It seems some disagreements are interminable: the Anabaptists versus the Calvinists, capitalism versus communism, the Hatfields versus the McCoys, or string theorists versus their detractors. It is the latter I will discuss here although the former may be more interesting. This essay is motivated by a comment in the December 16, 2014 issue of Nature by George Ellis and Joe Silk. The comment takes issue with attempts by some string theorists and cosmologists to redefine the scientific method by eliminating the need for experimental testing and relying on elegance or similar criteria instead. I have a lot of sympathy with Ellis and Silk’s point of view but believe that it is up to scientists to define what science is and that hoping for deliverance by outside people, like philosophers, is doomed to failure
To understand what science is and what science is not, we need a well-defined model for how science behaves. Providing that well-defined model is the motivation behind each of my essays. The scientific method is quite simple: build models of how the universe works based on observation and simplicity. Then test them by comparing their predictions against new observation. Simplicity is needed since observations underdetermine the models (see for example: Willard Quine’s (1908 –2000) essay: The Two Dogmas of Empiricism). Note also that what we do is build models: the standard model of particle physics, the nuclear shell model, string theory, etc. Quine refers to the internals of the models as myths and fictions. Henri Poincaré (1854 – 1912) talks of conventions and Hans Vaihinger (1852 –1933) of the philosophy of as if otherwise known as fictionalism. Thus it is important to remember that our models, even the so-called theory of everything, are only models and not reality.
It is the feedback loop of observation, model building and testing against new observation that define science and give it its successes. Let me repeat: The feedback loop is essential. To see why, consider example of astrology and why scientists reject it. Its practitioners consider it to be the very essence of elegance. Astrology uses careful measurements of current planetary locations and mathematics to predict their future locations, but it is based on an epistemology that places more reliance on the eloquence of ancient wisdom than on observation. Hence there is no attempt to test astrological predictions against observations. That would go against their fundamental principles of eloquence and the superiority of received knowledge to observation. Just as well, since astrological predictions routinely fail. Astrology’s failures provide a warning to those who wish to replace prediction and simplicity with other criteria. The testing of predictions against observation and simplicity are hard taskmasters and it would be nice to escape their tyranny but that path is fraught with danger, as astrology illustrates. The feedback loop from science has even been picked up by the business management community and has been built into the very structure of the management standards (see ISO Annex SL for example). It would be shame if management became more scientific than physics.
But back to string theory. Gravity has always been a tough nut to crack. Isaac Newton (1643 – 1727) proposed the decidedly inelegant idea of instantaneous action at a distance and it served well until 1905 and the development of special theory of relativity. Newton’s theory of gravity and special relativity are inconsistent since the latter rules out instantaneous action at a distance. In 1916, Albert Einstein (1879 – 1955) with an honorable mention to David Hilbert (1862 – 1943) proposed the general theory of relativity to solve the problem. In 1919, the prediction of the general theory of relativity for the bending of light by the sun was confirmed by an observation by Arthur Eddington (1882 – 1944). Notice the progression: conflict between two models, proposed solution, confirmed prediction, and then acceptance.
Like special relativity and Newtonian gravity, general relativity and quantum mechanics are incompatible with one another. This has led to attempts to generate a combined theory. Currently string theory is the most popular candidate, but it seems to be stuck at the stage general relativity was in 1917 or maybe even 1915: a complicated (some would say elegant, others messy) mathematical theory but unconfirmed by experiment. Although progress is definitely being made, string theory may stay where it is for a long time. The problem is that the natural scale of quantum gravity is the Planck mass and this scale is beyond what we can explore directly by experiment. However, there is one place that quantum gravity may have left observable traces and that is in its role in the early Universe. There are experimental hints that may indicate a signature in the cosmic microwave background radiation but we must await further experimental results. In the meantime, we must accept that current theories of quantum gravity are doubly uncertain. Uncertain, in the first instance, because, like all scientific models, they may be rendered obsolete by new a understanding and uncertain, in the second instance, because they have not been experimentally verified through testable predictions.
Let’s now turn to the question of multiverses. This is an even worse dog’s breakfast than quantum gravity. The underlying problem is the fine tuning of the fundamental constants needed in order for life as we know it to exist. What is needed for life, as we do not know it, to exist is unknown. There are two popular ideas for why the Universe is fined tuned. One is that the constants were fine-tuned by an intelligent designer to allow for life as we know it. This explanation has the problem that by itself it can explain anything but predict nothing. An alternate is that there are many possible universes, all existing, and we are simply in the one where we can exist. This explanation has the problem that by itself it can explain anything but predict nothing. It is ironic that to avoid an intelligent designer, a solution based on an equally dubious just so story is proposed. Since we are into just so stories, perhaps we can compromise by having the intelligent designer choosing one of the multiverses as the one true Universe. This leaves the question of who the one true intelligent designer is. As an old farm boy, I find the idea that Audhumbla, the cow of the Norse creation myth, is the intelligent designer to be the most elegant. Besides the idea of elegance, as a deciding criterion in science, has a certain bovine aspect to it. Who decides what constitutes elegance? Everyone thinks their own creation is the most elegant. This is only possible in Lake Wobegon, where all the women are strong, all the men are good-looking, and all the children are above average (A PRAIRIE HOME COMPANION – Garrison Keillor (b. 1942)). Not being in Lake Wobegon, we need objective criteria for what constitutes elegance. Good luck with that one.
Some may think the discussion in the last paragraph is frivolous, and quite by design it is. This is to illustrate the point that once we allow the quest for knowledge to escape from the rigors of the scientific method’s feedback loop all bets are off and we have no objective reason to rule out astrology or even the very elegant Audhumbla. However, the idea of an intelligent designer or multiverses can still be saved if they are an essential part of a model with a track record of successful predictions. For example, if that animal I see in my lane is Fenrir, the great gray wolf, and not just a passing coyote, then the odds swing in favor of Audhumbla as the intelligent designer and Ragnarok is not far off. More likely, evidence will eventually be found in the cosmic microwave background or elsewhere for some variant of quantum gravity. Until then, patience (on both sides) is a virtue.
Though the mills of science grind slowly;
Yet they grind exceeding small;
Though with patience they stand waiting,
With exactness grind they all.
 I have already broken my new year’s resolution to post no more philosophy of science blogs but this is the last, I promise.
 With apologies to Henry Wadsworth Longfellow (1807 – 1882)