概称句推理与排序

概称句推理可分为主要通过演绎方式的和主要通过归纳方式的。对于演绎方式进行的推理又可根据研究关注点的不同分为结论是事实句的推理和结论是概称旬的推理。由于概称句的作用，这三种类型的推理都是非单调推理。通过分别的考察，本文指出，这三种类型的概称句推理要想真正得以刻画，都需要引入前提集的排序的概念。进而，排序将是融合这三种类型推理的纽带。

The Reasoning Encoded Generics and Orders on Premise Set

The study of generics began in 1970s. Because of their importance in our thinking and language, generic sentences have been studied from many viewpoints such as linguistics, artificial intelligence, logic, philosophy, etc. Though these researches have provided some theories and interpretations for generics, the study of generics is still at the beginning stage, and not any theory can be accepted by most researchers. Logicians study generics for the reasoning encoded generic sentences. The reasoning encoded generics can be classified into two kinds： by deduction mainly and by induction mainly. For the kind of reasoning by deduction, we can divide them into the reasoning getting factual sentences and the reasoning getting generic sentences. In this article, we＇ll argue them separately, and we conclude that, for every kind of reasoning encoded generics, if we want to capture them, we have to int~;oduce the priority order on premise set. Then, to capture the non-monotonic reasoning encoded generics, we have to introduce the priority order on premise set. We begin with the reasoning which gets generic sentences by deduction. From ＇birds fly＇ and ＇sparrows are birds＇ getting ＇sparrows fly＇ （GAG type） is a representative instance of this kind of reasoning. Based on the logic system GD, we introduce the concept of deduction with a priority order on the premise set. We argue that, to capture the reasoning of getting generic sentences by deduction, we must intro- duce the priority order on premise set. After this, we give normal rules for priority order and two special rules for our reasoning, which are sub-category generics priority and generics priority. Our definition can pass the GAG type plainly. Next, the conclusions of reasoning encoded generics are factual sentences. In [7], Beihai Zhou and Yi Mao have given some priority order on the premise set and show that their definition can pass through benchmark examples like Nixon Diamond, Penguin Principle etc. At last, we go to the most difficult one, getting generic sentences by induction. By the famous instance of swan, we show that we cannot get the conclusion that ＂swan is white＂ because of the orders on the premise sets again! But here, it＇s category generics priority, Finally, we make a summary. Our daily life reasoning with generic sentences is the mixture of induction and deduction. We classify and study them respectively just for the technical reason. But our final goal is to combine the induction and deduction. The ＇broad-brush＇ order may serve as the best ligament among the three kinds of reasoning.