5. The cause makes a difference 305
6. Overlooking a common cause 305
7. Tracing the cause backwards 306
8. Criteria for cause and effect 307
9. Two mistakes in evaluating cause and effect 308
• Exercises for Section A 310
B. Examples 310
• Exercises for Sections A and B 315
C. How to Look for the Cause 317
• Exercises for Section C 318
D. Cause and Effect in Populations 320
1. Controlled experiment: cause-to-effect 320
2. Uncontrolled experiment: cause-to-effect 321
3. Uncontrolled experiment: effect-to-cause 321
• Exercises for Section D 323
Summary 329
Maria caused the accident. Smoking causes cancer. Gravity causes the moon to stay
in orbit. These are causal claims. We make lots of them, though they may not
always contain the word "causes" or "caused." For example, "Jogging keeps you healthy" or "Taking an aspirin every other day cuts the risk of having a heart attack."
And every time someone blames you, you're encountering a claim that you caused
something that was bad and, apparently, avoidable.
What does a claim about causes look like? How do we judge whether it's true?
301
302 CHAPTER 15 Cause and Effect
A. What is the Cause?
1. Causes and effects
What exactly is a cause! Consider what Dick said last night:
Spot caused me to wake up.
Spot is the thing that somehow caused Dick to wake up. But it's not just that Spot
existed. It's what he was doing that caused Dick to wake up:
Spot's barking caused
Dick to wake up.
So barking is a cause and waking is an effect? What exactly is barking? What is
waking? The easiest way to describe the cause is to say:
Spot barked.
The easiest way to describe the effect is to say:
Dick woke up.
Whatever causes and effects are, we can describe them with claims. And we know
a lot about claims: whether they're objective or subjective, whether a sentence is too
vague to be a claim, how to judge whether an unsupported claim is true.
So now we have:
Spot barked
-> Dick woke up
caused
What is this relationship of being caused?
It has to be a very strong relationship. Once Spot barked, it had to be true that
Dick woke up. There's no way (or almost no way) for "Spot barked" to have been
true and "Dick woke up" to be false.
We know about that relationship—it's the relationship between the premises
and conclusion of a valid or strong argument. But here we're not trying to convince
anyone that the conclusion is true: We know that Dick woke up. What we can carry
over from our study of arguments is how to look for all the possibilities—all the
ways the premises could be true and the conclusion false—to determine if there is
cause and effect. But there has to be more in order to say there's cause and effect.
SECTION A What is the Cause? 303
2. The normal conditions
A lot has to be true for it to be (nearly) impossible for "Spot barked" to be true and
"Dick woke up" to be false:
Dick was sleeping soundly up to the time that Spot barked.
Spot barked at 3 a.m.
Dick doesn't normally wake up at 3 a.m.
Spot was close to where Dick was sleeping.
There was no other loud noise at the time. . . .
We could go on forever. But as with arguments, we state what we think is important
and leave out the obvious. If someone challenged us, we could add "There was no
earthquake at the time"—but we just assume things are the way they "normally" are.
Normal conditions For a causal claim, the normal conditions are the obvious
and plausible unstated claims that are needed to establish that the relationship
between purported cause and purported effect is valid or strong.
3. Particular causes, generalizations, and general causes
Spot waking Dick is a particular cause and effect. This happened once, then that
happened once.
To establish the causal claim, we have to consider all the possible ways Spot
could have barked, under the normal conditions, and ask whether Dick would have
woken up. With a physical situation like this we could even do experiments to look
at some of the possible ways the cause could be true, say, getting Spot to bark at
3:23 a.m. on a cloudless night, or getting Spot to bark at 4:18 a.m. on an overcast
night. We need that every time Spot barked, Dick woke up. There has to be a
correlation: Every time this happens, that happens. So to establish a particular
cause and effect, we might try to establish a generalization.
Alternatively, we could generalize from this particular cause and effect to any
situation like it:
Very loud barking by someone's dog near him when he is sleeping causes
him to wake, if he's not deaf.
This is a general cause and effect claim: For it to be true, lots of particular cause
and effect claims must be true. The normal conditions for this general claim won't
be specific just to the one time Spot woke Dick, but will be general. Here, too, in
trying to survey the possible ways that the cause could be true, we might want to
establish a generalization: "Anytime anyone's encountered these conditions—the
barking, the sleeper, etc.—the sleeper woke up."
304 CHAPTER 15 Cause and Effect
Exercises for Sections A.1-A.3
For each exercise here, if appropriate rewrite the sentence as a causal claim, that is, one that uses the word "causes" or "caused." If it's a particular causal claim, describe the purported cause and the purported effect with claims. Here are two of Tom's exercises.
Your teaching made me fail this class.
Causal claim: Your teaching caused me to fail this class.
Particular or general? Particular.
Cause (stated as a claim): You taught badly.
Effect (stated as a claim): I failed.
You've got the idea. But why did you say the cause was "You taught badly"?
Maybe it should be "You taught well, but didn't slow down for unprepared
students." The problem is that the original sentence is too vague.
Drinking coffee keeps people awake.
Causal claim: Drinking coffee causes people to stay awake.
Particular or general? General.
Cause (stated as a claim): People drink coffee. No.
Effect (stated as a claim): People stay awake. No
Remember that with a general causal claim there isn't a cause and effect, but lots
of them. So there's no point in fitting in after "cause" and "effect." When we try to figure out a particular causal claim that this general one covers, we see the real
problem: Maria drank coffee yesterday, Maria stayed awake. How long did she
stay awake? What would count for making this true? It's still too vague.
1. The police car's siren got me to pull over.
2. The speeding ticket Dick got made his auto insurance rate go up.
3. Speeding tickets make people's auto insurance rates go up.
4. Because you were late, we missed the beginning of the movie.
5. The onion's smell made my eyes tear.
6. Dogs make great pets.
7. I better not get the pizza with anchovies, because every time I do, I get heart
burn.
8. Someone ringing the doorbell made Spot bark.
9. Your allowing me to take the final exam a day late made it possible for me to pass.
10. Coffee keeps me from getting a headache in the afternoon.
11. Penicillin prevents serious infection.
12. If it weren't for my boyfriend, I'd have no problems.
13. My hair looked nice today until I walked outside and the wind messed it up.
SECTION A What is the Cause? 305
14. Our airplane took off from gate number thirteen. No wonder we're experiencing so
much turbulence.
15. Tom: Hey, you want to be a ball player, you have to do better than that.
Lee: It was the sun that made me drop the ball.
16. The cold makes people shiver.
4. The cause precedes the effect
We wouldn't accept that Spot's barking caused Dick to wake up if Spot began
barking only after Dick woke up. The cause has to precede the effect. That is,
"Spot barked" became true before "Dick woke up" became true.
For there to be cause and effect, the claim describing the cause has to become
true before the claim describing the effect becomes true.
5. The cause makes a difference
We often need a correlation to establish cause and effect. But a correlation alone is
not enough.
Dr. E has a desperate fear of
elephants. So he buys a special wind
chime and puts it outside his door to keep
the elephants away. He lives in Cedar
City, Utah, at 6,000 feet above sea level in
a desert, and he confidently claims that the
wind chime causes the elephants to stay
away. After all, ever since he put up the
wind chime he hasn't seen any elephants. There's a perfect correlation here: "Wind
chime up on Tuesday, no elephants," Dr. E notes in his diary.
Why are we sure the wind chime being up did not cause elephants to stay
away? Because even if there had been no wind chime, the elephants would have
stayed away. Which elephants? All elephants. The wind chime works, but so would
anything else. The wind chime doesn't make a difference.
For there to be cause and effect, it must be that if the cause hadn't occurred,
there wouldn 't be the effect. If Spot had not barked, Dick would not have woken.
Checking that the cause makes a difference is how we make sure we haven't over-
looked another possible cause.
6. Overlooking a common cause
Night causes day.
This is just wrong. There is a common cause of both "It was night" and "It's now
day," namely, "The earth is rotating relative to the sun."
306 CHAPTER 15 Cause and Effect
Dick: Zoe is irritable because she can't sleep properly.
Tom: Maybe it's because she's been drinking so much espresso that
she's irritable and can't sleep properly.
Tom hasn't shown that Dick's causal claim is false by raising the possibility of
a common cause. But he does put Dick's claim in doubt. We have to check the
other conditions for cause and effect to see which causal claim seems most likely.
7. Tracing the cause backwards
So Spot caused Dick to wake up. But Dick and Zoe's neighbor tells them that's not
right. It was because of a raccoon in her yard that shares the same fence that Spot
started barking. So really, a raccoon entering her yard caused Dick to wake up.
But it was no accident that the raccoon came into their neighbor's yard. She'd
left her trash can uncovered. So really the neighbor's not covering her trash caused
Dick to wake up.
But really, it was because Spot had knocked over her trash can and the top
wouldn't fit; so their neighbor didn't bother to cover her trash. So it was Spot's
knocking over the trash can that caused Dick to wake up.
SECTION A What is the Cause? 307
But really, . . . . This is silly. We could go backwards forever. We stop at the
first step: Spot's barking caused Dick to wake up. We stop because as we trace the
cause back further it becomes too hard to fill in the normal conditions.
Compare what happened to Dick yesterday:
Dick is just wrong. The purported cause—Spot lying next to where Dick walked—
was too far away from the effect. But what does "too far away" mean? The
astronomer is right when she says that a star shining caused the image on the
photograph, even though that star is billions of miles away and the light took billions
of years to arrive.
"Too far away in space and time" is just a sloppy way to say that we can't see
how to fill in the normal conditions, the other claims that would make it obvious that
it's (nearly) impossible for the claim describing the cause to be true and effect false.
8. Criteria for cause and effect
We can collect everything we've learned about cause and effect so far. These are
necessary conditions for there to be cause and effect, once we describe the cause
and effect with claims.
Necessary criteria for cause and effect
• The cause happened (the claim describing it is true).
• The effect happened (the claim describing it is true).
• The cause precedes the effect.
• It is (nearly) impossible for the cause to happen (be true) and
the effect not to happen (be false), given the normal conditions.
• The cause makes a difference—if the cause had not happened
(been true), the effect would not have happened (been true).
• There is no common cause.
308 CHAPTER 15 Cause and Effect
9. Two mistakes in evaluating cause and effect
a. Reversing cause and effect
Consider what Tom said after the demonstration in front of the post office:
Tom: That ecology group is twisting their members' minds around.
Dick: Huh?
Tom: They're all spouting off about the project to log the forest on
Cedar Mountain. All in lockstep. What do they do to those guys?
Tom's got it backwards. Joining the group doesn't cause the members to become
concerned about the logging on Cedar Mountain. People who are already concerned
about ecological issues join the group. He's reversing cause and effect.
Suzy: Sitting too close to the TV ruins your eyesight.
Zoe: How do you know?
Suzy: Well, two of my grade school friends used to sit really close,
and both of them wear really thick glasses now.
Zoe: Maybe they sat so close because they had bad eyesight.
Even if Suzy had a huge sample instead of just anecdotal evidence, it would be
just as plausible to reverse the cause and effect. That doesn't mean Suzy's claim is
false. It just shows we have no good reason to believe that sitting too close to the TV
ruins your eyesight.
b. Looking too hard for a cause
Every Tuesday and Thursday at 1:55 p.m. a tall red-headed lady walks by the door of
Professor Zzzyzzx's classroom. Then he arrives right at 2 p.m. When Suzy says the
lady walking by the door causes Professor Zzzyzzx to arrive on time at his class,
she's jumping to a conclusion: It happened after, so that's the cause. We call that
kind of reasoning post hoc ergo propter hoc (after this, therefore because of this).
Zoe belched loudly in the shower with the bathroom window open, and she and
Dick haven't seen Spot since. He
must have run away because she belched.
That's just post hoc ergo propter hoc. A possible cause is being overlooked:
SECTION A What is the Cause? 309
Perhaps someone left the gate open, or someone let Spot out, or . . . . Post hoc
reasoning is just not being careful to check that it's (nearly) impossible for the cause
to be true and effect false. Jumping to conclusions about causes isn't a sign of a rich
imagination. ("Gee, I'd never have thought the red-haired lady caused Professor
Zzzyzzx to arrive on time.") It's a sign of an impoverished imagination.
We look for causes because we want to understand, to explain, so we can
control our future. But sometimes the best we can say is that it's coincidence.
Before your jaw drops open in amazement when a friend tells you that a piano
fell on her teacher the day after she dreamt that she saw him in a recital, remember
the law of large numbers: If it's possible, given long enough, it'll happen. After all,
most of us dream—say one dream a night for fifty million adults in the U.S. That's
three hundred and fifty million dreams per week. With the elasticity in interpreting
dreams and what constitutes a "dream coming true," it would be amazing if a lot of
dreams didn't "accurately predict the future."
But doesn't everything have a cause? Shouldn't we look for it? For much that
happens in our lives we won't be able to figure out the cause—we just don't know
enough. We must, normally, ascribe lots of happenings to chance, to coincidence, or
else we have paranoia and end up paying a lot of money to psychics.
Suppose two million Parisians were paired off and set to tossing coins in a game
of matching. Each pair plays until the winner on the first toss is again brought tc
equality with the other player. Assuming one toss per second for each eight-hou
day, at the end often years there would still be, on the average, about a hundred-
odd pairs; and if the players assign the game to their heirs, a dozen or so will stil
be playing at the end of a thousand years! The implications are obvious. Suppo;
that some business had been operating for one hundred years. Should one rule o
luck and chance as the essence of the factors producing the long-term survival ol
the enterprise? No inference whatever can be drawn until the number of original
participants is known; and even then one must know the size, risk, and frequency
Richard L Epstein Page 39
