Read AC Grayling unpacking the philosophy behind experimentation
If a scientific theory is elegant, and is consistent with known facts, does it need to be tested by experiment? Scientific knowledge is supposed to be empirical: to be accepted as scientific, a theory must be falsifiable—that is, it must be possible, at least in principle, to empirically disprove it. This argument was advanced in 1934 by Karl Popper, the philosopher, and is generally accepted by most scientists today as determining what is and is not a scientific theory.
In recent years, however, many physicists have developed theories of great mathematical elegance, but which are beyond the reach of empirical falsification, even in principle. The uncomfortable question that arises is whether they can still be regarded as science. Some scientists are proposing that the definition of what is “scientific” be loosened, while others fear that to do so could open the door for pseudo-scientists or charlatans to mislead the public and claim equal space for their views.
The question of whether highly theoretical scientific ideas can be subjected to experimental testing is an issue for the most advanced and powerful ideas in the world of physics. String theory and the idea of the “multiverse”—the existence of multiple universes—are two leading theories that attempt to explain the most fundamental characteristics of the physical world. Both ideas have immense theoretical appeal. String theory is not intrinsically untestable—but there has been no success yet. In experimental terms, one can imagine some future technology that is—in theory at least—capable of accelerating particles to what is known as the Planck energy scale. This is an energy level a thousand trillion times greater than what can be produced at the Large Hadron Collider (LHC) and the point at which the implications of string theory are predicted to be manifest. Multiverse theory presents apparently insuperable obstacles to experiment as other universes are intrinsically impossible to detect, although even here, physicists are suggesting ways to infer their existence.
Is physics moving towards an era in which elegance will suffice and into the domain of theories that are beyond the reach of experimental proof? Or will empirical evidence remain the arbiter of science?
String theory is an attempt to develop a unified theory of particles and forces, and it first burst onto the scene 30 years ago. The theory posits that miniscule one-dimensional entities—strings—exist in dimensions higher than those currently known to us, and that these strange high-dimensional phenomena underlie all of physics. Since its development, the techniques of string theory have been widely and successfully used by mathematicians. But the original motivation—to create a scientific theory that unifies the laws governing the behaviour of particles and forces—has stalled. According to the cosmologist George Ellis, a former professor at the University of Cape Town and a world authority on the physics of the cosmos, string theory is “an exploration of fascinating mathematical structures that may or may not relate to the physical universe. So in terms of its applicability to the real universe, it’s hypothetical science rather than testable science.”
Frank Wilczek, a professor of physics at the Massachusetts Institute of Technology and winner of the 2004 Nobel Prize in Physics, describes the present situation thus: “The string theory community contains many serious and gifted individuals who are trying to understand nature, and it would be crazy to rule them out of science. But to me, the parts of science that use a few assumptions to explain a lot about the world are the most impressive and important, and from that perspective string theory could use improvement.”
The challenge made by Wilczek and Ellis is whether evidence to support string theory could ever be found through experiment. One line of investigation concerns a central plank of the theory. This holds that each particle of matter, such as electrons or quarks, has a partner among the particles that transmit forces—“bosons” such as the photon and gluons. This property is known as supersymmetry. Wilczek remarks that it is “an important ingredient in string theory. So discovery of it, though not evidence [for string theory], would be encouragement.” However, evidence of supersymmetry has not yet been detected at the LHC, the particle accelerator on the French-Swiss border, which is the world’s highest-energy facility and so the best equipped to test this. In July 2012, the accelerator had one much-publicised triumph when it confirmed the existence of the Higgs boson, the particle that gives some fundamental particles their mass.
The discovery was of such significance because it completed the “standard model” of particles and forces: the core theory that physicists have developed about the fundamental building blocks of nature. Yet there is currently no empirical evidence to support any physics—such as supersymmetry—that lies outside the standard model. To follow Popper’s guidance on what is scientific we would have to say that string theory currently stands outside science.
However, physicists are optimistic that a breakthrough might soon be made in the search for a particular type of matter known as “dark particles.” Supersymmetry predicts the existence of particles whose properties might be consistent with those of dark matter. So if scientists could find evidence of dark matter, this would support the theory of supersymmetry and would count as a first step in providing an empirical basis for string theory.
Physicists have long observed that the motions of stars and the interactions of galaxies suggest that they feel more gravitational force than can be explained by visible stars. This missing gravitational pull is thought to be exerted by dark matter.
Wilczek is optimistic that the LHC might provide a breakthrough. His hopes are echoed by Rolf-Dieter Heuer, the Director-General of Cern, which operates the LHC. In his opinion, the higher energy of the refurbished LHC will “open a window to direct discoveries beyond the standard model.” In the opinion of Steven Weinberg, the Nobel laureate whose work has been central to the development of the standard model, the discovery of the particles of dark matter would be “the most exciting of all.”
In theory, then, evidence of the validity of supersymmetry could be found by science. The same goes for dark matter. Neither would confirm string theory, but they would be a first step. The idea is, then, in principle, open to empirical testing.
Multiverse theory, however, is more problematic. As there is no possibility of communication between us and other universes, there is no empirical way to test the multiverse theory. George Ellis makes the point explicitly: “In a general multiverse model, everything that can happen will happen somewhere, so any data whatever can be accommodated. Hence it cannot be disproved by any observational test at all.” By implication, the multiverse concept lies outside science.
“Mathematical tools enable us to investigate reality, but the mathematical concepts themselves do not necessarily imply physical reality”
For as long as humans have pursued science, they have tried to understand the universe. Wilczek says: “Modern physics implies that it is plausible that the physical world can exist in qualitatively different forms, similar in spirit to how water can exist as ice, liquid water or steam. These different forms… can in effect implement different laws of physics. If such diverse regions of space exist, then the ‘universe’ as we’ve defined it is not the whole of reality. We call the whole of reality the multiverse.”
Ellis and his cosmologist colleague Joe Silk, a professor at the Université de Pierre et Marie Curie in Paris, call this “a kaleidoscopic multiverse comprising a myriad of universes.” They, as proxy for many physicists, then pose the basic challenge: the suggestion that another universe need not have the same fundamental constants of nature as ours inspires the question of what determines the values in our universe. Of the variety of universes that could exist, the conditions for the narrow range of parameters for which intelligent life could exist are trifling. The odds that we exist are therefore so vanishingly small, that multiverse theory claims that there is a “landscape” of universes “out there” in which all possible values of these parameters exist. Thus one universe will exist somewhere with conditions just right for life, and we are the proof.
Weinberg accepts that the multiverse is unlikely to be confirmed by observations in our specific “sub-universe.” But he argues that this is not necessarily fatal to the theory’s scientific validity. “The multiverse idea is very speculative,” he says, “but it’s not an entirely unreasonable speculation. The existence of a multiverse might some day be confirmed by deducing it from a theory that is confirmed by the success of sufficient other predictions.”
In this vein, Wilczek points out that scientific theories can still be of use even when they are only partially understood. He says: “It is very common and successful practice to work with theoretical structures much vaster than what we can observe about them.” One example he cites is quantum theory, a basic tool in theoretical physics, which is full of concepts that appear to contradict our intuitive notion of how things behave. Many theorists, myself included, are uncomfortable with its foundations, yet manage to apply its mathematics with confidence and empirical success. The theory of quantum mechanics is science because it can in principle be disproved. It has survived innumerable tests, and made countless successful predictions. Ellis and Silk remind us that the multiverse may be a convenient mathematical device, but this does not require these universes to have “reality.” They drive this home by recalling the warning of the German mathematician David Hilbert: “Although infinity is needed to complete mathematics,” he said, “it occurs nowhere in the physical universe.”
This is the crux. Mathematical tools enable us to investigate reality, but the mathematical concepts themselves do not necessarily imply physical reality. Thus evidence in support of a theory has to be experimental or observational, not simply theoretical. Ellis and Silk make this point powerfully, and warn against the notion that “theoretical discoveries [can] bolster belief.” They remind us: “experiments have proved many beautiful and simple theories wrong.”
Wilczek gives an example of one such theory in his book A Beautiful Question. In the 17th century, the German astronomer Johannes Kepler became convinced that he had developed a model of the structure of the solar system. His “theory” had a seductive, geometrical beauty that convinced Kepler that he had stumbled on God’s plan. He wrote: “I feel carried away and possessed by an unutterable rapture over the divine spectacle of heavenly harmony.” But his theory was false—Kepler’s planetary model was eventually undermined, not least by the discovery of further planets. Yet as Wilczek reminds us, although Kepler was wrong in his description of the arrangement of the planets, he was accurate in his description of their motion—that planetary orbits are not circles but ellipses, and that the sun is not at the centre of the ellipse, but located at a “focus” of the ellipse. These insights inspired Isaac Newton to develop his law of gravity.
We might hope for a modern parallel: that the raptures over string theory inspire experimentalists at the LHC to the discovery of supersymmetry. This in turn may resolve the mystery of dark matter, whose existence was suspected by the apparent refusal of the motions of galaxies of stars to obey the rules of Kepler and Newton. Or, perhaps, supersymmetry and dark particles will refuse to appear at the LHC, because they do not exist. Being ruled out by experiment would be a setback, but it would be a scientific setback. In the great human project to understand better the physical laws that govern the behaviour of matter and forces, it would count as progress.