As we saw at the outset of section 3.9, the principle that should guide us in relation to the issue of retreat is the principle of proportionality. One optional framework for this issue (in ‘regular’ cases) is to allow the court to decide whether in the par- ticular case before the court the requirement of proportionality that exists in the law of private defence requires retreat, or not. However, even if the legislator takes this path, I think that in light of the doubts that were raised regarding the relevancy of the possibility of retreat, the legislator should explicitly provide that the possi- bility for a (safe) retreat should be taken into account in evaluating the existence of the requirement of proportionality.
Another option—that will be examined below—is to attempt to establish more detailed rules in the legislation for the instruction of the public in general, and the judges in particular. For this purpose, four typical cases are presented below:
(1) An attack that creates an existential danger for the person attacked (an injury to his life or bodily integrity) where the necessary defensive force for repelling the attack, if the attacked person does not exhaust an existing (safe) possibility to retreat, is deadly force (an injury to the life or bodily integrity of the aggressor);
(2) An attack that creates an existential danger for the person attacked where the defensive force that is necessary (if the possibility for retreat is avoided) is moder- ate (ie, not deadly); (3) An attack that creates a non-existential danger and that demands the use of deadly defensive force; (4) An attack that creates a non- existential danger and that demands the use of moderate defensive force. The fac- tors that should be taken into account with regard to each of these four typical cases are detailed in Table 2(A).
854In fact, among approximately 20 penal codes that were examined there was consideration of retreat in only one of them, and this one also concerns the exceptional case of the actor who by his own fault causes the situation that necessitates defensive action—see s 272 of the Penal Code of Queensland (Australia).
As a preliminary step for the clarification of Table 2(A), a table of the hypothetical factors that should be taken into account if (and only if) a possibility for safe retreat does not exist is presented in Table 2(B) at the following page.
Table 2(B) shall be explained first. This table relates to the situation where there is no route for a safe retreat. In such circumstances, the principle of proportional- ity may necessitate avoidance of the exercise of defensive force in one case alone—
when the defensive force necessary is deadly and the danger to the actor (ie, the TABLE2(A) THEINTERESTS IN THECONFLICTGIVEN THEPOSSIBILITY FORSAFERETREAT
The danger created by the attack
Existential Non-existential
The necessary defensive force The necessary defensive force
(1) Deadly (2) Moderate (3) Deadly (4) Moderate
The A The life The A Injury to The A The life The A Injury freedom U of the freedom U the freedom U of the freedom U to the of action T aggressor of action T aggressor of action T aggressor of action T aggressor
of the O of the O of the O of the O
person N person N person N person N
attacked O attacked O attacked O attacked O
M M M M
The Y The Y The Y The Y
honor honor honor honor
of the of the of the of the
person person person person
attacked attacked attacked attacked
The The The The
aggressor’s aggressor’s aggressor’s aggressor’s
guilt guilt guilt guilt
Slight Slight Slight Slight
danger to danger to danger to danger to
the person the person the person the person
attacked attacked attacked attacked
The social- The social- The social- The social-
legal order legal order legal order legal order
person attacked) is not existential (the case that is marked with the number 3 in Table 2(B)). In all of the four cases that Table 2(B) addresses, the expected injury to the person attacked—if he avoids the use of defensive force (in the left column of each case)—is weighed against the expected injury to the aggressor if the defen- sive force is used (in the right column). Case 2 is very simple, since the expected injury to the aggressor is less severe than the expected injury to the person attacked, and therefore it is clear that the exercise of defensive force does not create any difficulty. In cases 1 and 4 the expected injuries to the aggressor and to the person attacked are identical (or, at least, similar). Here three important additional factors should be taken into account, all of which act in the direction of justifying the exercise of defensive force: (1) The injury to the autonomy of the person attacked; (2) The aggressor’s guilt (that leads to a certain decrease in the value of his interests); (3) The factor of the social-legal order. Consequently, there is no problem in the exercise of defensive force even in these cases. The more problematic case is, as mentioned, case 3. Here the operation of the principle of proportionality finds expression. The severe injury to the aggressor (by deadly TABLE2(B) THECONFLICT OFINTERESTS IN THEABSENCE OF APOSSIBILITY FORSAFERETREAT
The danger created by the attack
Existential Non-existential
The necessary defensive force The necessary defensive force
(1) Deadly (2) Moderate (3) Deadly (4) Moderate
The life of The life The life of Injury to Injury to The life Injury to Injury the person of the the person the the person of the the person to the attacked aggressor attacked aggressor attacked aggressor attacked aggressor
The social- The social- The social- The social-
legal order legal order legal order legal order
The The The The
autonomy autonomy autonomy autonomy
of the of the of the of the
person person person person
attacked attacked attacked attacked
The The The The
aggressor’s aggressor’s aggressor’s aggressor’s
guilt guilt guilt guilt
force) does not meet the required correlation in relation to the slight injury to the person attacked,855and the abstract factors that appear in the balance are also insufficient to justify the exercise of the deadly defensive force. Moreover, the exercise of defensive force that greatly deviates from the principle of proportion- ality may actually injure the social-legal order, and certainly does not serve it (the expression of this in Table 2(B) is in the dashed square in the right column of case 3).856
This leads to Table 2(A), which relates to the situation where a route for safe retreat does exist. Firstly—two clarifications need to be made. The first—a com- mon mistake is to refer to Table 2(B), which was discussed above, as a correct description of the conflict of interests even when there is a route for safe retreat.
There are those, for example, who place the life of the aggressor against the life of the person attacked (in case 1 in the Tables 2(A) and 2(B)), while ignoring the fact that if the attacked person exhausts the route of safe retreat that lies open before him, he will not suffer any physical injury. This mistake, of ignoring the implica- tions of the retreat, leads to a mistaken negation of the duty to retreat.
It is important to emphasise here, that—in contradiction to the mistaken impression that is sometimes created—the necessary consideration of the possi- bility for retreat does not shift the point in time at which the balance of the com- peting interests is carried out until after the time of the actor’s action. The consideration of the danger to the life of the attacked person, and the determina- tion that he is defending his life while ignoring the possibility of a safe retreat that is open before him, are artificial and mistaken. According to my thinking as well, the point of time that should be considered in the weighing of the conflicting interests is the time of the actor’s action (and by no means later), yet if, at this stage, a possibility for a safe retreat already exists, it should not be overlooked.
The second clarification relates to the safety of the route of retreat. As men- tioned, even those who support the imposition of a duty to retreat on the attacked person only suggest imposing this duty when the retreat would be safe. If the retreat endangers the person attacked, there is a consensus that he has no obliga- tion to retreat.857However, there is no real certainty with regard to the future.
855 Here it deserves mentioning that the accepted definition of deadly force—in my opinion as well—is sufficiently broad and is not limited to force whose expected results are death, but also includes, eg, force whose expected results are severe bodily injury. Consequently—and correspond- ingly—the range of ‘the existential dangers’ that the above tables deal with is wide, so that the non- existential dangers that fall within the bounds of the subject of our present discussion (case no 3) do indeed relate to relatively lighter injuries to the attacked person. Thus, eg, there is no doubt that the danger of rape should be classified as an ‘existential danger’ that justifies deadly defensive force, and this is even if the life of the victim is not endangered at all. See with greater detail on the desirable con- tent of the proportionality requirement and on the definition of ‘deadly force’ section 3.8 above.
856 For an additional explanation of the possible influence of the factor of the social-legal order in the direction of negating the use of defensive force see Ch 1.6.3 above.
857 See the references in n 788 above. It should be noted that even danger to another person negates the safety of the retreat in a situation where if the actor does not retreat, he would protect the other per- son (the person attacked), a defence that is also justified within the framework of private defence.
Therefore, it should be clarified that the word ‘safe’ in the expression ‘safe retreat’
does not reflect certainty, but an approximation of safety. Usually, even in a retreat that is envisaged as safe, a certain danger for the person attacked nevertheless exists.858For example, when an aggressor threatens the person attacked with a knife from a distance at which the knife is not yet effective (50 metres, for exam- ple) and the attacked person can run faster than the aggressor and is armed with a gun, the way for retreat that is open before the person attacked appears safe, espe- cially if it is unreasonable to assume that the aggressor has an additional weapon.
However, even in such a situation there may be surprises, such as if the aggressor suddenly brandishes a gun that he had hidden in his sleeve (or if he takes a gun from someone else who is present there) and shoots the retreating attacked person. Thus, the requirement that the way of retreat should be safe should be understood (as a condition for viewing the way of retreat as relevant for the mat- ter under discussion) as a requirement for a relatively safe retreat. At the same time, the danger involved in the retreat (not large, but existing) for the person attacked must also be taken into account in the balance of the competing interests.
As can be understood from the comparison between the two Tables, there is no difference between them with regard to the interests of the aggressor. As to the inter- ests of the person attacked and of society (the interests that operate for the justification of the use of defensive force), here too there is no real difference with regard to the factors of the social-legal order and the aggressor’s guilt. The main dif- ference between the two Tables is therefore that in the place of expected physical injuries to the attacked person, which appear in Table 2(B) (in the absence of a pos- sibility for a safe retreat), there appear in Table 2(A) (given the existence of a possi- bility for a safe retreat) injuries to his secondary interests: the injuries to his freedom of action and to his honour, alongside a certain danger that is expected to him (that is—by the definition of the duty to retreat—danger at a low level of probability).
Cases 2 and 4 in Table 2(A) do not raise much difficulty. It seems that everyone would agree that the factors detailed in the table (on the left side of each case) are usually sufficient to justify the exercise of moderate defensive force against the aggressor. The difficulty arises with regard to cases 1 and 3 in Table 2(A)—in which the actor needs to use deadly defensive force. My opinion is that despite the injuries to the freedom and honour of the person attacked involved in a retreat, despite the (small) physical danger expected to the person attacked, despite the guilt of the aggressor, and despite the injury by the aggressor to the social-legal order—the life of the aggressor—from the vantage point of society at large—is nevertheless to be preferred.859Consequently, deadly defensive force cannot be justified, but rather,
858See the references in n 809 above.
859There is an argument that the retreat is completely contradictory to the public’s sense of justice.
However, Robinson claims that most people will view the retreat as a trivial sacrifice that is clearly out- weighed by the consideration of the life of the aggressor and his bodily integrity—see Robinson (1984), n 37 above, vol 2 at 94.
it should be determined that its exercise deviates from the principle of proportion- ality (a determination whose practical significance is, in effect, the imposition of a duty to retreat). I shall clarify my reasoning for this approach.
Between the two last cases, in which the actor needs deadly defensive force (cases 1 and 3 in Table 2(A)), the dispute focuses on the first of them, and there is an almost complete consensus that deadly defensive force should not be justified in case 3. This approach, which distinguishes between cases 1 and 3—between a deadly attack and an attack that is not deadly—may stem from one of two posi- tions. One possible position that leads to this approach is the treatment of case 1 as if—in the absence of use of deadly defensive force—a real danger to the life of the person attacked exists. As mentioned, I think that such a position (described, in effect, in Table 2(B) above) is incorrect, given that it overlooks the existing possibility of a safe retreat and the (great) influence of this possibility on the com- peting interests.
The other possible position that leads to this approach (that distinguishes between a deadly and non-deadly attack) is more convincing, although I do not support it. According to this position, even though there is no difference between the two cases (the subject of the present discussion) with regard to the interests of the attacked person that are injured by the retreat, there is a real difference between the cases with regard to the abstract interests—a difference that can form the foundation for the negation of the duty to retreat in case 1. According to this position, when the attack is performed with deadly force, the difference (between the two cases) with regard to the social-legal order is particularly striking. The injury by the aggressor to the social-legal order when he attacks with deadly force is so great that the defence of this order justifies the use of deadly defensive force, even though, given the possibility of a safe retreat, the expected physical injury to the attacked person is not great.
However, in my opinion, private defence is first of all privatedefence. Its prin- cipal purpose and justification relate to the interest of the person attacked, and first and foremost to his life and bodily integrity. There is no doubt that the pro- tection of the social-legal order that is achieved by the actor’s use of defensive force adds significant weight for the justification of private defence. Nevertheless, it should be remembered that defence of the social-legal order, prevention of crimes, deterrence of criminals etc. are imposed primarily on the state authorities (such as the police) and not on the individual—even within the framework of private defence.860Accordingly, if there is a route for safe retreat open before the actor, the social-legal order is insufficient to justify the exercise of deadly defensive force.
Such force will greatly deviate from the principle of proportionality and it should
860 A similar opinion, in this spirit—Bein n 539 above, at 11. See also Fletcher, n 72 above, at 306 (although I must express my reservation with regard to the presentation of private defence in that con- text as the ‘necessary evil’).
not be justified. The correct message that should be directed to the actor is not ‘Get up and strike the aggressor with deadly force!’, but ‘If you can avoid the use of deadly force by using a route for safe retreat, then you should avoid killing the aggressor’.
Here it is necessary to return to the essence of private defence as a ‘justification’, ie, the correct and justified action, from a moral point of view as well. The deter- mination that the actor must not use deadly force (a determination that, as men- tioned, implies the imposition of a duty to retreat), does not mean that criminal responsibility will necessarily be imputed to the actor if the actor avoids retreating and exercises deadly force. It seems that this possibility is what troubles the minds of those who oppose the duty to retreat. It should be remembered that alongside the justification of private defence, there is room for other similar defences—espe- cially of the excuse type, as well as for diminished responsibility.861Accordingly, viewing the defensive action by means of deadly force, accompanied by a refusal to retreat, as behaviour that it is possible to understand, must not lead to a decla- ration that this is the correct and justified behaviour. Such a declaration stems directly from the negation of the duty to retreat, and may direct the public to extremely violent behaviour, whose direct results are the loss of human life. This is instead of instructing the general public to act with restraint and self-control in cases in which it is possible to avoid deadly force and still protect the important interests—firstly of the person attacked, but also of the aggressor (his life), and indirectly also of society at large.
My above conclusion, with regard to the limited power of the factor of the social-legal order in justifying private defence, is also reinforced by the back- ground of existing reality in society. It is certainly possible to instruct the attacked person to retreat before using deadly defensive force, and to rely on the solution to the problem that is provided by the police and other authorities after the retreat.
However, if in a certain society and a certain period a situation is created where the attacks are so frequent that there is no possibility to carry on normal social life given the duty to retreat (such as a situation where it is impossible to walk down the road without having to frequently run away from aggressors), then there is no reason to negate the grant of a more elevated status for protection of the social- legal order within the framework of private defence.
As to the suitable legislative arrangement for the issue of retreat, the question is whether it is sufficient to establish explicitly that the possibility of a safe retreat should be taken into account in evaluating the existence of the requirement for proportionality, or if specific rules should also be added, which stem from the above discussion of the different typical cases (which are described in Table 2(A) above). The main candidate for this issue is the rule according to which the possible (safe) retreat negates (according to the principle of proportionality) the
861See the discussion of excessive defensive force, in Ch 5.3 below.