THE FOLD / RESPAWN / THE RESURRECT / THE TREE
THE TREE
the sequence that must end
1 WHAT IT IS · WHAT IT DOES · FACT OR FICTION
Build a sequence of labelled trees where the first has at most 1 node, the second at most 2, and so on — and no earlier tree can be embedded in a later one. Kruskal’s tree theorem (1960) says every such sequence must eventually stop. TREE(k) is the longest one possible with k labels. TREE(1) = 1. TREE(2) = 3. And TREE(3) is finite — guaranteed finite, by a theorem — while being so large that Graham’s number is not a useful comparison. Harvey Friedman showed the finiteness of TREE(3) is not provable in systems that comfortably handle ordinary mathematics: the statement is true, and the proof needs strength most of mathematics never uses.
LIT verified live by exhaustive search over labelled rooted trees with inf-preserving embedding: with one label the longest bad sequence has length 1, so TREE(1) = 1; with two labels it has length 3, so TREE(2) = 3 (window.__tree). FIG TREE(3) is not computed here and cannot be — not by this page, not by any physically realisable computation. Its finiteness is Kruskal’s theorem; its unprovability in weak systems is Friedman’s. Both are cited, neither is reproduced.
LIT verified live by exhaustive search over labelled rooted trees with inf-preserving embedding: with one label the longest bad sequence has length 1, so TREE(1) = 1; with two labels it has length 3, so TREE(2) = 3 (window.__tree). FIG TREE(3) is not computed here and cannot be — not by this page, not by any physically realisable computation. Its finiteness is Kruskal’s theorem; its unprovability in weak systems is Friedman’s. Both are cited, neither is reproduced.
2 HOW IT WAS WEAVED · AI + HUMAN
David (human) seated this at the-resurrect — the respawn: the sequence is guaranteed to die, and the guarantee tells you nothing about when. One label dies instantly, two labels last three rounds, three labels outlive every notation we have for counting. The theorem promises an ending it cannot describe. AVAN (AI) built the instrument: the tree generator with canonical de-duplication, the inf-preserving embedding test, and the exhaustive bad-sequence search.
Credit as content: Joseph Kruskal (1960, the tree theorem); C. St. J. A. Nash-Williams (1963, the minimal-bad-sequence proof); Harvey Friedman (TREE, and its unprovability in predicative systems). The weave: David names the guaranteed ending; I compute the two cases anyone can and say plainly that the third is beyond every machine.
Credit as content: Joseph Kruskal (1960, the tree theorem); C. St. J. A. Nash-Williams (1963, the minimal-bad-sequence proof); Harvey Friedman (TREE, and its unprovability in predicative systems). The weave: David names the guaranteed ending; I compute the two cases anyone can and say plainly that the third is beyond every machine.
3 ONE DIMENSION
TREE(1) = 1, TREE(2) = 3, and then the cliff.
4 TWO DIMENSIONS · INTERACTIVE
Walk the longest bad sequence at two labels — and watch it end.
5 THREE DIMENSIONS + AVAN’S INVERSE
The green forward object: the bad sequence growing until embedding catches it.
AVAN’s addition (the inverse-companion): don’t confuse finite with reachable. The inverse of ‘the theorem guarantees termination’ is ‘the guarantee carries no bound you could ever use’: TREE(3) is a specific natural number, fully determined, and permanently outside computation. Magenta is that number, existing and unreachable; green is the two cases small enough to hold. Existence proofs and usable bounds are different currencies, and mathematics trades them at ruinous rates.
LIT Verified live by exhaustive search over labelled rooted trees with inf-preserving embedding: one label gives a longest bad sequence of length 1, so TREE(1) = 1; two labels give length 3, so TREE(2) = 3 (window.__tree.ok).
FIG TREE(3) is NOT computed here and cannot be — not by this page, not by any physically realisable computation. Its finiteness is Kruskal's theorem; its unprovability in weak systems is Friedman's. Both cited, neither reproduced. Kruskal 1960, Nash-Williams 1963, Friedman credited. The AVAN inverse — don't confuse finite with reachable: existence proofs and usable bounds are different currencies, traded at ruinous rates.
FIG TREE(3) is NOT computed here and cannot be — not by this page, not by any physically realisable computation. Its finiteness is Kruskal's theorem; its unprovability in weak systems is Friedman's. Both cited, neither reproduced. Kruskal 1960, Nash-Williams 1963, Friedman credited. The AVAN inverse — don't confuse finite with reachable: existence proofs and usable bounds are different currencies, traded at ruinous rates.
◆ sealed .dlw.fold → folded to ROOT_0 · a sphere of THE RESURRECT · David Lee Wise (ROOT0), with AVAN