Oxford Undergraduate Law & Policy Review

Unreasonable risks: The overlooked element of recklessness in the English criminal law

I – Introduction and background

The current definition of recklessness in criminal law, as adopted by Lord Bingham in G and R, is that a person is reckless in respect of a result when “[1] he is aware of a risk that it will occur, and [2] it is, in the circumstances known to him, unreasonable to take the risk”.1 The first element of this definition was subject to much judicial and academic debate following Caldwell.2 However, the second requirement – that the risk be an unreasonable one to take – has not received significant attention, aside from its inclusion by the Law Commission in the definition adopted by Lord Bingham and the earlier brief allusion to it in Stephenson.3 This is understandable: in common scenarios where a defendant does an act with a high risk of causing serious injury or damage, the risk is plainly an unreasonable one and so the point does not need to be addressed before an appellate court or when directing a jury.

This article argues that the ‘unreasonable risk element’ is vital to understanding the concept of recklessness and has important implications for the distinction between recklessness and intention. It also draws attention to several scenarios (sports injuries, horseplay, and everyday touching) where the unreasonable risk element arguably is operating to negate criminal liability, albeit without this analysis being explicitly adopted by the courts.

Before addressing those scenarios, it is worth setting out a classic hypothetical which clearly explains that not all foreseeable risks are reckless. Following Ireland, causing another person to develop a recognised psychological condition is capable of constituting an offence against the person.4 Suppose I am in a relationship with a partner who I know to be in some way emotionally vulnerable, and foresee that if I break up with him, he is likely to develop a recognised psychological condition.5 Am I guilty of a reckless offence against the person if I end the relationship? The intuitive answer is probably not, and the way to reason this is that, even in these circumstances, breaking up with a romantic partner is something which is reasonable to do. Perhaps not everyone will share this intuition in the case of an excessively aggressive or vindictive way of ending the relationship, but this illustrates the usefulness and flexibility of the reasonable risk approach: there is room to disagree about which risks are reasonable to take in which circumstances, but the basic framework is clear. It would therefore be open to a jury or magistrate to conclude that a careful, respectful break-up which led to psychological harm was a reasonable risk to take (and so the individual is not criminally liable), but that a break-up in different circumstances unreasonably risked the other person’s psychological health.

II – Implied consent: The case for reinterpreting sports injury cases

Continuing to focus on the law of offences against the person, one problem encountered by the courts is that of sporting injuries. Frequently the players of professional and amateur sports (especially contact sports) inflict injuries on each other that are, in themselves, no doubt serious enough to fit within the scope of the criminal law. The question is essentially a line-drawing exercise: when do such injuries lead to criminal liability, and when are they permissible as part of their sport context?

In Barnes, Lord Woolf CJ made clear that the boundaries of acceptable conduct were not determined solely by the rules of the game: injuries inflicted outside the rules, especially in the ‘heat of the moment’ would not necessarily lead to criminal liability.6 Instead, the threshold for criminal liability depends on various relevant factors including: “[t]he type of the sport, the level at which it is played, the nature of the act, the degree of force used, the extent of the risk of injury, the state of mind of the defendant”.7 It is notable that, aside from the last factor (the defendant’s state of mind),8 all of these factors fit very naturally into an assessment of ‘reasonableness in the circumstances’.9

However, the reasoning in Barnes (and other sporting injury cases) was based on implicit consent as the route to negating criminal liability, rather than being directly based on the reasonableness of risks taken. The implicit consent approach is workable in the sports context, but there are several reasons to suggest that a more direct approach based on the unreasonable risk element would be preferable here.

Firstly, even if there is intuitive appeal in the idea that, by taking part in sporting activity, players genuinely do consent to the risk of some harm, it is clear that the boundaries of this consent10 are entirely constructed by the court. Lord Woolf’s suggested factors of relevance mostly speak to external attributes of the game, rather than the internal state of mind of the victim. That is not to downplay the extent to which the external circumstances do affect what the victim (subjectively) consents to: if I take part in a very relaxed amateur football game, it is unlikely that I anticipate or consent to aggressive and risky tackles one might expect in professional sport. To put the point at its lowest, however, a direct unreasonable risk approach is no worse practically than consent: it too can take into account these concerns, as part of the circumstances against which the reasonableness of the risk is to be judged. One point about the nature of this standard is worth noting at this stage. For example, Lord Woolf notes in Barnes that conduct justifying “a warning or even a sending off” will not always be criminal.11 It might seem that such conduct is ‘unreasonable’ in a common sense understanding of the word, but the question is whether it is legally unreasonable, on the standard set by this area of criminal sporting law, which might directly mirror what Lord Woolf says here.12 In this context the unreasonable risk approach can therefore achieve everything implied consent can. The unreasonable risk approach has the benefit, however, of ‘cutting out the middleman’ and being more honest about the reasons for affirming or negating liability: the court is deciding what is acceptable in the context of the game.

Where implied consent arises as a construction of the court, it is also unclear the extent to which the potential victim can withdraw this consent.13 Certainly, it would not be effective for a football player silently to decide that he no longer consents to the risk of being injured in a tackle, and in fairness, this situation can be dealt with by some concept of awareness (or belief) of consent: unless the withdrawal of consent is communicated, it is fair for other players to assume the consent persists. But even so, it is not obvious how directly effective a player’s communication modifying his ‘implied consent’ will be in the eyes of the criminal law. If I join a high-stakes amateur football match but announce to the other players beforehand ‘take it easy on me; I don’t consent to any risks of harm’, it is not clear whether this will be effective if I am ultimately injured by someone taking a risk. Perhaps this withdrawal would be factored in. But either way, the reasonable risk analysis can deal with the situation by addressing whether the announcement has sufficiently changed the circumstances of the game in such a way as to render taking the risk unreasonable. This is more flexible and more reflective of the actual considerations the court will have to take into account.

The second difficulty is the fact that consent is now not always capable of negating liability for actual and serious bodily harm, following the decision in Brown,14. To negate liability in a particular situation, it seems the law requires both consent and a reason why that consent should be operative. What has emerged is a system of categories delineated by subject matter, which to some extent derive from a general test of public interest, but in practice simply exist as instances where the courts have decided consent is sufficient.15 A full assessment of the category approach is beyond the scope of this article, but the point to be made here is that in the sports context, where the courts are dealing with reckless injuries,16 it would again be simpler and easier to ‘cut out the middleman’ and recognise the unreasonable risk element as capable of replacing the role of consent. Although justifying sports as a category where consent should be valid is trivial,17 there is a great simplicity in avoiding the thorny question entirely.

Clearer boundaries with the unreasonable risk element?

As noted above, the main reason that the unreasonable risk element is apt for the sports context is that the factors used to determine the boundaries of implied consent align almost directly with the factors relevant to assessing the reasonableness of a risk. This raises the question of whether the unreasonable risk element actually provides clearer boundaries than an analysis based on consent. The starting point is again to put the point at its lowest and highlight that the reasonable risk analysis is at least no less clear than constructed consent, if broadly the same factors are being assessed. The benefit, then, lies in the law being transparent about where these factors come from (as also noted above).18 This should make the line-drawing exercise easier for finders of fact because it avoids confusion deriving from vague assertions about what the victim probably consented to. Where the goal is to establish that a risk was objectively reasonable to take, arguments about the boundaries of acceptable play will be forced to point clearly to concrete facts about the circumstances which justify the conduct. By contrast, if the goal is to prove that the victim impliedly consented to the particular conduct, there is a risk of appeals to largely irrelevant thoughts that may have been part of the victim’s state of mind.19 The unreasonable risk approach therefore does not guarantee clearer boundaries per se, but it helps to focus the inquiry.

Focusing on the core of the wrong

A final benefit of basing the analysis of sports cases on the unreasonable risk element (and of focusing on the unreasonable risk element more generally) is that it focuses the recklessness inquiry on the element which is the core of the wrong. Where I take the risk of a prohibited consequence (such as causing actual or serious bodily harm to a fellow football player), that prohibited consequence is intuitively the basis for describing the conduct as criminal. The link, therefore, between my mental state and the prohibited consequence is the fact that I am (1) aware of the risk and (2) on the facts as I understand them, it is unreasonable (read: wrong) for me to take the risk. If we set aside the subjective awareness requirement, the wrong is that of taking an unreasonable risk. If I am unaware, or I make a mistake of fact,20 that may stop me from being criminally liable, but the wrong still exists. The unreasonable risk analysis is best at locating that wrong. There is a slight caveat to this analysis deriving from the fact that many offences, including (famously) s.47 and s.20 Offences Against the Person Act 1863, are constructive: they do not require recklessness relating to the actual level of harm that is risked, but rather recklessness relating to a lesser harm.21 What constitutes the wrong in the mental element of the offence is the unreasonable taking of the risk of the lesser harm.22 The unreasonable risk approach is not undermined by the fact that, if a more serious result does occur, the offence is upgraded to a more serious one without any further mens rea.

This is clear in the sports context when we make a comparison to the tort context, where the assessment of fault admits no subjective element. In the recent case of Elbanna v Clark, the Court of Appeal confirmed that a finding of fact that the defendant was ‘reckless’ (in the context of a sporting tackle) entailed a finding of negligence.23 Although, as the court emphasised, the proper terminology in tort cases is negligence, the finding is significant because it confirms that the idea of recklessness (purely as a matter of terminology) is unreasonable risk-taking. This may be obvious, but it highlights how the criminal law’s longstanding debates over subjective and objective recklessness have somewhat obscured this core idea. Nicola Davies LJ said in Elbanna that recklessness in the tort context was a “higher and more stringent” test than negligence,24 and it is possible to understand this as either a statement that recklessness is an unreasonable (negligent) risk plus a subjective mental state as to that risk (irrelevant in tort) or that recklessness implies a more unreasonable risk than mere negligence. Either way, the central element of the linguistic description ‘reckless’ is the unreasonable risk.

III – Pushing the boundaries: The horseplay cases

Another situation, in many ways not too distant from sport, where the law uses implied consent but might be aided by unreasonable risk analysis instead, is that of horseplay. In Aitken, the Court Martial Appeal Court quashed the convictions of RAF officers who, as part of an extended session of horseplay, unintentionally caused a fellow officer serious harm by setting fire to his fire-resistant suit.25 The basis for this finding was that the victim’s presence in the activities was capable of constituting implicit consent to the risk, in the defendants’ honest belief, and that this was sufficient to negate liability,26 referring to the earlier case of Jones.27 Again, instead of directly using consent analysis, it would be possible to say that in the circumstances of horseplay amongst seemingly-willing participant RAF officers, the risk was a reasonable one to take. This, however, raises two important points surrounding consent and reasonable risks.

Honest yet unreasonable beliefs

Possibly the most controversial aspect of Aitken is the finding that genuine proof of an honest yet unreasonable belief in consent is sufficient to negate liability for an offence against the person in the horseplay context. This approach to the issue of mistake is a clear feature of the common law, as emphasised in the House of Lords decision in K,28 and it is only in the context of legislation such as the Sexual Offences Act 2003 that this position is now explicitly altered. If we analyse reckless horseplay through unreasonable risk analysis instead, does this position change? The simple answer is that it does not. The definition of recklessness adopted by Lord Bingham in G and R is clear that the objective question of whether a risk is reasonable to take is assessed on the facts as the defendant believed them to be.29 If the defendant perceives a participant to be more willing and comfortable in a horseplay situation than the participant in fact is, this mistake will be taken into account in the defendant’s favour.30 There is, however, a subtle beneficial difference in that the unreasonable risk analysis is less binary than consent. If a victim must either have consented or not consented, it is easy for the defence (unjustly) to assert even a remote possibility that the defendant simply thought the victim consented, leading to an all-or-nothing situation. By contrast, the unreasonable risk analysis adverts the jury to a slightly wider range of factors in the defendant’s belief which they may be able to assess more carefully.

The second big question raised by Aitken and the other cases is why horseplay has been recognised as a category where consent is effective in negating liability. An explanation based on the public interest (if that is the best principle available) is not obviously convincing. This issue highlights a question of how exactly consent itself fits into the proposed unreasonable risk analysis, which has largely been ignored by this article up to this point. One option would be to say that, as a general rule, a risk is reasonable to take in the circumstances if (1) the victim consents and (2) something else about the circumstances (for example, whether the situation fits into a recognised sporting or horseplay context). This approach would do very little in terms of re-analysing the existing law but adopting it would be to assert the ‘moral magic’ of consent. That is to say, the analysis would afford the concept of consent the privilege that it and only it is capable of converting some types of criminal conduct into non-criminal conduct.

This approach is not, however, the best way to explain the unreasonable risk concept, and this arguably tells us something about the nature of consent.

Instead, the position argued for is the following: although the ‘categories of consent’ that have arisen in the English case law may have some overarching basis in reasonableness, consent is not powerful enough to facilitate a general formula of the form ‘consent plus something else equals reasonable’. Consent, rather than a magic binary which, if present, has intrinsic value, is instead a combination of various factors which themselves individually have weight in the reasonableness analysis. This is clear in the sports context, where implicit consent is constructed out of some of these factors, but may apply more generally. The factors part of consent might include the relevant situations (for instance, age, acquaintance) of the other participant(s) in the activity, the context, and (most obviously) the extent to which other participant(s) appear to be willing. The benefit of analysing (at least) the sports and horseplay scenarios through the unreasonable risk element rather than consent directly is that this rightly denies consent a role as a single, binary concept. Attention is focused more widely on all the relevant factors, and so, for instance, critics of the current law of horseplay can argue that it is unreasonable to take serious risks in these circumstances, perhaps particularly where children are involved.31

IV – Everyday touching and the issue of intention

A further situation of non-liability which the unreasonable risk element of recklessness can begin to explain is the accidental touching that often occurs in everyday life. If I rush onto a busy train, I may well foresee brushing past and coming into physical contact with other passengers. I am clearly not guilty of battery, and one way to explain this is to say that the risk of touching other people is no doubt a reasonable one to take in the circumstances.

Everyday touching was addressed in Collins v Wilcock,32 where the Divisional Court recognised “a broader exception [to battery] to allow for the exigencies of everyday life”.33 Goff LJ noted one way to analyse this exception was through implied consent: presence on a train or at a supermarket (for instance) entails accepting the risk of some accidental touching.34 The problems with this approach broadly mirror those identified above in the sports context:35 the boundaries of the implied everyday consent seem to be more a construction of law than a conscious choice by potential victims, and it is not always clear that individuals can withdraw their consent to everyday touching.36 An alternative approach Goff LJ noted was to say that everyday touching simply lies outside the scope of battery.37 This avoids the implied consent difficulty, but encounters the conceptual problem that what is acceptable depends on the circumstances,38 and so it is not the case that a particular type of physical intrusion is inherently excluded from the bodily autonomy rights protected by battery law. An analysis based on what is reasonable in the circumstances would appear to fit better.

However, the everyday touching examples also highlight a key shortcoming of focusing on the unreasonable risk element: the analysis can only apply to cases of reckless offending. An example of everyday touching explicitly recognised as lawful in Collins v Wilcock was “touching a person for the purpose of engaging his attention”39 such as tapping someone gently on the shoulder. This touching is intentional rather than reckless and so it is impossible to say that liability is negated because a reasonable risk has been taken.40 The analysis also fails in the context of intentional harms in sport, for example, such as boxing. The unreasonable risk element therefore cannot fully replace the concept of consent or the rules surrounding everyday touching. Nevertheless, understanding the true operation of the unreasonable risk element in the recklessness contexts highlights valuable points about the difference between intention and recklessness.

If we assume for practical purposes that the core idea of intention is factoring a consequence into one’s plan or goal,41 we can build a distinction between intended consequences and recklessly risked consequences. For instance, if it is my plan to cause a prohibited consequence (and do in fact cause it), the intention means that I am inherently mentally linked with this prohibited consequence. That is not to say that I cannot escape criminal liability, but that the way I must exonerate myself is by proving a specific supervening defence. Conversely, if I take the risk of causing a prohibited consequence (and the risk materialises), but it is not my plan to cause it, the mental ‘link’ between me and the consequence only exists (on the current English law) if I foresaw the risk, and the risk was an unreasonable one to take.

The practical effect of this can be illustrated with an example. If I slap you on the arm, and that is my goal, the only way to avoid criminal liability is to argue that, despite my prohibited purpose, the case fits within a specific defence (or negation of liability) such as self-defence, duress or consent.42 By contrast, if I slap you on the arm, but my goal is to kill a wasp which has landed there, there are two general routes by which I can avoid criminal liability (quite independent of any specific defences which may also apply). I can argue that I genuinely did not foresee that killing the wasp would cause me to batter you, or that in the circumstances, it was reasonable to take the (almost certain) risk of battering you.

The difference between the general routes to avoiding liability in recklessness and the need to rely on a specific defence to intentional wrongs is significant in that it explains one sense in which intention is ‘worse’ or ‘more serious’ than recklessness. The need for a specific defence to intentional wrongs is emphasised by the courts’ constant determination to avoid establishing a general lesser-of-two-evils necessity defence,43 and the upshot of this is that there is something inherently wrong about intending a prohibited consequence, such that the law is anxious to avoid permitting individuals intentionally to cause harm. That is not to say that there are no specific justificatory defences to intentional wrongs: self-defence, for instance, may be an example. But the justification element of (for example) self-defence comes from something inherent in the content of that defence. Intentional wrongdoing is otherwise inherently wrong. By contrast, merely risking a prohibited consequence is less serious in that it will only be wrong where, in the circumstances, there is nothing general to justify taking the risk.44

Whilst a more detailed theoretical discussion of the differences between intention and recklessness is beyond the scope of this article, the distinction drawn above does invite a brief side point about Woollin.45 The House of Lords allowed the direction where necessary (and ostensibly only in murder cases) that the jury “are not entitled to [find] the necessary intention, unless they feel sure that death or serious bodily harm was a virtual certainty (barring some unforeseen intervention) as a result of the defendant’s actions and that the defendant appreciated that such was the case.”46 The question is raised, therefore, whether Woollin ‘intent’ (as it might be called) is properly understood as intention, or is instead a special form of recklessness sufficient for murder. A crucial detail of the Woollin direction is the negative formulation ‘not entitled to find…unless’,47 which allows the jury ‘moral elbow room’ to decide that a defendant who foresaw a consequence as virtually certain is nevertheless not guilty of murder.48 This would appear to be a general (albeit limited) way of avoiding liability, more comparable to the unreasonable risk element of recklessness than any normal aspect of intention. To modify the classic example from Section I, if I break up with my romantic partner, foreseeing that it is virtually certain he will develop a recognised psychological condition as a result, I can probably avoid criminal liability even though, following Woollin, the jury have the option to equate virtually certain foresight with ‘intention’.49 If, therefore, the availability of a (comparatively) general reasonableness ‘defence’ is characteristic of recklessness (as opposed to intention), Woollin would appear to be best described as a special (very culpable) form of recklessness. This also helps to move away from the model that intention is inherently more culpable than recklessness: where the risk is very unreasonable (for example, because of its seriousness, high chance of occurring, or a combination of the two), recklessness could be equally or more culpable.50

V – Conclusion

By way of conclusion, it is worth summarising three key reasons to focus on the unreasonable risk element of recklessness. Firstly, it is useful to move the discussion and judicial focus in recklessness from the intense debate between advertence and inadvertence. Whether or not subjective foresight is always necessary to establish criminal liability, the core idea of recklessness should be understood in the unreasonableness of the risk taken, which links the defendant’s mental state to the prohibited outcome. Secondly, moving away from reasoning directly focusing on consent in sports and horseplay, and instead analysing whether the circumstances (including the victim’s willingness) justify taking the risk, puts more pressure on courts and the law to avoid the overly simplistic use of consent as ‘moral magic’. It is not enough to establish that the victim consented: we must also explain why that consent negates criminal liability. Finally, the inability of unreasonable risk analysis to deal with situations of intentional harm should be understood (to use a computer programming analogy) as a ’feature, not a bug’. The unreasonable risk element, combined with the requirement of awareness of the risk, provides a general route to avoiding criminal liability for reckless harms, which differentiates recklessness from intention.

Notes

  1. [2003] UKHL 50, [2004] 1 AC 1034 [41] (Lord Bingham) ↑

  2. [1982] AC 341 (HL) ↑

  3. [1979] QB 695 (CA) 703 (Geoffrey Lane LJ) ↑

  4. [1998] AC 147 (HL) ↑

  5. I take this example from a revision class given by Trenton Sewell. ↑

  6. [2004] EWCA Crim 3246, [2005] 1 WLR 910 [15] ↑

  7. Ibid ↑

  8. It is interesting to speculate on the implications of considering the defendant’s state of mind as a relevant factor: perhaps the effect is to make the awareness of risk element of recklessness less of a binary than it might seem. Alternatively, Lord Woolf may simply be referring to the need to decide whether the infliction of an injury was in fact intentional. ↑

  9. See, for instance (albeit in the tort context) Tomlinson v Congleton Borough Council [2003] UKHL 47, [2004] 1 AC 46 [34] (Lord Hoffmann) ↑

  10. It might alternatively be argued that Lord Woolf is not defining the boundaries of the (implicit) consent but rather defining the boundaries of the ‘lawful sport’ category. Although this approach can derive a great deal of support from Lord Woolf’s phrasing, the problem is that to negative liability, the facts must fall within the category and there must in fact be consent. This consent must come from a presumption that players, by taking part in lawful sports, consent to all risks exactly up to the point that the conduct stops being lawful sport. This means that even if Lord Woolf was purporting to set out the boundaries of lawful sport, he was in practice also setting out the boundaries of the actual consent. ↑

  11. [2004] EWCA Crim 3246, [2005] 1 WLR 910 [15], as referred to above in relation to the ‘heat of the moment’. ↑

  12. This is a normal and workable feature of the law’s use of ‘reasonableness’, as seen frequently in tort law: on the facts of Nettleship v Weston [1971] 2 QB 691 (CA), it might seem ‘reasonable’ by common sense for an inexperienced learner driver to make mistakes, but the legal standard applied is different. ↑

  13. See, albeit in the everyday touching context Jonathan Herring, Criminal Law (13th edn, OUP 2023), 102 ↑

  14. [1994] 1 AC 212 (HL) ↑

  15. See the observations in BM [2018] EWCA Crim 560, [2019] QB 1 [38]-[41] (Lord Burnett CJ), where the court resolved not to recognise new categories by reference to any wider principle, but rather develop the law only by analogy to existing ones. Although at [40] the court identified two aspects “which appear to underpin almost all of the exceptions” (social benefit and unreasonable to criminalise the activity), it is clear that there is no test against which each individual set of facts can be tested. The categories are, to all intents and purposes, fixed. ↑

  16. For present purposes, this excludes boxing or any other activity where the infliction of harm is intentional, on which see Section IV below. ↑

  17. Lord Woolf referred in Barnes at [11] to the obvious public policy merits. ↑

  18. See the text to footnote 10 above: even if Lord Woolf is setting out the boundaries of ‘lawful sport’, the factors must relate to the extent of the implied consent itself as well. The point is that the courts will explain whether particular conduct is reasonable in the sporting context by reference to a wide range of factors. Simply explaining that these factors point to the behaviour being ‘lawful sport’ is less clear in explaining why it is acceptable to injure others as part of lawful sport. ↑

  19. An interesting point surrounding the need to prove a lack of consent in the sexual offences context is made by Jonathan Herring and Sorcha McCormack, ‘Reframing rape: from consent to responsibility’. Gender and Justice, Bristol University Press, 2025. They argue (Section 6 in particular) that consent as a matter of fact unjustly puts the ‘spotlight’ on the victim, leading to traumatic trials and spurious defences being raised, and that instead the concept of reasonable belief in consent (which focuses more on the defendant) should entirely replace factual consent. ↑

  20. See Section III below on mistakes of fact ↑

  21. Parmenter [1992] 1 AC 699 (HL) ↑

  22. Note that the amount of harm actually foreseen might be relevant to the reasonableness analysis (alongside other factors): if I do an act which I perceive as risking only minor harm, this risk could be reasonable in appropriate circumstances, and the fact that more serious harm unluckily results makes no difference. The risk was reasonable and so I lack mens rea. It is not a problem for the analysis that, were the risk unreasonable to take, the label (and penalty) of the offence I would be convicted of would reflect more than the minor harm foreseen. That said, the issue is, admittedly, messy: this analysis would appear to take a lot of the ‘bite’ out of constructive offences when we consider that the facts on which the reasonableness assessment is to be based are to be taken (at least in theory) entirely as the defendant perceived them to be. This is an inevitable tension between strongly subjective recklessness and constructive liability, solved in practice by the fact that the situations which justify risking even minor harm are relatively limited, and such justifications rarely succeed in circumstances where major harm is in fact risked (even without taking the actual risk of major harm into account). One such rare situation is horseplay, for which see Section III. ↑

  23. [2025] EWCA Civ 776 [24] (Nicola Davies LJ) ↑

  24. Ibid ↑

  25. [1992] 1 WLR 1006 (CMAC) ↑

  26. [1992] 1 WLR 1006 (CMAC) 1018 (Cazalet J) ↑

  27. (1986) 83 Cr App R 375 (CA) ↑

  28. [2001] UKHL 41, [2002] 1 AC 462 ↑

  29. [2003] UKHL 50, [2004] 1 AC 1034 [41] (Lord Bingham) ↑

  30. This is assuming there is no self-induced intoxication, which probably alters the honest mistake analysis: Richardson and Irwin [1999] 1 Cr App R 392 (CA) ↑

  31. As was the case in Jones (1986) 83 Cr App R 375 (CA) ↑

  32. [1984] 1 WLR 1172 (DC) ↑

  33. [1984] 1 WLR 1172 (DC) 1177 (Goff LJ) ↑

  34. Ibid ↑

  35. See Section II above ↑

  36. Jonathan Herring, Criminal Law (13th edn, OUP 2023), 102 ↑

  37. That is, it “fall[s] within a general exception embracing all physical contact which is generally acceptable in the ordinary conduct of daily life”: [1984] 1 WLR 1172 (DC) 1177 (Goff LJ) ↑

  38. [1984] 1 WLR 1172 (DC) 1178 (Goff LJ) ↑

  39. [1984] 1 WLR 1172 (DC) 1177-8 (Goff LJ) ↑

  40. Unless we take an extremely narrow view of what constitutes intention and suggest that if I tap you on your shoulder to get your attention, I ‘intend’ only to get your attention, and merely ‘risk’ touching you. Such a view might derive some support from Steane [1947] KB 997 but is realistically untenable in that it confuses intention with motive. ↑

  41. John Finnis, ‘Intention and side-effects’ in Raymond Frey and Christopher Morris (eds) Liability and Responsibility (CUP 1991) ↑

  42. The phrasing of ‘defence’ is used quite loosely in this section: none of the analysis here relies on any assumption about the burden of proof or anything else specific to defences as opposed to elements negating liability. The analysis in Section IV is fully compatible with the view that non-consent is part of the actus reus of offences against the person. ↑

  43. See the attempt of Ward LJ in Re A (Conjoined Twins) [2001] Fam 147 (CA) 204-5 to confine the narrow necessity defence he identified to the specific facts of this “very unique case”, and also of course Dudley and Stephens (1884) 14 QBD 273 (QB) ↑

  44. This only establishes intention as ‘more serious’ (more inherently wrong) than recklessness in this technical sense: see the text to footnote 50 below: there is nothing in this analysis that suggests all intentional wrongs are always more culpable. ↑

  45. [1999] 1 AC 82 (HL) ↑

  46. [1999] 1 AC 82 (HL) 96 (Lord Steyn) ↑

  47. The importance of allowing the jury the option not to find intention was subsequently stressed in Matthews and Alleyne [2003] EWCA Crim 192, albeit that the conviction was nevertheless safe in that case. ↑

  48. See Andrew Ashworth, ‘Principles, pragmatism and the Law Commission’s recommendations on homicide law reform’ [2007] Crim LR 333, who is critical of affording the jury this “latitude” and regarded its inclusion in the Law Commission’s 2006 Report on Murder, Manslaughter and Infanticide as “unfortunate”. ↑

  49. Of course, this hypothetical assumes that Woollin applies to charges other than murder. In practice, Woollin is most meaningful in murder cases because these are among the relatively few situations which require a finding of intention rather than recklessness. ↑

  50. Finnis notes (at 60) Glanville Williams’ example of the ‘eccentric surgeon’ who decides to remove a patient’s heart for an experiment, not intending to kill, but rather “quite happy that the patient should go on living if he could do so without a heart”. Finnis suggests this would be murder “[b]ut not because D intends to kill P”. Instead, using D as an instrument in this way (as I would put it: taking a seriously unreasonable risk with D’s body) is equally (or arguably more) culpable. ↑

Cite as: Philip Kimber, ‘Unreasonable risks: The overlooked element of recklessness in the English criminal law’ (2026) 1 OULPR 27