In the debate over what 'social control' means, the link between control and measurement of that control is lost. Whether control is achieved depends on how the matter to be controlled is defined. To make the issues clear, it helps to examine the etymological root of 'control.' The late Latin origin of 'control' is 'contarotulus' or 'counterroll' or 'counter register.' The earliest cited keeper of counterrolls was the English sheriff, who began keeping copies of the coroner's roll in the early 13th century. In theory, the counterroll had three basic requisites of control: (1) measurement against an arbitrary standard of need for action, (2) action hypothesized more likely than not to move the measured trend in a preferred direction, and (3) repetition of the measurement to see whether the trend has shifted as hypothesized. Since an organization has conventional limits within which the bureaucrat is relatively immune from outside scrutiny of records, it is fallacious to treat the counterroll as a standard of control. In terms of crime measurement, it has been found that police offense rates understate crime according to victim survey figures and that victimization rates understate crime according to police records. Given the possibility of overreporting, each becomes a flawed criterion of the validity of the other. 'Confinement' is one definition of what crime and criminality data measure that seems to be tautologous with the measures we have and use. This means that all measures refer to persons or incidents believed by the reporters to be subject to containment (confinement) by criminal justice employees. Thirty-six references are listed.