This essay is a description and justification of my project Compass, which can be accessed here:
https://ethanedwards.github.io/
If you like the project, think it could be improved, or have ideas to discuss about other moral technologies, please get in touch.
Spiritual Exercises
How does one become a better person? We regularly acknowledge that some people are good and that are some are bad, and that good people do good things and bad people bad. This is a key part of life and typically we agree that one should try to be good and avoid being bad. But do we really try?
It’s not really polite to point out that someone is a bad person, or that they could be better, directly to their face. Discussions of moral quality are usually reserved for mean-spirited gossip, not for critique and growth. It is exceptionally rare for someone to say “I’m currently working on being a better friend” or “I’d like to treat people better” in the same way as “I’d like to eat healthier” or “I am working on my golf-swing.” Typically whether one is good or bad is understood implicitly to be fixed either at birth or just after adolescence - one is just that kind of person and there is no opportunity for real change. If someone has committed a grave wrong, true redemption is rare. Sometimes people will express an interest in getting more involved with charity or politics, important proxies for social good, but this is usually the limit.
This has been the state of secular Western society for a while now, but it need not be the case. Certain cultures in premodern China and the Mediterranean which are unusually well documented insisted that being a good person was a skill that could be practiced just like sports or a musical instrument. Students of Confucius or the Greek philosophers were supposed to be practitioners - performing rituals and spiritual exercises to bring them closer to the ideals of a good person. Such practices were certainly not practiced by the entire culture, and the evidence is ambiguous (but worth carefully studying) for how effective they were. But bringing practice back to morality is an idea worth trying.
Learning, study, and practice have all changed dramatically in the meantime. We have textbooks, apps, exercise-programs, and much more. While there are no longer teachers of morality, the teachers of music and exercise and science and history have produced a great many tools for one who wishes to learn. The global population is more educated than it has ever been, and there is a real opportunity to try, with gentle experimental progress, to help people become better.
This project is such an experiment, one of many that might be used by a contemporary aspirant towards better moral behavior. It is intended as a regular practice, similar to the nightly reflections done by Stoics or Confucians, in order to help broaden one’s area of moral concern. It is imperfect and flawed in many ways, but I hope it can be a starting point and that some are able to use it to refine their moral vision. I am not qualified to be a moral teacher, but I hope that by sharing such tools and discussing them with other aspirants, we might all together become a little better.
The Project Itself
The project, called Compass, simply asks the user to empathize with a potential person they have not met generated at random according to a flawed statistical method which simulates the human population of the world.
The intention is that by consistent practice, the user will get a better sense of who all the people in the world are, and therefore how their actions should be guided to assist them. I usually do not consider how my actions might impact a poor teenage girl in Burkina Faso, but hopefully that can change.
Empathy and Action
People naturally tend to narrow their field of moral concern to what is immediate. For most of history this has meant their immediate surroundings, the people local to them who are the people they know the most about and have the most possibility of affecting. But now in a globalizing world, information stretches far beyond local boundaries onto the national and international level. And with this information also comes the power to act morally and altruistically towards previously unknown others. An Ethiopian in the year 1000 would have no knowledge or potential to help a starving Chinese peasant. But now that the cost of transferring money and goods is so small and the disparities between regions so great a potential altruist could do more good for someone across the world than their comfortable neighbor. And certainly we can do a great amount of harm to far away people if we do not hold them in moral regard. To do good, we need to start with moral concern for people outside of our immediate circles of friends and families.
Empathy is one origin of moral action in our lives and can be a powerful tool for stimulating more moral action. But is it the right tool? The rise of mass-media has given many examples of people suddenly becoming motivated to help distant others - including examples as diverse as starving populations in far away lands and celebrities in contract disputes. Digital media, social media and VR, have promised to create even more empathy using technology.
But empathy suffers from many imperfections as an exclusive driver of moral action. It is momentary, sensational, asymmetric, and can motivate rash action which does little good. News stories of women in peril can help motivate a humanitarian military invasion which dooms millions to decades in greater peril, all while there was a much more solvable but less sensational plight somewhere else in the world.
Moral action, especially at a distance, requires a careful and rational attitude to decisions which does not always follow empathy. One would do better to follow statistics about who is suffering the most to guide action than go entirely on immediate emotions.
But are statistics enough? “The death of one man is a tragedy, the death of millions a statistic.” Human moral motivation and math are not easily compatible, and there are some deep problems with simply replacing empathy with statistics.
First off, statistics are flawed. They are frequently inaccurate in important and systematic ways and to truly understand statistics is an art and science that must be learned through careful study and practice. Our world is not fully knowable at present and guesses and assumptions are still necessary to get a reckoning of which problems need urgent solving.
Statistics are also extremely modest measures of the actual truth. To understand an epidemic for example, it is useful to know the number of infected and their geographic distribution, but do we also capture how far the diseases has advanced in each individual? Who has symptoms that are severe during the day but alright at night? Who will survive but will suffer for long months and never fully heal? Who had to see their family members perish from a similar disease several decades before? Statistics can and should be made to try to get closer to this truth, but right now they are not available.
Data-driven understanding presents a view from nowhere, a consciousness hovering above the world with all the facts right at hand. False confidence in statistics as the whole truth of a situation can make it a moral hazard at times. Mathematical facts must be enriched with large amounts of qualitative information and practical understanding of limitations and imperfections to be truly useful for moral action.
In terms of disinterested altruism, the perfect good person would combine the deep, highly motivated, intimate and personal care we have for our friends and family with the broad and systematic understanding of humanity given by rational and mathematical approaches. They would be able to know and care for each and every person in a deep and personal way. This is perhaps best described as agape, the divine love for all of creation in the Christian tradition. This runs up against human cognitive capacity, and might not even be logically possible. But small steps are worth pursuing.
Empathy and statistics are both imperfect and have many stark incompatibilities but combined in certain contexts they can bring us closer to this ideal.
The project attempts to make a very small steps on this path. Unlike news stories or VR, it will not show any captivating or urgent images. Instead it presents a statistical picture and asks the user to create a more human truth using their own resources. Given imperfect demographic data, it asks you to imagine a career, a life, a history, a set of desires and aspirations and fears. To do this correctly, one will have to research places and peoples they do not know about. One will make many mistakes and assumptions that do not correspond to the real people out there in the world, but hopefully through practice they will become better at this, and a little closer to the truth. When situations arise that require response, hopefully the practice of getting to know these imagined others will help guide a more moral response.
Data
This project groups people based on available statistics for several parts of identity, namely sex, age, country of residence, ethnicity, religion, employment status, location, and access to electricity, drinking water, and sanitation. These categories together cannot possibly define a human being, but they can give a hint. The system would be improved by adding more, and hopefully will be in the future.
The data is drawn from the CIA World Factbook, a source chosen for convenience more than accuracy. The political policies and ambitions of the United States government occasionally bleed through. Much of the data is inaccurate or out-of-date, and many of the decisions of how to read the data are interpretative rather than absolute. Part of successful use of this project will be learning these inconsistencies and learning to think beyond them.
The project itself uses the data imperfectly. It assigns a potential person to a country first and then largely uses the demographics of that country to determine the other categories, irrespective of their dependence on each other. A Hui person in China is given a religion based on the overall patterns of the Chinese population, rather than the Hui Chinese population, so they are more likely to be Buddhist than Muslim. This is a serious flaw and should be fixed.
And there is much that the data fails to record. It records nothing about gender of sexual orientation, nothing about familial status, nothing about illness mental or physical. The user should endeavor to learn about these, how they shape lives, and how for true understanding they must be added to the picture that comes from the data.
Future
This is a highly imperfect project, and this description leaves out many crucial questions of moral and political theory. But I hope that some will use it, and that some will find it helpful in guiding their thoughts and actions. I welcome any critiques or suggestions, since it is my sincere hope that this project can be part of my own progress towards becoming better.
Simply fusing empathy and rationality is not a clear line to true moral behavior, and to reach that later goal will require thinking and techniques that go in many other unexplored directions. There is much left for us to do.