Abstract: The existing linguistic knowledge classification system has been developed mainly based on education, book classification, project management, introductionary textbooks and academic journals. If Chinese linguistics is to truly open up the imagination of linguistics and continuously create new theoretical paradigms, it is necessary to jump out of this “formatting”, rethink “what is language” and then put forward different assumptions about “language”. In this way, a series of linguistic research paradigms can be recognised, discovered and created: “Linguistics based on the assumption of linguistic symbol system”, “Linguistics based on the assumption of the social system”, “Linguistics based on the assumption of the behavioural process”, “Linguistics based on the assumption of the neural functions”, “Linguistics based on the assumption of the numerical process”, “Linguistics based on the assumption of activity competence”, etc. These are “goal-problem-methodology” research paradigms that are both related to each other and quite different from each other, and these paradigms are not “ontological” and “applied” or “central” and “peripheral”, but rather embedded in, supportive of, and in competition with each other.