An Ontology of Chinese Ceramic Vases

Tong Wei

1,2 a

, Christophe Roche

1,2 b

, Maria Papadopoulou

1,2 c

and Yangli Jia

2 d

Condillac Research Group of LISTIC Lab, University Savoie Mont-Blanc, Rue du Lac Majeur, Le Bourget du Lac, France

The School of Computer Science, Liaocheng University, HuNan Road 1, Liaocheng City, China

Keywords: Cultural Heritage, Ontology, Ontology Building, Semantic Web, Chinese Ceramic Vases.

Abstract: Extensive collections of Chinese ceramic vases are housed in museums throughout China. They could serve

as rich sources of data for historical research. Although some data sources have been digitized, the vision of

heritage institutions is not only to display objects and simple descriptions (drawn from metadata) but also to

allow for understanding relationships between objects (created by semantically interrelated metadata). The

key to achieving this goal is to utilize the technologies of the Semantic Web, whose core is Ontology. The

focus of this paper is to describe the construction of the TAO CI (“ceramics” in Chinese) ontology of the

domain of ceramic vases of the Ming (1368-1644) and Qing (1644-1911) dynasties. The theoretical and

methodological approach adopted to construct the TAO CI ontology is term-and-characteristic guided, i.e., it

relies on a morphological analysis of the Chinese terms used in the domain, and respects the ISO principles

on Terminology (ISO 1087 and 704), according to which concepts are defined by means of essential

characteristics. The research presented in this article aims to publish the resulting structured data on the

Semantic Web for the use of anybody interested, including museums hosting collections of these vessels, and

to enrich existing methodologies on domain ontology building. To our knowledge, there are no comprehensive

ontologies for Chinese ceramic vases. TAO CI ontology remedies this gap and provides a reference for

ontology building in other domains of Chinese cultural heritage. The tool used is Protégé. The TAO CI

ontology is open access here: http://www.dh.ketrc.com/otcontainer/data/OTContainer.owl.

1 INTRODUCTION

China has a rich cultural heritage and has

concentrated on producing digital data under the first

wave of digitization. This is also true of the

knowledge domain of Chinese ceramic vessels. The

domain of Chinese ceramic vessels is rich, yet it lacks

knowledge representation models (ontologies) to

capture Chinese pottery concepts, express them in

Semantic Web compatible interchange formats, and

make them shareable and linkable to other data. As

there is no ontology in the domain of Chinese

ceramics providing the semantics of relevant data,

most heritage institutions in China have not yet

published cultural heritage data on the Semantic Web.

Furthermore, every institution accumulates its data in

its own traditional database system rather than linking

https://orcid.org/0000-0002-4159-6248

https://orcid.org/0000-0002-0756-0559

https://orcid.org/0000-0001-6451-8712

https://orcid.org/0000-0002-6849-1059

data through an open data policy. To bridge this gap,

this paper proposes the TAO CI ("ceramics" in

Chinese) ontology. In compliance to the ethos of

reuse recommended by the W3C, the TAO CI

ontology relates to existing ontologies and thesauri,

such as CIDOC CRM (Cidoc, 2003), EDM (Doerr et

al., 2010), and AAT (Soergel, 1995). First, the TAO

CI ontology aims to provide an important reference

for the publication of other cultural heritage

ontologies and to be conducive to more and more

Chinese heritage institutions publishing open cultural

heritage data and linking them. Second, the

theoretical and methodological approach adopted in

the construction of the TAO CI ontology is term-and-

characteristic guided, i.e. it adopts the ISO principles

on Terminology (ISO 1087 and 704), which define

concepts on the basis of their essential characteristics

for defining concepts. Last but not least, this work

Wei, T., Roche, C., Papadopoulou, M. and Jia, Y.

An Ontology of Chinese Ceramic Vases.

DOI: 10.5220/0010110600530063

In Proceedings of the 12th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2020) - Volume 2: KEOD, pages 53-63

ISBN: 978-989-758-474-9

aims at enriching existing methodologies of building

domain ontologies and suggests that taking into

account a term-and-characteristic guided approach,

makes ontology engineering less dependent on formal

languages and description logics as required

background.

The rest of the paper is organized as follows:

Section 2 describes the domain, section 3 introduces

the objectives, section 4 is dedicated to the state of the

art. Section 5 is the main section dedicated to our

contribution to ontology building methodology

relying on a morphological analysis of Chinese terms

and on the ISO principles on Terminology. Section 6

presents the TAO CI ontology in Protégé, and the last

section presents the evaluation of the TAO CI

ontology.

2 DOMAIN OF RESEARCH

2.1 Ming and Qing Dynasties

Chinese history goes back about 5,000 years. Chinese

ceramic vessels are among the most iconic objects of

Chinese cultural heritage. Changes in pottery styles

reflect the change in dynasties. In this paper, the focus

is on Chinese porcelain vases defined as “clay vessels

fired at high temperature used for decoration” of the

Ming and Qing dynasties (冯先铭, 2002).

Ming Dynasty (1368-1644) ceramics were

famous for the boldness of their form and decoration

and the varieties of design.

Already from the time of

the Tang dynasty and the Song dynasty, there were

many famous kilns and many different types of

ceramic vessels. From the beginning of the Ming

dynasty, the Jingdezhen kiln gradually became the

most important production place: ceramic vessels of

the Jingdezhen kiln represented the highest quality at

the time. Between 1350 and 1750, Jingdezhen was a

center of production for nearly all of the world’s

porcelain.

Qing dynasty (1644-1911) porcelain was famous

for its polychrome decorations, delicately painted

landscapes, and bird and flower as well as

multicolored enamel designs. The peak of Chinese

ceramics production took place in the reigns of

Kangxi (1661-1722), Yongzheng (1722-1735), and

Qianlong (1735-1796) during which improvement

http://factsanddetails.com/china/cat7/sub40/item258.html

#chapter-11

https://www.comuseum.com/ceramics/qing/

https://www.dpm.org.cn/Home.html

http://www.chnmuseum.cn/

was seen in almost all ceramic types, including the

blue and white wares, polychrome wares,

monochrome wares, etc.

During the Qing dynasty,

potters began using bright colors to adorn plates and

vases with meticulously painted scenes.

2.2 Collection of Vases

The first step of our work was to select the set of vases

to study. The set had to be enough representative of

the richness of the domain without being too big since

the main and first goal was defining the ontology

rather than populating it. In China, many museums

publish much information about ceramic vessels on

their websites. One hundred forty-nine objects were

selected from different museums in China. Ninety-

seven objects come from the Palace Museum

that

has the most important collection of ceramics.

Twenty-two objects come from the National Museum

of China

, twenty-four objects from the Guangdong

Museum

. Four objects come from the Shanghai

Museum

and two objects from the Capital

Museum

. For the selection of objects, we have

adopted the following three criteria. The two first

concern the selection of the museum, which had to

fulfil the following conditions: first, the collection of

ceramics had to be recognized as a reference in

ceramic vessels in China; second, the information

about the collection should be publicly available and

precise enough for the building of an ontology. The

third principle was to select objects as different as

possible, i.e., of different types according to their

shape, the technique of making, decoration, etc.

3 OBJECTIVES

The TAO ontology has two aims. The first one is to

build a knowledge representation of Chinese ceramic

of the Ming and Qing dynasties in the form of an open

ontology. The second one is to provide a bilingual

(Chinese-English) e-dictionary of ceramics vases.

The competency questions (Ren et al., 2014) that

were used to specify the requirements of the ontology

are shown in Table 1.

http://www.gdmuseum.com/

https://www.shanghaimuseum.net/museum/frontend/

http://www.capitalmuseum.org.cn/

KEOD 2020 - 12th International Conference on Knowledge Engineering and Ontology Development

4 STATE OF THE ART

The state of the art presented here features work done

in an effort to produce interoperable vocabularies for

the expression of Cultural Heritage data. It includes

W3C languages, semantic data models, thesaurus and

ontology resources, and ontology building

methodologies.

Opening cultural heritage on the web relies on

W3C standards such as OWL

(Web Ontology

Language), a Semantic Web language designed to

represent rich and complex knowledge about things,

groups of things, and relations between things,

SKOS

(Simple Knowledge Organization System), a

data model for sharing and linking knowledge

organization systems on the Web. SKOS can be used

to capture much of the semantics of existing thesaurus

of museums and other memory institutions thesauri.

Let us also quote DC (Dublin Core), a metadata

schema based on 15 essential properties to describe

online and physical resources

Semantic Data Models for the Cultural Heritage

domain have to be taken into account. In particular,

CIDOC-CRM, a meta-ontology for the representation

of concepts for the use of museum and cultural

heritage specialists (Cidoc, 2003). It provides a

semantic framework to building a mapping between

different cultural heritage resources reducing their

heterogeneity (Doerr, 2003). Our work not only aims

to build an ontology for museum publishing open

museum data, but also aims to build a multilingual

terminological knowledge base. From a

terminological point of view, we need to build a more

‘granular’ ontology for knowledge representation of

Chinese ceramic vases. Let us also quote EDM, the

common data model that was built in order to

harmonize data from different providers of

Europeana (Doerr et al., 2010). It is used for the

representation of concepts in the cultural heritage

domain. It is not a fixed schema that dictates the way

of representing data, but rather a conceptual

framework (or ontology) to which more specific

models can be attached, and interoperability between

them enhanced.

As far as ontological resources that the TAO CI

project can benefit, let us quote AAT (The Art &

Architecture Thesaurus), a structured resource that

can be used to improve access to information about

art, architecture, and other material culture through

rich metadata and links, hoping to provide (along with

other Getty vocabularies) a powerful conduit for

research and discovery in digital art history and

related disciplines

(Soergel, 1995). The AAT

comprises over 250,000 terms on architectural

history, styles, and techniques. Our ontology has been

linked with AAT in order to provide more

information for our terms in the ontology.

Kerameikos

is a “collaborative project dedicated to

defining the intellectual concepts of pottery following

the tenets of linked open data and the formulation of

an ontology for representing and sharing ceramic data

across disparate data systems.” (Gruber & Smith,

2014). Let us also quote Ontoceramic, which is an

OWL ontology for ceramics classification (Cantone

et al., 2015). Lekythos

is an another project that

aims at representing concepts in the domain of

ancient Greek pottery, but having natural language

terms in the domain as its starting point.

According to (Ushold, 1998), “An [explicit]

ontology may take a variety of forms, but necessarily

it will include a vocabulary of terms and some

specification of their meaning (i.e., definitions).” For

domain experts, identifying and defining concepts in

ontology also presents a challenge for which ontology

building methodology can bring useful aids.

Ontology building methods can be based on objective

criteria, e.g., clarity, coherence, extensibility, etc.

(Gruber, 1995), software engineering methods

(Fernández-López, 1999), text-based construction

(Zouaq & Nkambou, 2009), modular design approach

(Özacar et al., 2011), ontological engineering

(Suárez-Figueroa et al., 2012), unsupervised domain

ontology learning method (Venu et al., 2016) , based

on Formal Concept Analysis (Nong et al., 2019), etc.

Let us quote some methodologies focusing on the

stages which compose them. METHONTOLOGY

(Fernández-López et al., 1997) includes seven

stages: specification, knowledge acquisition,

conceptualization, integration, implementation,

evaluation, and documentation. On-To-Knowledge

Methodology (Sure et al., 2004) includes the

following phases: feasibility study, kick-off,

refinement, evaluation, and application & evolution.

NeOn methodology (Suárez-Figueroa et al., 2015)

provides nine scenarios for developing ontologies.

https://www.w3.org/TR/2004/REC-owl-features-20040210/

https://www.w3.org/TR/skos-reference/#notes

https://dublincore.org/schemas/ Schemas are machine-

processable specifications that define the structure and syntax

of metadata specifications in a formal schema language

https://www.getty.edu/research/tools/vocabularies/aat/

about.html

http://kerameikos.org/

http://o4dh.com/lekythos

An Ontology of Chinese Ceramic Vases

Table 1: The competency questions.

Competency Question

Class

Relation

What are the different types of vase?

Vase

vase-type is-a Vase

What material is vase the made of?

Vase, material

Vase is made of Material

What is the color and glaze of the vase?

Vase, Glaze-color,

Vase glaze-color is

Which dynasty is the vase?

Vase, Dynasty

Vase hasDynasty Dynasty

Which emperor is the vase?

Vase, Emperor

Vase hasEmperor Emperor

What are the Chinese and English terms of

the vases

Vase

Vase label string

What are the components of the vase?

Vase, Component

Vase hasComponent Component

What is the function of the vase?

Vessel, Function

Vessel hasFunction Function

Which dynasty does the emperor belong

to?

Emperor, Dynasty

Emperor belongTo dynasty

Where is the vase x collected in?

Vase x

Vase x is Collected In string

Which kiln is the vase x produced?

Vase x

Vase x is produced in string

What is vase x decorated with?

Vase x

Vase x is decorated by string

What is the image of vase x?

Vase x

Vase x image string

5 TAO CI METHODOLOGY

Ontology building follows a lifecycle made up of

several stages (Fernández-López et al., 1997). Some

of them have to be specialized, and others have to be

introduced to take into account the specificities of the

domain. The theory of concept underpinning the

ontology can also strongly impact the building

methodology. 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

” (ISO 1087, ISO 704), we

were led to adopt a “term-and-characteristic” guided

methodology derived from works carried out in

Digital Humanities (Roche & Papadopoulou, 2019).

Identifying essential characteristics becomes the

main goal to achieve.

The problem of identifying essential

characteristics, is a new and central phase of our

methodology. This phase is aimed both at identifying

differences between objects (vases with neck versus

vases without neck) and on a morphological analysis

of Chinese terms whose characters carry meaning in

relation to the denoted objects, e.g. the term "清德化

窑白釉堆雕花卉瓷碗" where the first character (清)

which represents the Qing dynasty and the last one

the type of vase (碗).

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(Guarino & Guizzardi, 2006) and to rigid properties in

The term-and-characteristic guided methodology

includes seven steps. Each of them aims 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.

5.1 Identifying Essential

Characteristics

There are two approaches to identifying essential

characteristics. The first one relies on identifying

differences between objects, e.g., in their structure:

vase with or without neck. The second one is based

on a morphological analysis of Chinese terms whose

characters directly express knowledge about the

denoted objects. We will adopt the following

notation: essential characteristics will be enclosed

between slashes, e.g. /garlic shape mouth/.

5.1.1 Differences between Objects

Identifying differences between objects is a useful

means towards identifying essential characteristics.

The differences can be functional (e.g., vase for

the OntoClean method (Guarino & Welty, 2004). Unlike

essential characteristics, which define the concept,

“descriptive characteristics” own values which describe the

current state of an object, e.g. weight, colour, etc.

KEOD 2020 - 12th International Conference on Knowledge Engineering and Ontology Development

decoration, vase for storing), material (in clay, in

bronze), structural (with or without foot). Thus, one

can rely on the part-of relationship between a whole

and its parts to understand the concept the object

belongs to (Gerstl & Pribbenow, 1996). The presence

or the absence of a component can be interpreted as

essential characteristics. For example, a Chinese

ceramic vase has a lid, a mouth, a neck, handle,

shoulder, belly, and foot (Figure 1).

Figure 1: Parts of a vase.

From the handle point of view, vases can be split into

vases with handles and vases without handles

corresponding to the essential characteristic /with

handle/ and /without handle/. The type of vases with

handles can be itself specialised according to the

different shapes of handles defining as many as

corresponding essential characteristics: /dragon-mask

handle/, /dragon-shaped handle/, /elephant-shaped

handle/, /fish-shaped handle/, /halberd shaped

handle/, /phoenix shaped handle/, /pierced handle/,

/ribbon shaped handle/, and /Ru-Yi handle/ (Figure

2). These characteristics are exclusive to each other.

5.1.2 Morphological Analysis of Chinese

Terms

The characters which compose the Chinese terms of

vase carry important information about the nature and

the description of the objects denoted by the terms.

Terms are composed of a set of characters of which

the last one corresponds to the type of vase and the

others, called modifiers, precise the type. For

example, the Nanjing museum adopts the following

order of modifiers for naming Tibetan ceramic (霍华,

1989): dynasty + kiln + glaze + colour + decoration +

shape + texture + type. The information conveyed by

the modifiers expresses knowledge of different

nature, either essential, such as shape, material, and

type, or descriptive, like glaze and color. For

example, the term “清雍正粉青釉凸花如意耳蒜

头瓷瓶” (for convenience of non−Chinese speaker,

we put spaces between modifiers) conveys the

descriptive characteristics of dynasty (“ 清 ” Qing

dynasty), emperor (“雍正” Yongzheng mark), glaze-

color (“粉青釉” powder blue glaze), and decoration (“

凸花” designed with flowers). It also conveys the

essential characteristics of handle (“如意耳” Ru-Yi

handle), shape (“蒜头” garlic-like head), material (“

瓷” porcelain), and type (“瓶” vase). The English

translation of the Chinese ceramic terms used by the

Nanjing museum does not follow the Chinese order

of modifiers, but the following order: glaze + colour

+ shape + texture + type + decoration + period + kiln.

Thus, the previous term “清雍正粉青釉凸花如意

耳蒜头瓷瓶” is translated as: “powder blue glaze

garlic porcelain vase designed with flowers and Ru-

Yi handles, the Yongzheng mark of Qing dynasty”.

This object belongs to the type of Garlic-head Vase

(“蒜头瓶”).

5.2 Combining Essential Characteristic

From the ISO point of view on terminology, a concept

is defined as a unique combination of essential

characteristics (ISO 1087). Nevertheless, not any

combination of essential characteristics defines a

meaningful concept from the expert’s point of view.

Figure 2: The essential characteristics of the analysis axis of the handle shape.

An Ontology of Chinese Ceramic Vases

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 organisational 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 Chinese term “蒜头瓶”, “garlic vase” in English,

denotes the following set of essential characteristics

{/vase/, /one mouth/, /garlic shape mouth/, /ring

foot/}. Based on this formal definition, the definition

in natural language is then: “Vase with a garlic shape

mouth with a ring foot”. We notice that the characters

“圈足” (“ring foot”) do not appear in the name of the

concept, an ellipsis owed to the fact that both types of

garlic vase (Garlic vase I and Garlic vase II) have a

ring foot.

5.3 Implementation

The following will present the implementation in

Protégé of our ontology building approach. Concepts

are represented as named classes in Protégé, and

objects as individuals. Terms are represented as labels

(skos:prefLabel, skos:altLabel, skos:definition).

Relations, e.g. ‘hasFunction’, ‘hasComponent’,

‘isMadeOf’, are represented as object properties. For

example, the object property ‘isMadeOf’ whose

domain is the Vessel class and range is the Material

Class, and the object property ‘hasComponent’ whose

domain is the Vessel class and range is the

Component class. Let us note that among the different

types of ‘part-of’ relationships, only the

‘Component/Integral Object’ relationship has been

taken into account (Winston et al., 1987). Descriptive

characteristics are attributes whose values describe

the current state of an object. The TAO CI ontology

includes 10 descriptive characteristics: dynasty,

emperor, kiln, color and glaze, height, diameter of

mouth, diameter of foot, decoration, museum, and

image. The descriptive characteristics are represented

either as data properties, if their value is a data literal,

or as object properties and classes, if the value is an

individual. For example, the decoration characteristic

is represented by the data property ‘isDecoratedBy’

whose domain is the Vessel class and the range is the

String data type. The dynasty to which a vase belongs

is represented by the object property ‘hasDynasty’

The ontology file is published on

http://www.dh.ketrc.com/otcontainer/data/OTContainer.o

whose domain is the Vessel class and the range is the

Dynasty class.

Implementing essential characteristics is a slightly

more complex process. Since essential characteristics

correspond to rigid predicates (Guarino & Guizzardi,

2006) they cannot be directly expressed into

Description Logic. Essential characteristics are

expressed as classes. Thus, essential characteristics

corresponding to parts of vase are subclasses of the

Component class: Lid class, Mouth class, Neck class,

Handle, Shoulder, Belly, Foot class, etc. Some being

themselves specialized into subclasses according to

the different types of parts: LongNeck and ShortNeck

subclasses of the Neck class, RingFoot and

SquareFoot subclasses of the Foot class, etc. Essential

characteristics corresponding to functions, such as

/for decoration/ are subclasses of Function class, etc.

Owning an essential characteristic for a concept

(class) is represented as a restriction of an object

property whose range is the class associated to the

essential characteristic. It means that the class

(concept) is a subclass of the anonymous class

defined by the restriction (see figure 5). For example,

owning the essential characteristic /long neck/ will be

translated into the following restriction of the

‘has_component’ object property: ‘has_component’

some LongNeck. The following restriction of the

‘has_function’ object property: ‘has_function’ some

FunctionForDecoration expresses the owning of the

essential characteristic /for decoration/.

Protégé relies on the open-world assumption,

which means that what is not known to be true is

unknown. In this vein, it is necessary to express

information corresponding to essential characteristics

such as /without handle/, /without lid/, /without foot/,

etc. The object property restriction allows to do that.

For example, owning the /without handle/ essential

characteristic will be translated into the following

object property restriction not (‘has_component’

some Handle).

6 TAO CI ONTOLOGY

The TAO CI ontology

contains 165 classes, 11

object properties, 8 data properties, 132 individuals,

and 3124 axioms. It is mapped with CIDOC CRM

(E4_Period, E21_Person, E22_Man-Made_Object,

E57_Material) and AAT.

http://www.cidoc-crm.org/cidoc-crm/

KEOD 2020 - 12th International Conference on Knowledge Engineering and Ontology Development

6.1 Class

Although our work focuses on Chinese ceramic

vases, The TAO CI ontology also includes other types

of vessels, such as Box, Bowl, Jar (Figure 3). The

Vase class is the core of our work. It includes 41

subclasses representing 41 different types of Chinese

ceramic vases (Figure 5).

Let us consider the following example. The

concept denoted by the term “arrow vase I”, whose

definition in natural language is “vase with a square

mouth, long neck, slanting shoulder, pierced handle,

bulge belly, and square foot”, is defined by the set of

essential characteristics: {/vase/, /square mouth/,

/long neck/, /slanting shoulder/, /pierced handle/,

/bulge belly/, /square foot/}. It is represented by the

ArrowVase_I class defined in OWL as subclass

(rdfs:subClassOf) of:

- ArrowVase

- hasComponent some SquareMouth

- hasComponent some LongNeck

- hasComponent some SlantingShoulder

- hasComponent some PiercedHandle

- hasComponent some BulgeBelly

- hasComponent some SquareFoot

6.2 Property

Object properties include belongTo

(domain:Emperor, range: Dynasty), hasFunction

(domain:Vessel, range:Function), hasComponent

(domain:Vessel, range: Component), etc.

Data properties include isDecoratedBy,

diameterOfFoot, height, isProducedIn,

diameterOfMouth, mouthNumber, etc.

6.3 Annotations

Annotations allows to enrich the description of the

ontology and thus facilitate its understanding and

reuse. The RDFS, DC, and SKOS vocabularies are

used to express metadata and the linguistic dimension

associated to a concept (dc:publisher, dc:license,

dc:creator, skos:prefLabel, skos:altLabel,

skos:definition, rdfs:comment) as well as to express

linking and mapping to external resources

(rdfs:seeAlso, skos:broadMatch, skos:exactMatch).

For example, the individual arrow vase 001 is

described as follows (Figure 4).

Figure 3: The Vessel class of TAO CI ontology.

Figure 4: The individual arrow vase 001.

An Ontology of Chinese Ceramic Vases

Figure 5: The Vase classes.

7 EVALUATION

The last stage of ontology building is ontology

evaluation whose main goal is “to assess the quality

and correctness of the obtained ontology” (Sabou &

Fernandez, 2012). We used two online platforms and

queried the ontology against the Competency

Questions defined in chapter 3.

The TAO CI ontology was submitted to OOPS!,

an online tool to detect some of the most common

pitfalls appearing when developing ontologies

(Poveda-Villalón et al., 2014). OOPS! has detected

only minor pitfalls for the TAO CI ontology (e.g. P08

“Missing annotations”, P13 “Inverse relationships not

explicitly declared”).

https://ontometrics.informatik.uni-rostock.de/wiki/in

dex.php/Schema_Metrics

The TAO CI ontology was also submitted to

OntoMetrics, an online platform to calculate more

advanced ontology metrics (Lantow, 2016). The table

2 shows some schema metrics and knowledge base

metrics results

in relation to ontology clarity and

conciseness (Denny, 2010).

Table 2: TAO CI advanced metrics.

Metric

Value

Attribute richness

0.048485

Inheritance richness

2.715152

Relationship richness

0.334324

Class/Relation ratio

0.245171

Average population

0.8

Class richness

0.321212

KEOD 2020 - 12th International Conference on Knowledge Engineering and Ontology Development

Most of the scores are very low. That is due to:

 the implementation of essential characteristics in

Description Logics. Essential characteristics are

translated as classes without any attributes

(attribute richness);

 the main goal of the TAO CI ontology is the

classification of vases; neither to represent

relationships between vases and other artefacts

(relationship richness, class/relation ratio), nor to

populate the ontology with individuals (average

population, class richness).

Evaluation of criteria strongly depends on the aims of

the ontology and the choices made in to regard to its

implementation: “a good ontology does not perform

equally well with regards to all criteria” (Denny,

2009).

Let us also note that, in regard to our objectives of

classification and terminology, the TAO CI ontology

well covers the domain in the sense that each

individual clearly falls into a concept (classification),

and each concept is clearly defined as a unique

combination of essential characteristics

(terminology).

Figure 6: The competency questions expressed in

SPARQL.

The last validation concerns the answers to the

Competency Questions. All of them are satisfied.

Figure 6 presents 3 competency questions translated

into SPARQL and figure 7 the returned results.

Q1: What are the types of a vase?

Q2: What are the Chinese terms and the English

terms for subclasses of the class Vase?

Q3: What are the Chinese terms, the English

terms, and the images of individuals?

8 CONCLUSION

The objectives of the TAO CI project are to provide,

first, a knowledge representation of Chinese ceramic

of the Ming and Qing dynasties in the form of a

domain ontology to be published in the LOD, and,

second, a bilingual (Chinese-English) e-dictionary of

vases based on the domain ontology. We 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”.

In order to identify essential characteristics, the

core task of our ontology building methodology, we

proposed to combine two approaches. The first one is

object-oriented. It consists of identifying differences

between objects, either structural or functional. The

second one is language-dependent. It relies on a

morphological analysis of Chinese terms whose

characters convey a lot of useful semantic

information about vases.

The implementation of the TAO CI ontology in

Protégé raised the problem of how to implement the

notion of essential characteristic. Since essential

characteristics cannot be directly expressed in

Description Logics, we proposed to represent them as

classes. Owning an essential characteristic for a

concept (class) is then represented as a restriction of

an object property whose range is the class associated

to the essential characteristic. This means that the

concept (class) is a subclass of the anonymous class

defined by the restriction.

The result is an open RDF/OWL ontology

accessible from the web site

http://www.dh.ketrc.com/. A bilingual (Chinese-

English) dictionary based on the TAO CI ontology is

also accessible from the same web site.

Further work is to be carried out in two different

directions. The first one consists in enriching the

linguistic dimension. Currently terms are reduced to

labels on class, some vocabularies such OntoLex-

Lemon would allow to represent the linguistic

An Ontology of Chinese Ceramic Vases

Figure 7: An excerpt of the results.

dimension. The second one aims to complete the

ontology by taking into account additional types of

ceramic vessels.

ACKNOWLEDGEMENTS

This work has received generous funding by the

China Scholarship Council (CSC) in the framework

of a PhD program between Liaocheng University and

University Savoie Mont-Blanc.

REFERENCES

Buitelaar, P., Cimiano, P., & Magnini, B. (2005). Ontology

learning from text: methods, evaluation and

applications (Vol. 123). IOS press.

Cidoc, C. (2003). The CIDOC Conceptual Reference

Model.

Corcho, O., Fernández-López, M., & Gómez-Pérez, A.

(2003). Methodologies, tools and languages for

building ontologies. Where is their meeting point? Data

& Knowledge Engineering, 46(1), 41–64.

Cantone, D., Asmundo, M. N., Santamaria, D. F., and

Trapani, F. (2015). Ontoceramic: an owl ontology for

ceramics classification. In CILC, pages 122–127.

Denny, V. (2010). Ontology Evaluation. PhD thesis,

Karlsruhe Institute of Technology.

Denny, V. (2009). Ontology evaluation. In Handbook on

ontologies (pp. 293-313). Springer, Berlin, Heidelberg.

Doerr, M. (2003). The CIDOC Conceptual Reference

Module: An Ontological Approach to Semantic

Interoperability of Metadata. AI Magazine, 24(3), 75–

75. https://doi.org/10.1609/aimag.v24i3.1720

Doerr, M., Gradmann, S., Hennicke, S., Isaac, A., Meghini,

C., & Van de Sompel, H. (2010). The europeana data

model (edm). World Library and Information

Congress: 76th IFLA General Conference and

Assembly, 10–15.

Fernández-López, M. (1999). Overview of methodologies

for building ontologies. IJCAI99 Workshop on

Ontologies and Problem-Solving Methods: Lessons

Learned and Future Trends, 430.

Fernández-López, M., Gómez-Pérez, A., & Juristo, N.

(1997). Methontology: from ontological art towards

ontological engineering.

Gerstl, P., & Pribbenow, S. (1996). A conceptual theory of

part-whole relations and its applications. Data &

Knowledge Engineering, 20(3), 305–322.

Gruber, E., & Smith, T. J. (2014). Linked open greek

pottery. 42nd Annual Conference on Computer

Applications and Quantitative Methods in

Archaeology, 205–214.

Gruber, T. R. (1995). Toward principles for the design of

ontologies used for knowledge sharing? International

Journal of Human-Computer Studies, 43(5–6), 907–

928.

Guarino, N., & Guizzardi, G. (2006). In the defense of

ontological foundations for conceptual modeling.

Scandinavian Journal of Information Systems, 18(1), 1.

Guarino, N., & Welty, C. A. (2004). An overview of

OntoClean. In Handbook on ontologies (pp. 151–171).

Springer.

Lantow, B. (2016). OntoMetrics: Putting Metrics into Use

for Ontology Evaluation. KEOD, 186–191.

Nong, J., Liang, C., and Xie, P. (2019). Research on

government domain ontology construction method

based on formal concept analysis. In 2019 IEEE

International Conferences on Ubiquitous Computing &

Communications (IUCC) and Data Science and

Computational Intelligence (DSCI) and Smart

Computing, Networking and Services (SmartCNS),

pages 230–235. IEEE.

Özacar, T., Öztürk, Ö., & Ünalır, M. O. (2011).

ANEMONE: An environment for modular ontology

KEOD 2020 - 12th International Conference on Knowledge Engineering and Ontology Development

development. Data & Knowledge Engineering, 70(6),

504–526.

Poveda-Villalón, M., Gómez-Pérez, A., and Suárez-

Figueroa, M.C. (2014). ‘Oops! (ontology pitfall

scanner!): An on-line tool for ontology evaluation’, Int.

J. Semantic Web Inf. Syst. IJSWIS, vol. 10, no. 2, pp. 7–

34.

Ren, Y., Parvizi, A., Mellish, C., Pan, J. Z., Van Deemter,

K., & Stevens, R. (2014). Towards competency

question-driven ontology authoring. European

Semantic Web Conference, 752–767.

Roche, C., & Papadopoulou, M. (2019). Mind the Gap:

Ontology Authoring for Humanists. 1st International

Workshop for Digital Humanities and Their Social

Analysis (WODHSA)-Episode V: The Styrian Autumn

of Ontology.

Ross, S. (2003). Towards a Semantic Web for Heritage

Resources, vol. 3 of DigiCULT Thematic Issue.

DigiCULT, Salzburg, May, 7–11.

Sabou, M., & Fernandez, M. (2012). Ontology (network)

evaluation. In Ontology engineering in a networked

world (pp. 193–212). Springer.

Soergel, D. (1995). The art and architecture thesaurus

(AAT): A critical appraisal. Visual Resources, 10(4),

369–400.

Suárez-Figueroa, M. C., Gómez-Pérez, A., & Fernández-

López, M. (2012). The NeOn methodology for

ontology engineering. In Ontology engineering in a

networked world (pp. 9–34). Springer.

Suárez-Figueroa, M. C., Gómez-Pérez, A., & Fernandez-

Lopez, M. (2015). The NeOn Methodology framework:

A scenario-based methodology for ontology

development. Applied Ontology, 10(2), 107–145.

Sure, Y., Staab, S., & Studer, R. (2004). On-to-knowledge

methodology (OTKM). In Handbook on ontologies (pp.

117–132). Springer.

Ushold, M., Gruninger, M. (1996). Ontologies: Principles,

Methods and Applications. Knowledge Engineering

Review, Vol. 11, n° 2, June 1996. Also available from

AIAI as AIAI-TR-191

Venu, S. H., Mohan, V., Urkalan, K., and Geetha, T. (2016).

Unsupervised domain ontology learning from text. In

International Conference on Mining Intelligence and

Knowledge Exploration, pages 132–143. Springer

Winston, M. E., Chaffin, R., & Herrmann, D. (1987). A

taxonomy of part‐ whole relations. Cognitive Science,

11(4), 417–444.

Zouaq, A. and Nkambou, R. (2009). Enhancing learning

objects with an ontology-based memory. IEEE

Transactions on Knowledge and Data Engineering,

21(6):881–893.

冯先铭. (2002).

中国陶瓷

(上海古籍出版社).

霍华. (1989). 南京博物院藏瓷器标准化命名方法.

东南

文化

, 198–201.

An Ontology of Chinese Ceramic Vases