Source / Research ethics judgment

The Nuremberg Code

The Nuremberg Code is the later name for ten principles of permissible human experimentation read by judges in Nuremberg on 19 August 1947. They formed part of the judgment in United States v. Karl Brandt et al., the Doctors' Trial of physicians and administrators implicated in coercive experiments and mass killing under National Socialism.

The text made voluntary consent, scientifically justified design, limits on risk, and the investigator's continuing duty matters of moral and legal responsibility. It was neither the first attempt to regulate experiments nor a self-executing international law; its historical importance lies in how a criminal court joined individual autonomy to accountability after survivors documented organized medical violence.

Reading the judgment

Consent begins the code, but it does not exhaust it.

The judges introduced the list as principles needed to satisfy “moral, ethical and legal concepts.” The period term “human subject” is retained when describing the judgment; “participant” is used here for people in research. Although later writers often call its first rule “informed consent,” that phrase does not appear in the 1947 text. The judgment says “voluntary consent” and then specifies what makes consent meaningful.

1. Consent requires capacity, freedom, and understanding

The first and longest principle calls voluntary consent “absolutely essential.” It requires legal capacity, freedom from force, fraud, deceit, duress, and other coercion, plus enough knowledge to make an “understanding and enlightened decision.” Before agreement, the participant is to know the experiment's nature, duration, purpose, methods, expected inconveniences and hazards, and possible effects. Responsibility rests personally on everyone who initiates, directs, or engages in the experiment; it cannot simply be passed to a recruiter, institution, or state.

2–4. Scientific validity is an ethical requirement

An experiment should seek socially valuable results unavailable by other methods and should not be random or unnecessary. Its design should rest on animal work and knowledge of the disease or problem, and anticipated results must justify proceeding. Avoiding unnecessary physical and mental suffering is therefore connected to method: badly designed research exposes people without producing the knowledge claimed for it.

5–8. Benefit does not license unlimited risk

The text rejects an experiment when there is prior reason to expect death or disabling injury, followed by an ambiguous exception for experiments in which physicians also serve as subjects. Risk must not exceed the “humanitarian importance” of the problem. Preparations and facilities should protect against even remote possibilities of serious harm, and scientifically qualified investigators must exercise the highest skill and care throughout.

9–10. Permission remains revocable

A participant may end the experiment when their physical or mental condition makes continuation seem impossible. The scientist in charge must also stop when good faith, skill, and careful judgment provide probable cause to expect injury, disability, or death. These paired duties make ethical authority continuous rather than something secured once by a signature.

The digitized judgment transcript preserves the sequence and wording, including typographical and optical-recognition errors in the online text. The ten points should be read with the paragraph that follows them: the judges said their immediate concern was those requirements sufficiently legal to help determine criminal responsibility, not the creation of a complete system of research administration.

Before 1947

Nuremberg did not invent concern about experimental consent.

Earlier rules existed. Public controversy over Albert Neisser's non-consensual syphilis research contributed to a Prussian directive in 1900, reconstructed by Jochen Vollmann and Rolf Winau. In 1931 the German Reich Ministry of the Interior issued more detailed Richtlinien, or guidelines, for new therapy and human experimentation. They distinguished therapeutic innovation from non-therapeutic experiment, addressed consent and prior animal work, and assigned professional responsibility. Hans-Martin Sass's translation and historical analysis shows that these regulations remained formally binding during the Nazi period.

Rules did not protect people when institutions embraced persecution. The existence of the 1931 guidelines disproves a simple “no ethics before Nuremberg” story; it does not show that German research was humane. Nazi racial ideology, wartime agencies, universities, the SS, armed-services medical organizations, and concentration camps supplied authority, personnel, facilities, and captive people. The judgment itself rejected the idea that the crimes were isolated acts by a few uncontrolled doctors, describing coordinated governmental, military, and party planning.

Victims challenged the experiments before and during the trial. Historian Paul Weindling documents protest, evasion, sabotage, survivor reports, and declarations by prisoner doctors after liberation. At Ravensbrück, Polish prisoners subjected to wound and sulfonamide experiments called themselves króliki (“rabbits”) and resisted further operations. Their agency matters: ethical condemnation did not begin only when Allied physicians or American judges supplied words for it.

Trial chronology

Survivor evidence, expert proposals, and judicial revision produced the ten points.

1945–August 1946: investigation before prosecution. Liberated prisoners and prisoner physicians reported experiments at Auschwitz, Bergen-Belsen, Buchenwald, Dachau, Ravensbrück, and other camps. Allied scientific-intelligence officers had to connect medical records, material evidence, institutions, and individual perpetrators. At a Paris meeting on medical war crimes on 31 July–1 August 1946, U.S. physiologist Andrew C. Ivy outlined consent, prior evidence, social usefulness, avoidance of suffering, qualified investigators, and a bar on predictably fatal or disabling research. This draft helped make an ethics standard part of the proposed case.

9 December 1946: the Medical Case opened. Twenty-three defendants—twenty physicians and three non-physician administrators—were prosecuted before an American, not a four-power international, tribunal. Charges included war crimes and crimes against humanity arising from high-altitude, freezing, infection, wound, poison, sterilisation, and other experiments, as well as participation in killing programmes. The categories were those of a criminal indictment; they neither encompass every Nazi medical crime nor recover every victim.

20 December 1946: injuries and resistance entered the courtroom record. Polish survivor Jadwiga Dzido testified that Ravensbrück prisoners were not asked to volunteer, that women resisted renewed operations, and that she still required treatment and special footwear she could not afford. She displayed her injured leg while prosecution medical expert Leo Alexander interpreted scars, muscle loss, impaired movement, and radiographs. The court transcript is indispensable testimony, but it is also a translated record structured by prosecutors' questions and Alexander's retrospective examination. His medical inferences should not be confused with everything Dzido herself knew or said.

December 1946–April 1947: proposed rules were revised. Ivy, appointed after nomination by the American Medical Association, and Alexander, a U.S. Army physician advising the prosecution, each developed formulations. Alexander completed an early memorandum in December and submitted a six-point version to the prosecution on 17 April. Defendants and expert witnesses argued about voluntariness, state authority, professional custom, and precedents in other countries. Scholarship has sometimes credited either Ivy or Alexander as sole author. The documentary reconstruction instead supports cumulative authorship: Ivy supplied an early framework, Alexander expanded it, courtroom evidence tested it, and the three judges issued and were responsible for the final ten points.

19–20 August 1947: principles, findings, and sentences. On 19 August Judge Harold L. Sebring read the “Permissible Medical Experiments” section. The tribunal then compared the proven experiments with the ten principles and found pervasive absence of consent and withdrawal, unnecessary suffering, inadequate safeguards, and disabling or lethal harm. Individual judgments were completed the next day: sixteen defendants were convicted and seven acquitted; seven received death sentences. The code was part of the court's reasoning, but guilt still required proof connecting each defendant to crimes under the tribunal's governing law.

Name, influence, and limits

A publicized judgment became a landmark without becoming a universal statute.

The familiar title is part of the afterlife

The judgment did not label its principles “The Nuremberg Code.” Weindling dates regular use of that name from about 1963. He also challenges the opposite myth that the principles were unknown before then: they circulated in English, French, and German and appeared in post-trial publications. Naming nevertheless helped turn one section of a judgment into a portable ethical document.

Publicity and practical influence are different questions

Research codes and histories often place Nuremberg before the World Medical Association's 1964 Declaration of Helsinki and later national regulation. That sequence should not be treated as automatic implementation. Jenny Hazelgrove's study of Britain from 1946 to 1973 finds that wartime experience, professional structures, and persistent medical paternalism shaped researchers more directly than the code. Different countries and institutions received and operationalised its principles unevenly.

Its direct legal status remained narrow

The principles were pronounced by a court applying criminal law to particular defendants; they were not a treaty or an enacted general research statute. Ulf Schmidt and Jonathan Moreno note that no government adopted the whole code as domestic law, while the U.S. Department of Defense partially adopted it for certain research in 1953. Courts, regulators, professional bodies, and ethics declarations have cited or echoed it, but moral authority, historical influence, and enforceable law are not interchangeable.

Ten principles are not an oversight system

The text says nothing about an independent review committee, fair selection of participants, compensation for research injury, community involvement, or public registration and reporting. Its language on legal capacity leaves unresolved how research involving children or adults unable to consent might ever be conducted, and its self-experiment exception to anticipated death or disability is terse and troubling. Later frameworks answered some of these questions differently. These gaps do not erase Nuremberg's force; they mark the limits of reading a short 1947 judicial statement as a complete contemporary rulebook.

Its most durable historical warning is institutional as well as individual. A signed form cannot create voluntariness inside captivity, dependency, deception, racism, or state violence, and a code cannot enforce itself. Consent, sound science, material safeguards, independent scrutiny, and accountable institutions have to work together.

Across the collection

Continue from the Nuremberg Code

History of medical ethics

Connect research ethics to professional duty, patient rights, public health, institutional power, and histories beyond a single Western genealogy.

History of clinical trials

Follow changing standards for comparison, consent, risk, independent review, registration, and reporting.

The Belmont Report

Compare Nuremberg's rules for permissible experiments with a later U.S. framework of respect for persons, beneficence, and justice.