Guidelines on authorship requirements are common in biomedical journals but it is not known how authorship is defined by journals and scholarly professional organizations across research disciplines. Prevalence of authorship statements, their specificity and tone, and contributions required for authorship were assessed in 185 journals from Science Citation Index (SCI) and Social Science Citation Index (SSCI), 260 journals from Arts & Humanities Citation Index (A&HCI) and 651 codes of ethics from professional organizations from the online database of the Center for the Study of Ethics in the Profession, USA. In SCI, 53 % of the top-ranked journals had an authorship statement, compared with 32 % in SSCI. In a random sample of A&HCI-indexed journals, only 6 % of the journals addressed authorship. Only 71 (11 %) codes of ethics carried a statement on authorship. Almost all journals had defined authorship criteria compared with 33 % of the ethics codes ( = 75.975; P < 0.001). The tone of the statements in the journals was aspirational, whereas ethics codes used a normative language for defining authorship ( = 51.709, P < 0.001). Journals mostly required both research and writing contributions for authorship, while two-thirds of the ethics codes defined only research as a mandatory contribution. In conclusion, the lack of and variety of authorship definitions in journals and professional organizations across scientific disciplines may be confusing for the researchers and lead to poor authorship practices. All stakeholders in research need to collaborate on building the environment where ethical behaviour in authorship is a norm.