Doing "good"
post by pchvykov · 2022-01-13T19:52:02.119Z · LW · GW · 7 commentsContents
7 comments
I just did my first substantial donation (informed by Effective Altruism, and my own work on complex social systems). Even though I have thought before about how and when it is best to donate, the act of actually giving spawned many new insights!
The biggest of these is that no matter how much research you do, and how much quantifiable metrics and statistics you collect, you never get any guarantees that your donation is making the world a better place. The trick is that in a system as complex as our society, any intervention will always send avalanches of consequences rippling through many branches of the system. While the most immediate and obvious effects may seem clearly beneficial, there will always be some secondary or tertiary effects that may be harmful. As it is impossible to account for all such effects, it becomes impossible to reliably quantify the overall benefit of the intervention -- and the more you think about it, the more your head hurts. We all know of the possible contradictory consequences of giving money to a homeless person on the street – but a similar argument could be made with apparently much more clear-cut causes such as sponsoring malaria treatments in Africa. Mosquito nets distributed in very poor regions are sometimes used as fishing nets in ways devastating for the local environment. Giving preventative malaria medicine to children may hinder the natural development of immunity and could thus increase mortality later in their lives.
But these are only the effects we can think of and quantify. When considering long-term and large-scale consequences, causal attribution becomes so challenging that no rigorous research or even logical reasoning is reliable. And so, knowing that the true cumulative impact of a donation is never certain, why did I still choose to donate?
The question seems rich and subtle to me. Are interventions in vastly different cultures we don't fully understand just another form of neo-colonialism? Do acts of kindness in my local community disempower people from learning to care for themselves? It seems that the only certain positive impact of donating, whether it be time or money, is the impact on myself. Making a decision and taking an action in full awareness of uncertainty turns out to be very empowering – a somehow “pure” opportunity to exercise free will. Doing the “right” thing in a context where reliable evidence clearly indicates what this “right” thing is, is more a social conditioning than a free decision. As such, it is only in the face of ambiguity that we can truly exercise our power to create something new - whereby our choice tilts an otherwise balanced scale.
As for the decision to donate in particular, for me it tilts the balance towards a personal free choice to try to do good – even when outcomes are unclear. And precisely because no single act can guarantee success, consistency then pushes me to choose good intention in all aspects of my life. Fundamentally, this can lead to a shift in personal identity: I can begin to identify as someone who strives to improve the world, and who trusts himself enough to act on it in the face of uncertainty. This in itself seems valuable enough for me.
7 comments
Comments sorted by top scores.
comment by Jackson Wagner · 2022-01-13T21:16:12.500Z · LW(p) · GW(p)
You might be interested in this 80,000 Hours podcast about the extreme moral uncertainty created by our complex world, and the (tounge-in-cheek) "moral case against ever leaving the house". I agree that it can be dizzying to think about how our deep uncertainty about the future (which philosopher Hillary Greaves calls "moral cluelessness") seems to potentially undermine all our efforts -- not just our altruistic endeavors, but what we seek to accomplish in our jobs, in our personal relationships, etc.
But the logic of expected value maximization tells us that we ought to go forth regardless, and make an effort according to our best judgement even in a shifting landscape that threatens to sometimes reverse the effect of our actions. To me, this embrace of uncertainty (as opposed to the obsession with blame, credit, and false guarantees that cloud much moral thinking) is a core part of what it means to be an effective altruist and a rationalist. Indeed, I think the embrace of uncertainty is a big part of the "edge" that allows effective-altruist donations to do such incredible amounts of good on average.
Here is the EA Forum tag page for "Moral Uncertainty" [? · GW], which collects a bunch of thoughtful posts on the subject you might be interested in.
Replies from: pchvykovcomment by Viliam · 2022-01-13T22:40:01.742Z · LW(p) · GW(p)
Uncertainty doesn't necessarily mean absolute uncertainty. If you save a drowning child, who knows, may grow up to be the next Hitler. But we still assume that saving drowning children is, at least on average, a good thing.
Replies from: pchvykov↑ comment by pchvykov · 2022-01-17T18:46:54.430Z · LW(p) · GW(p)
To me, the counter-argument to saving drowning children isn't the admittedly unlikely "Hitler" one, but more the "let them learn on their own mistakes" one - some will learn to swim and grow up more resilient, and some won't. The long-term impact of this approach on our species seems much harder to quantify.
comment by Dacyn · 2022-01-14T15:47:20.156Z · LW(p) · GW(p)
And so, knowing that the true cumulative impact of a donation is never certain, why did I still choose to donate?
Why did you phrase it like this -- what does being certain about the cumulative impact have to do with the choice of whether to donate or not? Why isn't the expected impact more important?
This is a serious question, I have the same psychological issue where certainty of a positive outcome is sometimes more important than positive expected value. But I don't yet have a rule which allows me to predict when this psychological effect will come into play.
Replies from: pchvykov↑ comment by pchvykov · 2022-01-17T18:51:28.185Z · LW(p) · GW(p)
oooh, don't get me started on expectation values... I have heated opinions here, sorry. The two easiest problems with expectations in this case is that to average something, you need to average over some known space, according to some chosen measure - neither of which will be by any means obvious in a real-world scenario. More subtly, with real-world distributions, expectation values can often be infinite or undefined, and median might be more representative - but then should you look at mean, median, or what else?