Why facts aren't enough

Pierre Maurice Marie Duhem 1861-1916

Studies abound.  Political and scientific discourses trade in statistics.  In public debates, some question scientific theories, such as Darwinian evolution or anthropogenic climate change, as insufficiently factual while others hold them up as indefeasibly factual.

At the same time, our popular culture is captivated by conspiracy theories.  From 9/11 and Pearl Harbor to the moon landing and the one whose wound has been more recently reopened, the assassination of President Kennedy, we are drawn to elaborate explanations of even tragically banal facts.

These two conflicting tendencies stem from the same error:  the failure to see that something more than facts is needed to get at the truth.  This holds especially for the most factual of human enterprises, scientific inquiry.

Anyone who has argued with a conspiracy theorist knows how difficult it can be to win the argument.  As the great Catholic apologist G.K. Chesterton once observed, "If a man says that men have a conspiracy against him, you cannot dispute it except by saying that all the men deny that they are conspirators;  which is exactly what conspirators would do."  Moreover, the conspiracy theorist will often have a better command of the facts than you.  "But did you know...," or "Well how do you explain...," and "If you're right, then why...," etc.

According to Chesterton the problem is that "his explanation covers the facts as much as yours."  The conspiracy theory "saves the phenomena," as the ancients would say.  We have what philosophers call empirically equivalent theories.  The empirical facts alone aren't sufficient to bring the argument to an end, since the theories all explain the facts.

How do we trace where the conspiracy theorist went wrong?  Not easy, says Chesterton.  The problem is not so much a flaw in his reasoning but that his whole reasoning process has become unmoored.  "The madman is not the man who has lost his reason.  The madman is the man who has lost everything except his reason."

Pierre Duhem, the great Catholic scientist and historian and philosopher of science, is famous for arguing for the "underdetermination" of theory by experiment.  The idea is that testing a single scientific theory by empirical observation always involves testing a bundle of other theories and their corresponding assumptions.  So if I am testing a given theory by observing through a microscope what certain molecules do under certain conditions, I am assuming a theory of optics (in order to make sense of what the microscope shows me) in addition to other theories from biology or chemistry.  I also assume what lawyers call ceteris paribus, idealized conditions, e.g., that my laboratory is a sufficiently closed system, that the temperature in the room is fixed, that the volume of physical bodies do not accordingly fluctuate, and so on.

Thus, for example, if the observation I predict turns out not to obtain, the empirical facts tell me that one of my theories or assumptions is incorrect — but which?  Perhaps my laboratory conditions were inadequate, perhaps there is a hitherto unknown physical force acting on the molecules, or perhaps the laws of optics are wrong.  I must choose.  But if the facts alone are insufficient to guide my choice, to what can I appeal?

Duhem's answer was reminiscent of Chesterton's:  good sense.  If I make an unexpected observation, one that would require modifying a theory I cooked up last week, good sense dictates that my theory, not the laws of electromagnetics, ought to be modified.  If my radio telescope detects the background radiation predicted by the Big Bang, not the Steady State model of the universe, it makes good sense to throw out the Steady State model — not the telescope.  But, Duhem would emphasize, it was not the empirical fact, but the empirical fact combined with a free choice about how to interpret it — a choice informed by good sense — that led you to accept the Big Bang.

Duhem readily admits such choices between theories are mostly constrained by other factors.  One such factor is simplicity:  too many recalcitrant observations make the old theory unworkable.  Other criteria are aesthetic, such as "elegance," which is especially prized by theorists and mathematicians.  Scientists are also similarly educated and socialized so it is often obvious which choice will be accepted as correct.  As a consequence, many philosophers have thrown out Duhem's appeal to "good sense" as unnecessary or vague.  But I think he was on to something.

To be sure, in many cases where there is a stark choice to be made, the various theories are not so different in merit that selecting one over the other would be akin to accepting a conspiracy theory.  But many choices faced by the scientist in the laboratory, whether they are recognized as choices or not, are much less stark.  In a sense, the scientist is always ruling out the existence of unknown forces perturbing the laboratory apparatus, or upending the laws of optics.  What licenses such choices?  Good sense, the (correct) assumption that to choose otherwise would simply evince bad judgment.

Duhem did not invent this concept of good sense.  The 17th century mathematician and philosopher Blaise Pascal once distinguished between l'esprit de géométrie (the geometrical mind) and l'esprit de finesse (the subtle or keen mind).  The former is a rational faculty necessary for logic and geometry.  By contrast, l'esprit de finesse, though rational, requires something more:  a subtleness or nimbleness of thought.  Duhem's idea was that this rational je ne sais quoi is a necessary part of scientific inquiry.

So this is not an argument for the irrationality of science or the marginalization of views that don't conform to what the majority accepts.  It is not as though scientific inquiry progresses according to the whims of scientists or by excluding unpopular theories.  Rather the inverse:  scientific knowledge progresses through the back and forth about how well a theory can or cannot explain the data.  But in addition to criteria such as empirical robustness, simplicity, and elegance, good sense should always play a role.

There were times, for example, in the heyday of phrenology or Eugenics, when good sense was precisely what was called for.  There may not have been good data, but then good sense ought to have put an end to it.  Today, we would be naïve to think that we are somehow immune to similar bad judgments and their political consequences.

Duhem thought that "underdetermination" applied only to certain experimental sciences such as physics and chemistry.  In fact, Chesterton's conspiracy theorist reminds us that these considerations come into play not only in the philosophy of science, but also in the dilemmas that face us in social, political, and moral life.  Contemporary culture holds scientific knowledge up as the final arbiter of Fact.  But if scientific inquiry is not itself a simple matter of accumulating indisputable facts, but a subtle process of interpreting, weighing, and explaining them, how could the messy and complex choices we face in our daily lives be any different?

When those in the public sphere hold up scientific theories as irrefutably factual in order to justify their goals, skepticism is in order.  Similarly, when those in the public sphere tear down scientific theories for being insufficiently factual in order to justify their goals, caution is in order.  Something more is needed, something that the scientific theory in question cannot provide, but is no less rational for that:  good sense.  More than ever in our own time, when the social, moral, and political consequences of scientific and technological change threaten to overwhelm us, this is a precious quality in both scientists and citizens.