Natural language processing (NLP) is a subfield of computer science, with a history of already more than 50 years. It concerns the interaction between humans and computers through natural language, and integrates specialties like speech recognition, speech generation, and natural language understanding (NLU).

This chapter considers language understanding aspects and traces essential moments in the development of NLU in relation to logic programming. It points out important contributions from David Scott Warren, for example, tabling as a tool to handle constraints and understanding the language structure. The presentation is focused on Prolog, a logic programming language for artificial intelligence (AI), initially intended for NLP. The chapter is also about Prolog evolutions such as the XSB programming language conceived to support tabled logic programming.

The starting point of this study is the definite clause grammar with its Prolog implementation, including its early drawbacks (looping and exponential backtracking). The presentation discusses the subsequent means to mitigate these problems, to augment the semantic content and the syntactic rules and improve grammar formalism. One development is the introduction of assumption grammars, which are able to handle long distance dependency phenomena, like interrogative clauses, anaphora, and so on, through global accumulators. A second step is the introduction of logical constraint handling rules (CHR) and the CHR grammars. Probabilistic logic programming for language processing, a more recent trend in NLU, is discussed together with its PRISM implementation, as it relies on an underlying tabling system.

The generously added historical preliminary notes are very interesting. Although the scope of this chapter does not allow for a very deep investigation of the entire subject, comprehensive bibliographic information is provided.