Enrichment lexical knowledge with interword based features

Throughout the generation or interpretation of textual materials, we need extra information to select specific terms or meaning based on context. This process needs to be supported by extra lexical information in terms of features expressing interrelation between lexical items such as (synonymy, polygamy, antonyms, etc.). In many cases the suitable term corresponding to the concept to be expressed escapes the author. Means to provide him/her with alternatives has to be provided, e.g. choosing between synonymy is usually based on recognizing the minor semantic difference in terms of an integrated hierarchy of lexical features. These features may include overall discourse parameters such as the global topic of the text and the target audience. Such features system should be devised to be easily accessible by both human and machine. This study falls in the intersection area between lexicography and text generation, it has a direct use in automatic translation, language learning and document categorization.