The decision theory argument isn't just about ability to retaliate - it's about ability to engage in reciprocal decision-making and honor agreements. Most animals can't make or understand explicit agreements or intentionally coordinate based on understanding others' choices. Maybe some corvids and a very few other nonhuman animals can try to imagine our perspectives and take actions based on predictions of what we're likely to decide, on levels of abstraction that might give us some basis for ongoing noninstrumentalizing cooperation.
This matters more in our current context because:
We're relatively early in cosmic time, with vast potential ahead
Our capacity for effective coordination and decision-making is precarious and needs strengthening
Given those facts, our priority ought to be preserving and improving our ability to make good individual and collective decisions. While animal welfare matters, compromising human coordination capacity to address it would be counterproductive - we need better coordination to address any large-scale welfare concerns effectively.
Humans are fundamentally an instrumentalizing species - that's how we solve problems. Animals suffer in factory farms not because we instrumentalize them, but because our capacity for instrumental reasoning is being turned against itself through broken coordination systems. Trying to fix animal suffering without addressing this underlying coordination failure seems like palliative care for a dying civilization.
If you are interested in cooperating with nonhuman animals - say, on the theory that cognitive diversity enables more gains from trade - it would make more sense trying to figure out how to trade more equitably and profitably with whales or corvids, than treating chickens as counterparties in a negotiation.
Your hypothetical seems to be anticipating an argument that farming mentally disabled humans would be repugnant even with slightly net-positive lives, so that therefore something similar must apply to animals. Let's consider a concrete case: Someone hires a mentally disabled person to provide warmth to their blanket-averse toddler at night, instead of the 'vegetarian' solution of turning up the heat (or 'vegan' if fossil fuels aren't involved). If they don't worry much about their employee's other living conditions as long as they seem willing to perform the service, we might say there's room for improvement but not that they're doing anything particularly wrong.
Consider another concrete case, closer to home for me. My aunt had Down syndrome. From the way my father talked about it, I didn't realize until my teenage years that Down syndrome isn't usually fatal in early childhood. The state advised my grandparents that it would be better for the family if they sent her to an institution, where she died of pneumonia (i.e. neglect) a couple years later. It seems like it would've been better for everyone involved if she'd been allowed to live until 20, and then some rich person had bought her organs.
You might object that efficient farming is different from neglectful institutionalization. But my landlord works in an Amazon warehouse, spending much of his waking time having value extracted from his labor at maximum efficiency, which can be seriously physically depleting work, and he seems to be pretty cheerful and happy with his life. Or consider Foxconn, where they installed nets to prevent worker suicides - arguably worse conditions than many factory farms - yet we all use smartphones.
This isn't about catching ethical vegans in hypocrisy - it demonstrates that we can't solve these problems by drawing bright moral lines around 'exploitation' or 'farming.' We have to consciously engage with complex trade-offs to get the goods we need while improving conditions where we can. The horror at 'farming humans' seems more about aesthetic revulsion at the framing than actual welfare - when we call similar value extraction 'employment' it becomes acceptable.
There are principal-agent problems here - if the person is mistreated enough in their off hours, they might not be a safe cuddle buddy for a toddler. These are real decision-theoretic concerns. But these problems are much less applicable to factory farming.
The clearest evidence of discontent is the unprecedented fertility decline across developed nations. Humans are increasingly choosing not to reproduce at replacement rate when given modern conditions. This isn't just about discomfort - it suggests our large-scale coordination systems (markets, governments, corporations, media) are actively hostile to the welfare of the governed in a way that factory farming isn't.
Nearly no one wants to torture broiler chickens at massive and increasing scale. If we're doing that, this suggests our interlocking coordination systems are already producing outcomes severely misaligned from individual human values and preferences.
Either we restore human agency enough to avoid relying on distasteful and repugnant systems like the worst aspects of factory farming, or we lose the capacity for meaningful ethical action entirely as our systems drift toward whatever our failing coordination mechanisms were optimizing for, or civilization collapses and takes factory farming with it (along with most humans and domesticated animals). Only the first path offers hope of addressing animal welfare systematically.
The decision calculus would be substantially different if we were near the end rather than the beginning of expansion through the universe, just as one should usually focus more on improving one's own capacities earlier in life and on contributing to others' development later on.
The framing of inequality as 'systemic violence' and altruism as 'control of aggression' assumes rather than demonstrates that wealth differences primarily reflect exploitation. This fails to address the central question I posed: whether and when inequality reflects productive allocation versus extractive behavior.
While moral sensitization may affect how we feel about others' suffering, this doesn't help us understand the causes of that suffering. Defining inequality as violence or aggression is effectively a stance in favor of violence, because it makes it impossible to discuss alternatives.
Wealthy people who primarily engage in wasteful consumption become less wealthy over time. Those who maintain or grow wealth must be doing something else with it. You brought up slavery; the antebellum South required massive coordinated violence to directly maintain internal power imbalances, and state-backed territorial expansion to support its economic growth. This illustrates why we need detailed models of how extractive systems actually operate, rather than reducing everything to market mechanisms.
Enclosure acts seem like the correct analogy. And I'd say the enclosure acts and 20th century Soviet modernization were along some relevant dimensions more similar to each other than either is to a decentralization of economic decisionmaking.
The distinction I'm trying to draw attention to in this post is one between unironically believing microeconomics and modern academic finance as descriptive theories that help one interpret the environment in which one lives and has real embedded experience of - treating them as stage 1 simulacra - and, on the other hand, treating those theories as stage 3 simulacra, a bad-faith substitute for interpreting one's environment that serves to entitle one's identity to false credit and the concomitant extraction of resources from less entitled groups. The former attitude would reason from evidence of inefficiency, to predictions about profitable deals one could strike with the locals. The latter would lead to the sort of thing that actually happened. This us approximately the difference between liberalism and neoliberalism.
My understanding of the term “privatization” is that it generally refers to the voluntary sale of state assets, by the state. That doesn’t seem like quite the same thing as the state expropriating and possibly selling assets that were previously understood to be owned and administered by some smaller community within the state. Am I missing some important detail here?
The applicability to animal welfare is relatively complex, because it has to do with biases in how we project our agency onto animals when trying to sympathize with them. The applicability to global development is relatively straightforward, as frequently success is defined in terms that at least partially include acceptance of acculturation (schooling & white-collar careers) that causes people to endorse the global development efforts.
You haven't addressed my question about how this post differs from other abstract theoretical work in EA. It's a bit odd that you're reiterating your original criticism without engaging with a direct challenge to its premises.
The push for immediate concrete examples or solutions can actually get in the way of properly understanding problems. When we demand actionable takeaways too early, we risk optimizing for superficial fixes rather than engaging with root causes - which is particularly relevant when discussing preference falsification itself. I think it’s best to separate arguments into independently evaluable modular units when feasible.
I'd still like to hear your thoughts on what distinguishes this kind of theoretical investigation from other abstract work that's considered EA-relevant.
I'd like to better understand your criteria for relevance. Are you suggesting that EA relevance requires either explicit action items or direct factual support for current EA initiatives? If so, what makes this post different from abstract theoretical posts like this one on infinite ethics in terms of EA relevance?
The decision theory argument isn't just about ability to retaliate - it's about ability to engage in reciprocal decision-making and honor agreements. Most animals can't make or understand explicit agreements or intentionally coordinate based on understanding others' choices. Maybe some corvids and a very few other nonhuman animals can try to imagine our perspectives and take actions based on predictions of what we're likely to decide, on levels of abstraction that might give us some basis for ongoing noninstrumentalizing cooperation.
This matters more in our current context because:
Given those facts, our priority ought to be preserving and improving our ability to make good individual and collective decisions. While animal welfare matters, compromising human coordination capacity to address it would be counterproductive - we need better coordination to address any large-scale welfare concerns effectively.
Humans are fundamentally an instrumentalizing species - that's how we solve problems. Animals suffer in factory farms not because we instrumentalize them, but because our capacity for instrumental reasoning is being turned against itself through broken coordination systems. Trying to fix animal suffering without addressing this underlying coordination failure seems like palliative care for a dying civilization.
If you are interested in cooperating with nonhuman animals - say, on the theory that cognitive diversity enables more gains from trade - it would make more sense trying to figure out how to trade more equitably and profitably with whales or corvids, than treating chickens as counterparties in a negotiation.
Your hypothetical seems to be anticipating an argument that farming mentally disabled humans would be repugnant even with slightly net-positive lives, so that therefore something similar must apply to animals. Let's consider a concrete case: Someone hires a mentally disabled person to provide warmth to their blanket-averse toddler at night, instead of the 'vegetarian' solution of turning up the heat (or 'vegan' if fossil fuels aren't involved). If they don't worry much about their employee's other living conditions as long as they seem willing to perform the service, we might say there's room for improvement but not that they're doing anything particularly wrong.
Consider another concrete case, closer to home for me. My aunt had Down syndrome. From the way my father talked about it, I didn't realize until my teenage years that Down syndrome isn't usually fatal in early childhood. The state advised my grandparents that it would be better for the family if they sent her to an institution, where she died of pneumonia (i.e. neglect) a couple years later. It seems like it would've been better for everyone involved if she'd been allowed to live until 20, and then some rich person had bought her organs.
You might object that efficient farming is different from neglectful institutionalization. But my landlord works in an Amazon warehouse, spending much of his waking time having value extracted from his labor at maximum efficiency, which can be seriously physically depleting work, and he seems to be pretty cheerful and happy with his life. Or consider Foxconn, where they installed nets to prevent worker suicides - arguably worse conditions than many factory farms - yet we all use smartphones.
This isn't about catching ethical vegans in hypocrisy - it demonstrates that we can't solve these problems by drawing bright moral lines around 'exploitation' or 'farming.' We have to consciously engage with complex trade-offs to get the goods we need while improving conditions where we can. The horror at 'farming humans' seems more about aesthetic revulsion at the framing than actual welfare - when we call similar value extraction 'employment' it becomes acceptable.
There are principal-agent problems here - if the person is mistreated enough in their off hours, they might not be a safe cuddle buddy for a toddler. These are real decision-theoretic concerns. But these problems are much less applicable to factory farming.
The clearest evidence of discontent is the unprecedented fertility decline across developed nations. Humans are increasingly choosing not to reproduce at replacement rate when given modern conditions. This isn't just about discomfort - it suggests our large-scale coordination systems (markets, governments, corporations, media) are actively hostile to the welfare of the governed in a way that factory farming isn't.
Nearly no one wants to torture broiler chickens at massive and increasing scale. If we're doing that, this suggests our interlocking coordination systems are already producing outcomes severely misaligned from individual human values and preferences.
Either we restore human agency enough to avoid relying on distasteful and repugnant systems like the worst aspects of factory farming, or we lose the capacity for meaningful ethical action entirely as our systems drift toward whatever our failing coordination mechanisms were optimizing for, or civilization collapses and takes factory farming with it (along with most humans and domesticated animals). Only the first path offers hope of addressing animal welfare systematically.
The decision calculus would be substantially different if we were near the end rather than the beginning of expansion through the universe, just as one should usually focus more on improving one's own capacities earlier in life and on contributing to others' development later on.
The framing of inequality as 'systemic violence' and altruism as 'control of aggression' assumes rather than demonstrates that wealth differences primarily reflect exploitation. This fails to address the central question I posed: whether and when inequality reflects productive allocation versus extractive behavior.
While moral sensitization may affect how we feel about others' suffering, this doesn't help us understand the causes of that suffering. Defining inequality as violence or aggression is effectively a stance in favor of violence, because it makes it impossible to discuss alternatives.
Wealthy people who primarily engage in wasteful consumption become less wealthy over time. Those who maintain or grow wealth must be doing something else with it. You brought up slavery; the antebellum South required massive coordinated violence to directly maintain internal power imbalances, and state-backed territorial expansion to support its economic growth. This illustrates why we need detailed models of how extractive systems actually operate, rather than reducing everything to market mechanisms.
Enclosure acts seem like the correct analogy. And I'd say the enclosure acts and 20th century Soviet modernization were along some relevant dimensions more similar to each other than either is to a decentralization of economic decisionmaking.
The distinction I'm trying to draw attention to in this post is one between unironically believing microeconomics and modern academic finance as descriptive theories that help one interpret the environment in which one lives and has real embedded experience of - treating them as stage 1 simulacra - and, on the other hand, treating those theories as stage 3 simulacra, a bad-faith substitute for interpreting one's environment that serves to entitle one's identity to false credit and the concomitant extraction of resources from less entitled groups. The former attitude would reason from evidence of inefficiency, to predictions about profitable deals one could strike with the locals. The latter would lead to the sort of thing that actually happened. This us approximately the difference between liberalism and neoliberalism.
I don’t think I stated or drew this conclusion. You might be confusing it with the bit about crop sales.
My understanding of the term “privatization” is that it generally refers to the voluntary sale of state assets, by the state. That doesn’t seem like quite the same thing as the state expropriating and possibly selling assets that were previously understood to be owned and administered by some smaller community within the state. Am I missing some important detail here?
The applicability to animal welfare is relatively complex, because it has to do with biases in how we project our agency onto animals when trying to sympathize with them. The applicability to global development is relatively straightforward, as frequently success is defined in terms that at least partially include acceptance of acculturation (schooling & white-collar careers) that causes people to endorse the global development efforts.
You haven't addressed my question about how this post differs from other abstract theoretical work in EA. It's a bit odd that you're reiterating your original criticism without engaging with a direct challenge to its premises.
The push for immediate concrete examples or solutions can actually get in the way of properly understanding problems. When we demand actionable takeaways too early, we risk optimizing for superficial fixes rather than engaging with root causes - which is particularly relevant when discussing preference falsification itself. I think it’s best to separate arguments into independently evaluable modular units when feasible.
I'd still like to hear your thoughts on what distinguishes this kind of theoretical investigation from other abstract work that's considered EA-relevant.
I'd like to better understand your criteria for relevance. Are you suggesting that EA relevance requires either explicit action items or direct factual support for current EA initiatives? If so, what makes this post different from abstract theoretical posts like this one on infinite ethics in terms of EA relevance?