This paper was prepared based on an initial overall framework paper developed for the TOEFL 2000 project by Jamieson, Jones, Kirsch, Mosenthal, and Taylor (2000). The paper applies concepts advanced in the overall paper to the modality of speaking. In doing so, this document presents an initial framework for research and development activities for the speaking component of the TOEFL 2000 test. The paper should be viewed as a work-in-progress, as research activities presently underway for TOEFL 2000 will undoubtedly bring about refinements to the contents of this document. The paper is made up of six parts. Part 1 provides an introduction to the overall document. In Part 2, oral discourse is discussed from a sociological perspective as well as in terms of speech act theory. Part 3 discusses the details of the speaking framework for the TOEFL 2000 test. Identification of the test domain and relevant task characteristics and variables is discussed, along with some of the factors suspected to influence the difficulty of speaking tasks. In Part 4, some of the technical issues involved in eliciting and capturing speech samples are discussed. Part 5 contains a list of relevant research activities that should be pursued as the project progresses. The final part considers ways in which the new TOEFL 2000 speaking component will improve upon the current version of the TOEFL test and the Test of Spoken English (TSE).