7

Why should science be falsifiable?

Furthermore, should disciplines - like astrophysics that are extremely hard to experiment on - be considered as a science in a standing point of philosophy?


I know what falsifiability is, but I would like to know the philosophical reason (if there is one).

I am thinking whether if falsifiability is a functional principle (to be given up should other more productive means appear) or falsifiability is essential to the nature of science (in other words, it is impossible for other better means to exist, science and falsifiability overlap precisely).

  • Science is not falsifiable (or at least that not what Popper et al. proposed); scientific hypotheses are. – Dave Aug 6 '15 at 20:23
  • @Dave, you're right of course. but it's a semantic difference. we could define our definition of "Science" to share the same properties or some of the same properties of scientific hypotheses, such as "falsifiability". – robert bristow-johnson Aug 6 '15 at 22:25
  • 1
    Shing, my answer is below. one thing i would say here is that there are many, many observations in astronomy that do support (and sometimes refute) predictions made in astrophysical theories. like we have a pretty good idea about what the cross-section of a star is (the temperatures at different levels) and we have never sent a probe into any to measure anything. but we have telescopes and they tell us quite a bit. but it does get more speculative when you get to cosmology. not totally, though. i would say that the Big Bang has supporting observational evidence. – robert bristow-johnson Aug 6 '15 at 22:30
  • actually i think it is the predictions should be falsifiable, rather than the hypotheses. Some hypotheses are partly definition, which is hardly falsifiable. (e.g. Newton's first law) – Shing Aug 7 '15 at 9:56
  • I can't quite wrap my mind around this question. If science were different to the extent that is didn't play by the rules of falsification, then it wouldn't be science - I'd say that's part of the definition and you can't remove that. Further (and maybe it's because I'm too much of a relativist) an answer to "why should..." is always just asking for opinions. I'd need a moral or economic argument to tell you why something should. If the question made sense to me (see part 1 of this response), then I'd not give you a reason because I don't need science to bow to my will. And I'm a scientist. – Nikolaj-K Aug 7 '15 at 11:27
11

In natural sciences like physics or astrophysics one cannot prove general theorems, i.e. theorems which hold for infinitely many cases. Because verification by experiment only allows to verify finitely many cases. Hence the possibility remains that cases have been left, where the statement in question does not hold.

That's the reason why one cannot prove scientific theories. Hence one needs other criteria to keep or to abandon a scientific theory.

As a consequence from this insight, the philosopher Karl Popper proposes not to verify but to falsify scientific theories. Because - theoretically - one single counter example suffices to falsify the theory. The fundamental book and a cornerstone of Critical Rationalism is

Karl Popper: The Logic of Scientific Discovery (1934)

This means: Science advances by eliminating errors, which prompts new hypotheses.

Added. Or a bit sloppy, but the bon mot of the philosopher and physicist Gerhard Vollmer hits the point:

"We err upwards."

  • I have edited my question for a few question I would like to know further, please feel free to answer them, I am not sure if it breaks the rules of this site to make it like discussion, please tell me if I did. I will correct that. – Shing Aug 7 '15 at 10:03
  • I agree to take falsifiability as a guiding principle for deciding about concurrent theories in science. One should only propose theories which are falsifiable - that's Popper's criterion of demarcation between science and metaphysics. Roughly speaking, the scientific process starts with collecting new data from experiment or observation. In order to explain the data, the investigator updates an existing theory or creates a new one, possibly with new concepts. A falsifiable theory then explains the results and predicts the outcome of further experiments. – Jo Wehler Aug 7 '15 at 20:31
  • In case the results do not confirm the prediction, the theory is falsified and must be changed. Popper did not detect a new scientific method. He only made explicit what is usually done in science. Of course one cannot prove that there does not exist a better method than falsification. But up to now, no one has found a better one. - Popper's view taken as description of the scientific progress has been criticized by other philosopher of science, notably by Thomas Kuhn. – Jo Wehler Aug 8 '15 at 13:15
  • It is very hard to choose from all the good answers, so I accept this one made things most simple but not simpler. plus "We err upwards." I love this quotation. – Shing Aug 15 '15 at 15:43
4

One way to suggest natural or physical sciences should be falsifiable or falliable is to ask what it would mean for them to be unfalisfiable or infalliable. This would mean that they would be incapable of making mistakes or of producing errors. This sounds like it would be beneficial, but it would be methodologically disastrous.

This would mean our science cannot say anything unless what it says were such that it could never be wrong. Which would consequently mean that science would have to remain silent or merely describe possible world physics for worlds that may or may not be similar to our own. Since these sciences are our understanding of our physical world, they need to reflect our limits.


The second part of your question relates to the boundaries of what is and is not a science. First, there's a linguistic problem. While we sometimes use "science" to refer to just the physical or natural sciences, the word science is the English derivative of the word scientia which means knowledge. Thus, the word is sometimes understood more broadly to refer to any form of knowledge. This sort of usage is also common in German sources where the German word Wissenschaft has a broader meaning.

To make everything simpler, we will limit ourselves, however, to those sciences which are falsifiable and describe our physical world,

As you point out, there are some things that call themselves "sciences" that do not do experiments. But doing experiments is not the only way to have falsifiability. Astrophysics for instance involves the creation of theories and accompanying equations about how celestial and planetary bodies work. It can then observe whether these things occur or not and thus remain falsifiable. Moreover, astrophysics and particle physics are intimately related such that the experiments of particle physics can falsify claims in astrophysics.

2

In my opinion, the reason why, ultimately, real science is eventually falsifiable is to differentiate it from philosophy in general. I have an answer to a similar question.

The whole purpose of developing a new scientific theory is that the new theory makes a difference from the existing or previous theory in some way. If the theory is never falsifiable, then it makes no difference whether the theory is valid (or more valid than the old) or not. If it makes no measurable difference and has no hope in ever making a measurable difference (from the old theory), then it is speculation about reality no different than philosophizing.

However, if the new theory says that in some specific scenario, reality will appear a little different than it would as expected from the old theory, then the new theory actually says something: it actually makes a difference.

Suppose general relativity (GR) made the same prediction about the procession of the perihelion of Mercury that the Newtonian theory of gravitation (and motion) does and if GR did the same for all other physical scenarios, then what point is it in existing? Why make use of (or believe in) a more complicated new theory of reality that has outcomes no different than the simpler older theories?

Only if a theory of reality makes a difference in the tangible experience of physical reality is it of consequence. Only such theories of consequence are what science is about. Otherwise it's undifferentiated from astrology, alchemy, religion, or philosophy.

If the theory does make a difference, that difference should be expressed explicitly, then someone can think up an experiment (or observation) of the physical reality that tests that difference. That's what falsifiability is. If the procession of the perihelion of Mercury was more as Newton would predict and less as Einstein would, then GR would have been tested and falsified and we would not be working on using that model to develop other models to explain things.

However, the observation (and a few other observations) went the other way, and Newton was falsified. But not so badly that we don't use Newtonian mechanics and gravitation for many, many applications with slower speeds and smaller gravitational fields. Newtonian gravitation and mechanics are a very good approximation to reality. Einsteinian GR is a better approximation. Maybe some day someone will show how GR fails some observation and the human species will be on to developing a better model than GR.

  • I have edited my question for a few question I would like to know further, please feel free to answer them, I am not sure if it breaks the rules of this site to make it like discussion, please tell me if I did. I will correct that. – Shing Aug 7 '15 at 10:03
1

The term falsifiable is not well agreed upon. The most well agreed upon version is statistical in nature (which is rather ironic, if you ask me). However, an answer to your question can be given if one translates falsifiable as "designed to be easily discredited if wrong."

The power of falsification is that it allows one to proceed on the assumption that the claim is probably true, knowing that, if anybody cares enough about the derived results, at a later date someone can falsify the entire chain of theories by taking the time and expense to contradict the original thesis. This has allowed science to take tremendous leaps forward at an unprecedented pace. At any time, if someone considers the results "wrong," (technically or ethically), they can pay the price to falsify it later.

As claims become more difficult to falsify, it becomes harder to build off of them, entrusting the displeasure of others will eventually test the underlying hypothesis.

There are other approaches, of course, besides falsifiability, so when you ask "why should science be falsifiable," it is only reasonable to look at alternatives. One is the "do no harm" approach of Traditional Chinese Medicine. The logic there is that, if a false claim does no harm, you can afford to have many false claims out there, and let entropy weed them out.

The different approaches have different strengths and weaknesses. For example, science primarily focuses on objective results, eschewing subjective results. When subjectivity is needed, that subjectivity is objectified first, and then tested. The approach of Chinese philosophy assumes everything is in the tiniest bit subjective, and seeks the least subjective bits to be written down as theories.

Which one satisfies one's needs is very much dependent on what they want out of the philosophy. For example, mob mentality is a hazy topic in psychology that we are still working on slowly within science. It is a fundamental aspect of Chinese philosophy, and they've understood it for thousands of years. Clearly there are cases where the alternative is superior.

However, the reason I would say science needs to be falsifiable, at least in some degree, is in the expense of testing it. Modern science requires laboratory equipment well out of the price range of 99.9% of the population (and the remaining 0.1% that can afford it is too busy making money). These resources need to be spent as efficiently as possible. Intentionally melding falsifiability into our hypotheses allows us to be as imaginative as possible as to how the universe could work, and efficiently weed out these possibilities as tests are done. For every test of the Higgs-Boson we hear about, there are probably a hundred theses which have made predictions about the universe that we tangentially falsified by one LHC test. Trying to do it the other way, tangentially verifying theories, is much harder, so we would not get as much bang for our buck.

1

edited: I am thinking whether if falsifiability is a functional principle (to be given up should other more productive means appear) or falsifiability is essential to the nature of science (in other words, it is impossible for other better means to exist, science and falsifiability overlap precisely)

The best means to explore this part of your question is science fiction. Suppose a more productive means exists. Say there is a drink that gave you and the entire human race perfect knowledge on a subject you concentrate on while you drink it. Setting aside the way it works, magic, technological, religious. So long as it confers and communicates correct knowledge this would be a better means (assuming no prohibitive drawbacks). If it existed and were adopted we could learn a lot. But that doesn't seem like science at all. It's seems closer to book learning. I do think it's possible that an undiscovered "better means" may exist. If it did though using it wouldn't be science.

The essential element of falsifiability is the challenge to "prove me wrong". Without that the only way to fight a bad idea is with a better idea. But now we're arguing over what makes an idea better. Which is not exactly a hard science.

While proving a negative is impossible, falsifiability does have a flip side. Make a prediction. A theory that makes predictions that others couldn't will gain respect for being useful.

0

This notion is Popper's, and it is a good one, but is not universally held.

During periods of normal science, or when total falsification can be justified without depending deeply on a theory, falsification gives the most statistically stable means of moving forward.

The entire mechanism of modern statistics is built around the 'null hypothesis', and stating up front exactly what it would mean for a theory to fail to account for experimental results. This boils 'partial failure' down into measurable numbers and allows us to accumulated them and trust what they mean in terms of our real likelihood of being wrong.

So when you can use this criterion and apply these methods, you can converge on the most reliable answer in a way that can be validated by one's peers, and can give them some notion of how much risk is involved in accepting the theory if it is marginal.

Of course, where there are people involved, the math is not completely trustworthy, because practice involves interpretation, etc., and scientists still have judgment calls to make. But when Popper applies, it is gives decisions a lot of accountability.


But it remains an open question whether all science should be approached in this way, even when it can. Many folks after Popper have contested the power of this method when things are less stable. Kuhn, for example has suggested that science regularly moves into 'revolutionary' periods, when determining falsifiability is not always possible, and therefore it cannot be required. (Feyerabend goes even further, claiming that science is constantly at least partially in such a state, and should attempt to remain so.)

When a science is trading out older paradigms whose creative energy has been spent, and deriving new paradigms, the old and the new can be incommensurable in a way that obscures our ability to know what exactly 'falsified' means in the context of the science in flux.

When a science is close to a revolutionary boundary, just descending into or just getting out of a paradigm switch, demanding falsifiability it is not always the best choice. It may shoot down new ideas based on habits and older ideas that are about to be shot down, themselves. Potential new paradigms need slack to establish a footing, and preserving parts of the old paradigm that still happen to remain clear in the new interpretation may undermine cohesiveness in the new foundation coming into practice, which will make for a complicated mess in the future.

If, like Feyerabend, we prefer every science should always have competing sets of underlying assumptions, and avoid 'hegemonizing', then falsifiability becomes something optional, if still very convenient, in a theory.


You can easily claim that the main parts of psychology are not supported by a single underlying paradigm, but that psychology is so young a science that it is continually in revolution, and will continue to have competing basic paradigms for the foreseeable future.

That excuses the fact that its basic motivating theories, from psychoanalysis to Cognitive Behaviorism are basically non-falsifiable, but should not simply be dismissed as non-science.

And you can see how good statistical practice is still useful, even though there is no ultimate chance of falsifying, e.g. the idea of the ego.

  • Reference for Kuhn claims that falsifiability is not applicable in periods of revolution? – Dave Aug 7 '15 at 16:20
  • @Dave There is a big gap between saying that in those periods it is 'not always possible' and saying it 'does not apply'. I feel you have done an unwarranted quantifier replacement. There exist statements during paradigm shift that can be correctly tested and proven false but still remain useful and ground new theories, because their meanings will change. That does not mean that all falsification during a whole time period is pointless. – user9166 Aug 7 '15 at 17:26
  • You can find discussions of incommensurability in The Structure of Scientific Revolutions, in his defenses of it in Criticism and the Growth of Knowledge. But it is well captured in the sections under its title in the SEP page on Kuhn plato.stanford.edu/entries/thomas-kuhn – user9166 Aug 7 '15 at 17:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.