What is the logical structure of this argument by a Hindu philosopher about causation?

Question

The Khandana Khanda Khadya is a work by the 12th century Hindu philosopher Sri Harsha which in some sense sets out to disprove epistemology - it seeks to show that all putative means of knowledge, whether sensory observation, inference, or scriptural revelation, are ultimately invalid. That's because Sri Harsha subscribed to a Hindu Vedanta philosophy called Advaita, according to which the physical world is an illusion, and thus the means of knowledge which seem valid in the world of appearance do not capture the true nature of reality.

Now in section 2 of the Khandana Khanda Khadya, Sri Harsha responds to the argument that certain entities postulated by logic, like premises, conclusions, and the like, must be real because they're the cause of all philosophical discussion. Sri Harsha's response is that he agrees that those entities are the cause of all philosophical discussion, but he denies that they must therefore be real. He first argues that "[t]he non-real can have causal efficiency", i.e. that just because something is a cause doesn't mean it's real. And then he goes one step further and argues that "[c]ausal efficiency cannot belong to that which has real being", i.e. that if something is real then it can never be a cause! Here's his justification for the latter statement:

Causal efficiency cannot belong to that which has real being.

I. ‘If a cause be that into the nature of which real existence enters as an essential element, then, for this very reason, the cause has no real being.'
II. 'If, on the other hand, real being does not essentially enter into the nature of the cause, then, for this very reason, the cause has not real being.'

The meaning of this stanza is as follows:—

I. If the nature of the cause be such that it implies as an essential element real existence, then to say that the generic character ‘real existence' belongs to the cause would involve the absurdity of something (real existence) residing partially in itself (i. e. that real existence which goes to constitute the nature of the cause). Even if the thing qualified by real existence (i.e. the cause with such existence as an essential element of itself) were considered as something different from real existence (so that the said absurdity would not arise), we could not accept the real existence (in the latter sense, i. e. the real existence which is predicated of the cause) to be the same with the real being that enters into the nature of the cause; for it is a recognised principle that no more than a thing can reside in itself, can it reside in that of which it already is an essential attribute. It would therefore be necessary to assume another existence as residing in the cause qualified by existence ; and as this would mean that existence does not enter into the nature of the cause, the cause would have to be regarded as ‘not really existing.' And if, in order to avoid this, we were to assume a series of existences, one after the other, there would be no end of such assumptions.

Nor will you escape from this predicament by taking the long step of assuming an infinity of different kinds of real existence. For if you assume different kinds of real existence, you relinquish the very foundation on which the generic conception of ‘existence' rests, and hence lose the idea of even the first existence. Seeking to establish the notion of existence you thus have lost the basis of it, and are worse off than before! ...

II. Let us then consider the second alternative stated in para. 44, viz., that that which really is constitutes the cause, without ‘real being' entering into it as an essential element.—On this view, we point out, that which has no real being also may be a cause, since real being and non-being equally do not enter into the nature of the cause.

Can someone please explain the logic of Sri Harsha's argument, especially the part in bold?

He seems to find some kind of contradiction between the statement "The nature of the cause is such that it implies as an essential element real existence." and the statement "The generic character ‘real existence' belongs to the cause." I can barely even tell these two statements apart. But apparently he thinks that if both statements were true, that would imply that something resides partially in itself, which is impossible. And so he concludes that they cannot both be true. And then he argues that if the first statement is false, then we have no compelling reason to believe that the second statement is true.

None of this makes much sense to me, especially because the two statements seem so similar to each other. So can someone explain to me the difference between these two statements, and elucidate the logical structure of this argument?

Note that I previously posted a question about this passage on Hinduism Stackexchange here, but that was more about examining the argument in its religious context. Here I just wanted to focus on the logical structure.

Answer 1

The boldface passage is fairly easy to interpret, the rest of the argument not so much. Śri Harsha seems to deal with a one-level ontology similar to scholastics', where attributes (essential properties) are things in their own right. His "nature" seems to correspond to scholastic "essence", the sum total of attributes. To untangle the argument let us abbreviate "real existence" by the scholastic Latinism esse, which, like most scholastics (except Thomists), Śri Harsha considers an attribute.

Boldface argument: What he says in the boldface passage is that cause=(esse,...), where I list the constituting attributes in parentheses, and we can not have esse=cause because then esse=(esse,...). And this is not allowed because a thing can not be its own attribute ("the absurdity of something (real existence) residing partially in itself").

"Recognized principle": Then it gets murky. Śri Harsha allows that the thing, which is the cause, is not the esse itself but something more: thing=(esse,...). We also have cause=(esse,...), and I see no problem with thing=cause=(esse,...). But he does. For some reason the same esse can not underlie the thing both as a thing and as a cause, or rather it has to underlie it/them twice, thing=cause=(esse,esse,...) ("It would therefore be necessary to assume another existence as residing in the cause qualified by existence"). But according to the "recognized principle" nothing can be an attribute of the same thing twice ("no more... can it reside in that of which it already is an essential attribute").

Accidental esse: After that it gets even murkier. Apparently, we are now envisioning non-essential properties, perhaps what scholastics called accidentals, so we have to refine the notation. I will list accidentals after the vertical line. So we have something like thing=cause=(esse,...| esse,...), and while esse can not underlie the thing as a thing and the thing as a cause both essentially, it can apparently underlie one essentially, and the other, say the cause, only accidentally. That is not good enough though, "but then the cause would have to be regarded as ‘not really existing'".

Multiple esses: Finally, the splitting of esse is allowed, say into esseT and esseC (actually, it seems that it was already done previously, and we tacitly had thing=cause=(esseT,...| esseC,...), but ok). Now we can have thing=cause=(esseT, esseC,...), which again I see no problem with. But Sri Harsha thinks that to get it to "really" exist we need to "unify" the two esses. So we get something like thing=cause=(esseT, esseC, esseTC...), and then an infinite regress. This does not deter him in itself, but even taking the "long step" of creating the requisite esses all at once does not work for another reason. Because the esse was supposed to be the "generic conception" of existence, and so had to unify all existences including all our specialized esses, which are thereby revealed as not esses at all, "seeking to establish the notion of existence you thus have lost the basis of it".

Summary: Try as I might, I seem to infer circularity at a key juncture of the argument. To apply the "recognized principle" Śri Harsha must claim (and appears to) that the same single esse can not be the attribute of the thing and the cause, both as such. But on his interpretation of real existence as an attribute this just rephrases the conclusion. Namely, that nothing really existent can be a cause, which is what he set out to prove. The rest of the argument does not seem to make up for it.

Answer 2

If the causation of something is what makes it real, then the thing is not real until the causation process is completed. At the same time, the causation process is not real without the thing, as then it would not be a causation process, as it would have nothing to cause. So the causation exists only as an attribute of the thing caused, and not as any independent thing.

But the causation process itself has attributes, among them, the thing it is supposed to be causing. But the value of that attribute cannot actually be the as-yet-nonexistent thing to be caused, it must be some special version of or reference to that thing, with a subordinate less-real status of existing.

So this leads to a hierarchy of degrees of subordinate existence. The thing being caused does not yet exist, yet it must exist already for the causation itself to be well-defined in any way, so it exists 'subordinately', and the cause exists subordinately to it, and so forth, and so on.

So if there is any existence involved, there are infinitely many separate layers of existence. Unfortunately, this is a descending tower, we cannot find the most basic way in which either the causation process itself or the thing it results in would exist. So how can any of this pile of causations get started?

So if causation exists, it is not a real thing, only an attribute of the caused thing. And thus all causation is the causation of the real by the unreal.