Actus reus

“Actus reus” is shorthand for the external elements of a criminal offence – everything the prosecution must prove happened in the world, as opposed to what was going on in the defendant’s head (which is the mens rea). The phrase comes from the maxim actus non facit reum nisi mens sit rea – “an act does not make a person guilty unless the mind is also guilty” – attributed to Edward Coke in the 17th century, though the underlying idea traces back to Augustine and canonical law.

The label is a little misleading. “Actus reus” suggests we are looking for a guilty act, but in reality it encompasses:

• Conduct (the act or, sometimes, the omission);
• Circumstances (surrounding facts that must exist – e.g. that property “belongs to another” in theft);
• Consequences (in result crimes – e.g. that death was caused in homicide).

Glanville Williams famously disliked the Latin and preferred “external elements.” Many modern textbooks (Ormerod & Laird’s Smith, Hogan & Ormerod; Simester & Sullivan) keep the Latin but flag the imprecision.

Key point: Actus reus is not just “the act.” It is whichever combination of conduct, circumstance, and consequence the particular offence requires.

2. Historical development

2.1 Medieval and early modern origins

Early English criminal law was heavily concerned with conduct that disturbed the King’s peace. The subjective element (mens rea) developed slowly; in the 12th – 13th centuries liability was often close to strict, with the focus on the harm done. Bracton (13th century) began importing canonical ideas about moral fault, and by the time Coke wrote his Institutes (1628–44), the act/mind distinction was reasonably settled.

2.2 The 19th Century: codification attempts and doctrinal clarity

The Victorian era is when actus reus theory really matures. James Fitzjames Stephen’s Digest of the Criminal Law (1877) and his failed Draft Criminal Code (1879) tried to systematise the external elements. The case law of the period also starts drawing sharper lines – for instance on causation (R v Dyson [1908] 2 KB 454, better cited for the proposition that accelerating an inevitable death may amount to causing it; the now-abolished year-and-a-day rule formed part of the background, and was later abolished by the Law Reform (Year and a Day Rule) Act 1996).

2.3 The 20th Century: Glanville Williams and the analytical turn

Glanville Williams’s Criminal Law: The General Part (1953, 2nd ed. 1961) is the single most influential 20th-century English-language work on actus reus. Williams pushed for analytical clarity – breaking offences into conduct, circumstances, and consequences; distinguishing voluntary from involuntary acts; and treating omissions as exceptional. The Law Commission’s Draft Criminal Code for England and Wales (1989) largely adopts this structure.

2.4 The modern period

Recent developments have been less about reinventing actus reus and more about refining hard cases: omissions liability (Miller, Stone and Dobinson, Evans), causation in medical and drug-supply contexts (Cheshire, Kennedy (No 2)), and automatism/voluntariness (Bailey, Hennessy, Coley).

3. The core requirement: a voluntary act

The classical position, articulated by Stephen and endorsed by Lord Denning in Bratty v AG for Northern Ireland [1963] AC 386, is that the actus reus must be a voluntary act – conduct the defendant willed.

Lord Denning in Bratty:

“No act is punishable if it is done involuntarily: and an involuntary act in this context… means an act which is done by the muscles without any control by the mind, such as a spasm, a reflex action or a convulsion; or an act done by a person who is not conscious of what he is doing…”

3.1 Automatism

Automatism is commonly analysed as negativing voluntary conduct, and so the conduct element of the actus reus; depending on the offence and facts, it may also prevent proof of mens rea. Classic examples include concussion, reflex action, hypoglycaemia caused by insulin administration, and the well-known obiter examples given by Devlin J in Hill v Baxter [1958] 1 QB 277, including the driver suddenly attacked by a swarm of bees.

The law distinguishes between non-insane automatism, caused by an external factor and leading to a complete acquittal, and insane automatism, caused by a “disease of the mind” and leading in the Crown Court to the special verdict of not guilty by reason of insanity, following which the court has a range of disposals under the Criminal Procedure (Insanity) Act 1964 (as amended), including a hospital order, a supervision order, or an absolute discharge: see R v Quick [1973] QB 910, R v Sullivan [1984] AC 156, R v Hennessy [1989] 1 WLR 287 and R v Burgess [1991] 2 QB 92. The modern procedure is governed by the Criminal Procedure (Insanity) Act 1964 (as amended).

The law distinguishes:

Type	Cause	Legal consequence
Non-insane automatism	External factor (blow to head, drug reaction)	Complete acquittal
Insane automatism	Internal “disease of the mind” (epilepsy, sleepwalking, hyperglycaemia caused by failure to take insulin)	Special verdict of not guilty by reason of insanity

This distinction, entrenched by R v Sullivan [1984] AC 156 and R v Hennessy [1989] 1 WLR 287, is widely criticised as producing absurd results (a diabetic is “insane” if he fails to take insulin, but not if he takes too much).

3.2 State of affairs offences

Some offences criminalise a situation rather than an act, and here voluntariness breaks down alarmingly. The textbook example is:

• R v Larsonneur (1933) 24 Cr App R 74 – a French woman was deported from Ireland and brought by police to the UK against her will. She was convicted of being “found” in the UK as an alien without leave. The fact she had no choice was irrelevant.
• Winzar v Chief Constable of Kent (1983) (often cited, though of doubtful authority) – D was taken to hospital drunk, asked to leave, and when he was found on the highway by police (who had put him there) he was convicted of being “found drunk on the highway.”
  

These cases are routinely described as affronts to the voluntariness principle, and modern commentators (Ashworth, Simester) argue they survive only because courts find them administratively convenient.

4. Omissions: the big doctrinal battleground

English law begins from the principle that there is no general duty to act. The much-quoted illustration: you can watch a baby drown in an inch of water and commit no crime, absent some duty. This is sometimes called the “bad Samaritan” position, contrasted with civilian systems (France, Germany) that have general duty-to-rescue offences.

4.1 When does liability for omission arise?

Liability attaches only where (a) the offence can be committed by omission, and (b) the defendant was under a legal duty to act. The recognised duty categories are:

1. Statutory duty – e.g. failure, without reasonable excuse, to provide a specimen when required under s.7 Road Traffic Act 1988; and wilful neglect of a child by a person aged 16 or over who has responsibility for that child under s.1 Children and Young Persons Act 1933.
2. Contractual duty – R v Pittwood (1902) 19 TLR 37: railway gatekeeper who left the gate open, causing a death, was convicted of manslaughter.
3. Special relationship – most clearly parent/child; possibly spouses or close family depending on the facts, though such cases often overlap with voluntary assumption of responsibility. R v Gibbins and Proctor (1918) 13 Cr App R 134: father and his partner starved his child to death.
4. Voluntary assumption of responsibility – R v Stone and Dobinson [1977] QB 354: the defendants took in Stone’s anorexic sister, began caring for her, and were convicted of manslaughter when she died of neglect. The assumption of care created the duty.
5. Creation of a dangerous situation – R v Miller [1983] 2 AC 161: the squatter who fell asleep with a lit cigarette, woke to find the mattress smouldering, moved rooms and did nothing. Lord Diplock held that creating the danger imposed a duty to take reasonable steps to counter it.
6. Public office – R v Dytham [1979] QB 722: police officer who stood by while a man was kicked to death was convicted of misconduct in public office.

4.2 The extended Miller principle: R v Evans

R v Evans [2009] EWCA Crim 650 extended Miller significantly. The defendant supplied heroin to her half-sister, who overdosed. She did not summon help. The Court of Appeal (Lord Judge CJ) held that where a person “has created or contributed to the creation of a state of affairs which [they know] … has become life-threatening,” a duty to take reasonable steps arises. Note: contributed to — not necessarily caused. This is arguably a significant widening, and it has been criticised for fuzziness and for effectively routing around the Kennedy (No 2) limitation on causation in drug-supply cases (below). It should be kept within its setting: Evans was a gross negligence manslaughter case, and should not be read as creating a universal omissions duty across all offences. The courts have not framed it as bypassing Kennedy; the distinction is doctrinally maintained as one between causation and omissions duty.

4.3 Withdrawal of Treatment: Airedale NHS Trust v Bland

Airedale NHS Trust v Bland [1993] AC 789 is the defining case on omissions at the edge of life. Tony Bland, crushed at Hillsborough, was in a persistent vegetative state. The House of Lords held that withdrawing artificial nutrition and hydration was an omission, not an act, and — given no prospect of recovery — there was no duty to continue. The act/omission distinction did a great deal of heavy lifting, and Lord Mustill openly acknowledged the moral incoherence: “The conclusion that I have reached will appear… to some to be almost irrational. How can it be lawful to allow a patient to die slowly… yet unlawful to produce his immediate death by a lethal injection?”

4.4 Continuing Act Doctrine

One way courts avoid the act/omission boundary is to treat conduct as a “continuing act.” Fagan v Metropolitan Police Commissioner [1969] 1 QB 439: D accidentally drove onto a policeman’s foot and then refused to move. He argued there was no actus reus at the moment he formed mens rea. The court held that driving onto the foot was a continuing act, with which mens rea coincided. This is widely seen as a somewhat artificial dodge, but it remains good law.

4.5 Points of Contention

• The act/omission distinction itself. Ashworth’s classic article (“The Scope of Criminal Liability for Omissions” (1989) 105 LQR 424) argued for broader omissions liability. Glanville Williams resisted (“Criminal Omissions — The Conventional View” (1991) 107 LQR 86).
• Moral inconsistency. Why does voluntarily assuming care expose you to liability, when refusing to help at all doesn’t?
• The Evans expansion and whether “contributed to” risks becoming unmoored from genuine fault.

5. Causation

In result crimes, the prosecution must prove the defendant’s conduct caused the prohibited consequence. English law uses a two-stage test: factual causation and legal causation.

5.1 Factual causation: the “But For” test

Would the result have occurred but for the defendant’s conduct? R v White [1910] 2 KB 124: D poisoned his mother’s drink intending to kill her; she died of a heart attack before drinking any of it. No factual causation — she would have died anyway. (He was, however, convicted of attempt.)

5.2 Legal causation

The core requirement, associated particularly with R v Smith [1959] 2 QB 35 and applied in R v Pagett (1983) 76 Cr App R 279, is that the defendant’s act must be an “operating and substantial” cause — not necessarily the sole or main cause.

5.3 Medical treatment

Poor medical treatment rarely breaks the chain. R v Smith [1959] 2 QB 35: a stabbed soldier was dropped twice on the way to hospital and received poor treatment. Conviction upheld — the wound was still an operating and substantial cause. Only “palpably wrong” treatment (R v Jordan (1956) 40 Cr App R 152 — an outlier) breaks the chain.

R v Cheshire [1991] 1 WLR 844 formulated the modern test: medical negligence breaks the chain only if it is “so independent of [the defendant’s] acts, and in itself so potent in causing death, that [the jury] regard the contribution made by his acts as insignificant.” A very high bar.

5.4 The thin skull rule

You take your victim as you find them – physical peculiarities, pre-existing conditions, and (importantly) religious or other beliefs. R v Blaue [1975] 1 WLR 1411: a Jehovah’s Witness stab victim refused a blood transfusion and died. The defendant argued her refusal broke the chain. The Court of Appeal held it did not; the thin-skull rule extends to the “whole person.”

5.5 Victim’s own conduct

Where the victim flees or reacts and is injured, the question is whether the response was within the range of reasonable responses. R v Roberts (1971) 56 Cr App R 95: V jumped from a moving car to escape the defendant’s sexual advances. Stephenson LJ: only if the victim’s response was “so daft as to make it [her] own voluntary act” is the chain broken.

5.6 The big modern difficulty: drug supply

R v Kennedy (No 2) [2007] UKHL 38 is the leading case. D prepared a syringe of heroin and handed it to V, who self-injected and died. Was D liable for manslaughter? Lord Bingham, for a unanimous House of Lords, held he was not: a fully informed, voluntary adult who chose to self-inject had performed a free, deliberate, and informed act that broke the chain.

“The criminal law generally assumes the existence of free will. The law recognises certain exceptions, in the case of the young, those who for any reason are not fully responsible for their actions, and the vulnerable, and it acknowledges situations of duress and necessity… But, generally speaking, informed adults of sound mind are treated as autonomous beings able to make their own decisions how they will act…”

This reaffirmed the free, deliberate, and informed intervening act doctrine. But as noted, Evans (above) has since allowed an omissions-based route to liability in similar facts.

5.7 Continuing tensions in causation

• The Kennedy/Evans tension. Kennedy closes the door on causation; Evans reopens it via duty-to-act. Is there really a principled difference between supplying a drug and supplying it plus leaving the user to die?
• “Rogue” treatment. Jordan survives doctrinally but is almost never applied. What counts as “palpably wrong”?
• Supplying/encouraging suicide. Supplying/encouraging suicide. R v Wallace [2018] EWCA Crim 690 (the acid attack case) held that causation should not have been withdrawn from the jury merely because the victim later chose lawful euthanasia in Belgium; the case is fact-specific and is better read as illustrating the potential flexibility of the causation inquiry than as a dramatic extension.
• Causation and joint enterprise post-R v Jogee [2016] UKSC 8, where the focus shifted back to intention but the causal contribution question remains live.

6. Coincidence of actus reus and mens rea

The actus reus and mens rea must coincide in time, but the courts have taken a flexible view:

• Fagan (continuing act, above).
• Thabo Meli v R [1954] 1 WLR 228: D attacked V intending to kill him, thought he had succeeded, and rolled the “body” off a cliff. V was alive and died of exposure. The Privy Council treated the sequence as “one transaction” — the mens rea at the start sufficed.
• R v Church [1966] 1 QB 59 and R v Le Brun [1992] QB 61 applied the same transactional reasoning to unplanned sequences.

The doctrinal price is some fiction: we pretend the mens rea persists or the act continues, when really the courts are avoiding a technical acquittal.

7. Possession offences: a curious hybrid

Possession offences (drugs, weapons, indecent images) have an actus reus that is really a state. The leading case is Warner v Metropolitan Police Commissioner [1969] 2 AC 256, which struggled with whether possession requires knowledge of the nature of what is possessed. In drug cases, the actus reus is possession or control of a controlled drug. Warner shows the difficulty of deciding how much knowledge is built into possession itself; under the Misuse of Drugs Act 1971, s.28 provides important statutory defences where D neither knew nor suspected, nor had reason to suspect, the relevant facts. In practice, therefore, it is not purely “bare” possession. Commentators note that possession stretches the concept of a “voluntary act” — you are criminalised for a continuing situation, not for anything you did at the charged moment.

8. Modern developments and continuing tensions

8.1 Cyber conduct and “act” in a digital world

How do we locate the actus reus when conduct is distributed across networks and jurisdictions? In classic cybercrime offences the “act” is often supplied by statutory language itself: under s.1 of the Computer Misuse Act 1990, the actus reus is that the defendant “causes a computer to perform any function” with the relevant intent and without authorisation. By contrast, the Online Safety Act 2023 is not built around a single equivalent formula; it is primarily a regulatory regime for online services, albeit one that also creates and amends certain criminal liabilities. The better point, therefore, is that digital offending puts pressure on traditional ideas of act, place, and voluntariness, rather than that one unified cyber actus reus has emerged across both statutes.

8.2 Coercive and controlling behaviour

Serious Crime Act 2015 s.76 criminalises a repeated or continuous course of conduct, not a single act. That already stretched orthodox actus reus analysis, because the gravamen of the offence is cumulative behaviour rather than one sharply defined event. The provision has since been amended by s.68 of the Domestic Abuse Act 2021, which removed the former cohabitation requirement and refocused the offence around whether the parties are “personally connected”. The actus reus point remains the same, however: this is an offence whose external elements are deliberately pattern-based rather than instantaneous. The amendment applies from 5 April 2023 and is not retrospective.

8.3 Corporate actus reus

The identification doctrine, classically associated with Tesco Supermarkets Ltd v Nattrass [1972] AC 153, has been progressively supplemented by statutory models that do not depend on attributing one directing mind’s conduct to the company. Examples include the Corporate Manslaughter and Corporate Homicide Act 2007, the Bribery Act 2010 s.7 offence of failure to prevent bribery, the Criminal Finances Act 2017, ss.45–46 failure to prevent facilitation of tax evasion offences, and the Economic Crime and Corporate Transparency Act 2023 offence of failure to prevent fraud. These “failure to prevent” models are, in substance, omission-based forms of corporate actus reus.

8.4 Omissions creep

The trajectory from Miller → Evans → Wallace suggests a steady expansion of omissions and causation liability. Whether this is principled development or doctrinal drift is contested.

8.5 The voluntariness puzzle in intoxication and mental health

R v Coley, McGhee and Harris [2013] EWCA Crim 223 grappled with the interaction between voluntary intoxication and automatism. In the sexual offences context, the harder issue is often not voluntariness in the Bratty sense but whether the complainant had capacity to choose. Here the relevant authority is R v C [2009] UKHL 42, concerning s.30 Sexual Offences Act 2003, which held that a complainant may be “unable to refuse” because of or for a reason related to mental disorder where that disorder removes the capacity to choose. That illustrates how, in consent-based offences, questions that look like mens rea or policy questions are often built directly into the external elements of the offence.

8.6 Consent as an absence-of-actus-reus question

In sexual offences, the complainant’s lack of consent is part of the prosecution case, with consent defined by s.74 of the Sexual Offences Act 2003 as agreement by choice where the complainant has the freedom and capacity to make that choice. Cases such as R v Bree [2007] EWCA Crim 804 and R v Lawrance [2020] EWCA Crim 971 show the conceptual strain: conduct that appears outwardly consensual may, on the statutory analysis, be non-consensual, while deception will negate consent only in limited classes of case. It is therefore accurate, but only with that caveat, to say that in sexual offences absence of consent functions as an actus reus element.

9. Summary table of leading cases

Topic	Case	Proposition
Voluntariness	Bratty v Attorney-General for Northern Ireland [1963] AC 386	Actus reus must be voluntary
State of affairs	R v Larsonneur (1933) 24 Cr App R 74; Winzar v Chief Constable of Kent, The Times, 28 March 1983	Voluntariness not always required
Omission — contract	R v Pittwood (1902) 19 TLR 37	Contractual duty founds liability
Omission — relationship	R v Gibbins and Proctor (1918) 13 Cr App R 134	Parent/child duty
Omission — assumption	R v Stone and Dobinson [1977] QB 354	Voluntary care creates duty
Omission — creation of danger	R v Miller [1983] 2 AC 161	Duty to counteract danger you created
Omission — extension	R v Evans [2009] EWCA Crim 650	Duty where you “contributed to” danger
Omission — withdrawal	Airedale NHS Trust v Bland [1993] AC 789	Withdrawal = omission, lawful if no duty
Coincidence	Fagan v Metropolitan Police Commissioner [1969] 1 QB 439; Thabo Meli v R [1954] 1 WLR 228	Continuing act / single transaction
Factual causation	R v White [1910] 2 KB 124	But-for test
Legal causation	R v Pagett (1983) 76 Cr App R 279	Operating and substantial cause
Medical treatment	R v Smith [1959] 2 QB 35; R v Cheshire [1991] 1 WLR 844	Very high bar to break chain
Thin skull	R v Blaue [1975] 1 WLR 1411	Take victim as you find them
Victim’s response	R v Roberts (1971) 56 Cr App R 95	Chain not broken unless V’s reaction was so unexpected/”daft” as to be V’s own voluntary act
Drug supply	R v Kennedy (No 2) [2007] UKHL 38	Free/deliberate/informed act breaks chain

10. Closing Thoughts

Actus reus looks like the “easy” half of criminal liability – the bit you can see and prove objectively – but it hides most of criminal law’s deepest puzzles: what counts as an act, when silence becomes conduct, when we can treat causes as legally significant, and when the law should override the moral intuition that people are responsible only for what they freely do. The doctrinal map drawn by Bratty, Miller, Kennedy and Evans is coherent but fragile, and it will keep being tested – especially as criminal law expands into coercive control, corporate failure to prevent, and online harms, where the classical “act” becomes increasingly elusive.