PersonLink: A Multilingual and Multicultural Ontology Representing

Family Relationships

Noura Herradi

1,2

, Fayc¸al Hamdi

, Elisabeth M

etais

and Assia Soukane

Cedric Lab, Conservatoire National des Arts et M

etiers (CNAM), Paris, France

Ecole Centrale d’Electronique (ECE), Paris, France

Keywords:

Semantic Web, Ontology, Linked Data, Vocabulary, Multicultural, Multilingual, Interpersonal Relationships.

Abstract:

Many existing open linked datasets include descriptions of real world persons, with the relationships between

them. For some traditional and/or emerging relationships, existing ontologies do not provide the adequate

links. This paper represents PersonLink, an ontology that deﬁnes rigorously and precisely family relationships,

and takes into account the differences that may exist between cultures, including new relationships emerging

in our societies nowadays. Moreover, the transition from one culture/language to another one cannot be solved

with a simple translation of terms, especially when concepts do not intersect in different languages; thus our

solution refers to a multicultural meta-ontology of concepts and associated mechanisms. A validation has been

performed on two linked datasets DBpedia and Freebase.

1 INTRODUCTION

In the context of Linked Data (Bizer et al., 2009)

(Berners-Lee, 2012), a huge and growing number of

data representing persons and relationships between

them are published (e.g., data available in DBPedia

Freebase

, Yago

, etc.). Different ontologies have al-

ready been proposed in order to describe and enable

sharing family relationships. Some of these ontolo-

gies such as FOAF (Brickley and Miller, 2010), Re-

lationship (Davis and Vitiello, 2005), Agrelon (Litz

et al., 2012), Bio

are widely used. FOAF for ex-

ample deﬁnes these relationships through the predi-

cate “foaf:knows” that links together two individuals.

However, this representation is very limited because

it does not provide more information about the nature

(e.g., family, friendship, etc.) of these links. Thus, it

is not possible to know what is the real relationship

between these two individuals. FOAF relationships

have already been extended as we will show in the re-

lated works section, but these extensions remain not

sufﬁcient. This is because relationships are evolving,

both in scientiﬁc and cultural dimensions. As a result,

many new relationships have emerged and therefore it

http://dbpedia.org/

https://www.freebase.com/

http://www.mpi-inf.mpg.de/yago/

http://vocab.org/bio/0.1/.html

becomes more complex or impossible to express them

with existing ontologies.

In addition to the lack of precision in the deﬁni-

tion of interpersonal relationships in existing vocab-

ularies, an important research issue is added, which

is the problem of the language used for this repre-

sentation which may not necessarily reﬂect the exact

meaning of the term used in another language/culture.

Currently, the majority of existing ontologies are in

English. We argue that several family relationships

cannot be expressed in this language or reversely can-

not be translated (samples are given in the motivat-

ing examples section). Interesting works have been

proposed for translating terms in ontologies (Hellrich

and Hahn, 2014). However, a simple translation re-

mains insufﬁcient, since the meanings of the concepts

depend not only on the language but also on the cul-

ture deﬁning them. Thus, the deﬁnition of the concept

may be different between cultures.

To deal with this issue, we propose a novel ontol-

ogy, called PersonLink, that considers the richness of

the cultures/languages it represents. For this purpose,

we conducted a study to deﬁne the interpersonal re-

lationships according to their presence or absence in

different cultures/languages. For this study, we chose

three languages: French, Arabic and English to show

how concepts may differ from one culture to another

and even between countries on the same culture. We

sought for each interpersonal relationship its concept,

Herradi, N., Hamdi, F., Métais, E. and Soukane, A..

PersonLink: A Multilingual and Multicultural Ontology Representing Family Relationships.

In Proceedings of the 7th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2015) - Volume 2: KEOD, pages 147-154

ISBN: 978-989-758-158-8

147

and if possible, the deﬁnition that has been assigned to

this concept in the three cultures/languages, as well as

the term that has been attributed to this concept in the

languages related to these cultures. Note that some

translation engines (e.g. wiktionary, google trans-

late, etc.) offer different translations for most terms

(e.g. surrogate mother in Arabic), but these transla-

tions have no meaning in the target culture.

The paper is organized as follows. In the next sec-

tion, we present some motivating examples to show

relationships differences in three cultures/languages.

In section 3 we present our PersonLink ontology and

in section 4 experiments that we performed in the con-

text of linked data. In section 5 we present some re-

lated works. Finally, we conclude and give some per-

spectives in section 6.

2 MOTIVATING EXAMPLES

Family links are completely dependent on the cul-

ture and the language. Thus, the terms corresponding

to the concepts they deﬁne depend also on this cul-

ture/language.

For instance, the surrogate mother carries the child

of a couple who gave its embryos.

• In the United States, there is no federal legislation

for Surrogacy. Each state has its own rules, based

on jurisprudence. So the Surrogacy is licensed in

14 states. The term used in English for this mother

is “Surrogate”;

• In France, the law n

94-653 of 29 July 1994

on the respect of the human body explicitly, pro-

hibits surrogacy. The term used in French for this

mother is “M

ere porteuse”;

• In Arabic countries, there is no law authorizing or

prohibiting the Surrogacy, and such a practice re-

mains marginal, due to the weight of tradition. In

Arabic, this relationship does not exist and there

is no term to express it.

Another example is about some concepts that may

not exist in certain languages, for example “God-

mother” does not exist in some cultures/languages.

Even if concepts are sometimes similar in differ-

ent languages/cultures, they may differ in their con-

straints. For instance, depending on the culture, the

“spouse” relationship may be deﬁned between 1 man

and 1 woman, or between 1 man and several women,

etc. Moreover, the accuracy in the deﬁnitions of con-

cepts can change from one language/culture to an-

other. We take for example the cousinship relation-

ship. In English the term used to deﬁne a cousin

relationship is “cousin of”. In French a degree of

precision is added to this deﬁnition, which is the

gender, so there are two terms deﬁning the relation-

ship: “cousin” for male and “cousine” for female. In

Arabic, another layer of precision joins the last one,

which deﬁnes the cousin (male or female) from the

side of the mother or the father. So, there are eight

terms deﬁning the relationship of cousinship in Ara-

bic:

• “È Ñ«



áK



”: the cousin (male), son of their fa-

ther’s brother;

• “È ÈA





áK



”: the cousin (male), son of their

mother’s brother;

• “È



éËA





áK



”: the cousin (male), son of their

mother’s sister;

• “È



éÔ



áK



”: the cousin (female), son of their fa-

ther’s sister;

• “È Ñ«







”: the cousin (female), daughter of

their father’s brother;

• “È ÈA











”: the cousin (female), daughter of

their mother’s brother;

• “È



éËA











”: the cousin (female), daughter of

their mother’s sister;

• “È



éÔ







”: the cousin (female), daughter of

their father’s sister.

Thus, It is not possible to carry out a translation

of terms to switch from a language/culture to another

one, since the concept could not exist or could have

another deﬁnition in the target culture.

3 THE PersonLink ONTOLOGY

Our main objective is to represent interpersonal re-

lationships in a precise manner and in different cul-

tures/languages in order to be both generic and adapt-

able to users. That is why the PersonLink ontol-

ogy represents and deﬁnes the concepts according to

the considered culture, and expresses them by using

terms of the appropriate language. In order to do so,

the ﬁrst step consists in considering culture in the def-

inition of the concept. For each culture, we look at

whether the concept exists or not. If it exists, we de-

scribe it using its deﬁnition in this culture/language.

So, in the ontology, only if the concept exists in the

culture, a term is assigned to it using the language re-

lated to this culture. We obtain as a result a kind of

sparse ontology that we have called “lace ontology”

because it contains many null values as we show in

Fig. 1. Then, from this precise deﬁnition of concepts

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

148

related to culture, we proceed to the formal represen-

tation of these relationships using the fragments of

OWL2 (Hitzler et al., 2012) corresponding to the de-

scription logic S R OI Q (D) (Horrocks et al., 2006).

Finally, we enrich these relationships by a set of DL-

safe rules (Motik et al., 2005) (ensuring decidability)

to have more inference possibilities.

3.1 The Lace Meta-ontology

The problem that arises in the representation of prop-

erties that describe interpersonal relationships, lies in

the description of the property in three different lan-

guages/cultures. Taking the example of the property

deﬁning cousins. In French there exist two speciﬁc

terms that represent this relationship according to the

gender of cousin and eight possible terms in Arabic

capturing in addition to the gender of the cousin, the

mother/father lineage as well as its gender. However

only a generic term exists to deﬁne this relationship

in the English language/culture whereas this generic

term does not exist in the French and Arabic cul-

tures/languages.

In the PersonLink ontology, we deﬁne each con-

cept with a unique number, so each number repre-

sents a concept deﬁning a relationship. This will al-

low us to have a hierarchy with multiple levels of ac-

curacy which combines different languages/cultures.

We can move from one concept to another in the level

of accuracy (vertically), and therefore from a cul-

ture/language to another. Besides, the true meaning of

the concept represented by a term for each language

(obviously, if it exists in the associated culture) is pre-

served. We get as a result a lace meta-ontology (be-

cause of the null values it may have) of concepts with

their representations in different cultures/languages.

The concepts represented in the lace meta-

ontology for the cousinhood relationship of a person,

shown in Fig. 1, have the following deﬁnitions:

• Concept #2: the descendant (regardless of gen-

der) of the uncle or of the aunt (both mother’s or

father’s side);

• Concept #2.1: the female descendant of the uncle

or of the aunt (both mother’s or father’s side);

• Concept #2.2: the male descendant of the uncle

or of the aunt (both mother’s or father’s side);

• Concept #2.1.1: the female descendant of the

mother’s male sibling;

• Concept #2.1.2: the female descendant of the fa-

ther’s male sibling;

• Concept #2.1.3: the female descendant of the

mother’s female sibling;

• Concept #2.1.4: the female descendant of the fa-

ther’s female sibling;

• Concept #2.2.1: the male descendant of the

mother’s male sibling;

• Concept #2.2.2: the male descendant of the fa-

ther’s male sibling;

• Concept #2.2.3: the male descendant of the

mother’s female sibling;

• Concept #2.2.4: the male descendant of the fa-

ther’s female sibling.

3.2 The Translation Algorithm

The main objective of our PersonLink ontology is to

express rigorously family relationships by consider-

ing the culture/language aspect. Nevertheless, note

that to express a source relationship in a different cul-

ture, and in order to get the term(s) used to express

this relationship in the target related language, we ap-

ply an algorithm using our meta-ontology. Given a

source language L

and a target language L

, given

a concept in the source language L

and C

a con-

cept in the target language L

, and given T

a term

in the source language L

and T

a term in the target

language L

. We proceed by cases:

• Case 1: T

0 , which means that in our meta-

ontology, the term corresponding to the concept

exists in the target language L

. So we use it;

• Case 2: T

0, which means that in our meta-

ontology, the term corresponding to the concept

does not exist in the target language L

– We go down in the meta-ontology to ﬁnd a

more speciﬁc concept C

deﬁning this relation-

ship in the target language L

(that may imply

to get further information about the person, e.g.

the gender of the person) and we use the term(s)

associated to this C

to express the T

in the

;

– If a more speciﬁc concept C

does not exist,

we go up in the meta-ontology to ﬁnd the ﬁrst

generic concept C

and we use the term corre-

sponding to this C

to express this relationship.

For our example of cousinhood, using the algo-

rithm based on the lace meta-ontology, we can ex-

press the “cousineDe” (female cousin) relationship

in English. We note that for this relationship, the

= (Fr) and L

= (En). In our meta-ontology, we

have T

0, so we go down in the meta-ontology to

look for the ﬁrst more speciﬁc concept C

, we found

it, but the T

is still =

0 (we are then in the case 2 of

the translation algorithm). We look, in this case, for

PersonLink: A Multilingual and Multicultural Ontology Representing Family Relationships

149

Concept #2

Label(En): cousinOf

Label(Fr): Ø

Label(Ar): Ø

Concept #2.1

Label(En): Ø

Label(Fr): cousineDe

Label(Ar): Ø

Concept #2.1.1

Label(En): Ø

Label(Fr): Ø

Label(Ar): !"!"#"$%&'

Concept #2.1.2

Label(En): Ø

Label(Fr): Ø

Label(Ar): !"()"$%&'

Concept #2.1.3

Label(En): Ø

Label(Fr): Ø

Label(Ar): !"$*"#"$%&'

Concept #2.1.4

Label(En): Ø

Label(Fr): Ø

Label(Ar): !"$+)"$%&'

Concept #2.2

Label(En): Ø

Label(Fr): cousinDe

Label(Ar): Ø

Concept #2.2.1

Label(En): Ø

Label(Fr): Ø

Label(Ar): !"!"#",&'

Concept #2.2.2

Label(En): Ø

Label(Fr): Ø

Label(Ar): !"()",&'

Concept #2.2.3

Label(En): Ø

Label(Fr): Ø

Label(Ar): !"$*"#",&'

Concept #2.2.4

Label(En): Ø

Label(Fr): Ø

Label(Ar): !"$+)",&'

rdfs: subPropertyOf

Figure 1: Excerpt of the meta-ontology for cousinhood in English, French, and Arabic.

Algorithm 1: Translation algorithm.

1: C

b= T

| T

≡ label(C

, L

)

2: T

= T

+ {label(C

, L

)}

3: if T

0 then

4: for all C

| C

v C

5: T

= T

+ {label(C

, L

)}

6: if T

0 then

7: T

= T

+{label(C

, L

)} | C

v C

the ﬁrst generic concept C

used to express this re-

lationship, and we use the term corresponding to it,

which is T

= cousinO f .

3.3 Formal Deﬁnition of the

Relationships

Interpersonal relationships in the PersonLink ontol-

ogy are represented in a structured way with the

OWL2 language that corresponds to the description

logic S R OI Q (D). The use of OWL2 is privileged

because it provides a high expressiveness and allows

us to represent relationships that we were not able to

represent in OWL1. In addition, OWL2, with this

logic description, allows semantic reasoners to ver-

ify the consistency of data, to derive new knowledge

or to extract information already present. Besides, the

reasoning in OWL2 is complete and decidable.

In predicate logic, the hierarchy represented in

Fig. 1 means that:

2(?x, ?y) ⇔ 2.1(?x, ?y) ∨ 2.2(?x, ?y)

and 2.1(?x, ?y) ⇔ 2.1.1(?x, ?y) ∨2.1.2(?x, ?y) ∨

2.1.3(?x, ?y) ∨ 2.1.4(?x, ?y)

and 2.2(?x, ?y) ⇔ 2.2.1(?x, ?y) ∨2.2.2(?x, ?y) ∨

2.2.3(?x, ?y) ∨ 2.2.4(?x, ?y)

This representation means that the most spe-

ciﬁc concept implies automatically the generic one,

for example whenever (x cousinDe y), it implies

that (x cousinO f y): 2.2(?x, ?y) ⇒ 2(?x, ?y) and

also whenever (z È Ñ«







w), we deduce that

(z cousineDe w) and (z cousinO f w) as following:

2.1.2(?z, ?w) ⇒ 2.1(?z, ?w) and

2.1.2(?z, ?w) ⇒ 2(?z, ?w)

In addition, in order to have more precision about

the relationship going from the most generic con-

cept (“cousinO f ” in our example) to a speciﬁc one

(“cousineDe”) (which means: a female cousin), we

have to get more information about the instance (fe-

male in our example) from the knowledge base. The

speciﬁc concept would be inferred from the SWRL

(DL-Safe) rules that we would have previously cre-

ated:

2(?x, ?y) ∧ Female(?x) ⇒ 2.1(?x, ?y)

Note that we need the information Female(?x)

to deduce that the type of the relationship 2(?x, ?y)

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

150

Table 1: Entities having parent, child, and sibling relationships in DBpedia.

Relationship DBpedia Relationship Number of Entities

Parent dbpedia-owl:parent 1675

Child dbpprop:children 5117

Sibling dbprop:sibling 4

Sibling’s child – 0

Cousin – 0

Table 2: Entities having parent, child, and sibling relationships in Freebase.

Relationship Freebase Relationship Number of Entities

Parent /people/person/parents 2000

Child /people/person/children 5155

Sibling /people/sibling relationship/sibling 1815

Sibling’s child – 0

Cousin – 0

(“cousinO f ”) is, in this case, the 2.1(?x, ?y) relation-

ship (in French “cousineDe”). However, if we want

to get a more speciﬁc concept for this relationship of

cousinhood (for example: in Arabic z È Ñ«







(which means that the instance is a female cousin on

the father’s side), in this case we need to get further

information that will indicate, in addition of the gen-

der information, if the person is a cousin on the side of

the father or on the mother’s one. This information,

could be deduced by using the DL-Safe rules previ-

ously deﬁned. It is worth noting that, for some rela-

tionships, the application of SWRL (DL-Safe) rules is

not necessary, because OWL2 allows us to deduce in-

formation through the objectPropertyChain function

as following:

subPropertyOf(objectPropertyChain(:sonOf :sib-

lingOf :parentOf):cousinOf)

The whole PersonLink ontology is available and

published in a dereferenceable manner

(and thus,

would be referenced by the Linked Open Vocab-

ulary) at: http://cedric.cnam.fr/∼hamdif/ontologies/

PersonLink.owl

4 VALIDATION

To test the validity of our reasoning and translat-

ing mechanism, we have taken as a sample persons

described in the context of the Linked Open Data

(LOD). We chose as datasets Freebase, which is a

large collaborative knowledge base built mainly from

data provided by its community members and DB-

Pedia, which is a knowledge base built from ex-

tracted and structured data from Wikipedia. We show

http://www.w3.org/TR/swbp-vocab-pub/

as relationship example, the cousinship relation pre-

sented in section 3. First, we searched in the Free-

base and DBPedia ontologies properties that could

express the “cousinO f ” relationship. In the current

versions, none of Freebase or DBpedia uses include

properties to express this kind of relationship. How-

ever, we found other relationships (parental and sib-

ling) that, combined, could be used to express implic-

itly the cousinship relation. Hence, we extract from a

subset (due to the user queries limitations) of DBpe-

dia and Freebase, entities that are linked to each other

by a parent and/or sibling relationships, as well as

the relationships themselves. This extraction process

is done automatically using scripts running SPARQL

(SPARQL Protocol And RDF Query Language) and

MQL (Metaweb Query Language) queries. The ob-

tained results are presented in Table 1 and Table 2.

From DBpedia, we extracted 1675 entities that,

having as type “person” and linked to other ones by

the “parent” relationship. We extracted also 5117 per-

sons having children and 4 having siblings. However,

as we explained above, no sibling’s child or cousin re-

lationships were found. From Freebase, we extracted

2000 entities “person” that have parent relationships.

From theses entities, 5155 children and 1815 are gen-

erated. There too, no sibling’s child nor cousin rela-

tionships were found.

In the result of these tests, we note that the sib-

ling’s children and the person children could be can-

didates to be cousins. Thus, we integrated entities

and relationships that we obtained, on the Person-

Link ontology to populate our knowledge base, then

we applied automated reasoning to get new relation-

ships. For DBpedia, the reasoner produces as ex-

pected “cousinO f ” relationships. The results are pre-

sented in Table 3.

PersonLink: A Multilingual and Multicultural Ontology Representing Family Relationships

151

Table 3: Inferred “cousinOf” relationships.

PersonLink Relationship DBpedia Entities Freebase Entities

cousinO f 3 16426

Table 4: Inferred relationships using Freebase properties.

FreeBase

Relationship

Number

of Entities

Inferred

Relationship

Number

of Entities

Null gender

value

/people/person/parents 2000

motherO f 717

f atherO f 1275

/people/person/children 5155

daughterO f 2100

269

sonO f 2786

– 5117

nieceO f 2069

230

nephewO f 2818

In the case of Freebase, the reasoner is able to in-

fer much more rigorous relationships. For instance,

the “ f atherO f ” and “motherO f ” relationships are in-

ferred by exploiting the “parent” and “gender” prop-

erties describing Freebase entities (DBpedia does not

provide these kind of properties). The “daughterO f ”

and “sonO f ” relationships are inferred by exploiting

the “children” and “gender” properties. We note that

there are some people with null gender values (this

is because data available on these datasets are not al-

ways complete). The results we got by reasoning on

data are presented in Table 4.

Among the 2000 Freebase person, 717 could be

represented using the “motherO f ” relationship and

1275 the “ f atherO f ” one. These results show that

our PersonLink ontology has the particularity of using

properties much more expressive than those present in

DBpedia or Freebase.

We also proceeded to translate relationships us-

ing our translation algorithm based on the lace meta-

ontology showed in Fig. 1. So, for the inferred Free-

base “cousinO f ” relationships (cf. table 3), the al-

gorithm speciﬁes, in the case of french translation,

if this cousin is a “cousineDe” (female cousin) or a

“cousinDe” (male cousin). The results we got for this

case are gathered in Table 5. This table, showed that

among 16426 “cousinO f ”, the translation algorithm

found 5986 relationships of “cousinDe” (male cousin)

and 4116 “cousineDe” relationships (female cousin)

in French. 6324 represents the number of cousins that

Table 5: “cousinOf” relationships translated to the French

language.

cousinO f (En) 16426

cousinDe (Fr) 5986

cousineDe (Fr) 4116

null gender value 6324

we cannot specify if they are male or females, due to

the absence of the gender property in their descrip-

tions.

5 RELATED WORKS

Different ontologies have been proposed to describe

family relationships in the web. The most famous

one is FOAF (Brickley and Miller, 2010) which de-

ﬁnes relationships between people through the predi-

cate “foaf:knows” that links together two individuals.

However, this representation is very limited because

it does not provide the nature (e.g., family, friendship,

etc.) of these links. That’s why some extensions have

been proposed for this property. For instance, the Re-

lationship

ontology (Davis and Vitiello, 2005) which

extends FOAF by introducing several sub-properties

to the property “foaf:knows”, that provide some terms

representing parenthood, childhood, siblinghood and

a generic term representing marriage “SpouseO f ”.

The Agrelon (Agent Relationship Ontology)

ontol-

ogy (Litz et al., 2012), designed in the context of

the CONTENTUS Project (Nandzik et al., 2013),

presents a more precise set of terms that distinguish

between the different types of relationships extending

the property “knows” of FOAF. For example, siblings

and half siblings can be distinguished by the two dis-

tinct properties “hasSibling” and “hasHal f Sibling”.

The Relationship and Agrelon ontologies bring more

clarity to the relationships. Nevertheless, they remain

very generic, lack precision, and they do not support

multiculturalism. The Bio

ontology, aims to describe

biographical information about people. In this ontol-

http://vocab.org/relationship/.html

http://www.contentus-

projekt.de/ﬁleadmin/download/agrelon.owl

http://vocab.org/bio/0.1/.html

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

152

Table 6: Excerpt of the comparison table of the representation of certain family relationships in different existing ontologies.

Family Relationship FOAF Relationship AgRelOn Bio

Surrogate Mother knows parentOf has parent mother

Half Brother knows / has half sibling /

common-law wife knows live with has cohabitee /

ogy, there are some relationships that may be interest-

ing, such as “Father”, “Mother”, etc. This ontology

gives more precise deﬁnitions than the two previous

ones. However, in addition of being incomplete and

not multilingual, it is intended to store events and not

links. Therefore, “Bio” can not be used to interpret

interpersonal relationships linking different individu-

als.

The table 6 shows an excerpt of the compari-

son between the four previously mentioned ontolo-

gies, depending on how they express different rela-

tionships. Generally, these ontologies use generic

properties (e.g. parentOf, hasParent etc. for the

“Surrogate mother” concept), also when there is no

generic property to deﬁne these relationships, the ex-

isting ontologies use approximative meanings’ prop-

erties (e.g. Agrelon:hasCohabitee when talking about

common-law wives). Moreover, they offer very short

and generic deﬁnitions to describe interpersonal rela-

tionships whereas this kind of relationships has to be

rigorously described to give exact meaning when dis-

playing information about interpersonal relationships.

In addition to the lack of precision, the majority

of existing ontologies about family relationships are

in English. However, ontologies should be used in

different cultures. Some projects aim to solve the is-

sue related to the “syntactic” translation (Chalupsky,

2000) (Dou et al., 2005). But as we saw with moti-

vating examples, a literal translation is not always rel-

evant. Besides that, other important studies targeted

the “cultural” translation issue, like the recent Euro-

pean project MONNET which has proposed a model

named LEMON to deal with this issue. The authors of

this work (Montiel-Ponsoda et al., 2011) developed a

translation module which is intended to link between

concepts coming from various language/culture on-

tologies. First, they provide a literary translation of

the concept in the target language. Then they seek

its equivalent in the associated target culture, when

the term used in the target culture is different from

the literary given one, they link between the source

concept and its equivalent one by using a “Cultur-

alEquivalentTranslation” class. But this solution does

not consider the absence of the equivalent term rep-

resenting the concept in the target culture and still

gives a literal translation to the concept even if it

may be incorrect since we consider culture and lan-

guage are completely bound when deﬁning interper-

sonal relationships. Moreover, the different languages

are not merged into a single ontology; their solution is

based on ontology alignment including cultural equiv-

alences.

Some ontologies have been introduced with the

purpose of structuring web resources including de-

scription of persons. YAGO (Suchanek et al., 2007)

is an ontology derived from Wikipedia, WordNet and

GeoNames. Its goal is to structure Wikipedia as a

linked Database. Thus, an important sub-thematic of

YAGO concerns relationships between persons. How-

ever, all the relationships are expressed in English and

are limited regarding expressiveness. Freebase and

DBpedia are two projects similar to YAGO (except

that Freebase gathers its information from the com-

munity of users), and thus present the same issue.

6 CONCLUSION AND FUTURE

WORKS

In this paper we have presented a new ontology

called PersonLink, that enables users and applica-

tions to represent interpersonal relationships. Person-

Link provides a precise deﬁnition for each relation-

ship and takes into consideration the culture/language

aspect. We have introduced the notion of lace meta-

ontology that facilitates the expression (by reﬁnement

of concepts and speciﬁcation of contraints), in multi-

ple cultures/languages of each relationship and allows

to switch between languages and ﬁnd the right terms

expressing the relationships. The particularities of our

solution compared to usual ontology translation meth-

ods are on the one hand that it merges all concepts

and terms in a single ontology rather than providing

ontology alignments and on the other hand the trans-

lation is perfectly reversible. This solution is partic-

ularly suitable for our domain application where con-

cepts are very closed but not similar in the different

languages, and depend on the context (culture, law,

gender of the person, etc.).

A set of inference rules allows to deduce new links

from the given ones, and to check the consistency of

inputs. The current version of PersonLink includes 3

classes (Person, Male and Female), 86 properties, and

582 SWRL rules.

PersonLink: A Multilingual and Multicultural Ontology Representing Family Relationships

153

We have experimented the reasoning mechanism

and the translation algorithm of the PersonLink on-

tology in the context of Linked Open data. Tests con-

ducted on DBpedia and Freebase datasets show that

the use of PersonLink enables inferring much more

rigorous relationships than those already present in

these datasets.

As the PersonLink ontology is available and pub-

lished in a dereferenceable manner and will be refer-

enced by the Linked Open Vocabulary) it can be used

by any Linked Data publishers who need family rela-

tionships and should comply with the W3C best prac-

tices. In our laboratory, two on-going applications are

using the PersonLink ontology:

• The ﬁrst one is a memory prosthesis called CAP-

TAIN MEMO devoted to persons with memory

impairments. This application need to store all

family relationships of the owner of the prosthe-

sis. Thus, an ontology with a ﬁne-grained rela-

tionships deﬁnitions is mandatory to allow storing

all possible links. Moreover, the reasoning mech-

anism contributes to check the consistency of the

inputs, that is a great help, especially for persons

with memory impairments and cognitive discor-

dances.

• The second one aims to integrate PersonLink in

the SIGIL electronic books editing system. SIGIL

allows to edit an electronic book in the Epub for-

mat. PersonLink can structure the metadata con-

cerning the genealogy of main characters. Con-

cerning this application, the main interest of Per-

sonLink is its availability to switch from one lan-

guage to another, by providing links correspond-

ing to different cultures and to ensure a smart

translation.

Future work will be mainly devoted to enrich the

ontology with convivial links between people (neigh-

bours, friends, care givers, etc.) and to take into ac-

count time variance.

REFERENCES

Berners-Lee, T. (2012). Design Issues: Linked

Data. Retrieved March 1, 2012, from

http://www.w3.org/DesignIssues/LinkedData.html.

Bizer, C., Heath, T., and Berners-Lee, T. (2009). Linked

data - the story so far. Int. J. Semantic Web Inf. Syst.,

5(3):1–22.

Brickley, D. and Miller, L. (2010). FOAF Vocabulary Spec-

iﬁcation 0.97. Namespace document.

Chalupsky, H. (2000). Ontomorph: A translation system

for symbolic knowledge. In Cohn, A. G., Giunchiglia,

F., and Selman, B., editors, KR 2000, Principles of

Knowledge Representation and Reasoning Proceed-

ings of the Seventh International Conference, Breck-

enridge, Colorado, USA, April 11-15, 2000., pages

471–482. Morgan Kaufmann.

Davis, I. and Vitiello, E. J. (2005). RELATIONSHIP: A vo-

cabulary for describing relationships between people.

Technical report, vocab.org.

Dou, D., McDermott, D. V., and Qi, P. (2005). Ontology

translation on the semantic web. J. Data Semantics,

3360:35–57.

Hellrich, J. and Hahn, U. (2014). Enhancing multilin-

gual biomedical terminologies via machine translation

from parallel corpora. In M

etais, E., Roche, M., and

Teisseire, M., editors, NLDB, volume 8455 of Lecture

Notes in Computer Science, pages 9–20. Springer.

Hitzler, P., Kr

otzsch, M., Parsia, B., Patel-Schneider, P. F.,

and Rudolph, S. (2012). OWL 2 Web Ontology Lan-

guage Primer (Second Edition). W3C Recommenda-

tion, World Wide Web Consortium.

Horrocks, I., Kutz, O., and Sattler, U. (2006). The Even

More Irresistible SROIQ. In Doherty, P., Mylopoulos,

J., and Welty, C. A., editors, KR, pages 57–67. AAAI

Press.

Litz, B., L

ohden, A., Hannemann, J., and Svensson, L.

(2012). Agrelon - an agent relationship ontology. In

Dodero, J. M., Palomo-Duarte, M., and Karampiperis,

P., editors, MTSR, volume 343 of Communications in

Computer and Information Science, pages 202–213.

Springer.

Montiel-Ponsoda, E., Gracia, J., de Cea, G. A., and G

omez-

erez, A. (2011). Representing translations on the

semantic web. In Montiel-Ponsoda, E., McCrae, J.,

Buitelaar, P., and Cimiano, P., editors, Proceedings of

the 2nd International Workshop on the Multilingual

Semantic Web, Bonn, Germany, October 23, 2011,

volume 775 of CEUR Workshop Proceedings, pages

25–37. CEUR-WS.org.

Motik, B., Sattler, U., and Studer, R. (2005). Query

Answering for OWL-DL with rules. J. Web Sem.,

3(1):41–60.

Nandzik, J., Litz, B., Flores-Herr, N., L

ohden, A., Konya,

I. V., Baum, D., Bergholz, A., Sch

onfuß, D., Fey, C.,

Osterhoff, J., Waitelonis, J., Sack, H., K

ohler, R., and

Ndjiki-Nya, P. (2013). Contentus - technologies for

next generation multimedia libraries - automatic mul-

timedia processing for semantic search. Multimedia

Tools Appl., 63(2):287–329.

Suchanek, F. M., Kasneci, G., and Weikum, G. (2007).

Yago: A core of semantic knowledge. In Proceed-

ings of the 16th International Conference on World

Wide Web, WWW ’07, pages 697–706, New York,

NY, USA. ACM.

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

154