[EDIT: No longer confident in this post, see here.]
Yesterday’s post reviewed research showing that animals’ intelligence seemed correlated with their number of cortical neurons. If this is true, we could use it to create an absolute scale that puts animals and humans on the same ladder.
Here are the numbers from this list. I can’t find chickens, so I’ve used red junglefowl, the wild ancestor of chickens. I can’t find cows, so I’ve eyeballed a number from other cow-sized ruminants (see here for some debate on this).
|Animal||# cneurons||% of human cneurons||# equivalent to human|
Some animal rights activists discuss the relative value of different species of animal. You have to eat a lot of steak to kill one cow, but you only have to eat a few chicken wings to kill one chicken. This suggests nonvegetarians trying to minimize the moral impact of their diet should eat beef, not chicken. But any calculation like this depends on assumptions about whether one cow and one chicken have similar moral values. Most people would say that they don’t – the cow seems intuitively more “human” and capable of suffering – but most people would also say the cow isn’t infinitely more valuable. Different animals rights people have come up with different ideas for exactly how we should calculate this.
I wondered how people’s intuitive ideas about the moral value of animals would correspond to their cortical neuron count. I asked Tumblr users who believed that animals had moral value to fill out a survey (questions, results) estimating the relative value of each animal, in terms of how many animals = 1 human. Fifty people answered, including 21 vegetarians and 29 nonvegetarians. Their numbers ranged from 1 to putting their hand on the 9 key and leaving it there a while, but when I took the median, here’s what I got:
|Animal||# = human (cneurons)||# = human (survey)|
The middle column in this table is the same as the third column on the one above – it’s how many of each animal it takes before the group has the same cortical neuron count as a human. The last column is the survey results on how many of each animal it takes to have the same moral worth as a human.
The two are pretty similar. For example, 53 cows have the same neuron count as one human, and respondents said that 50 cows would have the same moral value as one human. Respondents placed a little less value on elephants and chimps than their neurons warranted, and a lot more on lobsters. Carl S suggests by email that lobsters might have been a bad choice: they have fewer neurons than an ant (really!) but because they’re so big everyone assumes they must be pretty advanced. Despite this, overall people’s answers were pretty close. Here’s a log-log graph:
This is a better match for the moral value data than other ways of calculating animal intelligence. For example, a cow has an encephalization quotient of 0.5, compared to a human’s 7.5. That would suggest a human is worth 15 cows, which doesn’t match the survey-takers’ impressions.
I would like to see this repeated with a larger survey of a more representative population. But for now I think this adds at least a little credibility to intuitive ways of thinking about these problems.
Here’s another animal moral survey: http://datacolada.org/2013/11/25/8-adventures-in-the-assessment-of-animal-speed-and-morality/
I thought the most interesting part of the survey was this:
“Michael Jennions is a world renowned evolutionary biologist, [and] rated based on the complexity of the central nervous system (complex = evil).”
Have we been looking at animal morality all backwards?
When “morality” is defined as “not doing evil”, then of course the simplest organisms are the most moral. Inanimate objects are even more moral.
According to most major religions (including at least Hinduism, Buddhism, Christianity, & Islam), earthly life is inherently bad, evil, or at least of no value. The Buddhist says all life is suffering. The Christian says all life is imperfect and thus not good.
All these religions begin with a concept of “perfection”, and can measure the worth of real things only as how far they fall short of perfection. This gives a scale on which perfection has zero value, and all else has negative value.
This will always be the case for any ontology which consists of instances of eternal types, and in which the value of those instances is given by how closely they resemble the definitions of those eternal types. But this is basically the definition of spiritualism: the belief that temporal objects in the physical world are instantiations or manifestations of perfect, eternal objects in a non-physical (“transcendental”) realm.
My conclusion seems to be that spirituality leads to the conclusion that life is evil.
To have a concept of morality which has a positive definition of “good”, one which doesn’t mean merely “not evil”, requires not believing in essences, prototypes, Forms, or any other eternal definition of types, because such transcendental types are always taken as the source of value and meaning, and hence require an unattainable level of perfection just to achieve non-negative goodness.
I’m somewhat surprised that the survey takers’ intuition of how complex these animals are (and therefore how much moral value they have) track this intelligence metric so accurately, when most of them presumably haven’t interacted that much with these animals. My knowledge of these animals come from the occasional zoo visit, youtube video, and popular science article. I don’t have a good sense of what intelligent behaviors a pig is capable of, but I’d cached that they were pretty smart, so I put a lowish number for pigs in the survey. I hadn’t heard anything impressive about chicken cognition, so I guessed they were one tenth as morally valuable as a pig. Etc.
So – assuming I’m a representative survey taker, and the others were similarly lacking in direct experience with these animals – I’m startled that the impressions of animal intelligence that bubble out of popular science, and into people’s minds, track a significant-seeming metric like cortical neuron count so well. My expectation is that the intelligence estimates that bubble out are skewed by the reporting layer (people want to play up how intelligent farm animals are, or exaggerate the complexity of a social behavior of a cute animal for hits) and also the research layer (human scientists may not be good at recognizing complex animal behaviors that are unlike any complex human behaviors), and what ends up in people’s heads will be distant from the reality of animal intelligence/complexity.
I’d be interested to see a comparison to a survey about the relative cuteness of various animals.
SMBC did this analysis:
The unbearable accuracy of stereotypes strikes again!
Animal % of human cneurons
The answer is obvious.
Breed dumber cows.
We’re working on it. We call it “vat meat” or similar.
Or bigger ones, so you can eat more for one dead animal.
Plot twist, factory farming was the morally superior option all along.
I’m running this on MTurk with 1000 respondents. Results should be in within 15 hours or so. Unfortunately I missed the vegetarian/vegan option, so I guess we’ll just see what the results are.
Isn’t that really expensive?
You can set your own rates. I’m offering $0.03 per survey, so with Amazon’s tax that’s $40 for the whole thing. As an update, results are coming in quite slowly (probably because, at the average 1.5 min per survey, this rate works out to barely $1.20/hr). I have about 80 results, and the new projected end date is sometime on the 28th or 29th.
Medians so far:
Chicken = 40
Chimpanzee = 2
Cow = 5
Elephant = 1
Human = 1
Lobster = 100
Pig = 6
Unfortunately, my survey design was rushed and doesn’t map well onto yours: I had a checkbox for “do you think animals have moral value” which was required to continue the survey, instead of the much clearer opt-out at the beginning of yours. However, I would think that this design error would cause people who don’t believe in an animal-to-human conversion to put in really large numbers, so it’s odd that my numbers would skew lower. Anyway, elephants = people, confirmed. I’ll update with the complete results later.
I imagine you could be getting fairly low quality responses by only paying $1.25/hr and that this could explain some of the discrepancy. I’m a part of an effective altruism research organization called [Rethink Priorities](http://rethinkpriorities.org) and we’d be interested in potentially running a scaled-up version of this study at a larger budget if other people would think that would be useful.
Are you able to determine nationality or geographic location of the people taking the survey? I’ll bet $20:an admission of a study flaw that people in India (using geography as a proxy for religion as a proxy for moral beliefs) will rate cows at lest half an order of magnitude more morally important than people not in largely Hindu locations.
*Longer Answer Than You Wanted*
I. Update on results:
The response rate drastically fell off after the first few hours, but it’s been holding steady at about 3 per hour since then: responses vs time histogram.
I have 263 responses (data) and I think that’s enough to do a little analysis. Note: I made a few manual edits to my own data set – e.g. replacing “1-2” with “1.5”. I have not included those changes in the linked file.
The medians are low.
A small percent of the responses were snarky, e.g. a human is worth “1 except for the [people group that I wasn’t aware many people were racist about?] who will never equal REAL human beings.” Fortunately medians are pretty stable to this stuff.
II. Yes, demographics are important, and no, I didn’t control for this in any way.
DeciusBrutus gets his $20. I didn’t ask for demographic information in any way. Of course, Hinduism regards cows, elephants, and monkeys to be sacred. (Indeed, members of the [a href=”https://en.wikipedia.org/wiki/Cow_vigilante_violence_in_India_since_2014″>Cow Protection Movement are in the business of killing humans who kill cattle, so perhaps they’d rate cows as having _higher _value than humans, as several respondents did!) As a rough estimate, another MTurk survey I did recently, in which I did ask for geographic information, had 11% respondents from India and 77% respondents from the US, but there’s no reason to suppose this would hold constant.
But it’s not like we can just ignore responses from specific groups. If anything, we’d want to weight people’s responses according to the prevalence of their demographic group. This is what pollsters do. I’m not a pollster, and anyway I understand that this is as much of a science as an art. Perhaps @PeterHurford’s study can do this right?
III. High-variance responses
Some responses gave roughly equal value to each species. While this could represent a principled Hindu stance, it could also represent a low-effort response and/or trolling. So what if we only look at survey responses whose individual answers had a high variance?
If we take the top 50% surveys by intra-survey variance, the median responses were:
This gets us a lot closer to Scott’s results.
IV. The Ordering Tribes
Not everyone ranks the animals in the same order. There are 7! = 5040 possible orders, and an answer could be consistent with all of those, if all values were the same (12% of responses).
So, what happens if consider each possible ordering, and count the most popular orderings?
Here’s a list of the top 30, with the number of survey results consistent with that ordering.
97: Human Elephant Chimp Cow Pig Chicken Lobster
95: Human Chimp Elephant Cow Pig Chicken Lobster
82: Elephant Human Chimp Cow Pig Chicken Lobster
75: Human Elephant Cow Chimp Pig Chicken Lobster
75: Elephant Chimp Human Cow Pig Chicken Lobster
72: Chimp Human Elephant Cow Pig Chicken Lobster
72: Chimp Elephant Human Cow Pig Chicken Lobster
71: Human Elephant Chimp Pig Cow Chicken Lobster
70: Human Cow Elephant Chimp Pig Chicken Lobster
70: Human Chimp Elephant Pig Cow Chicken Lobster
68: Elephant Human Cow Chimp Pig Chicken Lobster
65: Elephant Cow Human Chimp Pig Chicken Lobster
63: Cow Human Elephant Chimp Pig Chicken Lobster
63: Cow Elephant Human Chimp Pig Chicken Lobster
62: Human Chimp Cow Elephant Pig Chicken Lobster
61: Elephant Human Chimp Pig Cow Chicken Lobster
58: Human Cow Chimp Elephant Pig Chicken Lobster
57: Human Elephant Chimp Cow Pig Lobster Chicken
57: Elephant Chimp Human Pig Cow Chicken Lobster
57: Elephant Chimp Cow Human Pig Chicken Lobster
56: Chimp Human Elephant Pig Cow Chicken Lobster
55: Human Elephant Cow Pig Chimp Chicken Lobster
55: Elephant Cow Chimp Human Pig Chicken Lobster
55: Chimp Elephant Human Pig Cow Chicken Lobster
55: Chimp Elephant Cow Human Pig Chicken Lobster
54: Elephant Cow Human Pig Chimp Chicken Lobster
54: Cow Human Chimp Elephant Pig Chicken Lobster
54: Cow Elephant Chimp Human Pig Chicken Lobster
54: Cow Chimp Human Elephant Pig Chicken Lobster
54: Cow Chimp Elephant Human Pig Chicken Lobster
Fun stuff. Of course, about 29 of all results were consistent with _any_ ordering. (The exact number depends on how you deal with malformed responses).
Ok that’s enough of this for today.
You’d ignore Hindus (or any group) where the response is based on religious teaching rather than just a gut feeling. The whole point is how people intuitively morally rank animals. People who don’t rank them intuitively (or at all) don’t really contribute data to that question.
SMBC has an amazing take on this http://www.smbc-comics.com/?id=2393
The sad thing is that beef is the WORST for emissions.
I’ve wondered about how to factor that in. Do you have any intuition here? Can AGW and animal suffering be compared on a single moral scale, and where would the cost/benefit of eating beef fall?
How much worse?
A few hours before Scott posted this, I asked in the last OT https://slatestarcodex.com/2019/03/25/ot124-opentatonic-thread/#comment-734570 and in my deluge of questions was trying to figure out just how much worse.
Say my total GHG emissions as a middle-class American is 800 units/year. If I change 1000 calories a week from chicken to beef, how many units does that change my GHG emissions? Up to 801? 810? 1000?
My rule of thumb is beef and lamb (ruminant meat): 16-17 kg CO2e per kg, pork: 4-5 kg, chicken less than that (2-3, maybe?), and fish less than that again. Beef values vary quite a bit, the numbers are for dairy cattle (so some emissions are amortized over milk and cheese), meat cattle is worse, and third world cattle much, much worse, due to long lives being used for labor, scrawnier animals, and low-grade feed (grass vs grain, more digestive methane).
If those numbers are accurate, the impact seems quite small, given that CO2e numbers are usually given in metric tons, and offsets are usually lowish double-digit $ per metric ton.
The carbon capture of the feed equals the carbon emissions of the livestock, by conservation of mass.
That’s a pretty huge multiplier for the emissions being methane instead of CO2.
Yes, about thirty times more potent, according to a quick search. Note there’s also emissions from fertilizer and land use (e.g., N2O) and transport, and so forth. But the main difference is methane.
Typical emissions per capita in western countries are 5-10 tons. Eating 50kg meat results in 0.2 (pork) to 1 ton of CO2, so from 2% to 20% of total emissions.
People will have to decide for themselves if and what they will do about emissions, but for me, switching from beef to other protein sources is a fairly low utility loss.
Edit: US consumption of meat appears to be roughly 25kg beef, 25kg pork, 50kg poultry.
US is 20 tons per capita, so maybe I’m anchored differently.
But really, my thinking was based less on proportions than on costs. If the social cost of a ton of CO2e equivalent is ~$50, the GHG emissions of 1kg of beef is ~$1. Even if you take the higher end estimates of ~$200/ton, the cost per kilogram is ~$4. If you compare prices for purchased offsets (and you believe they work as advertised, which I admit is a big ‘if’), the cost of offsetting it is even smaller.
And the sticker price of beef is probably $5/kg more than chicken. It makes a difference on the margin, obviously, but if you think the sticker price difference is insignificant compared to the animal welfare gain, the GHG emissions should be in the same boat.
I also found 20 tons when researching. (As of 2010.) This nice size chart includes diet:
I think beef is around $4/pound for me (I’m not sure, because I only look at it when it’s marked on special, and this is something that I have only started paying real attention to), while chicken is $2/pound. I saw Perdue has their “usadaprocessverified” seal which sells for $4.50 a pound.
I thought 50 cents would be worth it, but it may not actually get us anything useful:
This says that they aren’t cannibals, but cage-free just means not in cages, not necessarily with access to outdoors. So I’m not sure I’m getting anything better here.
Perdue has their “Harvestland” sublabel which is $5.50 per pound, which is the same “vegetarian-fed” combined with “free range.” Free range is not defined, even in a footnote. (It’s non-GMO certified, which I generally consider an active negative.)
This is interesting, and makes me wonder if I’ve been overweighting the difference in intelligence between cows and chickens in my own dietary decisions. But, I don’t think my own intuitions agree with a weighting where number of cortical neurons corresponds linearly to moral weight. Partly, my intuitions are calibrated on an approach to human moral value which I think many people would probably regard as highly offensive. But, while the value of a human life can probably treated as a specific individual value based on averages for most practical purposes, if I had to weight different individuals’ lives against each other on a pragmatic basis where I knew that my decisions would have consequences, but I wouldn’t be judged for my judgments, I think I’d have to say that the differences in moral value between people can be quite large. I know a number of people who’re mentally impaired in various ways to a sufficient degree that they’ll never be able to take care of themselves, or understand how other people feel about their interactions with them. Obviously, these people have a different instrumental value to society than more capable people, but if I had to judge the value I assign to them based on their inner life or spark, in a scenario where people were all being watched over by ever-loving machines or something, I’d have to admit I assign considerably less weight to them than to an average person. But, the difference in their number of cortical neurons compared to an average person is probably pretty small. How smart are they compared to elephants or chimpanzees? Honestly, probably smarter in some ways, dumber in others; a sufficiently dysfunctional human may be a lot less intelligent in some respects than a fully functional chimpanzee. But in some respects, they’re clearly a lot smarter.
It’s hard to say how much I should weight the difference between a chimpanzee and a human based on that gap in intelligence. 2.6 chimpanzees to an average human honestly seems too low. 10 to 1 feels possibly too high, but probably closer to appropriate based on my intuitions. Maybe around 8 to 1. But in that case, how much do I weight the much larger proportional gaps between something like a cow and a chicken? I honestly don’t know. I can’t calibrate that according to experiences with people as dumb as cows. I feel like the difference should probably be larger than the direct proportion of their number of cortical neurons, but it’s hard to say by how much.
You will be free once you realize that there is no actual right answer out there in objective reality and that these considerations live only in your mind.
Of course I’m aware of that, but that doesn’t generate conclusions on how I ought to behave based on considerations I care about.
Of course there is a right answer. Morality by definition tells you to do one thing and not another. It’s just that the right answer is hard, maybe impossible, to determine.
There is no objective morality.
Moral intuitions are the result of partially biological and partially cultural heuristics which evolved to produce functioning societies in a given environment.
Right. Weighting the moral value of animals starts from the premise that animals have moral value, a premise that people accept (fully or conditionally) or reject for essentially hand-wavy reasons. Some people are perfectly happy to stop there, naturally on a rationalist blog people are going to want to rationalize those reasons, but it doesn’t make them any less handwavy (for either side).
For half the article I was chanting in my head “log, log, log”. And than it maps nicely on a log function – differences in value are more than proportional to differences in neuron count. But… it seems not as much as you’d expect, which really sinks my boat: I was always in favor of eating chicken because, well, dumb dinosaur, while cow is smart cousin mammal. But it takes quite a bend in the log function to get to the almost 1000x difference in weight, and what I’m seeing here doesn’t support it.
So does this update estimates for how intelligent dinosaurs were? Do we know enough to be able to say whether their brains likely had similar neural density to modern birds?
I was under the impression birds have such a high density because they minimise flight weight.
Most dinosaurs wouldn’t have that constraints.
It would be interesting to compare flying to flightless birds.
Birds manage to pack so much intelligence in so little space because they are weight constrained, dinosaurs that didn’t fly but occupied similar ecological niches of the early birds (imagine some reptilian squirrel-lemur-monkey, which eventually evolved into archaeopteryx) would have been less weight constrained, but would have likely still benefited from high intelligence.
How does this add credibility to anything? How is the median of stranger’s impressions a salient data point?
Also, I think you’re catching an even stronger correlation, which is that organisms “care” more about their closer relatives. If you want to differentiate versus your hypothesis, you should survey feelings towards cuttlefish and other highly-intelligent species distantly related to humans
It would be interesting to include domestic pets in the survey. I suspect this will reveal most people’s inconsistent moral intuitions if we consider that dogs’ intelligence is similar to those of pigs. Which moral intuition is the ‘correct’ one? The intuition about the moral value of pigs or of dogs? What about horses vs cows?
Throw in ravens, octopi, and whales. Ideally, let people set error ranges to indicate their own uncertainty, but that takes a higher quality of survey participant than I expect.
At that point I think you would want to switch to forced-choice comparisons. I would distrust any attempt to elicit cardinal ratings from people for such an esoteric and abstract topic.
When graphing, show the middle quintile instead of the median value. ‘Chickens’ would be a bar from 2.3 to 3 (200 to 1000 or 10^2.3 to 10^3) on the X-axis.
Lobsters then go from 500-10,000, and overlap with chickens.
Strangely, the middle quintile was much more consistent among themselves about how various animals compared in moral worth to chickens than about how they compared to humans. When comparing to humans, the middle quintile thought that 30-100 cows was worth one human, but extrapolating their answers making the standard assumptions, the median quintile of ‘how many chickens is a cow worth’ ranged from 0.2 to 0.46. Pig:Chicken was even tighter; 0.1 to 0.15 pigs per chicken, but 18-100 pigs per person. Chimps and elephants were much more consistent with people than with chickens; 4-10 and 3-10 chimps per person and elephants per person, but .01-.038 and .01-.06 per chicken. And with lobsters there are 1.3 orders of magnitude difference of opinion within the middle quintile of people comparing lobsters to humans, but merely a 3x difference between the middle quintile of people comparing lobsters to chickens.
(Methodology: Acquire everyone’s implied valuation in terms of chickens rather than people by dividing the number of animals that they said were equivalent to a human by the number of chickens they said were equivalent to a human. For each implied valuation in chickens, sort and select the middle 8.)
Analysis: I think that people are more consistent at making comparisons between things that are closer together, and their disagreement grows more than linearly with differences in scale. There appears to be a consensus about among survey participants about the relative scale of farm animals to each other, nonfarm mammals to each other, and lobsters to chickens, even though there is no consensus about the scales that those groups have to each other.
> This suggests nonvegetarians trying to minimize the moral impact of their diet should eat beef, not chicken.
That might be the correct choice in the short-term if you’re only measuring how much intelligent matter you’re destroying to feed yourself today. But cows have a higher carbon footprint than chicken -> climate change -> future impact on more intelligent matter.
If we’re counting future impact, is there any metric more important than your total carbon footprint? Is “number of people tortured” a metric that is even on the same scale as ‘carbon footprint’?
There’s a match with intuitive moral perceptions, but it doesn’t really say that much about how it feels to be dumb and suffer. We tend to think that kids are worse off when it comes to emotional damage, which naturally has to do with the potential trajectories they will land on, but intuitively it also has the component of kids not being able to rationalize and integrate their suffering to a more coherent whole. Dumb animals can be thought to have similar issues.
Also, well… are we going to generalize this to humans of different intellect? With humans there are obvious differences in intelligence (at the ends of the curve, at least), and we don’t need proxies such as the number of cortical neurons to approximate a person’s moral value. This might have a profound impact on how we allocate scarce resources on, say, healthcare, education and social aid, as we could determine a person’s moral worth from their intellect.*
Or am I misunderstanding something?
(*I do not believe this is the case)
Pretty neat, though I’m not sure exactly what neuron-counting adds when we already have behavioural tests for specific mental abilities that are morally relevant.
Which cow-sized ruminants did you use? The only cow-weight ruminant I see on that list is the giraffe, which has 1.7 billion cortex neurons. Both giraffes and cows weigh roughly 1500-2000 pounds.
The greater kudu weighs a lot less (~500 pounds), but at least is in the right ballpark for head-and-body length (6-8 feet, compared to 8-8.5 feet for cows), and they have ~750 million cortex neurons.
Blesboks are much smaller in terms of both length (5 ft) and weight (160 lbs). Same with springboks (4.5 ft/100 lbs). But even these have 550m/400m cortex neurons.
I feel like given that the greater kudu is quite a bit smaller than the cow, a better estimate for the cow would be ~1 billion cortex neurons. This puts it back at ~1/15 of a human, in line with the encephalization quotient.
I used the springbok and blesbok, then assumed cows would be slightly dumber because domestication. I was encouraged to see this was slightly dumber than pigs, which fit my priors.
In retrospect, I could have used the kudu or giraffe, so now I’m not sure.
https://www.frontiersin.org/files/Articles/88283/fnana-08-00128-r2/image_m/fnana-08-00128-g003.jpg gives us a scaling rule for artiodactyls, but the pig is an outlier because it’s domesticated, so I would expect cows to be a similar outlier. Also, I am getting weirdly inconsistent results on how much cows weigh. But it doesn’t seem consistent with 1 billion plus.
EDIT: I found this blog post, which says: “Neuroscientist Suzana Herculano-Houzel has performed numerous studies to determine neuron count in various species, and argues that neuron count better tracks cognitive abilities than brain size. She provides an estimate of 85 billion neurons for a human brain, and 221 million neurons for the Red Junglefowl, a wild relative of domesticated chickens. She also studies scaling patterns of brain size and neuron count in various taxonomic groupings, and in her book estimates that a cow would have ~3 billion neurons.”
This is total neurons, not cortical. But if we assume related animals have the same total/cortical neural balance, then 3 billion neurons is between the springbok and the blesbok, so my use of them to estimate cows seems probably right.
Got it, thanks. Yeah, 3 billion total neurons would put it right under the blesbok (by 2%, so basically the same). Does domestication affect total neurons, or just cortex neurons? Given that cows are much bigger than blesboks but have the same number of total neurons, it seems like it affects both? I.e. you shouldn’t discount it again when calculating cortex neurons, so it feels like 550m would be a better estimate for a cow’s cortex neurons. But maybe I’m wrong.
Also, for pigs – this alternate source seems to say that pigs only have 307m cortex neurons: https://www.karger.com/Article/Pdf/437413 (Table 1, Sus scrofa domesticus, column “N, n”).
So I would guess that at a minimum, cows have more cortex neurons than pigs, although maybe not that much more.
Maybe this is obvious, or maybe it is irrelevant, but the process of domestication and breeding tends to change traits dramatically, behavior very much included. So I think it is likely that intelligence of domesticated animals can’t reliably be extrapolated from wild relatives.
We expect domestication to make them dumber, so estimating from wild relatives provides an upper bound.
Makes me wonder. From my totally anecdotal and unscientific experience, there is a large variance in the intelligence of dog breeds. Some are able to learn amazing feats, while some appear to have affectionate slobbering as their main if not only mode of behavior.
Is my intuition wrong? And if not, are there substantial anatomical differences in the brains of dog breeds?
And: what if we bred for intelligence…could we breed an animal species to become smarter than us?
It’s an interesting post, but I think the human tendency to equate how bad it is to mistreat an animal with its intelligence is deeply unethical. The right metric should be their capacity to suffer, and it’s not clear at all that this has much to do with intelligence. I don’t think anyone seriously argues that unintelligent humans systematically suffer less than intelligent ones when they are harmed. It’s just a convenient story people tell themselves to feel better about the misery we inflict on animals. I also think intelligence has little to do with a being’s capacity to feel pleasure/joy, which must certainly factor into its “moral value” as well.
Not sure capacity to suffer is the right metric either. Because in that case, we could just make eg humans without the capacity to suffer, and eat them.
That’s the reason for all the discussion of meditation on this blog.
I would argue that certain kinds of suffering require a certain amount of intelligence. For example, hopelessness requires the ability to foresee the future and to empathize with the future you.
Furthermore, lower intelligence makes it easier to prevent suffering due to boredom, since it makes it easier to entertain.
My intuition is to disagree, but I think you make an interesting argument about humans. I’m really unsure whether less intelligent humans might suffer less. But I suspect that below a certain intelligence, less intelligent species do suffer less. Would you agree that bacteria have very low capacity to suffer? How about non-living things like electrons?
My problem with this approach is it makes moral consideration into a scale. Instead, I believe it is a threshold condition. Animals that cross the threshold are morally-considerable, those which don’t aren’t, relative intelligence doesn’t enter into it. It makes much more sense for a question as fundamental ethically as “Does this thing have a right to live?” to be a threshold question and not a sliding scale.
Professor Richard Hanley suggested not killing animals which had future-oriented mental states (they have a concept of self and would actually miss living) and not torturing animals capable of pain (a much lower threshold of intelligence). That seemed like a reasonable framework to me. A convenient one, since it allows people to continue eating Western diets provided the farming is humane, but it at least justifies the decision on where to set the thresholds in a reasonable fashion.
Not sure I agree with this. Consider this thought experiment:
Suppose an adult chimp is smart enough to pass whatever your test for moral worth is. But an infant chimp is pretty dumb and probably won’t be.
Either there is some kind of gradient, or there is some specific point at a chimp’s life – let’s say when it turns twelve months old – when it becomes able to pass the test.
Would you kill 9999999 eleven-month-old chimps to save one twelve-month-old chimp?
(if this thought experiment doesn’t satisfy you for one reason, change bits around until it does)
I would like to say, as an aside, that your last line is an excellent extension of the principle of charity to debate (especially on the Internet), and should be doctrine throughout the world.
That said, it should be noted that infanticide is often a separate crime to murder, with a lesser penalty (though infanticide usually requires you be basically post-partum depressed).
I hope it isn’t CW to suggest that regardless of whether you agree with it, this is how it works in the abortion debate. At least in my jurisdiction, a fetus is considered to have no value (except to the mother) until 12 weeks, then it has moderate value (i.e., you can get an abortion if approved by a committee, which you basically get if you can provide any reason at all) until 18 weeks, and after than, it has full (or close to it) worth as a human being.
As to cj’s argument, I think an interpretation could be that worth is assigned to a species (based on some perceived average or representative properties), and all members of the species get the same value. This can be extended to species groups, e.g., whales are intelligent, therefore particular whale species is intelligent, therefore particular individual whale is intelligent.
Not saying it is rational, but humans like to stereotype things and organize into categories.
Scott’s objection is reasonable, but I favor either applying the moral status to the entire species, or granting moral status to any animal which will acquire a self-concept in the future (because killing it before it becomes self-aware deprives it of the value it would later derive from its existence when it becomes self-aware.) The species-categorical rule also covers infant humans without a concept of their continued existence. The “future potential” approach would cover most infants, but fail in the case of infants with major deformity who will never mentally develop.
One of my biggest objections to a sliding-scale of value w/r/t life is that some hypothetical superintelligent AI might come along and say “well humans have 0.001% the intelligence and comprehension that an AI does, so it’s worth killing 100,000 humans to benefit an AI”. If you have a threshold instead of a comparative-worth scale, then all animals meeting the basic threshold condition are morally-considerable and have a right to life, regardless of how superior one species might be to another. You can still have crazy lifeboat hypotheticals between morally-considerable individuals and species, but crossing the threshold should be enough to say that in a vacuum 1 member of species A isn’t “worth” 10 members of species B by relative merit, no more than a person with a 150 IQ is “worth” two with an 85. The purpose of ethics in the broader sense is to program society, and if the natural right to life isn’t equivalent across all individuals who possess it, that leads to some nasty things and nasty people.
Anatoly Karlin had a lengthy post about exactly this a few weeks ago.
So, out of curiosity, according to this metric, how valuable is a human infant compared to an adult?
I think the same – we’re born with most of our neurons, they just grow new synapses.
Does this commit you to any moral position regarding fetuses? Does anyone have data on how many neurons there are at different stages of development?
Edit: So according to not particularly reliable sources, fetuses produce about 250,000 neurons per minute from 8 weeks onwards (this gets you to around the correct range of 100 billion at birth). Let’s assume that 20% of those are cortical neurons. That means we are getting 72 million cortical neurons a day from 8 weeks onwards.
So based on this, a fetus reaches the moral value at the following times:
Lobster 8 weeks
Chicken/pig/cow 8-9 weeks
Elephant: 19 weeks
Chimpanzee: 20 weeks
This means that the distribution of abortions is heavily distributed towards lobster level moral harm, with almost none reaching above chimpanzee.
I think this post may hit the sweet spot of making everyone dissatisfied on both sides of the debate.
I think the criterion is how much utility per year does the human have. An infant may not accrue much utility at first but there’s the expectation that it will get much more during it’s adult life.
Note that we should have the same reasoning for cows: a calf doesn’t accrue as much utility as an adult cow, so the percentage of its life spent as a calf vs an adult is going to change the result.
Also relevant: I think the criterion that Scott is using is “how much total life utility are you removing from the word when you kill one animal”. And a common counter argument is that the cow wouldn’t be alive (and accruing utility) in the first place if it wasn’t needed for its meat.
So to maximise utility, you might want to raise an animal that has a lot of neurons, and only kill it when its about to die naturally.
… And I’ve just realised that my argument leads to eating old people.
This would be a nice solution to wealth redistribution, political turnover, and minimizing healthcare costs. But what does one do about prions?
Well, if you’re waiting until people are about to drop dead, you may as well wait a few moments and let them die naturally. And I don’t think there’s any utilitarian reason that it would be unethical to eat human corpses.
I’m not positive that counting future utility is a good idea. This relates to the repugnant conclusion, i.e. if you value potential human life so much, do you have a responsibility to make babies constantly? (Or maybe you accept the repugnant conclusion, in which case good luck.)
OK but I really would question whether people would really go through with this.
Say, you’re truly given the choice between 51 (or 54) cows dying and a human dying ? Do you really pick the human for death? Really? I wouldn’t. Similarly, assuming you eat 2 steaks/pieces of cow a week [conservative estimate but probably around what I am averaging in my specific present conditions], how much time does it take to reach 50 cows and thus invalidate my existence? Chicken’s equivalent might be even faster. Yet no way I’m committing suicide/jailing myself once I’ve reached that number…
I personally feel this whole line of enquiry, while interesting, is a dead end.
I eat [cow/chicken] meat b/c I find it pleasant – even if I know it’s an immoral choice. OTOH, I wouldn’t eat whale meat. B/c whales are key elements of keeping our oceans alive. Indeed their impact on the environment is so profound that I would gladly trade the lives of [random/not-personally-known-to-me] humans for a whale (maybe, up to 50 of them? No rational for that number, just “more would be greedy” and “possibly more of these humans might do enough good to compensate for the whale” – even if, in reality, it’s likely to be the opposite i.e. nothing can be better for Earth than getting rid of a maximum of humans).
You’ll likely eat around 10-15 cows total over your entire lifetime, so you won’t reach 50 cows.
Yeah, the chicken is what gets you in trouble…back of the envolope, a conservative estimate is about 1,500 chickens over a 60 year period for the average american, more when you consider that the weight of chicken meat consumed requires more than the equivalent weight in live chickens (I would guess close to double and again, 1,500 was intentionally conservative).
I guess the good news is that only comes out to murdering 5 people in your lifetime…
Hmm… I wear leather too.
TBH, IMHO, the only way out of this conundrum is either to become vegetarian or, assuming most of us can’t be arsed/find it too demanding, going for vat grown meat/leather in the near future.
It so happens that, in my professional capacity, I had the opportunity to invest in such start-ups. It’s too early to say if anything will come of it but that’s my solution to the ethical issues created by meat/leather consumption.
I think you’re observing that society is not based around this blog’s opinions.
I’m not sure whether society should treat 50-something cows as equal to a human life, but that’s because people are better at rule-based, intuition-inclusive ethics than utilitarianism.
I remember from reading @DavidFriedman that the weregeld for a death in many societies is actually on the order of magnitude of 50-100 cows.
Of course, that’s a utilitarian/commercial approach, not a moral one – but this makes me think… aren’t we answering the question from a utilitarian approach by habit? I mean, over many generations humans actually had to make this choice: do you save 1 human, or X cows? And losing 50 cows could easily mean losing more than 1 human long term, so you’d probably go for the cows.
That’s pretty neat. Though there’s still a difference between ’50 cows is repayment for 1 human life’ and ‘choosing to kill 1 human instead of killing 50 cows’. Killing any number of humans violates a sacred value of society, and I’d be nervous about messing with the sacredness of that value.
If you do this, then this actually makes the situation twice as bad, because you’ve just killed 50 cows and 1 human, so you’ve got the right idea. Sunk costs save the day!
Good point indeed. TBH, the same applies to climate change – no way am I shooting myself regardless of my carbon footprint… 🙂
I don’t think these are the values most people use when they actually have to make such decisions (weighting animals more heavily than real life). How many cows were killed on the chance that they might have contracted mad cow disease, on the chance that it could be spread to one human, for example? How much money does society spend to save the life of one individual human with a terminal disease, vs. one individual cow? These are the numbers that seem more likely to get at our real values to me.
Society isn’t even consistent about the amount it spend on human QALYs or lives.
See eg the recent outrage over some plane crashes, yet people still drive cars.
Another good point indeed. Matthias is also right but I think that goes also to the fact that we humans are piss poor at statistical reasoning and risk-based decision making i.e. we will overemphasize risks with tiny odds vs. ones with far bigger odds if the resulting deaths are seen as more spectacular and/or more outside of our own control (supposedly).
One way I think about car deaths vs. plane deaths is : 1- plane deaths tend to be all or nothing i.e. either everyone survives or everyone dies. Car crashes have a wider distribution of consequences. You can get away with being ‘just’ wounded. And 2- I can drive a car and thus do something about whatever situation arises. In a plane, I’m utterly powerless.
I think most people feel those 2 points above instinctively.
We should avoid jumping to conclusions. Most likely moral significance will ultimately be based on capacity to suffer and experience joy (echoing zooid’s comment above), which might be much more related to the limbic system, pleasure centers, and thalamus, than to cortical neurons. People’s perceptions and intuitions are likely to be biased towards over-valuing intelligence because that’s something that got ingrained into our perception of value due to sexual and kin selection.
If QRIs Symmetry Theory of Valence is correct, most intense and valuable/disvaluable experiences will be very simple (but high-energy). Complexity of neural architecture is correlated with potential for intelligence and complex states. But not necessary for emotionally powerful experiences. The cortex, after all, plays majorly an inhibitory role in the brain, whereas emotion centers are evolutionarily ancient and phylogenetically preserved across the animal kingdom. e.g. Even octopodes enjoy MDMA. As John Lilly’s studies indicate, thickness of the cortex is correlated with emotional control. A macaque in panic is more impulsive and violent than a chimpanzee in panic, than a human in panic. So, sadly, there is a good chance that pigs, dogs, birds, and cows experience emotions more, rather than less, intensely. Their emotions may not be as ‘subtle’ and multilayered, but what does that matter for ethics? Raw panic is worse than subtle poetic melancholia and other ‘valued (i.e. fetishized) human emotions’. That we overvalue such subtle emotions may be a side effect of the specific sexual selection pressures experienced in our recent evolutionary history.
I would offer the suggestion of enforcing pleasure center activation in factory farm animals as a precondition to having them for meat, the same for animal testing for things other than the brain. Why not wirehead rats who are being studied for kidney failure? Won’t matter for that area of research.
Another note concerns our bias towards cute mammals because of mirror neuron activation. Really we need what Mike Johnson has pointed at in his blog for ages: objective measures of species’ specific valence landscapes. Hunch: the life of some animals is simply not worth living no matter their living conditions. Crickets- the consciousnesses science of 2050 could reveal- are 100% driven by dissonance and never experience pleasure. If so, let’s fix them or phase them out.
Also consider: Cluster headaches are 10,000sX more painful than other conditions. Likely, there are kinds of ultra-high-intensity suffering lurking in species’ specific diseases. Morally, it ought to be key to identify these cases and treat them as the priority.
So, if these numbers are broadly correct, they seem to make a strong case for veganism.
Take chickens, for example. We’ll say that a chicken has 1/500 the moral value of a human, as per the numbers above.
There are plenty of stats on the internet describing how many chickens are killed each year in the USA — the number is around 9 billion. Here seems like a neutral source:
9 billion / 500 = 1.8 million. So, the US chicken insustry is killing the moral equivalent of 1.8 million people a year.
The same page puts the number of pigs eaten per year above 100 million. That’s more than 2 million people (at 1 pig = 1/35 of a human).
So, moral value of eating pork and chicken (in the US) comes to around 4 million human lives per year. That’s more than the total death rate of the US (2.7 million per year, source).
So, if these numbers are to be believed, ending the slaughter of pigs and chickens would have a greater moral value than eliminating death itself. Even if we err on the side of caution, and assume the values have been inflated by an order of magnitude (making a pig 1/350 of a person, instead of 1/35 of a person), then ending the slaughter of animals is roughly morally equivalent to eliminating cancer.
So, either the moral case for veganism is unassailable, or these numbers are completely inflated. (Personally I’m in the former camp.)
I think it’s a mistake to assume that “moral value” is equivalent to “unwillingness to slaughter for food”.
Consider a hypothetical conversation with the Prime Minister of Cows where we offer a deal. We’ll treat you and your kind well for two years—tasty food, good pasture etc—then at the end of two years, we get to (humanely) kill and eat you. The *alternative is that we don’t have cows any more. Would the cow take the deal?
Perhaps the cortical neuron number makes a case for ethical farming rather than no farming at all?
* The deal would be different for pigs and chickens because they might choose to live in the wild whereas cows don’t really have that choice.
It would be different again for antelopes: “We’ll keep you safe from lions for two years, etc…”. I expect the Prime Minister of the antelopes would take their chance with the lions.
Or the entire approach is a non-starter, because that’s not how people’s moral intuitions actually work…
I appreciate the potential importance of coming up with a coherent moral system that is friendly to human preferences; the trouble is that as far as I can tell the actual practice of the field is drunks looking under a lamp-post for keys that may not exist at all and if they do are so unfathomably complex we would have almost no chance of recognizing them as keys, much less using them. Occasionally, someone picks up a stick and is very excited about it.
Why should we be more willing to take as evidence of human values what people in the survey say, versus the actions people actually take? Based on the actions they take (eating meat), most people apparently place very little moral value on the preserving of an animal life. It seems to me the only reason society kills fewer cows than chickens is that the cows have more meat on them: for most people, based on their actions, the moral value of even a cow’s life is near zero. (They are very inconsistent in their behavior towards pets, though.)
By this same metric, most people also apparently place very little moral value on the preserving of a human life.
The main difference between how humans and non-human animals are treated appears to be the distinction we make between harming humans and allowing humans to come to harm (and the taboo against the former but not the latter) but this distinction itself is not compatible with the concept of moral value.
Right — this was kind of my point. There is a tension between the numbers above, and the actions people take in practice.
At face value, the numbers above suggest that ending animal consumption is a gigantic moral imperative lying somewhere between ‘curing cancer’ and ‘stopping the holocaust’. Even if we are far more conservative, and adjust the numbers downwards by a further two orders of magnitude (making a pig 0.01% of a human), ending animal consumption becomes merely a moral imperative on the level of ‘preventing all automobile-related deaths’.
Animals are slaughtered for food in such numbers that ascribing them even the smallest amount of moral value causes the utilitarian scales to tip massively in their favour. When asked in a non-food context, people are more than happy to allow animals some non-zero amount of moral value. But in practice, the moral value of food animals is set to zero.
It’s a little like the ‘drowning child’ argument for utilitarianism. You have the small-scale moral code everyone agrees to when spoken about out of context (you should jump in and save the child’s life, even at a small cost to yourself). But in practice everyone throws this utilitarian calculus out the window when it comes to dealing with the vast amount of suffering just over the horizon.
The EA approach is to argue that the small-scale moral calculation is correct, and the fact that this gets ignored in practice is a failure mode to be fixed.
This suggests that most people do not run on utilitarian morality. Dictionary sorting comes to mind: acts (or outcomes, or people) are judged based on their impact on the innermost moral circle, in the same way dictionaries are sorted by their first letter; if two courses are tied (or irrelevant) in the innermost circle, only then does the second circle out come under consideration.
This seems mostly accurate to me, although there’s definitely fuzziness when it comes to circle boundaries: “existential threats to family” > “existential threats to stranger” > “inconvenience to family” can be described as “family before strangers” or “life-or-death before inconveniences”.
Or your reasoning/morality doesn’t actually match how most other people value things.
In medicine they tried to account for this with the QALY, where quality of life is given weight, not mere existence.
The cows, sheep, etc that I see grazing outside seem to be fairly content with their lot in life, not showing any obvious signs of stress. I don’t feel that bad about eating them. The main thing I’d like to see changed is for them to be given more opportunities to find shade in the summer, as they do seem to quite dislike being in the sun all day, as well as improvements to transport conditions.
None of these considerations are reflected in a simplistic ‘life of being X is worth Y.’
The answer probably doesn’t change the math much, but how much of that is actual human consumption (vs. other uses like, say, pet food or other non-human food products)? My intuition is “most”, but my confidence is not terribly high.
The link to the questions in Scott’s survey doesn’t work. It goes to a page that just says the form is closed. Can you point us to a copy of the questions somewhere?
I can’t quite make a logical connection, but this post reminds me of a folk song that holds a calf responsible for its slaughter for failing to fly away:
It’s a beautiful song, whether or not one considers the Holocaust allegory. Here’s Joan Baez’s version: https://www.youtube.com/watch?v=j1zBEWyBJb0
Maybe it’s not “intuitive” moral weights, but to me it makes a lot more sense to use the “opinions” or “judgments” of people who have thought a lot about it than to use the general public perceptions. It sounds like something similar to the difference between expert consensus and public opinion.
Sure, but the there is a heavy bias in that sample. It depends on what question you are trying to ask and what you are hoping to do with the answer, I suppose.
Yes, I undestand there is bias. What I’m saying is that it might be the same kind of bias that happens when you ask the question “Is climate change man-made?” only to climate experts instead of asking to the general public. To me, it sounds like asking the question on the post only to people who think a lot about it is very similar to asking the question about climate change only to experts. It seems to me that it’s the good kind of bias.
Retracted? This didn’t seem serious enough to retract. I guess this would have been a good use for an epistemic status comment to communicate to me that Scott was serious about it.
Since the actual numbers aren’t necesarily all that meaningful – if person A thinks a human is worth 20000 chickens and 1000 cows, person B says 5000 chickens and 500 cows, I don’t think that necessarily is a meaningful difference. Instinctively (i.e., I haven’t thought it through before posting), I feel it might be a better approach to take the mean of the ratios between classes. The fact that A thinks a cow is worth 20 chickens, while B thinks a cow is worth 10 chickens seems more relevant than the precise numbers.
In that case, I find the following median ratios:
One chimp is equal to one elephant is equal to 5.33 pigs, a pig is equal to 1.38 cows, a cow is equal to 4.29 chickens, and each chicken is roughly equal to 3000 lobsters.
The cneuron count is 0.9 chimps per elephant, 14.6 pigs per chimp, 1.39 cows per pig, 6.04 chickens per cow, and 500 lobsters per chicken.
And now I wonder why I went to this trouble. Is it a better match? Seems to do quite well for chimps/elephants, and for pigs/cows/chicken. The discrepancy is that humans overrate other mammals vs chimps and elephants (maybe interpreted as a combined class of animals judged less valuable than humans irrespective of intelligence), and for lobsters (which perhaps are far enough out that they don’t count much at all – or maybe it’s just that very large numbers tend to be picked more randomly).
This is exactly and explicitly how I decide what animals I can eat. Though I also do not eat octopus or squid due to uncertainly about how to count them and them seeming smart.
I confess that I don’t really take ecology or morality into account regarding my diet and that me currently being vegetarian and semi-vegan has more to do with the cost of buying and storing meat than any sense that eating animal flesh is morally inferior to eating the flesh of plants(I can buy six cans of beans for the cost of one pound of meat from the deli, don’t need to refrigerate the beans, and the beans go further than the meat even eating larger portions, and I’m lucky if I can spare 100USD per month for groceries, which makes the canned beans the much more attractive option for protein that’s fairly tasty).
That said, I do find this debate kind of interesting and don’t feel like there’s the demonization like you get with stereotypical “meat is murder” type moral vegetarians.
Not sure I buy the argument someone made that the number of slaughtered cows and chickens is equivalent to 4 million murders. After all, those cows and chickens persumably make up a significant portion of the food the ~327 million people living in the US eat each year.
Admittedly, most Americans probably eat more than they need to survive, but if you take the percentage of food eaten in the US that comes from beef and chicken and multiplied it by the population, you’d get a first approximation to theh number of lives sustained by those slaughtered animals.
Though, has anyone actually tried calculating how long it takes the average person to eat a cow? I don’t know how much cows tend to weigh or how much of their waight is useable meat, but I would think it would take a while to eat a cow even if you ate beef at every meal, averaging three meals a day and a pound of meat per meal(which I understand to be a rather hearty appetite).
I also kind of wonder how many acres of grass it takes to feed the average cow from weening to slaughter and how many acres of various high-protein crops it takes to equal the protein content of one cow.
Also, in the event that everyone became vegan or the production of vat meat reached a point that vat meat was indistinguishable from bio meat, had a lower environmental footprint, and was cheaper than bio meat, I’d expect most livestock species to become endangered relatively quickly. Sure, some people would probably continue to keep chickens or sheep as pets and some livestock might be able to find a niche in the wild, but if we weren’t raising them for food, humans would have little incentive to protect them from predators, provide any kind of medical care, or even feed them, and many would probably be reclassified as pests. You might have fewer chickens being slaughtered, but I suspect feral chickens would, on average, have much less food an a much shorter life expectancy before being mauled by a wolf compared to the farm chcicken that can expect to be slaughtered. And that’s assuming domestication hasn’t rendered livestock completely incapable of living in the wild.
About 490 pounds of beef per cow, according to the Internet.
Assuming you ate 4 oz per day, it’d take you 5.36 years to eat a cow. Eat a full pound per day and it is, of course, only 1.34 years.
EDIT: as an example, Canada’s beef industry produces 1.3 million pounds annually (though that may not all be edible). Two thirds of that is consumed in-country. That’s approximately 2.86 billion pounds. About 87 pounds per person per year.
The older I get the less and less convinced I am but a EA/”shut up and multiply” mindset. Does anyone look at this line of analysis and change anything? Or do they just like to circle-jerk about whether the relationship between neurons and utilons is linear, and then continue on eating factory farmed meat just the same? The people who are convinced are those who have seen a video of slaughter and found it emotionally compelling enough to render meat distasteful. I tell people I’m “eating less meat”, but it feels like empty social signalling. I am a hypocritical, virtue-signalling monkey.
Is it possible to see the wording of the questions in the survey anywhere? The “questions” link doesn’t display them anymore now that the survey is closed.
If a cow has 1/50th the number if neurons of a human I would expect it to take about 50 billion cows to equal the moral worth of one human, at 1/500th the number of human neurons in a chicken I would expect the moral value of a human to be equivalent to around 3 solar masses of chickens, and at 1/160,000 the number of human neurons in a lobster I wouldn’t expect there to be enough mass in the universe to add up to the moral equivalent of a human if it were entirely lobsters.
Why in the world would this scale anything like linearly?
So what you are saying is that the universe is made of lobsters?? Is that what you are saying??
I’m very sorry if I sound like a jerk, but so much of this discussion seems to be riddled with a bias called motivated reasoning. I mean, most of the people here must understand that they don’t come up with these convenient explanations why it’s moral to keep eating all sorts of meat out of pure innocence.
I recall Yudkowsky claiming on FB that he had a parts list for what makes for a conscious animal capable of suffering in the moral sense of the word. When asked for this parts list, he pretty much responded that he is tired of explaining complicated things to people who cannot follow complicated reasoning, and that he’ll give the list assuming the person who asked the question had read this and that about evolutionary biology, IIRC.
All the same, what struck me was that whatever his parts list was, it’s quite a wonderful coincidence that by his definitions humans are capable of suffering, and perhaps monkeys – which, I assume, he wouldn’t care to eat. So it just happens that the very exact parts list which he assumes to give a creature the capability to suffer excludes those animals he likes to eat.
And Yudkowsky wrote a series of books on the subject of biases.
The whole premise here seems to be that intelligence serves as a good metric for determining the moral value of an entity. Does this idea not seem to have some very troubling implications to anyone else? Specifically, in what it implies about how much moral value different human beings have?
So… here’s a slightly different way of looking at things, trying to tease apart the confounding factors.
Imagine you are tasked with educating an AI railway points switch (to remove the immediate agency from your choice). You teach the AI to to throw the switch to protect the most valued entities from a runaway trolley (yes, it’s a trolley problem). Do you load in these tables so that the switch will protect a single human for anything less than 160,000 lobsters? Or less than 320 chickens? If not, why not?
I wonder how much of this variance is “similarity to humans” rather than “how intelligent are they”. Chimps are the closest, while lobsters are very different, chickens closer than lobsters but further than cows, etc.