Michelle Graham: How Evolution Can Help Us Understand Wild Animal Welfare
We must take evolution into account when we consider animal welfare — whether we’re thinking about which animals are sentient or how animals might respond to a given intervention. In this talk, Wild Animal Initiative’s Michelle Graham presents a brief introduction to the theory of evolution (she also recommends this video for more background), explains how understanding evolution can help us conduct better research, and discusses the ways misconceptions about evolution lead us astray.
We’ve lightly edited Michelle’s talk for clarity. You may also watch it on YouTube and read it on effectivealtruism.org.
Evolution is a commonly misunderstood — but also commonly discussed — concept. We often have intuitive notions about how traits are inherited and develop over time, but unfortunately, these notions don’t always match perfectly with reality. That makes it quite difficult to make good predictions about how traits develop and change over time.
Today I’ll go over some of the basic concepts of evolution and engage with some of the reasons why it is so important for wild animal welfare. Then, I’ll talk about some objections to working on wild animal welfare that rest on evolutionary concepts.
A review of evolutionary basics
First of all, I want to make sure I answer this question in my talk: What is fitness? It is the likelihood of a particular set of traits enabling the organism that has them to contribute offspring to the next generation. This is important to clarify because I think fitness is often misunderstood as representing health or strength. But in terms of Darwinian or evolutionary fitness, health and strength are only relevant in so far as they contribute to reproductive success.
Genes that tend to produce positive reproductive outcomes tend to continue through multiple generations. And generally, evolution is as much about genes as it is about organisms. So, although there’s more than one way to define evolution, perhaps the most common definition is based on allele frequency.
It’s commonly described as the process by which allele frequency changes in a population over time. (Just to clarify, an allele is a specific qualitative version of a gene. For example, we all have genes for hair color, but I express the allele for brown hair color.)
The changes in allele frequency that occur in evolution are accomplished through various evolutionary mechanisms.
The most famous is natural selection. But natural selection is not synonymous with evolution. There are multiple evolutionary mechanisms, including sexual selection, genetic drift mutation, and gene flow. And it’s important to note that some of these mechanisms are not necessarily adaptive.
Let’s take the example of gene flow, which is also known as migration.
This is the transfer of genes from one population to another. If we have, for example, two populations of dogs that are reproductively isolated, migration will occur when an individual from one population migrates and introduces its genes, primarily through sexual reproduction, into the new population.
The result of this population blending is a greater variety of alleles for particular traits — for example, for size or color.
Again, it’s important to note that this change in color may not lead to better outcomes for those animals, so it’s not necessarily adaptive. It’s just a fact that those new genes lead to variation. But when considering adaptive outcomes, we’re normally talking about natural selection, which is the process by which some sets of traits tend to lead to greater reproductive success than others.
We often think of natural selection as being primarily about survival, but ultimately, survival is only important to the extent that it leads to a larger reproductive output or success.
To clarify: In order for an animal to pass its genes on to the next generation, it has to survive long enough to reach a reproductive age. It has to find a mate and reproduce.
So, if there are brown and green ants, and for whatever reason the brown ants are more reproductively successful than the green ants, we’d expect the relative proportion of brown ants to increase over time.
It’s critically important to note that this is a relationship with the environment. The [determinants of] reproductive success are not inherent to a set of traits. In one context, the brown ants may be more successful, but when a predator that can see the brown ants much more clearly is introduced to the environment, being brown is probably no longer successful.
It’s also important to think about reproductive success or natural selection as coming about through the elimination of traits that don’t succeed, rather than the promotion of perfect traits. There are a lot of constraints on evolution; things that don’t succeed tend to disappear. And that’s different from things that are amazing, that just spring up out of nowhere.
There are several requirements for evolution by natural selection to occur.
1. There must be variation. If we only have white bears in a population, [there’s no way a bear could gain reproductive fitness based on its color]. Bears will just continue to be white.
2. Traits must be heritable. If, for some reason, being blue is a selective advantage — maybe, for example, other bears like the blue bear — but the bear is only blue because I painted it that color, the subsequent generation will remain white.
3. There must be differential reproduction. We would expect that if blue bears are more reproductively successful, there will be more blue bears produced in subsequent generations. That means a change in allele frequency. If, for whatever reason, there’s variation but each individual merely creates a replacement for itself [with identical alleles], the relative frequency of the alleles [within the population] will not change.
How evolution and wild animal welfare intersect
I just mentioned variation and heritability. We have some reason to think that things like brain structure and certain kinds of behaviors vary [between animals], and may be heritable to some extent. Therefore, we should anticipate that natural selection will act on these and other potentially welfare-relevant traits.
Evolution plays a really important role in three main questions of welfare biology:
A solid understanding of evolutionary biology will help us generate hypotheses related to the first question. This slide shows a phylogenetic tree.
The endpoints on the far side represent different species and the connections represent common ancestors. Let’s say that we know for certain that the ability to experience pain is a heritable trait and that we discovered that both of these species had it. There’s a method in evolutionary biology called “ancestral state reconstruction” that can lead you to reasonably hypothesize that the common ancestor of those two species also had this trait. That is a parsimony-based argument). It’s not impossible that the trait evolved multiple times [such that the ancestor didn’t have it even though both descendant species did]. But you would expect the common ancestor to have that trait as well.
When we want to generate hypotheses about the effects of an intervention, we also need to consider evolution very carefully because our actions intersect with evolution. An example of a situation in which human intervention led to significant evolutionary results is the introduction of cane toads into Australia.
In the 1930s, cane toads were introduced to try to control the cane beetle, which was a pest species on sugar cane crops. But cane toads have poisonous glands that you can see on the sides of their heads. Snakes unaccustomed to ingesting poisonous prey items would eat these toads. And what we’ve seen over just a few short decades is that the snakes vulnerable to these toads now have, on average, smaller head sizes and larger bodies. That’s because a snake is a gape-limited predator; the smaller a snake’s head, the smaller the amount of poison it is able to ingest, because it can only eat smaller frogs. At the same time, the bigger its body, the less likely a given dose of the toxin is to kill it. So its phenotype — the characteristics that snakes express — protects it from this threat.
Obviously, the cane toad intervention was not a welfare-based intervention, but it does show how evolutionary outcomes will accrue from our actions in nature [and why it is so important to carefully consider the potential effects of our actions].
Finally, when considering the welfare value of different experiences for different animals, it’s important to think about how evolution might play a role. A reasonable working hypothesis about the purpose of pain is that it tells us to stop doing things that will harm us. An animal that has no pain response to an injury is less likely, perhaps, to avoid injuries. That means they’re possibly more likely to die before they’re able to successfully reproduce. Again, this is a hypothesis. But it’s a commonly used one. And if this is the case, we would expect that animals with different survival risks or reproductive strategies might end up with different responses to a given stimuli.
Consider an aquatic frog, for example, whose life depends on the chemical content of the water around it. You might expect that it’s going to be more sensitive to changes to that chemical content — and more likely to experience pain if that content is not ideal. But if I go into a stream that happens to have a slightly less-than-ideal chemical content, I’m probably not going to notice, because it’s very unlikely to affect my survival.
Thinking about the kinds of experiences that animals have is very important to people who care about wild animal welfare. And there are just so many animals out there in the wild. I’m thinking about ants in particular. There are between ten thousand and one hundred thousand trillion ants.
To put that into a visual for you, if a line were to represent the population of humans on the planet — and there are about 7.6 billion of us — it would be about one inch. The line that represents ants would be between 20 and 200 miles long. That’s a lot.
These animals are different from us. They have different life histories and life strategies. They have different reproductive needs. They have different survival needs. And so it’s important within the wild animal welfare community to focus on research that helps us understand the needs of animals, so that we can address the problems they face. And given how many of them there are, we should expect addressing their problems to be really important.
Objections to working on wild animal welfare
I’ve talked a bit about how evolution interacts with wild animal welfare. And now I want to engage with a few objections to working on wild animal welfare.
One that I’ve heard is that evolution shouldn’t be interfered with because [non-interference] results in the animals with the greatest fitness.
When people make this argument, it’s not clear to me whether they think that fitness has intrinsic value. If I’ve defined fitness for you effectively, I hope you would wonder why it would have intrinsic value. It’s just reproductive success.
Or maybe they think that fitness is highly correlated with welfare. But for that argument to [make sense], it would have to be the case that any increase in wild animal welfare requires intervening in an evolutionary process. It would have to be the case that evolution always leads to maximum fitness. And it would have to be the case that high evolutionary fitness always leads to high animal welfare.
[None of those assumptions] is true. Some interventions do not influence reproductive output. Examples would be changing our fishing practices or changing pesticides to be more humane. Neither of those may influence which animals die. They just influence _how_ they die, and so it’s possible to [enable] higher-welfare deaths.
As for the concept that evolution always leads to maximal fitness, that also is incorrect. Evolution is constrained by history and our genetic baggage. There are many things that, because they are based on our past behavior, are inaccessible to us.
If the curve [on this slide] represents various traits, some of which are low fitness and some of which are high fitness, it could be the case that current conditions [put us] in a local optimum. But because small variations away from that local optimum are lower fitness, natural selection would continue to constrain to current conditions. It may well be that human intervention, through perhaps genetic modification or breeding, could move us toward a theoretical “best fit.”
Finally, we should be suspicious of the idea that high evolutionary fitness leads to welfare.
Rats and elephants might have different reproductive strategies. The rat could have many offspring, but only one survives. And the elephant could have a single offspring that survives. They have equal fitness; they each produced one [surviving] offspring. But many rats had to die — and that seems like a lower welfare result.
Could high evolutionary fitness correspond to high animal welfare in some cases? If that’s true, we want to investigate and understand it.
That brings me to the role of experimental research.
Evolution is really complicated. It’s not easy to make predictions about it from your armchair. Think of the example of a polar bear.
For many years, people thought that polar bears were white for only one reason: camouflage. But if you look closely at the hairs of polar bears, they’re see-through — and their skin is black. There are thermal advantages to that combination. These complicated paired effects, like camouflage and thermal advantages, make it difficult to predict what changes to the fitness environment might do (in the case of this example, to the polar bear’s color).
Tradeoffs are another factor that makes evolutionary outcomes difficult to predict.
Improvements in one trait can lead to detriments in a different trait. Exaptations are animal traits that evolved for one purpose, but are now used for another. For example, feathers evolved for thermal insulation, but are now used for flight.
Also, as I mentioned, there are multiple evolutionary mechanisms — and some of them are random, which makes things hard to predict. And finally, there’s individual variation. Evolution acts on species, not individuals. If we care about individual welfare, we must always be cognizant of the fact that an individual is not a perfect representation of its species. There will be individual variations that affect welfare.
The complicated nature of evolution is another argument against the tractability of working on wild animal welfare — the idea that it’s too complicated to do anything about.
But I want to say that “hard to predict” is not the same as “immeasurable.” And we can do welfare biology that will advance our ability to make good predictions and better inform [our views on how best to help] animals in the wild.
If you agree and want to support our efforts to do research into the space, I encourage you to go to the Wild Animal Initiative’s website, where you can sign up for our newsletter and donate.
Moderator: Thanks very much, Michelle. How do you currently view the field of wild animal suffering, and how has it grown? Where do you think more input is needed?
Michelle: I think it’s a difficult question to say exactly which aspects [of the field we should focus on], at what rates, and when. One set of [issues centers on the fact that] we just don’t know what’s going on out there and we need data. And [another issue] is that huge numbers of people don’t care about wild animals at all. I think we don’t want to completely ignore one [of those things in favor of] the other.
But at the moment, I do think the most value to be had is in getting [organizations that are] already considering these questions around what’s going on in the wild to frame them in a way that is more conducive to people caring about wild animal welfare.
A good example would be how conservationists already consider a lot of ethical questions tied to environmental outcomes. To get a conversation going about taking action to improve wild animal welfare at the individual level, rather than [just] the species level, fits into the overall framework.
Moderator: Have you seen any good examples of that yet?
Michelle: I can’t cite one right off the top of my head. But I think there are definitely conservationists who are interested in the [Wild Animal Initiative’s] projects and curious about what we’re doing.
One of our staff members just recently finished a master’s program in a conservation field. So it’s definitely not the case that every conservationist who hears about this concept [of improving wild animal welfare at the individual level] responds, “No, that’s not right.” There are definitely people who say, “Oh, that’s very interesting. I want to know more [about individual animals’ suffering].”
I think often the first time people hear about [individual wild animals suffering] it feels like a huge moment. I think it’s complicated. There are different ways that can happen. Some people are not [in a frame of mind] that allows them to care about animals at all, and so it’s a big shift for them just to start incorporating animals into their worldview.
For others, they’ve [absorbed the] pervasive cultural attitude that the way you conserve and care about wildlife is to preserve habitat and species. The shift from thinking about species as the target of interventions to the [individual] animals within that species is scary. What are their feelings? What are their needs? That was exactly my response. The first time I considered wild animal welfare, I thought, “Oh my gosh, there are so many out there and they could all be suffering, and I don’t want that to be true.” But it’s very important that we interact with that question and find those answers.
Moderator: I have a few questions from the audience. Are you concerned about how a focus on welfare might differentially impact our attention to r- versus K-selected species? (You can also define what that means.)
Michelle: Sure — r/K selection has fallen out of fashion slightly in life-history classification. But there are two ends of a spectrum. In the example I gave of the rats and the elephant, r-selected species would be the rat model, where animals produce a lot of offspring and maybe only a few of them survive. The K-selected model [entails] producing few offspring and investing really strongly in them.
I think a focus on welfare is [key]. And if that means that you care about animals that tend to have a lot of offspring who all die young, [you might want to focus on r-selected species]. But to some extent, the welfare of those different groups is an open question. And again, it’s a spectrum. It’s not like you have animals that only exist on one side or the other. So I’m not positive about what the impact of that would be. But if it is the case that animals toward the r-selected side of the spectrum have much lower welfare, I do want to care about those animals more.
Moderator: In your estimation, does having a prefrontal cortex and the ability to reflect upon our suffering increase or decrease our capacity to suffer relative to animals with less reflective capacity?
Michelle: That’s a very interesting question. I do not have a definitive answer to it. My own opinion tends to be that we are certainly, on average, over-focused on processing capacity as a lens for moral weight. I certainly don’t think we have strong evidence in favor of the idea that you can only suffer if you have strong intellectual capacity. In my personal experience, when I have been in pain and I’ve been able to say, “I know that this is going to end” and “I know what caused this,” it has actually easier for me to deal with [the pain]. Therefore, I can see that it’s a viable hypothesis that being less able to analyze your pain would make it more painful.
Also, if our general hypothesis about the evolutionary function of pain is correct, then it may be the case that a lower ability to extract a signal from noise means that pain signals need to be stronger for animals with less cognitive capacity in order for them to respond effectively. That’s another reason to think that my gut feeling of less cognitive capacity leading to greater pain [is correct]. But again, science needs to be done.