Just as I half-concentratedly read through Rong-zom-pa’s commentary on the Sarvabuddhasamayogatantra, this line made me pause a bit longer (RZ2: 595): srin skad ’di yang he ru ka’i gsang sngags su byin gyis brlabs so || (“Also the cry/sound of the demon has been empowered [by the Heruka] as a mantra of the Heruka”). We may not believe in a demon, a Heruka, or a mantra, but what does this line tell us regarding the idea of mantra (a kind of “magical formula”). In plain terms, it suggests (at least to me) that a mantra can be “devised” or “formulated” (although not by anyone). In other words, what has initially been a non-mantra has been “mantricized” or “mantrified.”
It is as though the attainment of Buddhahood via Mantric means is a huge soteriological project. For that, all kinds and classes of non-human beings (e.g. lha ma srin sde brgyad) are recruited (by hook or crook). They are asked or somehow forced to participate. Their tools and techniques came to be “empowered” and employed. The accounts of Padmasambhava’s coming to Tibet deal a great deal about the “Buddhicization,” “soteriologization,” “mantricization,” or “mantrification” of human and non-human forces. I can imagine how a gTer-ma apologist could use the idea of an ahistorical figure such as the Heruka “mantricizing” a demon’s sound and making it a mantra to justify emergence of new mantras in Tibet (or anywhere else). Of course, we could argue that Indian mantras are “genuine” and hence efficacious, whereas Tibetan mantras are simply “hocus-pocus” or “abracadabra.” Well, this would take us to the very heart of the issue: “What is a mantra and how is it supposed to work?” I would not dare answer these questions (yet)!