The Multilingual Morpheme Ontology (MMoOn) is a model that contains linguistic concepts and relations that are necessary to describe the morphemic elements of natural languages. This includes the description of elements on the word-level, such as lexemes and word-forms as well as elements on the sub-word level, such as roots, stems and affixes. Further, derivational relations between lexemes and inflectional relations between word-forms are provided and all linguistic elements can be specified for their meaning.

MMoOn is an ontological model that corresponds to the principles of the Web of Data. It makes use of RDF and OWL and hence allows every single data point to be represented as a directly accessible and interoperable resource on the Web.

The purpose of MMoOn is to serve as descriptive starting point for creating fine-grained morphological language data for inflectional languages, the so called MMoOn morpheme inventories.

Every language-specific MMoOn morpheme inventory consists of primary and secondary language data comprising the following three data levels:

Architectural setup of MMoOn morpheme inventories.

1. Language-independent schema level

The center that connects the morphemic data of all MMoOn morpheme inventories constitutes the MMoOn Core model. It is a language-independent schema representing the linguistic domain of morphology and which is cross-linguistically applicable.

2. Language-specific schema level

In order to describe the language-specific pecularities of the morphological elements present in a certain natural language, another schema is developed which is an instantiation of the MMoOn Core model. As a result, each MMoOn morpheme inventory can be described acording to the actual morphological realizations of a certain language. Usually, language-specific schemas are manually developed by the language experts when a new MMoOn morpheme inventory shall be created. Also, these schemas can be reused by others to represents the morphological data of the same language.

Together, the language-independent and the language-specific schema levels, will yield the Multilingual Morpheme Ontology which enables a cross-linguistic comparision of morphological characteristics and elements of all MMoOn language inventories described.

3. Language-specific data level

Finally, the language-specific data level encompasses the instance data. These are the lexemes, word-forms, morphs and morphemes of a certain language which are described adequately by a created language-specific schema  or by the reuse of an already existing schema.


At the current state the MMoOn Core is still in the creation phase. However, an attempt to describe Hebrew morphemic data has been already undertaken and is described in a paper accepted at the LREC conference 2016.

The development of the MMoOn Core model as well as the Open Hebrew MMoOn inventory can be found at Github.