The problem of causality has been discussed and studied by various scholars.Meanwhile,the theory of causality has also been enriched along with the development of the times.However,since traditional theories of causality in modern times cannot be applied for addressing the complicated social phenomena due to the progresses in advanced technologies such as artificial intelligence(AI),big data,and biomedicine,computer scientists and philosophers represented by Judea Pearl have subversively"revolved” the theory of causality to re-examine the process of people generating causal consciousness.Also,Bayesian networks and structural causal models have been also established in accordance with the laws of causal consciousness for solving complicated causal problems and cross-population causal problems in the fields of AI,medicine,law,and economics.On this basis,it is still difficult to handle complicated causal problems generated in the constantly developing technology and society.In that case,offering reasonable suggestions for the future development of the theory of causality is both a theoretical and practical problem.This paper was designed for studying the theory of causality using a research method combining comparative analysis,text interpretation,and case modelling.To begin with,representative perspectives of philosophers were selected from the development of traditional causality theories in different periods of modern times,including David Hume’s theory of empirical causality,Russell’s understanding and induction,and Searle’s theory of intentional causality,and summarized.It can be found these traditional theories of causality in modern and contemporary times are presented in a flat dilemma,a dilemma of unidirectional and linear means of causal interpretation and a dilemma of deficient mathematical quantitative basis in causal analysis when traditional causality theories in modern times are applied to deal with modern complicated causal problems in computer technology.Moreover,Judea Pearl’s theory of causality was briefly introduced,including the trapezoidal process of human generating causal consciousness-the ladder of causality.Based on this,it is proposed that quantitatively treating uncertain factors is an essential operation for analyzing the relationship between causality and probability.Hence,a Bayesian network was established to analyze uncertain factors of causality.Meanwhile,a more precise structural causal model was constructed based on the Bayesian network,so that the revolution was implemented in the theory of causality.Also,this study also indicated that realizations of visualization,multi-path analysis and mathematical physics of the theory of causality in the computer field were the "revolutions" of these theories to the traditional theory causality.At last,the value of Judea Pearl’s theory of causality was analyzed through introducing the application of related theories in medicine,economy,and law,as well as analyzing the dilemmas of Judea Pearl’s theory of causality,including insufficient consistency between extrapolation and experience in positive correlation,the to-be-improved adaptability of linear models in complicated causality,and the unclear definition of intervention conditions in causal models.Targeted at the above-mentioned dilemma,theoretical development suggestions were proposed for improving multiple paths for the same causal category,increasing the inspection path combining inverse probability and inverse correlation as well as presupposing analysis causal adjustment models and causal multi-dimensional models. |