Scott Weathers is a strategic partnerships manager at Compassion in World Farming USA, which advocates for the end of factory farming. In this talk, he rejects Robin Hanson’s “logic of the larder” argument — the idea that it is better to live and be slaughtered than not live at all — and explains why he finds it problematic. He also shares evidence for focusing efforts on chickens, and walks through the pros and cons of a new proposal for evaluating farm animal welfare from researchers Karolina Sarek and Joey Savoie.
A transcript of the talk, which we have lightly edited for clarity, is below. You can also watch Scott’s talk on YouTube or read its transcript on effectivealtruism.org.
I'm Scott Weathers. I work for the U.S. office of Compassion in World Farming, and I'm going to talk about when farmed animals will experience more happiness than suffering.
Just to give you a little bit of context on this presentation, I was speaking to Amy Labenz earlier about different topic ideas. I presented about four or five ideas to her. When I presented this one, she said, "Oh, I like that one. That's a spicy topic." In my head, I thought, "Wow. That is literally the hardest talk I could possibly give," and so as payback, I turned in my slides about an hour ago. I hope you're okay with that, Amy.
I want to take a really serious and hard-headed look at this and start with some cute animal photos.
I think the interesting part of this presentation is that's really what it's supposed to be about — the joy of animals is an important component of their experience. But since this is the last cute animal photo, things are going to go mostly downhill from here.
A little bit of context: I work for Compassion in World Farming and I run the Friendly Food Alliance, which is focused on reducing the numbers of animals in companies' supply chains.
One strategy that we use to achieve that end is to convince companies to offer more plant-based options. There are a lot of other strategies we use. For example, we try to convince companies to offer more blended products that have a certain percentage of animal protein and another percentage of plant-based protein, or to do other things, like reduce food waste.
After we work with them for a year on these strategies, we plan to award companies that are willing to commit to actual reduction targets in the number of animals that are in their supply chain. So, that's just a little bit of context on the work that I'm doing.
One of the central arguments that I'm going to engage with today is called the logic of the larder.
This is an argument that was propagated by Robin Hanson in 2002. So, he says, in rough terms, that it's better to live and be slaughtered than to not live at all, and that's the central argument that I'm going to dispute here. (If you didn’t know — and I had to research this — “larder” is just an old word for a cold place that you store food.)
So, in 2006, [Jason] Matheny and [Kai] Chan presented six conditions that they thought were necessary for the logic of the larder to be true. I want to review those and the implications that I think they have.
The first condition is that life could be beneficial. This is a pretty philosophical argument, but it basically gets into the idea that non-existence can't be compared with existence. I’m not going to go too deep into that one, but I think it's an important one to consider. Matheny and Chan say that it invalidates the logic of the larder.
The second condition — and I think the most important one, at least for the purposes of this talk — is that life is worth living. I'm going to argue that life on factory farms, at least for most of the animals under consideration, is not worth living.
The third criterion is that existence would not happen without consumption of an animal product. So, the reason that farm animals exist is that we're raising them for food.
The fourth condition is just that animals would be replaced after their slaughter. Compassion by the Pound: The Economics of Farm Animal Welfare is a book written by two agricultural economists. They address this argument pretty thoroughly, and conclude that when you purchase an animal product, that does cause the future creation of another sentient being.
The fifth condition is a little complex.
Matheny and Chan take the view that wild animals have net-positive lives, specifically focusing on mammals and birds. They say that this condition holds if the rearing of a factory farmed animal does not prevent a greater number of animals with net-positive lives from existing. So, if you thought that birds and mammals did have net-positive lives, then eating an animal product that causes the destruction of habitats would not make this condition hold.
The last condition is just that the purchase of an animal product doesn't produce greater moral value than other uses of money. So, for example, they argue that, even if you're an omnivore, it might be better for you to eat chickpeas or other cheap sources of plant-based proteins and then raise hundreds and hundreds of mice — and that would be an easier, more cost-effective way to create utility than eating meat. That's the argument there.
I want to get into what farmed animal quality of life is like today, and one thing that I want to clarify is that quantification is going to play a really important role here. And while a lot of people might rightfully object that we can't 100% precisely estimate the quality of life on factory farms, using quantitative estimates is an important tool. It allows us to continuously update our assumptions and explicitly debate those assumptions, and then refine our model as we incorporate new forms of evidence.
I think the slide below might actually be the most important one.
If you're a consequentialist, land animal welfare is, essentially, the welfare of broiler chickens and layer hens. The vast majority of animals that are being raised on factory farms are chickens. And so, when we're talking about farm animal welfare and we're talking about the logic of the larder, this is the thing we need to pay attention to. This is where we need to focus the vast majority of our attention.
Probably most of the people in this room are aware of a lot of this stuff, but just to give a quick description: These are broiler chickens being raised for meat on the left. One factor that I think really determines a lot of their quality of life is their poor genetics. So, they're selected for breeds of chickens that grow extremely quickly, and as a result of that quick rate of growth, they have bones that break very quickly, they have heart and lung ailments, and they're not designed to exercise very much. They sit around most of the time and they wear off their feathers on the bottom of their bodies as well as suffer infections, and they're killed after about 45 days. So, that's the quality of life for broilers.
Layer hens over here live their entire lives in cages, and this obviously causes extreme welfare complications. So, you see their feathers are worn off because the hens are pecking each other. They have very brittle bones, and so they suffer, obviously, from a lot of welfare complications as well.
So, the basic reason that I'm rejecting the logic of the larder is just that when you look at chicken quality of life — layer hens and broilers — it is very, very poor and certainly net-negative.
I'm going to go through one proposal for evaluating farm animal welfare that strikes me as promising. I want to say that Compassion in World Farming can't endorse this methodology. It's a very new one and the final results of it haven't been published. Our technical team hasn't been able to evaluate it yet, but I do think it's an interesting proposal. Once the answers are released, I think it will provide answers in a format that is useful for evaluating this question of whether farm animal lives are negative or positive.
This is [Karolina] Sarek and [Joey] Savoie.
This year, they published a post on the EA Forum. They proposed an index that uses reviewer-assessed animal reports. So, basically, they create descriptions of the lives of farmed and wild animals using objective facts, and then they ask reviewers to assess the animals’ quality of life according to those descriptions. They use a lot of different factors. They look at the death rate of different animals as well as the reason that those animals are dying. They also look at human preference from behind the veil of ignorance — that is, saying: "Given a set of facts about what an animal's quality of life is like, would you prefer to be in that state or not?"
This is what it might end up looking like. These are not the final results, but they might conclude that a developed-world human lives a pretty strongly net-positive life, a caged hen has a fairly net-negative life, and a wild insect has a fairly net-negative life. These are examples of what we might conclude, but again, the results haven't been published.
I think there are a lot of trade-offs of this approach.
I think one of the really major pros is that it clearly demarcates where an animal's life is net-positive or net-negative, using a numerical scale, and that's something that's fairly new in farm animal welfare and the measurement of farm animal welfare. It also measures wild and farmed animal species, which, to my knowledge, makes it the first index to try to do that, and it has very clear criteria with adjustable weighting. So, if you wanted to say, "I care about the pain of death twice as much as any other factor," you can adjust the weighting of those factors according to that belief of yours.
However, I do think this approach excludes a lot of ethical views. If you're a non-consequentialist, it would be pretty difficult to agree with this approach. But even within consequentialism, there may be folks who don't agree with it entirely. I also think this approach by Sarek and Savoie does rely on anthropomorphization, thought this isn't necessarily a bad thing. I think there are trade-offs to this approach. This is a critique that applies to any attempt to estimate farm animal welfare, but it is something to be aware of nonetheless.
A couple more pros and cons: I really like that this system maximizes the use of objective data. It tries to look at mortality statistics and other objective forms of data while permitting certain subjective judgments. Clearly, if we're trying to assess farm animal welfare, it's not as simple as plugging in a bunch of numbers and then looking at what the eventual answer is. You really have to make subjective calls here, and they are transparent about when they do that.
They also have a scoring system that really maximizes the reliability of raters. They provide very clear criteria for what range of numbers should be used if, say, the rater would like to be in a certain state as an animal. This maximizes the ability of reviewers to clearly indicate their preferences.
I also think that having an independent reviewer structure really improves the generalizability of results. They allow reviewers to read the animal reports without necessarily looking at the final conclusions, which helps a lot.
* For each of the criteria feeding into the eventual score of farm animal welfare, they have sub-criteria, and I think that weighting is not necessarily empirically derived. This gets back to objective judgment, but I think there are likely to be issues that some people take with the weighting that they've used.
* I think that using a consistent standard across species might neglect important species differences. So, a chicken is not going to think that the same things are important to its welfare as, say, a dolphin or another animal.
* Lastly, just a practical consideration: This approach requires a lot of accurate data, and that's just difficult to get in a lot of cases.
So, what does this imply for meat reduction?
Carl Shulman in 2013 argued that meat consumption might actually reduce wild animal suffering by displacing wild animal habitats, and so if you take the view of wild animal suffering that Oscar Horta presented earlier today, that wild animals do suffer, you might buy this argument. I think the assumptions that you make about invertebrate sentience matter a lot here, and so we really need to be careful about that. But I think this is an important consideration.
Compassion in World Farming doesn't necessarily support this exact view, but I want to present the implications of our work in terms of what meat reduction might imply for wild animal suffering. I'm not going to present the research here on the direct impact that focusing on chicken or eggs will make, but others have argued elsewhere that prioritizing chicken and eggs is likely to yield the greatest improvement for farm animal welfare because those animals are very tiny, and so you get a lot less meat from them than you would, say, a cow or another bigger animal. So, in terms of direct impact, it's very clear that focusing on chicken and eggs is the best prioritization.
However, I'm going to show here a table that, I think, demonstrates that even if you take the view that wild animal suffering is a serious problem, [focusing on chicken and eggs] might be wise.
So, this is a table from the Matheny and Chan paper that I presented earlier. Again, they take the view that lots of wild animal lives are net-positive. So, in this calculation, they're just looking at mammals and birds that they think experience net-positive lives.
So, over here [points to slide], we have farm animal life years gained per 20 kilograms of protein. As I said, this demonstrates that chicken and eggs have the largest consequences on farm animal welfare. Then, in the third column, it shows wild animal life years lost per 20 kilograms of protein. So, as you can see, 20 kilograms of chicken or eggs have a pretty minimal impact on wild animal suffering and other animal products have a much greater impact. So, you can see here, when you focus on the net impact, that focusing on chicken and eggs, when you take wild animal suffering into account, is likely to yield the strongest positive impact.
In conclusion, taking the Oscar Horta and Brian Tomasik view of wild animal suffering, you might still prioritize chicken or eggs in terms of meat reduction, and that's what most farm animal welfare organizations are doing right now.
This has implications for the work that we're doing at Compassion in World Farming.
The first implication is that if we really focus on layer hens, as well as (potentially) fish, that is likely to yield the highest direct impact, while minimizing the chances of having a negative impact on wild animal suffering. I think we do have a lot of uncertainty about the impact on all animal products, but that uncertainty is likely to be highest for beef and dairy.
One other consideration: I think it's possible that the long-term effects of this work dominate. For example, we may need society to get rid of speciesism in order to tackle wild animal suffering or other problems in the long run [and this could be easier if we’ve taken steps to recognize farm animal welfare as important]. That's an important consideration in any sort of model that we develop around this problem.
Some final thoughts: I think the logic of the larder doesn't hold, primarily because the welfare of broiler chickens and layer hens is really, really poor. Research into quantifying farm animal suffering and wild animal suffering is still really early, but it's critical that we get this research right, and quantifying farm animal welfare is an important task that is going to allow us to estimate the impacts of meat reduction in other programs.
Then, lastly, chicken and fish products have the worst direct effects, as well as the worst wild animal suffering effects, and so focusing on these products is likely to yield the greatest impact. Thanks so much.