The purpose of this paper is to compare the two major types of formalization strategies through the disambiguation of natural language textual ambiguities. The method is: The first step is to select the same text. Using poetry as an example, two types of formal strategies are used to resolve the ambiguities that exist. The second step is to analyze the limitations of the first formal path, at the same time, using traditional artificial intelligence methods and a new generation of artificial intelligence. The third step is to use the double-word board tools and methods to do the same thing. The result is that using the first path, whether based on rules (traditional artificial intelligence methods) or on statistical and machine learning, especially deep learning (a new generation of artificial intelligence methods), only local solutions can be obtained; With the checkerboard tools and methods, the overall solution can be obtained. This shows the unique advantages of the second path. Its significance lies in: using the double-word chessboard tool and method (second path) can solve the common problems faced by traditional artificial intelligence and new generation of artificial intelligence, and how to eliminate the ambiguity of natural language texts. The most important thing is that it has a new role. The most typical is to construct a knowledge base of the subject through the acquisition of knowledge and formal expression of experts, so as to gradually resolve a series of ambiguities between natural language (text) processing and formalized understanding.