I’m passionate about building a better food system and have been working at ProVeg since 2017 to help make that happen. In Berlin, I organise the Effective Animal Advocacy Meetups to bring together people who want to create a better world for animals. I’ve also taken a pledge at Giving What We Can and donate the equivalent of one monthly salary every year to effective charities. My academic background is in theoretical physics and business management, and I’m also very enthusiastic about evidence-based nutrition.
Chris Popa
Karma: 192
Thanks for bringing that post into the conversation! We really appreciate the cautious approach it advocates and agree that there’s a great deal of uncertainty when it comes to understanding the full welfare implications of wild-caught fish.
That said, we think the situation with sardines and anchovies might be somewhat different from many other species. These small pelagic fish are already being caught at or near maximum catch levels, primarily to produce feed for farmed salmon. So increasing human consumption wouldn’t lead to more fishing, but rather to a diversion of existing catch from salmon feed to direct human food. This shift could reduce the scale of salmon farming, thereby lowering both the direct suffering of farmed fish and the wider ecological harms associated with aquaculture.
Because these small pelagic fish are already being caught (unlike species like tuna, which are mostly caught for direct human consumption) the food web effects of this shift may be more or less neutral. And while anchovies do eat krill, which might be sentient, this concern may be outweighed by the positive ripple effects of reducing salmon farming.
On the human side, sardines and anchovies may help some people stick to a veganish diet by improving nutrient intake and reducing reliance on supplements or costly alternative proteins, which could further reduce harm overall.
Thanks Vasco, I really appreciate your thoughtful engagement!
I agree that feed conversion losses mean more than 1 kg of alternative feed would be needed to produce 1 kg of farmed fish. So if that alternative feed uses as much land as the replaced plant-based food, land use could indeed increase. However, many promising feed alternatives (based on microbial fermentation, insects, or algae) may have a much smaller land footprint than typical crops grown for human consumption. So the net effect on land use depends on what those alternatives are. That said, many current feed alternatives might be less efficient.
I think it’s important to also consider not just the calories or kilograms involved, but the nutritional value. Sardines and anchovies are exceptionally nutrient-dense, as discussed in the “Health and Nutrition” section. So even if land use were slightly higher per kilogram, the nutritional return per land area might still be better. It’s definitely a complex question, and I’d love to see more data on this.
Thanks for the suggestion! We think herring might differ from sardines and anchovies in several relevant ways. For more detail on our thinking, take a look at Elizabeth Crewe’s reply to the comment by Sagar K Shah.
After receiving feedback on the duration of the purse seine process, we’ve revised some of our earlier assumptions and updated the post accordingly. In light of that, we now think it’s more accurate to say that we simply don’t know whether death by purse seine fishing involves more or less suffering than typical natural deaths. The mechanisms differ, but without direct evidence comparing subjective experiences, it’s hard to draw strong conclusions either way.
Given that, I’d suggest that the indirect effects of food choices, particularly those related to land use, biodiversity loss, and climate change, may be more decisive from an animal welfare perspective. These broader impacts affect not only current individuals but also the long-term well-being and survival of many wild animals and ecosystems. So even if we can’t confidently compare direct suffering in one death scenario vs. another, reducing demand for foods that minimise land use change, such as sardines and anchovies, seems to be the preferable choice overall.
Clearly, though, there’s a need for more research.
Thanks for the clarification!
As mentioned in the “Fish Populations” section, sardines and anchovies are mostly used for aquaculture feed and pet food. Since total catch is capped, increased demand for human consumption would raise prices and thereby likely incentivise feed producers to switch to alternative proteins, rather than leading to more farmed fish overall.
I appreciate the focus on comparing direct animal suffering, although I’m somewhat uncertain about the extent to which soil nematodes, mites, and springtails should be considered sentient, but I haven’t looked into this in depth so far. However, I wonder whether this analysis fully accounts for the long-term and potentially irreversible consequences of increased GHG emissions, particularly the risk of triggering climate tipping points, large-scale biodiversity loss, and the associated risk of human extinction. Ecosystem collapse and permanent species loss represent a qualitatively different kind of harm than direct animal suffering, and may be ethically more significant in the long run. Though of course, there is a lot of uncertainty around all of this.
Thanks for your kind words and for raising this. It’s a really interesting comparison and I actually touched on mussels and oysters briefly in the response to someone else’s question: I think many of the arguments here also apply to them, and I see no ethical concerns with their consumption. However, there are some health-related reasons to be cautious about eating them too frequently. As filter feeders, they can accumulate heavy metals and microplastics, especially if sourced from polluted waters. Nutritionally, sardines and anchovies offer even greater benefits, particularly higher levels of EPA/DHA and calcium, which makes them a better overall choice in that regard. They also tend to be significantly more affordable.
Thanks for engaging with the post! You’re right that the total wild fish catch has been relatively stable for decades. However, I think there may be a misunderstanding: sardines and anchovies aren’t farmed as a matter of principle. As outlined in the post, their biological characteristics make farming them economically unviable. So if demand for sardines and anchovies increases for human consumption, the most likely outcome is a reallocation, with more sardines and anchovies being directed to human food rather than animal feed.
Also, just to clarify: while I appreciate the point about invertebrate welfare, a core motivation behind the post is to minimise land use, given its outsized role in driving biodiversity loss and climate change. From that perspective, increasing land use, even if it may benefit some invertebrate populations, is not a positive outcome in my view. Would love to hear your thoughts on that trade-off!
It’s true that because of the small body size, you’d need a large number of sardines or anchovies to match the calories from a larger animal like a salmon. However, it’s worth noting that salmon themselves feed on fish like sardines and anchovies, both in the wild and in aquaculture, where each salmon is typically fed hundreds of small pellagic fish throughout its life.
Regarding the number of crop deaths caused by plant agriculture, it’s difficult to draw firm conclusions based on the limited data currently available, as discussed in the section “Fishing vs. Crop Deaths”.
That said, as explored in the section “Fishing vs. Natural Deaths”, these comparisons may be less relevant if fishing sardines and anchovies is net positive, that is, if their deaths via fishing involve less suffering than they would have experienced otherwise in the wild. In addition, the long-term impacts such as biodiversity loss or climate change are very hard to quantify but could significantly shift the overall ethical calculus even further in favour of eating sardines and anchovies.
I agree this is a central question from an animal welfare perspective, and also a very tricky one.
As far as I know, there’s no direct research comparing the subjective experience of death by fishing versus natural causes like predation, starvation, and disease in sardines or anchovies. So we’re left making our best-educated guess, based on what we know about the mechanisms and likely duration of each kind of death. Being chased, caught and then dissolved in a predator’s stomach, or slowly dying from starvation or disease, is almost certainly prolonged and stressful. Purse seine fishing, by contrast, seems to involve a much shorter timeframe. While it’s obviously still bad, my current best guess is that it may involve less total suffering than most natural deaths in the wild.
That said, I’d love to see more research on this and am very open to revising that view.
Another relevant point from an animal welfare perspective is that choosing sardines and anchovies over foods that have a bigger environmental footprint helps preserve biodiversity and ecosystem resilience, which are crucial for the long-term flourishing of both wild animals and humanity.
It makes sense to ask whether the points about low bycatch, minimal seabed damage, and reduced plastic pollution from purse seine fishing apply more broadly than just the specific fishery mentioned in the cited study. As far as I can tell, they do. Because of the shoaling behaviour of sardines and anchovies, purse seining is ideal and by far the most cost-effective way to catch them at scale. That’s why the vast majority of the global sardine and anchovy supply appears to be caught this way. There may be rare exceptions, such as small-scale artisanal fisheries using line and hook fishing, which likely has an even lower negative impact, or rare instances of trawling, which may occur when the fish are unusually deep and in regions with weaker regulatory oversight, but these seem to account for only a tiny fraction of the global catch. That said, I’m not an expert on global fisheries and would be very happy to hear from someone who is.
Regarding mussels and oysters: I think many of the arguments here also apply to them, and I see no ethical concerns with their consumption. However, there are some health-related reasons to be cautious about eating them too frequently. As filter feeders, they can accumulate heavy metals and microplastics, especially if sourced from polluted waters. Nutritionally, sardines and anchovies offer even greater benefits, particularly higher levels of EPA/DHA and calcium, which makes them a better overall choice in that regard. They also tend to be significantly more affordable.
A Vegan Case for Eating Sardines and Anchovies
While Kant’s ethics doesn’t logically reduce to consequentialism, the categorical imperative seems to rest on assumptions about long-term outcomes. Kant’s insistence on a universal prohibition against lying appears grounded in the belief that a strict norm of truth-telling creates a more stable and morally reliable society, even if it leads to worse outcomes in rare cases. So while consequences aren’t the explicit justification, they seem to determine the principles we find reasonable and can will to become a universal law.
Parfit’s distinction between agent-neutral aims (e.g. maximise happiness) and agent-relative aims (e.g. care for your family) strikes me as more semantic than substantive. All moral reasoning depends on the agent’s situation, and identity (like being a parent) can be viewed as part of that situation. Take Peter Singer’s drowning child for example: the moral responsibility to act arises because you are standing there and you can save the child. That situational fact is decisive, much like being a parent is. In this sense, even utilitarianism relies on agent-specific facts, making it functionally agent-relative. I’m not sure any moral theory is truly agent-neutral in practice.
Thanks for this thoughtful piece. I really appreciate the pragmatic framing and the effort to look beyond the binary of vegan vs. non-vegan.
That said, I’d like to challenge the heuristic that the bigger the animal, the better, based on the idea that fewer individuals are killed per calorie. While it’s true that, for example, chickens’ small size increases the number of animals killed, higher-order consequences might shift the overall moral cost calculation.
Compared to chickens, cows live much longer, may suffer more intensely, produce more greenhouse gas emissions, and require far more land, contributing to biodiversity loss and causing additional animal deaths through crop production. So, when factoring in these broader effects, chickens might actually be the preferable choice, even putting aside that they arguably offer better nutrition per calorie.
Curious how you see these downstream consequences factoring into the argument.