Beware Shared Basic Value Changes

Most of our activities can be seen as nested plans, to achieve nested goals. For example, in a war an overall plan to win the war is composed of many more specific plans to start or end particular battles and win them. Particular battles are made of plans to move troops to particular places, to watch for enemy troop movements, and to engage or not when opposing troops collide. At a small scale, particular engagements have plans for particular soldiers to shoot at particular enemies and to avoid being hit by shots from particular sources. There are also plans to develop and deploy new weapons, made of plans to design them, test them, make them, and distribute them.

In general, each particular plan is of some general type, such as move or shoot, is evaluated by goals set by its higher level plan, and sets the goals for its component plans. And as we experience far more low than high level plan attempts, we have far more chances to learn how to achieve them. We can see far more variations in attempts to achieve lower plans, and have far more data on which variations worked well or badly.

We thus learn faster how to achieve lower level plans, which allows those plans to change faster. Which reasonably induces faster changes to their overall structure, duration, scope, supporting tech, supporting personnel, and subgoals. Thus we should expect, and not be especially concerned by, rapid changes in subgoals for low level plans, at least if such change rates roughly match those of other features of low level plans.

In organizations, higher level units enact higher level plans. Thus it is easier to judge the efforts of lower level employees. At higher levels promotions, demotions, etc. depend less on a concrete track record of success in achieving one’s division’s subgoals, and more on one’s social skills and the politics of varying alliances.

We can learn less about plans for individual lives, like “become a successful actor”, compared to sub-plans like “host a successful party”. Even so, if our world’s eight billion people would share their experiences across generations, we should be able to learn a lot about the effectiveness of various individual life plans. Alas we don’t share such info as much as we could.

However, it should be much harder to share info to learn about the plans of whole communities over generations. For example, consider large communities who share plans to get rich together, to live longer, to stay united, to deter enemy aggression, to preserve a heritage, or to end racism. Such plans tend to involve conformity pressures, to induce community members act similarly to achieve their shared goals. Communities can have far less data on which to base such plans, especially when the experiences of plans from long ago become less relevant due to changing conditions.

We should thus want our highest level and most widely shared plans and goals to vary the slowest, due to our having the least data on which of them work well when. And in a fast changing environment, we should want such plans and goals to be simple and robust, so that they might more plausibly retain validity and effectiveness in the face of such changes.

All of this suggests that you should be shocked and concerned by our modern era’s many rapid changes to key shared values, norms, and community plans. These changes mostly just can’t plausibly be the result of getting new data on the relative consequences of trying them out in different contexts. While the modern has seen huge changes to low level and small scale plans and goals, those mostly seem to be based on adequate data on their effectiveness.

While we should reasonably fear that our shared high level goals and plans have been undermined by these many recent low level changes, we can’t have much confidence in our particular high level goal changes. The main changes that would make sense would be to make them simpler and more robust, and less tied to the details of recent experience.

But what we actually see are big fast changes to key high level values, changes that are neither plausibly an adaption to recent specific changes in our experiences, nor plausibly attempts to make such values more robust across varied contexts. We have few reasons to believe, for example, that they would have worked better when averaged over the many environments seen in all of human history.

The reasons we have usually given for our recent changes in basic values and community plans is that they just feel more morally right to us. But unless we have some reason to think that such changing moral feels are tracking some key features of our changing environment, we have to expect them to be mostly maladaptive, in an evolutionary sense. (Both re DNA and cultural evolution.) Which should terrify you, if you understand the basics of natural selection.