Science and Truth

New Scientist vol 165 issue 2226 – 19 February 2000

Instruments such as microscopes or telescopes can reveal the physical world to us, showing a cell’s walls or a distant star system. To examine science itself, you need a different approach: you must think about thinking. The philosophy of science uses step-by-step examination to reveal good or bad science

DOES science tell us the truth? How do we tell the difference between science and non-science? If one group of scientists says that genetically modified foods are harmless and another says they are dangerous, who should we believe? To answer these questions we must think about the way scientists reach their conclusions.

Science’s goal is to discover the laws of nature, which we assume exist independently of humans. We find these laws by collecting facts and assembling new theories to explain them. Good science is conducted publicly. Scientists release their results in a way that allows others to scrutinise them and try to duplicate them or show that they are wrong. Few people seriously doubt that science works. It has been hugely successful in giving us explanations of the world around us. It has the power ultimately to explain all natural phenomena, even if in practice some problems are proving very difficult. Science has also allowed us to create technologies such as drugs to treat cancer or the laser in your CD or MiniDisc player.

WHAT IS SCIENCE?

Testing ideas

No one has yet defined what science is in a way that satisfies everyone. Science, for example, cannot give absolute proofs of the laws of nature because, although we can test an idea repeatedly, we can never be sure that an exception does not exist. Some religious fundamentalists and TV psychics exploit this difficulty, and claim that science is just another set of beliefs, with no more validity than any other. But while science may not give us absolute truth, this doesn’t mean we must give equal time to magicians and the like. Far from it.

To see why, we need to examine the philosophy of science. Like other branches of philosophy this involves thinking about thinking (the word originally meant “love of wisdom”). The philosophy of science uses similar methods to amathematical proof: a step-by-step examination of assumptions, data and conclusions.

A classic philosophical question is: “Do I exist? How do I know that I am not just a program in some immense supercomputer that is feeding me false sensations about a simulated world?” The French mathematician and philosopher René Descartes (1596-1650) answered this question with a proof involving the famous statement, “I think, therefore I am.” In other words, the act of doubting that we exist proves we exist; there must be something that thinks about the problem of proving existence.

The philosophy of science examines scientific method and asks what it can tell us. Science deals with empirical knowledge. This is knowledge about the Universe that we acquire by examining how it appears to our senses—enhanced, if necessary, by instruments such as microscopes or particle accelerators—rather than by sitting and thinking. Empiricism sounds like common sense, but as a way of learning about the world it is comparatively recent. It triumphed in the scientific revolution of the 16th and 17th centuries, when Galileo Galilei, Robert Boyle, Isaac Newton and others showed that facts gained from empirical observations could revolutionise our picture of the world.

This was where science parted company with magic. Although there was some overlap at the time—Newton was an enthusiastic alchemist, and mystical texts may have inspired him to think of gravity—there is a basic difference between science and magic. Science involves repeatable observations and open publication. There are no hidden or “occult” texts, and when an experiment does not work we do not blame the heavens, the experimenter’s lack of spiritual purity or—a favourite of today’s TV magicians—”bad vibes” from critical observers.

Empiricism creates its own philosophical problems, however. How do facts lead to theories and laws of nature? Imagine an experiment involving observations of apples. After watching apples fall from trees, and verifying that apples will also fall if dropped from the hand, or from the top of a tall building or other tall structures, we reason that a fundamental law is responsible. We call it gravity, and we predict that when we release an apple or any similar object in midair, it will fall to the ground.

When we make a prediction based on past experience, we are moving from statements based on our observations, such as “the apple fell to the ground”, to universal statements such as “all apples in the future will fall to the ground”. This leap from the singular to the universal is calledinductive reasoning.

Inductive reasoning appeals to common sense, but is logically flawed. The empiricist philosopher David Hume (1711-76) pointed out that there can be no logical connection across time. Just because something has happened many times in the past does not prove that it will happen in the future. Karl Popper (1902-94) pointed out that scientific verification doesn’t actually prove anything. No matter how many times we record in our notebooks the fact of observing a white swan, we get no closer to proving the universal statement that all swans are white. Popper decided that science finds facts not by verifying statements but by falsifying them. We may never be able to prove that all swans are white, but the first time we see a black swan we can firmly disprove it.

To reason in this way runs counter to intuition (see Figure 1). Logically, however, it is very powerful, and scientists make good use of this power. Popper said that science progresses by testing hypotheses. One scientist holds up a hypothesis for examination— for example, that gravity bends light waves. Colleagues or rivals then subject this hypothesis to experimental tests that could show it to be false. If the hypothesis survives repeated tests, it becomes accepted as scientific truth.

Figure 1

Popper’s ideas provide a link between theory and experiment. They tell us that no matter how many tests a hypothesis survives, we will never have a philosophical proof that it is true. Popper wrote: “There can be no ultimate statements in science . . . and therefore none which cannot in principle be refuted.” This makes a willingness to accept falsification central to science. Scientists must behave rationally and gracefully, by stating in advance what experimental observations would disprove their hypothesis, and if such findings do emerge, accepting that their hypothesis was wrong.

This was important to Popper, who was born in Austria and whose life was dominated by struggles against ideologies such as those of Nazi Germany, which tolerated no doubts. Popper also contrasted Albert Einstein’s theories of relativity with Karl Marx’s theories of history. While Einstein offered his followers tests, such as solar eclipses, which might have disproved his theories, Marxists were undeterred when history did not unfold according to prediction. Popper also attacked Freudian psychology and Darwinian evolution for what he saw as their unfalsifiability.

Most working scientists today would go along with the idea of falsification. But Popper’s ideas leave us with several problems:

– Falsification alone cannot distinguish science from non-science. The hypothesis that reindeer can fly is falsifiable by any scientist with access to a herd of reindeer, a high cliff and an unusually compliant ethics committee. No one, however, would describe the hypothesis as scientific.

– Where do hypotheses come from? One answer might be that they are merely the application of general principles. For example, they might be inspired by the principle—named after the medieval philosopher William of Occam—known as Occam’s razor: the simplest explanations are the best—or that the Universe everywhere obeys the same laws of physics. But this brings us back to the problem of induction.

– Science doesn’t progress through falsification. In a strictly Popperian system, we would have to abandon the laws of chemistry every time a school student got the wrong result in a chemistry practical. Clearly, we do not do this. We blame the student’s error, or if confronted with a run of anomalous findings, contaminated samples or faulty instruments. Sometimes this is wrong. Scientists rejected early evidence of a hole in the ozone layer over Antarctica because, rather than accepting such unexpected results, they assumed that the satellite collecting the data was faulty. This leads us to the next problem.

– How to explain scientific revolutions, discoveries which transform understanding? Leaps of genius like the theory of evolution by natural selection, or the theory of relativity, appear to be neither new bricks in the wall of knowledge nor the consequence of falsifying previous theories.

WAYS OF SEEING

Paradigm shifts

The last question was tackled by Thomas Kuhn (1922-96). In his book The Structure of Scientific Revolutions, published in 1962, Kuhn said that scientific revolutions need creative thinking of a kind that cannot grow out of the old order. He dismissed Popper’s picture. “No process yet disclosed by the historical study of scientific development at all resembles the methodological stereotype of falsification by direct comparison with nature,” he said.

Kuhn suggested that science does not develop by the orderly accumulation of facts and theories, but by dramatic revolutions which he called paradigm shifts. The worlds before and after a paradigm shift are utterly different—Kuhn’s term was “incommensurable”— and experiments done under the old order may be worthless under the new.

The switch between before and after is as dramatic as that which occurs when looking at a trickgestalt-switch picture (Figure 2). You cannot reject one view without replacing it with the other. Such switches are rare. Kuhn’s examples include the Copernican revolution, which adopted the idea that the Earth orbited the Sun and not the other way round, the discovery of oxygen, and Einstein’s theories of relativity. By contrast, most “normal” research takes place within paradigms. Scientists accumulate data and solve problems in what Kuhn called “mopping-up operations”.

Figure 2

Inevitably, some research throws up findings that do not fit the paradigm—perhaps an unexpected wobble in a planet’s orbit around the Sun. In Popper’s model these would immediately falsifythe paradigm’s central theory. But according to Kuhn, scientists prefer to cling to old paradigms until a new one is ready. The anomaly is either discarded or, preferably, worked into the existing paradigm. In this way the elegant model of an Earth-centred Universe developed in the second century BC by the Greek astronomer Ptolemy accumulated more and more subsidiary orbits to account for astronomers’ subsequent observations.

After a while, however, the anomalies build up into a crisis of confidence, and science stalls. Eventually a genius comes up with a new paradigm. Copernicus realised that the observed orbits of planets made sense when he placed the Sun, rather than the Earth, in the centre of the Solar System. Kuhn said that such leaps happen only in times of crisis.

In times of paradigm shift, hard scientific facts can become meaningless, or change their meaning entirely. For years, scientists made measurements on a substance called phlogiston, which they thought was given off when objects burnt. The discovery of oxygen rendered phlogiston meaningless. But chemists could not discover oxygen until they decided to treat it as a distinct gas. In other words, oxygen had to be invented as well as discovered.

Individual scientists are loath to make such leaps, Kuhn says. The revolution occurs only when practitioners under the old paradigm either die or retire. It takes a new generation to carry the torch of the new paradigm.

Many people have criticised Kuhn. They say his use of the word “paradigm” is imprecise. He chooses his examples overwhelmingly from physics, and they say other sciences may change in different ways. And scientists do not seem as reluctant to make paradigm shifts as Kuhn implies. The discovery of DNA’s double-helix structure utterly changed the way we think about biology, yet biologists accepted it with enthusiasm, replacing a model based on metabolism with one based on information. Did this make it less than a paradigm shift?

Likewise the discovery in the late 1980s of new materials that become superconductors at relatively high temperatures was eagerly pursued by scientists. Such breakthroughs must throw into doubt Kuhn’s distinction between “normal” science—the mopping-up of facts—and revolutionary science.

Finally, Kuhn does not tell us where revolutionary ideas come from. We enjoy the folklore of scientific breakthroughs happening by accident, as with Alexander Fleming and penicillin, or through the work of outsiders such as Einstein. Sadly for Hollywood, such stories are often myths. Although Einstein was working as a patent office clerk when he came up with his theories of relativity, he had steeped himself in contemporary work on physics. Fleming spotted penicillin’s effects because he was an expert in bacteriology, working in a laboratory. In science, chance favours the prepared mind.

Most worrying, if Kuhn is right, science is just a matter of fashions and a kind of crowd psychology, with nothing to distinguish it from pseudoscience. This problem concerned the Hungarian Imre Lakatos (1922-74), who refined some of Popper and Kuhn’s ideas in a way that makes such a demarcation clear. Instead of “normal” and “revolutionary” science, Lakatos drew a distinction between progressive anddegenerative research programmes. A progressive research programme is one that leads to the discovery of facts that were previously unknown. An example is Newton’s theory of gravity, which allowed Halley to predict the return of the comet that now bears his name. A degenerating research programme allows no such predictions; rather, it must itself be modified to cope with inconvenient facts. Lakatos cites Marxism, which although it describes itself as a science has a poor record of predicting a crucial phenomenon—political revolutions.

In progressive research programmes the appearance of awkward facts, such as unaccountable wobbles in a planet’s orbit, is not necessarily fatal to the core hypothesis. Scientists can ignore them if the central hypothesis is still coming up with “unexpected, stunning, predictions”, Lakatos says. Revolutions happen gradually as progressive research programmes replace degenerating ones. But even in progressive research, facts come after theories.

Theories are clearly made up by humans: they aresocially constructed in modern jargon. Does this mean that scientific facts are too? The idea that science is a social construct intrigues many people, especially those thinkers described as “postmodernists”. If, according to Popper, scientific laws are impossible to verify logically and, according to Kuhn, the same findings can mean different things before and after a scientific revolution, how can science claim to be any more objective than any other cultural pursuit?

No one would deny that culture, values and beliefs shape our choice of what science to do. Drugs companies began researching AIDS when it affected people who could afford to buy medicines rather than rural Africans. Military spending on research and development is responsible for similar biases. Scientists believe, however, that the basic facts of the Universe are there to be discovered, whatever the motivation for doing so. We spent billions of pounds developing nuclear weapons, and in the process learned a lot about some strange metal alloys. But we would have found the same facts in a race to build the ultimate ploughshare.

The “science wars” being fought out between academics, especially in North America, question whether this assumption is generally true. Philosophers such as Bruno Latour in Paris study science as a social phenomenon, and suggest its results are little more than social rituals.

Some scientists are horrified by the spectre ofrelativism, which holds that ideas are not universal or absolute but differ from culture to culture, individual to individual. A relativist would assert that science is only one way of discovering the nature of the physical world. The anarchist philosopher Paul Feyerabend (1924-94), perhaps mischievously, took the relativist argument to its logical conclusion: “There is no idea, however ancient and absurd, that is not capable of improving our knowledge.” In Against Method(1975) he defended the Church’s indictment of Galileo. It was rational, he said, because there was at the time no reason to suppose that Galileo’s crude telescopes could show the mountains on the Moon that he claimed to have seen. The Church believed that the Moon was a perfect smooth sphere quite unlike Earth.

TRUST AND TRUTH

Science and non-science

One vigorous defender of science’s special place is the biologist and author Richard Dawkins. He notes that when relativist philosophers fly to an international conference on postmodernism, they generally put their trust in a high-technology airliner rather than a magic carpet. And, of course, absolute relativism contains its own contradiction. “Those who tell you there is no absolute truth are asking you not to believe them,” says the contemporary philosopher Roger Scruton. “So don’t.”

One battle in the “science wars” is over Darwin’s theory of evolution (see “Evolution under attack”). Some assaults on evolution come from a particularly stormy debate over evolutionary psychology or, as it is sometimes called,sociobiology. This attempts to explain people’s patterns of behaviour—whether it be fear of snakes, or why we enjoy particular kinds of landscape gardening—solely in terms of evolutionary advantage.

Evolutionary psychology is controversial because it can be used to justify types of behaviour, such as violence, which are generally considered unacceptable. It is possible to challenge the science of evolutionary psychology without challenging evolution itself.

The phenomenon of consciousness is another problem area. Philosophers and scientists both stake a claim to holding the key to understanding consciousness. The fact that some computer scientists believe they can create artificial consciousness gives the debate extra spice. But any such project will first have to define what constitutes consciousness, which is probably a job for philosophy. Science and technology can then take over.

Finally, there is the question of exactly what science is. As we have seen, Popper’s falsification criterion alone is not enough to distinguish science from non-science. In fact, if we look at the whole array of science, from particle physics to cell biology to ecology to engineering, it is hard to find any single practice that they all have in common. Even openness is not always there: much research is kept secret for military or commercial reasons.

A way out is to use a concept developed by one of the most important 20th-century philosophers, Ludwig Wittgenstein (1889-1951), that of family resemblance. There are many groups of human activities that are impossible to define exactly. For example, it’s hard to say what a game is, but when we see a new game we have no trouble deciding that that’s what it is, because of the things it shares with other members of the games family. Likewise with science: all we can say aboutgood science is it has most of the qualities of other activities we call good science, including empiricism, peer review and openness to refutation.

Those who work in this family believe that truth is out there. Perhaps not always in the strictest philosophical sense, but enough for practical purposes and definitely enough to distinguish science from propaganda and muddled thinking. Scientists do not need to be shy of admitting that its laws are always provisional. That is not a weakness, but science’s greatest strength.

Evolution under attack

THE philosophy of science figures in heated modern debates, such as the relationship of science and religion. The idea that science must conflict with Christian religion is recent. Newton was a devout if unorthodox Christian who saw science as revealing the wonders of creation, not challenging them. Indeed what is known as theargument from design—that the world is too intricately created to have arisen by chance— was cited as a scientific proof that God existed. Darwin’s work smashed that consensus. Evolution by natural selection demonstrated that we do not need a divine creator to explain where human beings came from.

Today Darwinism itself is under fire, although the attacks are rarely overtly religious. The target is generally the philosophy of Darwinian evolution. Some critics cite Popper, who said that the theory of evolution by natural selection is unscientific because it is unfalsifiable. In one sense, this is correct: we cannot rerun the tape of the past 5 billion years. But thousands of biologists every day test evolution’s crucial components and processes. Fame and a large fortune in the biotechnology business await any scientist who can find a short cut to the slow grind of Darwinian evolution. No one has.

Of course, Darwinism is “only a theory”, but this does not mean that all other theories deserve equal respect.