How to trust what you read here
This project was written, intentionally, with the help of large language models, with me holding up a mirror at them at every obsessive iteration. But a tool that powerful used in this silly little manner is also the surest way for an author to fool one's self.
The danger is real and specific: earlier language models had a way of handing you back your own conviction dressed up as evidence, and it was easy to get pulled in. People got swept up that way, mistaking a model's agreement for the world's, and an idea can feel proven long before it is. This page tells you the rules I use to keep from doing that.
Or, as Mark Twain might say:
And subsequent to all foregoing declarations, provisos, caveats, and other ceremonial throat-clearings, be it known that this undertaking is an independent affair, animated by a central idea of uncommon vigor, and presently engaged in the hazardous pastime of reenacting the very self-trickery heretofore identified, notarized, and lamented.
1.The theory was built one careful step at a time
The original hypothesis, vaguely based on autocatalysis and cognitive mirroring, was formulated over decades. The work of putting it into form using a large language model began in early spring of 2025.
The seed of it was a single thought: that some things carry a pressure to return, to close back on themselves and reappear. I was sure this was a matter of nonlinear dynamics, but I could find no existing word in that field for what I was seeing. So I had to invent one. I started with recurcline, a portmanteau of recursion and incline, in the sense of a gradient, a tendency to lean back toward a shape. That was the initial seed, and everything grew from there.
The recording of the theory (Principia Attractum) did not arrive all at once. It was assembled slowly, one small addition at a time, and no addition was allowed to stay until it passed two checks. First, it had to be coherent with everything already in the theory: a new piece that contradicted an existing one was not admitted. Second, it had to be coherent with established empirical science: a piece that conflicted with what is already well understood in physics, chemistry, or biology did not survive either. Every increment faced both checks, every time. That work took the author (me) roughly ten hours a week for thirteen months, accreting the theory piece by piece into a single recorded whole. It sounds like the most tedious chore in the world, but it turned out to be the opposite. The constraints were guardrails, and guardrails are what let you put the pedal to the metal. They gave a direction that needed no second-guessing: the lanes were empty, the highway was wide open, and the work felt like rocket fuel. It was the thing that got me out of bed in the morning. Principia Attractum is that accretion.
2.The theory and the experiments are kept apart
There are two halves to this site, and they are not equal.
- Principia Attractum is the theory: the ideas, the vocabulary, and the rules they follow. This is the settled part.
- The Bio-Kernel Series is the workshop: small simulations and digital ecosystems built to test those ideas. This is the experimental part, and it is allowed to be wrong.
3.Looking stable is not the same as being real
The one question the workshop keeps asking is this: when a shape holds together on screen, is it holding itself together, or is the program underneath quietly holding it up? Those look identical from the outside, and the difference is the whole point. So I make the impressive answer earn it:
ASSUME THE BORING ANSWER Until proven otherwise, a structure that persists is assumed to be propped up by the program, not sustaining itself.
NOT PROOF "It stayed on screen." A pattern lasting a long time tells you nothing, by itself, about whether it is self-sustaining.
Before I'll call something genuinely self-sustaining, it has to pass a real test: knock it off balance, and it has to recover on its own steam rather than being reset by the program. Until it clears that bar, I won't describe it as self-sustaining, autonomous, or alive. Those words are earned, not assigned.
4.What I did with the failures
There were many failures along the way, and I documented them as I went. The honest problem is what became of that record: it is an enormous amount of data, I have no repository set up to hold it, and much of it is now lost. A good deal could be reconstructed, but it has not been, and I will not pretend the full trail is available when it is not.
One failure is documented clearly, and it is the one that matters most. The entire Bio-Kernel Series was rebuilt from scratch after I admitted the first version had a flaw: it ran on a single shared clock that was quietly feeding the simulations the very order they were supposed to be discovering on their own. That admission is why the rebuild is worth anything. If you never see a project's failures, you have no reason to believe its successes.
I am deciding whether to build a proper repository for the rest of this material, both the failures and the results that appeared to confirm the hypothesis. If that would be useful to anyone, that interest would help me decide. You can reach me at the address below.
The Bio-Kernel Series is the workshop side of this project: small cellular-automata worlds and digital ecosystems built to test one thing: whether a structure that looks like it is surviving on its own actually is, or is just being held up by the simulation around it.
A note, and a request. The author is unfunded and works alone. Verifying citations against primary sources matters a great deal to me, and I am doing my best at it, but I cannot keep buying paywalled PDFs one at a time just to confirm a single reference. If anyone in the research community is willing to help with access or verification, that support would be genuinely appreciated.