The rapid increase of computer-mediated communications (CMCs) in various forms such as micro-blogging (e.g. Twitter), online chatting (e.g. digital reference) and community-based question-answering services (e.g. Yahoo! Answers) characterizes a recent trend in web technologies, often referred to as the "social web". This trend highlights the importance of supporting linguistic interactions in people's online information-seeking activities in daily life--something that the web search engines still lack because of the complexity of this human behavior. The presented research consists of an investigation of the information-seeking behavior of digital reference services through analysis of discourse semantics, called dialogue acts, and experimentation of automatic identification of dialogue acts using machine-learning techniques. The data was an online chat reference transaction archive, provided by the Online Computing Library Center (OCLC). Findings of the discourse analysis include supporting evidence of some of the existing theories of the information-seeking behavior. They also suggest a new way of analyzing the progress of information-seeking interactions using dialogue act analysis. The machine learning experimentation produced promising results and demonstrated the possibility of practical applications of the DA analysis for further research across disciplines. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://bibliotheek.ehb.be:2222/en-US/products/dissertations/individuals.shtml.]