[T]here is another special kind of boldness – the boldness of predicting aspects of the world of appearance which so far have been overlooked but which it must possess if the conjectured reality is (more or less) right, if the explanatory hypotheses are (approximately) true. It is this special kind of boldness that I have in mind when I speak of bold scientific conjectures. It is the boldness of a conjecture which takes a real risk – the risk of being tested, and refuted; the risk of clashing with reality.
Thus my proposal was, and is, that it is this second boldness, together with the readiness to look out for tests and refutations, which distinguishes ‘empirical’ science from non-science, and especially from pre-scientific myths and metaphysics.
I will call this proposal (D): (D) for ‘demarcation’. (Karl Popper ‘The Problem of Demarcation’)
At the heart of Popper's proposal is the concept of falsifiability (he also saw this as the way around the problem of induction). He thought that the demarcation between scientific and unscientific theories should be based upon whether the theory in question is falsifiable or not - with falsifiable theories being considered scientific, and non-falsifiable theories unscientific. Popper's proposal excludes from the domain of science not unfalsifiable statements, but whole theories that contain no falsifiable statements.
Moreover, to be properly falsifiable, a theory should make clear, unambiguous, and bold statements that can be compared empirically against reality. If the theory’s predictions are vague and equivocal, then it will be difficult to falsify, since it is not clear what would constitute a failed prediction. Equally, if a theory merely predicts things that we already know to be true, then we have no good reason to favour this theory over any other that merely predicts the same observations.
For example, Einstein’s Theory of General Relativity predicts that the Earth will orbit the sun in an elliptical orbit. But this was already predicted by Newton’s theory of gravitation, so we have no reason to prefer Einstein’s theory over Newton’s existing one. However, Einstein’s theory also predicted that massive objects bend light – something that Newton’s theory does not predict, and which had not previously been observed. This bending of light was duly observed during an eclipse in 1919, and the fact that relativity’s bold prediction was not falsified counted strongly in its favour.
It should come as little surprise that physics, for example, would be classified as a science according to Popper’s proposal – since Popper no doubt formulated it with this in mind. Theories in physics generally make plenty of clear, unambiguous, and bold statements that are open to falsification. Whilst attempts would be made to rescue theories in the light of anomalous observations, these attempts would often make the theory even more falsifiable, not less. Theories that consistently make false predictions would be rejected in favour of better theories that don’t.
Now let’s have a look at what Popper’s proposal would have to say about astrology. Would it qualify as a science? Popper himself gave this as an example of something that clearly failed his falsifiability criterion. There are two problems with falsification in astrology. Firstly, the predictions that it makes are often sufficiently vague or ambiguous that it is very difficult to establish what would constitute a failed prediction. For example, tests have been carried out in which professional astrologers have been asked to cast a horoscope for a specific person. This reading has then been given to a large and disparate group of other people, purporting to be a reading done for them specifically. In almost all cases, the subjects who were given the horoscope rated it as highly accurate for them. This is because the types of statements made in such horoscopes are sufficiently ambiguous and open to interpretation that they can be made to fit many diverse people.
Secondly, if we do somehow manage to pin down astrological predictions sufficiently unambiguously to carry out a rigorous empirical test in an attempt to falsify it, then failed predictions don’t seem to be accepted as such by the astrologers in general. It might be reasonable to ignore or explain away a certain amount of anomalous data, as science does, but the astrological community seems prepared to accept nothing as constituting a falsification of their theory. So, it is also unfalsifiable in practice.
Let’s look at an example. Imagine that I am a Sagittarius, but that I don’t have typical Sagittarian personality traits. Would this qualify as a means of falsification? I think that astrologers would give little weight to this counter-example, as they would likely say that one’s birth sign gives nothing more than a tendency to have certain personality traits. There will always be cases of people not following that tendency, they would say. So, I think that this example is a very weak one, and really counts for little. Astrological theory allows for this kind of anomalous data, so it would certainly not be accepted in the astrological community as constituting some sort of a falsification.
How about a much stronger example? Astrological theory says that the accuracy of its predictions increases dramatically when we consider the personalities and lives of people born at almost the same time. Accordingly, there was a well-known and rigorous study done of 2000 people born within a few days of each other (70% born within 5 minutes of each other), in which their personalities and lives were analysed statistically to look for the types of correlations that would be predicted by astrology. However, no such correlations were found. A sure falsification you might think? No such thing I’m afraid – the astrological community ignored or attempted to explain away the findings by ad hoc means. Within the vague and equivocal framework of astrology, I can think of no more clear falsification of astrological theory than this (although there are plenty of other studies with similar results), but the astrological community refused to accept it as such. So, even when it is possible to pin down the types of vague and ambiguous predictions made by astrology, and show them to be failed, falsification is rejected. Therefore, falsification is impossible in practice within astrology.
So, to summarise, we have two problems with falsification in astrology. Firstly, the theory is sufficiently vague and imprecise that it is difficult to frame clear, unequivocal, and bold tests that would allow us to falsify astrology. Secondly, any reasonable attempt at such a falsification, as in the study cited above, will not be accepted as a falsification. So, in practical terms, astrological theory has been rendered unfalsifiable. This is in clear contrast to science. Whatever science’s epistemological limitations, scientific theories do almost always make clear and unambiguous predictions that can and do allow for their falsification. For example, as the biologist J.B.S. Haldane said, evolution would be disproved by the finding of "Fossil rabbits in the pre-Cambrian". This is a little simplistic, as scientists would naturally and justifiably look for alternative explanations for this anomalous data (justifiable in this case as evolutionary theory is supported by a huge amount of data, unlike astrology). However, there would come a point at which scientists would just admit that their theory is wrong. Nothing analogous ever seems to happen in the world of astrology.
Whilst scientists might and do resist such falsifications, the history of science shows countless examples of scientific progress through theories being falsified and replaced by better ones. So, the scientific method is a truth-seeking one that progresses despite the limitations of individual scientists and of other epistemological debates. By contrast, astrological theory has remained almost set in stone since Ptolemy’s day, despite huge problems with its methodology (no proposed physical mechanism for these supposed planetary influences, problem of the precession of equinoxes etc.) and its predictions. So, to summarise, astrology would clearly fail Popper’s criterion of falsifiability, and would therefore not be classed as a science.
But, at this point an interesting question to ask is whether falsifiability alone is sufficient to allow us to demarcate science from non-science. One problem with this idea is that science often seems to progress by verification, rather than falsification - scientists aren’t always looking to falsify theories but, rather, to verify them.
Nevertheless, I think that it is a fundamental part of the scientific method that physical theories should be falsifiable - even if this is not what physicists are inclined to do. Of course, physicists will we wedded to their preferred theories, and will look to verify these theories, and try to avoid any possible falsification - to the possible extent of ignoring or 'fudging' anomalous data. However, this is a human failing, and not a failing of the idealised method of physics, in which I think that properly constituted theories should offer some means of falsification, and falsification should be attempted. I think that one important aspect of the scientific method here is that, whilst physicists might seek to avoid falsification of their pet theories, other physicists will be attempting to achieve that very falsification, in order to push alternative theories i.e. we have peer review.
Of course, the theories might not falsifiable yet e.g. with the proposed existence of the Higgs boson as predicted by the Standard Model. However, even such esoteric cosmological theories as Turok and Steinhardt's colliding branes offers some means of falsification - the detection of gravitational waves from the creation event would falsify it, for example. I think that most physicists would not consider a theory to be a properly scientific one if it offered no way of ever being falsified, even in principle.
Of course, scientists are not just going to discard a theory when some anomalous data turns up. They will try to introduce some additional, possibly ad hoc, element in order to explain away the mismatch between theory and observation. This happened with the discrepancy between the observed orbit of Uranus and what was predicted by Newton’s theories. Scientists got around this discrepancy by proposing the existence of another, unknown, planet that was influencing Uranus. This actually turned out to be correct, and was the planet Neptune. So, you might ask, what is the difference between a scientist doing this and an astrologer (or creationist etc.) doing it? Well, the difference is that in the case of science, these ad hoc elements introduced to explain away a mismatch will likely make the theory more falsifiable, not less. Also, science doesn’t always do this, whereas pseudoscience seems to do nothing but introduce ad hoc elements into their theories in order to explain away discrepancies. By doing this, the pseudosciences are rendering their theories effectively unfalsifiable.
In the final analysis, I think that Popper was on the right track with his concept of demarcation based upon falsification, but I think that there is more to it than this. I believe that falsification is a necessary but not sufficient condition for a theory to be considered properly scientific, and I think that we need to add some criteria in order to properly demarcate between science and pseudoscience. For a theory to be considered scientific, I would suggest that we would wish it to have most or all of the following properties: consistency, parsimony, falsifiability, grounding in empirical evidence, reproducibility, tentativeness, and correctability.
1 comment:
An excellent and eloquent explanation of Falsification! Thank you for writing this. Do you have a facebook page as well?
Post a Comment