The Accreditation Council for Graduate Medical Education (ACGME) is encouraging medical residency programs to objectively assess their trainees for possession of six general clinical competencies by the completion of residency training. This is the thrust of the ACGME Outcome Project, now in its seventh year. As residency programs seek to integrate the general competencies into clinical training, educators have begun to suggest that objective assessment of clinical competence may be able to guide decisions about length of training and timing of subspecialization. The authors contend that higher-level competence is not amenable to assessment by the objective comparison of resident performance with learning objectives, even if such objectives are derived from general competencies. Present-day attempts at such assessment echo the uses to which medical schools hoped to put curricular learning objectives in the 1970s. Objective assessment may capture knowledge and skills that amount to the "building blocks" of competence, but it cannot elucidate or scrutinize higher-level clinical competence. Higher-level competence involves sensitivity to clinical context and can be validly appraised only in such a context by fully competent clinical appraisers. Such assessment is necessarily subjective, but it need not be unreproducible if raters are trained and if sampling of trainee performance is sufficiently extensive. If the ACGME approach to clinical competency is indeed brought to bear on decisions about training length and subspecialization timing, the present apprenticeship model for clinical training in the United States, a model both remarkably successful and directly descendant from Osler's innovations, will be under threat.