The publication of a critical edition and scholarly annotation of the apparent yoga pAda of the mAlinI-vijayottara-tantra (MVUT) by the learned western indologist Somadeva Vasudeva (done as part of his doctoral research) greatly facilitates the study of evolution of yoga within the realm of mantra-mArga. SV belongs to the lineage of Alexis Sanderson of Oxford who has also been a learned Sanskrit scholar diligently studying various tantric texts. This publication is an instance of the efforts of a western indologist making a positive contribution to Hindus in their “svAdhyAya”. We had earlier pointed out how the roots of classical sAMkhya lie plainly in the later layer of brAhmaNa texts, which encompass the upaniShad-s. Thus, a major component of vedAnta (in the literal sense as the terminal teaching of the veda-s) is sAMkhya. In this context it is interesting to note how the mAlinIvijayottara calls sAMkhya as vedanta. In chapter 16 the MVUT describes a meditation in order to be able to perceive the puruSha tattva. It states that after performing the yoga for 3 years one attains tulyatA with the regent of the puruSha-tattva. It then goes on to add that:
“etad vedAnta-vij~nAnaM samAsAd upavarNitam | kapilasya purA proktam etad vistarasho mayA ||” MVUT 16.24
This statement immediately raises a number of points: 1) Clearly, the tantra is describing a tattva of sAMkhya-yoga (i.e. puruSha) and how it is perceived via meditation. 2) Further, it calls this teaching regarding a sAMkhya-yoga tattva as the summary of the knowledge of vedAnta, which as earlier expounded by shiva at length to kapila, who is acknowledged from the earliest period of his mention as the promulgator of sAMkhya. 3) Further this statement from this rather evolutionarily derived tantra does actually remind one of the far more ancient and ancestral text the upaniShad of shvetAshvatara.
This last connection is of some interest because a number of points: 1) The shvetAshvatara is one of the earliest texts with a shaiva leaning [*1]. 2) It is also one of the earlier texts that clearly describes a form of yoga (SU 2), which is a predecessor of some of the subsequent yogic practices including those seen in the mahAbhArata, yoga-sUtra-s of pata~njali and the siddhAnta tantra-s of the Urdhva-srotas. 3) It is also one of the earliest texts giving the name sAMkhya in connection to its sAMkhya-yoga philosophy and mentioning its founder kapila (SU 5.2, second pAda).
“R^iShiM prasUtaM kapilaM yas tam agre j~nAnair bibharti jAyamAnaM ca pashyet ||”
The brahman (which has been earlier identified with rudra in the SU) is said to be generate kapila who bears the knowledge of the system.
Thus, the statement of the MVUT is completely consistent with the internal evidence from the SU, which is a bona fide, even if late, component of vedAnta. This link of a sAMkhya-yoga element between an ancestral and a highly derived text in the shaiva tradition raises some questions. Is it merely a lateral transfer of ideas at a time point closer to the redaction of the MVUT? Or does it actually represent an inheritance that has continuously been transmitted from the earliest shaiva traditions down to the much later trika tantra-s? A cursory analysis of this issue has some value for reconstructing the evolution of aspects of yoga.
As the learned editor Somadeva Vasudeva points out the MVUT represents a rather nuanced synthesis of two distinct systems of yoga present within the shiva-shAsana, namely that from the saiddhAntika tantra-s and the kula texts (i.e. successors of matsyendra). But this occurs in the context of a third distinct stream of the shiva-shAsana i.e. trika tantra-s to which the MVUT belongs. Our examination of the texts suggests that the core of the trika tantra-s appear to have evolved from the more ancient bhairava tantra-s of the dakShiNa-srotas (those emanating from the aghora face). MVUT tradition presents itself as the successor of an even earlier trika tantra the siddha-yogeshvarI-mata, via the mAlinIvijaya (of which it is the uttara tantra). However, the earlier evolution of yoga within the trika stream remains mysterious due to the loss of key texts from the earlier period. The surviving text of the siddhayogeshvarI-mata tantra (which seems partial) does not have any serious account of yoga. So it is not clear if the MVUT is the first to cause the synthesis or it is building on earlier lost synthetic attempts. Nevertheless, we can study the evolution of the yoga from the kula and saiddhAntika sources to look back at their origins. Very crudely, we can characterize the kula tantra-s as having major focus on the system of the accent of kuNDalini, whereas the yoga of the other systems like saiddhAntika-s tend to instead focus on other types of contemplations, more in line with those of the pAta~njala tradition.
Now the origins of both these elements can be traced back to the upaniShadic period. The proto-kuNDalini class of ideas are mentioned in the ChAndogya upaniShad 8.6.6 and maitrAyaNI upaniShad 6.21. The latter is especially clear in mentioning the suShuMNa channel by name and detailing a proto-type of the kuNDalini-raising procedure. This is further developed to a certain degree in the bhR^igu-smR^iti embedded in the shAnti-parvan of the MBH. Finally, we see it arriving in its full-fledged tantric form with the chakra-s in the kaula-j~nAna-nirNaya, an early kula text. On the other hand the beginnings of the yoga of dhAraNa (closer to the pAta~njala tradition) are seen in different early upaniShad-s including chAndogya but more explicitly in the shvetAshvatara (SU 2.8 onwards). It is further developed in the description of yoga in the mahAbharata (e.g. MBH-vulgate 12.195, 12.317) and finally systematized by pata~njali and his commentator vyAsa. Thus, the evidence of the yoga-sUtra-s and the MBH suggests that there was a long development of both these yogic streams independent of the shaiva context.
Does this mean the shaiva-s adopted them secondarily as the nAstIka-s of both the bauddha and jaina variety did? Evidently not because given the SU we know that the shaiva stream had a yoga association right from its origin. This further supported by the fact that the shaiva-s continue to show this association with yoga throughout their early evolution. The early post-vedic shaiva-s represented in the MBH clearly describe rudra as yogAtman and the goal of yogi-s. For example the ajAnana virachita rudra stuti (MBH-critical 12.47.35) states:
“yaM vinidrA jitashvAsAH sattvasthAH samadarshinaH |
jyotiH pashyanti yu~njAnAs tasmai yogAtmane namaH ||”
This verse implies the practice of yogic vigils, prANAyAma and also dhAraNa on the form of light by these shaiva-s. The MBH already recognizes pAshupata-s, and the pAshupata-s of the lAkulIsha stream (the lineage of lakulIsha) who follow the pAshupata sUtras appear to be the next recorded stage of shaiva development after those of the great epic. The paurANic account of pAshupata tradition records lakulIsha as preceded by an earlier lineage of pAshupata yogin-s. The pAshupata sUtra-s begin with the statement: “athAtaH pashupateH pAshupataM yogavidhiM vyAkhyAsyAmaH | (PS 1.1)”. Further it states: “ato yogaH pravartate (PS 5.23)”.
The pa~nchArtha bhAShya [*2] of kauNDinya on the PS explains that the yoga of the pAshupata includes yoking or unifying oneself with rudra (e.g commentary on PS 5.33). Thus, we find these early pAshupata-s strongly aligned with the practice of yoga and this in turn represents a continuity with their predecessors in the great epic.
When we look at the saiddhAntika-s, we find that their yoga appears to be aligned with that of the earlier shaiva-s (i.e. emphasis on dhAraNa and the sAyujya with sadAshiva) and more generally with the practices of the pAta~njala tradition. An examination of the earliest surviving layers of the siddhAnta tantra-s shows that they appear to subsume within them an older pAshupata/lAkulIsha base. While the present themselves as superseding the pAshupata/lAkulIsha doctrines in terms of their structure, they are generally respectful towards them and their practices as a more primitive branch of the shiva-shAsana. For example, the niHshvAsa tantra presents the procedure for performing the pAshupata vrata that is generally similar to the versions presented in the pAshupata system. Importantly, the saiddhAntika-s also share the pa~nchabrahma mantra-s with the ati-mArga. Thus, in light of these inheritances we see no reason not to see the yoga of the saiddhAntika-s as another inheritance from an ancestral pAshupata/lAkulIsha base [*3]. Thus, we may trace a continuous stream of dhAraNa-type yoga associated with the shaiva-s from their origins in the SU, through the early shaiva-s of the epic, the pAshupata/lAkulIsha ati-mArga shaiva-s and the saiddhAntika-s prior.
In the case of the kula system we can trace it amongst the shaiva-s of the bhairava tantra-s down to the earliest kula texts. The kaulaj~nAna-nirNaya mentions several different streams of kaula-s, suggesting that there was an extensive history prior to that fairly early text. Is it true that there is a long “ghost-lineage” of lost ati-mArga (or coeval equivalents) texts with kula-type yoga in parallel the ati-mArga pAshupata-s whose texts partially survive? This appears unlikely from the commentarial tradition of abhinavagupta and his cousin kShemarAja, and various saiddhAntika scholars from Kashmir that preserve a consistent account of various shaiva systems from a period closer to their heydays. Instead, it indeed appears likely that matsyendra was a revolutionary who actually introduced the “secular” precursor of the kula yoga that we mentioned above into the shaiva context. Some murky predecessors of his from an earlier yuga are mentioned, like khagendra – it is possible that these were even earlier developers of the system.
*1 nIlarudra and shatarudrIya proper are indeed much older. But both these texts are not ancestral shaiva-oriented upaniShad-s (as they were called in later shaiva literature) but actual mantra-s from the paippalAda AV and YV saMhitA-s respectively. Only after the shvetAshvatara they were used in the shaiva, as opposed to the purely vedic sense.
*2 the pAshupata sUtra-s have 5 sections corresponding to the 5 brahma mantras, which are given at the end of each section.
*3 This relationship between the saiddhAntika and pAshupata systems is a clear illustration of the principle we stated earlier of how structure of a text often preserves a signature of its evolutionary process incorporating earlier systems inside. In the case of the shaiva texts this signature might assume the form of loka-s or layers of emanations of rudra. The early pAshupata-s added rudra as a principle beyond the original 25 of sAMkhya. The later ones probably including the lAkulisha-s added a series of emanation of shiva above the basic rudra-s going all the way to dhruva. Then the saiddhAntika-s added other emanations going beyond all the way to sadAshiva. The trika and shrIkula tantric-s have added the preta-s of brahma, viShNu, rudra, ishvara and sadAshiva to the throne of their deities, thus creating a further level.