Jean-Philippe Bernardy

A textbook exploring predictive modes of linguistic development and analysis. During the last two decades, computational linguists, in concert with other researchers in AI, have turned to machine learning and statistical techniques to capture features of natural language and aspects of the learning process that are not easily accommodated in classical algebraic frameworks. These developments are producing a revolution in linguistics in which traditional symbolic systems are giving way to probabilistic and deep learning approaches. This collection features articles that provide background to these approaches, and their application in syntax, semantics, pragmatics, morphology, psycholinguistics, neurolinguistics, and dialogue modeling. Each chapter provides a self-contained introduction to the topic that it covers, making this volume accessible to graduate students and researchers in linguistics, NLP, AI, and cognitive science.

The main idea behind this book is to encourage readers to approach mathematical domains from a functional programming perspective: to identify the main functions and types involved and, when necessary, to introduce new abstractions; to give calculational proofs; to pay attention to the syntax of the mathematical expressions; and, finally, to organize the resulting functions and types in domain-specific languages. The book is recommended for developers who are learning mathematics and would like to use Haskell to make sense of definitions and theorems. It is also a book for the mathematically interested who wants to explore functional programming and domain-specific languages. The book helps put into perspective the domains of Mathematics and Functional Programming and shows how Computer Science and Mathematics are usefully taught together.