The “scientific method”, a needless stumbling block. With a note on falsification
Science does not have a separate special method for learning about the world, the “scientific method” as taught in schools is a damaging illusion, and the falsifiability criterion has itself been falsified
Below, R: How not to; “The Scientific Method”, as inflicted on Science Fair participants. Click to enlarge
Consider this, from a justly esteemed chemistry text:
Scientists are always on the lookout for patterns.… Once they have detected patterns, scientists develop hypotheses… After formulating a hypotheses, scientists design further experiments [emphasis in original]
Or this, from a very recent post to a popular website:
The scientific method in a nutshell:
1. Ask a question
2. Do background research
3. Construct a hypothesis
4. Test your hypothesis by doing experiments
5. Analyze your data and draw conclusions
6. Communicate your results [emphasis in original]
Then, if you find yourself nodding in agreement, consider this:
Since a scientific theory, by definition, must be testable by repeatable observations and must be capable of being falsified if indeed it were false, a scientific theory can only attempt to explain processes and events that are presently occurring repeatedly within our observations. Theories about history, although interesting and often fruitful, are not scientific theories, even though they may be related to other theories which do fulfill the criteria of a scientific theory.
If you are familiar with the creation-evolution “controversy”, you may well suspect that last example of being so much creationist waffle, intended to discredit the whole of present-day geology and evolutionary biology. And you would be right. This quotation is from Duane Gish, a major figure in the twentieth century revival of biblical literalist creationism, writing for the Institute of Creation Research.1
Such nonsense isn’t funny any more, if it ever was. The man who may very soon find himself President of the United States is an eloquent spokesman for creationism.
And yet Gish’s remarks seem to follow from the view of science put forward in the first two excerpts. What has gone wrong here? Practically everything.
Consider the first great accomplishment of modern science; working out the laws2 of planetary motion, and Newton’s explanation of those laws in terms of his theory2 of gravity. Copernicus, picking up on an idea that dates back to the ancient Greeks and was also well-known to the astronomers of Islam’s Golden Age, treated the Earth as a planet like any other, and had the planets circling the Sun. Kepler showed that the orbits were in fact, to a very good approximation, ellipses, and found out how a planet’s speed varied during each rotation, and how the length of a planet’s “year” depending on its distance from the Sun. Finally, Newton showed that Kepler’s laws could be explained using his theory of gravity and his laws of motion, and that the same set of laws explained the motion of the Moon, and the downward acceleration of falling bodies on the Earth.
So where are the experiments, said in the first two extracts to play an essential role in testing a hypothesis? Regarding falling bodies, such experiments were carried out by Galileo, but it is only within my own adult lifetime that we have been able to do what might by any stretch be called “experiments” on satellite and planetary orbits. However, even when we can’t do actual experiments, repeated observations will do very nicely instead.
For an embarrassing moment, I thought that the creationist had got that bit right. Then I re-read more carefully: “a scientific theory can only attempt to explain processes and events that are presently occurring repeatedly within our observations.”3 Of course, we can obtain better quality information if we can repeat our observations. Indeed it was only repeated observation that made it possible to work out the orbits of the planets. But the creationist has committed what I might call the logical fallacy of the misplaced epithet; it is not the processes and events that need to be repeated, but (if possible) the observations.
Below: Impact of Shoemaker-Levy 9 fragment G on Jupiter
In 1994, I like many millions of others sat enthralled in front of my television set while the fragments of Comet Shoemaker-Levi 9 made their death plunge into Jupiter, under intense scrutiny from Earth-and space-based telescopes, while from the far side of the planet the Galileo spacecraft observed the impacts in real time. Earlier calculations had shown that the comet had been broken into fragments in 1992 during a previous close encounter with the planet, while spectroscopic studies of the impact plumes showed the presence of previously unsuspected sulphur-containing molecules in the Jovian atmosphere. This was not an example of “processes and events that are presently occurring repeatedly within our observations.” It was a one-off. And yet if this was not science, what is?
I chose this example because it straddles the boundary sometimes drawn between “experimental”4 and “historical” science. Jupiter has been around for a long time, and, if science survives, we or our descendants will be learning yet more in due course about those sulphur-containing molecules and how they came to be there. So in that sense we are looking at observations that might be repeated. As for the impacts themselves, Comet Shoemaker-Levi 9 was unique in our experience, the only occasion on which we have seen such an event as it happened. And yet it can hardly be unique in the history of the Solar System. While preparing this piece, I learnt of evidence that a red stain across the surface of Pluto was caused by a giant impact. It is also now thought, on excellent evidence, that a major impact played a crucial role in the mass extinction that took place at the end of the Cretaceous, including the extinction of the ammonites (after a 350 million year run) and, famously, the dinosaurs.5 This of course takes us into historical science par excellence; the intertwined stories of our planet, the life it supports, and how both have changed over time. No wonder the creationists want to pretend that historical science is inferior to experimental science, if indeed it qualifies as science at all.
So what is science, what is the scientific method, and where do the boundaries lie? Here I would quote the Nobel prize-winning biologist, P.J. Medawar, who has thought more deeply about the practice of science and its philosophical underpinnings than almost any other scientist I can think of. It was Medawar’s work on immunology that made transplants possible. Medawar also wrote the invaluable Advice to a Young Scientist, which I read in my forties, and wish I had read much earlier. But here I offer two quotations from his 1967 collection, The Art of the Soluble:
Among scientists are collectors, classifiers, and compulsive tidy years-up; many are detectives by temperament and many are explorers; some are artists and others artisans. There are poet-scientists and philosopher-scientists and even a few mystics.
Is there such a thing as a “scientific mind”? I think not. Or the scientific method? Again, I think not… [A] scientist, so far from being a man who never knowingly departs from the truth, is always telling stories… stories which might be about real life but which have to be tested very scrupulously to find out if indeed they are so.
Here I side with Medawar. There are real differences between historical science and experimental science, but I hope I have convinced you that there is no sharp boundary between them. Still less is there a boundary between questions like “What killed the dinosaurs?” which is certainly a question in (historical) science, and “Who killed JFK?”, a question in history. There are often useful distinctions with fuzzy boundaries, such as the distinction between blue and green in the spectrum, but I do not think that the difference between “scientific” and “non-scientific” knowledge of the external world is of this kind. We can ask whether a statement appears to be true, what reasons there are to accept it, how much confidence we should have in those reasons, and what it would take to make us change our minds about it. These are useful questions. But having answered them as best we can, to turn around and ask “Is this statement scientific?” adds nothing to our understanding.
Worse. Imposing the hypothetico-deductive model (as it is called) on science in general actively hampers discovery. To quote Craig Venter, whose maverick approach has done so much for modern genetics,
NIH grants have to be based on hypotheses. Well, my hypothesis is, we know ‘diddly’ about most genes. So why not, ‘I want a grant to look at that gene and see what it does.’ But we don’t have the mechanism as a scientific society to deal with this. That’s why I started TIGR. We deny that biology is descriptive, but it is.
But what about falsifiability, a venerable criterion, dating back to Karl Popper’s work in the 1930s (it is no accident that Popper and Medawar were friends)? The creationists love this one, because it enables them to draw a sharp contour across the rolling landscape of reality, claiming that every unsolved puzzle, and every correction of an error, shows evolution to be false. Much better is the blogger’s “Test your hypothesis.” But the blogger does not tell us what to do if the hypothesis fails the test. Here’s my take on that.
Consider a trivial case, the hypothesis that “All swans are white”, which seemed for a long time to be a true but not very interesting statement about a kind of bird. So what do you do if you come across a black swan, close enough in all its features, except its colour, to be unmistakably swan-like? Exactly this happened to Willem de Vlamingh in 1697, when he was exploring what is now Western Australia. You can do one of three things. You can protect your hypothesis, by narrowing down your definition of “swan”, to exclude non-whites, in this case turning the statement into a tautology. Not very helpful. Or you can say that the hypothesis has been falsified, because you have found a counterexample. That just throws away a generalisation based on thousands of years of swan-watching; not very helpful either. The smart thing to do is to modify your hypothesis; all swans are white except the Antipodean black swans. You can then start asking (and nowadays, using the techniques of molecular biology, answering) interesting questions like, are the white swans and black swans more closely related to each other than to any non-swan-like birds, how long ago did they diverge from each other, and is there any reason why black swans should be favoured in Australia and white swans in Europe, or is it just a matter of random drift? Modification, rather than falsification, is the probable fate of any hypothesis worth its salt when faced with a counterexample.
Apply this to our very first example, the motion of the planets. Around 1820, it was realised that the orbit of Uranus deviated from Kepler’s Laws, even after allowing for the perturbing effect of the other known planets. This was not regarded as a falsification of existing planetary science, but as showing the need for a modification, in this case adding to the known solar system a new planet beyond Uranus, and in 1846 that planet (Neptune) was found where predicted.
Compare the much more troubling case of Mercury. The orbit is, as predicted, almost exactly an ellipse, and the direction of the ellipse changes, as predicted, under the gravitational influence of the other planets, but the speed of change of direction is slightly greater than could be accounted for by Newtonian mechanics. This anomaly had been known since 1859, and was only explained in 1912, when Einstein’s General Theory of Relativity showed how space and time themselves are distorted in gravitational fields. So Newtonian mechanics was used for over fifty years without demur despite having been falsified. The falsifying data were referred to as an anomaly, an exception to the laws, for which, it was hoped, an explanation would eventually be found, rather than as reason to reject the entire structure. And while it would eventually transpire that the problem lay at the very roots of Newtonian mechanics, almanac-makers still carry on using those mechanics to predict next year’s tides. And are right to do so.
And finally, back to historical science. My first three quotations belong to a tradition that regards it as inferior to experimental science, if indeed it qualifies as science at all. This, I think, is a mistake. Indeed, time and again, when historical and experimental science have come into conflict, it turned out to be the experimental science that required modification. More on this later.
1] Other creationists do recognise that the past is open to scientific investigation, but make great play of the fact that reconstructing the past involves interpretations and assumptions, as if that were not true for all our reasoning about the world. They do this in order to claim equal status for those statements that are, and those that are not, supported by material evidence.
2] In this context, a law is just a general statement, briefly and often mathematically expressed. Calling something a law, rather like calling something a theory, tells us absolutely nothing about how closely it matches reality. Newton’s laws of motion match reality very well provided things are not moving too fast and the local gravitational field is not too strong. Boyle’s Law, which Michael Gove, former UK Education Secretary, thinks is a fundamental principle, says that the volume of a gas is inversely proportional to pressure. This is only roughly true, and even then only if the temperature is not too low and the pressure is not too high. Bode’s Law, which says that each successive planet should be twice as far from its sun as its predecessor, is simply wrong.
I will not even attempt here to explore the morass of multiple meaning and malicious misdirection associated with the word “theory”.
3] Strictly, of course, we can never observe anything that is presently occurring, because the very act of observation takes time, meaning that all our knowledge is based on information about the past. And events on Jupiter can only be observed on Earth over half an hour later, because of the finite speed of light. But let that pass.
4] Sometimes called “nomative”, from the Greek word for “law”; science looking for general laws, rather than for specific explanations.
5] Or, some prefer to say, non-avian dinosaurs. Some people regard birds as dinosaurs, others merely as their descendants. Semantics, really.
Scientific method image from University’s guidance to Science Fair participants. Shoemaker-Levy impact, Keck telescope infrared image (NASA), public domain. Black swan image, swans introduced to Lisbon’s King dward VIII Park, photo by author
An earlier version of this piece appeared in 3 Quarks Daily. I thank Ken Pidcock, whose comment there alerted me to Venter’s remarks.
Posted on May 29, 2017, in Creationism, Education, History of Science, Philosophy and tagged Craig Venter, Experimental science, Historical science, Karl Popper, Kepler's Laws, Medawar, Mike Spence, Philosophy of science, Scientific method, Shoemaker-Levy. Bookmark the permalink. 12 Comments.