Multilingual Aspects of Monolingual Corpora

If someone would collect opinions among the computational linguists what had been the most important trend in linguistics in the last decade, it is highly probable that the majority would answer that it was the massive use of large natural language corpora in many linguistic fields. The concept of collecting large amounts of written or spoken natural language data has become extremely important for several linguistic research fields. The majority of large corpora used by linguists are monolingual, although there are several examples of bilingual corpora (e.g. Hansard corpus). This paper would like to present evidence that even the monolingual corpora can be useful for multilingual applications.