Unnecessary Self-defence

TL;DR

This paper argues that the legal requirement of necessity in self-defense is inconsistently applied and should be removed.

Contribution

The paper identifies inconsistencies in the necessity requirement and proposes its removal.

Findings

The necessity requirement in self-defense is formulated inconsistently.

Abandoning the necessity requirement would not create a problematic legal gap.

Necessity remains relevant in assessing the reasonableness of force.

Abstract

Self-defence is traditionally said to contain a necessity requirement, according to which defensive force is lawful only if it is necessary. But the necessity requirement is formulated inconsistently, and these inconsistencies substantially alter the scope of the defence. This article explains these inconsistencies. It concludes that it would be preferable to abandon the necessity requirement altogether. This would not leave a problematic gap in the law, because necessity would remain an important consideration when judging whether any use of force was reasonable.

Tables

Table 1.. Disambiguations of ‘the’ necessity requirement

	Less demanding	More demanding
How necessary?	Reasonable or necessary;Sufficiently close to necessary	Strictly necessary
(Strictly) Necessary to What End?	Vindicating civil rights	Self-defence
What Must be Necessary? (if force is required)	Some force	The degree of force used
What Must be Necessary? (if force is not required)	D’s action fell within the set of similarly justified actions, the taking of at least one of which was necessary	D’s precise action
Necessary in fact or belief?	Believed to be necessary	Factually necessary, relative to D’s beliefs as to the primary facts

Full text

Introduction

It is widely claimed that defensive force is lawful only if it is necessary.^1^ Call this the necessity requirement.^2^ The necessity requirement can be found at least as early as the 19th century, in which English criminal cases claimed that, to be lawful, force must be used in ‘necessary self-defence’,^3^ or when ‘absolutely necessary’.^4^ The leading 20th-century case held that, when directing juries, ‘All that is needed is a clear exposition … of the conception of necessary self-defence’.^5^ The necessity requirement finds continued support in modern English authorities,^6^ as well as enjoying broad international support, including in the German Criminal Code and US Model Penal Code.^7^ It is also widely defended by moral philosophers and legal scholars.^8^ According to Boaz Sangero:

The requirement that the action of the actor against the aggressor … must be necessary in order to repel the aggressor, is perhaps the most accepted condition of private defence in all the legal systems and from ancient history … Moreover, it is not only that the requirement of necessity is not under dispute, but that it is even widely accepted to see it as a fundamental and significant condition of private defence.^9^

Notwithstanding such claims, the thesis of this article is that the authority supporting a necessity requirement in English law is much weaker than it appears. Further, I will suggest that the law should abandon the language of the necessity requirement altogether.

There are two reasons why the apparently impressive weight of authority favouring a necessity requirement is weaker than it appears. Firstly, the necessity requirement forms no part of the ratio of any case. I have cited many judgments invoking it, but none offer a reasoned defence of the requirement after considering its merits.^10^ This might be surprising. But consider a different element of self-defence. For well over 100 years, the orthodox position was that a defendant claiming self-defence could rely on a mistaken belief as to the facts only if that mistaken belief was held on reasonable grounds.^11^ But the merits of this reasonable belief requirement were not considered directly by any appellate court until 1982.^12^ Once the requirement was finally questioned, it was then repeatedly contested in the highest courts, and ultimately abandoned.^13^ Today the necessity requirement is in a similar position to the reasonable belief requirement prior to 1982. It is repeated as the orthodox legal position without an appellate court ever considering its merits.

The second reason why the authoritative support for a necessity requirement is weaker than it appears is that the authorities which refer to a necessity requirement do so in different and incompatible ways. I survey these differences in section 2. It follows that, even on the highest view of the authorities, the law does not contain a single necessity requirement supported by a great weight of authority. Rather, it contains several competing necessity requirements, each backed by much slenderer authority.

Together, these reasons destabilise the apparently strong authority upholding a necessity requirement in English law. But should the law contain a necessity requirement? Given the several competing formulations, to defend ‘the necessity requirement’ requires selecting one to defend. In section 3, I argue that the most legally defensible formulations do not in fact make necessity a requirement of self-defence at all, while formulations which specify a genuine requirement have little legal support and minimal effect. Worse still, that (minimal) effect is regrettable. I therefore propose abandoning the language of a necessity requirement altogether.

Abandoning the necessity requirement might sound radical. But it is not. For I am not claiming that the necessity (or otherwise) of defensive force is irrelevant to its lawfulness. Defensive force must still be reasonable, and its necessity is an important factor bearing on its reasonableness. What I am claiming, rather, is that the necessity of defensive force is not and should not be a necessary condition—a requirement—of reasonableness and hence lawfulness. In suitable cases, unnecessary defensive force can yet be reasonable and therefore lawful. (By analogy: verbal agreement is an important factor bearing on the existence of valid consent but is not a necessary condition of valid consent.) This possibility is sometimes overlooked because ‘necessary’ and ‘reasonable’ are treated as synonyms. But, as I explain in section 3, the notion of necessity is stricter than the notion of reasonableness. If the true test of lawful defensive force is reasonableness, as I contend, then courts should not reformulate this test in terms of necessity.

If I am right that the authorities contain many substantively different formulations of the necessity requirement, then the courts should, as a minimum, attempt to specify more precisely which formulation should prevail. But what if I am right that the necessity requirement ought to be abandoned? Are courts at liberty to do so, given the litany of authorities cited above? In section 4, I claim that they may do so, and indeed that Parliament has already, if implicitly, acknowledged that necessity is but one (unnecessary) reasonableness factor, rather than a necessary condition of lawful self-defence. Nonetheless, clearer codification would helpfully clarify the roles of reasonableness, necessity and proportionality in evaluating the lawfulness of self-defence.

Disambiguating Different Necessity Requirements

Imagine that C is about to shoot D. D can avoid being shot only by:

Necessary: Shooting C first. Unnecessary: Either (i) shooting C first or else (ii) tasering C.

In Necessary, it is necessary for D to shoot C. In Unnecessary, it is not, because D had another, gentler, defensive option. These cases illustrate the comparative structure of necessity judgments: they involve weighing potential defensive options against each other. (By contrast, proportionality judgments weigh the defensive action against the attacker’s threat.)

But necessity judgments are ambiguous. Perhaps it was not absolutely necessary to shoot C as judged from an impartial perspective, but it was reasonably necessary to use force against C as judged from the defendant’s perspective. Which is the relevant test? Generalised, we can ask of any necessity judgment: How necessary? What must be necessary? From what perspective is necessity judged? This section demonstrates how different authorities have offered substantively different and incompatible answers to these questions when formulating ‘the necessity requirement’. It follows that, far from endorsing a single necessity requirement, the authorities actually endorse many substantively different requirements.

How Necessary?

A.

The first difference between different formulations of the necessity requirement concerns how necessary it must be to use defensive force.^14^ While some authorities refer to ‘necessity’, simpliciter,^15^ between others we can distinguish two views:

Strict necessity: Defensive force is lawful only if it was strictly necessary.^16^ Reasonable necessity: Defensive force is lawful only if it was reasonably necessary.^17^

The difference is significant. Consider an analogy. In a criminal trial, the prosecution must prove the defendant’s guilt to the criminal standard of proof.^18^ This requires a jury to be ‘sure’ of guilt. Directing juries on the standard of proof in any terms other than being ‘sure’ is likely to constitute a misdirection.^19^ It would be an obvious misdirection to direct a jury that they must be ‘reasonably sure’ of guilt. A jury might (rightly) think that being ‘reasonably’ sure was compatible with harbouring reasonable doubts. In other words, being unsure. Analogously, directing a jury that defensive force must be ‘reasonably necessary’ is very different from directing a jury that defensive force must be ‘necessary’, simpliciter. So we must choose between these formulations.

Strict necessity is too strict. Consider:

Choice: C is about to punch D. D can choose between two equally effective defensive options: (i) punching or (ii) kicking C. Punching is slightly more forceful than kicking.^20^

Punching is not strictly necessary, given the slightly gentler alternative of kicking. But denying self-defence on this basis, should D punch C, would make the law far too demanding. Nor is it how the law works in practice.^21^

That leaves reasonable necessity. But making sense of reasonable necessity is not straightforward. For ‘necessary’ is not a gradable concept. It does not come in degrees. Something is either necessary or it is not.^22^ In standard philosophical usage, ‘necessary’ means always the case, as distinguished from merely possible.^23^ A ‘necessary condition’ means just what it says. With this in mind, ‘How necessary?’ sounds like a malformed question and ‘Reasonably necessary’ a malformed answer.

One solution to this problem is to interpret ‘reasonably necessary’ as ‘reasonable or necessary’. This is the formulation offered by one leading textbook.^24^ As Sangero writes,

the requirement of necessity should not be interpreted in a strict and absurd fashion such that if the actor had another alternative way to repel the aggressor this would suffice to negate the existence of necessity. There may be situations in which the actor has several possible ways of action … [If D does not choose] the most moderate way … [but rather] chooses another reasonable way, his act is likely to comply with the requirement of necessity.^25^

The problem with this interpretation is that judging whether an action is reasonable is not the same as judging whether it is necessary. Consider:

Burglary: D catches C burgling his home. C is clearly unarmed, and much weaker than D, but will not leave. D picks up C and throws him out of the front door, causing scratches and bruising. D could have waited until C left of his own accord, or else removed him more delicately.

It is perfectly coherent to judge D’s use of force to be unnecessary—he had alternative, gentler, defensive options—yet reasonable. The standard of necessity is more demanding than the standard of reasonableness.^26^ If the necessity requirement only required D’s actions to be either necessary or reasonable, necessity would no longer be a necessary condition of lawful self-defence. The true test would be reasonableness.^27^

A different way to make sense of ‘reasonably necessary’ is to interpret ‘reasonably’ as qualifying ‘necessary’. If it was strictly necessary for the defendant to use (say) 1000 newtons of defensive force, then it may be reasonably necessary to use a degree of force which exceeds but is sufficiently close to that strictly necessary amount (say, 1500 newtons).^28^ But the difficulty with this interpretation is identifying what counts as ‘sufficiently close’. If only a narrow margin of error is permitted, it is likely to deny self-defence in worthy cases like Burglary. But if a wide margin of error is permitted, such as only forbidding force grossly exceeding the necessary amount, it would render the necessity requirement functionally redundant, for such grossly excessive force would be unreasonable even in the absence of a necessity requirement. Once again, it would effectively abandon the necessity requirement in favour of a generic reasonableness requirement which incorporated considerations of necessity.

Interpreting reasonable necessity as either ‘reasonable or necessary’ or ‘sufficiently close to strictly necessary’ is effectively to abandon a genuine necessity requirement in favour of necessity being just one factor going towards reasonableness. But there is another way to make sense of ‘reasonably necessary’. To see this, consider a problem with strict necessity: in one sense, defensive force is never necessary. Defendants always have at least one alternative to using defensive force. They could submit to the threat. Because submission is always an option, there is a sense in which defensive force is never strictly necessary. But this interpretation of strict necessity would make all defensive force unlawful. It would make the necessity requirement ‘strict and absurd’. Noticing this problem suggests a way to make sense of degrees of necessity even though the concept of necessity itself is not gradable.

Here is the thought: whether an action is necessary is assessed relative to some end.^29^ More actions are necessary to achieve a complex end than a narrow end. More soldiers are necessary to win a war than a skirmish. More musical notes are necessary to play a concerto than a chord. Likewise, more (and stronger) defensive actions are necessary to achieve the end of not conceding an inch (which might require aggressive physical measures), while fewer (and weaker) defensive actions are necessary to achieve the end of avoiding attack (which might require only retreat). We can therefore make sense of reasonable necessity as meaning ‘Necessary relative to a wider range of ends’ and strict necessity as ‘Necessary relative to a narrower range of ends’.^30^ Necessity judgments remain binary: for any given end, defensive force is necessary or not. But the necessity requirement can vary in strength by altering the range of ends against which necessity is judged.

This thought is obliquely reflected in the authorities, which often specify the end to which defensive force must be necessary. And different authorities offer different formulations. Many claim that force must be necessary for the defendant ‘to defend himself’. They imply:

The end of self-defence: Force must be necessary for D to defend themselves.^31^

(Mutatis mutandis for defence of another, defence of property, prevention of crime.)^32^  The end of self-defence is clearly more generous than the end of doing anything else. It does not demand submission: it permits defence. But it remains strict. For defence can often be achieved without force. Avoiding a threat is equally or more defensively effective than repelling it. It follows that the end of self-defence implies a duty to retreat. As Andrew Ashworth explains, ‘when an individual’s purpose in a threatening situation is to save himself from injury or death, it cannot be necessary for him to inflict harm on his assailant if there is a safe avenue of withdrawal open to him’.^33^

English law did historically contain a duty to retreat.^34^ But the duty to retreat was rejected by the courts,^35^ and ultimately Parliament, which clarified that: ‘[The possibility of D retreating] is to be considered (so far as relevant) as a factor to be taken into account [when judging the reasonableness of the force used], rather than as giving rise to a duty to retreat.’^36^

If defensive force was lawful only if necessary to the end of self-defence, this would entail a duty to retreat. But as lawful self-defence does not require retreat, it therefore does not require force to be necessary to the end of self-defence.

To which other end can we index the necessity requirement? First consider why English law rejected the duty to retreat. It did so because a duty to retreat inappropriately disempowers reasonable defensive actions. Consider:

Threat: C and D are both in the same pub. C threatens imminently to attack D unless D leaves. D knows that C will not attack if D leaves. But, instead of leaving, D responds by forcefully pushing away C.

A duty to retreat forces D to choose between relinquishing their civil rights (to remain in situ) or else being criminally liable for defending those rights. This was considered unfair. It was considered preferable that such citizens were afforded a defence if they used defensive force to vindicate their liberties, providing they did so in a reasonable way. This suggests a natural alternative to the end of self-defence, namely:

The end of rights vindication: Defensive force must be necessary to vindicate the defendant’s civil rights.^37^

Framing the necessity requirement in this way expands the availability of lawful self-defence to cover cases like Threat without abandoning necessity’s strict literal meaning. It makes sense of talk of ‘reasonable necessity’ without retreating to ‘reasonable or necessary’. It accommodates the absence of a duty to retreat in English law. And it finds implicit authoritative support in Field, a case with facts similar to Threat, in which the Court of Appeal permitted self-defence.^38^ As one textbook puts the point: ‘The question of sufficient necessity to use force should normally take, as its baseline, D’s right to the peaceful enjoyment of her autonomous choices.’^39^ If this is right, the necessity requirement does not require force to be necessary for self-defence. It is sufficient that force was necessary to vindicate one’s civil rights.

Interpreting the necessity requirement relative to the end of rights vindication undeniably weakens the necessity requirement, relative to the end of self-defence. We have a very wide range of civil rights. Many are trivial. If the vindication of any such rights qualified as a relevant end of the necessity requirement, then it would make the requirement very weak indeed.^40^ It would rule out self-defence only when the defendant used force to protect no legitimate interest whatsoever.^41^ If this is the law, then the necessity ‘requirement’ all but shades into one factor going towards reasonableness.

What Must Be Necessary?

B.

A second difference between different formulations of the necessity requirement concerns what it is that must be necessary; the object of necessity claims.

Is force necessary?

(i)

The most popular formulation is that it must be necessary to use force.^42^ But it is controversial whether self-defence is available only in relation to prima facie offending involving the use of force. Call this the force requirement.^43^ If self-defence contains the force requirement, then it would be unremarkable for the necessity requirement to follow suit and to demand that it was necessary to use force. But if self-defence does not contain the force requirement, then this would be bizarre, given that, in fact and in law, force was not (otherwise) necessary.

The force requirement was endorsed by the Divisional Court in Blake. The defendant protested the Gulf War, which he believed to be illegal, by writing a biblical quotation upon a concrete pillar with marker pen.^44^ The court rejected the defence of prevention of crime on the basis that this defence ‘is only available in relevant crimes committed by the use of force’ and Blake’s graffiti did not involve force. That self-defence contains the force requirement was subsequently regarded as ‘conventional’ by the Court of Appeal,^45^ and is presupposed by the relevant statutes, which are framed in terms of the use of force, specifically.^46^ The force requirement has also been defended by the Law Commission.^47^

Against the force requirement, JC Smith points out its potential absurdity: ‘if Mr. Blake had taken a hammer and chisel and cut the letters into the concrete he might have had a defence; but as he merely wrote them with a felt pen, he did not!’^48^ Generalising the point, one textbook claims that ‘if D can justify an act of force as a response to particular circumstances, a fortiori justification should be accorded to a lesser, non-forceful response to the same situation’.^49^ This view has found favour in the Divisional Court, albeit obiter, in Oraki.^50^

The Court of Appeal has avoided the most absurd implications of the force requirement by watering down the meaning of ‘force’ to mean any application of energy whatsoever.^51^ But this does not address truly non-forceful examples, like driving offences involving retreat (eg speeding or running a red light) rather than forceful attack. At any rate, the courts have never offered a principled defence of the force requirement. For these reasons, it is arguable that the courts should reject the force requirement.^52^

Whichever view we prefer, however, we face a further interpretative question. If force is necessary, then how much? And if force is not necessary, then what must be?

If force is necessary, then how much force is necessary?

(ii)

Assume for now that the law contains the force requirement. Two rival views remain as to how much force must be necessary:

Some force: It was necessary to use some force (versus none).^53^ Degree of force: It was necessary to use the degree of force which D used.^54^

The thought motivating some force is that once it is established that at least some force was necessary, as a first threshold, then questions as to the appropriate degree of force are more appropriately discussed under the rubric of proportionality or (other factors affecting) reasonableness.^55^

This thought can be doubted. Philosophers of self-defence standardly distinguish necessity from proportionality in the following way. Necessity concerns a relationship between the defender’s defensive harm and alternative defensive options open to the defender. By contrast, proportionality concerns a relationship between the attacker’s threatened harm and the defender’s defensive harm.^56^ To illustrate, consider:

Options: C is about to punch D. D can choose between exactly three equally effective defensive options: (i) shooting, (ii) punching or (iii) pushing C.

On the standard philosophical picture, shooting is both unnecessary (given the other options) and disproportionate (too extreme); punching is unnecessary (given the lesser option of pushing) but proportionate (equal to the threat); while pushing is both necessary and proportionate. This standard picture endorses degree of force. What makes shooting and punching unnecessary is not that no force was necessary—some force was necessary—but rather that this degree of force was unnecessary. Equally plainly, this picture still distinguishes the necessity requirement from the proportionality requirement, for punching was unnecessary but proportionate.

The idea motivating some force was that it helped to distinguish the necessity question from the proportionality question. The standard philosophical picture undermines this motivation: we can distinguish the two questions while endorsing degree of force. Does it follow that some force is mistaken?

Not necessarily. Philosophical theses do not translate directly into legal rules. We often better achieve our underlying aims through indirect rules (eg speed limits may cause safer driving than a rule directly commanding safe driving). The standard philosophical picture says that defensive force is morally permissible only if the degree of force was necessary. But it does not automatically follow that the law’s necessity requirement does or should duplicate that moral demand.

Remember: I do not dispute that the necessity to use the degree of force used is an important factor which bears on the reasonableness and hence lawfulness of any defensive force. Everyone agrees that, in Options, shooting would be unlawful. Degree of force says that this is because it violated the necessity requirement. Some force says that it does not violate the necessity requirement. But such force remains unreasonable, for the (distinct) reason that it is disproportionate. They get to the same answer, just via different routes. Of course, if all unnecessary degrees of force were automatically unreasonable, this would be a distinction without a difference. But that is not the case. Consider:

Burglary 2: D catches C burgling his home. C is clearly unarmed, and much weaker than D, but will not leave. D picks up C and throws him out of the front door, causing scratches and bruising. D could have evicted C by using slightly less force.

D’s degree of force was unnecessary because lesser force could have achieved the same end. According to degree of force, it would therefore be unlawful. But this seems too harsh. Unlike shooting in Options, this unnecessary degree of force still seems reasonable. We cannot reasonably expect people to escort burglars out of their houses using the minimum possible force. Even if D’s conduct is impermissible under some ideal moral theory, it would be astonishingly harsh on defenders to withhold self-defence on this basis. This is reflected in section 76 of the Criminal Justice and Immigration Act (CJIA), which says that ‘a person acting for a legitimate purpose may not be able to weigh to a nicety the exact measure of any necessary action’.^57^ This seems to envisage that an unnecessary degree of force could yet be reasonable in cases like Burglary 2.^58^ And this implies that the better interpretation of the necessity requirement is that it requires some force to be necessary, not the degree of force used.

Of course, this choice—between some force and degree of force—assumes that the law contains a force requirement. What if it does not?

If force is not necessary, then what is?

(iii)

While most cases claim that it must be necessary to use force, other cases are more precise. They suggest that what must be necessary was the specific action taken by the defendant: actions like resorting to a gun,^59^ detaining the attacker^60^ or even going on to a football pitch.^61^ But this specific action view is too specific. Consider:

Options 2: C is about to punch D. D can choose between exactly two equally effective defensive options: (i) punching or (ii) kicking C. Punching and kicking are equally forceful.

The specific action of punching is unnecessary because D could instead kick. And the specific action of kicking is unnecessary because D could instead punch. The specific action view implausibly concludes that both actions are unnecessary and hence unlawful. That cannot be right. Where defendants have a disjunctive set of defensive actions (A or B or C, etc), we do not want the mere existence of other possible actions to preclude lawful self-defence.^62^

The necessity requirement should not require defendants to take one specific necessary action. It should only require defendants to take a necessary action. One way to make sense of this idea is to say that what must be necessary is that the defendant’s action fell within the set of actions, the taking of at least one of which was necessary. That is: if D can take action A or B or C, then so long as action A was necessary absent actions B and C, and likewise for B and C, then D’s chosen action would count as necessary. Call this the set view. In Options 2, punching would have been necessary if kicking were not an option, and vice versa. Hence, the set view appropriately judges both punching and kicking to be necessary in the relevant sense.

Unfortunately, the set view fatally undermines necessity judgments. Consider:

Options 3: C is about to shoot D. D can choose between exactly two equally effective defensive options: (i) punching or (ii) shooting C.

There is a disjunctive set of actions (punching C, shooting C), where each action would be necessary absent the other. The set view therefore deems both punching and shooting to be necessary. But this cannot be right. Shooting when punching was equally effective is a paradigmatic example of unnecessary defensive force. (It is, in substance, the case I first labelled Unnecessary.) More generally, every defensive action would be necessary absent other options. The whole point of necessity judgments is to compare defensive actions against other possible defensive options. Yet the set view prohibits precisely such comparisons.

We want to say that punching and kicking are both necessary in Options 2, but that only punching and not shooting are necessary in Options 3. The reason, we might think, is that punching and kicking are equally justifiable, while punching and shooting are not. This can be achieved with the equally justified set view, which is as per the set view, but the disjunctive set of defensive options comprises only actions which are equally as justified as the defendant’s action. But consider:

Options 4: as per Options 2 (punching or kicking), but punching is marginally more harmful than kicking.

Punching is now (marginally) less justified than kicking. Punching therefore falls outside the set of equally justified actions and is therefore considered unnecessary on the equally justified set view. But this returns us to the problems associated with degree of force. The necessity requirement should not consider defensive force to be unnecessary (hence unlawful) merely because there was a marginally preferable alternative option.^63^

A plausible necessity requirement must permit non-optimal defensive actions whilst forbidding actions which are significantly worse than available alternatives. This implies that what must be necessary is that the defendant’s action fell not within a set of equally justified actions, but rather a set of similarly justified actions, the taking of at least one of which was necessary. Call this reasonable set view.

The reasonable set view solves the problems of the previous interpretations. Unlike both some force and degree of force, it permits non-forceful defensive actions (such as assault by threats or running a red light). Unlike the specific action view and the equally justified set view, it permits non-optimal defensive actions. And unlike the set view, it forbids highly sub-optimal defensive actions. The price of success, however, is significantly eroding the distinctive nature of necessity judgments. Consider:

Options 5: C is about to punch D. D can choose between exactly three equally effective defensive options: (i) punching C, (ii) causing C to suffer a mild allergic reaction or (iii) threatening to kill C (without intending to follow through).

I do not know whether a jury would deem these options to be sufficiently similarly justified. But what is clear is that this enquiry looks quite different from ‘classic’ necessity enquiries. We are no longer asking whether D’s force or action was necessary, but rather whether their action fell within a reasonable range. And this is functionally very similar to abandoning the necessity requirement altogether and replacing it with a generic assessment of reasonableness; of turning necessity into just one factor going to reasonableness.

Necessary in Fact or Belief?

C.

The third and final difference between different formulations of the necessity requirement concerns the perspective from which we judge necessity.

Some authorities judge necessity from an objective perspective. They claim that defensive force must be ‘necessary’, simpliciter.^64^ Other authorities judge necessity from a subjective perspective. They claim that the defendant must ‘honestly believe … that it was necessary to defend himself’.^65^ Call these views, respectively:

Factual necessity: Defensive force is lawful only if it was factually necessary. Believed necessity: Defensive force is lawful only if D believed that it was necessary.

(I refer to ‘defensive force’ for convenience, not because I endorse the force requirement.) To illustrate the difference, consider:

Mistake: D hears C approaching behind her. Believing she is about to be attacked, D hits C. In fact, C was harmlessly jogging.

D (subjectively) believed defensive force to be necessary despite it not being (objectively) factually necessary. These different formulations of the necessity requirement seemingly result in different outcomes.

But this appearance is misleading. As noted earlier, where defendants are mistaken as to the relevant facts, their conduct is judged relative to the facts which the defendant (subjectively) believed to exist, not the facts which (objectively) existed. It follows that factual necessity actually amounts to:

(Belief-relative) factual necessity: Defensive force is lawful only if D believed in facts, which, if true, would make force factually necessary.^66^

In other words, even if the necessity of using force is judged objectively, the relevant facts are judged subjectively.^67^ This brings the two perspectives closer together. In Mistake, D (subjectively) believed that force was necessary. And D believed this because D (subjectively) believed in facts (that C was about to attack her) which, if true, would (objectively) make it factually necessary to use force.^68^  (Belief-relative) factual necessity and believed necessity therefore result in the same outcome in Mistake.^69^

But while (belief-relative) factual necessity is closer to believed necessity than was factual necessity, they are not identical. For a defendant can be mistaken that defensive force is necessary without being mistaken about the underlying facts. Consider:

Mistake 2: D sees C approach. Correctly believing that C plans to snatch D’s phone, D punches C. D failed to consider that he could have avoided the theft simply by putting his phone in his pocket.

D believed his defensive force to be necessary. He satisfies believed necessity. This belief was mistaken. He could have avoided theft simply by hiding his phone.^70^ Crucially, however, his mistake derived from his failure to reason about alternative options—about the necessity of defensive force per se—rather than a failure to perceive the relevant facts. He did not believe in any facts which, if true, would have made defensive force necessary.^71^ He therefore fails to satisfy (belief-relative) factual necessity.

(Belief-relative) factual necessity permits reliance on mistaken self-defence only where D was mistaken about primary facts, as in Mistake. By contrast, believed necessity permits reliance on mistaken self-defence even where D was mistaken about the demands of necessity itself, as in Mistake 2. Believed necessity is therefore more lenient than (belief-relative) factual necessity. They imply substantively different necessity requirements.^72^

Which formulation is to be preferred? On the one hand, defendants cannot typically rely on mistakes about the very evaluative standards by which they are to be judged. Elsewhere in the criminal law, it is irrelevant that defendants believe their conduct to be reasonable or honest. What counts is whether their conduct was in fact reasonable or honest (albeit judged relative to whatever facts they (perhaps mistakenly) believed to exist).^73^ This general approach supports (belief-relative) factual necessity.

On the other hand, proponents of the necessity requirement do not claim it to be a sufficient condition of lawful self-defence. Even necessary defensive force must still be (objectively) reasonable, which includes considerations of proportionality.^74^ Given this further control mechanism, demanding even (belief-relative) factual necessity may be to demand too much. If defendants like D in Mistake 2 are (at least sometimes) worthy of self-defence—providing their response was reasonable—then believed necessity is to be preferred. This may well be appropriate. It would be harsh for self-defence to fail merely because the defendant, who acted reasonably, perhaps in the heat of the moment, failed to consider the availability of some alternative option. And this view is supported by Parliament, which clarified that if a defendant ‘honestly and instinctively thought [their actions were] necessary’—ie believed necessity—this ‘constitutes strong evidence that only reasonable action was taken’.^75^

Against the Necessity Requirement

I have distinguished several different versions of the necessity requirement. Some are substantially more demanding than others (Table 1).

At one extreme, the necessity requirement would demand that D believe in facts which, if true, would render their precise action strictly necessary to the end of self-defence. At the other extreme, the necessity requirement would demand only that D believe that their action fell within a set of similarly justified actions, the taking of at least one of which was necessary or reasonable to vindicate any right. These are very different propositions.

The most demanding versions of the necessity requirement are untenable as a matter of positive law. The law simply does not require defenders’ actions to be strictly necessary to the end of self-defence, for that is incompatible with the absence of a duty to retreat. Nor does it demand that defenders’ precise degree of force, or their precise action, was strictly necessary, for that would be incompatible with the leeway to take (somewhat) suboptimal actions implicitly acknowledged in the decided cases and Parliament’s suggestion that defenders may act lawfully despite ‘not be[ing] able to weigh to a nicety the exact measure of any [“]necessary[”] action’.^76^ Similarly, Parliament’s suggestion that reasonableness (in part) depends on whether defenders ‘honestly and instinctively thought [their actions were] necessary’ is not easily compatible with necessity requirements which claim that force must be factually necessary (on the facts which D believed to exist).^77^

By contrast, the least demanding versions of the necessity requirement identify uncontroversial legal propositions. It is clearly true that lawful defensive force must be ‘necessary or reasonable’, for the reasonableness requirement—unlike the necessity requirement—is not in dispute. Likewise, the requirement to take an action which falls within a set of similarly justified actions, the taking of at least one of which was necessary, functionally amounts to a (mere) reasonableness requirement. The only problem with these weaker versions of the necessity requirement is that they are misleadingly named. For they do not truly make necessity a requirement, a necessary condition, of self-defence. They merely highlight that defensive force must be reasonable, and that necessity is one factor (among others) going to reasonableness. And this was never in dispute.

Between those extremes, however, lie versions of the necessity requirement which are neither incompatible with the positive law nor reducible to a mere reasonableness factor. The law could require defenders to believe (in facts such) that some force was strictly necessary to the end of vindicating their civil rights. But no authority does so in these exact terms. This version of the necessity requirement is highly complex. And it has minimal legal effect: it prohibits virtually no instance of defensive force not already prohibited by a much simpler standard of reasonableness. These factors should give us pause as to whether the law ought to contain a necessity requirement at all.

Worse still, even the minimal effect of such a necessity requirement would be regrettable. Consider two more cases:

Mugging: While he is out jogging, two men threaten to beat up D unless he hands over his phone. D is a professional athlete and knows that he could simply run away. But D roughly knocks them to the floor as he leaves, to dissuade any pursuit. Arrest: D, a police officer, is lawfully arresting a drunken offender. The offender ineffectually swings his arms towards D. D correctly judges that the offender will continue to miss, and that D could slowly persuade the offender to calm down. However, to hasten the encounter, D instead decides to firmly restrain the offender.

These defendants knew that force was not strictly necessary to defend themselves or vindicate their rights. But should we deny them the benefit of self-defence, merely because they knew that their conduct (or any use of force) was not, strictly speaking, necessary? I think not. They acted eminently reasonably in defence of valid legal interests, in situations which could have escalated to become more serious or onerous. The law should not deny them a defence.^78^ But that would be the effect of even viable versions of the necessity requirement. For these reasons, I propose that the law abandon even these versions of the necessity requirement. It should acknowledge that necessity is but one, unnecessary, factor going to the reasonableness of any defensive force.

What makes the standard of reasonableness more appropriate than the necessity requirement? Broadly speaking, the standard of reasonableness assesses whether something is acceptable on the balance of reasons.^79^ It is a holistic standard, not restricted to a particular domain of reasons.^80^ Clearly, there are reasons not to use unnecessary force.^81^ But, in suitable cases, these reasons might be outweighed by reasons in favour of using such force. In Mugging, the reasons to use force (deterring pursuit) plausibly outweigh the reasons against (it being unnecessary), such that pushing the muggers was reasonable on balance.^82^ Likewise in Arrest: the reasons favouring force (strategically defusing and hastening a tense situation) plausibly outweigh the reasons against (non-necessity). This is why unnecessary force can yet be reasonable. Making necessity a necessary condition of reasonableness, rather than one factor in the equation, overstates the importance of reasons of necessity relative to other reasons.

The law of self-defence has already abandoned a once-central requirement: the reasonable belief requirement. Comparing the two cases only strengthens the case for abandoning the necessity requirement. Abandoning the reasonable belief requirement left no objective standard by which to judge defendants’ beliefs. It made no exception for blameworthy mistakes, such as those formed due to racial bias.^83^ By contrast, abandoning the necessity requirement leaves the rigorously objective standard of reasonableness by which to judge defendants’ actions. Abandoning the necessity requirement is therefore much easier to justify.

Of course, objections are possible. I will consider four. Firstly, even accepting that ‘reasonable’ is more permissive than ‘necessary’, it might be objected that the distinction is minimal; that judges and juries are likely to interpret these terms in much the same way. This certainly seems possible, given that judges have often treated ‘necessary’ and ‘reasonable’ interchangeably. But, for the reasons given above, necessity and reasonableness are distinct concepts. Using different terms in different cases creates a real risk that at least some juries will reach different conclusions depending on which word is used; that they will treat like cases unalike. (All the more so if prosecutors emphasise the literal meaning of ‘necessary’.) As noted above, asking juries to be ‘reasonably sure’ (versus ‘sure’) constitutes a misdirection on the standard of proof. Likewise, if the real evaluative test is reasonableness, as I have argued, then asking whether defenders’ actions were ‘necessary’ should constitute a misdirection on self-defence.

Secondly, abandoning the necessity requirement seemingly puts the law in conflict with the philosophical consensus that necessity is a requirement of morally justified self-defence.^84^ My response, raised above, is that we should not blindly translate philosophical theses into legal rules. Necessity as a reasonableness factor captures the moral significance of necessity in self-defence without suffering the practical downsides of enacting it as a bright-line legal rule. Further, however, the philosophical consensus in support of the necessity requirement, much like the legal consensus, is more apparent than real. For different philosophers offer strikingly different accounts of the necessity requirement’s basis and demands.^85^ Some accounts are especially ill-suited to legal enactment. For instance, one leading account explains the necessity requirement by reference to a deeper duty to rescue.^86^ Whatever the merits of that account, it is a poor fit with English law, which categorically rejects a duty to rescue outside of specific narrow categories. The philosophical consensus in favour of a moral necessity requirement does not justify a legal necessity requirement.

Thirdly, it might be objected that abandoning the necessity requirement will have collateral damage. For, while I have distinguished many different versions of the necessity requirement, I have not canvassed every thesis which might fly under that banner. According to one textbook,

Force cannot be justified to prevent a crime or to defend against an attack unless it was a sufficiently necessary means of doing so. This includes a causal intentional requirement, namely that D uses force in order to defend against attack or prevent crime, which excludes force used in revenge against previous attacks from V.^87^

This ‘causal intentional requirement’ is more simply described as the purposive requirement, according to which defensive force is lawful only if D acted in order to defend themself; only if D’s conduct was defensively motivated.^88^ The problem with this passage is that one can use force in order to defend oneself (the purposive requirement) without such force being (believed to be) necessary to defend oneself (a necessity requirement). They are distinct requirements. And this is important because, while I propose abandoning the necessity requirement, I do not propose abandoning the purposive requirement. If anything, I believe that the purposive requirement ought to be emphasised more strongly than is typically the case.^89^ So I agree that it would be a problem if, by abandoning the necessity requirement, the courts unwittingly abandoned the purposive requirement. But such collateral damage is easily avoided by simply distinguishing the two requirements.

Fourthly, and finally, one might object that judging defensive actions only by an undifferentiated standard of reasonableness is worrisomely vague. One virtue of the necessity requirement, the thought goes, is providing structure to fact finders’ evaluation of the defendant’s conduct. And abandoning the necessity requirement means abandoning that virtue. A simple reply to this objection is that the disvalue of the necessity requirement—denying a defence to those whose unnecessary-but-reasonable conduct warrants a defence—outweighs the benefit of adding structure to the evaluative exercise. But a better reply is that necessity can play an important structural role in evaluating self-defence without being a necessary condition of its lawfulness. As I have emphasised throughout, necessity would and should remain an important factor going to the reasonableness and hence lawfulness of self-defence. Judges can make this clear to juries in their directions: while the jury must ultimately evaluate whether the defendant used reasonable force, they should consider the extent to which the defendant’s actions were necessary as an important factor within that broader enquiry. Sensible legislation could also emphasise this point. It might even create a rebuttable presumption that unnecessary force is unreasonable.^90^ But it would not make necessity a necessary condition of lawful self-defence.

What Follows?

I have proposed abandoning the necessity requirement. But what does this mean in practice? Are courts at liberty to abandon it, notwithstanding the litany of authorities claiming that necessity is a condition of self-defence? Or is parliamentary intervention required?

I believe that courts are entitled to abandon the necessity requirement without parliamentary intervention. For, while many authorities claim that necessity is a necessary condition of lawful self-defence, in practice they usually interpret it in a way that amounts to just one reasonableness factor. Indeed, some authorities go further: they explicitly define the evaluative test for lawful self-defence solely in terms of reasonableness, without a distinct necessity requirement.^91^ Future courts may also draw on the persuasive authority of necessity-free formulations of self-defence from Law Commission proposals,^92^ foreign criminal codes^93^ and the Rome Statute of the International Criminal Court.^94^

But the strongest reason why parliamentary intervention is not required is that Parliament has already ‘clarified’ the law of self-defence in section 76 of the CJIA. This clarification assumed that the operative evaluative question is ‘whether the degree of force used by D … was reasonable in the circumstances’.^95^  Reasonable, not necessary. Section 76’s only action-guiding reference to necessity directs decision makers ‘that evidence of a person’s having only done what the person honestly and instinctively thought was necessary for a legitimate purpose constitutes strong evidence that only reasonable action was taken by that person for that purpose’.^96^

‘Strong evidence’ does not mean ‘necessary condition’. It means ‘an important factor in favour’. Parliament therefore seems to envision (believed) necessity as but one reasonableness factor: not a requirement, a necessary condition, of the reasonableness and lawfulness of self-defence.^97^

Admittedly, section 76 did not purport to replace the common law, only to ‘clarify’ it.^98^ If the common law contained a necessity requirement, as many authorities claim, then section 76 would not clearly override those authorities. But, once we acknowledge the frailty of the common law support for a necessity requirement, section 76 may provide sufficient textual cover for courts explicitly to abandon the language of a necessity requirement altogether.

Nevertheless, explicit statutory language to this effect would be preferable. When Parliament ‘clarified’ the law, it should have said that defensive force is lawful only if it is reasonable, and that necessity is (merely) one important factor bearing on reasonableness, and not a necessary condition of reasonableness.

Conclusion

I have argued that ‘the necessity requirement’ is not a single condition of self-defence, supported by a great weight of authority. Instead, it refers to many different possible requirements. Stronger versions are untenable as a matter of law, while viable versions would make the law worse. The dominant interpretation of the necessity requirement, in practice, is that defensive conduct must be reasonable, and that necessity is one factor bearing on reasonableness. This interpretation puts the law in its best light. But it does not make necessity a requirement, a necessary condition, of self-defence. For this reason, the law should abandon the language of a necessity requirement. It should no longer say, without clarification or caveat, that defensive force must be necessary to be lawful.

CMV Clarkson once pointed out that the courts have sometimes treated the defences of self-defence, duress and necessity as interchangeable, or as sub-species of the same defence.^99^ He described ‘an increasing realisation that clear demarcation between these defences is highly problematic’.^100^ His solution was to go further still, by unifying them into a single defence of ‘necessary action’. Predictably enough, I think it would be a mistake to subsume self-defence into a broader defence of ‘necessary action’. For I think that self-defence can be lawful even if unnecessary. Abandoning the necessity requirement allows us more clearly to delineate these distinct defences. For while self-defence should abandon the necessity requirement, a general defence of necessity could embrace it. As the name implies, if one hopes to rely on a defence of necessity, perhaps one’s actions must be truly necessary: strictly, literally, narrowly necessary.^101^ This is not to defend a general defence of necessity. That would take us too far beyond my topic. But rejecting the necessity requirement in self-defence at least opens up distinct space for a distinct necessity defence.^102^