My understanding of Benacerraf's identification problem breaks this problem into two subquestions. There is the format issue (the one that leads to the possible "junk theorems") and then there is the "first number" issue, i.e. whether to start from 0, 1, or maybe (with Brouwer?) 2, or maybe (with Christian philosophers like Cornelius van Til, who thought that the Trinity conceptually resolved the one-over-many problem) 3, or...
Now, in the Tao Te Ching there is a passage that reads:
The Tao gives birth to One. One gives birth to Two. Two gives birth to Three. Three gives birth to all things.
I read an article about Plato's Form of the Good that interpreted this as some kind of Form of Unity, something that seems to make Plato into a neo-Platonist before the fact, so to speak. That there could be a concept of unity that was not reducible to the number 1 was also indicated to me by the following passage from Kant's first Critique:
But the conception of conjunction includes, besides the conception of the manifold and of the synthesis of it, that of the unity of it also. Conjunction is the representation of the synthetical unity of the manifold. This idea of unity, therefore, cannot arise out of that of conjunction; much rather does that idea, by combining itself with the representation of the manifold, render the conception of conjunction possible. This unity, which a priori precedes all conceptions of conjunction, is not the category of unity (SS 6); for all the categories are based upon logical functions of judgement, and in these functions we already have conjunction, and consequently unity of given conceptions. It is therefore evident that the category of unity presupposes conjunction. We must therefore look still higher for this unity (as qualitative, SS 8), in that, namely, which contains the ground of the unity of diverse conceptions in judgements, the ground, consequently, of the possibility of the existence of the understanding, even in regard to its logical use.
Although while writing my essay about "deontic set theory," I had started out with an intro that tried to assimilate Plato to the tradition of philosophers who have tried to ground ethics in mathematics (as the first of these philosophers, even), it occurred to me a little while later that something very much else seems to have been going on: that Plato thought that mathematics could be grounded in ethics instead.
This would be cashed out in terms of saying that Plato and the Lao Tzu persona who wrote the Tao Te Ching, can be thought of as implicitly believing in a specific response to the second part of the identification problem for the natural numbers. The Tao itself as the "first number" for Lao Tzu is often mapped to emptiness and hence the number zero; but Plato is mapped to the number 1 as the "first number"; and yet the way in which the Tao gives birth to the One, or in which the Form of the Good transcreates the other Forms, makes it seem, to me, as if, for Plato and Lao Tzu, neither zero nor the One are the first numbers, but something else is.♡
Aside from being rather anachronistic (though in my defense, I would say that adapting the Tao to the empty set, or the Form of the Good to abstract monadicity, are already anachronistic interpretive maneuvers), how fair is this interpretation?
♡Some theoretical serendipity was involved on my end, here: for various reasons I had already foregone having zero be the first natural number in deontic set theory, but I was leery of having 1 play the role instead. I also wanted 1 to still be a set, though. So I axiomatized the following: ∃xy((x ∈ y) & ∀z((z ~= x) → (z ~∈ y)) & ∀z((z ~= y) → (x ~∈ z)); in other words, there exists an x that is (a) only an element of some y and (b) the only element of y. Let's call this u (with an eye towards its eventual name in English). To round things out, ∀x(x ~∈ u), i.e. u has no elements. So far, then, u might be a set, but empty. So finally add in ∀x(u ~∋ ~x), so u is not a set of no elements, either, so is not a set at all. Presto, u is an ur-element, and in fact here is called the ur-element, and the set with which it is so essentially correlated, is the number 1. So I had the template for the Tao/Form of the Good as the "first number" in place, and all that remained was to code u for a unique Boolean-styled value in the logic of my system, which was J as the fundamental justification value for the Boolean-style justification logic (as one level of deontic (tran)set theory).
The exciting immediate consequence was an albeit extremely slight resolution of the formatting side of Benacerraf's implementation question. In deontic (tran)set theory, since u is only an element of 1, it is not an element of 2, and hence a formatting of the natural numbers that produced a junk theorem according to which u is an element of 2, is ruled out. So even if we were to commence with von Neumann format, we could not do so until after we had the number 2 in Zermelo format instead. That is, we would have u, {u}, {{u}}, and only then, perhaps, {{u}, {{u}}}. "Sadly," this does mean that 2 is no longer the first ordered pair in the system, but 3 is. Perhaps that fact is grounds for revising the implementation further, i.e. to get a system in which either 2 is still an ordered pair somehow (I don't see how) or 3 also is not the first ordered pair, but maybe 4 is. At any rate, the broader attempt in the system, to address the identification/implementation problem, is to assign a strictly infinitesimal positive justification value to at least the elements of one countably infinite set of Benacerraf-theoretic notation styles (for the natural numbers in set theory), maybe the set of so-called "eventually writable ordinals" or at least some other large countable ordinal series of that "kind." Then deontic (tran)set theory's own proper implementation is supposed to sum over all those justification values. (Having said that those are "strictly" infinitesimal in scale, I mean to have said that I was using Conway's surreal infinitesimals, here. So let's say Zermelo format has n/ω-much justification, von Neumann format has m/ω-much justification, etc. Then "deontic format" has a justification value that is (n + m + ...)/ω. But n + m + ... = ω, after all, so the "deontic format" has a justification value of ω/ω = 1 (surreally), and we are adapting the unit interval's alethological role to the deontic scene and saying that 0 = unjustification and 1 = completely sufficient justification (though we allow x < 0 and > 1, and speak also of antijustification and hyperjustification). I am suspicious of this derivation (it's almost an exact copy of a major part of the derivation of CH in the system), but otherwise it seems like a promising way to paint the deontic picture.)