Skip to content
Create an account for full access.

From the system-of-interest through the methodology to the role of the creator, but not vice versa

Simple systems, such as those represented by the construct "hammer", have extremely low intelligibility: "intellectual characteristics" (more details on this will be covered in the "Intellect Stack" course). For human agents and AI, intelligence is stronger/more universal and intelligibility is much higher (they do not need to be "hardware-wise" reassembled in terms of changing the construct, they can be taught some behavior, given explanations – "reprogram the hardware to a new algorithm", not "redo the hardware in a new way"). Furthermore, such agents can use various affordances to incorporate systems into their composition; these affordances used by creators for carrying out some work according to their algorithms/theories as part of the creator system are called "tools" (for measuring instruments – instruments).

It should be reminded that "agent" has two meanings:

  1. Broad – any autonomous entity found in a non-equilibrium stable state (from atoms and molecules to cats, from cats to city agglomerations with a couple of million inhabitants and numerous buildings filled with household utensils and industrial equipment).
  2. Narrow – an agent in the broad sense, but one capable of planning actions leading to a sharp deterioration of its state in order to then improve its state (for example, spending all resources on planned creation of a bow and arrows to be more successful in hunting later). Usually, the term "agent" refers to humans and/or technical systems with sufficiently strong artificial intelligence (AI agents). And there are constant debates about whether the intelligence of AI systems is sufficient to call them agents in the narrow sense, or are they only agents in the broad sense.

Next, we will mainly talk about agents in the narrow sense: humans and AI. And since the discussion will mainly be about terminology rather than concepts, the "power of intellect" here will not be so crucial. You will understand that it is easier to talk about the function of driving nails for a hammer or the culture of driving nails with a hammer. Most likely, in the first case, the role will be played by the hammer, and in the second case, the role will be played by some agent, including a hammer as part of its composition. But even for the agent, we can say that its role is driving nails (whether with a hammer, a microscope, or an electric nailer, it's up to the agent). How to know what to say, how to know how to understand what is said? From linguistic intuition. Simply remember that the "function" for "agents" most often sounds like "method/way of work"/"type of labor/engineering," etc. But thinking about the role and the construct is the same! The role and its function – "why this agent is needed as a system," the construct – "what system-agent as a material/constructive taken by affordance for the role."

It is always necessary to clarify which specific "method/way of work" is chosen for performing the role of an agent with strong intelligence: due to the universality of the agent, there can be surprises. When we talk about "surprises," it usually means "unpleasant surprise," a deviation from expectations. For example, an inexperienced worker in the role of a carpenter may perform the "board fastening" work without distinguishing between screws, nails, or glue – including, for example, driving a screw with a hammer (and in the first few seconds after completing this work, it may even seem that the boards are securely fastened!). While using a microscope for a "nail remover" probably doesn't make sense for anyone, but for humans and AI, it's easy, even if they are not able to work according to a specific method. Patient/sick person: a role in a doctor's appointment is a role:

  • "Hello, doctor, I have a stomach ache."
  • "What did you eat yesterday?"
  • "I ate a pigeon."
  • "Did you eat a raw pigeon?"
  • "Yes, I caught it, plucked it, and ate a raw pigeon."
  • "Did you eat the pigeon's head?"
  • "Yes, I ate the pigeon's head."
  • "Did you eat the raw pigeon's head?"
  • "Well, I ate the raw pigeon's head. So what's the problem exactly?"
  • "Hmm... Honestly, I'm a speech therapist, but I did what I could – I've helped!"

It is impossible to understand roles without considering their function/behavior. And remember: behavior in terms of achieving changes in the surrounding environment for different agents will be called very differently: functions, practices, organizational processes, methods, ways of working, and so on.

If the "one called a doctor" (i.e., described by the classification as a "doctor") is considered, then a speech therapist may be seen as a "specialization of a doctor." This approach does not focus much on the mastery, as all doctors have some specializations, and the speech therapist is just a part of the doctors, specializing in speech therapy. Therefore, the sub-role in this case would be a "specialization." However, when it comes to roles that are part of the doctor, it would be more about the method of work. Thereby, the speech therapist specializes in improving speech, but also has an overall doctor's knowledge, so the doctor is a sub-role for the speech therapist. In the case of mastery, the relationship would reflect the composition; the speech therapist would require more training than the doctor, making the doctor a sub-mastery.

We recommend relying on the line of reasoning related to the function/"method/practice of work" of the role object (role), the mastery of carrying out this function by the constructive object (agent, for example, a human, or even an organization) – and therefore consider the "sub-role" as a node in the hierarchy concerning the composition/part-whole relationship for mastery. This is convenient for engineering (for example, in training planning).

You will often encounter classification descriptions in everyday life where the term "sub-" indicates specialization. There is another term that is also multi-meaning, that is "meta-," which has many different meanings in terms of some levels of relationships[1]. We use it to refer to meta-modeling in the sense of a "higher level of objects in a hierarchy concerning classification."

For the relationship of composition, we use "sub-" as in "subsystem," and "super-" as supersystem. Specifically, a sub-role would literally be a "subsystem," while a super-role would be a supersystem (systems are usually defined by their roles in the supersystem, subsystems by their roles in the system, and roles by their functions/behaviors). To determine whether or not to have separate prefixes for specializations, one could use "kind" to indicate a kind for the genus, for example, "speech therapist kind of doctor, don't confuse with a gastroenterologist," "salsa dancer kind of dancer, another kind – tango dancer, don't confuse," but classification schemes are more or less arbitrary when it comes to highlighting particularly important objects (but they are good for highlighting a variety of objects, not necessarily important objects[https://en.wikipedia.org/wiki/Animal_classification_(Borges)]), and we will use them rarely.

However, in everyday life, modifiers "sub-" and "super-" are often used, which can cause confusion. How to understand? Only from the context: when encountering "sub-" and "super-," ask yourself about the type of relations they denote. What kind of mastery will be required to perform the practice – is it more or less? Another criterion: when talking about specializations, it mainly refers to the choice of an exclusive role – for example, speech therapist and gastroenterologist. By definition, in composition/assembly, roles are compounded, functions are added: the functions of a gastroenterologist or a speech therapist are combined with the function of a doctor. And the mastery of a speech therapist and/or a gastroenterologist will necessarily be added to the mastering of a doctor, where the doctor will be an object of sub-mastery. In the case of the relationship of kind-species, there is nothing to combine, only to describe – and one can discuss the "choice": speech therapist type or gastroenterologist type for the doctor group. In any case, pay attention when dealing with sub-roles and super-roles – is it about a system (and there will be a composition relationship, everything is physical) or a type of system (and there will be some specializations, usually showing contrast). A "tiger" and a "mouse" are not part of the "beast"; they are just "species of the beast"!

In Krylov's fable "It's a problem when pie maker bakes shoes, and shoemaker makes pies," it is a typical metonymy. Let's clarify the types to specify what is being said: "baking pies" is a method of the "pie maker" role, "cobbling shoes" is a method of the "shoemaker" role. But for Krylov, the shoemaker is an "agent who has the mastery of a shoemaker," so the shoemaker is not a role but a construct/material for the pie maker role that is mistakenly utilized as an affordance for the "pie maker" role. Therefore, shoemaker is not a role! The same applies to the pie maker, who in Krylov's story is not a role but a construct/material for the shoemaker role.

In real work projects, there will be exactly the same confusion between roles and constructs: a construct that is a suitable affordance for one role may well be assigned to perform another role. And if discussions mix up roles and constructs in terms of terminology (which is perhaps unavoidable!), problems will arise. A good move here would be to switch from the language of roles to the language of functions (or, in the case of humans and AI as creators, the language of methods/practices/culture/style/ways of working, types of labor/activity/engineering, etc., as synonymy is very extensive here).

In any case, the conclusions about the target system and the methods applied to it for changes can be derived from a chain of reasoning:

  • First, understand the role and function of the target system in the environment, "what it does" (during the operation/use of the target system).
  • Then understand the construct/material that will be an affordance for embodying the role performing the function (during the creation of the target system).
  • Then understand the methods of making the construct/material (methods of changing the state of the construct/system from raw material to the state of the finished and working target system).
  • Then understand which roles of the creating systems will execute these methods.
  • Then understand which agents as construct/material for the creating roles (people, departments) will perform these methods of creating the target systems.

Never the other way around because if you do not know what the end result should be, i.e., you do not know what the system should do with its environment, then you do not know which materials you need to make the system and how to process them, and if you do not know what materials to use and how to process them, then you do not know who with the necessary mastery and tools can take these materials and process them to create a successful working target system.

This we have just mentioned is the "mantra of systems thinking," a recurring pattern guiding the thinking of a systems thinker in a certain direction. Use it as a checklist, check to see if all the steps in your thinking about the working project are being made according to this mantra.


  1. https://ailev.livejournal.com/1053878.html ↩︎