Freddie rejects the notion that data exists in a vacuum:
Empiricism exists within a framework of theory, and theory cannot be derived empirically. The fact-value distinction is real. (This argument of mine is illustrative of its own point: I take it as an empirical truth, not a normative statement, but its empirical claims are necessarily grounded in theoretical assumptions.) And fact-value problems exist for both the commission of empirical projects and the evaluation of empirical results.
While I sympathize with this stance, Freddie is wrong to think that journalists are going to–or should–reflect on the fact-value distinction. When scientists themselves routinely push a simplistic model of empiricism, it’s not fair to expect Ezra Klein to resolve that disagreement. Here’s the lengthy comment I left:
Great post. But methinks the problem doesn’t start with Klein as much as it does with academics. The delusion that empiricism alone can solve our problems started with us, and we continue to promote it. The Ezra Kleins out there are simply following our lead.
While I sympathize with your efforts…there are sound historical, institutional, and cultural reasons why social scientists are not the standard-bearers for empiricism and its connection to theory.
For better and for worse, it’s usually the natural and physical scientists who speak on these matters. Even more specifically, it’s often physicists. And trust me, they are not going to accept that “science exists within philosophy.” I have tried to raise this very point before, and I can’t begin to describe the ruckus it caused.
If your goal was to pick a fight, you couldn’t have chosen a better title. I can predict my friends’ response: “What? Are you trying to say that the mass of the electron depends on philosophy? That there are norms about the existence of gravity?”
These discussions always get reduced to something like gravity. Again, trust me because I’ve been there.
You’re trying to raise a very complicated idea. One that requires a ~1,000 word example to even get started. It’s a bit much to expect wonks to descend into this territory.
It’s especially harder when much public outreach about science and empiricism by academics stress that data alone can set us free, that empiricism is a way to free ourselves from our pre-conceived philosophies, and that thinking otherwise is the first step to destroying the Enlightenment.
If you want to change Ezra Klein, start with the physics professor down the hall from you.
Ashutosh Jogalekar explains why chemistry can never be reduced to physics:
To me the real problem with reductionism has not really been philosophical as much as practical. In some cases there is almost no tangible link between a phenomenon and its very deep underlying components. For instance, no one can reasonably draw a causative link between two people falling in or out of love or two countries going to war and the strong force between protons and neutrons. This is what we might term “strong reductionism” where the limitations of reductionist philosophy are stark and obvious. Philosophers love to discuss this kind. But I think that the type of reductionism that is far more relevant to the daily work of most scientists is what may be called “weak reductionism”. This type applies to phenomena which can actually be connected to other basic phenomena in a reasonable sense, but it’s just that explaining them in terms of these basic paradigms is utterly unhelpful at the level of the phenomena themselves.Let me state an extreme example which would make this case clear. To a first approximation a physicist might say that there is no difference between a banana, a human being and a suspension bridge since all of them are made out of protons, neutrons and electrons. The only difference is really in their numbers. Almost everyone (including physicists) would realize that non-explanation as absurd. While it does seem like the ultimate unifying elucidation, it says absolutely nothing about the very different functions performed by a human, a banana and a bridge. All three of these respectable entities would resent their reduction to varying numbers of subatomic particles.An example more familiar to chemists which is stated by Hoffmann also vividly illustrates the problem with reducing chemistry to physics. Consider the carbonyl functional group, a workhorse of chemistry. The most important reaction that this group undergoes is nucleophilic addition. How does physics explain this process? By essentially pitching electrostatics. Physics will tell us that the carbonyl carbon has a partial positive charge and the oxygen has a partial negative one, thus attracting nucleophiles to the carbon. But a chemist would find this simple explanation deeply unsatisfying. There is much complexity associated with addition to carbonyls which goes beyond merely electrostatic attraction. There’s the angle of attack of the nucleophile- the well-known Burgi-Dunitz trajectory- which maximizes orbital overlap. There’s coordination of positively charged counterions with the oxygen which can dictate the stereochemistry. There’s also the size of groups on the attacking nucleophile which can sharply tip the distribution of products through steric effects. Then there’s the gradation of reactivity of various nucleophiles based on their size and charge. And finally, there’s the all-pervasive solvent which can drastically change product ratios and stereochemistry through solvation effects.Now note that the truly fundamental underlying basis for all these factors (and indeed, virtually everything in daily life) is the electromagnetic force, and so yes, physics can purport to actually ‘explain’ all these factors by saying that they are all mediated through electromagnetism. Even steric effects are essentially electromagnetic in nature. Yet this would be akin to saying that wars happen because people get really angry at each other. The physics-based explanation is useless to the chemist since each one of the ingredients responsible for nucleophilic addition constitutes a unique chemical feature and conundrum which the chemist has to understand and predict. A unifying framework for these based on the physicist’s conception of electromagnetism does nothing to delineate the special role that each factor plays in controlling the reaction. To a carbonyl-enamored chemist, these determinants are as fundamental in their own right as protons and neutrons are to a physicist.Thus in such cases, reductionism fails not because there is no palpable connection between the chemical phenomenon and its physical underpinnings, but because the physics-based explanations tend to be useless at the level of chemistry. One of the simple tests for interrogating the utility of reductionism then consists of asking whether a reductionist approach can help truly explain a certain phenomenon at the same level that the phenomenon is embedded in its parent discipline. Sometimes it helps, more often than not it doesn’t. The underpinnings of the American Civil War and World War 2 are different. Chemistry is not physics.
According to Thompson-Reuters, eight-thousand and seventy-three science and engineering journals published just over one million peer-reviewed articles in 2010. Another two-thousand one hundred and seventy-six social science journals published over 200,000 papers. This works out to two peer-reviewed journal articles being published every minute of every hour of every day for the entire year.
At the start of the millennium, there were “only” seven-thousand three-hundred and eighty-three journals. So in a mere ten years, almost three-thousand new journals were formed. This trend will surely intensify as more nations invest in science. Brazil, China, Korea and Turkey are all publishing at least 10% more scientific papers every year. Even tiny countries like Tunisia and Qatar are getting in the game, more than tripling the share of GDP spent on research since 1996 (see here).
So what does it mean to be a scientist anymore? Is it fair to group theoretical physicists with observational ecologists? Should these groups have the same label when scientists in the same discipline don’t always overlap? I have attended several annual meetings of the American Geophysical Union. It always struck me that the soil scientists are separated—physically and intellectually—from the space physicists. The near-Earth space physicists are separated from those studying Mars, who are in turn separated from those studying Venus.
To clarify why I found Williamson’s and Rosenberg’s definition of naturalism so unsatisfying, let’s look at the former’s definition of the scientific method:
What is meant by “the scientific method”? Why assume that science only has one method? For naturalists, although natural sciences like physics and biology differ from each other in specific ways, at a sufficiently abstract level they all count as using a single general method. It involves formulating theoretical hypotheses and testing their predictions against systematic observation and controlled experiment. This is called the hypothetico-deductive method.
What does it mean to make a systematic observation? How about controlled experiment? These words are vague and imprecise, and can mean different things to physicists and psychologists. They can even mean different things to space physicists and atomic physicists. Systematic isn’t really possible when solar and magnetic field conditions change continuously, but can be if the entire experiment is run on a lab bench.
The discussion would have been much richer if they spent some time (dare I say it?) deconstructing these terms. Even at a “sufficiently abstract level” it’s not clear how to lump particle physics and observational ecology under a single general method. As I’ve said repeatedly, the various branches of science are disunified and heterogeneous.
Granted, these were relatively short blog posts and not a dissertation on naturalism. But given my longstanding desire to disaggregate science and demonstrate its intense diversity, I’m always disappointed when writers whose blogs are more widely read than mine don’t explore these themes.
If semiotics, existentialism, hermeneutics, formalism, structuralism, post-structuralism, deconstruction and post-modernism transparently flout science’s standards of objectivity, or if they seek arbitrarily to limit the reach of scientific methods, then naturalism can’t take them seriously as knowledge.
That doesn’t mean anyone should stop doing literary criticism any more than forgoing fiction. Naturalism treats both as fun, but neither as knowledge.”
While skewering Rosenberg on this point, Williamson’s response muddles two distinct questions. We can ask whether science requires naturalism as defined by Rosenberg. We can also ask whether the knowledge produced by science is the only type worth having. These are different issues and the entire discussion confuses them.
I’ve decided I can’t bring myself to care that much about the first. Science in practice is effectively naturalistic almost all the time and is probably going to stay that way. It’s even harder for me to care in this case because I don’t find their definition to be helpful. I suspect people like Paul Newall can marshal history and philosophy to reject all naturalism criteria. Although I might agree with him, I won’t spend any more time trying to decide either way.
The second question is the far more interesting one, and though Williamson is mostly on solid ground his argument doesn’t quite add up. Consider this passage from his intellectual dismembering of Rosenberg, where Williamson appears to believe history and literary criticism are valid forms of knowledge:
Rightly noting the successes of physics, [Rosenberg] says “We should be confident that it will do better than any other approach at getting things right.” What things? If he means questions of physics, what reasonable person denies that physics will do better than any other approach at answering those questions? But if he means all questions, why should we be confident that physics will do better than history at getting right what happened at Gettysburg?
I raised history and literary theory as test cases. According to Professor Rosenberg, naturalism treats literary criticism as fun, but not as knowledge. Does he really not know whether Mr. Collins is the hero of “Pride and Prejudice?”…
For Professor Rosenberg, it may turn out that “reality contains only the kinds of things that hard science recognizes.” By “hard science” he seems to mean something like physics…That physics does not show that there is such a thing as a debt crisis does not mean that physics shows that there is no such thing as a debt crisis: physics simply does not address the question. That is no criticism of physics; it has other work to do.
This line of reasoning will be familiar to long-time readers of this blog. Other than the chest-thumping that is distressingly routine among scientists, there is no general-purpose epistemological ranking. If you want to determine the acceleration due to gravity, use physics. If you want to analyze the protagonists in a Jane Austen novel, use literary criticism. Nothing that controversial here.
The confusion occurs because at the same time Williamson insists history and literary criticism produce useful knowledge on their own terms, he calls for a “broader conception of science that includes mathematics, history, much of philosophy, and the sensible parts of literary criticism, as well as the natural and social sciences.”
So which is it? Do we expand science to include practically all forms of scholarship–at which point science just means research Williamson happens to value? Or do we recognize other equally valid, but different, forms of knowledge production? Both answers can’t be correct.
This schizophrenia weaves throughout Williamson’s posts. On one hand, the “scientific spirit” includes both experiments and philosophical reasoning, implying an expanded definition of science. On the other, in that same paragraph, scholars who “know what they are doing” may use methods outside the natural sciences, implying different research methods can be equally valid. Even though mathematics is “one of the most spectacular success stories in the history of human knowledge”, it must be counted as a science to receive that title.
Perhaps I am violating my own rules by expecting consistency and rigor in mere blog posts. Nevertheless, I was dismayed to see Williamson’s indecisiveness. He can’t quite come out and say scientific knowledge is not always better. By sneaking in subjects with no clear connection to traditional science, he undermines the stronger part of his argument and loses some of the benefits of categorization.
Physics has more in common with chemistry and microbiology than it does with literary criticism. Literary criticism in turn has more in common with art history than it does with thermodynamics. As long as we are nonjudgmental and recognize that boundaries are fuzzy and overlap, it can be useful to sort these activities into different groups. But as nebulous as they may be, these boundaries do exist: literary criticism and particle physics are substantially different forms of knowledge. Pretending otherwise devalues both.
Writing in the Times, Timothy Williamson criticizes the dogma of naturalism:
Which other disciplines count as science? Logic? Linguistics? History? Literary theory? How should we decide? The dilemma for naturalists is this. If they are too inclusive in what they count as science, naturalism loses its bite. Naturalists typically criticize some traditional forms of philosophy as insufficiently scientific, because they ignore experimental tests. How can they maintain such objections unless they restrict scientific method to hypothetico-deductivism? But if they are too exclusive in what they count as science, naturalism loses its credibility, by imposing a method appropriate to natural science on areas where it is inappropriate. Unfortunately, rather than clarify the issue, many naturalists oscillate. When on the attack, they assume an exclusive understanding of science as hypothetico-deductive. When under attack themselves, they fall back on a more inclusive understanding of science that drastically waters down naturalism. Such maneuvering makes naturalism an obscure article of faith. I don’t call myself a naturalist because I don’t want to be implicated in equivocal dogma. Dismissing an idea as “inconsistent with naturalism” is little better than dismissing it as “inconsistent with Christianity.”
UPDATE: In the comments, Paul notes that my truncated quote somewhat misstates both him and Feyerabend. I should have been more careful as I know Feyerabend probably would not have insisted on methodological anarchy. My bad. Read Paul’s full post for more context.
I first turn to Paul Newall’s keen observation that Feyerband’s historical approach to understanding the progress of science itself violates his insistence on methodological anarchy:
The result was to place the epistemic systematists in the absurd position of advocating a methodology for science that would have killed the very progress allegedly brought about because early scientists followed the methodology.
Although this reductio succeeds, I want to suggest that Feyerabend’s historiography, and perhaps the historical approach in general, is somewhat paradoxical. The aim of his historiography is to free us from methodological or epistemological strictures but we find ourselves using the lessons of history to show that there are no lessons to be learned from history. This is too simplistic, though: what Feyerabend argued was not that there is and can be no methodology worth adopting but rather that all methods have their limits. Nevertheless, the paradoxical aspect comes from considering the use Feyerabend makes of history. Faced with a methodological rule, we can look to the history of science – and to apparently paradigmatic cases of good practice in particular – and show that an application of the rule would have been disastrous. However, it seems that this relies implicitly on a fixed interpretation of the events under consideration; after all, if it were possible for a rationalist or anyone else to recast the episode in a more favourable light for the rule at issue, we might be able to show that in fact its application would have worked then as now.
I now turn to a rather ancient Massimo Pigliucci post on why Occam’s Razor isn’t as useful as it’s sometimes made out to be:
The obvious question to ask about Ockham’s razor is: why? On what basis are we justified to think that, as a matter of general practice, the simplest hypothesis is the most likely one to be true? Setting aside the surprisingly difficult task of operationally defining “simpler” in the context of scientific hypotheses (it can be done, but only in certain domains, and it ain’t straightforward), there doesn’t seem to be any particular logical or metaphysical reason to believe that the universe is a simple as it could be.Indeed, we know it’s not. The history of science is replete with examples of simpler (“more elegant,” if you are aesthetically inclined) hypotheses that had to yield to more clumsy and complicated ones. The Keplerian idea of elliptical planetary orbits is demonstrably more complicated than the Copernican one of circular orbits (because it takes more parameters to define an ellipse than a circle), and yet, planets do in fact run around the gravitational center of the solar system in ellipses, not circles.Lee Smolin (in his delightful The Trouble with Physics) gives us a good history of 20th century physics, replete with a veritable cemetery of hypotheses that people thought “must” have been right because they were so simple and beautiful, and yet turned out to be wrong because the data stubbornly contradicted them.
The responses on The Dish makes me think they didn’t click through and read the full post.
One of the many remarkable features of Popper’s thought is the scope of his intellectual influence. In the modern technological and highly-specialised world scientists are rarely aware of the work of philosophers; it is virtually unprecedented to find them queuing up, as they have done in Popper’s case, to testify to the enormously practical beneficial impact which that philosophical work has had upon their own. But notwithstanding the fact that he wrote on even the most technical matters with consummate clarity, the scope of Popper’s work is such that it is commonplace by now to find that commentators tend to deal with the epistemological, scientific and social elements of his thought as if they were quite disparate and unconnected, and thus the fundamental unity of his philosophical vision and method has to a large degree been dissipated. Here we will try to trace the threads which interconnect the various elements of his philosophy, and which give it its fundamental unity.