Dear all,
This is my best effort to write about systems ontology. References are not in very academic manner (e.g. references to some explanations in Wikipedia that is not too rigorous) but I tried to be understandable.
Towards the ontology of third generation systems approach
Anatoly Levenchuk, ai...@asmp.msk.su
Systems engineering as a practice of changing the world for the better is based on a systems approach ontology (sometimes said "systems ontology", including all the concepts needed to implement a systems approach in thinking). There are several modern initiatives on the ontological commitments of such an ontology, which we consider the third generation of the systems approach, taking into account thermodynamically driven evolution. We propose ways to harmonize these approaches to create a systems ontology.
A systems ontology as a set of descriptions for conceptual directing of attention
Gruber, Borst and Studer [1] proposed that an ontology is a formal, explicit specification of a shared conceptualization. Let's define conceptualization as a specification of important objects of perception in the world, as a way of indicating objects that would be good to distinguish in the world for reliable active/embodied inference [2]. This allows us to formulate the task of creating an ontology in terms of attention management, and "explicit formal specification" here indicates that these objects of attention are not spontaneously singled out, but according to some explicit model as ontology specified by using another model as its formalism (foundational ontology).
We resolve the issue according to Popperian epistemology [3]: objects in ontology appear by guessing, the acceptability of these guesses for judgments about the world is questioned, but guesses that survive criticism are "taken seriously." That is, good guesses about how the world is made up of systems (what is the ontology of the system) and judgments that these guesses are useful for changing the world for the better can be taken from literature, and if we don't know their critique and can't immediately suggest falsification of those conjectures by reasoning and/or experiment, then we consider them our best theories (SoTA, state-of-the-art) about the world, take them seriously and then teach people (replicate memes of those theories) to make shared those conjectures about important objects.
Of course, we don't just tell people the ontology of the system, but we teach them to attend abstract and physical objects in the world, described by shared explicit and as formal as possible specification, i.e. ontology. Problems emerging within activity that was performed with this ontology as a theory of world means that ontology is falsified, but this is simply a reason to correct the ontology by correcting identified errors. Ontology of the systems approach thereby continuously evolves, it represents at each point in time the best we know about systems at that point. We [4] have implemented a curriculum in which several hundred people a week are now learning to identify in the world objects described/typed by systems ontology. The curriculum implements teaching students fully conscious at the start and then interiorized "automatic" conceptual guidance of students' attention to objects described/typed by systems ontology. The student is trained to:
- to direct one's own and other agents' attention to objects from an explicit specification (ontology), that is, to direct attention conceptually rather than spontaneously (Ontology and Communication course),
- to hold conceptually guided attention on a wide variety of time scales, including collective attention in such agents as a team and whole enterprise via leadership, or influencing on community or society (Self-Collectedness course),
- to direct attention to systems (Practical Systems Thinking course),
- to direct attention to the activities/practices that are carried out by constructor/enabling systems (Methodology course),
- structure the activities and roles that are necessarily held in the attention of systems engineering projects, (Systems Engineering course),
- to draw the attention of oneself and the team to the objects described by the systems ontology specialization for systems-creators such as organizations/businesses (Systems Management course).
In order to teach students a SoTA systems ontology, we have harmonized several particular conceptualizations of the concept of systems (i.e. we performed merge of partial systems ontologies), which after merging represent together ontology of the third generation of the systems approach. As usual, each generation of systems approach incorporates all the achievements of the previous generation, but adds something new.
The first generation of the systems approach: the system in its environment during operations time
The first generation of the systems approach emerged in the 1940s, mainly as a result of the work of von Bertalanffy [5]. The notion of a system as a subject separated from its environment appeared in physics a long time ago, but the systems approach as a consideration of the whole world as interacting systems appeared mainly after von Bertalanffy's works on general systems theory. "Approach" is the usual term for a situation in which ontology, successfully developed and tested in one domain, begins to be used in many different domains. The key here was the realization that systems are interacting holons (Koestler coined the term [6] to describe a part of the whole that itself made up of parts), plus these systems somehow appear in the world and then disappear from the world, i.e. undergo a life cycle (von Bertalanffy was a biologist and generalized among other things the success of the systems approach in biology, that is the life cycle is the birth-growth-breeding cycle, common in biology). The interaction of parts of the system during work/operations results in emergent properties of the whole. The gears in the clock do not yet shows the time, the clock shows the time, the house with the clock inside no longer shows the time. The key here was that the first generation of systems thinking operated on at least two different ways of partitioning depending on the time of consideration:
-- functional partitioning depending on the purpose of the system in the suprasystem and the purpose of the subsystems in the system. This viewpoint needed to think about the system at the time of its operations.
-- a constructive/modular partitioning needed to think about the system at the time of its creation/construction.
The difficulty in mastering systems thinking was mainly that people found it difficult to grasp the concept of selecting/isolating dynamic objects in the world with their attention -- right in the system operations time. More often than not, they imagine an "explosion-diagram" in their heads when mentioning decomposition into functional parts, at the same time losing the multilevelness of such a partitioning [7]. It is clear that one cannot discuss the emergent properties of stopped and broken down system into single-level physical constructive parts, because interaction performs in operations time and viewpoint must be functional, not constructional. But the partitioning into constructive parts is also important, for such a system must be created in construction (design, implementation, test, deploy, transfer to operations) time. For example, scissors functionally consist of a cutting unit and a handle, while structurally for construction they consists of two halves of scissors and a screw that fastens them together. The scissors user is concerned with the functional viewpoint, while the factory engineer is concerned with the constructional viewpoint.
Bertalanffy included systems engineering in his proposed set of systems disciplines. Then systems engineering rapidly developed. In systems engineering, the target systems were physical: this gave guaranteed grounding for all descriptions. At the same time the descriptions themselves as a system were not considered due to interaction (i.e. changes in time under the action of each other) of parts of abstract objects that comprise views in descriptions cannot show emergence, mereotopology of mental/abstract/mathematical objects that comprise views for the most part ended in failure.
The second generation of the systems approach: system-of-interest, which is created by constructor/enabling systems
The second generation of the systems approach relied mainly on the ideas of systems engineering: physical (including cyber-physical, including cyber-physical with people) systems-of-interest are created by people with tools, not reproduced themselves in a cycle, and these systems-of-interest are mostly non-living. This doubt the notion of the life cycle, which turned out to be neither life nor cycle, although the term for the concept remained in use. In the second generation there appeared constructor/enabling systems. That introduce, besides the part-whole relation, the construction/enabling relation and the viewpoints for activity of constructor systems and interaction of constructors and systems-of-interest (the most famous at introducing of this viewpoint was works of Checkland in the late 70's and early 80's [8]). In systems engineering was introduced the systems engineering management with viewpoint of construction/enabling of construction/enabling systems, thus appear chain of constructors/enablers (constructors of constructors of system-of-interest).
The diagrams of systems engineering emphasized the "waterfall" model of the life cycle, in which the system was conceived, designed, implemented, tested, operated and decommissioned, where the project usually ended. The living system did all this work for itself mostly by itself (with some conceive and design work done for the system by evolution, but this was not usually considered), but in engineering it was all done by the constructor as a separate system. Different roles (stakeholders) in the project as constructor chain had different interests/concerns with respect to the system and its project, which required different kinds of descriptions/views that were made with respective viewpoints as methods of system description that were well suited to reconcile the very different concerns of the very different project roles.
The ISO 42010 [9] and ISO 15288 [10] standards solidified the systems conceptualization for systems engineering, this ensured that the second generation systems engineering ontology of the systems approach was shared. A formalization of systems ontology based on 4D extensionalism (the idea that if two objects occupy the same place in space-time, they are the same object -- and it is a physical object) was proposed in ISO 15926-2 [11] in 2003, it drew on ideas from BORO [12]. Based on these 4D ontological commitment several other similar ontologies were proposed mainly for military applications, this work was led by IDEAS Group [13]. Special mention can be made of the work of Matthew West, who proposed the HQDM ontology based on the same ideas of 4D extensionalism [14], in which the concept of the system was made one of the central ontologies. All these approaches assumed not just conceptualization specification, but also formal expression of conceptualization in logical ontology description languages (such as very rarely used EXPRESS [15], but also appeared later more popular OWL[16]) to create data models in PLM systems databases [17]. The basic ontological premise was to use a 4D ontology, which gave a good and compact description of the changes that the system undergoes during its construction from parts [18].
The peak of this line of work on the second generation of the systems approach that evolved in systems engineering, came in 2008-2013, after which interest in such a logic formal systems ontology descriptions faded a bit: the development of ontologies as an explicit formal specification of shared conceptualizations was no longer seen as an advance in AI creation, and not even the rebranding of formal ontologies as knowledge graphs helped [19]. Semantic web was expected as an internet mainstream Web 3.0, but this idea was never realized, semantic technologies remain niche.
<b>Continuous everything in engineering as techno-evolution</b>
The limitation of the "waterfall model" [20] was the idea of a single unidirectional passage the life cycle as a set of works carried out constructors in accordance to life cycle practices. System-of-interest was considered as passively undergoing this one-time creation. Therefore, agile approaches developed in engineering, overcoming the concept of a "waterfall" life cycle, in 2001 the Manifesto for Agile Software Development appeared [21]. In 2017, the idea of evolvability as a core architectural characteristic reflecting "continuous everything" notion was not only entrenched in best engineering practice, but also entrenched in the literature [22], and architecture was finally separated from development into a separate domain, as the literature began to reflect the inevitable "productive conflicts" between architects and developers. Architecture worked on slicing the system into minimally interacting modules to maintain stability of the whole in the evolutionary time scale (maintain characteristics such as -ilities well known in engineering appear as architectural characteristics), while developers were concerned with maximizing functionality, which could entail temporary deterioration of architectural characteristics as an accumulation of "technical debt" (work aimed at maintaining selected architectural decisions, but not directly affecting functionality).
The idea of evolution came to engineering, but it differed from the idea of biological evolution because the systems memome as an analogue of the biological genome was not contained in the manufactured system, but was separately stored in a digital (i.e. implying exact multiple replications without accumulation of errors of analog representation) form somewhere in constructor systems. This allowed significantly speeding up the time of evolution, because for some techno-crab it was not necessary to wait until the whole crab died to replace the unsuccessful claw variant only together with the crab by mutations of genes forming the phenome of the claw. It was also possible to use smart mutations to techno-crab memome, getting results, i.e. in real life to try only claw variants that showed success in simulations in the virtual world -- and to change only the claw, but not the whole crab. This gave a high rate of techno-evolution and its results, unattainable only by a one-time offer of a set of new features (one-time "waterfall" engineering) or only evolution with random mutations [23]. To understand how successful the system is in operation, was adopted system modeling at operation time -- emerged concept of a digital twin reflecting not only the memome, but also the phenome of the system. So a systems ontology focused on two times (1. operation and 2. creation of a single increment/feature) was not enough. It was also insufficient to model a whole group of systems (product lines [24], sometimes system families) as different variants of the system coexisting at one point in time.
The way out was the idea of "continuous everything" [25] as infinitely ongoing system development with continuous delivering/commissioning of more and more variants of the target system with more and more changes, and speed up of slowing with growth of system-of-interest complexity development is supported both by the evolving architecture and the changing structure of the teams that perform this development (using Conway's inverse maneuver [26]). In software engineering this idea was developed within the DevOps/SRE/platform engineering approach [27], in traditional systems engineering this idea is discussed now on examples of aerospace systems, where testing of the next version engines now starts before the tested engines of the first version in the rocket perform their first flight, and each new instance of rocket hardware has an improved design. "Continuous everything" means moving engineering to techno-evolution projects, that is, evolutionary changes in the meme -- continuous improving the information model of the system design, not just the improving phenome by "field refinement" of an instance of the system. The continuous development as continuous evolving of a system tries to make it as close to endless survival as possible, adapting the system-of-interest as long as possible to changes both in the environment, in itself and in a chain of constructor systems. In biology and life sciences similar problems of approaching infinite in time development/techno-evolution are discussed as open-endedness [28].
Thus a requirement to reflect the ideas of evolution in systems ontology emerged. If there were a developed ontology of evolution, it could be used to develop a third generation systems approach that would not only account for construction/enabling systems as agents in their different project/activity roles, but would also work explicitly with three times: 1. operations time of the instance of a system with intended by design phenome, 2. construction time of an system instance with an increment change in design, i.e. phenome from memome implementation, 3. continuous changing of memome time (techno-evolution, continuous everything).
Scaleless descriptions of physical systems
In the discussion of evolution in biology, the systems approach manifested itself in reality in large evolutionary transitions: increasing complexity in systems from large molecules to cells, from cells to multicellular organisms, etc.. Approximately the same increase in the complexity of systems occurs in techno-evolution: transistors eventually composed to chips, chips become computers, computers evolve to data centers, data centers transit to global computer networks. The analogy seems to be clear, but it was required to get a scaleless theory that will be common for systems of any scale and origin, i.e. for the world of elementary particles in nuclear physics taking into account quantum phenomena, but also for macro objects, including living and non-living, as well as living conscious and collectives of living conscious beings (for example, mankind as a whole). Scaleless also meant scaleless in the fourth dimension, time: taking into account the system construction/growth time, the time of operations, but also the time of evolution.
The key was the transition to a formulation of physical phenomena as informational, and an explanation of the phenomenon of the stability of objects in the physical world - why some objects (such as a molecule or a person) maintain their shape in space-time. Fields, Glazebrook, Levin, and Friston proposed an ontological framework of panpsychism in minimal physicalism form to describe physically stable systems as implementing the principle of free energy minimization and covering the whole complexity spectrum from elementary particles to people and society [29]. Free energy is defined as an informational characteristic of a system rather than the traditional energy of mechanical work or the work of electromagnetic forces. These systems increase in complexity from elementary particles through molecules, through штуке matter bodies to living creatures, where they increase in complexity from unicellular to multicellular organisms and their populations. All of these types of systems reduce the Bayesian (or excess Bayesian to account for the quantum-like nature of computation in biology [30]) surprise of mismatch between expected measurements according to generative model of world and real measurements in the reality.
This scaleless ontology for physically stable systems has been mereologically formalized using category theory as a foundation ontology [31] and expressed in a form that allows describing quantum-like active/embodied inference [32]. This line of ontological engineering shows well how to think about the functioning of systems of very different evolutionary complexity, including how to apply stochastic methods to non-ergodic systems, that is, systems with memory. One of the strongest confirmations of such a theory was the creation of hybrot [33], which learned to play Pong: a sufficiently complex system capable of learning (such as a natural neural network) must exhibit, due to the principle of free energy minimization, behavior that minimizes unpredictability of the external environment -- and exactly this was shown by experiment [34]. In fact, in this line of research was achieved formalization and mathematization of ideas of the first generation of the system approach in form of the physics-based ontology. Due to the scaleless nature of this ontology, other time scales also can be covered: the time scale of system life cycle, and the time of evolution. The key to this was the notion of measurement, developed in quantum physics to solve the problem of the observer. Measurement is an interaction of systems, not just passive perception/observation. The reverse is also true: the interaction of systems in construction described as the inverse of measurement. This means that any interaction of even a molecule as a system with its environment can be considered as a measurement or construction: a molecule is a proto-agent that somehow perceives/cognizes/measures the world around it, somehow maintaining its stability in it due to the principle of minimizing free energy.
This line of reasoning (perception/measurement as inverse to changing/constructing and both are interaction of systems) was drawn in the constructor theory proposed by Deutsch and developed by various researchers of "quantum gravity" that needed scaleless theories [35]. A constructor is some physical device that can maintain its immutability for a long time while repeatably changing its environment by some pre-described sequences of operations (e.g. a catalyst molecule, or a robot with a universal computer, or a living being). Sufficiently advanced constructors can replicate themselves, among other things, to be the subject of evolution. Scaleless physical theories based on notions of information-related changes as computations have proven to be very productive.
Constructivism: systems perform operations on both physical objects and abstract objects
The next thing to be considered in the physics-based system ontology is the ontology of physics and mathematics itself as a foundation ontology. Such an ontology has been proposed in many works, but let us highlight the work of Deutsch [36] in which it is proposed to consider physics as a science of real objects, mathematics as a science of mental/abstract/mathematical objects, and computer science as the experimental science of proving that the behavior of physical objects can somehow reflect the behavior of ideal/abstract/mathematical objects, i.e. science of universal computers as physical devices capable of performing computation -- traditional electronic, quantum, etc., including biological and social computers such as human brain and even collectives of humans altogether with their computers.
These ideas imply a shift of the discussion from static models and passive data to universal computers as physical devices that interpret such data and change the state of the environment depending on these computations (input of raw data for computations and output of results in symbolic form are just a special case here. Perception of the environment and changing the environment and/or "self" as a computer/constructor device/system is just a more general case). Central to this approach will be the notion of constructor, a physical device that can interact with the environment, performing repeatedly some operations in it as described, while maintaining its stability (e.g. a catalyst molecule, or a robot, or a human). So, we have some approach to formal description of systems of the second generation of the systems approach: some systems of different degree of agency (from a dense matter, robots, intelligent agents and all kinds of their hybrids) construct according to some design and method data different types of systems-of-interest, which further operates in their environment.
One more attempt of integrating physics, topology, logic and computer science with a shift of attention to enactive operations/morphisms with objects instead of considering static objects in their static relations, was proposed by Baez and Stay in the Rosetta stone approach [37]. After moving from describing system interactions as processes in networks (electrical circuits, hydraulic networks, as well as networks of interactions in system dynamics, usually associated with functional representations of the systems, that is, representations of system in it environment in operations time, the first generation of the systems approach) Baez makes a proposal to use the formalism of symmetric monoidal category theory to describe not just processes, but open systems interacting with the environment [38]. Again we see a move to use category theory as a foundation ontology for systems ontology to representation ontology not as set of static objects and relations but morphisms as object changes. This move corresponds to a general move toward constructivism in mathematics, where we move from eternal classes and their relations to operations of construction [39]. This make it possible to reformulate of mereology as the central ontological discipline of the systems approach from research of the "part-whole" relations [40] to research of operation of construction parts to whole.
Another move in this direction of constructivist mereology is Fine's ideas about the mereology that includes abstract objects [41], which suggest operations of constructing the whole from parts, and this also applies to descriptions (constructing a set from its elements). Although Fine does not explicitly say so, defining construction operations (as well as category theory morphisms) performed as if by "nobody", but behind these operations one can easily see a physical device-constructor from constructor theory, and for abstract parts physical device, implementing a "universal computer" (in general it is the same constructor from constructor theory, including quantum computer and living mathematician -- they are computationally equivalent, and physically embodied to somehow get information about input and output data of computation). Fine's approach allows us to pose the question of the notion of a system extending beyond physically interacting parts. Interaction of physical parts driven by abstract objects can be inside computer device. Definition of a system as interacting parts with emergence persists, but interaction appears in the constructor/computer system rather than in the abstract system-of-interest. Along this line of reason it is possible to discuss systemicity in complex cases of communities that seem to exist in physical world, but which are difficult to discuss on the basis of the interaction of the members of these communities with each other: the interaction occurs by carrying out operations of reasoning about these communities in the constructor's (social engineer's) computer [42].
This line of work on a constructivist reformulation of second-generation systems ontology was continued with the creation of Core Constructional Ontology [43], which offers a constructivist perspective on the theory of parts, sets and relations and serves as one level of foundational ontology for systems descriptions in PLM systems of engineering projects. For core constructional ontology lower level foundational ontology is mathematics, and expressing 4D mereotopology of space-time in 4-Dimensionalist Top Level Ontology give us next ontology level as upper ontology for modeling of systems in contemporary systems engineering [44].
The use of some mathematics (most often in computational ontology these are some variants of first-order logics, but more and more often this means morphisms and, respectively, category theory) as the foundational ontology with the omission of all other (upper ontology, middle ontology) ontological levels is common in physics. Physics itself, in terms of systems ontology, deals with functional objects (e.g. "physical body"), whose roles are played by different objects of the physical world (mountains, stones, molecules, elementary particles). Then in conjectures and experiments the peculiarities of behavior of these functional objects are found out and they are expressed by mathematical relations. Structural similarities of formulas describing some different objects rise question about the general ontological nature of these objects. This can be made as analogical reasoning. This technique is also used for ontologies expressed logically as lattices (knowledge graphs), an example of which is VAE (VivoMind Analogy Engine) [45]. An example of the use of such a technique by physicists can be found not only in works on information theory ("free energy" in information theory have its name due to appearance at the same place of similar formulas in informatics theory that it appears in formulas of thermodynamics [46]). So there is a tendency to use mathematical/abstract objects and formulaic relations between them to formulate a third-generation systems ontology in order to get physical explanation of evolution leading to large evolutionary transitions lead to multiple system levels in stable objects.
The third generation of the systems approach: considering the evolution time
Vanchurin, Wolf, Koonin, and Katsnelson observed that if one continues to apply analogy reasoning to classical thermodynamics and information theory to quantities/quantities in thermodynamics, machine learning and evolutionary biology [47] this lead to surprirsing common framework/ontology for all of it. In other words, evolution could be described as learning, and learning is thermodynamic in nature, i.e. physical phenomenon. Based on these notes, they formulated theory of evolution as multilevel learning [48]. In this theory, the driving force behind evolution is frustrations arising from conflicts between objects of different system levels. This notion of frustrations was introduced into the systems language by studies of spin glasses as an example of the behavior of non-ergodic (with memory) systems and there it meant geometric frustrations as impossibility of stable geometry of spins in glasses [49]. Evolution is thus a learning process, which is reduced to solving the optimization problem of finding a minimum of the free energy by changing the structure of numerous system layers. In the course of this, evolution finds quasi-minima, but not an absolute minimum. From time to time there is a jump of complexity growth (appearance of another system level, another type of whole systems, e.g. multi-cell organism constructed from cells), which gives a sharp minimization of the free energy of the evolving system, but still this is just another quasi-minimum, not an absolute minimum. In [50] the physical/thermodynamical nature of the growing complexity of systems was shown with example of biological systems, but the reasoning there is scaleless, i.e. the living or inert systems or with the presence of consciousness or not do not influence the conclusions of the paper. So in systems ontology there are also should present notion of conflicts between system levels, and notion of the phenomenon of frustrations caused by these conflicts, and the rule (law, conclusion) about the inevitable growth of system complexity, i.e. inevitabile increase in the number of system levels of stable systems in the course of evolution.
All these researchers are in contact with each other (for example, there were discussions between Vanchurin and Friston, this is indicated at least in the acknowledgment to [47]), because most of these works are based on thermodynamics and the understanding that all systems implement the free energy minimization principle both during their existence, during their creation/construction, and during continuous evolution. The translation of the traditional formulas for expressing this ontology (now this ontology is expressed in the formulas familiar to physicists, traditional for thermodynamic calculations) into the constructivist form of category theory is a separate topic, but such a formulation of the problem is more or less usual for mathematicians and we can discuss the research program on this topic. The same can be said about the reformulation of the differential form into a quantum-like form, which could increase the accuracy of the physical modeling of biological evolution and techno-evolution (memetic evolution with separated memome) due to the quantization/digitization of most of the described phenomena [51].
Thus, the modern/SoTA (third generation systems approach) ontology of the system:
-- provides types of objects for layered directing of attention to ensure the evolution of target systems (continuous everything) by constructor systems
-- considers at least three times of system existence: operations, construction/evolving of phenome, evolution/development of genome/memome.
-- is based on physics, mathematics, and computer science
-- treats systems as stable entities within a minimalist physicalism (including systems active with respect to themselves and their environment, systems seeking minimum free energy by active/embodied inference, regardless of their level of "intelligence")
-- gives scaleless descriptions of systems (phenomena of quantum physics are thereby taken into account), explains the emergence of system levels (growth of complexity) due to multilevel optimization to achieve the minimum of free energy
-- mereology is no longer expressed through eternal classes and relations between them, but rather through morphisms and operations reflecting operations with physical systems in the course of their interaction as well as operations with abstract objects performed by constructor systems with a (universal in the sense of Turing machine equivalence) computers in them.
References:
[1] Nicola Guarino, Daniel Oberle, and Steffen Staab, What Is an Ontology?, in S. Staab and R. Studer (eds.), Handbook on Ontologies, Springer-Verlag, 2009
[2] Karl Friston, Embodied inference: or "I think therefore I am, if I am what I think" In W. Tschacher, C. Bergomi (Eds.), The implications of embodiment: Cognition and communication (pp. 89–125). Imprint Academic., https://www.fil.ion.ucl.ac.uk/~karl/Embodied%20Inference.pdf
[3] Karl Popper, https://plato.stanford.edu/entries/popper/
[4] Школа системного менеджмента
[5] von Bertalanffy, https://en.wikipedia.org/wiki/Ludwig_von_Bertalanffy
[6] Koestler, Arthur (1967). The Ghost in the Machine
[7] Bellami, Nikon F3-P Parts Diagram, https://www.japancamerahunter.com/2014/11/nikon-f3-p-parts-diagram/
[8] Peter Checkland, https://en.wikipedia.org/wiki/Peter_Checkland
[9] ISO/IEC/IEEE 42010, Systems and software engineering — Architecture description, https://www.iso.org/standard/50508.html
[10] ISO/IEC/IEEE 15288, Systems and software engineering — System life cycle processes, https://www.iso.org/standard/63711.html
[11] ISO 15926-2, Idustrial automation systems and integration — Integration of life-cycle data for process plants including oil and gas production facilities — Part 2: Data model, https://www.iso.org/standard/29557.html
[12] Peter El Hajj, Business Objects Reference Ontology, https://digitaltwinhub.co.uk/top-level-ontologies/business-objects-reference-ontology-r2/
[13] IDEAS Group, https://ideasgroup.org/
[15] ISO 10303-11:2004 Industrial automation systems and integration — Product data representation and exchange — Part 11: Description methods: The EXPRESS language reference manual, https://www.iso.org/standard/38047.html
[16] Web Ontology Language (OWL), https://www.w3.org/OWL/
[17] What is PLM (Product Lifecycle Management), https://www.oracle.com/uk/scm/product-lifecycle-management/what-is-plm/
[18] Ian Bailey, The simplification in integration architecture that 4D supports, https://gateway.newton.ac.uk/sites/default/files/asset/doc/2105/Ian%20Bailey.pdf
[19] John Sowa, Knowledge Graphs for Language, Logic, Data, Reasoning, https://www.youtube.com/watch?v=J9_kXpZAcQY (Part of the Ontology Summit 2020)
[20] Waterfall model, https://en.wikipedia.org/wiki/Waterfall_model
[21] Manifesto for Agile Software Development, https://agilemanifesto.org/
[22] Neal Ford, Rebecca Parsons, Patrick Kua, Building Evolutionary Architectures, 2017, https://www.oreilly.com/library/view/building-evolutionary-architectures/9781491986356/
[23] Joel Lehman, Jonathan Gordon, Shawn Jain, Kamal Ndousse, Cathy Yeh, Kenneth O. Stanley, Evolution through Large Models, 2022, https://arxiv.org/abs/2206.08896
[24] Software Product Lines matures into the next generation of Systems and Software Product Line Engineering, https://www.softwareproductlines.com/
[25] The Continuous Everything (CE) concept defined, https://en.itpedia.nl/2021/06/02/het-continuous-everything-ce-concept-gedefinieerd/
[25] Nicole Forsgren, Jez Humble, Gene Kim, Accelerate, 2018, https://www.oreilly.com/library/view/accelerate/9781457191435/
[26] Jonny LeRoy, Matt Simons, Dealing with creaky legacy platforms, 2011, http://jonnyleroy.com/2011/02/03/dealing-with-creaky-legacy-platforms/
[27] Aeris Stewart, How Is Platform Engineering Different from DevOps and SRE?, 2022 https://thenewstack.io/how-is-platform-engineering-different-from-devops-and-sre/
[28] Kenneth O. Stanley, Joel Lehman and Lisa Soros, Open-endedness: The last grand challenge you’ve never heard of, 2017, https://www.oreilly.com/radar/open-endedness-the-last-grand-challenge-youve-never-heard-of/
[29] Fields, C., Glazebrook, J. F. and Levin, M., Minimal physicalism as a scale-free substrate for cognition and consciousness. Neuroscience of Consciousness 2021, https://chrisfieldsresearch.com/min-phys-NC-2021.pdf
[30] Yukio-Pegio Gunji, Shuji Shinohara and Vasileios Basios, Connecting the free energy principle with quantum cognition, https://www.frontiersin.org/articles/10.3389/fnbot.2022.910161/full
[31] Fields, C. and Glazebrook, J. F. A mosaic of Chu spaces and Channel Theory I: Category-theoretic concepts and tools. Journal of Experimental and Theoretical Artificial Intelligence 31: 177-213, 2019, https://chrisfieldsresearch.com/mosaic1-pre.pdf и Fields, C. and Glazebrook, J. F. A mosaic of Chu spaces and Channel Theory II: Applications to object identification and mereological complexity. Journal of Experimental and Theoretical Artificial Intelligence 31: 237-265, 2019, https://chrisfieldsresearch.com/mosaic2-pre.pdf
[32] Fields, C., Friston, K., Glazebrook, J. F. and Levin, M., A free energy principle for generic quantum systems. Progress in Biophysics and Molecular Biology 173: 36-59, 2022, https://chrisfieldsresearch.com/qFEP-2112.15242.pdf
[33] Hybrot, https://en.wikipedia.org/wiki/Hybrot
[34] Brett J. Kagan, Andy C. Kitchen, Nhi T. Tran, Forough Habibollahi, Moein Khajehnejad, Bradyn J. Parker, Anjali Bhat, Ben Rollo, Adeel Razi, Karl J. Friston, In vitro neurons learn and exhibit sentience when embodied in a simulated game-world, 2022, https://www.cell.com/neuron/fulltext/S0896-6273(22)00806-6
[35] https://www.constructortheory.org/
[36] David Deutsch, The Beginning of Infinity, 2011, https://www.thebeginningofinfinity.com/
[37] John Baez and Mike Stay, Physics, topology, logic and computation: a Rosetta Stone (2011), in "New Structures for Physics", ed. Bob Coecke, Lecture Notes in Physics vol. 813, Springer, Berlin, 2011, pp. 95—174, https://arxiv.org/abs/0903.0340
[38] John Baez, Symmetric Monoidal Categories: a Rosetta Stone, https://johncarlosbaez.wordpress.com/2021/05/28/symmetric-monoidal-categories-a-rosetta-stone/
[39] Constructive Mathematics, https://plato.stanford.edu/entries/mathematics-constructive/
[40] Mereology, https://plato.stanford.edu/entries/mereology/
[41] Kit Fine, Towards a Theory of Part, 2010, https://as.nyu.edu/content/dam/nyu-as/philosophy/documents/faculty-documents/fine/accessible_fine/Fine_Theory-Part.pdf
[42] Kit Fine, The Identity of Social Groups, 2020 https://metaphysicsjournal.com/articles/10.5334/met.45/
[43] Salvatore Florio, Core Constructional Ontology (CCO): a Constructional Theory of Parts, Sets, and Relations, 2021, https://gateway.newton.ac.uk/presentation/2021-04-22/29947
[44] Chris Partridge, 4-Dimensionalist Top Level Ontology, https://gateway.newton.ac.uk/presentation/2021-04-22/29946
[45] John F. Sowa and Arun K. Majumdar, Analogical Reasoning, 2003, http://www.jfsowa.com/pubs/analog.htm
[46] Entropy in thermodynamics and information theory, https://en.wikipedia.org/wiki/Entropy_in_thermodynamics_and_information_theory
[47] Vitaly Vanchurin, Yuri I. Wolf, Eugene V. Koonin, Mikhail I. Katsnelson, Thermodynamics of evolution and the origin of life, 2022, https://www.pnas.org/doi/full/10.1073/pnas.2120042119
[48] Vitaly Vanchurin, Yuri I. Wolf, Mikhail I. Katsnelson and Eugene V. Koonin, Toward a theory of evolution as multilevel learning, 2022, https://www.pnas.org/doi/10.1073/pnas.2120037119
[49] Geometrical frustration, https://en.wikipedia.org/wiki/Geometrical_frustration
[50] Yuri I. Wolf, Mikhail I. Katsnelson, and Eugene V. Koonin, Physical foundations of biological complexity, 2018, https://www.pnas.org/doi/10.1073/pnas.1807890115
[51] Irina Basieva, Andrei Khrennikov, Masanao Ozawabc, Quantum-like modeling in biology with open quantum systems and instruments, https://www.sciencedirect.com/science/article/pii/S0303264720301994
Best regards,
Anatoly Levenchuk
Dear All,
I completely have rewritten the text «Towards a Third-Generation Systems Ontology» with the help of Grammarly. Find it attached. I guess that it needs to be published by me as a preprint to arxiv.org and I will do it soon if there are no substantial comments. Hope this time it will be more understandable.
Also, I collecting comments here: https://ailev.livejournal.com/1657040.html (this is the post in my blog with English text), original in Russian you can find here: https://ailev.livejournal.com/1656653.html
Best regards,
Anatoly Levenchuk
From: Anatoly Levenchuk <ai...@asmp.msk.su>
Sent: Friday, November 4, 2022 11:17 AM
To: 'ontolo...@googlegroups.com' <ontolo...@googlegroups.com>
Subject: Towards the ontology of third generation systems approach
Dear all,
This is my best effort to write about systems ontology. References are not in very academic manner (e.g. references to some explanations in Wikipedia that is not too rigorous) but I tried to be understandable.
Towards the ontology of third generation systems approach
Anatoly Levenchuk, ai...@asmp.msk.su
… the old text are deprecated (it was completely rewritten)
Best regards,
Anatoly Levenchuk
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/000e01d8f3cf%245816d1e0%24084475a0%24%40asmp.msk.su.
Ravi,
> The subject areas covered are quite diverse
Yes, that's intentional. The main goal was to harmonize (i.e., show the way of ontology merge) several partial ontologies, which together can be considered an ontology of the systems approach. Special attention was to consider scaleless approaches - especially to include the time of evolution and techno-evolution (not just one complete life cycle of one system, as in the second generation systems approach) and size (micro-objects with quantum phenomena and macro-objects with gravitational phenomena, so-called "quantum gravitation:" theories). So we had to go through physics (including scale-free and living problems), mathematics and informatics, then mereology (to deal with parts and whole), then various features of the systems approach itself (functional descriptions of the target system, descriptions of the practices of the life cycle of systems of creation). Quite a few questions remain unanswered in the text, but they are already answered in cited works or works that are easy to find. For example, you can find answers about the potential issue of determinism of the proposed system ontology and the demonstration of unpredictability under determinism (this is important for quantum scales, and determinism should not be confused with predictability) in https://royalsocietypublishing.org/doi/10.1098/rspa.2015.0883
> example when you talk of scaleless systems, these are pertinent in particle physics as demonstrated by the work of Wilczek QCD and quarks nobel prize. But to take it to life, human and social levels is beyond my understanding.
Yes, this is a relatively new direction: scale-free physical descriptions that go beyond the description of inert matter. Based on the recent appearance of these works, we can talk about the new third generation of systems ontology. Two teams did the bulk of the work. In one team, quantum physicists Wanchurin and Katznelson, who worked with biologists Koonin and Wolf, gave scale-free descriptions of evolution in biology as a physical process. They showed how the provisions of the systems approach extend to living objects, including a physical explanation of the emergence of new system levels (levels of complexity: molecules -- cells -- multicellular organisms -- populations). Another team is the neurophysiologist Friston, who is interested, among other things, in problems of consciousness and the "selfhood" of agents. Quantum physicist Fields, mathematician Glazebrook, and evolutionist Levin have worked with him. These two teams somehow communicate. They rely on the same foundational ontological descriptions (mathematics and mathematical formulations of physical principles). Deutsch (inventor of quantum computers) also works scale-free. His constructor theory is just for that. Therefore his examples of constructors are primarily a catalyst molecule, a robot with a universal computer, and humans.
I'm glad to give to the ontolog-forum community literature about these new developments in systems ontology. Most papers were published in 2021-2022 and have yet to be known to ontology people. My contribution is in harmonizing mentioned research within a unified ontological description that clearly shows the physic-based multi-levelness of the evolving world. I called all these harmonized researches a new, third-generation systems ontology.
> I also did not find any references to our Ontology Summit topics of presentations there, of course we need to move deliberations to foundational ontology questions as well as to what many describe as ontology commitment!
My text mentioned 4D dimensionalism that appears in multiple presentations at Ontolog Summit (e.g., by Chris Partridge and Mathew West, David Leal, and even me) and VAE (VivoMind Analogy Engine) that John Sowa presented. You are right. A few presentations were on topics on foundational ontologies and systems ontology. Many thanks to Matthew West, who pointed me to the works of 4D ontologists, based on the writings of Kit Fine, I added them too. I hope my work will draw attention to these topics.
All the references you can find in my text (51 references).
Best regards,
Anatoly Levenchuk
.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAAN3-5diR8wvdUyBr4Ank%2BevE2iN6YW4MJJVTeKoC6j5BT76_g%40mail.gmail.com.
--
Azamat,
I know Bunge's works; I regularly referred to them in my posts ten years ago. But now I have stopped referring to them. Now I am more interested not in the speculative separation of some obviously different system levels but in the explanation of evolution: how atoms, then molecules, then cells, then multicellular organisms, then populations, and so on, suddenly appeared as complex physical objects. What laws of physics govern it, and what concepts are needed to explain it? Then I wonder how Darwinian evolution differs from techno-evolution, where the meme is detached from the organism, and its replication is independent of the fate of the target organism. Techno-evolution is not Darwinian because mutations are smart rather than random and are provided by universal computers in the constructor systems. All this is in the works I cite but not in Bunge's works.
I would object to turning my 12 pages into a literary review of thousands and thousands of papers that talk about all sorts of systemic levels that "just are" with all kinds of ontological commitments. Say, Deutsch's book mentioned in my text explain how physics dealing with atoms is different from chemistry -- the properties of molecules are irreducible to the properties of atoms, and chemistry thereby studies emergent properties. What matters to me is a description of how systemic levels emerged -- and not in the researcher's mind, but in nature. Thus, I am interested in system levels thinking of physicists, not thinking of philosophers (e.g., Dooyeweerd -- https://www.dooy.info/ext/st.html. Many, many hundreds of systems thinkers! But all of them are thinkers of previous generations of systems approach).
I have very few references to philosophers, although there is a reference to Kit Fine, for example (his papers of 2011 and 2020). But there are many references to the works of physicists, mathematicians, biologists, and engineers that deal with systems and know what will be helpful to them in modern systems ontology.
Best regards,
Anatoly Levenchuk
From: ontolo...@googlegroups.com <ontolo...@googlegroups.com> On Behalf Of Azamat Abdoullaev
Sent: Wednesday, November 9, 2022 7:17 PM
To: ontolo...@googlegroups.com
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAKK1bf9eSPBgRVJX5yqsKdM2M7XMLBK6-DV7yyJkBFwsseYNew%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/002401d8f48b%24045aca40%240d105ec0%24%40asmp.msk.su.
Ravi
What is specifically wrong with a suggested ontology? What ontological understanding missed? Researchers with a huge Hirsh index in papers cited in my text answered many questions about it, and I believe their answers.
I wrote a textbook on systems engineering based on this system ontology (it was published in Russian a couple of month ago). Here are slides from the presentation I gave a month ago about contemporary systems engineering -- https://www.slideshare.net/ailev/contemporary-systems-engineering-oct-2022
It is based on many ideas I explained in my systems ontology paper.
Best regards,
Anatoly Levenchuk
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAAN3-5chdktvJU0dDtZVjWvTo4dyrJs2V2duwTmWpbGNvTw%3D8Q%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/002f01d8f4a7%24d9f3f7f0%248ddbe7d0%24%40asmp.msk.su.
Ravi
> I went through 17 slides, these cover a lot, but not everything!
I would be amazed if, in 17 slides for a half-hour presentation, I could talk about systems engineering and the systems thinking behind it. I, of course, have more extended stories, but they are still in Russian. I have written practical systems thinking, methodology, and systems engineering textbooks. They have a couple of thousand pages and links to many books for additional reading. Here is a page with my textbooks on Amazon -- https://www.amazon.com/Левенчук/e/B07VGGXLJB (we are preparing English translations of all these books now). Now I writing a textbook about systems engineering management.
If there are any particular questions about systems engineering concerning systems ontology, I am willing to answer them. We have a course called "Ontologics and Communication," which comes before our Systems Thinking course (ontology crash course is part of it). Prapion Medvedeva developed it, and we made a draft translation into English and will develop it little by little: https://eem.institute/open-endedness/ontologics-and-communication/.
We can discuss how we educate engineers and managers with the help of ontology course, but it seems off topic here. I have a couple of paragraphs in my systems ontology text about it: ontology help keep attention to important objects in engineering and managerial projects.
Best regards,
Anatoly Levenchuk
.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAAN3-5c7vdKzh3a5jYRYFm7B%2BMZOpwbr9VYWpj0DEVn3J%2BFqNw%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/002401d8f48b%24045aca40%240d105ec0%24%40asmp.msk.su.
Azamat
I completely agree that a systems ontology cannot be the only and complete one. Like any other theory, it is constantly evolving, sometimes wholly changing its foundations as the foundations of related theories also change. For example, the foundations of physics with classical mechanics have been replaced by a couple of others -- quantum physics and relativity theory, and now there are attempts to replace them with "quantum gravity" theories. My references to the works of physicists are just within this new "quantum gravity" generation of physics. So the systems ontology of the first two generations of the systems approach was one (and there were many different schools of thought in those generations), and now it is changed and become a third-generation one. Then it will change again, and in the third generation, there will also be many different schools of systems and ontology thought. But evolution and techno-evolution (continuous development, continuous deployment, continuous delivery) will be in all of them.
As for the empirical (inference of theories from experimental data) versus theoretical (theoretical = conjectures/hypotheses step, then critical and experimental falsification step, then non-falsified theories should be taken seriously) approach, my text does not take a route of empiricism but choose rationalism (https://plato.stanford.edu/entries/rationalism-empiricism/). I have an explanation about contemporary rationalism (sorry, in Russian): https://ailev.livejournal.com/1619025.html. There I give the place of rationality in the overall intelligence stack (it is explained as "Epistemology = research+rationality, rationality = explanation+decision+action"). But other names are essential: epistemology of Karl Popper with David Deutsch's corrections, Judea Pearl, and the quantum-like (not quantum as in physics but with quantum-like mathematics applied to modeling) decision theory of Andrew Khrennikov and others. All this rationality vs. empiricism topic remains outside the realm of systems ontology. Still, in my paper, it is easy to get to the clarifications through the mentioned works on active inference (Friston, Fields, and others). Active inference supposes that the agent/creator must have generative models of the world and the self to hold the stability of the self in a changing world.
The remark about the need to formalize not only the system description but also the very notion of description is correct. I think I had a lot of work touched on this topic in my paper: category theory in mathematics as a formal constructive description, the need for the physicality of the executor of these descriptions (a universal computer in the sense of Turing machine equivalence -- electronic computer, quantum computer, mathematical brain or computer network). Then a remark about the works of Kit Fine and Core Constructional Ontology, which formalize descriptions of parts and wholes (I don't write "relations of parts and integers" on purpose because a static "relation" is denied and a constructor for them is given) for the convenience of describing systems. And I have remarked that the work of Kit Fine can be a move to descriptions-as-systems (where the interaction of parts of descriptions is done in a computer that emulates these descriptions in reality as objects of the abstract world: this is the harmonization of works of Fine and Deutsch).
You can find all of these references in my paper. You must read the original works (not my short paper) referred to by me in the paper.
Best regards,
Anatoly Levenchuk
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAKK1bf8hKFLBZCWvS5UxduXYkV0v198VASA7ztD23EmhCe489A%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/006301d8f514%243fff0be0%24bffd23a0%24%40asmp.msk.su.
Anatoly and Azamat: you are so close, in your views, that I see you both cooperating on a text. Ontologies captive to determinism will always rehash the premise. The living cannot be explained in ONLY cause and effect language. Never mind the reductionist burden.
As usually, I stay away from arguing with my distinguished colleagues. But I do not hesitate when it comes to suggesting commonalities and what they can afford. For me the text subject to conversation does not justify my effort to point to errors (factual, logical).
Best wishes.
Mihai Nadin
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAKK1bf8XVR6O2NzLssY43pudW%2BrvPPga3_GWwzCSBOk2XqpzhQ%40mail.gmail.com.
Azamat,
We can discuss whether "causes" and "effects" are real (whether they exist in the physical world or whether these are only classes as a way of describing the world). I don't need to point out that all the theories I've discussed are formulated counterfactually and thus take causality into account. My paper mentioned Deutsch's work on causality; in my last reply, I also referred to the work of Judea Pearl on causality. In Prapion Medvedeva's course "Ontology and Communication", there is a part that deals with causality based on the work of Deutsch and Pearl.
In a short paper on systems ontology, I cannot raise every possible question of epistemology and make every possible reservation. I am not inclined to quote a complete philosophical encyclopedia here and demonstrate the epistemic choice in every possible question. I have a list of 51 papers that are mostly compatible with each other in the epistemic choices made in those papers. Rationalism and not empiricism, Deutsch's corrections to Popper's ideas about explanations as the only valid theories (explanations are about causality), Pearl's ideas about causal inference, the causal ladder and causal graphs, Gunji's ideas about quantum-likeness for excess Bayesian inference, and so on -- the literature list for my paper discusses all these choices quite a bit. No need to defend all of these epistemic choices in my paper.
I intentionally leave all this out and focus not on a "theory of everything" (not writing another all-encompassing ontology of everything). I purposely limit my scope to the ontology of the systems approach. My paper is about parts of some system that comprise a whole that is part of some even more complex whole. Then there can be different ways of dividing into parts, with some systems creating and developing other systems (I need techno-evolution, I need three times, not just a system in its environment as in the first-generation of systems approach).
We can discuss everything else in other papers if needed. For example, artificial intelligence problems are not considered in the system ontology. It is enough that the constructor system can include in it a universal computer (including the brain of a mathematician or ontologist, a computer device based on a variety of physical principles, a collective of people and computers on which a variety of AI programs are executed, etc.). We can discuss this work further in other papers (and we certainly will). For a system ontology, it suffices to point out that one constructor system (suggested by Deutsch explicitly and by Fine implicitly) can somehow create another system. And I point out in my paper that the computer implements embodied cognition.
You can easily find many other exciting ideas in the literature in my paper. Still, I want to focus on just system ontology, not ontology in general, not philosophy in general.
Best regards,
Anatoly Levenchuk
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAKK1bf8XVR6O2NzLssY43pudW%2BrvPPga3_GWwzCSBOk2XqpzhQ%40mail.gmail.com.
AL: I intentionally leave all this out and focus not on a "theory of everything" (not writing another all-encompassing ontology of everything). I purposely limit my scope to the ontology of the systems approach. My paper is about parts of some system that comprise a whole that is part of some even more complex whole. Then there can be different ways of dividing into parts, with some systems creating and developing other systems (I need techno-evolution, I need three times, not just a system in its environment as in the first-generation of systems approach).
.....................
You can easily find many other exciting ideas in the literature in my paper. Still, I want to focus on just system ontology, not ontology in general, not philosophy in general.
I read them in Russian. Indeed, they are the ideas worthy of time.
My concern is about how the greatest science of ontology as the theory of reality is massively misinterpreted or downgraded provided that science is after universal knowledge.
Real ontology is not some highly metaphysical "theory of everything", some top-abstract study of what exists, but what was/is/will be real.
Real ontology is about real-world things and relationships modelled in terms of the real-world categories formalized by the real-world variables and measured as the real-world data.
All life-significant quantities and qualities or data are causal variables, physical variables, environmental variables, chemical variables, biological variables, medical variables, economic variables...
Reality is the sum total of real-world entities, processes and relations, causes and effects. Or, the world is essentially ontological and causal or scientific, neither mathematical, nor logical nor statistical...
Again, the systems ontology is about the world of ontological systems and causative networks, neither formal systems, nor logical constructions from sets, types and relations, like in the core ontology construction project.
This is the reality of ontology, which could be avoided only at the prejudice of the truth.
Any intelligent entities, be it human researchers or AI machines, sense and affect real world things and processes to realize real-world systems, obeying constraints/rules/laws inherited from the real world.
Typically, important characteristics/features of real-world entities and relationships and their human-mind or machine-world representations are badly specified explicitly in thoughts or code, and important opportunities for detecting errors due to mismatches are lost.
Thus we have as many views, prospects, guesses and conjectures, hypotheses and theories as many as its holders, being lost in the chaos of academic noise and fake articles, thus creating antiscience in mass scale.
Bottom line.
Real-World Ontology (RWO) with its system ontology and its ontological/causal reasoning is crucial to how humans or machines understand, explain, and make decisions and interact with the world.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/000001d8f552%2474b015d0%245e104170%24%40asmp.msk.su.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/BL3PR01MB6897A5F3AC14FBE31FA18DB8DA019%40BL3PR01MB6897.prod.exchangelabs.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/BL3PR01MB6897A5F3AC14FBE31FA18DB8DA019%40BL3PR01MB6897.prod.exchangelabs.com.
Mihai,
I am pleased that I made system ontology deterministic because quantum physics had (and in some interpretations of quantum mechanics still has) big problems with determinism. That said, I will repeat: one should not confuse the determinism of the world with its predictability, and I will reiterate the link to the text where this is discussed (this is not in scope of systems ontology, this is one of the endless epistemological choices, so the link is not in my article, but I have given it here in the discussion): https://royalsocietypublishing.org/doi/10.1098/rspa.2015.0883
So systems ontology in the version described in my article describes the deterministic world, but fundamentally unpredictable due to laws of quantum physics in their scale-free version, i.e., applicable to macro-objects and living physical objects too. While we call both "non-determinism" and "unpredictability" the same word "non-determinism", it is a waste of time to discuss further.
Best regards,
Anatoly Levenchuk
.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/BL3PR01MB6897A5F3AC14FBE31FA18DB8DA019%40BL3PR01MB6897.prod.exchangelabs.com.
Causal is way more than the limited causality described in determinism.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAKK1bf8x0hUePSVoLctPHxyK1ro%3DReBrAC%2B%2BaSn%2B3GJCmort7g%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAAN3-5cgHLYC8_JPz_gSu53p0NN9QqpUcmHCmurZ2i7Y-ML3pQ%40mail.gmail.com.
Azamat,
Azamat,
I agree that a systems ontology should be based on many clearly stated epistemological choices. The essence of my work was precisely to select many research projects that would clarify epistemological choices (attitude to reality, attitude to causality and counterfactuality of explanations, attitude to determinism, attitude to empiricism and rationality, attitude to formality level, attitude to modeling behavior of real objects by ideal objects, attitude to agency and their rationality, and so on to the end of a complete encyclopedia of philosophy). I have cited literature (51 references) that goes into detail about all of this, hundreds of pages about this epistemic stuff.
My short (12 pages including references) text concerns the serious formulation of systems approach ontology. There is a short set of notions/concepts comprising the systems approach, which ontologists (e.g., me) should express somehow after accepting all these important epistemological choices already made by researchers in references in my paper. Systems ontology should be formulated in such a way that there are no notable contradictions (this was the primary goal of my paper).
Say, if we talk about "the constructor of parts into a whole, the constructor of membership in a set," then we should then avoid saying "the part-integer relation," for it is not constructive (until something else is shown). It is constructive ontology, with no relations between objects but constructor operations! If we say "reality", then we cannot say that some objects behave according to rules -- no, rules describe reality. They are derived as conjectures and tested by criticism, first logical and then experimental. And one must add that these "rules" are not just a model but a generative model. Otherwise, it won't be easy to talk about "explanations" and make counterfactual statements about possible worlds (however, possible worlds must also be explained, where they came from and what to do with them). It is necessary to explain the evolution of systems, not just "processes" in operation time or even life cycle. It is necessary to explain how exactly instances of systems are obtained in the course of systems evolution, in which many instances will take part.
There is no escaping from this. The authors of the papers I have mentioned in the literature to my paper have done great epistemological work. Thank them very much. We do not need to duplicate their work. We can stand on the shoulders of these giants, which is what I have tried to do: take some parts of their work and show that they constitute a whole systems ontology.
When we teach engineers (it's written in my text), we first teach them just the general view of the world without the concept of a system. Here is the syllabus for the ontology and communication course (only chapter names):
Chapter 1. Words and their meanings
Chapter 2. Complex descriptions and meta-languages
Chapter 3. Explanations Development
Chapter 4. Descriptions that address the requirements
Chapter 5. Goal-oriented communication
Chapter 6. Communication failures
Chapter 7. Relationships between objects and ontological rattle
Chapter 8. Causality and modeling
Chapter 9. Working with data and updating models
Chapter 10. Contexts and models of the addressee
Chapter 11. Descriptions by implicit request
Chapter 12. The Big picture and what's next
This course answers quite a few questions, but we need to go into more detail about a system. The concepts of a system approach (i.e., systems ontology) are covered in other courses -- but with many epistemic questions already answered. Here is a one-level deep table of contents for the current systems thinking course without part about system lifecycle:
1. About thinking.
Before doing systems thinking
The peculiarities of presenting the training material
Different kinds of thinking
The requirements for thinking
Is systems thinking a substitute for applied thinking?
The place of systems thinking among other thinking: intellect-stack
Readiness for (thinking) action
Variants of systems thinking
Systems Engineering
Systemic and systematic
Our variant of the systems approach
Basic concepts of the systems approach
Terminology
Words (terms) are important and unimportant.
Definitions: a coffin for a dead thought
How we chose the terms for our book
The formality of systems thinking
Systemic creativity.
The subject specializations of systems thinking
Can thinking be taught?
Metanoia
Stages of teaching thinking
Features of systems thinking instructional problem solving
Transition to using thinking
The temptation of complexity
How long does it take to master systems thinking?
Applicability/applicability of systems thinking
2. Embodiment and description of the system
Embodiment, description, and documentation of the system
Descriptions
How to negotiate: don't generalize but specify
Relationship of composition
Openings
Works and Actions.
Don't overuse process-as-systems
Computer programs
3. Roles
Type Machine.
Role objects/roles and play by role/function/assignment in the environment
Aristotelian physics in systems thinking
Physical and functional objects
The second generation of systems thinking
Methodology: agents, intentions, strategies, plans, roles, objects of interest, preferences
Methodological modeling.
Roles of analysts and engineers. Positions
External and internal project roles
Cultural conditioning of project roles
Theatrical metaphor.
Thinking about people: they are first and foremost performers of roles
Subjects of interest and interests/important characteristics and preferences
Don't glue together an important characteristic and preference
Methods of description from CPS PWG Cyber-Physical Systems (CPS) Framework
Talk not to people. Talk to roles
Position
Leadership
External and internal roles
Organizational Places, Organizations, Organizations
Bosses: they are jokers. Expect them to play any role
Rank and skill level
How many project roles in total
Mistakes in defining roles
4. System levels.
Not all systems what they are called
Systemic partitioning.
Emergence and meta-system transition
Reductionism
Synergy: it's not a systemic effect. It's not emergent
Purposeful system and collective systems thinking
How to depict system levels, how to name the systems that comprise them
System levels in systems engineering. An example of computational engineering
4D systemicity: patterns in space, time, shapes, and rhythms
Systemic levels are highlighted by attention. Example of social dance.
Intelligence stack: system levels of intelligence
Constructor systems
Our system.
Recursive application of systems thinking: recursive control of attention
A diagram of the essence of systems thinking
Needs, requirements, limitations
Examples of systems terminology
Division of labor and system levels
Systems of systems
People in systems
State building and state projects
The future and leading discipline of entrepreneurship
Thinking alike for simple and complex systems
Complexity and measures of complexity
5. The target system and its supersystem
Find the target system first
The target system depends on who is looking for it
Systems-products and services systems creation
Example of a barbershop as a service provider
Service orientation. The world of providers
Examples of services and their providers
You are a team member
Signs of a target system
The letter carrier principle
Common mistakes in target system identification
Naming the system
A supersystem: you have to find it too!
Systems approach: for all types of systems, not just the target system
6. How to describe systems
The transdisciplinarity of systems thinking
The four basic types of system decomposition
Functional analysis and modular synthesis
One system, but many descriptions, many names: that's okay!
Alternatives to the main types of system partitioning
Non-overlapping functional and structural partitioning of the system
Architecture creation: functional analysis and modular synthesis
Alphas and artifacts/products
Apples of life, apples of task
Alphas
System description
Project roles and system descriptions
Sub-Alphas
Distinguishing the subject of interest from the method of description
7. System modeling
Descriptions and description methods, models and meta-models
Multi-model and transdisciplinarity
Description method and mega-model
The notion of configuration
Functional descriptions: circuit diagrams and scenarios
Modular/product/constructive descriptions
Platforms and technology stacks
Functionality and mastery of the subject area
Organizations
The need for good modularity
Struggling with complexity in thinking
Enterprise engineering, non-engineering
8. Requirements and architecture
Requirements as a sub-phase of system description
Two understandings of the term "requirements"
Requirements and system partitioning
Target-oriented requirements engineering
Verification and acceptance
Notion of architecture
After eight chapters about the notion of a system, we will define the life cycle in terms of practices of the constructor system (this is the second time to consider: constructor time, first time was operations time. We have a "Мethodology" course dedicated to this. I will give you its content. There are also many important ontological and epistemic issues:
1. Methodology as a fundamental discipline
Methodology as a fundamental discipline.
The notion of the method
Why study methodology
Methodology and systems thinking
2. The non-life nor-cycle
Biological lifecycle
System lifecycle 1.0: works that change the states of the system-of-interest
Work execution by organizational units
Life cycle as a work (Life Cycle 1.0)
Project Lifecycle
Lifecycle 1.0 issues
Practices
Lifecycle 2.0
Operation as a dedicated life cycle stage
Creation Chains
Three lifecycle times
Notion of practices
Life cycle management and other management practices
The discipline of practice
Practice support technology
Improvement and development
Life Cycle Practices
Example: systems engineering lifecycle practices
Methodologies
3. Type of lifecycle
V-diagram
Model-centricity in the lifecycle
V-models as a system decomposition model
Hybrid lifecycle models
Moving away from "waterfalls" with gates
More about job management and lifecycle management
Types of work management practices
Agile lifecycle management methodologies and case management for work management
Trends in work management/operations management practices
Beyond the life cycle
Life cycle as an activity architecture
4. Project system diagram
Project system diagram
Modern notion of a project
Areas of interest of suprasystem, system-of-interest, constructor
Alphas - a general object to track in an organization/collective/team/cooperative project
Alpha of commercial opportunity
Alpha of external project roles
Alpha of target system description
Alpha of system embodiment
Alpha of operations (operations/service/operations) of the target system
Alpha of organization description
Alpha of project organization
Alpha of the project organization
What to keep track of in the project
Alpha states and artifacts/work products
How to work with the project system diagram
Alphas
Adapting a project system diagram to an enterprise subject area: example of a training course project
V-diagram and project system diagram
Digital twins: more subalphas as objects to track
Basic lifecycle diagram
Maturity models and technology readiness models
5. Practice as a first class object
The ontological status of practice
Practice as an object of the first class
Decomposition of practices: the example of making coffee
Communities of practice
The primary technique for defining practice
Evolution of practices
Systemic levels of the target system and choice of practices
What's next
We can already talk about systems engineering on this basis (at this point, students already understand systems thinking). Here is the content of "Systems Engineering":
1. Scaleless continuous systems engineering
Engineering
Engineering and evolution.
Ethical/political problems of natural evolution that can be solved by engineering
Specialization/concretization of systems engineering practices
Systems engineering's responsibility for the integrity of systems.
Transdisciplinarity of systems engineering
The role of the (systems) engineer.
Systems engineer as a technical leader
Engineering and science
Systems engineering methodology as an engineering science
The non-science of engineering. Heuristics
Research/science as "teaching the birds to fly"
Engineering Science.
Trends in modern systems engineering
2. Continuous development.
What do developers do?
The death of requirements engineering.
Who develops the concept of operations and the system concept?
Continuous refinement of the concepts
Design-neutral expression of system function: use cases
Case management and use cases
Invention: generating ideas for the concept
Decision-making/trade-off studies: going through the forks
Conceptual and architectural design
How to misunderstand the first subsection about "no requirements now"
Systems management and design
Error-free design, precision manufacturing
3. Continuous architectural decision making
Architecture: was "about hard to change", became "about facilitating change"
Architectural Characteristics
Architectural solutions
Executing of architectural trade-offs
Architectural styles
4. Continuous commissioning (DevOps)
DevOps, SREs, lifecycle management, and "putting the project in order"
DevOps vs. work management
Configuration management is needed for all systems in the project
Checklists in lifecycle management
6-, Multi-, poly-, ... D
Integration of lifecycle data
Digital Twin
Lifecycle management alphas, configuration and change management, engineering documentation, lifecycle data, digital engineering
Testing, verification and acceptance, assurance
What is justification in engineering
The rationale for assurance
The rationale for meeting standards
Engineering assurance is actionable. It takes resources
Testing: it is measurements.
DevOps and SRE as it relates to engineering assurance
Engineering rationale issues at the substantive, personal, community, and societal levels
Alpha of engineering reasoning: emphasis on features/incremental
After this will be «Systems management» course, but it is already too much for this thread )))
All this courses are already exist and working (yet in Russian, but we are actively working on translation project).
Best regards,
Anatoly Levenchuk
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAKK1bf90KDYe57%3Dn58j%2BPbOmd86WzOKGhsRcBZgdR0dNg2pBnw%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/002a01d8f62e%24005b46a0%240111d3e0%24%40asmp.msk.su.
Azamat,
Thanks for the prompt. I just had three posts in Russian on the topic of how to teach xGI. Let us call it xGI, where x is N (natural), A (artificial), C (collective), E (embodied), or everything else. GI is general intelligence, which we define after Chollet https://arxiv.org/abs/1911.01547.
My last post on teaching xGI the same way as we teach humans was https://ailev.livejournal.com/1600861.html. To speed up calculations of xGI we need it to calculate only with the account of the essential things. Science gives us a set of ontologies with objects and operations that serve as objects of attention for xGI preventing it of random calculations about any possible objects. It is called innate priors if embedded in hardware, but most often, it is called inductive bias if we speak about neural networks and software-defined bias (e.g., an inductive bias that is a result of the pretraining of foundation models). For people, such focusing on essential things inductive biases they got in schools (to be not Mowgli but educated humans). I suggest the usage of schools for xGI, not separating human GI and artificial GI. The same schools for any general intelligence!
Therefore suggested curriculum on ontologics and communications, systems thinking, methodology, and systems engineering, which I gave in my previous letter here, will be helpful for humans and AGI. Sure, the usual curriculum of the usual universities will not be as helpful as ours, you are right ;-)
Best regards,
Anatoly Levenchuk
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAKK1bf_hLV6M8EwMCLcWZ2rHtxkWoH832m6HG0ewaSgiOGLJuw%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/000001d8f93f%24968516b0%24c38f4410%24%40asmp.msk.su.