Before the great Śaiva tantric tradition — the mantramārga that produces Trika, Krama, Kashmir Śaivism, the Tantras you've heard of — there were two earlier ascetic orders. The Pāśupatas, founded (according to tradition) by Lakulīśa, brahmin-male only, focused on cultivating identity with Maheśvara through the vrata system. And a second order, going by several names — Kāpālavratins, Lākulas, Kālāmukhas — who pushed further: they took the Pāśupata core and elaborated a higher cosmological hierarchy, more rigorous initiation, more sophisticated metaphysics. Together these two orders form what Alexis Sanderson (drawing on the Svacchanda Tantra and the Niśvāsatattvasaṃhitā) calls the atimārga — the "surpassing path."1
This is the first organized Śaiva ascetic system in the historical record. White uses Sanderson's reconstruction to anchor the historical claim that Śaivism, like Vaiṣṇavism (see BhP 2.2 and the Śrīvaiṣṇava architecture), produced its own systematic operational soteriology before the tantric explosion of the 8th–12th centuries. The atimārga is the substrate the medieval Śaiva tantric tradition crystallizes from. Without the atimārga, the Trika is unintelligible. Without the Pāśupatas, Lakulīśa, and the Lākulas, the Bhairava cults that produce the paravedha and utkrānti technologies of mature Śaiva tantra have no historical antecedent.
The atimārga is defined by the Svacchanda Tantra as the path that "transcended the gnosis, detachment, and mastery (aiśvaryam) that were the hallmarks of the yoga system."1 It is beyond yoga in the early Śaiva taxonomy — yoga is preparatory; the atimārga is the renouncer-ascetic system that uses yoga but is not reducible to it. Atimārga initiates were full renouncers (saṃnyāsins) committed to lifelong ascetic practice. They were the Śaiva counterpart to the Vaiṣṇava virtuoso meditators of BhP 2.2.
The atimārga had two branches. Sanderson reconstructs both from textual sources.
Branch 1: The Pāśupatas. The order whose founder Lakulīśa appears in iconographic record from the 4th century CE (see the standing-Lakulīśa sculptures from the Kushan period and the seated-Lakulīśa lotus sculptures from the 5th-6th centuries). Pāśupata practitioners were brahmin males exclusively (in theory; the practical reality may have varied). Their soteriology aimed at identity with Maheśvara — becoming a "Rudra-Master" (rudreśa) at the time of death, dwelling eternally in Īśvara. Pāśupata initiation elevated the initiand to the sphere of Īśvara as the highest destination; the initiand could not attain deferred liberation from there. This is a terminal destination — the Rudra-Master rules a sphere but does not progress further. The Pāśupata yoga (see the Pāśupata Lakulīśa page for the eight-mastery system) is the operational substrate.
Branch 2: The Kāpālavratins / Lākulas / Kālāmukhas. The order(s) that elaborated beyond the Pāśupata terminus. The textual sources (the Niśvāsatattvasaṃhitā, the Niśvāsamukha, the early Lākula material) are 4th-6th century CE. The Lākula soteriology adds two higher worlds above the Pāśupata Īśvara: the world of the Rudra Tejīśa ("Lord of Splendor") and the world of the Rudra Dhruva ("Lord of the Pole Star"). The lower Tejīśa bestows the "state of yoga" that is the immediate antecedent to final liberation. The higher Dhruva affords liberation itself.1
This is a crucial architectural move. The Lākulas have out-cosmologized the Pāśupatas. Where Pāśupata initiation ends at Īśvara, Lākula initiation continues through Īśvara and up to Dhruva. The Lākula initiate is liberated through mystic knowledge of eleven realms from the Avīci hell to Dhruva. The initiation ritual transmits this knowledge: the preceptor meditates on the cosmic hierarchy, purifying it, transforming it into an instrument of liberation, and then initiates the disciple through the descent of Śiva's power. The disciple is guaranteed liberation at the time of death.
After initiation, the Lākula ascetic spends the remainder of his life in repeated visionary ascent through a gnosis-driven yoga up the ladder of principles or worlds. The ascent is rehearsed; at the moment of death, the ascent becomes actual.
The atimārga two-branch structure solves a doctrinal problem the early Śaiva tradition faced: how to integrate the terminal-Master soteriology (the Pāśupata Rudra-Master who rules a sphere) with the transcendent-liberation soteriology (the Lākula Dhruva-attainer who is fully released). Two operational endpoints, two ritual systems, one ascetic order frame.
The architecture parallels BhP 2.2's two-alternative structure. BhP 2.2 offers the embodied yogeśvara-gati journey through the cosmic spheres or the meditative immediate-realization path. The Pāśupatas correspond to the former (the embodied progression that ends in Master-rulership); the Lākulas correspond to the latter (the meditative transcendence that ends in liberation). The Vaiṣṇava and Śaiva traditions are arriving at the same architectural distinction by different routes.
White's reading: this is evidence of a shared substrate problem across Vaiṣṇavism and Śaivism in the 4th-6th century period. Both traditions are responding to the same Upaniṣadic gradualist vs immanentist tension (see ChU/BU Soteriology Tension) and producing parallel two-track soteriologies. The track-doubling is the integrative resolution both traditions arrive at: rather than choosing between gradual journey and immediate transcendence, the tradition offers both as valid options for different practitioner-types.
The Pāśupata-Lākula split also has a social dimension. The Pāśupatas were brahmin-male exclusive. The Lākulas (and the Kāpālavratins, Kālāmukhas) were more transgressive — wandering ascetics who courted disrepute, used skull-bowls, lived in cremation grounds, mixed with low-caste populations. The Lākula branch is more proximate to what becomes the kaula and Kāpālika practice of mature Śaiva tantra. The Pāśupata branch is more orthodox-renouncer in tone. The two branches encode different social positionings for the Śaiva ascetic vocation.
This page is the historical anchor behind every later Śaiva tantric tradition in the vault. Whenever a Trika, Krama, or Kashmir Śaiva text describes its lineage, its initiation structure, its cosmological hierarchy, or its visionary ascent — the page that explains the substrate the tantric synthesis crystallizes from is this one. The medieval Śaiva tantric tradition is not a fresh creation in the 8th-9th century; it is the synthesis and re-engineering of the atimārga's two-branch architecture.
It also gives the vault the structural parallel between Śaiva and Vaiṣṇava soteriological architecture. The page makes legible that the 4th-6th century period was producing parallel two-track soteriologies across both traditions, responding to the same Upaniṣadic substrate problem. The convergence is not coincidence; it is the natural resolution the Sanskrit corpus arrives at when the gradualist-vs-immanentist tension is engineered into systematic practice.
The most operationally distinctive feature of the Lākula branch is its eleven-realm cosmological hierarchy — from the Avīci hell at the bottom to Dhruva at the top. The Niśvāsamukha (4.95) states that mystic knowledge of these eleven realms, communicated through initiation, is the means of liberation.1 The Lākula does not just know about the eleven realms; he traverses them in his ongoing visionary ascent practice.
The eleven-realm structure becomes the substrate for the medieval Śaiva tantric bhuvana hierarchy. The Trika's twenty-three-bhuvana cosmology, the Saiddhāntika's thirty-six-tattva architecture, the various tantric cakra-hierarchy schemes — all of these are downstream elaborations of the basic move the Lākulas made: transform cosmological mapping into operational ascent practice. The eleven realms are not just a map; they are a ladder. The practitioner climbs the ladder mentally during life and finally during death.
The most striking detail of the Lākula architecture is what happens after liberation. The liberated Rudra-souls that govern the various planes of the cosmic hierarchy are omniscient but not omnipotent. They lack agency. The power to "activate the universe" is reserved for the supreme Śiva alone.1 This is parallel to the Vaiṣṇava soteriology of BhP 2.2 where the Masters of Yoga reside in their cosmic sphere but lack the cosmic agency of Viṣṇu Bhagavān himself.
White's reading: both traditions are reserving agentic divinity for the supreme deity while allowing liberated practitioners to attain gnostic but non-agentic divinity. This is a sophisticated theological move — it preserves divine sovereignty while opening the practitioner-path to authentic liberation. The architectural elegance is what makes these 4th-6th century systems load-bearing for the next millennium of Hindu sectarian development.
You read Sanderson's article on the atimārga (his "Religion and the State: Initiating the Monarch in Śaivism and the Buddhist Way of Mantras" or "The Śaiva Age" — both contain his atimārga reconstruction). You read the relevant passages from the Niśvāsatattvasaṃhitā and the Svacchanda Tantra in the translations Sanderson has provided.
What you notice: the atimārga is not the Śaiva tantra you have heard of. There are no cakras in the technical sense. There is no kuṇḍalinī. There is no guru-disciple lineage in the medieval transmission sense. There is initiation, there is cosmological mapping, there is visionary ascent, there is identity with Maheśvara — but the medieval tantric vocabulary that you associate with Śaivism has not yet been engineered.
This is the substrate. The medieval engineerings that produce Trika and Kashmir Śaivism are 8th-12th century. The atimārga is 4th-6th century. The gap is four hundred years of doctrinal development that you can now read as the crystallization of substrate atimārga material into tantric synthesis. The Pāśupata terminal-Rudra-Master becomes one strand of the tantric pati-sambandha doctrine. The Lākula eleven-realm hierarchy becomes the foundation of the tantric bhuvana cosmology. The atimārga vrata discipline becomes the tantric sādhana.
You sit with the architecture. The atimārga gives you the pre-tantric Śaiva position. The medieval tantric synthesis you may have studied becomes legible as synthesis: it is doing the integrative work on a substrate it inherited rather than inventing the architecture from scratch.
This is the practice the historical reading makes possible. Not a new technique. A re-historicizing of your relationship to the Śaiva tradition. The medieval Trika is younger than you thought. The substrate it draws on is older than you thought. The doctrinal development from atimārga to mature tantra is visible once Sanderson's reconstruction is in view.
You close the book. The Śaiva tradition has acquired a backstory it lacked before.
The failure mode is treating the atimārga as primitive Śaivism that medieval tantra progresses beyond. This is the same progressive-narrative failure that the Knowing Supplants Going page warned about. The atimārga is not primitive. It is prior. The medieval tantric synthesis is not a maturation of the atimārga; it is a different synthesis on shared substrate material.
Diagnostic sign: you read medieval tantra as superior to the atimārga because of doctrinal sophistication.
The honest reading recognizes both as sophisticated systems operating in different historical moments and producing different operational outcomes. The Pāśupata terminal-Rudra-Master soteriology is doctrinally complete on its own terms. The Lākula eleven-realm visionary ascent is operationally rigorous on its own terms. The medieval tantric synthesis is a third sophisticated system that does new work; it does not invalidate the prior systems.
Sanderson's reconstruction is based on the Svacchanda Tantra (6th c. CE), the Niśvāsatattvasaṃhitā (4th-6th c. CE), the Niśvāsamukha, the Mataṅgapārameśvarāgama (pre-9th c. CE), and various Pāśupata textual sources.1 The reconstruction is partially philological (close textual reading) and partially archaeological (working back from later tantric sources to earlier substrate material that the later sources presuppose).
The first tension is whether the Kāpālavratin, Lākula, and Kālāmukha names refer to one order or to three distinct orders. Sanderson's reading is that the names are partially synonymous (referring to different aspects or regional variants of one broad Śaiva ascetic movement) and partially distinct (the Kālāmukhas may have been a later organizational form distinct from the earlier Lākulas). The textual record is not crisp enough to resolve the question definitively.
The second tension is the dating. Some atimārga material is securely 4th-6th century CE (the Niśvāsatattvasaṃhitā). Other material is later but the traditions it preserves may be substantially older. Sanderson's general position is that the atimārga is no later than the 6th century and perhaps as early as the 4th. This dating places the atimārga firmly before the medieval tantric synthesis and contemporary with the formation of the classical Hindu Purāṇic corpus.
The third tension is whether the Pāśupatas and Lākulas were competing branches of the atimārga or complementary branches with different practitioner-bases. Sanderson reads them as institutionally distinct but operating in shared doctrinal space. The two-branch structure is robust in the textual record but the social relationship between the branches is harder to reconstruct.
White depends heavily on Sanderson for the Śaiva-historical material; the two scholars converge tightly on the atimārga reconstruction. White's contribution is to integrate Sanderson's Śaiva work into the broader yoga-history frame the Sinister Yogis project develops.
Samuel's two-period thesis sets the chronological frame.Samuel — Two-Period Thesis The atimārga sits in the late substrate-period / early crystallization-period transition — exactly where Samuel's thesis predicts the major institutional and doctrinal crystallizations should be occurring. The atimārga is Samuel-period-appropriate: it is the kind of systematic ascetic doctrine that would emerge from the substrate as the substrate begins to crystallize into Śaiva and Vaiṣṇava systematic traditions.
The parallel with BhP 2.2 (Vaiṣṇava soteriology, 4th-6th century crystallization) is structural and historical. Both traditions are responding to the same Upaniṣadic substrate problem; both are producing parallel two-track soteriologies; both are forming in roughly the same chronological window. The convergence is not coincidence but the natural solution-shape the period produces.
The convergence with Lakulīśa Pāśupata Reform anchors the Pāśupata branch in its founder's historical and iconographic record. The early Lakulīśa sculptures (4th c. CE Kushan-style standing figures, 5th-6th c. CE seated-on-lotus figures) provide the archaeological anchor that confirms the textual dating.
The structural insight — that a single ascetic movement can have two operationally distinct branches that share doctrinal substrate while differing in soteriological endpoint — translates across domains.
Behavioral mechanics: Two-Track Discipline System Design — Operational disciplines (military training, surgical residency, monastic orders, special-forces selection) frequently develop two-track internal structures: a primary track that produces functional operators and a secondary track that produces senior practitioners with capacities the primary track does not develop. The atimārga's Pāśupata-Lākula split is structurally identical: Pāśupata produces functional Rudra-Master ascetics; Lākula produces senior practitioners with the meditative-transcendence capacity. The Sanskrit case sharpens the design insight by showing that two-track architecture is the natural resolution of a discipline's internal tension between terminal-mastery and continued-development outcomes. Disciplines that try to force a single track tend to lose either the functional mastery (everyone aiming for the senior outcome and failing to develop functional capacity) or the senior development (everyone settling at functional mastery and never developing further). The two-track design preserves both.
Psychology: Practitioner Track vs Development Track Clinical Architecture — Clinical psychology faces the same architectural problem. Some clinicians are functional masters — competent therapists with stable practices, predictable outcomes, no ambition to develop further. Some clinicians are development-track — perpetually expanding capacities, working at the edge of clinical knowledge, treating the discipline itself as an unfinished project. Both tracks are necessary. The discipline that pushes everyone toward development-track loses its functional therapists. The discipline that pushes everyone toward functional-mastery loses its development edge. The atimārga's solution — two parallel tracks within one ascetic order — is the architectural insight: name both tracks, value both equally, route practitioners to the track matching their capacity and ambition, and let the development-track inform the functional-track through textual transmission while keeping the functional-track operating on stable doctrinal foundations.
The Sharpest Implication
If the atimārga's two-branch structure is the natural resolution of the gradualist-vs-immanentist tension in systematic Śaivism, then the contemporary Anglophone Śaiva community — which tends to identify with the medieval tantric synthesis (Trika, Kashmir Śaivism, Spanda) — is missing the substrate the medieval synthesis was built on. The implication that cuts: you cannot fully understand the medieval Trika until you have read what it was synthesizing from. The atimārga is not optional historical background; it is the architectural prerequisite for the medieval system that succeeds it. Modern Śaiva practitioners reading Abhinavagupta without reading Sanderson on the atimārga are operating with a half-built foundation.
Generative Questions