Capacity for Welfare Across Species 0192882201, 9780192882202

Table of contents :
Dedication
Acknowledgements
Contents
1. Introduction
2. The Difference-View
3. The Equality-View
4. Welfare across Time
5. Practical Implications
Bibliography
Index

Citation preview

Capacity for Welfare across Species

Capacity for Welfare across Species TAT JA NA V I Š A K

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Tatjana Višak 2022 The moral rights of the author have been asserted First Edition published in 2022 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2022941775 ISBN 978–0–19–288220–2 DOI: 10.1093/oso/9780192882202.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

To Miko

Acknowledgements Special thanks for helpful discussions and support throughout the years goes to Roger Crisp, Jeff McMahan, and Peter Singer. I am also grateful to everyone else—colleagues, friends, and my husband and kids—who showed interest in me and this book project. Thank you, Verena, for going through the whole manuscript and checking my English! The book profited from the valuable feedback from two anonymous referees and the helpful suggestions from my OUP editors. I dedicate this book to my dog Miko, who, familiar and mysterious, has been present throughout the whole writing process and has taught me a lot about the topic.

Contents 1. Introduction 1.1 Comparisons of Welfare across Species 1.2 Assumptions 1.3 The Difference-View 1.4 The Equality-View 1.5 Theories of Welfare 1.6 What Ought We to Do? 1.7 Outline of the Book

1 2 4 8 11 13 16 19

2. The Difference-View 2.1 Kagan’s Objective List Account of Welfare 2.2 Singer’s Whole Life Preferentialism 2.3 Mill on Higher and Lower Pleasures 2.4 McMahan’s Notion of ‘Fortune’ 2.5 Wong on Experiential versus Absolute Welfare 2.6 Budolfson and Spear’s Formula

21 21 28 31 36 40 54

3. The Equality-View 3.1 Animal Welfare Science 3.2 Welfare as Self-Fulfilment 3.3 An Evolutionary Perspective

60 60 65 74

4. Welfare across Time 4.1 Harm of Death 4.2 The Relevance of Different Lifespans 4.3 An Application of the Total-Duration View

88 88 96 99

5. Practical Implications 5.1 Moral Status 5.2 Promoting Welfare by Creating Welfare Subjects? 5.3 (Dis-)Counting Animals

112 113 130 137

Bibliography Index

143 151

1 Introduction Assessments and comparisons of individuals’ welfare are common in everyday life. We answer questions about how we are doing. We enquire about the welfare of others and do or do not believe the answers that we get. We compare welfare across time. It might, for instance, worry me that, due to some illness, my friend is worse off these days than she was last month. I  might, furthermore, wonder whether my dog is better off in my family than he used to be when he was still living in the streets of Romania. We also compare welfare across individuals. For instance, a father’s judgement that his daughter is currently worse off than his son might motivate him to give her some special attention. I might wonder whether my dog, with his playful and carefree life, is better off than I am. The latter would be a comparison of welfare not only across individuals of the same species but across individuals of different species. This book is mainly concerned with a central question related to cross-­ species comparisons of welfare. The book inquires whether welfare subjects of different species have the same rather than a different capacity for welfare. Without an answer to this question, welfare cannot be compared across species. There is currently no agreement and only a little discussion about the important question of cross-­species capacity for welfare. Before I further introduce the issue of cross-­species capacity for welfare (sections  1.3 to  1.6) and present an outline of this book (section  1.7), I will illustrate the practical relevance of comparisons of welfare across species (section 1.1) and mention some assumptions that underly such comparisons (section 1.2).1

1  Parts of chapters 1.1 and 1.2 are reproduced with permission from Palgrave MacMillan from Višak (2017).

Capacity for Welfare across Species. Tatjana Višak, Oxford University Press. © Tatjana Višak 2022. DOI: 10.1093/oso/9780192882202.003.0001

2 

Capacity for Welfare across Species

1.1  Comparisons of Welfare across Species Comparisons of welfare across species are of practical relevance. They may, and often do, inform considerations of consumers, policymakers, activists, and ethicists. Let me mention a few examples. Policymakers formulate animal welfare considerations in species-­specific rules. Since what one may or may not do to a member of an animal species is supposed to be related to the impact on the welfare of the individual in question, cross-­species comparisons seem to be necessary. For instance, in the Netherlands it is legally permitted to take calves but not piglets from their mother immediately after birth. Does this suggest that such a sep­ar­ ation is more harmful for piglets than it is for cows? Most probably, the reason for this difference is the economic interest in milking the cow but not the sow. But since such regulations are part of animal welfare legislation, one might justify or question the difference in terms of the animals’ welfare (Bracke 2006). Likewise, the Dutch decision to ban mink farms on animal welfare grounds may raise the question whether minks are worse off than, for ex­ample, hens in cages. Again, other considerations may explain the ban on mink farming, such as the public opinion that fur, as opposed to eggs, is a luxury product. But if a certain treatment is forbidden for one species but not for others in animal welfare legislation, one might justify or question such rules by an appeal to cross-­species comparisons of welfare (Bracke 2006). There are many examples of possible appeals to cross-­species comparisons of welfare: consumers, for instance, may decide which animal products to consume or to boycott based on such assessments. One might, for instance, reject the use of meat, but accept the use of wool based on a comparison of the welfare of the animals that are used for these products. Here as well, other considerations, such as the availability of alternatives to the products in question, may play an important role. Still, animal welfare considerations may at least be part of such a decision. Similarly, animal rights or animal welfare activists may prioritize their campaigns based on cross-­species comparisons of welfare. For instance, they may decide to campaign against the use of great apes in invasive biomedical research or they may campaign against the use of elephants in circuses, rather than against the use of other animal species that are expected to suffer less in these practices. Again, other considerations may be relevant, for instance the expected success of the campaign. I do not here want to defend the correctness of any of these cross-­species comparisons of welfare. The examples are just meant to illustrate their practical relevance.

Introduction  3 Ethicists as well appeal to comparisons of welfare across species. For instance, utilitarians aim at maximizing welfare. They need to know how to make the best use of their resources, such as time and money. So, they need, at least in principle, to compare the effects of various actions on welfare and this often involves comparing welfare across species. The same holds for effective altruists, who are not necessarily utilitarians, but who want to make the greatest positive difference with the time and money that they devote to charitable causes. Prioritarians, in turn, hold that the worst-­off individuals deserve our special attention and thus they need to determine how individuals fare as compared to each other. Egalitarians also care about how welfare is distributed and strive for an equal distribution of welfare. Thus, they cannot do without cross-­ species comparisons either. Other moral theories as well are likely to encounter situations where such comparisons are asked for. From a normative perspective, comparisons of welfare across species are important just in case the welfare of individuals from different species is morally relevant. That both human and non-­human animal welfare matters for what we have reason to do is widely accepted among ethicists and ­people in general. It is an assumption that I take for granted in this book. (In section 5.1 I take up the question how much the welfare of various species of animals matters.) However, the question about the capacity for welfare of different species as well as the issue of cross-­species comparisons of welfare can be considered interesting, independently of its relevance for what we have reason to do. Even though it is easy to see the practical relevance of cross-­species comparisons of welfare, surprisingly little has been said so far about how exactly to compare welfare across species. Mark Budolfson and Dean Spears (2020, 606) characterize what they call ‘the challenge of interspecies comparisons’ of welfare: Animal welfare is almost never included in policy analysis [. . .] increasingly also because we do not currently have good methods for quantifying animal wellbeing consequences and putting them on the same scale as quantified human wellbeing consequences.

Thus, Budolfson and Spears suggest that even if we wanted, for example, to assess the harm of climate change for both humans and non-­humans, we are lacking any principled guideline for how to do it. Not merely are we unsure about certain details, but we may have no clue about how to compare welfare across species even in principle. According to Budolfson and

4 

Capacity for Welfare across Species

Spears (2020, 606), it is possible and relatively easy for animal welfare scientists to find out how well off an individual human or non-­human welfare subject is relative to its potential or capacity for welfare. Even lay people may be able to assess, for example, whether a particular dog is leading a good or a bad life for a dog. They can make up their minds about the question to what extent the dog fulfils his potential for welfare. The much harder question, according to Budolfson and Spears (2020, 606), is getting grip on the relative potential or capacity for welfare of different species: [. . .] the most difficult problem that needs to be solved in connection with interspecies comparisons, [. . .] is how to estimate the wellbeing capacity (wellbeing potential) of members of a non-­human species relative to the wellbeing capacity of humans.

One may add that not only do we need to know how another species’ potential for welfare relates to that of humans, but we also need to compare the welfare potential of various non-­human species with each other. We need a grip on capacity for welfare across species if we are to count the welfare of different species on the same scale. This book aims at addressing this challenge and thus at illuminating what the relative capacity or potential for welfare of individuals from different species is. Comparisons of welfare will differ enormously, depending on what answer to the question about cross-­species capacity for welfare one gives. Before I introduce the main positions about cross-­species capacity for welfare, I will pause to make explicit some assumptions behind comparisons of welfare across species.

1.2 Assumptions Comparisons of welfare across species are based on various assumptions. First, it is assumed that animals of different species have a welfare level. This does not mean that all animals are subjects of welfare. Whether any animal is a subject of welfare depends on which theory of welfare is correct as well as on empirical facts about the animal in question. For instance, according to hedonism the amount of pleasure—­usually understood broadly as enjoyment minus suffering—­in an individual’s life determines how well off the individual is. According to this account of welfare, only those who can experience enjoyment and/or suffering are subjects of welfare. It is an empirical question what animals are capable of enjoyment and suffering. According

Introduction  5 to most accounts of welfare, non-­sentient animals are not welfare subjects. If any of these accounts of welfare is correct, we need to know which animals are sentient if we are to determine who qualifies as a subject of welfare. This is a question that empirical scientists need to answer, keeping in mind, though, that the information that scientists provide (e.g. about observable behaviour or brain structures) and the one that ethicists require (e.g. about subjective experiences) does not always fit together in a straightforward way. Scientists who work on demarcating the boarders of sentience often focus on the detection of pain. Pain needs to be distinguished from mere nociception, which is ‘the non-­ conscious detection of tissue damage’ (Browning 2019, 11). It is controversial how to distinguish the capacity of feeling pain from the capacity for nociception. Scientists propose different criteria (e.g. Sneddon et al. 2014). There is a broad consensus among scientists that at least all mammals and birds are sentient. These animals possess a similar nervous system and brain structure as can be found in humans and show similar reactions to stimuli (Low et al. 2012, 2; Proctor 2012). Several scientists argue in favour of sentience in fish, based on alleged behavioural evidence (Proctor 2012; Braithwaite 2010). Others are unconvinced and point out that fish lack the allegedly required brain structures and connectivity (Key 2016; Cabanac 2009,  1998). Fish may process pain in a structurally different way than mammals (Sneddon et al. 2014). The issue is unsettled, and different capacities may be available to different fishes (Allan 2011). However, there seems to be a growing consensus in favour of fish sentience and even in favour of sentience for some invertebrates, such as cephalopods and decapods (Birch et al. 2021). Whether animals such as snails, earthworms, or insects may be sentient is more doubtful. These animals also lack the structures that prod­ uce sentience in mammals (Proctor 2012). Nevertheless, some scientists hold that the function of sentience does not necessarily require these structures and that the animals in question seem to meet some behavioural or physiological criteria for sentience (Sneddon et al. 2014; Jones 2013; Ginsburg & Jablonka 2019, 351). On the other hand, which criteria they meet, if any, is still controversial and meeting only some criteria may not be enough. Currently, we do not have strong evidence for the sentience of these animals and the scientific consensus in favour of the possession of certain ethically relevant properties can be expected to be weaker than ethicists make it seem (Allen 2006). Since most animals are in fact insects, which might not be sentient, most animals may not be subjects of welfare. But nevertheless, many of them—­and certainly mammals and birds—­are.

6 

Capacity for Welfare across Species

A second assumption behind cross-­species comparisons of welfare is that an animal’s welfare level can in principle be measured on a (semi-) quantitative scale. Thus, we can at least rank an individual’s welfare level as low, medium, or high (Bracke 2006). It seems that we can describe welfare levels as negative, neutral, or positive and we can even provide estimates of welfare scores, such as 8 out of 10. A third assumption is that the welfare of different animals, including animals from different species, can be measured, and compared on a single scale. This means that we are talking about the same thing, welfare, in case of both humans and non-­human animals (Bracke 2006). Are these assumptions warranted? It seems that they are. After all, laypersons and experts alike have been found to compare welfare across species without hesitation, apparently using a common scale across species (Bracke 2006, 63). Even legislation, for instance about animal experimentation, requires weighing the discomfort of different species on a common scale. Furthermore, we use one word, ‘welfare’, for different species. Lastly, as we saw, many moral theories require cross-­species comparisons. These metaphysical, psychological, linguistic as well as political and ethical perspectives suggest that it is plausible to assume that these comparisons make sense in principle (Sandøe 1996). There is some controversy about the possibility of cross-­species comparisons of welfare based on preferentialism, which is a specific theory of what welfare consists in (Broome 1998; Scanlon 1991). But this discussion concerns possible difficulties for this theory of welfare, and not the impossibility of these comparisons in general. Nevertheless, if preferentialism were the correct theory of welfare and couldn’t account for cross-­species comparisons of welfare, the latter would not be possible. But this can be con­ sidered an implausible implication of preferentialism and thus an argument against the plausibility of this account of welfare rather than an argument against the possibility of comparing welfare across species. I am inclined to think that these comparisons are possible, even based on preferentialism. I will not discuss the alleged problem of preferentialism with cross-­species comparisons of welfare in this book. Instead, I go along with the widely accepted view among ethicists today, according to which these comparisons face no principled objections. Even though I assume that there is no principled objection to comparisons of welfare, there can, of course, be practical challenges. Welfare cannot be directly measured. To measure welfare, we first need to settle on a theory of welfare that tells us what welfare is. (Section 1.5 introduces theories of

Introduction  7 welfare.) Whatever welfare consists in is likely to be something that cannot be directly measured either. For example, assume that welfare consists in having pleasant mental states. We cannot directly measure how pleasant a mental state of an individual is, since we do not have direct access to other individuals’ experiences. In some cases, we may be able to ask the in­di­vid­ uals about their amount of felt pleasure and take their responses as indicators of their felt pleasure and thus of their welfare. In addition, or alternatively, we may find other indicators, such as behaviour, heart rate, or brain activity. Whatever indicators we chose, we can never be sure whether the same score of two individuals regarding the indicator really goes together with the same subjective experience. Assume that I offer two dogs the same treat and that they show different behavioural and physiological reactions. Does the one with the faster heart rate and tail wagging enjoy the treat more? Or is this animal merely showing a stronger reaction to the same subjective experience? How can we tell? As Browning (2019, 138) points out: Within-­ species differences in individual behavioral and physiological responses to positive and negative stimuli are common [. . .] and it is difficult in these cases to determine whether or not results imply a welfare difference.

This is what Browning (2019, 136) calls ‘the comparison problem’. Browning’s challenge does not concern the principled, theoretical, issue of comparisons of welfare across species. It concerns the identification of indicators for doing so. Those who are pondering the question whether the different heart rates of the two dogs upon receiving my treat indicate real differences in felt pleasure or merely different physiological reactions based on the same amount of felt pleasure, are already assuming that the dogs feel pleasure and that their felt pleasure, whatever it is, grounds their welfare. They are, thus, not casting doubt on the principled comparability of welfare. They are only pondering how we can practically do it. I say ‘only’ from a philosopher’s perspective here. For animal welfare scientists, the practical question of how to do the comparisons and what indicators to rely on is, of course, a major issue. This practical challenge concerns comparisons within species and even more so between species. This practical challenge of what indicators to rely on will not be addressed in this book (see Browning 2019). In one of the classical contributions to the philosophical discussion about welfare, Wayne Sumner (1996, 14) introduced what he called the

8 

Capacity for Welfare across Species

‘four cardinal virtues’ for theories of welfare. One of them, the virtue of generality, says that a theory of welfare should be applicable to all beings with welfare: We make welfare assessments [. . .] concerning a wide variety of subjects. Besides the paradigm case of adult human persons, our welfare vocabulary applies just as readily to children and infants, and to many non-­ human beings. It is perfectly natural for me to say that my cat is doing well, that having an ear infection is bad for her, that she has benefited from a change of diet, and so on. In making these judgments it certainly seems to me that I am applying exactly the same concept of welfare to my cat that I habitually apply to my friends. A theory of welfare will therefore [. . .] be incomplete if it covers only them and ignores her.  (Sumner 1996, 14)

I will take it for granted in this book that comparing welfare across species is possible in principle. Indeed, theories of welfare should be able to handle such comparisons. If we are to determine how to compare welfare across species, we cannot avoid the issue of cross-­species capacity for welfare, to which I will turn in the remainder of this book.

1.3  The Difference-­View Before getting into the philosophical details and controversies concerning capacity for welfare across species, it is worth asking what a common-­sense approach to the issue would be. Shelly Kagan describes and defends a view, which he thinks is widely accepted. According to this view, people are generally better off than animals (and chimpanzees, than mosquitos); they have lives of higher quality, providing greater benefits to the individuals whose lives they are. Putting the point in slightly different terms, by virtue of the tremendous differences in their capacities for welfare and flourishing a person generally has a significantly higher level of well-­being than a dog, say, or a cow; and a dog, in turn, has a significantly higher level of well-­being than a snake or a fish. It is because of facts like these that a loss of life, for example, harms a person more than it harms a mouse. The person would normally be incomparably better off than the mouse; so what she stands to lose is incomparably greater as well. (Kagan 2019, 46)

Introduction  9 ‘People’ or ‘persons’, as Kagan uses the term, are individuals with certain higher cognitive capacities, no matter of which species. Normal human adults are people in that sense. In Kagan’s terminology, very young humans are not people yet and some humans with severe cognitive disabilities may never be people. Meanwhile, some non-­humans, such as, perhaps, great apes or cetaceans may well be people. Thus, when Kagan contrasts people to animals in the above quotation (and in other passages that I will quote), ‘animals’ refers to non-­human animals that lack the capacities that are required for personhood. It is an empirical question what capacities particular animals have. Kagan’s view implies that animals with lower cognitive capacities are generally worse off than people. Kagan somewhat confusingly says that people are ‘incomparably’ better off, but this is not to say that their welfare cannot be compared on the same scale. Kagan assumes that the welfare of in­di­vid­ uals from different species can be compared on the same scale. It’s just that, according to this view, people are capable of, and generally have, much more welfare than dogs, which, in turn, are capable of, and generally have, much more welfare than mice. Kagan (2019, 46) assumes that ‘most people’ would find this view ‘uncontroversial’. Indeed, Kagan is not its only proponent. Even though I doubt his claims about the wide agreement with this view among the public, it is a prominent and perhaps the dominant view among philosophers who are writing on the issue. For example, Kagan (2019, 47) approvingly quotes Kitcher (2015, 298), who claims that a fly’s life doesn’t amount to a lot, even at the best of times.

Furthermore, Kagan (2019, 47) approvingly quotes Frey (1996, 13), who claims: The fullest mouse life there has ever been, so science would seem at the moment to suggest, does not approach the full life of a human; the difference in capacities, and what these additional capacities make pos­sible by way of further dimensions to human existence, is just too great.

John Stuart Mill (1963/2002, 13) famously exclaimed that ‘it is better to be a human being dissatisfied than a pig satisfied, better to be Socrates dissatisfied than a fool satisfied’. According to this common philosophical view different welfare subjects have different capacities for welfare, based on different cognitive capacities.

10 

Capacity for Welfare across Species

Some authors focus on emotional, rather than cognitive, differences between species. They may refer to the capacity for having more complex emotions, or more intense pleasures and pains. Typically, those who focus on emotional differences, or on hedonic differences, relate them to cognitive differences. Thus, they may believe that the complexity of one’s emotional life is linked to one’s cognitive complexity or that a greater number of  brain neurons allows for a broader range of enjoyment and suffering. However, in principle, cognitive and emotional capacities may come apart. For example, individuals may have a certain complexity of neural affective pathways or concentrations of neurotransmitters without having complex cognitive capacities. (I will discuss hedonic capacity in detail in sections 2.5, 2.6, and 3.3.) I will call the view, according to which different cognitive or emotional capacities are the basis for different capacities for welfare among welfare subjects the ‘difference-­view’ about capacity for welfare across species, DIF for short. DIF:  Welfare subjects, such as humans, dogs, and mice, have fundamentally different capacities for welfare, due to their different cognitive or emotional capacities. Kagan (2019, 46) considers an objection to DIF. According to this objection, the view expresses a mere prejudice that humans tend to have in favour of the value of their own lives. If we asked a dog or a squirrel, each of them would possibly rank the lives of its own species as the best. To this objection Kagan replies that if those animals could contemplate comparisons of welfare across species, they too would be people. Kagan also claims that the mere fact that only humans express the above-­mentioned view about the value of lives doesn’t disqualify it as prejudiced. I agree with both observations. Nevertheless, as Kagan admits, humans should take care to avoid prejudice and parochialism and should be open to new and developing empirical insights about the lives and capacities of non-­humans. Kagan (2019, 47) considers it ‘extremely unlikely’, though, that we will discover any empirical facts about the lives of animals that make us abandon the judgement that our lives are more valuable than theirs. I also do not expect that new insights about the capacities of animals will be likely to cause that change of mind. Yet, I seriously doubt the line of reasoning that leads Kagan to conclude that people are generally better off than dogs and that dogs are generally better off than mice. The problem with DIF, as I see it, is not the underlying assumptions about different cognitive or emotional capacities of

Introduction  11 different species of animals. I am happy to grant that people have a range of cognitive and emotional capacities that dogs and mice lack. However, I doubt their relevance for welfare. I hope that the philosophical investigation along with some empirical findings that I present in this book will show that DIF is on weaker ground than Kagan and likeminded authors make it seem.

1.4  The Equality-­View Even though DIF is a possible view about cross-­species capacity for welfare and apparently prominent among philosophers, not everyone seems to agree with it. It is particularly remarkable, I think, that animal welfare scientists hold views that may be incompatible with DIF. This is not to say that animal welfare scientists take an explicit stance on cross-­species capacity for welfare. They don’t. But what they have to say about animal welfare seems to be in tension with DIF. Animal welfare science is concerned with the scientific study of the welfare of animals in different settings, such as in zoos, in laboratories, on farms, as pets and in the wild. The world’s first professorship in this discipline was appointed by Cambridge University (UK) in 1986. Animal welfare science aims at investigating animal welfare using rigorous scientific methods. Nevertheless, in their conceptualizations and theories of welfare, animal welfare scientists enter the field of philosophy, specifically value theory. After all, to know what to measure and compare scientifically, they first need to settle on what welfare is. (I will have more to say about the views of animal welfare scientists in section 3.1.) Here is an example of an animal welfare scientist’s assessment of welfare. Mark Bracke (2006) compares the welfare of different species of animals that are kept in different husbandry systems. Among others, he compares the welfare of laying hens in battery cages and dairy cows on pasture. As depicted in Table 1.1, Bracke provides a list of needs—­such as food, water, and rest—­that the animals have. (Bracke adds many other needs that are not shown here.) He uses a ranking scale from 1 (worst) to 4 (best) that indicates to what extent each need is fulfilled for the animal in question. In Table 1.1, we see that the need for food and water is maximally fulfilled for both species in both systems. The need for rest is not entirely fulfilled in case of battery hens, since they have no perch and only a wire floor, which is not what they need to rest appropriately. Adding the scores for all needs gives us an animal’s welfare score. To refine the assessment, one can give more weight to more important needs, as Bracke (2006) points out.

12 

Capacity for Welfare across Species Table 1.1  Fulfilment of needs as indicator of welfare. Need

Laying hens in battery cages

Dairy cows at pasture

Food Water Rest …

4, ad libitum 4, clean nipple, easy to reach 2, no perch; wire floor …

4, ad libitum 4, clean trough in pasture 4, proper surface, pasture …

This results in a welfare score for dairy cows on pasture and for laying hens, which is the average score of the overall need fulfilment for each animal. These welfare scores can be directly compared. In case of cows on pasture the average score, based on all relevant needs that Bracke included (and that are not all shown here), was 3.19. In case of battery hens, it was 2.16. Bracke concludes, among other things, that dairy cows at pasture are better off then battery hens and he can explain why this is the case: they can fulfil their needs to a greater extent. Something along these lines seems to be a common way of comparing welfare across species among the public, even if it is not always done as explicitly as Bracke does it here. I assume that Bracke would have used the same procedure in comparisons of welfare between mice, dogs, great apes, and all other welfare subjects. This view on comparing welfare across species implies that welfare subjects have an equal capacity for welfare, despite different cognitive or emotional capacities. I will therefore call it the ‘equality-­view’ about capacity for welfare, EQU for short. EQU:  Welfare subjects have fundamentally equal capacities for welfare, despite their different cognitive or emotional capacities. So, we’ve got two opposed views: DIF and EQU. Even though I have presented only rough sketches of these views so far, it should be clear that they may have wildly different implications. According to DIF, even a barely happy human may be much better off than the happiest of pigs. Meanwhile, EQU implies that a barely happy human is probably worse off than the happiest of pigs. The barely happy human may have many unmet needs, while the very happy pig may be completely satisfied. Throughout this book, I will have much more to say about each of these approaches to find out which of them is more plausible. I will ultimately reject DIF and defend EQU. My rejection of DIF will not be restricted to Kagan’s version of it, but I will also discuss and reject other defences of such

Introduction  13 a view. I will argue in favour of a view along the lines of EQU, even though it need not be the one that Bracke proposes. Before I present the book’s outline, it will be helpful to introduce the notion of welfare and point out how welfare is discussed in the philosophical literature.

1.5  Theories of Welfare Substantive theories about welfare, i.e. philosophical theories about what welfare consists in, are concerned with personal good, which refers to the goodness of lives for the individuals whose lives they are. They aim at specifying what makes someone’s life good for this individual. A lot of things can be good for me, such as food and shelter. However, theories of welfare are not interested in instrumental goods, but in basic goods. So, for example, one would need to ask: What is it about food and shelter that makes these things good for me? Perhaps, for example, they ultimately make me happier. If so, happiness would be the basic (or final) good and food and shelter would be only instrumental goods (Zimmerman 2015). Theories of welfare are interested in basic personal goods and when I talk about goods in this book, I refer to basic goods, unless otherwise indicated. Theories of welfare have four tasks and answer one central question related to each of these tasks. The four tasks of theories of welfare are: 1. Identification: What states of affairs are goods? 2. Measurement: How good is each good? 3. Analysis: What are the kinds of goods (or basic goods) under which the goods can be subsumed? 4. Explanation: What is the good-­making property in virtue of which goods are good for the individual? (see Lin 2014) Table 1.2 provides an overview of various theories of welfare and what they view as particular goods, basic goods, and good-­making properties. A hedonist, for instance, typically holds that instances of S taking pleasure in something are the particular goods or good tokens. All instances of S taking pleasure in something are good. Thus, S taking pleasure in something is the good kind. The goods are good in virtue of S’s experiencing pleasure. So, S’s experiencing pleasure is the good-­making property. A preferentialist may hold that states of affairs that S desires are goods. The good kind consists in being a state of affairs that S desires. Any such state is good in virtue of S’s desiring this state.

Table 1.2  An overview of some accounts of welfare and what each of them views as particular goods, good kinds, and good-­making properties. Account of welfare

Particular goods/good tokens: The following particular goods are good for S:

Good kind(s): Good-­making property/ All of the following goods are good for S: properties: The particular goods are good for S in virtue of:

Hedonism

Particular instances of S taking pleasure in something Particular states of affairs that S desires to have Particular instances of S realizing knowledge, spirituality and mathematics

All instances of S taking pleasure in something All states of affaires that S desires to have

Preferentialism Pluralist account with knowledge, spirituality, and mathematics on the list of goods (Objective list account of welfare) Pluralist account with pleasure and desire-­satisfaction on the list (Subjective list account of welfare) Nature fulfilment theory Self-­fulfilment theory

S’s experiencing pleasure

All instances of S realizing knowledge, spirituality or mathematics

S’s desiring these states of affairs S realizing knowledge, spirituality or mathematics

Particular instances of S experiencing pleasure or satisfying a desire

All instances of S experiencing pleasure and all instances of S satisfying a desire

S experiencing pleasure or satisfying a desire

Particular instances of S fulfilling her nature Particular instances of S realizing her self-­fulfilment

All instances of S fulfilling her nature

S fulfilling her nature

All instances of S realizing her self-­fulfilment

S realizing her self-­fulfilment

Introduction  15 I will not defend any theory of welfare in this book. Instead, I will point out what follows from the main theories of welfare about the capacity for welfare of individuals from different species. When philosophers present and discuss substantive theories of welfare, as the ones presented in Table 1.2, they need to have at least a rough understanding of what ‘welfare’ means and thus of what the theories of welfare are theories about. In other words, they need a conceptualization of welfare. Most philosophers who write about welfare may not have a detailed concept of welfare in mind. Instead, they rely on a rough picture of what ‘welfare’ refers to. This rough picture includes, for example, the idea that welfare is related in specific ways to concern: if I am concerned about myself or others I am concerned about their welfare. Welfare, according to the rough picture, is also related to pity and to envy: I pity those who are doing poorly and envy those who are doing well. Furthermore, the rough picture of welfare appeals to the idea that sacrifices involve reductions of welfare. As I said, many philosophers take this rough picture to provide a sufficiently clear idea about the concept of welfare (Campbell 2016). Other philosophers present detailed and explicit conceptualizations of welfare. Conceptualizations of welfare aim at telling us what it means to say that something, p, is good for some individual, S. So, in contrast to substantive theories of welfare, conceptualizations of welfare do not make pro­posals as to what is good for S, but they just propose what ‘being good for S’ means. For example, the rational care analysis of welfare holds that ‘p is good for S’ means ‘if S is worthy of care, then there is reason to desire p out of care for S’. According to the locative analysis of welfare ‘p is good for S’ means ‘p is good simpliciter and is located in S’s life’. According to the positional ana­ lysis of welfare ‘p is good for S’ means ‘p contributes to the desirability of being in S’s position’. Lastly, according to the suitability analysis ‘p is good for S’ means ‘p is suitable for S in that it serves S well’ (Campbell 2016). I will not commit to a conceptualization of welfare in this book, but this little overview should suffice as a very rough indication of what the concept refers to. Accounts of welfare, such as those presented in Table 1.2, are commonly divided into those that consider welfare to be attitude-­dependent and those that consider welfare to be attitude-­independent. It is often claimed that this distinction tells us something deep and significant about the characterization and evaluation of accounts of welfare (Sumner 1996; Dorsey 2017, 199). In the relevant literature, attitude-­dependent/attitude-­independent distinctions go under different names, such as subjectivist/objectivist or

16 

Capacity for Welfare across Species

internalist/externalist distinctions (Singer & Lazari-­Radek 2014, 214–215; Lin 2014, 148, note 6; Sarch 2011, note 1; Haybron 2008a, 193–95). In fact, theories of welfare can be attitude-­dependent or independent in different ways (Fletcher 2013). While it is common to divide the field of accounts of welfare based on attitude (in)dependence, different authors seem to have different ideas about what the distinction is supposed to track and how best to draw it. In this book, I will leave these characterizations uncommented. For example, I will follow the common usage of calling some theories of welfare ‘objective list accounts’, as indicated in Table  1.2, without further commenting on what, if anything, is particularly ‘objective’ about them. We will, however, find out that different theories of welfare have different implications for the issue of cross-­species capacity for welfare. This can also be an alternative way of characterizing accounts of welfare: they may be more in line with DIF or rather with EQU. This book’s exploration of cross-­species capacity for welfare will have theoretical implications for the philosophical discussion about welfare. It is, after all, still unclear what the most plausible theory of welfare is. There are no knock-­down arguments against one theory or the other. Instead, evalu­ ations of the various theories rely on a careful assessment of each theory’s implications. For each theory, it needs to be assessed how plausible its implications are. This should include each theory’s implications on the issue of cross-­species capacity for welfare. Thus, this book’s exploration of cross-­ species capacity for welfare will inform evaluations of the plausibility of philosophical theories of welfare.

1.6  What Ought We to Do? In this book I defend EUQ: the view that welfare subjects of different species have an equal capacity for welfare. EQU has implications for comparisons of welfare across species, which, in turn, are likely important for what we have reason to do. Many readers may be eager to know: What follows from EQU (or indeed from DIF) for how we ought to act? This depends, of course, on the importance of welfare considerations for how we ought to act. It is an open question whether what we have reason to do is based on what is valuable. If it is, then to know what we have reason to do, we need to  know what is valuable. I take it to be uncontroversial that well-­being is a value. Many ethicists believe that well-­being or welfare is a value that grounds normative reasons for action. I agree. This means that at least

Introduction  17 sometimes facts about welfare are relevant for what we have normative ­reason to do. When I speak about what we ‘should do’ or ‘ought to do’, I am talking about what we have, all things considered, normative reason to do. I believe that the effects of our actions on welfare are at least part of what determines what we should or shouldn’t do. For example, the fact that smoking is bad for Tom may count, for Tom, against smoking. If this is so, the fact that smoking is bad for Tom is a prima facie reason for Tom not to smoke. The fact that Tom enjoys smoking may, in turn, be a prima facie reason for Tom in favour of smoking. There may be different prima facie reasons of different strengths, some of which may count against smoking and some in favour of smoking. These reasons at least partly come down to effects on Tom’s well-­being, and perhaps also to effects on the well-­being of others. Whether Tom has all things considered reason to smoke depends on the balance of these prima facie reasons. There are different theories about the nature of normative reasons for action and, for example, the relationship between normative and motivational reasons. Normative reasons for action concern what I am justified in doing. Motivational reasons for action, in contrast, concern my motivation to act in some way. These are, at the very least, different concepts. Perhaps one can have normative reasons for action independently of one’s actual motivation to act in this way. For example, Tom may have an all-­things-­considered normative reason to quit smoking, even though he lacks the motivation to do so. I do not take a stance on the relationship between normative and mo­tiv­ ation­al reasons for action in this book. I assume, though, that welfare is at least part of what constitutes normative reasons for action.2 Without this assumption, comparisons of welfare across species would not be practically relevant but could still be of theoretical interest. Welfarists about normative reasons for action hold that welfare, and only welfare, is the basis for normative reasons for action. There are different kinds of welfarists about normative reasons for action. For example, egoists about normative reasons for action hold that only the consequences for one’s own welfare determine what one has reason to do. Egoists are, thus, partial when it comes to the promotion of welfare.3 Utilitarians, on the other hand, believe that the effects of one’s actions on the welfare of all

2 For more about the relationship between normative and motivational reasons, see Alvarez (2017). 3  Egoism about motivation is a different kind of theory, which answers a different question. It is about what motivates us and not about what we have normative reason to do.

18 

Capacity for Welfare across Species

affected individuals must be taken equally into account. According to them we have reason to do what is impartially best. There are also proponents of dual-­source views, which accept both partial and impartial welfare-­based reasons for action. Within utilitarianism, act utilitarians evaluate actions as right just in case they impartially maximize welfare. Rule utilitarians, in contrast, consider actions right if and only if they are in line with an optimific set of rules. This is a set of rules that has the best consequences for everyone’s welfare. There are also forms of utilitarianism that can roughly be described as combining both act-­based and rule-­based reasons for action.4 Welfarists may not only focus on welfare as such, but also on how it is distributed. For example, egalitarians hold that it matters whether welfare is distributed equally, prioritarians want to make sure that the positions of the worst-­off individuals are as good as possible, and sufficientarians hold that everyone should have enough welfare.5 Non-­welfarists, in contrast, believe that welfare does not determine normative reasons for action, or at least that welfare is not the only thing that counts when it comes to what we should do. Non-­welfarists typically do not focus on the consequences of one’s actions in terms of welfare, but on different considerations, such as the principles from which one acts, rights, or virtues. Nevertheless, these other concepts may themselves be based on welfare considerations in some way. For example, one understanding of rights is that they serve to protect welfare. Furthermore, even non-­welfarists often accept that at least sometimes welfare is at least part of what grounds normative reasons for action. For example, non-­welfarists often hold, in contrast to welfarists, that there are moral limits to what can rightfully be done to individuals in the name of promoting welfare. They may also accept that moral agents should have options to do things other than promoting welfare. But even non-­welfarists tend to believe that welfare considerations provide at least defeasible normative reasons for action. This book is primarily concerned with the capacity for welfare across species. This issue is of theoretical interest and can be explored in­de­pend­ ent­ly of one’s views about the relevance of welfare for what we have reason 4  For an exploration of egoism and utilitarianism, see Sidgwick ([1907] 1981) and Singer & Lazari-­Radek (2014). For defences of a dual source view, see Crisp (2006a), Phillips (2011), and Parfit (2017). For a defence of rule utilitarianism and its contrast to act utilitarianism, see Hooker (2000). For a pluralist version of utilitarianism with elements of both act and rule utilitarianism, see Woodard (2019). 5  Kagan (2019) provides an accessible introduction to a variety of welfarist views. For a defence of egalitarianism, see Temkin (2003). For a defence of prioritarianism, see Parfit (1991), and for a defence of sufficientarianism, see Crisp (2006a).

Introduction  19 to do. Considerations about welfare as such are axiological considerations, i.e. considerations about value. In contrast, considerations about reasons for action are deontic considerations; they concern the question how we should act. As I said, one need not be a welfarist to accept that effects on welfare are at least sometimes relevant for what we have reason to do. I assume that the book is of practical interest for everyone who believes that at least sometimes welfare matters for what we have reason to do. I do not take a position about the different normative views about how we ought to act. If any axiological view has practical implications, these are always the implications of the axiological view in combination with a particular normative view about what we should do. Thus, I will not defend any claims about how we should act in this book but focus on the issue of cross-­species capacity for welfare. I will, however, occasionally point out what EQU or DIF imply when combined with certain normative (and empirical) assumptions, most extensively in chapter 5.

1.7  Outline of the Book In chapter 2, I criticize the difference-­view about capacity for welfare. For this purpose, I present and reject the major proposals in the philosophical literature that appeal to the idea of different capacities for welfare. I discuss Kagan’s (2019) defence of DIF based on an objective list account of welfare, Singer’s (2011) defence of DIF based on whole life preferentialism, and Mill’s (1998) classical appeal to higher and lower pleasures. I also evaluate McMahan’s (1996, 2002) appeal to ‘fortune’, Wong’s (2016) distinction between experiential and absolute welfare, as well as Budolfson and Spear’s (2020) take on cross-­species capacity for welfare. I conclude that despite the prominence of DIF in philosophical discussions about the topic, its foundations are weak. Chapter 3 is devoted to a defence of the equality-­view about capacity for welfare. The chapter starts with an exploration of how animal welfare scientists understand and assess animal welfare and points out that this seems to be in line with EQU. As an example of a philosophical theory of welfare that is in line with EQU, I then present the self-­fulfilment account of welfare. An evolutionary perspective on hedonic capacity, which can be seen as a proxy for capacity for welfare, lends further support to the equality-­view. I conclude that EQU has much going for it. Chapter 4 extends the discussion to capacity for welfare across time. This means that we are not only interested in the level of welfare that different

20 

Capacity for Welfare across Species

individuals experience at any given point in time (i.e. synchronic welfare), but we are considering the amount of welfare that individuals of different species can gather throughout their lives (i.e. diachronic welfare). In particular, the welfare-­loss due to a premature death and the relevance of different species-­specific lifespans for an individual’s capacity for welfare will be explored. Do species with longer natural lifespans have a greater capacity for welfare, simply because of their longer lives? I distinguish what I take to be the main views and arguments on this issue and conclude that considering welfare across time requires more exploration in further research. Chapter 5 addresses practical implications. The implications of EQU or DIF for what we ought to do depend on the correct answers to several normative and empirical questions. It depends a lot on the issue of moral status. I introduce the issue and argue against hierarchical views of moral status (Kagan 2019). It also depends on whether we have reason to promote welfare by creating more welfare subjects. I introduce the discussion and defend the view that we ought to promote welfare only by improving the lives of those who live or will live. The book ends with a brief overview of ways of counting (or discounting) animals, which places the debate about cross-­species capacity for welfare in a broader context of philosophical controversies.

2 The Difference-­View DIF has been defended based on specific accounts of welfare. Shelly Kagan defends DIF based on an objective list account of welfare (section 2.1). Peter Singer’s defence of DIF was (formerly) based on whole life preferentialism (section  2.2). A classic defense of DIF is John Stuart Mill’s distinction between higher and lower pleasures (section  2.3). Jeff McMahan’s take on  cross-­species capacity for welfare introduces the notion of ‘fortune’ (section  2.4). Kevin Wong distinguishes what he calls ‘experiential’ and ‘absolute’ welfare to capture species-­specific differences in welfare capacity (section  2.5). Mark Budolfson and Dean Spears use the number of brain neurons as a proxy for cognitive capacity, which they take to be related to welfare capacity (section 2.6). I will point out what I see as serious shortcomings of each of these defences of DIF. My aim is to show that DIF is far less plausible than it is often assumed.

2.1  Kagan’s Objective List Account of Welfare As we already saw in this book’s introduction, Kagan (2019) accepts DIF and argues that humans are typically better off than dogs, which, in turn, are typically better off than mice, due to higher cognitive capacities. Let us now check out in more detail, how Kagan defends this view. Kagan (2019, 47) says: To put the point simply, if a bit crudely, there are goods which people are capable of possessing—­ and which, in a reasonably full life, they do possess—­that animals either lack altogether, or, alternatively, possess in smaller quantities, or possess in less valuable forms. And something similar holds when comparing significantly different forms of animal life as well. If we think that dogs have more valuable lives than fish, say, that is because we think that the goods available to dogs are, overall, better (in quantity and quality) than the ones available to trout or salmon. Capacity for Welfare across Species. Tatjana Višak, Oxford University Press. © Tatjana Višak 2022. DOI: 10.1093/oso/9780192882202.003.0002

22 

Capacity for Welfare across Species

What goods befall individuals with higher cognitive capacities rather than or more than those with lower cognitive capacities? As examples of such goods, Kagan appeals to suggestions that can be found in the philosophical literature. Yet he doesn’t claim to be comprehensive or precise. The goods that he lists are (a) meaningful relationships, based ‘not just on caring and shared affection but on insight and mutual understanding as well’, (b) possession of greater and more valuable knowledge, ranging from self-­ knowledge to scientific theories, (c) achievements in the pursuit of a wide range of goals related to hobbies, culture, business, and politics, among other things, (d) a highly developed aesthetic sense, allowing for the appreciation of such things as music, painting, natural beauty, laws of nature, and mathematics, (e) a sense of morality, and (f) the capacity to be virtuous and to show virtuous behaviour, and finally (g) the capacity to experience the sacred and the holy and to have and engage for spiritual understanding (Kagan 2019, 48). These are certainly items that many people value highly and that play an important role in their lives. Kagan (2019, 48) sums up: The point is only that almost all of us recognize that our lives are richer and more valuable by virtue of our having goods like these in our lives. To have fewer such goods—­or to have less valuable instances of them—­is to live a less desirable life, a life at a lower level of well-­being.

I agree with Kagan that some goods along those lines are important building blocks of good lives for most people. I also grant that at least some of these goods are unavailable to at least some non-­human welfare subjects or are not available in the same quality and quantity to them as to us. However, are these items good in themselves, or merely instrumental goods? Or, one could also ask, are they good-­making features? Proponents of different theories of welfare would provide different answers to these questions. For example, a hedonist would say that these things are good as a means, if at all. They are good if and only if and because and to the extent that they bring us enjoyment and bad if and only if and because and to the extent that they give us suffering. According to hedonism, that something is an instance of a meaningful relationship is not a good-­making feature. The only good-­making feature is that something is pleasant. Thus, for a hedonist, the things on Kagan’s list are only instrumental goods. So-­called ‘instrumental goods’, in a way, aren’t really goods at all. They are, at best, means to

The Difference-View  23 bring us goods. According to hedonism, the only ultimate good is pleasure. Pleasure is the only thing that counts when it comes to the goodness of my life for me. It is completely irrelevant from the perspective of personal good, whether an individual derives pleasure from meaningful relationships or from being alone, from contemplating laws of mathematics or from bathing in mud. From a hedonist perspective, the list that Kagan provides can at best give helpful advice for beings that are sufficiently like us and that want to lead pleasant lives. A hedonist would judge Kagan’s list of suggestions utterly worthless for beings that derived no pleasure from these things. Yet, even such a very different being need not be worse off than most of us, according to hedonism. This depends on the overall amount of pleasure that one experiences. Very similar things could be said about Kagan’s list from the perspective of nearly all other accounts of welfare, for example preferentialism or nature-­fulfilment accounts. (For an introduction of the­ or­ies of welfare, see section 1.5.) Kagan (2019, 49) does not appreciate this important point. Instead, he says, Of course, different theories of well-­being will accommodate this thought in slightly different ways. On some theories of well-­being, for example, the sorts of things I have just listed count as the very constituents of well-­ being, the possession of which directly constitutes being better off. On other theories, in contrast, they are only necessary components of well-­ being, and must fit into one’s life properly if one is to be well off. And on still other theories they are simply the essential means to well-­being, instrumental to having the very best sort of life but not valuable in and of themselves. These points of theoretical disagreement are philosophically interesting, but for our purposes they needn’t detain us. What is important is only the recognition that it is because our lives can have these things and other goods like these to a significant degree that our lives are more valuable than the lives of animals—­that we are better off than birds, and frogs and cows.

I think that Kagan gets this wrong. It is true that on objective list theories these or similar items may be on the list of ultimate goods and of good-­ making features. But on all other theories of welfare, these things are simply not good in themselves. They are, at best, means. Kagan admits that on some theories, these alleged ‘goods’ might be mere means. But he says that

24 

Capacity for Welfare across Species

on these theories the items on his list are ‘the essential means to wellbeing, instrumental to having the very best sort of life’. But what does he mean by ‘essential means’? Does he want to say that without these things, individuals cannot get desire satisfaction, nature-­fulfilment, or pleasure, or at least not much of these ultimate goods? That would be at the very least in need of justification. Why would a solitary animal need meaningful relationships in order to fulfil her desires or realize her nature? Why would knowledge of mathematical laws be an essential means for a pig in search of pleasure? These are not philosophical questions that ‘need not detain us’. They are central in an evaluation of Kagan’s claim that people are better off than animals due to their possession of the items on his list. Kagan’s claims hold only on the basis of specific objective list accounts of welfare. In fact, Kagan’s support of DIF rests on acceptance of an objective list account of welfare with some such items on the list. Proponents of all other accounts of welfare would disagree, and not only slightly. Or at the very least they would require significant further argumentation, which Kagan doesn’t provide. Before I turn to the most prominent lines of argumentation that aim at reconciling other theories of welfare with DIF, I want to say a bit more about the plausibility of the kind of objective list account of welfare that Kagan appeals to in his support of DIF. After all, if his assumed ob­ject­ ive list account of welfare would be the most prominent or the most plaus­ ible theory in this field, the fact that his proposal depended on it could be considered less of a problem. However, the theory that Kagan’s defence of DIF relies on does not look to me like a plausible theory of welfare at all. An objective list account of welfare supports DIF only if it has two features. First, there must be one and the same list of goods for all species, rather than different lists for different species. It is at least thinkable that different lists of goods apply to different species. But then, in order for these differences not to be ad hoc, there would need to be some underlying justification why particular items are good for particular species. Those who think that this is obvious, since different species simply have different natures, may be drawn to a nature-­fulfilment account of welfare rather than to an objective list account. If, instead, one really holds that the items are good in themselves and simply having them in one’s life makes for a better life, then it makes sense to have one and the same list of goods for all species. So, this first additional feature doesn’t require much in terms of further justification, once an objective list account, rather than, say, a nature-­ fulfilment account is accepted. Second, the items on the list must be such that humans have them to a significantly greater degree than non-­humans.

The Difference-View  25 For some items that could conceivably be on an objective list of welfare, such as ‘health’, ‘positive emotions’, or ‘getting what one wants’, it would not hold that humans can realize these goods to a greater extent than non-­ humans. So, not all objective list accounts would support DIF. One can, of course, come up with an objective list account that has these specific features and that therefore supports DIF. In that case, it would be important to evaluate how plausible such an account of welfare is. It is remarkable that among the authors drawn towards an objective list account of welfare, there is no agreement about any list of ultimate goods. In fact, many authors proceed like Kagan and merely hand-­wave in the direction of some possible list of goods without committing to any specific item on the list and without ruling out that additional goods may belong on it. So, whenever a particular item on the list is criticized as not being a basic good, or whenever an addition to the list is suggested, these authors are in a pos­ ition to say: ‘Well, you might be right. All I am saying is that some such list of goods is correct’. In order to claim that some objective list account is the correct theory of welfare, it suffices indeed to establish that one list or the other will be correct. But not all lists will support DIF. Since Kagan’s list would support this view, it is worth taking a closer look at the items that he suggests. I will discuss the items on his above-­mentioned list in reverse order. Consider first the capacity to experience the sacred and the holy and to have and engage for spiritual understanding. This formulation suggests that there is something spiritual, sacred, or holy about the world and our existence that can be experienced and understood. Many people would deny that our existence has such a dimension. This is not to deny that people have certain experiences, which they label ‘spiritual’ or the like: but it is to deny that there is a spiritual, sacred of holy fact of the matter about the world. If there is no such fact of the matter, there is nothing in that regard to be understood or experienced. People have very different and incompatible spiritual beliefs and at the very best only a tiny fraction of these beliefs could qualify as knowledge. It is hard to believe that having false beliefs about the world in and of itself makes one’s life better. This holds for false beliefs about spiritual issues just as for false beliefs about mathematics or other topics. Typically, spiritual experiences, false as they may be, are at least pleasant in some way. If they are unpleasant, it is hard to believe that these experiences are good at all. So, I think that the item derives any plausibility that it has from convergence with other items on the list, such as knowledge, achievement, or perhaps aesthetic experiences, or from potential goods that are not on the list, such as pleasure.

26 

Capacity for Welfare across Species

Consider next the capacity to be virtuous. There is much disagreement among humans about what counts as virtuous. There may be more agreement about special cases, such as some historic heroes, but these are the standards that most people never reach. Typically, being virtuous promotes happiness or other possible goods, and if it wouldn’t do so, it would be hard to see why it should be good at all. We can, in contrast to non-­humans, break our heads about what would be virtuous, but does this, in itself, really improve our lives? If two people led the exact same lives, also in terms of their own emotional states and effects on others, would it really make one of them better off if he tried, perhaps even unsuccessfully, to be virtuous? I think that trying to be virtuous can be good as a means. It might make the individual feel better about himself and it might positively affect others. But I do not think that virtue is a basic personal good. The same, I take it, holds for morality. People can aesthetically appreciate things, such as music, paintings, landscapes, or even maths or science. Appreciating these things gives these ­people pleasure, what else would it mean to ‘appreciate’ them? It seems to me that all sentient beings that have preferences can appreciate certain things in the sense of deriving pleasure from them. Of course, different species and individuals appreciate different things. An alternative understanding of what it means to appreciate something involves understanding the value of something. In case of aesthetic appreciations, this could be the aesthetic value, i.e. some kind of beauty. But what is beauty? It seems to me that beauty is whatever we appreciate in particular ways. Other animals have different tastes, as dog holders can probably confirm. My dog seems to highly appreciate the smell of different pieces of excrement or waste. If my dog were able to have such thoughts, he would think that I completely fail to appreciate the aesthetic value of such things. What members of any species appreciate is related to this species’ evolutionary history. It is what helped members of the species to survive and reproduce. Ideas about beauty with regard to the opposite sex, for example, can be traced back to this evolutionary criterion in all species, as well as many other aesthetic preferences and tastes. But why should it be good for members of other species to appreciate the things that happened to help humans survive and reproduce, or that (as our appreciation of mathematical laws) are by-­products of things (such as our reasoning capacity) that helped us survive and reproduce? Or, to put it differently, why should the things that we happen to find attractive improve the lives of members of other species? By the way, I have never appreciated the beauty of a mathematical law, and I don’t think I’m worse off for not

The Difference-View  27 having done so, as long as I, just like my dog, derive as many pleasant ex­peri­ences from contemplating other things as Kagan may derive from contemplating mathematical laws. What about knowledge? I do not think that knowledge is valuable as such. Knowledge can be useful. Or it may simply feel good to know a lot about certain things that interest us. Or engaging with knowledge may amount to nature-­fulfilment for some species. But I do not think that knowledge in itself adds value to a life or is a good-­making feature. Imagine two lives that are equal in all respects except that one life contains add­ition­al knowledge of some facts that are neither useful nor in any way nice to know. I wouldn’t say that the individual leading the second life is better off. Lastly, consider meaningful relationships. We humans are social mammals that evolved in order to flourish in groups. When we are born, our survival depends on the care and support of others. We need at least some meaningful relationships, at least during some parts of our lives, to secure our mental and physical health. And even beyond that bare necessity a lot of the things that humans do (and did for ages in the course of evolution) depend on meaningful relationships. As a species we learned to empathize with others and to read others’ minds to a certain degree, since this helped us to survive and reproduce in our social groups. Therefore, it is unsurprising that relationships mean a lot to us and that we even are inclined to ­consider them objectively good. As Kagan (2019, 50) claims: Friendship is a tremendous good, and if some animals are incapable of friendship, then that is simply one more way in which it is better to be one of us than to be an animal of that kind.

I am inclined to say that friendship is a good for humans, because it gives us deeply pleasant feelings to maintain friendships and it is in our nature to be social. Typically, we chose the people as our friends that are like us in im­port­ant ways and that confirm our values and choices. Interacting with these people feels good. Besides friends, we also have other relationships. Many of us have neighbours or mothers-­in-­law, for example. But interacting with these people is not always pleasant and often rather the contrary. Otherwise, we would probably have proclaimed objective value for having neighbours or mothers-­in-­law. So, it’s not the relationships as such that are good for us. In fact, some relationships are a pain in the derrière and our lives would be better without them. Even though, these terrible relationships may nevertheless be ‘meaningful’, as Kagan uses the term. (He doesn’t say what

28 

Capacity for Welfare across Species

exactly makes a relationship ‘meaningful’.) I think that certain relationships are in some way good for us, perhaps only instrumentally good, but they need not be good for members of other species. In contrast to social mammals like us, members of species that are solitary by nature aren’t worse off for not having our kinds of relationships with conspecifics. I conclude that Kagan’s defence of DIF depends on acceptance of a specific and controversial account of welfare and is therefore on much weaker philosophical ground than he makes it seem. Nevertheless, defending DIF on the basis of an objective list account of welfare is internally consistent, provided that the goods on the list are unavailable to non-­humans or available to them only to a weaker extent. The overall plausibility of such a view depends on the plausibility of the underlying account of welfare. There are no knockdown arguments against such an account of welfare, but I have indicated some reservations. In the end, we will need to compare this way of thinking about cross-­species capacity for welfare to rival proposals.

2.2  Singer’s Whole-­Life Preferentialism One version of preferentialism, which is called ‘whole-­life preferentialism’ has been used to justify DIF. In Practical Ethics Peter Singer (2011, 90–93) asks the reader to imagine a being that can at one time be fully a horse and at another time be fully human. Then, from a third perspective, this being can remember how it was to be a horse and how it was to be a human, and would, according to Singer, favour being a human. Singer (2011, 92) concludes: In general, it does seem that the more highly developed the mental life of a being, the greater the degree of self-­awareness and rationality and the broader the range of possible experiences, the more one would prefer that kind of life, if one were choosing between it and a being at a lower level of awareness.

A first set of challenges relates to the feasibility of Singer’s thought experiment. We, the readers, are humans, and we are in fact only familiar with human life and the human perspective. Therefore, it is questionable to what extent we can put ourselves in the shoes (or, better, the hooves) of the horse. While this is a challenge indeed, I take it that we can gain relevant know­ ledge about other beings’ needs and experiences and empathize with them, at least to some degree that might be sufficient for most practical purposes.

The Difference-View  29 The next difficulty is choosing between these lives from a third, neutral perspective. I think it is impossible to make sense of such a neutral perspective within this particular preferentialist framework. After all, preferentialism needs to assume given preferences and defines welfare as the extent to which these preferences are fulfilled. In that sense, there cannot be a neutral perspective for a preferentialist without any given set of preferences. A second set of challenges relates to accounting for the alleged preference for the human’s life. If we ended up preferring the human’s life, we would probably do so based on reasons unrelated to personal value. As human beings, we generally care strongly about many other things, besides our own welfare. We often care strongly about those near and dear to us. Whether this is the same for other animals depends on their species. As humans, we may also care about moral issues, politics, or scientific discoveries. Perhaps we care about truth and beauty. After we have become acquainted with these concerns when we adopted the human perspective, we might, from the ‘neutral’ perspective, still see the importance of these things. If so, our choice in favour of the human life might testify that we consider other things, except our own welfare, important. It does not necessarily prove that human life has more personal value. One might find it better simpliciter, even if it were worse in terms of personal value. This suggests that not everything that one desires, overall, for one’s life may be based on its personal value. Whole-­life preferentialism has been criticized for being unable to account for self-­sacrifice (Heathwood 2011). I make a self-­sacrifice just in case I willingly make myself worse off for the sake of others. According to whole-­life preferentialism, whatever I most want to do with my life is best for me. If killing myself in order to benefit my children is what I most desire to do, this is, per definition, best for me, according to whole-­life preferentialism. This understanding of my action is incompatible with ‘willingly making myself worse off overall, for the sake of others’ and thus with the standard account of self-­sacrifice. So, from the perspective of whole-­life preferentialism, even what seems to be a self-­ sacrificial act needs to be described as an act that is best for me. This just illustrates my second point about Singer’s thought experiment: It is not always the case that our (informed and rational) choices among possible lives indicate what is better for us. A further problem for Singer’s underlying preferentialist theory is that it cannot be applied to babies and young children, even though we clearly consider them to be subjects of welfare. This is because the theory holds that ‘benefiting me’ means ‘fulfilling my desires for my life as a whole’. Thus,

30 

Capacity for Welfare across Species

it seems that in order to be a welfare subject, a being needs to have desires for her life as a whole. Babies and young children do not have a notion of a life as a whole that they would prefer, since they lack the capacity to envision possible future lives or to evaluate them in the relevant way (Lin 2017). Appealing to what a baby or young child (or a non-­human animal or a mentally disabled human, for that matter) would prefer under some idealized non-­actual conditions, such as rationality and full information, is not helpful either, since all one could do is make such a choice on behalf of that individual. By doing so, one would have to stipulate what would be good for that individual on grounds other than the individual’s current desires about that issue. One would thus need to decide on the basis of an alternative account of welfare. As Kraut (2007, 107) observes: When parents nurture a young child, they shape her desires, goals, and plans, and they do so for her good.

If a being doesn’t have the relevant desires, whole-­life preferentialism seems to lack the resources to say what would benefit that being and indeed seems to imply that nothing would benefit or harm the individual in question. Other versions of preferentialism avoid these problems (Heathwood 2011). On these versions, it is not the case that whatever life I would (under idealized conditions) choose would be better for me. Instead, what life would be better for me depends on the balance of desire satisfaction and desire frustration that I would experience living that life. The more and the more important of one’s preferences are satisfied and the fewer and the less important of one’s preferences are frustrated at a time, the better off one is at that time. The father in our example could be said to forego future preference satisfaction for the sake of his kids and thus make a self-­sacrifice. And even beings without the capacity of having desires for their lives as wholes can be said to be welfare subjects, provided that they have any preferences at all. All sentient beings have preferences, such as the preference for certain food or surroundings, or the preference not to suffer. Since the content of one’s preferences is considered irrelevant, these versions of preferentialism do not support DIF. In fact, they support EQU. On one rather negative and less common interpretation of preferentialism, while frustrated preferences count negatively towards one’s welfare, the satisfaction of a preference does not bring about positive welfare, but merely neutralizes the negative entries that would otherwise be in place due to unsatisfied preferences (Fehige 1998). On this version of preferentialism,

The Difference-View  31 the best one can hope to achieve is to have as few as possible and as weak as possible unfulfilled preferences. This account would not in any way rule out the possibility that the welfare of a cognitively or emotionally weaker developed individual is as great as the welfare of a cognitively or emotionally stronger developed individual. To the contrary, an individual with only a few easily fulfilled preferences is more likely to reach the neutral level or come close to it than an individual with more complex and more difficult to fulfil preferences. Again, this is in line with EQU. Summing up, Singer’s defence of the claim that higher species-­specific cognitive capacities are connected to higher welfare relies on a particular version of preferentialism that focuses on preferences for lives rather than preference satisfaction within lives. This version of preferentialism has implausible implications, such as its failure to account for self-­sacrifice and its inapplicability to many individuals to which we normally ascribe welfare. Furthermore, it seems impossible to make sense of Singer’s thought experiment from the perspective of such a theory, since only given preferences determine what is good for an individual. Thus, one cannot choose among lives from a perspective without such preferences. Other versions of preferentialism reject DIF and avoid these problems. These other versions of preferentialism are in line with EQU.

2.3  Mill on Higher and Lower Pleasures Mill is known for his specific version of hedonism. Mill (1998) draws a distinction between the quantity of the enjoyment and its quality, and he argues that both are important: It would be absurd that while, in estimating all other things, quality is considered as well as quantity, the estimation of pleasures should be supposed to depend on quantity alone.

Mill takes it that what he calls ‘higher pleasures’ are superior to what he calls ‘lower pleasures’. Higher pleasures are associated with more intellectually demanding activities and lower pleasures are supposed to come from less intellectually demanding activities. The quantity of an enjoyment, according to Mill, is determined by its intensity and duration. In addition, an enjoyment can have a higher quality than a second one. An enjoyment of a higher quality,

32 

Capacity for Welfare across Species

according to Mill, is superior to an enjoyment of a lower quality, no matter what the quantity of these enjoyments is. If this is true and if the in­tel­lec­ tual­ly demanding activities are unavailable to certain individuals, it explains why the life of these individuals contains less welfare. It is controversial how to interpret Mill’s proposal and whether such a view can plausibly be defended. On the standard interpretation, Mill faces a dilemma: Either the higher pleasures are higher because they are more pleasurable or enjoyable, in which case no special distinction between higher and lower pleasures can be drawn [. . .]; or they are higher for some other reason, such as their being more ‘noble’, in which case Mill has abandoned hedonism by allowing non-­hedonistic values into his formal theory. (Crisp 2006b, 631)

The traditional hedonist view of Bentham and Sidgwick embraces the first ‘horn’ of this dilemma and rejects Millian superiorities. However, some contemporary defences of Millian superiorities have been brought forward. Roger Crisp defends an improved version of Millian superiorities. Mill, according to Crisp, was confused, because he followed the Empiricists of his time and accepted that enjoyment is a kind of sensation. Sensations—­such as warmth or headaches—­ can indeed vary in intensity and duration. However, according to Crisp, enjoyment is not such a kind of sensation. Enjoyment, according to Crisp is a determinable, while particular enjoyments are determinates, just like colour is a determinable and particular colours, such as red and blue, are determinates. Red can, in turn be a determinable relative to the determinates consisting in different shades of red, such as scarlet (Wilson 2021). Particular experiences can vary in intensity and duration and this might influence the enjoyment of the experience. The enjoyment itself, however, can only vary in the sense of being more or less enjoyable. So, Crisp rejects Mill’s quantity/quality distinction for enjoyment. Crisp instead justifies the distinction between higher and lower pleasures in a slightly different way. According to Crisp, there is no objective way of determining how enjoyable an experience is for some subject. The only measure is the subject’s judgement as to how enjoyable an experience was or is. Therefore, Crisp claims, if a subject considers some experience more enjoyable than any amount of a second experience, that first experience is a higher pleasure. Here is Crisp’s (2006b, 633) example of such a case:

The Difference-View  33 So imagine someone who has just drunk a cool glass of lemonade and has also completed her first reading of Jane Austen’s Pride and Prejudice. If we ask her to rank, on a scale of enjoyableness, the experience of drinking the lemonade against that of reading the novel, she may well rank the novel higher than the lemonade. Why? There is much more to this judgment than mere duration. There is nothing to prevent our judge’s claiming that it would not matter how long the experience of enjoyable drinking could be prolonged: She would never enjoy it as much as she enjoyed the novel. For what she enjoyed in the novel was its wit, its beautiful syntax, and its ex­quis­ite delineation of character. The loss of such enjoyments (that is, enjoyable experiences)—in the context of her own life—could never be compensated for, in terms of enjoyment alone, by any amount of lemonade pleasure.

How is Crisp’s argument supposed to work? It can, of course, be granted that no matter how long one would continue the experience of drinking lemonade, it would never be as enjoyable as reading the novel. That makes sense, since the experience of drinking lemonade has declining marginal utility: at least after a couple of emptied glasses, drinking any additional lemonade won’t be all that enjoyable anymore. But this simple fact is not the point that Crisp is making here. Crisp does not simply imagine prolonging the experience of drinking but the experience of enjoyable drinking. Perhaps we are to imagine, unrealistically, that every additional glass is just as tasty, is just as refreshing, eases our thirst just as much and doesn’t make us want to vomit. Even if that were all granted, we might just get bored with refreshing ourselves in a tasty way. If this were the case, declining marginal utility would show up again. But what if it is not only the tastiness and other characteristics of the drinking experience that are prolonged, but the sheer overall enjoyableness that drinking the first glass of lemonade gave us? What if the prolonged experience is meant to give us more and more of this en­joy­ able­ness, making the experience increasingly enjoyable? Then it seems, intuitively, that there must come a point at which the enjoyableness of this experience surpasses the enjoyableness of reading the novel, at least if we assume that enjoyable experiences can be added up across time. In that case they would not be as enjoyable at any given point in time, but a year with regular lemonade pleasure may contain more enjoyment overall than a year with a few days of novel pleasure and then many months without any pleasure. My claim that there must come such a point in time at which the aggregated lemonade pleasure surpasses the novel pleasure rests on the

34 

Capacity for Welfare across Species

assumption that lifetime welfare is measurable on a continuous additive scale that satisfies the so-­called Archimedean property for real numbers: for any positive number x, no matter how small, and for any number y, no matter how large, there exists an integer n, such that nx ≥ y. This means that any two (positive) levels of welfare are commensurable, i.e. their ratio is not infinite.  (Klint Jensen 2008)

Defenders of Millian superiorities reject this assumption. For instance, Griffin (1986, 340 note 27) writes: [. . .] it treats well-­being as measurable on a single continuous additive scale, where low numbers, if added to themselves often enough, must become larger than any initial, larger number. But this seems not true in prudential cases [. . .].  (See also Riley 2008, 63)

Rejecting the continuity of personal value seems to be plausible for value pluralists. Dale Dorsey (2009), for instance, holds that hedonic goods matter for well-­being, but no amount of hedonic goods can be as valuable as the superior good of achieving one’s projects. Parfit (1984, 414) proposes such a lexical view as well and he admits: ‘it is more plausible when applied . . . to things which are in two quite different categories’. For purposes of illustration, he refers to the belief that both pain and sin are bad, but no amount of pain is as bad as even a tiny amount of sin (Parfit 1984, 414). Even though it seems less plausible for monistic accounts of prudential value, Millian superiorities are still being defended. In spite of his concerns, Parfit (1984, 414) does not rule out such a view: Such a view can plausibly be held when we compare certain human lives to the lives of the lower animals. [. . .] When they are not factory-­farmed, the lives of pigs are probably worth living. But we can plausibly claim that, even if there is some value in the fact that these lives are lived, no amount of this value could be as good as the value in the life of Socrates.

This is a mere restatement of Mill’s intuition and as such doesn’t add anything to its justification. In his own justification of higher and lower pleasures, Mill refers to preferences:

The Difference-View  35 If one of the two is, by those who are competently acquainted with both, placed so far above the other that they prefer it [ . . . ] and would not resign it for any quantity of the other pleasure which their nature is capable of, we are justified in ascribing to the preferred enjoyment a superiority in quality, so far outweighing quantity as to render it, in comparison, of small account.  (Mill 1863/ 2001, 11)

It is controversial what to make of these claims. As we saw, Crisp (2006b, 633) takes a purely subjective position and proposes to base judgements on the individual’s reports of how enjoyable various experiences are. But if every individual’s judgement must be taken at face value, it seems that Crisp cannot defend any general claims about what a higher or lower pleasure is. Riley’s understanding of Mill’s test is that the determination of what is a higher and lower pleasure is not purely subjective. According to Riley (2008, 63), when ‘most if not all’ of those who are familiar with two pleasures agree that the first is superior to any amount of the second, that indicates that the first is objectively superior. Thus, Riley’s interpretation entails, contrary to the spirit of hedonism, that those who hold a minority position about the pleasantness of various pleasures get it wrong. On standard accounts of hedonism, while a cognitively higher developed individual might derive pleasure from certain intellectual activities that are unavailable to a cognitively lower developed individual, the latter might derive pleasure from other sources. In general, a being’s physiology and psych­ology determine what he or she finds enjoyable. The sources of pleasure are irrelevant. What matters, according to the standard version of hedonism, is only how pleasant an experience is for the individual. Sidgwick already criticized Plato’s argument (restated by Mill) that the life of the phil­ oso­pher is more pleasant than the life of the sensualist: But who can tell that the philosopher’s constitution is not such as to render the enjoyments of the senses, in his case, comparatively feeble? While on the other hand the sensualist’s mind may not be able to attain more than a thin shadow of the philosopher’s delight.  (Sidgwick 1907/1981, 148)

Those who are familiar with different kinds of pleasures do not necessarily consider simple pleasures less pleasant than more sophisticated ones. What do most people do in their holidays or weekends? Things like eating, drinking, sunbathing, resting, running, swimming, spending time with friends

36 

Capacity for Welfare across Species

and family, spending time in nature, having sex, etc. Most of these things even cognitively less advanced animals can do. Furthermore, there is no such thing as ‘novel pleasure’ that can be compared in abstract to ‘lemonade pleasure’. What gives me pleasure always depends on internal and external circumstances. If I am very thirsty after a long walk through the desert, a glass of lemonade will give me far more pleasure at that time than any novel can do. The people that are said to prefer intellectual pleasures to the pleasures of food and drink tend to do so only once they have had enough food and drink to be able to enjoy the intellectual pleasures in the first place. In later work, Roger Crisp, together with his co-­author Morten Kringelbach, appeals to current insights from neuroscience to argue against Millian superiorities. They say with regard to higher and lower pleasures: If there are two such sensations, we might expect them, purely on the basis of supervenience considerations, to differ significantly in their underlying physiological features. If we found that the brain states underlying the pleasantness of higher pleasures were radically different from those underlying that of lower pleasures, this would be major piece of evidence in favour of the higher/lower thesis.  (Crisp and Kringelbach 2017, 213)

Unfortunately for supporters of Mill’s thesis, neurological evidence suggests the contrary. Neuro-­imaging studies show that ‘the neural substrate of pleasure in quite different kinds of activity is quite similar’ (Crisp and Kingelbach 2017, 213). Different kinds of pleasures seem to be processed via the same networks. It therefore shouldn’t surprise that we react similarly to different kinds of pleasures, for example with a smile. So, Mill’s defence of DIF rests on weak evidential ground. As I will argue in detail in section 3.3, Mill’s view of higher and lower pleasures is also incompatible with current insights from evolutionary psychology.

2.4  McMahan’s Notion of ‘Fortune’ According to McMahan (1996, 7) cognitively impaired humans are badly off mainly for the following reason: For the range of forms and levels of well-­being that are in principle ac­cess­ ible to an individual is determined by that individual’s cognitive and

The Difference-View  37 emotional capacities and potentials. The more limited an individual’s capacities are, the more restricted his or her range of well-­being will be. There are forms and peaks of well-­being accessible to individuals with highly developed cognitive and emotional capacities that cannot be attained by individuals with lower capacities.

McMahan, just like Kagan, accepts an objective list account of welfare. He holds that the exercise of a certain higher capacity, talent, or skill is itself a good, quite independently of any of its effects.  (McMahan 1996, 8)

Those to whom these activities are unavailable are worse off. Therefore, McMahan (1996, 8) continues: If [. . .] it is a misfortune to be natively endowed with cognitive and emotional capacities and potentials of the sort possessed by the cognitively impaired, then nonhuman animals with comparable capacities and potentials must be unfortunate as well.

So, according to McMahan, non-­human animals have a low level of well-­ being, because of their rudimentary cognitive and emotional capacities, which prevent them from performing certain intrinsically valuable ac­tiv­ ities, such as playing the piano, writing poetry, and engaging in science. McMahan (1996, 9) is well aware that this ‘may seem absurd’. In order to make the implications of his unadjusted pluralist account of welfare more palatable, McMahan distinguishes the notion of ‘well-­being’ on the one hand from the notions of ‘being well off ’, ‘being fortunate’ and ‘flourishing’ on the other hand. So, McMahan (1996, 9) specifies that a dog, in spite of his comparatively low level of well-­being, may not be badly off, but may instead be flourishing and have ‘what counts as a good life for a dog’. This move is very surprising, since normally the notion of well-­being just is the notion of how an individual’s life is going for that individual. In contrast to that convention, McMahan (1996, 9) wants this later notion, which he calls ‘fortune’, to express ‘a relation between an individual’s level of well-­being and a standard against which well-­being is assessed’. This move raises the problem of finding an appropriate standard against which to judge an individual’s welfare in order to determine how well off she is. McMahan (1996, 10) describes this task as follows:

38 

Capacity for Welfare across Species

Why is it that, for example, a normal dog is not unfortunate despite its low level of wellbeing? And is the explanation compatible with the common-­ sense view that a human being congenitally endowed with comparable cognitive capacities is unfortunate? To answer these questions, we require an analysis of the notion that I have called ‘fortune’ that indicates what the standard is against which an individual’s level of well-­being is assessed in order to determine whether that individual is fortunate or unfortunate, or faring well or badly.

McMahan (1996) considers the species-­ norm account, the individual-­ possibility account and the peak-­capacity account and rejects them all because of counterintuitive implications. He finally accepts the native potential account, even though he acknowledges that this account faces some unsolved problems as well. According to the native potential account, whether an individual is well or badly off depends on the gap between her actual welfare and her native potential for welfare. The greater the gap, the worse off the individual is. Since both the normal dog and the severely cognitively impaired human may have a welfare level that is in line with their native potential for welfare, they may be well off according to McMahan even though their welfare is poor. McMahan’s (1996) normative conclusion is that since none of them is unfortunate, none of them requires any compensation on grounds of distributive justice. So, he connects normative reasons for action (at least in some cases) to fortune rather than welfare. What McMahan calls ‘welfare’ is not relativized, but absolutist. He accepts an objective list account of welfare, and he holds that a human’s life can contain (and typically does contain) more goods than a dog’s life, so the human’s welfare level is higher. Yet, a dog who is as well off as he can be is not unfortunate. Instead, this dog is more fortunate than a human who realizes only half of his native potential, even though the dog’s welfare level is lower. So, at this point in McMahan’s theory a relativizing move comes in: welfare ­levels are relativized to native potential (or intrinsic capacities). As depicted in Figure 2.1, one can think of McMahan’s view based on an analogy with measurement cups. ‘Welfare’ indicates the absolute amount of liquid in the cup. ‘Fortune’ refers to the degree to which the cup is filled. ‘Native potential for welfare’ refers to the size of the cup. The human, in Figure 2.1, would be better off, but less fortunate. McMahan (1996) is aware of an odd implication of this account. It implies that of two equally cognitively impaired individuals, one should be judged well off and the other badly off, depending on whether the

The Difference-View  39 size of cup = native potential for welfare extent to which it is filled = fortune dog

human

total amount of liquid = welfare

Figure 2.1  A simplified picture of the difference between fortune and welfare, according to McMahan.

individual had a native potential for more welfare and lost it right after birth without ever using it, or whether he never had the potential for more welfare in the first place. The individual who lost his native potential before having developed it is badly off, the individual who never had that potential in the first place is well off, even though, in terms of welfare, both are, have been and will be in the same condition. What to make of McMahan’s account? Even though it is certainly in­geni­ ous and captures a range of intuitions that many people may have, I think it is implausible for at least three reasons. First, saying that a perfectly happy and flourishing dog has a low level of welfare, because she is incapable of playing the piano or writing poetry is counterintuitive. Many people would certainly want to say about a flourishing dog who is happy and healthy, that he has a high level of welfare. Second, making a conceptual distinction between welfare on the one hand and ‘being well or badly off ’ on the other hand is theoretically undesirable. It really seems that the notion of welfare indicates how well off an individual is. Third, the implication that of two individuals with exactly the same impairment and mental life one is well off and the other is badly off is hard to swallow. In spite of these criticisms, McMahan’s proposal is internally consistent, and it is an interesting suggestion for how to compare welfare across species. In the end, it will have to be compared to alternative proposals in order to judge its overall plausibility. Just like Kagan, McMahan accepts an objective list account of welfare. I evaluated Kagan’s objective list account in section 2.1 and my main line of criticism applies to McMahan’s account as well. But unlike some other proponents of these theories, McMahan is unwilling to accept that the low level of welfare that, say a healthy and happy dog has according to this account, provides us with reasons to have special concern for the dog from an egalitarian or prioritarian perspective. The dog, according to McMahan, needs no special concern despite his low level of welfare, because he is not unfortunate. However, when it comes to the question of saving the life of a human or the life of a dog, McMahan would probably want to say that we should

40 

Capacity for Welfare across Species

save the human’s life, since death would harm the human more than the dog. After all, the human, on his account of welfare, has more welfare to lose. So, here it would not be fortune but rather welfare that grounds ­reasons for action. I find it ad hoc to vary in this way between fortune and welfare when it comes to the basis for (some of our) reasons for action.

2.5  Wong on Experiential versus Absolute Welfare Kevin Wong (2016) distinguishes two notions of welfare, which he calls ‘experiential’ and ‘absolute’ welfare. ‘Experiential welfare’ refers, roughly, to how well off the being is at a time, relative to its species-­specific potential. It is like McMahan’s notion of fortune (see section 2.4). In contrast, the notion of ‘absolute welfare’ is not relativized to species potential. The capacity for absolute welfare is related to cognitive or hedonic capacity, according to Wong. Absolute welfare is the notion that matters, according to Wong. To introduce the notion of ‘experiential welfare’, Wong (2016, 16) asks the reader to assume that a human has an experiential welfare score of 3 and a dog has an experiential welfare score of 10. This means, according to Wong (2016, 16) ‘that the dog lives a better life for itself, than the human being lives for himself or herself’. Typically, the notion of well-­being is understood as the value of life for the individual in question. So, if the human has an experiential welfare score of 3 and the dog an experiential welfare score of 10, this sounds as if this denotes the well-­being levels of these individuals. But Wong says that this is not meant to be their well-­being levels, but it just indicates how good their lives are for them, as if these were different issues. So, he seems to deviate from the standard understanding of ‘welfare’ as ‘goodness of an individual’s life for the individual in question’. Instead, Wong calls the goodness of a life for the individual the individual’s ‘experiential welfare’. This raises the question whether what he calls ‘total’ or ‘absolute’ welfare does not indicate the goodness of the life for the individual and, if not, what else it does indicate. I think that Wong’s proposal can be interpreted most charitably as follows. Experiential welfare indicates the welfare of an individual relative to the range of well-­being that is possible for the individual in question. Even though an individual may have a high score in experiential welfare, because it is close to its maximum attainable level of welfare, it may still have only little welfare in absolute terms. It may score lower in terms of absolute welfare than a different individual with only little experiential welfare, but a

The Difference-View  41 broader range and higher peak of attainable welfare. This is an intelligible story. But it is, of course, in need of justification. So, Wong claims that one unit of experiential well-­being for a human is worth more in absolute terms than one unit of experiential well-­being for a dog. I think that this is a crucial assumption and one that is difficult to justify properly. Let us be clear about what it means. The way I just framed it is somewhat misleading. Wong does not want to make a claim about moral status. Unlike Shelly Kagan (see section 5.1) Wong does not want to say that one unit of well-­being of a human matters more than one unit of well-­being of a dog. Instead, Wong introduces a discounting factor at an earlier stage. According to Kagan’s hierarchical view on moral status (as discussed in section 5.1), first there is welfare, and then we need to weight it differently to determine our reasons for action. In contrast, according to Wong, first you have experiential welfare and second you bring in a weighing factor and only then you’ve got your unit of absolute well-­being, which you then, if you like, and as Wong suggests, can give equal consideration, no matter whose unit it is. So, Wong does not reject the principle of equal consideration of well-­being. The crucial point in the evaluation of Wong’s proposal is whether and, if so, how the weighing factor that comes in between ex­peri­ en­tial well-­being and total well-­being can be justified. To be sure, Wong does not defend any particular weighing factor. He merely argues that some such factor is required. Wong accepts a relativizing account of experiential welfare. He relativizes welfare to the individual’s native potential or something along those lines. On top of this, he brings in some non-­relativizing or absolutist con­sid­er­ ations. The latter are meant to justify the weighing factor. As in Figure 2.2, the experiential welfare scale of a dog can be pictured as a medium-­sized measurement cup. In Wong’s example, the dog’s cup of experiential welfare is full: the dog has an experiential welfare score of 10 out of 10. The human’s experiential welfare scale can be pictured as a larger measurement cup. The human’s cup is not full, but less than half full (only 3 out of 10). Yet, due to the sizes of the cups, one experiential welfare unit of a human is bigger than an experiential welfare unit of a dog. Wong wants to take this absolute difference into account when it comes to measuring bene­fits and harms. Furthermore, Wong wants to avoid committing on what the max­ imum welfare level for a human is. Is it really 10, or rather 20 or 100? But Wong believes that the human’s maximum may be higher than that of the dog. And he wants to take this consideration into account. So, to say it in terms of my picture, Wong does not want to take a stance on the height of the cups.

42 

Capacity for Welfare across Species extent to which cup is filled = experiential welfare

dog

human

total amount of liquid = absolute welfare

Figure 2.2  A simplified picture of the experiential welfare and absolute welfare, according to Wong.

Let us now explore in some more detail how Wong proposes to put to work his discount rate. Wong (2016, 16) explains the purpose of the discount rate that he proposes to implement between experiential welfare and absolute welfare as follows: This unit [i.e. the discount rate—­ T.V.] undergirds the intuition that humans and dogs are differently equipped to flourishing lives, and so should be assessed on their own terms.  (Wong 2016, 16)

It would be easy to agree if this just meant that a flourishing life for a dog is different from a flourishing life for a human. But this is not what Wong means. Instead, he wants to say that dogs’ lives cannot be as flourishing as humans’ lives. Wong (2016, 16) continues: It’s no misfortune, for example, for a dog to lack the cognitive capacity of a human being.

Again, one could take this to mean that fewer cognitive capacities do not prevent a dog from living a flourishing dog life, even though they would perhaps prevent a human from living a flourishing human life. But, again, this is not Wong’s claim. He does not want to make the obvious point that dogs and humans need different things in order to flourish. Instead, he wants to say that higher cognitive capacities (or more generally: typical human as opposed to, say, canine capacities) allow for more well-­being in absolute terms, as McMahan and Kagan, among others, also assume. But when it comes to experiential welfare, rather than total well-­being, these differences between species are not important, according to Wong. He agrees with McMahan that it’s no misfortune for the dog to lack the relevant capacities of the human being. But McMahan would say that the dog is worse off, yet not unfortunate. In contrast, Wong says that the dog is not

The Difference-View  43 unfortunate and therefore doesn’t have a lower experiential welfare level, while he does have a lower absolute welfare level. Wong (2016, 16) continues: This is because the benchmark for canines is naturally lower than for humans, and the evaluation of EW (experiential welfare, T.V.) ought to reflect this asymmetry.

The fact that Wong talks about a benchmark here suggests the following picture: There is a third measurement cup, besides the above-­mentioned different-­sized cups for humans and dogs that indicate their experiential welfare, as depicted in Figure  2.3. You pour the content of the human’s ex­peri­en­tial welfare cup into this third cup. And then you empty the third cup and pour the liquid of the dog’s cup into the third cup. In absolute terms, the human’s cup contained more liquid. Yet, the third cup has a benchmark that indicates flourishing lives for humans and the liquid does not reach this benchmark. The liquid from the dog’s cup was much less in absolute terms. Yet, when poured into the third cup, this liquid reached the benchmark for flourishing lives for dogs. This benchmark, after all, is much lower than that for humans. Thus, Wong’s above-­quoted claim may be understood as: We need different-­sized cups to measure experiential welfare, due to the absolute difference in the height of the benchmarks in the third cup. To be more precise, Wong is not interested in welfare levels, but in welfare gains or losses. This is because he takes the normative perspective of ef­fect­ ive altruists, which are interested in doing as much ‘good’ as possible with a given unit of resources. This means, for example, that they want to make a maximum difference in terms of well-­being per dollar spent. For this purpose, one need not care about welfare levels as such, but instead one is flourishing life for humans

flourishing life for dogs dog

human

Figure 2.3  The absolute content of the dog’s cup and that of the human’s cup can be measured with the help of a third measurement cup, which has different benchmarks for flourishing lives for humans and for dogs.

44 

Capacity for Welfare across Species

interested in the size of benefit (or harm) that an intervention brings about. More precisely, one needs to compare possible benefits (or harms) that one could bring about in order to identify what to do. In that sense, Wong (2016, 17) clarifies: Crucially, what we’re interested in here is not a particular being’s total level of EW, that is, an absolute value of EW relativized to a minimum or max­ imum standard of wellbeing. Rather, we are concerned with the measurement of the welfare increase promoted by a particular intervention.

In the example that Wong uses to introduce the notion of EW, he imagines a human with an EW of 3 and a dog with an EW of 10. But these are just arbitrary units. In fact, Wong is not interested in whether the individual is at a 3 or a 10, but he is interested in the difference that an intervention makes. Ultimately, he is interested in the difference in terms of total or absolute well-­being. But the difference in terms of experiential welfare serves as a first step in the calculation. So, for example, if one can bring a dog from 8 to 10 or a human from 3 to 4, what matters is not the numbers as such, but the difference in terms of the number of units that one can make. Here the choice would be between two dog units or one human unit of experiential welfare. A dog unit, according to Wong, is smaller than a human unit. Perhaps two dog units are smaller, in absolute terms, than one human unit. Wong (2016, 17) continues: Even more narrowly, we are concerned with the relative differences in efficacy between two (or more) interventions, employed either within or across species.

So, ultimately, as I said, the question is how to make the greatest difference in welfare per dollar spent. I see that this is enough for effective altruist purposes. Of course, from other normative perspectives, one might be interested in other details about cross-­ species comparisons of welfare. For example, if one is an egalitarian or prioritarian and wants to know who is worst off in absolute terms, one will be interested in absolute welfare levels and not only in the sizes of benefits or harms. So, it would be desirable if some story about how to compare welfare across species could be put to work for different normative purposes. But, of course, if it were easier to know the size of benefits and harms than to determine absolute levels, the former would be a reasonable place to start.

The Difference-View  45 Indeed, Wong suggests that focusing on benefits and harms requires fewer controversial assumptions. Let us see how he supports this claim. Wong (2016, 17) provides an example, where one human, Blue, receives a pleasant hot stone massage, while another human, Red, receives an anti­ retro­viral treatment to combat his HIV. Wong claims that it is clear that Red benefits more from the medical treatment than Blue from her massage. I agree that we can intuitively make such comparisons and in cases like the one that Wong mentions they are pretty uncontroversial. Even though it is difficult to say precisely how much greater Red’s benefit is, I agree with Wong that it is reasonable to assume that he benefitted at least ten times more from the medical treatment than Blue from the massage. Wong (2016, 18) also offers a second example, where Blue’s massage is compared to the benefit that accrues to a hen that is rescued from a factory farm. Again, Wong claims that the benefit in terms of experiential welfare for the hen is at least an order of magnitude greater than that for Blue. I agree. But what are these examples supposed to show? According to Wong (2016, 19): By limiting our inquiry to the differences between two interventions’ effects on intraspecies well-­being, we make our task again more tractable than a straight assessment of how well a life goes for a being simpliciter. This is because we need not be worried about defining a maximum well-­ being against which we could index our measurement. Such a task would be quite difficult, as it is hard to make sense of what a ‘maximally flourishing’ organism might look like. Nor do we have to rely on a minimum of some sort, which sidesteps a similar problem.

So, Wong suggests focusing on the difference that some intervention makes in terms of experiential well-­being. This is what he means by ‘intraspecies well-­being’ in the above quotation: we need not immediately compare well-­ being across species. We only need to assess what difference the intervention makes for the being in question. And this is what he takes experiential well-­being to capture. How are we going about assessing the size of the difference that an intervention makes in terms of experiential well-­being? How do we arrive at the intuitive judgement that being saved from a life in a factory farm is a larger benefit for a hen than a massage is for a human? I, for one, proceed as follows: (1) I imagine how good life would be for the hen in more accommodating circumstances. (2) I imagine how bad the hen’s life would be in a

46 

Capacity for Welfare across Species

factory farm. (3) I make a rough estimate of how much better the hen’s life would be if saved from the factory farm. In this case, the hen’s life outside the farm would, I think, be at the very least twice as good. At this point (4) I imagine, in a similar way, how much better the human’s life is, due to receiving the massage. I am sure that the human’s life is not twice as good, due to receiving the massage. Instead, the benefit for the human is much smaller. Considering the human’s lifetime welfare, the massage would only amount to a slight improvement. This leads me to the conclusion that in our ex­ample the hen profits considerably more from the intervention than the human. Thus, indeed, we do not need to know what a human’s or a hen’s max­ imum or minimum welfare level would be in order to estimate the beneficial (or harmful) effects of some intervention. We only need to compare the difference that the intervention is likely to make, as compared to the counterfactual scenario without the intervention. If, instead, I wanted to know the individual’s actual welfare level, I would need to (1) imagine what a worst life for the individual might look like, (2) imagine what a best life for the individual might look like, and (3) place the individual’s actual welfare level at some point on a scale that reaches from this minimum to this maximum. Thus, I would need to guess the difference of the individual’s actual life to these endpoints of the scale. Wong is right that his own proposal spares us the challenging task of defining the endpoints of the welfare scale. However, as I said, this advantage comes at a cost. After all, for some normative purposes and from some normative perspectives, we are interested in an individual’s welfare level rather than merely in the comparative size of possible differences in welfare that we could make. Is it really so hard to define a maximum and a minimum, and even, perhaps, a neutral point, on an individual’s welfare scale? It is certainly true that in these matters we can only help ourselves to rough judgements. But this is also true for Wong’s proposal, since we can only make rough guesses about particular welfare gains and losses. But is it really significantly easier to make a rough judgement about how much better off a hen would be outside as compared to inside of a factory farm, than it is to roughly indicate a hen’s welfare level outside as compared to inside of the factory farm? If one roughly knew how the hen’s life in both scenarios would be—­something that we need to know anyway, even for determining the size of the benefit from being saved from a life in the factory farm—­I think one could also estimate the hen’s welfare level. Let’s say, outside of the farm in reasonably good conditions, it would possibly be an 8 out of 10. Inside the farm, it would possibly be close to zero or below. For the human, depending on the

The Difference-View  47 precise circumstances, the welfare level (or EW level, if you like) would probably be estimated at roughly a 7 out of 10 and the stone massage wouldn’t make a big enough difference to be visible on such a rough scale of lifetime welfare. Of course, if one had a more fine-­grained scale, or if one considered the human’s welfare at a day rather than her lifetime welfare, the stone massage would make a visible difference. Thus, making rough claims about welfare levels (rather than merely the sizes of benefits or harms) seems possible to me. From the perspective of particular accounts of welfare, setting the maximum, minimum, and neutral point on the welfare scale is straightforward. For example, according to preferentialism it can be said very roughly that a being that has all its preferences fulfilled would be maximally well off and a being with all its preferences frustrated would be maximally badly off. Alternatively, one might consider the individual’s overall emotional state, success in important projects, fulfilment of basic needs or some such things, depending on one’s underlying account of welfare. If it were possible for human welfare (or that of other species) to potentially rise far above what has hitherto been shown, perhaps by means of some yet-­to-­be-­developed drug or technology, this would indeed make it impossible for us to fix a maximum welfare level now, given our current ignorance of human (or other species’) true potential. From an evolutionary perspective on the origin, nature, and function of our hedonic capacity, as explained in section  3.3, I do not consider this to be possible, though. So, I think that we can estimate welfare levels as points on a welfare scale without major problems. We now have an idea of the overall structure of Wong’s proposal for how to compare welfare across species. First, he assesses the difference that a particular intervention makes for an individual. For example, some intervention may increase an individual’s lifetime welfare level, relative to this individual’s potential for welfare, only slightly or very significantly, or something in between. For this assessment, one compares the situation (without the benefit) to the counterfactual situation (with the benefit). This indicates the size of the benefit in terms of what Wong calls ‘experiential welfare’. But then, if one needs to compare such benefits across species, one needs, in a second step, to bring in a weighing factor. The weighing factor is based on species-­typical hedonic capacity. An individual’s absolute benefit is the experiential benefit discounted by the weighing factor. So, for ex­ample, a big experiential benefit for a dog may be smaller in absolute terms than a minor experiential benefit for a human, according to Wong’s line of argumentation. This is due to differences in species-­typical hedonic capacity. So,

48 

Capacity for Welfare across Species

Wong’s proposal depends not only on hedonism but also and importantly on the assumption that there are species-­typical differences in hedonic capacity. This assumption will be explored in the reminder of this section. The first thing to say about Wong’s appeal to what he calls ‘hedonic capacity’ is that this shows that he assumes hedonism as the underlying account of welfare. Wong’s proposal is supposed to be useful primarily for effective altruists. But effective altruists do not as such commit to hedonism (MacAskill 2019, 14). This makes Wong’s proposal somewhat restricted in scope. I think this is not necessarily a problem. After all, hedonism may be the correct theory of welfare. And, anyway, it may be impossible to provide an account of cross-­species capacity for welfare that remains ecumenical as to what the correct account of welfare is. Nevertheless, I think it is im­port­ ant to make such crucial assumptions explicit. Wong doesn’t do so. Since it doesn’t make sense to say that humans have a higher capacity for ‘self-­ fulfilment’, ‘nature-­fulfilment’, or ‘preference satisfaction’ than dogs, it is clear that other theories of welfare would not provide a basis for the weighing factor that Wong needs. A dog, after all, has a different nature, but it is no less possible to fulfil his nature. The same holds for preference satisfaction and self-­fulfilment. For a hedonist, in contrast, it is at least conceivable to claim that certain individuals have a higher capacity for pleasant ex­peri­ ences. For an objective list theorist who accepts the same list across different species, as McMahan (see section 2.4) and Kagan (see section 2.1) do, it also makes sense to speak about some capacity to realize the goods in question that varies for different species. But from the perspective of an objective list account, it would be clearer to talk about a ‘welfare capacity’ rather than a ‘hedonic capacity’. I explored the strategy of objective list theorists in section 2.1 and will focus on hedonic capacity here. Are there species-­typical differences in hedonic capacity or hedonic potential? I call the claim that such differences exist the ‘different-­hedonic-­ capacities claim’. I distinguish two versions of the different-­hedonic-­capacities claim. The first, narrow, version of the claim links hedonic capacities to cognitive capacities in particular. Different-­hedonic-­capacities claim, narrow version: Species with higher cognitive capacities have higher hedonic capacities.

The second version of the different-­hedonic-­capacities claim simply states that among the beings endowed with hedonic capacities, some have higher such capacities than others. The difference, according to this second,

The Difference-View  49 broader, version of the different-­hedonic-­capacities claim, need not be based on different cognitive capacities. Different-­hedonic-­capacities claim, broader version: Species have different hedonic capacities, not necessarily due to different cognitive capacities.

What are the relevant capacities that determine hedonic capacity? Wong (2016, 23) mentions ‘characteristics that facilitate pleasurable experiences’ and ‘properties related to experiencing pleasure’. I can see, of course, that one needs to have the property of being sentient in order to have hedonic capacities at all. But this is not what Wong has in mind here, since we are talking about differences among beings who are granted to have hedonic capacities. Instead, Wong talks about properties that he takes to enable beings to experience a greater amount of pleasure. I take it to be uncontroversial that (a) different individuals suffer from and enjoy different things and (b) different individuals experience different amounts of pleasure or suffering from any given source, and (c) there can be species-­typical differences in these regards. For example, in cases of captivity, depending on the details, cognitively more advanced beings may suffer more (since they may fully realize how terrible their situation is), or cognitively less advanced beings may suffer more (since they cannot make sense of their situation and cannot see that it is only temporary and to their overall benefit). Animals have different needs, different preferences, and different sources of pleasure. Therefore, any given situation can have different impacts on the well-­being of different animals. This, as such, does not establish that animals with ‘more sophisticated abilities’ (Wong 2016, 24) have a greater capacity for pleasure. By the way: What counts as a more sophisticated ability? Is it the dog’s olfactory sense, the bat’s sense of echolocation, the bird’s capacity to fly? There are so many capacities that non-­human animals have and that we lack. These capacities allow them to flourish. We also have capacities that they lack. These capacities are important for us and allow us to flourish. I think it is problematic to consider the things we are good at ‘more sophisticated’. And I also don’t see why the things that allow us to flourish should

50 

Capacity for Welfare across Species

facilitate flourishing in general. A dog would probably say that higher olfactory capacities facilitate flourishing in general and a bat would say that echolocation facilitates flourishing in general. From an evolutionary perspective it doesn’t make sense to say that currently living humans are more advanced than currently living dogs or bats. So, talk about ‘more sophisticated beings’ is problematic. Be that as it may, the crucial question is whether individuals or species that possess a hedonic capacity possess it in different degrees. In section 3.3 I will argue that this is not the case. My argument will be based on an evolutionary understanding of the origin, nature, and function of hedonic cap­ acity. For now, I will simply evaluate the arguments that Wong presents in favour of the different-­hedonic-­capacities claim. So, let us take a closer look at Wong’s understanding of ‘hedonic capacity’ and his argument for ‘fundamental’ inter-­species differences in this regard (Wong 2016, 23). These differences in hedonic capacity are supposed to justify the ‘scaling factor’, i.e. the discounting of some species’ experiential welfare in the calculation of absolute welfare (Wong 2016, 23). As I said, Wong does not propose any specific discount rate, but he argues that some discount rate is required on the basis of differences in hedonic capacity. What, then, determines hedonic capacity, according to Wong? He says: It is worth noticing that the qualities relevant to a particular animal’s hedonic capacity involve both the characteristics that facilitate pleasurable experiences, and the faculties that increase the acuity of a being’s suffering. (Wong 2016, 23)

So, what characteristics may facilitate pleasurable experiences or suffering? Wong mentions an example in which the same conditions of confinement are worse for species that suffer not only the physical harm but also the harm from deprivation of a social life in conditions of confinement. I think it is uncontroversial that the same treatment may cause different levels of suffering for different species or individuals, due to their different needs. But this observation does not justify talk about different hedonic capacities, unless individuals of species A consistently suffer more than those of species B, no matter what their common treatment is. Wong’s example doesn’t show this. Wong (2016, 24) speaks about ‘the constituents of hedonic capacity’:

The Difference-View  51 The capacity to realise pleasurable experiences consists in several more specific faculties, each of which accounts for at least some part of the experiences that we like to think make up a flourishing life. (Wong 2016, 24)

Wong thus proposes to focus on specific capacities that make up or impact our hedonic capacity. In particular, he focuses on (1) sentience, (2) higher cognitive capacities, and (3) sociality. Wong (2016, 24) considers these things to be ‘proxies’ for hedonic capacity. Let us thus see what he says about each of these capacities in turn. Wong is not very specific in his definition of sentience, the first constituent of hedonic capacity that he discusses. On the one hand, Wong understands ‘sentience’ as the capacity to have subjective experiences. This, thus, refers to a particular understanding of consciousness in terms of ‘qualia’: the capacity to experience an internal ‘what-­is-­it-­like-­ness’ (Wong 2016, 24). On the other hand, Wong refers to the understanding of ‘sentience’ in terms of simply the capacity to experience pleasures and pains. Wong claims that sentience, thus understood, is relevant for hedonic capacity, because it is its precondition. In his further remarks on sentience, Wong shares the common view that many non-­human animals are sentient, and he explains how scientists assess the presence of sentience and the relative subjective strength of various preferences. Wong says: In each of these studies, animals demonstrated clear inclinations to secure particular ends over others, pointing strongly to their ability to have specific preferences, a manifestation of sentient capacity.  (Wong 2016, 26)

Here, Wong links the capacity to choose among competing motivations with the ‘sentience capacity’. All this sounds plausible. However, with regard to sentience, Wong does not even try to defend differences in hedonic cap­ acity between sentient species! But this would be required in support of his different-­hedonic-­capacities claim. Wong (2016, 27) discusses ‘higher cognitive capacities’ as the second constituent of hedonic capacity. Here Wong basically sides with Mill. He believes that there are higher and lower pleasures and those who are familiar with both sorts of pleasure can surely tell us what the higher pleasures are.

52 

Capacity for Welfare across Species

On balance, however, I am inclined to believe that Mill’s appeal to the notion of a ‘higher arbiter’ is compelling. (…) The fact that such people prefer sophisticated pleasures to baser ones, I am convinced, indicates that the former must bring about greater quantities of hedonic experience. (Wong 2016, 28)

This suggests that Wong’s proposal for comparing welfare across species relies on Millian superiorities, which are, even among hedonists, not widely accepted and highly controversial. I criticize Millian superiorities in section 2.3. Cognitive capacities come in degrees, for sure. It is plausible that species and individuals with different cognitive capacities derive pain and pleasure from different sources. But do cognitively more advanced individuals have a capacity for greater pleasure and pain? I can’t see why this should be the case. A dog would probably say that a being with greater olfactory capacity has a higher capacity for pleasure. Even if cognitively more advanced humans have a greater capacity for pleasure than cognitively less advanced humans (which I doubt), why should cognitively more advanced humans have a greater capacity for pleasure than otherwise more advanced dogs? Wong (2016, 27) accepts what he calls the ‘hedonic connection thesis’. It says that the capacity to realize a broader range of pleasures is accompanied by an associated capacity to realize greater quantities of pleasure.

But why should this be the case? How are we supposed to individuate different kinds of pleasures? Is, for example, echolocation just one kind of pleasure? Or are there more fine-­ grained kinds of pleasure, related to different uses of echolocation? The same holds for our cognitive capacity: Do they count as one kind of pleasure, cognitive pleasure, or should we individuate more fine-­grained cognitive pleasures, such as reading books, listening to lectures, etc.? Should we separate different kinds of pleasures in other ways, for example musical pleasures, visual pleasures, etc.? Or are these simply aesthetic pleasures? Perhaps for a dog, smelling food is a different kind of pleasure from smelling conspecifics. I think that we are likely to be biased in favour of humans when it comes to individuating kinds of pleasures and identifying the range of pleasures that a being can have.

The Difference-View  53 Even if we can individuate pleasures in an unbiased way, the hedonic connection thesis is implausible, because humans are not good at multitasking. We typically cannot give our full attention to two activities that demand our attention at the same time. So, if we enjoy a great novel, music, and sex at the same time, the overall pleasure is not necessarily more (and perhaps rather less) than if we focus on one such thing at a time. Since our focus is limited, the experience is not better if we do more good things at a time. Perhaps we (as humans) would be bored if we devoted ourselves to one kind of pleasure all the time. But other animals may not have that problem. So instead of enjoying different kinds of pleasures in a row, as we tend to do, they may enjoy the same kind of pleasure for a longer period. As long as they are still enjoying themselves, this may deliver the same amount of enjoyment. Thus, the hedonic connection thesis does not seem obvious. The third constituent of hedonic capacity that Wong discusses is sociality. He points out that not only for humans, but also, for example, for the animals that we keep in factory farms, a rich social life is very important for their flourishing. As Wong (2016, 30) points out: For these beings, social interaction textures nearly every facet of their existence: it mediates the nurturing relationship between mother and ­offspring, grounds the affection between male and female mates, and ­promotes cooperation among animals that live in groups.

This is certainly true. According to Wong (2016, 31), this suggests ‘that at least among naturally social animals, a key constituent of hedonic capacity bears on this ability to have relationships’. Here Wong seems to allow for the possibility that the hedonic capacity of different animal species has different constituents. After all, he suggests that sociality may not be such a constituent for animals that are not naturally social. This suggests that what Wong labels ‘proxies’ for, or ‘constituents’ of, hedonic capacity are just things that tend to be important for the happiness or pleasure of a given species. This, however, does not support the different-­hedonic-­capacities claim on which his argument relies. To conclude, Wong offers an elaborated account of how to compare welfare across species and of cross-­species capacity for welfare, but he does not offer any support for it, except, perhaps, for a brief mentioning of Millian superiorities. Even though Wong’s work is less influential than that of other authors that I discuss in this book, I engaged with it at some length. This is because

54 

Capacity for Welfare across Species

Wong spells out ideas that other people may share, and his work gets attention from leading effective altruists. Peter Singer supervised Wong’s thesis and was exceptionally enthusiastic about it. Independently of its value as a thesis, I think it is instructive to point out the shortcomings in Wong’s argumentation to advance our thinking about cross-­species capacity for welfare.

2.6  Budolfson and Spear’s Formula Mark Budolfson and Dean Spears (2020, 606) start their discussion of how to compare welfare across species with a reference to Wong’s work (which I discuss in section 2.5): Recent work by Kevin Wong has highlighted the most difficult problem that needs to be solved in connection with interspecies comparisons, which is how to estimate the wellbeing capacity (wellbeing potential) of members of a non-­human species relative to the wellbeing capacity of humans.  (Wong 2016)

This would indeed be a practical challenge, provided that one bought into all the controversial philosophical and empirical claims that Wong needs in order to uphold his overall line of argumentation. Budolfson and Spears offer some hand waving in the direction of possible ways to estimate the cognitive capacity of different animal species. But we only need an estimate of cognitive capacities for purposes of welfare assessments, if we assume that cognitive capacities are related to well-­being capacity. This, however, is far from clear, as I argue in my discussion of Wong in section 2.5. Here, I want to explore Budolfson and Spears’ take on this issue. Budolfson and Spears provide an example of the work that knowledge of species-­based well-­being capacity could do. Their example concerns harms due to one additional degree of climate change for humans and birds. They ask us to imagine that one million life years of a particular bird species are lost, due to one additional degree of climate change. How could we compare this welfare loss to the loss of the same amount of human life years? According to Budolfson and Spears, the answer would be simple if we knew the bird’s species-­based well-­being capacity. After all, in that case, we could simply multiply the amount of lost bird life years by the bird’s well-­being capacity, which would be a fraction of the human well-­being capacity. Budolfson and Spears take it that well-­being capacity is based on cognitive

The Difference-View  55 capacity. If, for example, the birds’ welfare capacity was half that of humans, we would need to multiply the number of lost bird life years by 0.5 in order to compare the welfare loss of humans and birds on the same scale, or so Budolfson and Spears argue.1 The authors do not discuss different possible ways of accounting for species-­specific differences in lifespan. I will say more about this issue in section 4.2 and ignore it here. After suggesting how to calculate the bird’s welfare loss on the same scale as that of humans, Budolfson and Spears add that this holds under the assumption that the relevant amount of global warming ‘does not change the quality of life of those birds’. They go on saying that if the climate change does diminish the quality of life of the remaining birds of that species, we can simply multiply the number of remaining bird species life years by a further quality of life adjustment term that is itself an empirical impact estimate from zoological experts and the like. (Budolfson and Spears 2020, 606)

This is to say that after the discounting on the basis of lower cognitive capacities the total number of lost life years of the relevant species of bird would have to be further discounted in order to account for the fact that the additional life years that the birds would otherwise have had, would have been life years with a reduced quality of life. This would make sense, if the authors were aiming at calculating the harm of death for these individuals. But this is not their task here. After all, the author’s stated aim was to take into account the harm that climate change causes, both to humans and to birds. Thus, we need to compare the individual’s welfare in case of climate change with their welfare in the closest possible world without climate change. And in the closest possible world without the additional degree of climate change, the bird’s welfare level is not reduced by an additional degree of climate change, since there is no such additional degree of climate change in this counterfactual world. Thus, the second discount factor is out of place here. However, it would make sense to further discount the lost

1  As an aside, Budolfson and Spears do not specify whether these are lost life years due to premature death or due to never coming into existence. Depending on one’s view in population ethics, this could make a huge difference when it comes to our reasons for action. I will assume here that they are talking about premature death, in order to make the example acceptable for proponents of different theories in population ethics. I will further discuss this issue in section 5.2.

56 

Capacity for Welfare across Species

years of life of the birds, if their quality of life was higher or lower than it typically is for some reasons unrelated to climate change. Budolfson and Spears present an equation for the average lifetime well-­ being experienced by a member of a species s. They present this as a function of the average well-­being capacity per unit of time of members of s relative to humans, multiplied by the average duration of a life of a member of s, multiplied by a quality-­of-­life adjustment term. So, the function is supposed to yield a welfare level, namely the average lifetime welfare level of a member of some species. What are the factors that we need to multiply, to arrive at the average lifetime welfare level of a member of some species? Let us first do the calculation for a human. According to Budolfson and Spears, we first need to know the average well-­being capacity per unit of time for humans. (We don’t know yet exactly what this is, but we will find out soon.) Then we need to know the average duration of a human life. Multiplying this obviously yields the average lifetime welfare level, assuming that lifetime welfare is the sum of temporal welfare levels. For example, if a human’s average temporal welfare level (i.e. the average for members of the species) is 5 and if an average human being lives for 70 years, then the human’s average lifetime welfare level is 5 times 70, thus 350. This holds, unless we have reason to assume that human welfare in the case under consideration was above or below average. In that case, we would have to multiply 350 by a quality-­of-­life adjustment term, which would presumably be bigger or smaller than 1, depending on the divergence from the average synchronic quality of life. For a member of a species with lower cognitive capacities, we would also need to discount the result by some factor, namely by the average well-­being capacity per unit of time of members of s relative to humans. Now, we know what Budolfson and Spears have in mind when they talk about well-­being capacity. It is the average well-­being capacity per unit of time of members of s relative to humans. For beings with lower cognitive capacities than humans, this would be a number smaller than 1. Budolfson and Spears suggest that this discount factor might (as a first shot) be based on the extent to which a typical member of the non-­human species under consideration has fewer neurons than a typical human in some specified part of their brains. It is important to carefully examine the assumptions that underlie that proposal. Contrary to McMahan’s notion of fortune, Budolfson and Spears’ notion of well-­being capacity as they use it here does not link cognitive capacities to the species-­based maximum welfare level that an individual has

The Difference-View  57 the potential to reach. So, it’s not about the highest heights (or the lowest lows) that an individual could in principle attain. I think it is wrong, but at least conceivable that the welfare level that an individual could attain under ideal circumstances depends on that individual’s species-­based cognitive capacities. It is conceivable that the size of the well-­being glass, as it were, is different for different species. But that is not Budolfson and Spears’ claim. And contrary to Wong, Budolfson and Spears’ claim is not about the absolute size of a unit of temporal experiential welfare. I think it is wrong, but at least conceivable, that a unit of so-­called experiential welfare of beings with higher cognitive capacities is bigger in absolute terms than a unit of ex­peri­en­tial welfare of beings with lower cognitive capacities. Conceivably, this is because larger cups might come with different scales. But, again, this is not Budolfson and Spears’ claim. The authors do not make theoretical suggestions concerning the size of the cups or the size of the units on the scales. Instead, they want to suggest that while the cups may or may not have the same size and the same scales, what differs is the average extent to which the cups are filled. Budolfson and Spears do not tell us any story along the lines of Wong or McMahan. They do not tell us why we should think that cognitive capacities matter for welfare. To be sure, stories can be told about the relationship between the number of neurons in some part of the brain and an individual’s cognitive capacity. That is something I am willing to grant. But why should we think that individuals of cognitively stronger species are better off on average? Do they experience more pleasure? Are they happier? Are they more successful in fulfilling their species-­specific nature or their individual projects? Are they more successful in getting what they want? How would I know whether an average human has less frustrated desires or more fun than an average dog? And how is cognitive capacity supposed to be rele­vant in that regard? In order to defend such claims, you need a story that could at least conceivably make sense. I think that McMahan and Wong have at least come up with something. The problem is that their stories are not entirely convincing and even if they were, they wouldn’t allow us to make the claims that Budolfson and Spears want to make. As Singer and Lazari-­Radek (2014, 265) point out in a different context: Perhaps it can be argued that self-­aware beings are capable of greater pleasure than beings lacking in self-­awareness, but presumably they are also capable of greater misery. It isn’t clear that the surplus of happiness over misery of self-­aware beings is really greater than that of beings without self-­awareness.

58 

Capacity for Welfare across Species

Note that I don’t believe that self-­aware beings are capable of greater happiness. But my point is that even if one believes they are, it doesn’t follow that they are better off on average. Budolfson and Spears are not ‘wedded’ to the metric of brain neurons and, as I said, my criticism of their approach does not concern this specific metric. It concerns the underlying story or rather the lack thereof. Budolfson and Spears (2020, 607) say: When greater accuracy is required for specific species or individuals, n can be set equal to a more complex metric based on expert analysis of empirical properties that are best correlated with different levels of wellbeing (which might differ according to different substantive theories of wellbeing)—e.g. the number of neocortex-­like neurons, cortisol levels, sociality, or other leading factors identified by the scientific community and philosophers as most closely correlated with wellbeing capacity.

There are many problems with this suggestion. The problems are much more fundamental than the choice of the complexity of the metric. First of all, in order to choose a metric, we need to determine what we want a metric for. Is it for cognitive capacity? Or for hedonic capacity? Or for capacity to be social? Second, in order to know what we want a metric for, we need to know what is (or at least what we take to be) the correct theory of welfare. After all, we can only know what determines well-­being capacity if we know what determines well-­being. It makes a huge difference, whether you are a hedonist, a preferentialist, an objective list theorist, or a nature-­fulfilment theorist, to mention just a few options. Third, for most of these theories of welfare, we do not need to discount based on welfare capacity, because according to these theories of welfare, species do not have different welfare capacities. In the quoted passage, Budolfson and Spears mention examples of ‘leading factors’ that both the scientific community and philosophers allegedly identified as ‘being most closely related to wellbeing capacity’. And then they provide nine references to the work of philosophers and scientists. Unfortunately, as far as I could see, none of these papers or books mentions the concept of ‘wellbeing capacity’ or any related concept at all. The publications are concerned with empirical claims about capacities of animals, such as the capacity for consciousness or the capacity to feel pain, or so-­called ‘higher cognitive capacities’. But the concept of ‘wellbeing capacity’ is a philosophical construct. It is not evident at all, how it should be related to these empirical capacities. Just as empirical scientists cannot determine

The Difference-View  59 what the correct theory of welfare is, they cannot determine the correct way of comparing welfare across species. This is a question that value theory needs to answer. And normative ethics needs to inform us about the link between welfare considerations and normative reasons for action. Budolfson and Spears conclude their paper as follows: In sum, the method outlined here allows interspecies comparisons based on empirically available estimates of species population dynamics and within-­species quality of life adjustment, together with empirical proxies for wellbeing capacity n that can be calibrated . . . to reflect normative uncertainty about the connection between those empirical proxies and wellbeing capacity.

I agree with the authors that there is currently normative uncertainty about the comparative well-­being capacity of humans and other species and that this issue deserves further consideration. However, I think that the normative uncertainties are more profound than the authors seem to realize. Mark Budolfson (2020) emphasizes that his main point is the proposal of the formula as such, rather than any specific numbers to be fed into it. If we understand the formula in this way, it is compatible with all kinds of substantive positions on cross-­species capacity for welfare, including EQU. I see the value of having such a formula, since it can serve as a framework for discussion about what numbers to feed into it. To make these decisions, philosophical and empirical questions need to be addressed. Before we can accept brain neurons or anything else as a proxy for welfare capacity, we need at least a plausible story about the relationship of that empirical measure to capacity for welfare. Budolfson and Spears do not provide such a story at all.

3 The Equality-­View In the present chapter, I want to point out that EQU has a lot of intuitive appeal and seems, all things considered, at least as plausible as DIF, if not more so. For this purpose, I will first provide an overview of what animal welfare scientists have to say about measuring and comparing welfare and show that it is in line with EQU (section 3.1). Then I will present the theory of welfare as self-­fulfilment as an example of a philosophical theory of welfare that is in line with EQU. This illustrates the relativizing structure that all such theories share (section 3.2). Lastly, I present an evolutionary perspective on animal capacities, which lends additional support to EQU (section 3.3).1

3.1  Animal Welfare Science Ruth Harrison’s book Animal Machines (2013), first published in 1964, which criticizes modern animal farming, led to public and political interest in animal welfare. In 1965, this resulted in the UK government commissioning Professor Roger Brambell to lead an investigation into the welfare of intensively farmed animals. The Brambell Report stated: An animal should at least have sufficient freedom of movement to be able without difficulty, to turn round, groom itself, get up, lie down and stretch its limbs.  (Brambell Committee 1965)

The content of these so-­called ‘five freedoms’ has been changed throughout the years. Veterinarians and many organizations, such as the World Organisation for Animal Health, the Royal Society for the Prevention of Cruelty to Animals, and the American Society for the Prevention of Cruelty to

1  Parts of section  3.2 are reproduced with permission from Philosophy and Public Policy Quarterly from Višak and Balcombe (2013). Capacity for Welfare across Species. Tatjana Višak, Oxford University Press. © Tatjana Višak 2022. DOI: 10.1093/oso/9780192882202.003.0003

The Equality-View  61 Animals accept the five freedoms account. Currently, the five freedoms are formulated as follows: 1. Freedom from hunger and thirst by ready access to fresh water and diet to maintain health and vigour. 2. Freedom from discomfort by providing an appropriate environment including shelter and a comfortable resting area. 3. Freedom from pain, injury or disease by prevention or rapid diagnosis and treatment. 4. Freedom to express normal behaviour by providing sufficient space, proper facilities, and company of the animal’s own kind. 5. Freedom from fear and distress by ensuring conditions and treatment, which avoid mental suffering.2 The five freedoms are meant to be a minimum welfare standard for animals under human control. They are not to be understood as a full-­fledged account of welfare and their applicability to, for example, free-­living animals is questionable. If we wanted to use it to compare welfare across species, we would need to check to what extent an animal has the five freedoms. In principle, animals of different species can possess the five freedoms to the same extent, suggesting that they have the same capacity for welfare, according to this account. The five freedoms approach has been criticized and alternative approaches have been suggested by authors who want to assess animal welfare in general and not only focus on some minimum human duties towards animals under their care. For example, it has been argued that the five freedoms approach is overly concerned with ‘freedom from’ and should also or rather consider ‘freedom to react to’. The idea is that animals, and in particular free-­living animals, regularly encounter challenging situations and experience aversive feelings such as hunger, thirst or discomfort. As long as the animal can appropriately react to these feelings, for example by searching for food and water, and ultimately eat and drink, the welfare of the animal is not at peril. For example, Frauke Ohl and Franz Josef Van der Staay point out that so-­called ‘negative’ states are part of normal behaviour. They reformulate the five freedoms account and talk about the animal’s

2  See www.Aspcapro.org, accessed 4 September 2020.

62 

Capacity for Welfare across Species

[. . .] freedom adequately to react to hunger, thirst or incorrect food, thermal and physical discomfort, injuries or diseases, fear and chronic stress, and thus the freedom to display normal, species-­specific behavioral patterns and adapt to changing living conditions up to a level which it perceives as positive.  (Ohl and Van der Staay 2012, 15–16)

Both Ohl and Van der Staay’s proposal and the more classical five freedoms account talk about welfare in terms of the absence of suffering and the ability to perform species-­specific behavioural patterns. This, again, suggests support for EQU. After all, in principle all welfare subjects have an equal capacity to be free from suffering and perform species-­specific behaviour. Donald Broom is one of the most influential animal welfare scientists. Broom (1998, 394) presents his understanding of ‘welfare’ with references to the work of other animal welfare scientists and to his own earlier work as follows: Welfare is a term restricted to animals, including humans, and hence not used for other organisms or inanimate objects. It is used in science and in legislation and therefore must have a meaning precise enough for such use. Welfare refers to a characteristic of an individual rather than to something given to it, it must be measurable in a scientific way, and it must vary over a scale from very good to very poor (Curtis 1986, Duncan 1987, Broom 1988, 1991, 1996, Broom and Johnson 1993, 74–76). The original use of the word welfare, meaning how well an individual fares or goes through life, is followed by Broom (1986): the welfare of an individual is its state as regards its attempts to cope with its environment. Its state includes how well or badly it is coping and how much difficulty it is having in coping. As emphasized by Broom (1991, 1996) and Broom and Johnson (1993, 80–82) the feelings of the individual are an important part of that state. The assessment of how good or how poor the welfare is depends on a wide range of measures of behavior, physiology, brain functioning, immune system functioning, pathology, injury, and life expectancy.

In this passage, Broom mentions the importance of feelings as parts of coping systems and thus of welfare. This can be seen as a reaction to his critics, in particular Ian Duncan (1996), who defends a hedonistic position in animal welfare science. Duncan argues that animal welfare is wholly a matter of positive or negative mental states. As such, he urges proponents of the five freedoms account to consider the importance of positive emotions,

The Equality-View  63 rather than merely the absence of negative ones. And he challenges Broom not to forget the role of the emotions. After all, Broom’s approach focuses completely on coping. Broom (1998, 394–395) reacts to this challenge, when he says [. . .] although feelings are part of coping systems, they do not make up all of them, so if the concept of welfare is to be usable it must refer to all aspects of coping systems and not just feelings (Broom 1993). The sep­ar­ ation of feelings from all other aspects of coping systems in biology is unsound and impractical when welfare assessment is attempted.

This shows that Broom is not in favour of hedonism as an account of welfare. Instead, he views the feelings as only one aspect of welfare, besides ­others. Or perhaps the criteria that he mentions are indicators, rather than aspects of welfare. They indicate how well an animal is coping with its en­vir­ on­ment. The ‘environment’ in this context is everything outside the individual’s brain (Broom 1998, 395). If welfare consists in coping with one’s environment, as Broom has it, and if feelings are indicators of how well an individual copes, it follows that welfare subjects of different species have the same capacity for welfare. There is a controversy between animal welfare scientists who focus more on subjective experiences and those who propose more objective criteria. Some authors propose to understand welfare in terms of the animal’s survival and reproduction (Barnard 2007, 7). This goes against a more dom­in­ ant view that focuses on the animal’s subjective experiences (Browning 2019, Mellor and Beausoleil, 2015, 242). Marian Stamp Dawkins focused on animal feelings as well (Dawkins 1990, 1), but more recently, she argues that animal welfare assessments should be based on the animal’s health and preferences rather than on their subjective feelings, since the former are easier to understand and assess (Dawkins 2017, 7). Other animal welfare scientists understand welfare in terms of the fulfilment of the animal’s needs (Bracke 2006). This leaves it unclear whether needs are simply considered to be strong desires or whether desires are thought to be imperfect indicators of needs. Yet other animal welfare scientists think of welfare in terms of nat­ ural behaviour, health, or some combination of factors, such as natural behaviour, positive affective experiences, and normal biological function (Fraser et al. 1997). Peter Sandøe (1996, 12) understands animal welfare completely in terms of preference-­satisfaction:

64 

Capacity for Welfare across Species

A subject’s welfare at a given point in time, t1, is relative to the degree of agreement between what he/ it at t1 prefers . . . and how he/ it at t1 sees his/ its situation—­the better agreement the better welfare.

Some animal welfare scientists use preference tests to find out what an animal’s preferences are. They ‘ask’ the animal in an indirect way by assessing how much the animal is willing to work in order to gain access to a particular resource or to avoid a particular threat (Dawkins 1990, Duncan 1993, 1996, Broom and Johnson 1993, 145–157). Preference tests are used in order to make a choice among different conditions for kept animals and to evaluate different procedures involving animals, for example in husbandry systems. As with humans, a non-­human’s preferences may not always be in line with what is good for the individual. For example, humans would probably signal a greater ‘need’ for ice cream in comparison to more healthy food than is actually good for them. So, their willingness to invest effort in order to get something is not always a good guide to what is good for them. Preferences may depend on previous experiences and addiction, among other things (Mendl 1990, Duncan 1987). This is true for humans and for non-­human animals. Therefore, as Broom (1998, 397) points out: The preference studies do not provide comprehensive information about the welfare of the individuals but should be followed up by studies using other welfare indicators to make sure that what is preferred does not lead to poor health or other problems.

This is also why Dawkins (2017, 7) proposes to combine preferences with indicators of health when it comes to measuring welfare. How do these various accounts of animal welfare relate to the philo­soph­ ic­al theories of welfare as presented in section 1.5? Relating the accounts of animal welfare scientists to philosophical theories of welfare is not straightforward, because it is not clear for many of the approaches in animal welfare science, whether they intend to provide theories of welfare or rather indicators of welfare. Duncan, Browning, as well as Mellor and Beausoleil seem to assume hedonism about welfare, while Sandøe defends a preferentialist theory. Elements of a nature-­fulfilment theory can be found in approaches that focus on the capacity of performing species-­specific behaviour, on natural behaviour, and on coping with one’s environment. Typically, these concerns are combined with concerns about the animal’s subjective experiences, which shows hedonistic concerns as well. These approaches may alternatively be understood as listing indicators of welfare. Similarly, those who focus on

The Equality-View  65 objective criteria, such as survival, reproduction, and health may either present objective list accounts of welfare or merely indicators, without explicitly defending any theory of welfare. Since the focus in animal welfare science is on practical measurement of welfare, the theoretical basis often remains implicit. Animal welfare scientists typically leave it unclear whether the criteria that they mention are meant to be ultimate personal goods or instrumental goods. Are they meant to be aspects of welfare, or merely indicators of welfare? And how exactly do animal welfare scientists understand central concepts, such as ‘needs’? I will not here try to spell out or defend any of these views that are commonly assumed in the field of animal welfare science. I leave this somewhat messy picture as it is. I will try to make sense of these views from an evolutionary perspective in section  3.3. The purpose of this little rather unsystematic glimpse into the field of animal welfare science merely serves one purpose: showing that the approaches of animal welfare scientists seem to be closer to EQU than to DIF. None of the perspectives in animal welfare science that I discussed—­and I take this to be a representative selection of perspectives—­provides any support for DIF. Rather, the opposite is the case. If, for example, welfare measures the extent to which one is coping with one’s environment, it seems to follow that different species have the same capacity for welfare: animals can, in principle all cope to a greater or lesser extent, depending on the individual circumstances. They can, similarly, fulfil their needs or preferences to a greater or lesser extent and complete fulfilment equals maximum welfare. It follows that a welfare score of 8 out of 10 for a chicken is the same as an 8 out of 10 for a mouse, because welfare is nothing but the extent to which an individual is coping or fulfilling her preferences or needs. Approaches that focus on subjective experiences, may, assuming greater hedonic capacities, be compatible with DIF. But they are in line with EQU if one assumes equal hedonic capacities. (I argue against different hedonic capacities and for equal hedonic capacities in sections 2.5, 3.2, and 3.3.) Thus, even if animal welfare scientists may not explicitly compare welfare levels or capacity for welfare across species, their approaches seem to be compatible with EQU.

3.2  Welfare as Self-­Fulfilment Most of the theories of welfare listed in section 1.5 (Table 1.2), which are the main theories on offer, support EQU. In this section, I will introduce the self-­fulfilment account of welfare to take a somewhat closer look at a theory

66 

Capacity for Welfare across Species

of welfare that supports EQU and to illustrate its typical, relativizing structure. Daniel Haybron’s (2008a, 177–199) self-­fulfilment account of welfare is a eudemonistic account of welfare, because it defines welfare as nature-­ fulfilment. Roughly, nature-­fulfilment accounts of welfare hold that an individual is well off to the extent to which it lives according to its nature, where ‘nature’ is typically related to the individual’s species. A challenge for these accounts is determining what the nature of some species is. Aristotle, for example, held that a good life for a human is the life of a philosopher, since the capacities for rational thinking and contemplation are characteristic for human lives. All eudemonistic accounts of welfare are in line with EQU. After all, an individual mouse can in principle be as good an exemplar of a mouse as an individual human can be an exemplar of a human. In contrast to traditional Aristotelian eudemonism, Haybron (2008a, 177–199) defines welfare as self-fulfilment. Traditional Aristotelian accounts of welfare are perfectionist. They link welfare to being a good exemplar of one’s kind, for instance being an excellent human being. Traditional Aristotelian accounts of welfare are also externalist. What it means to be an excellent human being is not wholly dependent on the personal characteristics of the individual in question, but on general ideas about what a good human being is. In contrast, the self-­fulfilment account of welfare is non-­ perfectionist and internalist: it defines what is good for an individual primarily based on the individual’s constitution as an individual. According to Haybron, welfare consists primarily in self-­fulfilment, which, in turn, consists of two aspects: (1) happiness and (2) success in identity-­related projects. Haybron conceives of the ‘self ’ in a naturalistic way. It comprises, firstly, the individual’s emotional nature or ‘emotional self ’. The individual’s ­emotional nature is an important aspect of who the individual is. One’s emotional nature determines, for instance, what makes one happy. Thus, happiness is an important aspect of self-­fulfilment, because it is the fulfilment of one’s emotional nature. Secondly, the individual’s ‘self ’ comprises the individual’s social identity. This concerns her social role and how others see her. Thirdly, it comprises the individual’s character, which concerns morally relevant aspects of the individual. Fourthly, the ‘self ’ comprises the individual’s temperament, for instance whether she is generally cheerful and extraverted. Lastly, it comprises the individual’s self-­understanding, which refers to her understanding of her life, ideals, projects, and relationships (Haybron 2008a, 184). These aspects of the ‘self ’ influence what makes the individual happy and what the individual’s identity-­related projects are.

The Equality-View  67 Therefore, self-­fulfilment, according to Haybron, consists in happiness and success in identity-­related projects. Happiness, on this account, is a central aspect of welfare, because it is the fulfilment of one’s emotional ‘self ’. Happiness results from living in accordance with one’s disposition to be happy in certain circumstances and not in others. Haybron defends an emotional state theory of happiness and conceives of happiness in terms of a broadly positive emotional reaction to one’s daily life. Happiness thus understood is a state of psychic affirmation or psychic flourishing. According to Haybron, an individual’s emotional flourishing encompasses her emotions, moods, and mood propensities, where the latter are inclinations to be in a certain mood. Happiness is thus a central emotional state and can be contrasted with states of anxiety, alien­ ation and depression. Even though Haybron calls his account of welfare the ‘self-­fulfilment account’ and even though he takes welfare to consist primarily in self-­ fulfilment, as explained above, he also tentatively suggests that there is a third aspect of welfare, besides happiness and success in identity-­related projects. He calls this third aspect of welfare the fulfilment of the individual’s sub-­personal nature. This refers to aspects of the individual’s nature that, according to Haybron, are not related to her personal idiosyncrasies or her ‘self ’. Haybron mentions health, vitality and bodily pleasure as examples. Some things, according to Haybron, benefit us simply because of the sort of creature we are. So, strictly speaking, only emotional flourishing and success in identity-­related projects are aspects of self-­fulfilment, as depicted in Figure 3.1. The focus on self-­fulfilment distinguishes Haybron’s account of welfare from other varieties of eudemonism, which define welfare in terms of nature-­fulfilment. These other forms of eudemonism do not focus on the individual and their idiosyncrasies at all, but rather on what it means to flourish as, for example, a human or as a dog. Haybron’s account of welfare applies to welfare subjects of different species. After all, non-­human welfare subjects as well can live in accordance with their emotional nature (and thus be happy) and they as well can pursue what they care about. As already mentioned in chapter 1.2, there is broad consensus that mammals and birds, at least, experience pleasure and pain (Rollin 1998, Panksepp 2004, Balcombe 2006), and some evidence that these capacities may extend to all vertebrates (Stoskopf 1994, Braithwaite 2010) and perhaps beyond (Sherwin 2001, Elwood 2011). More broadly, animals of different species seem to experience enjoyment and suffering. For example, rough-­and-­tumble play in rats has been shown to raise their

68 

Capacity for Welfare across Species Welfare

success in identity-related projects happiness health, vitality, and pleasure

Figure 3.1  According to Haybron, welfare consists in (1) success in identity-­related projects, (2) happiness, and in (3) health, vitality, and pleasure.

levels of the ‘pleasure hormone’ dopamine (Siviy et al. 1996), and experienced rats will hurry to the hand of a trusted human to be tickled, which induces high levels of ultrasonic chirps associated in other contexts with positive affect (Burgdorf and Panksepp 2001; Panksepp and Burgdorf 2003). Chickens inflicted with joint pain prefer water adulterated with an an­al­ gesic, switching to pure water when their pain subsides (Danbury et al. 2000). And the degree of cage impoverishment in captive mice and rats correlates with their self-­administration of an anxiolytic drug (Sherwin and Olsson 2004), or amphetamines, respectively (Bardo et al. 2001), pre­sum­ ably because these substances provide relief from negative affective states (e.g. anxiety, frustration), as they do for humans. It is a common practice, therefore, to apply hedonistic accounts of welfare to non-­human animals. Many species of non-­human animals also experience the deeper emotions, moods, and mood propensities that Haybron’s emotional state account of happiness refers to. Studies with rats and with starlings have documented pessimistic response biases when the subjects are subjected to days of unstimulating confinement, and more optimistic biases by conspecifics kept in enriched environments (Harding et al. 2004, Bateson and Matheson 2007, Matheson et al. 2008). Persecuted chimpanzees and elephants are vulnerable to psychological breakdown (Bradshaw et al. 2005, Brüne et al. 2006), and baboon mothers whose infants have died show patterns of hormone changes that parallel those of bereaved women, and they seek therapy by expanding their social (grooming) networks (Engh et al. 2006).

The Equality-View  69 Few animal studies have sought to address pleasure, let alone happiness, but there are numerous accounts of animals showing joyous affect in such circumstances as play, mischief, liberation, and some even show humour (see Balcombe 2006). Moreover, capacities for such negative moods as bereavement, depression, and pessimism suggest that positive moods are also present. I will not say much about identity-­related projects here, but I suppose that human as well as non-­human welfare subjects typically do have projects that matter to them. Examples of such projects are finding food, hunting, finding a mating partner, nest building and raising offspring. The third aspect of welfare besides self-­fulfilment, which Haybron more tentatively suggests, is the fulfilment of the individual’s sub-­personal nature. According to Haybron, how well off an individual is may depend not only on the extent to which the individual fulfils her emotional self and succeeds in her identity-­related projects. It may also depend on how the individual fares with regard to some other aspects that do not concern her personality. What Haybron has in mind here are criteria such as health, vitality and physical pleasure. What is healthy for an individual, according to Haybron, does not depend on her personality. It is simply determined by the sort of creature the individual is. His idea is, for instance, that exercise is healthy for humans, simply because of the sort of animal we humans evolved to be. Haybron refers to these aspects of our natures as our ‘nutritive’ or ‘animal natures’, indicating that certain things are good for us simply because of the sort of animal we are. This aspect of Haybron’s account of welfare can straightforwardly be applied to non-­human animals. As we saw in section  3.1, the theories of animal welfare scientists often include the criterion that animals are better off to the extent that they can live according to their species-­specific nature (Webster 1994, Ohl and Van der Staay 2012). For instance, since the pig evolved as a terrestrial forager, living on an earthen substrate contributes to her flourishing, and it is part of the volant bird’s nature-­fulfilment to fly. Indeed, regarding non-­human animals, concern with their welfare has focused nearly exclusively on sub-­personal nature-­fulfilment. There has been very little attention to individual idiosyncrasies and self-­fulfilment, properly speaking. Only recently, and mainly regarding pets, individual idio­syn­cra­sies are being mentioned in relation to animal welfare (Yeates 2013). What are the implications for the generality of Haybron’s account of welfare? Physical pleasure, health and vitality are applicable to all animals, including humans at various stages and levels of development. Physical

70 

Capacity for Welfare across Species

pleasure does not apply to non-­sentient beings. However, health and vitality are applicable to non-­sentient individuals as well. It is conceivable to speak about the health and vitality of a non-­sentient embryo and fetus, the health and vitality of a bug or spider, and the (assumingly poor) health and vitality of a permanently comatose patient. Furthermore, one may even speak about the health and vitality of plants. Thus, simply ignoring the aspects of welfare that are not applicable to any individual and focusing on those that are applicable seems to lead to the conclusion that Haybron’s account of welfare is applicable to trees, bugs, embryos, or comatose patients. To the extent that at least some of the aspects of welfare apply to any individual, this individual is, according to that interpretation, a subject of welfare. I am not sure whether Haybron would happily embrace the implication that even non-­sentient beings can be welfare subjects, which certainly deviates from many prominent accounts of welfare, such as hedonism and preferentialism. Sumner (1996) explains that unless there is something that it is like to be creature X, creature X cannot have welfare. On the other hand, Aristotelian eudemonistic accounts of welfare or objective list accounts of welfare with items such as health and vitality on the list of goods, do not exclude non-­sentient welfare subjects either. Acceptance of non-­sentient welfare subjects may also be in line with common-­sense psychology. After all, many people normally speak about the (poor) welfare of a comatose patient, about a bug being made worse off by losing a leg, or even about the flourishing of a plant. Dawkins (2017) recent focus on health and (even non-­conscious) ‘preferences’ also seems to make room for non-­conscious welfare subjects. Here is a possible revision of Haybron’s account of welfare that avoids the  conclusion that non-­sentient individuals are subjects of welfare. One may not include the fulfilment of the individual’s sub-­personal nature as a ­sep­ar­ate aspect of welfare. Instead, sub-­personal nature-­fulfilment can be included, in as far as possible, as a part of emotional flourishing and the fulfilment of identity-­ related projects. To the extent that sub-­ personal nature-­fulfilment does not bear on these two things, it can be left out of the aspects of welfare. Vitality and health, for instance, may count towards an individual’s welfare only to the extent that they impact the individual’s emotional flourishing or her success in identity-­related projects. In that sense, only beings that can flourish emotionally or have identity-­related projects, and thus only sentient beings, are subjects of welfare. This would deviate from Haybron’s current position, since Haybron currently holds that sub-­ personal nature-­fulfilment directly contributes to an individual’s welfare.

The Equality-View  71 For instance, he considers being healthy good for a person, in and of itself, independently of whether it makes the person happy or helps him achieve his identity-­related projects. It may seem odd to consider health only instrumentally valuable for welfare. However, this view of the value of health is in line with hedonist and preferentialist accounts of welfare, among others. If one’s ill health does not in any way—­neither directly nor indirectly, neither on the short nor on the long run—­negatively affect one’s experienced quality of life, then arguably one’s welfare is not negatively affected. It is very unlikely that ill health does not affect one’s feelings or preference-­satisfaction, though. So, health is certainly an important instrumental good and it may also be a good indicator of welfare, because health typically affects the realization of basic goods, such as happiness and success in identity-­ related projects. While Haybron currently holds that sub-­ personal nature-­ fulfilment counts directly towards welfare, doing away with this assumption may improve his theory in various ways. After all, the distinction between self-­ fulfilment and sub-­personal nature-­fulfilment cuts across happiness and success in identity-­related projects. So, positing sub-­personal nature-­fulfilment as a third aspect of welfare, besides these two, seems not to get things right. Instead, the fulfilment of an individual’s sub-­personal nature is an im­port­ ant precondition of her happiness and of her success in identity-­related projects. Support for this revision of Haybron’s theory can be found in the following passage where Haybron (2008a, 124–125) refers to bears in an analogy for pointing out the limitations of current happiness research: A further limitation of most research [. . .] is the homogeneity of the populations studied. To an Amish farmer or San hunter, or the fisherman on the island mentioned [. . .], the affluent Westerners who mostly get studied may seem to be leading pretty near identical ways of life. If all your subjects live in similar environments, then of course the role of environment in determining happiness is going to seem limited. It is as if one were to run a series of studies on zoo bears and circus bears, find not much difference in well-­being between the groups, and conclude that it doesn’t matter very much what environment you put bears in.

Here, Haybron seems to suggest that certain environments make bears happier and certain environments make humans happier. If this is true,

72 

Capacity for Welfare across Species

happiness partly consists in nature-­fulfilment rather than only in self-­ fulfilment. Thus, happiness and fulfilment of one’s animal nature are not separate aspects of welfare, but the latter partly contributes to the former. Another piece of support can be found in Haybron’s discussion of the benefits of exposure to natural environments, notably trees, for human welfare. Haybron (2011) cites evidence that exposure to nature contributes positively to our emotions, moods and mood propensities. If so, it holds because of the sort of creature we have evolved to be, rather than because of individual idiosyncrasies. Thus, emotional flourishing partly consists in the fulfilment of our sub-­personal natures. Last but not least, Haybron’s motivation for studying happiness was evoked by his experience that the people on an island where he spent much of his childhood were leading happier lives than the mainlanders among which he usually lived. This, as well, suggests that certain environments and lifestyles seem to enhance our happiness and welfare, simply because of the sort of creature we have evolved to be: The fulfilment of our sub-­personal natures contributes significantly to our emotional flourishing rather than being a separate aspect of welfare. Likewise, what we take physical pleasure in expresses both our sub-­ personal natures and our individual idiosyncrasies. Haybron lists physical pleasure under sub-­ personal nature-­ fulfilment. However, what we take physical pleasure in does not only depend on our sub-­personal natures. Even for broccoli—­Haybron’s example of what humans take pleasure in simply because of the sort of creatures we are—­it does not hold that all humans like it. The same is true for other physical ‘pleasures’, such as cold showers, sunbathes, saunas, massages, and various sexual practices. Instead of listing physical pleasures under sub-­ personal nature-­ fulfilment, one might conceive of them as an aspect of emotional flourishing. A simplified picture of welfare as self-­fulfilment might view the ‘selves’ of different individuals as jars of different forms and sizes, as depicted in Figure 3.2. An individual’s welfare level then is the extent to which its individual jar is filled. This picture has the advantage of making clear that what matters is the extent to which a jar is filled rather than the total amount of, say, liquid in a jar. The sizes of the jars are, thus, irrelevant. A disadvantage of this picture is that it doesn’t seem to be able to account for negative welfare, but only for different extends of positive welfare. This defect can be remedied if we adopt a different picture. In a more complex picture different objects of different forms and sizes symbolize the different individuals’ ‘selves’. The colour of the objects varies

The Equality-View  73 different cups = different selves welfare = extent to which it is filled dog

human

Figure 3.2  Different individuals have different selves and welfare consists in the extent of self-­fulfilment. darker color = more welfare

dog

human

Figure 3.3  Different colours indicate the extent to which an individual’s welfare is negative (white), neutral (grey), or positive (black). The shape and size of the selves does not matter.

from white, via grey to black, where shades of white stand for the extent to which the individual’s welfare is negative and shade of black for the extent to which it is positive. Some shade of grey would symbolize a neutral welfare level. One could either show the relevant colour for each aspect of the individual’s self. In that case, the objects might have spots of different shades of black, grey and white. Or else one might not differentiate different aspects of the self and just have the relevant colour and shade summarize the individual’s overall welfare level. In any case, the extent to which the object is completely black (rather than white) would indicate the individual’s degree of self-­fulfilment and thus its welfare level, as depicted in Figure 3.3. What the simple and the more complex picture have in common is that the size and shape or total content or surface of the object does not matter. This makes sure that the account of welfare is in line with EQU rather than DIF. The more complex picture can also be adopted for other accounts of welfare that are in line with EQU, such as preferentialism, nature-­fulfilment accounts, subjective list accounts, and standard versions of hedonism. (A detailed discussion about hedonism’s compatibility with this relativizing picture will be presented in section 3.3.) Thus, we see how the self-­fulfilment account of welfare and other common theories of welfare can be naturally interpreted as being in line with EQU.

74 

Capacity for Welfare across Species

3.3  An Evolutionary Perspective In this section, I will present an evolutionary argument in defence of the equality-­view about cross-­species capacity for welfare. I will argue that welfare subjects of different species have the same hedonic capacity. I assume that an individual’s hedonic capacity, i.e. the possible range of subjective valance of one’s conscious experiences is (at least) an acceptable proxy for capacity for welfare. My argument that welfare subjects of different species have the same hedonic capacity is based on considerations about the nature and function of hedonicity. ‘Hedonicity’ refers to welfare subjects’ capacity to value experiences on a subjective scale ranging from various degrees of pleasant, via neutral, to various degrees of unpleasant. One can call the hedonicity scale of welfare subjects a ‘pleasure scale’, or ‘pleasantness scale’, keeping in mind that it captures neutral and unpleasant experiences as well as pleasant ones. In what follows, I will first make three important points about the nature and function of this pleasure scale. Then I will point out how this supports the view that welfare subjects of different species have an equal capacity for pleasure. The chapter ends with a refutation of some possible criticism of this line of argumentation, followed by a summary. The first thing to understand from this evolutionary perspective is that pleasure or hedonicity is not a sensation or experience, but it is a dimension of sensations. It can be thought of as a scale for rating sensations or experiences in terms of their pleasantness. This understanding differs from the common philosophical understanding of pleasure. Philosophers tend to think of pleasure as a feeling. Internalists and externalists quarrel about what characterizes this feeling. According to internalists, all sensations of pleasure share a common feeling tone. According to externalists, there is no such common feeling tone. Instead, different feelings of pleasure may have nothing in common, except for a certain attitude that we take towards the underlying feeling: we are pleased by it (Crisp and Kringelbach 2017, 212). There is some truth in both philosophical views about pleasure. If hedonicity is a dimension of experiences, if we automatically rate experiences along our subjective pleasure scale, then this has some similarity with taking an attitude towards the experience in question, except that ‘attitude’ suggests something cognitive, while the rating along the hedonicity dimension happens on a pre-­cognitive level. The idea about the common feeling tone gets right that our evaluation of how pleasant an experience is takes place on a pre-­cognitive level. But we do not feel pleasure, strictly speaking. We feel

The Equality-View  75 other things, such as sweetness and automatically rate the experience as more or less pleasant. Sensations can be said to have a particular duration, intensity, quality, and hedonicity. For example, if I taste some ripe fruit, the taste sensation may be there for a minute (duration), it may be mildly (intensity) sweet (quality). The sensation may be pleasant for me at the time (hedonicity). Cabanac (1996) sums up: Sensation is a conscious experience in response to a stimulus. Sensation is quadri-­dimensional, each dimension describes one property of the stimulus.

Concerning the property that the hedonicity dimension captures, Cabanac (1996) explains: All sensations are either unpleasant, indifferent or pleasant. Incidentally, this includes pain, a sensation most often unpleasant, but sometimes indifferent or even, but rarely, pleasant.

Thus, pleasure is not itself a sensation. Instead hedonicity is one of the four dimensions that all sensations have in our conscious perception. The second thing to understand is that pleasure has three characteristics (Cabanac 1996). Let me talk about sensory pleasure first, to illustrate these characteristics. First, pleasure is contingent. This means that a stimulus can arouse pleasure or displeasure, depending on its nature as well as on the internal state and history of the subject. For example, a certain environmental temperature may feel pleasant, unpleasant, or neutral, depending on my body temperature at the time (Cabanac 2000). Second, sensory pleasure describes a stimulus’ usefulness for correcting physiological trouble or deficit and thereby improving short-­term and—­in combination with learning—­also long-­term survival value. For example, the environmental temperature that feels pleasant is always the one that directs my body temperature towards its optimal level. Sensory pleasure thus indicates the usefulness of a stimulus and motivates behaviour accordingly. Furthermore, we often need to choose between different things that we may be motivated to do, since, for example, we cannot work and sleep at the same time. The brain therefore needs a ‘common currency’ to rank the motivations (Cabanac 1992). This common currency allows motivations to converge towards the behavioural common path (McFarland and Sibly

76 

Capacity for Welfare across Species

1975, Cabanac 1992). Cabanac (1992) proposes pleasure as this common currency. He proves this point in a variety of empirical experiments, where subjects’ behavioural adjustments based on pleasure correlate with a variety of objective bodily measures that indicate homeostasis. For example, they adjust the steepness of a walking mill to the temperature of the room, or the other way around based on what feels best and this happens to direct their bodily functions towards homeostasis (Cabanac 1992, 2000). So, pleasure indicates a stimulus´ objective usefulness at the time. Third, sensory pleasure is transient since the usefulness of the stimulus and therefore its pleasantness changes over time. As soon as the physio­ logic­al deficit or trouble has been solved and the subject is in a normal state, a stimulus that felt pleasant is no longer pleasant. For example, when one is full, eating more of the food is no longer pleasant after a certain point. If you are cold, entering a room with a normal temperature feels pleasantly warm, but once you warmed up, the room temperature feels neutral. So, pleasure is inherently dynamic, due to its contingence, transience, and relation to the usefulness of the stimulus. The third thing to understand is that pleasure is bounded. This follows from its contingence, its relation to a stimulus’ usefulness, and its transience. Since the aim of bodily pleasure is to direct us towards homeostasis, this aim provides a clear boundary to how pleasant something can get. Homeostasis is a state of the absence of trouble. If everything is in balance, that is as far as you can get. There cannot be less than no trouble. Homeostasis can be defined as follows: In biology, homeostasis is the state of steady internal, physical, and chem­ ical conditions maintained by living systems. This is the condition of optimal functioning for the organism and includes many variables, such as body temperature, and fluid balance, being kept within certain pre-­set limits (homeostatic range). Other variables include the pH of extracellular fluid, the concentrations of sodium, potassium and calcium ions, as well as that of the blood sugar level, and these need to be regulated despite changes in the environment, diet, or level of activity. Each of these vari­ ables is controlled by one or more regulators or homeostatic mechanisms, which together maintain life. (https://en.wikipedia.org/wiki/Homeostasis)

If we reach homeostasis, we feel a subjective state of comfort. The state of comfort feels rather neutral. It feels good only in comparison to states of

The Equality-View  77 deficit. Comfort can potentially last longer than pleasure, but only provided that the environment and the subject remain in stable conditions. This is usually not the case for a very long time in living welfare subjects, since inner and outer conditions tend to change over time and require adaptation. We talked mainly about bodily pleasure so far, but human beings are not only bodies and do not only have bodily needs. The same holds for non-­ human welfare subjects. According to Cabanac (1996), consciousness in its broader sense may have evolved from sensation and may have retained sensation’s quadri-­dimensional structure. In the same way as sensory pleasure tags the usefulness of a stimulus, joy tags the usefulness of any other conscious experience.

The characteristics of (1) contingency, (2) focus on usefulness, and (3) transience apply not only to the sensory pleasure that different stimuli evoke, but also to the joy that different conscious states bring about. In the same way there are two different elements in sensation: sensory pleasure highly positive but transient, and comfort indifferent but stable, it is possible to recognize two elements in the affectivity of global consciousness: positive and transient joy, and indifferent but stable happiness. Happiness is to joy what comfort is to pleasure.  (Cabanac 1996)

If the subjective state of comfort is related to the objective state of homeo­ sta­sis, what objective state is the subjective state of happiness related to? Cabanac links pleasure and comfort to the level of sensations, while he relates joy and happiness to the broader level of consciousness. Therefore, while homeostasis concerns the state of dynamic balance between an individual’s body and its physical environment, the state that happiness tracks may be a state of dynamic balance between an individual as a conscious being and its environment, including its social environment. We humans and many non-­human welfare subjects are, after all, not merely physical bodies in a physical environment and for us to fare well, bodily homeostasis is not enough. As far as I know, Cabanac doesn’t make it explicit, what the counterpart to homeostasis is on the broader level of consciousness. I suggest that it might be the fulfilment of one’s needs. This gives us the following picture, as presented in Table 3.1: Thus, welfare subjects feel pleasure and joy if they approach states of homeostasis or, more broadly, fulfilment of needs. These latter states are

78 

Capacity for Welfare across Species

Table 3.1  Pleasure is to comfort and homeostasis what joy is to happiness and the fulfilment of needs. Level of (bodily) sensation Subjective feeling when moving towards state of balance Subjective feeling when state of balance is attained State of balance

Broader level of consciousness

Bodily pleasure (intense ‘Joy’, or pleasure in a broader but transient) sense (intense but transient) Comfort (potentially Happiness (potentially more more stable but neutral) stable but neutral) Homeostasis (fulfilment Fulfilment of one’s needs of bodily needs)

characterized by the absence of trouble. Due to their contingence, transience, and their relation to evolutionary usefulness pleasure and joy are bounded. The bigger picture that I take from this is the following. From an evolutionary perspective, so to say, our hedonic capacity (broadly understood now, as referring to bodily pleasure/displeasure and to joy/suffering more generally) is just a means. It provides a rough and ready aid for picking out the actions that tend to facilitate physiological homeostasis, fulfilment of needs, and survival. This holds for the environment in which we evolved, so assumingly for our gatherer-­hunter environment. Hedonicity is a common currency for deciding among competing motivations that is easily available to us. It doesn’t require rationality or higher cognitive involvement. In contrast, from the individual perspective, high hedonicity presents itself as an aim, as the very thing worth achieving and worth maximizing. Hedonicity works well as a means for evolutionary purposes, so to say, precisely because we consider pleasure and joy worth having for their own sakes from our individual perspective. What follows from this picture for the question about cross-­species capacity for welfare? If welfare subjects had such a hedonicity scale with the explained origin, nature, and function, how would it be calibrated? What range of pleasantness would individuals from different species in principle be able to experience? Would it make sense to assume that certain welfare subjects can experience a broader range of welfare, higher heights, or lower lows? Or would it make more sense to assume that the possible range of hedonic experiences is roughly the same for all welfare subjects? In what follows, I will argue that the latter is the case. My argument is based on taking a closer look at the signalling function of pleasure.

The Equality-View  79 My pleasure-­based rating scale exists to convey to me the relative survival value for me of various actions that I could pursue. As Cabanac (1996) puts it: Thus this axis is the seat and the source of permanent prerational computation. (…) The advantage of consciousness is thus primarily in the af­fect­ ive dimension, the dimension that makes living brains different from computers and robots. (…) If (…) the role of consciousness is to provide an infinite flexibility to the decision-­making process by allowing selection of new priorities and to rearrange plans in unexpected situations, then this function is fulfilled by affectivity.

This emphasizes the important function that the hedonicity dimension plays. We seem to have it primarily to assist decision-­making in situations that require flexibility. In these situations, instinctive, automatic stimulus-­ response behaviour is not enough, because it lacks flexibility. If pleasure is an incentive for decision-­makers for choosing the evolutionarily most advantageous actions, one can ask what an effective use of this incentive would be, from an evolutionary perspective. How would evolution, so to say, put this incentive to work, for the intended aim, in an ef­fect­ive and economic way? The economists Rayo and Becker have something to say about this question. Instead of talking about a ‘hedonicity scale’, they use the word ‘happiness function’, but this should not confuse us. The authors view ‘happiness as a decision-­making device that allows the individual to rank alternative courses of action’, so they are, in fact, referring to our hedonicity scale (Rayo and Becker 2007, 303). They say: The theoretical problem of finding the fitness-­ maximizing happiness functions can be conveniently stated as a metaphorical principle-­agent problem. As is customary in the literature, the principal represents the process of natural selection, and the agent represents an individual carrying a set of genes. In the present context, the principal designs the innate happiness function of the agent, with the goal of maximizing the propagation of the agent’s genes. Importantly, the happiness function is only a means to this end: the principal does not directly care about the agent’s happiness level. The agent, on the other hand, is born with the happiness function designed by the principal and, via his actions, seeks only to maximize his level of happiness. In the process, however, he inadvertently serves the principal’s goals.  (Rayo and Becker 2007, 304)

80 

Capacity for Welfare across Species

Rayo and Becker (2007, 302) emphasize that the hedonicity scale, or what they call the ‘happiness function’, ‘measures the individual’s success in relative terms’. This is to say: The optimal function is based on a time-­varying reference point—­or performance benchmark—­that is updated over time in a statistically optimal way in order to match the individual’s potential. [. . .] This updating results in a volatile level of happiness that continuously reverts to its long-­term mean.

This means that the hedonicity function is calibrated with the individual’s potential, since it aims at helping the individual to choose among its options the action that best fulfils its needs. Rayo and Becker take their view on the hedonicity scale to explain and accommodate cases of adaptation, such as the observed fact that people, at least above a certain income level, do not lastingly feel better after a pay rise or after winning a lottery and they may even adapt to becoming severely bodily handicapped. According to Rayo and Becker, such events may change the reference point or performance benchmark since these depend on the individual’s current potential. Basically, Rayo and Becker argue that the hedonicity axis indicates to what extent one realizes one’s actual potential. Rayo and Becker (2007, 303) use economic tools to argue that this calibration of the hedonicity scale with the individual’s current potential is ‘evolutionarily advantageous in the sense of improving the individual’s ability to propagate his genes’. The authors’ argument is based on two key assumptions: We assume that two alternative choices x1 and x2 cannot be ranked by the agent whenever the difference between his expected happiness levels E[V|x1] and E[V|x2] is smaller than some minimum threshold. The second constraint is a bound on the range of values that V can take. (Rayo and Becker 2007, 305)

Physiological constraints about the function of nerve cells as well as laboratory experiments on rats motivate these assumptions. The authors show that, due to these imperfections the agent will be capable of a more accurate ranking across choices when he measures the fitness associated with each choice relative to the fitness of alternative options, as opposed to measuring fitness in absolute terms. (Rayo and Becker 2007, 305, emphasis mine)

The Equality-View  81 Rayo and Becker draw an analogy with a specific feature of human vision, where the eye adjusts to the environment’s luminosity. As in man-­made measurement instruments, such as voltmeters, rescaling of the incoming information in accordance with the instrument’s effective range allows for a more accurate discrimination across stimuli (Rayo and Becker 2007, 305). The authors also draw a parallel with the problem of optimal incentives under moral hazard, where the happiness level (V) that follows on an agent’s choice parallels a performance payment for the agent. In both cases, the principal who designs V seeks to maximize the signal value of this function.  (Rayo and Becker 2007, 306)

Thus, the aim of the hedonicity scale is to make us choose the most advantageous action among our options. The most effective way for the hedonicity scale to work, the most effective calibration, and the maximum signal value of this costly device can be achieved if it is relativized to the individual’s current potential for fulfilling his needs. As Rayo and Becker (2007, 306) put it: In any given period, the agent’s happiness depends exclusively on the difference between his output y and an endogenous reference point ŷ, which serves as a performance benchmark (the difference y − ŷ is the carrier of happiness). This reference point ŷ is positioned according to the current opportunities faced by the agent and is updated over time in tandem with changes in these opportunities.

Rayo and Becker are not discussing comparisons of welfare across species. Yet their conclusion about humans is that the hedonicity axis functions to guide an individual towards the best actions relative to their potential at the time. There is no reason to assume that this is different for non­humans. Indeed, there is both behavioural and neurological evidence of the relative nature of preference. For example, rats who have been trained to run towards a reward run faster if the reward is exchanged for a bigger one than if they receive the bigger reward all along (Zeaman 1949). Monkeys who have a preference among rewards according to which, say, reward A is better than B, which is better than C, have been presented with rewards A and B and in another situation with rewards B and C. The higher reward resulted in more neural firing and the firing was based on relative, rather than absolute attractiveness of the rewards (Tremblay and Schulz 1999). It has been

82 

Capacity for Welfare across Species

shown in studies with both human and non-­human welfare subjects that ‘attraction is relative, not absolute’ (Watanabe 1999). All this suggests that hedonicity, whose function it is to steer preferences, is relative to one’s opportunities rather than absolute. The signalling function of pleasure is maximized if the range of the hedonicity axis is calibrated to the individual’s current potential. This supports the equality-­view about capacity for welfare across species: Evolutionary psychology suggests that different welfare subjects have the same hedonic capacity. In other words, it is not the case that humans have a higher potential for experiencing pleasure or joy than non-­human welfare subjects. One may object and argue instead that welfare subjects of different species differ in hedonic capacity. After all, they have different capacities for smelling, hearing, and seeing. Why should they not also have different hedonic capacities? This objection misses the first point that I made in this paper about the hedonicity scale. Cabanac’s picture makes clear that it is a category mistake to mention hedonic capacity alongside the capacities for smelling, hearing, and seeing. ‘Hedonic capacity’, after all, most plausibly refers to the hedonicity axis, as Cabanac describes it. This is a dimension along which sensations can be valued as more or less pleasant. Hedonicity, to repeat, is not itself a sensation. Their capacities of hearing, seeing, and smelling allow individuals of different species to survive in their ecological niches. It makes sense, for example, for wolves to have a greater capacity for smell than humans have and for humans to form social bonds and to be able to think. But is it plausible that some species have a hedonic capacity that ranges only from ‘mildly pleasant’ to ‘mildly unpleasant’, while others have one that ranges from ‘extremely pleasant’ to ‘extremely unpleasant’? The hedonicity axis is supposed to track usefulness, and usefulness here is a relative notion: useful for the survival of the individual in question and ­useful relative to alternatives. The only case I can imagine of a species with a limited hedonicity range would be a species for whose members it does not matter what they do. But consciousness is costly and would not have evolved in such a species at all. It evolved to allow flexible decision-­making in complex situ­ations in which it matters what the individual does. Some animals may have only relatively dim sensations, as compared to us. Their sensations may not be intense. Perhaps these relatively dim sensations are sufficient for them to survive in their ecological niche, just as our relatively poor sense of smell, as compared to a wolf, is sufficient for us. Yet, intensity and hedonicity of sensations are different dimensions and independent from each other. If the relevant species has a hedonic dimension of

The Equality-View  83 consciousness at all, the fact that the species has only dim sensations does not entail that the range of their hedonic dimension is narrower than ours. A second objection may be that a different number of neurons fire in different animals’ brains when they experience pleasant sensations. Pleasure only does its motivating job if it is subjectively felt. It doesn’t matter what it takes for pleasure to come to the forefront of an individual’s mind and to capture an individual’s attention. Perhaps for some individuals it takes more firing of neurons than for others. This is irrelevant. What matters is only the subjective valance of the experience that the firing causes. Pleasantness consists in this subjective valance, not what it takes to physiologically bring it about. One may object, thirdly, that there may be higher and lower pleasures and some animal species may be capable only of lower pleasures. To be brief about this point, neurological evidence suggests that there is only one pleasure circuit for all kinds of pleasures. Furthermore, it is unclear how hedonists can make logical sense of this Millian idea of higher and lower pleasures (as argued in section 2.3). Lastly, it may be objected that the evolutionary view I presented here is implausible, because it considers happiness to be a neutral state. Many ­people take happiness to be a positive state. I think there is an explanation for this inclination that is compatible with the view I presented. Note that happiness is typically depicted in literature in one of two strikingly different ways. Jules Renard, for example, captures the potentially stable aspect when he writes ‘we are not happy; our happiness is the silence of our unhappiness’. Proust, in contrast, focuses on the transient aspect when he writes about ‘this extension, this possible multiplication of ourselves that is happiness’ (Cabanac 2000). According to Cabanac (2000, 7): The first one is indifference, nirwana-­happiness (or, simply, happiness). The second one is joy-­happiness (or, simply, joy).

As we saw in 3.2, Daniel Haybron (2008a), one of the most influential contemporary philosophical authors writing about the nature of happiness, distinguishes what he calls three ‘faces’, or indicators, of happiness. Two of the indicators of happiness that Haybron mentions are in line with Cabanac’s nirwana-­happiness state. One of these indicators of happiness is the state of feeling at home in one’s body and life, of letting one’s defences down and being at ease. The other indicator is the state of engagement with one’s activities that might culminate in a state of flow. Interestingly, flow has been

84 

Capacity for Welfare across Species

established in happiness research as one of the greatest states to be in and it is a state of self-­forgetting immersion in what one does (Csikszentmihalyi 1990). In that sense it is a neutral state, characterized by freedom from ­trouble and deficit at the time. Joy, according to Haybron, is a further indicator of happiness and one that is more in line with Cabanac’s joy-­happiness. This suggests that people merge joy-­happiness and nirwana-­happiness when they talk about happiness. A second explanation why people are depicting happiness as a positive state is that even nirwana-­happiness is a relatively positive state in comparison to states in which not all our needs are fulfilled. Third, Cabanac’s claim is that the neutral state of happiness can be stable only if internal and external states remain unchanged. But this is not what happens in real life. In real life, nirwana-­happiness can never be realized for long and joy-­happiness is part of the best state we can be in, so it is understandable that people do not tend to separate the two. Living beings are dynamic organisms that can never just stay as they are. They need to constantly adapt to changing inner and outer conditions. We constantly move between hunger and satiation, between a need for activity and a need for rest, and between a need for social interaction and a need for a break from it. There is no such thing as complete and lasting satisfaction. And we are lucky that it is like this. After all, without feeling hungry again, we wouldn’t be able to enjoy another meal, without feeling needs for social interaction, activity, or rest, we wouldn’t be able to experience the joy of fulfilling these needs. Remaining in a constant nirwana would deprive us of our opportunities for joy. Unfulfilled needs are not a bad thing if we are able to fulfil them. Thus, happiness, in our world, is a dynamic state in which we can fulfil our various needs when they occur. I take this to hold for humans as well as for sentient non-­humans. This means that the sharp distinction that Cabanac draws is theoretically enlightening, but in practice things are not that clearly separable. When sheep or cows are being let out on a meadow after having spent the winter months in a dark barn, they feel a lot of pleasure and joy and jump around in a way that we rarely observe in these species. Once they are used to the meadow, they do not keep behaving in this way. The relaxed grazing on the meadow that can be witnessed during the summer is likely to feel subjectively less intense than the first steps out of the barn. Yet, it is certainly pleasant, time and again, to be able to eat when hungry, to feel the sun on one’s back when cold, and to find a place to rest in the shade when it is getting too warm and one’s stomach is full. If they could reflect upon it, the

The Equality-View  85 cows and sheep would assumingly be willing to do without the intense but fleeting joy of seeing the meadow again if that would allow them to experience their summertime life throughout the whole year. ‘Happiness’, they would judge wisely with a smile on their snouts, ‘is not fleeting ecstasy, but the satisfaction that comes from having our needs fulfilled for a while and the joy of being able to fulfil our needs when they arise’.

3.4 Conclusion In section 3.3 I have presented an argument based on the nature and evolutionary function of hedonicity, i.e. welfare subjects’ capacity for sub­ject­ive­ly valuing experiences along the dimension of pleasantness. Pleasure is contingent and transient. Its function is to make us select the most advantageous actions among our options. To have a maximum signalling function, it is relative to a welfare subject’s options rather than absolute. This suggests that welfare subjects of different species have the same hedonic capacity, which I discussed here as a plausible proxy for capacity for welfare. Throughout this chapter I have presented approaches to cross-­species capacity for welfare that are in line with the equal-­capacity-­for-­welfare view. No matter whether one focuses on pleasure, happiness, desire-­satisfaction, self-­fulfilment, or nature-­fulfilment, coping or fulfilment of needs, a happy mouse can in principle be just as well off as a happy human. These theories are inherently relativizing in some way. They focus on the extent to which an individual’s desires, ‘self ’, or nature (whatever they happen to be) are fulfilled, the extent to which an individual copes or fulfils its needs. Or they focus on pleasure or happiness, which, as I argued, are also best understood as inherently relative notions, since they are related to the individual’s homeo­sta­sis or to a broader state of fulfilment of needs of conscious beings. Approaches in animal welfare science as well as philosophical accounts of welfare can be understood as emphasizing different aspects of the evolutionary principal-­agent picture that I presented in section 3.3. For example, the various versions of the five freedoms account focus on freedom from deficit and trouble and the freedom to react to various challenges in order to approach comfort and happiness. Broom defines welfare as an individual’s state regarding its attempts to cope with its environment, where ‘en­vir­ on­ment’ refers to everything outside the individual’s brain. So, this can be understood as a striving for homeostasis and fulfilment of needs. Broom also emphasizes the role of feelings, to which he ascribes a useful function

86 

Capacity for Welfare across Species

in search of homeostasis and fulfilment of needs. Those who emphasize the individual’s evolutionary success, survival and reproduction as aspects or indicators of welfare can be seen as emphasizing the principal’s rather than the agent’s perspective. In contrast, those who focus on feelings and preferences focus on the agent’s perspective. Based on the evolutionary picture presented in section 3.3, what would we, humans, from our perspective, think welfare consisted in? What would we come up with as proposals for the ultimate personal good or goods? I can easily imagine that some would bring forward pleasure or happiness as the ultimate goods. After all, pleasure and happiness are the sensations that we seek, which the ‘principle’ linked to successfully doing what we are ‘supposed’ to do. Pleasure and joy are closely linked to our motivation. I can imagine that others would claim that the ultimate good consisted in getting what we want, fulfilling our needs or our preferences. This is, after all, what we are motivated to do and what feels good for us. I can also imagine that some would rather mention a list of specific things that we tend to want and that tend to bring us pleasure and happiness, such as, perhaps, good health, friendship, or useful knowledge. Still others would, perhaps, take a step back from our perspective as agents and focus more on the principle’s underlying aims, such as enabling our homeostasis, survival, and reproduction. If we could ask dogs or other welfare subjects what welfare consisted in, they would, if they could, naturally provide similar answers. What is more, in the circumstances in which we and other species evolved, these different perspectives on what welfare consists in yield identical judgements about particular cases. Friendship, for example, is what we need and want, what gives us pleasure and contributes to our happiness. Since we evolved as a social species, those without friends were less likely to survive and to spread their genes. It is therefore very natural indeed that we value friendship as a personal good, and perhaps even fail to imagine that it may not be a personal good for different animal species that naturally live solitary lives. No matter whether we focus on what brings us pleasure or on what fulfils our needs and desires: these things would lead us towards the same behavioural common path in natural circumstances. Nowadays and perhaps even more so in futuristic thought experiments, our environments are relevantly different from those in our evolutionary past and this holds not only for humans. Consider, for example, intensive animal husbandry. The animals that we keep in that way haven’t evolved under those conditions. As we saw, Dawkins (1990) is concerned that these animals experience pain, boredom, and frustration. At the same time, due

The Equality-View  87 to our interference, these animals spread their genes more successfully than any comparable free-­living species. In an imagined experience machine, we could have all kinds of pleasant experiences while doing and achieving nothing at all. Thus, in these unnatural circumstances, reproductive success is decoupled from living out species-­specific behaviour, and homeostasis is decoupled from positive feelings. So, focusing on an evolutionary perspective can explain both our agreements and our controversies about welfare. Is there some best answer as to what the correct account of welfare is? Perhaps there isn’t anything over and above that evolutionary story. In any case, the evolutionary perspective on hedonic capacity, pleasure, and happiness supports EQU rather than DIF.

4 Welfare across Time One may be interested in the welfare of an individual at some point in time. For example, I may wonder how well off my dog is right now. In that case, one is interested in synchronic welfare. Alternatively, one may be interested in welfare across some longer timespan. For example, I may wonder how good the life of some chicken was until now, how good my own future life will be, or how good my dead grandfather’s life was for him. In that case one is interested in diachronic welfare. There are different accounts of how one gets from welfare at a time to welfare across time. By far the dominant view is that diachronic welfare consists simply in the sum of synchronic welfare. This is called the simple additive view. According to this view, the goodness of a whole life consists in the sum of the goodness of its moments. This chapter discusses some issues concerning diachronic welfare. Welfare across time plays a role in assessments of the harm of death (section 4.1). It is also relevant when we want to compare the value of whole lives or time-­extended benefits and harms across species (section  4.2). In inter-­species assessments and comparisons of diachronic welfare, the relevance of differences in species-­ specific lifespans needs to be settled (section 4.3).1

4.1  Harm of Death This section introduces the main philosophical views about the harm of death, but what are these views about? When we talk about the harm of death for an individual, we do not consider effects on others. Furthermore, in discussions about the harm of death, we contemplate a life that ends at a particular time, say t1, rather than at some later time, say t2. Thus, we do not compare mortality to immortality, but dying sooner rather than later. We 1  Parts of section 4.1 are reproduced with permission from Palgrave MacMillan from Višak (2013). Parts of section 4.3 are reproduced with permission from Springer Nature from Višak (2018a). Capacity for Welfare across Species. Tatjana Višak, Oxford University Press. © Tatjana Višak 2022. DOI: 10.1093/oso/9780192882202.003.0004

Welfare across Time  89 also do not talk about the process of dying, bus simply about the fact that a life ends at some point in time rather than at a later point in time. The suffering that may be involved in the process of dying does not belong to the harm of death, thus understood, but it affects the individual’s lifetime welfare. Lastly, we assume that the state of being dead is not as such good or bad or neutral for an individual. Rather, the welfare subject has ceased to exist, so there is nobody for whom it can be good, bad, or neutral to be dead. It is important to consider what determines how harmful having a shorter rather than a longer life is for an individual. In this section I will first present the frustration view and the deprivation view and then the time-­relative-­interest account. One influential theory, which we can call the ‘frustration view’ on the harm of death, holds that death harms a being to the extent that it frustrates the fulfilment of her desires. Imagine that I have the desire to go to Paris tomorrow and I die tonight. In that case, the harm of death, according to the frustration view, consists in preventing me from fulfilling my desire to go to Paris and in preventing me from fulfilling all other desires that I had for my future before I died and that remained unfulfilled due to my death. Dying later rather than earlier is not necessarily less harmful for an individual in that view. But it typically is less harmful since people tend to have fewer and weaker desires for the future towards the end of their natural lifespans. Peter Singer accepted this frustration view on the harm of death throughout much of his career (Singer 2011, 76). ‘Frustration’ here is a technical term for an unfulfilled desire. What matters is that a desire remains unfulfilled. It does not matter whether an individual feels frustrated in the sense of being sad and angry about it. In fact, having died precludes the possibility of being sad or angry about anything. Yet, death may prevent individuals from fulfilling their desires, which thus remain unfulfilled. Beings that have a conception of their own existence across time usually have plans and projects for the nearer and further away future, along with the general desire to go on living. In contrast, beings that lack a conception of their own existence across time lack a desire for continued life. They may have immediate and short-­term desires, such as the desire to escape frightening or painful situations, the desire to eat, and the desire to rest. Thus, according to the frustration view, death is typically a greater harm for normal adult humans than for sentient non-­humans, since the former typically wish to continue living and have all kinds of plans. Most ­non-­human animals, in contrast, do not have a conception of their own existence across time. They live in the moment. So, death counts as a

90 

Capacity for Welfare across Species

significantly lesser harm for them in this view. Unborn and newly born babies, as well as some mentally severely impaired humans, are comparable to non-­human animals in this regard. Singer (2011, 120) argued that death is, for this reason, not particularly harmful for these individuals. In this view, death at some point in time is good for an individual just in case continued life would frustrate more of her current desires than death. Most proponents of the frustration view hold that the relevant desires are ­idealized desires rather than actual desires. For example, the relevant desires must be well informed and rational. Based on this account of the harm of death, Singer argued that death is not a great harm for most non-­human animals. He also argued that abortion and infanticide do not harm the baby in question, since it has no future-­directed desires that death frustrates. The question what desires particular individuals have is an empirical one. It is very relevant when it comes to the practical implications of the frustration view of the harm of death. It does seem true, though, that most sentient non-­humans, as well as human babies, and some severely mentally impaired humans, live in the present. If so, death does not harm them according to the frustration view. Acceptance of the frustration view may be (and usually is) based on metaethical subjectivism about normative reasons for action. According to metaethical subjectivism our normative reasons for action (or what is morally right or wrong) cannot be based on any objective value. This is because metaethical subjectivism denies the existence of objective values. Therefore, welfare cannot be an objective value either and the fact that killing diminishes an individual’s welfare cannot as such make it wrong. According to this subjectivist view about value, valuing subjects confer value on things. So, welfare, just like anything else, is valuable if and only if and because some subject values it, perhaps under some idealized conditions. The badness of death, in this view, cannot consist in depriving an individual of what would have been objectively good for him or her. Instead, the badness of death consists in preventing the subject of getting what he or she wants, such as, perhaps, going to Paris tomorrow or simply continuing to live. Thus, subjectivism in metaethics goes together with the frustration view about the harm of death. In fact, subjectivism is incompatible with the alternative view on the harm of death, the deprivation view. This is because the latter presupposes that welfare is good for the subject and the more of it, the better, independently of the subject’s valuing attitudes and desires. Peter Singer, for example, was a metaethical subjectivist throughout most of his

Welfare across Time  91 career. As explained, he also accepted the frustration view on the harm of death. When he became a metaethical objectivist, he abandoned the frustration view and accepted the deprivation view on the harm of death instead (Višak 2014). If value and (thus) normative reasons for action are based on such preferences in one way or the other, then it is relevant what preferences animals have. If animals have a preference not to be in pain but lack a preference for continued existence, then this can explain why killing them as such is not a problem, while causing them suffering is. The so-­called deprivation view on the harm of death, in contrast, does not focus on the frustration of what the individual wants. Instead, it focuses on the loss of what would have been valuable for the individual in question. It conceives of the harm of death as the deprivation of the well-­being that the individual would otherwise have experienced in the future. So, this view does not consider how much a being wants his future, but rather how much value the future would have had for the individual in question (Bradley 2009, 2016). So, for example, if a zebra dies at a certain point in time, the animal is thereby deprived of the good and bad experiences that it would otherwise have had, had it lived longer. Depending on whether its future would have been good, or bad, or neutral for the zebra, the fact that its life ends when it ends is harmful, beneficial, or neutral for it. Of course, we do often not know when an individual would otherwise have died and what would have happened until then. The deprivation view only tells us that, depending on the relevant counterfactual scenario, death at a particular time is neutral, harmful, or beneficial for the individual. In practice we may only be able to make informed guesses about what would have been the case. In this view, death at a particular point in time is harmful for all and only individuals who would otherwise have had a future life with a positive welfare level. This includes individuals that do not have future-­directed desires, such as young babies and certain non-­human animals. Death is good in this view for individuals that would otherwise have had a bad future. Killing such an individual would be beneficial for it and would thus be a case of euthanasia. (Whether and under what conditions euthanasia should be legally allowed is a separate question.) Since the deprivation view compares the value of lives (or of actual and counterfactual lives), it is also called the ‘life comparative account’. So, we see that whether death harms an animal that would otherwise have had a future with a positive welfare level depends on the correct view about the harm of death. According to the deprivation view, death would

92 

Capacity for Welfare across Species

harm the individual in this case. According to the frustration view, death would only harm the individual if it frustrated any of the individual’s desires. The latter view implies that death is not harmful for beings that do not have any future-­directed desires. The fact that even these beings typ­ic­ al­ly have the desire not to be in pain explains why causing them pain is problematic. Thus, if the frustration view on the harm of death is correct, it can explain why kicking an animal may be more problematic than (painlessly) killing it. As explained, this view also implies that death is not harmful for babies and for those other humans who lack future-­directed desires. For this and other reasons I find the frustration view more plausible, but I do not go further into this issue here. The time-­relative-­interest account (TRIA), which is most prominently defended by Jeff McMahan (2002), is typically discussed in the philo­soph­ ic­al literature as a further account of the harm of death. As such it would be a rival to both the deprivation view and the frustration view. However, it is unclear how best to understand TRIA’s nature and implications. Upon closer inspection, it does not answer the same question as the de­priv­ation view and the frustration view and it is, therefore, not a rival to these views. McMahan’s support for TRIA is based on the intuition that a death might sometimes be more tragic than another death, despite depriving the individual in question of fewer happy years of life. Imagine, for example, that a doctor in an emergency room can only save one of two lives: that of a two-­ year old or that of a fifteen-­year-­old. Many people believe that the fifteen-­ year-­ old should be saved. TRIA is supposed to justify that intuition. McMahan argues that the older child has stronger psychological connections with her future self. For example, she might have many plans and desires for the future, in contrast to the younger child, who might not have such future-­directed intentions, or only a few of them. Furthermore, the future self of the two-­year-­old would not remember his life at age two, while the future self of the fifteen-­year-­old would remember her life at age fifteen. And thus, the older child would lose a future that would be hers in a stronger sense. This might explain the intuition that the doctor should save the older rather than the younger child, or so McMahan argues. TRIA is also supposed to align with existing theories in the field of metaphysics. These theories concern the question about the metaphysical basis for concern with one’s personal good. One possible metaphysical basis for concern with one’s personal good is individual identity across time. This view implies that if some future harm occurs to my future self, and thus to

Welfare across Time  93 me, it harms me, and thus it gives me a self-­concerned reason for concern that I do not have about harm that occurs to others. Theories of identity across time answer questions, such as ‘What, if anything, makes it the case that the child that was born to my mother on my birthday and that carries my name and I as I am now, are different “time slices” of the same individual?’ Of, course I am different from that child in various ways. The relevant sense of identity is not qualitative identity, but numerical identity, though. For example, a banana in front of me may be the numerically same banana as the one that was on my desk a week ago, even though it started to ripen and change its colour and smell. Numerical identity can persist despite qualitative change. Major views about the basis of numerical identity across time for individuals, such as you and me, are the body view and the personality view. Roughly, proponents of the former would track the body through time and space to determine whether numerical identity was maintained. Proponents of the personality view, in contrast, hold that certain psy­cho­ logic­ al connections are the basis for individual identity instead. According to some authors, psychological connections provide the metaphysical basis for concern with one’s personal good, no matter whether they are also the metaphysical basis for identity across time. If, for whatever reason, psy­cho­logic­al connections are the metaphysical basis for concern with one’s personal good, the harm of death depends on the strength of the relevant connections. That line of thought about the ­prudentially relevant metaphysical unity relation provides another motivation for McMahan’s support of TRIA. TRIA, after all, aligns well with the thought that certain psy­cho­logic­al connections provide a metaphysical basis for that kind of concern. Hilary Greaves suggests that TRIA can best be seen and should be seen not as an account of the badness of death in an axiological sense, but as an account of the badness of death in an emotional-­reaction sense. She distinguishes the two senses as follows: Badness in the axiological sense: A is more bad than B in the axiological sense iff a world that contains A as component is worse[-overall] than one that contains B as component but that is equal in all other relevant respects.  (Greaves 2019, 195) Badness in the emotional-­reaction sense: A is more bad than B in the emotional-­reaction sense iff the negative emotional reaction that it is fitting to have to A is stronger than the negative emotional reaction that it is fitting to have to B.  (Greaves 2019, 196)

94 

Capacity for Welfare across Species

Greaves provides three lines of justification for her proposal to understand TRIA as a theory about the badness of death in the emotional-­reaction sense and not in the axiological sense. First, she quotes McMahan in support of that suggestion. McMahan writes: [N]ote that my claim is not that it is illegitimate to evaluate a death in the  way required by the Life Comparative Account [i.e., the deprivation view - TV]. On the contrary, the Life Comparative Account requires us to notice the difference that a particular death makes to the amount of value that a life, and therefore the world, contains, and that may be an important consideration. My claim is only that it is not the basis of our intuitive comparative evaluations of different deaths.  (McMahan 2002, 106)

Second, McMahan often describes TRIA as picking out the ‘more tragic’ death, or the death that is ‘a greater misfortune’. But how tragic a death is may depend on other things than the amount of value that it takes from the individual or the world. Third, Greaves points out that TRIA, specified as a theory of axiological badness, is implausible. In support of her claim that TRIA is implausible as a theory of axiological badness, Greaves mentions various kinds of objections that also other authors have brought forward against TRIA. Here, for example, is a case in which TRIA, understood as an axiological theory, would have counterintuitive implications: Delayed Choice: A baby (A) and a 30-­year-­old (B) are each such that, unless they are treated now, they will die in 30 years’ time. The doctor only has the resources to treat one; whoever is treated will live until age 80.

In this case, based on the individual’s current time-­relative interests, TRIA would now require treating B, since A now is only weakly connected with her later self. But in thirty years’ time, the situation will be different. A will by then be thirty and would, if not treated as a baby, lose fifty years of life, while B would be sixty and lose only twenty years. Considering the interests of A and B at that time, TRIA will imply that it would be better if A survived, and thus that it would have been better to treat A. Therefore, TRIA has been criticized as being incoherent (Broome 2004, 249–251, Broome 2019). Perhaps one should not take the individuals’ current interests into account, but rather the interests they will have, depending on how one acts,

Welfare across Time  95 or the interests they will have, independently of how one acts, or some such thing. But each of these interpretations of TRIA comes with its own problems. Importantly, according to Greaves (2019), if TRIA is not to be understood as an account of badness in the axiological sense, ‘then it has no direct relevance to the question of what we ought to do’, at least for welfarists. This is because she takes it to be ‘non-­sensical’ to base our normative reasons for action on folk intuitions about degree of tragedy, since these often do not track the value loss that is at stake. I side with Greeves on this issue. We may find it the more tragic the closer the shot of our favourite soccer team misses the goal. But for the result of the game, it doesn’t make any difference how closely the goal was missed. Similarly, we may find some circumstances of death more dramatic than others. This may well matter for what concerns the adequacy of our emotional reaction, but it doesn’t bear on the size of the harm or benefit that is at stake. Our intuition that it is better to save the fifteen-­year-­old rather than the baby in the original emergency-­room case may be explained without appealing to TRIA. For example, the older child may suffer more, since it has a better grasp of its terrible situation. In addition, friends and family of the older child may suffer more, since they may have known the individual for longer and may have invested more in the individual’s welfare or their relationship. They may have a clearer sense of what the individual may other­wise have achieved and experienced and for these and similar reasons they might experience a greater sense of loss. These differences between the cases, if they exist, seem to me morally relevant even though they do not concern the harm of death as such. They matter for what we ought to do. But they should not guide our choice between rival theories of the harm of death. Just like the frustration view on the harm of death, TRIA may imply that the death of a non-­human animal is less harmful than the death of a normal adult human for the individual in question. This is because non-­humans may have a weaker psychological connection across time and thus the future that death deprives them of would be their future only in a weaker sense. This is controversial, though. Even if true, TRIA or the frustration view would not justify a general discount of the welfare of individuals with weaker cognitive or emotional capacities. They would only justify discounting the harm of death for individuals who are less psychologically connected with their future selves. Anyway, it seems to me that the deprivation view is a more plausible view on the badness of death in the axiological sense.

96 

Capacity for Welfare across Species

I think it is often worse if a human being dies prematurely than if, for example, a pig does. This is due to a variety of contingent considerations. For example, the human may have more awareness of life and death and therefore suffer more upon realizing that his life is about to end prematurely. Furthermore, humans have longer lifespans than pigs and therefore premature death is likely to deprive them of more well-­being. (I will discuss the relevance of different lifespans in section 4.2.) Finally, humans are likely to have a larger network of friends and family that may suffer from the death as well. They may, furthermore, suffer about the loss of a loved one in more profound and enduring ways, simply because of differences in imagination and cognitive capacities. Thus, even if cognitive and emotional capacities may not influence harm of death per se, there are various facts that are likely to make it the case that the premature death of a human does more harm than that of a pig. But this need not always be the case. Things become even more complicated if one considers indirect considerations, such as the fact that premature death is likely to deprive human beings of countless opportunities for affecting others in a positive or negative way, be it on purpose or inadvertently. For example, humans may use resources, pollute the environment, and consume animal products. The harm of death itself for the being that dies is, therefore, not the only thing to consider if we are to determine how much benefit or harm a premature death brings about. And whether our normative reasons for action depend only on (that kind of) welfare considerations is a further question. So, there is room for the position that the premature death of a cognitively more advanced individual often causes more harm than the premature death of a cognitively less advanced one. Whether this is true depends on the details of the case.

4.2  The Relevance of Different Lifespans According to the simple additive view, lifetime welfare is the sum of all moments of temporal welfare. Thus, for example, my dog’s lifetime welfare is the sum of his welfare scores at any point in time throughout his life. It follows that if my dog will have a long, happy life with fourteen happy years, his lifetime welfare will be higher than if he lives only ten happy years and then dies. This implication seems plausible, since we usually try to save lives, also for the sake of the individuals who then hopefully can experience additional happy days and years. We think that it is often better for an

Welfare across Time  97 individual not to die prematurely, since it will benefit from additional welfare if it lives longer. So, we seem to focus on the total duration of a life in comparisons of lives that are equally good at each point in time. We seem to think: the longer, the better (assuming equal and positive synchronic welfare). In assessments of lifetime welfare, how should we facture in different natural lifespans of individuals from different species? In some situations, differences in lifespans seem to be irrelevant. This holds in cases in which welfare subjects of different species could gain or lose an equal number of life-­years. Does a dog gain as much from one additional year of life as a human (assuming equal and positive synchronic welfare)? I think the answer is ‘yes’. Similarly, a dog loses as much from dying one year earlier than he otherwise would die as a human would lose in the same situation. In other cases, differences in natural lifespan are relevant, because they determine the relevant counterfactual. Imagine that a twelve-­year-­old dog and a human of the same age die in an accident. Here the human very likely loses more than the dog. This is because the human would otherwise have had many more additional happy years of life than the dog, due to their different natural lifespans. Similarly, saving the human in this case would bestow a larger benefit. (Indirect benefits and harms are left out of consideration here, but in a complete evaluation of what to do, they are likely to be relevant as well.) But if the situation was thus that saving the dog would gain him more years of life than saving the human would gain the human, perhaps because the human was much older at the time of the accident, being saved would be a greater benefit for the dog than for the human. So, here natural lifespan is relevant for what a reasonable counterfactual would be. Does this way of assessing diachronic welfare across species yield a counterintuitive implication for egalitarians? This view implies that dogs have worse lives than humans, due to the dog’s shorter lifespan. This implies that an egalitarian who wants to equalize lifetime welfare should bestow a specific benefit on a dog rather than on the human. After all, the dog is worse off, due to its shorter lifespan. I find this implication of egalitarianism counterintuitive. I would not find it counterintuitive to make the same choice between two humans, or two dogs, one of which falls short of his natural lifespan. But I do find it counterintuitive for the egalitarian to pity that dog more than the human, or to provide a benefit to the dog rather than the human, merely because of the dog’s shorter natural lifespan. Thus, I seem to share McMahan’s intuition that the dog with high synchronic welfare who

98 

Capacity for Welfare across Species

lives up to his natural lifespan is not unfortunate. This intuition that the dog is not unfortunate, upon closer inspection, is compatible with the view that the dog’s twelve-­year-­long life contains less diachronic welfare than the eighty-­year-­long life of the human. There is a sense in which the dog is worse off than the human, due to its shorter lifespan. This axiological sense of being better or worse off arguably counts in ethical decision-­making. If so, it is an interesting implication of egalitarianism that individuals with shorter natural lifespans are, for this reason, worse off in a way. It is, after all, better to have a longer happy life than to have a shorter happy life. So, the total-­duration view about diachronic welfare may not be counterintuitive after all. To avoid the implication that welfare subjects with shorter natural life­ spans are worse off in terms of lifetime welfare than welfare subjects with longer natural lifespans, one would need to relativize the actual duration of an individual’s life to its species-­specific lifespan. On that basis, a dog who lived to the end of his natural lifespan wouldn’t be worse off than a human who did the same, despite differences in the length of natural lifespans. This may initially sound plausible, but perhaps this is only the case because we confuse the emotional reaction sense with the axiological sense (see section 4.1). Furthermore, relativizing welfare gains and losses to an individual’s natural lifespan implies that a dog who loses or gains one year of life loses or gains more than a human who loses or gains one year of life. On some normative views, this would imply that we should save the dog rather than the human, at least if we leave indirect effects on the welfare of others out of consideration. It seems counterintuitive to me that a dog gains more than a human from an additional year of life (given equal synchronic welfare). So, I do not find this relativize-­to-­natural-­lifespan view about cross-­species comparisons of diachronic welfare plausible. Both the relativize-­to-­natural-­lifespan view and the total-­duration view are in principle available to proponents of both DIF and EQU. Proponents of DIF are likely to find the relativize-­to-­natural-­lifespan view unattractive. Furthermore, it would be very complicated in combination with DIF, because, on the one hand, a lost year for a mouse would count for more, since it was a greater chunk out of the mouse’s life, while, on the other hand, it would count for less, since it was (typically) much less valuable at a time, due to the mouse’s smaller potential for synchronic welfare. Proponents of DIF are probably more attracted to the total-­duration view, and they would adjust the duration of an individual’s life based on species-­specific potential

Welfare across Time  99 for synchronic welfare. Proponents of EQU would probably feel more attracted to the relativize-­to-­natural-­lifespan view than proponents of DIF, since EQU itself can be seen as a relativizing view. Nevertheless, I think that even in conjunction with EQU, the total-­duration view has more going for it. I will further defend and apply the total-­duration view in section 4.3.

4.3  An Application of the Total-­Duration View In this section, I apply the total-­duration view about how to take different lifespans into account in comparisons of welfare. (I introduced the total-­ duration view in section  4.2.) The section also engages with McMahan’s work on capacity for welfare across species and his notion of fortune, which I introduced in section 2.4. In his paper ‘Eating Animals the Nice Way’ Jeff McMahan (2008) explores whether there are ways of routinely using non-­human animals for human consumption that are morally acceptable. McMahan ends his paper with the following consideration: The only form of benign carnivorism that is possible now—­raising animals humanely and killing them painlessly—­ seems morally unjustifiable because the interest the animals would have in not being killed would decisively outweigh the interest people would have in killing and eating them. It does not, however, seem morally objectionable to eat an animal that has died of natural causes, which suggests that it could be permissible to use techniques of genetic modification, when they become available, to create animals that would die naturally on a predictable schedule and in good health. It is hard to see what could be wrong with this practice . . . . (McMahan, 2008, 75)

So, what could be wrong about creating individuals with a shorter natural lifespan? I will refer to the practice of benign animal agriculture that McMahan considers to be available now as the ‘mundane practice’ and I will refer to the practice that may become available in the future as the ‘en­gin­ eer­ing practice’. McMahan says about the engineering practice that it is hard to see what would be wrong with it. He means this literally: while he would intuitively judge that the engineering practice is no improvement as

100 

Capacity for Welfare across Species

compared to the mundane practice, he finds it very difficult to bring up a plaus­ible justification for this intuitive judgement.2 Indeed, McMahan does not offer any such justification. So, he ends his quest for acceptable ways of routinely killing and using animals for human consumption with his sketch of the engineering practice and it remains an open question whether this practice would be an improvement as compared to the mundane practice. Since the engineering practice may become available in the future, we better be prepared and have a well-­justified moral verdict ready. McMahan is a philosopher who refuses to commit himself to any moral theory and who rather addresses problems on the basis of what he takes to be common-­ sense intuitions. If it is true, as I think it is, that many are inclined to share McMahan’s intuitions, finding out where they lead us is not only of exe­get­ ic­al relevance but also of broader interest. We need to find out whether and, if so, how we can reject both the mundane practice and the engineering practice as being incompatible with our duties toward or regarding the involved animals. What does it take to dismiss the engineering practice for those who share McMahan’s intuitive assumptions? How can I reject both the engineering and the mundane practice based on animal welfare considerations or our duties toward or regarding the involved animals? If we are to evaluate these practices, we need to compare different possible scenarios or different possible worlds. We need to compare the possible world in which we have the mundane practice with the different possible world in which the en­gin­eer­ ing practice is realized. These possible worlds, which we assume to differ only with regard to the different practices, can be considered to be different outcomes of our available actions: we could either realize the mundane practice or the engineering practice. We are ignoring further options in order to evaluate just these two and in order to find out which one, if any, would be better. How to explore which of the two practices, if any, is better? We will, first, need to take a closer look at the two practices of so-­called benign animal husbandry. Then we will have to turn to population ethics, which deals with the question of how to act if depending on what one does different individuals exist. This is necessary, because it holds for each of the practices that different animals will exist within this practice, depending on what one does. As we will find out, none of the standard theories in population ethics supports McMahan’s intuition that both practices are equally wrong. 2  McMahan, personal communication, November 2016.

Welfare across Time  101 Argumentative support of McMahan’s intuition that the engineering practice would be no improvement as compared to the mundane practice not only requires a yet-­to-­be-­found plausible theory in population ethics. It also depends on a contentious further assumption about welfare or about something closely related to welfare that gives us reasons for action. One needs to assume that agents have reason to bring longer-­lived rather than shorter-­ lived animals into existence. This, as I will point out, runs counter to a common-­sense intuition about cases in which different individuals come into existence, depending on what one does. In the relevant cases, which are called ‘non-­identity cases’, it seems counterintuitive that we have reason to bring longer- rather than shorter-­lived individuals into existence. Elsewhere McMahan captured this or a similar common-­sense intuition when he argued that individuals are not unfortunate if they fulfil their native potential for welfare (see section 2.4). An individual’s native potential for welfare is the potential that the individual has at birth. I will argue that we should not trust our intuitions concerning non-­identity cases that involve differences in welfare that are due to differences in species- or breed-­related life expectancy. Independently of whether both fulfil their native potential for welfare, animals with a longer life and thus, all else equal, with a higher absolute level of welfare are better off. This insight, in conjunction with a yet-­to-­be-­specified plausible theory in population ethics, will allow me to justify the common-­sense judgement that both practices of ‘benign animal husbandry’ are equally wrong. As I understand McMahan’s proposal, the animals in the engineering practice will live just as long and will be just as well off at every moment of their lives as the animals in the mundane practice. In current practices of animal husbandry, pigs are usually slaughtered at six months, while their natural lifespan is six to ten years. Chickens are slaughtered at six weeks, while their natural lifespan is one to eight years, depending on the breed. Cattle are killed for beef at eighteen months; dairy cows are slaughtered at four to five years. Their natural life span is eighteen to twenty-­five years or more. Sheep are slaughtered at six to eight weeks or up to one year and their natural lifespan is twelve to fourteen years.3 This, I take it, would be roughly the same in the mundane practice of animal husbandry that McMahan

3  See http://www.four-­paws.us/campaigns/farm-­animals-­/farm-­animal-­life-­expectancy/ (accessed 21.11.2017). Note that I speak about ‘(natural) life span’ and ‘life expectancy’ inter­change­ably here, since the average life span of members of a particular species or breed of animal in question equals their life expectancy from birth.

102 

Capacity for Welfare across Species

takes to be possible now. In this practice of so-­called benign animal ­husbandry, the animals would have better lives than in common practices. But they would not live significantly longer. In the engineering practice ­animal breeds like the current breeds of pigs, chickens, cattle, and sheep would exist. They would have a natural life expectancy that is just as short as the lives of the animals in the mundane practice until slaughter. The only difference between the mundane and the engineering practice would be that in the latter, instead of being slaughtered, the animals dropped down dead all by themselves. In a way, it is not hard to see what would be wrong with the engineering practice. Most standard arguments against animal farming would, after all, still apply: Animal farming would still be a major cause of climate change, pollution and deforestation, as well as a waste of resources, to mention just a few problems. McMahan (2008, 67, 76) knows these arguments against animal husbandry would still apply and would probably be decisive. For the sake of argument, however, he ignores these effects and focuses on animal welfare considerations or on our duties toward or regarding only the directly involved animals themselves. From this restricted perspective, he claims that it is hard to see what would be wrong with the engineering practice. After all, in the engineering practice the animals’ welfare is not reduced by an act of killing. If there were, therefore, nothing wrong with it from the animal-­welfare perspective, the engineering practice, despite its remaining problems, would at least be an improvement as compared to the mundane practice. It would be better in one respect, to wit in terms of animal welfare or, more broadly, in terms of our duties toward or regarding the animals in question. The common-­sense intuition seems to be that both practices are roughly equally wrong with regard to our duties toward or regarding the involved animals. However, as McMahan points out, it is hard to see how this verdict can be justified. In what follows, I will first point out why it is so hard to justify the common-­sense intuition and then I will go some way toward pointing out how it might be done. The three most common theories in population axiology are (1) the impersonal total view, (2) the person-­affecting view, and (3) theory Q. These are theories about the goodness of outcomes rather than about what one has reason to do. I will, however, discuss the respective deontic theories, since I am interested in the ethical question of which of the two practices of animal husbandry, we have reason to bring about. I talk loosely and inter­change­ ably about actions that we have reason to (not) bring about and actions that

Welfare across Time  103 are right (wrong). As we will see, the first two theories of population ethics are incompatible with McMahan’s intuitions. The third is unsuited for addressing the case at hand but may be improved to do so. One of the most common theories in population axiology is the impersonal total view (Parfit 1984, 387). This theory evaluates an outcome simply based on the amount of welfare that it contains. A related deontic view would state that we have reason to bring about the outcome that contains more welfare overall. Proponents of this view would not find anything wrong with the mundane practice. True, the practice involves killing animals that could otherwise have continued to live lives with a positive welfare level. A longer life with a positive welfare level contains more welfare than a shorter such life. Therefore, the outcome in which the animal is killed contains less welfare, all else being equal. This seems to suggest that it is wrong to kill the animals in the mundane practice. But all else is not equal. Other animals whose lives contain as much welfare as the future lives of the killed animals would have contained replace the killed animals. These animals would otherwise not have lived. Provided that the killing does not have any uncompensated negative side effects, this replacement compensates for the welfare that would otherwise be lost in the outcome in which the animals are killed.4 Given this replacement, killing the animals, as it happens in the mundane practice, is permissible, according to the (deontic version of the) impersonal total view. In the engineering practice, the same thing happens, except that the animals are not killed but they die all by themselves at an equally young age, due to the genetic modification. Again, other animals replace the dead ones so that the overall amount of welfare is the same as it would be in an outcome in which animals with a longer nat­ ural lifespan were to live. Thus, from the perspective of the total amount of welfare that an outcome contains, neither practice involves doing anything wrong, since—­given replacement—­neither killing nor reducing the animals’ lifespans diminishes welfare. This shows that, according to the impersonal total view, the engineering practice is no improvement. So, the impersonal total view supports this part of McMahan’s intuitions. However, McMahan cannot avail himself of the impersonal total view. After all, it implies that, given replacement, there is nothing wrong with the mundane practice. McMahan, in contrast, judges the mundane practice to be wrong, because it involves killing and thereby 4  For a discussion of the replacement argument, see Singer and Lazari-­Radek (2014, 374), Singer (2011, 94–123) and Višak (2013).

104 

Capacity for Welfare across Species

harming animals. Even if the humans’ gustatory pleasure could in principle outweigh this welfare loss to the animal, the animal in fact loses much more welfare than the consumers of its flesh and its other products gain. McMahan does not bring the welfare of the animal that replaces the killed one into the equation. This is, he explains, because an unborn animal does not have any interest in being born, while an existing animal has an interest in not being killed (McMahan 2008, 71). Furthermore, he does not assume that there is any impersonal reason to cause animals to exist, simply because their lives would be good (McMahan 2008, 72). All this implies that McMahan does not accept the impersonal total view, and those who share his intuition that the mundane practice would be wrong cannot do so either. A second common theory in population axiology is called the ‘person-­ affecting view’. It says that an outcome is worse if and only if and to the extent that particular individuals are worse off overall in this outcome. An outcome is better if and only if and to the extent that particular individuals are better off overall in this outcome (Parfit 1984, 370). In its deontic version this theory implies that the mundane practice is wrong. This is because killing an animal makes this particular animal worse off than he or she would have been in the alternative outcome in which he or she would not be killed. The replacement does not improve the situation, since the newly created animal is not better off than he or she would otherwise have been. He or she would, after all, otherwise not have existed and one can only compare an individual’s welfare level in two possible worlds if he or she has a welfare level in each of these worlds. This is not the case if in one of these worlds the individual does not exist at all. So, it follows from the person-­affecting view that the mundane practice is wrong: It is wrong because it involves some harm (the killing) and no benefits that could compensate this harm. Note that the person-­affecting view would not imply that the mundane practice was wrong if the newly created animal benefited from being brought into existence. After all, in that case there would be a benefit that compensated the harm. But McMahan argues, correctly as I think, that existing as opposed to never existing cannot benefit an individual. Furthermore, he takes it that the animals’ interests ground our duties toward them, and he believes that unborn individuals do not have an interest in being born. So, he does not avail himself of the view that we have a reason to bring well-­off animals into existence.5 5  For further arguments against the view that existence can be better or worse for an individual than never existing, see Višak (2016).

Welfare across Time  105 One might claim that while the killing harms the animal in question, the mundane practice as such does not. After all, by killing the animals, one only deprives them of something that one gave them in the first place. McMahan (2008, 71) argues: That justification would allow parents to kill their children. Whatever good the practice has bestowed on animals up to this point cannot be cited as credit from which the killing can now be debited.

Even though the animal would not otherwise have existed, killing the animals harms them and therefore the practice involves doing something wrong. The mundane practice, according to the person-­affecting view, is wrong; but what about the engineering practice? While the mundane practice involves harming particular animals, this cannot be said of the engineering practice. Even if having a shorter natural life expectancy is worse than having a longer one, the engineered animals have pleasant lives and would not be better off if different, longer-­lived animals existed instead. So, no particular animals are harmed in the engineering practice. Thus, according to the (deontic version of the) person-­affecting view, the engineering practice is acceptable, while the mundane practice is wrong. This shows that McMahan and those who share his relevant intuitions cannot avail themselves of the person-­affecting view in support of their intuitive position. They intuitively believe, after all, that the engineering practice is wrong. Can McMahan appeal to the third of the common theories in population ethics, the same-­number-­quality claim, also known as theory Q? According to (a deontic version of) theory Q, we ought to bring about outcome A rather than outcome B if and only if the individuals who exist in A are better off than the individuals who exist in B. These need not be the same individuals (Parfit 1984, 369). In both practices, an animal is worse off due to having a shorter life and according to theory Q it does not matter whether the same animal could have lived longer or whether a different, longer-­lived animal could have existed. This seems to allow McMahan to say that both practices involve doing something wrong and that the engineering practice is no improvement as compared to the mundane practice. But such a verdict would not take replacement into account. If we assume replacement, we need to compare in both practices an outcome in which, say, ten short-­lived animals exist with an outcome in which one longer-­lived animal exists. These are different-­number cases. Theory Q, as it stands, does not apply to different

106 

Capacity for Welfare across Species

number cases. As Parfit (1984, 370) explained, a complete, plausible theory in population ethics may imply theory Q, but it needs to go beyond theory Q to be applicable to different number cases. The complete and plausible theory in population ethics, which Parfit labelled theory X, is not available yet and theory Q is not applicable to different number cases. So, given replacement, McMahan cannot avail himself of theory Q. Which theory in population ethics could those who share McMahan’s intuitions appeal to in order to dismiss both the mundane practice and the engineering practice as (equally) wrong? We saw that the impersonal total view is incompatible with the rejection of the mundane practice, while the person-­affecting view is incompatible with the rejection of the engineering practice. Theory Q, in turn, is completely inapplicable to the cases at hand, since they are different number cases. This confirms that for those who share McMahan’s intuitions about the mundane practice it would indeed be hard to justify the verdict that the engineering practice is no improvement. While a more encompassing and plausible theory in population ethics is yet to be found, we can already establish several desiderata that this theory should meet. A theory that meets these desiderata will give us the desired verdict in the cases at hand. We could say, firstly, that theory X should imply theory Q in same number choices. We could secondly say that the desired theory in population ethics should not simply judge outcomes based on the total quantity of welfare that they contain. Thus, it shouldn’t accept replacement. A theory that satisfied these desiderata would imply that the mundane practice is wrong, because killing an animal harms this animal. The engineering practice, it would imply, is also wrong, provided that the short-­ lived animal that exists is worse off than the different longer-­lived animal that would otherwise have existed. That there would be a greater number of these short-­lived animals, due to replacement, would not be considered an acceptable compensation. Granted, it would be difficult to find a view that combines these desiderata. On the one hand, this theory would not be wholly impersonal since it wouldn’t accept replacement. On the other hand, it would not be wholly person-­affecting either since it would consider it irrelevant that the short-­ lived animal in the engineering practice couldn’t possibly live longer. Perhaps some such theory could be spelled out and defended. I don’t claim that it is impossible, but it is certainly not easy to see how it could be done. Baderism, which I introduce in section  5.2, is a promising theory along

Welfare across Time  107 these lines.6 What I said about population ethics confirms that it is indeed hard to support the common-­sense verdict and it also indicates some desiderata for the yet-­to-­be-­found plausible theory in population ethics that would be required in order to support the judgement that both practices are (equally) wrong. The desired theory in population ethics would dismiss the engineering practice as being no improvement as compared to the mundane practice only if we made one further assumption. The necessary further assumption is that our reason to bring into existence the longer-­lived rather than the shorter-­lived animal is just as strong as our reason not to shorten the life of an existing animal. But this assumption can be challenged, as I will show in the remainder of this section. The following story may cast doubt on the assumption that bringing about the shorter-­lived animals in the engineering practice is wrong. Let us assume that you have the choice between bringing into existence a small dog or a large dog, each of which would be perfectly happy and healthy. Small dogs tend to live a couple of years longer than large dogs. Let us say that a small dog would live thirteen years whereas a large dog would live ten years. You do not mind which of them is brought into existence since you consider them both very adorable. On the one hand, if your dog lived longer, you would be able to enjoy his company for a couple of more years. On the other hand, if your dog’s natural lifespan were shorter, that would allow you to travel more after he died, which you would enjoy as well. As you know, neither your own welfare nor that of any third individual would be affected by your choice between the dogs. There are no dogs waiting in shelters that you should adopt instead. You will not have any further pets after this one. In order to decide between the potential dogs, you throw a coin. You welcome your new family member: a large dog. Did you make a mistake? Should you have chosen the small dog? I think that many, perhaps most, people have the intuition that your choice was acceptable and that the longer life expectancy of smaller dogs does not provide any normative reason at all for bringing into existence small dogs rather than large dogs. Many or most people, I think, would also say that there would have been nothing wrong with taking the large dog, even if the 6  There are some further interesting proposals, each of them with its own problems. See, for example, Meacham (2012) and Johann Frick ‘Conditional Reasons and the Procreation Asymmetry’, unpublished manuscript.

108 

Capacity for Welfare across Species

alternative—­just as attractive for you—­would have been breeding a cat. Cats, to be sure, tend to live even longer than small dogs: their natural life expectancy is about fifteen years. Nevertheless, it seems to be a common-­ sense intuition that the differences in life expectancy between different breeds or species make no relevant difference in non-­identity cases. As long as all else is equal, and thus the animals’ synchronic welfare levels during their lives are equally high and nobody else is affected, common sense would say that it does not matter which animal to bring into existence. In a different paper, McMahan (1996) captured this or a very similar intuition, when he said that a happy dog, even though she may—­in some absolute sense—­have fewer goods in her life than a happy human, is not unfortunate. McMahan uses the notion of welfare to indicate the absolute amount of goods in an individual’s life. He uses the notion of fortune, in contrast, to indicate welfare relativized to some standard. The standard that he proposes is the native potential of the individual in question. This refers to the potential for welfare that the individual in question has at birth. Since a happy dog fulfils his native potential for welfare, he is, according to McMahan, not unfortunate (as explained in section 2.4). In some cases, McMahan considers fortune, rather than welfare, relevant for what we have reason to do. For example, he appeals to the notion of fortune to argue that egalitarians or prioritarians do not need to be particularly concerned about happy dogs. Even though they are much worse off than happy humans absolutely speaking, they are not unfortunate (McMahan 1996). In the example in which you must choose between bringing a large dog or a small dog into existence, both dogs would be equally fortunate since they would both fulfil their native potential for welfare. If fortune rather than welfare is supposed to ground reasons for action in this case, we have no reason to bring one rather than the other into existence. Similarly, in the engineering practice, we would have no reason to bring the longer-­ lived rather than the shorter-­lived animals into existence. So, we would not do anything wrong if we did the latter. This would mark a relevant difference between the engineering practice and the mundane practice, because in the latter practice prematurely killing an animal would be wrong. McMahan’s response to this line of argumentation may be that welfare rather than fortune should ground our reasons for action in the choice between the engineering practice and the mundane practice. Even though the engineered animals would fulfil their native potential and therefore be fortunate, they would realize fewer goods in their lives absolutely speaking. Therefore, McMahan may insist, we should bring the longer-­lived animals

Welfare across Time  109 into existence and, thus, the engineering practice does involve doing something wrong. After all, when we have the choice between rescuing a well-­off dog and rescuing a well-­off human, McMahan (2008, 66) would also say that we should save the human, because he has more to lose by death. So, in choices about saving lives, welfare rather than fortune matters, according to McMahan.7 The same may be true in choices about creating life. But this is not obvious, given McMahan’s insistence that fortune matters in choices about improving lives. A rationale that explained which of these notions is the basis for our reasons for action in what kind of case would be desirable. This shows that the yet-­to-­be-­found theory in population ethics can only justify McMahan’s intuition that both the mundane and the engineering practice are wrong under the additional assumption that we have reason to bring individuals with a longer rather than a shorter natural life expectancy into existence. Considering common-­sense intuitions about what we have reason to do in non-­identity cases that involve differences in natural life expectancy—­and considering McMahan’s allegedly common-­sensical ideas about the relevance of fortune—­the required further assumption is not obviously true. So far, I only confirmed McMahan’s verdict that it is hard to justify the common-­sensical view that the engineering practice would be no improvement as compared to the mundane practice. Can we make any progress? Here is my proposal. I suggest that we need to give up the intuition about the dogs and accept that, all else being equal, we have more reason to bring the longer-­lived rather than the shorter-­lived dog into existence (assuming that we ought to promote welfare). Most of us would probably say that if a human couple had the choice between bringing into existence a child that, due to some genetic condition, lived only for twenty years, or a different child that lived for eighty years, and the children would be equally well off during each point of their lives, and nobody else’s welfare would be affected, they had all-­things-­considered reason to bring the longer-­lived child into existence. After all, the parents do more ‘good’ if they bring the longer-­lived individual into existence: not only in an absolute sense but also for their child. Should this intuition make us doubt our intuition in the dog case? As a potentially relevant difference, one might point out that the short-­lived human but not the short-­lived dog 7  The larger absolute quantity of goods in the human’s future life is not the only reason for rescuing the human rather than the dog, according to McMahan. But we can ignore the other considerations here. (See McMahan 2002)

110 

Capacity for Welfare across Species

falls short of what is normal for his kind. But what would ‘his kind’ be? The short-­lived human in our example has a genetic condition that causes him to drop down dead at age twenty. So, his life expectancy is normal for individuals with his genetic condition. The same holds for the short-­lived dog. Both, the short-­lived human and the short-­lived dog, fulfil their native potential. Yet, the shorter-­lived individuals are worse off in absolute terms. What can explain the different intuitions about within-­species or -breed as compared to between-­species or -breed differences in life expectancy in non-­identity cases? We seem to be concerned about a human whose life expectancy falls short of that of other humans. Similarly, we tend to pity a cat if her life expectancy falls short of that of other cats. But we seem to be less troubled by the idea that cats in general have shorter lives than humans. Does this mean that a normal cat is not worse off than a normal human in this respect? I suggest that the cat is worse off. We simply tend not to worry about this comparative disadvantage of cats, because we are used to it and there is nothing that we can do about it. Our emotional reactions, such as concern, are a scarce resource. We could in principle be concerned about many things, such as the fact that our human lifespans are rather limited and most of us die much earlier than we would want. But since this is the normal situation and we are used to it, have never experienced anything else, and cannot do anything about it, we tend not to be very concerned about this fact of life. We just consider it a normal, unchangeable background condition.8 However, the considerations that influence our distribution of scarce emotional resources such as concern shouldn’t blind us to the fact that cats are worse off than humans due to their shorter life expectancy. Even though we do not usually pity cats for their shorter life expectancy, cats would be better off, all else being equal, if they lived longer rather than shorter happy lives. That G is a good for S just means that if S were to possess G, S would benefit from it.9 This holds no matter whether S actually possesses G, wants to possess it, or can possess it. Even though cats cannot live longer in the actual world and may not even desire a longer life, it holds that if they had a longer lifespan in a different possible world, they would be better off in that world (assuming equal and positive synchronic welfare).

8 Kahane and Savulescu (2012) made a similar point about the relevance of statistical normalcy. 9  Eden Lin makes this point in ‘Welfare Invariabilism’ (unpublished manuscript, available at https://www.academia.edu/22820234/Welfare_Invariabilism, accessed on 14 February 2017).

Welfare across Time  111 If we focus on welfare in an absolute sense, as I suggest, does this mean that prioritarians and egalitarians should be particularly concerned about happy cats, who—­due to their shorter life expectancy—­are so much worse off than happy humans? I wouldn’t say that prioritarians and egalitarians have a reason to be concerned, since such concern wouldn’t do any good in a world in which there is nothing one can do to prolong the natural life expectancy of cats. Furthermore, such worked-­up concern would detract energy from more fruitful projects. However, if they could bestow more pleasant years on cats, prioritarians and egalitarians would have at least a prima facie reason to do so. More generally, benefiting cats rather than welfare subjects with a longer lifespan should probably have a certain priority for egalitarians who care about equalizing lifetime welfare (again, also depending on the individual’s synchronic welfare). Where does this leave us regarding our evaluation of the engineering practice? Based on some yet-­to-­be-­specified plausible theory in population ethics that strikes a middle ground between the person-­affecting view and the total view, we could argue that in the mundane practice we have reason not to kill the animal in question, because killing would make this particular animal worse off than it would otherwise have been. In the engineering practice we bring animals into existence, which—­due to their shorter life expectancy—­are worse off than different animals would have been. In both practices our interference, be it the killing or the engineering, makes it the case that things are worse for animals than they would otherwise have been, and equally so. This would allow us to maintain the intuitive verdict that both practices are equally wrong. This discussion of McMahan’s paper allowed me to apply and explore the total-­duration view. If the total-­duration view is correct, then the capacity for diachronic welfare of mice, dogs, humans, and other welfare subjects depends on their natural life-­expectancy. EQU in combination with the total-­duration view yields an equal capacity for synchronic welfare across species, in conjunction with a different capacity for diachronic welfare across species, based on different natural life-­spans.

5 Practical Implications Imagine the following criticism to EQU. One might claim that EQU is absurd since it implies that climate change ought to be promoted. After all, climate change benefits insects and if insects have the same capacity for welfare, it follows, due to the large number of insects, that we promote welfare overall by bringing about climate change. But the implication that we ought to bring about climate change is absurd, or so one could argue, and therefore we must reject EQU. In that sense the insects-­and-­climate-­change argument, which leads to the conclusion that we ought to promote climate change, can play a central role in a reductio-­ad-­absurdum of EQU. But one need not reject EQU to avoid the absurd conclusion that climate change is a good thing. One can also reject one or more of the argument’s other premises. For example, the argument seems to assume that we have reason to neutrally maximize welfare. Thus, it seems to assume utilitarianism as a normative theory. That normative premise (as well as the underlying metaethical assumption that there is something we ought to do) can be questioned. Perhaps welfare is not to be maximized and maybe insect welfare counts for less than the welfare of other welfare subjects. The argument also assumes that insects are welfare subjects. Whether they are welfare subjects depends not only on the correct account of welfare, but also on the cap­aci­ ties of insects, which is a contested empirical issue. Since the major benefit to insects is supposed to consist in their larger number, one can wonder whether coming into existence can benefit an individual or, from a utilitarian perspective, whether adding additional welfare subjects should count as a way of maximizing welfare. Thus, what follows practically from any view about cross-­species capacity for welfare depends on the correct answers to several further questions, mainly normative and empirical ones. EQU as well as DIF have implications for what we ought to do only in conjunction with various normative and empirical assumptions. In this book I argued in favour or EQU and against DIF. This book is not the place to settle further normative or em­pir­ ic­al issues that are relevant for these views’ practical implications. So, the Capacity for Welfare across Species. Tatjana Višak, Oxford University Press. © Tatjana Višak 2022. DOI: 10.1093/oso/9780192882202.003.0005

Practical Implications  113 answer to the question what follows from any of these views can only be: it depends. While the book could end here, I will add some comments on a couple of normative issues that are relevant for determining practical implications of DIF and EQU. (I already commented on the empirical issue of animal capacities in section  1.2.) Section  5.1 discusses the issue of moral status, section 5.2 explores whether bringing additional happy lives into existence should count as a way of maximizing welfare, and section 5.3 provides an overview of ways of counting or discounting the suffering and death of non-­humans.1

5.1  Moral Status This section addresses a basic deontic question, the question what entities we ought to consider at all in moral considerations. This question is addressed in the philosophical debate about who has moral status. The focus of this section is on the moral status of individuals from different species of welfare subjects. If members of certain species of welfare subjects had a lower moral status, their welfare would count for less, even if they had the same capacity for welfare. According to common hierarchical views of moral status effects on cognitively more advanced individuals generally matter more than effects on cognitively less advanced individuals. This section argues against hierarchical views of moral status. I first focus on the influential argument from marginal cases, which is an argument against hierarchical views of moral status. I quickly review and criticize the major responses to this argument. I then criticize Kagan’s (2019) hierarchical view of moral status in more detail. The issue of moral status merits attention in a book about welfare cap­ acity not only because views about moral status are relevant for the implications of any view about welfare capacity. Looking at the philosophical discussion about moral status is also important because many authors consider the very same capacities relevant for an individual’s moral status that are also taken to be relevant for its capacity for welfare. Specifically, an individual’s or group’s cognitive capacities are often said to determine either their moral status or their capacity for welfare or both. For example, Budofson 1  Parts of chapter  5.1 are reproduced with permission from Springer Nature from Višak (2018b).

114 

Capacity for Welfare across Species

and Spears, as we saw in section 2.6, propose to use species-­specific number of brain neurons as a proxy for an individual’s capacity for welfare. This yields the conclusion that the welfare of cognitively less advanced in­di­vid­ uals counts for less. Some people may find this conclusion plausible. However, what drives this intuition may not be the firm conviction that welfare cap­acity is linked to cognitive capacity (which I take to be implaus­ ible), but rather the idea that the welfare of cognitively less advanced in­di­ vid­uals matters less when it comes to our normative reasons for action. So, the underlying intuition may be one about moral status. I think we should distinguish these issues and carefully consider whether cognitive capacities are relevant for welfare capacity or moral status, or for both or for none. I pointed out in this book that I take a welfare subject’s capacity for welfare to be independent of its cognitive capacity. I haven’t seen convincing arguments for cognitive capacities as a basis of differential moral status of sentient beings either. We will see that Shelly Kagan defends a hierarchical view of moral status to avoid counter-­intuitive implications that DIF would otherwise have in conjunction with egalitarianism. As I illustrated regarding the insects-­and-­climate-­change argument, Kagan’s moves show that one can play around with different normative and empirical assumptions to avoid unwelcome implications of DIF or EQU. It is often argued that people deserve some special moral consideration, due to certain properties that they have. For example, Immanuel Kant is generally understood as arguing that the capacity to understand and act upon moral principles convers dignity and a higher moral status upon humans. However, the properties that are supposed to justify a higher moral status for humans are properties that not all human beings share. There are, for example, humans who cannot understand and act upon moral prin­ ciples, such as very young humans or those with limited cognitive cap­aci­ ties. The ‘problem of marginal cases’ refers to a challenge that many theories about who has moral status face: showing whether and, if so, on what basis humans who lack the relevant properties share the higher moral status. The problem of marginal cases is not restricted to questions about moral status, though. It shows up whenever someone claims that all members of certain groups are due some special status or treatment based on characteristics that not all members of the group possess. Consider, for example, the claim that women should be excluded from a particular profession, because they typically lack the required physical strength. Those who subscribe to this claim need to point out how to deal with women who, perhaps un­typ­ic­ al­ly, possess the required strength, or with men who happen to lack it.

Practical Implications  115 There are basically two options. It may ultimately be individual properties that count. In this case the strong women would be allowed for the profession and the weak men would be excluded. Or else there should be a plausible justification for focusing on group membership rather than on individual properties. In addition, of course, one also needs to check whether the property that is claimed to be relevant is relevant and whether it is true that membership of the groups in question is related to possession of the rele­vant property. Thus, for example, it needs to be checked whether physical strength is a relevant criterion for performing the job in question and whether men are typically physically stronger than women. The problem of marginal cases does not primarily concern the identification of the morally relevant property, but the challenge for those who appeal to the relevance of any property to point out how to deal with so-­called marginal cases (as in our example physically strong women and weak men). This challenge exists, because in ethics we ask about normative reasons for action and thus we need justifications for acting in one way or another. If someone claims that it is okay to treat two individuals, A and B, differently, or to take into moral consideration only A’s interests but not B’s, then such a claim needs to be justified. Arbitrary discrimination is, per definition, unjustified. Alastair Norcross provides an example to point out that group membership is not necessarily a plausible criterion for unequal treatment. Imagine that ten famous people face St. Peter at the door to heaven. May they enter the kingdom of eternal bliss, or are they directed down to hell? Of five ­people that are found guilty of crimes against humanity, four are send to hell and one is admitted into heaven. Of five innocent individuals, in turn, four are allowed into heaven and one is told to go down to hell. The innocent who is supposed to go to hell complains and requires a justification. Petrus explains: ‘Isn’t it obvious, Mr Gandhi? You are a male. And the other guilty people—­Adolf Hitler, Joseph Stalin, George  W.  Bush and Richard Nixon—­are male as well. It is thus typical for the group of males to be guilty. In contrast, the women who showed up here were typically innocent. That is why Margaret Thatcher was sent to heaven in your place’ (Norcross 2004, 240). In animal ethics the argument from marginal cases is often applied to challenge those who, like Kant, grant humans a higher moral status than non-­humans on the basis of allegedly typical human properties. If it holds that the suggested properties are not present in all humans and/or are present in some non-­humans, the question pops up how to deal with such

116 

Capacity for Welfare across Species

marginal cases. This question as such is not an argument yet. The appeal to marginal cases can be part of different kinds of arguments in animal ethics. These are all called ‘arguments from marginal cases’. ‘Marginal case’ is a technical term and not meant to be derogative. It refers to individuals, which, in a way that is supposed to be untypical for the group, lack or possess the characteristics that are designated as morally rele­vant for the issue at hand. In this sense, the innocent Gandhi or physically strong women are marginal cases in the examples that I mentioned. Since it can nevertheless sound derogative to refer to individuals who are thought to be untypical in some respect as ‘marginal cases’, some people prefer to avoid this label. An alternative name for the argument from marginal cases in animal ethics is ‘argument from species overlap’. The line between those in­di­vid­uals who possess the allegedly relevant capacities and those who don’t does not follow the species boundary. Instead, there is an overlap. Peter Singer’s Animal Liberation, first published in the 1970s, counts as a source of inspiration for modern animal ethics and the animal welfare movement. In this book Singer uses the term ‘speciesism’ in analogy to ra­cism and sexism to refer to unjustified discrimination based on species. Unequal treatment may be compatible with equal moral status. After all, individuals with equal moral status may have different interests. According to Singer, equal moral status comes down to equal consideration of interests. ‘Interests’ refers to whatever benefits the individual in question. For example, that women are paid less for equal work is unjustified sexism, since there isn’t a good justification for that difference in treatment. Analogously it would be speciesist to harm a pig for purposes that did not justify a similar harm to humans. Singer appeals to the argument from marginal cases in the following way. To escape Singer’s charge of speciesism, those who allow the killing of pigs but not of humans for culinary purposes may point to certain morally rele­ vant characteristics that they take to distinguish humans from pigs. For example, one might claim that pigs are less rational than humans and cannot act morally. However, there are humans whose rationality and capacity to act morally doesn’t surpass that of pigs. This holds, among others, for very young and for cognitively impaired humans. No matter what criterion one appeals to in an attempt to justify the unequal treatment in question, the line between those who possess these properties and those who do not doesn’t follow exactly the species boundary (Singer 1995, 19). If one treats differently even those who are equal regarding the relevant capacity, such as, for example, human babies and pigs, this behaviour may be disqualified

Practical Implications  117 as speciesist. Unjustified discrimination takes place just in case there is no plausible justification for the differential treatment. This version of the argument from marginal cases shifts the argumentative burden back to those who wanted to get rid of the reproach of speciesism. In this way Singer and others use the argument from marginal cases in animal ethics. In that case the argument doesn’t serve the defence of one’s own position, but it is meant to put pressure on rival positions by reproaching them of speciesism. The argument from marginal cases is sometimes also used as support for one’s own position. Some authors argue that human babies and seriously cognitively impaired humans deserve moral consideration and that this shows that more demanding cognitive capacities cannot be the relevant criteria for moral status. Instead, it might be sentience or some other criterion that even these cognitively less developed individuals possess. It follows that even some non-­ human animals are morally considerable. Tom Regan (1979) and Evelyn Pluhar (1996, 63), for example, apply the argument in this manner. These authors do not only appeal to consistency con­sid­er­ ations. Thus, they do not merely say that if babies count then some non-­ humans count as well. Such a conditional would allow for the possibility that neither babies nor non-­humans had moral status. Instead, Regan and Pluhar do not question the moral status of sentient individuals. So, Pluhar claims that in search for consistency, one shouldn’t throw out the human baby with the non-­human bath water. Consistency alone is insufficient. Pluhar argues that all sentient individuals have moral rights and that the status of non-­humans should be adjusted upwards to match that of babies and cognitively impaired humans, rather than the other way around. While Pluhar and Regan mention the argument from marginal cases only in passing as additional support for their positions, Daniel Dombrowski (1997) uses it as the very basis of his defence of animal rights. Controversial discussion in animal ethics concerns what the morally rele­vant property should be; but the basic structure of the argument from marginal cases is widely accepted as legitimate. What is considered the morally relevant property determines which individuals count as marginal cases. Regarding the question about moral status, human babies and cognitively impaired humans are marginal cases for Kant but not for Singer. Singer grants the same and full moral status to all sentient beings. Human babies and the cognitively impaired are therefore not marginal cases for him when it comes to moral status. Singer’s former view on the harm of death, however, used to imply that these groups of individuals are marginal

118 

Capacity for Welfare across Species

cases in this regard. Singer took the fact that these individuals do not have future-­directed desires to be relevant for the question to what extent death is harmful for them. Singer thought that without future-­directed desires these individuals lack relevant interests that normal adult humans typically have. Therefore, he thought that death is a lesser harm for them. (This view, the frustration view, was discussed in section 4.1.) Currently Singer accepts a different theory of the harm of death, according to which death can be harmful for all sentient individuals, independently of their future-­directed desires. Thus, human babies and the cognitively impaired are no longer marginal cases according to Singer’s current view of the harm of death. (Singer can still hold that there are differences when it comes to how bad the death of different beings is, for various reasons as explained in section 4.1.) The argument from marginal cases is based on moral individualism, the idea that actual intrinsic properties of individuals are relevant for questions about these individuals’ moral status (Rachels 1990). This fundamental premise can be questioned in at least two ways. Some authors deny that the relevant properties must be intrinsic. These authors suggest extrinsic, especially relational properties. Others argue that the morally relevant properties need not be actual properties. These authors argue that potential or so-­called modal properties may be relevant instead. (As will become clear in the discussion of Kagan’s hierarchical view on moral status, Kagan has a different understanding of moral individualism. According to him, denying that actual and intrinsic properties matter is compatible with individualism. So, he does not reject individualism, as he understands it, while he denies that actual and intrinsic properties matter and thus rejects individualism as defined here. The verbal question, how to understand moral individualism, does not concern me here. I will discuss substantive views about what the relevant properties are in what follows and evaluate Kagan’s view more closely at the end of the chapter.) Let me first turn to the objection that not only intrinsic properties matter morally. Carl Cohen, for example, insists that it is enough to be a member of a group whose normal members have the relevant property. He says: The capacity for moral judgment that distinguishes humans from animals is not a test to be administered to human beings one by one. Persons who are unable, because of some disability, to perform the full moral functions natural to human beings are certainly not for that reason ejected from the moral community. The issue is one of kind . . . .  (Cohen 1986)

Practical Implications  119 Alan White (1984) and David Schmidtz (1998) present similar arguments. Alastair Norcross defends the individual approach against such views with the above-­mentioned example of Gandhi’s talk with Petrus at the door to heaven. Since we generally do not rely on group membership in judgements of individuals, appealing to group membership seems ad hoc. As Norcross complains: A particular feature, or set of features is claimed to have so much moral significance that its presence or lack can make the difference to whether a piece of behavior is morally justified or morally outrageous. But then it is claimed that the presence or lack of the feature in any particular case is not important. The relevant question is whether the presence or lack of the feature is normal. Such an argument would seem perfectly preposterous in most other cases.  (Norcross 2004, 240)

Some authors appeal to feelings of connectedness or relationships in their claims that human marginal cases in contrast to non-­humans should get full consideration. For example, Bonnie Steinbock explains: I doubt that anyone will be able to come up with a concrete and morally relevant difference that would justify, say, using a chimpanzee in an ex­peri­ment rather than a human being with less capacity for reasoning, moral responsibility, etc. Should we then experiment on the severely retarded? Utilitarian considerations aside, we feel a special obligation to care for the handicapped members of our own species . . . Here we are getting away from such things as ‘morally relevant differences’ and are talking about something much more difficult to articulate, namely, the role of feeling and sentiment in moral thinking.  (Steinbock 1978)

Regarding Steinbock’s appeal to given preferences and sentiments, one can wonder whether this provides enough of an ethical justification for the differential treatment in question. A mere appeal to people’s ­inclinations does nothing to justify them. At least Steinbock would need an argument of why these feelings are supposed to hint at what is morally right. Furthermore, sympathy is a fragile basis for moral consideration. In the past, members of suppressed groups couldn’t necessarily count on sympathy from the part of the dominant group and the same is true today. The second line of criticism against the argument from marginal cases has it that individuals need not actually possess the relevant properties,

120 

Capacity for Welfare across Species

such as rationality or moral agency. It is sufficient, according to this objection, to have the potential for the properties in question. Thus, for example, it is true of babies that even though they lack the relevant properties, they possess the potential to be rational moral agents. This means that babies can have these properties in the future, and they will have them if all goes well. The moral ­relevance of potential properties is highly controversial and much debated, specifically in connection with abortion. Jeff McMahan, for example, notes: There remains the familiar question about the potential for rationality as a basis for this moral status: namely, why should an immature human being’s internal directedness toward the development of a rational nature affect how we ought to treat that individual now? There is, of course, a good answer to that question in cases in which how we treat the individual now will affect her for better or worse later, when she actually has a rational nature. But there are certain types of acts the effects of which are limited entirely to the present, including—­crucially—­acts of killing. Why should the morality of an act of killing be governed by the kind of respect that is appropriate for a nature that the individual killed does not have now but may have later, though only if it is not killed? Why should an act of killing not be governed instead by due consideration for the nature of the individual at the time of action (or of the death, if it occurs later)?  (McMahan 2008, 88)

Furthermore, since many marginal cases do not even potentially possess the relevant property, an appeal to potential cannot be a satisfactory response to the argument from marginal cases, according to many authors. A further option involves consideration of modal properties. These are properties that the individual could have realized but cannot realize anymore. David Schmidtz (1998) might have such properties in mind when he claims that in contrast to non-­humans, marginal humans are ‘logically pos­ sible’ subjects of rights and that it is ‘a misfortune, but not a tautology’ that they cannot claim these rights. Bonnie Steinbock (1978) as well might aim at modal properties when she justifies granting rights to marginal humans in the following way: It makes sense to think that one might have been born retarded, but not to think that one might have been born a monkey.

Shelly Kagan presents an elaborated defence of ‘modal personism’, which will be discussed in more detail at the end of this chapter.

Practical Implications  121 Some proponents of natural rights theories argue that members of certain species have the species-­specific rights by nature, independently of whether they possess the morally relevant properties as individuals. Rahul Kumar argues, for example, that claims about species are not statistical generalizations. Rather, what they concern is the essential nature of a living kind, revealing facts about the normal life-­cycle of that kind of living thing. The use of ‘normal’ here is unashamedly normative. . . . What respect for the value of a living thing requires will depend on the characteristic life-­cycle, or nature, of members of that [individual’s] species.  (Kumar 2008)

Thus, according to Kumar, it isn’t merely a statistical fact that most humans are rational. Rather being rational is an aspect of human nature. Against this line of argumentation, it can be objected that without appealing to stat­ is­tic­al facts about the species in question it is hard to determine what a particular species’ nature is (McMahan 2008). Given the difficulties with non-­speciesist ways of justifying a higher moral status for humans, can’t one directly appeal to the relevance of species instead? Bernhard Williams (2009) belongs to the very few authors who do just that. Williams accepts subjectivism about normative reasons for action. He claims that our fierce conviction that species membership is morally rele­vant provides normative reason for us to act accordingly. Imagine that some alien species wants to eradicate humans, because it correctly calculated that a world without any humans would be better for all. In such a case, Williams claims, it would be morally justified to prefer the survival of members of our species simply because we are humans. He asks the rhet­ oric­al question: ‘Which side are you on?’ Williams, thus, considers spe­cies­ ism to be justified in principle. To this one can object that Williams’ rhetorical question sounds implausible in a conflict between humans of different races or sexes. So, why should it be plausible when it comes to species? Contrary to the situation in Williams’ thought experiment involving the aliens, our bare survival is typically not at stake in our daily interactions with non-­humans. However, our actions often affect the life and welfare of non-­humans. In this conflict choosing the side of the humans, simply because we are humans, seems unjustified. The question cannot simply be: ‘Which side are you on?’ Instead, the question must be how to live up to the ethical challenge to justify our actions. In this pursuit, the argument from marginal cases plays a central role and hasn’t been convincingly rebutted so far.

122 

Capacity for Welfare across Species

I will now take a closer look at Shelly Kagan’s (2019) defence of a hier­ arch­ic­al view on moral status and point out some of the implications that, I think, make it implausible. I also comment further on the relationship between the issue of moral status and the issue of cross-­species capacity for welfare. Kagan is concerned about counter-­intuitive implications that DIF would have in conjunction with non-­hierarchical views of moral status. He defends a hierarchical view of moral status at least partly to avoid these implications while maintaining DIF. What bothers Kagan are the implications of DIF for those welfarists who hold, as he himself does, that the distribution of welfare, rather than only its overall amount, determines the value of an outcome. Many of those who evaluate outcomes in terms of welfare believe that the distribution of welfare matters. Consider, for example, prioritarianism. Prioritarians hold that how much a given quantity of welfare improves an outcome depends on whether the welfare goes to individuals who already have a lot of it, or to individuals who have less of it. Per definition, the personal benefit that people derive from one unit of welfare is the same: one unit of welfare is one unit of welfare. Different people may need different things to get a welfare increase of one unit. For example, for rich people it may take more money to give them a welfare increase of one unit than it takes for poor people. But once they get a one unit increase, this benefits them just as much as a one unit increase would benefit poor people. That is true per definition and prioritarians do not deny it. What they deny is that the welfare value of an outcome is solely a matter of the personal value of welfare. They take into account other things besides effects on individuals. They thus value welfare in an impersonal way. According to DIF healthy and happy non-­human animals may be much worse off than only mildly satisfied humans. Combining DIF with prioritarianism about reasons for action implies that it is more important, say, to make an already happy and healthy mouse even happier than it is to benefit an only mildly satisfied human. Kagan, along with probably most other ­proponents of DIF, finds this implication unacceptable. Even though he agrees that animal welfare matters in principle, he is convinced that helping the happy mouse in question is certainly not more pressing than helping the only mildly satisfied human. Given DIF, the human in our example, despite his barely positive emotional state, is doing much better than the happy and healthy mouse, though. If we nevertheless have reason to benefit the human rather than the significantly worse off mouse, this must be, according to Kagan, because of the human’s higher moral status. The differences in status

Practical Implications  123 are so great, according to Kagan, that prioritarians need not prioritize the mouse even though it is much worse off than the human. DIF in conjunction with a hierarchical view on moral status it thus a combination of views that yields the intuitively plausible result in the case at hand: the barely happy human deserves more of our concern than the very happy mouse. I agree with Kagan that helping the human rather than the mouse is the plaus­ible verdict in this case. But, unlike Kagan, I think this is so because the human is worse off at the time. Kagan’s acceptance of a hierarchical view of moral status is thus closely related to his acceptance of DIF on how to compare welfare across ­species. Kagan defends a hierarchical view on moral status at least partly with the aim of avoiding some of DIF’s counter-­intuitive implications. I  agree with Kagan that DIF has counter-­intuitive implications in conjunction with certain normative views, such as prioritarianism. But as a way of avoiding these problems I would rather suggest abandoning DIF. I would want to maintain a non-­hierarchical view about moral status. There are, thus, two rival proposals on the table: (1) DIF in conjunction with a hierarchical view on moral status and (2) EQU in conjunction with a non-­hierarchical view on moral status. It will be interesting to compare the strengths and weaknesses of the different combinations of views. I  think that accepting a hierarchical view can indeed take away some counter-­intuitive implications of DIF. But it does so at too high a cost, since the hierarchical view on moral status is itself implausible upon closer inspection (as is DIF). According to the hierarchical view, an individual’s moral status depends on its psychological capacities. The relevant capacities are cognitive and emotional capacities. For example, normal human adults are rational and self-­conscious, capable of abstract thought and deliberation, aware of themselves as having a future and possessing detailed preferences for that future. These capacities are characteristic of persons. There may also be non-­ humans who possess these capacities to a similar extent as normal adult humans. These non-­humans would then count as persons as well. Probably most non-­humans that possess the relevant capacities have them to a significantly lower degree than normal adult humans. These non-­humans may be called ‘minimal persons’. Many other non-­humans are non-­persons. Roughly, the extent to which an individual possesses these and similar capacities, determines the individual’s moral status, according to the hier­ arch­ic­al view under consideration. Thus, a person has a higher moral status than a non-­person.

124 

Capacity for Welfare across Species

Kagan suggests, at least as one possible specification of such a view, that precisely the individuals that tend to be better off according to DIF may be accorded a higher moral status. This may be the case, because the exact same capacities that allow an animal to have a higher amount of welfare may also be the capacities that give these animals their higher moral status. As Kagan (2019, 121) puts it: So lives that are more valuable by virtue of involving a greater array of goods, or more valuable forms of those goods, will require a greater array of psychological capacities, or at least more advanced versions of those capacities. And the thought, then, is that it is precisely these differences in capacities that are relevant for welfare that are also relevant for fixing status. More advanced capacities make possible more valuable forms of life, and the more advanced the capacities, the higher the moral status grounded in the possession of those very capacities.

The thought is thus that an individual’s status increases along with its potential for welfare since the same underlying capacities are responsible for both. This allows Kagan to avoid the implausible implications that DIF would otherwise have in conjunction with prioritarianism. And, as Kagan is aware, it’s not only prioritarianism that has implausible implications in conjunction with DIF, but also other views that take distribution into account, such as egalitarianism and sufficientarianism. Kagan seems to hold that it is not the capacities as such that determine an individual’s status, but rather the range of well-­being that an individual can attain. This suggests that status depends on the range of well-­being that an individual is capable of. Kagan (2019, 122) remarks that the psychological capacities relevant for fixing status are relevant by virtue of their potential impact on one’s wellbeing, whether for good or for ill.

If this is indeed what Kagan wants to say, then his hierarchical view on moral status depends on the truth of DIF! The hierarchical view on status is thus, at least in part, motivated by the desire to avoid counter-­intuitive implications of DIF and it is also premised on its truth. According to this line of reasoning, both views stand or fall together. Kagan wants to avoid an overly narrow approach when it comes to the psychological capacities that make welfare possible. He says (122):

Practical Implications  125 Welfare is the result of desiring, pursuing, acquiring, and enjoying pos­ sible goods (and avoiding possible bads).

I cannot really make sense of that. What does Kagan mean by ‘possible goods’ here? Are these the final goods? I think that is what Kagan means, since he also refers in this context to the ‘constituents’ of welfare (Kagan 2019, 123). But according to hedonism, enjoyment is good for me, even if I do not desire or pursue it. It wouldn’t make sense for a hedonist to say that to have welfare, which is understood in terms of enjoyment, I must desire, pursue, acquire, and enjoy enjoyment. And for a desire theorist, the basic good is getting what I want. It doesn’t matter whether I enjoy it. But even for proponents of an objective list account of welfare, welfare simply consists in having certain objective goods in one’s life. It doesn’t seem to matter whether I desire, pursue, or enjoy these goods. So, it is not clear to me in what sense Kagan takes exactly these psychological capacities to make welfare possible. Kagan (2019, 122) asks the reader to consider cases in which a given individual lacks some psychological capacity that is needed for the obtaining or enjoyment of a given good, but nonetheless recognizes that very lack, and wishes it were the case that they had the requisite capacity (or capacities) so that they could try to acquire the missing good. [. . .]

He then suggests that the very ability to have such aspirations may well be relevant to fixing one’s status as well. What matters isn’t just the kind of welfare that one can in principle achieve (given one’s capacities), but the kind of welfare to which one can aspire.  (Kagan 2019, 122)

It is not clear to me what Kagan means by ‘kind of welfare’. Perhaps he refers to different objective goods on an objective list account of welfare. But if the capacity to aspire to such a good is also relevant for fixing an individual’s status, even if the individual cannot attain that good, an individual’s status matches neither its capacity for welfare not its actual welfare. Kagan is also inclined to accept that being a moral agent confers a special moral status. It is, however, as Kagan admits, controversial whether moral agency makes one’s welfare possible. According to most accounts of welfare

126 

Capacity for Welfare across Species

there is no such link. Kagan thinks that being capable of moral agency makes (more) welfare possible, since he takes moral agency to be one of the goods on his list of objective goods. This shows that Kagan’s claim that moral status is based on the capacity for welfare assumes not only DIF but also a rather specific objective list account of welfare. (I point out in detail in section 2.1 that Kagan’s defence of DIF assumes a specific objective list account of welfare.) As an alternative proposal as to which capacities ground status, and one that he prefers, Kagan suggests that status may be based on agency. Kagan (2019, 124) remarks about some advanced cognitive capacities: Even if these very same capacities all turn out to play a role in making possible the having of lives with a particular kind of well-­being, I suspect that the ultimate reason that these capacities enhance our status is by virtue of the role they play in constituting the particular kind of agency that we possess.

To the extent to which capacity for welfare and agency come apart, Kagan’s hierarchical view on status will perform worse in remedying DIF’s counter-­ intuitive implications. Imagine for example a machine that possesses agency. (Kagan is willing to ascribe agency to machines.) The machine cannot realize many of the objective goods in its life because it cannot grasp, pursue, or enjoy them. This machine would be badly off, according to an objective list account of welfare. It wouldn’t suffer at all or miss anything, we can assume. But nevertheless, since it is an agent to some significant extent, we may have strong reasons to improve the welfare of this machine, according to Kagan’s line of argumentation. This is just to make the principled point that to the extent that agency and capacity for welfare can come apart, an appeal to status does not take away DIF’s counter-­intuitive implications. Before further evaluating Kagan’s hierarchical account of status, I should note another important feature of it. Actual capacities are not all that matters for moral status, according to Kagan. If an individual has some moral status based on its actual capacities, this status can be augmented by the individual’s potential and modal capacities. As already mentioned in this chapter, potential capacities are those that the individual does not yet possess but will possess in the future if all goes well. For example, a human baby is a potential person. Modal capacities, in turn, are those that an individual does not and will never possess but could have possessed. For example, a

Practical Implications  127 human adult who lost her potential for personhood due to some accident during birth is a modal person. Similarly, a kitten that can still develop into a normal adult cat is a potential adult cat. And an impaired adult cat that could have developed into a normal cat but hasn’t done so and can’t do so anymore is a modal cat. As Kagan makes clear, modality and potential do not provide status all by themselves. But they can augment any status that already exists based on an individual’s actual capacities, such as, for ex­ample, its sentience. The extent to which modal and potential capacities augment status depends on the likelihood with which the individual in question will possess or could have possessed the relevant higher capacities. For example, the likelihood that a mouse will develop into a person is very low and even negligible. It would take some yet unavailable technology to let a mouse develop into a person. In contrast, the likelihood that a normal human child will become a person is rather high, since this is what will happen in the normal course of events. The extent to which potential and modality augment status also depends on what the potential or modal capacities are. For example, the status of a potential normal adult human is higher than that of a potential normal adult mouse. Lastly, the more status there already is, the more it can be augmented. Thus, the individual’s actual capacities also determine how much its potential and modal capacities can add to the actual status. Kagan believes that the status of, say, a modal person should be lower than that of an actual person, but higher than that of the individual’s psy­cho­logic­al peers that lack this modal feature. Psychological peers are individuals whose actual psychological capacities are at a similar level. According to Kagan, the modal person’s status should be about halfway that of an actual person and that of her psychological peers who lack modal personhood. Here is what I find most counter-­intuitive about Kagan’s hierarchical account of status. Let us assume that the fact that someone is in pain gives us a prima facie normative reason to take away this pain. Let us also assume that two different individuals, A and B, experience an exactly equal amount of pain and that we can help only one of them by taking away this individual’s pain. There are no effects whatsoever on third individuals and we have no special relation to any of them. According to Kagan, we have a stronger reason to take away the pain of a normal adult human rather than that of a severely mentally handicapped human. This is not because the former suffers more. She doesn’t. It is, after all, stipulated that both individuals suffer

128 

Capacity for Welfare across Species

equally. It is simply because of the normal adult human’s higher cognitive capacities that we ought to help her, according to Kagan. I cannot see why cognitive capacities should be relevant for our reasons for action (or the value of the outcome) in this case. Intuitively, whether one individual is more intelligent, say, is as irrelevant as their hair colour or the amount of money on their bank account, when it comes to the value of alleviating their pain. It doesn’t affect the extent of their suffering or the benefit that they would derive from my help. Therefore, it seems to me irrelevant in the case at hand what the cognitive capacities of each of the individuals are. It doesn’t matter whether the choice is between two normal adult humans, one slightly more cognitively advanced than the other, or whether the differences are larger. And it doesn’t matter whether the choice is between in­di­ vid­uals of different species. It simply doesn’t seem relevant to me in this situation how intelligent the individuals are. As explained, Kagan tries to soften the force of the argument from marginal cases by granting relevance to potential and modal status. Yet, there is the fundamental worry how to justify the moral relevance of modality (DeGrazia 2015). Kagan merely notes that we regret it more if a good event could have happened but failed to be realized than if it was clear from the beginning that it could never have happened. DeGrazia (2015, 3) agrees with that psychological observation. He explains how much he regrets it if his favoured football team could have prevented losing the match if only the shot at the goal in the last minute was successful. If the ball in this case fails to reach the goal by a small margin, this is regrettable indeed. But, as DeGrazia points out, this doesn’t change the fact that the team in question loses the match. It doesn’t matter for the result of the match how closely the ball missed the goal: modal goals do not count. The burden of proof is therefore upon Kagan, who needs to argue what makes modal properties morally relevant. It seems odd to allow future or past possibilities for attaining a higher status to augment one’s actual status, while not allowing future or past possibilities for a lower status to reduce one’s actual status. Kagan (2019, 141) provides no justification for this asymmetry, as he admits. Of course, it may be too extreme to claim that negative potentials and negative modal statuses don’t count at all. But at the very least they do seem less significant than correspondingly strong positive potentials and modal statuses, and that fact remains puzzling.

Practical Implications  129 Let me come back to my example of two individuals who are experiencing an equal amount of pain. I already stated my intuition that intelligence doesn’t matter when it comes to whose pain to relieve. Now we can assume that both individuals are entirely equal in their actual psychological cap­aci­ ties. Let us say that both individuals are cognitively disabled human children. Does it matter that one of them got her handicap during birth, while the other got the exact same handicap due to some genetic defect during conception? According to Kagan, it matters. Due to the different likelihoods that one or the other could have become a person, we have a stronger reason to alleviate the pain of the individual that got her handicap during birth rather than the one that got it during conception. For me, this is counter-­ intuitive. Both children do not only suffer equally; they are also similar in all other actual (and even potential) characteristics. I find it hard to believe that how they came to be handicapped, both of course through no fault of their own, matters for our reason to alleviate their current pain. Imagine that both individuals in the example are cognitively severely disabled human infants. They are equal in all respects, except that one is a bit older than the other. The older one is a modal person. After all, she could now have been a person, if things had gone differently and she weren’t handicapped. The younger one is not a modal person yet. After all, due to her younger age, she would not have been a person yet, even if she weren’t handicapped. Only in a year’s time, it will be possible to call her a modal person. She is, thus, merely a potential modal person. Does this make a difference regarding whom to help? According to Kagan, this depends on whether modal persons have the same status as modal potential persons, an issue that he doesn’t settle. I can only, once again, register my intuition that I find it implausible that similar pains provide reasons for action of different strengths in the way that Kagan suggest. Summing up, at least several welfarist views, such as prioritarianism, egalitarianism, and sufficientarianism have implications in conjunction with DIF that even proponents of these combinations of views consider implausible. For example, DIF in combination with prioritarianism may imply that we should devote resources to making a happy mouse (which DIF takes to be relatively low in welfare) even happier rather than to improving the welfare of an only mildly satisfied (but, according to DIF, better off) human. To avoid these implications, proponents of DIF may either reject these normative views or they may combine DIF with a hierarchical view on moral status.

130 

Capacity for Welfare across Species

I think that acceptance of a hierarchical view on moral status is a very high cost for those who are drawn towards the above-­mentioned welfarist views on normative reasons for action. An alternative would be rejecting DIF in favour of EQU. For example, prioritarianism in conjunction with EQU yields the plausible implication that one should devote one’s resources to the mildly happy human rather than the very happy mouse, because the former is worse off. (EQU in combination with prioritarianism, egalitarianism or sufficientarianism would have some remarkable implications in cross-­species comparisons of diachronic welfare, due to species-­specific differences in lifespans. I discussed this issue in section 4.2.)

5.2  Promoting Welfare by Creating Welfare Subjects? One assumption of the insects-­and-­climate-­change argument against EQU is that climate change benefits insects. The idea is that there will be more insects on the planet if the temperatures rise and that these insects are well off. Even if this was true (and I am not convinced that insects are welfare subjects at all) whether this counts in favour of bringing about climate change depends largely on whether we have reason to promote welfare by making it the case that a greater number of well-­off welfare subjects exist. As explained in section  4.3, there are different views in population ethics. I mentioned the impersonal total view, which asks us to maximize welfare either by making individuals better off or by creating more well-­off in­di­vid­ uals. These two ways of maximizing welfare are on a par, according to the impersonal total view since what counts is only the total quantity of welfare in the universe that an action brings about. I also mentioned the person-­ affecting view, which considers an outcome to be better than another one if and only if and to the extent that it benefits individuals. This view cannot say whether one of two outcomes is better or worse if only different in­di­vid­ uals exist in each outcome. If some individuals existed in both outcomes, the outcomes could be compared based on only these individuals’ welfare. A third view in population ethics would be one that entails theory Q, i.e. the claim that outcome A is better than outcome B if and only if and to the extent that outcome A is better for the A-­people than outcome B for the B-­ people. These need not be the same individuals. Ralf Bader (2022) defends an interesting (and, I think, plausible) theory along these lines, which I will sketch in the remainder of this chapter. Bader discusses the

Practical Implications  131 theory’s implications, assuming (if only for the sake of argument) that we have reason to maximize welfare. Bader (2022) distinguishes two ways in which one can care about welfare. According to impersonal views, impersonal welfare is morally significant. Meanwhile, welfare as a personal good is only considered to be ethically relevant and it is connected to welfare as an impersonal good via a conversion function, as will be explained below. Personal good is a particular kind of good. It is what is good from a particular perspective: the perspective of an individual welfare subject. ‘Impersonal good’, in contrast, refers to what is good from the impersonal perspective, or, as it is often called, from the point of view of the universe. Of course, the universe doesn’t literally have a point of view. Thus, one shouldn’t take talk about ‘perspective’ literally in this context. But these are two fundamentally different types of good. Some proponents of impersonal views make it explicit that they acknowledge only impersonal good. They reject the notion of personal good as unnecessary. It is more common, however, for proponents of impersonal views to accept both kinds of good, but to consider only impersonal good to be ethically significant.2 Personal good, then, must be converted into impersonal good via a conversion function. This is analogous to converting one kind of currency into another. Even if the conversion rate were 1:1, we still had two different currencies. For example, the conversion function from personal good to impersonal good that the impersonal total utilitarian accepts is linear: one unit of personal good counts for one unit of impersonal good. To clarify this, it may help to consider prioritarianism, by way of contrast. Prioritarianism can be viewed as an impersonal theory with a different, namely strictly concave, conversion function: The more units of personal good one already has, the smaller the contribution of each additional unit of personal good to impersonal good. Yet, the value function at the level of impersonal good is the same for both theories: they add up the impersonal good to determine the goodness of the outcome. The graphs in Figure 5.1 depict personal good on the x-­axis and impersonal good on the y-­axis. The conversion function for utilitarianism is shown on the left, for prioritationism on the right (Bader 2022). The value function at the level of impersonal good is simply additive. The aggregate amount of value of a distribution is determined by adding up 2  Impersonal views may also hold that both personal and impersonal good are morally significant. Such views are scarcely accepted in the current debate and are implausible for various reasons.

132 

Capacity for Welfare across Species Impersonal good

impersonal good

personal good

personal good

Figure 5.1  The conversion function from personal to impersonal good, according to utilitarianism (left) and prioritarianism (right).

the impersonal value in the distribution. Alternatives to the total view are the average view, variable value views, and critical level views. These are all impersonal views. For most of these impersonal views, it is obvious that personal good and impersonal good can come apart in the evaluation of outcomes. For example, on averagism, if only miserable people exist, it can make an outcome better to add a group of somewhat less miserable people, since this raises the average level of welfare. Yet, the latter outcome is not better in any way for people. In case of impersonal total utilitarianism, it may be less clear how these things can come apart. I will therefore point this out in more detail in what follows. To recap, personal good is only ethically relevant, but not significant, according to impersonal views. What really matters, namely impersonal good, is related via some conversion function to personal good. Welfare can thus be counted as either a personal good or an impersonal good. Proponents of person-­affecting views consider it to be the main problem of impersonal views that they count welfare only as an impersonal good. One could think that just as more happy moments make a life better, more happy people make a distribution (i.e. an outcome) better. But while a life is a fundamentally important axiological unit, a distribution is not. Ralf Bader (2022) puts this criticism of impersonal views as follows: [. . .] impersonal views treat the relation between a person and a distribution as being analogous to that between a time-­slice and a person, thereby failing to recognise the importance and axiological significance of the personal point of view.

And further: Intuitively, it seems that we need to take persons as well as personal good more seriously. Persons are not mere containers of goodness, i.e. locations

Practical Implications  133 where impersonal good can be instantiated, but are beings that matter and that are normatively significant. We have reasons to help others because it makes their lives go well. Promoting their happiness is something that we have reason to do because their happiness is good for them, not because there being more happiness makes the world a better place. The fact that their lives go well is of direct ethical significance.

This discrepancy about significant axiological units lies at the basis of the distinction between impersonal and person-­affecting views. I am not sure how to settle the dispute about what the fundamentally relevant axiological units are. Perhaps one reaches intuitive bedrock here. Yet, it should be noted that this is not just some intuition about a particular case or about a set of similar cases. It is, instead, a fundamental intuition about what matters. That individuals are mere containers of goodness according to the impersonal total view most clearly comes to the fore in the theory’s implications in different number cases, such as the repugnant conclusion and the replaceability argument. Very roughly, the impersonal total view implies that an outcome in which a billion very well-­off individuals exists is worse than an outcome in which a much larger number of individuals exist whose welfare levels are barely above the neutral point. Perhaps these individuals’ welfare levels are so low that a mere pinprick would bring them below the zero level. As Parfit (1984, 388) puts it, the largest amount of milk may be found in a heap of bottles, each of which contains only a single drop. Furthermore, the impersonal total view implies that an outcome in which a life ends prematurely is not worse than an outcome in which the individual continues to lead a happy life, provided that if the life ends prematurely, another individual comes into existence that would not otherwise have existed and whose life contains at least as much welfare as the future life of the dead individual would have contained (Singer 2011, 106). Alternatives to impersonal views are person-­affecting views. According to person-­affecting views, personal good is morally significant. In fact, person-­affecting views reject impersonal good and hold that all goodness is (reducible to) personal goodness. This is the fundamental axiological difference between person-­affecting views and impersonal views. All person-­ affecting views hold that lives are the fundamental axiological units. Lives are the things that have a particular (personal) value: they can be good or bad, better, or worse, for the individuals whose lives they are. Person-­affecting views come in different versions, not all of which are equally plausible. Wide person-­affecting views can be distinguished from narrow person-­affecting views. Narrow person-­affecting views hold that

134 

Capacity for Welfare across Species

outcome A is better than outcome B if and only if the individuals are better off in A than they are in B. Narrow person-­affecting views imply that outcomes consisting of completely different individuals are either equally good or non-­comparable. Outcomes with partly overlapping populations are evaluated based on the welfare of only the individuals that exist in both outcomes. Wide person-­affecting views, in contrast, hold that outcome A is better than outcome B if and only if A is better for the individuals living in A than B is for the individuals living in B, where the populations in both outcomes may or may not overlap. Wide person-­affecting views can compare outcomes in different-­people cases, but not in different-­number cases. The difference between these views is that narrow as opposed to wide person-­affecting views consider identity to be axiologically relevant. They reject impartiality. Narrow person-­affecting views have counter-­intuitive implications. For example, they imply that outcomes in which different individuals with different welfare levels exist are non-­ comparable (or equally good) even though, intuitively, one outcome is clearly better than the other. This is known as the non-­identity problem. Wide person-­affecting views avoid the non-­identity problem, as well as other problems of narrow person-­affecting views. Ralf Bader’s version of a wide person-­affecting view, ‘Baderism’, as I call it, is based on non-­comparativism, the view that existing and never existing are non-­comparable in terms of personal good. Outcomes containing different numbers of people are non-­comparable in terms of personal good, and thus are non-­comparable tout court, according to Baderism, because it only counts personal good. Equinumerous outcomes, in contrast, can be compared by matching everyone in one outcome to one individual in the other outcome and then assessing for each matched pair of individuals, who is better off and to what extent. The differences are then aggregated, as depicted in Table 5.1. This results in an overall verdict about what outcome is better and to what extent. It doesn’t matter whom to match to whom. Different matchings wouldn’t change the result of the evaluation, i.e. the betterness relation between the outcomes. Gains and losses to individuals are, thus, balanced in an impartial way. Identities do not matter. It doesn’t matter who loses and who gains. It might seem odd to talk about gains and losses when different people exist in the different outcomes. There are, after all, no individuals that are better or worse off in one outcome or the other. But since Bader does not consider identities to be axiologically relevant, but only lives as such, different lives can be compared in terms of the personal goodness that

Practical Implications  135 Table 5.1  Horizontal balancing as a method for comparing the value of outcomes. Outcome A

Outcome B >1 2 A effects on welfare → normative reasons for action

Figure 5.4  Discounting based on harm of death.

follow this line of argumentation, as depicted in Figure 5.4, claim that premature death is less harmful for individuals with weaker cognitive capacities. These authors may, for example, accept the frustration view or a time-­ relative interest account of the harm of death. I discussed these views in section 4.1. A third line of argumentation intervenes between effects on welfare and normative reasons for action. More specifically, it builds in an extra step between welfare and normative reasons for action. Jeff McMahan (1996) does so, and he calls this extra step ‘fortune’. The discounting on the basis of weaker cognitive capacities takes place between welfare and fortune, as indicated in Figure 5.5. Good or bad fortune, rather than welfare, according to this view, grounds (some of our) normative reasons for action. (Whether and to what extent such a view would discount effects on non-­humans would depend on further details, such as one’s account of welfare and in what cases one takes fortune to be relevant for normative reasons for actions. The view does not necessarily discount non-­humans.) I discussed McMahan’s arguments in section 2.4. A fourth line of argumentation discounts between different notions of welfare. For example, Kevin Wong (2016) distinguishes between what he calls ‘experiential welfare’ and ‘absolute welfare’, as Figure  5.6 shows. His view implies that even though a dog may lose 5 points in experiential welfare and a human only 3, the human may lose more welfare in absolute terms. This way of discounting is an example of the difference view about ­cap­aci­ties for welfare across species, DIF. I discussed various versions of DIF, along with McMahan’s above-­mentioned closely related proposal, in sections 2.1 to 2.6. I think that the arguments in favour of discounting that are on offer in the philosophical literature are far weaker than they may initially seem. We may wish that one or the other of these arguments be successful, but this

140 

Capacity for Welfare across Species

Premature death and suffering → effects on welfare ---> effects on fortune → normative reasons for action

Figure 5.5  Discounting based on fortune. empirical facts → effect on experiential welfare ---> effect on absolute welfare → normative reasons for action

Figure 5.6  Discounting based on the distinction between experiential and absolute welfare.

doesn’t mean that they are successful. If people agree with the result of some argument, they may be less critical about the strength of the argument itself. But I think it does a service to philosophical reasoning if we scrutinize the arguments whose conclusions suit us as carefully as those whose conclusions do not suit us. I think this is not always done when it comes to what we owe to non-­human animals. Even many of us who are strongly in favour of a better and fairer treatment of non-­human animals are often reluctant to follow the arguments where they seem to lead. We are eager to find ways to circumvent some uncomfortable conclusions. So, we should be particularly careful, as I said, about the quality of our philosophical reasoning in these areas. I hope that this book can point out weaknesses in the philosophical reasoning about how to count animals, specifically in views about cross-­ species capacities for welfare. In this book I focused on the question of cross-­species capacity for welfare and argued that contrary to what seems to be the dominant view among those who are writing on the issue, EQU is at least as plausible as DIF. I, thus, aimed at strengthening the view according to which a well-­off mouse is roughly as well off as a well-­off dog and a well-­off human. The average benefit that we can bestow on a mouse at a time is not necessarily smaller than the one that we can bestow on a dog or a human. A mouse that loses a year of life may well lose as much welfare as a dog or a human who lose a year of life. At the very least, it is far from settled that the contrary is the case. If I must choose between granting a mouse or a human an additional year of happy life, does EQU imply that my normative reasons for doing each of these things are equally strong? No, EQU, of course, is not a view about normative reasons for action. It only implies that the benefit for both individuals would be roughly equally great. Intuitively, most of us would probably consider prolonging the human’s life more important.

Practical Implications  141 This in­tu­ition is not necessarily at odds with EQU. EQU is only a view about cross-­species capacity for welfare. To know what we ought to do, we need to appeal to a theory about our normative reasons for action. We thus need to turn to normative ethics, and even metaethics. Assume that we have normative reasons for action, and that they are based on the effects of our action on welfare in the sense that we ought to neutrally maximize welfare. Still, it need not follow from the fact that the mouse would lose as much welfare as the human that it doesn’t matter whom to save. It could, after all, be the case that saving the human would prevent more suffering, since, due to her greater cognitive capacities, the human, in contrast to the mouse, is aware of her situation and of what she stands to lose. Furthermore, the death of the human may cause more harm for others, due to the numerous and close relations that humans may have, including, perhaps, longer memories of adverse events. These and similar considerations are relevant for our welfare-­based reasons for action. The implications of my defence of EQU also depend on one’s metaethical assumptions. Assume that there is objective moral truth. If so, our in­tu­ itions, feelings and motivation may not necessarily track moral truth. I take it that most of us are much better in empathizing with humans than with mice. Similarly, I care more about some discomfort of my own child than about the death of some strangers. This subjective evaluation doesn’t necessarily track moral truth. If so, the counter-­intuitiveness of EQU in conjunction with normative views may not count against the truth of any of these theories. This line of argumentation is commonly brought forward in ­ethics. I think there is something to it: one’s intuitions, feelings, and motivations need not always be an infallible guide to moral action. On the other hand, I think that an ethical theory or demand that, even after careful consideration, remains completely out of touch with human intuitions, feelings and mo­tiv­ ations should not be trusted. But assume, alternatively, that there is no objective moral truth. In that case there is nothing to guide us except for our subjective (or inter-­ subjective) judgements, those that result from initial intuitions and re­flect­ ive scrutiny. In his fascinating book Humankind Rutger Bregman notices our tendency to care more about individuals who are like us, but also our great capacity for compassion. He claims, ‘evil does its work from a distance’ (Bregman 2020, 390). The more we come to know others, history shows time and again, the easier it is to feel compassion and the harder acting cruel­ly becomes. But, he says, it may be asking too much for us all to abandon our family and friends to help strangers more effectively:

142 

Capacity for Welfare across Species

As humans, we differentiate. We play favorites and care more about our own. That’s nothing to be ashamed of—­it makes us human. But we must also understand that those others, those distant strangers, also have fam­ ilies they love. That they are every bit as human as we are.

We are both partial and able of compassion, even beyond the border of our own species. Just like us, many non-­humans have their own character and desires and can experience frustration and fear, sadness, and happiness. At least to some extent we have the freedom to locate and relocate ourselves in the space of human possibility between partiality and compassion. Whether we arrive at that conclusion by moral reasoning or by empathy alone: it can hardly be denied that welfare matters. Unless and until we find more convincing arguments for discounting animal welfare, we need to take non-­human welfare much more seriously than we currently do. We must admit that all our argumentative moves to the contrary are weak. It is far from settled whether non-­human animals have a lower moral status than we do, that death is a lesser harm for them, or that they have a weaker capacity for welfare. So, we may have to take their premature deaths and suffering (as well as options of prolonging and improving their lives) much more seriously. This would require a change not only in our treatment of non-­humans, which is widely agreed to be disastrous anyway. It would, in addition, require a change regarding our ideals or aspiration. Thus, not only should we count the welfare of non-­humans in the first place; we should also stop discounting it.

Bibliography Allen, C. (2006). ‘Ethics and the Science of Animal Minds.’ Theoretical Medicine and Bioethics 27 (4): 375–394. Allen, C. (2011). ‘Fish Cognition and Consciousness.’ Journal of Agricultural and Environmental Ethics 26: 25–39. Alvarez, M. (2017). ‘Reasons for Action: Justification, Motivation, Explanation.’ In Zalta, Edward  N. (ed.), The Stanford Encyclopedia of Philosophy. https://plato. stanford.edu/archives/win2017/entries/reasons-just-vs-expl/. Bader, R. M. (2015). ‘The Neutrality of Existence.’ Unpublished manuscript and lecture presented at the Conference on Theoretical Population Ethics. Oxford University, 22 November 2015. Bader, R. M. (2022). ‘Person-Affecting Utilitarianism.’ In Arrhenius, G., Bykvist, K, Campbell, T., and Finneron-Burns, E. (eds). Oxford Handbook of Population Ethics, 251–270. Oxford: Oxford University Press. Bader, R. M. (2022). ‘The Asymmetry.’ In MacMahan, J., Campbell, T., Goodrich, J., and Ramakrishnan, K. (eds). Ethics and Existence: The Legacy of Derek Parfit, 15–37. New York: Oxford University Press. Balcombe, J. P. (2006). Pleasurable Kingdom: Animals and the Nature of Feeling Good. London: Macmillan. Bardo,  M.  T., Klebaur,  J.  E., Valone,  J.  M., and Deaton,  C. (2001). ‘Environmental Enrichment Decreases Intravenous Self-Administration of Amphetamine in Female and Male Rats.’ Psychopharmacology (Berlin) 155: 278–284. Barnard, C. (2007). ‘Ethical Regulation and Animal Science: Why Animal Behavior is Special.’ Animal Behavior 74: 5–13. Bateson, M. and Matheson, S.  M. (2007). ‘Performance on a Categorisation Task Suggests that Removal of Environmental Enrichment Induces “Pessimism” in Captive European Starlings (Sturnus Vulgaris).’ Animal Welfare 16 (1): 33–36. Birch, J., Burn, C., Schnell, A., Browning, H., and Crump, A. (2021). Review of the Evidence of Sentience in Cephalopod Molluscs and Decapod Crustaceans. London: LSE Consulting. Bracke, M. B. M. (2006). ‘Providing Cross-Species Comparisons of Welfare with a Scientific Basis.’ NJAS - Wageningen Journal of Life Sciences 54 (1): 61–75. Bradley, B. (2009). Well-Being and Death. Oxford: Oxford University Press. Bradley, B. (2016). ‘Is Death Bad for a Cow?’ In Višak, T. and Garner, R. (eds), The Ethics of Killing Animals, 51–64. New York: Oxford University Press. Bradshaw, G.  A., Schore, A.  N., Poole, J.  H., Moss, C.  J., and Brown, J.  L. (2005). ‘Elephant breakdown.’ Nature 433: 807. Braithwaite, V. (2010). Do Fish Feel Pain? Oxford: Oxford University Press. Brambell Committee (1965). ‘Report of the Technical Committee to Enquire into the Welfare of Animals kept under Intensive Livestock Husbandry Systems.’ Command Report 2836. London: Her Majesty’s Stationery Office.

144 Bibliography Bregman, R. (2020). Humankind—­A Hopeful History. London: Bloomsbury. Broom, D.  M. (1986). ‘Indicators of Poor Welfare.’ British Veterinary Journal 142: 520–526. Broom, D. M. (1988). ‘The Scientific Assessment of Animal Welfare.’ Applied Animal Behaviour Science 20: 5–19. Broom, D.  M. (1991). ‘Assessing Welfare and Suffering.’ Behavioural Processes 25: 117–123. Broom, D. M. (1993). ‘A Usable Definition of Animal Welfare.’ Journal of Agriculture and Environmental Ethics & Supplement 7: 15–25. Broom, D.  M. and Johnson, K.  G. (1993). Stress and Animal Welfare. London: Chapman & Halt. Broom, D. M. (1996). ‘Animal Welfare Defined in Terms of Attempts to Cope with the Environment.’ Acta Agriculturae Scandinavica. Animal Science Supplement 27 (3‒4): 22–28. Broom, D. (1998). ‘Welfare, Stress, and the Evolution of Feelings.’ Advances in the Study of Behavior 27: 371–403. Broome, J. (2019). ‘The Badness of Dying Early.’ In Gamlund, E. and Solberg, C.  T. (eds), Saving People from the Harm of Death, 105–115. Oxford: Oxford University Press. Brüne, M., Brüne-Cohrs, U., McGrew, W.  C., and Preuschoft, S. (2006). ‘Psychopathology in Great Apes: Concepts, Treatment Options and Possible Homologies to Human Psychiatric Disorders.’ Neuroscience & Biobehavioral Reviews 30: 1246–1259. Broome, J. (1999). Ethics out of Economics. Cambridge: Cambridge University Press. Broome, J. (2004). Weighing Lives. Oxford: Oxford University Press. Broome, J. (1998). ‘Extended Preferences.’ In Fehige, C. and Wessels, U. (eds), Preferences, 171–187. Berlin: De Gruyter. Broome, J. (2019). ‘The Badness of Dying Early.’ In Gamlund, E. and Solberg, C.  T. (eds), Saving People from the Harm of Death, 105–115. Oxford: Oxford University Press. Browning, H. (2019). ‘If I Could Talk to the Animals: Measuring Subjective Animal Welfare.’ Doctoral thesis, submitted to Australian National University, December 2019. Budolfson, M. and Spears, D. (2020). ‘Public Policy, Consequentialism, the Environment, and Non-human Animals.’ In Portmore, D.  W., The Oxford Handbook of Consequentialism, 592–615. New York: Oxford University Press. Budolfson, M. (2020). ‘Methods for Quantifying Animal Welfare, and Implications for Population Ethics.’ Invited lecture at the Workshop Animals and Climate Change, Institute for Future Studies, Stockholm, 4 December 2020. Burgdorf, J. and Panksepp, J. (2001). ‘Tickling Induces Reward in Adolescent Rats.’ Physiology & Behavior 72: 167–173. Cabanac, M. (1996). ‘On the origin of consciousness: a postulate and it corollary.’ Neuroscience and Biobehavioral Reviews 20: 33–40. Cabanac, M. (1992). ‘Pleasure: the common currency.’ Journal of Theoretical Biology 155: 173–200.

Bibliography  145 Cabanac, M. (2000). ‘What is happiness?’ Invited lecture at Fechner Day 2000, International Society of Psychophysics at Strasbourg, January 2000. https://www. researchgate.net/publication/339068662_WHAT_IS_HAPPINESS. Cabanac, M., Cabanac, A. J., and Parent, A. (2009). ‘The emergence of consciousness in phylogeny.’ Behavioral Brain Research 198: 267–272. Cabanac, M., Brondel, L., and Cabanac, A. (2000). ‘Alliesthesia. Up-Date of the Word and Concept.’ American Journal of Biomedical Science and Research 8 (4): doi: 10.34297/AJBSR.2020.08.001293. Cabanac, M. and Laberge, F. (1998). ‘Fever in goldfish is induced by pyrogens but not by handling.’ Physiology & Behavior 63: 377–379. Campbell, S.  M. (2016). ‘The Concept of Well-Being.’ In Fletcher, Guy (ed.), The Routledge Handbook of Philosophy of Well-Being, 402–413. New York: Routledge. Cohen, C. (1986). ‘The Case for the Use of Animals in Biomedical Research.’ The New England Journal of Medicine 315 (463): 865–870. Crisp, R. (2006a). Reasons and the Good. New York: Oxford University Press. Crisp, R. (2006b). ‘Hedonism reconsidered.’ Philosophy and Phenomenological Research 73 (3): 619–645. Crisp, R. and Kringelbach, M. (2018). ‘Higher and Lower Pleasures Revisited: Evidence from Neuroscience.’ Neuroethics 11: 211–215. Csikszentmihalyi, M. (1990). Flow. The Psychology of Optimal Experience. New York: Harper and Row. Curtis, S.  E. (1986). ‘The Case for Intensive Farming in Food Animals.’ In. Fox, M.  W. and Mickey, L.  D. (eds), Advances in Animal Welfare Science, 245–255. Washington D.C.: The Humane Society of the United States. Dawkins, M. S. (1990). ‘From an Animal’s Point of View: Motivation, Fitness, and Animal Welfare. Behavioral Brain Sciences 13: 1–9. Dawkins, M. S. (2017). ‘Animal Welfare With and Without Consciousness’, Journal of Zoology 301 (1): 1–10. Danbury, T.  C., Weeks, C.  A., Chambers, J.  P., Waterman-Pearson, A.  E., and Kestin, S. C. (2000). ‘Self-Selection of the Analgesic Drug, Carprofen, by Lame Broiler Chickens.’ Veterinary Record 146: 307–311. DeGrazia, D. (2015). ‘Moral Personhood and Moral Status: A Reply to Kagan’s Proposal.’ Journal of Applied Philosophy: doi: 10.1111/japp.12166. Dombrowski, D.  A. (1997). Babies and beasts: the argument from marginal cases. Champaign: University of Illinois Press. Dorsey, D. (2009). ‘Headaches, Lives and Value.’ Utilitas 21 (1): 36–58. Dorsey, D. (2017). ‘Idealization and the Heart of Subjectivism.’ Noûs 51 (1): 196–217. Duncan, I. J. H. (1987). ‘The Welfare of Farm Animals: An Ethological Approach.’ Science Progress 71 (3): 317–326. Duncan, I.  J.  H. (1993). ‘Welfare is to do with what Animals Feel.’ Journal of Agricultural and Animal Ethics 6, Supplement 2: 8–14. Duncan, I.  J.  H. (1996). ‘Animal Welfare Defined in terms of Feelings.’ Acta Agriculturae Scandinavica. Section A, Animal Science, Supplementum 27: 29–35. Elwood, R.  W. (2011). ‘Pain and suffering in Invertebrates?’ ILAR Journal 52: 175–184.

146 Bibliography Engh, A. E., Beehner, J. C., Bergman, T. J., Whitten, P. L., Hoffmeier, R. R., Seyfarth, R. M., and Cheney, D. L. (2006). ‘Female Hierarchy Instability, Male Immigration, and Infanticide Increase Glucocorticoid Levels in Female Chacma Baboons.’ Animal Behaviour 71: 1227–1237. Fehige, C. (1998). ‘A Pareto Principle for Possible People?’. In Fehige, C. and Wessels, U. (eds), Preferences, 508–543. Berlin: De Gruyter. Fletcher, G. (2013). ‘A Fresh Start for the Objective List Theory of Well-Being.’ Utilitas 25 (2): 206–220. Fraser, D., Weary, D.  M., Pajor, E.  A., and Milligan, B.  N. (1997). ‘A Scientific Conception of Animal Welfare that Reflects Ethical Concerns.’ Animal Welfare 6: 187–205. Frey, R. G. (1996). ‘Response: Autonomy, Animals, and Conceptions of the Good.’ Between the Species Winter & Spring 1996, 8–14. Ginsburg, S. and Jablonka, E. (2019). The Evolution of the Sensitive Soul: Learning and the Origins of Consciousness. Cambridge: MIT Press. Greaves, H. (2019). ‘Against “the Badness of Death” .’ In Gamlund, E. and Solberg, C.  T. (eds), Saving People from the Harm of Death, 189–202. Oxford: Oxford University Press. Griffin, J. (1986). Well-Being: Its Meaning, Measurement and Moral Importance. Oxford: Oxford University Press. Harding, E. J., Paul, E. S., and Mendl, M. (2004). ‘Animal Behavior—­Cognitive Bias and Affective State.’ Nature 427: 312–1312. Harrison, R. (2013). Animal Machines: The New Factory Farming Industry. First published 1964. Wallingford: CABI Publishing. Haybron, D.  M. (2011). ‘Central Park: Nature, Context and Human Wellbeing.’ International Journal of Wellbeing 1 (2): 235–254. Haybron, D. M. (2008a). The Pursuit of Unhappiness. The Elusive Psychology of WellBeing. New York: Oxford University Press. Haybron, D.  M. (2008b). ‘Why Hedonism Is False.’ In Cahn, S.  M. and Vitrano, C. (eds), Happiness. Classic and Contemporary Readings in Philosophy, 173–179. New York: Oxford University Press. Haybron, D.  M. (2008c). ‘Happiness, the Self and Human Flourishing.’ Utilitas 20 (1): 21–41. Heathwood, C. (2011). ‘Preferentialism and Self-Sacrifice.’ Pacific Philosophical Quarterly 92 (1): 18–38. Hooker, B. (2000). Ideal Code, Real World. New York: Oxford University Press. IPCC (2018). ‘Summary for Policymakers.’ In: Global Warming of 1.5°C. An IPCC Special Report on the Impacts of Global Warming of 1,5°C Above Pre-Industrial Levels and Related Glabal Greenhouse Gas Emission Pathways, in the Context of Strengthening Global Response to the Threat of Climate Change, Sustainable Development, and Efforts to Eradicate Poverty. https://www.ipcc.ch/2018/10/08/ summary-for-policymakers-of-ipcc-special-report-on-global-warming-of-1-5capproved-by-governments/, accessed 11.4.2021. Jones, R. C. (2013). ‘Science, Sentience, and Animal Welfare.’ Biology & Philosophy 28 (1): 1–30.

Bibliography  147 Kagan, S. (2019). How to Count Animals, More or Less. New York: Oxford University Press. Kagan, S. (2015). ‘What’s Wrong with Speciesism?’ Journal of Applied Philosophy 33 (1): 1–21. Kahane, G. and Savulescu, J. (2012). ‘The Concept of Harm and the Significance of Normality.’ Journal of Applied Philosophy 29 (4): 318–332. Key, B. (2016). ‘Why Fish Do Not Feel Pain.’ Animal Sentience 1 (3): 1–34. Kitcher, P. (2015). ‘Experimental Animals.’ Philosophy and Public Affairs 43 (4): 287–311. Klocksiem, J. (2012). ‘A Defense of the Counterfactual Comparative Account of Harm.’ American Philosophical Quarterly 49 (4): 285–300. Klint Jensen, K. (2008). ‘Millian Superiorities and the Repugnant Conclusion.’ Utilitas 20 (3): 279–300. Kraut, Richard (2007). What is Good and Why. Cambridge, Mass.: Harvard University Press. Kumar, R. (2008). ‘ “Permissible Killing and the Irrelevance of Being Human.” ’ The Journal of Ethics 12: 57–80. Lin, Eden (2014). ‘Pluralism About Well-Being.’ Philosophical Perspectives 28 (1): 127–154. Lin, E. (2017). ‘Against Welfare Subjectivism.’ Noûs 51 (2): 354–377. Low, P., Panksepp, J., Reiss, D., Edelman, D., van Swinderen, B., and Koch, C. (2012). ‘The Cambridge Declaration on Consciousness’. Presented at the Francis Crick Memorial Conference, Cambridge, England. https://fcmconference.org/img/ CambridgeDeclarationOnConsciousness.pdf. MacAskill, W. (2019). ‘The Definition of Effective Altruism.’ In Greaves, H. and Pummer, T. (eds), Effective Altruism: Philosophical Issues, 10–28. New York: Oxford University Press. Matheson, S.  M., Asher, L., and Bateson, M. (2008). ‘Larger, Enriched Cages are Associated with “Optimistic” Response Biases in Captive European Starlings (Sturnus vulgaris).’ Applied Animal Behaviour Science 109: 374–383. McFarland, D.  J. and Sibly, R.  M. (1975). ‘The Behavioral Final Common Path.’ Philosophical Transactions of the Royal Society, Biological Sciences 270: 265–293. McGlone, J. (1993). What is Animal Welfare? Journal of Agriculture and Environmental Ethics 6, Supplement 2: 26–26. McMahan, J. (2002). The Ethics of Killing. New York: Oxford University Press. McMahan, J. (2008). ‘Eating Animals the Nice Way.’ Daedalus Winter 2008, 137 (1): 66–76. McMahan, J. (1996). ‘Cognitive Disability, Misfortune, and Justice.’ Philosophy and Public Affairs 25 (1): 3–35. McMahan, J. (2008). ‘Challenges to Human Equality.’ The Journal of Ethics 12: 81–104. Meacham, Christopher J. G. (2012). Philosophical Studies, 158: 257–287. Mellor, D. J. and Beausoleil, N. J. (2015). ‘Extending the “Five Domains” Model for Animal Welfare Assessment to Incorporate Positive Welfare States.’ Animal Welfare 24 (3): 241–253.

148 Bibliography Mendl, M. (1990). ‘Developmental Experience and the Potential for Suffering: Does “Out of Experience” Mean “Out of Mind”?’ Behavioral and Brain Sciences 13: 28–29. Mill, John Stuart (1863/2001). Utilitarianism. Kitchener, Ont.: Batoche Books. Mill, J. S. (1998). Utilitarianism, Crisp, R. (ed.). Oxford: Oxford University Press. Mill, J. S. (1969). Utilitarianism. In Robson, J. (ed.), Collected Works, vol. 10. Toronto: University of Toronto Press. Norcross, A. (2004). ‘Puppies, Pigs and People: Eating Meat and Marginal Cases.’ Philosophical Perspectives 18 (1): 129–245. Ohl, F. and Van der Staay, F.  J. (2012). ‘Animal Welfare: At the Interface Between Science and Society.’ Veterinary Journal 192 (1): 13–19. Panksepp, J. (2004). Affective Neuroscience: The Foundations of Human and Animal Emotions. New York: Oxford University Press. Panksepp,  J., and Burgdorf,  J. (2003). ‘ “Laughing” Rats and the Evolutionary Antecedents of Human Joy.’ Physiology & Behavior 79: 533–547. Phillips, D. (2011). Sidgwickian Ethics. New York: Oxford University Press. Parfit, D. (2017). ‘Future People, the Non-Identity Problem, and Person-Affecting Principles.’ Philosophy and Public Affairs 45: 2. Parfit, D. (1991). ‘Equality or Priority?’ In Clayton, M. and Williams, A. (eds), The Ideal of Equality, 81–125. New York: Palgrave. Parfit, D. (1984). Reasons and Persons. New York: Oxford University Press. Pluhar, E. (1996). Beyond Prejudice: The Moral Significance of Human and Nonhuman Animals. Durham: Duke University Press. Proctor, H. (2012). ‘Animal Sentience: Where are We and Where are We Heading?’ Animals 2 (4): 628–639. Rachels, J. (1990). Created from Animals: The Moral Implications of Darwinism. Oxford: Oxford University Press. Railton, P. (2008). ‘The Problem of Well-Being: Respect, Equality, and the Self.’ Lecture presented to the International Society for Utilitarian Studies, Berkeley, October 2008. Rayo, L. and Becker, G. S. (2007). ‘Evolutionary efficiency and happiness.’ Journal of Political Economy 115 (2): 302–337. Regan, T. (1979). ‘An Examination and Defense of One Argument Concerning Animal Rights.’ Inquiry 22: 189–219. Riley, J. (2008). ‘What are Millian Qualitative Superiorities?’ Prolegomena 7 (1): 61–79. Rollin, B. (1989). The Unheeded Cry: Animal Consciousness, Animal Pain, and Science. New York: Oxford University Press. Sandøe, P. (1996). ‘Animal and Human Welfare – Are they the Same Kind of Thing?’ Acta Agriculturae Scandinavica. Section A, Animal Science, Supplementum 27: 11–15. Sarch, A. (2011). ‘Internalism about a person’s good: don’t believe it.’ Philosophical Studies 154: 161–184. Scanlon, T.  M. (1991). ‘The moral basis of interpersonal comparisons.’ In Elster, J. and Roemer, J.  E. (eds), Interpersonal Comparisons of Well-Being, 17–44. Cambridge: Cambridge University Press. Schmidtz, D. (1998). ‘Are all Species Equal?’ Journal of Applied Philosophy 15 (1): 61.

Bibliography  149 Sen, A. (1979). ‘Utilitarianism and Welfarism.’ The Journal of Philosophy 76 (9): 463–489. Sherwin, C. M. (2001). ‘Can Invertebrates Suffer? Or, How Robust is Argument-byAnalogy?’ Animal Welfare 10: 103–118. Sherwin,  C.  M. and Olsson,  I.  A.  S. (2004). ‘Housing Conditions Affect SelfAdministration of Anxiolytic by Laboratory Mice.’ Animal Welfare 13: 33–39. Sidgwick, H. (1907/1981). The Methods of Ethics (7th ed.). London: Macmillan. Singer, P. and Lazari-Radek, K. (2014). The Point of View of the Universe. Sidgwick and Contemporary Ethics. New York: Oxford University Press. Singer, P. (2011). Practical Ethics (3rd ed.). New York: Cambridge University Press. Singer, P. (1995). Animal Liberation (2nd ed.). London: Pimlico. Siviy, S.  M., Fleischhauer, A.  E., Kerrigan, L.  A., and Kuhlman, S.  J. (1996). ‘D2 Dopamine Receptor Involvement in the Rough-and-Tumble Play Behavior of Juvenile Rats.’ Behavioral Neuroscience 110: 1–9. Sneddon, L. U., Elwood, R. W., Adamo, S. A., and Leach, M. C. (2014). ‘Defining and Assessing Animal Pain.’ Animal Behavior 97: 201–212. Stoskopf,  M.  K. (1994). ‘Pain and Analgesia in Birds, Reptiles, Amphibians, and Fish.’ Investigative Ophthalmology & Visual Science 35 (2): 775–80. Sumner, L. W. (1996). Welfare, Happiness and Ethics. New York: Oxford University Press. Steinbock, B. (1978). ‘ “Speciesism and the Idea of Equality.” ’ Philosophy 53 (204): 247–256. Temkin, L. (2003). ‘Egalitarianism Defended.’ Ethics 113: 764–782. Tremblay, L.  and Schultz, W. (1999). ‘Relative Reward Preference in Primate Orbitofrontal Cortex.’ Nature 398: 704–708. Višak, T. (2018a). ‘Engineering Life Expectancy and Non-Identity Cases.’ Journal of Agricultural and Environmental Ethics 31 (2): 281–293. Višak, T. (2018b). ‘Argument der Grenzfälle.’ In Ach, J.  S.  and Borchers, D.  (eds), Handbuch Tierethik, 149–154. Stuttgart: Metzler Verlag. Višak, T. (2017). ‘Cross-Species Comparisons of Welfare.’ In Woodhall, A. and Garmendia de Trindade, G. (eds), Ethical and Political Approaches to Non-Human Animal Issues, 347–364. New York: Palgrave MacMillan. Višak, T. (2016). ‘Do Utilitarians Need to Accept the Replaceability Argument?’ In Višak, T., and Garner, R. (eds), The Ethics of Killing Animals, 117–135. New York: Oxford University Press. Višak, T. (2014). Review of The Point of View of the Universe. Sidgwick and Contemporary Ethics [Lazari-Radek  K.  and Singer, P., New York: Oxford University Press, 2014]. Ethical Perspectives 21 (3): 469–471. Višak, T. (2013). Killing Happy Animals. Explorations in Utilitarian Ethics. New York: Palgrave MacMillan. Višak, T. and Balcombe, J. (2013). ‘The Applicability of the Self-fulfilment Account of Welfare to Non-human Animals, Babies and Mentally Disabled Humans.’ Philosophy and Public Policy Quarterly 31 (2). Watanabe, M. (1999). ‘Neurobiology: Attraction is Relative Not Absolute.’ Nature 398: 661–663. Webster, J. (1994). Animal welfare: A cool eye towards Eden. Oxford: Blackwell Science.

150 Bibliography Wilson, J. (2021). ‘Determinables and Determinates.’ In Zalta, Edward N. (ed.), The Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/archives/spr2021/ entries/determinate-determinables/. Wong, K. (2016). ‘Counting Animals: On Effective Altruism and the Prospect of Interspecies Commensurability.’ PhD thesis, Princeton University. White, A. (1984). Rights. New York: Oxford University Press. Reprinted in Regan, Tom and Singer, Peter (eds), Animal Rights and Human Obligations (2nd ed.), 120 (Prentice Hall, 1989). Williams, B. (2009). ‘The Human Prejudice.’ In Schaler, J. A. (ed.), Peter Singer Under Fire: The Moral Iconoclast Faces His Critics, 77–96. Illinois: Carus Publishing Company. Woodard, C. (2019). Taking Utilitarianism Seriously. New York: Oxford University Press. Yeates, J. (2013). ‘How Happy Does a Happy Animal Have to Be (and How Can We Tell)?’ RSPCA Australia Scientific Seminar, Melbourne, 27 February 2013, Melbourne University. Zeaman, D. (1949). ‘Response Latency as a Function of the Amount of Reinforcement.’ Journal of Experimental Psychology 39: 466–483. Zimmerman, Michael (2015). ‘Intrinsic Versus Extrinsic Value.’ In Zalta, E. N. (ed.), The Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/valueintrinsic-extrinsic/.

Index For the benefit of digital users, indexed terms that span two pages (e.g., 52–53) may, on occasion, appear on only one of those pages. activism 2 animal husbandry  11, 64, 86–7, 99–111 animal welfare science (see also: measuring welfare)  11, 60–5 asymmetry 136–7 attitude (in-)dependence of welfare  15–16 axiology (see also: population axiology)  11, 18–19 Bader, R.  130–7 basic versus instrumental goods  13, 22–3, 25–8, 65, 71 birds  5, 54–6, 67–8 Bracke, M.  11–12 Browning, H.  7 Budolfson, M. and Spears, D.  3–4, 8 Cabanac, M.  75–9, 83–5 conceptualization of welfare  11, 15 consumption of animals  2, 99–111 common currency  75–6, 78 common-sense approach to capacity for welfare 8–13 comparability of welfare across species  6–9 cognitive capacities  8–11, 25–8 Crisp, R.  31–6 deontology 18 deprivation view  90–2 dual-source views  17–18 egalitarianism  3, 18, 97–8, 108, 111, 113–14, 124, 129 egoism 17–18 emotional capacities  10–12, 36–7, 95–6, 123 fish 5 friendship (see: social relationships) frustration view  89–91

genetic engineering  99–111 Greaves, H.  93–5 happiness  66–9, 77–8, 83–5 Haybron, D.  65–74, 83–5 hedonic capacities  10, 40–59, 74–85 hedonicity scale  74–85 hedonism  4–5, 13–14 homeostasis 75–8 impersonal total view (see: total view) indicators of welfare  6–7 insects  5, 112 insects-and-climate-change argument  112 instrumental goods (see: basic versus instrumental goods) invertebrates (see also: insects) 5 joy 77–8 Kagan, S.  8–11, 21–8, 113–30 mammals 5 McMahan, J.  36–40, 92–6, 99–111 measurement of welfare  6–7 metaphysics  6, 92–3 modal personism  118–20, 126–9 native potential for welfare  38–9 neurons (see: neuroscience) neuroscience  10, 36, 56–9, 83, 113–14 nirwana-happiness 83–5 nociception 5 non-comparativism  134, 136 non-identity problem  133–4 normative reasons for action (see: reasons for action) pain  5, 67–8, 75, 127–9 Parfit, D.  34, 105–6, 133

152 Index people (see persons) person-affecting view  102–6, 130–7 persons  9, 126–7 pleasure 74–85 population axiology  102–7, 130–7 potential 126–9 preferentialism  6, 14, 28–31, 47 prejudice 10–11 principle-agent problem  79 prioritarianism  122–4, 130–2 Rayo, L. and Becker, G.  79–81 reasons for action  16–19 relationship  22–4, 27–8, 53, 119 relativizing  38, 41, 65–6, 79–80, 85, 98–9 relativize-to-natural-lifespan view  98–9 relevance of comparisons of welfare across species  1–4, 16 replaceability argument  103–6, 133 repugnant conclusion  133, 135–6

sentience  4–5, 51 Sidgwick, H.  32, 35 Singer, P.  28–31, 53–4, 57, 89–91, 116–18 Subjectivism  90–1, 121 subjects of welfare  4–5, 29–30, 70–1 sufficientarianism  18, 124, 129 Sumner, W.  7–8, 70 theory Q  102–3, 105–6, 130–1 time-relative-interest account  92–6 total view  102–4, 106, 130–7 TRIA (see: time-relative-interest account) utilitarianism  3, 17–18, 131–2 value theory (see axiology) welfarism 17–18