Conceptual Modeling and Classification of Events

Basic Information

• Paper ID: 2501.00276
• Title: Conceptual Modeling and Classification of Events
• Author: Sabah Al-Fedaghi (Kuwait University)
• Classification: cs.SE (Software Engineering)
• Publication Date: 2025
• Paper Link: https://arxiv.org/abs/2501.00276

Abstract

This paper is a continuation of an ongoing research project on the graphical methodology "Thinging Machine" (TM). TM was originally proposed as a foundation for conceptual modeling (such as conceptual UML) and applied to fields including requirements engineering. Conceptual modeling involves high-level representations of real-world systems, integrating various components and refining them into more concrete (computer) executable forms. The TM project has evolved into a more comprehensive approach through applications across multiple research domains and expansion of its theoretical and ontological foundations. The first part of the paper addresses enhancements to certain characteristics of TM in structuring existential events, such as missing events. The second part focuses on how to classify events and the types of relationships that can be identified between events. The concept of events occupies a central position in modeling, influencing computer science as well as multiple disciplines including linguistics, probability theory, artificial intelligence, physics, philosophy, and history. In TM, events are defined as "thimacs" (thing/machine) with temporal characteristics, which inject dynamism into static thimac descriptions called "regions." Regions are graphical specifications based on five universal actions: create, process, release, transfer, and receive.

Research Background and Motivation

Problem Background

1. Challenges in Conceptual Modeling: Existing conceptual modeling approaches encounter difficulties in handling entity types, generalization hierarchies, relationship types, attributes, and cardinality
2. Complexity of Event Classification: The concept of events occupies important positions across multiple disciplines, yet lacks a unified classification framework
3. Fragmentation of Static and Dynamic Modeling: Traditional approaches struggle to effectively handle the relationship between static structures and dynamic behaviors

Research Motivation

• Develop more comprehensive conceptual modeling methodologies, particularly in handling various forms of existence
• Propose a novel event classification method based on graphical structures rather than semantic classification
• Enhance the expressive capacity of the TM model for complex real-world systems

Core Contributions

1. Enhanced the Theoretical Foundation of TM: Introduced a dual-layer structure of potentiality/actuality, adopting Stoic philosophical concepts of reality
2. Proposed the Concept of Missing Events: Provided a new theoretical framework for modeling negative propositions
3. Established a Graph-Based Event Classification Method: Proposed replacing traditional semantic classification with graphical classification based on TM model diagrams
4. Enriched Conceptual Modeling Methods: Particularly in handling the diversity of existence, such as missing events corresponding to negative propositions

Methodology Details

Task Definition

This research aims to:

• Establish a unified conceptual modeling framework capable of handling static regions and dynamic events
• Provide formal representations for various types of events, including missing events
• Develop an event classification system based on graphical structures

TM Model Architecture

Core Concepts

Thimac (Thing/Machine): The fundamental unit of the TM model, possessing dual nature of both thing and machine:

• As a thing: can be created, processed, released, transferred, and received
• As a machine: capable of creating, processing, releasing, transferring, and receiving other things

Five Basic Actions

1. Create: Reveal a thing within a machine, i.e., register a thimac as an ontological unit
2. Process: A thing is changed, processed, and examined, but does not produce new thimacs
3. Release: A thing is prepared for transfer outside the machine
4. Transfer: A thing crosses boundaries as input or output of the machine
5. Receive: A thing arrives at and enters the machine

Dual-Layer Structure

• Static Layer (Potentiality): Regions - describing possible event configurations
• Dynamic Layer (Actuality): Events - the combination of regions with time

Technical Innovations

1. Relationship Between Regions and Events

• Regions and events are inseparable but distinct in reality
• Regions are static components of events; events are the realization of regions in time
• This relationship resembles Platonic form theory but does not require the concept of the soul encountering forms

2. Theory of Missing Events

• Missing Events: Events that model negative propositions
• Example: "John is not in the room" represents an existing event where John is a missing sub-event
• Missing events exist but their regions cannot execute actions

3. Graphical Representation

• Rectangles represent thimacs
• Arrows represent action flows
• Cylinders represent storage
• Dashed arrows represent triggers

Experimental Setup

Case Studies

The paper validates the effectiveness of the TM model through multiple cases:

Case 1: Propositional Logic Modeling

Modeling of complex logical expressions (a ∨ (¬a ∧ b) ∨ c) → d, where:

• a: Computer is within the local network
• b: Computer has valid login ID
• c: Computer is under administrator use
• d: Computer can access the internet

Case 2: Event Classification Verification

TM modeling of four event types from Vendler's scheme:

• Activities: such as "running," "walking"
• Accomplishments: such as "running a mile," "painting a picture"
• Achievements: such as "John kissed Mary"
• States: such as "on a ship," "loving"

Modeling Methods

1. Static Models: Showing structural relationships between thimacs
2. Dynamic Models: Displaying temporal execution of events
3. Event Chronologies: Describing temporal ordering of events
4. Temporal Diagrams: Specifying extensions for event durations

Experimental Results

Main Findings

1. Unified Modeling Capability

The TM model successfully provides a unified representation framework for various types of events, including:

• Traditional positive events
• Missing events corresponding to negative propositions
• Complex composite events

2. Feasibility of Graphical Classification

Through modeling verification of multiple event classification schemes, the feasibility of graph-structure-based classification methods was demonstrated:

• Structural differences between activities and accomplishments are primarily reflected in specifications of implicit termination conditions
• Different event types can be distinguished through topological features of graphs

3. Effective Modeling of Missing Events

Successfully modeled various negative expressions, such as:

• "Fake gold" as a metal thing event lacking the "gold" attribute
• "John is not in the room" as a missing sub-event of John in a room event

Case Analysis

Propositional Logic Case

Through modeling of network access control logic, demonstrating how TM handles:

• Conditional branching logic
• Bidirectional communication flows
• Complex control structures

This case contains 15 marked points, completely describing the bidirectional verification process from computer to internet and from internet to computer.

Event Classification Case

Modeling of "Terry builds a house" demonstrates how TM distinguishes:

• Initial concept (E1): Terry's existence
• Project initiation (E2): House's existence as an incomplete project
• Building process (E3-E5): Continuous building and completion checking
• Final state (E6): Completed house

Related Work

Event Ontology

• Whitehead: "Everything is an event; the world is composed of events"
• Aristotle: Distinguishing states and events, endpoints and continuous processes of events
• Vendler's Scheme: Classification framework for four verb types

Conceptual Modeling Methods

• UML: Limitations as a conceptual modeling language
• Entity-Relationship Model: Insufficiencies in handling dynamic behaviors
• Temporal Logic: Development of specification languages such as TimeML

Uniqueness of TM Model

Compared to existing approaches, the advantages of the TM model lie in:

• Unified thing/machine duality
• Clear separation of static/dynamic
• Formal treatment of negative events

Conclusions and Discussion

Main Conclusions

1. Theoretical Contribution: The TM model provides a richer theoretical foundation for conceptual modeling, particularly in handling the diversity of existence
2. Methodological Innovation: Proposes a novel event classification method based on graphical structures rather than semantic classification
3. Practical Value: Validates the effectiveness of the TM model in complex system modeling through multiple cases

Limitations

1. Complexity: TM diagrams may become extremely complex, affecting readability in practical applications
2. Tool Support: Lacks mature tool support for automatic generation and verification of TM models
3. Empirical Validation: Requires validation from more large-scale real-world projects

Future Directions

1. Graph-Theoretic Foundation: Develop structured classification systems based on graph theory (such as line graphs, reflexive graphs)
2. Tool Development: Develop automated tools supporting TM modeling
3. Application Extension: Apply TM methods to more domains, such as artificial intelligence and data science

In-Depth Evaluation

Strengths

1. Theoretical Depth: The paper provides profound philosophical and ontological foundations, particularly in handling existence and absence
2. Methodological Innovation: The dual nature (thing/machine) and dual-layer structure (static/dynamic) of the TM model provide new modeling perspectives
3. Rich Examples: Multiple detailed cases demonstrate the applicability and effectiveness of the method
4. Interdisciplinary Integration: Successfully integrates concepts from philosophy, logic, and computer science

Weaknesses

1. Representational Complexity: TM diagrams may become difficult to understand and maintain when handling complex systems
2. Lack of Quantitative Evaluation: The paper primarily validates the method through qualitative analysis, lacking quantitative performance assessments
3. Engineering Practice: The transformation path from theory to practical software engineering practice is unclear
4. Comparative Analysis: Insufficient systematic comparative analysis with existing conceptual modeling methods

Impact

1. Academic Value: Provides new theoretical perspectives for the conceptual modeling field, particularly in event modeling
2. Practical Potential: Has application potential in requirements engineering, system analysis, and other fields
3. Methodological Contribution: Graph-based classification methods may inspire development in other modeling domains

Applicable Scenarios

1. Complex System Modeling: Particularly suitable for systems requiring handling of static structures and dynamic behaviors
2. Requirements Engineering: Provides more precise representation methods in requirements analysis and specification phases
3. Event-Driven Architecture: Provides theoretical foundation for designing event-driven systems
4. Knowledge Representation: Has application value in knowledge representation for artificial intelligence

References

The paper cites 27 important references, covering:

• Conceptual modeling theory (Rosenthal et al., 2023)
• Event ontology (Casati & Varzi, 2014; Whitehead, 1964)
• Temporal logic (Pustejovsky et al., 2022)
• Philosophical foundations (Research related to Heidegger, Aristotle)
• Author's prior work (Al-Fedaghi series of papers)

This paper provides an innovative theoretical framework and methodology for the conceptual modeling field. While further development is needed in engineering practice, its theoretical contributions and interdisciplinary integration characteristics give it significant academic value and application potential.