One recurrent theme in moral philosophy is the conflict between theory and practice. It is, inter alia, prevalent in the features on which one focuses when doing metaethics, it is prevalent in conflicts between theory-driven approaches – e.g.  (neo-)Kantian or utilitarian, that focus on formulating abstract principles – and practice-driven approaches – e.g. neo-Aristotelian ones, that focus on evaluations of actions and characters in some particular circumstances – in normative ethics, and just how to handle the conflict, whenever it arises, is an enormous methodological issue in applied ethics.

I have previously defended, and no doubt will continue to defend, a theory-driven approach to first-order ethics – i.e. normative and applied ethics as well as political philosophy – in general. I think this theory-driven kind of thinking has two main dimensions: first, metaethical implications in normative theory count for a lot, and second, in so far as one does  the normative stuff, one should work abstractly, preferring going from principles to cases, rather than the other way around. Apart from metaethical arguments for this view,[1] normatively, I think the practice-driven views are non-starters: to specify what is good or bad, right or wrong, just or unjust in particular situations you need the general theories in the first place. And, furthermore, Ross-style intuitionist pluralisms tend to strike me as cheating: why care about, not to mention how adjudicate between, particular intuitions unless you have some general systematic theory to fall back on? Surely a more consistent theory is preferable to another, and if it is based on some general value, it can give some support about what to do in many cases of adjudication between various principles. So pluralists have to choose between conflicting intuitions, on pain of irrelevance in comparison to more systematic approaches.

On such a theoretical approach, however, it is often very hard to tell what one is to do in particular cases. If you are an act-utilitarian, for example, how would you know which action maximizes value? Or if you are a rule-utilitarian, which are the rules you should follow (and how do you know which ones to apply in a given case)? Similarly, for a Kantian, what does it really mean to respect people as ends-in-themselves? There are, of course, substantial literatures on these questions, but far from any consensus, and I think a sin among moral philosophers – in fact, among philosophers working on normative matters more generally, since the same problem arise for many theories about epistemic justification[2] – is this preference for defending general theoretical strategies rather than also working out their details.

So. What on earth should you do? Well, here is one suggestion. Undoubtedly there are others, and I would be very grateful if a reader would provide one. But here, the idea is this: Distinguish first between Moralität – construed as developing and defending overarching principles, and Sittlichkeit – construed as the question about what is fitting to do in some particular situation, with or without background knowledge.

Within Sittlichkeit, there is a further distinction: the Kantian one of acting for a reason, or acting in accordance with it.[3] You act for a reason if you act by recognizing some particular consideration[4] – for example, that you shoot yourself because you are Hitler and shooting Hitler would save millions of lives – and you act in accordance with it if you get the same results but from another motiv(ation) – for example, you shoot yourself because you are Hitler and you are depressed and, as an unforeseen consequence, doing so saves millions of lives.

Now, if your moral thinking is very theoretical, it cannot be assumed or presumed that most people will be able to act for the reasons – or, rather, principles – that you recognize to be the most important ones, because they do not know of, or do not accept, the theory you think is justified. But that does not at all preclude others from acting in accordance with your views.

One could then try to see which people share the intuitions that one thinks supports, or are supported by, the larger theoretical framework, and see which people share those intuitions. If those people are also very active, there seems to be a way in which one can be guided by knowing what they do: for example, if you are a Kantian and you know someone who seems to think that respecting others is important and that person works for Amnesty International, there is a prima facie reason for believing that you should do that as well. In short, you can take moral advice from people who act in accordance with your reasons, even if they do not act for those reasons, because if they share the same intuitions and therefore similar motivations, one could expect, ceteris paribus, such persons to act in the same ways that the principles would recommend. To be sure, the advice is prima facie and defeasible, but it is advice nevertheless.[5]

In this way, even heteronomous agents could well be brightly burning moral beacons, and they have certainly been such for me: the Kantian-working-for-Amnesty case is literally speaking taken from my own life; I have become an activist not least in virtue of applying not least by applying the kind of methodology for moral advice suggested here upon noticing certain people with whom I share intuitions, if not theory,[6] doing such work.

And so should you.



[1] Some of them are found here, but they need some qualification, since I also doubt that you can do second-order metaethics without there being any first-order ethics, either as a natural or philosophical phenomenon, to do them about. They should thus not be construed as general anti-first-order arguments, but rather as  potential defeaters for non-theoretically informed first-order ethics inside some general web of beliefs.

[2] I am in particular thinking of reliabilists with their generality problem, and coherentists with the general question of just making sense of what coherence is.

[3] Actually, things get messier, since there are probably ways of acting for a reason that need not be the paradigmatic cases: you could act for a reason without acknowledging it to be a reason, for example. Some philosophers – e.g. Audi, Robert (2005). The Good in the Right, Princeton University Press, and Skorupski, John (2010), The Domain of Reasons, Oxford University Press – have noticed this point and draw further distinctions, but I will stick with the original Kantian one for the sake of simplicity.

[4] This is also extremely under-specified, and I apologize for that.

[5] Interestingly, this epistemology inverts the one defended by Shafer-Landau, Russ (2003). Moral Reality: A Defense of Moral Realism, Oxford University Press. He posits original intuitive knowledge possessed by (morally and epistemically) virtuous agents as reliable, and then wants to develop that into theory by the ordinary reflective equilibrium procedures. I say the opposite: you start with theory, then see who is virtuous, then apply it.

[6] But n.b. that I do not mean to suggest that I defend the Categorical Imperative or something else along those lines; rather, I am, if anything, a neo-neo-Kantian, who thinks themes worked on by various neo-Kantian writers (including contemporary Harvard neo-Kantians such as John Rawls and Christine Korsgaard or Oxford neo-Kantians such as R.M. Hare or John Skorupski) are the most promising approaches to moral theory.