October 5, 2018
Should certain highly intelligent animals, like chimpanzees, orangutans or elephants have legally recognized rights? And if they did, would it improve human morality? Tyler John thinks so, and in this talk from Effective Altruism Global 2018: San Francisco, he explains why. A transcript of Tyler's talk is below, followed by questions from the audience.
So I take it all of us are seasoned veterans of EA, and so all of us know that the strongest arguments are based on nothing more than flimsy anecdotes. So I'd like to start with one of those. In May of 2018, New York state's highest court denied the Nonhuman Rights Project leave for appeal, refusing to review an earlier decision made by a lower court that did not grant a writ of habeas corpus to two chimpanzees, Tommy and Kiko. Following a five year campaign between the Nonhuman Rights Project and each of New York's four intermediate appellate regional courts in which these courts, too, refused to grant a writ of habeas corpus relief to Tommy and Kiko, two chimpanzees, which would release them from confinement and move them into sanctuary. And so this means that the Nonhuman Rights Project needs to push their case again before the court.
Tommy and Kiko have to wait even longer to be moved from captivity into sanctuary, and moreover New York's case law remains unchanged. However, along with this decision came an opinion from Judge Fahey, which was really quite striking in its content, and the opinion stated that the decision not to grant leave to appeal to the Nonhuman Rights Project was not based on the merits of the Nonhuman Rights Project's claim. In fact, moreover, he argued that the legal system really needs to grapple with questions of whether nonhuman animals can be granted a fundamental right to liberty and can be graned a writ of habeas corpus for relief from confinement. And more striking still, he argued that the status of nonhuman animals as property under the law is a manifest injustice and that while it may be arguable that a chimpanzee is not a person, there's no doubt that it's not merely a thing.
This came after an earlier decision in the intermediate appellate courts from Barbara Jaffe, which was also striking in its content, arguing that whether one is a human is not relevant to whether one is a person under the law. And that ultimately questions about personhood will be decided not by facts about biology, but facts about public policy and principle. These opinions create precedent within the court system and greatly increase the probability of the success of future personhood initiatives to grant fundamental legal rights for chimpanzees and other animals.
So what I'd like to talk about today is how we can evaluate the altruistic efficacy of campaigns such as these, and how we can start to think about comparing them to other kinds of interventions we can form that have caused more direct measurable change in the world, as well as other interventions which have a much more long-term systemic structural impact in the world, just like these kinds of campaigns. There's a lot of inherent uncertainty when we're thinking about shaping the far future with structural distributed interventions. And I think that there's room for a lot of reasonable disagreement with these issues, but that by reasoning qualitatively about some of the considerations that are in play, we can make progress and think about how to evaluate some of these kinds of interventions.
So, I will be reasoning qualitatively for the case for long-termism, the case from long-termism to moral circle expansion and the case from moral circle expansion to personhood initiatives.
The key players right now that are working on personhood initiatives can be subdivided into two different kinds of approaches. Some go through the court systems which are seeking to pursue writs of habeas corpus, which release individual, particular animals from situations of confinement and establish their fundamental rights, and then legislative initiatives which seek to change the laws in a whole region for a whole group of animals, directly. And so, on the court side of things, the Nonhuman Rights Project, who I mentioned at the beginning, is working in the US. In addition to the work that they've been doing for chimpanzees, they've recently launched an initiative in Connecticut, trying to gain a writ of habeas corpus for three elephants who are in the Connecticut zoo.
The AFADA in Argentina recently, I believe in 2015, won a writ of habeas corpus for an orangutan and later won a writ of habeas corpus for a chimpanzee. Currently the state is trying to decide what the content of these rights will be, and how to enforce them, so there's some uncertainty about how things will proceed. It looks like a very promising move for at least these particular animals, if not for the majority of them in the country. And currently AFADA is pursuing another writ for three chimpanzees who are at an eco-park in Argentina.
The Federation for Indian Animal Protection Organizations only recently started moving on personhood initiatives in India, but they are gearing up to pursue personhood for elephants, and look to be using litigation strategies as well. On the legislation side,
Sentience Politics, which is a spin-off organization from the original larger Sentience Politics working primarily on political campaigning in Switzerland, is pushing for the fundamental rights to bodily liberty and mental integrity for chimpanzees in the state of Basel, and they're doing so through a citizens' initiative. So, like, a ballot initiative where citizens go and vote on an issue on the ballot, to decide whether they want to establish these basic rights for chimpanzees throughout the state.
So, legislation, as I mentioned, directly shifts laws regarding particular groups of animals in particular districts. Litigation, however, only concerns particular animals. And so what litigation does is it sets precedent for further litigation and for legislation, and it also legitimates positive legislative ideals, sort of blessing them before the law so that the law can take further steps. And what both of these strategies do, in addition, is they force the state to think less simplistically and more complicatedly about rights law in general, clearing up the confusion about what the purposes of rights are, what they're doing, how they protect individuals, that can pave the way for much more nuanced and expansive protection in the future. That can cover a wide variety of sentient beings in a wide variety of circumstances. It allows the state to digest some of these ideas, which, which will allow for the progress in the future.
For those who don't know, long-termism is the view that the primary determinant of the value of actions we take today is the effect that those actions have on the very long-term future. And I think that most people in this audience, most people in the world, when they reason on what they care about, are committed to some form of long-termism, because the vast majority of beings who our actions will affect are going to be in the future, and because future beings have welfare that is no less important than that of present day beings. If you don't think your values commit you to that, that'd be surprising. We can talk about that in the Q&A or talk in office hours. But I think that's something that many of us are going to find ourselves committed to.
Moral circle expansion is an alternative or complimentary strategy to AI alignment, which is often also motivated by long-termist considerations. I say it's complimentary because in order to get the best outcomes that we want, we need to pursue both of these strategies. We need AI alignment to ensure that artificial intelligence has human values. But we need moral circle expansion to ensure that human values are good values that we actually want to see instantiated in an extremely powerful being who could greatly shape the trajectory of the future. You might work on moral circle expansion if you think that the far future is not already likely enough to be positive in expectation. You also might if you think that for reasons of tractability, you, a particular group or all of us could do more to improve the value of the far future by improving its expected quality rather than by extending its expected longevity.
Many EAs who work on moral circle expansion are particularly concerned with risks of astronomical suffering, or s-risks. These are events that would bring about suffering on an astronomical scale, vastly exceeding all suffering that's ever existed on earth so far. These include possibilities such as creating a large number of suffering sentient beings for recreation, labor for scientific experiments and prediction models, blackmail, revenge, justice, religion, and even pure sadism. All the, like, nightmare scenarios that you can possibly cook up are thrown in here. Many but not all s-risks have much higher probability if we have an artificial general intelligence that is insufficiently aligned with values that would be ideal, the kinds of values we try to cultivate within the effective altruism movement. But again, not all of these outcomes are contingent upon any kind of AGI takeoff. And some of them can exist without that.
People who are pursuing moral circle expansion try to shape the future to decrease the probability of these s-risks and extremely bad outcomes, and increase the probability of very good outcomes by expanding and improving the values that humans and future generations will have.
Rights legislation can perform three different functions. It can perform a deterrence function, it can set precedent, and it can also change values through a process that, I'll call it institutional signalling. And I'm going to talk about each of these in turn.
Deterrence is the standard boring role that we see the law playing, and that all of us believe that the law plays at least to some degree in some contexts, which is disincentivizing people from doing bad things, by threatening them with punishment. If we had the right kinds of legislation, it could disincentivize people from harming sentient, nonhuman beings. And while this is a boring humdrum kind of role for the law to play, it is an important one. Because if AI take off is sufficiently slow, there may be opportunities for significant state regulation of AI. And if that's the case and we have sufficient moral circle expansion work, we could ensure that the kind of regulation that we have protects more sentient beings than just humans, and more sentient beings than it otherwise would protect if we did not seek moral circle expansion. Also, seemingly any future action which could cause the risk of astronomical suffering could be disincentivized through deterrence efforts.
Particular legislative acts in the law, and particular legal interventions, also set precedent. In common law systems like the United States, all published decisions in judicial opinions become part of the body of law that constrains the way that judges rule. And this is why the precedents, or the opinions I discussed the beginning of this talk, are so important: because these too become part of the body of law, and the body of decisions that constrains how judges can rule. This is true even at the highest levels. So, for example, in the US, the Supreme Court's decisions are significantly shaped by precedent. The Supreme Court decides which cases they're going to hear, and will not hear cases where they think that there's been a sufficiently good ruling on them already. They won't pick up the cases, dive into them, analyze them, discuss them amongst themselves. And so legal precedent can shape what the Supreme Court hears and does, and it can also make certain laws immutable with increasing time, institutional reliance and political acceptance. Some laws just become too deeply ingrained to be very easy to do something about, through precedent setting.
And of course, the decisions of courts and other branches of government significantly shape each other's agenda. What one branch does can change what's actionable for another branch, and change what's relevant for another branch. In fact, when courts decide in favor of ascribing certain rights to individuals, such as in the case of Sandra the orangutan in Argentina, oftentimes what the court system will do is they'll hand off the legislation to another government agency who will decide what content this law should have, and how they should employ it. And so this then can change the legislature's priorities as they think about what the structure of these rights should be. And it can also produce positive legal content for the legislatures to use in their legislation.
And so in all these cases, whether we're using legislative actions or litigatory actions, decisions that courts and other branches make will set precedent for future decisions and allow us to have better, more progressive, more effective laws in the future.
Finally, and I think most importantly for our purposes, the law can provide institutional signaling that can greatly shape human moral attitudes and behavior. To see if institutional signalling matters, we should ask whether common sense morality is better than it once was because the laws changed, or whether the laws changed because common sense morality is better. Do people discriminate less today because discrimination is legally prohibited? Or does the law prohibit discrimination because we discriminate less, and our values have gotten better? So I think obviously, the the question as posed is not a good one. We would expect the relationship between the law and morality to be bidirectional. Morality influences the law and law influences morality. But I'm going to be arguing, giving some reasons to favor the latter thesis, to think that the law can significantly change people's morality.
So one way in which the law can shape human morality and moral behavior, is by being a persuasive source of morality for people to uptake. Because society is diverse and different people are offering you different moral judgments all the time: your peers, your parents, your faith communities and political parties, and everyone around you seems to disagree about morality, the law may be a particularly powerful source for shaping and sustaining moral norms. Since law is a common denominator that holds across all citizens, unless a particular issue is very divisive and contentious in the law. But to the degree that it is a common denominator among citizens, the law may be a very powerful source for shaping morality. And if this thesis is right, then the informational influence of the law is likely to be a heuristic process, affecting our system two reasoning, such that the informational content of the law is invoked in our intuitive common sense moral judgments that we make in everyday cases.
Now, I think this view is plausible on reflection, but not a lot of research has been done to understand it and see whether this is actually a way that morality is shaped. The evidence is pretty thin. The two best known empirical studies, which are studies in which people were asked to judge the morality of a bunch of unsavory behaviors, are highly inconclusive. In these cases, what the study participants were asked is whether certain behaviors were right and wrong, and then they were told the action was either legal or illegal, and there was some change in people's perceptions of the morality of these actions based on whether they were legal or illegal. They were more likely to judge them as immoral if they were illegal, as one would expect. But as Nigel Walker, one of the study authors himself has pointed out, a lot of this data might be explained in terms of the fact that there were some conscientious law abiding citizens in the room who thought that some actions that are illegal can be wrong in virtue of the fact that they are illegal. So, when certain people are told that an action is illegal, they then decide it's morally wrong because breaking the law itself is morally wrong.
So, we don't have a lot of evidence for the persuasive source of morality view of the law. But these particular studies, and a bunch of other studies, also do show a much stronger effect of perceived peer opinion on morality. So one thing that these studies did show was that when they, in a separate variable condition, when participants were told that their peers strongly believed that certain actions were moral or immoral, this greatly updated people's perceptions of the morality of these actions. And so I think that the law can play a very effective role in shaping moral attitudes and behavior by representing group attitudes as being a particular way.
Because social norms strongly anchor human moral attitudes and behavior, the law can shape morality by shaping what's perceived as group norms and attitudes. This can happen directly when group members infer that a certain level of momentum must exist in order to support the change that has happened. We can see this in two separate studies that Tankard and Paluck recently performed on Supreme Court rulings on gay marriage. Here they found robust evidence that the Supreme Court ruling legalizing gay marriage was associated with significant shifts in perceived present and future social norms in support of gay marriage.
So, these two studies, the second study was a study in which they performed five separate Mturk studies. The first three studies happened before the Supreme Court ruling in favor of gay marriage and the last two Mturk studies happened after it, and they found that study participants evaluating the morality of gay marriage and evaluating the positivity of gay rights updated strongly on pure opinion after the Supreme Court ruling on gay marriage. They strongly improved their belief that the public supported gay marriage and was moving in a direction that was more supportive of gay marriage.
A different study used a different technique, and what they did was they gave fake articles to study participants that were either predicting a really strong landslide win in favor of gay marriage or one really strongly against gay marriage and evaluated candidates' attitudes towards gay marriage and gay rights in this context. And this one found that people who were given a positive ruling in these articles not only perceived a much greater status quo norm in favor of gay rights and much stronger directional norms in favor of it, but their personal attitudes towards gay marriage were also much more positive, and their feelings towards gay people on a feeling thermometer were also much more positive.
So this study strongly suggests not only that the law can shape what's perceived as pure opinion and group norms, but also people's own moral attitudes. Flores and Barclay similarly found a massive increase in pro gay-rights attitudes immediately following an earlier Supreme Court ruling, which importantly didn't have any backlash either. And I won't go through this study very much, but you can see looking at the transitional probabilities that in a number of states there were very, very large updates from people who are opposed to gay rights to becoming ambivalent about gay rights. In some states, 24 percent of people who were opposed became ambivalent, and in some 47 percent who opposed became ambivalent. And there was also a minority of participants who updated in favor from being ambivalent about gay rights to favoring them as well. And so here we see just a really massive transition in pro-rights attitudes following a Supreme Court ruling that legalized gay marriage in some states and increased protections for people in gay partnerships.
The law can also represent group attitudes more indirectly: when people change their opinion, change their behavior to accord with law, their cognitive processes are opaque. People don't know why they're acting this way, but, but they assume automatically that the reason people were acting this way is because they want to. Because that's what good people, responsible people do. And so when people discriminate less, or they eat less meat, or perform any behavior that may become incentivized by the law, people assume that they're doing that because they want to, and that's what good people do. And over time people begin to see a world free of certain behaviors as normative, not because the law says so, but because others are doing that.
I won't go through all of these conditions under which norms and behavior shifting seem more powerful, but some of these I think are particularly important for our purposes here. So, the law can be more influential on norms and behavior when, number three, individuals' personal views are closer to the new normative information, which I think points positively in favor of a goalpost shifting strategy, which starts with more low hanging legal fruit. When we update society's morality in a direction that's favorable to the rights of nonhuman animals, we can then pursue increasingly more progressive legal initiatives. Taking people's personal views iterated closer and closer to the more progressive legal wins. For when new normative information is widely shared, which I think suggests the importance of mass media mobilization surrounding legal campaigns, and six, when political movements coexist to mobilize individuals following a decision, which I think suggests the importance of large amounts of grassroots mobilization around legal initiatives.
And so all of this suggests that we should be pluralistic in our approach to advocacy, employing legal interventions alongside grassroots, mass media mobilization, and through a combination of precedent setting and norm and attitude change, iterated person initiatives and other legal interventions can progressively shape the law and moral attitudes, establishing increasingly progressive basic rights for sentient beings. And we can expect this in turn, for the reasons I've discussed, to increasingly expand humanity's moral circle. So I've got some conclusions here, but I've already run over by a little bit. So I think I'll just start to move towards Q&A, and let you read the conclusions on your own. Thanks very much.
Question: Do you think that these initiatives work in concert with incrementalist legal approaches like the current California ballot initiative, which maybe you know about or should they be prioritized above other laws?
Tyler: I think it's not a binary sort of question. I mean, my suspicion is that they do work in concert. Welfare laws do increase people's moral sentiments towards nonhuman animals and create momentum for further legal wins. I think there's increasing consensus in the effective animal advocacy movement that that is indeed the case. And for that reason, we can expect these to increase the likelihood of success of establishing basic legal rights for nonhuman animals. So I think they do work in concert. But there's still a question of which of these strategies should we be prioritizing more heavily. And that's a question that I feel ill-equipped to answer. But we should be prioritizing both of them much more heavily than we currently are.
Question: Are there risks with personhood being legally expanded to include animals, if there are times when significant welfare improvements for animals would come into conflict with those laws? For example: say you give personhood rights to dolphins, then maybe the conditions surrounding that would require you to factory farm more fish to feed them.
Tyler: Okay. I think you're going in a different direction with this. A lot of these basic legal protections are in favor of things like autonomy and bodily liberty. If we give a large amount of autonomy and bodily liberty to nonhuman animals, where you leave them too much to their own devices, to make their own decisions, it gives us fewer grounds for paternalism and supporting their welfare against their own preferences. And so, liberal rights could also come into conflict with welfare in that case. I think if this is the kind of idea we have here, we do need to balance the autonomy we give for indirect moral circumspection kinds of reasons, and the direct welfare benefits of paternalistic interventions. And I think this is a question just to be worked out in the, in the decision about what kinds of laws we codify and how much paternalism we allow, given the basic rights that we ascribe to nonhuman animals.