{
  "path": "/pub/3mh7imf6ztv22/thinking-in-systems",
  "site": "at://did:plc:4m3kouplb7s7xozjd3whinvl/site.standard.publication/3mdteq7ldoc2i",
  "$type": "site.standard.document",
  "title": "📘 Thinking in systems",
  "publishedAt": "2026-05-05T00:00:00.000Z",
  "textContent": "📘 Thinking in systems\n\nDonella H. Meadows - 2008\n\nWhat the book is about\n\nWhat I find magical is the resilience between systems. This book interests me a lot because it offers a vision of the world built on a set of systems that influence each other: which actors and which interactions produce this result? What a fascinating question! Understanding the complexity of the world rather than settling for simplistic explanations. That's the value of this book.\n\nUnderstanding the world around us means understanding the complex interactions between systems.\n\nUnderstanding a system happens through modeling it. Let's practice together how to model them.\n\nThe systems lens\n\nRelationship between structure and behavior.\n\nExterior events suppress and release.\n\nA system is a set of things connected with characteristic behaviors.\n\nWar, hunger, poverty. No one wants the behavior but these behaviors persist nonetheless.\n\nThe elephant and the blind people. Touching it, knowing parts of it, does not tell us how the system behaves.\n\nPart 1 - system's structure and behavior\n\nThe basics - More than a sum of its parts\n\nSystems are made of:\n\n1. Elements,\n2. connections,\n3. function or purpose deduced from behavior (not from rhetoric).\n\nSand on the road, on the other hand, is not a system because it lacks a purpose. Remove some sand from the road, and it's still sand on the road.\n\nInflow, stock and outflow change slowly. The stock — material or immaterial — can be what we need or what we want less of. The simplification schema can be a good way to share the new direction we want. Time lags allow us to maneuver, to adapt.\n\nSystem thinkers think in flows, stocks and feedback.\n\nDiscrepancy = Divergence\n\nResponse delays drastically change how systems behave. Responding too quickly will make the system overreact and make the situation worse (stock, buying inventory). Like shower faucets where the pipes are long.\n\nWhy systems work so well\n\n1. Resilience: the ability to recover after some pressure. Lots of feedback loops, with different mechanisms, temporality, redundancy. Meta meta resilience creates systems that are more and more resilient depending on the environment, like the human body. Just-in-time deliveries reduce cost by limiting inventory, but they are more vulnerable to traffic flow and other glitches. Air pollution breaks the resilience of forests. Resilience is a plateau where it can play, a big plateau. But losing resilience makes the plateau shrink.\n2. Self-organisation: making its own structure more complex. A network. It requires freedom and experimentation and a little bit of disorganization. With simple rules, you can build a sophisticated structure.\n3. Hierarchy: subsystems on subsystems on subsystems, from a cell to a town. This is hierarchy. Self-regulation within a subsystem, where they take care of themselves while working for the greater good when the larger system coordinates them. Ora and tempus, the watchmakers. These coordinations may be quickly forgotten. Too much central control can end up killing the whole system. Enough central control for coordination toward the main goal, enough autonomy to keep subsystems self-organizing and flourishing.\n\nWhy do systems surprise us?\n\nBecause we're ignorant. We know a lot of things, but we are unaware of a lot more.\n\nEverything we know about the world is a model. The model will never be the true nature.\n\nOur model falls short of representing the system.\n\nEvents are outputs; they are unpredictable when seeing the system as a black box. But, with some information, the tendency of a river can be useful to know when it'll overflow. A winning streak of a team.\n\nThe behavior of a system is its performance over time. News with context will be better for us to evaluate events. They're clues to understand systems. The structure of a system helps us understand what events occur and why. The structure is the combination of interlocking stocks, flows and feedback loops.\n\nThe triptych Event - Behavior - Structure is what we should study to better understand the world around us. Events are always surprises. But, looking from a bit further away, by following behaviors we realize that many things can be explained. Imagine a \"System media\" newspaper that would offer an overview in addition to the recent event — one more point on a curve, tracking a trend. The sensational would be less present.\n\nSystem traps\n\nRule beating effect\n\nThe misuse of a subsystem's rules makes the system less performant — they work against the spirit / intent of the system. For example: intermediate indicators, measuring effort without measuring impact, spending the entire budget for fear that the same budget won't be allocated the following year, protecting land where an endangered species lives — which leads landowners to hunt that very species.\n\nLeverage points - the levers to modify a system\n\nOnce we know all this, what are our levers to effectively modify a system? Without falling into the silver-bullet trap, we can try to identify a root cause that allows minimum effort for great impact.\n\nMost of the time these levers are not intuitive — on the contrary, a person who thinks they know the system well is very likely to propose simplistic solutions.\n\nThese change points are not magic recipes in a complex system. The author even warns us that she may have spent months or years rigorously studying a system, proposed an intervention lever, and seen this proposal rejected because it seemed so crazy.\n\nFinally, when we find a relevant lever, we still need to push it in the right direction! Financial markets, for example, are all the more volatile because information loops have shortened.\n\n12/12 The weakest lever: numbers\n\nFor the author, modifying parameters is the last possible option to improve a system. Its impact is too unpredictable and too small. And yet that's what we look at in 99% of cases.\n\nIt's like rearranging the deck chairs on the Titanic — it's cute but useless.\n\n11/12 Buffer\n\nAdding stock to a capital makes the system more stable. But it's not necessarily a good thing, since it also makes it more inflexible, harder to adapt, and creates substantial cost. See JIT.\n\n10/12 Stock and flows structure, physical structure and their nodes of intersection\n\nGenerally, the structure is hard to modify and is more of a constraint we must adapt to.\n\n9/12 Delay\n\nBuilding a nuclear plant can never be a solution to short-term changes given a construction delay of several years.\n\nDelays are rarely a lever that can be activated — they are usually endured.\n\n8/12 Balancing feedback loops - the strength of the feedback loops relative to the impact they try to correct\n\nFrom the physical system we move to the parts about information and controls, where we have more room for action.\n\nEvery feedback loop needs a goal: the thermostat that regulates a temperature. These feedback loops, if they are only informative (visualizing a performance is only interesting if it triggers countermeasures to deviations), can be useless without an associated corrective action.\n\nA mistake we can make — because it takes time or is costly — is removing the prevention loops (warnings, information, etc.). The perverse effect is that the change brings immediate effort reductions without an apparent counterpart; there's little risk that the system stops being useful right away. But it only takes one failure for it not to have been worth it in the long run.\n\nA feedback loop has a strength: its ability to keep the capital / stock at — or close to — its goal within a small delta and a short time frame.\n\n7/12 Reinforcing feedback loops\n\nA balancing feedback loop is self-correcting.\n\nThe more a reinforcing loop is used, the stronger it gets. The more it is used, the more it will be used. It's a chain reaction. It is the source of erosion, explosion, growth, and collapse. This loop ends at one point or another. Weakening a reinforcing loop is often more effective than strengthening a balancing loop. The inheritance tax, for example, is more effective.\n\n6/12 The information flow\n\nThe structure of who does or doesn't have access to information. Giving feedback to an environment that wasn't aware of it before. The tragedy of the commons happens when there isn't enough information. Lack of accountability is a source of malfunction, and there is a natural tendency not to take responsibility for one's own choices (the politician who declares war without volunteering on the front line). That's why this countermeasure is so popular with the masses but less so with leaders.\n\n5/12 Rules: incentive, punishment, constraint\n\nRules are very powerful for shaping a system: sports rules, commercial rules, transparency obligations.\n\n4/12 Self organisation\n\nThis is what we call evolution in nature. This ability is what allows a system to be resilient to external changes. Experimentation and testing against reality are what allow a system to evolve. But it starts from a culture's desire to learn. If this culture feels superior to others, that's when it is most vulnerable. Encourage variability, experimentation and diversity. It's a loss of control — but a necessary one.\n\n3/12 Goals\n\nConforming to the goal is what allows all the levers mentioned above to be used in service of achieving it. That's why it's so powerful. Is genetic engineering a good thing? It all depends on the goal in the end — ethics is tied to the goal.\n\nInitiating a new goal for a system is possible at the leaders' level. Less so at the team-member level.\n\n2/12 Paradigms\n\n> [!NOTE]\n> A paradigm is a representation of the world, a way of seeing things, a coherent model of the world based on a defined foundation.\n\nIf a company systematically communicates about _how productive it is_, then it will be more productive but probably not _more resilient_.\n\n1/12 Transcend paradigms\n\nParadigms are only extremely limited representations of the real world. The paradigm that all paradigms are wrong is the truest paradigm o",
  "canonicalUrl": "https://remanso.space/pub/did:plc:4m3kouplb7s7xozjd3whinvl/3mh7imf6ztv22/"
}