To the Editor:
Professor Herbert Heidelberger, of the University of Massachusetts at Amherst, needs no defense from me or anybody else. But the treatment his letter received in the October issue [Letters from Readers]1 may well have provoked in him that peculiar blend of irritation and boredom in which one simply says to oneself, “Drop it!” And it shouldn’t simply be dropped
Professor Heidelberger was obviously attempting to show that a certain argument purporting to prove that marijuana is dangerous is an invalid argument. His procedure was this: he set about to construct a second argument, of the same form as the original, whose premises are all true, and whose conclusion is false. A piece of “sophomoric sophistry.” as Norman Podhoretz called it? Nonsense. This is an eminently respectable procedure for showing that an argument is invalid, in use by philosophers and non-philosophers alike at least since the time of Socrates.
What can Samuel McCracken be referring to by the expression, “Mr. Heidelberger’s exercise in pure statistics”? Just what is it that he is calling a chestnut in writing, “Mr. Heidelberger’s exercise in pure statistics is a chestnut going back at least to Mark Twain”? Professor Heidelberger’s second argument? To my knowledge, it appears nowhere in the writings of Mark Twain, or anybody else. Professor Heidelberger’s procedure for showing that the original argument was invalid? And are Aristotle’s procedures for assessing validity in syllogistic arguments also chestnuts?—after all, they go back at least to Aristotle.
Moreover, so far as I can see, Professor Heidelberger’s use of his procedure was entirely successful, even elegantly so. So far as I can see, his second argument does have true premises and a false conclusion, and is of the same form as the original argument; if I am right in thinking this, then the original argument really was invalid. If opposition to marijuana had to rest on such support as the original argument gives it, those of us who think marijuana dangerous, and who oppose legalizing it, as I do, would be in serious trouble.
COMMENTARY is often insulting, but it is not often both insulting and silly. Professor Heidelberger is owed an apology.
Judith Jarvis Thomson
Department of Humanities
Massachusetts Institute of Technology
Samuel McCracken writes:
I could claim that Professors Heidelberger and Thomson have fallen prey to the seductive force of von Manteuffel’s Paradox, but that would be the coward’s way out. Let me deal with the problem head-on.
There is, first, a certain amount of cognitive dissonance going on here. I responded to Professor Heidelberger’s second argument as if it were an approximation of one advanced, I think by Mark Twain, to the effect that water must be intoxicating, since you get drunk whether you mix it with whiskey, brandy, gin, or rum. (If Professor Thomson has searched the Collected Works for this in vain, then she has the advantage of me; I at any rate recall hearing it ascribed to Twain.)
In either argument, the reasoner (whether Pseudo-Twain or Professor Heidelberger) uses true premises to argue to obviously false conclusions. The form of argument, Professor Thomson declares, must therefore be invalid, and I can do no other than agree. Why did I not react thus to Professor Heidelberger’s letter? Because it did not then occur to me that this was all he was saying, and that he actually believed Norman Podhoretz to have made an argument so classically an example of the fallacy post hoc ergo proptey hoc. Professor Heidelberger’s imputation of such a blunder is, pace Professor Thomson, quite as insulting as any remark made in rejoinder to it. Now, as a matter of fact, Mr. Podhoretz did not say that marijuana use causes heroin use, but made the quite different statement that it can lead to heroin use. The two claims are distinguishable, and I do not see that either Professor Heidelberger or Thomson has made the distinction.
One difficulty with the water-leads-to-heroin formula is that it is not strictly analogous to marijuana-leads-to-heroin, for water (including in such mixtures as whiskey and milk) is consumed by all human beings, a fact not true of marijuana. If we are talking, and we are, about statistical correlation as a proof of causation, it will be the case with regard to water that whatsoever human behavior we may tentatively attempt to correlate with it, the correlation will emerge, inasmuch as the first correlative is universal human behavior. The presence, therefore, of constant conjunction is of no particular importance. Nor can there be any proper test of concomitant variation, for there is no way to remove the consumption-of-water factor to see if it is associated with a concomitant consumption of heroin. (That the deprival of water it extended long enough will be associated with suspended consumption of all else is of no interest either; death after extended inanition is likewise universal human behavior.)
It is quite otherwise with marijuana and heroin. Inasmuch as not all human beings use marijuana, its conjunction with heroin becomes an object for further analysis. At this point, the argument funs into difficulties for lack of statistics. Is it true that all those that have used heroin have used marijuana? If so, it becomes the case that in the most striking variation one can imagine, that of presence as opposed to absence, the absence of marijuana use is concomitant with the absence of heroin use. Among those who do use heroin, furthermore, two populations can be found, A and B. For the former, marijuana use is concomitant with heroin use, for the latter, not. There being no clear way of categorizing the naive into A and B, for any one person the decision to use marijuana may entail the decision to use heroin. No one, I think, would employ the fact that some drinkers, even a majority, can drink regularly without addiction as proof that alcohol is harmless for all those who cannot. We have been for some time accustomed to discerning two populations among drinkers, A and B, for, each of whom the drug is functionally entirely different. None of this argument, to be sure, is meant to demonstrate conclusively that marijuana does lead to heroin use, merely to suggest the possibility that it can.
Finally: my characterization of Professor Heidelberger’s argument as an exercise in “pure statistics” meant no more than that his argument ignored relations between the use of marijuana and heroin other than the purely temporal, such as the extent to which the former constitutes an induction into the drug culture, in both its social and assumptive natures. But to accept Professor Heidelberger’s difficult reading of Mr. Podhoretz’s argument would indeed force one to agree that his procedure is successful, even elegant. It may be doubted whether it is very illuminating.
1 Professor Heidelberger’s letter commented on Norman Podhoretz’s Issues column, “Seducer of the Innocent,” and Samuel McCracken’s article, “The Drugs of Habit & the Drugs of Belief,” both of which appeared in June.—Ed.