Corpus Parole (3 milions words)

The PAROLE project (Preparatory Action for Linguistic Resources Organization for Language Engineering) has produced a set of harmonized corpora and lexicons for a large number of European languages. Each corpus, made up of 20 million words, was built up as reference corpus for Human Language Technology applications, to provide full information about a large variety of text types in the language considered, to represent the use of contemporary language and to become the first nucleus of an electronic text library. The texts have been stored using a common format following the standards recommended in the CES (Corpus Encoding Standard), according to flexibility and multifunctionality criteria. The texts belong to a wide range of media and genres, selected in proportions aimed at reflecting their prominence within the society, classified according to medium, genre, topic and time of production.