the boundary, things are not so sharp. But before we get into how real science (and real non-science) might depart from Sir Karl's image of things, I think it's important to look more closely at the distinction he's trying to draw.

A central part of Karl Popper's project is figuring out how to draw the line between science and pseudo-science. He could have pitched this as figuring out how to draw the line between science and non-science (which seems like less a term of abuse than "pseudo-science"). Why set the project up this way? Partly, I think, he wanted to compare science to non-science-that-looks-a-lot-like-science (in other words, pseudo-science) so that he could work out precisely what is missing from the latter. He doesn't think we should dismiss pseudo-science as utterly useless, uninteresting, or false. It's just not science.
The big difference Popper identifies between science and pseudo-science is a difference in attitude. While a pseudo-science is set up to look for evidence that supports its claims, Popper says, a science is set up to challenge its claims and look for evidence that might prove it false. In other words, pseudo-science seeks confirmations and science seeks falsifications.
There is a corresponding difference that Popper sees in the form of the claims made by sciences and pseudo-sciences: Scientific claims are falsifiable -- that is, they are claims where you could set out what observable outcomes would be impossible if the claim were true -- while pseudo-scientific claims fit with any imaginable set of observable outcomes. What this means is that you could do a test that shows a scientific claim to be false, but no conceivable test could show a pseudo-scientific claim to be false. Sciences are testable, pseudo-sciences are not.
So, Popper has this picture of the scientific attitude that involves taking risks: making bold claims, then gathering all the evidence you can think of that might knock them down. If they stand up to your attempts to falsify them, the claims are still in play. But, you keep that hard-headed attitude and keep you eyes open for further evidence that could falsify the claims. If you decide not to watch for such evidence -- deciding, in effect, that because the claim hasn't been falsified in however many attempts you've made to falsify it, it must be true -- you've crossed the line to pseudo-science.
This sets up the asymmetry in Popper's picture of what we can know. We can find evidence to establish with certainty that a claim is false. However, we can never (owing to the problem of induction) find evidence to establish with certainty that a claim is true. So the scientist realizes that her best hypotheses and theories are always tentative -- some piece of future evidence could conceivably show them false -- while the pseudo-scientist is sure as sure as can be that her theories have been proven true. (Of course, they haven't been -- problem of induction again.)
So, why does this difference between science and pseudo-science matter? As Popper notes, the difference is not a matter of scientific theories always being true and pseudo-scientific theories always being false. The important difference seems to be in which approach gives better logical justification for knowledge claims. A pseudo-science may make you feel like you've got a good picture of how the world works, but you could well be wrong about it. If a scientific picture of the world is wrong, that hard-headed scientific attitude means the chances are good that we'll find out we're wrong and switch to a different picture.
A few details are important to watch here. The first is the distinction between a claim that is falsifiable and a claim that has been falsified. Popper says that scientific claims are falsifiable and pseudo-scientific claims are not. A claim that has been falsified (demostrated to be false) is obviously a falsifiable claim (because, by golly, it's been falsified). Once a claim has been falsified, Popper says the right thing to do is let it go and move on to a different falsifiable claim. However, it's not that the claim shouldn't have been a part of science in the first place.
So, the claim that the planets travel in circular orbits wasn't an inherently unscientific claim. Indeed, because it could be falsified by observations, it is just the kind of claim scientists should work with. But, once the observations show that this claim is false, scientists retire it and replace it with a different falsifiable claim.
This detail is important! Popper isn't saying that science never makes false claims! What he's saying is that the scientific attitude is aimed at locating and removing the false claims -- something that doesn't happen in pseudo-sciences.
Another note on "falsifiability" -- the fact that many attempts to falsify a claim have failed does not mean that the claim is unfalsifiable. Nor, for that matter, would the fact that the claim is true make it unfalsifiable. A claim is falsifiable if there are certain observations we could make that would tell us the claim is false. So, the claim that Mars moves in an elliptical orbit around the sun could be falsified by observations of Mars moving in an orbit that deviated at all from an elliptical shape.
Another important detail is just what scientists mean by "theory". A theory is simply a scientific account (or description, or story) about a system or a piece of the world. Typically, a theory will contain a number of hypotheses about what kind of entities are part of the system and how those entities behave. (The hypothesized behaviors are sometimes described as the "laws" governing the system.) The important thing to note is that theories can be rather speculative or extremely well tested -- either way, they're still theories.
Some people talk as though there's a certain threshold a theory crosses to become a fact, or truth, or something more-certain-than-a-theory. This is a misleading way of talking. Unless Popper is completely wrong that the scientist's acceptance of a theory is always tentative (and this is one piece of Popper's account that most scientists whole-heartedly endorse), then even the theory with the best evidential support is still a theory. Indeed, even if a theory happened to be completely true, it would still be a theory! (Why? You could never be absolutely certain that some future observation might not falsify the theory. In other words, on the basis of the evidence, you can't be 100% sure that the theory is true.)
So, dismissing Darwin's theory as "just a theory" as if that were a strike against it is misunderstanding what science is up to. Of course there is some uncertainty; there is with all scientific theories. Of course there are certain claims the theory makes that might turn out to be false; but the fact that there is evidence we could conceivably get to demonstate these claims are false is a scientific virtue, not a sign that the theory is unscientific.
By contrast, "Creation Science" and "Intelligent Design Theory" don't make falsifiable claims (at least, this is what many people think; Larry Laudan* disputes this but points out different reasons these theories don't count as scientific). There's no conceivable evidence we could locate that could demonstrate the claims of these theories are false. Thus, these theories just aren't scientific. Certaintly, their proponents point to all sorts of evidence that fits well with these theories, but they never make any serious efforts to look for evidence that could prove the theories false. Their acceptance of these theories isn't a matter of having proof that the theories are true, or even a matter of these theories having successfully withstood mant serious attempts to falsify them. Rather, it's a matter of faith.
None of this means Darwin's theory is necessarily true and "Creation Science" is necessarily false. But it does mean (in the Popperian view that most scientists endorse) that Darwin's theory is scientific and "Creation Science" is not.
*See Laudan, "Science at the Bar -- Causes for Concern", in Michael Ruse, But Is It Science?

13 responses so far

  • revere says:

    Janet: Thanks for the lucid discussion of Popperism and issues related to it. Popper is a favorite of epidemiologists, but for reasons I'll explain in a subsequent post, most epidemiologists are very bad Popperists. I'll wait to see if some of our colleagues weigh in on this and do a round up of responses later in the week.
    In theory, I will, anyway. I'm not sure it will turn into a fact, but you've given everyone a reason not to expect it.

  • Macht says:

    "(in the Popperian view that most scientists endorse)"
    Are you saying that most scientists claim that the Popperian view is what makes something science or are you saying that most scientists practice the Popperian view of science? (Or both?)

  • Sean Carroll says:

    Great post. You're right, I think, that most scientists are closet Popperians. As a rigorous way to demarcate science from non-science, falsification completely fails; theories can rarely be unambiguously falsified, experimental results are interpreted within theoretical contexts, other criteria play a big role in theory-choice, etc. But informally, it does capture something important about the scientific attitude -- we propose hypotheses without being wedded to their truth, and then try to separate what works from what doesn't.
    I'm perfectly willing to count Intelligent Design as science, and just say that it's incredibly bad science. It proposes ill-defined and metaphysically cumbersome solutions to problems that already have much simpler and more promising answers.

  • That parenthetical "in the Popperian view that most scientists endorse" is meant, as Sean also points out, to capture an attitude that scientists think is central to doing good science -- namely, being willing to kiss even your favorite theory goodbye if the evidence stacks up against it.
    Frankly, I doubt that many scientists have read any Popper. But if they had posters in their lockers (assuming they had lockers), instead of Che Guevara, they'd have Sir Karl.

  • Kait says:

    Hi, great post.
    Have you read Laudan's 'Demise of the Demarcation Problem'?

  • Hi. Nice discussion of the issues of falsifiability in science. However, I have a philosophical question that I don't believe is actually answered by Popper. How does one make scientifically based decisions?
    To give an example, what should be done about global warming? The evidence that supports global warming is good, falsifiable science. However, a claim of the form: "It is necessary to do X to prevent catastrophe" or "It is sufficient to do Y" may very well be falsifiable, but we certainly don't want to wait to find out. We have to make decisions before all the evidence is in.

  • Stephen says:

    I would suggest that non-science and pseudo-science are actually two different things. Much of what you find in management books for example is non-science: it isn't science and doesn't pretend to be science, although it does claim to be words of wisdom.
    Creation science / ID is an example of pseudo-science: it isn't science, but it pretends to be science.
    I realise of course that this introduces another demarcation problem ...

  • SteveG says:

    I've long thought that Popper was exactly the wrong club to pull out of the bag for the ID shot for reasons of both necessity and sufficiency. On the one hand, banking on the impossibility of falsifiable propositions from intelligent design is put one's eggs in a poorly made basket. I think Laudan is right here and there are some very smart (albeit misguided) folks on the ID-team working very hard with things like information theory to show how one could have derivable results that are non-tautologous. On the other hand, there are certainly aspects to evolution theory which are placed beyond falsification. As Duhem pointed out long, long ago, it is always possible to save any given part of a theory if you are willing to pfutz with the rest of it and we do take central axioms and put them out of bounds.
    All of this was pointed out by Popper's student Imre Lakatos who presents what I would contend is a much better way of looking at the evolution/ID debate. He argues that all scientific theories have an unfalsifiable hard core and a protective belt of auxilliary hypotheses. On this view, both evolutionary biology and ID would be scientific -- BUT, theories can be judged on the way they have developed historically and whether they account for a broad spectrum of phenomena with little significant alteration (he deems there progressive research programmes) or whether they need major renovations and ad hoc patches to account for things as they pop up (he terms these degenerating research programmes). We can fully allow that, like phlogiston theory, for example, ID is scientific, it is just incredibly degenerate and that we have an amazingly progressive research program as a competitor. I see little problem letting ID into the ring, when there is no way it wins a fair fight.

  • Otheus says:

    Science does not have a monopoly on knoweldge (or "true knowledge" or whatever). It is not merely "science" that benefits from Popperist epistemology, but knowledge itself. Whether or not water boils at 100() C is interesting to Science, but it's also very interesting to the non-scientist. Whether or not "God answers prayer" is also subjectable to falsification. When God does not answer prayer, this theory has been falsified.
    Falsification is a critical piece of the demarcation puzzle, but it is only one piece. Other pieces include: rigorous adherence to terminilogical consistency; rigorous adherence to (deductive) logic; repeatability of observations. Though some of these are somewhat obvious, and may fall under the umbrella of "falsification", there is one lynch-pin that does not: Precision. As long as what is being measured or tested involves an element of precision, then the degree to which the theory has been falsified is also subject to that degree of precision. Consider the theory: "God always answers prayer" combined with "One hears the answer only 45% to 55% of the time." Is this falsifiable?
    Because knowledge involves an element of precision (or imprecision, or quantified uncertainty), the line of demarcation is, in fact, a shadowy line. Theories falling too close to this line will not be objectively verifiable as 'science' or 'non-science'. These theories will continue to be the subject of debate ... at least until the theories evolve into more precisely falsifiable ones.

  • Andrew Wade says:

    By contrast, "Creation Science" and "Intelligent Design Theory" don't make falsifiable claims (at least, this is what many people think; Larry Laudan* disputes this but points out different reasons these theories don't count as scientific).

    I'm mostly with Larry Laudan on this one, many of the Creationist's claims are testable, have been tested, and have failed the tests (as he puts it). For the most part, creationists[1] haven't cared. Now some of their theories do run up against the falsification criteria; the definition of "irreducable complexity" is now hopelessly vague. (The earlier, less-vague definition didn't do so well; scientists had no trouble explaining how such systems could evolve; and had an annoying habit of finding evidence of such evolution in the exemplars given.)
    [1] By "creationist" I mean those that believe in the special creation of humans. There are plenty of people that believe God created the universe who have no problems with science.
    I think part of the problem stems from viewing Creation Science as a coherent whole; it's really not. The theories are crap, but the reasons they are crap vary quite a bit.
    Now, I do agree somewhat that pseudo-science seeks confirmations and science seeks falsifications. But only somewhat. I really doubt that many scientists are seeking to falsify the theory of Evolution through Natural Selection (NS); there are much more promising lines of inquiry. But there is nonetheless something to the statement above; for if evidence appeared that appeared to falsify NS, I think it would generate considerable interest among scientists. For example, the discovery that some bacteria increase their mutation rate in response to stress seems to have generated some interest (I heard about it after all); it challanges a common assumption associated with NS; albeit not a necessary one. In contrast, creationists are notorious for not caring what the facts are.
    "Evilutionists" would be interested in appearant falsifications of NS not so much because such falsifications are at all likely to be true; but because scientists are likely to learn something new from such evidence. This I think is where the fundamental difference between scientists and creationists lies; scientists are interested in learning from nature, whereas creationists seem to be at best indifferent[2]. But that is not the fundamental difference between their theories; it is merely why creationists are willing to believe theories that are crap. The fundamental difference between creationist theories and NS is that the latter theory happens to be, on the whole, true. (As far as we can tell).
    [2] This indifference I think is part of a broader anti-intellectualism. Hence the poor definitions, weak logic, and "cargo cult" math that also go along with much of pseudoscience.

  • Periphrasis says:

    Otheus - although I understand the general thrust of your post, I must take issue with your statement: "Whether or not "God answers prayer" is also subjectable to falsification. When God does not answer prayer, this theory has been falsified." As is commonly pointed out in religious circles, it may be that God always answers prayer, but that the answer is usually "no."
    Given a framework in which answers are played out by action (or the lack thereof), rather than direct (verbal, written, telepathic or otherwise linguistic) communication (which, though not everyone's preferred method of communication, is certainly not unusual. If you ask someone to hand you something nearby, there's a significant chance that the person will simply hand it to you without saying "Yes, I will hand this to you" beforehand), the observation of this claim may not be falsified in the manner you suggest.

  • demi says:

    not sure how applicable popperian paradigm is for contemporary scientific industries. the institutionalization of pharmaceuticals, for example, it's dependency on funding, government grants, public dollars, may very well have redirected those industries toward corroboration. after-all, governments , corporations and the public want to see evidence, not falsifications.

  • Will Nelson says:

    Personally I think that mathematizability is the only possible demarcation between science and non-science. Scientific theories are those which could, in principle, arise from an underlying fully mathematical structure. This obviously includes evolution which is arises inevitably from the chemical underpinnings of life, which in turn arise from the mathematical theory of elementary particles.

    By contrast, any theory which involves a god is inherently not reducible to mathematics. Indeed, that could almost be taken as the definition of a god.

    It is, unfortunately, possible that our universe could be described by mathematics which is not fully testable. This is too bad, but after all the universe was not designed for our convenience, nor to fit our preconceived philosophies.

    If, on the other hand, the universe is not described by mathematics at the deepest levels, then there's no point even talking about science because anything is possible at any time. Any regularity that we observe at any time is just the whim of the gods. Personally, I think this is impossible, because regularity is necessary for existence, and regularity comes only from mathematics. But that's just me.