Core Principles of Criminal Liability

Actus reus is not always "an act" in the everyday sense — it is better understood as whatever conduct element the specific offence demands, and that element takes different shapes depending on how the offence is drafted.

• A voluntary act: the defendant's willed physical movement (a punch, a squeeze of a trigger).
• An omission: a failure to act, but only where the law recognises a duty to act (covered below).
• A state of affairs: the offence is defined by circumstances existing, not by anything the defendant did.
• A prohibited consequence: the offence requires a specified result to occur in the world.

That last distinction — consequence versus conduct — reappears throughout the criminal law syllabus, so it is worth fixing precisely. Result crimes (murder, criminal damage, most non-fatal offences against the person) require proof that the defendant's conduct actually caused a specified outcome — death, damage, injury. Conduct crimes, by contrast, are complete the moment the prohibited act is performed, with no need to prove any further consequence flowed from it. This matters practically: for a result crime you must always run a causation analysis (see below); for a conduct crime you never need to.

Voluntariness and automatism

The starting requirement for actus reus is that the relevant conduct be voluntary — a product of the defendant's own will. Involuntary muscular movements do not satisfy actus reus at all, because the defendant's mind was never controlling the body. This is the doctrine of automatism: a state in which bodily movements occur without any control by the conscious mind, as in a reflex spasm, a convulsion, or a genuine state of unconsciousness. If a defendant's arm shoots out during an epileptic seizure and strikes someone, there is, in principle, no actus reus at all — the "act" was never the defendant's in the relevant legal sense.

The state-of-affairs trap

The requirement of voluntariness has one famous — and famously harsh — exception, illustrated by R v Larsonneur (1933) 24 Cr App R 74. Larsonneur was a French national whose leave to remain in the UK had been revoked; she left for the Irish Free State, was then forcibly deported from there by the Irish authorities, and was returned to the UK entirely against her will. She was nonetheless convicted of the offence of being "found" in the UK as an alien to whom leave had been refused. She had not chosen to return — she was brought back in handcuffs — yet the actus reus was satisfied regardless. Larsonneur is the textbook illustration of a state of affairs offence: where the offence is defined by the existence of a circumstance (being present in a place, being in charge of a vehicle while unfit), the prosecution need not prove any voluntary conduct by the defendant at all. It sits awkwardly with the general voluntariness requirement, which is exactly why examiners like it — it tests whether you understand that voluntariness is a default, not an absolute rule.

The common law's general instinct is that you should not be punished for what you failed to do — only for what you did. The general rule, therefore, is that there is no criminal liability for a pure omission unless the law imposes a specific duty to act. Advising a client on liability for inaction is really an exercise in checking whether their situation falls into one of a closed list of recognised duty categories. Six are examined regularly:

1. Statutory duty. Parliament can simply criminalise a failure to act directly (for example, failing to provide a specimen of breath under road traffic legislation).
2. Contractual duty. A duty can arise out of the defendant's job. In R v Pittwood (1902) 19 TLR 37, a railway level-crossing keeper left the gate open, a cart was struck by a train, and the keeper was held liable for the resulting death — his employment contract obliged him to close the gate, and his failure to do so was treated exactly as if he had acted.
3. Family relationship. A parent owes a duty to a dependent child. R v Gibbins and Proctor (1918) 13 Cr App R 134 is the grim illustration: a father was convicted of murder after his child starved to death, the parental relationship itself generating the duty to feed her.
4. Voluntary assumption of responsibility. If you take on the care of someone who cannot look after themselves, you cannot simply walk away. In R v Instan [1893] 1 QB 450, a niece who lived with, and was financially supported by, her elderly aunt failed to feed her or fetch medical help as the aunt's condition deteriorated; her voluntary assumption of care generated a duty, and its breach made her liable for the death. R v Stone and Dobinson [1977] QB 354 extends the same logic to a vulnerable relative who moved in with the defendants: once they had voluntarily undertaken some care for her, their failure to summon help as she deteriorated was itself a criminal omission.
5. Creating a dangerous situation. If you inadvertently create a danger, you must take reasonable steps to counteract it once you become aware of it. This is the Miller principle, from R v Miller [1983] 2 AC 161: a squatter fell asleep smoking, woke to find his mattress smouldering, and — instead of putting it out or raising the alarm — simply moved to another room and went back to sleep. His inaction after he became aware of the risk was held to satisfy the actus reus of arson. Notice the structure: the initial act of falling asleep was not itself blameworthy, but the omission to act once aware of the danger was.
6. Public office. Holding an office that carries responsibility for protecting others can itself generate a duty. In R v Dytham [1979] QB 722, a police officer stood by and watched while a man was fatally beaten nearby, doing nothing to intervene; he was convicted of misconduct in public office precisely because his office imposed a duty to act that ordinary bystanders do not owe.

A final, clinically important illustration sits slightly apart from this list. In Airedale NHS Trust v Bland [1993] AC 789, the House of Lords held that withdrawing artificial nutrition and hydration from a patient in a persistent vegetative state was a lawful omission, because continuing treatment was no longer in the patient's best interests. Bland shows that even where a duty to act would otherwise exist (here, the duty of medical carers), the law may excuse an omission where continued action serves no legitimate purpose for the patient. For SQE1 purposes, treat Bland as confirming that the "duty to act" framework is not absolute — it yields to a best-interests medical judgment.

Because result crimes require proof that the defendant's conduct caused the prohibited outcome, causation is where actus reus analysis gets its most exam-tested texture. Causation has two limbs, and a good answer always works through both in order.

Factual causation asks the "but for" question: would the consequence have occurred but for the defendant's conduct? R v White [1910] 2 KB 124 supplies the clean negative example — White put poison in his mother's drink intending to kill her, but she died of an unrelated heart attack before the poison could act. However appalling his intention, he was acquitted of murder (though convicted of attempt) because factual causation was not established: her death would have happened regardless of what he did.

Legal causation asks something more normative: was the defendant's conduct an operating and substantial cause of the consequence? Crucially, R v Smith [1959] 2 QB 35 confirms the defendant's act need not be the only cause of death — it only has to remain operating and substantial at the time death occurs, even where other factors (in Smith, poor and mistaken medical treatment) also contributed.

The most heavily examined wrinkle in legal causation is the novus actus interveniens — an intervening act or event that may break the chain of causation altogether. Medical treatment is the classic battleground:

Negligent medical treatment will only break the chain of causation where it is so independent of the defendant's act and so potent in causing death that the defendant's original act becomes insignificant by comparison.

R v Jordan (1956) 40 Cr App R 152 is the (unusual) case where treatment did break the chain: the victim's stab wounds had largely healed, and medical treatment described as "palpably wrong" — including a drug the victim was known to be intolerant of — was what actually killed him. Contrast R v Cheshire [1991] 1 WLR 844: the defendant shot the victim, who later died from complications of a negligently managed tracheotomy; the Court of Appeal held the chain was not broken, because the gunshot wound remained a substantial and operating cause of death even though the medical negligence also contributed. The practical lesson: Jordan is the exception, Cheshire is the rule — courts are reluctant to let poor medical care let a defendant off the causation hook.

The thin skull rule completes the picture: a defendant must take their victim as they find them, including any pre-existing physical or psychological vulnerability. R v Blaue [1975] 1 WLR 1411 is the striking illustration — the defendant stabbed a Jehovah's Witness, who then refused a blood transfusion on religious grounds and died. The defendant remained liable for her death: the rule that you take your victim as you find them extends to their belief system, not merely their physiology. You cannot argue that a "reasonable" victim would have accepted the transfusion.

Source: Egg-rmh, CC BY 2.0.

Mens rea is the mental state the offence requires — and different offences demand different degrees of mental fault. Ranked roughly from most to least culpable, the SQE1 syllabus expects you to distinguish intention, recklessness, negligence, and strict liability.

Intention

Intention is generally regarded as the most culpable form of mens rea, and it comes in two forms.

Direct intention is straightforward: the consequence is the defendant's aim, purpose, or objective — they wanted it to happen and acted in order to bring it about.

Oblique intention is subtler and does the real work in exam problems: the defendant did not desire the consequence, but foresaw it as virtually certain to result from what they were doing anyway. The governing authority is R v Woollin [1999] 1 AC 82, where the House of Lords held that a jury is entitled to find intention where death or serious harm was a virtual certainty as a result of the defendant's actions, and the defendant appreciated that this was so. Woollin threw his three-month-old son onto a hard surface in frustration; he did not want the baby to die, but the House of Lords confirmed the jury could infer intention from his appreciation that serious harm was virtually certain.

A precise point that examiners test directly: the Woollin direction is an evidential rule, not a legal definition of intention. It tells the jury when they are entitled to find intention (foresight of virtual certainty, appreciated by the defendant); it does not say that virtual certainty of a consequence automatically is intention as a matter of law. Keep this distinction sharp — conflating "evidence from which intention may be inferred" with "the definition of intention" is a common and heavily penalised SQE1 error.

Recklessness

Recklessness is a lesser form of fault than intention: it involves the conscious taking of an unjustified risk, rather than aiming at or foreseeing a consequence as near-certain. The governing test is subjective, and its history is worth knowing because it was genuinely contested.

R v Cunningham [1957] 2 QB 396 established the original subjective test: the defendant must have personally foreseen the risk of the prohibited consequence occurring. For a period, R v Caldwell [1982] AC 341 displaced this with an objective test, under which a defendant could be reckless simply by failing to give thought to an obvious risk, whether or not they personally foresaw it. This objective approach was widely criticised for convicting defendants (including children and those with reduced capacity) who genuinely had not appreciated risks that a reasonable adult would have spotted. R v G [2003] UKHL 50 finally overruled Caldwell and restored the subjective Cunningham approach as the law.

The modern test, distilled, has two components you should apply as a checklist:

The defendant must have personally foreseen the risk of the relevant consequence, and it must have been unreasonable, in the circumstances known to the defendant, for them to take that risk anyway.

Negligence and strict liability

Negligence, where it forms the basis of criminal liability (as in gross negligence manslaughter), is judged objectively — by comparing the defendant's conduct to that of a reasonable person, not by asking what the defendant personally foresaw. This is the conceptual opposite of recklessness's subjective foresight requirement, and the contrast is a favourite examiner pairing.

Strict liability offences dispense with mens rea for at least one element of the actus reus altogether — the prosecution need not prove any culpable state of mind regarding that element, only that the conduct occurred.

Motive is not mens rea

A final, precise distinction: motive — the defendant's underlying reason for acting — is legally distinct from mens rea and need not be proved at all. A defendant who steals medicine to save a dying relative still has the mens rea for theft (intention to permanently deprive), however sympathetic the motive. Do not let a sympathetic motive in a problem question tempt you into thinking mens rea is absent; motive explains why someone acted, mens rea concerns what they intended or foresaw about the act itself.

The final piece of the architecture is timing. The general rule is that actus reus and mens rea must coincide — exist at the same moment — for criminal liability to arise. If a defendant's guilty act is complete before the guilty mind forms (or vice versa), liability should, in principle, fail. Two doctrines soften this rule where the facts genuinely warrant it.

The continuing act doctrine treats certain conduct as an ongoing act rather than a single instant, so mens rea formed at any point during that continuation coincides with the actus reus. Fagan v Metropolitan Police Commissioner [1969] 1 QB 439 is the illustration: Fagan accidentally drove his car onto a police officer's foot (no mens rea for battery at that instant), but then — once told to move — deliberately left it there. Because the actus reus (the car resting on the foot) was ongoing, his later-formed intention coincided with it, and he was liable for battery.

The single transaction principle performs a related but distinct trick for a sequence of separate acts rather than one continuous act: where a series of acts forms a single overall transaction, mens rea formed at any point in that transaction can coincide with the actus reus, even if the specific act that actually caused the result happened after the guilty mind had already been present at an earlier stage. Thabo Meli v R [1954] 1 WLR 228 is the classic (if unsettling) example: the defendants attacked a man intending to kill him, believed him dead, and then threw what they thought was a corpse off a cliff to fake an accident — the medical evidence showed he actually died from exposure after being thrown off the cliff, not from the initial attack. The Privy Council held the whole sequence was one transaction, so the mens rea formed during the initial attack coincided with the actus reus that actually caused death.

Source: Judicial Committee of the Privy Council court room 3 in United Kingdom Supreme Court by Andrew Le Sueur, CC BY-SA 3.0.

A final doctrine addresses cases where the defendant's mens rea was aimed at one target but the actus reus lands on another. The doctrine of transferred malice allows a defendant's mens rea to be transferred to an unintended victim, provided the offence committed is of the same type as the one intended. R v Latimer (1886) 17 QBD 359 illustrates the doctrine operating successfully: the defendant swung a belt at one man, missed, and struck a woman standing nearby instead; his intention to injure the first man was transferred to make him liable for the injury actually inflicted on the woman.

The doctrine has a hard limit, however: transferred malice does not apply across different types of offence. R v Pembliton (1874) LR 2 CCR 119 shows the doctrine failing for exactly this reason — the defendant threw a stone intending to hit a person, missed, and broke a window instead. His intention to commit an offence against the person could not be transferred to found a conviction for criminal damage to property, because assault and criminal damage are offences of different types. When you see a "missed target" fact pattern in a problem question, always check first whether the intended offence and the actual offence are of the same kind — that single check decides whether transferred malice rescues the prosecution's case or not.

Every problem question on this topic collapses into the same four-step method: identify the actus reus required by the offence (and, for result crimes, run the full causation analysis); identify the mens rea required; check whether the two coincide in time, applying the continuing act or single transaction doctrines if the facts are a poor timing fit; and, if the actual victim differs from the intended one, check whether transferred malice bridges the gap. Every case above is a worked example of one step in that method — learn them as illustrations of the method, not as isolated trivia, and the analysis will hold together under exam pressure.