One problem with professional philosophy—and this holds for some of the sciences too, like physics and biology—is that the subject matter is difficult to master and require a great deal of time and technical training.
This does not, however, stop philosophical concepts from spilling over into popular discourse, where they are usually poorly understood, or even more commonly, completely misunderstood.
When I hear the terms “falsification” or “falsificationism” thrown around wildly, I experience something much like what I imagine a biologist experiences when he hears the term “evolution” being wildly and recklessly misapplied in contexts where it is misleading or meaningless.
What is falsificationism? It is a specific answer to a specific philosophical question given by a specific philosopher to solve a specific problem—and it is failed attempt at that, which left in its wake a sometimes useful methodological tool, and an unfortunate extra-philosophical cult following.
The philosopher in question is Sir Karl Popper, and the problem he was trying to solve was known in his day as “the demarcation criterion problem.” At the time, it was taken for granted that some knowledge is scientific and some knowledge is not scientific, that some human disciplines of inquiry are sciences and some are not. So, the question arose By what criterion can one demarcate a real or genuine science from a non-science, or more urgently, a pseudo-science claiming to be a science? In other words, How do you tell a science from what isn’t science?
Popper’s thinking was informed by the intellectual milieu of fin-de-siècle Vienna, in which many exciting intellectual revolutions were underway simultaneously. Einstein was threatening to shake the heretofore seemingly unshakable foundations of Newtonian physics, mathematics seemed, as per Hilbert’s famous speech, about to be completed (before Gödel put an end to that idea once and for all), Freudians claimed to have discovered the key to the depths of the human psychē, and Marxists to have unlocked the inner mechanisms of human history. It was a heady time, in which people were not afraid to make bold, often reckless, claims.
At the time, the universally accepted answer to the ‘demarcation problem’ was verificationism. Something would count as scientific knowledge if its claims could be empirically verified, and the more verifications, the more scientifically sound the view was supposed to be. This theory is very easy to understand and has a direct appeal to common sense: if a theory keeps being proven correct, it is probably a correct theory, right?
Unfortunately, Popper noticed something, something which today we would call “confirmation bias.” He noticed it particularly among the Marxists and the Freudians, both groups of which insisted that their respective disciplines of dialectical materialism and psychoanalysis were indeed SCIENCES. He noticed it personally in his friend Alfred Adler, the Freudian heretic who had ‘discovered,’ contra Freud, that sexual repression was not the key to the human psychē, but the inferiority complex instead.
What Popper noticed is that when a Marxist or a Freudian attempted to give a scientific account of a historical event in terms of class conflict or a psychological phenomenon in terms of sexual repression and neurosis, they were able to do this, no matter what the circumstances were. A Marxist could, and did, explain anything and everything by means of class conflict. A Freudian could, and did, explain anything and everything by means of sexual repression and neurosis. And each “explanation” which was given was considered by each as a “verification” of either Marxism or Freudianism, thus further proving each theory correct and scientific. Nor could one criticize Marxism or Freudianism: it was obvious that the critic of Marxism was a bourgeoise apologist, which proved Marxism’s class conflict theory, and the critic of Freudianism was sublimating his sexual repression into a neurotic attack on Freudianism, which proved Freudianism. (As you can imagine, when a Marxist got into a head to head argument with a Freudian, it was like a perpetual motion machine of pointless back and forth!)
Popper regularly dined with his friend Adler, with whom he often quarreled about this issue. One time, Popper discovered a psychological case that seemed clear to him could not be explained by means of Adler’s go-to explanation for everything, the inferiority complex (which had the same self-confirming structure as the Freudianism from which it sprang; e.g. A man sees a drowning child: if he fails to jump in the water to save her, it was due to his feelings of inferiority; if he does jump in to save her, he was overcompensating due to his feelings of inferiority, etc.).
Popper waited for his moment and sprang his example case on Adler, hoping to baffle him. Instead, to his astonishment, Adler immediately analyzed the case in terms of an inferiority complex. Popper managed to sputter “But … how do you know this?” To which Adler replied, in a magisterial tone, “Based on my thousandfold experience.” To which Popper, having recovered, replied “And I suppose now your experience is a thousand-and-one-fold!”
So, the scene is half set. We need one more piece of the puzzle. Einstein’s theory of special relativity was being hotly debated by physicists. Some were intrigued by it, some thought it was complete lunacy. As it happened, an eclipse of the sun was due to happen on a certain day, which would allow something not usually possible: the direct observation of the planet Mercury (Mercury is usually overwhelmed by the sun’s radiance and cannot be directly observed). Einstein’s theory predicts that light will be affected by sufficiently massive bodies (such as the sun) whereas Newton’s theory does not; that is, according to Einstein, light should bend around the sun due to its mass. This allowed a very interesting test case. Since we cannot normally see Mercury, we do not know exactly where it is (or more accurately, where it appears to be). But at the time of the eclipse, it would be possible to look and see exactly where Mercury appears to be: in essence, Einstein’s theory predicted Mercury would be at point E, due to the gravitational bending of light and Newton’s theory predicted Mercury would be seen at point N. All that was needed was to wait for the eclipse and see whether Mercury was at E or N. And as I suppose you all know, the eclipse came, and Mercury was observed to be at point E, just as Einstein had predicted, and the relativistic revolution began in earnest.
What impressed Popper so much about this was that Einstein was willing to “go out on a limb” so to speak and make a PREDICTION, which, if it didn’t pan out, would have significantly undermined the credibility of his theory; but at the same time, if it did pan out (as it did) would significantly boost the credibility of the theory. In other words, Einstein was willing to make a prediction which, if it failed to hold, would FALSIFY HIS THEORY.
Marxists and Freudians and Adlerians were NOT willing to make predictions which would, if they failed to happen, cause them to abandon their theories. NOTHING can make a convinced Marxist or Freudian (or feminist) abandon his or her theory.
So Popper hit on the idea of REVERSING the demarcation criterion: instead of VERIFICATION (he argued) truly scientific reasoning risks FALSIFICATION. It makes CONJECTURES and risks REFUTATIONS. A theory that cannot under any circumstances be falsified is not, according to Popper, a scientific theory.
Thus was born the idea of falsificationism, namely, the idea that what makes a discipline scientific is its willingness to risk being refuted by making risky predictions.
As with verificationism, there is a kind of common sense appeal to falsificationism. It is not a worthless idea—but in the event, it can’t do the work Popper hoped it could, and it certainly can’t do the work that Popper candidly admitted it could not: namely, serve as a criterion for all legitimate knowledge as such.
To start with the latter, as an epistemological principle, falsificationism is not a scientific statement. It is a philosophical statement (or a “meta-scientific statement” as Popper would have said). It was a statement about science, but not a scientific statement. Part of the whole problem it is meant to solve is to mark off scientific knowledge from other sorts of knowledge, such as philosophical knowledge. If one were to attempt to apply the falsification criterion beyond its intended scope of scientific knowledge, applying it to, say, all knowledge, the result (of which Popper was well-aware) would be disastrous:
- The principle of falsification: no statement which cannot be empirically falsified can count as legitimate knowledge, and must therefore be rejected.
- Statement 1, the principle of falsification, is a statement which cannot be empirically falsified.
- Therefore, the principle of falsification does not count as knowledge, and must be rejected.
Any attempt to take Popper’s principle of falsification as a universal standard applying to all knowledge fails on its face, because the principle is SELF-DEFEATING. Applied to itself, it fails its own test. And Popper was well-aware of this, since the same holds true for its predecessor, verificationism. Verificationism cannot be verified and falsificationism cannot be falsified. So any attempt to extend either principle beyond its narrower, scientific scope fails.
And neither principle succeeds in its more limited purpose of demarcating “scientific knowledge” from non-scientific knowledge either (to give the game away, most philosophers today are content to say that “scientific knowledge” is a Wittgensteinian family resemblance concept and not to think there even IS anything specific that exactly demarcates science from non-science).
Verificationism fails for the reasons that so upset Popper. Countless pseudo-verifications can be generated almost at will for any theory. This is why Popper wanted falsification to be clear, cut and dried. IF THIS PREDICTION DOESN’T HAPPEN, THE THEORY IS FALSIFIED AND SHOULD BE ABANDONED.
Why doesn’t this work for science at least? Didn’t it do a good job distinguishing Einstein’s theories from Marx’s or Freud’s? Yes, in that interesting and unusual case, it did. But unfortunately, that isn’t how science actually works nearly all the time.
Since the Einstein example that so impressed Popper involved planets, let’s talk about those. Consider the planet Uranus. When you look for it in the night sky, it isn’t where it is supposed to be, according to Newton’s theory.
BAM! Clear, cut and dried, remember? Newtonian physics is FUCKING FALSIFIED. Stop using it! Right now! It’s wrong! No, really, quit it!
Except … that’s now how things played out. Newtonian physics was too damned useful to throw out as Popper demanded. Instead of regarding Newtonian physics as falsified by Uranus not being where it should be according to the theory, the scientists decided that there was probably some other factor X which was causing the (correct) theory not to match up with the (would-be falsifying) observation. What could it be? Well .. if there were another planet out beyond Uranus … of THIS mass … and right about HERE … THAT would explain Uranus not being where it is supposed to be according to the theory.
Lo and behold, telescopes were pointed and Neptune was discovered. Far from being a falsification of Newtonian physics, the failed prediction of the location of Uranus led to the discovery of Neptune, which was then regarded as a stupendous confirmation of Newtonian physics!
Well … except for one little problem. Now that we know about Neptune .. it turns out that it keeps not being where Newton’s theory says it should be. So, once again, Newtonian physics is falsified, right? We abandoned it? Nope. Once again, scientists did what scientists actually do (and not what Karl Popper would later say that SHOULD do), and they decided that the mismatch between the theory and the observation was due not to a defect in either one, but due to some complicating factor X. What could it be? Well .. if there were yet another planet beyond Neptune .. of THIS mass .. right about HERE .. and so we discovered Pluto. And another crucial scientific discovery was made due to the fortunate fact that Popper hadn’t written yet, and no scientist in his right mind considered Newton “falsified.”
Now it had been known for a long time that Mercury also doesn’t do what it should do, according to Newton’s theory. And once again, scientists hypothesized that neither the theory nor the observations were wrong, but that there was a complicating factor X which was the cause of the discrepancy. If there were another planet HERE, of THIS mass … well, it’s too close too the sun to actually SEE (yet) but let’s assume it’s there. They even named it: Vulcan, the planet closer to the sun than Mercury, which disturbs Mercury’s orbit.
Alas for planet Vulcan! Sorry, Mr. Spock! We mourn for thee. It turns out that in this case, the theory was wrong. Newton wasn’t right, and Einstein’s theory explains perfectly well why Mercury moves as it does without the need to postulate a planet Vulcan.
But I hope you see the problem already from the examples given. Assume there is a mismatch between a THEORETICAL PREDICTION and an EMPIRICAL OBSERVATION. Then one of the following is true:
- The theory is wrong / falsified.
- The observation is erroneous in some way.
- Neither the theory nor the observation are erroneous, but there is an unknown factor X causing the discrepancy.
In brief, the problem with Popper’s theory is that it requires us to know in advance that 2 and 3 can be safely ruled out. Now, sometimes 2 can be ruled out to a high degree of probability, sometimes not. But 3 is the real sticking point. 3 can NEVER be ruled out as a possibility since, by definition, it is an UNKNOWN FACTOR X. It is simply not possible to say “I know with certainty that there is not something I do not know about interfering in this case.”
This alone would be enough to sink Popper’s falsificationism as a comprehensive theory of scientific knowledge (let alone all knowledge, which it never was nor could have been), but matters get even worse for falsificationism.
Again, I’m not going to give you the detailed philosophical arguments behind all this, but two philosophers advanced very similar ideas, one earlier and one better, but similar to the extent that the general view is known as the Quine-Duhem Thesis … or the Duhem-Quine Thesis … whichever sounds better to you, I guess (I like the first one, even though Duhem was before Quine in making the case).
Basically the Quine-Duhem Thesis states that it is impossible to test any scientific idea or theory or hypothesis in isolation, because any possible test one can devise already requires implicit and explicit embedded background assumptions. In something like Quine’s terminology, knowledge, including scientific knowledge, forms a wholistic system or web of beliefs and assumptions, and there is simply no way to ask IN ISOLATION OF THE SYSTEM AS A WHOLE whether one particular belief or idea is true or not. It is a kind of Einsteinian relativity applied to epistemology. What Quine argued, in effect, is that one can, with rational consistency, continue to hold any belief whatever, so long as one is willing to sufficiently adjust one’s entire conceptual web.
I have some personal experience with this, since I have on more than one occasion backed an atheist into a logical corner in which reason required him to admit the existence of God. To my amazement (and horror) I have found that many atheists are willing to take the leap into absurdity and reject reason outright rather than admit God. I have actually been told “If reason proves there is a God, then reason isn’t reliable after all.” And how could I persuade them not reject reason at this point? Not with reasons, obviously. Non credo, quia absurdum.
And I’m sure that many sciency-types have encountered fundamentalist Christians who are more than willing to sacrifice their belief in science to preserve their literal reading of the Bible. I imagine this drives them (the sciency-types) to distraction, because it seems so willfully irrational. The problem is, of course, the Quine-Duhem problem: it seems willfully irrational on the basis of certain background assumptions that your interlocutor does not share; what seems manifestly irrational to you, is not necessarily self-inconsistent: it is only inconsistent with some of your assumptions your interlocutor does not share.
So where does all this leave us? Well, it leaves us without any cut-and-dried test of what is or is not scientific knowledge, or more generally, what is or is not valid knowledge as such. We may know a lot of things, but we do not necessarily know what or how or why we know. And yet, the world endures, somehow—even if we cannot with 100% certainty say what is or isn’t “science.”
Maybe it suggests we should cultivate a bit more Socratic humility and Socratic ignorance.
And it certainly tells us that idiots on the internet should stop invoking falsification as if it were some kind of Holy Scripture or Harry Potter-esque magic spell: Wingardium leviosa! Expecto patronum! Falsificatio popperium!
It was never a theory of knowledge as a whole. It was a limited theory meant to explain how science is different from non-science. And it didn’t even WORK. So WHY are you bringing it up, as if it were some sort of trump card, you idiots?
I’ll close this with the most appropriate thing I can think of, a pair of quotes from the philosopher of science Paul Feyerabend:
And even more succinctly