I assert then, in plain and distinct terms, that in a properly constructed building, with a sufficient number of suitable attendants, restraint is never necessary, never justifiable, and always injurious, in all cases of lunacy whatever" (1). Thus declared "house surgeon" Robert Gardiner Hill of the Lincoln Asylum in an oft-quoted public lecture delivered in Lincoln, England, in 1838. By this point he was immune to incredulous responses, which he caricatured as "What! Let loose a Madman! Why he will tear us to pieces!" For him, the proof of the proverbial English pudding was in the eating. In 1833, a total of 12,003 hours and 1,109 instances of restraint were recorded for 44 of 87 patients in the asylum at Lincoln (1). Five years later, with an expanded census of 148, no instances of restraint were recorded for the entire year.

The use of restraint has continued to be both a source of controversy and a barometer for commitment to therapeutic reform since Hill's day. In 19th-century America, asylum physicians regularly debated the role and value of restraint devices. Many of their charges had indeed torn attendants to pieces. On the other hand, asylum leaders were concerned that the often poorly paid and poorly educated attendants Hill relied on might pose greater dangers in modifying patients' behavior than did the restraints themselves. Caught between the need to control certain behaviors, the realities of growing asylum populations and inadequately prepared staff, and the contradictions between the use of restraint and the reigning philosophy of moral insanity, leaders of treatment of the insane were often resigned to seeing restraint as a necessary evil (2).

Again in the early 20th century, and variably since then, the issue of restraint has assumed a prominent place as the prism through which the nature and ethics of the therapeutic purpose of mental health care is often scrutinized. From emergency departments to inpatient units, accreditation and regulatory attention has focused anew on moving beyond the need to use restraints safely and under certain conditions of necessity after other approaches have failed. There can also be heard again the echoes of Hill's declaration that "restraint is never necessary." Several state mental health departments have begun to take that task on as a specific goal of policy, with real results in child and adolescent settings in the state of Massachusetts, for example.

But as in Hill's day, and through the subsequent centuries of debate over the purposes of the asylum, the hospital, and the acute-stay crisis or emergency service, the moral resonance that surrounds discussions of restraint is very much tied to larger issues of therapeutic identity, purpose, and possibility. Hill and his fellow adherents to moral treatment—those self-described disciples of Pinel and Tuke who set out to reconsider fundamental assumptions about the nature of mental illness and the possibilities for intervening to change its course—tied their advocacy for the end of "instruments of restraint" to a theory of mental illness that had implications for the larger conduct of institutions and the care of persons with mental illness.

They sought, at least in theory and often in reality, to replace physical constraints on behavior with an asylum community that exercised and strengthened internal controls by setting up expectations of right conduct. These expectations were established through vigilant modeling and mirroring of ideal behavior. Again, Hill: "In short, what is the substitute for coercion? The answer may be summed up in a few words—classification—watchfulness—vigilant and unceasing attendance by day and by night—kindness, occupation, and attention to health, cleanliness, and comfort" (1).

Moral psychiatry and the efforts of Hill and his colleagues are usually portrayed within psychiatry as a great leap forward in attempting to replace external controls with internal ones through respectful vigilance and opportunities for self-improvement and respect. However, moral psychiatry has received more skeptical treatment from some historians of psychiatry. Most notable among such criticisms is Michel Foucault's charge that psychiatry, in the guise of removing one set of restraints, only succeeded in offering a new one. Chains and manacles were traded in for constant scrutiny, objectification within a scientific gaze, and loss of unique identity as recovery became the expectation of conforming to idealized norms of behavior established and elaborated by new institutions. Appeal to internal controls was perhaps, on second look, the imposition of norms through shamed displacement of responsibility for behavior (3).

Although Foucault's position has been criticized for its historical overgeneralization and simplification of the experience of illness and recovery, historians have nonetheless explored in great detail the idea that psychiatric progress has regularly had to deal with or conceal the degree to which therapeutic advance has involved merely reconstituted forms of control. The varied use and criticism of restraint reflect the tension surrounding the degree to which psychiatric work involves coercion and authority over the actions of others, a tension that is much harder to remove or eliminate than the restraints themselves. The risk of replacing external controls with somatic interventions that produce the same outcome can yield results that in retrospect court our censure. Historian and psychiatrist Joel Braslow notes, "With nearly every new therapeutic innovation, doctors rang the death knell of physical restraint, claiming they no longer had need of such a crude and nontherapeutic intervention. Physicians made this assertion with the introduction of hydrotherapy, the shock therapies, lobotomy, and antipsychotic drugs. However, reports of its death were greatly exaggerated. Like an unwanted dinner guest who refuses to leave, restraint remained a necessary, albeit disliked, adjunct to therapeutic practices…. In spite of heaping disdain on physical restraint, doctors ascertained a given remedy's effectiveness by how well it measured up against these simplest methods of bodily control" (4).

Like everyone else, mental health professionals are inclined to believe that we live in a particularly advanced time in history, and we view previous methods of managing problems as less sophisticated. However, it is at least arguable that current forms of treatment for much of the illness we treat are no more effective and no less steeped in the accepted science of the time than those that we ridicule from decades past (5,6,7). Comparing effectiveness over broad swaths of time is treacherous, because disease paradigms, social context, and expectations for—and the environment of—recovery all vary as well. The point is that vigilant self-reflection should guide the way scientific rationales are used to spin the alternative clinical formulations and routines of scrutiny that may replace mechanical restraint. This means facing, not hiding within unexamined claims about better scientific evidence, the ways in which our work struggles with what often seems an inevitable bit of exertion of discipline over the mind and the degree to which that work is "measured up against these simplest methods of bodily control."

Dr. Belkin is assistant professor of psychiatry at Harvard Medical School and assistant in psychiatry at Massachusetts General Hospital in Boston. Send correspondence to him at the Office of the Medical Director, Erich Lindemann Mental Health Center, 25 Staniford Street, Boston, Massachusetts 02114 (e-mail, gbelkin@partners.org). Douglas H. Hughes, M.D., is editor of this column.