Hayley Clatterbuck

109 karmaJoined


It's possible that invertebrate sentience is harder to investigate given that their behaviors and nervous systems differ from ours more than those of cows and pigs do. Fortunately, there's been a lot more work on sentience in invertebrates and other less-studied animals over the past few years, and I do think that this work has moved a lot of people toward taking invertebrate sentience seriously. If I'm right about that, then the lack of basic research might be responsible for quite a bit of our uncertainty. 

Hi weeatquince,

This is a great question. As I see it, there are at least 3 approaches to ambiguity that are out there (which are not mutually exclusive).

a. Ambiguity aversion reduces to risk aversion about outcomes. 
You might think uncertainty is bad because leaves open the possibility of bad outcomes. One approach is to consider the range of probabilities consistent with your uncertainty, and then assume the worst/ put more weight on the probabilities that would be worse for EV. For example, Pat thinks the probability of heads could be anywhere from 0 to 1. If it's 0, then she's guaranteed to lose $5 by taking the gamble. If it's 1, then she's guaranteed to win $10. If she's risk averse, she should put more weight on the possibility that it has a Pr(heads) = 0. In the extreme, she should assume that it's Pr(heads) = 0 and maximin. 

b. Ambiguity aversion should lead you to adjust your probabilities
The Bayesian adjustment outlined above says that when your evidence leaves a lot of uncertainty, your posterior should revert to your prior. As you note, this is completely consistent with EV maximization. It's about what you should believe given your evidence, not what you should do. 

c. Ambiguity aversion means you should avoid bets with uncertain probabilities
You might think uncertainty is bad because it's irrational to take bets when you don't know the chances. It's not that you're afraid of the possible bad outcomes within the range of things you're uncertain about. There's something more intrinsically bad about these bets. 

Hi Edo,

There are indeed some problems that arise from adding risk weighting as a function of probabilities. Check out Bottomley and Williamson (2023) for an alternative model that introduces risk as a function of value, as you suggest. We discuss the contrast between REV and WLU a bit more here. I went with REV here in part because it's better established, and we're still figuring out how to work out some of the kinks when applying WLU.

Thanks for your comment, Michael. Our team started working through your super helpful recent post last week! We discuss some of these issues (including the last point you mention) in a document where we summarize some of the philosophical background issues. However, we only mention bounded utility very briefly and don't discuss infinite cases at all. We focus instead on rounding down low probabilities, for two reasons: first, we think that's what people are probably actually doing in practice, and second, it avoids the seeming conflict between bounded utility and theories of value. I'm sure you have answers to that problem, so let us know!

Thank you so much for this comment! How to formulate hierarchicalism - and whether there's a formulation that's plausible - is something our team has been kicking around, and this is very helpful. Indeed, your first suggestion is something we take seriously. For example, suffering in humans feeds into a lot of higher-order cognitive processes; it can lead to despair when reflected upon, pain when remembered, hopeless when projected into the future, etc. Of course, this isn't to say that human suffering matters more in virtue of it being human but in virtue of other properties that correlate with being human.

I agree that we presented a fairly naive hierarchicalism here: take whatever is of value, and then say that it's more important if and because it is possessed by a human. I'll need to think more about whether your second suggestion can be dispatched in the same way as the naive view.

Thanks for your comments, Nick. 

On the first point, we tried to provide general formulae that allow people to input their own risk weightings, welfare ranges, probabilities of sentience, etc. We did use RP's estimates as a starting point for setting these parameters. At some points (like fn 23), we note important thresholds at which a model will render different verdicts about causes. If anyone has judgments about various parameters and choices of risk models, we're happy to hear them! 

On the second point, I totally agree that welfare range matters as well (so your point isn't nitpicky). I spoke too quickly. We incorporate this in our estimations of how much value is produced by various interventions (we assume that shrimp interventions create less value/individual than human ones). 

On the third point, a few things to say. First, while there are some approaches to ambiguity aversion in the literature, we haven't committed to or formally explored any one of them here (for various reasons). If you like a view that penalizes ambiguity - with more ambiguous probabilities penalized more strongly - then the more uncertain you are about the target species' sentience, the more you should avoid gambles involving them. Second, we suspect that we're very certain about the probability of human sentience, pretty certain about chickens, pretty uncertain about shrimp, and really uncertain about AIs. For example, I will entertain a pretty narrow range of probabilities about chicken sentience (say, between .75 and 1) but a much wider range for shrimp (say, between .05 and .75). To the extent that more research would resolve these ambiguities, and there is more ambiguity regarding invertebrates and AI, then we should care a lot about researching them!