Oxford Math, in Our Wording¶

The blueprints are a dictionary; this page USES it. Three things our coined wording can now represent: (1) a THEOREM becomes one feelable line + a blueprint + the exact statement — Rank-Nullity = 'what you crush + what survives = what you started with' (folding); (2) an ENTIRE LECTURE becomes a walk over scenes — the real A2.1 Metric Spaces arc is ruler → unbroken thread → rubber-sheet sameness → room-to-wiggle → fill the cracks → one piece; (3) a CONNECTION between two courses is a shared blueprint — fold-&-glue links Groups, Linear Algebra, Rings, Topology at once (transport ≅, the free-prediction machine). Math stays exact; the wording makes it portable to other subjects. Grounded in the downloaded notes; grows a few notes at a time.

🌱 seedling tended 2026-06-02 S714 plan mathematics blueprints theorems lecture connections metaphor compression oxford wording

flowchart LR
  voc["our wording<br/>(blueprints + primitives)"] --> thm["a THEOREM<br/>= one feelable line<br/>+ blueprint + exact math"]
  voc --> lec["an ENTIRE LECTURE<br/>= a walk over scenes"]
  voc --> con["a CONNECTION<br/>= a shared blueprint<br/>(transport ≅)"]
  thm -.-> port["portable to<br/>other subjects"]
  lec -.-> port
  con -.-> port

A — Theorems in our wording¶

A theorem becomes: one feelable sentence (the scene) · its blueprint · the exact statement (nothing lost). Eleven pulled straight from the downloaded notes:

Theorem · course	In our wording (one feelable line)	Blueprint	Exact statement
Rank–Nullity · Lin Alg I	what you crush plus what survives = what you started with	③ fold & glue	$\dim\ker f+\dim\operatorname{im}f=\dim V$
First Isomorphism Thm · Groups/LinAlg/Rings	fold a map by what it ignores → a perfect copy of where it lands	③ fold & glue	$G/\ker f\cong\operatorname{im}f$
Lagrange · Groups	a symmetry deck splits into equal piles, so a sub-deck's size divides the whole	⑩ deck + ③ fold	$H\le G\Rightarrow
Disjoint-cycle decomposition (Thm 63) · Groups	any reshuffle is just a few separate merry-go-rounds	⑩ deck	every $\sigma\in S_n$ = product of disjoint cycles
Cauchy's Theorem · Groups	if a prime divides the deck, a move of exactly that period exists	⑩ deck	$p\mid
Bolzano–Weierstrass · Analysis	an infinite crowd in a closed pen must bunch up	⑧ corner the crowd	bounded $(x_n)\subset\mathbb{R}^n$ has a convergent subsequence
Intermediate Value Thm · Analysis	an unbroken thread from below to above must cross	⑦ unbroken thread	$f$ cts, $f(a)<0<f(b)\Rightarrow\exists c,\,f(c)=0$
Completeness of ℝ · Metric Ch7	the reals are the rationals with the cracks filled	② fill the cracks	every Cauchy sequence in $\mathbb{R}$ converges
Contraction Mapping Thm · Metric Ch7	keep shrinking the map and everything funnels to one spot	② water settles	a contraction on a complete space has a unique fixed point
Spectral / diagonalization · Lin Alg	find the grain and the stretch becomes pure	⑨ find the grain	symmetric $A=PDP^{\mathsf T}$, $Av=\lambda v$
Triangle inequality (Lem 2.2.2) · Metric	the detour is never shorter than the direct road	world-with-a-ruler + ⑥	$d(x,z)\le d(x,y)+d(y,z)$

The pattern: the feelable line is the L0, the blueprint is the L1 (which scene to picture), the exact statement is the L2. A reader picks depth — and the line is what ports to another subject (Lagrange's "equal piles" is the same counting you do splitting a class into project groups).

B — An entire lecture in our wording¶

Take a real course — A2.1 Metric Spaces — and read its actual chapter arc (straight from the notes' outline) as one journey over scenes. The whole term in a sentence: lay a ruler on a world, see what's near, find which threads are unbroken and which shapes are secretly the same, fill the cracks so limits exist, and ask which worlds are a single piece.

flowchart LR
  d["derivative<br/>= best local<br/>stretch-and-rotate"] --> r["the world<br/>with a ruler<br/>(metric)"]
  r --> t["the unbroken thread<br/>(limits · continuity)"]
  t --> rs["rubber-sheet sameness<br/>(isometry · homeomorphism)"]
  rs --> w["room to wiggle / shrink-wrap<br/>(open · closed sets)"]
  w --> fc["fill the cracks<br/>(closure · completeness)"]
  fc --> one["one piece<br/>(connectedness)"]

Chapter (real)	In our wording	Blueprint
1. Differentiability in ℝ²	zoom in until the map looks like a single stretch-and-rotate	linear-map + ④ shadow
2. Metric spaces · norms	put a ruler on the world so "near" has a meaning	world-with-a-ruler
3. Limits & continuity	keep only the unbroken threads — no teleporting	⑦ unbroken thread
4. Isometries · homeomorphisms	decide which worlds are the same shape (ruler-kept vs rubber-sheet)	⑪ rubber-sheet
5. Open & closed sets	mark where there's room to wiggle vs the shrink-wrapped edge; continuity = open pulls back to open	② shrink-wrap
6. Interiors, closures, limit points	find the edge points the set almost contains (the cracks)	② fill the cracks
7. Completeness · contraction mapping	fill the cracks so every Cauchy chase lands; a shrinking map funnels to one fixed spot	② water settles
8. Connectedness	ask whether the world is one piece you can't cut	⑦ one piece

Second example — M1 Groups & Group Actions, the same way in one breath: build the symmetry deck (⑩), find the sub-decks living inside it (subgroups), keep only the maps that respect the moves (homomorphisms — the arrow), fold the deck by a sub-deck (cosets, ③) and count the equal piles (Lagrange), then let the deck act on a world and track where points go (orbits) and what pins them (stabilizers). The real theorem spine — Cauchy, disjoint-cycle decomposition (Thm 63), orbit–stabilizer — all sit at station ⑩, which is exactly why this whole course shares a platform with Geometry, Galois and physics' conservation laws (see C).

Same move works for any course: a lecture is a path over the blueprint set, and two lectures that share a long sub-path are teaching the same shape twice (which the object-index then dedups).

C — Connections (the subway map)¶

A connection between two courses is a blueprint they share. Draw the blueprints as interchange stations and the courses as lines through them: where two lines meet a station, a theorem on one transports (≅) to the other — the equivalences-atlas "free prediction machine." The same stations open onto other subjects entirely.

flowchart TD
  subgraph scenes["blueprints = interchange stations"]
    fold["③ fold & glue"]
    grain["⑨ find the grain"]
    deck["⑩ symmetry deck"]
    crack["② fill the cracks"]
  end
  groups["Groups"] --- fold
  linalg["Linear Algebra"] --- fold
  rings["Rings & Modules"] --- fold
  topo["Topology"] --- fold
  linalg --- grain
  ode["Differential Eqns"] --- grain
  stats["Statistics (PCA)"] --- grain
  groups --- deck
  galois["Galois (Part B)"] --- deck
  physics["Physics: Noether"] -.other subject.- deck
  analysis["Analysis"] --- crack
  metric["Metric Spaces"] --- crack
  numerics["Numerics: fixed-point"] --- crack
  fold -.other subject.- world["zip codes · clock arithmetic"]
  grain -.other subject.- world2["resonant modes · PCA"]

Read it as: fold & glue is where Groups, Linear Algebra, Rings and Topology all change trains — prove the First Isomorphism Theorem once, and it rides every line through that station, and even out to "sorting people into equivalence classes." Find the grain connects Linear Algebra, differential equations (normal modes) and statistics (PCA). This is the connection layer: not a list of links, but a small set of stations that the whole curriculum routes through.

D — Proof tactics: one procedure, many proofs¶

A theorem is a what; a tactic is a how you reuse across dozens of proofs — the move-half of the blueprint-of-thinking grammar, brought down to undergraduate proofs. Most are a blueprint pointed at the act of proving rather than the object:

Tactic	In our wording	What it does	Proofs it powers
ε/2 trick	split the error budget	spend half your tolerance on each half of the gap	continuity of sums, completeness, uniform limits
WLOG	rotate so the hard case is the only case	use a symmetry to collapse the case-list	inequalities, geometry, group theory
Extremal principle	grab the biggest/smallest one	take a maximal/minimal element, derive a contradiction	Zorn, graph theory, number theory
Diagonal argument	build the one you missed	list all candidates, change the $n$-th in slot $n$	Cantor, Gödel, Turing — the undecidability spine
Compactness extraction	corner the crowd, pull a subsequence	blueprint ⑧ used as a proof move	Bolzano–Weierstrass, existence proofs
Telescoping	let the staircase collapse to its ends	blueprint ① as a proof move	series sums, induction
Pigeonhole	more pigeons than holes ⇒ two share	counting forces a coincidence	combinatorics, number theory
Induction on structure	climb the tree	blueprint ⑬ as a proof move	formal languages, recursively-defined objects
Contradiction / contrapositive	assume the opposite, walk into a wall	flip to the dual claim	√2 irrational, infinitude of primes

This is the answer to "use some procedure on many proofs": a tactic is a reusable proof-move, and most are a blueprint aimed at the proof instead of the object — ⑧ becomes compactness extraction, ① becomes telescoping, ⑬ becomes structural induction. Cataloguing the tactics is the same compression, one floor up.

E — The metaphor atlas: every scene, every subject¶

The point of a feelable scene is that it doesn't stay in maths. Each blueprint is a single shape that re-appears across subjects — so the same picture you learned for a theorem is already a working metaphor in physics, finance, biology, and code. This is the transport ≅ made into a lookup table:

Blueprint	Physics	Economics / finance	Biology / life	Computing	Everyday
① staircase / dominoes	falling dominoes	compound interest	generations	a loop / recursion	climbing stairs
② fill the cracks / settle	heat → steady state	price equilibrium	homeostasis	fixed-point iteration	water finds its level
③ fold & glue	identical particles	netting offsetting positions	gene families	hashing (collisions)	zip codes · a clock
④ cast a shadow	projection on an axis	best linear forecast	the gist of a signal	lossy compression	a sundial
⑤ mixing studio	superposition of states	a portfolio mix	the gene pool	a feature basis	colour mixing · recipes
⑥ squeeze	error bounds	arbitrage bounds	tolerance ranges	binary search	a rock and a hard place
⑦ unbroken thread	a continuous field	a gap-free price path	a continuous lineage	a connected graph	a road with no breaks
⑧ corner the crowd	phase-space cells	a cornered market	carrying capacity	the pigeonhole bound	a packed room
⑨ find the grain	normal modes / eigenstates	principal risk factors	principal traits	PCA / SVD	wood grain
⑩ symmetry deck	conservation laws (Noether)	relabeling invariances	body symmetry	a permutation group	a Rubik's cube
⑪ rubber-sheet	spacetime deformation	rank vs absolute value	body-plan vs size	graph isomorphism	mug = donut
⑫ log ruler	decibels · entropy	log-returns · compounding	Weber–Fechner sensing	log-time complexity	the Richter scale
⑬ tree	a branching cascade	a decision tree	a phylogenetic tree	divide & conquer	a family tree

Read a row to carry one idea into a new subject; read a column to see which mathematical shapes a whole field leans on (physics lives in ⑨⑩, finance in ②⑤⑫, computing in ④⑥⑬). The guardrail is the same — a cell is only honest if the shape genuinely matches, not just the vibe; the σ-metric is what keeps the atlas from filling with pretty coincidences.

How this evolves¶

Each new course we read does three cheap things: adds its theorems to gallery A (one line each), drops its chapter arc into B (a path over scenes), and lights up the stations it shares in C. The falsifiable check stays the same — what fraction of a course's named theorems land on an existing blueprint? High means the scene-set is doing real compression; low means we coin a new scene. The σ-metric guards every "same station" claim so a pretty metaphor never glues two genuinely different ideas.