[Please read Part I of this post first]
Part III: What now of Popper?
Maybe, Popper was a philosophical dead end. Maybe, falsifiability was never central or important for demarcating science. Might falsifiability have been naught but its own bold conjecture, since superseded?
For those who follow Popper, scientists should never be devout, but always questioning, always approaching knowledge as naught but today’s best guess. The radical character of this position becomes clear if we set aside Popper for a moment and ask, How else ought we to approach true knowledge, other than with complete assent and perfect conviction? Truth cannot be provisional; the word provisional applies to guesstimates, not to knowledge. Absent Popper, to be devoted to the truth becomes once again a respectable scientific attitude. Darwin is fact. Climate change is real. Deal with it.
Maybe Popper’s philosopher colleagues were right: Popper was an ideological warrior, and not someone who pioneered a new understanding of science and how it proceeds.
To press the argument: in what sense was Newton proven wrong? After Einstein, do billiard balls, struck at a particular angle with this much force, no longer land in the pocket? Does distance travelled no longer equal initial velocity, plus one-half of acceleration multiplied by the square of time elapsed?
Not at all; most Newtonian predictions continue to be accurate, in the midrange between the cosmological and quantum scales—the domain where they were initially subjected to rigorous attempts at falsification. Outside that midrange, Newton’s meta-theoretical framework has some problems and failures, which Einstein’s alternative does not. Inside that midrange, Newton’s theories are as effective as ever in helping you lob an artillery shell over a rampart. Nothing that worked before Einstein, ceased to work; it’s only the explanation for why it worked that changed.
On the face of it, then, scientific revolutions are not overturns but supersets. Einstein could predict everything Newton predicted, and more besides; he introduced a better, more comprehensive solution, not a different one. Newton’s metatheory was flat out wrong, from the beginning; space and time are not the separate and empty vessels he depicted. But his midrange theories, within their sphere of applicability—calculating distance, velocity, acceleration and mass for human-scale objects—were and remain correct, for pragmatic purposes at human scale.
Likewise, were Einstein’s meta-theories of space and time to someday be shown wrong, they would always have been wrong; but the calculations we successfully make using Einstein’s theory—adjustments to Global Positioning Satellite data are notable among its pragmatic applications —will continue to be successful.
Switching gears, even if Darwin’s theory were some day to be found incomplete, everything we can do within the current neo-Darwinian synthesis—manipulating genes to bring out or suppress desired and undesired traits, exerting selection pressure to shift population values of some trait—would continue to be workable. We would simply explain our success on different grounds, with a new meta-theory that allowed us to predict outcomes and perform actions we could not, within the current neo-Darwinian synthesis.
On this first gambit, Popper was both right and wrong: right about some elements of how scientific knowledge is gained, wrong about other aspects. The parts of science that have direct pragmatic consequences don’t change; these pragmatic parts aren’t provisional. Successful tests stay that way. It’s the portions of theory that have not been tested, and / or the deep infrastructure of theory, which can’t be directly and unambiguously tested, that turn out to be provisional and endlessly subject to revision.
A good scientific theory is always both ahead of its tests, when first elaborated, and never completely tested, no matter how much testing has occurred. We may call this kind of good theory Big Theory. Whatever can be completely tested is not big, but merely instrumental, a micro-theory, a component theory. Whatever is Big—more than immediately instrumental—can never be completely tested, which means that it can always fail its next test. That future failure will trigger replacement with a different arrangement of the deep structure; or alternatively, introduction of a superset of the existing postulates.
This first gambit suggests one way we might resolve the question of whether the science of anthropogenic climate change, and its propositions concerning carbon in the atmosphere, should or should not be made subject to Popperian doubt. We could examine the portions of climate theory which have been repeatedly tested, and accept these unreservedly; we’d also identify the portions that represent deep explanatory structure, not easily or readily tested, and reserve any doubt for the latter.
Ask yourself: which parts of climate change theory have been repeatedly tested? Are there any parts that have not yet been tested, and which are those, exactly?
As I write, there appear to be two parts of climate change theory which have been tested, and one part that has not.
Tested:
- The planet has been getting warmer, and is now warmer than it has been for centuries.
- This increased temperature, and many other changes in the past, both increases and decreases in temperature, as far back as we have ice cores, have been associated with increases (decreases) in the levels of atmospheric carbon dioxide.
Not tested:
- If we slow down our current rate of addition of carbon to the atmosphere, temperature change will slow down accordingly; and if we ultimately succeed in reducing the levels of carbon dioxide to those prevailing before the current increase in temperature began, global temperatures will also revert to their baseline levels.
The third, untested proposition is of a different order than the first two. It represents the kind of knowledge claim that classically has been subjected to experimental testing. In fact, as stated, it describes an experiment: let’s raise and lower carbon, manipulate its level, and see how planetary temperature responds. But it also represents an experiment not yet conducted.
I’ll return to this third tenet of climate change theory. First, we need to pick up the thread with Popper. We’ve seen one way of rescuing both Popper and the non-provisional truth of (some) scientific claims; but there is at least one more gambit to the same end that we should consider.
Two kinds of science?
The first gambit consisted of distinguishing two levels of scientific claim, one always open to doubt, the other not.
On the second gambit, some scientific domains, at all levels, would be Popperian, where ideas about falsifiability are applicable and illuminating, and other scientific domains would not be Popperian. Macroeconomic history (Marx) and personality psychology (Freud) might be two examples of Popperian domains. Claimants to scientific status in these domains need to treat their theories as provisional, no matter how many tests have been conducted. There’s something about these domains that make an accumulation of successful tests less probative than in other domains. Perhaps tests here must be notably indirect and partial; maybe there’s a never a really good test, because there are too many intermediary links between Theory and what can be observed.*
*Sometimes known as the Duhem-Quine thesis, here located in and confined to certain scientific domains, rather than treated as a general critique.
Other domains, tied more closely to repeatable experiments at human scale, such as much of Newtonian kinematics, are not Popperian: knowledge claims here, either because more mathematical in their expression, or more limited in their scope, are not subject to doubt.
That’s still pretty vague, so let me flesh out the idea of a Popperian scientific domain by offering a prototypical instance: human paleontology. More specifically, let’s consider theories about how humans evolved from non-human forms, based on fossil evidence, and associated theories about the genesis of key evolutionary changes, as in the old thesis of “man the hunter,” which was used to explain bipedalism and the opposable thumb.
Paleontology is the quintessential Popperian science because here a theory is never more than one fossil find away from its overthrow. The entire theoretical structure of how a population of primates eventually produced a population of humans rests on a few dozen skull fragments. One day a finger bone is found in a Siberian cave, and Bam! an unsuspected third early modern human species, the Denisovans, takes its place besides Neanderthals and Homo Sapiens. A few footprints are found preserved in lava, and Lucy pushes the origins of bipedalism back a million or two years.
Here, it helps to be old. I’ve been reading paleontological accounts for 50 years, since the mid 1960s. The best available explanations have changed quite a bit over that span. Neanderthals were the predecessors of early Homo Sapiens. No, they died out without issue. No, they interbred with our ancestors; except if your recent ancestry is African rather than European. Or Asian, in which case you may have Denisovan ancestry instead. Our species, Homo Sapiens Sapiens, appeared about 40,000 years ago. No, they were in Australia by 60,000 years ago. The first evidence of modern cultural capabilities is found in Europe. No, that’s just Euro-centrism in another guise.
You get the picture. “Provisional” is an apt descriptor for any paleontological knowledge claim. And paleontology is likely to stay provisional for generations to come. It does not seem likely, no matter how good our satellite imaging or underground radar, that we will ever be in a position to claim, “all the fossils that will ever be discovered, have been discovered.” Hence, paleontological knowledge claims will always remain provisional. We might find a fourth pre-modern species; we might find that tool use preceded bipedalism, rather than vice versa; that language predates Homo Sapiens, or not; that Neanderthals made art, or did not; and so forth. In paleontology, everything is exactly as Popper said.
Now the question becomes, is climate theory more like Newtonian physics, where meta-theoretical postulates were thrown out, but all kinds of pragmatic outcomes remain undisturbed; or is climate change more like paleontology, where every statement, at every level, is ineluctably provisional? If climate theory has already produced predictive equations like kinematics, focused on the effect of carbon release into the atmosphere, then it doesn’t matter whether something at the meta-theoretical level later changes. We can act with confidence that reducing carbon release, now, is going to produce the expected results. But if climate theory is more like paleontology, then the idea that reducing our carbon release will arrest global warming is of the same order as the idea that Neanderthals are (are not) among our genetic ancestors, or that the advent of the genus homo dates to just over 2 million years ago. New evidence may appear that casts doubt on the particular dosage model currently advanced. The experiments haven’t been done.
Returning to the first gambit, global temperature change has been tested many times. The evidence is really strong that temperatures have increased and continue to increase, and that carbon dioxide levels have increased in synch. But there is another part of climate change which has not been tested: the effects of a reduction in carbon release outside of any naturally recurring cyclic reduction. That test has not been performed. No one gives you a population of planets to play with, which would enable you to run repeated trials of the effect of artificially adding and subtracting carbon to a planet’s atmosphere.
The untested dosage model that lies beneath debate about climate change is even more specific: that if we reduce the release of carbon in 2017, relative to 2016, and in 2018 relative to 2017; and reduce it further, each year thereafter; then, global warming will begin to slow down. Moreover, if we start now and reduce carbon enough, global warming will stop at whatever point it has reached in 2025, 2035 or whenever our carbon reduction efforts get far enough along; and catastrophic warming will be averted. In short, the scientific hypothesis is that anthropogenic carbon, carbon on the margin, is tightly coupled to planetary temperature, hence fully reversible. Our carbon release, added to natural carbon sources, acts like a toggle. Toggle the switch back, and we can cool the planet back down as readily as we warmed it up.
Again, has this precise dosage model been tested? Of course not. There can be no prior or preliminary test of a planetary dosage model. The only possible test is in the doing.
Disagreement about global warming turns out to be beside the point. That part’s easy. We may soon have, or may already have, overwhelming evidence of excess planetary warming, in lock step with anthropogenic carbon release—successful tests. Those measures of temperature may be like the many successful measures and predictions concerning the movement of planets, made following the theoretical innovations of Newton and Leibniz. There need be nothing provisional about climate theory predictions of 21st century global warming.
Once again, I snuck a rhetorical device into the previous paragraph: “in lock step with.” Had I instead written “correlated with anthropogenic carbon release,” readers might have been reminded of such tropes as “correlation is not the same as causation,” which would bring Popper to bear again on climate theory. Lock step has a pleasing connotation of certainty, suggesting an almost mechanical linkage, with temperature changes locked to changes in atmospheric carbon. This would validate the dosage model: planetary temperature, if locked to carbon levels, can certainly be made to move up or down by manipulation of carbon levels.
But lock step is rhetorical overstatement. We don’t have that kind of evidence for the dosage model; rather, we have increasingly well-elaborated measures of how, in the past, temperature and atmospheric carbon dioxide have co-varied, moving up and down together.
- For instance, if you plot planetary temperature over the past century or two against levels of carbon dioxide, you get a strong association.
But also:
- If you plot consumption of high fructose corn syrup over the past few decades against the incidence of obesity, you likewise get a strong association.
- And if you plot serum levels of cholesterol against the incidence of heart disease, you also get a strong association.
Conceptually, it’s the same sort of evidence in all three instances: observed at the macro-level, one quantity is observed to co-vary with another. But their treatment in contemporary discourse is very different.
The first correlation, of carbon and temperature, has become a litmus test: are you a science denier, lacking an intellectual leg on which to stand? Adherence to the truth of that association and unquestioning extrapolation to the dosage model are required, lest you be scorned.
By contrast, the second correlation, concerning the uniquely harmful character of high fructose corn syrup, among other dietary sugars, can be doubted with impunity. At present it’s a bold conjecture, and interesting to test. Until those tests accumulate, few citizens would make negative aspersions about either the concerned parent, who decided to minimize HFC in their child’s diet, or the nutritionist, who counter-argued that fructose is fructose, carbohydrates are carbohydrates, and at the level of isolated chemical substances taken up by the body, none is any worse than the other. The neutral position is viable; the agnostic stance is acceptable.
The third correlation is the most interesting, because clearly connected to an underlying dosage model, same as carbon and temperature. Once the association between cholesterol and cardiac risk had been repeatedly observed, medical science—not an oxymoron, yes?—undertook a search for a drug dosage which could reduce serum cholesterol levels. Statins were found and tested repeatedly. The tests showed conclusively that administering statin drugs would lower cholesterol in the blood. Accordingly, statins began to be prescribed to large numbers of patients. The science was clear: doctors could reasonably state they we know, scientifically, that statin drugs lower cholesterol.
But does lowering cholesterol improve mortality? A fierce controversy rages on this point. You can search the Internet to learn more, and decide for yourself whether the scientific evidence favors expensive and life-altering steps to reduce your own cholesterol levels. Or not.
Oh, wait: you don’t have an MD and are not trained as a clinical researcher. Doctors—people who do have those credentials—disagree. So what are you going to do? Do you feel qualified to evaluate the evidence for cholesterol on your own? Here is a place to start.
Hey—you are not going to resort to that dodge beloved of Republican politicians, are you? Did I hear you say, “I’m not a medical researcher?”
Many people would have some sympathy for the patient who moaned, “I don’t know what to do. Should I start an expensive life-long course of statins, because my cholesterol levels have exceeded the standard, according to my doctor?” By contrast, no Democrat has sympathy for a person who expresses the same degree of uncertainty about whether expensive and life-altering measures to reduce atmospheric carbon are well-advised.
How did a particular extrapolation from correlational evidence—the historical association between temperature and carbon, leading to the hypothesis of a reversible dosage model, in which diminution of carbon release will reverse recent temperature rises—become an article of faith?
Perhaps it was the bad faith of climate deniers, with their motivated reasoning based on pecuniary interest, which has so warped the debate. Whatever the source, the cost to conceptions of science is high. There is no evidence for the dosage model, nor any evidence against it, nor can there be, until the experiment is run.
Nowhere else outside of climate change would we see the phrase “correlation is not causation” rejected so vehemently.
Science, faith, and politics
The conflict between science and faith continues, especially here in America. That fact bothers and annoys some portion of the scientifically inclined. Debates over Darwin in the schools have provided a focus for that annoyance. Among many secular individuals, separation of church and state represents one of the great triumphs of American democracy. The fight to preserve that separation is the good fight. But citizens on the other side are also energized by the stakes around whether Creationism will be taught in the schools. After all, for some people of conscience, the question might be framed as, Shall godlessness be proselytized by the state?
I am reminded of the still heated debate over whether a democracy can successfully fight a totalitarian and absolutist regime, without becoming itself more totalitarian and absolutist, at least until the contest has been won (see: Manzanar). It seems to me that scientifically inclined people, faced with the unyielding resistance of the religious when it comes to questions of Origins, face the same risks as democracies confronting absolutist opponents. Theism begets atheism; the agnostic position gets squeezed out. The scientific knowledge claims of the day, the whole body and not just the well-tested pieces, gets treated as an Absolute. I think that’s a mistake, and a distortion of the truth about science. Such rigid conviction will inhibit rather than promote future scientific advances.
Now dial forward to the contemporary debate over anthropogenic climate change. I view the matter in political terms. A faction of the citizenry opposes pollution. I belong to that faction: I think dirt and poison are bad—I don’t want them spewed into my living space. But pollution isn’t sin; it occurs because, historically, polluting technologies—think coal—have been cost-effective to use. They’ve been cost-effective in part because of bad cost accounting: the utility isn’t charged for the dirt and poison it releases into the commons in the course of burning fuel. In turn, consumers, who are also voters, pay that much less for electricity. The price signal, on which the whole capitalist regime relies, gets distorted.
In a democracy, disputes like these are the proper terrain of politics: who may do what to whom. A carbon tax, or effective cost-accounting for pollution, can’t happen unless a large enough political coalition comes together to make it happen. That coalition has not yet coalesced; those who would suffer from a carbon tax, and from better cost-accounting for pollution, have successfully resisted. Just as many Christians have refused to accept the mandate that evolution be taught to their children, a cost they don’t wish to bear.
Confronted with the same sort of resistance, secular elites have responded in the same way: as Darwin was a made a matter of faith, so also climate change. Why, this is not a mere political dispute over cost-shifting: our civilization itself is at stake. Won’t heed my arguments for reducing pollution, and making polluters bear the costs of their dirt? I’ll hit you over the head with scientific certainty, the ultimate weapon of discourse.
It’s easy for me to justify a carbon tax and most of the other measures associated with reigning in global warming. I don’t know whether the dosage model is correct, and I don’t need to know; rather, I know that pollution imposes costs on others, and I deny that anyone has the right to pollute as they choose, without paying up. I am as certain of the rightness of a carbon tax as I am certain that “your freedom stops at the tip of my nose.” But I don’t claim that the “tip of the nose” principle represents scientific knowledge; it’s a product of reason, not science.
Ahh, but in what I just said, I put knowledge of the limits of human freedom beyond the reach of science. Should you claim that the tip of my nose poses no limit on your freedom of action, I can’t prove you wrong, in the mathematical and scientific senses of proof. But I am certain that the tip of the nose principle is correct, and I will adduce reasons for it if challenged.
Alas, too few have the patience for conviction-based political struggle anymore. Frustrated at the ballot box and in the legislatures, they seek to enlist science in their cause. This produces a need to insist that global warming the dosage model is scientifically proven. It’s not. Can’t be; the dosage model hasn’t been tested yet. We do not know, and cannot know in advance, whether bending the curve of carbon release will have a timely and proportionate effect on global temperature. That’s the scientific stance: you have to do the test. You don’t know what’s true until you put claims to the test.
Continue to Part IV
Be First to Comment