Introduction: the NAP

One of the founding ethical principles of many libertarian philosophies is the NAP: Non-Aggression Principle. This principle is so essential that you'll hear phrases such as «All you need is NAP» (nah-nah-nanananah) almost routinely, even (or especially) in lightweight contexts.

As a philosophical principle, the NAP solves the arguably biggest paradox of that pacifism often espoused by (certain) anarchists: the self-defense paradox. Indeed, interpreted in its strictest sense, pacifism does not allow for self-defence where violence would be required: safe for the few cases where a “soft” approach can be effective (think: deflection), therefore, it puts the holder of pacifist ideas into a position where they may not be able to defend themselves without resorting to that same violence that their pacifist ideas reject.

The NAP solves this problem by clarifying that ethical behavior is to not initiate violence (i.e. don't be the aggressor), but doesn't condemn the use of violence in self-defence (in fact, one could even argue in some extreme sense that the whole point of the NAP is to justify the use of violence in self-defence; but let's not get carried away.)

On a superficial level, the NAP makes perfect sense, and has the appeal of the simplicity and effectiveness that would justify its adoption as founding philosophical principle. In some sense, it's the negative rights version of The Golden Rule: instead of «Do unto others as you would have them do unto you» it's a «Don't do unto others as you would have them not do unto you». And it's a principle that —if followed by all— would (assumingly) lead to a peaceful society, and without the «turn the other cheek» weakness of hardcore pacifism.

In its simplicity, however, the NAP is plagued by an insurmountable issue: the unclear, and subjective, definition of aggression. As we shall see, this poses some significant limitations on the sufficiency of NAP as a founding principle1.

Am I under attack?

We'll start soft before moving to the core issue. As mentioned, the NAP allows for violence in self-defense:

«Somebody tries to kill you, you try to kill'em right back!»

Captain Malcom Reynolds. In Firely, episode 6: Our Mrs. Reynolds

The most obvious, if superficial problem of the NAP is reminiscing of the limitations of the Mutual assured destruction nuclear weapons military doctrine, i.e. the issue of deterrence, escalation and retaliation.

To put it simply, when is it right to exert violence?

On one side of the timing, if your reaction time is too slow, you wouldn't be able to defend yourself, so you want to exert your reactive violence as soon as possible —but not sooner, otherwise it wouldn't even be a reaction, but you'd be the one initiating the violence instead. The NAP doesn't help you in any way in the assessment of the acceptable threat level for escalation.

This introduces a significant degree of subjectiveness in adhering to NAP, leading to situations when “who started it” may be unclear: if I believe I am not posing a threat, but somebody else believes I am, they may feel in their right to resort to violence (i.e. escalate) in defence, and I may see their escalation as initiation of attack and thus feel in my right to actually resort to violence. Even with both of us behaving “ethically” with respect to NAP, violence would ensue.

On the other side of the timing, the issue is for how long after the aggression has begun we can still react and call it a defensive reaction to that aggression, rather than an entirely new episode of violence. Is revenge, «a dish best served cold», justified by the NAP?

More in general, is perpetuating violence justified by the NAP? Let's simplify by saying that throwing the first punch is a violation of the NAP, and counterattacking to defend yourself in a fist fight is not. What if the aggression was a single punch —such that violence would have ended there if there had been no response— and the only reason why the situation degenerated into a fist fight in the first place was the retaliatory punch thrown by the victim of the first punch?

Again, in the absence of clarification, subjective (and situational) judgement is necessary to assess the morality of the action and reaction, and the potential for neverending feuds. Ultimately, even with the best intentions, the NAP does not solve the problem of violence.

And finally, there's the related issue of proportionality: does the NAP justify any kind of violent defensive or retaliatory action in response to any kind of aggression, or is violence in self-defence only morally acceptable if commensurate to the (actual or perceived?) (threat of) aggressive violence? In the former case, it would justify the killing of entire families for the most trivial of offenses, in the latter it again requires additional (potentially subjective) criteria for the ‘measure‘ of the threats and violence.

The deeper issue: what constitutes aggression?

All of the questions raised in the previous section are actually aspects of a much larger issue that affects the NAP, and that's the definition of aggression.

Asking «what constitutes aggression?» may seem like an idle question, since it's easy to come up with positive (and even negative) examples that are likely to be universally (for some sufficiently restrictive meaning of the word) accepted as example of aggression (or not); it's however much harder (if possible at all) to come up with a universally (as above) discriminating answer, i.e. with a definition that would allow “anyone” to determine, given sufficient knowledge of any particular event, whether that specific case was a case of aggression or not.

There are two possible, relatively easy (and frequent, in my experience) answers that I am going to debate now —mostly because rather than actually providing an answer, they simply shift the subjectiveness of the definition to something else.

Aggression is the initiation of violence (?)

It's easy to see why this answer is just a cop-out, since then one would just ask: what is violence?

Again, this is far from being an idle question.

Most if not all the libertarians I've had the opportunity to discuss this with, for example, have an extremely restrictive concept (or at least range of example) of violence, limited essentially to the (ab)use of physical force or, slightly more generally, to the (threat of) physical harm.

Yet there are other forms of violence and threats that do not entail any form of physical force. Psychological abuse, for example, is another well-accepted, and sadly all-to-frequent, form of violence. But does it stop there?

The common (and arguably the defining trait) of (threat of) violence is the abuse of power in an unbalanced relationship: any time a party has power over another party, and (threatens to) abuse(s) such power to inflict (not necessarily physical) harm on the other other party, we can talk about violence.

One cannot exert physical violence on another unless they are stronger (either on their own, or by auxiliary means, e.g. weapons) and can act without particular risk of (effective) retaliation. Similarly, in a relationship, it's the emotionally/psychologically stronger party that can exert emotional or psychological abuse on the other.

Note that who's responsible for the power disparity is irrelevant. For example, I have power over someone who is enamoured with me regardless of whether I actively tried to seduce them, or even if it just happened because of reasons outside of my control. What matters is what I do with the power I have over them, and in particular whether I abuse such power to (intentionally, but we'll get to this point later on) harm them. In fact, it doesn't even matter how much they consciously realize they are being harmed (a frequent trait in abusive relationships is how hard it is for the abused partner to even recognize the abuse patterns); this is for example the rationale behind legislation against the exploitation of the unable or incapacitated.

Violence as abuse of power to inflict harm covers not only the “commonly” accepted forms of violence and power abuse (such as physical and psychological violence), but also forms that are either not formally acknowledged as such, or are so only in limited contexts.

For example, someone living upstream of someone else has power over them, because what they do with the water affects who is living downstream. Polluting the water or diverting its course will harm those living downstream, and may be considered an act of (“environmental”) violence.

Another example is economic abuse, which is only generally acknowledged in the specific context of personal, intimate partnerships, but also exists in the more general sense of harming others economically, taking advantage of one's superior power (e.g. forcing others to accept unfavorable conditions under threat of even worse action, such as driving them out of the market or enacting killer acquisitions).

Whether any of these types of violence is acknowledged as such or not is largely subjective (in fact, one of the leading differences between anarchism and libertarianism is precisely their view on economic violence), which already clouds the “completeness” of the NAP; but even worse, the existence of multiple forms of violence beyond the physical raises further question, such as whether it is acceptable or not to respond to one type of violence with another, and if so in what measure and to what extent (does responding to (threats of) environmental or economical violence, or to emotional abuse, with physical violence violate the NAP or not? if yes, how far can one go with the physically violent response to a non-physical threat?)

Aggression is the (unprovoked) violation of a negative right (?)

A slightly more elaborate cop-out answer to the question “what is aggression?” shifts the lack of clarity to a different concept, that of negative rights.

Note that this is still a cop-out. Even if arguably better defined than the concept of violence (to the point where one may define violence itself in terms of the violation of negative rights), the concept of negative right leaves plenty room for interpretation, since the distinction between a negative and a positive right is much less clear cut than the supporters of the distinction like to claim —if actually possible at all— and so is their violation.

The general distinction between positive and negative rights is framed in terms of action from a second party: a negative right is one that requires someone else to act for it to be violated, while a positive right is one that requires someone else to act for it to be enforced. A dual perspective presents the argument in terms of negative and positive moral duties for the other party: the dual of somebody's negative rights are moral duties for everybody else to not act against those rights (thus, negative duties), whereas the dual of somebody's positive rights are a moral duty for everybody else to act in support of my right (and thus a positive duty).

The classic example is the distinction between a negative and positive right to life: a negative right to life implies that others may not act in a way that endangers your life (and thus the right is violated when someone tries to kill you, for example), while a positive right to life would require others to assist you when your life is in danger (e.g. by providing you with food and water if they come across you in the desert, or by assisting you against a third party that is attempting at your life). In a similar fashion, a negative right to liberty implies that others may not deprive you of your liberty, and a negative right to ownership or private property is the dual of the “do not steal” moral duty (commandment).

Negative rights are generally considered to be first generation rights, and by their negative nature to be “compatible” with each other, in the sense that one will not violate somebody else's negative rights by exerting their own negative rights.

The inadequacy of the distinction between negative and positive rights can be remarked from a variety of different perspective.

We have mentioned that the dual to one's own negative right is a negative moral duty for others (a duty not act in a way that violates that right): but what does a negative right say about one's own action? It is generally argued that protection of one's own negative rights justifies one's own actions, even if this ends up violating someone else's right. Indeed, this is in many ways the “spirit” of the NAP: I should not violate others' negative right to life, but if I happen to kill someone while protecting myself from their attempt at violating my negative right to life, my actions are morally justified. The limit of this interpretation is that it still needs a grading of the rights that goes beyond the mere negative/positive dichotomy.

The simplest example is the question: can I kill someone for stealing (or trying to) from me? Is the negative right to life (violated by killing the thief) less, more or equally important as the right to property (violated by the theft, and in defence of which the killing happens)?

But there's more convoluted examples.

Consider a scenario in which someone is hoarding natural resources beyond what it's needed for their sustenance, and in doing so is depriving others of the essentials needed for survival. Such actions directly threaten the negative right to life of the others (not unlike the environmental or economic violence mentioned before).

Are the others then morally justified in “stealing back” the resources needed for their survival, violating the hoarder's negative right to private property to protect their negative right to life? Are they morally justified in killing the hoarder to protect their negative right to life?

What are the answers to those same questions if the scenario doesn't involve a hoarder, but a natural scarcity, where there simple aren't enough resources for everybody's survival? Are people morally justified in killing each other to protect their negative right to life against others' negative right to life?

But not only is the distinction between negative and positive rights insufficient to clarify the NAP, the distinction itself, as we mentioned, is not well defined. In some sense, there is a semiotic issue: it is possible to rephrase or reframe rights in such a way that they are considered either a positive or a negative right, without changing the practical implications in the respect of the right itself. This implies that “negativity” or “positivity” isn't an intrinsic property of the right, but depends on the frame of reference (and is thus at least partially subjective).

A concise example can be found in the negative right to private property, reframed as a positive right to have your property claim recognized (which it might not be, for example because you are considered a thief, or ownership was transferred to you under unacceptable circumstances, e.g. coercion). More aggressively, your negative rights to life and liberty can be reframed as the positive right to not be considered a threat to the life and liberty of others (since otherwise they would be in their right to suppress your life, as per the NAP).

The ambiguity is even more evident when looking at the dual viewpoint of negative versus positive moral duties. Consider for example the following scenario: I'm walking or standing outside of somebody's private property, marked by a wall, but due to lack of proper upkeep, the wall falls down on me, hurting or killing me. My negative right to life/good health has then been violated not by a direct action, but by inaction: in the specific case, failure from the wall owner to maintain the wall in good condition so that it wouldn't endanger passers-by. Similar examples would be for example the failure to secure one's domestic animal (dog's leash/muzzle) or failure to wear a mask in public while potentially infectious.

More in general, the moral duty of not endangering others (to not violate their negative rights to life and good health) covers both negative aspects (duty to not take actions that would endanger others) and positive aspects (duty to secure others from the danger posed by things one is responsible for).

A counterargument to this remark is that the positive duty is consequence of a previous actions (building or taking ownership/control/responsibility2 for something that can endanger the lives of other if left unchecked), but this simply confirms that the positive/negative view depends on the framing, and presents other aspects of the conflict, such as the right to private property entailing positive personal duties (e.g. to not let one's property endanger others). Is it then morally justifiable to act in violation of someone's negative right to private property, if they shy away from the consequent duties?

(The right to private property, in this sense, is one of the most fascinatingly debatable rights, but this is something that deserve its own discussion.)

Intentionality, awareness, ignorance, inaction.

The last point of subjectiveness in the NAP, whatever definition of aggression is chosen, is the matter of intentionality, here used in the “common language” meaning of acting with deliberate purpose.

We can consider four degrees in the undertaking of a harmful action:

  • intentionally harmful actions are actions undertaken with the express purpose of inflicting harm to another party;
  • a lesser degree is the awareness of the harmful nature of the action, even though harm is not the purpose of the action itself, with the action being undertaken because we value the benefit it gives us as more important than the harm it inflicts others as a side-effect;
  • even lower, we have ignorance of the harmful nature of the action, which leads us to undertake it because we are unaware of its harmful nature;
  • and finally, of course, inaction, as not undertaking the action in the first place.

The NAP clearly covers the highest degree, but arguably it covers the next one as well: even if harm is not the purpose, undertaking the action knowing that it will harm others shows a conscious (thus, intentional) disregard for their safety that can be considered an unprovoked act of violence or a violation of their negative rights to safety.

Or is it? The semantic choice of the word “aggression” carries with itself an implication of intentionality, as debated for example by Kinsella and Tinsley, even though this is not meant to imply that other degrees are exempt from punishment or reprisal, although it would influence the degree to which they would be.

From the perspective of the harmed (i.e. the one that by the NAP would be in their right to exert their right to act violently against the aggression), however, it really makes no difference if there was an intent to do harm, or “just” an intentional disregard for the safety of the others: even if might not considered an aggression in the strict sense, retaliation would be considered morally justified.

The distinction between the second and the third degree (ignorance) is again one that highlights the insufficiency of the NAP. There is wisdom in not escalating violence, which in this case might mean trying to communicate with whoever (presumably unaware of the fact) is acting threateningly or in a harmful manner , making them aware of the threat or harm they pose to you, and hopefully settling the situation without further violence.

Differentiating between the second and third degree however requires a subjective interpretation of the motives or knowledge behind the harmful actions (not unlike the distinction between the first and the second degree, but with a different outcome to the assessment). While it would be wise (with the aim to avoid escalation) to always assume the third degree, there are circumstances where this either requires a massive suspension of disbelief, or a considerable trust in the others, that due to other circumstances we may be unwilling or unable to concede. Even more so, there may be no way or time to contact the actor and inform them about the threat they pose or harm they are causing —especially considering the possibility that, despite our optimism, it might really be a case of disregard for others' safety (at best).

All of this thus requires subjective judgement, and further highlights the inadequacy of the NAP, since the “wise” approach to “assume no malice” is external to the NAP, and its own principle, better known as Hanlon's razor:

Never attribute to malice that which is adequately explained by stupidity.

or at least a variant thereof where ignorance is assumed instead of stupidity, since the latter may preclude any possibility of “enlightenment” and thus the undertaking of corrective actions —and of course, that's assuming we aren't in the previously mentioned situation, where the two parties might have different opinions or perceptions on whether a specific behavior is or is not harmful to others: do we still get to act according to the NAP if even after the confrontation the other party refuses to acknowledge the harm or potential harm (threat) posed by their actions?

And finally, in the fourth degree, we have the lack of action, which should be undoubtedly exempt from the NAP —unless of course, it's the kind of inaction that causes harm to others, when such a view is accepted.

The third party involvement issue

The last obstacle to the universality of the NAP is in reference to the subjective application of it to third party involvement, which can be summarized in the following question: can a third party intervene in a violent conflict while respecting the NAP? In other words, if a party A attacks a second party B, would a third party C intervening in B's defense against A be acceptable within the NAP, or would it be a violation of the NAP for C against A?

The answer to this question is both contextual and subjective.

Contextually, it is generally accepted that the NAP is not being violated if the intervening third party C has a relationship with the offended party B that would “command” their intervention: the typical example is a contract binding C to protect B.

The subjectiveness of the answer is tied both to the subjectiveness of the applicability of the NAP itself, previously discussed, as well as to other ethical considerations.

Consider for example the case where the protection contract that would (contextually) justify the third party intervention has as a trigger the unprovoked violence against the protected party (and is thus essentially just an extension of the NAP to the third party): all of the considerations made about the subjectiveness of the concept of violence that have been discussed so far will come into play. Does the contract trigger based on the protected party's idea of violence (thus being an extension of the NAP “from the perspective of the protected”), or based on the protector's idea of violence (thus being an extension of the NAP “from the perspective of the protector”), or their union, or their intersection?

Of course, this specific example isn't as much an issue for the NAP itself, but just one of the aspects of the contract that need a clear and detailed specification due to the subjectiveness (and thus inadequacy) of the NAP; it also poses an ethical conundrum for the protector, that might find themselves in a situation where they'd need to violate the NAP to satisfy the terms of the contract —but of course, that's up to the prospective protector to decide.

The situation becomes more subjective when one starts to consider that there are circumstances under which one may be compelled to intervene not because of a “legally binding” contract, but due to a specific relationship with one of the parties involved in the conflict that carries a moral (if not legal) obligation to intervene in their defence. This could be for example a blood relationship (parent/offspring/sibling), a close friendship, or being member of one's inner circle or other tightly-knit community. But why stop there? If one's ethos can be summed up as

Homo sum, humani nihil a me alienum puto
I'm human, I think nothing human is alien to me
Cremes, in Plautus' Heautontimorumenos

it would be ethical to intervene in any conflict, even if just to defuse or de-escalate it, but especially in defense of any obviously victimized party, all without —subjectively— violating the NAP.

A core NAP?

A way to escape the extent to which the NAP becomes applicable due to the subjective interpretation of the concepts it relies on is to limit the NAP to its “core”, its “stricter” interpretation: do not initiate physical violence, but accept its usage as a defence to a physical attempt at your (and nobody else's) life, liberty or private property.

This may be considered an “objective” core of the NAP, agreement on which could be reached by all; a “necessary” part, so to say. This being the case, though, is it still actually sufficient? If so, what are the implications of such a stance with respect to the morality of e.g. psychological and emotional manipulation, such as gaslighting or child grooming, which by such a “core NAP” would not be covered?

Conclusions

Assuming you've managed to get this far, it should be clear that none of my remarks invalidate the NAP per se. They do however highlight its insufficiency as an ethical founding principle, and the need for further clarifications, due to the highly subjective underpinnings of its application.

Beware then of anybody claiming that all you need is NAP, they obviously haven't thought this through.


  1. which is to say, even assuming that the NAP is an acceptable founding principle —which is actually debated even in libertarian circles— it is not true that it's sufficient, i.e. it's not true that «all you need is NAP». ↩

  2. and that's before even getting to the moral dilemma about responsibility and inaction, enshrined in the dubiously attributed maxim «for evil men to accomplish their purpose it is only necessary that good men should do nothing» and that culminates in the trolley problem, that will be discussed too. ↩