In this section, we discuss a CAS modeling approach of ecosystems targeting the pragmatic aspect when grasping a certain situation through semantic role modeling. The approach enables SoS thinking and will be exemplified by a concrete case from an institutional education. Shchedrovitsky (2014) in his analysis on the engineering nature of organization, leadership, and management of work pinpointed to conveying a specific meaning according to a situation and, thus, grasping situations according to semantics (p. 42 ff):

What is ‘meaning’? It is a tricky question. Really, there isn’t any meaning. Meaning is a phantom. But here’s the trick. I can say a sentence, like ‘The clock has fallen off the wall’ in two situations with two completely different meanings: ‘The clock fell’ and ‘The clock fell.’ The change of accent corresponds to two fundamentally different situations. Imagine this: when I am lecturing, I have got used to the fact that there is a clock here on the wall. At some point, I turn, I see an empty space, and someone in the audience says, ‘The clock fell off the wall.’ They might simply have said ‘it fell’ because, in this instance, the word ‘clock’ carries no new information. I look at the clock, I have got used to it and everyone in the lecture hall has got used to it. We look at that place and someone says ‘it fell off the wall’, and that phrase provides new information.

But now imagine a different situation. I am giving a lecture and all of a sudden there is a crash behind me. What has made it? I am told, ‘The clock fell off the wall.’ The situation is entirely different because what is new in this instance is the message about the clock. I heard something fall—that is a given—and I am told that it is the clock that fell. We pin this down in terms of ‘subject’ and ‘predicate’ in their functional relationships: in the first case, the clock is the subject, and in the second case the subject is the falling. We carry out syntactical analysis and highlight a difference between the two oppositions ‘noun–adjective’ and ‘subject–predicate’. The distinction between subject and predicate is this: when we have a text, the subject is what we are talking about and the predicate is the characteristic that we ascribe to it. So when I hear any text, I understand it through an analysis: I work out what is the subject. Why do I work it out? I relate it to the situation.

The subject might be an action. In an algorithm I always treat actions as items, to which characteristics are ascribed. So I am always doing a particular sort of work: I parse the text syntactically, identify its syntactical organisation, its predicate structure, and map this onto the situation. This is a process of scanning, of relating the text to the situation. When you understand my text now, you carry out this complex relational work. You are constantly identifying what is being talked about and what I am saying about it. This is the standard work that goes on automatically, you understand what is being said to the extent that you can find these objects and relate the text to them.

These paragraphs reveal several insights that are not only relevant when trying to capture a situation at hand but also when aiming to represent or modeling it. Hence, providing information needs to be considered a context-dependent process itself. Simply by focusing on a specific part of a sentence, like shown above for “The clock has fallen off the wall,” different meanings can be conveyed, and thus, different situations and adjacent work practices could be revealed.

Shchedrovitsky considers ascribing meaning to a situation as relational work. It requires an active entity identifying elements of concern (perceived) information can be assigned to. When we think about selecting a specific meaning out of possible meanings, the initial provision on information for an active system element or actor can have significant influence on the subsequent behavior of this actor and all relations (i.e., the entire system). Hence, in the following, we reflect on modeling the ecosystem in terms of CAS actors using subject-oriented concepts (see Part I of this book) to describe a perceived situation and their use for taking an actor perspective which relates to the semantic and pragmatic quality of models.

We then provide some triggers to rethink how developers elicit and represent situation-sensitive knowledge. A model of eliciting and structuring perceptual knowledge of actors in a certain situation is proposed based on analyzing behavior that could be used as blueprint for engineering digital selves. The model contains several perspectives helping to structure individually perceived situational information for further operation. Each perspective can be enriched with another one leading to a cascade of perspectives, finally allowing to create digital behavior models.

1 Basic Model Generation

Entities in terms of networked active elements, termed subjects, act in parallel. Since each of those actors or subjects can be described in terms of its behavior and has the capability to exchange messages, a federated choreographic ecosystem is established:

  • Federation means a form or single unit, within which each actor or subject or organization keeps some internal autonomy.

    • This form or single unit identifies the perceived part of the world that is considered relevant to describe a specific situation. It sets up the universe of discourse or context space for representation and action.

    • Keeping some internal autonomy at some point requires to be more concrete: The “some” is dedicated to the level of abstraction considered representative for the stakeholders or modelers, both with respect to functional or technical activities and interaction or communication with other subjects.

  • Choreographic ecosystem refers to recognizing concurrent, however, synchronized processes and activities

    • In a community of interacting elements and their environment

    • When considered as networked or interconnected system

According to this perspective, ecosystems operate as autonomous, concurrent behaviors of distributed subsystems or actors. A subject is a behavioral role assumed by some entity that is capable of performing actions. The entity can be a human, a piece of software, a machine (e.g., a robot), a device (e.g., a sensor), or a combination of these, such as intelligent sensor systems.

Since subjects represent systems with a uniform structure, they can be used to define federated systems or System-of-Systems (SoS), featuring autonomy, coherence, permanence, and organization (cf. IEEE-Reliability Society Technical Committee on Systems of Systems, 2014). SoS subjects can execute local actions that do not involve interacting with other subjects (e.g., a clock providing the time in a classroom) and communicative actions that are concerned with exchanging messages between subjects, i.e., sending and receiving messages, e.g., triggering ringing a tone. Figure 10.1 shows a set of federated systems or subjects, Clock, Facility Management, and Clock Producer, that could be considered of relevance for “The clock has fallen off the wall.” The directed links denote the interaction pattern for message exchange.

Fig. 10.1
figure 1

Sample universe of discourse for “The clock has fallen off the wall”

Full size image

As already mentioned, each setting or situation can be structured in subject-oriented behavior modeling as a set of individual networked actors or systems, such as facility devices, encoded in subject diagrams according to their communication with each other. Figure 10.2 shows the corresponding for that pattern (cf. Part I on behavior modeling). The rectangles denote the messages that the systems exchange.

Fig. 10.2
figure 2

Sample interaction pattern for “The clock has fallen off the wall”

Full size image

Figure 10.2 shows a Subject Interaction Diagram (SID). SIDs provide a global view of a SoS, comprising the subjects involved and the messages they exchange. The SID contains a maintenance support process. It comprises several actors (subjects) involved in communication: the Facility Management coordinating all maintenance activities, a Clock Producer taking care of providing a working clock, and the Clock providing scheduling support in classroom management. They exchange messages in case of operational problems as shown along the links between the subjects (rectangles).

Subject Behavior Diagrams (SBDs) provide a local view of the process from the perspective of individual actors (subjects). They include sequences of states representing local actions and communicative actions including sending messages and receiving messages. Arrows represent state transitions, with labels indicating the outcome of the preceding state (see Fig. 10.3). The part shown in the figure represents a service request to the Clock Producer subject from the Facility Management subject.

Fig. 10.3
figure 3

Sample Behavior Synchronization (SBD)

Full size image

Given these capabilities, representations are characterized by (1) a simple communication protocol (using SIDs for an overview) and, thus, (2) standardized behavior structures (enabled by send-receive pairs between SBDs), which (3) scale in terms of complexity and scope. They enable CAS behavior pattern specification.

Subject-oriented models are designed to probe representations for operation: once a SBD, e.g., the Facility Management subject, is instantiated, it has to be decided (1) whether a human or a digital device (organizational implementation) and (2) which actual device is assigned to the subject, acting as technical subject carrier (technical implementation). Typical subjects are devices and their process-specific services, including smartphones, tablets, laptops, healthcare devices, etc. Subjects can also be role carriers controlling or executing tasks. Both types of instantiations can be supported by subject-oriented runtime engines. For a recent overview, see Krenn et al. (2017). These engines provide services linked to some ICT infrastructure.

Once the runtime engine is tightly coupled to model representations, ad hoc and domain-specific requirements can be met dynamically. The situation-sensitive formation of systems and their behavior architecture need to be validated before being executed without further transformation. Hence, models can be adapted according to the SoS their models are part of. In case of existing knowledge based on successful, re-occurring patterns, e.g., for routine tasks, they could be integrated to improve the overall process performance, e.g., including the processing of complex events.

Overall, a subject-oriented representation of any setting can come close to the “reality” as perceived and pictured by humans, both, in terms of its elements as behavioral entities including their set of activities and interactions and in terms of its description as natural language, can directly be used in conveying the content of subject-oriented representations. In order to facilitate practical application (cf. Fleischmann et al., 2015; Neubauer & Stary, 2017), subject-oriented development requires situation-sensitive modeling support from organizational management.

2 Capturing Situations

In this section, we sketch behavior modeling for a specific situation, as introduced in Stary (2020b). We detail the methodological approach and exemplify it in a role-specific way.

When actors perceive situations, they start with spotting relevant elements according to their current perspective leading to assigning a specific meaning: “… meaning is a particular structural representation—a sort of freezeframe—of the process of understanding. … But see what we actually do. Here is a movement. For example, something falls. It leaves a trail. Now we begin to slice this trail into sections, we get parts of the trail and we transfer it to the movement. … We divide it into stages and phases, but in order to do this we have to find and register the traces (the trail) of this process” (Shchedrovitsky, 2014, p.43).

The “trail” may range from realizing the trigger event for the clock’s falling down to watching how the broken glass spreads over the floor in the glass room. Evaluating this trail allows to scope the entire scene in terms of all relevant elements involved, e.g., the holder went off the wall, the clock fell down, and the clock fell apart when touching the floor. Hence, meaning could be action triggered which in turn is relevant for the stakeholder in the room. Assuming that nobody got hurt through the event, for the students in the room, it may be an event of low complexity, as they do not have to care about the time and are able to watch their steps when avoiding to step in the clock’s broken parts. For the teacher, it is a major event, as he/she needs to take care about the time and the safety of the students.

As we can see, each actor constructs meaning through some role-specific glass. It may require immediate action or reaction to an event. The teacher may take action through interrupting the process of teaching and switching to the role of caretaker of classroom safety when warning the student when leaving the classroom. From the teacher’s perspective, in a second step, the time problem needs to be addressed, assuming classes are structured along time slots. The teacher needs to interact with somebody from the class or facility management to ensure correct timing, in case he/she relies on an external source of information with respect to time. Finally, the facility management needs to be addressed for taking care of all the damage. From a representational perspective, several entities are involved to make meaning out of a situation:

  • The event itself—being an action itself (falling off the wall ending another operation, namely, the time ticking) or “sliced,” a set of small actions or events

  • The role—student, teacher, caretaker, and facility management

  • Actions and interactions, such as teaching and warning the students

  • Concerned objects, the clock and the classroom

Each of these elements is constitutional to model representations. Subjects denote roles, encapsulating behavior in terms of doing, sending, and receiving messages. Finally, the concerned objects are addressed in or passed through messages exchanged between subjects.

Conveying meaning to others as another self

Situations trigger not only certain behavior but also need to be documented and transferred to others, e.g., to guide further behavior. It could happen that communication is not documented explicitly. Enforcing to think in terms of communication and interaction of actors enables taking further perspectives and further behavior specifications. For instance, the teacher subject (i.e., an actor role) activates the caretaker which in turn activates the facility management.

Aligning selves through goal-oriented behavior abstractions

In order to handle a certain situation, it may not be sufficient to develop a chain of interactions from a single perspective. For instance, administration, technically not involved into the clock falling off the wall, needs to be activated to ensure whether the classroom can be utilized by students for the next class. “The work of organisers, leaders and managers has the character of engineering work: it is structural and technical. Organisers, leaders or managers must always be one step ahead; they have to come up with something new” (Shchedrovitsky, 2014, p.7f).

An alignment scheme for individual and organizational activities alignment according to Shchedrovitsky (2014, p.11) is based on actor-specific goals and interaction relations, i.e., to know whom to involve in which way for further operation. As we will see in the following, the goal can help in identifying intentional actor performing self-contained tasks according to the perception of a situation. In addition, the means of organizing work could be subject orientation which needs to be probed by applying the model. Perspectives on the situation trigger:

  1. 1.

    Technical entities encapsulating behavior by focusing on activities that need to be performed to achieve an objective or implement an intention (usually referring to some task) and, thereby, establishing some functional role

  2. 2.

    Communication acts identifying which entity needs to be interacting with another

  3. 3.

    The mutually adjustment of encapsulated behavior specifications, as it plays a crucial role not only for acting as a collective in a specific situation but also to complete processes or reach intended goals

Accordingly, the model contains several perspectives helping to structure individually perceived situational information for further operation. Once started with an individual perspective, actors can enrich its result with another one and so on, thus leading to a cascade of perspectives.

Figure 10.4 shows the model serving as a frame of reference of building system capacity based on individually perceived situations. It instantiates Shchedrovitsky’s approach in terms of structuring behavior in a goal-oriented way. The left part shows the cascade of perspectives that finally captures the evidence of a specific stakeholder when perceiving and reflecting on a situation:

  • Perspective 1—Individual Actor View: This perspective captures a set of individual roles in which this stakeholder can act and thinks about in a specific situation. For instance, assuming the clock has fallen off the wall in a classroom with a teacher and students, the teaching role of the teacher addresses all duties related to classroom teaching, whereas the safety-responsibility role of the teacher concerns the physical safety of students in the classroom. Since humans are intentional beings, we can assume that each actor has at least one role or objective to (inter)act that constitutes an actor view. This role or a set of roles corresponds to the individual (task) profile of a person or an artifact. Each role refers to a specific behavior that has a driver, namely, an intention. For instance, the driver of the teaching role is increasing the level of competence of students, whereas the driver of the safety-responsibility role is ensuring the safety of all students in the classroom. Since each role has an intention, each actor can pursue a set of specific goals in a situation, depending on the set of roles.

  • Perspective 2—Individual Interaction View: This perspective looks on the same situation, but builds upon the results from taking perspective 1 and the further identified roles. It keeps the considered role/objective/intention at the center of interest, but additionally captures a set of individual interactions based on that previously defined intentional behavior set(s). Hence, the set of interactions also depends on the roles in which this stakeholder can act and thinks about in a specific situation. For instance, we assume the stakeholder identifies the role of the teacher (addressing all duties related to classroom teaching) and the safety-responsibility role (ensuring the physical safety of students in the classroom). Then, from this perspective, the stakeholder needs to think about interactions between these two roles. In case the teacher interrupts the class due to the clock’s falling off the wall, the safety-responsibility role takes over to ensure the safety of the students in the room. It may lead to ending the class, once the teacher cannot guarantee the safety of the students in this situation, as perceived by this stakeholder. In case the safety-responsibility role does identify safety risks, the safety-responsibility role informs the teaching role to continue teaching. In each case, the stakeholder can provide and specify a set of interactions, for sending and receiving information on a certain topic, involving relevant objects, such as safety measures.

  • Perspective 3—Organizational Interaction View: This perspective analogously builds upon existing results, this time from taking the previously described perspectives 1 and 2. They already include roles and interactions, however both from an individual perspective. This perspective captures a set of roles this stakeholder perceives to be relevant for a specific situation in addition to the ones he/she can act him-/herself, e.g., taking a community or network perspective. It concerns a set of roles the stakeholder having perspective 1 and 2 cannot take or has no privilege to take. For instance, assuming the clock has fallen off the wall in a classroom with a teacher and students and has been damaging some interior, neither the teaching nor the safety-responsibility role is sufficient to continue with giving a lecture in this classroom. Like from perspective 1, another individual actor view is driven by an intention. In the sample case, the goal could be to keep the classes running that are assigned to this room. Then, the interior needs to be restored, which brings in facility management. Its specific behavior needs to be coupled to the safety-responsibility role, in order to accomplish the respective tasks. Finally, there may be several perspectives related to the “We,” e.g., evolving from an internal community of practice to formal department, networks, regions, and global connections.

Fig. 10.4
figure 4

Cascading perspectives

Full size image

Since each perspective builds upon a previous one, a cascade of perspectives evolves in the course of specifying behavior-relevant information. The middle part of Fig. 10.5 reveals the evolving complexity according to refined and networked behavior specifications. The generation of actors and their interaction relations are based on a set of questions that trigger the definition of subjects and their interactions.

  • Initial set of subjects: The Individual Actor View leads to a set of intentional actor roles that allow stakeholders performing goal-oriented activities. The stakeholder at hand identifies the initial set of behavior abstractions (subjects) by dealing with the question “What can I do now?”. This question targets those behavior abstractions that a stakeholder can name, once a goal to be achieved in this situation becomes evident. For instance, in case the clock falls off the wall of the classroom, the ultimate goal of a teacher is to ensure the students’ safety before proceeding with the lecture. In order to achieve that goal, the stakeholder can perform a set of technical activities.

  • Interacting initial subjects: The Individual Interaction View leads to a set of intentional actor roles that synchronize their behavior. The stakeholder at hand identifies all those interactions between the initial set of behavior abstractions (subjects) by dealing with the question “How do ‘I’ interact?” when having identified more than one role for handling a specific situation. For instance, in case the clock falls off the wall of the classroom, the safety-responsibility role interrupts the teacher to ensure the students’ safety before signaling him/her to proceed with the lecture. Hence, the interactions are defined, in order to achieve the stakeholder goal determined upfront.

  • Collective of interacting subjects: The Organizational Interaction View leads to a set of intentional actor roles and synchronization of their behavior beyond the stakeholder at hand. This time, he/she needs to answer the question “How do ‘We’ need to interact?” when embedding further actor roles for handling a specific situation. For instance, in case the clock falls off the wall of the classroom, the safety-responsibility role informs facility management, in case he/she cannot ensure the students’ safety. Every interaction with the facility management needs to be defined, in order to achieve the upfront determined stakeholder goal.

Fig. 10.5
figure 5

Sample representation

Full size image

Figure 10.5 exemplifies the cascaded perspective. In this case, the stakeholder has identified “teaching” and “safety-responsibility” as role representatives for perspectives 1 and 2 which need to interact sensitive to the safety of the students. For the repair of the clock and classroom restoring, this stakeholder activates facility management through respective interactions.

The “We” perspective can be extended to bring in additional stakeholders, e.g., authorities managing school infrastructures, that are contacted in case needed, e.g., by facility management, to improve the interior. Hence, the number of cascaded perspectives depends on the intention and goal of the stakeholder and results in a systemic view. The schema allows on the one hand focusing on a perceived part of a situation while on the other hand extending perspectives limiting contextual or systemic thinking by enabling interaction links to actor roles valid from other perspectives.

Both elements are essential, as they allow handling complex situations or events without reducing the complexity itself, but rather offering a multi-partite structure. This structure facilitates handling (complex) situations

  1. 1.

    By starting with familiar, since ego-centric behavior encapsulations (roles), and then

  2. 2.

    Stepwise enriching this set of roles by

    1. (a)

      Sets of interactions between ego-centric behavior encapsulations

    2. (b)

      Including non-familiar behavior encapsulations (roles)

    3. (c)

      Coupling them through sets of interaction to all other behavior encapsulations

Hence, without pre-determining the number of perspectives and the number of modeling elements (behavior encapsulations, interactions), a stakeholder can be encouraged to express his/her perception of a situation based on interacting behavior elements. These elements represent subjects allowing to detail pragmatic information in terms of role-specific (internal) behavior. The latter is represented in Subject Behavior Diagrams (SBDs). Given the interaction between the subjects, a SID and, thus, a stakeholder can create a coherent pragmatic model of a situation.