The Queen's New Clothes

In 1781, Immanuel Kant announced that metaphysics, once queen of the sciences, had been dethroned.

Natural science was the usurper. This is the Enlightenment's own Creation story: metaphysics was banished for having unlawfully tasted the fruit of the tree of knowledge, making room for scientific knowledge. We have not shaken this myth.

The idea that science and metaphysics are totally distinct -- what Stephen Jay Gould called non-overlapping magisteria -- is comforting. Scientistic thinkers, such as the New Atheists, take comfort in the notion that their knowledge is purely empirical (and not metaphysical). The religious, meanwhile, take comfort in the notion that religious truths are immune to empirical disconfirmation -- precisely because they are metaphysical and not scientific.

But the truth is messier.

The fact is the queen was never dethroned, but merely disguised. Metaphysics, as Kant knew, could never be eradicated, only redirected. Like the monarch who welcomes constitutional reform in lieu of revolution, the queen of the sciences subjected herself to a new law: the mercurial tribunal of experience. Experimental science became the new face of metaphysics -- and this meant that metaphysics became something quite different.

Once upon a time we asked philosophers for answers to certain questions: "what is time?" "is the universe eternal?" At that point, science and philosophy were not distinct disciplines; physics was philosophiae naturalis, subordinated to metaphysics or theology -- what Aristotle called "first philosophy." 'Scientia,' after all, simply means 'knowledge,' knowledge of the surest sort, and it was through philosophy that one attained it. Philosophers had plenty of answers: "time is the moving image of eternity" (Plato) "time is the measure of motion" (Aristotle), "time does not exist" (Chrysippus). Philosophers still have answers, about which they still disagree, including whether philosophy can or even ought to answer such questions.

So we stopped asking philosophers when scientists promised something more. Scientific explanations -- "time is the fourth dimension of a semi-Riemannian manifold," "the universe began rapidly expanding from a state of high density approximately 13.8 billion years ago" -- have many virtues that philosophical answers do not. They exhibit precision (thanks to their mathematical nature) -- in particular, they are precise even about uncertainty -- fecundity (thanks to their predictive power and technological implications), and partial consensus (thanks to agreed upon experimental procedures).

But that's not all: for better or for worse, such explanations now provide the backbone for our modern metaphysical image of the cosmos.

How can this be? We learn in elementary school that scientific theories are formulated according to the "scientific method:" a presuppositionless inquirer formulates a hypothesis based upon pure observations and goes on to verify it via experiment. Hypotheses are just a necessary evil in a procedure that, unlike the "soft" sciences or the humanities, eschews extra-empirical reasoning in favor of strictly "verifiable" claims. But this, like many things we are told in elementary school, is a fairy tale.

Nevertheless, the fairy tale is persistent. When college-level physics students learn the classical laws of motion, they are often told that Aristotle's theory of motion was blatantly wrong because it was based on pure reasoning rather than observation. The implication is that "our" theory of motion is correct because it is based on observations. (Let's leave aside the fact that, in this case, "our" theory means the Newtonian theory, which is, nowadays, understood to be an approximation at best).

Of course, in saying that Aristotle didn't make observations, the teacher commits an historical error. Aristotle took all knowledge to be rooted in sensation and gathered an enormous range of empirical data himself. But, more importantly, the teacher commits a philosophical error. As Thomas Kuhn pointed out, mere observation confirms Aristotle's theory of motion to a great extent. In ordinary circumstances, heavy bodies do fall faster than light ones. Aristotle was not wrong because he failed to make observations; he was so because he only made observations.

Galileo's great innovation, in formulating the law of inertia that overthrew the Aristotelian theory of motion, was not that he observed more carefully than his Greek predecessor; rather he grasped, as Kuhn put it, a "more essential, but hidden, aspect of motion," one actually masked by ordinary observations. We do not ordinarily observe bodies moving at constant velocity and can never, strictly speaking, observe a body moving with constant velocity for all time (we don't have "all time"). As the historian Alexandre Koyré argued, Galileo's supposition that physical bodies obey the law of inertia was rooted not in observation so much as a metaphysical conviction that nature is mathematical -- that the path of a physical object could be described precisely using mathematical language.

To be sure, nobody would have taken Galileo's theory of inertia seriously if it had only been a Platonic idea. To be scientific it had also to be testable. Modern science "puts nature on the rack," as Sir Francis Bacon put it. But the conviction that led Galileo to utter "epper si muove" was of an altogether higher -- metaphysical -- order. Modern science is a metaphysics rooted in experiment.

The problem with trying to eliminate metaphysics from science is that, as the French philosopher Émile Meyerson argued a century ago, history never exhibits such a tidy distinction between "metaphysics" and "science."

Newton invoked divine intervention to explain his concept of an invisible force acting instantaneously across great distances, considered space to be an attribute of God, and dedicated much of his life to alchemy. It is often said that Kepler merely inferred his famous laws -- one of which states that planetary bodies follow elliptical trajectories -- from empirical data gathered by his Danish predecessor, Tycho Brahe. This is not quite true; Kepler was a Platonist at heart. His research was conducted on the assumption that Plato's "solids," ideal mathematical objects, helped explain the movement of celestial bodies. He even sought a mathematical account of the music of the ether -- the perfect harmonies produced by the movement of the stars through the celestial ether.

One could always conclude that the scientific brilliance of these men somehow pierced through the mystification and superstition of their unscientific Christianity and Platonism. On the contrary, the lesson of history is that, from Aristotle to Einstein, scientific theories are formulated within a metaphysics and for the purposes of answering metaphysical questions. As Meyerson said, and American philosopher W.V.O. Quine later echoed, "ontology is integral with science itself, and cannot be separated from it."

For agreeing to do the work of metaphysics, however, science demands something in return.

First, science demands that we give up the hope of definitive answers to its questions. Newton argued that light was made up of particles; Fresnel showed that it was made up of waves; quantum physics tells us that it is both. It is possible that this latter theory will never be overthrown. But that would be a first.

Scientific answers are, by definition, never definitive; they only hold until the next theory. This is the mercurial tribunal of experience: no matter how intellectually satisfying, a scientific theory can never become dogma. Science, though metaphysical, does not furnish scientia -- knowledge that holds come what may. Scientific theories provide images of the world that, as Meyerson observed, continually change.

Second, science has ground rules: in order to answer metaphysical questions precisely, science first specifies the conditions under which they can be asked. For example, the law governing electrical interactions, Coulomb's Law, describes the behavior of mathematical abstractions that scientists call "point charges." Similarly, Newton's universal law of gravitation describes the interaction between "point masses," which are immeasurably small objects. Point charges and point masses are idealizations; they do not exist in nature.

In other words, scientific laws do not describe ordinary objects, but symbolic objects that stand in for them. It's not that scientific laws are false, or artificial, or even that they "lie," as the philosopher Nancy Cartwright provocatively puts it. Rather, science provides conceptual models that we use to interpret reality; science paints pictures. If the theory is good, the picture is said to be "true," but only under specific, well-defined and highly abnormal conditions -- and never once and for all.

This is analogous perhaps to the way in which we claim that a beautiful painting, say one of Monet's Water Lilies, is true. We do not mean that the painting is identical with reality or that nature is reducible to it. We mean only that the painting successfully captures reality.

Similarly, we understand that there can be other aesthetic images that successfully capture that same reality. Monet himself is famous for his series paintings, in which multiple paintings represent a scene under various light conditions, times of day, and so on. If anything, these varying images make his artworks more true.

So too with science. When we ask it to explain reality, it responds with a picture - and another, and another. It thereby captures reality in a way that traditional metaphysics never could. In other words, science tells us how to think about reality, and it does so with incomparable precision.

But the scientific image of the world is always only an image. Reality is no more reducible to it than to Monet's impressionism. Nature, you might say, stands behind the pictures science paints of it. Thus, the metaphysical image of the cosmos that science furnishes may be true, but never complete.

There are, and always will be, more things in heaven and earth, than are dreamt of in its philosophy.