Difficult and important topic, so thanks for trying, despite “no conclusion” at the end.
My own naive reasoning about good and evil is that in short term evil is more profitable, but there are a few mechanisms that work against evil in long term, albeit each of them unreliably. In random order:
People instinctively play some approximation of the “tit for tat” strategy. If you defect habitually, you get a big payoff in the first round, and small ones in the remaining rounds. (Doesn’t work if the evil person can aim a gun at you and say “press the Cooperate button, or I shoot you”.)
Violent people can terrorize their peaceful neighbors, but once in a while they get in a conflict with another violent person, and then violence escalates dramatically. The mobster you paid racket to ten years ago is probably no longer alive by now. (Doesn’t work if the evil person is smart enough to only pick easy targets. Also, the old mobster may be dead, but so what, now you pay racket to a new one.)
If there is freedom of association, good people will prefer to associate with other good people, who will accept them in turn. The evil people will have to settle for other evil people, or stay alone. The “karmic reward/punishment” works by being surrounded by people like you. (Doesn’t work without the freedom of association. Also, evil people may guilt/trap/confuse the good ones to make leaving difficult.)
I also suspect that empathy towards other people might be related to empathy towards your future selves, which may be other mechanism how evil people get “karmic punishment”. -- What you said about “being with yourself” seems related to this.
The balance between good and evil behaviors probably depends a lot on environment. For example, in a big city you don’t have to care about your prestige so much, because you can always find fresh victims. On the other hand, escaping from abusive relationships may be easier. Freedom of association, or acceptability of violence, depends on culture. There may be ideologies that tell you that playing “tit for tat” is bad. Sometimes prevalence of evil makes evil strategies easier: good people do not have enough resource to fight all evil ones. (This would suggest that there are good societies and bad societies, with different stable equilibria.) Sometimes prevalence of evil makes evil strategies harder: psychopaths probably benefit from interacting with people who have no previous experience of interacting with psychopaths. (This would suggest that the balance between good and evil converges to some natural “predator:prey” ratio.) It’s complicated.
According to Solzhenitsyn’s famous quote, “the line dividing good and evil cuts through the heart of every human being”. In the original context, however, the idea is much less relativistic: it kinda says that people who keep striving (and failing) to be good, are the good ones. And then there are the evil ones, who don’t worry about these things, because they have outsourced their morals to an ideology. -- This also resonates with some of what you wrote, except instead of people too stupid or too busy to think about morality, it posits people who believe they already have the answer, and who don’t bother evaluating this answer anymore.
The reason for benevolence, I think, is quite simple: it is a value; it “feels good”. That is all there needs to be. Thinking in terms of “why is it better to do good? maybe because it helps you to profit from better cooperation...” is at risk of confusing evolution with human mind. From evolutionary perspective, sure, the ability to desire good for others is probably somehow an evolutionary advantage, probably through some combination of tit-for-tat and signalling abundance. But that is not necessarily how it feels from inside. It’s like saying that sex is good because it allows us to reproduce and spread our genes, while completely failing to consider that sex is also enjoyable. And then you would wonder why people have protected sex, and perhaps conclude that they must be confused, because there is no logical reason for such irrational behavior. Similarly, people who help others without getting anything back are not necessarily confused; they just follow their urges.
Which is not contradicting the evolutionary explanation, nor denying that a lot of benevolent behavior actually invites reciprocation which ultimately benefits the do-gooder. Sure it does; that is why evolution allowed this trait to exist. But a person who would do good only in situations where they predict a mutually beneficial exchange, would probably miss many non-obvious opportunities. (So could a sufficiently smart person update on this, and strategically keep doing good even when the benefits are not obvious? Sure they could; but an easier way for evolution to achieve the same outcome is to create a person who simply wants to do good.) It still makes sense to study the evolutionary perspective, because it can tell us which situations are more likely to trigger the urge to do good, etc. We just shouldn’t confuse “they do it because (evolutionary reason)” with “they do it because (mental process)”. Just like there is “sex because copying genes” vs “sex because it feels good”, there is “benevolence because coalitions and signaling” vs “benevolence because I love my neighbor”.
Then there is this further complication that some people do not feel the urge to do good, like never, at all. Well, yeah, they don’t; but this is a fact about them, not necessarily some “deep truth” about the rest of humanity. (Just like the fact that some people are asexual, doesn’t mean that the rest of us don’t “truly” enjoy sex. Or the fact that some people are colorblind doesn’t mean that when the others talk about colors, they merely engage in some socially enforced mutual deception.) For the rest of us, the urge to do good is turned on and off depending on our mood, circumstances, etc. Yes, this is all true; but again, just because we don’t feel an emotion 24 hours a day, doesn’t make it “not real”.
The thing with an identity of “good person” is tricky. If one believes that people are either inherently good or inherently evil, then classifying yourself as inherently good is more healthy than classifying yourself as inherently evil. But the problem is that these are both poor choices. Also, we are often classified by others. In your example with “white fragility”, well, if you happen to live in an environment where people label others as “racists” and “non-racists”, and then insist that the racists must be punished, it obviously makes sense to insist that you are a non-racist, regardless of the facts. You are just trying to avoid the likely punishment.
But it is true that identifying as a “good person” puts an emphasis on what you “are”, instead of what you do, which seems like a wrong perspective. (If what you “are” is just a sum of what you do, then it gets it completely wrong. But even if what you are is a generator of what you do, still, considering the human propensity for self-deception, it would be better to judge the generator by its actual outputs, rather than by advertising. In reality, it is probably somewhere in between: your mental setting shapes your actions, and is in turn shaped by them.) Also, seeing yourself as a “good person” may reflect your lack of ability to notice your demons. And making it a part of your identity may prevent you from noticing them later. A better approach would be something like “I am an ambiguous person, and sometimes I try to do good, and sometimes I wish I would do more good”.