Site MapHelpFeedbackChapter Overview
Chapter Overview
(See related pages)

Two kinds of explanation are addressed in Chapter 11: physical causal and behavioral causal. When physical phenomena need explaining, the hypothetical causes are drawn from a physical background, e.g. "The reason the paper is on the lawn is because the wind blew it there." When behavioral phenomena need explaining, the hypothetical causes are drawn from a psychological background, e.g. "The reason she went to the Dairy Barn was because she desired ice cream."

            An explanation's adequacy is relative to what one is looking for. Nevertheless, it shouldn't be unnecessarily complicated, circular, inconsistent, incompatible with fact or theory, vague, or untestable in principle. It shouldn't generate meaningless predictions, false predictions, or no predictions at all.

            The general strategy for forming causal hypotheses (or educated guesses) is called "inference to the best explanation." Observing an association, or co-variation, between two events can serve as a beginning. Inferring a causal connection, however, requires rigorous application of the Methods of Agreement and Difference while being guided by background knowledge of causal mechanisms. Finding a hypothesis that adequately explains the facts can be like diagnosing a disease or solving a crime. The Best Diagnosis Method gathers as many "symptoms" as possible, tries to sift out the irrelevant ones, and tries to find the strongest connections.

            Controlled cause-to-effect experiments are the most direct way of confirming a causal hypothesis. By repeating an experiment and systematically eliminating other possible causes and getting the same effect, the hypothesis becomes confirmed. Indirect methods of testing causal hypotheses are more appropriate for human populations, for practical and ethical reasons. These studies compare a group of people who exhibit the effect under investigation with a control group who do not have it. Animal experiments are another way to avoid testing humans directly. The results may be applied to humans by analogical reasoning.

            There are many, and varied, ways of making mistakes in causal reasoning. Some of the most prominent involve believing there is a causal connection between A and B when actually the relationship is: coincidental, a result of a third underlying cause, or reversed.

            The law relies on establishing a causal relationship between an action and the resulting harm. Whether or not someone is held liable for something depends on whether or not the harm can be traced back to his/her action as the proximate cause.

  1. Explanations are different than arguments. There are two kinds of explanations: physical causal and behavioral causal.
    1. You use arguments to support a statement but explanations to elucidate the reasons why some event happened.
      1. When we give a reason for doing something, we are presenting an argument for doing it.
      2. When we cite an individual person's reason for doing it, we are explaining why s/he did it.
    2. If the causal explanation of an event refers to physical background information, then the explanation is physical causal, e.g. "The main cause of global warming is an increase in the concentration of greenhouse gases resulting from human activity."
    3. If the causal explanation of an event is explained in terms of motives or reasons, then the explanation is behavioral causal, e.g. "The reason for poverty in capitalist societies is class struggle."
  2. Explanatory adequacy is a relative concept.
    1. What counts as an adequate causal explanation depends on our circumstances and needs. At minimum, an explanation should be consistent, not conflict with established fact or theory, be testable, and not have logical errors, such as circularity or unnecessary complexities.
    2. Testability is the most important feature of an adequate explanation.
      1. Causal explanations generate expectations, or predictions. Non-testable explanations generate meaningless predictions, no predictions, or false predictions. The test for explanatory adequacy is seeing if the predictions turn out to be true.
      2. Some predictions cannot be tested because of practical limitations, but others are untestable even in principle. It is only the latter that should be abandoned.
      3. Some explanations are not adequate because they are circular, i.e. they simply restate themselves and so generate no meaningful predictions. Others are not adequate because they are unnecessarily complex and so contain elements in which there is no reason to believe.
  3. Methods of forming causal hypotheses include the Method of Difference, the Method of Agreement, and the Best Diagnosis Method, all of which are guided by background knowledge of Causal Mechanisms.
    1. The Method of Difference identifies an event X as the only relevant difference (or one of the relevant differences) that has brought about the effect Y.
      1. More precisely, we say that one item has a feature that other items lack (the feature in question), and that only one relevant difference (the difference in question) distinguishes the item with the feature from the items without the feature; the difference in question then causes the feature in question.
      2. To make such an argument we need to know about at least two circumstances, one in which Y occurs and one in which it does not. If X is present along with Y and absent when Y is, then X might cause Y.
      3. "I ate my usual breakfast today, but with bacon instead of my usual sausage, and now I feel thirsty. Bacon tastes saltier than sausage, so I think the bacon made me thirsty."
        1. The bacon with breakfast is being put forward as the only occurring difference.
        2. Notice that the speaker is claiming some relevance to this difference. Bacon tastes saltier than sausage, and we know that salty foods can make us thirsty.
      4. Arguments from an only relevant difference can be as conclusive as any kind of reasoning we know.
        1. If you walk into a room, flip a switch beside the door, and see the lights go on, you conclude that flipping the switch caused the light to go on the basis of relevant difference reasoning.
        2. Even less indubitable arguments about the only relevant difference can provide as much certainty as ordinary experience ever provides, as long as the difference in question is truly relevant.
    2. The Method of Agreement links a cause to the feature in question on the grounds that it is the only (or one of the only) relevant common feature(s) among possible causes of Y. A correlation between two events provides a good starting point for hypothesizing.
      1. In such an argument, we begin by noticing that the feature in question (Y) occurs more than once, and that some common feature X is present on every occasion.
      2. Such reasoning requires that we know of more than one circumstance in which Y occurs.
      3. Co-variation is what happens when changes in one phenomenon are accompanied by changes in another phenomenon. Co-variation suggests, but does not confirm, that causation may be present.
      4. Fallacies of logic occur when thinking that correlations or co-variation prove causation.
        1. Cum hoc, ergo propter hoc is the fallacy of assuming a cause and effect relationship between co-varying phenomena.
        2. Post hoc, ergo propter hoc is the fallacy of assuming a cause and effect relationship between phenomena just because one event occurred before another.
        3. These are both fallacies of causal reasoning because they do not eliminate the possibility of coincidence, an underlying cause, or confusion between cause and effect.
    3. Causal Mechanisms are the interfaces between a cause and its effect. Background knowledge of "what causes what" is necessary to decide what is relevant to consider as a possible cause and what is not.
      1. We should be guided by our background knowledge of how things work, i.e. causal mechanisms, but humble enough to abandon a hypothesis when the relationship between one event and another proves to be something other than causation.
      2. Utilizing knowledge of causal mechanisms, the Method of Difference, and the Method of Agreement together facilitate hypothesis formation.
    4. The Best Diagnosis Method is a way to find a hypothesis, much like solving a crime or diagnosing a medical condition.
      1. Coming up with a hypothesis or "diagnosis" by this method involves assembling "symptoms" or "clues" and looking for patterns of association.
      2. The next step is to ascertain which are the relevant symptoms and the strongest associations between them and possible causal mechanisms. The best hypothesis is the one that eventually gets confirmed.
  4. A causal claim says that one thing causes another; a hypothesis is an initial speculation about a causal claim involving Inference to the Best Explanation.
    1. The general strategy for forming a hypothesis is "Inference to the Best Explanation." We begin by making the best guess and some alternative guesses.
    2. If we were reasoning deductively, Inference to the Best Explanation would constitute the fallacy of affirming the consequent. In inductive reasoning, however, on the basis of assumed probabilities of alternative explanations, it is not a fallacy to form an initial hypothesis this way; in fact, it may be all that we can do.
  5. Confirming a causal hypothesis consists in rigorous application of the above methods. Once the Method of Agreement or Disagreement suggests a hypothesis we can begin to eliminate other possibilities
    1. Controlled cause-to-effect experiments try to show directly that the presence and absence of a suspected cause, C, yields different frequencies, d, of the observed effect, E.
      1. By the Method of Agreement we might form a hypothesis as to what caused something to happen, e.g. that heat caused the water to boil, but the relationship between heat and boiling could just be a coincidence. The next step is to eliminate the other possibilities.
      2. Repeating the experiment many times is the backbone of the scientific method. If the water continues to boil every time we apply heat, we are closer to confirming the hypothesis. Using the Method of Difference, we eliminate the possibility that the type and composition of the pan it is boiled in (aluminum, iron, Teflon, etc.) had any effect on the outcome. When repeated trials, regardless of the composition of the pan, result in the same effect, we can confirm that heat was the cause of boiling.
      3. In essence, hypothesis confirmation is really just careful application of the Method of Agreement combined with the Method of Difference. Observing that the water boils when heat is applied is the Method of Agreement, i.e. that heat and boiling always occur together. Observing that the only difference between its boiling and not boiling is the application of heat is the Method of Difference.
    2. Causation in human populations differs significantly from causation among specific events and needs alternate argumentative strategies.
      1. nonexperimental cause-to-effect study tries to establish causation in populations, but with methods and standards that avoid direct experimentation on subjects.
        1. This type of study reasons forward from a possible cause, C, to the observed effect, E. People subjected to the suspected causal agent are compared to a "control" group who has not been subjected to the same suspected causal agent in order to see if the frequency, d, of a possible effect is greater in the first group. If the frequency is significantly higher in the target population than the control group, we conclude that C caused E.
        2. Because we can never be sure that factors other than the hypothesized cause contribute significantly to the effect, these studies are not nearly as conclusive as controlled experimental studies.
      2. nonexperimental effect-to-cause study also avoids direct experimentation on subjects.
        1. This type of study reasons backwards from an existing effect to its possible cause (or to one causal factor). This time investigators begin with a given effect, E; they select an experimental group that exhibits E and a control group not exhibiting E. Members of both groups are inspected for exposure to C, the suspected cause. If the frequency of C in the experimental group significantly exceeds the frequency in C, we call C a cause of E in the target population.
        2. The same cautions about nonexperimental cause-to-effect studies also apply here. As before, the members of the experimental group may differ relevantly from the rest of the target population. If we begin by studying people with arthritis, we must first recall that they will be older than an average member of the population. Again, we adjust the control group to resemble the experimental group. Again, if you can think of other factors that could have influenced C, make sure the control group was adjusted to reflect them.
        3. One final alert about effect-to-cause studies: They are less useful in making causal predictions about the population. Effect-to-cause studies show only the probable frequency of the cause in cases of a given effect, not the probable frequency of the effect in cases of a given cause. Therefore, they don't permit us to say what percentage of the target population would display E if everyone were exposed to C. Ideally, we would follow such a study with a cause-to-effect study, watching people with C over a long period to see if they develop E.
      3. Experimenting on animals is another method of testing causal hypotheses that avoids experimenting on humans, but the experimental results apply to humans by analogical reasoning (Chapter 10).
  6. Mistakes in causal reasoning compromise our ability to make predictions and hold reasonable expectations. Some examples of causally defective reasoning follow.
    1. Both cum hoc, ergo propter hoc and post hoc, ergo propter hoc fail to establish the improbability of the following three possibilities: that the connection between C and E is due to coincidence, that C and E both result from a third underlying cause, and that E caused C rather than the other way around, i.e. confusing effect with cause.
    2. Confusing effect with cause in medical tests is a common, and serious, mistake. Often the chances of actually having the condition are erroneously exaggerated. e.g. What are the chances that you have a disease if you test positive for it? Testing positive is the effect, E, of having the disease, C. Testing positive on a test that is 90% accurate means that 90% of those who have C will have E. Say the base rate (frequency of occurrence in the population at large) of the disease in question is 1%, and 10% who do not have C test positive (false positive), and you test positive. If you conclude (or your doctor tells you) that your chance of having the disease is 99%, then you'd be wrong. You'd have switched C and E. Your chances are actually only 8% of having the disease.
    3. Overlooking statistical regression causes errors in causal reasoning. The statistical property of measurements of mean values of populations, called "statistical regression," could be responsible for some effect, rather than causation. Sometimes the explanation for why a basketball player, for instance, returns closer to her average performance after an extremely good (or bad) performance is simply regression to the mean.
    4. Proof by absence of disproof is a defect of reasoning because the absence of the disproof of a causal hypothesis does not increase the likelihood that the hypothesis is true. The reasons for believing the hypothesis in the first place are left intact with the discovery of no disproof, but that absence doesn't create an additional reason for thinking that the hypothesis is true.
    5. Appeal to anecdote is really just the fallacy of hasty generalization of post hoc reasoning. Generalizing from an anecdote doesn't really show anything and can easily be countered by finding one example that disproves the generalization.
    6. Confusing explanations and excuses is a fallacy based on making the erroneous assumption that an explanation of an action is a justification for it. An expert may try to explain the psychological-sociological-political motivations behind the 9/11 suicide attacks on the World Trade Center without, in any way, trying to excuse the act.
  7. Causation in the law is the connection between action and harm.
    1. Under the law, if your action causes (or attempts to cause) harm (or contributes to its cause), you are held responsible for that harm.
      1. Conditio sine qua non, or "but for" causes (Y would not have happened but for X's having happened) are important because it wouldn't be fair to punish someone for causing harm Y by doing X when Y would have happened even if X had not been done.
      2. Proximate cause is a restrictive version of "but for" causes, combining fact and policy. With a policy that indicates what's relevantly important, we can trace the harm caused back to the action against a "causal background", or common recurrent feature in the environment. It can then be argued that some links in the causal chain of events were, or were not, part of the defendant's responsibility.
    2. Contravening events factor into legal argument.
      1. Establishing causation is essential in establishing liability, e.g. if Jeff unintentionally fails to completely extinguish his campfire and it burns down the forest, then he is liable to a certain degree. If Jennifer finds the smoldering fire and pours gasoline on it and it burns down the forest, then she is liable to a much greater degree than Jeff was. Jennifer's voluntary intervention contravenes Jeff's causal role.
      2. Proximate cause helps clarify when someone should not be held liable for being the cause of harm, e.g. if Jamie's muffler unexpectedly falls off his well-maintained car while driving, and causes a spark, which causes a fire that burns down the forest, Jamie cannot be blamed even though he was part of the causal chain.







Critical ThinkingOnline Learning Center

Home > Chapter 11 > Chapter Overview