Exploring mathematics: a classroom journey from proofs to arithmetic geometry

The basket on the table

The classroom door opens.

The reader has a seat near the window. Five children sit in a half-circle:

Ana    asks, “Can we try it?”
Malik  asks, “How do we know?”
Noor   asks, “What pattern repeats?”
Jo     asks, “What connects these worlds?”
Lina   asks, “Could this fail?”

I carry in a basket.

I put one apple on the table.

🍎

Then I put down one more.

🍎 + 🍎

“How many apples?” I ask.

“Two,” says Ana.

I write:

“Good. But today an answer is only the beginning.”

Malik leans forward. “What else is there?”

I draw four doors:

🎯 answer       where did we arrive?
🖼️ picture       why does it look right?
📜 proof        why must it be right?
🤖 checker      can another process check it?

“These doors are different,” I say. “A correct answer is not automatically a proof.”

The first question.

Ana calls out, “Four!”

I do not laugh. I put a red card beside four.

4  ❌

“A wrong guess is useful,” I say. “It tells us one thing that the answer is not.”

Noor tries three.

We place one counting circle beside each apple:

🍎 → ○
🍎 → ○

There are two circles, not three.

3  ❌

Malik asks, “If someone guesses two next, have we proved it?”

“No. A lucky guess reaches the answer. A proof gives a route that must work.”

The class repeats the first lesson:

finding the answer  ≠  proving the answer

The box of possible answers

Noor raises her hand.

“What if we make a complete list, then test every item?”

“That is a proof idea,” I say. “What must the list contain?”

“Every possible answer,” says Noor. “Nothing important may be missing.”

The two apples make a small counting game. The only candidates we need to inspect are:

I draw a box around them:

D = [ 0 | 1 | 2 ]

“D is just the box’s name,” I say. “The braces list what is inside.”

The checker tries each number.

0  ❌  no circle can match both apples
1  ❌  one apple is left over
2  ✅  every apple gets one circle

The test is fair:

every apple gets exactly one circle
every circle gets exactly one apple
nothing is left over
nothing is used twice

“What did the checker actually show?” asks Malik.

I write the tiny certificate:

D = {0, 1, 2}
0 fails
1 fails
2 passes

“It showed that two works and that no other candidate in this complete box works.”

The exact sentence is:

We read it slowly:

∃! n in D
there is exactly one candidate n in D

V(n) = 1
the checker says that candidate passes

The symbol V is the checker. The number 1 means pass. The number 0 means fail.

The class has now proved a bounded counting statement. It has not yet proved the result from the rules of arithmetic. That will be another door.

When the symbols ask for names

Noor points at:

“What does that say?”

“It has three useful voices,” I answer.

standard voice    n is an element of D
friendly voice    n is one thing in the box D
exact voice       n is one object collected by D

“People also say ‘n belongs to D’ or ‘n is in D.’ Those are common readings. They are not the only readings.”

Lina asks, “Does it mean that n is a subset of D?”

“No. That is a different question.”

n ∈ D       one object is in a collection
A ⊆ D       every object in one collection is in another

“The word ‘within’ is fine for a quick picture,” I say, “but it can hide the difference between one object and a whole collection. In mathematics, element of and subset of are safer names.”

Two roads between statements.

Malik writes:

“The standard reading is: P implies Q.”

“A second reading is: if P, then Q.”

“The meaning is: every allowed case in which P is true must also make Q true.”

P true + Q false = the one forbidden case

“The arrow does not by itself say that P causes Q. It says that the truth of P guarantees the truth of Q.”

Jo adds the return road:

“The standard reading is: P if and only if Q.”

“You may also say: P exactly when Q.”

The class draws both directions:

P ──implies──> Q
P <──implies── Q

“One road is not two roads,” I say. “That is why ⇒ and ↔ must not be confused.”

A map arrow is a different arrow.

Jo draws:

“This one is not a logical implication. It says that f is a function from A to B.”

each input in A  ──f──>  exactly one output in B

“The arrow tells us where the function may take inputs and outputs. It does not promise that every object in B is reached.”

The class keeps a small reading card:

∈       one object is an element of a collection
⊆       every element of one collection is in another
⇒       implies
↔       exactly when, in both directions
→       maps from one place to another
∀       for every
∃       there exists at least one
∃!      there exists exactly one

“The standard phrase is useful,” I say. “The meaning is more important than the phrase. When two phrases are equivalent, we can teach both.”

The staircase named Peano

I place a green tile on the floor.

🟢

“This is zero.”

Then I place one tile after it.

0 → 1 → 2 → 3 → 4 → ...

“The next-tile button is called S, for successor.”

1 = S(0)
2 = S(S(0))
3 = S(S(S(0)))

Ana walks two steps.

“That is two,” she says.

“Now we give the plus button two rules.”

“Adding no steps changes nothing.”

“Adding one next-step to the second pile puts one next-step around the whole answer.”

Now the proof is tiny:

“Each line is allowed by a public rule,” I say. “That is why this is a proof.”

Malik compares the two routes:

🍎 + 🍎                  1 + 1 = 2
picture                  rule-by-rule rewrite
good explanation         formal proof

The Peano rulebook.

The staircase needs promises:

0 is a natural number
every natural number has a next number
no next number is 0
different numbers have different next numbers

The last promise says the next-step button does not squash two different tiles together.

Induction is the staircase’s reusable proof:

✅ property works at zero
✅ whenever it works at n, it works at S(n)
--------------------------------------------
✅ it works at every natural number

“Checking four tiles is not induction,” Lina says.

“Correct. Induction proves the reusable step for an arbitrary tile.”

Two arithmetic rooms.

PA, Peano arithmetic
the natural-number rulebook
addition + multiplication + induction

Presburger arithmetic
the natural-number room with addition
no multiplication of two changing numbers

Presburger arithmetic has a decision machine that can finish every statement in its language. Once multiplication of two variables is allowed, no algorithm can correctly answer every statement built from those operations, equality, and “for every” or “there exists.”

The proof hallway

The class walks down a hallway. Each door has one proof shape.

➡️ direct          follow the licensed path
🁢 induction       first tile + reusable next-step rule
🗂️ cases           split a complete list of possibilities
🚫 contradiction   assume the opposite, reach an impossibility
🧰 construction    build the object, then test it
🔍 exhaustive      check every item in a finite box

“These are methods, not magic words,” I say. “A method is only a proof when its promises are met.”

The class’s formal-methods lesson is now visible:

answer       endpoint
explanation  picture or reason
proof        finite public route
checker      independent replay

The teacher hides the answer

The next morning I turn the board around. There is no pile of apples.

There is only a covered box and a question:

“This time,” I say, “x is an integer. Find it, then prove that your answer is the only one.”

Ana says, “Nine.”

“How do you know?”

She writes:

Now Malik checks the answer:

The class has done two jobs:

✅ existence    9 works
✅ uniqueness   no other x can work

“The first line finds a candidate,” I say. “The second line checks it. The reversible steps show why no rival can survive.”

This is a better proof prompt than “What is x?”:

find x
check x
prove only x works

The parity question.

I write:

“Is an even number plus an odd number always odd?”

Noor tests examples:

2 + 3 = 5
4 + 7 = 11

“Examples are clues,” I say. “The question says always, so we need a proof for every allowed pair.”

We give the words exact shapes:

even number = 2a
odd number  = 2b + 1

Now the proof fits on one line:

The answer still has the form “twice something, plus one.” Therefore it is odd.

Lina nods. “We did not check every pair. We showed that every pair has the same shape.”

The impossible question.

I write:

“Can an integer or natural number satisfy this?”

Ana tries zero. Then one. Then a very large number.

“Trying numbers could go forever,” Malik says.

“So suppose one works,” I answer:

That is impossible in the ordinary integer and natural-number systems. Therefore no allowed x works.

assume a solution
↓
derive an impossibility
↓
no solution exists

The forever question.

I draw an endless row of tiles:

1, 3, 5, 7, 9, ...

“Prove that the first n odd numbers always add to n².”

The claim is:

Noor proves the first tile:

Then she assumes the claim works at k. The next odd number is 2k+1:

So the truth moves from k to k+1. The first tile and the reusable step cover every tile.

first tile ✅
reusable next-step ✅
all tiles ✅

The counterexample question.

Lina writes:

“Every odd number is prime.”

“Disprove it,” I say.

“Nine,” she answers:

9 is odd
9 = 3 × 3
9 is not prime

One valid counterexample defeats a statement that claims “every.”

The students make a proof-making card:

“find”       may need a construction
“only”       needs uniqueness
“every”      needs a general proof
“cannot”     needs contradiction or an invariant
“disprove”   needs one counterexample

The word invariant means a feature that does not change while the allowed moves happen. It will become useful when the class studies puzzles, symmetries, and algebraic maps.

The algebra machine

The next morning I roll in a machine with an empty basket and buttons.

“What must we describe before we can use the machine?” I ask.

The children answer:

🧺 what objects may enter?
🔘 what buttons may we press?
📜 what promises must the buttons keep?
↔️ which translations preserve the promises?

The first three make a structure. The last one leads to the map room later.

One button, six worlds.

The blank combine button is ⊙. It is a placeholder, not multiplication.

Set.

🧺 {🍎, 🚗, 7, 🐈}

A set is a collection. It has no button yet.

Magma.

Add one two-input button:

The output stays in the basket. That is closure.

Semigroup.

Add the rule that regrouping does not change the answer:

“The parentheses may move,” says Noor.

Monoid.

Add a do-nothing object:

Examples:

0 for addition
1 for multiplication
empty word for joining words
do nothing for composing actions

Group.

Add an undo button for every object:

action + undo = do nothing

Abelian group.

Add the swap promise:

“Abelian” is the name for a group whose combine order does not matter.

The ladder is:

set
 ↓ add a button
magma
 ↓ regrouping is safe
semigroup
 ↓ add do-nothing
monoid
 ↓ add undo
group
 ↓ add swapping
abelian group

Two buttons cooperate.

Now the machine has addition and multiplication.

➕ addition
× multiplication

The multiplication button must distribute across addition:

Semiring.

Natural numbers are the child’s first example:

0, 1, 2, 3, ...
add ✅
multiply ✅
subtract inside the world ❌

A semiring, in the convention used here, has addition, multiplication, zero, one, and distributivity. It does not require negative numbers.

Ring.

Add an opposite for every addend:

5 + (-5) = 0

A ring, in the convention used here, has an additive group, a multiplicative identity, and a compatible multiplication. Multiplication does not have to commute.

Commutative ring.

Multiplication can swap:

The integers are a commutative ring.

Integral domain.

A commutative ring with no zero made by multiplying two nonzero things:

Field.

A commutative ring in which every nonzero multiplication can be undone:

semiring       add and multiply
ring           also subtract
domain         no nonzero × nonzero = 0
field          divide by every nonzero element

The arrows show a common ladder, not the whole mathematical forest. Some properties branch sideways.

Matrix city.

Jo brings in a grid:

“A square matrix is a rectangle of numbers,” I say. “Square matrices of the same size can be added and multiplied.”

The identity matrix acts like one:

But order can matter:

for many matrices.

“So matrices make a ring,” says Jo, “but usually not a commutative ring.”

“Exactly. A matrix ring remembers order.”

Side doors from the ring room.

Not every useful structure fits one ladder.

vector space  vectors scaled by field elements
module        vector-like objects scaled by ring elements
algebra       a module that also multiplies inside itself
ideal         a ring-room region stable under ring multiplication
lattice       objects with a common-up and common-down operation
Boolean       logic with AND, OR, and NOT
graph         objects remembered only by connections
topology      nearness remembered without exact distance

Module.

ring scalar × module object → module object

A vector space is a module whose scalars come from a field. A module lets algebra work even when division is unavailable.

Ideal.

An ideal is a special subcollection of a ring. It contains zero, is closed under addition and additive inverses, and stays inside the subcollection when any ring element multiplies one of its members. In symbols, for an ideal I of a ring R:

The additive rule matters: a pile can be stable under multiplication and still fail to be an ideal if it is not closed under addition. Ideals let us make quotient rings, which are rings where selected differences count as zero.

Boolean algebra.

The logic costume:

∧  AND
∨  OR
¬  NOT
0  false
1  true

The ring costume:

x² = x
× means AND
+ means XOR

The same pattern can wear two costumes.

Universal algebra.

Lina opens a recipe book.

“Could we study all these machines at once?”

“Yes. Universal algebra asks only:

what objects?
what operations?
what equations?

Groups, rings, lattices, and Boolean algebras are different recipes in the same recipe book.

The map room

Jo opens a second door. Behind it are baskets connected by arrows.

“A structure tells us what happens inside one world,” Jo says. “A map tells us how to translate one world into another.”

Homomorphism.

A homomorphism is a map that keeps a button honest. It preserves the operation:

The picture is:

combine, then translate
        =
translate, then combine

For rings, a ring homomorphism preserves addition and multiplication. For groups, a group homomorphism preserves the group operation.

Category.

The map room keeps two things:

objects  the worlds
arrows   the legal maps

Arrows can compose. If f is followed by g, the combined route is written g∘f. The order of composing three arrows does not matter:

A ──f──> B ──g──> C
A ──────g∘f─────> C

Every object has a do-nothing arrow:

A ──id_A──> A

That is a category. Its arrows also have do-nothing laws and associative composition. It may contain sets and functions, groups and homomorphisms, rings and ring homomorphisms, or many other kinds of objects.

Functor.

A functor translates a whole map room into another map room:

object A  ──F──>  object F(A)
arrow f    ──F──>  arrow F(f)

It preserves:

The forgetful functor is the child’s first functor:

field
 ↓ forget division
ring
 ↓ forget multiplication
abelian group
 ↓ forget the operation
set

Nothing is destroyed in the objects. The translator simply stops paying attention to some rules.

Natural transformation.

Suppose two functors, F and G, translate the same room in two ways. For a map f:A→B, the comparison looks like this:

F(A) ──F(f)──> F(B)
  │ η_A          │ η_B
  ↓              ↓
G(A) ──G(f)──> G(B)

A natural transformation is a coherent set of little arrows such as η_A and η_B comparing the two translations. “Natural” means the square agrees no matter which route is taken.

Child compression:

homomorphism             map between structures
functor                  map between map rooms
natural transformation   map between functors

The universal doorway.

Sometimes an object is important because every other valid construction reaches it in one unique way.

many possible routes
          ↓
one route that is forced

That is a universal property. It describes an object by the maps it receives or sends, rather than by listing all its internal furniture.

An adjunction is a paired promise:

build in one direction
      ⇄
forget or test in the other direction

The free group on a set is a classic example. It adds the least group structure needed to accept a function from the set.

Yoneda’s question.

Yoneda asks:

How does this object interact with every other object?

If two objects have exactly the same map behavior, category theory can identify them up to isomorphism. The object is understood through its relationships.

The equation telescope

The classroom lights dim. I open a telescope marked:

equations + numbers + maps

“This is arithmetic geometry,” I say.

An equation makes a landscape.

Start with:

Over the real numbers, the solutions form a circle:

real solutions  →  a smooth round shape

Now ask for rational solutions:

The equation is the same. The allowed number world changed.

same equation
different number world
different question

Algebraic geometry studies shapes cut out by equations. Arithmetic geometry asks what those shapes do over number systems such as the integers, rationals, finite fields, and local fields. A local field is a number system designed to study one prime at a time.

The elliptic curve station.

The guide writes:

Over the real numbers or rational numbers, the usual nonsingularity condition is:

when it holds, the curve has no sharp self-crossing in this form.

Two rational points can be combined by drawing a line:

point P + point Q
      ↓ draw the line
third intersection
      ↓ reflect
new point P + Q

With a chosen point at infinity as zero, the rational points form a group. The equation has become an algebraic structure.

“So geometry can hide a group inside a shape?” asks Ana.

“Yes. That is one of the great bridges.”

Diophantine questions.

A Diophantine question asks for solutions in a restricted number world:

integer solutions       x,y ∈ ℤ
rational solutions      x,y ∈ ℚ
mod-p solutions          x,y ∈ 𝔽_p

The same curve may have:

many real points
few rational points
different points modulo each prime

The restriction is part of the problem. “Has a solution” is incomplete until the number world is named.

The prime microscope.

Take a prime number p. Reduce the equation modulo p. The equation gets a finite-field view:

integer equation
       ↓ look through prime p
finite equation

Different primes reveal different shadows. Some shadows are smooth. Some collide or become singular. Those exceptional primes carry information about the original equation.

Schemes, told simply.

Lina asks, “Why not just collect all the points?”

“Because a point is not the whole story,” I say. “A scheme remembers:

the visible point
the functions near the point
the arithmetic that can vanish there

Prime ideals act like arithmetic lenses. A scheme keeps the global equation and its local neighborhoods together.

This is not merely a bigger set of dots. It is a space whose points carry local algebra.”

Fibers and base change.

Suppose a family of equations is drawn over a number line:

total family
      ↓ choose a number system
fiber over that system

Choosing the real numbers, the rationals, or the field with p elements gives a different fiber. Base change means carrying the same family into a new number world and watching its geometry change.

Galois symmetry.

An equation may have roots that are not visible in the starting number world. Galois symmetries move the hidden roots while preserving every rational relation between them.

hidden roots  🎭
symmetries    move them
rational facts stay fixed

Galois theory turns “how roots move together” into a group.

Local and global views.

Arithmetic geometers compare:

global view    the whole rational or integer problem
local view     what happens near one prime
finite view    what happens modulo p

The local views can reveal obstructions to a global solution. A solution that survives every local test may still fail globally, so the tests are evidence, not an automatic guarantee.

The board at the end of the lesson

The children cover the board with four routes.

Route 1: prove.

picture → question → rule → certificate → checker

Route 2: build a structure.

objects
  + operations
  + laws
  = algebraic structure

Route 3: translate structures.

structure → preserving map
map room  → functor
functors  → natural transformation

Route 4: study equations arithmetically.

equation → shape
shape + number world → arithmetic question
prime lenses + local views + symmetries → arithmetic geometry

Malik asks the final question:

“What did we really learn?”

I answer:

An answer tells where we landed.
A picture lets us see the landing.
A proof gives a public path.
A checker replays the path.

A structure keeps chosen rules.
A map keeps chosen structure.
A functor keeps whole map rooms.
An equation becomes geometry when its solutions are studied as a space.
Arithmetic geometry studies those spaces through number worlds,
prime lenses, local neighborhoods, and symmetries.

Noor points at the first apple.

“And one plus one?”

The class answers:

🍎 + 🍎 = 🍎🍎

1 + 1
= S(0) + S(0)
= S(S(0) + 0)
= S(S(0))
= 2

I close the rulebook.

“The deepest mathematics did not replace the apple. It explained how the apple, the number machine, the algebra machine, the map room, and the equation telescope can belong to one connected story.”