A review and critique of the first section of the volume The Ethics of Giving: Philosophers' Perspectives on Philanthropy, edited by Paul Woodruff, with an emphasis on issues relevant to the decision making of Effective Altruists.
Hill is a distinguished Kant scholar who looks at what Kantian theory has to say about giving. He starts with a good overview of the principles of Kantian ethics. He then underlines the fact that we have a general duty towards beneficence, making others' happiness our own end. He states that Kant's system places beneficence as an imperfect duty, which implies that it is not as important as refraining from violations of stronger duties like justice and respect. He also quotes Kant to point out that being partial towards some people does not violate the duty of beneficence.
But one of the first things that Hill does is to distinguish a special and more demanding case of beneficence, the duty to aid, which is for "those in distress, for example, those needing immediate help to meet basic needs for survival and tolerable existence." Hill says that "it seems presumptively wrong, contrary to the dignity of humanity, to let a person starve to death or live in mind- numbing squalor when one can easily prevent this." Unfortunately, he writes as if this is a special case of unusual emergency scenarios, like drowning children in ponds. Yet, for the hundreds of millions of people who are facing fatal diseases and severe poverty, this need for aid is already the state of affairs, and we can help them through existing charities. It is also the case for perhaps every animal in factory farms and the people who would die in the potential future extinction event that one might only solve through immediate efforts. Since this state of affairs is not going to go away for the foreseeable future, it's not clear how there is any room at all for one to be benevolent in other ways. Any beneficence towards less serious causes is going to take place at the direct expense of our duty to aid the worst off. Hill does not address this objection.
Elsewhere, Hill answers a typical drowning child type scenario by invoking a different duty, the principle of mutual aid, wherein we must help those in dire distress when the cost to ourselves is low and there are no overriding moral concerns. He does not explain how this is different from the duty to aid, and again he does not explain why donating to charity wouldn't fall under the duty of mutual aid.
However, these issues don't matter anymore if giving is a matter of justice, because justice is a perfect duty that supersedes all others. Hill argues that charitable giving falls under the duty of justice rather than beneficence if the wealth was acquired through unjust means. Hill goes on to point out that Kant viewed charity from the very rich as something that is owed, rather than something that is meritorious, but beneficence is a duty too, even if it is an imperfect one, and Hill does not explain how classifying giving as something that is owed implies that it is a matter of justice; Kant's imperfect duties are perfectly capable of generating moral obligations. Hill also references Kant's statement that giving is often an obligation rather than a gift because the wealth often comes from injustices of the government. It is not clear to me if this implies that a just person, who was able to obtain wealth as a result of the government's injustice, has a judicial obligation to give. Needless to say, the question of which kinds of wealth acquisition are or aren't just is a huge one, where I suspect there is plenty of room to dispute the extent of injustice in contemporary business and government.
Regardless, when giving is a matter of justice it cannot be compromised in favor of other duties. For instance, our imperfect duties for self development are always subservient to the imperative to donate our money. Likewise our imperfect duties to aid those in distress are always subservient to our duty to make justice-based reparations.
It seems clear to me that only particular causes might constitute a reparation of injustice. If the injustice was done to a specific person - for instance, I ripped off the co-founder of my startup or something of the sort - then naturally I have an obligation to repay her. If the injustice was done to all my employees then I should repay them. Similarly, we might think that if the injustice is something systemic, e.g. the structure of the whole global economy is unjust and it is wrong to participate in it, then I should donate to any poor people in the world who I can help. But it seems to me that if a wealthy person is responsible for the poverty of all those who have been wronged by the system that made them wealthy, then they are also responsible for the animals who have been wronged by that system, as well as the deaths and foregone births of those who are the victims of extinction events that are caused or allowed by that system. For instance, if we think that the capitalist system is globally unjust so that it is responsible for all poverty, we should note that it is responsible for factory farms and risky technological experimentation as well. Similar arguments could be made for a multitude of non-EA causes. So if someone has wealth that comes from injustice, they may be obligated to repay one particular person, or they may have an open mandate to donate to almost any cause, or anything in between - depending on the kind of injustice and how you interpret its resulting obligations.
But cause selection for the reparative donor can be constrained by effectiveness concerns just as it is for a consequentialist donor. Perfect duties must always be fulfilled to the "fullest extent possible" (Hill 1971), so giving-as-justice seems to entail an obligation for effectiveness. Hill denies that there are any values in Kant's system, even lives, that one can aggregate and maximize. But, does this mean that it literally does not matter how many people are saved? Such a view would be preposterous. Unfortunately Hill's following statements are extremely vague and it is not even clear if he is trying to elaborate upon this point or if he is simply moving on to making a different one. But elsewhere in the chapter, Hill clearly makes a distinction between saving all the drowning people from saving a few of them and quitting, and mentions how actions can be wrong in virtue of indicating that people have the wrong priorities. While the aggregate number of lives saved isn't a moral object in its own right for Kant, it still seems to be a fact that people can respond to and care about in their fulfillment of moral duties. Extending this logic, it seems clear that donating to an ineffective charity shows that one has other priorities than saving people. If one's duty to give is imperfect - ordinary beneficence - then it is permissible to be partial and relaxed about it, as Hill points out, as long as one sincerely values the happiness of all other people. But if one's duty to give is perfect - justice - then such partiality would constitute a failure to prioritize one's perfect duty.
The principle of mutual aid and the duty to aid are both imperfect duties, so we don't have to fulfill them to the fullest extent possible. However, Hill (and Kant, it would seem) considers them to be stricter than ordinary beneficence. The exact nature of this middle ground is left unclear, and this is very disappointing given that it probably constitutes the majority of our obligations to give. I am likewise disappointed that Hill does not spell out the details of how wealth should be given out if it was earned via injustice, though in this case I find it easier to figure out the logical implications of Hill's system.
In summary, I believe these are the correct takeaways from Hill's framework:
* If wealth was gathered via injustice, donate to whatever effective charity is most effective in serving as reparation to that injustice, up until the point that further donations would be counterproductive to your future ability to make more reparations.
* If wealth was not gathered via injustice, donate it to a charity that helps those who are in dire need. You must give at least some moral priority to effectiveness, and donate quite a bit of money, doing your best to balance these goals with other moral duties.
Meanwhile, it is worth pointing out that David Cummiskey (1996) has an alternative, less mainstream interpretation of Kant wherein we are obligated to follow an agent neutral consequentialism, promoting the good as much as possible. It is much more strict than Hill's conception of ordinary beneficence in terms of both prioritization and demandingness, but perhaps not more strict than the triad of duties towards justice, aid and mutual aid that I investigate here. In this chapter, Hill defends his view from some of Cummiskey's arguments, but does not go deep into the dispute.