The operation of voluntary processes can contaminate the study of Pavlovian conditioned responses in humans. The problem of voluntary control had apparently been solved by about 1940, particularly in human eyelid conditioning. Nonetheless, the problem returned in the early 1950s, calling forth a variety of methodological procedures for removing voluntary responses and/or voluntary-responding subjects from eyelid-conditioning data. During the 1960s, disagreement arose regarding the efficiency and comparability of alternative data-correction procedures; the rationale for data-correction; and whether, and under what experimental (...) conditions, eyelid-conditioning data ought to be corrected for voluntary responding. Circumstances that supported voluntary-response methodology are described, and points at which the methodology conflicted with the technical background of Pavlovian conditioning are noted. A diagnosis is offered for the failure, between 1950 and 1980, to solve again the problem of volition. (shrink)