Textual entailment has been proposed as a unifying generic framework for modeling language variability and semantic inference in different Natural Language Processing (NLP) tasks. This paper presents a statistical method for recognizing Chinese textu