A method of creating translingual text representations takes in documents in a first language and in a second language and creates a matrix using the words in the documents to represent which words are present in which language. An algorithm is applied to each matrix such that like documents are placed close to each other and unlike documents are moved far from each other.