The previous essay established the category error. Criminal law claims authority over the kṣetrajña — the knowing, conscious, morally responsible person — while possessing instruments only for the kṣetra — the body, the social identity, the behavioural surface. It punishes the field while imagining it is settling accounts with the knower of the field. This error is not a policy design flaw. It is constitutive of retributive justification.
That diagnosis, however complete, does not yet answer the question it forces. If retribution is philosophically incoherent and rehabilitation as currently practised is structurally inadequate, what is punishment? Not what punishment claims to be — but what it coherently can be, given an accurate understanding of the person it addresses and the causal structure it is trying to engage?
Classical Indian jurisprudence asked this question twenty-five centuries ago. It did not have the benefit of the Advaitic metaphysics in its full elaborated form, but it possessed the same philosophical culture from which that metaphysics grew. Kauṭilya’s Arthaśāstra, the Manusmṛti’s concept of pratyāścitta, and the Bhagavad Gītā’s dispositional analysis of the wrongdoer together constitute a theory of punishment that Western jurisprudence has never developed and modern Indian jurisprudence has forgotten.
This essay reconstructs that theory, tests it against the framework the series has built, and proposes from it a three-level account of what punishment could coherently be.
I. The Question Law Has Not Asked#
Modern jurisprudence has a theory of when punishment is justified (desert, deterrence, rehabilitation), and theories of how much punishment is proportionate (cardinal and ordinal proportionality, guideline sentencing). What it does not have is a theory of what punishment is — of the kind of practice it is, at the level of the human being it addresses, in relation to the causes of the conduct it responds to.
This is the absence that the classical Indian framework fills. To understand why the absence matters, consider the difference between the following two questions: Does this person deserve to suffer this much? and Does this practice actually engage the level of causation at which the harmful conduct originated? The first is the question modern criminal jurisprudence asks. The second is the question the classical Indian framework asks, and that the Advaitic analysis the series has been building makes mandatory.
The first question, taken alone, produces a system that can be internally consistent — proportionality metrics can be calibrated, desert criteria can be specified — while remaining systematically indifferent to whether it actually reduces harm. The second question forces engagement with the empirical question that determines whether any practice is worth undertaking at all: does it work at the level where it needs to work?
II. Daṇḍa in the Arthaśāstra — Punishment as Statecraft#
The Framework#
Kauṭilya’s Arthaśāstra is frequently misread as a manual of cynical power politics. It is, more precisely, a systematic account of the conditions under which a polity can sustain itself and its members can pursue the four aims of life (puruṣārthas). Within this account, punishment — daṇḍa — has a specific structural function.
Daṇḍa is one of the four instruments of statecraft: sāma (conciliation), dāna (inducement), daṇḍa (punishment), and bheda (division and disruption of hostile combinations). These four are applied in sequence, with daṇḍa reserved for situations in which the first three have failed or are inappropriate. Daṇḍa is coercive force — the rod — and the Arthaśāstra is explicit about both its necessity and its dangers.
The necessity follows from what the text calls matsyanyāya — the law of the fish, in which the strong eat the weak. Without the credible threat of punishment, coordinated social life is impossible: the powerful will always take from the less powerful, contracts will not be honoured, property will not be secure, and the conditions for economic and social flourishing will be systematically destroyed. Daṇḍa is what prevents this. It is not a response to moral desert. It is a condition of social possibility.
The danger follows from daṇḍa’s corrupting potential. Kauṭilya warns at length against punishment administered from the king’s egoic passions — anger, greed, vanity. He specifically identifies apradāna (punishment without proper basis) and atidāna (excessive punishment) as productive of hatred and potentially more destructive of social order than the original offence. The measure of daṇḍa is always its effectiveness in maintaining the conditions for flourishing, never a retributive calculation.
The Alignment With the Advaitic Analysis#
This is the first and perhaps most important point of alignment between classical Indian jurisprudence and the Advaitic framework the series has developed. Kauṭilya’s daṇḍa operates entirely at the vyāvahārika level. It manages the conditions of the transactional order — the empirical world of bodies, property, social relationships, and institutional arrangements. It makes no claim to paramārthika moral settlement. It does not pretend to balance cosmic scales or to reach the kṣetrajña. It is, explicitly, field management.
This is far more honest than modern retributivism, which claims paramārthika authority (the offender deserves to suffer — this is a claim about ultimate moral reality) while possessing only vyāvahārika instruments (prison cells, fines, constraints on bodily freedom). Kauṭilya’s theory is honest about what it is doing and what it cannot do. This honesty is itself a philosophical achievement.
The Limit#
But daṇḍa, as Kauṭilya theorises it, has a clear limit. It operates on the sthūla śarīra. It contains harm. It deters through the threat of physical and social consequences. It incapacitates through restriction. It does none of the work that the Advaitic analysis identifies as actually necessary — none of the work at the sūkṣma level where the saṃskāras that generated the criminal conduct reside. Kauṭilya’s theory is honest about what punishment is but silent about what the problem actually requires.
The Arthaśāstra knows how to build a wall. It has no theory of what happens inside.
III. Pratyāścitta in the Dharmaśāstra — Punishment as Expiation#
The Structure#
The Manusmṛti’s concept of pratyāścitta — expiation, literally from prāyaś (largely, predominantly) and citta (directed toward, going back to the beginning) — is structurally different from daṇḍa in every important dimension. The differences are not superficial. They represent a fundamentally distinct understanding of what a response to wrongdoing is supposed to accomplish.
Daṇḍa is externally imposed. Pratyāścitta is self-assumed. The wrongdoer who undertakes pratyāścitta does so not under compulsion but in recognition of a structural fault in themselves that the wrong act expressed. The court does not sentence the person to pratyāścitta; the person, recognising what they have done and what its root was, undertakes the expiation because the root needs to be addressed.
Daṇḍa is measured against the harm done — its proportionality criterion is external. Pratyāścitta is measured against the fault in the doer — its adequacy criterion is internal. This is the critical distinction. It is not asking how much suffering is proportionate to this harm? It is asking what practice is adequate to address the structural misalignment in this person that produced this harm?
The Taxonomy of Wrongs and Their Expiations#
Manusmṛti distinguishes categories of wrongdoing and specifies forms of pratyāścitta appropriate to each. The forms include: fasting and austerities (prāyaścitta through tapas), charitable giving and service (dāna), study and recitation (svādhyāya), pilgrimage (tīrtha-yātrā), and in certain cases the payment of compensation to those harmed (pratipatti). The selection among these is not mechanical — it is guided by consideration of the nature of the wrong, the disposition from which it arose, and the kind of practice most likely to address that disposition.
The austerities function not as suffering-for-its-own-sake but as practices that loosen the attachment structures — the saṃskāras — that gave rise to the harmful action. Fasting reduces the grip of sensory craving. Service cultivates the capacity to consider others’ wellbeing above one’s own convenience. Study introduces viveka — discriminative intelligence — as a counter-force to the avidyā that generated the wrong. Pilgrimage detaches the person from the habitual environmental context in which the saṃskāric patterns were reinforced.
This is not theological decoration. It is a systematic attempt to address the sūkṣma śarīra — to reach the layer of conditioning that daṇḍa leaves entirely untouched.
The Critical Questions Pratyāścitta Poses to Modern Jurisprudence#
The Manusmṛti acknowledges that pratyāścitta is most effective for wrongs committed without full deliberate intention — for wrongs that arose from ignorance, weakness, or reactive emotional states. For deliberate, repeated, and severe wrongdoing, its efficacy is described as limited to making the person once again fit for social interaction, with the deeper karmic consequences unresolved by any external expiation.
This is not a weakness of the classical theory. It is its honesty. Pratyāścitta acknowledges its own limit at the kāraṇa level. The deepest dispositional roots of habitual, deliberate wrongdoing are not amenable to resolution by any scheduled practice. This is precisely the acknowledgement that the Advaitic framework demands: the aspiration toward kāraṇa-level transformation is not a legal project.
The question pratyāścitta poses to modern secular jurisprudence is whether its functional logic can be preserved without its theological scaffolding. The answer the series has been building toward is yes — but only if the system is honest about what it is doing. An institution that reframes incarceration as a practice environment for genuine dispositional repair, that measures its success by actual shifts in saṃskāric patterns rather than by the infliction of proportionate suffering, and that creates conditions in which the wrongdoer actively participates in the work of their own structural repair — that institution is secularised pratyāścitta. Restorative justice practices, properly designed, begin to approach it. None of the major carceral systems in the world, including India’s, currently approaches it at scale.
IV. BG Chapter 16 — The Dispositional Taxonomy of the Wrongdoer#
The Two Dispositions#
The Bhagavad Gītā’s sixteenth chapter opens with a passage that has no parallel in any Western jurisprudential tradition:
abhayaṃ sattvasaṃśuddhir jñānayogavyavasthitiḥ dānaṃ damaś ca yajñaś ca svādhyāyas tapa ārjavam
Bhagavad Gītā 16.1
‘Fearlessness, purity of being, steadfastness in knowledge and yoga, generosity, self-restraint, sacrifice, study of scripture, austerity, uprightness…’
ahiṃsā satyam akrodhas tyāgaḥ śāntir apaiśunam dayā bhūteṣv aloluptvaṃ mārdavaṃ hrīr acāpalam
Bhagavad Gītā 16.2
‘Non-violence, truthfulness, absence of anger, renunciation, tranquillity, absence of slander, compassion toward beings, absence of greed, gentleness, modesty, absence of fickleness…’
These constitute daivī sampat — the divine disposition. Against it, the Gītā places āsurī sampat:
dambho darpo’bhimānaś ca krodhaḥ pāruṣyam eva ca ajñānaṃ cābhijātasya pārtha sampadam āsurīm
Bhagavad Gītā 16.4
‘Ostentation, arrogance, self-conceit, anger, harshness, and ignorance — these are the portion of one born to the demonic inheritance.’
The Causal Analysis, Not the Moral Condemnation#
The crucial point about these dispositions is what Kṛṣṇa does not say about them. He does not say that the person with āsurī sampat is morally depraved and therefore deserves to suffer. He identifies the disposition, traces its causal origin, and analyses its consequences. The causal analysis comes earlier in the text:
kāma eṣa krodha eṣa rajoguṇasamudbhavaḥ mahāśano mahāpāpmā viddhy enam iha vairiṇam
Bhagavad Gītā 3.37
‘It is desire, it is anger, born of the quality of passion (rajas) — all-devouring and greatly sinful — know this to be the enemy here.’
The āsurī disposition is produced by the predominance of rajas and tamas in the individual’s guṇa-composition. The guṇa-composition is not a choice. It is the downstream expression of prior conditioning — saṃskāras accumulated through past actions and shaped by the environments in which those actions occurred. Desire driven by rajas, delusion deepened by tamas, the ego’s desperate grasping in the absence of genuine sāttvic stability — these are causal mechanisms, not moral failures freely chosen.
This means that what criminal law calls the ‘criminal mind’ — the mens rea, the intention, the deliberation — is, on the Gītā’s analysis, the expression of a guṇa-composition that the individual did not freely choose and cannot directly alter through an act of will. The ‘choice’ to commit the wrong arose from a configuration of the subtle body that was already in place before the moment of choice.
The jurisprudential consequence is not that wrongdoers bear no responsibility. The consequence is that responsibility must be reframed. What the person is responsible for, at the sūkṣma level, is their ongoing relationship with their own conditioning — whether they actively participate in practices that move the guṇa-composition toward sattva or whether they remain passive in conditions that reinforce rajas and tamas. This is a more demanding form of responsibility, not a lighter one.
The Dispositional Taxonomy’s Institutional Implications#
Chapter 16’s dispositional analysis implies a set of questions that no modern criminal justice institution asks in any systematic way.
The first is a diagnostic question: what guṇa-composition is expressed in this conduct, and what in the person’s history and environment produced that composition? The legal system asks a narrower and less useful question: did this person form the requisite mental state at the moment of the act? The Gītā’s question reaches back through the moment of the act to the dispositional substrate that made it possible.
The second is a design question: does the response to this conduct move the person’s guṇa-composition toward sattva, or does it reinforce rajas and tamas? This question is almost entirely absent from modern penological theory, which evaluates practices by their capacity to impose proportionate suffering (retributivism) or to reduce reoffending through incentive structures (deterrence/rehabilitation). Neither criterion asks whether the practice actually reshapes the dispositional substrate. The Gītā’s framework makes this the primary criterion.
The third is an institutional self-assessment question: what is the guṇa-composition of this institution? What is the guṇa-profile of a maximum-security prison? The environment is characterised by constant threat, suppression of vulnerability, status hierarchies enforced by violence, idleness alternating with regimented activity, sensory deprivation and overstimulation in alternating cycles, and the systematic humiliation of anyone who shows the qualities associated with daivī sampat — gentleness, openness, trust, intellectual curiosity. The institution is rajasic-tamasic in its essential character. It produces, reliably and predictably, the guṇa-composition most conducive to further criminal conduct. It is a saṃskāra factory for āsurī sampat.
The failure of rehabilitation in carceral settings is not mysterious. It is the inevitable result of administering sāttvic interventions — mindfulness programmes, education, cognitive therapy — within a tamasic-rajasic environment that continuously and overwhelmingly reinforces the conditioning patterns the interventions attempt to dissolve. The environment always wins. This is what the guṇa analysis predicts, and this is what the recidivism data confirms.
V. Niṣkāma Karma as Institutional Design Principle#
The Gītā’s most practically applicable teaching for institutional design is not found in Chapter 13 or 16 but in Chapter 3: the doctrine of niṣkāma karma.
niyataṃ kuru karma tvaṃ karma jyāyo hy akarmaṇaḥ śarīrayātrāpi ca te na prasiddhyed akarmaṇaḥ
Bhagavad Gītā 3.8
‘Perform the action that is required; action is superior to inaction. Even the maintenance of your body would not be accomplished by inaction.’
And the defining characterisation:
māphaleṣu kadācana mā karmaphalahetur bhūr mā te saṅgo’stv akarmaṇi
Bhagavad Gītā 2.47 (partial)
‘Never to the fruits alone; let not your motive be the fruits of action; nor let there be any attachment to inaction.’
Niṣkāma karma — action without ego-attachment to outcome, performed from duty and offering rather than from the ego’s investment in receiving a particular result — is the Gītā’s prescription for action that does not generate new binding karma. It is, simultaneously, a description of action from which the āsurī qualities of pride, greed, and reactive anger have been progressively removed. And it is a practical account of what institutionally sustained practice actually looks like when it succeeds in shifting guṇa-composition.
An institution designed around niṣkāma karma as a principle would look specific and unfamiliar. Structured work performed not for the institutional benefit of the prison but for the genuine benefit of others — service to the community, contribution to a collective good, care for those more vulnerable. Accountability practices that focus not on extracting confession but on developing the capacity to clearly see the harm done and the conditions that produced it. Relational engagement — not the managed distance of the professionally bounded relationship between therapist and client but the actual fabric of community, with its genuine demands and genuine supports. Contemplative practice not as stress management but as the systematic cultivation of the capacity for discriminative intelligence. These constitute the environment in which niṣkāma karma can gradually become the operative mode — in which action arises less from the ego’s fearful grasping and more from the witness’s clear seeing.
This is not utopia. It is a description of what happens in the small number of practice-based environments that have been allowed to develop within or alongside carceral systems, and whose outcomes consistently outperform those of the standard model. The obstacles to scaling are political and philosophical, not empirical.
VI. Toward a Theory of Dharmic Punishment — Three Levels#
From the conjunction of Kauṭilya’s realism, the Manusmṛti’s pratyāścitta, and the Gītā’s dispositional analysis, a coherent three-level theory of punishment emerges. It is not soft. It is not forgiving in the sense of refusing to engage with wrongdoing. It is harder on the causes of crime than any retributive system, because it refuses to close the account by punishing the surface.
Level One — Sthūla: Containment of Harm#
At the gross, transactional level, the state has a legitimate obligation to prevent harm. The body that commits violence can, and in cases of serious and persistent threat must, be constrained. Incapacitation of the gross body — incarceration, supervised release, restrictions on movement and access — is legitimate at the vyāvahārika level. It needs no retributive justification. It requires only the honest empirical claim that this constraint prevents harm that would otherwise occur, and that the quantum of constraint is the minimum necessary to serve that prevention.
This level answers to Kauṭilya’s daṇḍa: pragmatic, instrumental, operating on the sthūla śarīra without pretending to address anything deeper. The crucial discipline at this level is the renunciation of the retributive overlay — the refusal to claim that the constraint constitutes moral closure, that the person deserves to suffer, that the infliction of suffering settles any account beyond the immediate practical one of preventing further harm.
The conditions of the constraint are not indifferent at this level. Even purely at the sthūla level, constraint that is administered with deliberate humiliation, chronic deprivation, or the systematic elimination of the conditions for basic physical dignity is constraint that exceeds the instrumental justification and enters the domain of cruelty. Kauṭilya’s own analysis supports this: atidāna produces hatred and further social destabilisation. The measure of daṇḍa is always its functional effectiveness in maintaining conditions of social possibility, not the maximisation of the offender’s suffering.
Level Two — Sūkṣma: Conditions for Genuine Transformation#
At the subtle level, the guiding question is not what does this person deserve? but what conditions are required for the saṃskāric restructuring that addresses the actual causal root of the harmful conduct?
This level answers to pratyāścitta’s functional logic, secularised: the practices of acknowledgment, structural repair, and gradual dispositional transformation that the Manusmṛti describes in the context of voluntary expiation are translatable into institutional design even in the absence of their theological framing. The translation requires four elements that are consistently absent from standard carceral practice:
The first is genuine acknowledgment. Not the formal admission of guilt that the legal process extracts but the experiential recognition — felt at the level of the subtle body, not merely stated by the ego — of what was done, what it caused, and what in oneself made it possible. This is not achievable through coercion. It can be enabled by sustained practice and supported by genuine relational accountability, but it cannot be mandated.
The second is structured difficulty that is oriented toward transformation rather than suffering. The distinction is not between pleasant and unpleasant — genuine sūkṣma-level work is often profoundly difficult and uncomfortable. The distinction is in the orientation: is the difficulty in service of the dissolution of the patterns that produced the harm, or is it punishment-as-suffering with transformation claimed as a side-effect? Fasting, intensive service, sustained contemplative practice, the discipline of learning a craft or a skill — all of these can be genuinely difficult and genuinely transformative. The standard prison regime is genuinely difficult and transformatively neutral at best, actively harmful at worst.
The third is relational embedding. The saṃskāras were formed in relationships and environments. They are reshaped in relationships and environments. Isolated individual therapy within a threatening institutional context operates against the very conditions that make sūkṣma-level change possible. The social fabric — genuine community, with genuine accountability and genuine support — is not an optional enhancement to the transformation process. It is the medium without which the process cannot occur.
The fourth is preparation for reintegration as an integral part of the process, not an afterthought. The moment of release from carceral constraint is, on any sensible analysis, the most critical period in the entire arc of response to criminal conduct. It is also the period to which the least institutional attention and resource is typically directed. A pratyāścitta-logic institution would treat the reintegration as the goal toward which the entire process is oriented — not a ceremony at the end of punishment but the endpoint that gives the process its direction.
Level Three — Kāraṇa: Honest Acknowledgment of Limit#
At the causal level, the honest position is that no legal or punitive institution reaches the kāraṇa śarīra. The deepest dispositional seeds — the fundamental orientations toward existence from which the entire pattern of saṃskāra-accumulation grew — are not accessible to institutional intervention. They are addressed, if at all, by the kind of sustained transformative practice that belongs to the path toward liberation described in Essay VI.
This acknowledgment is not defeat. It is the discipline that prevents institutional hubris — the dangerous tendency of punitive systems to claim ultimate authority over the deepest dimensions of human personhood. The state’s legitimate authority is transactional: it manages the conditions of gross-level coexistence and, at its best, creates sūkṣma-level conditions for genuine transformation. It does not and cannot claim the paramārthika authority that retributivism implicitly invokes.
The practical implication of this acknowledgment is a doctrine of appropriate limits. Some things that the penal system is currently tasked with producing — the genuine reformation of deep character, the healing of intergenerational trauma, the resolution of the kāraṇa-level roots of persistent violent disposition — are not achievable by any legal or penal instrument. They require the slow, unglamorous work of family, community, cultural practice, and the kind of spiritual infrastructure that no state can mandate and no court can administer. The state’s contribution is to create the social conditions in which that work can occur. This is a civilisational project, not a criminal justice one.
VII. Reframing Responsibility#
None of this abandons responsibility. The three-level framework does not say that the person who committed the harm bears no accountability. It says that accountability means something different, and something more demanding, than the retributive picture allows.
At the gross level, the person is accountable in the simplest sense: the harmful action occurred, it caused real damage, and the prevention of further harm may require constraint. This form of accountability does not depend on establishing that the person is the ultimate, uncaused author of a free choice. It depends only on the connection between the person’s action and the harm — a causal connection, not a desert connection.
At the subtle level, the person is accountable in a more demanding sense: for their ongoing relationship with their own conditioning. The saṃskāras were not chosen. The guṇa-composition was not chosen. But the person’s engagement with their own dispositional structure — whether they allow it to run unchallenged or whether they participate actively in practices that move it toward sattva — is something for which accountability is meaningful. This form of accountability does not ask whether the person deserved to be in the situation their conditioning created. It asks whether they are willing to take responsibility for moving in a different direction. This is harder than retributive accountability, not easier.
At the causal level, the appropriate response is compassion — not in the sentimental sense of excusing harm, but in the precise sense of recognising that the deepest roots of the person’s condition lie beyond their control, beyond the state’s reach, and beyond any individual’s capacity to resolve by will alone. This recognition is not a counsel of passivity. It is a counsel of appropriate scope — a refusal to claim that the legal system can accomplish what only a civilisational practice of wisdom can accomplish.
VIII. What the Dharmic Theory Is Not#
Before closing this essay and the first cycle of the series, a clarification is necessary.
The three-level theory of dharmic punishment described here is not soft on crime. It is the phrase that critics will reach for, and it is wrong. Retributive punishment is easy. It requires no understanding of the wrongdoer, no engagement with the causes of conduct, no sustained effort at transformation. It simply inflicts proportionate suffering and declares the account closed. It is, in the Gītā’s typology, a tamasic response: the inertia of inherited institutional form disguised as moral seriousness.
The dharmic framework demands far more. It requires understanding the saṃskāric structure that produced the harmful act. It requires creating the institutional conditions for genuine dispositional transformation, which is slow, expensive, and uncertain. It requires the epistemic humility to acknowledge what law cannot accomplish and to direct resources toward what it can. It requires dismantling the satisfying retributive fiction — the fiction that justice is done when proportionate suffering is inflicted — and replacing it with the uncomfortable truth that justice requires addressing the conditions that made the harm possible.
This is not lenient. It is demanding in exactly the ways that retributive punishment refuses to be.
The Arthaśāstra’s daṇḍa contains harm without pretending to settle ultimate accounts. The Manusmṛti’s pratyāścitta repairs the saṃskāric damage without the illusion that suffering is its own justification. The Gītā’s analysis of āsurī sampat identifies the causal mechanisms of criminal conduct without reducing them to simple moral failure. Together, these constitute the framework of an honest jurisprudence — one that knows what it is doing, acknowledges what it cannot do, and orients itself toward what is actually at stake.
The body is not the self. The law that operates on the body while claiming authority over the self has been punishing in the wrong category. Classical India’s jurisprudence knew this. The question is whether the legal systems that emerged from other traditions can learn it before the recidivism data and the philosophy of mind and the epidemiology of early adversity force the learning upon them without the conceptual framework to make it productive.
The inquiry does not close. It moves forward.
Post 10 in the series, and the conclusion of the first cycle. Posts 1–8 examined the metaphysical presuppositions of criminal liability. Posts 9 and 10 examined what those presuppositions were concealing about the body that is punished and the practice of punishment itself. The second cycle will apply this analysis to specific areas of criminal doctrine: sentencing, parole, juvenile justice, and the jurisprudence of the death penalty in India.