Ontology-Driven Cultural Heritage Conservation: A Case of The Analects of Confucius

Abstract

Confucianism, recognized as the belief system of Chinese, is one of the most important intangible cultural heritages of China. The main ideas of its founder, Confucius, are written in The Analects of Confucius. However, its scattered chapters and the obscurity of ancient Chinese have prevented many people from understanding it. In order to overcome this difficulty, it needs some modern ways to reveal the vague connotation of Confucianism. This paper aims to describe how to construct the Lunyu ontology in which all concepts are abstract within the core scope, i.e., morality of Confucianism. The key task of this project lies in identifying essential characteristics, a notion that is compliant with the ISO principles on Terminology (ISO 1087 and 704), according to which a concept is defined as a combination of essential characteristics. This paper proposed an approach in the practice of identifying essential characteristics of abstract concepts from different meanings of its Chinese terms in The Analects of Confucius. With this work, Lunyu ontology established a semantic, formal, and explicit representation system for concepts of Confucianism, and the new proposed approach provides a useful reference for other researchers.

1. Introduction

Cultural Heritage is a civilization’s memory of a country or a society, as well as the wealth shared by people all over the world. It records the thoughts and activities of the ancients who lived in a certain area and at a certain time. Cultural heritage is divided into the tangible cultural heritage and intangible cultural heritage (ICH). According to Convention for the Safeguarding of the Intangible Cultural Heritage [1] reported by UNSECO, “Intangible Cultural Heritage means the practices, representations, expressions, knowledge, skills—as well as the instruments, objects, artifacts and cultural spaces associated therewith—that communities, groups and, in some cases, individuals recognize as part of their cultural heritage.” As the most important one among the rich ICH of China, Confucianism has been transmitted from generation to generation for over 2200 years and has been interpreted by thinkers in different eras, including thinkers in the present. Literally, the reason to recognize it as the most essential ICH of China is that Confucianism is the cultural foundation of Chinese society since the Qin dynasty (221 BCE–207 BCE) until the beginning of the 20th century. The main content of Confucian culture is condensed in the book named The Analects of Confucius (Chinese: Lunyu), which was written during the Spring and Autumn Period and the Warring States Period (770 BCE–221 BCE). The difficulty of understanding The Analects of Confucius lies in two aspects. One is that this book is a collection of dialogues; thus, it is difficult to understand its meaning without context. Second, there are great differences between ancient Chinese and modern Chinese. Therefore, compared with the requirement of other cultural heritages, which is to be preserved, Confucianism needs to be represented in more modern ways such that people with different cultural backgrounds can understand it better.

People in different dynasties of China all faced the problems mentioned above more or less and attempted to solve them. In Chinese history, many scholars annotated The Analects of Confucius. One of the most famous scholars is Zhu Xi, who lived in the Southern Song Dynasty, and his contributions will be referenced in this study in later chapters. In the modern era, the form of research on The Analects of Confucius varies from annotation to illustration, translation, etc. However, although much effort has been paid, it is still hard to clearly define Confucianism. Since the challenge to comprehend The Analects of Confucius derives from the illogicality of the text and the barrier of the ancient Chinese, this study attempts to reshape its core content—the morality system of Confucianism, based on ontology to propose an approach to solve this problem. Ontology is a modeling tool that could be used to describe concepts of information systems on semantic and knowledge levels [2]. Conversely, the Confucian morality system is embodied in the concept of some single characters that are independent in form but related in connotation. From the perspective of requirement and function, the formalization and clarity of ontology are exactly in line with the purpose of reshaping ideas in The Analects of Confucius. The outcome of this study, Lunyu ontology, aims to devote contributions to both providing a new way to learn about Confucianism and expanding the application of this methodology.

The rest of this paper is organized as follows: Section 2 briefly introduces the history of Confucianism and describes the domain of research, Section 3 puts forwards the objectives, Section 4 reviews the state of the art, Section 5 is dedicated to our contribution to ontology building methodology relying on a meaning analysis of Chinese terms and the ISO principles on Terminology, Section 6 presents the Lunyu ontology in Protégé, and the final section presents the evaluation of the Lunyu ontology.

2. Foundation of Research

2.1. Brief History of Confucianism

Confucianism is one of the most influential philosophies in the history of China, and it has existed for over 2500 years. It is a belief system rather than a religion, which focuses on the importance of personal ethics and morality. Confucius, the creator of Confucianism, is the most famous philosopher and teacher in China history and his thoughts on ethics, good behavior, and moral character were written down by his disciples in several books, the most important being The Analects of Confucius [3,4].

After Confucius, Mencius (Chinese: Mengzi), a fourth-century BCE Chinese thinker, enriched Confucianism by his theory that all human begins share innate goodness that either can be cultivated through education and self-discipline or squandered through neglect and negative influences, but never lost altogether [5].

When time came to the Southern Song Dynasty, another famous Confucian Zhu Xi created the supreme synthesis of Son-Ming Dynasty (960–1628 CE) Neo-Confucianism, which is known as lixue or daoxue. He redefined the Confucian tradition and brought the focus of Confucianism back to the moral cultivation of people which is the original focus of Confucianism by selecting the essential classical Confucian texts including The Analects of Confucius then compiling them with annotations [6]. This is why we referenced his contribution.

About 400 years later, Wang Yangming, who studied Zhu Zi seriously in his teenage years, developed his contribution to Neo-Confucianism. Wang Yangming’s thought broke through the limitations of lixue by arguing for a different philosophical interpretation and cultivation of the xin or mind-heart, such that his distinctive philosophy is called xinxue or the teaching of the mind-heart. Xinxue and lixue are the two most significant movements of Neo-Confucianism. Until the end of China’s feudal dynasty, Neo-Confucianism dominated the political field of China [7,8].

2.2. Description about Data Resource

Obviously, all content of Confucianism is too complex and profound; thus, this study selected the most core part of it, the morality system, to promote our work. The domain of this research can be outlined as follow four constraints:

• All concepts, therefore, all terms, come from The Analects of Confucius.
• Every concept is related to Confucian morality.
• Meanings of terms which points to the concepts are referenced from Shisanjing Dictionary [9] and The Analects of Confucius Variorum [10].
• Meanings of terms irrelevant to Confucian morality are not considered.

The first constraint is based on the fact that Confucian culture experienced a process of continuous derivation and scholars in different times made different interpretations of Confucianism. The Analects of Confucius mainly records the dialogues between Confucius and his disciples, some of the events of his life, and disciple’s comments on Confucius; thus, it is not a philosophical book that logically introduces the thought system of Confucius. If any book other than the Analects of Confucian was involved, there may be ambiguity in the connotation of a specific concept. Even if the scope was limited to such a book, there is no worry if some important concepts have been missed, because The Analects of Confucius comparably concentrated embodies the political views, ethical thoughts, moral concepts, and educational principles of Confucius and Confucian school. The Analects of Confucius seems to be short in length, but actually, as mentioned above, it covers many aspects of Confucian Culture.

Confucius advocated telling rather than writing; thus, The Analects of Confucius did not describe or explain Confucius’s thoughts, and even if they recorded what Confucius said, they did not record the context and background of the dialogues. Therefore, each chapter of The Analects of Confucius is only a few sentences long and almost independent from each other. This is why the second limitation is that concepts should be related to morality.

In particular, ancient Chinese is dominated by monosyllabic words, that is, a word represents an independent and complete meaning; thus, all concepts in our ontology appear in the form of a single word in The Analects of Confucius. These particularities of this book caused another problem for this study, that is, it is impossible to find the characteristics of these concepts in the original text. It led to the third condition that Shisanjing Dictionary and The Analects of Confucius Variorum provide meanings of terms. Actually, the entries in the Shisanjing Dictionary already contain almost all the meaningful words (i.e., characters) in The Analects of Confucius and list all their explanations that occurred in this book. While some of them are still not detailed or comprehensive enough and were replenished by Zhu Xi’s annotation in The Analects of Confucius Variorum. Of course, meaning items of terms referring to moral concepts have been eliminated if they are not within the scope of Confucian morality, as the fourth constraint says.

To sum up, we obtained all terms from The Analects of Confucius with the guide of experts. In order to ensure the accuracy of the comprehension of Confucianism, we acquired the meanings of each term from the Shisanjing Dictionary, a dictionary specially compiled for 13 Confucian classics, and Zhu Xi’s The Analects of Confucius Variorum of The Southern Song Dynasty (960–1279) as a complement.

3. Objective

The goal of this study includes (1) to build a knowledge representation of Confucianism at the aspect of morality and ethic in the form of an open ontology in a W3C standard and (2) to build a bilingual (Chinese-English) terminological e-dictionary for users who want to know the thoughts of Confucianism. This article aims to present the first goal that aims to build a knowledge base based on ontology for the second goal. It is an attempt to express Chinese philosophical ideas in a new way from traditional ones where Chinese philosophy tends to use vague and suggestive language when expressing ideas. With the building of Lunyu ontology, thoughts of Confucianism will be represented in a formal and explicit form. Moreover, we hope to accumulate some experience in the process of constructing the ontology to contribute to the methodology.

Competency questions are used to specify the requirements of the ontology [11]. A question such as “What are the Chinese terms of concept and their equivalents in English?” raises the problem of modeling the linguistic dimension of the terminology. Questions such as “What are the concepts that express character?” stress the hierarchy of classes. While “What are the definitions of the “filiality” in English and Chinese?” particularly affects ontology modeling because it directly relates to the focus of the ontology. Answering these questions requires an explicit representation of essential characteristics.

Table 1. The three competency questions.

CQ	Competency Questions	Class	Relation
1	What are the Chinese terms of concept and their equivalents in English?	Concept	Rdfs:subClassOf, skos:prefLabel
2	What are the concepts that express character?	Concept, Character	Rdfs:subClassOf
3	What are the definitions of the “filiality” in English and Chinese?	Concept, Filiality	Rdfs:subClassOf, skos:definition

presents some of the competency questions taken into account for the implementation of the Lunyu ontology. The table specifies the classes and relations required to answer these queries.

4. State of the Art

This chapter will present some current work about ontology and its application in cultural heritage. A general introduction of ontology and ontology building methodologies is to be made initially. Then we are going to take some examples to show how ontology is applied in cultural heritage conservation. In the last part, standards and models available will be introduced.

Ontology originated in philosophy and extended to other fields with the rapid development of computer science. After Nech et al. proposed the definition of ontology in artificial intelligence [12], its definition was expanded and refined to the point where it is now the most common definition which is “an ontology is a formal, explicit specification of a shared conceptualization” including four layers of meaning: conceptual model, explicit specification, formal, and shared [13,14]. Ontology construction is a systematic engineering, but there is no absolutely correct methodology for developing ontologies. The characteristic of domain knowledge is also a key factor to be considered during ontology design. Nonetheless, ontology construction still follows certain basic steps. There are various methods to develop a domain ontology, including Skeletal Methodology [15], IDEF5 [16], TOVE [17], SENSUS [18], and so on. The Ontology Development 101 [19] method developed by Stanford University School of Medicine is the most mature and widely used ontology development method at present.

Ontology has a wide range of application scenarios in the domain of cultural heritage and could be used in different stages in cultural heritage protection. Ontology for capturing and processing 3D cultural heritage objects could enhance the operability of 3D scanning and the control of data quality [20]. Mohammed Maree et al. proposed Holy-Land ontology, which attempts to encode knowledge about Palestine’s cultural heritage, aiming to provide multi-language semantic retrieval support for their cultural heritage system [21]. Compared to tangible cultural heritage, ICH is abstract and requires more detailed consideration in ontology design. Correspondingly, an appropriate ontology model can be of great benefit in representing abstract transactions. Ontology of Terengganu Brassware Craft completely records the knowledge about this craft and the process of constructing it presenting the experience in developing a domain ontology for ICH [22]. Similarly, ontology-based knowledge and understanding obtained from experts and knowledgeable members improved the inheritance of culture [23]. By building ontologies beforehand, entities and relationships could be extracted from the structured or unstructured text about ICH [24,25]. In some cases, existing ontologies can be reused to organize and classify new cultural heritage data collection. The degree of reuse depends on what class and properties those ontologies have. As for an example of representing an idea system with ontology, we have not yet found one, although there are not many studies that introduce ontology into ICH.

The digitization of the ICH is much more complicated than the general cultural heritage; thus, it requires the extension of the standard ontology for digitization of cultural heritage [26]. For this reason, it would be helpful to reference some public resources when developing a specific domain ontology of ICH. The W3C standard OWL (Web Ontology Language) [27] can be used to explicitly represent the meaning of terms in vocabularies and the relationships between those terms with good machine interpretability. SKOS (Simple Knowledge Organization System) [28] is a data model for sharing and linking knowledge organization systems on the Web. It captures the similarity of many knowledge organization systems and makes it explicit; thus, it is widely used in knowledge organization. DC (Dublin Core) [29] is a metadata schema (Schemas are machine-processable specifications that define the structure and syntax of metadata specifications in a formal schema language.) based on 15 essential properties to describe online and physical resources and it is also cited in this study. OTContainer proposed a term-guided method for ontology construction, which is applied in TAO CI ontology [30]. We will adopt this method in our research. There are also specific standards for cultural heritage. The CIDOC CRM [31] ontology is an official standard whose ISO standard can be found at ISO 21127: 2014. It provides the basic classes and relations devised for the cultural heritage. EDM [32] defined formal specification of the classes and properties that could be used in Europeana which develops expertise, tools, and policies to empower the cultural heritage sector in its digital transformation, whereas both are not suitable for our ontology modeling because we need to build a “granular” ontology to represent concepts of The Analects of Confucius, which means that we extract knowledge at a much deeper and more detailed level and build the ontology from it. The knowledge is specific and has the particularity of the field. Nevertheless, classes CIDOC and EDM defined are too general and common to be transplanted to our ontology.

5. The Methodology of Lunyu Ontology

Ontology building follows a lifecycle of several stages [33]. Some of them have to be specialized, others have to be newly introduced to take into account the specificities of the domain. The theory of concept underpinning the ontology can also strongly impact the methodology of engineering it. Following the ISO principles on Terminology where “a term is a verbal designation of a concept” and “a concept is a unique combination of (essential) characteristics” (An “essential characteristic” is a characteristic (abstraction of a property) of a concept and is indispensable to understanding that concept (ISO 1087). Essential characteristics correspond to rigid predicates in DL [34] and to rigid properties in the OntoClean method [35]. Unlike essential characteristics, which define the concept, “descriptive characteristics” own values which describe the current state of an object, e.g., weight, color, etc.) (ISO 1087, ISO 704), we adopt a “term-and-characteristic” guided method derived from previous work [36,37,38]. Identifying essential characteristics becomes the first goal to achieve.

The problem of identifying essential characteristics is a new and central phase of our methodology. For concrete concepts, this phase is aimed at identifying differences between objects (chair with leg versus chair without a leg). For abstract concepts, this phase is aimed at identifying differences between meanings of concepts because the concepts do not have a physical entity corresponding to it, e.g., the meaning of the term “礼”(ritual) is a hierarchical system that regulates the external behavior. Thus, the essential characteristics were {/system/, /hierarchical/, /external/, /behavior/}.

The term-and-characteristic guided methodology includes seven steps, each aiming at different tasks. Step 1: identify the scope of the domain and the objectives; Step 2: identify terms and objects; Step 3: identify essential characteristics; Step 4: define concepts; Step 5: build ontology using one of the available tools; Step 6: integration of other resources; Step 7: evaluation.

This chapter is dedicated to step 3 (identifying essential characteristics) and step 4 (combining essential characteristics into concepts). The first step has been presented in the previous chapter 3. There is no phase of identification of terms and objects, since the experts know the terms to be defined. The Lunyu ontology engineering in the OWL phase as well as the evaluation phase will be presented in dedicated chapters.

5.1. Identifying Essential Characteristics of Abstract Concepts

Identifying differences between meanings is a useful means toward identifying essential characteristics. Because this paper focuses on the concepts related to personal morality in The Analects of Confucius, the differences can be an object that is described by a concept (e.g., for oneself, for others), object’s identity (parents, brother, friends, monarch), concept’s type (character, ability, need, attitude, system, and morality). Thus, the presence or the absence of differences can be interpreted as essential characteristics. For example, the meaning of concepts could be divided into analysis axis: person, resource, identity, binding, range, action, hierarchy, performance and type (Figure 1).

From the concept type point of view, the concepts could be divided into concept for attitude, concept for character, concept for ability, concept for need, concept for system, and concept for morality corresponding to the essential characteristic/character/, /ability/, /need/, /attitude/, /system/, and /morality/ (Figure 2). These essential characteristics are exclusive to each other.

5.2. Defining Concepts Based on Essential Characteristics’ Combination

From the terminology point of view, a concept is defined as a unique combination of essential characteristics (ISO 1087). However, not any combination of essential characteristics defines a meaningful concept from the expert’s point of view. For the expert, concepts of interest are those that are named in a natural language. Hence, a concept is a set of essential characteristics stable enough to be named in a given language (even if some concepts, without any designation in natural language, can be introduced for organizational purposes of the conceptual system). Terms can be then considered as guidelines for identifying domain concepts to be defined from the expert point of view. For example, the term “信”, “trust” in English, denotes the set of essential characteristics {/morality/, /realize/, /promise/, /others/, /honesty/}. Based on this formal definition, the definition in natural language is then: “Morality that realizes promise to others and expressed as honest”.

6. The Lunyu Ontology

The specific implementation method of the “term-and-characteristics” guided method in Protégé was presented in the literary [37,38]. For example, concepts are defined as named classes in Protégé. Terms are represented as labels (skos:prefLabel, skos:altLabel, skos:definition). Relations, such as “hasRange”. “hasObject” are represented as object properties. Implementing essential characteristics is a slightly more complex process. Since essential characteristics correspond to rigid predicates, they cannot be directly expressed into description logic. Reference [39] introduced different ways to implement essential characteristics in Protégé. The first way is to implement essential characteristics as classes. For example, the essential characteristics corresponding to the action of concepts are subclasses of the Action class. Owning an essential characteristic for a concept (class) is represented as a restriction of an object property whose range is the class associated with the essential characteristic. For example, owing the essential characteristic/control/will be translated into the restriction of the “has action” object property: “has action” some Control.

However, the essential characteristics of type are special in our ontology. They were defined as the subclasses of the Concept class. In other words, the essential characteristics of types were translated as classes, which is an “is-a” relationship between owning essential characteristics for a concept (class) and essential characteristics of types (classes).

6.1. Class

The Lunyu ontology includes 100 classes, 8 object properties, 204 logical axioms. Although The Analects of Confucius includes different concepts, this paper focused on the concepts related to morality.

Let us consider the concept denoted by the term “礼” (ritual), denotes the set of essential characteristics: {/external/,/hierarchy/,/normalize/,/behavior system/}, which was defined in OWL as following (Figure 3):

Li ≡ BehaviorSystem ⊓ ∃hasAction. Normalize ⊓ ∃hasRange. External ⊓ ∃hasHierarchy. Hierarchy.

6.2. Properties

There were eight object properties, which included hasResource, identityIs, has Action, hasBinding, hasHierarchy, hasPerformance, hasRange, isRelatedTo (

Table 2. The object properties.

Object Properties	Domain	Range
hasResource	Concept	Resource
identityIs	Concept	Identity
hasAction	Concept	Action
hasBinding	Concept	Binding
hasHierarchy	Concept	Rank
hasPerformance	Concept	Performance
hasRange	Concept	Range
isRelatedTo	Concept	Person

). There are no data properties because the concept is an abstract concept, whose attributes are defined as the essential characteristics. Thus, we did not need data properties to describe the concept in Lunyu ontology.

6.3. Annotation

In Lunyu ontology, we reuse the RDFS, DC, OTContainer, and SKOS vocabularies to express metadata information. For example, dc:publisher, dc:license, dc:creator, skos:prefLabel, skos:altLabel, skos:definition, rdfs:comment. OTContainer:conceptName.

7. Evaluation

Evaluation is the last stage of building ontology, whose goal is “to assess the quality and correctness of the obtained ontology” [40]. In this paper, we used the OOPS! [41], OntoMetrics [42] and queried the ontology against the Competency Questions defined in chapter 3.

When we submitted Lunyu ontology to OOPS!, it only detected minor pitfalls (e.g., P08 “Missing annotations”, P13 “Inverse relationship not explicitly declared”, P36 “URI contains file extension”). The minor pitfalls are not a problem.

When we submitted Lunyu ontology to OntoMetrics,

Table 3. The Lunyu ontology advanced metrics.

Metric	Value
Attribute richness	0.0
Inheritance richness	1.84
Relationship richness	0.05641
Class/relation ratio	0.512821
Average population	0
Class richness	0

shows the result of schema metrics and knowledge base metrics about ontology clarity and conciseness.

Most of the scores are low, that is due to:

• The implementation of essential characteristics in Description Logics. In our work, the essential characteristics were translated as the classes without any attributes (attribute richness). Thus, the value of the attribute richness is a “0”.
• The goal of Lunyu ontology is the classification and definition of concepts; they did not represent the relationships between concepts in The Analects of Confucius and concepts defined in other literature (relationship richness, class/relation ration).
• The Lunyu ontology is about the abstraction concept, which does not correspond to the physical entity in the real world. Thus, we could not define any individuals in Lunyu ontology (average population, class richness). Depending on the calculation formula of the average population and class richness in OntoMetrics, the values of the average population and class richness are “0”.

Evaluation of criteria strongly depends on the aims of the ontology and the choices made to regard to its implementation: “a good ontology does not perform equally well with regards to all criteria” [43].

We should note that, about our objectives of classification and terminology, the Lunyu ontology well covers the domain concept (classification), and each concept is clearly defined as a unique combination of essential characteristics (terminology). It should provide the domain knowledge to a bilingual terminology e-dictionary (http://dh.ketrc.com/lunyu/index.html, accessed on 25 November 2021). Thus, the low values of OntoMetrics do not affect the construction of a bilingual terminology e-dictionary.

The last validation concerns the answers to the Competency Questions. All of them are satisfied. The three competency questions were translated into SPARQL as following. Figure 4 is the returned results.

• What are the Chinese terms of concept and their equivalents in English?
  • PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
  • PREFIX skos: <http://www.w3.org/2004/02/skos/core#>
  • PREFIX owl: <http://www.w3.org/2002/07/owl#>
  • PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntaxns#>
  • PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
  • PREFIX otc:<http://www.dh.ketrc.com/otcontainer/data/otc.owl#>
  • PREFIX lunyu:<http://www.dh.ketrc.com/ChineseCulture/data/lunyu.owl#>
  • PREFIX skos:<http://www.w3.org/2004/02/skos/skos#>
  • SELECT ?name_en ?name_zh
  • WHERE {
    • ?x rdfs:subClassOf* skos:Concept.
    • ?x skos:prefLabel ?name_en.
    • ?x skos:prefLabel ?name_zh.
  • FILTER(Lang(?name_en) = ‘en’)
  • FILTER(Lang(?name_zh) = ‘zh’)
• What are the concepts that express character?
  • PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
  • PREFIX otc:<http://www.dh.ketrc.com/otcontainer/data/otc.owl#>
  • PREFIX lunyu:<http://www.dh.ketrc.com/ChineseCulture/data/lunyu.owl#>
  • PREFIX skos:<http://www.w3.org/2004/02/skos/skos#>
  • SELECT DISTINCT ?Name
  • WHERE {
    • ?x rdfs:subClassOf lunyu:Character.
    • ?x skos:prefLabel ?Name
    • }
  • ORDER BY ?Name
• What are the definitions of the “filiality” in English and Chinese?
  • PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
  • PREFIX skos: <http://www.w3.org/2004/02/skos/core#>
  • PREFIX owl: <http://www.w3.org/2002/07/owl#>
  • PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntaxns#>
  • PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
  • PREFIX otc:<http://www.dh.ketrc.com/otcontainer/data/otc.owl#>
  • PREFIX lunyu:<http://www.dh.ketrc.com/ChineseCulture/data/lunyu.owl#>
  • PREFIX skos:<http://www.w3.org/2004/02/skos/skos#>
  • SELECT DISTINCT ?definition_en ?definition_zh
  • WHERE {
    • ?x skos:prefLabel “trust”@en.
    • ?x skos:definition ?definition_zh.
    • ?x skos:definition ?definition_en.
  • FILTER(Lang(?definition_zh)=‘zh’)
  • FILTER(Lang(?definition_en)=‘en’)

In our research, concepts referring to morality in The Analects of Confucius are translated into a clearly-defined format, and then a bilingual terminology e-dictionary is developed. This work will provide practical value on at least two levels. For anyone who wants to learn the thought in The Analects of Confucius, they can avoid the difficulties caused by the chaotic organization of the text and easily gain access to the connotation of each concept and the potential relationship between them. Especially for those who are not familiar with ancient and Confucianism, a clear definition with modern expression can help them quickly understand the concept. Conversely, the bilingual terminology e-dictionary could support cross-language retrieval such that even people who do not speak Chinese can accurately acquire knowledge about Confucianism. With all the work we have performed, we hope the Lunyu ontology, with the e-dictionary based on it, will be helpful to anyone interested in Confucianism.

8. Conclusions

Our publication proposed Lunyu ontology to formally represent concepts concerning morality in The Analects of Confucius. In this research, we explored how to develop a class hierarchy from discrete and abstract ideas. We also adopted a term-and-characteristic guided methodology derived from taking into account the ISO principles on Terminology, whereby “a term is a verbal designation of a concept” and “a concept is a unique combination of (essential) characteristics”.

The problem of identifying essential characteristics is a new and central phase of our methodology. Usually, the essential characteristics of the concrete concept are extracted from differences between objects, but for abstract concept, they come from the differences between meanings in this study. This method may provide reference for developing ontology based on abstract concepts.

Lunyu ontology was published at: http://www.dh.ketrc.com/ChineseCulture/data/lunyu.owl, accessed on 25 November 2021.

In our future work, we would like to enrich our work in two aspects: (i) expand the scope of concepts to other topics, such as politics, learning and life; (ii) develop a multilingual e-dictionary for our ontology. At present, we limited the range of concepts to the ethics topic, which reflects only a single slice of Confucianism. It will make our ontology more valuable to enrich the content of it. The multilingual e-dictionary will provide a user-friendly entry point to access our outcome. Another issue that deserves to be researched is that is there a common method or process or workflow to identify essential characteristics of abstract concepts. It will be a methodological discussion.