The devI purANa is one of the several dedicated shAktA purANa-s such as devI-bhAgavata, mahAbhagavata and kAlikA. The sections of other purANa-s like the proto-skanda purANa (the vindhyavAsinI section), the mArkaNDeya purANa (the devI mAhAtmya section) and the lalitopAkhyAna of the brahmANDa purANa can also be considered mini shAkta purANa-s. The dedicated shAkta purANa-s and the lalitopAkhyAna section contain distinctive tAntrika material beyond the core Pauranic story frame. This material is greatly developed in the kAlikA purANa and the lalitopAkhyAna. Both of these have a regional flavor with the locus of kAlikA being eastern India centered on the kAmAkhya pITha and that of the lalitopAkhyAna is south India centered on kA~nchi. In a sense the lalitopAkhyAna is similar to the Tantric vindhyAchala mAhAtmya, but found is way into an older mahApurANa as the uttara bhAga. Its main objective is incorporation of the kAdividyA lineage of the upAsana of tripurasundarI and the nityA-s from the shrIkula tAntrika system inside a smArta system mainly followed by the south Indian taittirIyaka-s. The kAlikA appears to be more plainly tAntrika expounding the upAsana of the shrIkula goddesses, durgA-chaNDikA and vaiShNavI-mahAmAyA. The devI purANa is centered mainly on the upAsana of chaNDikA-vindhyavAsinI the slayer of the asura ghora who assumed the form of a buffalo. She is described as riding a lion comprised of the combination of all other deva-s rather than being formed by all the deva-s (as in the devI mAhAtmya). She stations herself in the vindhya-s as a beautiful virgin and is sought by ghora. In course of a fierce battle ghora assumed the form of buffalo and was beheaded by the goddess. So her name became mahiShamardini. Thus, it appears that the devI purANa preserves a parallel version of the famed shAkta myth of the killing of the buffalo demon. The central tAntrika vidyA taught in the devI purANa is the padamAla mantra of chaNDikA, which has many different mantra pada-s for different viniyoga-s (objectives).
Interestingly, the devI purANa also preserves a small mantra saMhitA of complete Vedic mantra-s (not just pratIka-s) for the performance of the daily Vedic ritual (chapter 56). This mantra saMhitA is primarily composed of R^ik-s but also contains some yajuSh-es and a late brAhmaNa-like terminal portion. This saMhitA is of interest because it appears to present several variant readings from those seen the well-known Vedic R^ik saMhitA-s: the shAkala saMhitA, the AV-S 20, RV-khila, the AshvalAyana saMhitA which incorporates the khilAni into the main R^ik saMhitA equivalent to the shAkala collection, the hautra parishiShTha of the Apastamba taittirIyaka-s, the small R^ik saMhitA of the hiraNyakeshin taittirIyaka-s and the saMaveda saMhitA pATha-s. I must state that I have not examined any complete manuscripts of the devI purANa but the editor used at least 4 of them from eastern India and some of these variant readings in the R^ik-s are uniformly seen across all the manuscripts used in the printed edition. The devI purANa was also known in late medieval South India among shrIkula adepts. It is mentioned by the great bhAskararAya in his saubhAgya-bhAskara (commentary on LS 200). However, I have not been able to locate any South Indian manuscript or publication of this purANa – this would have been useful in determining of those variant readings were also prevalent in the south.
Given that these reading drastically differ from all the extant homologs of the given mantra we may postulate the following distinct but not unrelated explanations:
1) These readings were acquired from a now extinct shAkhA that was isolated from the mainstream shrauta tradition early on, thereby diverging drastically from the others.
2) The readings are from the known shAkala text but due to memetic drift arising much later isolation, which prevented correction through meme conversion with the mainstream texts, the readings greatly diverged (analogs of the evolutionary processes occurring in genes). This isolation could have been spatial, i.e. geographical isolation, perhaps in eastern India, beyond the zone with a high concentration of vaidika practitioners. Or else it could have been traditional i.e. being a tAntrika tradition and text the devI purANa mantra saMhitA transmission was not subject to correction from the mainstream vaidika tradition.
While the grammatical irregularities in this mantra saMhitA might support the less interesting alternative #2, we still suspect that these differences might in part indeed reflect a distinctive Vedic tradition, even if not a distinct shAkhA tradition. Evaluating this is hampered by the fact the only RV tradition that survives is the shAkala shAkhA. More recently collection supposed to be the AshvalAyana shAkhA was published. However, this saMhitA is identical to the shAkala saMhitA, except that it has the kAshmIra pATha of the khila inserted into the saMhitA itself. From various sources it was clear that the bAShkala shAkhA showed differences in the arrangement of the upamaNDala-s of the first maNDala (i.e. the gautama-kakShIvAn-paruchChepa-kutsa-dIrghatamas order) and had an additional samj~nAna sUkta in the 10th maNDala which ended with the R^ik “tach ChaMyor AvR^iNImahe …”. Of course there is also the traditional variability between these R^ik collections in terms of the inclusion or not of kANva-s of the vAlakhilya section of maNDala 8. What we really do not know is whether the readings of the cognate R^ik-s in bAShkala had substantial differences from those in shAkala. Now some workers have taken the reference to bahvR^icha by AchArya-s like kumArila bhaTTa to mean that there was an eponymous RV shAkhA. There is no support for this view – 1) we do not find this name as a shAkhA in the charaNavyUha-s, which typically only name seven recensions. 2) pata~njali says: “ekaviMshatidhA bahvR^ichyAm” i.e. there are 21 shAkhA-s of the RV – he means RV by bahvR^icha. Yet the term bahvR^icha might point to a connection between certain RV practitioners and the shAkta system because there is a shrIkula upaniShad known as the bahvR^icha upaniShad. Whether these were connected to the compilers of the mantra saMhitA in the DP remains to be seen.
Some examples of variant readings:
indraM vishvA avAvR^iShaM samudravyachasaM giraH |
rathItamaM rathInAM vA rAjAnAM shatapatiM patim ||
nAke suparNam upatyAnta havA revanto abhya chakShata tvA |
hiraNyAkShaM varuNasya dUtaM yamasya bhuraNyuM yonau shakunaM ||
varuNaM bodhipAdam R^ichAm indraM havAmahe |
pari vrajech cha bAhvor jagandhaso svarNavam naraM ||
~AV-paippalAda 20.5.8 [~RV khila 184.108.40.206]:
AdityaM devaM savitA manyoH kavi kratum archAmi |
satyasavaM ratnadhAm abhi priyam atitaram kaviM ||
~AV-paippalAda 20.5.9 [~RV khila 220.127.116.11]:
UrdhvA yasyA matibhA Adidyutat savImani |
hiraNyapANir tvamimIte sukratuH kR^ipApayaH ||
somaM rAjAnam bhavate agnim anvArabhAmahe |
AdityaM viShNuM sUryaM brAhmaNaM cha bR^ihaspatim ||
gaurIr mimAya salilAni takShaty ekapadI dvipadI sA chatuShpadI |
aShTApadI navapadI bhUmIM sahasrAkSaM vA parame vyoman ||
mA nastoke tanaye mA na AyUShi mA no goShu mA no ashveShu rIriShaH |
mA no vIrAn rudra bhAmito vadhIr haviShmantaH sadamit tvA havAmahe ||
These examples (by no means the entire set of paTha bheda-s in this collection) clearly illustrate the grammatical irregularities and the meaning-deforming mutations that are likely indicative of reflect poor preservation. Nevertheless, it should be noted that the DP tradition was founded by a genuine vaidika tradition with knowledge of a R^ik saMhitA, including the khila mantra-s, as well as a yajuSh tradition. In light of the changes in word order found in many of the mantra-s it is possible that these represent genuine pATha-bheda-s in this tradition with respect to the mainstream shAkala tradition. However, we would be cautious, in light of the above discussed difficulties, about proposing that they inhere from a lost eastern Vedic recension.
After providing the core mantra-s for the sthAlIpAka oblations the section has an appendix imitating brAhmaNa-like prose which makes this tradition distinct:
evaM vediko.agniH saMtuShTho bhavati | athAtaH paristaraNa devatAH kathyante | parisamUhane kAshyapaH | upalepane vishve-devAH | ullekhane mitra-varuNau | ullikhane pR^ithvI | ambukShaNe gandharvAH | agniH sAdane sharvaH | parisamUhy opalikhy ollikhana karma vidhir uchyate | dakShiNAsAdane brahmA | uttarataH pUraNe sAgarAH | AstaraNe trasAH | athAvasAdane shatakratuH | pavitra bandhane pitaraH | prokShaNI saMskAre mAtaraH | juhvate sruche sruvApAsh cha brahmA-viShNu-maheshvarAH | AjyAtapane vasavaH | adhishravaNe vaivasvataH | paryagni-karaNe marutaH | udvAsane skandaH | utpanna pratyutpavane chandraH dyauH | AjyAvekShaNe dishaH sarvAH | pavitra dhAraNe praNItAH | yAmumA devI | idhme cha lakShmI | vishvasya vishvAbhUtAni ||