Best explanation vs. ad hoc explanation. A sign that you may not be seeking the best theory surrounding the truth of a matter, and instead protecting a favored theory—one you want to be true over all others—is if you’re doing a lot of ad hoc reasoning in response to objections to it. We might call this “ad hoc-ing.” Ad hoc is Latin for “for this explanation or moment only” or “add here for a special purpose,” as when a business or bureaucracy forms an ad hoc committee to address an unforeseen situation. Ad hoc reasoning is deployed in, as it were, unanticipated or emergency intellectual situations; situations where a thesis or claim has failed to foresee an important objection or is at an impasse. When you engage in ad hoc explanation, you’re trying to save a favored thesis from pointed and skeptical questions—or new and disconfirming data or competing theses—by seat-of-the-pants rationalizing that cannot be generalized to other theses of the same type: “My psychic powers failed today because the audience had too many doubters in it.” “UFOs exist, but they never land on the White House lawn because they prefer not to be seen.” These explanations save the theses in question (the claims that psychic powers and UFOs are real), but at the expense of adding additional claims to them (psychic powers are real plus they don’t work in the presence of doubters; UFOs are real plus they don’t want to be known).
With the addition of a premise (a claim supporting another claim) to each thesis, the theses become less simple and probable—though still logically possible. In ad hoc premise adding, skeptical questions and new data are not really being anticipated or naturally incorporated into a theory or claim, but deflected with an additional claim or premise that is logically possible, but maybe not subject to empirical verification (reality testing). The ad hoc explanation, with each new premise added to the original claim, thus renders the explanation less plausible. Because there are lots of logically possible ways that the world can be that cannot be verified by evidence—we may, for example, be a dream in the mind of a butterfly—if you are willing to believe things absent evidence, you can engage in a lot of “ad hoc-ing” to save your favored beliefs, theories, explanations, and behaviors from skeptical inquiry and new data. Ad hoc premise generating can also be a sign that someone is arguing in bad faith (their motivations for posing such explanations are something other than the truth or the good).
Occam’s razor. One way to push back against ad hoc rationalizers is to deploy in response Occam’s razor, stated by William of Occam, c. 1324, in this manner: “No more things should be presumed to exist than are absolutely necessary.” That is, if you can keep things simple, do it. Don’t multiply premises unnecessarily. Ad hoc explanation’s seat-of-the-pants rationalizing multiplies premises, Occam’s razor shaves them off: “Maybe psychic powers fail in the presence of skeptics—and UFOs don’t land on the White House lawn—because neither psychic powers nor UFOs exist.”
Often the simplest explanation is best. And in some instances, mathematics can be deployed in support of Occam’s razor. As a matter of sheer probability, for example, two inductions being true are always going to be at least slightly less likely than a single induction being true. Each time a premise is added to an inductive thesis, the odds of the combined premises being true must necessarily come down (induction, recall, is about probabilities, not certainties). If, for instance, you’re 90% certain a particular woman is a Democratic voter, 90% certain she’s vegetarian, and 90% certain she signed your friend’s animal rights petition, the odds that you are actually right about all three of these in combination drops statistically to 72% (.9 x .9 x .9 = .72). If you also infer, say, with a confidence of 65%, that she’s a feminist, then the odds that she’s all four of these things come down further still (.72 x .65 = .468 or 46.8%). With just four inferences or assumptions surrounding the woman in question, your odds of being right about her have plummeted to under 50% (assuming you scaled your levels of confidence surrounding each claim accurately in the first place). Occam’s razor helps us get our odds of being right back up—proceeding with caution in the multiplication of premises.
But a caution to simplicity (Occam’s razor) as a criterion for evaluating truth is offered by the mathematician and philosopher, Alfred North Whitehead: “Seek simplicity and distrust it.” Why distrust it? One reason is that humans tend to find comfort in things they can control, and simple models or explanations might thus serve an emotional bias for control, distorting the complexity of the matter at hand, as in “the politician’s behavior can be summed up in one word: greed.” If you accept this simple heuristic (rule of thumb, model) for the politician’s behavior, you’ve got a lot of control over the processing of news you encounter about him or her, and you don’t have to expend energy thinking about it, but you may judge their words and actions wrongly, or fail to anticipate how their actions might impact your life. Simplicity as a criterion can arrest a deeper inquiry, and can open people up to such things as the availability heuristic (landing on the nearest and simplest rule of thumb, model, or map that comes to mind for explaining a problem or situation). Occam’s razor, used too casually, can signal lazy thinking.
Ideally, the critical thinker wants to locate heuristic rules of thumb, models, and maps that, like Goldilocks in search of her soup, chair, and bed, are just right. (That is, useful, attendant to the right signals in the noise, neither more nor less complicated than necessary.)
Prometheus Unbound 