Exploring the Lexical Organization of English

This book presents two empirical studies that explore the possible relationship between the semantic similarity of two lexical items and their collocational range. Confronted with the seemingly arbitrary rules of language, the translator continually seeks effective strategies to assist him in overcoming the daunting problems posed by collocations in language encoding situations. Foremost, the translator is uncertain if he may replace the base word in a known collocation with a semantically-similar word without violating the rules.
The New York Times online archive provides a sound basis for an initial illustration of semantic-based collocation research which analyzes both the usage and collocability of six English verbs of thinking. Subsequently, a text corpus is compiled, consisting of more than 37 million words derived from a full year of published news from The Washington Post. This corpus is analyzed with regard to the usage and collocability of four English semantic fields (love, result, trouble, and work), which are represented by twelve nouns each. The method of card sorting, combined with an agglomerative cluster analysis, assists in gaining insights into the meaning structures of semantic fields. Two significant themes emerge: The collocation reveals itself as an intangible entity, and the collocational ranges of base words far exceed the translator's expectations. Finally, COLLCATION ANALYZER is presented: a new and fully-functional electronic collocation retrieval and verification tool containing all of the collocations extracted for the four semantic fields. It is a powerful tool designed to fulfill the translator's urgent need for a reliable electronic collocation dictionary.