The Anatomy of Systems
A transdisciplinary description of the inner working of systems
Epistemic Status: Medium confidence that this system anatomy can describe most systems at a high level, low confidence I have covered all fundamental concepts necessary to describe the inner workings of systems.
Epistemic Effort: Low-medium effort. I did not do extensive research into previous work on this topic since my experiences studying and contemplating complex manufacturing, biological, and AI systems revealed a common system anatomy that is transdisciplinary. I wrote an outline of my ideas and then used Notion AI to expand on each, with some manual revisions. I created the diagram using LucidChart.
Systems are all around us. From the human body to the global economy, systems are responsible for a wide range of phenomena. Understanding how systems work is not only important for scientists and engineers but also for anyone who wants to build life-operating systems. In this post, we will explore the anatomy of systems.
A Brief Definition of Systems
Before we dive into the anatomy of systems, let's define what a system is. According to Ludwig von Bertalanffy, a system is a set of interrelated components that work together to achieve a specific goal or function. A system can be anything from a simple machine to a complex ecosystem. Understanding how systems work is crucial because it allows us to predict their behavior and design more effective systems.
System Anatomy
Inputs
Inputs are the information atoms that trigger an outcome or change in the system. Inputs can be either exogenous or endogenous. Exogenous inputs originate external to the system, from its environment or context. For example, the environment a child is raised in is a conglomeration of external influences, such as parenting, culture, and education quality. Endogenous inputs originate internal to the system, from one or more of the components within the system itself, and are fed to one or more components within the same system.
In the diagram, X1 and X2 are exogenous inputs. At the same time, AB is an endogenous input to component B, AA is an endogenous input to component A, and BC is an endogenous input to component C.
Outputs
Outputs are simply the responses to inputs, determined by the structure of the system, the system state, and the intensity and duration of the inputs. Endogenous outputs are responses that change the system's state or become endogenous inputs to another component in the system. Exogenous outputs are a product of the system, something that leaves the system and enters or influences the environment.
The diagram shows three exogenous outputs–Y1, Y2, and Y3. AA and AB are endogenous outputs of component A, while BC is an endogenous output of component B.
Components
A system can be broken down into parts called components, which work together to achieve the system's design goal. To understand the system’s behavior and how to improve it, one must consider how each component works and how they all relate to each other. Typically, the components of a system are actually systems themselves, so-called subsystems. For example, the human body is structured hierarchically, with cells making up tissues, and tissues making up organs.
There are three components in the diagram–A, B, and C.
States
Systems are parameterized, and these parameters can be modified. The parameters also determine how the system behaves. The same set of inputs might produce wildly different outcomes if the system state changes.
I should also make a distinction between states and inputs. Inputs are fed to a system, whereas states represent the internal configuration of a system. However, I can seek to manipulate inputs to achieve a desired system state (this is the premise of medicine, process engineering, etc.). For example, I can manipulate the condition of my body by controlling the dominant inputs like food, exercise, and sleep. A system receives an input, there is a response, and the subsequent configuration of the system after this interaction becomes the new state.
Each component in the diagram has an associated set of parameters, which come together to make up the state of the system.
Feedback Loops
Feedback loops are mechanisms that redirect system outputs as inputs for the future. Positive feedback loops occur when system output leads to an increase in input, which in turn leads to an even greater increase in output. These loops are often associated with fragility and can lead to runaway behavior. On the other hand, negative feedback loops attempt to reverse trends in the outcomes rather than reinforce them. These loops are often associated with stability and can help to regulate a system.
Constructive feedback loops tend to be negative feedback loops, such as the body’s thermoregulation mechanism, and are core mechanisms in self-correcting systems like machine learning algorithms and practitioners of personal development. Positive feedback loops tend to be destructive, the classic example being drug addiction.
But this is not always the case. One constructive positive feedback loop I have noticed is the relationship between learning and curiosity. A curious mind wants to learn, and upon learning, one discovers more things to know and becomes more curious.
There are several feedback loops in the diagram, demonstrating the different possibilities. The output of component A becomes an endogenous input to itself and combines with the feedback from Y3 to manipulate the X2 input. Additionally, Y3 feedback also connects with the endogenous output of component B to form the input to component C.
Example: The Anatomy of a Project Management System
Now that we understand the anatomy of a system, let's apply it to a real-world example: a project management system. A project management system can be broken down into several components: a project database, a task database, a goals database, and methodology templates. Inputs to the system include projects, tasks, project goals, deadlines, and methodologies. The output of the system is project completion and all the benefits that come with it. The state of the system includes the status of tasks, projects, and goals. A feedback mechanism could be project retrospectives, where lessons learned from the project are recorded and used to update the methodology templates, tasks, goals, and time management for subsequent projects.
Conclusion
Thinking about system anatomy forces us to be purposeful in designing our systems. We must consider what inputs and components are relevant, and how to incorporate constructive feedback loops to turn our systems into self-improving systems. To optimize a system, one requires a mental framework for defining the inner workings of that system. Understanding the anatomy of systems also provides a powerful mental model for comprehending complex topics in systems design, theory, and science. This, in turn, helps improve system proficiency and enables us to build more effective systems to elevate the well-being of conscious creatures.