Academic publishing is undergoing profound changes that shape the conditions of knowledge production and the way research is communicated, prompting a lively debate on how the various activities of those involved can be adequately acknowledged in publications. This contribution aims to empirically examine the relationship between authorship regulations in journal policies, the disciplinary variance in authorship practice and larger concepts of academic authorship. Analyzing: (1) editorial policies; and (2) data from an interdisciplinary survey of scientists, we examine to what extent disciplinary variances are reflected in the policies as well as in researchers' individual understandings. Here we find that the regulation of authorship qua policies is primarily effected at the level of the publishers. Although considerable disciplinary variations of journal policies are sometimes suggested in the literature, we find only minor differences in authorship criteria. The survey data however show that researchers' understandings of authorship exhibit significant, discipline-specific differences, as well as differences related to the characteristics of the research practice. It hence becomes clear that discipline-specific conditions of knowledge production with the resulting differences in authorship practices are hardly reflected in authorship policies. We conclude that the regulatory ambitions of authorship policies mostly focus on the prevention and elimination of deficits in the quality and integrity of scientific publications. Thus, it seems questionable whether authorship policies in their current form are suitable instruments for mediating between diverse authorship practices and normative ideals of legitimate authorship.