摘要: Modem digital libraries contain loads of multilingual information. They often refer to lexicons for building indices and retrieving information. However, many Asian languages lack lexicons or even comprehensive public electronic d... 展开 Modem digital libraries contain loads of multilingual information. They often refer to lexicons for building indices and retrieving information. However, many Asian languages lack lexicons or even comprehensive public electronic dictionaries usable for multilingual information retrieval. This study has first done a historical study on current lexical ontology models. Finally it proposes a cognate based approach for building a Sinhala-English lexicon. Within the Sri Lankan multicultural context, multilingual information retrieval is necessary. An understanding between these cultures will lead to greater ethnic harmony, ending decades old ethnic war. From previous studies, it is evident that parallel-corpora based methods of multilingual information retireval are far more expensive than dictionary based systems. Therefore we propose a cognate based approach and algorithm for construction of a Sinhala-English lexicon. Because English-Tamil lexicons are already available, this step would lead to building up a Tamil-Sinhala lexicon as well. Sinhala and Tamil are the main local languages in Sri Lanka. The historical study reveals different models of lexicons, and lexicon building approaches. It is the first study ofits kind since 1996. The suggested cognate based approach is an automatic means of translation, and is an innovative approach. It selects words adjacent to cognates, based on frequency counts as candidate translations between English and Sinhala In future, these translations can be added to the ontology as instances. For example, English WordNet can be extended to incorporate Sinhala terms, making it multilingual. 收起