Thursday, 6 December 2012

The Overreach of Stupid Science, Part II

The Abdication of the Philosophers

[Part II of  The Folly of Scientism by  Austin L. Hughes
Originally published in The New Atlantis ]

If philosophy is regarded as a legitimate and necessary discipline, then one might think that a certain degree of philosophical training would be very useful to a scientist. Scientists ought to be able to recognize how often philosophical issues arise in their work — that is, issues that cannot be resolved by arguments that make recourse solely to inference and empirical observation. In most cases, these issues arise because practising scientists, like all people, are prone to philosophical errors.

To take an obvious example, scientists can be prone to errors of elementary logic, and these can often go undetected by the peer review process and have a major impact on the literature — for instance, confusing correlation and causation, or confusing implication with a biconditional. Philosophy can provide a way of understanding and correcting such errors. It addresses a largely distinct set of questions that natural science alone cannot answer, but that must be answered for natural science to be properly conducted.

These questions include how we define and understand science itself.
One group of theories of science — the set that best supports a clear distinction between science and philosophy, and a necessary role for each — can broadly be classified as “essentialist.” These theories attempt to identify the essential traits that distinguish science from other human activities, or differentiate true science from nonscientific and pseudoscientific forms of inquiry. Among the most influential and compelling of these is Karl Popper’s criterion of falsifiability outlined in The Logic of Scientific Discovery (1959).

A falsifiable theory is one that makes a specific prediction about what results are supposed to occur under a set of experimental conditions, so that the theory might be falsified by performing the experiment and comparing predicted to actual results. A theory or explanation that cannot be falsified falls outside the domain of science. For example, Freudian psychoanalysis, which does not make specific experimental predictions, is able to revise its theory to match any observations, in order to avoid rejecting the theory altogether. By this reckoning, Freudianism is a pseudoscience, a theory that purports to be scientific but is in fact immune to falsification. In contrast, for example, Einstein’s theory of relativity made predictions (like the bending of starlight around the sun) that were novel and specific, and provided opportunities to disprove the theory by direct experimental observation. Advocates of Popper’s definition would seem to place on the same level as pseudoscience or nonscience every statement — of metaphysics, ethics, theology, literary criticism, and indeed daily life — that does not meet the criterion of falsifiability.

The criterion of falsifiability is appealing in that it highlights similarities between science and the trial-and-error methods we use in everyday problem-solving. If I have misplaced my keys, I immediately begin to construct scenarios — hypotheses, if you will — that might account for their whereabouts: Did I leave them in the ignition or in the front door lock? Were they in the pocket of the jeans I put in the laundry basket? Did I drop them while mowing the lawn? I then proceed to evaluate these scenarios systematically, by testing predictions that I would expect to be true under each scenario — in other words, by using a sort of Popperian method. The everyday, commonsense nature of the falsifiability criterion has the virtue of both showing how science is grounded in basic ideas of rationality and observation, and thereby also of stripping away from science the aura of sacred mystery with which some would seek to surround it.

An additional strength of the falsifiability criterion is that it makes possible a clear distinction between science properly speaking and the opinions of scientists on nonscientific subjects. We have seen in recent years a growing tendency to treat as “scientific” anything that scientists say or believe. The debates over stem cell research, for example, have often been described, both within the scientific community and in the mass media, as clashes between science and religion. It is true that many, but by no means all, of the most vocal defenders of embryonic stem cell research were scientists, and that many, but by no means all, of its most vocal opponents were religious. But in fact, there was little science being disputed: the central controversy was between two opposing views on a particular ethical dilemma, neither of which was inherently more scientific than the other. If we confine our definition of the scientific to the falsifiable, we clearly will not conclude that a particular ethical view is dictated by science just because it is the view of a substantial number of scientists. The same logic applies to the judgments of scientists on political, aesthetic, or other nonscientific issues. If a poll shows that a large majority of scientists prefers neutral colors in bathrooms, for example, it does not follow that this preference is “scientific.”

Popper’s falsifiability criterion and similar essentialist definitions of science highlight the distinct but vital roles of both science and philosophy. The definitions show the necessary role of philosophy in undergirding and justifying science — protecting it from its potential for excess and self-devolution by, among other things, proposing clear distinctions between legitimate scientific theories and pseudoscientific theories that masquerade as science.

By contrast to Popper, many thinkers have advanced understandings of philosophy and science that blur such distinctions, resulting in an inflated role for science and an ancillary one for philosophy. In part, philosophers have no one but themselves to blame for the low state to which their discipline has fallen — thanks especially to the logical positivist and analytic strain that has been dominant for about a century in the English-speaking world. For example, the influential twentieth-century American philosopher W. V. O. Quine spoke modestly of a “philosophy continuous with science” and vowed to eschew philosophy’s traditional concern with metaphysical questions that might claim to sit in judgment on the natural sciences. Science, Quine and many of his contemporaries seemed to say, is where the real action is, while philosophers ought to celebrate science from the sidelines.

This attitude has been articulated in the other main group of theories of science, which rivals the essentialist understandings — namely, the “institutional” theories, which identify science with the social institution of science and its practitioners. The institutional approach may be useful to historians of science, as it allows them to accept the various definitions of fields used by the scientists they study. But some philosophers go so far as to use “institutional factors” as the criteria of good science. Ladyman, Ross, and Spurrett, for instance, say that they “demarcate good science — around lines which are inevitably fuzzy near the boundary — by reference to institutional factors, not to directly epistemological ones.” By this criterion, we would differentiate good science from bad science simply by asking which proposals agencies like the National Science Foundation deem worthy of funding, or which papers peer-review committees deem worthy of publication.

The problems with this definition of science are myriad. First, it is essentially circular: science simply is what scientists do. Second, the high confidence in funding and peer-review panels should seem misplaced to anyone who has served on these panels and witnessed the extent to which preconceived notions, personal vendettas, and the like can torpedo even the best proposals. Moreover, simplistically defining science by its institutions is complicated by the ample history of scientific institutions that have been notoriously unreliable.

Consider the decades during which Soviet biology was dominated by the ideologically motivated theories of the geneticist Trofim Lysenko, who rejected Mendelian genetics as inconsistent with Marxism and insisted that acquired characteristics could be inherited. An observer who distinguishes good science from bad science “by reference to institutional factors” alone would have difficulty seeing the difference between the unproductive and corrupt genetics in the Soviet Union and the fruitful research of Watson and Crick in 1950s Cambridge. Can we be certain that there are not sub-disciplines of science in which even today most scientists accept without question theories that will in the future be shown to be as preposterous as Lysenkoism? Many working scientists can surely think of at least one candidate — that is, a theory widely accepted in their field that is almost certainly false, even preposterous.

Confronted with such examples, defenders of the institutional approach will often point to the supposedly self-correcting nature of science. Ladyman, Ross, and Spurrett assert that “although scientific progress is far from smooth and linear, it never simply oscillates or goes backwards. Every scientific development influences future science, and it never repeats itself.” Alas, in the thirty or so years I have been watching, I have observed quite a few scientific sub-fields (such as behavioral ecology) oscillating happily and showing every sign of continuing to do so for the foreseeable future. The history of science provides examples of the eventual discarding of erroneous theories. But we should not be overly confident that such self-correction will inevitably occur, nor that the institutional mechanisms of science will be so robust as to preclude the occurrence of long dark ages in which false theories hold sway.

The fundamental problem raised by the identification of “good science” with “institutional science” is that it assumes the practitioners of science to be inherently exempt, at least in the long term, from the corrupting influences that affect all other human practices and institutions. Ladyman, Ross, and Spurrett explicitly state that most human institutions, including “governments, political parties, churches, firms, NGOs, ethnic associations, families ... are hardly epistemically reliable at all.” However, “our grounding assumption is that the specific institutional processes of science have inductively established peculiar epistemic reliability.” This assumption is at best na├»ve and at worst dangerous. If any human institution is held to be exempt from the petty, self-serving, and corrupting motivations that plague us all, the result will almost inevitably be the creation of a priestly caste demanding adulation and required to answer to no one but itself.

It is something approaching this adulation that seems to underlie the abdication of the philosophers and the rise of the scientists as the authorities of our age on all intellectual questions. Reading the work of Quine, Rudolf Carnap, and other philosophers of the positivist tradition, as well as their more recent successors, one is struck by the aura of hero-worship accorded to science and scientists. In spite of their idealization of science, the philosophers of this school show surprisingly little interest in science itself — that is, in the results of scientific inquiry and their potential philosophical implications. As a biologist, I must admit to finding Quine’s constant invocation of “nerve-endings” as an all-purpose explanation of human behavior to be embarrassingly simplistic. Especially given Quine’s intellectual commitment to behaviorism, it is surprising yet characteristic that he had little apparent interest in the actual mechanisms by which the nervous system functions.

Ross, Ladyman, and Spurrett may be right to assume that science possesses a “peculiar epistemic reliability” that is lacking in other forms of inquiry. But they have taken the strange step of identifying that reliability with the institutions and practitioners of science, rather than with any particular rational, empirical, or methodological criterion that scientists are bound (but often fail) to uphold. Thus a (largely justifiable) admiration for the work of scientists has led to a peculiar, unjustified role for scientists themselves — so that, increasingly, what is believed by scientists and the public to be “scientific” is simply any claim that is upheld by many scientists, or that is based on language and ideas that sound sufficiently similar to scientific theories.

No comments: