One might ask: Is a 1995 semantics textbook still relevant in the era of large language models (LLMs) like ChatGPT and BERT?
The answer is a qualified yes. While computational semantics now relies on distributional vectors and neural embeddings, Lyons’ insights into sense relations (synonymy, hyponymy) are directly encoded in WordNet and similar knowledge graphs. His warnings about circularity in defining meaning anticipate current debates about whether LLMs truly understand or merely pattern-match.
Moreover, his insistence on speaker-meaning over sentence-meaning remains a necessary corrective to purely formal approaches. For anyone building or critiquing AI systems, Lyons provides the conceptual vocabulary.
The "work" of John Lyons in this volume is significant for several reasons:
This report provides an overview of John Lyons’ seminal work, Linguistic Semantics: An Introduction. As one of the most distinguished linguists of the 20thth century, Lyons wrote this book as an accessible yet rigorous entry point into the field of semantics—the study of meaning. Unlike his earlier, more technical two-volume set Semantics (1977), this single-volume work distills complex theories for students and general readers. The text is renowned for establishing semantics as an autonomous branch of linguistics, independent of philosophy or psychology, while acknowledging their intersections.
