Belief in Hell might ignite fallacies
Headline for an article at the St. Louis regional Federal Reserve website: “Fear of Hell Might Fire Up the Economy”. Good news for all those Christians who think that faith ought to be rewarded here on earth. The authors, Kevin L. Kliesen and Frank A. Schmid, conclude that “Religious factors can also help explain variations in economic growth […] In particular, in countries where large percentages of the population believe in hell, there seem to be less corruption and a higher standard of living”. The article thus attempts to explain an asserted positive correlation between “belief” and wealth.
In the article is a table listing for thirty-five countries the percentage of people in each who “believe in hell” and per capita income. A simple linear regression on the data yields a weak negative correlation between belief and income. And because the data points are widely scattered, belief, however correlated with income, is a poor predictor of it.
In short: there‘s nothing to explain! And even if there was, the conclusion of the paper would be tendentious. I suspect that belief in evolution too correlates positively with income. What then: “Secular Humanism Might Fire Up the Economy”? Voltaire thought so.
Even if the explanandum were granted, the argument is flawed. If a variable X correlates positively with a variable Y, and Y with a third variable Z, there need be no causal relation between X and Z at all.1 Approval of Bush may correlate positively with smoking; smoking certainly does correlate positively with increased mortality. But even so I would hesitate to say that approval of Bush makes you more likely to die. A common cause would be creating the correlations. The article, in fact, acknowledges some of the difficulties. But you would never guess it from the headline or the lead paragraph.
It’s not clear by what criteria the thirty-five countries in the list were chosen. There is a preponderance of the relatively well-off.2 The poorest is Nigeria, with a rather high percentage of believers (51%) and the lowest income in the list. I wonder whether the inclusion of more countries as badly off would not tip the balance quite strongly in favor of a negative correlation between belief and income.
Philosophers have written a great deal about “method”. A method is a set of precepts for doing good science. What counts as good science is of course open to debate, as are the precepts conducive to it. Nevertheless I think one would find some agreement on rules of thumb–HempelCarl Gustav Hempel (1905–1997). Educated at Göttingen, Heidelberg, and Berlin (where he studied under Reichenbach, Planck, von Neumann). Met Rudolf Carnap in 1929, and moved to Vienna to work with him; became a member of the Vienna Circle. In 1934 he emigrated to Belgium, and in 1939 to the US where he taught at, among other places, the City College of New York, Yale, Princeton, and Pittsburgh. Notable works include “Studies in the logic of confirmation” (1945), “Studies in the logic of explanation” (with P. Oppenheim, 1948), Aspects of scientific explanation (1965), Philosophy of natural science (1966), and “The meaning of theoretical terms” (1973). With Carnap, Hempel originated what came to be known as the standard or received view of scientific explanation: to explain an event is to deduce it from well-confirmed laws and initial conditions. ’s requirement of total evidence, for example, construed modestly as the maxim that one should not ignore any available or easily-gotten data relevant to one’s claims. Here we have a list of countries biased toward the comparatively wealthy. Another maxim is to consider alternative hypotheses. In the present instance the possibility of a common cause has been ignored. Still another maxim is (roughly) that to promote one causal factor at the expense of other, stronger, factors is a mistake, especially if you want to intervene causally in the production of the effect. If levels of education are more strongly correlated with income than belief in hell, then inducing belief in hell rather than improving the schools is poor policy—unless you think that improving public education is hopeless.
Following the maxims will not infallibly yield truth, nor even save us always from error. But the maxims do help mitigate known tendencies to error. The requirement of total evidence combats our tendency to overlook data contrary to our hypothesis. It is an antidote to bias and to wishful thinking. We may wish that in this world virtue should be rewarded. Art often responds to, or caters to, that wish. Siegfried is tall & handsome, Mime short & ugly. But nature, considered independently of human action & production, may be entirely indifferent to virtue, and ignoring that may cost us. Wishful thinking about AIDS has already caused harm.
The maxims of method set forth an ideal. Actually existing science does not adhere to that ideal, & there are sound practical reasons why it should not (just as there are sound practical reasons to use extrinsic criteria, like institutional affiliation, to decide whether a paper in your field is worth reading). Kant wrote of enlightenment as an intellectual coming of age (“What is enlightenment?”, 1785). The mature mind is tolerant, critical, and above all autonomous. How does one become autonomous in thought? By not accepting any proposition on the basis of authority—including that of beliefs whose force derives from desire, however benevolent.
Voltaire thought that intellectual maturity meant the end of superstition and the tyranny of priests. But it did not entail the end of belief in God, nor of faith, considered as a avenue to knowledge. Kant was more circumspect about established religion, but for him too superstition had no place in the life of the enlightened, nor was religion to force itself upon the mind by evoking “pathological” feeling. Hellfire and brimstone are out. He acknowledged the intellectual drive to find order in nature—some indication or sign that the moral law could be fulfilled—but the affirmation of such an order goes beyond what reason, governed by maxims of method, can affirm. Intellectual maturity does not require the abandonment of hope. But it does enjoin us not to foster illusions about the fulfillment of hope.
1. On probabilistic causation, see
John Collins, Ned Hall, & L.A. Paul, eds., Causation and Counterfactuals.
Ellery Eells, Probabilistic Causality (1991).
Christopher Hitchcock, “Probabilistic Causality”, Stanford Encyclopedia of Philosophy (1997, 2002).
Vaughn McKim & Stepher Turner, eds., Causality in Crisis? (1997).
Judah Pearl, Causality: Models, Reasoning, and Inference (2000).
Patrick Suppes, A Probabilistic Theory of Causality (1970).
2. My wife tells me that the original study included the thirty-five countries in the list. So Schmid and Kliesen had only those countries to work with in correlating belief with income. Nevertheless, given the new purpose to which the data is being put, they would have done well to ask themselves whether, with respect to the argument they were making, those thirty-five countries, on the whole rather well off, did not constitute a biased sample—i.e., one in which countries low in income but high in belief are underrepresented.
TrackBack URL for this entry:
Weblogs that refer to “Belief in Hell might ignite fallacies”: