Semantics: From meaning to text by Igor A. Mel’čuk (review)

This book is the first volume of Igor Mel’čuk’s long-awaited monograph building on his Meaning–Text Theory (MTT) (developed with Alexander Žolkovskij) more than four decades after its launch. MTT was conceived in the context of incipient machine translation research back in the Soviet Union of the 1960s. This theory gave the world the revolutionary idea of lexical functions and allowed for a new description of language in terms of lexemes and their combinatorial qualities. Aimed at developing interlingua for the purposes of machine translation, this research sought to present the meaning of utterances in a non-linear way in order to break free from the syntactic rules of specific natural languages. Further, MTT pursued working out a mechanism to solve the problem of the distinctive divisions of reality that exist between human languages. The description of lexemes and their lexical functions not only gave rise to Explanatory–Combinatory Dictionaries (ECDs) of Russian (Mel’čuk and Zholkovsky 1984) and French (Mel’čuk et al. 1984), among others, but also paved the way for the identification of multiple lexical functions across languages, which nowadays form part of machine translation technology (Apresian et al. 2003). This volume is intended as the first of three volumes that form part of an ambitious project where the author plans to integrate not only his own work but also that of his collaborators, including Alain Polguère and David Beck.

The book starts with Beck’s Foreword, which is followed by “Abbreviations and notations”, “Organization of SMT [Semantics: From meaning to text]”, and “General introduction”, which outlines the boundaries of MTT and provides a brief description of its history.

Chapter 1, “Meaning–Text Approach and Meaning–Text Models”, provides definitions for a series of basic linguistic notions like speaker, meaning, utterance, paraphrase, and other concepts more specific to MTT.

Chapter 2 is concerned with the notion of linguistic paraphrasing. It addresses three types of meaning: propositional, communicative, and rhetorical, of which only propositional meaning is responsible for two phrases to be considered paraphrases. To extract this meaning a set of (quasi-synonymous) paraphrases is used (Mel’čuk 1999 [1974], Milićević 2007). The author presents a typology of paraphrases classified according to four parameters. First, paraphrases can be obtained either vertically from deeper structures or horizontally by a special component of the MTT model called the “Paraphrasing System” (referred to as generator in Mel’čuk 1999:193 and to be described in detail in the second volume). Second, a distinction should be made between linguistic and cognitive paraphrases. MTT does not allow for the generation of the latter. Third, paraphrases can be exact or approximate. To deal with the latter, the author introduces the concept of semantic neutralization. Fourth, depending on the level of linguistic representation at which paraphrases are produced, another distinction should be made between semantic, deep-syntactic, and surface-syntactic paraphrases.

Chapter 3 is of paramount significance within the book as it presents the MTT with its main postulates as well as the restrictions assumed by the author. The first postulate, Language as a meaning-text correspondence, describes a natural language as a finite set of rules mediating between meanings (SemR) and texts (SPhonR). The second postulate, Linguistic description as a functional model, claims that text generation should rest on a system of linguistic declarative static rules and the related system of procedural dynamic rules. The third postulate, Multiple levels of linguistic representation, introduces two intermediate levels: SyntR and MorphR, each of which further split into deep (D) and surface (S) sublevels. The relation between SemR and DSyntR is regarded as the main object of this book. While SemR is studied in depth in this volume, a thorough discussion of DSyntR is left for the next volume.

Chapter 4 addresses the purely semantic structure of the utterance or, more accurately, the propositional structure of the set of all utterances that express the same propositional...