Violence Is Not Hiṃsā, Non-Violence Is Not Ahiṃsā

I think somewhere I read, maybe Jiddu Krishnamurti said it, that one has to really think about the word non-violence. Not accept it. Think about it. And I do not remember what he said, but something in that direction stayed with me. The more I thought about these words, the less sense the standard usage made. What we call violence, what we call non-violence, seemed to be pointing at something other than what I was trying to understand.

What do we actually mean when we call something violent? Consider what the word is asked to cover: a soldier killing in battle, a surgeon cutting tissue, a parent spanking a child, a protestor breaking a window, a predator catching prey, a government executing a prisoner, a mob lynching a stranger, a nation carpet-bombing civilians. These acts share almost nothing in common except that physical force is involved and someone, somewhere, disapproves. The word does not describe a natural category. It describes disapproval projected onto force.

Non-violence inherits the same incoherence. Sometimes it means pacifism, the refusal to use force under any condition. Sometimes it means passive resistance, absorbing harm without retaliation. Sometimes it means legal protest, avoiding crimes while demonstrating. Sometimes it just means the absence of armed conflict. A Jain sweeping the ground to avoid stepping on insects, a Quaker refusing military service, a civil rights marcher taking police beatings, a diplomat negotiating a treaty: all get called non-violent as if they were doing the same thing. They are not. The word papers over distinctions that matter. This is not sloppy usage. The confusion is structural. The terms were never coherent to begin with.

Modern English inherits its moral vocabulary from Abrahamic theological categories filtered through European legal and political philosophy. This genealogy is not incidental; it shapes what the words can and cannot mean. In Abrahamic traditions, violence is a moral instrument. It can be commanded by God, sanctified by obedience, suspended by decree, or condemned as sin. The moral status of harm depends on authorization, divine or legal, not on the inner state of the agent or the relationship to necessity. You can kill with hatred and be righteous if God commands it. You can kill with compassion and be sinful if unauthorized.

This framework divides violence into legitimate violence (commanded, permitted, or justified by divine or legal authority) and illegitimate violence (unauthorized, sinful, criminal, rebellious). And in secular adaptations: just war versus unjust war, lawful force versus criminal violence. In all cases, violence is treated as a thing-in-the-world that external authority endorses or condemns. The agent’s psychology, whether motivated by hatred or duty, revenge or protection, is secondary. What matters is alignment with command.

Out of this moral universe, violence arrives in English already loaded. The Latin violentia carries connotations of vehemence, excess, transgression against order. To call something violent is not to describe it; it is to condemn it. The word does descriptive work while smuggling in evaluation. This creates an impossible situation. If violence is by definition wrong, then justified force cannot be violence, which is why legal systems speak of lawful force versus criminal violence, and why states claim monopoly on legitimate coercion while condemning others’ violence. Everyone uses the same word while meaning different things. No resolution is possible because the dispute is encoded in the vocabulary.

Non-violence then emerges not as a description but as a counter-stance, a moral opposite to condemned harm. It is a posture, a refusal, a protest. It exists only in relation to the moralized concept it negates. This entire framework externalizes the question. It asks: Is this act authorized? Not: What motive drives it? What cognitive machinery produces it? Does it exceed necessity? Those questions never get asked because the vocabulary has already foreclosed them.

The fundamental error is this: modern discourse treats violence as a property of actions rather than a judgment about motives, contexts, and necessity. But actions do not carry moral weight in themselves. A knife cutting flesh is the same physical event whether performed by a surgeon, a murderer, or a chef. The moral status depends entirely on: Who is acting? Why? Under what constraints? With what relationship to necessity? These questions cannot be answered by looking at the action alone.

When violence is treated as inherently wrong, two equally absurd positions follow. Absolutist pacifism demands the impossible: life without injury, which requires either denying the reality of threats or accepting that one should die rather than defend oneself or others. Cynical realism accepts the intolerable: since force is sometimes necessary, moralizing about it is naïve, which abandons all constraints on how force is used. Both positions fail because both inherit the same flawed premise, that violence is a thing-in-the-world that must be either rejected entirely or accepted pragmatically.

Before going further I need to establish a boundary that will otherwise cause confusion.

A lion hunting a deer is not violent. Not because the deer survives, or because the lion is gentle, but because the lion lacks the cognitive apparatus that makes violence possible. The lion acts from necessity. When its belly is full, it stops. It does not kill for killing’s sake. It does not invent enemies. It does not sacralize aggression. It does not justify slaughter with stories about destiny, purity, or divine command. It does not remember grievances abstractly or transmit hatred across generations.

Animals cause harm. They do not commit violence.

Someone will object: What about surplus killing, foxes slaughtering chickens beyond need? What about chimpanzee coalition killings? Do these not show animal violence? No. Surplus killing in foxes is stimulus misalignment, not symbolic action. The foxes are not making a point. Chimpanzee territorial raids are power dynamics, not moralized destruction. No chimp kills to send a message or purify the group. These behaviors lack narrative reinforcement. There is no conceptual surplus.

Violence, in any meaningful sense, requires at least one of the following: symbolic meaning, imagined future, moral justification, identity reinforcement, or abstract enemy construction. Animals lack these layers. Even the most aggressive primate does not possess the cognitive machinery to transform harm into violence. This matters because it reveals where the problem actually lies. Violence is not an extension of animal aggression intensified by human capacity. It is a new phenomenon that emerges only with higher-order cognition. It is harm plus narrative.

What makes human violence different is not strength or weaponry. It is imagination, language, memory, and abstraction, the same capacities that enable philosophy, art, and science. The same cognitive surplus that lets us plan cities and write poetry also lets us construct enemies that do not exist, nurture grievances across centuries, and justify atrocity with stories about purification.

Human violence emerges when harm becomes detached from survival and reattached to symbols.

Ideology allows injury to be praised. Religion allows killing to be sacred. Nationhood allows murder to be honorable. Race allows extermination to be hygienic. Class allows oppression to be natural. Honor allows execution to be noble. Purity allows genocide to be cleansing. Destiny allows conquest to be inevitable. None of these require biological pressure. A well-fed, secure human can kill for ideas. An entire nation can mobilize for abstractions. Language does not merely describe this harm; it amplifies it, legitimizes it, transmits it across generations. Memory preserves grievances long after threats vanish. Imagination makes hypothetical enemies feel more real than present persons. Abstraction enables killing at scale without perceptual immediacy.

Genocides require speeches, not hunger. Lynching requires stories, not survival pressure.

Animals cannot do these things. They cannot invent the category of enemy that ideology creates. They cannot sustain the multi-generational hatred that language enables. They cannot abstract individual persons into despised types. Violence, in the distinctively human sense, is not a failure of restraint. It is a failure of meaning. The cognitive machinery that should distinguish reality from imagination instead convinces us that symbols are worth killing for.

This is the axis that matters: not whether harm occurs, for harm is inevitable in embodied existence, but whether harm is driven by necessity or by surplus psychological forces. The lion kills to eat. The human kills to prove a point, defend an abstraction, or satisfy an identity wound that exists only in imagination.

How does this happen? The machinery is precise and can be traced. It begins with ahaṅkāra , the I-making faculty, which constructs a self that must be maintained, defended, and expanded. This self is not the body; the body has needs but not grievances. The self is a narrative, a story told to itself about what it is and what it deserves. From ahaṅkāra arises mamakāra , mine-making: my family, my tribe, my nation, my religion, my honor. These extensions of self inherit its defensiveness. An insult to my flag becomes an injury to me, though no hand has touched my body.

Now rāga and dveṣa enter: attachment to what supports the narrative, aversion to what threatens it. I cling to those who confirm my story and recoil from those who challenge it. But rāga and dveṣa are unstable; they require continuous feeding. The self must find threats to defend against, enemies to define itself against. If no real threat exists, one must be invented. The enemy is not discovered; the enemy is constructed by the same cognitive machinery that constructed the self.

This is why violence reproduces itself. The act of harming another to defend the self-narrative does not satisfy; it reinforces the narrative’s reality and importance. Each injury inflicted proves the enemy was real, the threat was genuine, the defense was necessary. The self grows more solid, more defended, more requiring of defense. Grievance accumulates. Memory stores the record of insults, real and imagined. Language transmits the grievance to children who never experienced the original injury but inherit the hatred as identity.

The surplus is not biological; it is cognitive. The body’s needs terminate with satiation. The self’s needs never terminate because the self is a process, not a thing. It must keep producing itself through opposition. Violence is not the failure of this machinery; violence is its proper functioning. The machinery was built to generate enemies, and it performs exactly as designed.

Now I can introduce what I actually wanted to examine: hiṃsā and ahiṃsā .

In Sanskrit, hiṃsā is a bare term. It derives from the root √hiṃs, meaning to injure, to harm, to strike. The noun hiṃsā means simply injury, harm, hurt. Nothing more. The word does not encode cruelty. It does not encode excess. It does not encode immorality. It does not encode intention. It is an action-descriptor, not a moral judgment.

This austerity is deliberate. Sanskrit possesses other roots when it wants heavier semantic weight. √han means to kill, slay, put to death. Vadha means slaughter, murder, execution. Ghāta means destruction, killing with force. These terms carry the terminal, moralized connotations that English packs into violence. But the ethical pivot in Indian thought is hiṃsā , injury, not han or vadha. And hiṃsā itself is morally neutral at the lexical level.

Why does this matter? Because it forces the ethical work to happen elsewhere. Since hiṃsā does not mean immoral harm, the tradition must specify when hiṃsā is permitted or forbidden. Context must be examined: ritual, battle, punishment, self-defense. Agent must be considered: householder, warrior, renunciate. Motive must be interrogated: hatred, duty, protection, ego. The elaborate Dharmaśāstra machinery specifying who may harm whom under what conditions only makes sense if hiṃsā itself is not already moralized. If the word meant immoral violence, all that specification would be redundant.

This vocabulary structure encodes a different understanding of the problem. Sanskrit does not confuse act with intention. It does not collapse harm, motive, and justification into a single overloaded word. It keeps them analytically separate. Hiṃsā names injury as fact. Rāga and dveṣa name the motivational engines, attachment and aversion. Dharma and adharma name the evaluative frame. Ahiṃsā names the discipline. Modern English collapses all of this into violence, making it impossible to think clearly about the actual problem.

Ahiṃsā is simply the privative negation of hiṃsā : a- (not) plus hiṃsā (injury) equals non-injury. Etymologically, nothing more is claimed. But philosophically, ahiṃsā is never presented as an ontological opposite battling hiṃsā in the world. It is not a cosmic force. It is not a political doctrine. It is not a state of affairs. It is a discipline imposed on the agent, not on reality.

And here is the point that gets missed entirely: ahiṃsā does not exist as an independent thing. It exists only because we create hiṃsā that exceeds necessity. Ahiṃsā is not a positive substance opposing violence. It is the name for what remains when surplus harm is no longer generated. If there were no surplus harm, there would be no need for the word. Ahiṃsā is not the opposite of injury; it is the absence of the cognitive distortion that transforms injury into something pathological.

This is why ahiṃsā cannot be ontologized as non-violence. Non-violence pretends to be a thing, a stance, a position one adopts. But ahiṃsā is not a thing. It is what happens when the machinery that produces surplus harm stops running. You cannot practice non-violence the way you practice piano. You can only stop doing what creates violence. The discipline is negative, not positive. It removes causes rather than adding effects.

Patañjali makes this precise in the Yoga Sūtras . Ahiṃsā is listed as the first yama , a universal restraint, a self-binding rule. The practitioner regulates their own impulse to injure. This is individual, internal, disciplinary. It has nothing to do with state policy or collective action.

Yoga Sūtra 2.35 states:

This is not magic. It is recognition that ahiṃsā is primarily an inner condition with outward effects. The practitioner who has dissolved their own motivational roots of harm does not provoke hostility in others. The cause has been removed; the effect ceases. This is not magic. It is recognition that ahiṃsā is primarily an inner condition with outward effects. The practitioner who has dissolved their own motivational roots of harm does not provoke hostility in others. The cause has been removed; the effect ceases.

This is not passivity. It is not refusal to act. It is not allowing oneself to be killed rather than defend. It is the disciplining of surplus motive, so that action arises from necessity and dharma rather than from rāga , dveṣa , and ahaṅkāra .

The Bhagavad Gītā makes this unmistakably clear. Krishna commands Arjuna to fight. There is no pacifist hedging, no hand-wringing about violence. The war is dharmic; Arjuna ’s svadharma as kṣatriya requires it.

What Krishna rejects is not action but egoic appropriation of action. Action performed without rāga or dveṣa does not generate the psychological residue that perpetuates harm. Action performed as identity-assertion does.

The Gītā describes the one established in wisdom:

This is not passivity. This is action without the cognitive distortion that transforms it into violence. The kṣatriya fights, but without hatred. The harm occurs, but without surplus. The hiṃsā is bounded by dharma , not amplified by ego.

The Mahābhārata encodes this in its elaborate rules of war: no striking the unarmed, no killing after sunset, respect for the fallen, dignity for the enemy, restraint toward non-combatants. These rules exist precisely to prevent dharma-yuddha from collapsing into raw violence, to remove the surplus motives of revenge, cruelty, humiliation, and indiscriminate slaughter. War bounded by such rules involves enormous hiṃsā . But it is not violence as I am using the term, because it has stripped away the symbolic, egoic, ideological overlay that transforms necessary harm into pathological excess.

Ahiṃsā paramo dharmaḥ — Non-injury is the highest dharma . But this does not mean never harm. It means: allowing harm to exceed necessity is the highest ethical failure. Do not let ego hijack action. Do not let symbols replace reality. When harm is no longer surplus, when it serves only necessity bounded by dharma , it ceases to reproduce itself.

But what about the hardest case? The dharma-yuddha framework assumes bounded conflict, rules that both sides nominally accept. What happens when one side operates entirely within the surplus-harm paradigm, driven by ideology, identity, sacred hatred, while the other attempts to restrain itself? Does ahiṃsā require submission to annihilation?

No. The framework handles this case, though it is often misread as if it cannot. Dharmic bounded force remains applicable precisely because the boundaries are self-imposed, not negotiated. The kṣatriya who fights without hatred, who does not torture, who does not slaughter non-combatants, who stops when the threat is neutralized, is practicing ahiṃsā even against an enemy who recognizes no limits. The restraint is not for the enemy’s benefit; it is to prevent the defender from becoming what he fights. The asymmetry of ethics is not a weakness; it is the point. If restraint depended on reciprocity, it would not be restraint but bargaining.

The danger in asymmetric conflict is not that ahiṃsā makes defense impossible. It is that the defender, facing an enemy who sacralizes violence, may be tempted to adopt the enemy’s framework: to begin hating, to begin seeing the enemy as subhuman, to begin enjoying the killing, to begin believing that this war is holy. That temptation is the real threat. The moment the defender crosses into surplus harm, the enemy has won regardless of who holds the battlefield. Two parties operating in the violence-paradigm will reproduce violence indefinitely. One party maintaining ahiṃsā at least preserves the possibility of termination.

To translate ahiṃsā as non-violence is not approximation. It is structural falsification.

Non-violence presupposes that violence is a moral entity, something that can be embraced or rejected as a matter of stance. It frames ethics as opposition rather than discipline. It exists only in relation to what it negates. Ahiṃsā does not work this way. It is not defined against anything. It is the discipline that dissolves the cognitive roots of surplus harm. It applies to individuals, not nations or movements. It cannot be transferred to collectives because collectives do not have minds capable of rāga , dveṣa , or ahaṅkāra .

The translation error imports Abrahamic moral grammar into Dharmic conceptual space. Once this happens, ahiṃsā is mistaken for pacifism, restraint is mistaken for inaction, and dharma -guided force is mistaken for ethical failure. The entire framework shifts from internal discipline to external posture.

Gandhi’s passive resistance exemplifies this collapse. I am not critiquing Gandhi’s actions or purposes. He faced a specific political situation, an empire that could not be defeated militarily, and he needed a method that could mobilize masses without provoking massacres. His reasons were sound; his strategic intelligence was considerable; his personal courage is not in question. What I am addressing is what happened to the word. By framing ahiṃsā as passive resistance for global communication, Gandhi created a new ontological category that took on life independent of his intentions. Non-violence became a thing, a stance, a transferable political posture. This new entity, once born, could not be controlled. It was adopted by movements Gandhi never endorsed, applied to contexts he never imagined, and eventually fed back into Indian discourse as if it had always meant what the English word suggested. The word ahiṃsā was stolen, replaced by an impostor wearing its clothes. Gandhi may not have intended this theft, but the theft occurred, and we live with its consequences. Ahiṃsā in classical thought is not a method for changing others. It is a discipline for preventing self-corruption through surplus harm. The two have almost nothing in common.

Violence is not hiṃsā . They are not even adjacent concepts. Non-violence is not ahiṃsā . They operate in different dimensions. The confusion is not linguistic carelessness. It is civilizational misunderstanding.

I need to be direct here, because the structural problem is real and consequential.

Abrahamic theology makes violence a moral category, something that can be sanctified, commanded, or redeemed by divine authorization. This is not an abuse of the framework; it is its logical consequence.

When violence can be morally good if commanded by God, three things follow. First, externalization of responsibility: the agent no longer owns the harm; God or law does. The killer becomes an instrument rather than an author. Inner scrutiny becomes irrelevant or even sinful, for who are you to question divine command? Second, immunity to introspection: if harm is authorized, the motivational roots need not be examined. One can kill with hatred, revenge, or sadistic pleasure, and as long as the killing itself is permitted, the psychological surplus goes unaddressed. Third, scalability without termination: violence can be multiplied indefinitely as long as authorization persists. There is no intrinsic stopping condition. Crusades can last centuries. Inquisitions can burn thousands. Genocides can proceed until the last target is gone. Nothing in the framework itself calls a halt.

This is cognitively destructive: it trains the mind to bypass self-examination. It is socially destructive: it licenses harm without proportionality or natural limit. It is ethically destructive: it replaces restraint with obedience.

I am not making a sociological claim about whether Abrahamic followers are more violent. I am making a structural claim: a framework that moralizes violence itself is inherently unstable and dangerous, regardless of how many people follow it or how gently individual practitioners behave. A gentle pope does not fix Christian theology. A compassionate imam does not fix Islamic metaphysics. A benevolent rabbi does not fix the structural problem. Individuals do not launder frameworks. At best, they locally buffer consequences.

The Dharmic framework avoids this error categorically. Hiṃsā is descriptive, not moral. Ahiṃsā is disciplinary, not cosmic. Dharma regulates motive, context, and proportion. Vedānta dissolves the cognitive machinery that generates surplus harm. At no point is injury itself declared morally redemptive. This difference is not minor. A framework can be misused; a framework that licenses misuse by design is a different beast.

There is a final level to this analysis, and it is the most radical.

At the pāramārthika level, from the standpoint of Brahman -realization, the distinction between self and other collapses entirely. When avidyā dissolves, the cognitive machinery that constructs enemies, nurtures grievances, and defends identity simply stops running. There is no one to hate and no one to defend.

This requires argument, not assertion. Recall the machinery traced earlier: ahaṅkāra constructs the self-narrative, mamakāra extends it to mine, rāga and dveṣa fuel its defense, and the enemy is fabricated to give the self something to define against. Every component of this machinery depends on a prior error: the belief that the self is a real, bounded entity requiring protection. This belief is avidyā , not mere ignorance but active misperception, taking the constructed for the given, the narrative for the fact.

What happens when this misperception is seen through? The self does not die; it is recognized as never having been what it claimed. The narrative continues to run, as narratives do, but it is no longer believed. And a narrative that is not believed cannot generate rāga and dveṣa , cannot produce enemies that feel real, cannot motivate surplus harm. You can still think “this is my country” while knowing the thought is conventional, not ontological. The thought without belief is inert; it moves nothing.

This is why jñāna (knowledge) dissolves violence at the root rather than suppressing it by effort. Suppression addresses symptoms; knowledge addresses cause. The one who sees through avidyā does not struggle to be non-violent. The struggle presupposes a self that must restrain itself, and that self is precisely what has been seen through. What remains is action without the surplus, response without the narrative overlay, force when necessary without the hatred that makes it self-perpetuating.

This is not pacifism achieved through willpower. It is the spontaneous cessation of the motivational roots of harm. Ahiṃsā at this level is not a discipline; it is a description of what remains when delusion ends.

At the vyāvahārika level, in the lived world of apparent multiplicity, ahiṃsā functions as sādhanā , a practice that attenuates the roots of surplus harm in advance of full realization. The yama framework is exactly this: regulatory constraint for embodied life.

This two-level structure explains something that otherwise seems paradoxical. Ontologically, there is no independent entity called non-violence opposing violence. The categories dissolve. Ahiṃsā does not exist as a thing; it is only the absence of the surplus that we create. But psychologically and ethically, ahiṃsā names a practice that removes the causes of pathological harm. And spiritually, in jñāna , it becomes spontaneous, not because one chooses pacifism, but because the inner machinery that generates surplus harm no longer operates.

From this vantage, the modern discourse looks hopelessly confused. It treats violence and non-violence as moral opposites locked in eternal battle. But there is no battle. There is only the presence or absence of ignorance-driven motivation. Clear that up, and the vocabulary problem solves itself.

Preventing violence cannot be achieved by laws, treaties, or slogans. These may limit damage, but they do not address causes. Nor can it be achieved by moralizing action without examining cognition. The absolutist who forbids all harm and the realist who permits any harm are both trapped in the same confusion, both treat violence as a thing-in-the-world to be rejected or accepted, rather than a cognitive phenomenon to be understood.

The task is subtler: cultivate clarity about where harm becomes surplus, where necessity ends, and where imagination begins to lie to itself. Education must shift from moral posturing to cognitive literacy. Societies must learn to distinguish survival-based force from ideology-driven harm. Language must be treated with suspicion when it sacralizes injury. Symbols must be interrogated before they are defended. Identity must be held lightly, not weaponized. None of this requires denying conflict or force. It requires denying their mythologization.

Violence, as I have been using the term, is harm that exceeds necessity and is driven by surplus motive: ideology, identity, revenge, abstraction, symbol. This is what humans uniquely produce. This is what animals never do. This is what Dharmic thought diagnoses and Vedāntic realization dissolves.

The English words violence and non-violence cannot capture this. They were built to externalize judgment, not to interrogate cognition. They moralize action without examining mind. Hiṃsā and ahiṃsā operate differently. Hiṃsā is neutral, injury as fact. Ahiṃsā is discipline, the cessation of surplus harm. Neither is an ontological substance. Ahiṃsā exists only because we create the problem it names the solution to. If we did not generate surplus harm, there would be no ahiṃsā to practice. The word would be meaningless.

This is not a cultural preference or a religious claim. It is a structural observation about vocabulary and its consequences.

Truth is not democratic. Prevalence does not correct category errors. And a framework that moralizes violence itself, that treats harm as a moral instrument rather than a cognitive phenomenon, will continue to generate the pathologies it cannot name.