Lethal Self-Defense, the Quantum of Proof, the “Duty to Retreat,” and the Aggressor Exception

The tragic death of Trayvon Martin has led to more discussion of self-defense and the “duty to retreat.” I wanted to offer a few thoughts about the broad issues.

1. The costs of the legal acceptance of lethal self-defense: To begin with, I think both self-defense enthusiasts and self-defense skeptics have to acknowledge that the law’s accepting lethal self-defense poses very serious risks, not just to criminal attackers but to the law-abiding. (When I say self-defense, by the way, I generally also include defense of others.) First, there is a risk of accidents, where a well-intentioned defender erroneously assumes he is being attacked.

Second, there is a risk of false claims of self-defense being used as a cover for murder. When D kills V in alleged self-defense, there is often only one witness left: D himself. He can say, “I heard V threaten to kill me, saw him reaching for his waistband, and was sure that he was going to shoot me”; and unless the jury concludes beyond a reasonable doubt that D — again, the only witness — is lying, D will be off the hook, even if D was well aware that V was no threat at all and just wanted to kill him. (This could happen either if D was planning to kill V all along, if D and V are involved a dispute and D kills V in a moment of anger, or if D and V are involved in a fight that doesn’t suffice to authorize deadly force, but D uses deadly force in any case.)

Third, there is a risk — especially with guns — of an innocent bystander being hit. And, fourth, there is a risk that even a justifiable killing will lead to a cycle of retaliation, including not just against the defender but also against others in the defender’s family or community, perhaps with innocent bystanders being accidentally killed as a result. Blood feuds, race riots, and gang wars often happen this way.

As I said, this is something that bears on the arguments of both self-defense enthusiasts and self-defense skeptics. The former group has to acknowledge that the legal acceptance of lethal self-defense is a mixed blessing, with real costs (and, again, not just to criminals). The latter group has to acknowledge that our legal system (and, to a large extent, most of the world’s legal systems) protects self-defense despite these costs. The costs of forbidding lethal self-defense, or cutting down on it in dramatic ways (e.g., accepting only the use of lethal force against an immediate threat of death, and not, for instance, of serious bodily injury or rape), are seen — in my view, rightly so, as exceeding the costs of accepting it.

2. What the legal system does about these costs: The question, then, is: Given our acceptance of lethal self-defense, what does the legal system do and what should it do in an attempt to minimize the costs? Let’s briefly discuss several approaches.

a. Bans on possession of deadly weapons: One possible solution is to ban the possession of various kinds of deadly weapons — whether for all people or for some, and whether just in public or both in public and at home. This is indeed the American approach with regard to felons and some other prohibited categories, and the approach of about 10 states as to guns in public places. (The remaining 40 or so states generally allow any law-abiding adult to carry a gun in public places; a few don’t requite a license, and the rest do require a license but give it to pretty much all law-abiding adults who apply.) I don’t want to rehearse the arguments on this in detail here, but they are in many respects similar to some of the cost and benefit arguments related to self-defense: The availability of a gun can make a mistaken killings more likely, or can increase the likelihood of situations where someone (who is otherwise law-abiding enough to comply with a gun carry ban, if such a ban was in place) kills in anger but then is able to cover it up with a false claim of self-defense. The unavailability of a gun can make it harder to resist those attacks that people should be able to resist. Again, this is familiar territory, which I leave to others.

b. Allowing convictions unless the defendant proves self-defense by a preponderance of the evidence: This appears to be the old common-law rule: While a murder defendant’s identity, intentional action, and the like have to be proven by the prosecution beyond a reasonable doubt, the affirmative defense of self-defense has to be proven by the defense by a preponderance of the evidence. If the jury thinks that, say, there’s a 40% chance that the defendant is telling the truth when he says he heard the victim threaten to kill, and reach for his waistband, then the defendant would be convicted of murder (or manslaughter, if that’s what he’s charged with).

This approach was upheld as constitutional in Martin v. Ohio (1987), but is now the law in only one state, Ohio. In all other states, once the defense introduces some evidence of self-defense, the prosecution must rebut that with proof beyond a reasonable doubt. If the jury thinks there’s a 40% chance that the defendant is telling the truth when he says he heard the victim threaten to kill, and reach for his waistband, then the defendant would be acquitted. The theory that “better that [say] 10 guilty men go free than one innocent man be convicted” extends to self-defense claims: “Better that 10 men who committed murder and lied about the need for self-defense go free than one man who killed in justifiable self-defense be convicted.” The worry about fake claims of self-defense has not been seen as serious enough to retain the old common-law/now just Ohio rule.

c. The duty to retreat: Another means of trying to preserve the right to self-defense while minimizing the costs of lethal self-defense has been the so-called “duty to retreat.” This is something of a misnomer — it’s not a true duty, such as the duty to testify when subpoenaed, but rather a rule that

  1. even if D reasonably believes that he is facing an imminent threat of death, serious bodily injury, rape, kidnapping, or, in many states robbery and some other crimes from V,
  2. so long as D can escape this threat with complete safety
  3. by leaving the altercation
  4. except when D is in his own home (or, in some states, in his workplace),
  5. D does not have the right to engage in lethal self-defense because such lethal self-defense is no longer necessary given the availability of a safe retreat.

I hope to say more about the duty to retreat in a separate post, but for now let me observe a few things. First, most American states rejected the duty to retreat even before the recent flurry of new “Stand Your Ground” laws. That’s what the LaFave & Scott 1986 treatise reports, and I have no reason to doubt it. Now that many states like Florida have, in recent years, have rejected the duty to retreat, the “no need to retreat” rule appears to be the supermajority rule in the United States, though there is still a substantial minority of the states that adopt a duty to retreat.

Second, the duty to retreat has always been, at least in principle, a narrow doctrine. There is generally no duty to retreat from an assailant with a gun, since any retreat in such a situation would be quite likely to increase your risk of being shot as you’re retreating (compared to what the risk would be if you shoot first). A typical “duty to retreat” scenario would instead be when there’s a fistfight, and you fear serious bodily injury since even a fistfight can cause such injury, but you can easily leave (for instance, this is right outside your home, or a friend’s home, and you can go inside and close the door, or you’re in your car and can just drive away). The same may apply to someone who is at some distance from you, and is threatening to attack you with a knife (that is not likely to be thrown) or a stick, and you can get away without risk that the person will run up and indeed attack you.

Third, if the worry is that someone will lie about the alleged need for self-defense in an “only two people were there, and now one of them is dead” scenario, the duty to retreat doesn’t really do much about this worry: The killer’s lie could just as easily rebut the applicability of the duty to retreat, for instance if the killer claims that the victim seemed to be reaching for a gun.

Fourth, even when there are some witnesses, the question will often turn on what the facts were, and neither story would yield a different result in duty-to-retreat vs. no-duty-to-retreat states. I don’t want to speak in detail about the Martin/Zimmerman case, because the facts are so unclear. But if the dispute is between “Martin was trying to leave, and Zimmerman started shooting” and “Martin was on top of Zimmerman beating him, and Zimmerman reasonably feared serious bodily injury so he started shooting,” the duty to retreat would be irrelevant under either scenario. It would only be relevant if the facts show that (1) Zimmerman did reasonably fear death or serious bodily injury (or some other forcible felony) — since absent such a fear, lethal self-defense wouldn’t be allowed — but (2) the fear didn’t stem from a worry about Martin possibly pulling out a gun and shooting him (since in such a situation, one can’t retreat with complete safety), and (3) Zimmerman could have avoided the threat of death or serious bodily injury with complete safety by leaving. That’s possible, but there’s only a pretty narrow range of possible circumstances that fits all these criteria.

[UPDATE: I meant to include this paragraph, but neglected to in the first version.] Note that the duty to retreat is generally seen as arising only when the threat of death or serious bodily injury, and therefore the need to use deadly force in self-defense, arises. So even in duty-to-retreat states, a person does not lose his right to self-defense by simply doing something that’s potentially dangerous, such as approaching or following a potentially dangerous person, or going into a dangerous park. Instead, he loses his right to self-defense only if, once he sees the mortal peril, he is able to retreat with complete safety but doesn’t. See, e.g., Matter of Y.K. (N.Y. 1996). (He might also lose his right to self-defense on the different basis that he is the initial aggressor, but that’s a separate doctrine with separate requirements; I discuss it below.)

d. The aggressor exception: The duty to retreat, then, is a much less important factor in self-defense cases than many seem to assume, partly because it’s a minority rule and partly because it applies only a narrow set of cases. There is, however, another long-established means that the law has used to limit lethal self-defense: The principle that someone can’t use lethal self-defense (or, in some situations, any self-defense) if he was the aggressor in the first instance, at least unless he tries to withdraw from the fight first.

Sometimes, this principle can stem from the basic principle that self-defense can only be used against unlawful force. If someone tries to kill me, and I take out a gun and shoot him, he doesn’t then have a legal right to take out his gun and shoot me in supposed “self-defense.” My action would be legal because it was defense against unlawful force, but his wouldn’t be because it is defense against lawful force. (Note that this is something of an oversimplification. For instance, if you reasonably mistakenly believe that you need to inflict serious bodily injury to defend against me, or if you’re a small child or insane or acting under duress, you might not be acting unlawfully in your use of deadly force; but I can still defend myself against you. If you’re really interested in that, check out Model Penal Code § 3.11, which deals with this by defining “unlawful force” as nonconsensual force “the employment of which constitutes an offense or actionable tort or would constitute such offense or tort except for a defense (such as the absence of intent, negligence, or mental capacity; duress; youth; or diplomatic status) not amounting to a privilege to use the force.”

But the aggressor principle also has some independent scope as well. Here is the Florida version:

The [self-defense defense] is not available to a person who:
(1) Is attempting to commit, committing, or escaping after the commission of, a forcible felony; or
(2) Initially provokes the use of force against himself or herself, unless:
(a) Such force is so great that the person reasonably believes that he or she is in imminent danger of death or great bodily harm and that he or she has exhausted every reasonable means to escape such danger other than the use of force which is likely to cause death or great bodily harm to the assailant; or
(b) In good faith, the person withdraws from physical contact with the assailant and indicates clearly to the assailant that he or she desires to withdraw and terminate the use of force, but the assailant continues or resumes the use of force.

So in principle the duty to retreat does exist under Florida law for those people who have “[i]nitially provoke[d] the use of force.” (Recall that the duty to retreat arises only as to lethal self-defense, which is available only when the person reasonably believes he’s in imminent danger of death or great bodily harm, so that clause of (2)(a) does nothing there.) And it’s possible that, depending on the factual circumstances — and on how Florida courts interpret “initially provokes” — Zimmerman’s actions might have been a provocation. (Say, for instance, that Zimmerman started unjustifiably chasing Martin in a way that made Martin worry that he himself was going to be attacked, or started a fistfight with Martin, or perhaps even grossly insulted Martin.)

And this might possibly be a stronger duty to retreat than the norm; for instance, the question is whether the provoker “exhausted every reasonable means” to escape the danger, and not whether he could escape the danger with “complete safety.” On the other hand, courts might read “exhausted every reasonable means” as referring only to means that allow a retreat with “complete safety.” If that’s so, then, as I noted above, the duty to retreat would likely make a difference only in very narrow circumstances.

* * *

In any case, these are some thoughts that I had on the subject. I hope to blog a bit more about the “duty to retreat” soon, and perhaps on other matters, though I hope to stay away from debates about what actually happened in the Martin/Zimmerman situation, given how contradictory the various factual accounts are.