Ontology Matching: Terminology

As can be observed from what we have presented so far, in the area of ontology matching, different authors including ourselves use different terms to refer to similar concepts, and, vice versa, sometimes different concepts are referred to by the same term. This is especially confusing since these terms can be used for describing both an action and its result, e.g., mapping. We reproduce here the glossary of terms provided in the second edition of the Ontology matching book (pp38-40). The entries are stable and have not changed since its first edition, though several new entries have been added, such as networks of ontologies and data interlinking.


is the process of finding relationships or correspondences between entities of different ontologies.


is a set of correspondences between two or more (in case of multiple matching) ontologies (by analogy with molecular sequence alignment). The alignment is the output of the matching process.


is the relation holding, or supposed to hold according to a particular alignment, between entities of different ontologies. These entities can be as different as classes, individuals, properties or formulas. Some authors use the term mapping instead, however, it will not be used in this sense in this book.


is a correspondence between two ontology entities, which are usually required prior to performing other tasks, such as reasoning and also matching.


is the oriented, or directed, version of an alignment: it maps the entities of one ontology to at most one entity of another ontology. This complies with the mathematical definition of a mapping instead of that of a general relation. The mathematical definition would in principle require that the mapped object is equal to its image, i.e., that the relation is an equivalence relation. A mapping can be seen as a collection of mapping rules all oriented in the same direction, i.e., from one ontology to the other, and such that the elements of the source ontology appear at most once.

Mapping rule

is a correspondence which maps an entity of one ontology into another one from another ontology.

Ontology merging

is the creation of a new ontology from two, possibly overlapping, source ontologies. The initial ontologies remain unaltered. The merged ontology is assumed to contain the knowledge of the initial ontologies, e.g., consequences of each ontology are consequences of the merge. This concept is closely related to that of schema integration in databases.

Ontology integration

is the inclusion in one ontology of another ontology and assertions expressing the glue between these ontologies, usually as bridge axioms. The integrated ontology is assumed to contain the knowledge of both initial ontologies. Contrary to merging, the first ontology is unaltered while the second one is modified.

Bridge axioms

or articulation axioms are formulas, in an ontology language, that express the alignments such that it is possible to integrate the entities of an ontology within one another. Bridge axioms are the basis for ontology merging when the ontologies are expressed in the same language.

Ontology translation

is the process of transforming an ontology from one ontology language to another. By extension, it is a program for translating ontologies.

Ontology transformation

is the process of expressing the entities of an ontology with respect to the entities of another ontology, i.e., relations between entities of the first ontology and those of the second one are added to the first ontology. So the initial consequences of the first ontology are still consequences of the transformation result. The two initial ontologies are unaltered and a third ontology, the transformation result, is created. By extension, it is a program that transforms ontologies.

Data translation

is the process of transforming data or instances from one ontology into corresponding data or instances expressed with respect to another ontology. By extension, it is a program that translates data.

Data interlinking

is the process of establishing explicit links, e.g., equivalences, between instances from different data sources.


consists of interfacing two software components by dynamically altering the information stream between these. By extension, a mediator is a program performing mediation. In web service composition, a mediator translates the output of a service into the input of another one: it thus performs data translation. In query answering applications it is a dual pair of translations that transforms the query from one ontology to another and that translates the answers back.

Networks of ontologies

are made of a set of ontologies and a set of alignments between these ontologies. The ontologies may be related for several reasons: they may be complementary; they may be two independent domain ontologies, e.g., sales and tyres, refinement; there may be a domain ontology specialising a top-level ontology; or they may be supplementary, e.g., a version replacing another version or two ontologies about the same domain.

Linked data,

also known as web of data, is a global data space, which is made of structured data sources published on the web, e.g., in RDF, that are explicitly interlinked.

Ontology version

of an ontology is the ontology resulting from the application of modifications to this ontology.

Ontology reconciliation

is a process that harmonises the content of two or more ontologies, typically requiring changes on one of the two sides or even on both sides. In this case, there is no merging of the ontologies but co-evolution. Ontology reconciliation can be performed for the purpose of merging two ontologies or for the purpose of making them independent.

Reproduced from Jérôme Euzenat, Pavel Shvaiko, Ontology matching, 2nd edition
Copyright © Springer-Verlag, 2013.
Reproduction forbidden without authorisation.