I submit therefore that this passage opens up a very different way of looking the Candrakirtian treatment of the ālayavijñāna. Since we ought to accept that Jayānanda possessed the oral lineage of interpreting this text, I think we can safely say that this passage means we really ought to carefully rethink whether Candrakirti so thoroughy rejects ālayavijñana as some Tibetan Madhyamaka scholars seem to think he does.
It seems to me that the passage actually does not do a great deal to alter the way Chandrakirti describes the alaya, although an interesting distinction is introduced. Chandrakirti and Jayananda both clearly do not accept the 8th consciousness that is the repository of karmic seeds, but say that the term "alayavijnana" is being used to teach the ultimate, emptiness. Jayananda introduces the idea that this term can also be used to indicate the consciousness knowing the ultimate, becasue these two are non-dual.
This is not unique to Jayananda as Nagarjuna does the same, although less explicitly, in his Bodhicittavivarana.
In conclusion it is as clear as ever that Chandrakirti completely rejects an Alaya which is of a different nature from the six consciousnesses, although he accepted the use of Alaya understanding the referent object to be the ultimate, and on some occasions, the mental consciousness.
This discussion arose over a flap with Gorampa saying that although such are rejected ultimately, this does not mean that we should not accept entities like the alaya, the manas and self-cognizers conventionally.
I agree with his point that just because something is rejected ultimately does not mean that it is conventionally non-existent. However, I believe this point is misapplied in his passage, becasue as it turns out the three examples are things which are utterly non existent in both truths. The fact that we can use the terms "alaya" to refer to the sixth consciousness does not really bear on that point, becasue it seems clear in the context of the original discussion through the inclusion of the seventh consciousness posited by the cittamatrins and the self-cognition clearly refuted by Shantideva and others, that Gorampa was not referring to this.
Further, I think Gorampa's arguments and consequences here do not add up, mainly because the example of the butterlamp dispelling darkness is undifferentiated in Nagarjuna, he shows that the lamp does not dispel darkness ultimately in order to prove that arising from self and other do not exist ultimately.
However just as arising from other does
exist conventionally, so too a lamp dispels darkness conventionally. Arising from self, however exists in neither truth, and in a similar way self-cognition is faulty in both truths.
Gorampa makes some poor arguments here. The fact that I am a follower of Je Tsongkhapa does not mean I am rejecting Gorampa's points for that reason, and it is not the case that great and renowned scholars like Gorampa never make mistakes. I think he is wrong here, and elsewhere, and have good reasons to think so.
In other words, for Gorampa there is no doubt that these things like ālaya-vijñāna, the afflicted mind, and so on cannot bear analysis -- but as conventions we leave them alone with all their myriad internal contradictions just as we leave the convention that lamps remove the darkness of rooms alone.
This is not the case. Some things that cannot bear analysis are also dispelled conventionally, taking into account the contradictions that arise if they were to be allowed t stand conventionally. Consult Shantidevas chapter 9 for a flawless description of how self-cognition is non-existent even conventionally. He does not say "this does not exist ultimately, but let's allow it to stand conventionally even though it is contradictory." He demolishes it.
The fact that we leave a convention like "a lamp lights up a room" alone arises from the fact that it does precisely that, and cannot be shown to be false conventionally. It is conventionally non-contradictory. Examining how the light and darkness meet and so on is the examination for the ultimate, and so is irrelevant here.