A system is something that is composed of parts that have interrelations between them.
Each system has purposes, and those purposes may have nothing to do with stated goals. For example, if a government says it wants to protect the environment but allocates few resources to that goal, its purpose is not to protect the environment, but something else. A university may have the purpose of teaching students, while the components of the university have different goals: the student is trying to get good grades, the professor is trying to produce good research, and the administrator is trying to balance the budget.
Replacing the parts of a system does not usually change the function of the system. Changing the environment or interconnections will change it dramatically. Sometimes, changing an element will change the rules or connections, resulting in an outsize effect on the purpose or functioning of the system, like if you replace the coach of a team.
Stocks and flows are how to think about resources and how they’re used in the system. All models are simplified versions of the system. People tend to focus more on stocks than flows, and more on inflows than outflows: people think about finding more oil fields, rather than improving efficiency of vehicles. If you understand the current stock, you understand what leeway the system has when things change.
Delays in information and delays in physical processes both influence the kinds of decisions that need to be made. The necessary change is often unintuitive (e.g. extending the observation window instead of shortening it).
Resilience is baked into systems, by the fact that they continue to exist. Rigidity breaks down mechanisms of resilience, as seen when ecosystems become managed. Institutions limit self-organization.
Defining the useful boundary of a system
There are always multiple factors to an effect. When we do event-based analysis, we miss the systemic causes.
She blames our “bounded rationality” not on culture or whatever, but on imperfect information. Adam Smith was wrong about selfishness being the best, from obvious things like coordination problems.
If you are put into the position you critique, you’ll likely perform the same actions. Stanford prison experiment. I disagree with this. Aside from the issues with the Stanford prison experiment, I think in real life people select themselves into roles that work for them, rather than that the roles entirely shape the person.
It seems like under her view, you just need to zoom out enough and include all the relevant factors and how they relate to the issue you’re studying. She ignores the conceit of quantification.
Often, a policy will act in a certain direction, but other entities in the system with other goals will simply strengthen efforts in their direction. She compares pro-natal policies in Hungary, Romania, and Sweden, and said the most successful one was the Swedish policy, which aligned the interests of the community with the interests of individuals by focusing on the goal of every child being wanted and loved. This involved policies that on their own might be considered anti-natal, like fostering the use of contraception, but on the whole led to a healthier community. The Romanian policy (banning abortion below the age of 45) led to a huge population landing in the orphanages, and the Hungarian policy alleviated cramped housing for larger families but did not address other factors involved in the number of kids people chose to have.
- tragedy of the commons: A shared resource is not properly managed because it’s in each individual’s best interest to maximize their usage. There are three ways to combat this: tradition and appeal to morality, privatization, or regulation (“mutual coersion, mutually agreed upon”). Many cultures manage shared resources for a long time using tradition and morality, but economists/political scientists think modernity can’t handle that. Privatization can work if we are willing to let some people ruin themselves, but doesn’t work for resources that can’t be divided (e.g. the air). Regulation, regulation, regulation!
- continued degradation due to eroding goals: The fact that drift occurs leads people to laxity. You can keep past performance as a standard instead. Use the same structure to drift up instead of down.
- escalation: The increase in one stock prompts the further increase of another stock and vice versa. Examples include arms races, price wars, and holier-than-thou escalation of morality. This especially contributes to the continual consolidation we see under capitalism, which must be fought with anti-trust laws, regulation, progressive taxation, etc.
- addiction or dependence on an intervener: A system uses a band-aid to fix a symptom rather than addressing the root cause, and this results in the root cause deteriorating further. (This characterization can be used to paint things in a negative light, like if you think government involvement in a system is bad.) Real world examples include vaccination instead of community immunity, use of pesticides rather than working on polycultures, or our dependence on oil.
Here’s a useful list of definitions frequently used in this book, including the systems archetypes above.
I think a system can be used to explain any data post-hoc, by highlighting or de-emphasizing certain contributors to the system.
Diversity (biodiversity, diverse product markets, diverse human culture, etc.) increases resilience by allowing the overall system to be more self-organizing. Subjecting systems to centralized planning meets our wish for control, but at the cost of adaptability.It’s interesting that she acknowledges the need to let go of control, while simultaneously ignoring how using “systems thinking” to understand the world is a form of control as well.
I need to read a written copy of the last two chapters of this book. She’s got some eyebrow-raising quotes about how everything is a paradigm and we should try to reject as many paradigms as possible, but also recognize that it’s impossible to escape them completely. It’s sort of contradictory to the rest of the book in some ways, but I want to articulate why.