Guest Post: Self-Defence – Weapons, Lawful Commands, Duty to Retreat and Summary

Guest Post: Self-Defence – Weapons, Lawful Commands, Duty to Retreat and Summary

[Ian Henderson is a group captain in the Royal Australian Air Force and is currently posted as the Director Military Law Centre and Deputy-Director Asia-Pacific Centre for Military Law. Bryan Cavanagh is a squadron leader in the Royal Australian Air Force and is currently posted as a legal training officer at the Military Law Centre and Asia-Pacific Centre for Military Law. This note was written in their personal capacities and does not necessarily represent the views of the Australian Government or the Australian Department of Defence. This is the third in a four-part series. The first post can be found here (along with a response here) and the second post here.]

This is the third in a series of four posts that address the relationship between self-defence and LOAC.  In this post we compare how LOAC and the law of self-defence deal with a number of discrete issues like use of prohibited weapons, obedience to lawful commands, and a ‘duty’ to retreat. It also provides a table which summarises the main points in the first three posts.

Can you use a poisoned bullet to protect yourself in self-defence?

The Law of Armed Conflict (LOAC) prohibits the use of certain weapons. Under the Rome Statute and the Australian Commonwealth Criminal Code, it is a war crime to employ poison or poisoned weapons, prohibited gases, or prohibited bullets.In contrast, the law of self-defence does not specifically address the means of response to a threat, but rather merely requires the response to be necessary, reasonable and proportional.

Under the Australian Criminal Code and the Rome Statute, there is no limitation on pleading self-defence only to crimes relating to the use of force. Therefore, the use of a prohibited weapon would be consistent with self-defence analysed under the Australian Criminal Code and the Rome Statute provided that a person’s actions were a necessary, reasonable and proportionate response to the threat.

In some jurisdictions, for example New Zealand, self-defence operates to exclude criminal responsibility for use of force. It is possible in these jurisdictions a combatant could not successfully plead self-defence in relation to weapons offences which are separate and distinct to offences relating to the actual use of force. We did not come to any conclusion on this issue.

Nowhere to run

Under LOAC, not unsurprisingly there is no requirement to retreat from an attack. The position under self-defence varies from jurisdiction to jurisdiction. We found Leverick’s categorisation of the different approaches useful:

a) An absolute retreat rule. The accused must make an attempt to retreat before using force in self-defence regardless of the circumstances.

b) A strong retreat rule. The accused must make an attempt to retreat before using force in self-defence only if an opportunity to do so actually exists.

c) A weak retreat rule. Retreat is not treated as an independent variable, but rather as one factor that is taken into account in deciding whether the accused’s actions were necessary or reasonable.

d) No retreat rule. There is no duty on the accused to take an opportunity to retreat. The victim of an attack has the right to stand their ground and meet force with force.

Both the Australian Criminal Code and the Rome Statute apply the weak retreat rule. The interaction between the weak retreat rule and military duty can raise difficult questions when a member is given a task that either explicitly or impliedly requires them not to retreat from a threatening situation. For example, it appears that an Australian combatant could prima facie plead self-defence when, contrary to orders, he or she retreated or abandoned post in the face of the enemy during an international armed conflict. Conversely, a court might take into account the nature of the military duties when assessing the reasonableness of the member’s response if the member did not retreat.

Command Restrictions

There is nothing in LOAC that prevents a military commander from imposing restrictions on the use of force. Such restrictions are common and are regularly promulgated through Rules of Engagement (ROE). The situation under self-defence is less clear. The issue can be expressed as follows: can a commander issue an enforceable order to a soldier not to shoot in situations where the soldier would otherwise be able to shoot in self-defence? In other words, can a soldier rely on the defence of self-defence when charged with the offence of failing to comply with an order?

This issue will not arise under the Rome Statute as there is no offence of failing to comply with a lawful order. Under the Australian Criminal Code, self-defence is not limited to violent crimes. There appears to be no reason why self-defence could not be successfully pled as a defence to a charge of Disobedience of lawful command, or Failing to comply with a general order if the conduct by the soldier was a necessary and reasonable response to a threat. We would be very interested to hear what others have to say about the situation in other countries.


The table below provides a brief summary of our conclusions so far.

Issue LOAC Self-Defence
In what circumstances does the legal regime apply? When a military member is killing or wounding the enemy. Where a military member is subject to an unlawful use of force and is not otherwise authorised under LOAC to respond to the use of force.
Is use of lethal force permissible? Yes, against enemy combatants, dissident armed forces, other opposition organised armed groups and civilians taking a direct part in hostilities. Yes, but only if use of lethal force is necessary and reasonable to protect life.
Is the level of force in response to an attack measured in relation to the threat posed by the attacker? No (eg, can attack an unarmed combatant, as long as not hors de combat).
Is there are requirement to take precautions in attack, such as taking all feasible precautions to avoid/minimise collateral damage to property? Not explicitly – may be taken into account in assessing the reasonableness of a response.
Is it lawful to cause injury to bystanders when using force? Yes, if the expected injury to bystanders is not excessive to the military advantage anticipated from the use of force. Unclear, but the better view is no where such injury was reasonably foreseeable.
Is it permissible to use weapons prohibited by LOAC? Yes, if such use is necessary and reasonable.
Is there a duty to retreat? Not as such, but whether or not retreat was an option is a factor in considering whether use of force is necessary and reasonable in the circumstances.
Required to obey lawful commands? No, if disobeying is necessary and reasonable.

Tomorrow, in our last post, we will look at the issue of unit self-defence.

Notify of

Is Australian law as non-sensical as that of the ICC re: “proportionate”? e.g., that a soldier defending 5 cannot kill 50 advancing enemy soldiers, much less the 50 plus 50 civilian human shields while attempting to fire at the 50 enemy?


regarding the ICC, I see a future defense claim: despite the primacy of the statute in the statute, the ICC must not deny due process under customary international law, that a “proportionate” limitation is not part of customary international law and its use would deprive my client of a fundamental right to due process under customary international law.

Bryan Cavanagh
Bryan Cavanagh

Jordan, We may not have made ourselves clear. As LOAC would be applied in Australia under our war crimes legislation, you can kill 50 advancing enemy soldiers regardless of whether you are defending zero, five or five hundred. The enemy is a lawful target. However, say there is one enemy soldier – and let’s make him a private. It would be very likely to be disproportionate to kill 50 civilians in the vicinity of the enemy private. If we make it five enemy soldiers, it might well be proportionate to kill one civilian, but still not 50. This is a LOAC issue (excessive collateral damage under art 57 AP I) and has nothing to do with self-defence. Although the soldier is acting defensively but his use of force against the advancing enemy soldiers remains an ‘attack’ for the purposes of LOAC (art 49 AP I “attacks means acts of violence against the adversary, whether in offence or defence). It is probable that the protection of the civilians would add to the military advantage to be gained from the attack. As for human shields, the law is unsettled on whether it makes a legal difference whether they are voluntary or involuntary.… Read more »


I beg your pardon Sirs, but I do not understand the sense of your posts. During an armed conflict (the factual situation in which LOAC applies) a use of force consistent with LOAC raises no issues. Why mentioning self-defence? I think that LOAC rules on the conduct of hostilities and the law on self-defence belong to separate, incommensurable layers.

Ian Henderson
Ian Henderson

Federico, Thank you for your comment and sorry we had not made ourselves clearer. These posts are an attempt to summarise our book chapter, in which we are hopefully clearer on the reasons for the comparison. To give you three reasons: 1. There may be circumstances inside an armed conflict when LOAC is not the governing legal paradigm. For example, military forces might be confronted by an armed criminal gang who are completely unassociated with the belligerents. 2. Some States say they will apply LOAC as a matter of policy in non-armed conflict military operations. We wanted to explore the legal consequences of that type of policy, as the soldiers might operate consistent with their training under LOAC rules, but in fact LOAC would not be the governing legal paradigm. 3. It is not uncommon for Rules of Engagement (ROE) to state that nothing in the ROE limit a soldier’s inherent right of self-defence (note, we are not saying there is such an inherent right, but rather that is what is stated in the ROE). We wanted to explore the consequences of a soldier purporting to act in compliance with self-defence rather than under LOAC. I hope this clears up… Read more »


Bryan: thanks and I am in general agreement. It seems that you and Ian have identified an important area of concern, esp. Ian’s last reply regarding criminal gangs. What seems clear from all of the responses and postings is that there are serious issues regarding potentially 200 different types of domestic law regarding self-defense, whether and how international law (including the laws of war and any human rights law regarding self-defense, e.g., the ECHR) is incorporated into a domestic legal process as treaty law and as customary international law, and whether the laws of war would have primacy domestically if they are incorporated into a domestic legal process. For the average soldier, this appears to remain a significant problem.

Walt Slocombe
Walt Slocombe

I was puzzled by the discussion of an ROE limiting a soldier’s options for self-defense. (A real example: As I understand it, in some situations US soldiers in Iraq were required to use only aimed shots in response to small attacks — defined in the relevant order. Consider the case of a soldier who is subject to what is admitted to be a “small” attack, but believes in good faith that an aimed shot will not be an effective defense against the attack. He fires unaimed shots, kills someone, and is prosecuted for the killing. Assume further that the unaimed shots, while unaimed, were not excessive in terms of proportionality, etc., i.e., if there had been no restrictive ROE, his reaction would have been permissible. The original post seems to suggest that the soldier, having acted in personal self defense, would be able to argue that he was entitled to ignore the ROE. (If I have misunderstood, I apologize, and please ignore the rest of this comment.) But that position seems to be inconsistent with the basic concept of self-defense as a tailored exception to the laws against killing. The right of self defense exists only to the extent the… Read more »