Science not only helps us to understand the natural world around us, but correct predictions can be usefully applied to harness nature and adapt to the environment. Whether it involves particle physics, evolution, celestial mechanics or general relativity, good scientific theories have what is known as the power of prediction.
Two weeks ago, I wrote about the predictions of doom that attended the population explosion panic of the 1960s, which reverberate even today. Those predictions were alarmingly specific, but have been incontrovertibly falsified when the future arrived and turned out to be better – not worse – than the past. Although many of those predictions were made by scientists, what they told the world was not science.
Much of the power of the scientific method lies not merely in its ability to describe nature, or even to explain it. The most valuable science, and its most rigorous test, lies in its predictive power.
One might think of the scientific method as having four steps. The first is to accurately observe or describe a phenomenon or group of phenomena. Second, one forms a hypothesis – not to be confused with the term “theory” – which aims to explain these phenomena. Third, the hypothesis is used to make predictions, either of quantitative values or the existence of new phenomena. Fourth, experiments are devised to test such predictions, and these experiments must be replicable by independent researchers.
Once a hypothesis or group of hypotheses have been confirmed by repeated experiments to be sound, they may attain the status of a theory. Not all theories are uncontested, and not all theories stand the test of time, but in general, scientific theories form the basis of our present understanding of the world. Until results are discovered which do not fit the existing theoretical framework, a theory will stand. When such a result is discovered and verified, an inconsistent theory must be modified, or even entirely abandoned.
You may have noticed, from this procedure, that a theory can never be proven. It can only be disproven. Each prediction and experiment which confirms a theory strengthens it, but it takes only one observation that contradicts the theory to collapse the entire house of cards. Granted, it will take a very well-validated observation or experiment to bring a well-established theory down, but the basic principle remains: science is not the process of prediction. Science is the process of trying to disprove predictions. This principle is so important that philosophers of science consider “falsifiability” a prerequisite for any statement to be considered scientific at all.
The recent discovery of the Higgs boson makes an excellent example. Existing theory, known as the standard model of particle physics, predicted that such a thing must exist. Postulated by Peter Higgs in 1964, this was the last great unresolved question in the modern era of quantum physics, and would explain a number of oustanding mysteries, such as why some particles have mass. The trick was to devise an experiment that would demonstrate that this particle does, in fact, exist. The Large Hadron Collider at CERN in Europe was that experiment, and on 4 July 2012, the discovery of a new particle that fit the description was announced.
Ironically, although the Higgs boson was the greatest discovery in modern physics, scientists would have been far more excited if their 48-year search had instead shown that the Higgs boson did not exist. As spectacular as this confirmation was, merely confirming an age-old theory by finding something that fit perfectly with predictions is just the ordinary course of science. The theory-shattering discovery that the predictions of the standard model were wrong would have been far bigger news to scientists.
There are other classic examples of the predictive power of scientific theories. One is the case of an orchid, Angraecum sesquipedale, endemic to Madagascar. When a specimen was sent to Charles Darwin, he noted its exceptionally long spur, at the bottom of which its nectar could be found. “I have just received such a Box full from Mr Bateman with the astounding Angræcum sesquipedalia with a nectary a foot long,” he wrote. “Good Heavens what insect can suck it?”
The theory of evolution predicted that a moth with a foot-long proboscis must exist. In an 1862 publication on orchids, he developed the hypothesis that the orchids and moths co-evolved in a sort of “arms race”, as orchids with shorter spurs would give up their nectar without achieving pollination, and moths with shorter probosces would fail to reach the ever-lengthening spur. Alfred Russel Wallace supported Darwin’s hypothesis, writing: “That such a moth exists in Madagascar may be safely predicted; and naturalists who visit that island should search for it with as much confidence as astronomers searched for the planet Neptune – and they will be equally successful!”
Several anti-evolutionists disparaged these predictions, claiming that the species had supernatural origins. In 1903, 41 years after the prediction, and 21 years after Darwin’s death, a moth known as Morgan’s sphinx moth was discovered, which fit the bill. It was named Xanthopan morganii praedicta, in honour of the prediction.
Since then, the theory of evolution by natural selection has produced a number of specific predictions which validate its status as the over-arching theoretical framework for all of modern biology. Darwin himself, in On the Origin of Species, noted a number of observations that his theory could not explain. These objections, he wrote, “might be justly urged against the views maintained in this volume”.
Each of these can be restated as a prediction. For example, one of these questions was the notable absence of intermediate or transitional forms in the fossil record, of which Darwin argued the number must be “inconceivably great”, and for which he blamed the “extreme imperfection of the geological record”. While it seems a somewhat facile excuse for his inability to explain the problem, discovery of such fossils would verify Darwin’s theory. And, as it turns out, his surmise was correct.
Many transitional fossils have since been found. Perhaps the most famous is Archaeopteryx, which is an intermediate between reptiles and modern birds. There are only seven known fossil specimens of these toothed, clawed and winged animals. One of them, which had poorly preserved feathers, was initially described as a small bipedal dinosaur. It also has a strong resemblance to its dinosaur ancestors, including the famous velociraptors, of Hollywood fame.
Cetaceans, including whales, are part of the ungulate mammal group, which means they must at one point have had legs with hooves. Whether or not whales are in fact even-toed ungulates, like hippos and pigs, known as artiodactyls, was the cause of a controversy known as the “whippo” debate. It was only recently settled, by genetic evidence and the discovery of new fossils, which show that “the earliest cetaceans were artiodactyl-like terrestrial running animals with the long, slim limb bones and pulley-like ankle bones of artiodactyls”. Ancestral relatives of whales, such as Dorudon and Basilosaurus, have both tail fins and vestigial hind legs. The development of whales from land mammals also implies that its oldest ancestors must have been freshwater species, before they developed and moved to a marine environment. Modern animals, such as marine and river dolphins, can be told apart from the oxygen isotopes bound inside the animal’s bones. Tests on fossils of ancient cetaceans, in particular Pakicetus and Ambulocetus, show that in addition to having had hind legs, they also lived in freshwater environments. Finally, since modern whales have either baleen for filter-feeding, or teeth for preying, evolution also predicts the existence of ancestral relatives with both. Such an ancestor has also been found.
In this manner, all species that exist on earth today can be related, according to Darwin’s theory. And indeed, they can all be classified into a single, hierarchical “tree of life”.
Darwin also predicted that despite the absence at the time of pre-Cambrian fossils, life must have been plentiful before the Cambrian era, since the many diverse Cambrian species could not all have appeared in a very short time. This problem, he wrote, “may be truly urged as a valid argument against the views here entertained”. Time has proved Darwin to have been correct. In 1953, Stanley A. Tyler discovered the first pre-Cambrian microbial fossils. It turns out that although small, pre-Cambrian life was abundant.
The predictive power of the theory of evolution is so good that the existence of transitional forms can be predicted, and pinpointed to rock layers of a particular age. A great example is the recent discovery of Tiktaalik roseae, a 375-million-year-old intermediate between fish and amphibians, believed to be an ancestral relative of all four-legged terrestrial animals, including reptiles, birds and mammals. Scientists went to look for this fossil, knowing exactly in which rock layer they would find it, and did. “It’s one of those things you can point to and say, ‘I told you this would exist, and there it is’,” said Cambridge University scientist Jennifer A. Clack, as quoted in New Scientist magazine (paywalled).
Evolution does not appear to be very good at predicting the future. Random mutations are exactly that: random. Yet this perception turns out not to be true, either. Despite the obstacle of randomness, it turns out evolutionary biology has many applications. According to a journal paper from 2001: “Darwin’s ‘evolution by natural selection’ is being used in many contexts, from the design of biotechnology protocols to create new drugs and industrial enzymes, to the avoidance of resistant pests and microbes…” It has been used in genetic engineering to predict and prevent the evolution of poison- or drug-resistant species, for example.
As Wallace’s comment about the orchid and its co-evolutionary moth suggests, the idea of the predictive power of science was not new, nor was it limited to the theory of evolution. It had been applied successfully to deduce the existence and position of a major unknown planet from the observed orbital perturbations of Uranus.
Pierre Simon de Laplace, a renowned French mathematician, published a major Treatise of Celestial Mechanics in 1822, in which he proposed equations for using Newtonian physics to calculate how one planet’s orbit was affected by the gravitational pull of others.
Uranus had been rediscovered in 1781 by William Herschel, and it fell to Alexis Bouvard, a student of Laplace, to use these equations to calculate tables that would predict the future movement of the giant planets Jupiter, Saturn and Uranus. Unlike with Saturn and Jupiter, however, he could not reconcile the planet’s modern movements with older observations of what must have been Uranus. The difference might have been explained by observational error, but it certainly was not explained by perturbations resulting from the orbits of other planets.
As new observations of Uranus accumulated, it became clear that the perturbations were not simple matters of observational error, either. By 1845, Bouvard expressed what had by then become a common notion, that it was “entirely plausible … that another planet was perturbing Uranus”.
Independently, Urbain Le Verrier in France and John Couch Adams in England inverted the problem. They began calculations to derive the orbit of a postulated new planet, entirely from the perturbations observed in the orbit of Uranus. If found, such a planet would prove the predictive power of Laplace’s equations. If not, Laplace, and perhaps even Newton, would have to be corrected.
These calculations led directly to the discovery of Neptune by Johann Gottfried Galle in 1846, almost exactly between the points where Verrier and Adams concluded it must be. Francois Arago wrote that Verrier had discovered Neptune “at the tip of his pen”, but to whom the credit really belongs remains a matter of dispute. It does not matter, however. The existence of Neptune had been entirely predicted by mathematics. This was an outstanding vindication of the predictive power of 19th century science, on a par with the discovery of the Higgs boson in the 21st.
An even more famous prediction, this time in the 20th century, was that of Albert Einstein. Based on his 1915 theory of general relativity – which describes the geometry of gravity in space-time – he predicted that a mass in space would bend the path of light. As predictions go, this was unintuitive and exceptional. If confirmed, it would lend incredible weight to Einstein’s theory.
Arthur Eddington, an English astrophysicist, travelled to Sobral in Brazil, and the island of Principe in the Atlantic Ocean – two points near the equator, but separated by more than 45° in longitude – to make observations of stars near the solar disc during the eclipse of 26 May 1919. His purpose was to determine whether starlight, when it passes close to the sun, remains unaffected by gravitation, is affected according to Newton’s law of gravity, or is affected as much as Einstein had predicted. In 1920, he published his account of the expedition in the Philosophical Transactions of the Royal Society. His observations were a resounding confirmation of Einstein’s theoretical predictions.
Since so much of this column was devoted to Charles Darwin’s implicit predictions, it may be appropriate to close with another of his statements. In On the Origin of Species, long before the celebrated philosopher of science Karl Popper popularised the notion of “Science as Falsification”, Darwin wrote: “If it could be proved that any part of the structure of any one species had been formed for the exclusive good of another species, it would annihilate my theory, for such could not have been produced through natural selection.”
In the 157 years since he wrote that sentence, nobody has demonstrated a single case that would disprove his theory. That is the power of scientific predictions. DM