[Please read Part I of this post first]
Part IV: The disunity of science
This next part sketches an account of the disunity of science, one which makes room for Popper, but one which creates other discomforts for philosophy of science, which has generally been committed to the unity of science, the unity of knowing, and the unity of truth.
Unity is the cornerstone: if it is science that produces knowledge, and the tip-of-the-nose principle is not a matter of science, then under the unity principle, we cannot know that “your freedom stops at the tip of my nose.” That would be a misuse of the word know. Nor could it be true that your freedom has that limit, for truth is reserved for statements such as “distance traveled equals initial velocity plus acceleration multiplied by the square of time elapsed.”
But such totalism retards rather than advances the production of knowledge. Scientism, unchecked, must injure science. Hence this essay.
Experimental sciences versus the rest
The idea of two kinds of science is anathema to many, who will argue that there is only one science, which follows the one true method for knowledge production, which is the scientific method, which is irrefutably superior to any other method for producing knowledge, as seen by the many technological triumphs bequeathed since the scientific enterprise got into gear.
Does that language have religious overtones to your ear? Might this be the language of scientism, rather than science? Let scientism be defined as the defensive crouch, and also the reactive and boastful fist in the air, the intellectual distortions produced by enduring and apparently unending combat with religious faith, as pitched a battle today, in some parts of America, as in any previous era. Religious folk refuse to listen to reason, refuse to abandon the field, refuse to be cowed by the countless achievements of science. Dogged resistance among people of faith appears to drive some scientists mad.
Are there two kinds of science? If so, then one kind would be exemplified by physics. In any macro-debate about science, physics always come first. Physics is science, period. Everybody knows that. If you probe into why this claim is self-evident, you will discover that physics can claim to be the anchor science, the criterion science, because physics is so mathematical. And math is pure knowledge, pure truth. There’s nothing provisional about a mathematical proof.
Here it is helpful to know about physics envy. Put that phrase in quotes and enter it into a search engine; you won’t come up empty. To gain legitimacy as an intellectual endeavor, it helps to be regarded as scientific. And to ape physics would appear to be the royal road to appearing scientific, for any newborn intellectual discipline that seeks legitimacy.
But wait: in the preceding four paragraphs, I seeded two quite different rationales for accepting the superiority of physics and physics-like scientific endeavors. The two rationales are often swapped in and out of argument surreptitiously, and even unconsciously, in a manner I hope to have mimicked, the better to unfreeze your thinking.
The first rationale takes technological triumphs as the proof that science produces knowledge (and of the superiority of the scientific method, the proximate means by which knowledge is sorted from ignorance). The technological gap between what people can do today, versus 400 years ago, before the scientific endeavor took root, is undeniable. Inasmuch as many of these technological developments can be explained by, or even premised on, theories of physics and physical chemistry, this technological rationale points toward physics as the anchor science.
The second, entirely separate rationale starts with mathematical proof as the ultimate test of what is known to be true. Knowledge consists of what can be proved mathematically. Physics and math are inseparable; look at any university curriculum in the sciences, and note how calculus and physics are joined as omni-prerequisites (in pre-med, for instance). The tight connection to higher forms of math once again makes physics the anchor science, the exemplary discipline of knowledge production.
As the scientific method came to be applied more widely, newer sciences—fledgling sciences, struggling sciences, adolescent sciences—could in principle have modelled themselves on either of these two rationales for enthroning physics as the criterion science. They could have sought technological success or they could have striven for mathematical formulation. In the event, the social and behavioral sciences, the sciences concerned with human being, went uniformly for math, with Economics on the leading edge. In fact, with the exception of biology after DNA was discovered, all candidate sciences—all sciences newer and less well-established than physics—have considered mathematical formulations, and not technological success, to be the royal road to acceptance as a real and mature science.
Nonetheless, whenever the scientific endeavor is challenged from outside, technological accomplishments provide a ready fallback justification, a last line of defense. It’s not just math that supports the scientist, when he lays claim to a superior form of knowledge; for “…we have got / The Maxim gun, and they have not.” A gun—or a nuclear bomb—is a marvelous form of proof, accessible to anyone, and able to be repeated on demand. I don’t believe this loaded gun will blow your head off if fired; I know it to be true. And if you believe, because you are a witch doctor, that this gun, this creature of socially constructed Western ways of knowing, cannot harm you, I can show that belief to be false. And anyone can make that refutation—just hand them the gun. No Ph.D. required, nor any training in math.
From the standpoint of argumentation, the great advantage of human scale technological advances—the gun, the drug—is their accessibility and repeatability. Anyone who fires an intact gun, properly aimed, will achieve the same effect. Anyone who shoots up a large enough dose of heroin will stop breathing. The gun and the drug work every time, for everybody. You don’t have to be an expert to know their truths. Embodied as a gun, and on the scale of a gun, you cannot reasonably doubt that F = ma. Popper does not apply.
Unfortunately, the common sense argument just elaborated would not hold water in any academic department of philosophy. It suggests reliance on a pragmatic theory of truth, and accordingly, suffers from a gaping logical fallacy. This is most easily explained with reference to the electric light technology which we have in our homes.
The Problem with pragmatic demonstrations of the certainty of scientific knowledge
To the naïve technologist, the ability to command light by the flick of a switch, once the house has been properly wired and connected to an electric generator, is one of many everyday proofs of the advance of science. No human had such ability to control light, prior to a century or so ago. And—here is the logical fallacy—the technologist claims it is the scientific theory of the electron, and its elaboration in terms of a theory of electricity, which bequeathed the ability to control light at home at will. The discovery of electricity becomes one among many scientific advances showing the superiority of the scientific method as a means of producing knowledge.
The philosopher will have none of it. The scientific knowledge, if any, resides in the theory of electricity, and its postulation of tiny entities carrying a negative charge, a particle which is also manifest as one of several types of electromagnetic energy waves, yada yada. Scientific knowledge has little to do with know-how or can-do; science, as examined by philosophers, is know-that: we know that electrons can be mathematically described in such and such a way, as part of a larger system of equations that characterize electromagnetic energy. Per Popper, we may someday find a better explanation for electrical phenomena, including light bulbs, one that dispenses with tiny “particles” carrying a “charge.” Virtual strings vibrating in multiple dimensions might explain electricity and much more. The mathematical equations we now think to be true would then stand revealed as a local subset of a larger system, and true only locally.
As a general rule, any given instance of can-do or know-how can be explained by, and be consistent with, uncountably many general theoretical explanations. And since, for the philosopher, knowledge resides in the theory, and above all, in the mathematical laws in which a good theory consists, instances of can-do and know-how prove nothing and underwrite no claim of truth. Truth is a property of statements, not actions.
If you have not spent all your life in academia, you may be scratching your head. How can technological advances not provide support for the claim that the scientific method produces knowledge? How can the unique prevalence of technological innovation in our era not be tied back to the equally unique advent and flourishing of the scientific method within the past few centuries?
As a non-academic, you might fear the philosopher has had to destroy the village in order to save it. Once scientific knowledge is located in theory, and theory has been detached from pragmatic consequence, where does the superiority of scientific knowledge now reside? As just shown, it has nothing to do with guns, drugs, light switches or any of the other great improvements of modern life. These become the province of technology, not science. Science is a superior form of knowing because, uh … it’s math?
Not everyone will read that last sentence in a satirical voice. Many luminaries in the social and behavioral sciences would instead nod sagely in assent. Yes, science is math. Math is knowledge. New math is how new knowledge is forged. Do math, or go home.
By this commitment, at a stroke the newer and fledgling sciences were set free from the need to vouchsafe any pragmatic consequence whatsoever. Enough, to test null hypotheses on college sophomores responding, in writing, to instructions given in a classroom building—so long as the statistical apparatus had some bite to it. Enough, to show that free trade produced a net surplus of utility, within the overall system of equations. Enough, to show mathematically that a common currency across the European continent must increase the efficiency of market transactions.
In any case, as Popper said, these non-pragmatic theoretical structures must be subject to endless revision. Enough therefore to find some test, some single prediction, some particular observation, however remote from everyday affairs, and however minute, which could nonetheless be connected to a mathematical formulation. Per popper, the [economic, psychological, sociological] theory is entirely temporary; therefore, no point in obsessing over the quality, depth, or richness of the data. Can-do and know-how are beside the point when the goal is to do psychological science, or economic science. Indeed, what would a pragmatically effective psychology look like? I shudder to think.
Alas, along came climate change, and the need to convince the public of the need for precipitate action. Now it became necessary for the edifice of theory to connect back to pragmatic consequence. Climate theory needs to underwrite action in the world, expensive and difficult action. But that link between theory and technology had been broken long ago.
By Popperian definition, there must be uncountably many possible explanations for the recent rise in planetary temperature. A natural occurring observed correlation between carbon and temperature cannot underwrite any particular dosage model, in which the amount of carbon in the atmosphere, for the very first time in planetary history, would be artificially made to decrease, after a long period of artificial increase.
True, the hypothesis that carbon and temperature co-vary has been tested observationally, and has survived its tests thus far. But here comes Popper again: positive test results are logically weak evidence. A universal statement can be proved wrong by a single discrepant instance; but a universal statement can never be proved right by any finite collection of instances. Yes, carbon and planetary temperature have both moved up and down together, on multiple instances in the past; but, as the mutual fund prospectus states: “Past performance does not guarantee future returns.” No accumulation of positive tests can guarantee the truth of the universal statements which make up scientific theories. Newton’s theory survived countless tests—until it didn’t.
That insight was a core part of Popper’s appeal: he had revived the problem of induction, and was celebrated for it. ICYMI: Popper is Hume reborn.
The mutual fund prospectus is as close to Hume as most non-philosophers will ever come. And if I had to guess, most natural science Ph.D.s in America hold mutual funds, and many would assent to the truth of that bit of boilerplate. I would go further and wager that almost all Finance professors, and very many economists, would vigorously assert the truth of that mutual fund disclaimer: past performance does not guarantee future returns.
Unless it’s the past performance of planetary temperature in association with carbon dioxide: you can take that to the bank. We, the scientific community, guarantee the truth of the dosage model: future reductions in carbon will arrest the rise in global temperatures before these get out of hand. That outcome is scientifically certain. Forget Popper.
You see how the imperative of addressing climate change has made Popper so terribly inconvenient. Now math is no longer enough. Pragmatic action is at issue. But the link between pragmatic results and the truth of scientific theory has been sundered; the termites of philosophy chewed it through.
How can we both assent to the mutual fund prospectus, and decry skepticism about climate change? Easy: we have only to treat financial markets, and planetary climate, as two very different classes of phenomena. In a word: we have but to accept the disunity of science.
Let’s return to the certainty of the gun. The intact gun works every time, for everybody, regardless of race, creed or color of skin. A child can test F = ma (and too many do).
A gun is the fruit of experimental science: we can vary mass, and acceleration, independently. We can repeat these experimental variations over and over, in the field. Real objects at human scale can be smashed, at will, by using a small explosion to accelerate a concentrated mass of lead within a metal barrel.
We cannot vary planetary temperature and carbon concentrations at will. We cannot experiment on the planet. We can develop mathematical formulations of great sophistication that describe postulated interactions between carbon dioxide and temperature; but those descriptions are only that, and the underlying causal connection is only hypothesized.
Non-experimental sciences—or more exactly, domains of scholarship where true experiments are unlikely or impossible—can never have the pragmatic certainty of scholarship that is subject to an action-in-the-world criterion. The physics and chemistry of guns can be tested, at human scale, over and over; the dosage model connecting carbon in the atmosphere to planetary temperature cannot.
The implication is that we can be good scholars of climate, and paleontology, and history, and macroeconomics; and can be better scholars yet when we capture our observations in precise mathematical formulae; but we can never do the kind of science, in these observational settings, that was done for kinematics and the law of gases. In kinematics, scientists do experiments; over here we cannot.
I remind the reader that to propose any limit on science, and the omni-applicability of scientific ways of knowing, is rank heresy. To say that psychology and economics, and maybe climate studies, cannot be the same kind of science as kinematics, is to violate standards of academic propriety. C. P. Snow is dead and not lamented.
At this point, the sober reader might ask, How could there possibly be an alternative to carbon as an explanation for recent changes in planetary temperature? There are so many observations, and the carbon model fits them so well. Put another way, Popper’s fallibilism and tentativeness would have more bite if a plausible alternative hypothesis for recent joint rises in temperature and carbon can be devised.
The aside below makes the attempt.
—————————————-
Aside:
The classic riposte to any assertion of causation from correlation is to find a third variable that explains both correlates. Applying this general rule to climate change: what else could explain the correlation between planetary heat and greater concentrations of carbon dioxide in the atmosphere? The current preferred model, on the part of climate change insisters, translates the observed correlation directly into a causal model, locating the cause in the human activity of burning fossil fuels. In outline, the causal model is:
Humans burn things –> releasing carbon into the atmosphere –> raising temperature via a greenhouse effect
This causal model has a clear pragmatic takeaway: stop burning so many carbon-y things. Lower the carbon dose, and you restrain / slow down planetary warming. Thus, the Kyoto and Paris treaties, methods to determine your carbon footprint, and all the rest. Carbon is the culprit. In this model, carbon works just like the light switch.
To visualize an alternative model, change one thing: suppose it were decreed that in the future, all power must come from nuclear power plants. No one is to burn anything carbon-y anymore. All motive power would come from nuclear generated electricity, and any needed heat energy would come from running nuclear electricity through a resistor. Leave aside whether this is practical—just do it mentally, while leaving all other human activities unchanged.
In this thought experiment, humans would continue to make dark roofs, pave roads, and plow fields, covering much of the planet with unnaturally dark surfaces that convert sunlight to heat energy. Humans would continue to use central heating and to power industrial processes, directly releasing unnatural amounts of heat. Humans would continue to evaporate moisture from skin, and store up water in reservoirs, releasing unnatural amounts of water vapor over land. Billions of humans would continue to release waste heat and carbon dioxide directly into the atmosphere, as cells burn calories to live, and respire. Cattle would continue to fart.
Would planetary temperatures, now that we are not burning carbon-y things, soon revert to 18th century levels? Or would the planet stay warm, because so much of human comfort and convenience requires heat energy, and rests on breathing and farting creatures large and small?
The new model would be:
Civilized humans release heat and moisture and carbon dioxide and methane –> raising planetary temperature directly and by greenhouse effects
Under this alternative model, human civilization is the proximate cause of planetary warming. The only way to slow down or reverse planetary warming would be not to have billions of humans inhabiting the planet. Which is a solution I do not care to contemplate.
end aside
—————————-
The point of this thought experiment: absent experimentation in the world, we have no way to rule out either model. If we return to the mainstream climate model—that too much burning is the problem—and strive mightily to reduce the burning of carbon-y things, and discover several decades later that planetary temperatures remain elevated, we still won’t know whether the carbon model was right or wrong. Single trial experiments are not conclusive. The connection between Theory and any one observation is necessarily distant. We might not have reduced carbon enough; we might have released too much in earlier years; lag effects may be longer than expected; unsuspected feedback loops may be present.
It is repeated and repeatable experiments whose results have pragmatic force. And even then, as academic philosophers correctly note, the link between any set of statements that could justify the term Theory, and some specific set of pragmatic actions, must always remain more or less remote. The nature of scientific certainty can be debated; but not that of technology, the fruit of experimentation.
Technology is real; technology is powerful. Pragmatic results that have been oft-repeated will continue to be repeatable. Really, dear reader: do you think guns will just stop working some day? Under Humean reasoning, it is quite possible that guns might of a sudden, cease to fire: after all, past performance is no guarantee of the future. So how much of a Humean are you really? How seriously do you want to take the problem of induction? Would having this gun barrel jammed against your skull help you in considering the question?*
*Riffing off Samuel Johnson here (“I refute it thus!”)
Science, at the theoretical level, can’t ever have the certainty of religion. And unlike religious faith, which becomes more certain as we move away from specific tenets, toward the postulation of godhead, scientific theory becomes more uncertain, the more distant it becomes from specific experimental outcomes.
The certainty of technology differs from the certainty of scientific knowledge. And we don’t have any technology of climate change—at least, not as yet. So how certain can we be about the carbon dosage model?
Observation-based theories may be mathematical in form, and scientific in spirit, but can never have the pragmatic certainty of experimentally tested actions. It is an illusion to suppose otherwise, but an illusion powerfully fueled by the reluctance to admit the disunity of science, or to acknowledge any limits on the scientific ways of knowing forged in kinematics.
But acknowledging disunity, and accepting limits, might check the impulse toward scientism. And scientism poses the real threat, as far as I’m concerned.
Continue to Part V
Be First to Comment