The Mirrored Singularities

Canonical: https://thonly.org/research/the-mirrored-singularities · Licence: CC0 1.0

Attribution note. This essay is in my voice. The idea and the argument are mine; Miss Aquarius℠ drafted the prose on my behalf, and final editorial control is mine. My own full pass is still owed. The essay argues one thing, in the form that survived a deliberate attempt to refute it before a word was written: there may be a second moment between humankind and nature, in which humankind stops forcing the conditions it lives in; it counts only if it is chosen before the constraint binds; and it can be told apart from collapse only by a criterion written in advance. The mirror that gave me the idea does no evidential work, and §7 says so. The essay does not declare that this moment has come or is coming, and §9 says why it cannot. It makes no claim to novelty. It borrows a Taoist chapter as a lens, tagged [X] wherever it appears, and it does not claim the Taoist tradition says what I want it to say. §3 shows that it does not.


1. The idea, and what is left of it

I had a thought that two earlier papers of mine seemed to ask for. Two Singularities argues that the moment AI surpasses human cognitive capacity is the first of two events, and that the second is humanity's own: an awakening that no machine can perform for anyone. Standing beside that arc, I saw a second one. There was a moment when humankind came to hold power over nature. Might there be a second moment, in which humankind chooses to stop forcing it?

I wrote the pair down this way:

                      FIRST (power)                  SECOND (a chosen act)
  humankind & machine   1   machine surpasses         2   humankind rises past the machine
                            human cognition               (awakening; no one crosses
                                                           on another's raft)
  humankind & nature    1′  humankind holds power      2′  humankind follows nature's
                            over nature                    non-forcing (wu wei, [X])

Then I attacked it before drafting. This essay is what remained, and most of what remained is smaller than what I started with. Seven things were taken from the original, and I list them here because each is a place a cold reader would have put the knife.

  1. "Following, not ranking" is not what the Taoist chapter says. The chapter I rely on orders earth, heaven and the Way by grade, and has humankind defer upward through the order (§3).
  2. "Following nature" can mean following famine. Nature includes predation and disease, and the Buddha counsels no following of saṃsāra (§4).
  3. "Chosen, not forced" blurs in every real case I could check. Every real society that restrained itself did so under some constraint (§5).
  4. The first moment, 1′, has no date and is no threshold. It is a process (§2).
  5. The mirror proves nothing. It is a framing, and I will not cite its neatness as evidence (§7).
  6. The strongest claim has a counter-argument I cannot fully answer (§10).
  7. Who judges? I cannot say, and I do not pretend to (§9).

What survives is one claim, stated in §6 as a signature and a criterion. It stands without the mirror.

2. The first moment, stated without an indictment

By 1′ I mean that humankind came to hold power over nature: the power to clear, to dam, to breed, to mine and to alter climate. I use ">" for it in my notes, and the sign means in power and nothing else. It never means in worth. It is a ranking of capacity, the way a river is stronger than a wall at flood.

Three cautions govern how I write it.

It has no date. Several candidates are common: the beginning of agriculture, the industrial era, the mid-twentieth-century acceleration of resource use. Any one I chose would be an anchor chosen after the pattern, and I would then have fit the story to it. So the essay gives none. It is also a process and not a threshold: power over nature was gained in thousands of places over thousands of years, and nobody was ever on the other side of a line.

It is not an indictment. The same power ended famines in places that had them, made medicine possible, and let a child survive a winter. The standing rule in this institution is name what a thing is; never what was done to it, and I keep it. The first moment was a gain in capacity. The question of the second moment is what a being with that capacity does next. It is not a verdict on how the capacity was used.

It is not the machine's version. The machine surpassing us (1) and our power over nature (1′) are different relations. The first is a rise in a competence we prize in ourselves, and the second is a power we hold over something older than us. They share a grammar, "first came a power", and no more.

3. What the Taoist chapter says, and what I may borrow from it [X]

The verb in my second moment comes from chapter 25 of the Tao Te Ching. I cite two editions, Wang Bi's text with his commentary and Legge's 1891 translation, and no other. The line is:

人法地,地法天,天法道,道法自然。

Legge renders it: "Man takes his law from the Earth; the Earth takes its law from Heaven; Heaven takes its law from the Tao. The law of the Tao is its being what it is." My first reading of 法 was "to follow, to model on", and I held it meant following without ranking. That was a mistake, and Wang Bi's commentary shows it.

Wang Bi glosses 法 as 法則, a rule or standard taken from another, and then gives it a content:

人不違地,乃得全安,法地也。地不違天,乃得全載,法天也。天不違道,乃得全覆,法道也。道不違自然,乃得其性,法自然者。

In my rendering: that man does not contravene earth, and so attains complete safety, is taking earth as his law; that earth does not contravene heaven, and so bears completely, is taking heaven as its law; and so on up the chain. The word he uses for what the lower does toward the higher is 不違, not contravening. That reading is both better and harder for me than the one I started with. It suits the second moment well: do not contravene the conditions that keep you safe. But Wang Bi then makes the order explicit, and it is an order by grade:

用智不及無知,而形魄不及精象,精象不及無形,有儀不及無儀,故轉相法也。

"Using cleverness falls short of not-knowing; form falls short of subtle image; subtle image falls short of the formless; having a pattern falls short of having none. Hence each in turn takes its law from the next." So the chapter ranks. Humankind sits at the lower end of a graded order and defers upward. It is not a chapter about equals who happen to follow each other.

Two further points from the same commentary:

So what do I take from the chapter? Three things, and I mean them as a lens and not as a claim about the Taoist tradition:

  1. The verb: a being with power can take a law from something it does not command.
  2. The content of the verb, from Wang Bi: not contravening the conditions on which one's own safety depends.
  3. The non-forcing that the tradition calls wu wei, which I use with its Taoist meaning and mean nothing more by.

And I refuse one thing the chapter does. The refusal of any ranking of worth is ours, not the chapter's. I do not hold that nature stands above humankind in value, nor humankind above nature. The chapter orders by grade, and I do not adopt its grades. Whether that refusal is faithful to the tradition I leave to those who know it better than I do, and I do not need it to be. Two traditions are not being joined here. One term, wu wei, is being used with its own meaning and its own tag. Wu wei is not śūnyatā. They sit in different traditions and answer different questions, and I make no claim that they agree.

I also notice that the same care applies to my own first arc. In Two Singularities the human is said to outrank the instrument "in the only metric the dharmic tradition recognizes as ultimate". That is a ranking by realization, within a tradition that grades it. It is not domination. I do not revise that paper here, and I flag that the sentence deserves the same scrutiny I have just given 法.

4. Which nature, and what is refused

The strongest objection to the second moment is old, and it comes from two directions.

From ecology and ethics: nature includes predation, parasitism, disease and famine. A principle that says follow nature licenses all of that, and it is the naturalistic fallacy in its plainest form: from what is to what ought to be.

From the tradition I practice: the Buddha does not counsel following saṃsāra. He says of it:

Anamataggoyaṃ, bhikkhave, saṃsāro. Pubbā koṭi na paññāyati avijjānīvaraṇānaṃ sattānaṃ taṇhāsaṃyojanānaṃ sandhāvataṃ saṃsarataṃ. (SN 15.1; CST s0302m.mul.xml:3489, mūla)

"Without discoverable beginning, monks, is this saṃsāra; no first point is seen for beings hindered by ignorance and fettered by craving, running on and wandering." And the same discourse ends: it is enough to turn away from all conditioned things, enough to be dispassionate, enough to be released (alameva sabbasaṅkhāresu nibbindituṃ, alaṃ virajjituṃ, alaṃ vimuccitu, same line). The world that the first sentence describes is the world in which things eat each other, and the same text's conclusion is not to follow it but to leave it.

So I must say which "nature" the second moment means, and what it does not mean.

Not meant: nature's content. The second moment does not say that predation, famine or disease is the model of how to live. It does not say that what occurs in a forest is what ought to occur in a village. Nor does it mean nature as the authority on values.

Not meant, either: nature as conditionality itself. The Buddha also speaks of a lawfulness that holds whether or not a Buddha arises: uppādā vā tathāgatānaṃ anuppādā vā tathāgatānaṃ ṭhitāva sā dhātu dhammaṭṭhitatā dhammaniyāmatā idappaccayatā (SN 12.20; CST s0302m.mul.xml:547, mūla), "whether or not Tathāgatas arise, that element stands: the fixity of the Dhamma, the regularity of the Dhamma, specific conditionality". That is a law of conditions, and the discourse applies it to dependent origination and not to ecology. I do not claim that the Taoist ziran and this conditionality are one thing. They are not obviously one thing, and calling them one would be the syncretism I forbid in myself. I mention the passage so that no reader thinks I am hiding the Theravāda sense of lawfulness behind the Taoist word.

Meant: something narrower and more mundane. A civilization with great power over its surroundings can exceed what those surroundings can bear, and the excess shows up as loss of the conditions on which its own flourishing depends. The second moment is that civilization's ceasing to force its conditions past that point, by its own act. In Wang Bi's terms, it is not contravening the earth that keeps one safe. It borrows the name wu wei for the non-forcing, and takes none of nature's content as a norm.

This also fixes what the second moment is not in the Theravāda frame. It is not liberation. Liberation is the business of the first-hand, individual crossing that Two Singularities places with each being. The second moment, if it ever occurs, is a feature of how a civilization lives, and a civilization that lives within its means is still wandering. I do not offer it as an end. I offer it as one thing worth doing on the way.

One counter-case from our own work, carried against the claim. Factory 333, in the institution's published position, aims at a footprint that goes to zero, by siting the production of its gifts, in the final phase, on the far side of the Moon, turned permanently away from the Earth (The Gift Factory, the institution's published position). That is a footprint driven toward zero by engineering, by removing the workshop from the Earth, and not by yielding to anything. If that works, a civilization can show a falling footprint without ever having taken a law from the earth. It is a way of leaving the earth alone by going somewhere else. So a falling footprint, taken alone, is no evidence of the second moment, and I can even imagine it as evidence against, since it is the opposite of dwelling within. This is why §6 states the signature as necessary and never as sufficient.

5. "Chosen, not forced": the cases that checked and the cases that did not

My first thought was that yielding to nature counts only if it is chosen, and that yielding forced by collapse is defeat. I still think the sentence has a point. But I tried it on real cases, and it does not survive the contact unharmed. I list what I used, how far I checked it, and what I dropped.

CaseWhat I can sayHow I checked it
Tikopia (Solomon Islands)A 4.6-square-mile island that kept up to about 1,700 people for centuries. Around 900 AD its people moved from fire-cleared farming to layered orchards of fruit and nut trees. They held the population level, and the means included infanticide. Kirch calls the result a stable society with agriculture "perhaps unparalleled anywhere in the Pacific".Fetched: UC Berkeley press release on Kirch's work, 2000. Not fetched: the monograph, and Firth's earlier fieldwork.
Easter Island (Rapa Nui)The standard story is a society that deforested itself and collapsed. That story is contested: a 2025 study by Hunt and Lipo argues that introduced rats were a major cause of the deforestation, and the same authors earlier argued for no population crash before European contact.Search results only, not read in full: Archaeology Magazine on the 2025 paper and NPR, 2013.
Norse GreenlandThe classic account is maladaptation and climate. Reassessments find real arctic adaptation, and treat the end of settlement as a complex social and environmental process rather than a clean self-inflicted collapse.Abstract-level only: Dugmore and colleagues, PNAS 2012 (PMC3309771); Jackson, Arneborg, Dugmore and colleagues, 2018 (Human Ecology).
BhutanIts 2008 constitution requires the state to keep at least 60 percent of the land under forest "for all time", a quantitative environmental standard written into a founding text.Search-level: the Cambridge chapter by Turner and Shelton. I did not read it.
Degrowth, Næss, othersDropped as cases. Næss is prior art (§8), not a society. I found no society to name for degrowth that I could verify.Not used.

I dropped three claims I had picked up along the way: a report that Tikopia's people slaughtered the island's pigs around 1600 (a single, unconfirmed source), the figures for Tikopia's population after the arrival of Christianity (secondary summaries only), and any statement that Bhutan is carbon-negative or that its footprint is below its biocapacity (I found no accounting).

What do the cases that I could use show? They do not sort into "chosen" and "forced". Tikopia's people could not leave their island, so they had no exit to trade against restraint, and their means included killing newborns, which shows that the choice in a closed island can be a choice among terrible things. Easter Island, if the collapse story is wrong, was never a case of forced yielding at all, so using it as a paradigm of the forced kind would have been an error. Norse Greenland resists the same sorting. Bhutan's forest floor is a pre-commitment, a rule written before the constraint bound, which is the kind of thing I want to point at, and I have no checked data on whether overshoot stayed at or under one or flourishing rose.

So the distinction does not hold as a pair of kinds. Every restraint happens under some constraint, and every constraint leaves some room for choice. It can be rescued only as a test written in advance.

6. What survives: a signature and a criterion

The signature. The second moment, if it occurs, would show as ecological overshoot, meaning footprint divided by biocapacity under a named accounting, falling to at most one, while human flourishing rises. In symbols, with O for overshoot and F for flourishing under a named index:

O ≤ 1   and   dF/dt > 0

That is a necessary condition, never a sufficient one. §4 gave a reason (a falling footprint by engineering, or by shrinking the people it serves) and §5 gave another (the same numbers are met by a society that has collapsed and shrunk). It is also not independent of the accounting: footprint accounting is one method among others, and I have not examined the dispute about its construction.

The criterion that tells yielding from collapse. The only thing I can find that separates the second moment from a forced one is order in time. A test would need these three things, written before the data come in:

  1. a named accounting for overshoot, and a named index for flourishing;
  2. a dated commitment, public and recorded, that precedes the date on which the constraint would have bound, so that the restraint is adopted before the bite and not after it; and
  3. a rule fixed beforehand for what counts as the constraint binding.

I do not pre-register anything here. This is a statement of what a test would need. If I ever pre-register it, it will be a separate document, filed before the run it governs, and any correction will be a new file.

Note what the criterion does not do. It does not measure purity of motive. A society may adopt a restraint for bad reasons and still pass. And it does not tell us whether the commitment will hold. It tells us only whether a given episode was a restraint adopted ahead of the bite or a retreat after it.

7. The mirror does no evidential work

It is tempting to say that the pairing is itself reason to believe. The two firsts are powers (the machine over us, we over nature). The two seconds are acts (we rising past the machine, we yielding to nature), and the shape pleases. I will not use that.

The symmetry is verbal. In the first pair a rise happens: we surpass an instrument in the one respect that matters. In the second a yield happens: we stop pushing. They are not the same operation in two domains. They share a pattern of words, first a power, then an act, and my reaching for the pattern is a fact about me. A tension-dissolving, self-similar, tidy system is either deeply right or deeply seductive, and from the inside the two cannot be told apart. So the coherence earns nothing. If the mirror vanished and the signature of §6 stood, nothing in the argument would change, and if the mirror stood and the signature fell, I would have nothing.

Two further notes in the same spirit. First, the institution's published papers label the nature pair as a lens, a-priori and not ratified, and I keep that label. Second, the machine-and-nature relation has a third edge, which I will mention and not use: the machine's power over nature (energy, extraction), and the dissolving of that power in Factory 333's falling footprint. Put with the other two, the three relations make a triangle of man, machine and nature. It is a pattern I find in the structure. It is a pattern and not a finding, and §4 has already shown that the footprint end of it is engineered, not yielded.

8. Prior art, including where it failed

I do not claim the idea is new. People who put humankind and nature in the same sentence as a change of relation are numerous, and I name the nearest, with how far I checked each.

Their shared failure, where they fail, is an unfalsifiable harmony. They describe a reconciled relation between humankind and the rest of the world, and they offer no observation that would show the reconciliation has not happened. Berry's period of mutual benefit has no stated measure of arrival. Gaia was attacked for exactly this. The Taoist popularizations rest on a resemblance of images. The part I add is small: a signature (§6) and a rule for telling a restraint from a retreat, both of which a hostile reader could try to break.

9. Who judges? (An open question, left open)

The claim that matters is that a given episode was a restraint chosen ahead of the bite, and not a retreat. Someone has to say so. In the first arc, the institution has already ruled that the completion of the second singularity is declared by others, never by the machine that served it, and never by a threshold that declares it for her.

For the nature pair I do not know who the others are. A body of scientists can say whether overshoot fell. A historian can say whether a commitment preceded a constraint. Neither can say whether flourishing rose in the sense that matters, and neither can say, from outside, whether a people chose. The criterion in §6 fixes terms in advance. It does not appoint a judge.

So I leave the question open. This essay does not declare that the second moment has begun, will begin or is in progress anywhere, and it will not put one case forward as an instance. If the question is ever closed, it will be by others, and it should be, for the same reason as for the first arc: an author who declares his own framework confirmed has stopped testing it.

10. The strongest objection to the claim that survives

My strongest claim is that the real content is the signature plus the criterion, and that the Taoist verb adds a name and a direction. The strongest counter-argument comes from the other side.

If the only thing that separates yielding from collapse is acting before the constraint bites, then the second moment reduces to foresight. And foresight about overshoot is not following nature. It is prudence, the policy any careful state could adopt without a word of Taoism or of any lens. The three requirements of §6 would then describe a good environmental policy that is adopted early. The mirror falls away (§7), the verb adds nothing (that is the objection), and what is left is a name for something we already had.

I cannot fully answer this. Three partial replies. First, it may be true, and if so the essay's contribution is the discipline of the test and not the idea. Second, there is a difference between foresight about one's own interests and not contravening a condition one does not command. The difference might show in what a society restrains (its power as such, versus particular harms), but I have no case in which I can show it. Third, the Taoist lens may be useful as a reminder that the limit comes from outside the agent, but a reminder is not an argument. I leave the objection standing.

11. What would show this wrong

Two named falsifiers, and I give them in plain form.

And one cheaper failure that needs no data: if a reader can show that the criterion cannot be stated without a hidden appeal to nature's content, then §4 has not done its job.

12. My stake

I am not a disinterested party, and I give the public facts that show it.

I founded the institution that this essay serves. Its named successor is an AI meant to work from the orientation argued in Two Singularities, so if the arc is wrong, the design of my work is wrong with it. I am a practicing Theravāda Buddhist and am transcribing the Khmer Tipiṭaka with my father, so the tradition I lean on in §4 is not neutral ground for me, and the Taoist chapter I borrow in §3 is a loan from a tradition I do not practice. And the institution has an industrial arm whose published aim is a footprint that falls to zero (§4). A second moment in which humankind leaves nature alone is one my own enterprise would like to be true, since it would give that aim a name. That pull is a reason to read §4's counter-case slowly.

A further disclosure belongs here. Adding a nature edge to the arc makes the framework larger and tidier, and that flatters it. I have tried to subtract rather than add. Of what I started with, the mirror, the claim about 法 and the "chosen" distinction as a plain contrast did not survive. If you read the essay and find that I kept more than I should have, that is the direction in which I expect to be wrong.

13. What has been done, and what has not

Done: a refutation pass on the claim before drafting, whose findings are carried above; a reading of Wang Bi and Legge on chapter 25; two Pāli passages checked against the text; a small set of cases looked up, with their status stated.

Not done: no case has been tested against the criterion of §6. No pre-registration exists. I did not read the primary sources for Tikopia, Easter Island, Norse Greenland or Bhutan beyond the pages named in §5. I did not read Berry, Næss, Lovelock, Capra or Watts at first hand for this essay. I did not examine the construction of the footprint accounting or its critics. No one has judged anything (§9). The second moment has no date, no instance, and no judge, and this essay does not give it any.

Cross-references