This paper first briefly reviews what Halliday and Hasan (1976) said about "(the) same." The paper then examines the understanding of this form by qualitatively analyzing 259 naturally occurring spoken tokens of "(the) same" in their discourse contexts. It focuses on the following questions with reference to the data: (1) What determiners can occur before "same" besides "the"? (2) When can the determiner "the" be deleted? (3) What is the structure of tokens where "(the) same" occurs without a Head noun? (4) How is "(the) same" used in comparative reference and nominal substitution? and (5) Why is Conversation Analysis (CA) potentially very important when analyzing cohesive devices in conversation? The paper proposes that, based on data from conversation, the most frequent grammatical function of "same" in conversation is adjectival within a nominal group with a Head noun and not, as Halliday and Hasan's account suggests, occurring alone in an elliptical noun head in a nominal group. It calls for more work to be done in terms of linking a conversation analytic framework with issues and concerns in cohesion analysis. Selected transcript from "The Nixon Watergate Tapes" (1974) is appended. (NKA)