web analytics

Most people are stunningly irrational; falsifiability is a test and a tool that helps us to hew away unfounded mental models and leave behind what works (for now).

The criterion that a theory ought to tell us how to show its falsehood is a wonderful filter, holding back pseudo-scientific claims about nature and forcing us to be honest. Those of us who can’t or won’t open ourselves to falsification are in a freeloading relationship: kept fed and comfortable on the couch of those who do.

Perhaps it’s time that we struggled off the couch and got a job, opening up our claims about business, the economy, politics: what would you need to see to convince you that you were wrong about one of your beliefs in these or other areas that relate to the natural world? If we can’t do this, how do we expect others to believe that we’re thinking honestly and not acting on blind faith or deluding ourselves?

What Is Falsifiability?

Falsifiability is a characteristic of certain statements, meaning that they can be shown to be false. The statement, “All swans are white.” is falsifiable, as finding a black swan shows its falsehood. Lack of falsifiability in claims about nature or reality is the mark of pseudo-science; for example, all astrological claims are so vague that they can never be shown to be wrong.

It’s important to note that falsifiability is a property of well-formed theories of nature, but not necessarily of other more abstract domains like certain philosophies, aesthetics, metaphysics, etc. For example, mathematically, the statement that 2+2=4 is unfalsifiable; one can experiment, however, with piles of beans in the real world, and prove it false by combining 2 and 2 beans and getting 3.

I’m not being facetious: I can’t think of any possible way that you’d ever get anything other than 4 beans; yet, this claim is falsifiable to the extent that you can describe a condition in which it would be wrong. One shouldn’t forget that less fundamental but no less strongly held claims have been falsified: the intuition that time flows at the same rate universally being just one example.

Why Does Falsifiability Matter?

I’m going to proceed with the help of some popular misapprehensions about science which serve as the other part of my dialectic.

The first is the idea that theories in science are proven: “proof” in this sense belongs to the domain of mathematics, in which mathematicians use logic to show that their premises lead to a certain conclusion. For example, we can prove that 0.999… (“zero point nine recurring,” or zero followed by an infinite series of nines) is equal to 1:

  • Premise 1: if a pair of numbers is non-identical, one can enumerate the difference and even find a number in between (e.g. the difference between 1 and 3 is 2, and 2.3 lies between them);
  • Premise 2: one cannot name a number between 0.999… and 1 (try it);
  • Therefore: 0.999… and 1 are identical, or 0.999… = 1.

Such pure, logical expressions aren’t possible in the real world.

Secondly, theory as in the theory of relativity is not the same thing as when people say, “in theory there’s enough gas in the car to get us to Memphis.” Rather, in science, theory is the actually the highest level of certainty that we apply to descriptions of reality. As Karl Popper put it in his book that pioneered our understanding of falsifiability: “The demand for scientific objectivity makes it inevitable that every scientific statement must remain tentative for ever.”

The theory of relativity can never be proven, but it is the best we have given that it has withstood repeated attempts at falsification; it will almost certainly be shown to be false, but that new theory must contain all the explicative and predictive utility that relativity provides (such as time dilation, gravitational waves) and more.

Watch out for the converse of this: pretended theories from outside science that take on this honourable moniker without accepting the risk of falsifiability. For example, the Church of Scientology claims a theory of theta, which describes how metaphysical beings inhabit human bodies and form our souls: despite the fact that, if the Church actually made a testable claim about their theta and were shown to be correct, they would make history, they won’t. Can you guess why?

Finally, “fact” is another such misnomer. People often make statements such as “Evolution is a fact.” This is crossing the streams in that theories, like Evolution, explain and predict certain facts  — facts being incontrovertible observations based on repeated empirical study. Theories change, not least, because our facts do.

Falsifiability and falsification are useful because it is so easy to verify things. If one had a hypothesis that all cars are black one could continuously find new examples to support the claim. Falsifiability protects, also, against the habit of charlatans and fools to create models (which, again, are often unfairly called theories) which are malformed to the extent they explain anything, predict nothing useful, and are endlessly re-interpretable.

Christopher Hitchens was once asked by his creationist debating opponent to give an example of an organism that gave the appearance of having been created by evolution and not by God. Hitchens, in a response that would have been impressive for a normal person, gave the example of the blind salamanders, such as the Comal blind salamander, which (living in total darkness) have eye sockets but no eyes.

Why would God create a creature with eye sockets and no eyes? Surely it’s clear that these salamanders evolved from ancestors which could see, and since then lost their vision through generations in darkness. The creationist responded by saying that God created salamanders with eyes, but then they fell into the caves and “micro”-evolved to lose their vision. If you believe that, you could believe anything.

A recent article in Scientific American made a concerted case against falsifiability. I worry that I missed the point of this piece: please, dear reader, if you caught it, let me know.

The thrust appears to be this:

But the field known as science studies (comprising the history, philosophy and sociology of science) has shown that falsification cannot work even in principle. This is because an experimental result is not a simple fact obtained directly from nature. Identifying and dating Haldane’s bone involves using many other theories from diverse fields, including physics, chemistry and geology. Similarly, a theoretical prediction is never the product of a single theory but also requires using many other theories. When a “theoretical” prediction disagrees with “experimental” data, what this tells us is that that there is a disagreement between two sets of theories, so we cannot say that any particular theory is falsified.

“Haldane’s bones” here, refers to J.B.S. Haldane’s possibly apocryphal statement, in response to a request for some piece of evidence that would falsify the theory of evolution, that a rabbit skeleton in the Precambrian strata would suffice.

We actually face questions like this all the time. One case with which I’m quite familiar is the occasional experimental results that appear to violate Einstein’s universal speed limit: no particle with non-zero mass can travel at or faster than the speed of light.

Every now and then, however, a research group will come up with a result that disagrees with this principle. In cases like this, we assume that because of the extent to which relativity has withstood attempts at falsification, the new result is probably wrong. In the case of the OPERA experiment, for example, further study and falsification attempts by other teams invalidated their original faster-than-light result.

The question of which theory is falsified  is answered first by trying to falsify the result, then the weakest of the theories, and so on until we figure it out.

The article continues, wondering why we kept using Newton’s laws, despite their being inconsistent with “the motions of the perihelion of Mercury and the perigee of the moon.” This appears, however, to be a misstatement of what happened: in the second case, people puzzled with curiosity over why Newton’s laws worked so wonderfully in a preponderance of other cases but not this one. It turned out, like in the OPERA experiment, the discrepancy came from a mistake, not a problem with Newton.

In the first case, this inconsistency turned out to show the advantage of Einstein’s relativity of Newton’s laws. Relativity, though a geometry that is beyond me, explained the perihelion of Mercury perfectly, and since it was set forth has repeatedly been corroborated by experiment. The writer makes it seems as though scientists, confronted with a result that contradicted the theory, simply covered their ears and ignored it, or tried to wish it away. Rather, they seem to have said: well, either the result is wrong or the theory is wrong or some other theory, let’s try to puzzle it out. 

It’s easy to forget, also, that the best theories we have: relativity and quantum theory, contain stunning unknowns and inconsistencies, and are almost certain to be replaced.

The Freeloaders

To return to that Hitchens moment, people like the creationist whose worldview can rationalize away any contradiction are essentially freeloading off people who make falsifiable predictions and admit when they are wrong.

Putting aside the purely theological, the aesthetic and the metaphysical, the first group brings us faith healing and purported miracles, astrology (still brought to you by New York Magazine), witchcraft, and fanatical (usually totalitarian) political theories; the latter group keeps the former group connected (without Relativity our satellites would go off course), replete with computer technology (without Quantum Mechanics our circuit boards would be useless), fed (by the Haber process) and well (guided by the germ theory of disease).

We owe our civilization to the  group of people and organizations that are actually capable of thinking and acting scientifically. And you and I, dear reader, aren’t going to get away with putting ourselves, smugly, among the scientists. When was the last time you claimed something or made a prediction that was falsifiable? Amos Tversky and Daniel Kahneman showed us quite decisively how stunningly irrational most all humans are. The scientific method and its ability to falsify is what helps irrational people do rational things.

In my teenage radical days, when challenged on the colossal failure of the USSR, I would reply, as many do, that the Soviet system was not a true test of Socialism. The follow up that I never got but really should have was: “Well Oliver, in that case, you must describe a scenario that you would consider a true test of Socialism and how it might fail; you must define a scenario that would falsify historical materialism.”

The crux is, essentially, this: if you can’t name something that would convince you to change your mind, or at least lower your level of certainty, are you even thinking? How are you checking your ideas against reality?

This is common in politics and matters of personal belief, of course, but it happens elsewhere, too. If you’re in business, and recommend a particular marketing strategy, what would it take to convince you that it was a failure? In government, if you recommend a new program or policy, what needs to happen in order for you to call for its cancellation?

I want to make it clear, also, that scientists don’t come with superhuman rationality, as people: this is true from the chain-smoking medical doctor to Isaac Newton, with his quack beliefs and religious fanaticism. Science isn’t a state of being, it’s a posture. And, if you force my hand, I might say that falsifiability is too good to leave to the scientists.

As I mentioned above, most of us in society and the world are freeloading off people and institutions that make falsifiable claims; we’re on deck, they’re in the engine room. It’s only fair that we incorporate some of this ethic of intellectual honesty into our lives. Can you think of anything bad that would come of it?

Of course, we are then faced with the question of how to marshal the booty that science generates for us and how to decide what is of value: a preference for falsifiable claims is of limited help in answering those questions, but it might keep us alive long enough to answer them.

2 thoughts on “Falsifiability and Its Freeloaders”

  1. Ruoji says:

    Lovely post! A couple things for discussion:

    “For example, mathematically, the statement that 2+2=4 is unfalsifiable; one can experiment, however, with piles of beans in the real world, and prove it false by combining 2 and 2 beans and getting 3.”
    — I remember an undergraduate epistemology class where we used precisely this example. If you combined 2 beans with 2 more beans, and every time, a bean disappeared… does this mean 2+2 now equals 3? Is this a property of the beans or of the mathematical statement? Kant would call these truths a priori — not empirical. I still find this a useful way to think about things.

    As to the question “you must describe a scenario that you would consider a true test of Socialism and how it might fail; you must define a scenario that would falsify historical materialism.”
    I have another question for you: are theories of history the same thing as theories of “nature”? Are theories of how history happens the same thing as how things work? — Most of us would say no. History happens, but it doesn’t mean that it couldn’t have happened any other way. (I also think you’re allowed to have theories without betting your shirt on every single one.)

    On historical materialism:
    The power of historical materialism, many academics would argue, is in its work against ideology. One of the virtues of utopian thinking is that it does not demand material realization as such (it’s not a blueprint for assembling lego blocks). Thinking through its contours releases you from the ideology of what is self-evidently the case — and that’s already some solid political work! Nothing makes us more blind than the wisdom of the present.

    For instance, it’s just as absurd to claim that because socialism failed in the Soviet state, that socialism does not work — that it must always fail. We simply do not have that kind of historical insight. Part of the “utopian” aspect of socialism is also that it thrives in what is “yet to come,” an idea that the future can be different from the past — that the future can redeem the past (similarly to Christian eschatology). In that respect, it has a certain degree of built-in resiliency.

    Finally, ideas do not need to be universal to be helpful— they can be empirical yet live on provisionally. Creationism is a problem because it demands that practices of faith acquire the rigor of science. But that creationism is stupid doesn’t preclude the fact that things can appear to have been made by God without the categorical claim that they were. Something like this lives in the space of aesthetic judgments that may seem weak, but are often quite important to how we live our lives.

    1. Oliver says:

      Thank you! And thank you so much for the comment!

      Regarding the beans: I might need to re-write this part, what I was trying to do was to show that even though combining 2 and 2 beans and getting something other than 4 is impossible, the empirical claim that 2 and 2 beans is falsifiable, to the extent that you can describe the condition wherein it is falsified (the falsification condition doesn’t need to be possible in order for a statement to be falsifiable). Besides, the beans example is perhaps one of the things of which we can be most sure; there are other phenomena that share a similar, though lesser, feeling of a priori which have actually surprised us, empirically. The physics of tiny things supplies us with many such examples: “empty space is empty” — falsified by vacuum energy; “the mass of two quantities combined is equal to the sum of the two quantities added together” — falsified by nuclear fusion and fission.

      Re historical materialism: to the extent that history is part of nature, I’d say yes. I’m not, however, saying that if we created a system that satisfies all the conditions of system X but fails that that is a categorical falsification of the merits of system X — we have to take into account chaos and outside factors (even natural disasters). However we simply must be hard on people who say “it’ll be different next time” every time something fails, without giving a scenario in which they would actually admit to being wrong (otherwise, we have to assume that they’d say the same thing in any scenario).

      Indeed, as you identify, one can’t determine that “socialism has failed” categorically, merely from the example of the USSR; but, if one can’t describe a scenario that would suggest that it had failed, what’s to distinguish that attitude from the doomsday predictors that are wrong every time, but say that they will be right next time. We pick on socialism a great deal, but one can make this claim any political system in the real world: capitalism, anarchism, theocracy, etc.

      Indeed, re creationism: Gould said that religion and science are non-overlapping, but creationists make them overlap. And 100%, re universality; none of our best theories currently are universal: relativity and quantum mechanics are limited by scale, for example.

      Thank you again for the kind comment!

Leave a Reply