10

The “if/then” relationship of this statement suggests that it is setting up causal reasoning. While such a statement can be persuasive, the question at issue is: Is it sound reasoning?

Sometimes, the speaker is only citing a correlation between event A and event B, and on that basis alone expects one to believe that A causes B. But correlation does NOT establish a true cause-effect relationship. There may well be a relationship between two events, but there are other relationships that can explain correlation besides a causal one. So, reliable causal reasoning must demonstrate that the relationship between the two events is indeed a causal one and rule out alternative explanations.

The main task of a causal argument then, is to establish a causal hypothesis by distinguishing causally related events from mere correlations.

Post Hoc Ergo Propter Hoc

The Post Hoc Ergo Propter Hoc fallacy occurs in a causal argument when the only support given for the conclusion that event X causes Y is the fact that X preceded Y. The fallacious nature of this reasoning is seen easily enough where this is an isolated conjunction of events. However, even were the two events seen to be repeatedly conjoined, this is still fallacious reasoning. That is because there are (at least) three other possibilities for the relationship between the correlated events. (See below.)

Further, it is often appropriate to soften the bold claim that “A caused B” to the more modest claim “A influenced B.”

Causal thinking is often used when looking to blame something or someone, as can be seen in the following example:

It’s the governor’s fault that the economy hasn’t recovered more.

But here we must consider the nature of the overall phenomenon for which the speaker is trying to provide a causal explanation. While such a statement may garner a speaker some political capital, it is not based on solid reasoning. Economic and political processes are too complex to distill to such a simple cause-effect relationship. A speaker would need to use more solid reasoning, perhaps inductive reasoning through examples, to build up enough evidence to support that a correlation exists and a causal relationship is likely.[1]

Causal Argument Structures

Causal arguments usually have the following structure:

1. A correlation premise that claims the existence of a correlation between two events or states of affairs X and Y. This typically involves

a. temporal sequence (i.e. X before Y) and

b. constant conjunction (i.e. X always followed by Y. Y never unpreceded by X)

2. And

a. a plausible causation story

and/or

b. an elimination premise or premise set which seeks to prove that X causes Y by seeking to exclude the other possible relationships between X and Y that could account for the observed temporal sequence and constant conjunction.

There are three main alternative explanations that the elimination premise or premise set must address:

1. That X and Y have a common cause.

If there is some event “W” which is the cause of BOTH X and Y, then the constant conjunction can be explained without X being the cause of Y.

2. That the correlation is simply coincidental.

Sometimes, despite a high degree of correlation the proposed causal relation is so implausible as to make coincidence more reasonable. And the less the degree the correlation, the MORE reasonable attributing it to mere coincidence becomes.

3. That Y is actually the cause of X and the proposed causal relation is exactly backwards.

(e.g. Did the professor being impressed by the student’s performance on the exam (X) cause the professor to believe that the student was unusually bright (Y)? Or did the professors belief that the student was unusually bright (Y) cause the professor to be impressed by the student’s performance on the exam (X)?)

Plausible Causation Story (Three General Forms)

A plausible causation story is a plausible explanation of the specific relationship between X and Y. Plausible causation stories can be used to substantiate the idea that X causes Y. This alone may be so plausible that it is sufficient to advance the causal explanation being offered, even without the elimination premises.

For example, we could use a plausible causation story to argue that “Exercise prevents colds.”

1. Exercise stimulates the immune system.

2. Stimulated immune systems are better equipped to resist getting a cold.

Therefore:

3. Exercising regularly (to the extent necessary to stimulate the immune system) will prevent one from getting a cold.

Or one can also use plausible causation stories to argue for a part of the elimination premise or premise set..

Causal Argument Form 1

The first type of causal argument tries to establish a singular causal claim, either directly (via plausible causation story or causal study results) or by using an elimination premise (which itself can be supported via plausible causation story or causal study results).

1. Premise 1: X is correlated with Y (This must be backed up by data.)

2. Premise 2: Either

a. a plausible causation story or

b. an elimination premise of the form:

1. it is likely that X and Y do not have a common cause

2. it is likely that X and Y are not coincidental

3. Y does not cause X.

3. Conclusion: Therefore, X causes Y (a singular causal statement).

Causal Argument Form 2

A second type of causal argument reasons from general causal statements to singular causal statements.

1. Premise 1: It is known (generally established and accepted) that events of kind X tend to cause events of kind Y (a general causal statement).

2. Premise 2: the case in question is case of an event X and an event Y.

3. Conclusion: Therefore, in this instance, event X causes event Y (a singular causal statement).

Causal Argument Form 3

A third type of is actually an “anti” causal argument which attempts to show that X does NOT cause Y, by arguing for a different relationship between the correlated events X and Y.

1. Premise 1: X is correlated with Y

2. Premise 2: it is likely that either

a. X and Y have a common cause

b. X and Y are coincidental

c. or Y causes X.

3. Conclusion: Therefore, X does NOT cause Y (a singular causal statement).

Note that a plausible causation story can play a role in this type of argument as well. One might, for example, argue that X and Y have a common cause by presenting a plausible causation story. For instance, one might argue that depression (X) and diabetes (Y) have a common cause by explaining how both might be caused by a certain genetic trait combined with environmental factors.

I) Causal Statements vs. Causal Arguments

A Causal Statement asserts the cause of some event.

The floor is wet because the toilet is leaking. (e.g. The Leaking toilet is the cause of the wet floor.)

Alternatively,

Because the toilet is leaking, the floor is wet.

What is the effect being explained (explanandum)? (The wet floor.)

What is the purported cause (explanans)? (The leaking toilet.)

A Causal Argument is an effort to persuade one to accept a purported cause for some event.

I know that the toilet is leaking because the floor is wet.

While this too is a single sentence, it differs from the first in that it's offering a conclusion (That the toilet is leaking) as the best explanation for some fact in evidence (that the floor is wet). Thus it invites us to infer an explanans from the explanandum.

The first example (A) simply states the cause of the wet floor. As such, it is a bold, unsupported assertion.

The second example (B) gives evidence for the claim that the toilet is leaking. It is an argument inviting us to conclude that the toilet is leaking based on the evidence of the wet floor and positing a leaking toilet as the cause.

II) Establishing the Truth of Causal Statements:

Causal arguments are offered to establish the truth of causal claims. Here is an example of such an argument:

Causal Claim: The floor is wet because the toilet is leaking.

Example of a Causal Argument to Establishing the Causal Claim:

Premises:

· We turned off the toilet, dried the floor, and waited.

· The floor stayed dry.

· Then we turned the toilet back on and waited.

· Then, there developed a puddle on the floor.

Therefore:

The conclusion of this argument (C) is a causal statement. The rest are the reasons/premises offered for thinking the causal statement is true.

III) Forming Causal Hypotheses

A statement that X causes or caused Y is often a hypothesis. A causal hypothesis is a tentative claim, a statement offered as true, but nevertheless inviting further investigation or testing. Normally, when we are concerned with the cause of something, our reasoning falls into two parts:

1. forming an hypothesis and

2. confirming the hypothesis.

These are distinct activities (though they involve overlapping principles).

Brief Aside: Karl Popper

This is similar to the divide we see in scientific inductive reasoning between the "context of discovery" and the “context of justification.” Philosopher of science, Karl Popper (1902 – 1994) argues that science progresses by means of formulating bold hypotheses, and then trying to refute (disprove or falsify) them.

Popper believed that:

"Bold ideas, unjustified anticipations, and speculative thought, are our only means for interpreting nature: our only organon, our only instrument, for grasping her."[2]

He makes this point more specific in a 1953 lecture, where he argues that, if we aim to explain the world, then:

"... there is no more rational procedure than the method of trial and error – of conjecture and refutation: of boldly proposing theories; of trying our best to show that these are erroneous; and of accepting them tentatively if our critical efforts are unsuccessful. From the point of view here developed, all laws, all theories, remain essentially tentative, or conjectural, or hypothetical, even when we feel unable to doubt them any longer."[3]

Popper emphasized the virtue of “bold” hypotheses. A "bold" hypothesis is a new scientific idea which, if it was true, would be able to predict and/or explain a lot, or a lot more, about the subject being theorized about. The "boldness" of a hypothesis depends mainly on:

· Its scope – the number and variety of phenomena which it could explain, if it is true (its "explanatory power").

· Its novelty or originality – the extent to which the hypothesis is a genuinely new departure from the received scientific ideas.

· Whether it enables new and novel predictions ("predictive power").

· Whether it stimulates new, innovative research ("heuristic power").

· Its degree of applicability or usefulness for scientific research ("utility").

· The effect or impact it has on existing scientific thinking, if it is true.

Once a bold hypothesis has been offered, Popper argues, scientists try to investigate and test how well the bold hypothesis can stand up to the known evidence, with the aim of finding counter-arguments and counter examples which would refute or falsify the bold hypothesis. In this process of testing and criticism, new scientific knowledge is generated. Even if the bold hypothesis turns out to have been wrong, the testing itself generates new knowledge about what can and cannot be the case. Often it stimulates new research.

Inversely, if a hypothesis lacks the quality of boldness, then it would make very little difference to what scientists already know. It is not "a big deal", i.e. it is not very significant for the theory which exists already. It can contribute rather little to advancing scientific progress, because it does not expand or add to scientific understanding very much.

Nevertheless, there is a potential tension between Popper’s insistence on bold hypotheses in science and how we normally think in causal reasoning. For instance, when trying to explain the wet floor it is more appropriate to consider a potentially leaking toilet then aliens from outer space beaming down water using their transporter device from a spaceship in orbit. In everyday causal explanations, one generally has a set of explanatory resources immediately at hand. These are the first to be considered when forming a hypothesis. This is similar to the principle entitled “Occam's Razor[4].” This is a principle of theory construction or evaluation according to which, other things equal, explanations that posit fewer entities, or fewer kinds of entities, are to be preferred to explanations that posit more. This is also referred to the theoretical virtue of “Economy.”

Back to Forming Causal Explanations

First we need to identify the thing to be explained (the explanandum). The problem must be clearly stated.

“The car won't start.”

Then come up with an explanation (the explanans):

1. Form hypotheses: (i.e. Think of most likely possible causes- context of discovery – bold conjecture… or maybe not so much)

2. Confirm hypotheses. (i.e. test – context of justification)

III (a-c) Three Principles used to arrive at causal hypotheses:

III a) Forming Causal Hypotheses: Paired Unusual Events Principle

The "paired unusual events principle" of arriving at a causal hypothesis is pretty straightforward: If something unusual happens, look for something else unusual that has happened and consider whether it might be the cause.

If you wake up one morning with an unusual splitting headache (one such as you are NOT normally prone to getting), and you remember doing something unusual the night before that might be a cause, such as reading in poor light or having a bit too much to drink, you hypothesize that it has something to do with the headache.

Here is another example of how a pairing of unusual events can suggest a causal hypothesis:

As soon as my throat got scratchy I took Zicam. My sore throat went away and I never caught a cold.

Therefore:

Maybe Zicam prevents colds.

Now, the fact that two unusual things happened around the same time is, at best, only grounds for hypothesizing causation. It does not establish causation.

The previous argument does NOT establish that Zicam prevents colds or kept you from catching one. You don't know what would have happened if you hadn't taken Zicam. Take out the word "maybe" and you have the fallacy known as post hoc, ergo propter hoc.

To arrive at causal hypotheses, one uses common sense and background knowledge of what causes what and how things work. Yes, if your scratchy throat goes away, you might recall that a raccoon had crossed your path before that happened. But it isn't plausible to think that a raccoon crossing your path could make a sore throat go away. Why isn't it plausible? Because, given normal experience and common sense, one cannot see how a raccoon crossing one's path could make a sore throat go away. One cannot conceive of a mechanism by which this might happen.

Now, of course, it may be that there is an hitherto undiscovered relationship between raccoons crossing one's path and scratchy throats abating. But this is not part of our common sense knowledge at this time. So, this would not be where to begin one’s causal hypothesizing. Other more standard explanations would need to be ruled out first. Extravagant hypotheses would only be worth considering if less extravagant alternative hypotheses had been ruled out. There is an old adage in medical diagnosis that if you hear hooves, think horses not zebras. (Although that may not apply if one is in Africa.)

III b) Forming Causal Hypotheses: Common Variable Principle

The second principle for arriving at a causal hypothesis is also straightforward: A variable common to multiple occurrences of something may be related to it causally.

For example:

To be explained (explanandum):

Several people in Kearney complained of intestinal distress.

Possible Causal Hypothesis (explanans):

When several people in Kearney complained to their physicians about acute intestinal distress, health officials investigated and found that all of them had eaten tacos at the county fair.

Therefore:

Perhaps the tacos caused the distress.

Why consider this? Because each of the instances of a resident experiencing intestinal distress was also preceded by an instance of the same resident eating tacos at the County Fair. It would of course weaken the inference if only some of the residents who became ill had eaten the tacos and/or if other residents who ate the tacos were unaffected. The closer the covariance the stronger the inference.

Covariance: measures the direction of the relationship between two variables. A positive covariance means that both variables tend to be high or low at the same time. A negative covariance means that when one variable is high, the other tends to be low.

A high covariance of eating the county fair tacos and subsequently experiencing intestinal distress supports the hypothesis that tacos caused the intestinal problems suffered by Kearney residents. The conjecture might then be confirmed or disconfirmed by other means, such as testing the tacos for salmonella bacteria.

Here is another example of how identification of a common variable in multiple occurrences of something suggests a causal hypothesis:

Some years the azaleas bloomed well; in other years they didn't bloom at all. In the years they didn't bloom, I fertilized heavily, but I didn't do that in the years they bloomed well.

Therefore:

Perhaps heavy fertilizing caused them not to bloom well.

Here, the variable common to the years the azaleas didn't bloom is also said to be absent in the years they did bloom. It is reasonable to hypothesize that heavy fertilizing prevented the azaleas from blooming.

III c) Forming Causal Hypotheses: Double Covariation Principle

The third principle for arriving at a causal hypothesis works like this. When a variation in one phenomenon is accompanied by a variation in another phenomenon, we have a covariation or correlation. For example, if crime rates go up as gun sales increase, the two increases are correlated. And likewise if crime decreases at the same time as gun sales decrease they are doubly corelated. Of course, correlations don't prove causation. Indeed, even if these two variables are doubly covarient, this alone would not indicate which one causes the other, if either. It may be that rising crime rates cause people to engage in gun purchasing or it could be that engaging in gun purchasing causes people (enable people) to engage in more criminal behavior. And, of course it could simply be coincidental. Both hypotheses would need to be tested further.

Nevertheless, when guided by common sense and background knowledge of what causes what, correlations can offer a reason for hypothesizing causation may exist, as in this example. As with the earlier causal arguments we have looked at, this is only a tentative causal hypothesis standing in need a further testing.

· Over the past few years, online instruction has increased at San Diego State.

· During the same time the average GPA of San Diego State students has increased.

Therefore:

(Maybe) The increase in GPA may be due to the increase in online instruction.

The correlation does not "prove" causation, but it suggests a hypothesis.

Another example:

· When meat consumption in Holland went up after the Second World War, so did the rate of prostate cancer in that country.

Therefore:

(Perhaps) Eating meat causes prostate cancer.

Again, the information doesn't establish that meat consumption in Holland after the Second World War caused the increase in prostate cancer there, but it warrants that hypothesis. That is, it suggests a cause and effect connection.

To repeat, correlation does not prove causation.

The cause(s) suggested by correlation are only possible causal links and the degree to which the correlation should be taken seriously as suggesting a causal link will rely on the background beliefs available to us.

· A girl's hair grows longer as she learns the multiplication table. But this correlation suggests no causal link between the two things.

However, from our background knowledge, we can see how meat consumption might be related to prostate cancer, but we cannot see how hair length could have anything to do with learning arithmetic.

Sometimes background “common sense” beliefs which would have us dismiss a correlation as merely coincidental have turned out to be wrong. For instance, it was known that students with demonstrable musical talents often perform better in mathematics. But no one quite could imagine what causal connection there might be to explain the correlation. However, eventually further research was done to show that music instruction actually improves individuals mathematical skills. There does seem to be a causal link, but even now that cause a link is not well understood.

Correlation Due to Common Cause

One frequently seen situation where we find a high degree of correlation despite a lack of direct causal connection is when both series of events are the result of a single, common cause.

· Ex: Skiing accidents increase as Christmas sales pick up.

But we have no reason to suppose that there is a causal connection here between Christmas sales and skiing accidents. Why? Well, because a more reasonable explanation for the correlation is a “common cause” that is, the onset of winter explains the approach of Christmas holidays and the increase in skiing vacations with an increase in the opportunities for skiing accidents.

Another example of common cause:

There is a positive correlation between having nicotine-stained fingers and developing throat cancer. But it would be incorrect to assume that the nicotine-stained fingers is what causes the throat cancer. A more reasonable explanation is a common cause. The cause of the nicotine-stained fingers (i.e. smoking cigars) is also the cause of the throat cancer.

So, the principles we have discussed for developing causal hypotheses require “common sense” and preexisting knowledge of what causes what and how things work. But also creative imagination.

IV Confirming Causal Hypotheses

Confirmation of many cause-and-effect hypotheses consists in trying to show that the hypothesized cause is the condition "but for which" the effect in question would not happen. A condition but for which the effect would not happen is expressed in the Latin legal phrase conditio sine qua non ("a condition without which not").

For example:

· We dried the floor, turned off the toilet, and waited. The floor stayed dry. (No running toilet/ No Wet Floor)

· Then we turned the toilet back on and watched. Now there is a puddle on the floor. (Running toilet/ Wet Floor)

Therefore:

The turned-on toilet is a conditio sine qua non and thus the leaking toilet caused the puddle.

The hypothesized cause is the leaking toilet. The effect in question is the wet floor. The argument gives a reason for thinking that, but for the toilet, the floor would have stayed dry.

Now, a critic might complain that the toilet would not have been leaking but for the fact that…

1. there is water in the toilet and

2. there wouldn't be water in the toilet, but for the fact that water exists on Earth and

3. there wouldn't be water on Earth, but for the fact that Earth is just the right distance from the sun to have water.

Is it proper to say that the distance of the Earth to the Sun is the cause of the wet floor?

No. Don’t lose sight of “the issue.” We were concerned only with whether the puddle was caused by the leaking toilet. Since that question has nothing to do with these other things, we logically can ignore them. This is related to the legal principle of the “last proximate cause.”

Last Proximate Cause

In law and insurance, a proximate cause is an event sufficiently related to an injury that the court will determine to be “the event” that is “the cause” of that injury. There are two types of causation in law:

1. Cause-in-fact

2. Proximate (or legal) cause

Cause-in-fact is determined by the "but for" test discussed above. But for the action, the result would not have happened. For example, but for the defendant’s car running the red light, the collision would not have occurred. The action of running the light is a necessary condition for the accident occurring. But it was not a sufficient condition for the resulting collision and injury. In fact, very few circumstances exist where the “but for” test is all that is needed to establish causality in a legally relevant sense. Note in this case there might be many extraneous factors that “but for” the event would not have happened. Had the accident victim simply slept in that morning, he would not have been in the intersection. But for the victim waking up on time, the accident could not have occurred. While this shows the victim’s waking on time may have been necessary, it does not show it to be sufficient. So, more needs to be said then that the factor was necessary (but for) when trying to establish “the cause” of the accident, especially for legal purposes.

Therefore, a second test is used to determine if a certain “cause-in-fact” is close enough to a harm in a "chain of events" to be legally valid. This test is called proximate cause. Proximate cause is a key principle in law and matters of insurance and is concerned with how the loss or damage actually occurred. There are several competing theories of proximate cause, and these can be quite controversial. (If I am the last one to sell a drunk driver gasoline, am I a proximate cause and thus legally responsible for the subsequent accident?) But we need not go into details here. Nevertheless, for an act to be deemed to be “the” cause of a harm, both tests must be met. “Proximate cause” then is a legal limitation on a “cause-in-fact” (a necessary but not sufficient cause covered by the but for test).

V) Weighing Evidence

Coming up with a causal hypothesis involves weighing evidence.

1. Unusual Occurrence that needs to be given a causal explanations:

The car isn't starting. Why?

2. Other Unusual Occurrences.

Well, we heard funny clicking sounds when we tried to start it, the kind associated with having a battery that is almost dead.

We also noticed gas fumes, like when an engine is flooded.

We noticed other anomalies: we had just filled up with a new brand of gas; the steering wheel won't unlock; it is unusually cold out; and so forth. We had just installed a new radio, too.

In real life, forming a hypothesis is not as simple as the preceding three principles suggest. We have to weigh things.

For example, the association between clicking sounds and almost-dead batteries is more significant than the fact that the steering wheel won't unlock, or even that gasoline could be smelled.

The smell of gasoline, which often accompanies engine flooding, might be explained by the fact we just filled up.

We'd check battery connections and hope for the best.

Here is another example of weighing evidence:

You go to a physician about numbness in a leg. The doctor asks a series of questions.

1. Exactly where in the leg is the numbness?

2. When did it begin?

3. Did it begin suddenly?

4. Did something unusual happen?

5. Is it worse at some times of the day?

6. Do you experience it in the other leg?

7. Does it depend on your activities or the position of the leg?

8. Do you smoke?

9. Do you have high blood pressure?

10. Are you experiencing other unusual symptoms?

The point of the doctor asking you these questions is the determine whether or not there are other unusual circumstances or events or whether there is potentially relevant correlations to circumstances and events. The investigation may disclose additional symptoms. Some of them might be associated with a neurological condition, another with an orthopedic condition, perhaps another is a psychiatric condition, and so forth.

The physician considers which symptoms are most important, and diagnoses your condition accordingly.

The diagnosis is the physician's causal hypothesis. She doesn't arrive at it through any straightforward, formulaic strict application of the previous three principles. She is, however, looking for associations and correlations between symptoms and medical conditions, and she is looking for unusual events that might have accompanied the onset of numbness. She is also drawing upon her general knowledge of medicine and past experience as her source of explanatory resources.

"Did something unusual happen?" is a test to see if the “Paired Unusual Events Principle” applies. She, being a physician, is in the best position to gauge the comparative significance of our answers.

VI) Randomized Controlled Experiments

Here is another example of reasoning that tries to show that something wouldn't have happened, but for the hypothesized cause.

In an experiment, 50 willing volunteers were infected with a cold virus and then randomly divided into two groups. The subjects in one group were given a Zicam treatment, as per the instructions on the box. Two weeks later, the number of people with colds in both groups was compared. Eighteen of the subjects who had not been treated with Zicam had colds, and only 10 of the subjects who had been treated had colds, a difference that is statistically significant.

Therefore, Zicam probably reduced the duration of colds in the experimental group.

VII) Four Types of Causal Arguments:

Causal arguments have conclusions alleging that one thing does or does NOT cause another. This might be specific to a particular incident or case or it might be general to a population. So, besides being positive (argument to confirm) or negative (argument to disconfirm), the conclusion may be about a specific case or about what generally happens. The first sort is a cause of a particular case and the second sort is a cause in a population. (Technically, arguments about populations will be a special type of argument that generalizes.) Therefore, there are four types of causal arguments:

1. Argument to confirm a particular cause.

2. Argument to disconfirm a particular cause.

3. Argument to confirm a cause in a population.

4. Argument to disconfirm a cause in a population.

1. Argument to confirm a particular cause.

Example conclusion: The alarm clock woke Pat.

2. Argument to disconfirm a particular cause.

Example conclusion: The alarm clock did not wake Pat.

3. Argument to confirm a cause in a population.

Example conclusion: Smoking (X) is a cause of lung cancer (Y).

Standard form for this type:

1. X is correlated with Y.

2. Elimination Premises

The evidence is from a controlled/uncontrolled study, thus eliminating that, they share a common cause, they are only coincidentally correlates and that cancer is not the cause of the smoking.

Therefore:

3. X is probably a cause of Y.

4. Argument to disconfirm a cause in a population.

Example conclusion: Regular exercise is not a cause of lung cancer.

Standard form for this type:

1. X is not correlated with Y.

2. Elimination Premises

The evidence is from a controlled/uncontrolled study.

3. X is probably NOT a cause of Y.

VIII) Evaluation of Causal Arguments:

As with any other inductive argument, first determine whether the premises are true. Since correlations are demonstrated through sampling and generalizing, we should not accept the claim of correlation until we agree that the process of generalizing would be sound. In other words, we must evaluate the claim of a correlation as being an inductive generalization from samples.

But agreeing to the correlation is not enough to make the causal conclusion sound. By itself, a correlation never proves a cause. (Example: The craze for pet rocks may have coincided with the peak year for sales of disco music, but nobody thinks that disco causes pet rocks or that pet rocks cause disco music.) The arguer must take care to rule out the causal fallacies that often lead one to postulate a cause when there's nothing but a correlation.

Causal Fallacies

Four fallacies pose problems for arguments that try to demonstrate a cause in a population. The presence of any of these four problems will make a causal argument weak.

The four are:

5. The fallacy of Post Hoc

6. The fallacy of overlooking a common cause

7. The fallacy of coincidental correlation

8. The fallacy of reversing cause and effect

In Sum:

When using causal reasoning, the critical thinker must present evidence that shows the following:

1. the cause occurred before the effect

2. the cause led to the effect

3. it is unlikely that other causes produced the effect.

Review of Types of Reasoning

Inductive: Arguing from examples to support a conclusion; includes reasoning by analogy. Examples should be sufficient, typical, and representative to warrant a strong argument.

Deductive: Deriving specifics from what is already known; includes syllogisms. Premises that lead to a conclusion must be true, relevant, and related for the argument to be valid.

Causal: Argues to establish a relationship between a cause and an effect. Usually involves a correlation rather than a true causal relationship.

Epilogue: What is differential diagnosis?

Differential diagnosis is a process wherein a doctor differentiates between two or more conditions that could be the cause of a person’s symptoms. When making a diagnosis, a doctor may have a single theory as to the cause of a person’s symptoms. In this case, the doctor would then then order tests to confirm the suspected diagnosis. However, in practice, there is rarely one single laboratory test that can definitively diagnose the cause of a person’s symptoms. This is because many conditions share the same or similar symptoms, and some diseases present in a variety of different ways. To make a diagnosis, a doctor may need to use a technique called differential diagnosis.

What is differential diagnosis?

A doctor may perform a differential diagnosis when there is no single laboratory test to diagnose the cause of a person’s symptoms. Differential diagnosis involves making a list of possible causes of a person’s symptoms. The doctor will base this list on information they gain from:

· the person’s medical history, including their self-reported symptoms

· the results of a physical examination

· (initial) diagnostic testing

o This last requires that the doctor has at least one theory about the cause of a person’s symptoms and test for that one condition.

However, many direct causes of disease share some of the same or similar symptoms. This makes distinguishing among the possible causes the actual cause difficult to diagnose. To this end the doctor will utilize a “nondifferential diagnostic approach.”

A differential diagnostic approach is helpful when there may be multiple potential causes to consider.

The goals of differential diagnosis are to:

· narrow down the working diagnosis

· guide medical evaluation and treatment

· rule out life threatening or time critical conditions

· enable the doctor to make the correct diagnosis

What are the steps?

Differential diagnosis can take time. For a doctor to determine the correct diagnosis, they will follow the steps below.

1. Take a medical history

When preparing for differential diagnosis, a doctor will need to take a person’s full medical history. Some questions they may ask include:

· What are your symptoms?

· How long have you been experiencing symptoms?

· Do you have a family history of certain conditions?

· Have you traveled out of the country recently?

Naturally, it is important that a person answers all questions honestly and in as much detail as possible.

2. Perform a physical exam

Next, a doctor will want to perform a basic medical examination. The examination may include the following:

· taking the person’s heart rate

· taking their blood pressure

· listening to their lungs or examining other areas of the body from which symptoms may be originating

· Conduct diagnostic tests

After taking a medical history and performing a physical examination, the doctor may have some ideas as to what may be causing a person’s symptoms.

3. The doctor may order one or more diagnostic tests to rule out certain conditions. Such tests may include:

· blood tests

· urine tests

· diagnostic imaging tests, such as:

o ultrasound scan

o X-ray

o MRI scan

o CT scan

o endoscopy

4. Send the person for referrals or consultations

In some cases, the doctor may feel that they do not have the specific expertise to diagnose the exact cause of a person’s symptoms. In such cases, they may refer the person to a specialist for a second opinion. It is not uncommon for multiple doctors to review one patient during differential diagnosis.

Examples of differential diagnoses

Below is an example of common differential diagnosis process.

Symptom: Chest pain

Chest pain is a symptom that can have many causes. Some are relatively mild, whereas others are serious and require immediate medical attention. If a person is experiencing chest pain, a doctor will need to ask questions to determine certain factors, such as the location, severity, and frequency of the pain.

These questions may include the following:

· How do you feel? Describe the sensation.

· Where does it hurt?

· Does the pain extend to any other part of your body?

· Did anything trigger the pain?

· How long has the pain lasted?

· Has anything made the pain better or worse?

· Have you experienced any other symptoms?

· Have you experienced these symptoms before? If so when and how frequently?

By asking these questions, the doctor will hopefully be able to categorize the chest pain as one of the following types:

· Cardiac: These conditions relate to the heart. Examples include unstable angina and heart attack.

· Pulmonary: These conditions relate to the lungs. Examples include:

o pulmonary embolism

o pulmonary hypertension

o pneumonia

· Gastrointestinal: These conditions relate to the digestive system. Examples include gastroesophageal reflux disease, which can lead to Barrett’s esophagus, and peptic ulcers.

· Musculoskeletal: These conditions relate to the muscles, bones, and connective tissues. Examples include fractured ribs and other trauma to the chest wall or sternum.

· Miscellaneous: This category describes other potential causes of chest pain, such as:

o anxiety

o panic attacks

o lymphoma

Once the doctor has narrowed down the type of pain, they will order diagnostic tests to determine the potential cause of the pain. These tests may include:

· electrocardiogram (EKG)

· echocardiogram (echo)

· endoscopy

· X-ray

Differential Diagnoses: Summary

Notice as with other forms of causal inference and scientific inquiry the doctor begins in the imaginative and creative phase by hypothesizing possible causes and then proceeds through a methodological elimination process of possible candidates via the process of falsification.

If P then Q

Therefore:

However, it is important the doctor not start with too many (fanciful) possible options, but start with the most likely contenders. This is not only to conserve energy and resources, but also time. It is very possible that the condition needs to be addressed promptly. Too much time spent eliminating “wild goose chases” is not in the interest of the patient. Nevertheless, that is an art, rather than a formulaic process, sometimes requiring hunches and intuitions. Still, this remains a sort of process of elimination such that the last man standing is the likely cause of the symptoms. This is again an instance of inference to the best explanation.

“When you have eliminated all which is impossible, then whatever remains, however improbable, must be the truth.” ~ Arthur Conan Doyle, The Case-Book of Sherlock Holmes

[1] Causal Arguments attempt to establish a relationship between a cause and an effect. These kinds of arguments are very important if I am recommending a particular course of action based on what I allege will be the positive consequences that would result. Likewise, if I am arguing that a course of action is unwise and should be avoided because of the negative consequences, it will be necessary for me to establish that the action would cause certain (undesirable) effects. A little later in this course we will be looking at moral reasoning. One kind of moral reason is a “consequentialist reason.” It suggests that we ought or ought not to do something based on the causal results of that behavior in question. That is, it suggests we ought to do things which have good consequences and avoid things which have bad consequences. But these are most often empirical causal claims and so for the moral argument to be successful, in addition to the moral argument itself, the proponent must provide the causal argument which supports the consequentialist reasoning.

[2] Popper, Karl, The Logic of Scientific Discovery. London: Routledge, 1992, p. 280.

[3] Popper, Karl, British Council lecture given at Peterhouse, Cambridge, in Summer 1953. Published under the title "Philosophy of Science: a Personal Report" in C. A. Mace (ed.), British Philosophy in Mid-Century: a Cambridge Symposium. London: Allen & Unwin, 1966.

[4] The idea is frequently attributed to English Franciscan friar William of Ockham (c. 1287–1347), a Scholastic philosopher and theologian, although he never actually frames it this way.