What is the meaning of a principle like “causing needless pain is bad” or “lying is wrong”? These principles are not always right–think about the pain of an athletic event or lying to the Gestapo. Various explanations have been proposed for the relationship between such principles and their exceptions. Maybe lying is wrong if certain conditions are met, and those conditions are common. Or maybe lying is really the union of two concepts–“mendacium” (mendacious untruths) and “falsiloquium” (blameless misleading), to use medieval concepts. Or maybe lying and pain-causing are always bad “pro tanto”–as far as that goes. They are always bad but their badness can be outweighed.
Mark Norris Lance and Maggie Little have another theory: “defeasible generalization.”* The following are defeasible generalizations taken from science: Fish eggs turn into fish. A struck match lights. These assertions are certainly not always true. In fact, very few fish eggs actually turn into fish, and I rarely get a match going on the first try. Nevertheless, a fish egg turns into a fish unless something intervenes. Even though the probability of its reaching the fish stage is low, to do so is its nature. The privileged cases are the ones in which the egg turns into a fish and the struck match catches fire. All the other outcomes, even if they are more common, are deviant. To understand that something will normally or naturally turn into a fish is to realize that it is a fish egg.
Lance and Little make a close analogy to moral issues: “Many key moral concepts–indeed, the workhorses of moral theory–are the subjects of defeasible moral generalizations. … Take the example of pain. We believe it is important to any adequate morality to recognize that defeasibly, pain is bad-making.” In other words, it is correct that causing pain is bad, even though there are exceptions that may turn out to be common. “To understand pain’s nature, then, is to understand not just that it is sometimes not-bad, but to understand that there is an explanatory asymmetry between cases in which it is bad and cases in which it is not: it is only because pain is paradigmatically bad-making that athletic challenges come to have the meaning they do, and hence provide a kind of rich backdrop against which instances of pain can emerge as not-bad-making, as not always and everywhere to-be-avoided.” Moral discernment is grasping the difference between paradigm cases and aberrant ones. We learn this skill, but it is not just a matter of applying rules. It may not be codifiable.
This seems plausible to me. But I do not think that every moral issue works this way. Take the absolutely crucial concept of love. We might say, as a defeasible generalization, that love is good. We know that in some cases love is bad. Adultery, obsessive love, and lust are common examples (although each of these bad categories admits counter-examples that happen to be good). But maybe it is true to say that love is good just in the same way that it is true to say that fish eggs turn into fish. This principle (arguably) reveals an understanding of the concept of love even though many cases are exceptional.
Here is my worry. I do believe, as a statistical generalization, that most cases of love are good. However, I also believe that we have a tendency to overlook the bad side of love, especially if we are the subject or object of it. We have biases in favor of love that presumably arise from our biological desires for sex and companionship and from the legacy of a million stories, poems, paintings, movies, and songs in which the protagonists fall in love and are admired for it. So the principle that love is good, if treated as a defeasible generalization, a default position, or a rebuttable presumption, is likely to mislead.