Strategies and language trends in learning success and failure of negotiation

Abstract

In negotiation by electronic means, language is an important deal-making tool which helps realize negotiation strategies. Negotiators may use language to request information, exchange offers, persuade, threaten, as well as reach a compromise or find prospective partners. All this is recorded in texts exchanged by negotiators. We explore the language signals of strategies—argumentation, persuasion, negation, proposition. Leech and Svartvik’s approach to language in communication gives our study the necessary systematic background. It combines pragmatics, the communicative grammar and the meaning of English verbs. Language signals become features in the task of classifying those texts. We employ Statistical Natural Language Processing and Machine Learning techniques to find general trends that negotiation texts exhibit. Our hypothesis is that language signals help predict negotiation outcomes. We run experiments on the Inspire data. The electronic negotiation support system Inspire was gathering data for several years. The data include text messages which negotiators may exchange while trading offers. We conduct a series of Machine Learning experiments to predict the negotiation outcome from the texts associated with first halves of negotiations. We compare the results with the classification of complete negotiations. We conclude the paper with an analysis of the results and a list of suggestions for future work.