Create a classNoun

Singular sentence

Click on a row to asign the Object

Plural sentence

Click on a row to asign another Object

Lemon pattern code


Create a state verb


Click on a row to asign Subject and Object

Lemon pattern code


Create a relational noun


Click on a row to asign Subject and Object

Lemon pattern code


Create a intersective adjective

Create a relational adjective

Instances (A-Box)

Classes and properties (T-Box)

Welcome to Lemon Assistant

This application allows you to create Lemon Patterns, the most typical patterns of the Lemon model. These patterns allows you to create ontology 'lexicons', that is, lexicalizations of ontologies.

For instance, let us assume that you want to lexicalize the property “writer” of the DBpedia ontology. This is the lemon pattern code of the state verb “to write”:

StateVerb ("write", dbpedia:writer,
           propSubj = DirectObject,
           propObj  = Subject)

But there is another pattern for relational nouns. This is the lemon pattern code for the same property:

               propSubj = PrepositionalObject("of"),
               propObj  = CopulativeArg)

In order to help you to check the correcteness of the code, Lemon Assistant creates a sentence in natural language for each pattern. These sentences are created by using multilingual grammars built with Grammatical framework.

You have to provide some instance of the RDF triples in which this property appears. Let us say that, for the previous examples, we see that “Shakespeare” appears as the object in the triple, and “Macbeth” as the subject. These are the natural language sentences generated:

For the state verb example:

Shakespeare writes Macbeth

For the relational noun example

Shakespeare is the writer of Macbeth


  • Multilingual (English, Spanish, German). You can create lexicons in these three languages for any ontology.
  • Patterns supported: class noun, state verb, reational noun, intersective adjective, relational adjective.
  • This web application can be used in tablets or smartphones (also in desktop machines).

Technical details

  • This application uses a R package named 'Rlemon' oriented to join linked data and linguistic grammars (gf).
  • This web application is reactive, that is, there are no 'submit' buttons. For an introduction to reactive web applications visit the Shiny homepage.


Background images from Pixabay:

Made by Mariano Rico and Christina Unger, 2015. Supported by projects LIDER (EU FP7 project No. 610782) and MINECO’s JdC Grant (JCI-2012-12719) and INFRA (UNPM13-4E-1814).