Insanity and automatism

Insanity and automatism are best understood as two legal mechanisms for denying responsibility where the defendant’s conduct, though apparently satisfying the offence definition, is said not to be the product of a responsible agent. They sit at the intersection of actus reus, mens rea, incapacity, and public protection. That is why they generate such persistent doctrinal strain: the criminal law wants to avoid punishing involuntary or radically disordered conduct, but it also wants to manage the risk posed by those whose conditions may recur.

The first distinction is between automatism and insanity. Automatism, in its strict form, is a denial of voluntary action. If D’s bodily movement was not controlled by the mind, there is no actus reus. A successful plea of non-insane automatism therefore results in a complete acquittal. Insanity, by contrast, is a special verdict: not guilty by reason of insanity. It does not deny that D did the act; rather, it says that at the time D acted under a legally recognised mental incapacity. The consequence is not ordinary conviction, but neither is it always a simple acquittal; the court has disposal powers under the statutory regime.

The principal legal test for insanity remains the M’Naghten Rules. D must be labouring under a defect of reason, caused by a disease of the mind, so that he either did not know the nature and quality of his act, or did not know that what he was doing was wrong. Each phrase is technical. “Defect of reason” excludes mere confusion, absent-mindedness or impulsiveness. “Disease of the mind” is a legal, not medical, category. “Wrong” has been treated as legally wrong, not morally wrong. The result is narrow, archaic, and often criticised; but it is still the foundation of English law.

The principal legal test for automatism asks whether there was a total loss of voluntary control, and whether the cause was external and not self-induced. A partial impairment of control will not suffice. This is why the contrast with intoxication, covered in Week 12, matters: a defendant cannot usually escape liability by relying on a state he culpably brought about. The diabetic cases illustrate the structure. Hypoglycaemia caused by insulin may be an external cause and so support non-insane automatism. Hyperglycaemia caused by diabetes itself is internal and therefore insanity. This is not medical taxonomy; it is a legal classification driven by recurrence risk and disposal consequences.

For Tripos purposes, the subject is disproportionately examinable because it tests whether students can move cleanly between levels of analysis: voluntariness, mens rea, legal incapacity, burden of proof, evidential foundations, and policy. In a problem question, do not simply announce “insanity” because D has a psychiatric diagnosis. Ask: was there voluntary conduct? was mens rea present? was there a total loss of control? what caused it? was that cause internal or external? did D know the nature and quality of the act or that it was legally wrong? was the condition self-induced? In an essay, the best answers resist sloganising. The law is not merely “out of date”; it reflects a real tension between individual desert and public safety. The difficulty is that the present doctrinal tools express that tension through categories which are blunt, medically misleading, and morally unstable.

The modern law of insanity is conventionally traced to M’Naghten’s Case in 1843. Daniel M’Naghten, suffering from persecutory delusions, shot and killed Edward Drummond, private secretary to Sir Robert Peel, apparently believing him to be Peel. His acquittal on grounds of insanity provoked public and parliamentary controversy. The House of Lords then put questions to the judges, whose answers became the M’Naghten Rules. The historical point matters. The rules were not the product of a systematic criminal code; they were reactive, shaped by political anxiety, Victorian psychiatry, and a narrow conception of cognitive incapacity.

Earlier English law had treated madness through broad and often moralised language: the “wild beast” test in Arnold’s Case is the familiar, though crude, example. M’Naghten refined the inquiry, but only to a limited extent. It fixed on cognition: did the defendant know what he was doing, and did he know it was wrong? It did not recognise impaired volition as such. Thus a defendant who knows what he is doing and knows it is legally wrong, but cannot control himself because of mental disorder, generally falls outside insanity. This is one reason why diminished responsibility, considered in Week 4, became necessary in homicide. Diminished responsibility partially mitigates murder to manslaughter where responsibility is substantially impaired by an abnormality of mental functioning; insanity, by contrast, remains an all-or-nothing special verdict available across offences but confined by M’Naghten.

Automatism developed later as the law sharpened the requirement of voluntary action. The idea that criminal liability requires conduct attributable to the will is elementary, but cases involving concussion, epilepsy, diabetes, sleepwalking, and dissociative states forced the courts to decide when a defendant’s movements were not legally voluntary. In Hill v Baxter the Divisional Court accepted that exceptional events, such as being attacked by a swarm of bees or being struck by a stone, might render driving involuntary. Bratty then gave the canonical statement that an involuntary act is one done by the muscles without control by the mind, or by a person not conscious of what he is doing.

The history of automatism is also the history of a classificatory choice. If the involuntary conduct is caused by an “external” factor, D may receive a complete acquittal. If it is caused by an “internal” condition, the law tends to route the case into insanity. This division was consolidated in cases such as Quick, Hennessy, Sullivan and Burgess. It is often justified on the ground that internal conditions may recur and therefore require protective disposals. But the classification produces difficult results. Epilepsy, sleepwalking and diabetes are treated not according to ordinary medical language but according to the law’s concern with continuing danger.

The procedural consequences also changed. Historically, insanity could result in indefinite detention, making defendants understandably reluctant to plead it. Modern legislation provides more flexible disposals, including hospital orders, supervision orders and absolute discharge, though murder still attracts compulsory consequences where the statutory conditions are met. Nonetheless, the stigma and practical implications of the special verdict remain substantial.

This history explains the present law’s oddity. Insanity is not simply a defence, nor simply a denial of mens rea. Automatism is not simply a medical condition, nor simply a loss of memory. Both doctrines preserve fundamental principles of responsibility, but they do so through inherited categories that have not been comprehensively modernised. A strong Cambridge answer should make that historical origin do analytical work: M’Naghten’s cognitive frame, the protective function of the special verdict, and the later emergence of automatism together explain why the modern law is both principled in aspiration and incoherent in detail.

Begin with voluntariness. Criminal liability ordinarily requires a voluntary act or omission. The requirement is not usually litigated because most bodily movements are plainly voluntary. Automatism is exceptional: it alleges that D’s apparent conduct was not the product of conscious control. The courts require a total destruction of voluntary control, not merely reduced control, panic, excitement, impulsiveness, or impaired judgment. This is an important examination point. A defendant who says “I could not resist” is normally making a volitional excuse not recognised by automatism; a defendant who says “my body moved without the control of my mind” is closer to the doctrine.

Non-insane automatism requires an external cause. Classic examples include a blow to the head, the effect of anaesthetic, an unexpected reaction to prescribed medication, or hypoglycaemia caused by insulin. The external cause must explain the loss of control. The law is cautious because non-insane automatism leads to complete acquittal. It is not enough that D has no memory of the incident. Amnesia may be evidence, but it is not automatism. Nor is it enough that D acted in a dreamlike, dissociated or emotionally overwhelmed state. The jury must be left with evidence capable of supporting total involuntariness.

Insane automatism is the label often given where the loss of control is caused by a disease of the mind. This is not a medical term. The phrase covers any disease or disorder which impairs the mental faculties of reason, memory and understanding in the relevant way. The authorities have treated epilepsy, arteriosclerosis, hyperglycaemia and sleepwalking as capable of amounting to disease of the mind. That may sound extraordinary to a modern reader. The explanation is that “disease of the mind” is a legal category concerned with the source of impairment and risk of recurrence, not with psychiatric nosology.

M’Naghten’s Case remains the necessary starting point. It supplies the governing formula: defect of reason, disease of the mind, and lack of knowledge of nature and quality or wrongness. The case should not be treated as a psychiatric authority. It is a legal test allocating responsibility under conditions of cognitive disorder. Its limitations are equally important: it does not recognise irresistible impulse, and its “wrongness” limb has been interpreted as legal wrongness.

Kemp illustrates the breadth of “disease of the mind”. The defendant suffered from arteriosclerosis which caused a temporary blackout during which he attacked his wife. The condition was physical in origin, yet it affected the mental faculties and was treated as insanity. The case is crucial because it rejects any simple equation of insanity with mental illness. A bodily disease may be a disease of the mind if it impairs reason, memory or understanding.

Bratty is the leading general authority on automatism. Lord Denning’s formulation emphasises muscular movement without control by the mind, or conduct by a person not conscious of what he is doing. But Bratty also shows judicial suspicion of vague automatism claims. The courts require evidence of complete involuntariness and insist that some conditions which produce automatic behaviour may fall within insanity.

Quick and Hennessy are the diabetic pair and should be learned together. In Quick, hypoglycaemia caused by insulin was treated as an external factor, capable of founding non-insane automatism. In Hennessy, hyperglycaemia caused by diabetes itself was treated as internal, and thus insanity. The distinction is examinable because it is doctrinally orthodox and conceptually vulnerable. It shows the law classifying by source and recurrence risk rather than medical labels.

Sullivan confirms that epilepsy may be a disease of the mind. The House of Lords rejected the argument that the short duration of the epileptic episode prevented insanity. Permanence is not required. This is a central point: transitory impairment can still be insanity if caused by an internal condition. The case also demonstrates the breadth of the special verdict and the public-protection rationale.

Burgess applies similar reasoning to sleepwalking. The defendant attacked a friend while apparently asleep. The Court of Appeal held that sleepwalking could amount to insanity where the cause was internal and there was a risk of recurrence. The case is often criticised because lay and medical understandings of sleepwalking do not naturally correspond to insanity. But doctrinally it follows Sullivan: a disorder of the mental faculties arising internally may trigger M’Naghten.

Hill v Baxter is useful for examples and limits. It recognises that driving may become involuntary in exceptional circumstances, such as a sudden attack by bees or being struck by a stone. But it also affirms that ordinary lapses, drowsiness or inattention are not enough. The case is a warning against over-extending automatism into mere poor driving.

R v T is more difficult. It accepted that post-traumatic stress following rape might, on evidence, support automatism. It is best used cautiously. It does not mean that emotional disturbance automatically becomes automatism. The question remains whether there was total loss of voluntary control and whether the cause can properly be treated as external. It is valuable in essays because it exposes the instability of the internal/external distinction where trauma operates through psychological mechanisms.

See practice questions section below.

See further reading section below.