A dominant philosophical practice is to design hypothetical scenarios that elicit strong moral intuitions which run against a particular moral theory. Critics of consequentialism rely heavily on this methodology.
Bernard Williams, for example, devised the famous scenario which strikes a chord of discontent among many people, who nevertheless may sympathize occasionally with the consequentialist idea that best outcomes matter. Imagine that you are an explorer and wander into a village. There you will find a military official, named Pedro, who is about to shoot 10 Indians, randomly selected. Since you’re an honored visitor from another land, Pedro offers you the guest’s privilege of killing one of the Indians with the firm promise that the other Indians will be let off. If you refuse, then Pedro will kill all of them. Surely, you’re not happy to be the decision maker in the scenario. You can save 9 lives, but the thought of killing one person is excruciating and it’s likely you’ll find it repugnant. I myself tend to feel the same way.
However, an interesting shift may take place under the pressure of “the new science of morality”, as it has been labeled. Drawing upon data from neuroscience, Joshua Greene contends that deontological judgments are preferentially supported by automatic emotional responses which, in many cases, respond to irrelevant moral features, while characteristically consequentialist judgments are preferentially supported by conscious processes of cognitive control. What seems a strong methodological weapon may turn out an Achilles’ heel.
Yet, non-consequentialist philosophers do engage in highly demanding conscious reasoning. But what if the effort of philosophical theorizing is an exercise in moral confabulation to polish off track emotional responses? Greene speculates that if you put together the fact that most of the time we are driven by strong emotional responses with the tendency to fabricate plausible sounding stories in order to justify or explain these responses, you get deontological moral philosophy.
As a philosopher with sympathies for the Kantian tradition, am I confabulating? In my paper, recently published in Neuroethics, I argue, hopefully non-motivated, that the evidence used by Greene does not support the confabulation hypothesis and that even if we accept it we should not be too worried about its debunking power.
One suspicion I start with is that paradigmatic cases of confabulation do not seem to fit the relevant model for Greene’s ambitious attack on deontology, namely what I call alarm-like emotion based confabulation. Since established cases tend to favor a neutral model, which is not committed to a particular content of behavioral causes (cognitive/emotional), it is puzzling to expect outright alarm-like confabulations in philosophical theorizing.
This puzzle leads to a deeper reason as to why the confabulation hypothesis is problematic. Why is the case that paradigmatic cases are not driven by alarm-like emotions? By looking at the conducive conditions for confabulation, I argue that there is an inherent resistance on the part of alarm-like emotions to be subject to confabulation. A confabulation is likely to occur when stimuli are not salient and are not plausible causes of belief or action. And vice versa, a confabulation is unlikely to occur when stimuli are salient and plausible causes.
But alarm-like emotions are highly salient, blunt, simple and almost forces one to issue strong commands such as “Don’t do it!”. Thus, it is unlikely to expect people to have alarm-like emotions that are activated by the tragic conditions of moral dilemmas, used to pit deontological judgments against consequentialist judgments, and not know what the tragedy is about. Understanding what is conducive to confabulatory behavior suggests that it is resistant to Greene’s profile of the psychological “essence” of deontology.
Now suppose that the justification of a particular deontological judgment is, indeed, a confabulation. How worried should deontological theory be? The deployment of knowledge in particular cases is ill-grounded in confabulation tendencies, not the content of the justifications or explanations in general. Once we make the distinction between application and content of knowledge, we can see more clearly that confabulation data can support an argument that deontology is incorrectly applied in particular cases, not that it is faulty theory in general.
And if a deontological confabulation is to sound like a plausible story, then it has to involve some valid features, because non-pathological confabulation works by picking out content from shared knowledge and norms which are endorsed in general. Ironically, admitting cases of deontological confabulation as rational appealing stories implies accepting that, in general, deontology has some epistemic merit from which confabulations get their prima facie plausibility.