Hide table of contents

The importance, tractability, neglected framework: while working on important, tractable issues makes sense, what does neglectedness tell us?

The common response I hear in the community is that there exist diminishing marginal returns on working on an issue; however, I believe this explanation alone does not capture the full story. (And neither do a lot of EA organizations.)

Here, I explain what I believe to be a better catchphrase to explain the importance of neglectedness: contingency.

  • (1) 80/20: There are diminishing marginal returns on the n-th person working on an issue because all of the “low-hanging fruits” – e.g. easy research, development, and implementation of solutions to a cause area – will be taken early. This is the common argument for neglectedness, as it is understood – the 80/20 rule.
  • (2) Initial Momentum: The “founders” of a cause area take the time to set up the infrastructure for a new cause area – such as talent searching & recruiting; institution-building; securing funding; marketing; and building the worldview, values, and norms of a space. Helping to get the “ball rolling” – e.g. for neglected research fields within the AI alignment or biosecurity space – feels extraordinarily impactful, and distinct from pure diminishing marginal return from direct progress.[1]
  • (3) Meta-Level Reminder: Neglectedness serves as an acknowledgement that the most tractable thing to do may not be in the most “popular fields” or the ones that come directly to mind, serving as an epistemic reminder to be more open-minded and venture out further for the most effective solutions.

These arguments are important to differentiate. Unfortunately, pure diminishing returns tends to be visualized as an extension of direct tractability (1) by newcomers

However, especially for longtermists, (2) and (3) are significant reasons for working on neglected fields; working on extremely neglected issues should thus be viewed as having a cascading, systemic impact. Bringing a field to prominence and looking for new fields to start entrepreneurial projects in is extremely important. Longtermists usually find neglected fields to establish systems – build new infrastructure, bring in funding, bring together stakeholders, and gradually establish a proper research field with its own literature and institutions – in a cascading butterfly effect where more people, talent, and ideas get brought in.

Community-builders: how do we capture this, in a quick phrase? How do we replace "diminishing marginal returns" with something better?

The better argument in favor of neglectedness

Furthermore, diminishing marginal returns on a cause area tends to hold true, but has its exceptions. For example, what if you're able to leverage resources in a larger cause area? One may be able to climb to a powerful position & redirect resources from a popular cause area, helping to push them toward more tractable ways of working on the cause area. 

This is generally a special exception, but speaks to how "what can the marginal person do?" is too economic theory-esque, doesn't directly relate to actionability, and isn't generalizable.

Therefore, in explaining EA to newcomers, I've found the need to rephrase and connect "neglectedness" into "contingency."

Having a contingent impact is just a better explanation for the importance of neglectedness: in the absence of your intervention, how would have things have played out? 

  • If the same, you likely didn't make that big of an impact.
  • If differently, you did.

Newton & Leipzig invented calculus at around the same time, so the invention of calculus may not have been contingent & was bound to happen anyways. However, certain events & movements likely were, especially key historic elections, formational institution-building, Cold War-era nuclear policy, and (hopefully) longtermist AI alignment efforts. 

To tie it all together definitionally, very contingent efforts can be found in interventions that are neglected in the status quo.

We can then rephrase looking for neglected sub-areas of popular cause areas into looking at which points of leverage you can apply the most contingent impact, which generalizes the point more cleanly.


Edit Update: The contingency part was inspired by What We Owe the Future's significance, persistency, contingency framework. At the time of this article's posting, however, I didn't read the appendix ofWhat We Owe the Future where Will MacAskill suggests renaming neglectedness to leverage. I suppose we were both thinking similarly!

  1. ^

    Even if this effect is captured in the overall "diminishing marginal returns" argument in economics, it feels important enough to differentiate as as separate reason.

29

0
0

Reactions

0
0

More posts like this

Comments6
Sorted by Click to highlight new comments since: Today at 1:19 AM

Good post; interesting point with that the impact of the founder effect is probably higher in longtermism and I would tend to agree that starting a new field can have a big impact. (Such as wild animal suffering in space, NO FISH ON MARS!)

Not to be the guy that points something out, but I will be that guy; why not use the classic EA jargon of counterfactual impact instead of contingent impact?

Thanks a ton for your kind response (and for being the guy that points something out). :)

"Counterfactual" & "replaceability" work too and essentially mean the same thing, so I'm really choosing which beautiful fruit I prefer in this instance (it doesn't really matter).

I slightly prefer the word contingent because it feels less hypothetical and more like you're pulling a lever for impact in the future, which reflects the spirit I want to create in community building. It also seems reflect uncertainty better: e.g. the ability to shift the path dependence of institutions, the ability to shape long-term trends. Contingency captures how interventions affect the full probability spectrum and time-span, rather than just envisioning a hypothetical alternate history world with and without an intervention in x years. Thus, despite hearing the other phrases, it was the first word that clicked for me, if that makes sense.

That makes sense, and I would tend to agree that the framing of contingency invokes more of a "what if I were to do this" feeling which might be more conducive toward people choosing to do more entrepreneurial thinking which in turn seems to have higher impact

Nice text, I think it's always valuable to question words and their meaning. However, I don't agree in this case. The whole English-not-as-first-language part of the community took time and effort to squeeze the word "neglectedness" into their active memory. Replacing it now with another strange and highly ambiguous term might lead to more confusion than anything else.

All joking aside,  I do think this can cause some misunderstandings. "Contingency" in logic/philosophy (according to the English and German Wikipedia) defines states that are neither always true nor always false or even a non-necessity of existence. The former, IIUC, is kind of what you want to express—the impact can but doesn't have to be big. But the latter one is definitely too far away from what you describe in (1), (2), and (3): is a cause's existence is not necessary if it is contingent? That makes it seem so unimportant in general and overall! What I would rather like to get out of such a definition is 1) something more like "It's important but other people are already working on it" and 2) "It's important but not right now". To me, it seems as if "contingency" doesn't have such a changeable time dimension, a state described as "contingent" to me seems more permanent. "Neglectedness", however, does seem to be evaluated at a specific point in time and therefore is able to be subject to change.

Lastly (and you might just want to disregard this point as not scalable to other languages but maybe other speakers find similarities:), the almost same word "Kontingenz" in German means more something like "subsequent events". It's closer to the Latin "contingere" which stands for "being close" or even "touch" rather than the Latin "contingentia" for "possibility/chance". "Subsequent events" would of course not make any sense but, again, confuse a lot. (This argument's epistemic status is "my best guess" as I don't know Latin very well. Would be glad for anyone who's more familiar here).

However, there is one dimension I do agree on here with you: In terms of being respected by other research fields. Similar to the effect of the word "longtermism", "neglectedness" is very specific to the EA community. "Contingency" on the other hand, is a term used by many other disciplines. Especially for new people, an interdisciplinary term is easier to process, find relations to, and maybe even remember. I think this might still hold if the person doesn't know the term beforehand. Alone the possibility of drawing parallels to other fields might make it more intuitive.

Alternatively, if you're trying to devise a reliable framework for evaluating decisions, you could just abandon INT in favor of things like COILS (disclaimer: this is my own personal work), which is actually specifically tailored for decision evaluation (rather than "evaluating a cause area to indirectly inform decisions") and is, in my view, actually a relatively reliable framework.