Tuesday, July 30, 2013

Vegan advocacy and pessimism about wild animal welfare

Summary: Human meat consumption seems to seriously reduce wild animal populations. This has previously been considered an additional bad effect of meat consumption. However, some animal advocates claim that wild animal populations' aggregate welfare is so negative as to dwarf domestic animal suffering. Some of these advocates also favor spreading veganism, which on this view would seem to have the immediate effect of greatly increasing animal suffering. This tension should be addressed by advocates in cost-effectiveness estimates and research.

Gaverick Matheny, in his information-packed paper “Human Diets and Animal Welfare: the Illogic of the Larder” argues against the view that the demand for meat increases animal welfare through the creation of more animal lives worth living, the so-called "Logic of the Larder" argument. Part of the paper argues that most farm animals' lives are not worth living, and another part argues that if one counts animals other than birds and mammals, the land use changes required to supply grain for livestock would reduce wild and total animal populations if counted inclusively. Animal husbandry displaces wild ecosystems and reduces wild animal populations Agricultural land used to grow crops has a lower average rate of primary production than fertile wild ecosystems. Wikipedia provides this table:
ProducerBiomass productivity
RefTotal area
(million km²)
RefTotal production
(billion tonnes C/yr)
Swamps and Marshes2,500[3]
Tropical rainforests2,000[41]816
Coral reefs2,000[3]0.28[42]0.56
Algal beds2,000[3]
River estuaries1,800[3]
Temperate forests1,250[3]1924
Cultivated lands650[3][43]1711
Open ocean125[3][43]31139
Converting land to agriculture reduces the biomass available to feed animals by changing the mix of plants (where clearing forest, with minimal effects in grasslands) and diverting it to humans and farm animals. These data seem to support Matheny's claim that animal husbandry reduces animal populations. Most animal biomass is found in (invertebrate) wild animals Because of this productivity, and the fact that wild animals are mostly "cold-blooded" poikilotherm invertebrates, and convert food into tissue more efficiently than do farm animals, the total biomass of wild animals is larger than that of domesticated animals. And because wild animals are smaller on average, their population count is far larger. Under this metric effects on wild animals would dominate. [Image from Vaclav Smil's "Harvesting the Biosphere"]
However, population count is a questionable measure in light of large differences in nervous systems among the animal kingdom. Some animals may lack neurological features necessary for morally relevant experience, particularly the invertebrates that make up most animal biomass. Moreover, they may also differ drastically in quantity of experience. In human split-brain patients the hemispheres can experience and act quite independently without common knowledge or communication. If one thinks it implausible that the quantity of experience is doubled simply by reducing communication between brain regions, then one should think that positive and negative experience can occur in substructures of brains, and not only whole brains. So if a human brain includes a million structures as complex as an insect's brain, undergoing reinforcement learning in response to positive and negative reward, it is plausible the human brain will produce drastically more positive or negative experience.
Nervous system scale differs drastically among animals, from 5000 g for an elephant, 1300-1400 g for a human, 400+ g for a cow, a few grams for chickens or ducks, substantially less than a gram for fish, amphibians, and reptiles, with additional orders of magnitude to reach insects and other relatively large-brained invertebrates, and another several orders of magnitude to reach flatworms and similar. Weighting by nervous system scale greatly reduces the ratio between wild and domestic animals but the total of wild nervous system tissue remains large, almost entirely via invertebrates. While poikilotherms and invertebrates have smaller brains for a given body size, smaller animals tend to have higher brain:body mass ratios, e.g. some of the tiniest ants have been found to have brains making up approximately 15% of their body mass, where human brains make up ~2% of our body mass (this is a result of the brain being the bottleneck in miniaturization, other ants have lower percentages). So it seems that there may be more nervous system tissue found among wild animals than in the human sphere (of which the great majority is human, followed by cows). However if one weighted by nervous system and thought small invertebrates did not undergo relevant experiences (a common and plausible view), the conclusion would reverse. Pessimism about wild animal welfare resuscitates the Logic of the Larder argument Some argue that suffering exceeds happiness in the lives of wild animals in the aggregate, and in light of the above facts further argue that that the total negative welfare of wild animals dwarfs that of factory-farmed animals. On a this view, the immediate effect on animal welfare of expanded meat consumption would be positive, as the destruction of wild habitat outweighs conditions on farms. Proponents of this pessimistic view cite the short average life-span of animals before death. On the other hand, even for short-lived animals the vast majority of their lives are not spent dying, so if welfare is positive at other times, pain accompanying deaths may be dominated by other experiences. Also, the shortest-lived animals have the least developed nervous systems, lowering the probability and expected quantity of welfare involved.
I remain agnostic on the question, but for the pessimists it appears to reverse the analysis made by Matheny, which worked with the assumption that the lives of wild animals were worth living. A tension for some animal advocates I have spoken with a number of advocates for animal welfare who both a) take the above pessimistic view of wild animal welfare, and b) argue that the short term welfare of farm animals makes vegan outreach the most cost-effective available charitable intervention. But given those views the short term effects of vegan advertisements would be to increase wild animal suffering much more than farm animal suffering is reduced. These advocates might (and sometimes do) revise their argument along the following lines:
Yes, this intervention will greatly increase animal suffering in the short run, but we hope that some of the new vegans will eventually lead to more concern with wild animal suffering, and reduce it in the long run by more than we have increased it in the short run.
However, those who take up this revised view cannot claim the intervention as one with a proven track record of reducing animal suffering, or one that stands separately from speculative claims that the intervention will have more positive effects on civilization's long run future than other interventions.

It seems to me that this issue should be more prominently addressed by animal advocates and organizations in their cost-benefit analyses and advocacy, particularly in analyses of unadjusted "years of life prevented," a metric where the problems would be even more severe than under other metrics, e.g. life-years weighted by nervous system.

For instance, if the pessimistic view is right, it would also mean that if poverty reduction and human population growth increase meat consumption to some degree this would not on balance worsen animal welfare in any direct way, but rather improve it.

On the other hand, if the pessimistic view about wild animal welfare is wrong, then the loss of welfare by wild animals should be taken as a serious harm of human meat consumption, aligning the welfare of wild and domesticated animals, and deserving of a more prominent place in discussion.

Better estimates of animal population density by land use might be a tractable empirical research topic for animal advocates. The results may be sensitive to the weightings of different sorts of animals: Matheny sees factory farming of chickens as increasing bird populations at the expense of more numerous poikilothermic vertebrates.


  1. The case of wild animals strikes me as a reductio ad absurdum of pure hedonism.

    I remember that Bernard Williams had a video that, although I did not find his specific argument that convincing, certainly did mention this in a memorable way.


  2. Alex,

    All plausible moral theories assign value to positive and negative experiences; where these theories differ from hedonism is in recognizing additional sources of value. But if the number of wild animals is large enough, or their nervous systems sufficiently complex, wild animal experiences will dominate even for non-hedonistic theories of value.

  3. Pablo, yes if the values of individual creatures are combined additively, but one can also have bounded value functions which aggregate subadditively.

    Larry Temkin and Richard Chappell have discussed such views, and I think there is a lot to be said for them (a bounded utility function can always prefer more happiness to less, and address many other standard objections other than a demand for additivity through the appropriate choice of terms).

  4. Fair enough. I thought Alex was objecting to hedonism as such, not to hedonistic utilitarianism.

  5. That's an interesting graphic and book from Smil, by the way. I've been looking for something like that.

    But my reaction with wild animals is that it may be mistaken to assign value to individual experiences of any kind--rather, what we should value is normal functioning of animals and the whole ecosystem. This is deep ecology. I guess this is not compatible with utilitarianism in general, actually. But it could be consequentialist.

    Utilitarianism would apparently assign no value to wild plants and other nonsentient organisms. Do you think that is OK?

  6. Great post, Carl!

    I've had the same concerns since 2006.

    I'm less confident than you that meat consumption reduces wild-animal suffering on balance because climate change can also be a big factor, and its sign is very unclear. Also, the relevant question is what land is used for marginal agriculture, and at least in the US, that seems more likely to be grassland than forest. Temperate grassland has about the same primary productivity per hectare as crop land, although more of it may be eaten by small organisms.

    I do think it's >50% likely that crop cultivation reduces wild-animal suffering. My probability for climate change is basically 50%-50% at this point. So, yes, it does seem like farmed-meat consumption has a slightly greater than 50% chance of reducing wild-animal populations, and given the vast numbers of wild animals, this dominates the impact on farmed animals, especially for someone like me who's not sure about brain-size weighting.

    Also note that the effect is different depending on the type of meat -- e.g., wild-caught fish doesn't have these effects at all. Beef seems possibly best in terms of habitat destruction (especially if it's farmed in the rainforest).

    I do make the argument that veg*ism is net good because of the attitudes toward animal welfare it promotes. I don't actually believe that life-year-per-dollar calculations for veg outreach are the right metric to use for those charitable interventions, just as I don't believe QALYs/$ are the right metric for human charities. The reason to use those numbers is (1) to get on the same playing field as people who do use the QALYs kinds of numbers (which themselves ignore major ripple effects, just like the veg calculations do) and (2) to motivate people who care about the more tangible short-term stuff to do things that ultimately are beneficial for wild animals because of flow-through memetic impacts.

    I like discussing these issues openly, and I think this post is a nice contribution to the debate. :) I would just point out that it's less certain that meat reduces wild-animal populations than was suggested.

    Alex, deep ecology can be consequentialist, but holding that view means more wild animals suffer, so I hope you don't hold it. :)

  7. Brian, yes, climate change may have complicating effects that increase uncertainty over time.

    The original inspiration for the post was discussions with interlocutors "argu[ing] that the short term welfare of farm animals makes vegan outreach the most cost-effective available charitable intervention," or a robustly demonstrated intervention, while simultaneously proclaiming wild animal pessimism.

    This also shows up even in cases where those memetic arguments are largely inapplicable, e.g. individuals advocating against increasing the wealth of poor humans on the grounds of meat consumption while embracing wild animal pessimism.

  8. Maybe by "short term welfare" they mean "easily quantified" rather than literally "near in time to the present." The broad effects of veg*ism on the future are arguably more predictable than its effects on wild animals today.

    In terms of temporally short-term effects, I'm pretty unworried about the poor-meat-eater question because of wild animals: If more humans means fewer wild animals, that's worth the factory-farming cost. (I think climate change is a primary source of uncertainty about whether more humans is net good for wild animals. Also adding uncertainty are eutrophication, fertilizer use, etc., though these are probably less significant.)

    I think memetic side-effects probably dominate on poor-meat-eater just like on veg, though those side effects may be somewhat more complicated to compute.

  9. "It seems to me that this issue should be more prominently addressed by animal advocates and organizations in their cost-benefit analyses and advocacy"

    I agree.

    I think to promote anti-speciesism is pretty better than to promote only veganism. Anti-speciesism makes easier to accept another issues (wild-animal suffering and so on), and it has less linked risks than only veganism.

    Unfortunately, anti-speciesism is not very present in American animal charities. But several Spanish and South-American charities have a more focused anti-speciesist message, and some of their members take wild-animal suffering into account.

  10. This is an excellent post, Carl. I began to wonder about this very issue when I encountered Matheny's article this last weekend. Thank you for helping me think this through.

    I'm left wondering why for some "veg*n outreach" is the preferred mechanism for inspiring concern for animal suffering toward the end of getting people concerned with wild animal suffering. There are many other ways in which non-wild animals suffer that can be used to inspire concern for animal suffering and that don't demand such difficult lifestyle changes. Or is the difficulty of becoming veg*n versus say boycotting certain cosmetics somehow important or is perhaps the severity of the suffering of farmed animals somehow important in generating the right level of commitment or concern to get people interested in what do about wild animals? It seems like these "veg*n outreach" advocates should be able to produce good reasons for their chosen strategy particularly if the net effect of increasing the number of veg*ns is uncertain.

    I would guess that most veg*ns I know have a far more idealized view of animals' lives in nature than non-veg*ns. I now want to test that.

  11. Jason, the analysis only applies to those who think that wild animals welfare is very negative relative to farm animal suffering. I think many supporters of vegan outreach would disagree with that conclusion.

    That might be because of optimistic or agnostic views about the sign of wild animal welfare, but there are also other ways to that conclusion.

    The attribution of conscious experience of pleasure and pain as we understand it in insects and small invertebrates is a questionable step that is essential to many arguments for wild animal welfare dominating farm animal welfare.

    If you take a look at the Smil infographic, wild vertebrates make up a small portion of the land vertebrate biomass, which is mostly composed of humans and farm animals. If one considers land vertebrate nervous system tissue (or vertebrate nervous system tissue generally), humans make up the great majority, with domestic animals making up the great majority of the remainder.

    If one restricts attention to birds and mammals, as Gaverick did in the first pass analysis in the paper, one gets the conclusion that chicken meat consumption increases even unadjusted population count, while other meats reduce population.

    Such criteria would leave effects on domestic animals more important than effects on land use.

    Also, meat consumption involves a closer causal and means-end connection to the suffering of farm animals, evoking deontological and common sense moral concerns.

    I have spoken to some animal activists involved in the vegan ads who took the position that they would want to reduce the suffering and killing involved in factory farming even if that led to worse situations for animals overall (via considerations like those discussed in the post), on the grounds that factory farming involves active harm rather than passive failure to help.

  12. You make a great point, Jason! I agree that while veg outreach is likely positive for wild animals considering memetic effects, it's not optimal, and it's possible veg memes could even be negative for wild animals, because as you say, veg*ans are often more supportive of environmentalist intuitions. That said, it's important to remember that correlation is not causation, and probably a newly created veg*an is less environmentalist than existing veg*ans on average.

    The reason I supported veg outreach in the past was historical: Before this year, there was no charity working on wild-animal suffering, so my best option for donating was veg outreach. Now the world's first charity focused on wild-animal suffering, Animal Ethics, has recently been created. We're still getting our website up, but if you're interested, I can give you more info. :)

  13. Hi Jason.

    "Or is the difficulty of becoming veg*n versus say boycotting certain cosmetics somehow important or is perhaps the severity of the suffering of farmed animals somehow important in generating the right level of commitment or concern to get people interested in what do about wild animals?"

    I think that is the key.

    It's possible to speak directly about wild-animal suffering, but it's very difficult to find non-vegans very involved into the wild-animal suffering issue. Why? I think because vegans tend to think a lot about animal suffering, what is important for becoming a involved activist.

    I know that there are a lot of vegans against interventions in the wild. But I think this is the problem of promoting veganism ignoring anti-speciesism.

  14. The infographic on biomass is great! It definitely increased the importance I place on domestic animals and changed my general imagination of wild-life. Is it in accordance with [url=http://www.utilitarian-essays.com/number-of-wild-animals.html] Brian's estimates [/url]? I noticed that wild land-vertebrates seem more important in Brian's estimates. It could be useful to get this right if one wants to approximate the short-term effects of veg-advocacy better. It's especially relevant for people who don't count invertebrates.

    For example if both estimates were correct then land vertebrates (amphibians, reptiles, mammals, birds - land vertebrates in descending size of population) would weigh on average just 0.46g-0.046g (dry). I took the biomass of 0.01Gt and divided it by the number of wild land vertebrates in Brians list. This figure seems impossible given that even the Etruscan shrew, the lightest mammal on Earth, weighs 1.8g and the very smallest reptiles and amphibians are little frogs and chameleons. Am I missing something or are these data significantly conflicting?

    A note on reptiles though (I'm writing this here since utilitarian-essays doesn't allow comments). Reptiles and amphibians seem very important since in Brian's data they are more than mammals, birds, livestock and humans combined in number. A foot note on the page says:

    "Examining 25 m2 quadrats, the researchers found an average of 0.2559 reptiles per quadrat = 10,240 reptiles per km2 (p. 413). Assuming this is a typical density of reptiles in tropical rainforest, I naïvely divide this number against the Gaston et al. (2003) figure of 1,250 birds per km2 of tropical rainforest, yielding ~8 times as many reptiles as birds. I extrapolate this world population."

    I am not a biologist, but since reptiles are cold-blooded I would expect much more of them in warm climates, especially the tropics. For instance, the cold forces every reptile to hibernate in temperate regions. The same may be true for amphibians.

    [url=http://www.thegatesnotes.com/Books/Energy/Harvesting-The-Biosphere] Smil also says[/url] that global biomass production has almost halved in the last two millennia. Likely due to agriculture. Interesting. Simulataneously the share of biomass production from crops has increased to a very rough 17%, which suggests that agriculture does indeed produce less biomass. Not necessarily on marginal land though.

    ] Global warming gives tropical lizards energy problem [/url] for an example of climate change influencing animal populations.

  15. Correction: Most of the 45% drop in global biomass production must be caused by factors other than agriculture, since more than 80% of it happened before 1900 according to Smil's data. Natural climate change (http://en.wikipedia.org/wiki/Milankovitch_cycles) could be a reason.

  16. Soren,

    The period before 1900 saw massive deforestation and expansion of agricultural land.


    I agree Smil's estimates look to be in tension, with Brian's. I would tend to think that the expert is right, but I do plan to figure out what drove the divergence.

  17. "For example if both estimates were correct then land vertebrates (amphibians, reptiles, mammals, birds - land vertebrates in descending size of population) would weigh on average just 0.46g-0.046g (dry). I took the biomass of 0.01Gt and divided it by the number of wild land vertebrates in Brians list. This figure seems impossible given that even the Etruscan shrew, the lightest mammal on Earth, weighs 1.8g and the very smallest reptiles and amphibians are little frogs and chameleons. Am I missing something or are these data significantly conflicting?"

    Yes, Smil's figures are for dry weight.

    Brian's figures are dominated by estimates for amphibians, especially rainforest amphibians which do weigh in the vicinity of a few grams when wet. They should also have a higher body water percentage (less bone, more water).

    And rainforests are unusually amphibian-friendly: Brian is plausibly extrapolating too much from studies in rainforests.

  18. This comment has been removed by the author.

  19. This comment has been removed by the author.

  20. Cultivated land is not only used for food production, but also other forms. For instance, biofuels, wood for housing, potentionally new bioplastics or other materials that can be turned into all kinds of non-suffering consumer goods.

    Is there any evidence to what point an increase in general GDP per capita decreses wild-animal populations, independent from meat consumption? I would assume that meat consumption levels off at some point as people become wealthier beyond poverty, since even the rich eat only so much food. However, you can sink many more natural resources into other forms of energy and biomatter consumption, without creating neural tissue capable of suffering.

    If true, then an increase in average wealth per capita may acually decrease WAS even more than it would from increased meat consumption alone - what else do people buy when they are relieved from poverty, and how does it affect land use? And can this be affected by advocacy/wealth transfers as well?

  21. Hedonic Treader, yes, it seems rather likely that more GDP reduces WAS in the short run. In addition to the effects you cited, there are parking lots and buildings covering land. Some effects of human activity might be negative, possibly including climate change or eutrophication, though those are both really uncertain.

    But presumably the far-future impacts of economic growth are more significant, and there the sign is less clear IMO. In contrast, the sign of veg*ism for the far future is likely positive.

  22. "the sign of veg*ism for the far future is likely positive"

    This is based primarily on memetic flow-through effects against speciesism/for suffering-prevention, correct?

  23. Yes, that's primarily what I had in mind. A secondary reason is reducing climate change, which I'd expect to slightly decrease risks of conflict.

  24. Are the conclusions of this post (that veganism is likely to increase short term suffering if one holds the pessimistic view of wild animal suffering) still relevant 2 years later, or has new info changed that conclusion? Thanks!

  25. Horizontal Integration,

    In my view the main conclusion of the post is that assessments of the impacts of agriculture based on animal-years without large adjustments for the characteristics of the animals are missing the vast majority of the effects they should be looking at on their own terms. One shouldn't both say that chickens are collectively overwhelmingly more important than cows because of their numbers in assessing agricultural impacts while completely ignoring far more numerous wild animals impacted by agriculture. Once one considers those impacts, then there is a lot of room for ethical and empirical study and debate.

    I thought and still think the quality of the data for all wild animal population effects, including crop residues, food waste, climate change, variation in marginal cropland, and so forth mean one should not be very confident about whether wild animal populations will go up or down from the limited data in the OP. Marginal things I've learned at the last 2 years have pushed towards lower population effects, but I haven't done an in-depth review.

    Brian Tomasik has spent more time looking at some of those factors than I have subsequent to this post:


  26. Here's a follow-up to the discussion with Sören and Carl on global wild-animal estimates.

    First, I took your suggestions to heart about not extrapolating herpetofauna outside rainforests. In the latest version of "How Many Wild Animals Are There?", I estimate the lower bounds for reptiles and amphibians based only on tropical land area.

    Second, I got a copy of Smil's book and have begun reading the details. I don't see an estimate of the global biomass of all terrestrial vertebrates. It looks to me like Smil only talks about mammals, in which case the discussion about the herpetofauna numbers yielding too small "mass per animal" values wasn't relevant after all. My mammal population estimates seem quite consistent with Smil's numbers, as you can see in the "Mammals" section of my piece.