This essay was the winner of the Graduate category of the 10th National Oxford Uehiro Prize in Practical Ethics 2024. Written by Jakob Lohmar.
Many people believe that we sometimes ought to produce one larger benefit rather than any number of smaller benefits. For example, many believe that in a choice between saving a human life and alleviating headaches, one always ought to prioritize the life, no matter the number of headaches.[1] Call the benefits that (no matter their number) cannot outweigh the saving of one human life “minor benefits” and the wellbeing differences that minor benefits amount to “minor wellbeing differences”. Now consider a hypothetical species with a wellbeing (or “welfare”) range that is as large as such a minor wellbeing difference. That is, the difference between the maximum and minimum level of wellbeing that members of that species can (under any realistic circumstances) achieve equals a minor wellbeing difference.[2] Call species with such minor welfare ranges “low welfare species”. The question that I will discuss in this essay is whether benefits for low welfare species can in principle outweigh even the most significant benefits for humans. Using the saving of a human life as a proxy for the most significant benefit for humans, we can state this question more precisely as follows: Is there any number N and benefit B such that one ought to confer benefit B to N members of a low welfare species rather than save one human life?
The question is not just of theoretical interest. There are many actual species that presumably have welfare ranges that correspond to what we tend to think of as minor wellbeing differences (in the stipulated sense). If they have any welfare at all, invertebrates, small mammals, and many fish, reptiles and birds may well be low welfare species.[3] An answer to my question would accordingly tell us whether benefits for many actual species could in principle outweigh even the largest benefits for humans. This would not be practically relevant if we could help humans more effectively than animals anyway, so that even if benefits for low welfare species could in principle outweigh all benefits for humans, we should prioritize humans under the actual circumstances. But it seems that we can in fact help animals very effectively. That goes in particular for animals in factory farms whose living conditions we could greatly improve. But also the unmeasurable suffering of wild animals provides an opportunity for us to potentially do a lot of good.[4] If benefits for low welfare species could in principle outweigh all benefits for humans, it may well be, therefore, that they (sometimes) actually do outweigh the largest benefits for humans.
Here is a straightforward argument for a negative answer to my question. The welfare range of an individual’s species restricts by how much individuals of that species could be benefited: increasing their wellbeing from the minimum to the maximum wellbeing level of that species is the most that could be done for them. The welfare range of a low welfare species is, by definition, as large as the wellbeing difference that a minor benefit amounts to. Hence, the most significant benefit to individuals of a low welfare species is no larger than a minor benefit. By definition, no number of minor benefits can outweigh one human life. If no number of a certain benefit can outweigh one human life, no number of an equally (or less) large benefit can outweigh a human life either. Thus, no number of even the most significant benefit to members of low welfare species can outweigh one human life.
One way of resisting the argument is to reject the implicit assumption that there are any minor benefits in the stipulated sense: perhaps all benefits can, if numerous enough, outweigh a human life.[5] This is a possibility that should be taken seriously. However, as alluded to in the beginning, such unrestricted aggregation has counterintuitive implications that many find unacceptable. Let us therefore assume, for the sake of the argument, that there are indeed minor benefits no number of which can outweigh one human life. Is the argument sound under that assumption? I think that it is not. We should reject the premise that if no number of a certain benefit can outweigh one human life, no number of an equally (or less) large benefit can outweigh a human life either. The more general principle that motivates this premise is that the comparative magnitudes of benefits determine whether they can be traded off against each other: the smaller benefit can outweigh the larger benefit iff its magnitude is sufficiently close to the magnitude of the larger benefit. This principle, which I will refer to as “Compare Absolute Magnitudes”, has some plausibility. It can explain why headaches can outweigh slightly stronger headaches but not human lives, and why human lives can be outweighed by cases of constant paralysis. However, while the principle gives us the right results in common intra-species comparisons, it fails in cases of inter-species comparisons.
To see this, consider a hypothetical “high welfare species” with a welfare range that is much larger than the welfare range of humans. Perhaps individuals of that species live much longer or are able to experience much more intensely than humans, for example. Assume that we could increase the wellbeing of one individual of that species from a neutral level to the maximum wellbeing level of that species by saving that individual’s life. This seems to be a morally weighty consideration, but it also seems that the greatest goods for humans are still an important moral factor that can be morally decisive. According to Compare Absolute Magnitudes, however, there is no number of any benefits for humans such that one ought to prioritize them, given that the maximum wellbeing level of the high welfare species is high enough. This seems to be the wrong result. Surely, for example, there is some number of human lives such that we should save them rather than the one individual of the high welfare species.
Compare Absolute Magnitudes should be rejected, then. But perhaps there are other (related) principles that don’t share this problem but equally suggest that we ought to prioritize a single human life over any number of any benefits for individuals of low welfare species? In what follows, I will present two arguments that we should reject all such principles. These arguments will also illuminate why it seems wrong to prioritize some benefits for members of high welfare species over any number of any benefits to humans.
The first argument is based on the value of fairness. As far as this value is concerned, it seems plausible that we should give the welfare of members of low welfare species special consideration to compensate for their low natural capacity for welfare. At the least, we should give their welfare considerable weight. There are different reasonable explications of what considerable weight amounts to here. On no reasonable explication, however, would we give their welfare considerable weight if we followed a principle according to which no number of even the largest benefits attainable to them can outweigh a single instance of some human goods, such as one human life. In a world like ours in which we could virtually always use our resources to save human lives with statistical certainty (e.g. by donating to effective charities), this would mean that we should never prioritize them, even if we could help them incredibly effectively. This would be an extreme form of discrimination that is incompatible with any plausible notion of fairness.[6]
The second argument can be considered an argument from empathy. Unlike the argument from fairness, this argument is not based on an inter-species welfare comparison but focuses on the perspective of members of the species with the smaller welfare range.[7] The point is that from their perspective the largest goods attainable to them are significant even if they are much smaller than the goods that are attainable to us. Perhaps, for example, the joy that they can experience is much less intense or much shorter than the joy that we are able to experience. Still, if that is the most joy that they can experience, experiencing it would constitute a highlight in their lives that is worth aspiring to. Or take their lives: from our perspective they may seem insignificant because they have, for example, only a tiny fraction of the length of our lives. But for them, their lives are all that they have, and preventing their lives from being cut short is a worthy goal. None of that is to say that their capacity for welfare is not in fact much lower than our capacity for welfare, or that this difference is morally irrelevant. However, this change in perspective is supposed to show that the goods to low welfare species can have a significance that makes it inappropriate to consider them irrelevant whenever important goods for us are at stake. It would be arrogant and perhaps disrespectful not to count their largest goods at all whenever our largest goods are at stake.
I have argued that we should reject any principle that asks us to prioritize a single human life over any number of any benefits to members of low welfare species. How should we decide, then, in which cases some number of benefits of one kind can outweigh benefits of another kind? Establishing such a principle is beyond the scope of this paper, but I will (tentatively) argue for some features of an acceptable principle that the previous arguments suggest. It seems that instead of comparing the absolute magnitudes of benefits we should compare, at least among others, the magnitudes of benefits relative to some factor that correlates with the welfare range of the respective species. Which factor? The most obvious choice is the species’ welfare range itself. However, I think it is clear that species’ welfare ranges only derivatively matter. At least, they matter only insofar as they correlate with the individual’s welfare range. This correlation is usually strong, but an individual’s welfare range can also significantly come apart from the welfare range of its species, as in the case of some chronic diseases. Benefits to such individuals should not be considered irrelevant merely because they are small relative to the species’ welfare range. It could be suggested, furthermore, that individuals’ welfare ranges only matter insofar as they correlate with the individual’s actual welfare.[8] Whether this suggestion is correct is unobvious to me. It is plausible that the actual welfare of an individual matters, but questionable if an individual’s capacity for welfare (as represented by its welfare range) does not also matter independently.
Whatever the exact factor is relative to which the magnitude of benefits should be measured, I think that a principle on which we should compare such relative magnitudes is promising. Such a principle (“Compare Relative Magnitudes”) would give us the same plausible results as Compare Absolute Magnitudes for common intra-species comparisons since the relational factor has in these cases (per common assumption) the same value for all benefits and therefore cancels out. At the same time, Compare Relative Magnitudes can avoid the problems discussed above for Compare Absolute Magnitudes arising in inter-species comparisons. By comparing relative magnitudes of benefits, it does not discriminate against species with small welfare ranges and does justice to the perspective of members of low welfare species. Most importantly, Compare Relative Magnitudes can explain why the largest attainable benefits for a species, whether for humans or a low welfare species, are always relevant: these benefits have the largest possible relative magnitude.
Since principles that prioritize a single human life over any number of any benefits for members of low welfare species face severe problems, and there is a promising (type of) principle that can explain why some benefits for members of low welfare species are relevant even when human lives are at stake, I conclude that some benefits for low welfare species can in principle outweigh even the most significant benefits for humans.
References
Dorsey, Dale (2009), “Headaches, Lives and Value”, Utilitas, 21(1): 36-58.
Horta, Oscar & Dayron Teran (2023), “Reducing Wild Animal Suffering Effectively”, Ethics, Policy and Environment 26(2): 217-230.
Horton, Joe (2018), “Always aggregate”, Philosophy & Public Affairs 46.2: 160-174.
Kamm, Francis (2013), “Nonconsequentialism”, In H. LaFollette (Ed.), The Blackwell guide to ethical theory. Oxford: Blackwell Publishing.
Kamm, Francis (2015), “Cost effectiveness analysis and fairness”, Journal of Practical Ethics 3 (1).
Norcross, Alastair (1997), “Comparing Harms: Headaches and Human Lives”, Philosophy & Public Affairs, 26(2): 135-67.
Fischer, Bob (2023), „Rethink Priorities’ Welfare Range Estimates”, URL: https://rethinkpriorities.org/publications/welfare-range-estimates.
Scanlon, T. M. (1998), What We Owe to Each Other, Cambridge, MA: Belknap Press of Harvard University Press.
Voorhoeve, Alex (2014), “How Should We Aggregate Competing Claims?” Ethics, 125(1): 64-87.
[1] This case, which is known as “lives for headaches”, is often used to motivate restrictions on aggregation (e.g. Dorsey 2009). If benefits of one kind cannot be aggregated, no number of them can outweigh a more significant benefit. Advocates of restricted (or “partial”) aggregation include, for example, Kamm (2013), Scanlon (1998), and Voorhoeve (2014).
[2] Which circumstances count as “realistic”? This is hard to specify. Surely not all metaphysically possible circumstances, but perhaps also not quite all nomologically possible circumstances. For an approximation, we can think of the minimum wellbeing level as the wellbeing level resulting from a life full of the suffering that the species is capable of, and the maximum wellbeing level as the wellbeing level resulting from a life in full health with (virtually) no unsatisfied needs or desires.
[3] See Rethink Priorities’ pioneering research on the welfare ranges of other species, in particular Fischer (2023). Note that on their definition, welfare ranges measure the momentary welfare rather than the total welfare of species.
[4] See Horta & Teran (2023) for different promising ways of benefitting wild animals effectively.
[5] See, for example, Horton (2018) and Norcross (1997) for arguments that aggregation should not be restricted at all.
[6] Compare this with Kamm’s (2015) concern that a utilitarian approach to global health might have the unfair result that we should save the life of a healthy person rather than the life of a disabled person. This is a valid concern, but there seem to be reasonable notions of fairness that are compatible with the utilitarian verdict. By contrast, it is clearly unfair to prioritize one healthy life over any number of disabled lives. Analogously, it is unfair to prioritize one human life over any number of lives of low welfare species.
[7] This does not need to be their actual perspective. We can consider this perspective a hypothetical one which it would be appropriate to have given the welfare structure of members of a low welfare species. My argument therefore does not require that individuals of low welfare species have the cognitive capacities to contemplate the significance of their goods.
[8] This would make the principle similar to how contractualists usually test the relevance of claims (e.g. Voorhoeve 2014). A contractualist framework might therefore have similar results if, against contractualist orthodoxy, all animals are assigned claims as well.