Abstraction is a thinking process that extracts characteristics, functions, structures, relationships, and other elements relevant to a particular purpose from concrete objects or phenomena and represents them as concepts or models that are not dependent on specific instances. It is used in concept formation, problem-solving, scientific research, technological development, system design, analogical reasoning, and many other forms of intellectual activity.
In everyday life, we recognize similarities among individual experiences and understand them through general concepts. For example, we can recognize objects of different shapes and designs as “chairs.” We can also classify chairs and tables under the broader category of “furniture.”
However, Abstraction is not limited to classifying concrete objects into higher-level categories.
For example, a vibration sensor used in a factory can be understood through the broader category of “measurement equipment.” It can also be described functionally as “a device that detects changes in physical conditions.”
Furthermore, if we abstract its structure as “acquiring signals representing changes in an object’s state and detecting deviations from normal conditions,” we can explore similarities with technologies and phenomena that are quite different from vibration sensors.
Abstraction thus allows us to understand familiar objects from different perspectives and identify possibilities for comparison and application that would otherwise remain unnoticed.
An important role of Abstraction is to move beyond the characteristics of individual objects and discover features and relationships that may also apply to other objects or phenomena.
However, Abstraction necessarily involves leaving out some information. If essential conditions are omitted, the resulting generalizations or analogies may not hold in reality.
The critical challenge is therefore not simply reducing information, but determining what should be preserved.
Abstraction and Human Cognition
Humans do not perceive every object they encounter as something entirely new.
For example, even when we see an unfamiliar design of chair, we may recognize it as a chair by understanding its structure and function as something intended for sitting.
This recognition involves identifying common characteristics among different objects.
Philosophers have long examined the relationship between individual objects and general concepts. Aristotle’s classification theory, philosophical discussions of universals, and questions concerning concept formation in modern philosophy are all related to Abstraction.
Cognitive psychology has also investigated how humans form categories and recognize different objects as members of the same class.
In their 1976 paper “Basic Objects in Natural Categories,” Eleanor Rosch and colleagues examined the hierarchical organization of categories and its relationship to human cognition.
For example, “furniture,” “chair,” and “rocking chair” represent different levels of abstraction.
The Basic Level refers to a level of categorization that tends to provide a useful balance between generality and information content in everyday human cognition.
However, organizing objects into hierarchical categories represents only one aspect of Abstraction.
In technological development and scientific research, it is often more important to extract functions, structures, operating principles, and causal relationships than merely to identify a higher-level category.
Different Directions of Abstraction
Abstraction is often described as moving from concrete concepts toward more general concepts.
However, different kinds of concepts and models can be produced depending on what is abstracted.
Consider a vibration sensor used in a factory.
| Direction of Abstraction | Abstracted Representation |
|---|---|
| Category | Vibration sensor → Sensor → Measurement equipment |
| Attributes | High sensitivity, low power consumption, compactness, non-contact operation, etc. |
| Function | Detect changes in the state of an object |
| Operating principle | Convert physical changes into measurable signals |
| Structure | Measurement target → Sensing element → Signal processing → Judgment |
| Relationship | Changes in the object’s state are observed as changes in signals |
| Pattern | Detect abnormalities by identifying deviations from normal conditions |
These representations cannot simply be arranged along a single scale from low to high abstraction.
For example, “measurement equipment” is a broader category that includes vibration sensors, but it conveys relatively little information about anomaly detection.
In contrast, the pattern “detecting deviations from normal conditions” may apply to technologies other than vibration sensors.
It may also extend beyond physical measuring instruments to biological sensory mechanisms and software monitoring systems.
Abstraction therefore involves more than moving upward through a hierarchy of categories.
It can also involve changing the perspective from which an object or phenomenon is understood.
Category Abstraction and Generalization
One basic form of Abstraction is grouping objects according to shared attributes.
For example, apples, oranges, and bananas can be classified under the general category of “fruit.”
This involves setting aside differences among individual objects and recognizing a shared category.
Such operations are closely related to Generalization.
However, Abstraction and Generalization are not completely identical.
Generalization involves identifying properties shared by individual cases and treating them as potentially applicable to a broader class of objects or situations.
Abstraction, on the other hand, does not necessarily aim to extend the number of objects to which a concept applies.
For example, when modeling a manufacturing process, we may omit equipment shapes and physical layouts while retaining only material flows and processing sequences.
This does not necessarily mean that more manufacturing processes are being included.
Rather, information relevant to a particular analytical objective is selected while other details are temporarily disregarded.
Abstraction therefore includes not only the generalization of categories but also the selective representation of important characteristics and relationships.
Differences Between Abstraction, Simplification, and Idealization
Several related concepts should be distinguished from Abstraction, particularly Simplification and Idealization.
Simplification reduces complexity to make an object or problem easier to understand or process.
Abstraction selects characteristics relevant to a particular purpose and disregards others.
Idealization may introduce assumptions that do not strictly hold in reality to construct models that are easier to analyze.
For example, physics frequently uses models involving frictionless surfaces or point masses with no physical dimensions.
A perfectly frictionless surface does not exist in ordinary reality, but such idealizations make it easier to analyze fundamental relationships governing motion.
| Operation | Main Purpose | Example |
|---|---|---|
| Abstraction | Extract characteristics or relationships relevant to an objective | Extract material flows and process relationships from manufacturing equipment |
| Generalization | Move from individual cases toward broader concepts or rules | Define a maintenance approach common to several types of equipment |
| Simplification | Reduce complexity to facilitate understanding or processing | Create a process diagram that omits operational details |
| Idealization | Introduce assumptions that make analysis easier | Model a system without friction or heat loss |
| Formalization | Express concepts and relationships using explicit symbols or rules | Represent process-state transitions with equations or graphs |
These operations are not mutually exclusive.
Scientific models may combine Abstraction, Simplification, Idealization, and Formalization.
However, if the conditions omitted through Abstraction or assumed through Idealization are not made explicit, applying a model to real situations may lead to incorrect conclusions.
Functional Abstraction
Functional Abstraction is particularly important in technological development and new product development.
Products and devices have specific characteristics such as shapes, materials, structures, and manufacturing methods.
However, different technologies may perform the same function.
For example, the function of transporting liquids can be achieved using pumps, gravity, capillary action, pressure differences, and other mechanisms.
These technologies differ in their configurations and operating principles.
Nevertheless, when we focus on the function “moving a liquid from one location to another,” they can be compared as alternative solutions to the same problem.
Furthermore, if we abstract the function to “transporting materials to required locations,” the search may extend to mechanisms involving gases, powders, or particles.
However, broadening the abstraction also introduces many technically unsuitable alternatives.
Properties of the material being transported, required flow rates, transportation distance, response time, and energy consumption may strongly influence which mechanisms are feasible.
Functional Abstraction therefore requires distinguishing between the objective to be achieved and the conditions under which it must be achieved.
Abstraction of Operating Principles
Functional Abstraction makes it possible to search for different mechanisms that achieve the same objective.
Abstraction of operating principles instead focuses on the mechanisms through which particular effects are produced.
Consider the function of removing contaminants from a material surface.
This function may be achieved through mechanical friction, dissolution using solvents, chemical reactions, ultrasonic effects, and other mechanisms.
The function is common, but the operating principles are different.
Conversely, the same operating principle may be applied to different purposes.
For example, ultrasonic phenomena can be used in cleaning, measurement, and machining.
Distinguishing between functions and operating principles makes two different types of exploration possible.
The first is searching for alternative operating principles that achieve the same function.
The second is searching for new applications of an existing operating principle.
These correspond to two important directions in technological innovation: exploring applications for existing technological capabilities and exploring technologies capable of satisfying particular needs.
Structural and Relational Abstraction
Another form of Abstraction focuses on relationships among components rather than on the attributes of individual objects.
For example, systems belonging to different technological fields may share a relationship structure such as:
Observe a state → Compare it with a reference → Detect a deviation → Respond
This structure may be found in industrial control systems, information system monitoring, and biological regulatory mechanisms.
However, similarity of structure should not be confused with identity of operating principles.
For example, biological homeostatic mechanisms and industrial feedback control systems may share relationships involving observation and regulation.
Nevertheless, actual biological systems involve numerous biochemical and physiological processes, and an industrial control model cannot necessarily be transferred directly to a biological system.
Even so, abstracting their relationships provides a basis for comparing phenomena across disciplines and exploring common problems or alternative solutions.
Relationship with Structure-Mapping Theory
A cognitive theory closely related to Abstraction is Structure-Mapping Theory, proposed by Dedre Gentner.
In her 1983 paper “Structure-Mapping: A Theoretical Framework for Analogy,” Gentner explained analogical reasoning in terms of the correspondence between relational structures in different domains.
The theory emphasizes how relationships among elements correspond, rather than relying primarily on similarities in the superficial attributes of objects.
For example, two systems may have components that look completely different but nevertheless exhibit comparable relationships among those components.
An analogy may then be established through the correspondence of their relational structures.
Furthermore, it is important that relationships form a coherent structure rather than merely coinciding in isolated ways.
This principle is known as Systematicity.
For example, the fact that two technologies both “transmit information” provides only limited evidence of structural similarity.
However, if their processes of information generation, route selection, transmission, reception, and feedback correspond in a coherent manner, a more substantial structural comparison becomes possible.
Nevertheless, Abstraction and Structure Mapping are not the same operation.
Abstraction extracts characteristics or relationships from individual objects.
Structure Mapping establishes correspondences between the relational structures of different objects or domains.
An abstracted structure can be used to search for correspondences across domains, but whether those correspondences are meaningful must be examined separately.
Abstraction and Analogy
Analogy is a thinking process that uses knowledge of one object or situation to understand another or solve a different problem.
Analogical reasoning plays an important role in technological development that draws on knowledge from different domains.
For example, a supply network used in one industry might provide insights into material transportation problems in another.
However, when two technologies are compared directly, their common structures may be difficult to recognize.
Abstraction can help by extracting and clarifying what the systems have in common.
In their 1983 study “Schema Induction and Analogical Transfer,” Mary Gick and Keith Holyoak demonstrated that extracting a common problem-solving structure from multiple concrete examples can facilitate analogical transfer to another problem.
However, providing an abstract explanation does not automatically produce successful analogy.
The same research also identified conditions in which abstract descriptions derived from a single example were insufficient to facilitate transfer.
Abstraction is therefore an important support for analogical reasoning, but comparison across examples and examination of specific correspondences remain necessary.
Example: Supplying Repair Agents in Self-Healing Materials
Consider the relationship between Abstraction and cross-domain exploration using self-healing materials as an example.
Some self-healing materials contain capsules filled with repair agents.
When damage occurs, the capsules rupture and release their contents into the damaged region.
However, because the repair agent is stored locally and consumed when the capsule ruptures, such systems may have difficulty responding to repeated damage at the same location.
If the problem is defined as “improving capsules containing repair agents,” exploration may concentrate on capsule materials and configurations.
Abstraction allows the problem to be expressed differently:
“How can a system repeatedly supply required substances to the locations where they are needed?”
With this formulation, the search is no longer restricted to capsule-based mechanisms.
For example, biological vascular systems, material transportation in plants, and industrial fluid distribution networks may offer potentially relevant structural principles.
The abstracted problem can be represented as follows.
| Perspective | Abstracted Representation |
|---|---|
| Objective | Supply required substances to damaged locations |
| Limitation of the existing approach | Local storage makes repeated supply difficult |
| Required functions | Supply, transportation, distribution, replenishment |
| Structure | Supply source → Transportation pathway → Required location |
| Temporal characteristic | Repeated operation rather than one-time release |
| Pattern to explore | Networks that continuously supply substances to distributed targets |
However, biological vascular systems and artificial supply networks embedded within materials differ in many respects, including materials, dimensions, pressures, fluid properties, and manufacturing methods.
The existence of a similar structure in biological systems does not mean that it can be directly implemented in artificial materials.
Relevant considerations include the viscosity of the repair agent, the formation of internal transportation pathways, replenishment of the supply source, and preservation of pathways after damage.
In this example, Abstraction creates opportunities to identify new subjects for exploration.
Actual technological feasibility must still be investigated through research and experimentation.
A Practical Procedure for Abstraction
When abstracting a technology or problem, the following process can be used.
- Describe the concrete subject. Organize what is known about the product, technology, phenomenon, or problem being examined.
- Clarify the purpose of Abstraction. Determine whether the objective is classification, problem understanding, causal analysis, or exploration of new applications.
- Extract characteristics and components. Distinguish attributes, functions, operating principles, structures, relationships, and constraints.
- Select information relevant to the objective. Preserve important characteristics and relationships while temporarily omitting less relevant details.
- Create an abstract representation. Express the subject in terms of functions, structures, or patterns without depending on specific product or technology names.
- Compare it with other examples. Investigate whether similar functions or structures occur in different objects or domains.
- Return to concrete conditions for verification. Reintroduce omitted conditions and examine whether the proposed correspondence or application is feasible.
It is important not to assume that a useful abstraction can always be produced in one attempt.
For example, the initial abstraction “transporting materials” may be too broad for a particular problem.
It could be refined to “continuously supplying materials from an external source to small damaged regions.”
Conversely, if search results remain concentrated within the original technological field, removing more domain-specific terminology or implementation details may broaden the search.
Abstraction and Concretization can therefore be repeated until an appropriate set of exploration conditions is identified.
The Relationship Between Abstraction Level and Search Scope
The degree and type of Abstraction influence the scope of objects that can be explored.
Consider a vibration sensor.
| Representation | Characteristics of the Search Scope |
|---|---|
| Industrial vibration sensor | Restricted to a particular technological category |
| Device for measuring equipment vibration | Includes a wider range of vibration measurement methods |
| Mechanism for detecting changes in equipment conditions | Includes measurement methods beyond vibration |
| Mechanism for detecting deviations from normal conditions | Includes anomaly detection beyond physical measurement |
| System that observes states and responds to changes | Extends to control systems, monitoring, biological regulation, and other domains |
Generally, more abstract expressions may include a greater variety of objects.
However, a broader search scope also increases the number of irrelevant candidates.
For example, “a system that observes states and responds to changes” could describe an enormous variety of phenomena.
On the other hand, restricting the search to “industrial vibration sensors” may exclude useful solutions from different technological fields.
A higher level of Abstraction is therefore not always better.
Effective Abstraction requires balancing the expansion of the search scope with the preservation of conditions that are important to the problem.
It should also be noted that the representations above do not necessarily form a single category hierarchy.
They combine movement toward broader categories with changes in functional and structural perspectives.
In actual exploration, several different abstractions may be used in parallel.
What Should Be Preserved and What Should Be Omitted?
The central challenge of Abstraction is deciding which information to preserve and which information to disregard.
For example, when exploring applications of a manufacturing technology in another industry, the original product’s external shape or material name may be irrelevant.
However, temperature, pressure, chemical stability, reaction speed, and other conditions may be essential to the technology’s operation.
Information can therefore be examined according to the following distinctions.
| Type of Information | Treatment in Abstraction |
|---|---|
| Superficial attributes | May be omitted if unrelated to the objective |
| Functions | Often preserved to clarify what the system accomplishes |
| Operating principles | Important when they determine transferability |
| Structures and relationships | Important for identifying cross-domain correspondences |
| Necessary conditions | Omitting them may lead to invalid analogies |
| Constraints | Help narrow the range of feasible candidates |
| Implementation details | May be omitted during initial exploration and reintroduced during verification |
However, these distinctions are not absolute.
For example, the color of a material may be irrelevant to many engineering problems but essential to optical measurement or appearance quality.
What counts as an essential characteristic cannot always be determined uniquely from the object itself.
Abstraction is a purpose-dependent selection of relevant information, and there is not necessarily one uniquely correct abstraction of an object or phenomenon.
From Abstraction Back to Concretization
Discovering a new possibility through Abstraction does not mean that the original problem has been solved.
To apply an abstract concept to a specific technology or product, the conditions of the actual application must be considered again.
This process can be described as Concretization.
For example, suppose the abstract function “repeatedly supplying substances to required locations” leads to the idea of using a fluid distribution network.
Applying this idea to self-healing materials requires detailed examination of internal transportation pathways, repair agent properties, supply pressure, manufacturing methods, and durability under repeated operation.
Abstraction broadens the range of potential solutions.
Concretization examines whether those solutions can operate under real conditions.
Combining these processes makes it possible to use knowledge from different fields in practical problem-solving.
However, Concretization may reveal that an initially promising structural correspondence does not actually hold under the necessary operating conditions.
In that case, alternative technologies may need to be explored, or the original abstraction may need to be revised.
The movement from Abstraction to Concretization is therefore not a simple one-way procedure.
It is an iterative process of exploration and verification.
Abstraction and Concept Formation
Abstraction plays an important role in Concept Formation.
When recognizing similarities among objects or phenomena and grouping them under a common concept, it is necessary to compare their characteristics and identify what they share.
However, Concept Formation is not limited to discovering established categories.
When new phenomena or technologies emerge, existing concepts may be insufficient to describe them.
For example, similar technological mechanisms might begin appearing independently in several industries.
If each example is classified exclusively within its established industrial category, the shared structure may remain unnoticed.
Comparing these examples and abstracting their common characteristics or relationships may lead to the formation of a new concept.
However, the mere presence of similarities does not automatically make a new concept useful.
Its value should be examined according to whether it can explain phenomena previously considered separately, help identify additional examples, or support further investigation and prediction.
Relationship with Concept Exploration
Abstraction is an important thinking operation within Concept Exploration, an approach proposed by Mindware Research Institute.
Concept Exploration emphasizes discovering emerging concepts and investigating unknown objects through their characteristics, functions, structures, relationships, and patterns, rather than relying exclusively on known names or disciplinary classifications.
Conventional information retrieval generally requires users to know relevant keywords or categories in advance.
For example, searching for an established material name or technological method can retrieve information associated with that known subject.
However, a technology with a similar function may exist in another industry under an unfamiliar name.
Without knowing that terminology, it may be difficult to include the technology in a conventional keyword-based search.
Abstraction can address this limitation by describing familiar technologies through their underlying functions and structures.
These descriptions can then become search conditions that are not limited to the original terminology.
For example, describing “repeatedly supplying repair agents to damaged regions of a material” may support exploration beyond the category of self-healing materials.
It could provide a starting point for identifying similar mechanisms in other industries or biological systems.
Comparing the discovered examples may then reveal additional common structures, allowing new search concepts to be formed and exploration to continue.
In this sense, Abstraction within Concept Exploration has at least two important directions.
The first involves comparing multiple observed examples and abstracting their shared characteristics or relationships to form new concepts.
The second involves abstracting functions, structures, or patterns from a known technology or problem and using those abstractions to search for unfamiliar objects.
The second direction has developed into Cross-Domain Concept Exploration.
However, Abstraction alone does not discover unknown technologies.
It produces representations that can be used for exploration. Identifying candidates and verifying whether the proposed correspondences are valid require additional processes.
Abstraction with AI and Machine Learning
Although Abstraction is an important human cognitive activity, AI and machine learning can support it.
For example, LLMs can summarize common characteristics across documents or reformulate product and technology descriptions in more general terms.
They can also generate explanations of a technology from different perspectives, such as attributes, functions, operating principles, structures, and relationships.
A vibration sensor, for example, might be described in terms of its higher-level category, purpose, function, sensing principle, and potential applications.
However, an abstract representation generated by an LLM is not necessarily technically appropriate.
When generalizing a mechanism, the model may omit physical conditions essential to its operation.
It may also group different technologies under the same concept merely because they perform superficially similar functions.
It is therefore desirable to retain links between the abstract representation and its source information, allowing users to verify which characteristics were preserved and which were omitted.
Embedding Vectors and Conceptual Structures
Machine learning can represent textual information using Embedding Vectors and organize it according to semantic similarity.
Such representations can sometimes position semantically related texts close to one another even when their surface vocabulary differs.
Dimensionality reduction methods such as UMAP and structural learning methods such as Growing Neural Gas (GNG) can also be used to organize large collections of information into conceptual maps.
However, embedding vectors do not automatically provide explicit representations of concepts or operating principles.
For example, two descriptions of technologies may be semantically similar even though their actual operating principles or structures differ.
Conversely, technologies from different disciplines may use substantially different terminology despite sharing important operating principles.
In such cases, embedding-based similarity may fail to reveal the underlying correspondence.
Semantic similarity analysis using embeddings should therefore be distinguished from the explicit abstraction of functions and structures.
Embeddings and concept maps can help identify and organize information worth comparing.
LLMs or human analysts can then examine that information to extract common functions, structures, and relationships, enabling more explicit conceptual analysis.
Limitations of LLM-Based Abstraction
LLMs can generate abstract descriptions using natural language.
However, generating an abstract description does not establish that the concept can be applied to real objects or systems.
For example, describing a technology as “a system in which distributed elements cooperate to achieve an overall function” may encompass a wide variety of technologies and phenomena.
Yet such a description may be too general to establish a technically meaningful correspondence.
Excessively broad abstractions can apply to many irrelevant examples.
Conversely, retaining too many characteristics specific to the original technology may prevent useful cross-domain discoveries.
LLM-based Abstraction should therefore be guided by constraints relevant to the exploration objective.
For example, a prompt might ask:
“Analyze this technology and extract its functions, operating principles, and relationships among components while excluding product names, material names, and other domain-specific attributes. Distinguish the conditions essential to the function from those that could be replaced by alternative mechanisms.”
Such instructions may produce abstract representations that are more useful for cross-domain exploration than simple general summaries.
However, the extracted conditions should be verified against the original technical documents and other relevant evidence.
Evaluating the Quality of Abstraction
There is not always one correct result when abstracting an object or phenomenon.
Nevertheless, the usefulness of an abstraction can be evaluated relative to its purpose.
For example, if Abstraction is intended to support cross-domain technological exploration, several criteria can be considered.
| Evaluation Criterion | Key Question |
|---|---|
| Relevance to the objective | Does the abstraction address the problem being investigated? |
| Functional clarity | Is it clear what the system is intended to accomplish? |
| Structural consistency | Are relationships among elements represented coherently? |
| Preservation of necessary conditions | Have important constraints and operating conditions been retained? |
| Exploration scope | Does the abstraction permit exploration beyond established categories? |
| Discriminative power | Does it avoid including too many unrelated objects? |
| Verifiability | Can the proposed correspondences with discovered candidates be examined concretely? |
For example, “detecting abnormalities” is a useful broad functional abstraction.
However, because many different technologies perform this function, the search may produce too many unrelated candidates.
A more specific expression might be:
“Detecting abnormalities by observing small periodic changes without physical contact and comparing them with normal conditions.”
This representation specifies the function and important measurement conditions more clearly.
However, its greater specificity may also reduce the exploration scope.
Comparing several alternative abstractions can help determine which representations lead to useful discoveries.
Limitations of Abstraction
Abstraction has several important limitations.
First, characteristics considered important depend on the objective. Information that can be omitted for one purpose may be essential for another.
Second, Abstraction may remove conditions necessary for a technology to function. Material properties, dimensions, temperatures, pressures, and time scales can strongly influence technological feasibility.
Third, superficial similarities may be overestimated. Technologies performing similar functions may rely on different operating principles and constraints.
Fourth, increasing the level of Abstraction can reduce discriminative power. Extremely general concepts may fail to distinguish the relevant technologies or phenomena.
Fifth, the resulting abstractions depend on the analyst’s knowledge and assumptions. Organizing information according to familiar categories may prevent the formation of genuinely new concepts.
Sixth, Abstraction alone cannot establish implementation feasibility. Discovering a similarity across domains does not demonstrate that a technology can be transferred successfully.
Seventh, Abstraction involves information loss. It may not be possible to reconstruct the complete original object from an abstract representation.
It is therefore useful to preserve the original information alongside the extracted characteristics and omitted conditions.
Abstraction is not intended to explain every aspect of an object. It selectively represents what matters for a particular purpose, enabling new understanding and exploration.
Abstraction in ThinkNavi
ThinkNavi regards Abstraction as a fundamental thinking method for extracting characteristics, functions, operating principles, structures, and relationships from concrete objects or phenomena and using them to support problem-solving and knowledge exploration.
ThinkNavi Free could provide a worksheet for examining technologies or problems from several perspectives.
For example, users could describe a specific technology and manually organize its attributes, functions, operating principles, components, relationships, and constraints.
They could then create multiple abstract representations based on that information.
When the objective is classification, users could examine broader categories and shared attributes.
When exploring new applications, they could focus on functions and operating principles.
When developing cross-domain analogies, they could extract relationships among components and patterns of change over time.
A tool that also records the movement from abstract concepts back to specific technologies or applications could help users understand the relationship between idea generation and verification.
An LLM-assisted version could generate alternative abstractions of the input information from functional, structural, and relational perspectives.
It could also suggest representations at different levels of abstraction.
However, rather than presenting one LLM-generated description as the correct abstraction, the system should allow users to compare multiple alternatives.
It would also be useful to preserve the relationship between source information and abstracted representations so that users can examine what was retained and what was omitted.
Further integration with ConceptMiner R&D is also possible.
ConceptMiner R&D uses characteristics and patterns extracted from technological information to support Cross-Domain Concept Exploration.
In this process, how the original technology is abstracted and which conditions guide exploration can substantially influence the results.
For example, multiple abstraction candidates could be generated from the same technological element.
Each abstraction could then be used to conduct cross-domain exploration.
Comparing the results would make it possible to investigate which abstractions produced useful candidates.
For each discovered candidate, the system could also examine shared functions and structures, as well as differences in operating conditions.
Accumulating such analyses could help improve the exploration process by revealing relationships between abstraction methods and subsequent discoveries.
However, an appropriate abstraction does not guarantee that a discovered technology can be implemented successfully.
Technical constraints, required modifications, and actual feasibility must be investigated separately.
The significance of Abstraction lies not simply in replacing concrete objects with general terminology, but in discovering new possibilities for comparison and exploration among objects and phenomena that were previously considered unrelated.
Achieving this requires both expanding the range of potential discoveries through Abstraction and returning to concrete technological conditions to verify whether the proposed correspondences are valid.
Abstraction is a fundamental thinking operation underlying Analogy, Structure Mapping, Concept Formation, and Concept Exploration. It also has an important role to play in AI-assisted exploration of knowledge across disciplinary boundaries.
Primary English Term: Abstraction
Related Terms: Generalization, Concept Formation, Conceptualization, Idealization, Functional Abstraction, Relational Abstraction, Structural Abstraction, Concretization, Analogical Transfer
Related Thinking Methods:
- Generalization,
- Concept Formation and Categorization | Change the Categories, Change What You See
- Analogy — Connecting Knowledge Across Domains to Generate New Ideas and Hypotheses
- Structure-Mapping Theory | Understanding Relational Correspondences and Generating New Inferences Across Domains
- Logical Thinking — Building Sound Arguments, Communicating Reasoning, and Achieving Mutual Understanding
- Systems Thinking — Understanding Interdependence, Feedback, and Change Over Time
- Concept Maps — Representing Relationships Between Concepts and Understanding Knowledge Structures
- Functional Analysis,
- Morphological Analysis | Systematically Exploring New Possibilities Through Combinations of Elements
- TRIZ | A Systematic Approach to Inventive Problem Solving Through Technical Contradictions
- Concept Exploration | Discovering Emerging Concepts and Unknown Possibilities
- Cross-Domain Concept Exploration,
- R&D Theme Exploration and Generation
References and Further Reading
- Eleanor Rosch, Carolyn B. Mervis, Wayne D. Gray, David M. Johnson, and Penny Boyes-Braem, “Basic Objects in Natural Categories,” Cognitive Psychology, Vol. 8, No. 3, pp. 382–439, 1976. A foundational study of category recognition and basic-level categories.
- Dedre Gentner, “Structure-Mapping: A Theoretical Framework for Analogy,” Cognitive Science, Vol. 7, No. 2, pp. 155–170, 1983. A foundational theory explaining analogy through correspondences between relational structures.
- Mary L. Gick and Keith J. Holyoak, “Schema Induction and Analogical Transfer,” Cognitive Psychology, Vol. 15, No. 1, pp. 1–38, 1983. Research examining the relationship between extracting shared problem-solving structures from examples and analogical transfer.
- Dedre Gentner and Arthur B. Markman, “Structure Mapping in Analogy and Similarity,” American Psychologist, Vol. 52, No. 1, pp. 45–56, 1997. A discussion of structural alignment in similarity and analogy.
- Dedre Gentner and José Medina, “Comparison and the Development of Cognition and Language,” Cognitive Studies: Bulletin of the Japanese Cognitive Science Society, Vol. 4, No. 1, 1997. Research on comparison, structural alignment, and cognitive development.
- Arthur B. Markman, “Structural Alignment in Similarity and its Influence on Category Structure,” Cognitive Studies: Bulletin of the Japanese Cognitive Science Society, Vol. 4, No. 4, 1997. Research on structural comparison and category formation.
- Stanford Encyclopedia of Philosophy, “Models in Science.” An overview of abstraction, idealization, and representation in scientific models.
- Stanford Encyclopedia of Philosophy, “Scientific Representation.” Philosophical perspectives on the relationships between scientific models and their targets.
- Mindware Research Institute, “What Is Concept Exploration?.” An introduction to the concept and background of Concept Exploration. (Japanese)