The evolution of information theory： from classical to modern

A review is presented of the evolution of information theory from its classical foundations to modern developments, with emphasis on its scientific content and global significance. We begin by reviewing Shannon’s classical information theory, elucidating its probabilistic framework for syntactic information measurement, including key concepts such as entropy, mutual information, channel capacity, and rate-distortion functions, along with the three fundamental coding theorems that underpin modern communication systems. The deep connections between information theory and physics—particularly thermodynamics and quantum mechanics—are explored, from which information is revealed as a fundamental descriptor of the physical world. Based on this, we introduce an innovative framework for semantic information theory, proposing“synonymous mapping”as a bridge between syntax and semantics. A new measurement system is constructed, featuring the entropy, mutual information, channel capacity, and rate-distortion functions of semantics. The three semantic coding theorems are demonstrated, showcasing how semantic communication can transcend classical limits to achieve more efficient meaning-oriented transmission. Finally, the paper discusses the integration of semantic information theory with cutting-edge fields such as artificial intelligence and task-oriented communications, paving the way for intelligent, efficient, and interpretable future information networks. From the trajectory of scientific development it is shown how information theory deepens our understanding of the nature of information, and also continuously expands the boundaries of theory and applications.