Geoffrey Nunberg challenges a widespread assumption that the linguistic structure of written languages is qualitatively identical to that of spoken language: It should no longer be necessary to defend the view that written language is truly language, but it is surprising to learn of written-language category indicators that are realized by punctuation marks and other figural devices.' He shows that traditional approaches to these devices tend to describe the features of written language exclusively by analogy to those of spoken language, with the result that punctuation has been regarded as an unsystematic and deficient means for presenting spoken-language intonation. Analysed in its own terms, however, punctuation manifests a coherent linguistic subsystem of 'text-grammar' that coexists in writing with the system of 'lexical grammar' that has been the traditional object of linguistic inquiry. A detailed analysis of the category structure of English text-sentences reveals a highly systematic set of syntactic and presentational rules that can be described in terms independent of the rules of lexical grammar and are largely matters of the tacit knowledge that writers acquire without formal instruction. That these rules obey constraints that are structurally analogous to those of lexical grammar leads Nunberg to label the text-grammar an 'application' of the principles of natural language organization to a new domain. Geoffrey Nunberg is a researcher at Xerox Palo Alto Research Center.

In the last decade, computational linguistics has produced a revival of the interest in the mathematical study of the various levels of human language. This volume contains a selection of recent research papers approaching mathematical and computational topics in natural languages, with a special attention being paid to syntax and semantics. According with their main focus, the papers are distributed into four parts: Syntax, Semantics, Natural language processing and Varia, which cover a vast range of problems. The book may be of interest to all those who intend to know which kind of mathematics is used when giving account of natural language, as well as to people working on computational issues involving human-machine interaction.

Anyone writing texts in English is constantly faced with the unavoidable question whether to use open spelling (drinking fountain), hyphenation (far-off) or solid spelling (airport) for individual compounds. While some compounds commonly occur with alternative spellings, others show a very clear bias for one form. This book tests over 60 hypotheses and explores the patterns underlying the spelling of English compounds from a variety of perspectives. Based on a sample of 600 biconstituent compounds with identical spelling in all reference works in which they occur (200 each with open, hyphenated and solid spelling), this empirical study analyses large amounts of data from corpora and dictionaries and concludes that the spelling of English compounds is not chaotic but actually correlates with a large number of statistically significant variables. An easily applicable decision tree is derived from the data and an innovative multi-dimensional prototype model is suggested to account for the results.

American English Grammar introduces students to American English in detail, from parts of speech, phrases, and clauses to punctuation and explaining (and debunking) numerous "rules of correctness," integrating its discussion of Standard American grammar with thorough coverage of the past sixty years’ worth of work on African American English and other ethnic and regional non-Standard varieties. The book’s examples and exercises include 500 real-world sentences and longer texts, drawn from newspapers, film, song lyrics, and online media as well as from Mark Twain, Stephen King, academic texts, translations of the Bible, poetry, drama, children’s literature, and transcribed conversation and TV and radio shows. Based on twenty years of classroom testing and revision, American English Grammar will serve as a classroom text or reference that teaches students how to think and talk not only about the mechanics of sentences but also about the deep and detailed soul and nuance of the most widely used language in human history.

The completely redesigned Grammar of Spoken and Written English is a comprehensive corpus-based reference grammar. GSWE describes the structural characteristics of grammatical constructions in English, as do other reference grammars. But GSWE is unique in that it gives equal attention to describing the patterns of language use for each grammatical feature, based on empirical analyses of grammatical patterns in a 40-million-word corpus of spoken and written registers. Grammar-in-use is characterized by three inter-related kinds of information: frequency of grammatical features in spoken and written registers, frequencies of the most common lexico-grammatical patterns, and analysis of the discourse factors influencing choices among related grammatical features. GSWE includes over 350 tables and figures highlighting the results of corpus-based investigations. Throughout the book, authentic examples illustrate all research findings. The empirical descriptions document the lexico-grammatical features that are especially common in face-to-face-conversation compared to those that are especially common in academic writing. Analyses of fiction and newspaper articles are included as further benchmarks of language use. GSWE contains over 6,000 authentic examples from these four registers, illustrating the range of lexico-grammatical features in real-world speech and writing. In addition, comparisons between British and American English reveal specific regional differences. Now completely redesigned and available in an electronic edition, the Grammar of Spoken and Written English remains a unique and indispensable reference work for researchers, language teachers, and students alike.