I used to spend a lot of time thinking about formal ethics, trying to figure out whether I was leaning more towards positive or negative utilitarianism, about the best courses of action in light of the ethical theories that I currently considered the most correct, and so on. From the discussions that I've seen on this site, I expect that a lot of others have been doing the same, or at least something similar.

I now think that doing this has been more harmful than it has been useful, for two reasons: there's no strong evidence to assume that this will give us very good insight to our preferred ethical theories, and more importantly, because thinking in those terms will easily lead to akrasia.

1: Little expected insight

This seems like a relatively straightforward inference from all the discussion we've had about complexity of value and the limits of introspection, so I'll be brief. I think that attempting to come up with a verbal formalization of our underlying logic and then doing what that formalization dictates is akin to "playing baseball with verbal probabilities". Any introspective access we have into our minds is very limited, and at best, we can achieve an accurate characterization of the ethics endorsed by the most verbal/linguistic parts of our minds. (At least at the moment, future progress in moral psychology or neuroscience may eventually change this.) Because our morals are also derived from parts of our brains to which we don't have such access, our theories will unavoidably be incomplete. We are also prone to excessive rationalization when it comes to thinking about morality: see Joshua Greene and others for evidence suggesting that much of our verbal reasoning is actually just post-hoc rationalizations for underlying moral intuitions.

One could try to make the argument from Dutch Books and consistency, and argue that if we don't explicitly formulate our ethics and work out possible contradictions, we may end up doing things that work cross-purposes. E.g. maybe my morality says that X is good, but I don't realize this and therefore end up doing things that go against X. This is probably true to some extent, but I think that evaluating the effectiveness of various instrumental approaches (e.g. the kind of work that GiveWell is doing) is much more valuable for people who have at least a rough idea of what they want, and that the kinds of details that formal ethics focuses on (including many of the discussions on this site, such as this post of mine) are akin to trying to calculate something to the 6th digit of precision when our instruments only measure things at 3 digits of precision.

To summarize this point, I've increasingly come to think that living one's life according to the judgments of any formal ethical system gets it backwards - any such system is just a crude attempt of formalizing our various intuitions and desires, and they're mostly useless in determining what we should actually do. To the extent that the things that I do resemble the recommendations of utilitarianism (say), it's because my natural desires happen to align with utilitarianism's recommended courses of action, and if I say that I lean towards utilitarianism, it just means that utilitarianism produces the least recommendations that would conflict with what I would want to do anyway.

2: Leads to akrasia

Trying to follow the formal theories can be actively harmful towards pretty much any of the goals we have, because the theories and formalizations that the verbal parts of our minds find intellectually compelling are different from the ones that actually motivate us to action.

For example, Carl Shulman comments on why one shouldn't try to follow utilitarianism to the letter:

As those who know me can attest, I often make the point that radical self-sacrificing utilitarianism isn't found in humans and isn't a good target to aim for. Almost no one would actually take on serious harm with certainty for a small chance of helping distant others. Robin Hanson often presents evidence for this, e.g. this presentation on "why doesn't anyone create investment funds for future people?" However, sometimes people caught up in thoughts of the good they can do, or a self-image of making a big difference in the world, are motivated to think of themselves as really being motivated primarily by helping others as such. Sometimes they go on to an excessive smart sincere syndrome, and try (at the conscious/explicit level) to favor altruism at the severe expense of their other motivations: self-concern, relationships, warm fuzzy feelings.

Usually this doesn't work out well, as the explicit reasoning about principles and ideals is gradually overridden by other mental processes, leading to exhaustion, burnout, or disillusionment. The situation winds up worse according to all of the person's motivations, even altruism. Burnout means less good gets done than would have been achieved by leading a more balanced life that paid due respect to all one's values. Even more self-defeatingly, if one actually does make severe sacrifices, it will tend to repel bystanders.

Even if one avoided that particular failure mode, there remains the more general problem that very few people find it easy to be generally motivated by things like "what does this abstract ethical theory say I should do next". Rather, they are motivated by e.g. a sense of empathy and a desire to prevent others from suffering. But if we focus too much on constructing elaborate ethical theories, it becomes much too easy to start thinking excessively in terms of "what would this theory say I should do" and forget entirely about the original motivation that led us to formulate that theory. Then, because an abstract theory isn't intrinsically compelling in the same way that an emphatic concern over suffering is, we end up with a feeling of obligation that we should do something (e.g. some concrete action that would reduce the suffering of others), but not an actual intrinsic desire to really do it. Which leads to the kinds of action that are optimizing towards the goal of stop feeling that obligation, rather than the actual goal. This can manifest itself via things such as excessive procrastination. (See also this discussion of how "have-to" goals require willpower to accomplish, whereas "want-to" goals are done effortlessly.)

The following is an excerpt from Trying Not To Try by Edward Slingerland that makes the same point, discussing the example of an ancient king who thought himself selfish because he didn't care about his subjects, but who did care about his family, and who did spare the life of an ox when he couldn't face to see its distress as it was about to be slaughtered:

Mencius also suggests trying to expand the circle of concern by beginning with familial feelings. Focus on the respect you have for the elders in your family, he tells the king, and the desire you have to protect and care for your children. Strengthen these feelings by both reflecting on them and putting them into practice. Compassion starts at home. Then, once you’re good at this, try expanding this feeling to the old and young people in other families. We have to imagine the king is meant to start with the families of his closest peers, who are presumably easier to empathize with, and then work his way out to more and more distant people, until he finally finds himself able to respect and care for the commoners. “One who is able to extend his kindness in this way will be able to care for everyone in the world,” Mencius concludes, “while one who cannot will find himself unable to care for even his own wife and children. That in which the ancients greatly surpassed others was none other than this: they were good at extending their behavior, that is all.”

Mencian wu-wei cultivation is about feeling and imagination, not abstract reason or rational arguments, and he gets a lot of support on this from contemporary science. The fact that imaginative extension is more effective than abstract reasoning when it comes to changing people’s behavior is a direct consequence of the action-based nature of our embodied mind. There is a growing consensus, for instance, that human thought is grounded in, and structured by, our sensorimotor experience of the world. In other words, we think in images. This is not to say that we necessarily think in pictures. An “image” in this sense could be the feeling of what it’s like to lift a heavy object or to slog in a pair of boots through some thick mud. [...]

Here again, Mencius seems prescient. The Mohists, like their modern utilitarian cousins, think that good behavior is the result of digital thinking. Your disembodied mind reduces the goods in the world to numerical values, does the math, and then imposes the results onto the body, which itself contributes nothing to the process. Mencius, on the contrary, is arguing that changing your behavior is an analog process: education needs to be holistic, drawing upon your embodied experience, your emotions and perceptions, and employing imagistic reflection and extension as its main tools. Simply telling King Xuan of Qi that he ought to feel compassion for the common people doesn’t get you very far. It would be similarly ineffective to ask him to reason abstractly about the illogical nature of caring for an ox while neglecting real live humans who are suffering as a result of his misrule. The only way to change his behavior—to nudge his wu-wei tendencies in the right direction—is to lead him through some guided exercises. We are analog beings living in an analog world. We think in images, which means that both learning and teaching depend fundamentally on the power of our imagination.

In his popular work on cultivating happiness, Jonathan Haidt draws on the metaphor of a rider (the conscious mind) trying to work together with and tame an elephant (the embodied unconscious). The problem with purely rational models of moral education, he notes, is that they try to “take the rider off the elephant and train him to solve problems on his own,” through classroom instruction and abstract principles. They take the digital route, and the results are predictable: “The “class ends, the rider gets back on the elephant, and nothing changes at recess.” True moral education needs to be analog. Haidt brings this point home by noting that, as a philosophy major in college, he was rationally convinced by Peter Singer’s arguments for the moral superiority of vegetarianism. This cold conviction, however, had no impact on his actual behavior. What convinced Haidt to become a vegetarian (at least temporarily) was seeing a video of a slaughterhouse in action—his wu-wei tendencies could be shifted only by a powerful image, not by an irrefutable argument.

My personal experience of late has also been that thinking in terms of "what does utilitarianism dictate I should do" produces recommendations that feel like external obligations, "shoulds" that are unlikely to get done; whereas thinking about e.g. the feelings of empathy that motivated me to become utilitarian in the first place produce motivations that feel like internal "wants". I was very close to (yet another) burnout and serious depression some weeks back: a large part of what allowed me to avoid it was that I stopped entirely asking the question of what I should do, and began to focus entirely on what I want to do, including the question of which of my currently existing wants are ones that I'd wish to cultivate further. (Of course there are some things like doing my tax returns that I do have to do despite not wanting to, but that's a question of necessity, not ethics.) It's way too short of a time to say whether this actually leads to increased productivity in the long term, but at least it feels great for my mental health, at least for the time being.

Edited to add in April 2017: A brief update on the consequences of this shift in thought, three years later.