“…Although Qadir and Riloff"s [12] design of speech act features was based on the text corpus of message board posts, the features were comprehensive enough to cover speech act properties in other text areas. For instance, the applicability of this data analysis approach has been verified in previous research for extracting commissives and directives in verbal maritime communication [8]. Following the approach proposed by Qadir and Riloff [12], the list of commissive classifiers was developed and presented in Table II. 保证, 承诺, 答应, 诺言, 许诺 "pledge" 保证, 承诺, 发誓 "commit" 承诺, 致力于 "guarantee" 保障, 保证 "plan" 计划, 规划, 安排, 打算, 方案 "First person pronoun" 我, 我们 "future tense" 将, 会, 将会 "modals" 愿, 愿意, 一定 "future" 将来, 未来, 明天 "later" 以后, 之后, 后期, 晚些…”