Units of Action
Unit of action is a term I have been using internally to be more specific when thinking about groups of people. This post is for fleshing out and clarifying my thinking for myself, and seeing if it would be useful to anyone else. Also it feels like there really ought to be a term for this already, and I might be able to find more information if someone knows it.
A unit of action is a group that takes actions, as a group.
I take the word unit from the military, and also from unit of analysis, reflecting my belief that this is the correct level of analysis for big-picture problems. Action is in the colloquial sense of “did something on purpose.”
To be more concrete, the kinds of things I am pointing at are like families, corporations, and government agencies. The kinds of things I am pointing away from are like race, class, gender, and religion.
Causal vs correlational
One way to see the dividing line is between groups that cause things to happen, and groups to whom things happen. Families take vacations; corporations launch products; government agencies sue people for breaking regulations. The examples I pointed away from are demographic—they don’t do anything as a group, but races may be segregated, genders have different bathrooms, and classes get tax breaks.
Another heuristic is whether the group can plausibly be modeled as a single agent. Following game-theoretic intuition further, the unit of action can be broken down into other units of action the same way agents can be broken apart into multiple agents. We might model one firm as one agent when looking at the behavior of firms in competition, but when looking at the behavior of one firm model the different departments within the firm as different agents. Sometimes, as in the case of the military, these components are very explicit.
Units of action are stable, successful-at-some-rate coordinators. I have a vague intuition that we can think of them as coordination strategies which propagate in the same way organisms have reproduction strategies, but the analogy is hardly precise; reproduction propagates genetic information, and units of action mostly propagate relationships.
Hierarchy is usually sufficient, but not necessary, to indicate a unit of action.
There is no taxonomy; it is defined by actions, not org type.
If a given group stops being a unit of action, I expect it to fall apart soon (split up, go bankrupt, etc).