Refuting Criticisms of Utilitarianism and Effective Altruism


This piece is a response to Robert Martin’s piece critiquing Peter Singer’s views concerning utilitarian ethics and Effective Altruism (EA). I do not address every point raised in this article, but restrict my response to four key lines of argument. First, I argue that Martin’s response presumes a binary conception of morality (moral versus immoral) which utilitarianism itself denies, and as such the criticisms he levels on the basis of this assumption have little relevance to utilitarianism. Second, I consider Martin’s argument that EA ethics inevitably leads to its attempted practitioners experiencing unbearable guilt, and argue that this falsely presupposes both that guilt has any place in a utilitarian ethic, and also that a perfect ideal needs to actually achievable in order to have merit as an ideal. Third, I argue that contra Martin’s argument, it is actually the EA supporter, and not the EA critic, who is more loving and caring towards his neighbour. Fourth, I argue that Martin’s critique of EA fails to adequately come to grips with the fact of opportunity costs in the use of resources, while in contrast EA very naturally and deliberately takes opportunity costs into consideration when making ethical judgements.

Note that the quotes at the beginning of each section are taken from Martin’s original article.

Binary Thinking about Morality

“To be truly objective the maxim, ‘to do the most good we can’ would be binding on all people regardless of whether we believe it or not. Therefore at any point if one is not ‘doing the most good we can’ we are actually acting immorally!”

“Hence justifying simply ‘moving in the right direction’ is inconsistent because it means that you don’t actually need to ‘do the most good we can’. The ethic is reduced to, ‘do the most good you feel you’re able to afford.”

“Effective altruism and the consequentialist ethic of Peter Singer reduces ethics to a kind of communist race to the communal bottom. Everyone is equal and if one person has utility above the lowest, then it becomes unethical.”

“My point is that given the claim of the objectivity of this particular ethical system it becomes immoral to do anything which does not save lives of those in extreme poverty.”

Utilitarian ethics has little place for binary notions like “moral” and “immoral”. At best, these may be useful as heuristics to guide behaviour in the face of uncertainty or insufficient time to fully consider the likely outcomes of a particular action in greater depth. They may also serve as shorthand to be used in particularly extreme cases (murder, robbery, rape, gross abuse, etc). In general, however, utilitarianism considers the morality of essentially all actions to be one of degree: action A is morally preferable to action B insomuch as the expected consequences of A serve to increase total utility more than the expected consequences of action B.

Under such an ethical framework, it makes no sense (other than in the purely heuristic sense as outlined above), to assert in any absolute, unqualified way, that an agent has acted “immorally” when they take an action which produces lower expected utility than some possible alternate action. Rather, what they have done is take an action which does less good than another action they may have performed – no more, and no less.

References to non utility-maximising actions as being ‘immoral’ thus exhibit a misunderstanding of the nature of the ethical claims made by utilitarians. Such statements simply fail to say anything non question-begging with respect to the suitability of utilitarianism as an ethical framework; for in criticising utilitarianism for pronouncing every action other than the very best possible one as being ‘immoral’, they are necessarily importing binary absolutist notions of ‘moral’ and ‘immoral’ which utilitarianism itself rejects. In order to proceed with this line of critique, therefore, it would be necessary to make an argument as to why incorporating such a binary, absolutist notion of ‘moral’ and ‘immoral’ actions is necessarily in order to provide a suitable ethical account. Absent some such plausible account as to why this is in fact the case, however, this line of attack on utilitarianism fails.

Effective Altruism and Guilt

“Ethical altruism has some helpful contributions to make in assessing how scarce resources be allocated, but my criticisms would be less savage if Singer didn’t claim it as an ‘objective’ system. If consequentialism and ethical altruism is objective then we are all condemned under a brutal loveless, ethical system which will lead to social improvement in the developing world but at the cost of an ascetic guilt-ridden hypocrisy.”

“In this ethical framework there is nothing to avoid the slide into a guilt-ridden (how can I ever enjoy chocolate again?) asceticism. Nothing beyond the basics could ever be enjoyed because they would be declared objectively ‘immoral’.”

“There is no forgiveness in ethical altruism, if you eat a chocolate for yourself, you are condemned under the objective guilt of knowing that lives could have been saved elsewhere in the world.”

The argument here seems to be that Effective Altruism is unliveable as an ethical system because it is too demanding, meaning that no one can live up to its dictates, and since no one can live up to its dictates, all those who try will inevitably be subject to a great deal of guilt and anxiety over their perceived moral failings.

My first response takes the form of a question: in what way does this constitute a refutation of EA as an ethical framework? EA says, in essence, that 1) it is morally right to produce as much utility/benefit/happiness/etc as possible, 2) certain courses of action, according to our best evidence, produce much more utility/benefit/happiness/etc than others, therefore 3) it is morally good for us to undertake those courses of action. How is this argument in any way undermined by the fact that it may be difficult, or even impossible, to carry out to its fullest extent? It seems even if the EA ethic is unliveable and tends to produce a great deal of guilt, that in no way casts doubt on any of the statements 1)-3). Thus this objection merely comes down to an assertion that the EA framework is inconvenient for us, as we would rather avoid all the bother and potential guilt. Needless to say, this does not constitute a philosophical argument of any substance for the inadequacy of effective altruism as an approach in applied ethics.

My second line of response is to say that this line of rebuttal seems to presuppose that effective altruism is only valid or relevant as a moral principle if it is possible to be a perfect, completely effective altruist. As far as I can see, this principle is totally unfounded and without any basis. One is a better EA to the degree that one accords one’s actions with EA principles. This is a matter of degree, and not a binary decision. This is hardly a radical concept: essentially all normative systems incorporate ideals that are unattainable in their pure form, but which nevertheless constitute a valuable ideal to strive towards, and to focus our thoughts and efforts around, even if we know we will never reach them. A cook my strive to make “the perfect dish”, even if they know such a thing is in reality impossible. In science, philosophy, and the legal system, we often speak of epistemic virtues like objectivity, rationality, and impartiality. Everyone accepts that such virtues, in their pure, idealised form, can never be achieved by any actual person in any real situation. We do not, however, conclude on that basis that the notions or theories themselves are flawed, or that therefore everyone is everywhere and always being “irrational” or “partial”. We accept that these virtues are only ever be exercised in greater and lesser degrees, and that the impossibility of the actualisation of their perfect ideal form does not somehow undermine the concept in its entirety.

A third line of response would be to point out that notions of guilt have very little relevance to either a utilitarian ethic in general, or an EA framework in particular. Guilt is simply of no interest to the EA supporter, except insomuch as it may be relevant to ethical outcomes, either by promoting giving, or inhibiting action by leading to despair or discouragement. The EA supporter views guilt as a real and important aspect of human psychology which one needs to seriously consider. It does not, however, play any critical or central rule, motivating or otherwise, in a utilitarian ethical theory. As such, it is simply false to assert that a person who chooses an action which yields less than maximal utility is “condemned under the objective guilt”. Likewise the notion of forgiveness – this notion just has no place in a naturalistic, utilitarian ethic. Arguing that the utilitarian/EA ethical framework is defective because it has no place for forgiveness is simply to beg the question against utilitarianism, because precisely the point of utilitarianism is that such notions about binary abolute moral/immoral decisions, guilt, and forgiveness are largely irrelevant to the question of morality, which is instead concerned with degrees of goodness determined by the consequences of different possible actions. A cogent critique of utilitarianism as an ethical theory cannot proceed by simply pre-supposing aspects of morality which utilitarianism itself rejects, as this is to simply beg the question.

Misconstrual of Love

“Indeed love is absent from the brutal consequentialist system advocated by Singer.”

“All good things are to be seen as gifts of God and to be received with thanksgiving (1 Tim 4:4). This means I can enjoy a chocolate cake!”

“Yet the imperatives also broadens the concept of ‘neighbour’ to include not just our global neighbours, but also our local ones, meaning we can build a school hall to the betterment of our local society and love our neighbours with cancer and perform research to help them. Therefore caring for the ‘good’ of our neighbours is achieved through both the Christian ethic and consequentialism, but the Christian ethic is more nuanced and sophisticated.”

The sincere Effective Altruist strives to do as much good for their fellow man as possible, knowing that they will never succeed completely, but always attempting to do better, and endeavouring to use the best reason and evidence available to seek out new and better ways to do the most good with the limited resources at their disposal. They seek to serve as many of their neighbours as possible, not discriminating by race, class, distance, or convenience, but deciding purely on the basis of how much help they can do to their fellow man.

The EA critic, it seems, is content to eat chocolate cake, donate to their local school hall, and then maybe also donate some money to EA charities as well, justifying this to themselves by saying that one could never be truly and completely effectively altruistic anyway, and also by pretending, through various logical contortions, that somehow the resources and time spent on their chocolate cake and local school hall could not have actually been used to help the world’s poor and needy anyway. They seek to serve their neighbour, but with a special preference for neighbours who are conveniently located close by (note: I hope this is not taken as a personal attack against anyone – it is not intended as such, I’m just trying to make a point).

I ask the reader in all sincerity: which now of these two, thinkest thou, was most loving?

Ignoring Opportunity Costs

“If Singer and the effective altruism ethic is correct, then virtually every economic, social and moral choice made in Australia today is ‘immoral’”

“This is because when these decisions are compared with saving lives of people in extreme poverty then on the simple consequentialist metric outlined by Singer, saving lives of those in extreme will always ‘win’ i.e. they will always be morally preferable. Therefore when posed with the question, ‘should we build a new road in Melbourne? The answer under effective altruism will be ‘no, because this money could save lives of people in extreme poverty’. Should I eat a chocolate cake on my birthday? ‘no, because this money could save lives of people in extreme poverty’ Should we build a new school auditorium? Should we treat an injured knee? Should I treat my friend’s cancer? The answer to all these questions is the same – ‘no, because this could save lives of people in extreme poverty’.”

“Moreover other decisions which would have enormously beneficial outcomes for the extreme poor are also rendered ‘immoral’. For example this ethical framework would preclude funding Ebola virus research because the net ‘utility’ of lives saved in developing countries would be greater by providing Malaria nets or immunisation compared with lives saved through Ebola research.”

It is unclear to me what these sorts of statements are attempting to accomplish. If we consider the tripartite core EA argument which I outlined above, which of the three propositions are these arguments supposed to address? They seem to be total non sequiturs. To take the Ebola research example, why would it be a bad thing for EA to recommend that we ought to put resources into bed nets and vaccinations rather than Ebola research, if it is true that the former will save more lives than the latter? Is it because Ebola research will save more lives in the long run, or have other indirect benefits that we haven’t considered? If this is the case, then we have simply denied the premise that vaccinations and bednets will actually do more good than Ebola research, in which case the effective altruist would support the Ebola research as well, so there is no disagreement. On the other hand, if it is agreed that the Ebola research will do less good than vaccinations and bednets, even when factoring in future benefits and side-effects, etc, then what possible justification can there be for preferring the Ebola research over the bednets and vaccinations? How is it a defect of the EA framework for coming to this conclusion?

I wish also to say a few words regarding resource use in developed countries. Taken at face value, the EA ethic would seem to imply that since building roads, medical expenditure – indeed most public expenses of any sort in developed countries – are not as effective uses of funds as donating to the leading EA charities, then we ought not do them. The first point to say here is that it is simply a fact that resources have opportunity costs. Instead of building a new road or paying a doctor’s salary or whatever else, that money could have been used to save lives in the developing world. This is a fact about reality. It has nothing to do with one’s ethical framework, or the worldview one is operating under. Opportunity costs exist, and (needless to say) they don’t go away merely because we don’t like the sound of them, or thinking about them makes us feel uncomfortable about the difficult tradeoffs we must make.

The second point, however, is that it is necessary to exercise some care when making statements like “we should donate money to EA charities rather than build a new road”, because there is in fact no moral agent to which such collective pronouns apply. “We” are not a moral agent; individuals are moral agents. “We” don’t have any money or any ability to choose how it is spent, so it makes little sense to ask how “we” should spend our money as a nation or a community or whatever. What makes sense from a moral framework is to ask how should you and I spend our money, as individual moral agents who can take particular moral actions. So rather than asking what “we” should do, we should be more careful in our thought and speech, and consider exactly who we are saying should do this or that with the resources they have at their disposal.

The third point to make about this comparison is that, as an attempted reductio against EA, it is a very poor one. The reason is because, if EA were applied ‘universally’, or even in a much more systematic way by many more people and organisations, there would be no need at all to redirect money from road building or hospitals (or whatever else) to fund EA charities, because all such charities would already have been fully funded many times over through funding sourced by forgoing other expenses. Every effective charitable cause could be fully funded many times over with the enormous amount of money that could be diverted from non-essential spending by westerners (I leave it to the reader to imagine precisely what is included in this category), without any need to sacrifice truly important things like roads, schools, and hospitals.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s