Recursive application of systems thinking: recursive attention management ​

Understanding that every system is part of a system breakdown and belongs to a certain system level allows a systems thinker to apply the same systems thinking recursively: conducting the same reasoning for each system level, for each subsystem at this system level, going down from the target, and for each supersystem when moving up the system levels. It doesn't matter what system it is - it can (and should!) be subject to comprehensive systems thinking (as prescribed in the system's mantra). After thoroughly considering the situation with a specific system, one can return to the system's breakdown as a whole (during usage time) and the creation graph to select the next system for consideration - and so on with various systems from different system levels and locations in the creation graph.

System breakdown structure is primarily a means of attention management. Attention captures one object-figure for detailed consideration, while everything else remains in the background, no matter how vast or diverse the "rest" may be. Attention allows simplifying the complexity of the world by temporarily ignoring insignificant details - leaving only the important in discussion. Systems thinking constantly focuses on the most crucial aspects, overlooking the unimportant details - attention focusing. It also ensures not forgetting the whole when focusing on parts - attention defocusing. The main focus in systems thinking is "up" the system levels: not from the target system to its subsystems, but rather from the target system to its environment, to the supersystem! If you have a figure in your attention, its surroundings will be in the background. Attention defocusing, shifting focus to the background, turns out to be even more critical than ignoring the background, attention focusing!

Consciousness in attention management is required; consciously concentrate and defocus attention, consciously shift attention from the object to the background in moments of defocusing (many objects in the field of attention, high alertness to new appearances in a wide field of attention), and focus attention, tracking objects in its focus on the ignored background. Systems thinking manages attention, describing its movement in the concepts/types of a systemic approach (system, function/method, role, constructive, emergence, etc.) using these concepts as types of a meta-meta-model to identify objects of these types in the surrounding world (physical world and world of descriptions).

You know what to look for in the physical world - objects of different types of systems (target, supersystem, subsystem, ours, creator). How these types relate to each other (what types of relationships exist between them) is described in our course "Systems Thinking"; this is the meta-meta-model expressed in the course text in natural language.

Next, actively search for objects of these types in life, as you already have expectations about these objects. Being active involves questioning the status of these objects, devising and creating such objects if it's realized they don't exist, verifying hypotheses through rational reasoning and even experiments.

The concentration-defocusing of attention in system modeling corresponds to the movement of attention across system levels (levels of system size), and the focus of attention is the selection of a specific target system, its subsystems and supersystems, the systems at a link in the creation chain, identifying "our system" among them.

A systems thinker is well-oriented in a complex world: they never lose sight of the context/environment of the object under consideration for a moment, remaining capable of discussing both the smallest screw in the smallest device and large-scale planetary systems. They do not lose their minds over these "scale jumps"; concentrating attention on increasingly smaller parts of the world, ignoring details from the surroundings of this part, or vice versa, defocusing the attention on increasingly larger parts of the world, losing the details of the small parts of this larger part.

The Systems thinker selects a system, considering it within the supersystem and the larger system environment ("in context," during the operation of the ready target system), then can further investigate this system as a set of parts - "zoom in" to another level of detail, enhancing the scrutiny of this part, like in modern cameras. Indeed, they say "attention cameras" when discussing attention work:

• This work is active: first, aim the camera at the piece of the world you intend to examine, then adjust it to the required zoom. This pre-measurement change: aim the camera at the object! You prepare the measurement before you take it! To describe what's around the corner, you have to approach it and peek around the corner, actively take action. You must understand what you see there, whether you spot the silhouette of a big mountain or a grain of sand on that mountain, focus the lens - this involves engaging knowledge, a meta-meta-model.
• Only after this can you observe what is captured by this camera and select specific objects in its field of view, assigning types. Recognition of the presence or absence of objects of certain types can only be done after conducting the measurement and having knowledge about these types (knowledge of the meta-meta-model).
• In general, the agent's cameras can be multiple: attention is "multithreaded," each camera has a flow of information, and the operation of each camera needs to be adjusted and the meta-model used, along with integrating the results of the cameras' work in thinking.

The Systems thinker can easily choose the required scale of observation for each attention camera, select the necessary systemic effects (emergent properties) as subjects of interest at the chosen system level. The Systems thinker does this consciously: they well understand that they navigate through the system levels and with each meta-system transition (from one system level to another), new systemic effects/emergences appear in their considerations.

Here's an example of system levels analysis for a multimodal transport system^[1]:

In a transport system, we can initially discuss multimodal transport and the competition of independently operating monomodal transport systems. For example, pipeline transport competes with railway transport in oil transportation - for their owners, they are rival competitors in the operational environment, but for someone willing to transport oil from one point of the world to another, they are part of one multimodal transport system (keep in mind that different roles identify systems differently as convenient for their activities. Although for teamwork on a project, they must agree). When discussing transport systems - planetary scales or scales of a country.

In one of the subsystems of the transport system, you can choose to discuss the railway system - trains, the railway's energy, train traffic management, and so on. Taking one subsystem of the railway - the railway station system, you can further explore its sub-systems - passenger boarding system, information station system, passenger catering system, ticket sales system. Part of this ticket sales system - its ticket vending machines. Each of these machines can also be viewed as separate systems. The screws that secure the controller's circuit board to the housing of this machine - are also systems. And even within the screws, different parts can be found - the head with slots for various types of screwdrivers, threads.

In this way, in a couple of paragraphs and a small picture, we traverse the situation from planetary or national scales to a small screw, yet maintaining a coherent thought process, clearly understanding the subject of discussion and the problem scales. This does not change even when reversing the process, from screw parts to the entire transport system. The movement across different system levels, attention focusing on different systems in them, attention defocusing - this is an extremely powerful thinking tool.

It is meaningless to consider a screw in a ticket vending machine as an immediate component of the transport system - this may be correct from a formal logic standpoint, but entirely pointless, like "a herd of cows' tail." The systemic approach, introducing system levels, makes reasoning meaningful: all individuals get the opportunity to discuss problems only at their "own" system level (the level for which they have interests, preferences, intention to act to change the situation for the better based on these role preferences) while considering issues of adjacent levels - those higher than their target systems and lower. Thus structured with division into system levels and different roles, collective thinking is a significant achievement of civilization.

Similarly, we can discuss the establishment and development of a multimodal transport system, indicating the different levels of subsystems entering it: the who:: creators and how:: methods who are constructing the railway system, the who:: creators and how:: methods who manufacture the fastening screw of the controller's circuit board to the ticket vending machine's casing, and who and how screws the fastening screw (this will be a different creator than the creator producing the screw). Everything is discussed as "systems," and creation systems are discussed in relation to the environmental systems because if you don't know what:: "target system or our system" you are creating - then you don't know which creating systems manufacture it, making the discussion impossible. First the environment, then the target system and its device, creation methods ("how we make the target system"), and creation systems (who:: roles, who:: agents play them, who:: managers organize them) - will be considered last.

A Boeing 747-8 consists of 6 million independent types of parts (forming the target system), manufactured by half a million people in 5400 factories (creation systems), 783 million plane parts were ordered in one year (the last plane was released in December 2022)^[2](http://787updates.newairplane.com/787-Suppliers/World-Class-Supplier-Quality){target="_blank"}:

And now the surroundings of these Boeings, the aircraft operation time: airports with air traffic control systems and passenger boarding-disembarking systems, air corridors (again systems! Equipped with radars and other infrastructure for tracking them. They are tangible, occupying space). The aircraft itself outside all this is useless, like a cork without a bottle, or a bottle outside its usage situation.

In modern systems, the number of individual elements that need to be coordinated (in design) and often created from scratch (in construction) reaches tens of millions in "hardware" systems, and if it's about electronic systems, then trillions: on one Cerebras electronic chip, the number of individual transistors is 2.6 trillion pieces with each transistor having its unique purpose within the chip, performing its unique function https://www.anandtech.com/show/16000/342-transistors-for-every-person-in-the-world-cerebras-2nd-gen-wafer-scale-engine-teased. Also, these chips are not the highest system level, above them is the circuit board level, and above it is the supercomputer based on these boards.

Hoping to create such complex objects without hierarchical examination and collective and personal attention management through a documented system hierarchy (that is, the hierarchy of systems in terms of composition) is futile. Managing the collective attention of creators by linking this attention to system levels distinguished in the hierarchy in terms of composition - the most crucial part of systems thinking. This collective attention is also directed at creation relationships between creators and the systems being created, where we speak about the creation graph.

System levels arise due to the complexity of systems, a property of evolution (biological/Darwinian, memetic, techno-evolution). The complexity of systems will continue to grow; it's an endless growth. At the same time, each creator-team and creator-enterprise can be relatively simple (although they are also systems, the complexity of creators also grows - enterprises join in eco-systems, grow into super-holdings, the number of system levels there also increases as part of evolution). Still, dealing with complex system creation projects (including the creation systems themselves) is possible because no individual creator or even their parts (for example, a project team within an enterprise) has to deal with the entire system on all system levels - no, each creator works on some parts of the system, significantly simplifying creation. Someone creates the rocket engine, someone creates the rocket computer, someone creates the rocket body, but there isn't a firm that made all of this - and even smelt steel for rocket body or purify silicon for the computer chip.

Different companies specialize in different crafts, but organized according to a creation graph - together they create remarkably complex target systems operating in a remarkably complex environment. All this is possible thanks to systems thinking, implemented by systems engineers and managers of these creator enterprises.