2011—2013 强化:推理边界四维核对
约 50 分钟
能力目标:控制推理边界
合理推理是在原文命题上走一小步,不添加新的主体、目的或必然性。把每个选项拆成对象、关系、范围和程度,可以把“感觉合理”变成逐项核对。
自拟短文
Three neighborhood clinics began sending appointment reminders in two languages. Missed visits declined over the next six months, most clearly among patients who had previously requested translation help. The clinics did not compare the program with other reminder methods, and the trial involved no rural sites.
题目:Which inference is best supported?
A. Bilingual reminders may help some patients keep appointments.
B. Translation needs are the only cause of missed visits.
C. The program will work equally well in every rural clinic.
D. Bilingual messages are superior to all other reminders.
四维核对
A 保留对象 some patients、关系 help keep、程度 may,且与观察结果一致。B 添加唯一因果;C 把无农村样本推广到所有农村点并改成必然;D 原文未做其他方法对照,无法比较优劣。
未研究不等于相反结论
“没有农村样本”只能限制外推,不能推出农村一定无效;“没有比较其他方法”只能说优越性未知,不能推出它更差。考试常把证据缺口包装成否定结论。
一小步原则
原文是观察性变化,最稳妥推论用 may / might / suggest;若要使用 cause / prove / always,必须有随机对照、排除替代解释等更强证据。
独立变式
若选项写 Patients needing translation benefited most from the reminders,是否完全等价?
**答案:**不完全。原文说下降最明显的是“此前请求过翻译帮助”的患者,这是可观察分组;needing translation 可能扩大到未请求帮助但有需求的人。对象范围发生变化。
独立验收
为 B、C、D 分别指出对象、关系、范围或程度中至少一个越界点,并把 A 改写成另一句保守推论。
Practice
本课练习
先独立作答再提交;编程题会在隔离沙箱中真实编译、运行并对拍。