The foundational arithmetic library#
This part of the book is the readable front end to a native first-order Peano arithmetic library. It begins with ordinary equality and semiring laws, builds division, relational gcd, balanced Bézout, Gauss cancellation and Euclid’s lemma, then constructs finite factor sequences and their products using Gödel-β codes. The endpoint is a checked, β-coded Fundamental Theorem of Arithmetic.
Two cumulative editions
The Stable library contains 432 closed native theorems, including
factorization existence, extensional uniqueness, their combined FTA, and a
constructive theorem producing a prime above every supplied bound. The newest
Alpha quadratic-reciprocity campaign includes parity, constructive
residue decision, finite folds, factorial and power algebra, modular units,
sign and half-range bridges, β swap/reindex, finite pigeonhole, replacement
balance, and exact swap-last product invariance. Twenty-five strict-HA entries
now expose canonical remainder, congruence, bounded modular inverses, relational
gcd/LCM compatibility, LCM existence and uniqueness, and the gcd–LCM product
law. The exact 23-row generalized-CRT dependency closure is now Stable: it
covers the all-modulus solvability criterion, classification modulo a
relational LCM, the zero/nonzero canonical boundary, certified obstruction,
and raw-input total decision. New reviewed layers enter Alpha first and
move to Stable only after closure, compilation, dependency, resource, and
release audits. Sealed Alpha v2 adds seventeen body-checked K3C rows for
valid list codes, membership, unique in-range lookup, extensional code
equality, and unique outer-cell decomposition. Sealed Alpha v3 adds the first
twenty-one Bertrand rows, and sealed Alpha v4 adds forty-two Round-2 rows for
exact valuation multiplication, integer envelopes, ceiling/floor-square
arithmetic, and the quotient budget. Sealed Alpha v5 preserves that complete
965-row ledger and appends seven body-checked FactorialVal rows. Sealed
Alpha v5 therefore has 972 rows, sealed Alpha v6 has 993, sealed Alpha v7 has
1,017, sealed Alpha v8 has 1,055, sealed Alpha v9 has 1,076, sealed Alpha v10
has 1,085, and sealed Alpha v11 has 1,123. Historical Alpha v12 preserves the
entire v11 ledger and appends 180 body-checked Bertrand rows in nine exact
twenty-row microbatches. Alpha v13 adds the four-square and Lucas campaigns;
v14 adds Kummer’s carry theorem; v15 adds the supplementary laws and complete
two-square classification. Current immutable Alpha v16 preserves all 1,673
v15 statements while promoting exactly 315 independently proved
quadratic-reciprocity results, including the final theorem, to checked use.
Stable remains the unchanged default. See
Alpha and Stable library editions for the exact scopes and lifecycle.
The generated Stable snapshot has ordered root
4d02dc439d53533e8992a471b26ee34059fb6001f822041e42c56b2cc0a7a079.
Every entry is reconstructed from its authored script and checked from the
empty context. Names, summaries and hashes organize the library; none of them
grant proof authority. Its graph has 1,185 direct dependency edges; the
Book exposes 432 theorem cards, while the synchronized vault contains
531 notes and 5,377 resolved links.
The current Alpha v16 graph has 1,673 theorems, 5,615 direct edges, and 53
dependency layers. Its mixed evidence is intentional: 432 Stable-closed and
453 Alpha-closed rows have actual checked-use authority, while 788 body-only
rows remain visible without being treated as empty-context facts. No pending
row remains. The historical Alpha v12 graph has 1,303 theorems, 4,302 direct
edges and its immutable evidence partition still includes 732 body_checked
rows. Every Alpha v1–v15 parent remains sealed; Stable remains 432. The exact
contract and opt-in API are on the edition page.
Historical Alpha v12 retains the Alpha-v11 B4 and B5-support tranches, then
adds the dependency-closed B6 support and B5–BP02 completion chain. Its
frozen partition was 432 stable_closed, 138 alpha_closed, 732
body_checked, and one pending_layered_closure; checked use was 570. The
current v16 partition is 432 stable_closed, 453 alpha_closed, 788
body_checked, and zero pending rows. Only the completed
quadratic-reciprocity closure was promoted; Bertrand, Lucas, four-square,
Kummer, supplementary-law, and two-square body evidence has not acquired
checked-use or Stable authority.
The mathematical metro map#
The exact dependency graph has 432 vertices and 1,185 edges and is useful to machines, but a human first needs the stations. Each box below is a link into the guided tour.
The exact generated graph remains available as an immutable
dependency-graph.mmd.
The Stable theorem atlas gives a readable local
neighborhood instead of attempting to draw all 1,185 edges at once. The
Alpha QR proof explorer adds permanent theorem tags,
numbered tactic-line targets, and the larger quadratic-reciprocity campaign
slice. It is not the complete Alpha or Stable catalog.
Its parallel definition-aware edition renders
the same 557 specifications with a 40-entry conservative-definition registry
(38 definitions occur) and exact native replay lines; it does not change the
proof graph or admission status.
Its interactive dependency graph draws short or
critical premise chains, start-to-target corridors, and complete transitive
cones.
The complete Bertrand proof explorer applies
the same interface to all 544 nodes in the final strict theorem’s closure.
Choose your route#
If you want to… |
Begin here |
Then move to… |
|---|---|---|
understand the two release editions |
canonical counts, checked-use boundary, promotion lifecycle, and graph legend |
|
understand the mathematics |
the focused theorem links inside each stage |
|
inspect every Stable proof |
exact statements, complete scripts, dependencies and dependents |
|
follow the Alpha QR slice line by line |
permanent tags, linked lemma references, informal outlines and source receipts |
|
read expanded formulas through linked names |
persistent |
|
curate the next conservative edition |
P0/P1/P2 definitions, API completeness, and paired-source gates |
|
see how theorems depend on one another |
short and critical premise chains, route corridors and complete prerequisite cones |
|
understand soundness |
||
study division and congruence |
||
study encoded lists and lookup |
the compact Alpha/Stable graph, exact proof sources, and WMI receipt |
|
use list validity and membership |
the seventeen-row interface, exact body receipts, and append/restriction gate |
|
study primes and factorization |
the FTA and |
|
follow the reciprocity campaign |
parity, residue-decision, and finite-fold cards in the atlas |
|
follow the Bertrand campaign |
constructive interval search, valuations, the central-binomial route, and exact risk gates |
|
inspect the complete Bertrand proof |
all 544 nodes, exact tactic bodies, and the interactive 1,917-edge dependency map |
|
train a proof-producing model |
the snapshot, corpus and vault links below |
|
audit provenance |
catalog source mappings and the separate Lean companion |
What the native FTA says#
At the readable level, the endpoint is
Here \(F\) and \(G\) are not primitive lists. BetaAt decodes a bounded entry
from two natural-number codes, a second β code records prefix products, and
equality is extensional on the selected finite prefix. The exact statement is
ordinary first-order PA over 0, S, +, * and equality.
Native endpoint |
Nodes |
Depth |
Cuts |
|---|---|---|---|
|
43,973 |
98 |
1,328 |
|
29,789 |
82 |
854 |
|
73,767 |
99 |
2,184 |
The top-level authored proof is deliberately only three commands:
split
exact prime_factorization_existence
exact prime_factorization_uniqueness
That contrast is central. A short modular script is not a hidden axiom: its
dependencies are embedded as self-contained Cut nodes, producing the
73,767-node closed certificate checked by the kernel. Read the complete card
for fundamental_theorem_of_arithmetic.
Prime unboundedness is a separate constructive theorem#
FTA is not used to prove that there is a prime above every bound. The native proof obtains a nonzero common multiple \(c\) of every positive number through \(n\), takes a prime divisor \(p\) of \(S(c)\), and excludes \(p\le n\): such a \(p\) would divide both \(c\) and \(S(c)\), hence one.
The theorem prime_unbounded has 4,595 nodes, depth 82 and 146 Cuts. Its full
84-command authored proof is embedded in the
prime_unbounded atlas card.
The trust path#
readable tactic bodies
complete dependency proofs embedded
original closed formula
The checked FTA and prime_unbounded use only PA1–PA6 and ordinary induction;
their audits find no double-negation elimination. The object language gained
no primitive division, remainder, gcd, prime, list, product or factorization
symbol.
Four synchronized views#
One theorem name identifies the same object in four places:
the executable
TheoremSpecand checked certificate;the generated snapshot and dependency graph;
its card and narrative links in this Jupyter Book;
its atomic Obsidian vault note.
The current synchronized surfaces are:
There is exactly one deliberately unproved catalog boundary: the conventional integer-coefficient Bézout interface. Peano Lab quantifies only over naturals. The separately named balanced four-natural Bézout theorem is checked and is the interface used by Gauss cancellation and Euclid’s lemma.
Start reading#
Continue with the guided route. It alternates intuition, exact formulas, proof anatomy and immutable proof links. When you want to move backward or forward through the dependency DAG, open the theorem atlas and focus the theorem you are reading.