post hoc ergo propter hoc

after this, therefore because of this(Latin)
Logical fallacy based on confusing correlation with causation.

This inaccurate phrase offers an interesting source for a fallacy noted by Milan Kundera in his novel Immortality. In all Romance languages, evolved as they are from Latin, and in English as well, the word "reason" denotes both "cause" and "the process of reasoning," suggesting that any event's cause must have a logical connection to its result.

This produces a number of problems, not least of which is the utter inability of most Western nations to understand socio-cultural phenomena in anything other than a hyper-Hegelian, rational, systematized context. School shootings are an excellent example: we strive to locate their cause in myriad logical ways, negelecting to consider the possibility that something about the modern era, something neither logical nor rational precipitates violence and destruction. In searching for "reasonable" sources for such behavior, we inevitably miss the point: something irrational, unreasonable, something related to broad psychosis in the contemporary American, could be at fault.

A notable exception to this confusion: the German language. Kundera notes that the German word for "cause" is "Grund," a word derived from the word for earth or ground, which suggests that some events have as their cause something buried, intrinsic and irreducible, illogical and unintelligible. Kundera thinks that we all behave in our lives in manners related to our individual "Grund," our base identities, rather than according to "reasonable" psychological predictions or other logical systems of understanding.

We struggle to trace causal factors, to understand the "reasons" behind events, but the presence of a cause does not entail the presence of rationale.

Lest this 'hyper-Hegelian' view seem utterly moronic, it is plausible (though not certain) that, at least at the macroscopic level (to avoid problems with quantum physics), all things are caused. This doesn't mean that nothing is mysterious, but it does mean that mystery can come in two major varieties: stuff we don't know, but about which there was a fact of the matter (for example, the number of hairs on Abraham Lincoln's head at the moment he was shot is a mystery, but only in this relatively mundane sense); and stuff which involves more than merely understanding the physical causes. For example, even knowing exactly what physical cause produced the avalanche that killed your dog, you might still wonder why things had to be so arranged that your dog died. My general feeling is that mysteries of this second type are actually meaningless--it is possible, for example, for there to be no plan, and there would then be no meaningful answer to give other than an account of the physical goings-on. That's a guess, though.

In any case, there seems to be no reason that the assumption that all things are caused is wrong, and it seems to be a fairly optimistic, humanistic belief. Rather than presuming that there are things we could not possibly understand (a seemingly defeatist attitude), it assumes that nothing is beyond us, in principle. It also allows us to DO something about our problems--if we can know why school shootings happen, we can prevent those circumstances from coming about.

There's an additional question, with this particular issue, of why so little attention is paid to those who make the shooter feel ostracized and tormented. I don't know the answer to that--Hegel seems to be failing us. :-)

I read an editorial, Vaccination Undermined*, with greater interest than I expected to. It gave voice to a conundrum which has been a personal thorn for some time -- specious correlation. The idea that we assume that because one thing follows another that the one thing was caused by the other. This very simple fallacy, which, in our society, causes a great deal of wasted resources, is one for which we seem to have an inexplicably opaque blind spot.

Some colleagues at work and I played with this problem a few months ago. We considered that one way to address it would be to introduce a catchy saying into the lexicon. Sayings are neat, codified cultural reminders of complex concepts. For example a "big fish in a little pond" is a seven syllable allusion to a rather subtle and involved aspect of status. Another example is how even a single word, less than a saying, schadenfreude, serves the Germans to describe a rather nasty form of quiet pleasure of which we English-speakers are too ashamed to label.

Like the Germans, Rome or the Church, had a means we don't have of reminding itself of this correlation fallacy. (I was interested to note that an episode of The West Wing was dedicated to it. In addition Google finds it 9,270 times. (Interestingly, 'big fish in a little pond' gets only 2,200 hits.))

My friends and I worked away at it for a good 2 minutes.

We didn't do a very good job. "Breast milk isn't mercury" was one alluding to the fact that every single person who starts off life drinking breast milk eventually succumbs to death. The other was "The nose does not cause the tail" derived from this Alan Watts fragment:

"Again, this is a problem which comes from asking the wrong question. Here is someone who has never seen a cat. He is looking through a narrow slit in a fence, and, on the other side, a cat walks by. He sees first the head, then the less distinctly shaped furry trunk, and then the tail. Extraordinary! The cat turns round and walks back, and again he sees the head, and a little later the tail. This sequence begins to look like something regular and reliable. Yet again, the cat turns round, and he witnesses the same regular sequence: first the head, and later the tail. Thereupon he reasons that the event head is the invariable and necessary cause of the event tail, which is the head's effect. This absurd and confusing gobbledygook comes his failure to see that head and tail go together: they are all one cat."

So I throw down the gauntlet here. Can anyone come up with a cute, English saying to remind us not to assume if A follows B, B caused A?

*http://www.sabin.org/PDF/111703_01.pdf

From Act I of Travesties, by Tom Stoppard (the acts are not divided into scenes, but if you would like to look it up, this passage appears on page 36 of the 1975 Faber and Faber edition):

CARR (stiffly): I believe it is done to drink a glass of hock and seltzer before luncheon, and it is well done to drink it well before luncheon. I took to drinking hock and seltzer for my nerves at a time when nerves were fashionable in good society. This season it is trenchfoot, but I drink it regardless because I feel much better after it.

TZARA: You might have felt much better anyway.

CARR: No, no—post hock, propter hock.

TZARA: But, my dear Henry, causality is no longer fashionable owing to the war.

I am therefore tempted to suggest, in response to E Propter Hoc's call for a catchy slogan refuting this fallacy, "It could be the seltzer." But obscure allusions will only carry us so far. In practice, it seems that the closest thing to a popular anti-fallacy catchphrase is the rather literal-minded "Correlation is not causation"—which, if you want to get picky about it, is really directed against the closely related "Cum hoc ergo propter hoc" fallacy.

These seem like such obvious fallacies that, if one considers them in the abstract, it may be difficult to imagine how anyone falls into them. But people do, all the time. This happens quite frequently in situations in which the supposed cause and the supposed effect are really two effects of some third factor which has been overlooked.

Here is an example. Francesco Rutelli, the mayor of Rome, is quoted on page 57 of the May 2, 2005 issue of The New Yorker as saying:

I think cities are like languages. If a language doesn't change, grow, and evolve, it dies. It is the same with cities—a city must be transformed from time to time.

Rutelli knows (I hope) more about the survival of cities than I do, but his "Evolve or die" view of language is simply preposterous. The fact is, if a language has a sufficient number of speakers to ensure its continued existence, then it will inevitably continue to change. Little differences from idiolect to idiolect, shifting social pressures, contact with speakers of other languages, and the need to describe new discoveries will all exert their various influences on the language, eventually shaping it into something quite different from what it had been. On the other hand, if a language has so few speakers that it does not change, then it is almost certainly doomed, not because of its stasis, but because people find it less useful than languages that will enable them to communicate with larger numbers of fellow-speakers. So sure, "if a language doesn't change ... it dies," but that doesn't mean that you can save a dying language by fiddling with its syntax or revamping its vocabulary. In fact, that will do about as much good as drinking a glass of hock and seltzer.

The expression post hoc ergo propter hoc, Latin for "after this, therefore because of this,” is a well-worn logical fallacy based upon the mistaken notion that simply because one thing happens after another, the first was necessarily a cause of the second. This conclusion is, of course, incorrect. Many events follow in sequence without being causally related. A baseball player wears a particular shirt one day and hits a home run, for example. Although someone following post hoc logic might believe the two events to be causally related, they’d be wrong. Coincidences happen.

The obvious flaw in such post hoc reasoning, however, has yielded a peculiarly damaging result with respect to federal laws forbidding retaliation against employees engaging in “protected activities” such as filing a complaint of sexual or racial discrimination. The typical scenario goes something like this. Employee files a charge with Company alleging sexual harassment by Supervisor. Charge disappears into the bowels of Company’s Human Resources Office. One week later, Employee finds herself miserable and alone working the graveyard shift.

If Supervisor moved Employee to the graveyard shift because of the harassment charge, he has unlawfully retaliated against her. When Employee tries to file her complaint, however, the only evidence she’s going to have is the temporal sequence of events. “I filed a charge and was moved to graveyard one week later.”

Company’s lawyer, if he’s worth anything, will move for dismissal, arguing that Employee’s “post hoc ergo propter hoc” logic is insufficient. And in most federal courts, particularly the more conservative ones, he’ll win, because the Court will jump to the equally incorrect conclusion that because temporal sequence alone doesn’t prove causation, no causation occurred. Sort of a “two logical wrongs don’t make a right” kind of deal.

BrevityQuest 2007