What Can Science Really Know?

Two book reviews on philosophy of science appeared in the leading general-science journals Nature and Science last week.  Both of them downplayed the oft-told triumphalist portrayal of science as a progressive path toward infallible knowledge – the picture most students get in school.
    In Nature,¹ N. David Mermin (Cornell) gave a surprising reprimand to an icon of triumphalist science: Alan Sokal (see 06/03/2008 commentary).  Sokal’s famous hoax against the postmodern deconstructionists in 1996 embarrassed them soundly and signalled the approaching end of the Science Wars of the 1990s.²  Sokal emerged as a champion of scientific realism.  His bold trick made him a darling of the scientific establishment.  The hoax’s value as a victory for scientific claims to epistemic superiority is debatable, though.  Some viewed it more as a boyish taunt than a serious conflict over ideas.  Among them is the reviewer of Sokal’s new collection of essays, Beyond the Hoax: Science, Philosophy, and Culture (Oxford, 2008).  Mermin, a physicist and colleague of Sokal, did not have much good to say about this book.  He considered Sokal’s treatment of critics of the triumphalist spirit of science as ill-informed, dismissive and shallow.  He found much to agree on, “But Sokal’s unwillingness to expand his frame of reference to accommodate legitimately different points of view undermines his effectiveness as a scourge of genuine rubbish,” he concluded.  “I would like to think that we are not only beyond Sokal’s hoax, but beyond the science wars themselves.  This book might be a small step backwards.”
    In Science,³ Kim Sterelny (philosophy program, Australian National University and Victoria University of Wellington) reviewed William C. Wimsatt’s essay compendium, Re-Engineering Philosophy for Limited Beings: Piecewise Approximations to Reality (Harvard, 2007).  Wimsatt, whom Sterelny considers “among the most creative, original, and empirically informed philosophers of our day,” dealt with the problem of finite humans trying to form idealized conceptions of natural phenomena that are too complex to grasp in their entirety.  These idealizations are heuristic devices.  Scientists believe by faith they can be refined with further research and become better approximations to reality.  In turn, higher-order complex phenomena can then be reduced in terms of their simpler components: i.e., biology reduces to chemistry, which reduces to physics.  Scientific explanation becomes organized into hierarchical domains of increasing complexity.  So much for the triumphal picture:

It is common ground between Wimsatt and his targets that these ideas about science are idealizations, perhaps even extreme ones.  But Wimsatt argues that they are unhelpful idealizations.  For they idealize away from what we most need to explain: the cognitive success of limited beings.  Treating science as ideally rational is like a developmental biologist using preformationism to model development: the subject matter of the discipline has been idealized away.

Wimsatt respects the success of science but understands that causes and effects are not simple.  There exists a “causal thicket” because elements at one level are not always influenced by adjacent levels.  If you understand the chemistry of the atomic bonds in DNA, for instance, how much do you really understand DNA translation?

The world is messy.  We are fallible and bounded.  Yet science progresses with great reliability.  Wimsatt’s conception of science is organized around these three facts.  Like science itself, his account is partial and incomplete, an approximation organized around the idea of a heuristic.  Many questions are left open, and much could be challenged.

Perhaps one question that arises immediately is how a messy, heuristic approach to epistemology could yield reliable knowledge.  In a new lecture series on the solar system,⁴ professor Frank Summers (Southwest Research Institute) stated openly that concordance with reality is not important in science.  If a theory has good explanatory power and makes good predictions, that’s what matters.  With that in mind, he had surprisingly good things to say about Ptolemy’s earth-centered model of the solar system.  It explained complex motions in terms of simple geometrical shapes, and helped its users make predictions to sufficient accuracy for 1500 years – hardly an achievement to sneeze it, whether or not the model corresponded to the way things “really are.”

1.  N. David Mermin, “Science wars revisited,” Nature 454, 276-277 (17 July 2008) | doi:10.1038/454276a.
2.  The Sokal Hoax episode is discussed at length in a lecture series by the Teaching Company, “Science Wars: What Scientists Know and How They Know It” by Steven L Goldman, Lehigh University.
3.  Kim Sterelny, “Philosophy of Science: Addressing Complexity,” Science, 18 July 2008: Vol. 321. no. 5887, p. 344, DOI: 10.1126/science.1156895.
4.  Frank Summers, New Frontiers: Modern Perspectives on Our Solar System, The Teaching Company, 2008.

As these book reviews show, it’s overdue to dispense with the triumphalist, progressivist view of science.  None of these factors guarantee science has a grip on truth: (1) it gets a lot of money, (2) a lot of smart people practice it, (3), it appears to be successful, (4) the textbooks portray it as victorious over superstition, (5) it wins Nobel Prizes, (6) it has a “scientific method” (whatever that is), (7) geeks major in it at school, (8) it’s hard and uses a lot of math, (9) it explains things, (10) it uses a peer review system, (11) it has big organizations and publishes impressive journals, or (12) it owns lots of big buildings and museums.
    Clearly science seems “on to something” because of its practical successes in medicine, electronics and the space program, but even then, how much of the success is due to trial and error?  How much is due to practical engineering?  How much do we assume is true simply because it works according to the best theories of the day?  One only need look at history to see many examples of practical success using theories we now believe are wrong.
    The “hard sciences” like physics and chemistry arguably have the best case to make.  They give us practical benefits like lasers, computers and robotic spacecraft that arrive at distant planets on schedule.  But arriving at Saturn and taking pictures of its rings is different from explaining how Saturn got there in the first place.  Even physics gets pretty far out when it comes to the Copenhagen interpretation of quantum mechanics, or the many-worlds interpretation, or cosmology.  How much more the storytelling that goes on in psychology and evolutionary biology and uniformitarian geology?  It’s not beyond belief to foresee today’s leading theories being tomorrow’s pseudosciences.
Look at how many of science’s claims to epistemic priority have been undermined by philosophers of science:

1. There is no one scientific method.
2. Even if there were a scientific method, exercising a method cannot be done without making assumptions and judgments.
3. There are no demarcation criteria between science and pseudoscience that can reliably keep the good stuff in (physics) and the bad stuff out (e.g., astrology).
4. Scientific discovery follows no rules: it can come from hard work, tacit knowledge, accident, or even dreams.
5. Scientific hypothesis-making follows no rules.  Few scientists arrive at a hypothesis out of raw data without some hunch or intuition of what to look for.  Many scientists have their hypothesis before looking at any evidence at all.
6. Scientific explanations are fraught with logical pitfalls.  Reducing a complex phenomenon into simple principles runs the risk, for instance, of “explaining away” the very thing to be explained. 
7. Induction is subject to serious criticism.  It begs the question that patterns in past experience will continue in the future.
8. Prediction is no reliable guide to good science.  Astrologers and other pseudoscientists often succeed at predictions.  Predictive success runs the risk of affirming the consequent: a logical fallacy.  That’s why Karl Popper denied predictive success has any role in scientific justification.
9. Falsification rarely succeeds in overcoming a paradigm’s web of belief.
10. Scientific reasoning may differ in diligence but not in substance from other kinds of reasoning.
11. The requirement for natural laws is fraught with pitfalls.  Are laws descriptive or normative?  Statements that sound like laws may be nothing more than accidental generalizations.  Laws also make claims that far outrun experience; on what basis can they be justified?  Should natural laws be permitted that have zero instances?  Newton’s did – e.g., “A body in motion acted on by no external forces will continue in a straight line forever.”  Some legitimate sciences, like biology, employ few natural laws, and the ones they do employ are often plagued by exceptions.
12. Science is a vague term with wannabees trying to latch onto the prestige of the word.  Where does one draw the line?  At social science?  political science?  economics?  Christian Science?  Scientology?  The speculations even within a “hard science” like physics are arguably just as unproveable as those of a “soft science” like psychology.

What are you left with?  “What works for our needs right now.”  We call something scientific if it gives us some nice feelings with its explanations, allows us to make useful predictions, or gives us some practical control over the world.  Its grip on reality or truth is tenuous at best.
    Once we get past triumphalist science, we should take a more informed look at other avenues of human knowledge.  History and the humanities might want to re-assert some of their claims in the marketplace of ideas.  Likewise, philosophy and theology have been footstools of imperial science for too long.  Provided that scholars in other fields apply sound principles of reason, use thorough research methods, interact where theories are analyzed and different points of view are considered, and build on prior knowledge, are their methods really so different from those of science?  Aren’t these good practices for any kind of research?
    At the end of the 19th century, science was king.  Two world wars later, and a century of revolutions in philosophy of science later (with no clear winners), it’s time to re-evaluate science’s claims to special epistemic status and cultural priority.  Take another look at that quote by Dr. Daniel Robinson at the top right of this page.