Ārali and Rigi - an odd pair (I)
According to Bu ston's classification (in the Rgyud sde rnam bzhag rgyas pa) these two tantras form a sub-class of their own under the broader family of the Heruka-tantras, the other sub-classes being the vast corpus of Śamvara-tantras along with those of the Hevajra, Buddhakapāla and Mahāmāyā.
When I first came across these tantras a few years ago, I was under the impression that they were fabricated in Tibet. I have no idea why I thought this, perhaps because of the odd names of the main deities. I should have suspected that there must have been some kind of evidence with Bu ston (since he considers it 'original').
Since then - with the help of Prof. Sanderson - I came across some evidence that at least the Vajrārali was known on the Indian subcontinent. A manuscript fragment from Cambridge (Or. 158, uncatalogued by Bendall) dating from 1162 AD has on its title page "vajrāmṛtatantra || vajrāralitantra || buddhakapālatantra ||". Unfortunately the fragment has bits and pieces of the Vajrāmṛta and the Buddhakapāla only. Since then, at a lecture held in Oxford, Prof. Harunaga Isaacson has signaled the identification of one folio with the actual text, also kept in Cambridge.
The first thing that strikes one as odd about this small corpus are the titles themselves. The expression 'ārali' is seen earliest in the yoginītantras (as far as I know) in a mantra in the Sarvabuddhasamāyoga. In a mahāyoga context, Āryadeva uses it more than once in his Caryāmelāpakapradīpa and all seem to mean a kind of 'play' (rol pa) associated with the Buddhas. Wedemeyer, as far as I can remember, always translates it as 'extensive play'. The etymology remains uncertain, but it is to be noted that all permutations of a/ā and l/ll seem to occur.
It should also be noted that this is not the interpretation that Bu ston seems to have preferred. We may surmise this only indirectly however. In a closing note after his synopsis of the Vajrārali he states that in some Tibetan collections the title is 'Heruka rol pa'i rgyud' (*Herukāralitantra?), but this he considers a total lie: ''dir rgyud 'bum kha cig tu He ru ka rol pa'i rgyud ces bya ba bris pa yod de | de ni rdzun ma yang dag zhig go ||.'
'Rigi' is even stranger. In this corpus the word occurs only in the Rigi-ārali-tantra, there 'Rigi' is the main goddess - far more than a mere consort to the Heruka 'Ārali', she is actually the teacher of the tantra. There seems to be one occurrence where 'rigi' is glossed as simply 'ḍākinī': ri gi mkha' 'gro mar ni gsung (D 179a). As far as I can tell, 'rigi' occurs only once in the yoginītantra corpus as the name of an entity, namely in a nebulous introductory line of the Catuṣpīṭhatantra: 'rigīnāṃ jñānam īśvaram'. However, Bhavabhaṭṭa glosses it not as 'ḍākinī' as we would expect from the line above but as 'buddhas': rigīnāṃ buddhānāṃ jñānam īśvaraś ca.
It is quite evident that the Vajrārali is much earlier that the Rigi-ārali. As I said above, the consort appears only in the second tantra, and this scripture betrays close connections with other yoginītantras. Such Hevajra-specific elements as the 'four blisses' and so on are missing from the Vajrārali but are there in the Rigi-ārali. Sacred places (pīṭha, kṣetra, cchandoha, etc. and their upa- varieties) from the Śamvara-tantras also turn up here. I believe that any future, closer study should take into account the possibility that the Rigi-ārali is a recycled variant of an older cult of Vajrārali.