Brief summary: whether someone’s aggregate welfare is positive or negative is morally relevant. If welfare is negative, life is not worth living and it would be better not to exist (or better to be euthanized). However, especially when it comes to non-human animals, it is difficult to tell whether their lives have an aggregate positive or negative welfare. What if we cannot determine whether farm animals or wild animals have a positive or negative welfare? Einstein’s theory of special relativity is used as an analogy to understand the possible indeterminacy of the sign of someone’s welfare. The level of someone’s welfare depends on the welfare frame, but there does not exist an absolute welfare frame. This indeterminacy or relativity of welfare has implications for our altruistic choices to do good. In particular I argue that animal welfare relativity implies farm animal abolitionism (abolishing animal farming with new food technologies such that no new farm animals are born) and wild animal welfarism (helping nature with new technologies to increase the welfare levels of wild animals). Farm animal welfarism and wild animal abolitionism would be ethically more risky.

Introduction

Doing good, in an altruistic sense, means doing what other individuals want (benefiting others by improving their welfare), and not doing what they do not want (not harming others by not decreasing their welfare). The former is called positive altruism, the latter negative altruism. In familiar cases, respecting positive altruism automatically implies respecting negative altruism, and violating negative altruism implies violating positive altruism. However, the two kinds of altruism become uncoupled in population ethics, when our choices determine the individuals who will exist in the future. Adding an individual with a negative welfare (i.e. a life not worth living, whereby the individual has stronger negative than positive experiences or would have preferred non-existence) is always bad, and it can be said that the individual is harmed when brought into existence. But bringing into existence an individual with a positive welfare (a life worth living) is not necessarily always good, according to asymmetric accounts of altruism. Not bringing into existence a person with a positive welfare does not harm that person, because that person does not exist.

When deciding whether bringing an individual into existence is permissible, a crucial question becomes: when does a life become worth living? This question becomes relevant when we study the problems of farm animal and wild animal suffering. Is the life of an animal born in a factory farm worth living? Does a wild animal have more negative than positive experiences such that one would prefer not being born as that wild animal? People who reflect on the lives of farm and wild animals, often believe that many of those animals have lives not worth living, i.e. with negative welfare. But uncertainties remain, especially for wild animals.

In this article, I use Einstein’s theory of special relativity as an analogy to understand the issue in population ethics. It is possible that a welfare level around zero is not absolute in the sense of well-defined for everyone. Welfare could become relative.

When does a life become worth living?

In my personal situation, I can say that my life is worth living, that it has a positive welfare level. I can also imagine a situation in great pain and suffering, where I would rather be dead. Hence, somewhere there should be a zero level of welfare. This can be determined with a rather abstract thought experiment.

Suppose there are three situations. Situation A is the current situation, situation B is the same as A, except that my brains will change in such a way that I will have a lower welfare level, but my behavior remains exactly the same as in A. No-one is able to tell that my welfare is low (when asked how high it is, I will lie and say that it is as high as in situation A). Hence, my lowered well-being will not induce feelings of compassion in other people. Everyone except me remains equally happy as in situation A. Finally, situation C is the same as situation A, except I will be a mental zombie. That means I will do the same things as in situation A, but I will not have any conscious experiences. As a zombie, I am mentally dead or non-existent, but no-one will know whether I am mentally dead. Hence, no-one will grief about my death. Everyone except me remains as happy as in A.

When in situation B, my welfare is lowered, at one point I will become indifferent between situations B and C. I will become indifferent between keeping on living a conscious life in situation B, or mental euthanasia, becoming a zombie in situation C. The welfare level in situation B, when I become indifferent, is my level zero, neutral reference point. (Note that this level zero is relative to situation A. When situation A changes, for example when a loved one starts suffering, my zero welfare level could change.)

The above thought experiment is very abstract, and hence some individuals might not be able to perform it. This is in particular the case for non-human animals. For those animals, two things are possible. Either they are inherently unable to determine their zero levels of welfare, as if their zero levels are not well-defined. Or they do have well-defined zero levels, but are not able to tell them. Even scanning their brains with the best brain scanning technologies, we remain uncertain about the true zero welfare levels of animals. Hence, we are able to tell whether an animal in situation X is happier than in Y, but not able to tell whether the welfare levels in X and Y are higher or lower than zero, i.e. higher or lower than in situation Z where the animal is a mental zombie. Only the differences in welfare count, not their absolute levels.

How should we deal with this indeterminacy or uncertainty about the zero level? The theory of special relativity could shed some light on this issue. So let us briefly digress and look at this strange theory.

Simultaneity in special relativity

Consider five events: nodding my head, stamping of my left foot, stamping of my right foot, snapping of my left fingers and snapping of my right fingers. These events are well located in space and happen at a specific time, and hence can be described as five space-time points, denoted by P0, P-, P+, P’ and P’’. Now the question is: which of those events happen simultaneously? The notion of simultaneity is problematic according to special relativity.

Let’s say that event P0, the nodding of my head, happens at time zero (i.e. time zero is defined in this way). This is the default time with respect to which we want to determine whether other events happen in the future or the past of P0. Suppose event P-, stamping of my left foot, happens before P0. How can we know that event P- happens at a negative time? When I stamp my left foot, light signals can travel from my foot to my head at the speed of light. Crucially, the speed of light is finite, and all observers measure the same speed of light. Hence, the speed of light is absolute. If the light reaches my head before I nod my head, then stamping my left right foot (event P-) happens in the past of time zero, i.e. before P0, for all observers. Everyone agrees that P- occurs before P0, so this time difference is absolute. Technically: in a space-time diagram, the point P0 has a future and a past light cone (light moving away or towards space-time point P0), and points within the past light cone are in the past of P0 (i.e. an earlier time according to all observers).

Similarly, when at the moment that I nod my head, light starts to travel from my head to my right foot and reaches the foot before I stamp that foot, the third event P+ happens after P0 according to all observers. Event P+ is in the future light cone of P0 and is registered at a positive time by all observers.

Now consider light traveling from my head to my left hand. When event P’, the snapping of my left fingers, happens before the light reaches my left hand, different observers will disagree whether or not P’ happens before or after P0. In this case, P’ is outside of P0’s light cone, and it is possible to choose a reference frame (e.g. measured relative to a non-accelerating rocket) such that P0 and P’ have the same time coordinate and hence are simultaneous space-time events: the person in the rocket sees the two events happening at the same time according to his personal clock. According to this reference frame, both P0 and P’ happen at time zero.

Finally, suppose that event P’’ (snapping of my right fingers) is in the future light cone of event P’, which means that light from my left hand to my right hand reaches the right hand before the right fingers snap. Then everyone agrees that event P’’ happens after event P’. If P’ happens at time zero, P’’ happens at a later, positive time.

Now we can change the reference frame, by considering a second rocket moving at a very high constant speed relative to the first rocket. This relative speed is close to the speed of light. (Technically, this change in reference frame is a Lorentz transformation of four-dimensional space-time, which is somewhat comparable to a rotation in three-dimensional space: the Lorentz transform ‘mixes’ or ‘interchanges’ a spatial and a temporal dimension, just like a rotation ‘mixes’ two spatial dimensions.) According to the person in the second rocket, P0 and P’’ happen simultaneously. The clock of this observer measures event P’ happening at a negative time, i.e. before event P0 which we defined as time zero.

The time when something happens, is relative in special relativity. There are two observers, one in each rocket. According to the first coordinate system (the reference frame of the first observer), t measures the time of event P’, the snapping of my left fingers, which was zero. After a Lorentz transformation, the time t’ in the new coordinate system becomes negative, which means that the snapping of my left fingers happens before the nodding of my head according to the observer in the second rocket. Hence time zero for events P’ and P’’ is not absolute or well-defined for all observers.

Special relativity is a bit counterintuitive, because we are used to transitivity: when A is bigger than B and B bigger than C, than A is bigger than C. Or when A is equal to B and B is equal to C, than A and C are equal. However, this latter kind of transitivity is not valid in special relativity: P’ is simultaneous to P0 (according to the reference frame of the first rocket), P0 is simultaneous to P’’ (according to the reference frame of the second rocket), but P’’ happens later than P’ (according to all reference frames). This intransitivity does not make space-time inconsistent or irrational, neither does it make the notion of time incomprehensible.

Relativistic welfare ethics

Now let’s move again to welfare ethics. Just like different observers measure time in special relativity, different observers can estimate the welfare level of an animal, and these estimates can differ. The welfare of the life of a person w is analogous to t/x in special relativity, where t is the time and x is the spatial distance (e.g. the distance between hands and feet). The welfare can be positive or negative, just like time can be positive or negative. However, the welfare value is not always absolute: it can be positive according to one welfare frame and negative according to another, just like time can be positive or negative according to different reference frames. The welfare frame allows to measure welfare, and is analogous to the reference frame or coordinate system in special relativity that allows to measure spatial distances and time intervals. A change in welfare frame corresponds to a change in reference frame, which is represented by a Lorentz transformation. This Lorentz transformation for the welfare between different welfare frames can be written mathematically as: w’ =(w-v/c²)/(1-vw), with w the welfare according to the first welfare frame, w’ the welfare according to the second welfare frame, v a parameter that measures the change between the welfare frames, and c a constant. (The derivation goes as follows: w’ is analogous to t’/x’, with t’ and x’ the time and space coordinates in the new reference frame or coordinate system. v is now interpreted as the velocity of the second rocket relative to the first rocket, and c is the constant speed of light. Using the Lorentz transformation, we can write w’=t’/x’ in terms of w=t/x.) This equation is valid for welfare levels w between -1/c and +1/c, and the parameter v is between -c and +c. Note that if w=1/c, then w’=w.

According to the first welfare frame, the welfare w of an animal can be zero, whereas a change in welfare frame gives a welfare w’ lower than zero. Hence, the zero welfare level of an animal is not unique or well-defined. When the welfare level w is between 1/c and +1/c, the sign of the welfare level is relative, i.e. dependent on the welfare frame. On the other hand, when w is very high (higher than +1/c), the sign is absolute: all welfare frames give a positive welfare. That means everyone agrees that the individual has a positive welfare.[1]

The constant c (the speed of light in special relativity) measures how well-defined a zero welfare is. For me, doing the above abstract thought-experiment is possible, which means a zero welfare is rather precise and well-defined, which means for me c is very large, almost infinite. But for a non-human animal, c can be finite and small. When c is small, the animal has a wide welfare range between -1/c and +1/c, and within this range a zero welfare is not well-defined. For example, when an animal has welfare w=0, we cannot conclude that this animal is indifferent between having a life at this welfare level and non-existence, because we can change the welfare frame such that the animal has a welfare w’ anywhere between -1/c and +1/c. Similarly, the welfare of an animal with welfare w=0 cannot be compared to the welfare of another animal with welfare w’=1/2c, because there is no absolute zero level.

Implications for altruism and animal ethics

If a zero welfare level is not clearly determined, at least for animals, we cannot always say whether the existence of an animal is good or bad. The existence would be bad if the welfare is negative, but if the welfare is higher than -1/c, after a Lorentz transformation (a change in welfare frame) we can get a positive welfare, which means the existence is no longer bad. Asking the animal which welfare frame (reference frame) she prefers, is impossible: the animal is not able to answer this question. In a sense, the animal is indifferent between the different welfare frames.

For simplicity, suppose we can choose between five options C0, C+, C-, C’ and C’’, just like we discussed five space-time events in the special relativity example above. Option C0 is the situation where no animal exists. This option serves as the default option, just like P0 defined the default time zero. In the other situations, animals exist. When animals are clearly better-off in C+ than in C0, the animals in C+ have a positive welfare w+. Choosing C+ respects positive altruism: it can be considered as a kind of help whereby we do something that someone else (the existing animal) wants. Similarly, when animals are worse-off in C- than in C0, they have a negative welfare w- in C-. Choosing C- means doing something bad and violating negative altruism: we do something that someone else does not want. The tricky issue is: there may be choices C’ and C’’ in which animals exist who does not exist in C0, and we cannot simply say that the animals are worse-off or better-off in C’ or C’’ than in C0.

Let us first consider farm animals and then wild animals. We can choose between five options: C0 is no animal farming, which involves e.g. eating vegan, producing cell-based meat,… This option means not breeding and using animals. If we do not exist, option C0 is automatically chosen, because we will not breed and use animals. Hence, C0 is our default option.

Option C+ means bringing fully happy animals in existence. These animals are not used for animal farming: the animals are not confined and slaughtered, but instead are cared for as much as possible such that they lead happy, healthy long lives. These animals are given everything they want and are not used for our purposes. Suppose that everyone agrees that these animals have a positive welfare. In that case, those happy animals have welfare levels above 1/c.

Option C- means animal farming in factory farms with current, minimal animal welfare legislation. This implies breeding, confining and slaughtering the animal for food. Most people strongly agree that animals in factory farms have very negative welfare levels. Because of this strong agreement, we can say that those animals have welfare levels likely below -1/c.

Option C’ is animal farming with more animal welfare laws, i.e. breeding and slaughtering animals, but taking more care of them, avoiding extreme confinements,…. Suppose we cannot determine whether these animals have positive or negative welfare. These farm animals have a welfare between -1/c and 1/c.

Finally, option C’’ is like C’, but with one extra measure that benefits the farm animals, such that the animals in C’’ are slightly happier than in C’.

If we choose option C-, we clearly violate negative altruism, because we do something that the animal does not want. First, the animal has a negative welfare, and second, the animal always prefers any other option above C-. The welfare of the animal is lowest in option C- according to all welfare frames, just like all observers agree that event P- happens at a negative time.

If we choose option C+, we clearly respect positive altruism: we do something that the animal in situation C+ prefers: an existence with the highest, positive welfare.

The cases for C’ and C’’ are mixed. There exist welfare frames according to which choosing C’ or C’’ violate negative altruism, because the welfare levels in those situations are negative. Furthermore, C’ and C’’ involve animal farming, whereby the animal always has to do or undergo unwanted things, such as being slaughtered or confined. Hence choosing C’ or C’’ does not guarantee respecting negative altruism. Neither does it guarantee respecting positive altruism.

Here we see a difference between farm animal abolitionism versus farm animal welfarism. The former involves choosing C0, the latter means choosing C’’. For example, if we become vegan ourselves, we always directly respect negative altruism because we do not choose situations C-, C’ or C’’ that someone else (the animal) does not always want. If we support research and development of animal free livestock products such as cell-based meat, we indirectly respect negative altruism, because with these animal free foods we allow other people (e.g. future living meat eaters) to respect negative altruism by avoiding animal farming. The development of new food technologies such as cell-based meat is an abolitionist strategy.

Next to abolitionism, some animal advocates prefer the welfarist approach. This approach basically means choosing option C’’: animal farming with animals that have higher welfare levels than the farm animals in situations C- and C’. However, this welfarist approach is neither always respecting negative nor positive altruism, which means that the welfarist approach is weaker in terms of altruism.

Furthermore, the welfarist approach faces a non-identity problem. Suppose a new farm animal welfare law is in place, which means animal farming changes from situation C’ to C’’. From that moment, new farm animals are born who automatically have welfare w’’. Without that law, other animals would have been born who would have had welfare levels w’ in situation C’ or w- in situation C-. The question is: did we help the animal in situation C’’ by choosing C’’? If situations C- or C’ were chosen, farm animals would be worse of, but they will not be the same farm animals who exist in situation C’’. The farm animals in situation C’’ do not exist in situations C0, C- and C’. If the animals in situation C’’ were helped, they by definition have a higher welfare level than in the situations without the help. But in the situations without help (i.e. situations C-, C’ or C0 without the new welfare law), those animals do not exist. And the welfare w’’ is incomparable to the zero welfare level that corresponds to the welfare level of non-existence. Hence, with these considerations, it becomes difficult to claim that the animals in C’’ were really helped. The farm animal welfarist approach is not clearly altruistic.

Finally, we move towards wild animal welfare. Again, we can consider five situations D0, D-, D+, D’ and D’’. Suppose animals in the wild live in situation D’ with welfare level w’. In that situation we cannot say whether their lives are worth living or not, because w’ is incomparable to the welfare w0 in situation D0 which contains no wild animals. The latter welfare level w0 equals 0.

Just like the abolitionist choice of C0 in the case of animal farming, we can choose to respect negative altruism by choosing D0 for wild animals. However, there is an important difference between the cases of farm animals and wild animals. Choosing D0 instead of D’ for wild animals typically means interventions in nature that could decrease the welfare of already existing wild animals. The transition from D’ to D0 means that the welfare of animals existing in D’ drops below w’. This transition could for example involve killing all wild animals such that no new wild animals can be born. When an animal has a clearly negative welfare, as in situation D-, killing that animal (i.e. euthanasia) would be helpful and good. But when the animal has a welfare w’ higher than -1/c in situation D’, the welfare is only negative in some welfare frames, which means killing that animal is not clearly permissible and may very well be harmful.

In contrast, choosing C0 for farm animals means stopping breeding farm animals. The latter does not generally decrease the welfare of existing farm animals. In other words, the transition from C’ to C0 does not decrease the welfare of farm animals existing in C’. Hence, choosing C0 for farm animals respects negative altruism (it is not harmful for existing farm animals), but choosing D0 for wild animals not necessarily respects negative altruism (it may be harmful for existing wild animals).

Therefore, when it comes to wild animal welfare, the welfarist approach of changing the situation from D’ to D’’ (or ideally D+) becomes more worth considering. That change from D’ to D’’ means helping nature to improve the welfare of wild animals. It is possible that the animals in D’’ would not have existed in D’, which means choosing D’’ does not involve positive altruism (the animals in D’’ are not really helped, just like the farm animals in C’’ were not really helped). However, the transition from D’ to D’’ does not necessarily involve a strong decrease in welfare of the existing animals in D’. That means choosing D’’ instead of D’ could respect negative altruism more strongly than choosing D0 instead of D’. And with sufficient research in welfare biology to improve the welfare of wild animals, it may become possible to choose a situation D+ where wild animals have clearly positive welfare levels (just like most humans in modern-day societies have clearly positive welfare levels).

Conclusion: farm animal abolitionism and wild animal welfarism

In summary, in the case of farm animal suffering, the abolitionist approach of choosing the situation without farm animals (C0) is the safest and most feasible option to respect negative altruism. The welfarist approach of choosing the situation with happier farm animals (C’’) is less good because it is riskier in terms of respecting negative altruism. Also, choosing the situation with clearly happy farm animals (C+) is not so feasible.

In the case of wild animal suffering, the abolitionist approach of choosing the situation without wild animals (D0) is risky because it quickly implies harming existing wild animals (i.e. decreasing their welfare). The welfarist approach of choosing the situation with happier wild animals (D’’) is probably the safest and most feasible option. With more research, choosing the situation with clearly happy wild animals (D+) may become feasible.

[1] In ethics, to make interpersonal comparisons of well-being possible, we can set w’=w when w is higher than +1/c or lower than -1/c, i.e. no Lorentz transformation. In special relativity, w’ =(w-v/c²)/(1-vw) for all w large and small, positive and negative.