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.

The Talk

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.


1000 Michelle Graham UPDATED.003

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.
1000 Michelle Graham UPDATED.004 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.
1000 Michelle Graham UPDATED.005
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.

1000 Michelle Graham UPDATED.006

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.
1000 Michelle Graham UPDATED.007
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.

1000 Michelle Graham UPDATED.008

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.

1000 Michelle Graham UPDATED.009

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

1000 Michelle Graham UPDATED.011

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:
1000 Michelle Graham UPDATED.012 A solid un­der­stand­ing of evolu­tion­ary biol­ogy will help us gen­er­ate hy­pothe­ses re­lated to the first ques­tion. This slide shows a phy­lo­ge­netic tree.

1000 Michelle Graham UPDATED.014

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.

1000 Michelle Graham UPDATED.016

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.
1000 Michelle Graham UPDATED.018

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.
1000 Michelle Graham UPDATED.019
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.

1000 Michelle Graham UPDATED.021

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.1000 Michelle Graham UPDATED.022 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.

1000 Michelle Graham UPDATED.023

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.
1000 Michelle Graham UPDATED.024

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.

1000 Michelle Graham UPDATED.025

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.
1000 Michelle Graham UPDATED.026
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.

1000 Michelle Graham UPDATED.028

Evolu­tion is re­ally com­pli­cated. It’s not easy to make pre­dic­tions about it from your arm­chair. Think of the ex­am­ple of a po­lar bear.
1000 Michelle Graham UPDATED.029

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.
1000 Michelle Graham UPDATED.030
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.

1000 Michelle Graham UPDATED.032

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.