Post Hoc Reasoning

One of the rules of causal arguments is that the cause must precede the effect in time. In other words, for A to cause B, it is necessary for A to precede B in time. But it is not sufficient. Just because A precedes B in time--and even if A precedes B every time--does not prove that A causes B. Arguing that "A preceded B, and therefore A caused B" is a fallacy called post hoc or false cause reasoning. The former term is short for the Latin phrase, post hoc ergo propter hoc, meaning "after this, therefore because of this." Consider the following examples:

  1. Whenever Fyodor strikes the flint with iron, he makes a spark.
  2. Whenever John thinks he is going to hiccup, he takes a deep breath.
  3. Whenever Nkrumah enters this line of code, his program crashes.
  4. Alison always wins whenever she wears her lucky headband.
  5. The barometer drops whenever it is going to rain.
In the first case, the apparent cause (striking the flint with iron) occurs before the apparent effect (the spark), as is true in both causation and post hoc reasoning. To argue that this is not merely post hoc, then, requires some causal connection between the striking and the spark. Since one of the physical properties of flint is that it produces a spark when struck by iron, we can conclude that striking the flint with iron caused the spark. Notice that this is only valid if we assume that we have accounted for all relevant details. If, for example, we also know that a live electric wire is arcing near the iron, the cause of the spark may be in doubt.

In the second case, it may seem at first as though the hiccup is causing John to take a deep breath, and therefore that the effect (breath) actually precedes the cause in time. But what really causes John to take that breath is his thinking that he is about to hiccup.

The third case may be our first example of post hoc reasoning. Inserting the code precedes the crash, but to know that it causes the crash Nkrumah would have to have a relevant explanation of how the crash occurs. Otherwise, it is just as possible that a bug somewhere else in the program is causing the crash, but that the crash only occurs once this line is entered because the only way that the section with the bug in it is accessed is by this line.

Alison may be superstitious, but is she wrong to believe that her headband causes her to win? Since there is probably no likely physical explanation of the causal link between winning and the headband (such as, "she can see better because her hair is out of her eyes"), we may be inclined to consider this a post hoc fallacy. And it remains a post hoc fallacy even if we consider a psychological explanation: like Dumbo's feather, Alison's headband gives her confidence, and that confidence enables her to win. Because such a psychological explanation seems secondary, it needs to be discussed explicitly before the conclusion, that the headband caused the winning, can be accepted, and even then it would probably be one of many "indirect" causes.

Notice that many superstitions are post hoc fallacies, and are often phrased so vaguely that they will almost inevitably be fulfilled in the normal course of events. Bad luck, for example, may happen to someone who walks under a ladder, but good and bad things happen to almost everyone with regularity, and there is nothing to link the act (walking under the ladder) and the supposed consequence (some particular instance of misfortune).

The reason the barometer drops is that the atmospheric pressure it measures has dropped. And an atmospheric low often leads to rain. So there is a common cause for the barometer drop and the rain, but it would be a post hoc fallacy to argue that the barometer change causes the rain. If that were true, we could avoid rain by physically forcing a barometer's indicator higher.


Exercises

1. In the fallacy of post-hoc reasoning, X is assumed to be the cause of Y only because:

Y preceded X in time.

X preceded Y in time.

X and Y happened at the same time.

X and Y happened at different times.

2. Which of the following is most likely to be a case of post-hoc reasoning?

The lightning struck at 10:02, and the power disruption occured at 10:03, so the lightning must have caused the power disruption.

Bennett was Secretary of Education when large drops in national achievement scores were recorded, so Bennett's policies must have been responsible for students across the country doing worse.

A faulty traffic light caused the accident. I hit the gas when the light should have turned green, but it didn't turn green on time, and then another car hit mine.

Raising the minimum wage means that the money available to pay salaries will be split up among fewer workers. So raising the minimum wage will cause employment to go down.

1. In the fallacy of post-hoc reasoning, X is assumed to be the cause of Y only because (you answered):

Y preceded X in time.

If Y preceded X in time, then X cannot possibly be the cause of Y, because a cause must always precede its effect in time. But that is only one criterion for establishing causation. In post-hoc reasoning, X is assumed to be the cause of Y only or mostly because it precedes Y in time, and no other criteria are considered.

1. In the fallacy of post-hoc reasoning, X is assumed to be the cause of Y only because (you answered):

X preceded Y in time.

Correct!

A cause must always precede its effect in time. But that is only one criterion for establishing causation. In post-hoc reasoning, X is assumed to be the cause of Y only or mostly because it precedes Y in time, and no other criteria are considered.

1. In the fallacy of post-hoc reasoning, X is assumed to be the cause of Y only because (you answered):

X and Y happened at the same time.

A cause must always precede its effect in time, so two things that happened at the same time cannot be the cause and effect of each other. Remember that time is only one criterion for establishing causation. In post-hoc reasoning, X is assumed to be the cause of Y only or mostly because it precedes Y in time, and no other criteria are considered.

1. In the fallacy of post-hoc reasoning, X is assumed to be the cause of Y only because (you answered):

X and Y happened at different times.

"Different times" isn't specific enough. A cause must always precede its effect in time, so X must occur before Y does, and not the other way around. Remember that time is only one criterion for establishing causation. In post-hoc reasoning, X is assumed to be the cause of Y only or mostly because it precedes Y in time, and no other criteria are considered.

2. Which of the following is most likely to be a case of post-hoc reasoning? You answered:

The lightning struck at 10:02, and the power disruption occured at 10:03, so the lightning must have caused the power disruption.

Since the lightning preceded the power disruption in time, and it is certainly possible that lightning could cause a power disruption, this doesn't seem to be fallacious reasoning--though it could, for instance, if the lightning struck in Oregon, and the power disruption was in Florida. But we have no reason to think that is the case here.

2. Which of the following is most likely to be a case of post-hoc reasoning? You answered:

Bennett was Secretry of Education when large drops in national achievement scores were recorded, so Bennett's policies must have been responsible for students across the country doing worse.

Correct!

It's hard to believe that a Secretary of Education's policies could have such quick and dramatic results; it's more likely that Bennett is being blamed for the drop just because he took office before that drop occured. In other words, this seems to be a case of saying that X caused Y only because X (Bennett as Secretary of Education) preceded Y (the drop in scores).

2. Which of the following is most likely to be a case of post-hoc reasoning? You answered:

A faulty traffic light caused the accident. I hit the gas when the light should have turned green, but it didn't turn green on time, and then another car hit mine.

It's not the light itself, faulty or not, but the speaker's anticipation of the light, which caused the accident. Indeed, since the light did not change before the accident happened, it cannot be the cause of the accident. Even if the light was faulty, if the fault did not cause the light to change prematurely, the fault cannot be blamed for the accident. So this seems to be a bad causal argument, but not an example of a post-hoc fallacy, because in the fallacy something is blamed as the cause only or mostly because it precedes the supposed effect in time, but here the cause (the traffic light) apparently does not precede the supposed effect (the traffic accident) in time.

2. Which of the following is most likely to be a case of post-hoc reasoning?

Raising the minimum wage means that the money available to pay salaries will be split up among fewer workers. So raising the minimum wage will cause employment to go down.

Economists disagree about whether this is a good causal argument or not, but even if you consider it a bad argument, you can't consider it a fallacy. In a post-hoc fallacy, one thing is assumed to have caused another only or mostly because the first precedes the second in time. And raising the minimum wage here precedes a drop in employment. But we also are given an explanation of the causal relationship between the two. After study, we may or may not think this a strong causal relationship, but there is certainly more to it than can be considered a fallacy.

Congratulations! You have finished the section on post-hoc reasoning.