Thursday, June 10, 2004

Post Hoc Ergo Propter Hoc

One way to improve one’s chess is by studying the traps associated with the different opening systems. One way to improve one’s reasoning is by studying the ‘traps’ into which the inexperienced reasoner is likely to fall. These ‘traps’ are called fallacies. We study them in order to avoid them in our own thinking, and in order to be able to diagnose the ‘pathology’ in the thinking of others. Without a typology of fallacies, one might sense that there is something wrong with the reasoning of one’s interlocutor without being able to say what exactly is wrong with it. To correct an ailment (whether medical or ratiocinative), however, one must first diagnose it properly.

A rather common fallacy is post hoc ergo propter hoc, which is Latin for ‘after this, therefore because of this.’ This fallacy is committed by someone who takes temporal sequence as sufficient for causation. Suppose event A occurs, and then event B. It does not follow that A is the cause of B. To think otherwise is to commit the fallacy in question. If Jack sneezes and Jill jumps, it does not follow that the sneezing caused the jumping. Jill’s seeing of a mouse might have caused her jumping.

In a more general sense, the post hoc ergo propter hoc fallacy is committed by someone who takes any temporal relation as sufficient proof of causation. Suppose process P1 occurs simultaneously with process P2. It does not follow that P1 is the cause of P2. Here is an example of the generalized post hoc ergo propter hoc fallacy. It is from a delightful old book by Georges Herment entitled The Pipe, trans. A. L. Hayward (Simon and Shuster, 1957, p. 150:

So far as longevity is concerned it has been proved by statistics that the greater the smoker the longer he lives. A number of centenarians have been fiends for smoking. Statistics, again, prove that the expectation of life has been longer since the introduction of tobacco and its use by the ordinary man. Taking the example of France: in 1830 the average duration of life was no more than twenty-eight years; in 1953 it is forty-five years, the consumption of tobacco in the population having trebled in that period.

Now you might think that Herment is joking, but the context shows that he is not. What he is doing is committing the fallacy of post hoc ergo propter hoc in the generalized sense lately explained. An increase in longevity occurs at the same time as an increase in tobacco consumption, whence it is inferred that the latter caused the former. Non sequitur! The two phenomena are not causally related. The increase in longevity was most likely due to medical advances, while the increase in tobacco consumption was most likely due (in large part) to the increased number of people available to consume tobacco products. Hence the increase in longevity most likely occurred despite the increase in tobacco use.

Knowing some logic, I was inoculated against Herment’s bad reasoning. What’s more, the analysis of the above passage made possible by my logical savvy afforded me great pleasure. The pleasures of analysis, like the pleasures of the intellect generally, are definitely superior to those of the senses, a point that J. S. Mill granted but could not justify within his ethical system. But that’s a topic for another occasion.