提出一种基于归结的并有关于背景知识及示例的一致特化理论，该理论给出了最大一般特化假设的归结构造方法，可将其作为一种蕴涵意义下的一般理论特化框架．基于该理论，进一步提出ｋ 一般特化概念以解决特化的可计算性问题，并相应地给出特化算法．有关实验表明，该理论与算法能够正确并有效地进行一阶理论特化 A theory of homogeneity and specialization based on the background knowledge and examples is proposed. This theory gives a generalization of the generalization hypothesis, which can be used as a generalized theoretical framework of implication. Based on this theory, the concept of k generalization is further proposed to solve the problem of special computability, and the specialization algorithm is given accordingly. The experiments show that the theory and algorithm can correctly and effectively make the first order specialization