Thank you for this high-quality and thoughtful post. I found myself agreeing with most of the key claims.
Overall, I’d actually rather use the term “moral circle” less, and instead focus on finer-grained consequences of work to benefit groups like nonhuman animals or artificial sentience. [...] These all point to important flawed assumptions in standard framings of the moral circle, and may merit retiring the term “moral circle” altogether to prevent confusion.
I strongly agree here. I think it’s a little dangerous that the “moral circle” was introduced as a metaphor, but it has been unwittingly applied as an actual political strategy (“moral circle expansion”) without going through the intermediate process of rigorously testing whether a “moral circle” is actually a consistently measurable thing that exists (whether in an individual’s mind or society’s broad attitudes).
For example, I’ve noticed in my work that some people assume that “moral circle expansion” is a benefit of some animal advocacy campaigns (e.g. fish welfare) and not others (e.g. dog welfare). For any proposed psychological or social construct, wouldn’t we need to subject that construct to rigorous testing before making it the target of political and advocacy strategies? It might be more constructive to discuss these views (as you suggest) in terms like “advocating for fish welfare, rather than dog welfare, might have longer-term effects on people’s attitudes/behaviours towards other animals like shrimp”, which is more clearly a hypothesis that can be debated, tested, and falsified.
I’ve noticed in my work that some people assume that “moral circle expansion” is a benefit of some animal advocacy campaigns (e.g. fish welfare) and not others (e.g. dog welfare).
I think the main difference is fish are not considered worthy of significant moral concern by most people, who view them more as living objects. With companion animal species, at least in many communities it is understood that their interests are very important. This doesn’t prevent there from being serious welfare concerns involving them, but I think these are usually more a symptom of insufficient awareness and action to address those concerns, rather than a denial those concerns are valid. So if we fix people’s current valuation of animals’ interests but consider a world in which we are much better able to put our values into effect (as may be the case in some futures), then companion animal species but not fish would hopefully be fairly well off. Therefore, if values might be locked-in at the level of which species matter, it seems important that we act to extend concern to as many species as possible (ignoring backfire risk).
Caveats to the above:
In some communities it is less common to view typical companion animal species as worthy of significant moral concern.
Some campaigns to improve fish welfare might not result in significantly more moral concern for them. Probably depends on the exact target and conduct of the campaign.
If we’re not expecting even a “soft lock-in” at the level of which species matter (where “soft lock-in” = it becomes way less tractable but not impossible to spread concern for neglected species in the future), then maybe this work is not time-sensitive.
If instead a lock-in occurs at the level of more abstract values, we might prefer to spread a value of “care about all of sentient life”, and as long as evaluations of sentience are not extremely value-laden this may be enough to ensure good worlds in situations where human values dominate the future. Then spreading concern for fish is important mostly insofar as it reinforces this abstract value of sentientism. Maybe advocacy for companion animal species can further sentientist values as well.
As you point out there is also the potential for secondary transfer effects where expanding concern to one additional species/type of entity increases concern for others. My impression is that the significance of this effect as regards nonhumans is debatable, but it’s been studied a little in the psychology literature (maybe see this review).
That said, I probably prioritise companion animal welfare more than most EAs! Relative to farmed animals, I think humanity might have slightly more of a deontological duty to companion animals; we have higher confidence that companion animal species are sentient in most cases; and advocacy for companion animal species seems less likely to backfire. I also care about it more from a partial perspective. Given the current distribution of resources in animal advocacy, I’d rather marginal resources go to farmed/wild animals unless there’s a particularly good opportunity to help companion animal species, but I think I endorse some level of disproportionality in spending (but a good deal less than the current level of disproportionately we see).
Yep I agree with all of this. I think the important thing for discussions like these is that, as you propose in your article, retiring the term “moral circle” in favour of more specific hypotheses (as we’re discussing) will facilitate more rigorous evaluation of claims and therefore better decision-making.
Thank you for this high-quality and thoughtful post. I found myself agreeing with most of the key claims.
I strongly agree here. I think it’s a little dangerous that the “moral circle” was introduced as a metaphor, but it has been unwittingly applied as an actual political strategy (“moral circle expansion”) without going through the intermediate process of rigorously testing whether a “moral circle” is actually a consistently measurable thing that exists (whether in an individual’s mind or society’s broad attitudes).
For example, I’ve noticed in my work that some people assume that “moral circle expansion” is a benefit of some animal advocacy campaigns (e.g. fish welfare) and not others (e.g. dog welfare). For any proposed psychological or social construct, wouldn’t we need to subject that construct to rigorous testing before making it the target of political and advocacy strategies? It might be more constructive to discuss these views (as you suggest) in terms like “advocating for fish welfare, rather than dog welfare, might have longer-term effects on people’s attitudes/behaviours towards other animals like shrimp”, which is more clearly a hypothesis that can be debated, tested, and falsified.
I think the main difference is fish are not considered worthy of significant moral concern by most people, who view them more as living objects. With companion animal species, at least in many communities it is understood that their interests are very important. This doesn’t prevent there from being serious welfare concerns involving them, but I think these are usually more a symptom of insufficient awareness and action to address those concerns, rather than a denial those concerns are valid. So if we fix people’s current valuation of animals’ interests but consider a world in which we are much better able to put our values into effect (as may be the case in some futures), then companion animal species but not fish would hopefully be fairly well off. Therefore, if values might be locked-in at the level of which species matter, it seems important that we act to extend concern to as many species as possible (ignoring backfire risk).
Caveats to the above:
In some communities it is less common to view typical companion animal species as worthy of significant moral concern.
Some campaigns to improve fish welfare might not result in significantly more moral concern for them. Probably depends on the exact target and conduct of the campaign.
If we’re not expecting even a “soft lock-in” at the level of which species matter (where “soft lock-in” = it becomes way less tractable but not impossible to spread concern for neglected species in the future), then maybe this work is not time-sensitive.
If instead a lock-in occurs at the level of more abstract values, we might prefer to spread a value of “care about all of sentient life”, and as long as evaluations of sentience are not extremely value-laden this may be enough to ensure good worlds in situations where human values dominate the future. Then spreading concern for fish is important mostly insofar as it reinforces this abstract value of sentientism. Maybe advocacy for companion animal species can further sentientist values as well.
As you point out there is also the potential for secondary transfer effects where expanding concern to one additional species/type of entity increases concern for others. My impression is that the significance of this effect as regards nonhumans is debatable, but it’s been studied a little in the psychology literature (maybe see this review).
That said, I probably prioritise companion animal welfare more than most EAs! Relative to farmed animals, I think humanity might have slightly more of a deontological duty to companion animals; we have higher confidence that companion animal species are sentient in most cases; and advocacy for companion animal species seems less likely to backfire. I also care about it more from a partial perspective. Given the current distribution of resources in animal advocacy, I’d rather marginal resources go to farmed/wild animals unless there’s a particularly good opportunity to help companion animal species, but I think I endorse some level of disproportionality in spending (but a good deal less than the current level of disproportionately we see).
Yep I agree with all of this. I think the important thing for discussions like these is that, as you propose in your article, retiring the term “moral circle” in favour of more specific hypotheses (as we’re discussing) will facilitate more rigorous evaluation of claims and therefore better decision-making.