Awesome article Naryan!

Some additional Complexity / Systems Thinking Resources:

• Systems Innovation
• Complexity Weekend
• Santa Fe Institute

+1 - I also see this as an area deserving of investigation.

Thanks for the article!

It seems right to me that strategic coordination and better communication among various EA nodes is essential. I'd love to see more thinking and action on the "improving movement coordination" like you are demonstrating, so thank you. We could host some collaborative events with complexity folks at EA Toronto maybe? Though I imagine knowing something about what higher complexity strategic thinking is currently being done is a wiser first step. What does Systems Change in EA (on FB) think? The CEA def has some thoughts: https://www.centreforeffectivealtruism.org/strategy/ - or is your sense that the thoughts are still lower on the abstraction ladder than they could be?

I also think it is appropriate that many bits of the movement are only vaguely connected through the highest  order question of EA - "how do we do the most good, act on that, and learn from our actions" - yet free to propose their own subquestions and seek answers. Since much of the meta framing itself is an open question, there feels to  me to be leverage in more bottom-up processes too - a sort of "everyone get some basic skills, now go do whatever you think is best, then we'll regroup and reflect, then go out and try again" kind of approach. And of course, not everyone can or wants to be part of more involved/meta reflections. Yet overall I feel you raise a dang good question to bring up regularly. 

I'd speculate the CEA's focus on reducing risky actions in interactions with individual EA commmunities is in part a reflection of a more bottom-up orientation. If you want a group to experientially explore without anchoring them too much with your framework, you can give them tools and then step back until/unless they come to an answer you have good reason to believe is not the right one.

I particularly appreciate your thought that people can not share information or not feel their work is meaningful if not aware of a broader community strategy though. I wonder how often this happens and how... Maybe a future EA Survey could ask "have you worked on a project which you believe is helpful but would not classify as EA-adjacent". 

Thanks for this wonderful article! I absolutely agree that it would be highly beneficial to have a community that is at the intersection of EA and Complexity. I recently participated in an event, where I actually found several other EAs interested in Complexity but unfortunately I couldn't spend enough time to network with them further (I got involved in another project there).

I have also been thinking about how we may use the tools of Complexity to make EA better although I haven't been able to concretely land on anything. Here are some vague thoughts I have. I am not entirely sure if any of these thoughts are worth pursuing so tug at these threads at your own peril!:

1. I wonder if there is a possibility of creating an Agent-Based Model to understand Global Catastrophic Risks although I am not entirely sure how to go about doing this. This talk by Luisa Rodriguez here might be a good place to start. She is not building an ABM (atleast going by what she said in that talk) but the way she talks about it made me feel like an ABM could help.
2. Complexity has some roots in Philosophy (A quick Google search took me here). I wonder how the philosophy of EA and that of Complexity would work together.
3. I wonder if we can deal with flow-through effects better if we had a Complex Systems view. Is this a network shaped problem?

But these are all mostly at a 'wondering-if' stage and one would definitely need help from cleverer people to actually start some concrete work. So having a community around EA & Complexity would be highly beneficial.

Thank you for sharing your thoughts Naryan. I agree with your main point that EA approaches are mostly limited to the complicated domain of the Cynefin framework. I have felt frustration with EA often focusing on complicated solutions that are easier to quantify and implement, rather than considering the complexity of the issue and taking a more preventative approach and intervening at a system level. And I think you’re right that complexity is needed “effective coordination in service of higher-leverage goals.”

However, I think both the Cynefin framework and the MHC will be useful in achieving this aim. 

The difference between these two approaches is that the Cynefin framework is a model that helps to classify systems and respond to systems, while MHC focuses on the cognitive ability to understand complexity. Obviously, people’s level on the MHC will impact their ability to accurately determine what kind of system they’re dealing with. 

Hence, MHC is a theory to describe and classify people’s differing ability to understand and conceptualise complexity. It is not an action orientated framework like doesn’t detail what to do about complexity, like the Cynefin model does. I think this is the issue you run into in your post when the education example, you ask “Which of these levels is ‘doing the most good’?” 

Cynefin can help you to understand more clearly the area you’re trying to operate within.  To see how the Cynefin framework can be applied, it might be useful to take a look at the EU Fieldbook Managing Complexity (and Chaos) in Times of Crisis. 

Regarding your point about that “not everyone is suited to managing each level”, within the Cynefin approach people are asked ‘what can you change at your level?’ so that they take action at the level at which they operate. For example, an employee taking action within their own team, someone taking action within their local community.  

Sounds like what you’re suggesting is that we can do the most good by helping people to see things at a higher level of complexity than they might be currently disposed to. Is that correct? This is a key aim outlined in Hanzi Freinacht’s The Listening Society.

Also, I believe Dave Snowden has critiqued MHC for being information/algorithmic-centric (I guess that’s why it’s attractive to the EA scene with it’s strong emphasis on measurability at the expense of other approaches) and using complexity in the dictionary definition which does not take complex adaptive systems (CAS) into account. But I can’t find much other info on this, can anyone help?

Great post, and you have arrived at the beginning. Yes, in most cases, or perhaps all cases, in the developing world, the most effective thing one might do is to effect system change. Even in the First World, one might see a far greater return on efforts and investment if one could find and understand basic principles of human group psychology and how complex systems like those we live in function.

That said, the danger of causing unintended harm is obvious. The last 114 years have seen a number of well meaning efforts to change the system in various nations cause vast amounts of harm and the deaths of perhaps a hundred million people all told, combining the fascist and communist utopian efforts. 

Additionally, for all the talk and intention to be open minded and willing to change our views in EA, it is far less psychologically costly to change our minds on peripheral issues like mosquito nets that we likely had no opinion on before anyway than it is to change our core political beliefs which define who we are to ourselves and our friends. There will be a lot more resistance to system change in EA than there is to anything else, and if system change gains any traction it is basically a given that the movement will fracture because of it. Especially because a real grasp of system change is an entirely new political paradigm and neither left nor right nor centrist.  

Yes, we need to understand complex systems, but we need to be specific here; there are many kinds of complex systems. What we need to understand are complex systems with multiple poorly related selfish agents and nested overlapping subgroups, and to do that we need to understand how selfish agents compete with each other in complex systems, and how and why the system benefits from that competition. 

Saying that selfish are selfish seems to be no insight, but for decades evolutionary theorists wondered why selfish failed to overrun and replace all the altruists in a group of altruists, because they forgot this. Groups of altruists are host bodies for selfish, they are fitness limiting resources for those pursuing the strategy of selfish defection, and selfish, being selfish, do not want to share the group of altruists with other selfish. In fact, too many selfish in the group causes group collapse, as seen in many hundreds of communes, and there actually is no such thing as a group of selfish humans (though there can be sub-groups of selfish): that is a herd or a flock. 

A selfish actor who finds themselves alone in a group of altruists will maximally exploit them. However, when multiple selfish actors are in a group, they compete with each other, attempt to remove each other from the group, attempt to win the group over to their side against the others (for example, one such process is what we refer to as "politics"), and attempt to thwart the ability of the other to selfishly exploit the group in a process that has been labelled "selfish punishment." The result is that the selfish strategy is self-limiting; selfish do not overrun groups of altruists because selfish are selfish and compete with each other. 

Understanding selfish competition, we should expect that more competition between selfish will be beneficial for the group, and less competition between selfish will be harmful for the group. When we look at other complex systems with multiple poorly related selfish agents, this is what we see over and over. Gut microbiota with more diversity and more competition result in healthier humans who live longer on average.  Ecosystems with competition within niches are more stable and fertile and robust. Economies with more diversity and more competition within industries are more stable and grow faster. Political systems with more competition for power are more stable and provide more benefits for the society. This is why Democracy is beneficial, this is why the division of powers is beneficial, this is why anti-trust laws are beneficial. On the other side, a lack of sufficient competition is dependably harmful. Monopolies and cartels are harmful, dictatorships are harmful, systems that combine political and economic power are harmful, monocultures are harmful, invasive species are harmful, economies dominated by one industry or by one resource such as oil are harmful (and usually not Democracies, competition between economic and political forces is extremely important), and this is why. 

A property of complex  systems with multiple poorly related selfish agents then is that more competition is beneficial to the system or group, and too little competition is reliably harmful. Based on this knowledge we can see why Communism and Fascism and Monarchies and dictatorships and Anarchist and Libertarian ideas should be dismissed from the political debate; they will always result in harm because they reduce or eliminate competition in either political or economic spheres, or both. We can also see that efforts to increase economic diversity and competition in developing nations, for example, could result in long term system change in the most positive way, as diverse economic interests seeking to protect themselves from political power is how Democracy arises. 

This then could be the strongest and most beneficial lever to grasp for: it is hard to envision the developing nation that would object to efforts aimed at helping to create a more thriving economy, and long term that thriving economy will tend strongly to create the fertile ground for Democracy. One way we can do this is to help protect developing nations from the exploitation and dominance of large multi-national corporations, and this would be a highly effective project, in my opinion, for EA. Yes, this is the opposite of "free trade," but this is where the facts and theory lead us, in my opinion. 

For much more details and proper citations and support for my ideas, read my paper on competition here: theroadtopeace.blogspot.com

In contrast to "Do the most good", which I see alongside "Compassion" and "Loving Kindness", I see clearly the two wisdom traditions I have leaned into: Zen, the Soto School of Dogen Kigen (aka Dogen Zenji) and the Vajrayana of Karma Kagyu. (More precisely, the Mahayana on which those practices rest.)
What I see, rather than the high virtues of heroism, is a simpler set:
First, "Cease from evil" ... only then "Do good" and only ultimately "Do good for others".
And then, in the tradition of prajna/upaya  and bodhisattva aspiration: solidarity/empathy.

So much easier to markety glory and fame! so much better to adopt the practicality of reasonable humility.

--Karma Chöpal (Ben Tremblay, from FB)

Addendum: "Don't be lucid and ironic. People will turn that against you, saying 'You see? I told you he wasn't a nice person!'"

p.s. "Aim to explain"? you mean rather than explore and inquire? lecture, rather and discourse? how tragically revealing ...