What constitutes valid evidence from medical education research is typically grounded in the scientific paradigm of proof through experiment. Here, explanation through single meaning is privileged over exploration of multiple presentations of phenomena--short, interpretation eclipses appreciation. This approach is challenged as reductive by naturalistic qualitative methods such as rich ethnographic field reports, presented as narratives. Contemporary ethnographic approaches have entered medical education by a back door--disguised as a stable of 'social learning theories'. Communities of Practice theory, Activity Theory and Actor-Network-Theory (ANT) all serve as research practices forming identifiable contemporary ethnographies. 'Evidence' is conceived as exploratory rather than explanatory, through baroque descriptions of innovations in learning organizations, including medicine. ANT is then both a theory of innovation in organizations and an ethnographic method, where practice and theory coincide. ANT is interested primarily not in epistemologies, but in how a phenomenon such as an 'illness' is conceived across differing practices as multiple ontologies (experienced meanings), each meaning generated and suspended within a particular network of effects. How such networks are initiated and developed has significance for rethinking the nature of 'evidence', restoring faith in the value of a good story.