Argument Frame Extraction and Term Clustering from an English-French Bilingual Aligned Corpus