The terms "citizen science" and "citizen scientist" have at least three meanings: (1) the participation of nonscientists in the process of gathering data according to specific scientific protocols and in the process of using and interpreting that data; (2) the engagement of nonscientists in true decision-making about policy issues that have technical or scientific components; and (3) the engagement of research scientists in the democratic and policy process. Looking just at the first definition, proponents of citizen science argue that it engages nonscientists in the scientific process, making them direct participants in the creation of reliable knowledge about the natural world. From an S&TS perspective, many statements in the preceding paragraph pose problems. What is meant by "engagement"? Can we distinguish between the "technical," "scientific," and "policy" components of a decision? How much social process is hidden by the phrase "specific scientific protocols"? What is meant by "the scientific process"? What constitutes "reliable knowledge"? What is the "natural world" and how does it differ from other conceptions such as the "social world"? In this paper, I will use attempts to define the "outcome" of several specific citizen science projects at the Cornell Laboratory of Ornithology as a case for exploring the meanings of these various phrases. I will suggest that defining "success" for citizen science projects requires, in part, that the scientific community that supports citizen science must leave behind its Mertonian ideals about the independence of science and adopt instead S&TS-inspired conceptions of the social embeddedness of scientific knowledge.