Time and again in ethics and politics, (and even epistemology and metaphysics) we ask whether statements should be justified by pragmatism or by principle. While people do in fact have principled ethical beliefs, we can see that ethical and political principles rest on a pragmatic foundation as a result of imperfect knowledge, the function of ethics in terms of social cooperation and the requirements of consistency.
Pragmatism, also known as instrumentalism or consequentialism, is the idea that some statement or system should be judged in some manner according to its results or consequences. Pragmatism stands in contrast to principle, (a.k.a. essentialism or idealism) the idea that a statement or system should be judged in some manner according to its intrinsic or inherent properties, without regard to its results or consequences.
Care should be taken not to confuse this sense of pragmatism with the narrower notion of expediency, the idea that some statement or system should be judged on its immediate or superficial effects. Although pragmatism is used in informal speech in the sense of "expediency", the latter term is restricted to the more narrow sense, and pragmatism is best reserved for the broader sense.
In ethics especially we see the tension between pragmatism and principle: Is some action inherently bad? Or is it bad because it produces bad effects? The question becomes complicated because it is manifestly true that people do have ethical principles: they condemn some actions without regard to their consequences.
Empirical investigation into the trolley problem adequately demonstrates the human adherence to principle. There are two cases of the trolley problem:
A trolley is running out of control down a track. In its path are 5 people who have been tied to the track by a mad philosopher. Fortunately, you can flip a switch which will lead the trolley down a different track to safety. Unfortunately, there is a single person tied to that track. Should you flip the switch?
As before, a trolley is hurtling down a track towards five people. You are on a bridge under which it will pass, and you can stop it by dropping a heavy weight in front of it. As it happens, there is a very fat man next to you - your only way to stop the trolley is to push him over the bridge and onto the track, killing him to save five. Should you proceed?
Studies have shown that, although the consequences are identical, people who would "flip the switch" would not "push the fat man" in front of the trolley. The obvious interpretation is that people have a principled objection to the latter case that they do not have to the former case. On the other hand, the former case clearly shows a pragmatic evaluation: people typically flip the switch not because flipping a switch is inherently good but because flipping the switch leads to better consequences.
What are we to make of this apparent contradiction? Some philosophers argue* that the trolley problem shows that people are inherently irrational about ethics. I think this view is misguided. After talking about religion for the better part of a decade, I'm well aware that people do have irrational beliefs. But one must make some effort to find the underlying rationality for a position before declaring that position irrational.
*"Some people say..." should be a red flag. See this comment.
One problem with pragmatism in the broadest sense is that all actions and states of affairs are related; it is impossible to discover all the effects and consequences of any action, however well- or ill-intentioned. We have imperfect knowledge of the consequences of individual actions. But just because we have imperfect knowledge doesn't mean we have no knowledge at all. Even if our best isn't perfect, we can still do our best, indeed we all feel that we ought to do our best to predict the consequences of our actions.
One artificial condition of both trolley problem cases is that the subject is assumed to to have perfect knowledge of the consequences of his action. He knows, beyond a shadow of a doubt, that flipping the switch or pushing the fat man onto the tracks will definitely save five at the cost of one.
In reality, we don't know with absolute certainty that flipping the switch will indeed change the course of the trolley. But if flipping the switch doesn't work, there is no harm done: the effects would be the same as not doing anything. Likewise, we don't know that pushing the fat man onto the tracks will really stop the trolley. However, if it is ineffective, then six will die instead of five. There's no risk to just flipping a switch, but there is a risk to pushing the fat man onto the tracks.
The trolley problem by its structure ignores the justification of principle by imperfect knowledge. It is misleading to probe our ethical beliefs that are formed in response to imperfect knowledge with thought experiments that assume perfect knowledge. Our ethical beliefs are typically not explicitly reasoned out step by step by each individual. They are the immediate result of childhood indoctrination and socialization and the result in a larger sense of tens of thousands of years of social evolution and hundreds of millions of years of biological evolution.
Another flaw in the trolley problem is that it considers only the expedient results of the various actions, the consequences isolated from larger social considerations.
But our ethical beliefs are formed in the context of a cooperative society, they are strategies for getting along with each other. A critical feature of ethical strategies is that people must know what actions will earn the approval or disapproval of their neighbors. But however well we define the "best" result, we cannot judge any action on just its result; because of imperfect knowledge, we cannot always know the effects of our actions.
Another important component of ethics in terms of social cooperation is consistency. We want to approve or disapprove of actions with a high degree of consistency, regardless of the expedient consequences of an individual action. We consistently disapprove of driving drunk even if the driver manages to avoid any accidents, because we know in a larger pragmatic sense that driving drunk creates an unacceptable risk of accident.
Combining the notion of imperfect knowledge, consistency, and ethics as a strategy for social interaction gives us a good rational account of not only the existence of ethical principles (i.e. ethical beliefs applied directly to actions rather than their consequences), but their not-perfectly-consistent application.
An ethical principle in this account is a generalization of the larger pragmatic consequences of specific kinds of action. If some action will typically or usually produce bad results, we disapprove of the action even if in some unusual specific circumstances it might have a good result.
In the trolley problem we know (or at least think we know) that in general it is bad to place someone at risk who is not already at risk. In the first instance, all six people tied to the tracks by the mad philosopher are at risk, so the principle of not putting people at risk in not operative. In the second case, only five people are at risk; pushing the fat man onto the tracks places someone at risk who is not already at risk. Even though it is (or appears to be) expediently justified in the specific case, it is not generally justified. (Keep in mind too that the action is expediently justified only by unrealistically assuming perfect knowledge about the effects of fat men on trains.) We really don't want other people putting us at risk just because they believe, however sincerely, that harming us would save many others. That's a decision I demand I reserve for myself.
An important argument for this view of ethical principle is the observation that one can come up with scenarios that undermine the absolute universality of any principle.
I consider lying to be wrong on principle, but there are cases (e.g. the canonical lying to the Nazis about hiding Jews) where I would lie with a perfectly clean conscience. If telling the truth were an absolute ethical principle then, as Kant notes, it ought to be wrong to lie in any circumstances, regardless of the outcome.
(An alternative would be to show that some principle always had a good outcome, regardless of the particular circumstances. However, because the principle is still being justified by the outcome, i.e. it is still be justified pragmatically, the operative distinction between principle and pragmatism is erased.)
Finding an exception, however unusual, to a principle shows that the principle is a generalization, not a universal. There's nothing wrong with generalizations when properly used, but just the fact that there are exceptions to every ethical principle shows that there must be some other foundation for justifying ethical principles than the inherent or intrinsic nature of the action.
Thus we can conclude that ethical principles are justified by generalizations of their larger pragmatic consequences, and that the employment of principles is itself a pragmatic response to our imperfect knowledge and need for consistency in their employment for social cooperation.