Although the Modern Language Association and other style manuals describe in exquisite detail ''how'' to cite the literature, explicit tutorials on ''when'' to cite are nonexistent. Most journals provide instructions to authors but also fail to give explicit guidance on when to cite. In spite of numerous studies of citation behavior and the wide recognition by editors of the need to acknowledge intellectual debts, authors and referees need explicit reminders as to when formal references or acknowledgments are appropriate. Since referencing is both subjective and culturally based, there can be no absolutes about when to cite. Hence, it is unlikely that algorithmic documentation of texts can ever meet the competing requirements for relevance, selectivity, and comprehensiveness. What is common wisdom in one domain may be new or unique in another. A three-year experiment involving graduate students demonstrated the varying perceptions of the need for documentation of terminology, ideas, methods, and so forth. A tentative tutorial is suggested for journal editors that should be modified in each scholarly context.