Intelligent Design and Falsifiability, #2

Yesterday I discussed falsifiability as a general criterion for knowledge. It turns out that demarcation arguments -- standards like verification or falsification, which tend to elevate science and marginalize other disciplines and areas of thought -- when taken in an all-encompassing sense, turn out to be self-defeating. That’s not a good thing. And it’s a pretty obvious logical flaw, but there was a time when people as qualified as A. J. Ayer (the Wykeham Professor of Logic at Oxford University) and his colleagues advanced verification as the death blow to religious and moral knowledge. The logical positivists got a lot of mileage out of it, despite its problems. Most people familiar with philosophy aren’t impressed with it anymore, but the mood of scientism it left behind -- that science is the only way to secure knowledge about the world -- is still pervasive. This is one reason that mentions of falsification (as in Mr. Simberg’s recent post) make me cringe. But there are other reasons, too.

There are a number of significant issues with falsifiability. The word choice, I think, is misleading. Without some familiarity with the difficulties of falsification, one might assume that falsification is just as simple as posing a test, seeing it fail, and then affixing a label to the theory to which the test belongs: FALSIFIED. But it’s nothing like that in principle or in practice. It’s nowhere near that simple. Undoubtedly there is some testing going on, but ‘falsify’ and its derivatives seem much too naïve and final to me.

For one thing, you can’t test a theory in isolation. Any test encompasses the whole system, which includes a huge body of assumptions about environmental and instrumentation conditions. If a test fails, has the theory failed, or a piece of equipment? Or were the initial conditions wrongly configured?

Furthermore, to subject a theory to test, one has to employ an array of auxiliary hypotheses. Again, if a test fails, has the theory failed, or just one or more auxiliary hypotheses? Philosopher of science Alex Rosenberg gives a specific example of the above difficulties in action:

To see the problem more clearly consider a test of PV = rT. To subject the ideal gas law to test we measure two of the three variables, say the volume of the gas container and temperature, use the law to calculate a predicted pressure, and then compare the predicted gas pressure to its actual value. If the predicted value is identical to the observed value, the evidence supports the hypothesis. If it does not, then presumably the hypothesis is falsified. But in this test of the ideal-gas law we needed to measure the volume of the gas and its temperature. Measuring its temperature requires a thermometer, and employing a thermometer requires us to accept one or more complex hypotheses about how thermometers measure heat, for example the scientific law that mercury in an enclosed glass tube expands as it is heated, and does so uniformly. But this is another general hypothesis - an auxiliary we need to invoke in order to put the ideal gas law to the test. If the predicted value of the pressure of the gas diverges from the observed value, the problem may be that our thermometer was defective, or that our hypothesis about how expansion of mercury in an enclosed tube measures temperature change is false…

 

[Rosenberg, Philosophy of Science: A Contemporary Introduction (New York: Routledge, 2004), p. 113.  Emphasis in original]

Now, sometimes this problem is relatively trivial, but as theories are erected on more and more hypotheses, rely on more instrumentation or procedures, and become more complex, the possibility for error or malfunction somewhere is multiplied. The whole system, and not just a single aspect of it, is what is being tested. Worse yet, perhaps, is the fact that auxiliary hypotheses may take all the heat for failures; they can be sacrificed and reformulated to protect the core theory, even when the core theory is mistaken.

Rosenberg raises other theoretical problems with falsification that I won’t go into now. So does William Alston, in his criticism of verificationism, with equal application [see Alston, “Religious language and verificationism,” in Copan & Moser, eds., The Rationality of Theism (New York: Routledge, 2003).]  As have other philosophers of science. It might be better, though, to show that a simplistic notion of falsification doesn’t work in a lot of cases, historically, that are otherwise considered paradigm cases of scientific practice.

For example, Richard Swinburne, considering the falsification criterion, cites the case of Newton’s theory of gravitation:

The most celebrated of all scientific theories is Newton’s theory of gravitation. The orbits of the heavenly bodies and the phenomena of terrestrial gravitation could all be deduced from the magic formula that bodies attract each other with force potential to mm'/r² -- with one exception. The observed movement of the moon was other than what was predicted by the theory. The theory predicted that the forward movement of the apse of the moon in one revolution of the body in its elliptical orbit should be 1º 31' 28". In fact, a sentence inserted into the second and third edition of Principia admitted, “The apse of the moon is about twice as swift”. For sixty years science lived without too much embarrassment with a theory which by the standards of the falsifiability principle had been conclusively falsified. But the scientists nevertheless believed that the theory gave the true explanation of the phenomena of gravitation.

[Swinburne, “Falsifiability of Scientific Theories,” Mind, New Series, Vol. 73, No. 291 (July, 1964), pp. 434-436; available online via JSTOR. Emphasis mine.]

One of Popper’s favorite cases of the application of falsifiability dealt with the “risky” eclipse observations that emerged out of attempts to confirm Einstein’s theory. But Einstein screwed up initially in his math, as Amir D. Aczel relates in his book God’s Equation: Einstein, Relativity, and the Expanding Universe:

[A]s fate would have it, by that time -- 1919 -- Einstein’s theory of general relativity would be complete, and he would have corrected his error about the true magnitude of the deflection of light rays by the sun’s gravity. This correction happened on November 18th, 1915, when Einstein announced the expected bending of a light ray just grazing the edge of the sun as 1.75 arc-seconds -- twice the amount he had predicted in 1914. A philosophical question comes up: What would have happened had history allowed Freunlisch to carry out his experiment, detecting a shift of 1.75 arc-seconds, plus or minus an experimental error, instead of 0.87 arc-seconds as Einstein had predicted. Actually, Einstein had made an additional arithmetic error, obtaining the value 0.83 arc-seconds. Would general relativity then be judged correct, or false? It should be pointed out that the smaller quantity, 0.87 arc-seconds, when computed without the math error, corresponds to the shifting of light when a light ray is considered a particle, and when one therefore uses the old Newtonian theory. It was the true incorporation of relativity that would lead to a doubling of the value. So quite possibly, had Freunlisch’s experiment worked, general relativity may not have been considered proven by the scientific community.

[The quote above was transcribed from the unabridged audio book.]

If a simplistic, strict criterion of falsification had been adopted, and if the original experiment planned by Einstein had been carried out, Einstein’s theory -- the paradigm case for Popper’s standard -- would have been discarded.

So what does all this mean?  Minimally, it means that charity is required.  If Newton's theory of gravity and Einstein's theory of relativity would have been wiped out by falsification, then we ought to cut emerging theories (e.g. I.D.) some slack.  I think there may be something else to consider, though.  Remember the original argument regarding demarcation criteria (e.g. strong verificationism, falsificationism). I noted that the scientism they engendered still carries over today. Then I suggested that the word falsification is misleading as well, conveying a simplistic notion of testing scientific theories. We’ve seen that there are substantial theoretical and practical difficulties with falsifiability, and that a strict version of it would have shot down some of the most fruitful theories in history. So there is a marked disparity between the rhetorical effectiveness of ‘falsification’ and the reality. In light of that disparity, I have to wonder if falsification doesn’t, in fact, function as more of a rhetorical device for the advancement of an entrenched viewpoint, rather than a core principle of scientific practice. I have to wonder if it doesn’t boil down, in practice, to little more than a helpful heuristic that fits into a larger strategy of inference to the best explanation (a strategy common to many forms of inquiry, by the way, such as philosophical and theological reasoning). And if that is the case, then attempts at formulating a new theory -- in this case, Intelligent Design -- should not be discounted or dismissed by that criterion.

Don’t misunderstand me. I haven't suggested that Intelligent Design theory cannot be subjected to falsifiability, in some form or another; only that falsifiability does not seem to be the determinative feature of scientific practice that many people, such as Mr. Simberg, take it to be.

__________________________

*Edited for clarity, 12/29/04.

*Added links to follow-up posts (below), 12/30/04.

Other bloggers on the same issue:

• John Mark Reynolds (Eidos) - see here, here, and here
• David M. (A Physicist's Perspective)
• Joe Carter (Evangelical Outpost)

Special thanks to Hugh Hewitt and Jim Lindgren (Volokh Conspiracy) for the link to my earlier post.  Further posts on this subject: ID & Falsifiability #3, #4.