What exactly is stabilized when phenomena are stabilized?

In their seminal article " Saving the Phenomena " , James Bogen and James Woodward argue that phenomena, not data, are the explananda of scientific theories (Bogen & Woodward 1988). By suggesting that phenomena are " relatively stable and general features of the world " , (Woodward 1989, 393), which are " detectable by means of a variety of different procedures " (Bogen & Woodward 1988, 317), the authors draw attention to the role of data as evidence in the process of " [u]nambiguously establishing the existence and character of a certain range of phenomena " (Woodward 1989, 446). This move is significant because it puts on the philosophical agenda questions about the production of experimental data in contexts that are not comfortably described as " theory-testing ". As the authors suggest, claims about the existence of phenomena can be – and frequently are – confirmed even if there is (a) no theory that explains the phenomenon, and (b) no theory that explains how the phenomenon produces the data. Bogen and Woodward's analysis is representative of a particular line of argumentation within the philosophy of experimentation, which has attributed to experiments a certain independence of theories (thereby countering claims about theoryladenness), while emphasizing that knowledge can be produced by means of various experimental strategies (Hacking 1983, Franklin 1993). While the distinction between data and phenomena has a high degree of prima facie plausibility, the notion of a phenomenon remains contentious in the literature. In particular, the question has been raised of whether phenomena are really investigator-independent (McAllister 1997) and whether the data/phenomena dichotomy really captures a distinction relevant to scientific practice (Glymour 2000). In my paper I will argue that the question of whether Bogen and Woodward may help themselves to an investigator-independent notion of phenomenon is ill posed, and that the data/phenomena-distinction does indeed capture a distinction that is important to scientific practice. I will make some suggestions on how to sharpen the distinction. My paper will proceed in three steps: First, I argue that the issue of whether Bogen and Woodward are entitled to an investigator-independent notion of phenomenon conflates two issues, namely (i) whether they are entitled to the metaphysical presupposition that there are investigator-independent phenomena in the world, and (ii) whether they are entitled to the epistemological claim that scientists always, or usually, succeed in identifying them. On my reading, the former assumption is simply presupposed, …