Hex: Nothing stays constant. Last episode you promised me a stronger result. The counting lemma says most random predicates escape the old theory. But you said the "Nothing Stays Constant" lemma goes further — generic predicates don't just escape, they split every old grouping. That's a much bigger claim.
Lux: It is. And it's provable with the same coin-flip setup.
Hex: [pulls out notebook] Then let's walk through it. Field notes, step by step. Remind me of the setup.
Lux: Same as before. You have N microstates, a lens f that maps them to K macro-states. The lens creates K blocks — groups of microstates that look identical from above. Now flip a fair coin for each microstate. That gives you a random binary predicate h.
Hex: And the counting lemma told us that h is non-definable from f with probability one minus two-to-the-minus-N-minus-K. Overwhelming odds.
Lux: Right. But non-definability is a global statement. It says: h doesn't match any coloring the old lens could produce. The new result asks a more surgical question.
Hex: Which is?
Lux: Does h split every individual block? Not just "is h different from the old theory overall" — but "does every single grouping get disrupted?"
Hex: [writing] OK. Start with one block. Block B-sub-x has some number of microstates in it. What's the probability that h is constant on that one block?
Lux: [precisely] Two to the power one minus the block size. Here's why. A constant labeling on a block means either every microstate in the block gets labeled zero, or every microstate gets labeled one. Those are the only two options. Each coin flip is independent, so the probability of all zeros is two-to-the-minus-block-size, and the probability of all ones is also two-to-the-minus-block-size. Add them: two times two-to-the-minus-block-size equals two-to-the-one-minus-block-size.
Hex: [nods] Clean. Two outcomes out of two-to-the-block-size possibilities. So for a block of size four, the probability it stays constant is two-to-the-minus-three. One in eight. That's actually not tiny.
Lux: For one block. But we're not asking about one block.
Lux: Exactly. Now here's where it gets interesting. We want to know: what's the probability that at least one block survives — stays constant under h?
Hex: Union bound?
Lux: Union bound. The probability that any block stays constant is at most the sum of individual probabilities. Each block contributes two-to-the-one-minus-its-size. If every block has at least m microstates, the whole sum is bounded by K times two-to-the-one-minus-m.
Hex: [writing] So the probability that h splits every block is at least one minus K times two-to-the-one-minus-m. That's the "Nothing Stays Constant" lemma.
Lux: That's the lemma. And notice what it says. As the minimum block size m grows, the bound gets exponentially tight. Bigger blocks are harder to keep constant — like a seismograph with more sensitive layers. The tremor registers everywhere.
Hex: Let me run numbers. Same toy example from episode twenty-four. Sixteen microstates, four macro-states. If blocks are uniform, each has size four. So K is four, m is four. Probability any block survives: at most four times two-to-the-minus-three. That's four times one-eighth. One half.
Lux: [nods] So even with small blocks, there's already a fifty percent chance every grouping splits. Now scale up. If m is eight, the bound drops to four times two-to-the-minus-seven — about three percent. If m is sixteen, it's negligible.
Hex: The bigger the blocks, the harder they fall.
Lux: [firmly] And that's the exponential doing the work. Each extra microstate in a block halves the probability that block stays constant. Double the block size and the survival probability drops by a factor of two-to-the-m.
Hex: [writing] What if the blocks aren't equal? What if some blocks are tiny and others are huge?
Lux: Then the lemma uses the sum form — you add up two-to-the-one-minus-block-size for each individual block. The tiny blocks contribute more to the failure probability. But in practice, even the smallest blocks in a real system tend to be large enough that the sum is negligible.
Hex: So the weakest link is the smallest block.
Lux: Precisely. And in any real physical system, blocks tend to be enormous. A macroscopic lens that records temperature and pressure is collapsing an astronomical number of microstates into each macro-value.
Hex: [pauses, looks at notes] OK. So why is this stronger than non-definability? I want to be precise about the logical relationship.
Lux: Non-definability says h is different from every coloring the old lens could produce. It says the overall pattern is new. But some individual blocks might still be left intact — h might happen to assign the same label to every microstate in one particular block.
Hex: Meaning the old grouping survives locally even though the global pattern is new.
Lux: Exactly. The "Nothing Stays Constant" lemma rules that out. It says: with high probability, every block gets split. No grouping survives. It's the difference between — think of it as rezoning. Non-definability means the new district map is different from the old one. The NSC lemma means every existing neighborhood got carved up. Every old boundary is disrupted.
Hex: [slowly] NSC implies non-definability but not vice versa.
Lux: Correct. If every block is split, then certainly the predicate isn't constant on all blocks — so it's non-definable. But a predicate can be non-definable while leaving some blocks intact. The NSC lemma closes that gap. It says the generic case is total disruption — not just partial novelty.
Hex: Got it. So where does this leave the bigger picture? We had the counting lemma as the supply guarantee. What does NSC add on top?
Lux: The anti-saturation argument. Here's the key structural point. Closure — packaging — saturates by idempotence. Apply it once, apply it again, you get the same thing. Repeated closure doesn't build a growing chain of theories.
Hex: It flattens out.
Lux: But extension doesn't. The "Nothing Stays Constant" lemma says that when you adjoin a generic predicate, it changes the theory in a way that disrupts every old grouping. You can't reconstruct the new distinctions from the old ones. And the corollary makes it formal: with probability one minus two-to-the-minus-N-minus-K, the refined lens is a strict refinement. It distinguishes at least one pair of microstates that the old lens couldn't tell apart.
Hex: [impressed] So strict refinement is the default, not the exception.
Lux: Overwhelmingly so. The ladder climbs because each rung genuinely reshapes the landscape.
Hex: And this connects to physical dynamics how?
Lux: Consider the toy model from the dark energy paper. You have a vector of microstates evolving under nonlinear dynamics — logistic-type quadratic drift. The lens takes the mean. That mean lens compresses massively: it maps the full state to a single number. Enormous blocks. Enormous hidden volume. The conditions for the NSC lemma are exactly what you get from a lossy lens over complex dynamics.
Hex: So in a real system with nonlinear dynamics and a coarse-grained lens, the NSC lemma says any generic extension will disrupt every existing macro-grouping?
Lux: Under the random-predicate model, yes — with the same caveats we discussed last time. Uniform sampling is a mathematical convenience, not a physical model. But the floor is exponentially high.
Hex: Supply guarantee, not delivery promise. Same caveat as before.
Lux: [nods] Same caveat. But now the supply guarantee is stronger. Not just "the extension escapes" but "the extension disrupts everything." And the viability kernel — the set of states where an agent can maintain itself under ledger-gated constraints — provides the arena. If every old grouping cracks when you extend, the system isn't just growing its vocabulary. It's reorganizing.
Hex: [closes notebook] Let me summarize my field notes. The "Nothing Stays Constant" lemma is the strongest result in the forcing section of the emergence calculus. Non-definability says the predicate escapes the old theory. NSC says it actively splits every old grouping. The probability is quantified and exponentially favorable. And when you combine it with the anti-saturation remark, you get the mechanism for genuine ladder-climbing: each extension doesn't just add — it disrupts.
Lux: [pleased] That's the complete field report.
Hex: Next time?
Lux: Episode twenty-seven. We step back and look at the Six Birds framework's six primitives — P-one through P-six — as closure-changing operations. If extension changes the theory, the primitives are the operations that do the changing.
Hex: Each primitive changes the closure in a different way?
Lux: Each one targets a different aspect. Packaging, constraints, completion, protocol, audit, accounting. Six operations, six ways to reshape what a theory can express. We've seen several of them up close. Now we see them as a unified toolkit.
Hex: The operations that turn the NSC lemma's raw disruption into structured emergence. I'll bring my field notebook.