Friday, May 16, 2008

Karl Popper and 21st century enemies of science

Nude Socialist has printed another venomous attack against theoretical physics that was written by an individual named Robert Matthews. He argues that science has to be "redefined" but unfortunately every single sentence written in his text is profoundly incorrect.


At the beginning, he claims that the trouble with astrology is that it is not falsifiable. However, astrology is easily falsifiable. It has already been falsified many times. The real problem is that some people haven't yet noticed.

For example, when Antonín Baudyš, a top Czech astrologer and a former minister of defense, used celestial bodies to predict that George W. Bush would die in 2003, it was a prediction. This prediction was falsified on January 1st, 2004. Even if you decided to weaken the statements of astrology and say that it is "substantially more often right than what you would expect by chance", this statement is also falsifiable and has also been falsified, either by direct observations or by indirect arguments based on theories that are supported by other observations.

The actual problem with astrology is not that it is unfalsifiable. The problem with it is that it is wrong. Being wrong is the only problem one can find on any hypothesis that makes material but arbitrarily indirect - and arbitrarily detached from everyday experience - statements about the real world. The process of learning that a hypothesis is wrong is one of the key procedures of the scientific method. Comments that a physical statement is or isn't pleasing, politically correct, falsifiable, or whatever are not really a part of the scientific method.

Science primarily talks about hypotheses' being right or wrong. Whoever wants other adjectives to dominate in science is an enemy of science.

Physicists vs philosophers

These days, many of the "critics" with a dramatic brain limitation similar to that of Mr Matthews like to worship the name of Karl Popper. But give me a break: Karl Popper was just a philosopher. I don't think that there exists a sane physicist who thinks that the opinions of a philosopher are important for science. I have never worshipped any philosopher and the physicists whom I consider good haven't really worshipped philosophers either. The concept of a Karl Popper as a unifying figure above all of science is something beyond me. It looks as silly as an Al Gore above all of climate science. Only popperazzi (or climate whackos) see science (or climate science) in this way.

As Martin Gardner puts it, Karl Popper tried to deny the existence of induction in science. But induction is very important in science - and a combination of induction and observations has been considered the pillar of science for many centuries. Popper was also naively imagining that in science, wrong ideas are being 100% falsified all the time: when a scientist finds something new, it always completely kills an idea. That's of course silly, too.

Real science is non-binary

In science, we are collecting additional experimental and theoretical evidence that makes some hypotheses more likely and others less likely. But in reality, almost no hypothesis is really known to be 100.0000...% wrong much like no hypothesis is 100.0000...% correct.

, the new evidence falsifies a hypothesis almost completely. Such an outcome is more likely when the hypothesis was more hopeless to start with. But it is completely silly to think that every advance in science must be of this character. More typically, new insights help to refine old hypotheses and they change the probabilities of various hypotheses only quantitatively.

And the changes can go in both directions; Popper's idea that they always go down was a sign of his irrational bias: falsification and confirmation are two sides of the same coin. The likelihood goes "mostly" down if most of your working hypotheses are wrong. For the promising ones (that are getting established right now), the likelihood obviously goes up.

And it may be very hard to falsify a very promising theory - almost by definition. A promising theory is a theory that seems consistent with all/most things we can see right now.

The idea that new evidence must always push the probability of some hypotheses all the way down to zero - and that this kind of a "miraculous" killer evidence is even needed for an idea to be considered as science - is absolutely crazy and no actual scientist could believe such a silly fairy-tale because this is simply not how science works (or how it can work). The goal of science is to distinguish the right and wrong hypotheses and the right theories are not guaranteed to be testable by devices we can imagine today, especially not by easily constructable gadgets.

Thinking otherwise would be a textbook example of a wishful thinking. It would actually be an oxymoron because the truly interesting and new science is usually located, almost by definition, on the "boundary" where things start to be hardly testable. After all, that's why we call the location "cutting edge". We often have to wait for new technological breakthroughs before the cutting-edge insights become accessible to our experience.

Should we love Karl Popper more than e.g. Rudolf Carnap or another 20th century philosopher of science because of his particular oversimplified statements? I have no idea. Both of them were just philosophers and both of them have been wrong in many cases.

Popper has promoted some extreme ways to look at science that were different from the previous ways. But the differences have nothing to do with the actual 20th century scientific results. Popper has just made his rules up. I think that Popper's rules, if used as universal rules, have been shown invalid but if you disagree, could you please tell me what I can do to falsify Popper's assertions? You should be able to tell me, right? ;-)

Even physicists are occasionally defending wrong philosophies of physics. Philosophers are much more likely to do so. And they are actually doing so.

Philosophers' work: segments

Philosophers write a lot of ill-defined nonsense and a lot of dogmas that are detached from the details of the real world and/or from the cutting-edge scientific research. About one third of their production is a completely incoherent babbling, another third is shown to be partially correct at this moment, and the remaining third is already known to be fundamentally incorrect. The middle third is paradoxically the most dangerous one because it may inhibit the progress in the future.

Some of their assertions only reflect the philosophical underpinnings of an approximate physical theory and as soon as a more complete physical theory is found, the philosophical assertion has to be abandoned which they are often unwilling to do. Unfortunately, the scientists sometimes join the Inquisition that is defending philosophical dogmas against new scientific insights.

Popper was no different. What did Mr Popper do so important that physicists should pay attention to his words?

Incidentally, no living pure philosopher is defending the principles that are actually likely to be universally valid - such as the superposition principle and other postulates of quantum mechanics. Such things are too technical for them. They always prefer some easily understandable rules inherited from our everyday experience and these rules are never universally valid.

Popper and Weinberg

Amazingly enough, Mr Matthews - after sketching his completely flawed explanation what's wrong with astrology - uses another authority, namely that of Steven Weinberg, to defend his thesis. Matthews tells us that Popper has been "lauded" as the greatest philosopher of science by the likes of Steven Weinberg.

I don't know whether Popper has been lauded by Weinberg but what I do know perfectly is that Weinberg has said many times that philosophy applied to science hurts science. The only good thing that some philosophers have done for scientists was that they have protected them against the preconceptions engineered by other philosophers. This quote can be found at the beginning of the chapter linked below.

In other words, Weinberg agrees with me that philosophical dogmas - and certainly dogmatic requirements about "falsifiability" - are not relevant for science and they have never been relevant. No redefinition of anything is necessary. Given this fact, Matthews' statement that "Weinberg considers Popper to be the most important blah blah blah" sounds misleading, if you want me to describe this vicious misinformation diplomatically.

Because all obvious statements are controversial these days due to the overabundance of aggressive simpletons, I must also offer you a proof of the previous sentence. Well, in his "Dreams on a Final Theory", Weinberg wrote an excellent chapter called
"Against Philosophy".
I invite you to click at the link - the most important external link of this article - and read the whole chapter. You may enjoy it.

Here is Martin Gardner's summary:
Although philosophy may have played a positive role in the work of Einstein, Heisenberg, and others, it has, in Weinberg's opinion, done more harm to science than good. Particularly baleful, he argues, was an early, crude version of positivism that prevented many eminent physicists from regarding unobservable entities as "real." The most flagrant example was the refusal of the Austrian physicist Ernst Mach, who had so strong an influence on Einstein, to admit the reality of atoms or to accept the unobservable fields of general relativity.
Read Weinberg's chapter in its entirety. It contains the word "positivism" 23 times and none of the occurences is really positive. ;-) If something is unobservable, it is true that a scientific theory is not required to talk about it. But it is certainly allowed to talk about it if the concept is useful to make it work smoothly.

Weinberg says very clearly that we shouldn't expect philosophy to be relevant for a modern scientist and, in fact, most philosophers acknowledge this fact. Weinberg also quotes Paul Feyerabend who realizes that most philosophers of science use such a narrow definition of a "scientific explanation" that science - and especially modern science - hasn't really found almost any explanations that would satisfy their criterion. ;-)

Weinberg's - and Gardner's - observation is precious. Positivism might be a nice philosophy - and it may be helpful when we're lucky and it protects us against other wrong ideas - but when the fundamentalist principles of positivism are taken literally as universal rules, they are completely incompatible with the scientific progress. And they have always been incompatible with the scientific progress.

If someone claims that all theories must lead to practically verifiable observational consequences for them to be considered science - and there are numerous individuals who have parroted the same nonsensical thesis at least 687 times - then he simply misunderstands all of physics and all of its history.

Ideas that have been outlived

Virtually every new important advance in the history of physics was connected with a concept that looked "unreal" to the people who haven't yet understood the breakthrough or that violated some of the seemingly essential principles of science of the previous era.

As Weinberg says, the mechanistic way of thinking inherited from the ancient Greece made it harder to accept the action at a distance. It has also led to an unmotivated research of the luminiferous aether and to the grave mistakes of Marxism. It has slowed down the quantum field theory synthesis of waves and particles and so on. In Weinberg's words, it is "foolhardy to assume that one knows the terms in which the final theory will be formulated."

Ernst Mach and others thought that the atoms were not science because they were too small and couldn't have been observed. Mach also believed that the vacuum couldn't carry any "invisible" fields such as the metric tensor because "invisible" fields are not scientific and thus cannot exist: this formulation by Gardner is a very good way to explain the real, deeply flawed motivation behind Mach's principle. Sorry, Mr Mach, but it is up to Nature to decide what is real and what is not. A related but equally wrong philosophical prejudice is currently being sold by various third-class scientists as relationism or background independence (a term that means something completely different in science than what it does in Lee Smolin's pseudoscience).

People couldn't have imagined that the electromagnetic fields, much like the metric tensor, can live in an empty space, without any "material" substance that carries them i.e. without the aether. If there is no "matter" (aether), how could there be any waves in space? This is how the 19th century people used to think - and of course, many people outside the actual cutting-edge research think so even today. Weinberg dedicates several pages to the confused dogmas about the "reality of space and time" and about their beginning.

Others couldn't (and still can't) accept a theory (quantum mechanics) that only predicts probabilities but not a "real" ("mechanist") description of the world, including the degrees of freedom that will "decide" where the particle is observed a moment later. But the objects of quantum mechanics are equally "real" as the old concepts, unless you are dogmatic and wrong about the meaning of the word "real".

There exist hundreds of similar examples and Weinberg explains many of them. All those critics have been wrong throughout the history and their classification of an important concept as "unreal" (or "unscientific") was just a reflection of their ignorance, dogmas, and intellectual limitations, not a valid observation about reality.

Experiments always need a theory

Weinberg also emphasizes that the attempts to only allow "pure" observations to decide about the fate of theories are naive because we always need a theory to interpret our perceptions (at least a theory that explains the relationships between our perceptions and more "external" aspects of reality). Observations can never be freed of theory - all experimental data are theory-laden, as he says - and even most historians of science seem to realize this elementary point. Many philosophers and critics of science don't.

Weinberg offers us many detailed stories about relativity and quantum physics in which physicists were afraid to speak about things that seemed "unobservable". In most cases, this fear was irrational and wrong. In some special cases, it turned out to be justified. But one can never know in advance. Quarks are mentioned as the most dramatic abandonment of the principles of positivism. They are essential to understand nuclear physics but they just can't be observed in isolation: whenever you try to isolate them, you do something completely different (produce mesons). Ernst Mach would surely hate them.

In fact, Weinberg also writes down the very same thesis I wrote many times, too: we are reformulating the physical theory in terms of concepts that are more and more fundamental and at the same time further and further from everyday experience. It seems unlikely that the positivist attitude will be of much help in the future, Weinberg says.

Of course, he was right at least for 15 more years. The only thing that these positivist philosophical dogmas have led to since the mid 1990s is the previously growing - and now already and hopefully dying - anti-scientific tumor around Peter Woit and tons of his equally idiotic peers.


The rest of the chapter in Weinberg's book focuses on philosophical relativists who try to deny the objective character of scientific insights - i.e. postmodernists of various types. Unfortunately, the list includes ex-president Václav Havel. Although Weinberg has written some nice things about Thomas Kuhn - e.g. that he didn't really agree with the concept of a non-objective science - he describes his work as a revolution that didn't happen elsewhere.

But back to Kuhn's non-objective heirs. While science as a human activity is a social phenomenon, it simply doesn't imply that the results are social constructs.

Finally, Weinberg conjectures that the postmodernists are partly motivated by a mutated positivism, partly by their desire to (erroneously) feel superior when they look at the scientists, and partly by their general hostility to the Western civilization. The latter motive is becoming irrelevant for the criticism of science because science is getting truly global which is, in Weinberg's view, a good thing.


  1. You cant really be an "enemy of science" because science is just a set of ideas and postulates, and science as a set of abstract and potentially useful ideas cannot be the enemy of anyone.

    Rather than being hysterical about the "enemies" of science, it would be far more productive to understand what science is really all about---and what its relation to culture is altogether.

  2. Quite on the contrary, Sue, Science is not primarily a disjoint union of separate insights. It is predominantly the methods and compact package of principles that are necessary to get new insights. And one can be an enemy of this thing, indeed. There are many people who unquestionably deserve this label.

    Concerning the relation between science and culture, in Feynman's words, the image of the world as painted by state-of-the-art science is the main part of the genuine culture of our epoch. That's why the enemies of science are uncultural barbarians, too.

  3. Nice work,linked to you in my blog here:

  4. Lumo, have you ever read Feyerabend's 'Against Method'? I was wondering, he claims some weird stuff in there, like "brownian motion refutes de second law" and "general relativity don't agree with the precession of mercury perihelion", did anyone bother with that?

  5. Tonight I came upon Popper's fat book 'Conjectures and Refutations' in a used bookshop. I wanted to buy it, but didn't fortunately.

    The elegant idea behind all of science is so easy to understand and is so sensible that only an obtuse philosopher could question it:

    You make a hypothesis 'a,' do a relevant experiment, and either increase or decrease the probability that 'a' is true—never being able, only in principle, to make the probability of 'a' either 0 or 1.

    The hypothesis, e.g., 'Quantum mechanics [QM] makes incorrect predictions' can't be ruled out in principle. But in practice it can be. For, since we have so much evidence that QM makes correct predictions, we can say, for commonsense's sake, that the probability of the truth of the hypothesis 'QM makes incorrect predictions' is 0.