Guilty While Asleep?

Automatism, Sleepwalking, and Criminal Liability: A Comparative Analysis

Perspectives from Indian and Scots Criminal Law

Consider the following scenario. A person takes a lawfully prescribed sleeping pill, goes to bed, and falls asleep. Several hours later, their body rises, walks out of the house, gets behind the wheel of a car, and drives into traffic, all without any conscious awareness. They wake up in police custody, charged with a criminal offence, with no memory of anything that occurred between falling asleep and being arrested.

This is not a hypothetical. Cases of precisely this nature have been documented and litigated across jurisdictions, from Canada to England, from the United States to Scotland. They raise a question that goes to the very foundation of criminal law: can a person be guilty of an offence they did not choose to commit, were not conscious of committing, and cannot remember?

As a practitioner qualified across Indian and Scots law, two systems that share deep common-law roots yet have diverged in important ways, we find this question both practically significant and intellectually rich. This article examines the doctrine of automatism through the lens of both traditions, drawing on landmark cases from multiple jurisdictions to map where the law stands and where it ought to go.'

The Foundational Principle: No Voluntary Act, No Crime

Criminal law across virtually every tradition rests on a deceptively simple proposition: an act is not punishable unless it is voluntary. The Latin maxim actus non facit reum nisi mens sit rea: the act does not make one guilty unless the mind is also guilty, is as fundamental to the Indian Penal Code as it is to the criminal law of Scotland or England. Without a voluntary act, there is no actus reus. Without awareness, there is no mens rea. And without either, there should, in principle, be no crime.

Automatism sits precisely in this gap. It refers to a state in which the body performs actions without conscious control, a spasm, a reflex, a somnambulistic episode. The person is, in a real sense, not “there.” As Justice Stephen observed as far back as 1889, a person who commits what would otherwise be a crime while sleepwalking would plainly be entitled to acquittal, for the simple reason that they did not know what they were doing. The law’s difficulty lies not in accepting this principle but in applying it: who gets to claim they were asleep, how do courts separate the genuine sleepwalker from the calculated fraudster, and what should happen to someone who may pose a genuine risk of harming others during unconscious episodes?

Insane and Non-Insane Automatism: The Critical Divide

The common law has traditionally divided automatism into two categories, and the distinction carries enormous consequences. Insane automatism arises from an internal cause, often a “disease of the mind”, and results in a special verdict of not guilty by reason of insanity, potentially followed by hospitalisation or supervision. Non-insane automatism arises from an external cause (such as involuntary intoxication, a blow to the head, or a sudden medical episode) and results in a complete acquittal. The person walks free.

Where does sleepwalking fall? This seemingly straightforward question has divided courts for decades. In England and Wales, the Court of Appeal in R v Burgess [1991] 2 QB 92 classified sleepwalking as insane automatism, an internal condition amounting to a disease of the mind under the M’Naghten Rules. This has the uncomfortable consequence of labelling a person who is not, by any medical definition, “insane” as legally insane. As the authors of Smith, Hogan and Ormerod’s Criminal Law have observed, applying the term “insanity” to a normal person who happens to sleepwalk is profoundly misleading.

The Landmark: R v Parks and the Canadian Approach

The most significant judicial treatment of sleepwalking in criminal law remains the Canadian Supreme Court’s decision in R v Parks [1992] 2 SCR 871. The facts are striking. In May 1987, Kenneth Parks, a 23-year-old with no history of violence and no psychiatric illness, rose from his bed in Pickering, Ontario, drove over 20 kilometres to the Scarborough home of his in-laws, and attacked them both. His mother-in-law was killed. His father-in-law was seriously injured. Parks then drove to a police station.

Five expert witnesses, none contradicted by the Crown, testified that Parks had been sleepwalking throughout the incident. His family had a documented history of sleep disorders. He had an excellent relationship with his in-laws and no discernible motive. The Supreme Court upheld his acquittal, holding that sleepwalking constituted non-insane automatism. Justice La Forest, writing for the majority, examined whether the condition arose from an internal “disease of the mind” or an external cause, and whether there was a continuing danger of recurrence. On both counts, the evidence pointed to a transient, non-recurring episode that did not justify channelling the defence through the insanity provisions.

The decision was controversial. Critics argued that it opened the door to acquittals for violent offenders. Supporters countered that criminal law’s requirement of voluntariness is non-negotiable, and that the rarity of genuine somnambulistic violence made a “floodgates” concern unrealistic. Parks remains the global reference point for courts confronting this issue.

Scots Law: Automatism as a Special Defence

Scotland’s approach to automatism has its own distinctive character. Automatism is recognised as one of six special defences in Scots criminal law, alongside alibi, coercion, incrimination, insanity, and self-defence. The leading modern authority is Ross v HM Advocate [1991] JC 210, where the High Court of Justiciary accepted external-cause automatism in the context of involuntarily spiked drinks. The court defined automatism as a condition whereby a person unconsciously loses control of their body through no fault of their own.

The earlier decision of HM Advocate v Cunningham [1963] SLT 345 had appeared to narrow the scope of internal-cause automatism significantly, holding that mental or pathological conditions short of insanity were relevant only to mitigation, not to the question of guilt. This created an uneasy space for sleepwalkers: their condition is internal but not, in any meaningful medical sense, a “disease.”

The Criminal Justice and Licensing (Scotland) Act 2010 modernised the insanity framework by replacing it with a “mental disorder” defence, requiring that the accused was unable, by reason of mental disorder, to appreciate the nature or wrongfulness of their conduct. Scotland has notably retained a broader conception of “wrongfulness” than England, encompassing moral wrongfulness, not merely legal wrongfulness. However, as several commentators have observed, the 2010 reforms addressed only the insanity side of the equation and left external-cause automatism underexplored. The case of Finegan v Heywood [2000] JC 444, where sleepwalking precipitated by alcohol was treated as externally caused, demonstrates the continuing doctrinal uncertainty in this area.

The practical consequence is that a sleepwalker facing charges in Scotland today would enter an uncertain doctrinal space: is their condition a “mental disorder” warranting the reformed defence (with disposals ranging from absolute discharge to hospitalisation), or does it qualify as involuntary, non-culpable automatism justifying a straightforward acquittal? The answer may depend more on the facts of the individual case than on any settled principle, a state of affairs that is neither satisfactory nor sustainable.

Indian Law: From Section 84 IPC to Section 22 BNS

Indian criminal law presents a starkly different landscape. The Indian Penal Code, 1860, itself drawn from the same M’Naghten well as Scots and English law, addressed involuntary acts through Section 84: nothing is an offence committed by a person who, by reason of unsoundness of mind, is incapable of knowing the nature of the act, or that they are doing what is wrong or contrary to law. Indian courts have consistently interpreted this as requiring proof of “legal insanity” as distinct from “medical insanity,” a distinction emphasised by the Supreme Court in Hari Singh Gond v State of Madhya Pradesh and reaffirmed in Surendra Mishra v State of Jharkhand.

The fundamental difficulty is that the IPC never explicitly contemplated automatism as a standalone defence. A sleepwalker is not classically “of unsound mind” in the way Section 84 envisaged, they are not suffering from a delusion, nor are they unable, in their waking life, to distinguish right from wrong. They are simply unconscious. Indian jurisprudence has acknowledged the concept of automatism in academic commentary, and the expression has been recognised by Indian courts as encompassing involuntary movement caused by illness, intoxication, or somnambulism. But the country has not developed a robust body of case law around it. A person who commits a violent act while genuinely asleep would struggle to fit neatly within Section 84, yet might have no other statutory defence available.

The Bharatiya Nyaya Sanhita (BNS), which replaced the IPC with effect from 1 July 2024, offers a tentative step forward. Section 22 of the BNS replaces “unsoundness of mind” with the broader term “mental condition” and adds the element of incapacity to understand the “consequences” of one’s actions. This is a wider formulation. The shift from “unsoundness of mind” to “mental condition” potentially encompasses a broader range of states, including, arguably, parasomnia and medication-induced unconsciousness. Whether Indian courts will interpret Section 22 expansively enough to accommodate genuine automatism remains an open question, but the legislative language is now more accommodating than it has ever been.

That said, the BNS has missed certain opportunities. It does not explicitly codify automatism as a distinct defence. It omits any provision for diminished criminal responsibility outside the binary of full liability and full exoneration. And the failure to link the term “mental condition” to the definitions established under the Mental Healthcare Act, 2017, leaves an interpretive gap that future litigation will inevitably have to fill.

The Pharmaceutical Dimension: Medication-Induced Parasomnia

A growing category of automatism cases involves prescription medication, in particular, sedative-hypnotic drugs such as zolpidem (widely marketed as Ambien). Regulatory bodies including the US FDA have mandated warnings on these medications acknowledging the risk that users may perform complex activities, including driving, cooking, making phone calls, and engaging in sexual activity, while not fully conscious, with no subsequent memory of doing so. Reported incidents involving these medications number in the thousands.

These cases create a distinctive doctrinal tension. If a person takes a lawfully prescribed medication in good faith and experiences a known pharmacological side effect that causes them to act without awareness, the argument for criminal liability weakens considerably. The conduct is not the product of a voluntary choice to become intoxicated or to take a risk. It is, rather, the consequence of following medical advice.

Yet a paradox emerges in practice. For offences requiring proof of intent, such as murder or assault, the automatism defence may succeed because the prosecution cannot establish that the accused was aware of their actions. But for strict or absolute liability offences, such as driving under the influence, the prosecution typically need only prove impairment and the act of driving, not conscious intention. The result is that a person who causes a fatal accident while medication-induced sleepwalking may be acquitted of vehicular manslaughter but convicted of the lesser DUI charge. This asymmetry is difficult to justify on principle.

Under Scots law, the concept of prior fault would be central to the analysis. If a person is aware that their medication carries a risk of parasomnia and fails to take reasonable precautions, for example, by securing their car keys before going to bed, the defence of automatism may be weakened. The same reasoning would likely apply under Indian law, where the accused’s knowledge of the risk and failure to mitigate it could undermine any claim under Section 22 of the BNS.

Evidentiary Challenges and the Role of Expert Testimony

Courts in every jurisdiction have recognised that automatism claims require expert medical evidence. As Lord Devlin observed in Hill v Baxter [1958] 1 QB 277, a layperson cannot safely distinguish the genuine case from the fraudulent one without scientific assistance. This remains the settled position across common-law jurisdictions.

Modern sleep medicine offers increasingly sophisticated diagnostic tools: polysomnography (which monitors brain waves, breathing, and movement during sleep), electroencephalography (EEG), structured psychiatric assessments, and detailed sleep and family histories. Leading researchers in forensic sleep medicine have proposed standardised diagnostic protocols for cases of suspected sleep-related violence, including criteria for confusional arousals: the subject must have been asleep long enough to reach deep sleep, the arousal stimulus must be identifiable, the abnormal behaviour must commence immediately upon arousal, and the episode is typically brief unless alcohol is a contributing factor.

These developments are welcome. As diagnostic rigour increases, the risk of fraudulent automatism claims diminishes, and the case for accepting the defence as a legitimate part of criminal law becomes correspondingly stronger. Both Indian and Scottish courts would benefit from the adoption of such standardised protocols, which would bring greater consistency and predictability to an area of law that currently relies heavily on the quality of individual expert witnesses.

Towards Reform: A Standalone Verdict?

Perhaps the most thought-provoking proposal to emerge from the academic literature is the introduction of a standalone verdict: “not guilty by reason of sleep disorder.” Advocates of this approach, including several leading forensic sleep medicine researchers argue that it would eliminate the awkward fit between sleepwalking and the traditional insanity framework, remove the stigma of the “insanity” label from people who are not mentally ill, and allow courts to impose appropriate risk-management orders (including mandatory treatment, monitoring, and lifestyle restrictions) without criminalising the truly involuntary.

In England and Wales, the Law Commission has proposed a reformed defence of “not criminally responsible by reason of a recognised medical condition,” which would cover both mental and physical disorders and require that the defendant wholly lacked the relevant criminal capacity. This proposal would bring English law closer to the Canadian position. Scotland’s Scottish Law Commission would do well to consider a parallel reform, addressing the gap left by the 2010 legislation’s silence on external-cause automatism.

For India, the moment is especially opportune. The BNS is new legislation, still in the earliest stages of judicial interpretation. A considered series of High Court or Supreme Court decisions recognising automatism as falling within the scope of “mental condition” under Section 22 could establish a principled framework without the need for further legislative intervention. Alternatively, the legislature could amend the BNS to include explicit provisions for automatism, drawing on the comparative experience of Canada, Scotland, and the Law Commission proposals in England.

Conclusion

The law of automatism sits at the intersection of criminal doctrine, neuroscience, sleep medicine, and pharmacology. It forces us to confront a question that most legal systems prefer to assume away: what does it mean to act, and what does it mean to choose? When the body moves but the mind is absent, the entire architecture of criminal responsibility, built on assumptions of voluntariness, awareness, and intent, is tested at its foundations.

As practitioners working across two legal traditions, we believe the answer lies not in retreating to rigid doctrinal categories inherited from the nineteenth century but in building frameworks that are medically informed, procedurally fair, and honest about the limits of what the law can know about the sleeping mind. The cases will keep coming. The question is whether our legal systems will be ready for them.

Disclaimer: This article is intended for informational and educational purposes only and does not constitute legal advice. The views expressed are those of the author in their personal capacity. For advice on any specific matter, please consult a qualified legal professional in the relevant jurisdiction.

© 2026 Vibhuti Vasisth, Advocate and Solicitor. All rights reserved. No part of this article may be reproduced, distributed, or transmitted in any form or by any means without the prior written permission of the author, except for brief quotations in critical reviews and certain non-commercial uses permitted by copyright law. The analysis and opinions expressed herein are original work. For permissions or enquiries, contact legal@aparlaw.com / vibhuti@aparlaw.com.