the second hundred · metaphor 154
Some things you fail at because you didn't try hard enough. Others you can never reach at all — not for lack of effort, but because the moves you're permitted simply don't compose into that destination. How do you tell an unfinished climb from an impossible one?
His goal was unreachable with the tools he was permitted. Not difficult — unreachable, the way an odd number is unreachable if you may only take even-sized steps. He could search forever, combine his allowed moves in every order, work through the night, and never once land on it. The barrier was not in his stamina. It was a fact about the structure of the moves themselves.
This is the hardest kind of limit to accept, because effort is our answer to everything, and effort here is beside the point. The mathematician Évariste Galois, dead at twenty, gave us the tools to prove that certain targets lie outside the reach of certain operations forever — that some doors are not locked but absent. Below is an honest, hands-on shadow of that idea.
The idea
Fix a starting point and a set of permitted moves. The reachable set is everywhere you can get by chaining those moves — no matter how many, in any order. Some destinations are inside it; the rest are outside it forever. The line between the two has nothing to do with how hard you look. It is decided the moment you fix the moves.
What proves a target unreachable is a conserved invariant: a quantity that every permitted move leaves unchanged. If you may only move along diagonals, each step changes x + y by an even number, so the parity of x + y never budges. Start on an even cell and you are trapped on the even-coloured half of the board — the other colour shares none of your parity, and no search, however exhaustive, can flip it. The invariant is a wall you can prove exists without ever walking to it.
This is the shape of Galois' great discoveries, made concrete. Trisecting an arbitrary angle with compass and straightedge, doubling the cube, squaring the circle, solving the general fifth-degree equation by radicals — each is a target that the permitted operations provably cannot reach, guarded not by difficulty but by an invariant of structure. The tools generate a certain closed world, and the goal lives outside it.
What to try
The shading is a genuine breadth-first search running in your browser: from the origin it spreads outward, but only along the moves you've permitted, one ring of the frontier at a time. Nothing is pre-coloured — a cell lights up exactly when a real chain of permitted moves reaches it. Press Flood and watch the reachable set fill. The red target sits on the wrong side of the invariant, and the flood washes right up to it and around it, forever leaving it dark.
Click any cell to drop the target somewhere new: cells that share your parity light up green when the flood arrives; cells on the other colour read NO no matter how long you wait. Then press grant an extra tool. One new move is added to the permitted set — enough to break the invariant — and suddenly the whole board is reachable and the target turns green. Nothing about your effort changed. The toolset did. That is the entire lesson: impossibility here is a property of the permitted moves, and the cure is a new move, never a harder search.
The mapping
Read the permitted moves as the means you actually have — the rules you must obey, the resources you were handed, the moves your position permits. The reachable set is then everything those means can ever produce, however long you grind. And the conserved invariant is the structural fact that quietly bars a goal: a parity, a symmetry, a conserved quantity that no permitted move can touch, so that any target on the wrong side of it stays out of reach by construction.
The uncomfortable, clarifying move is to ask, of a goal you keep failing at, which kind of failure it is. Is the target merely far — reachable in principle, waiting on more effort — or is it across an invariant your permitted moves conserve, so that effort is the wrong currency entirely? A rigged process, a credential you're structurally barred from, a contradiction built into the rules: these are Galois-unreachable, and the honest response is not to try harder but to change what is permitted — to acquire the move that breaks the invariant. Naming a limit as structural is not defeat; it's the redirection of effort from a wall to a door.
Read as life lessons
"Is it possible?" is the wrong question; possible with what? is the right one. Reachability is always relative to a permitted set. Name the moves you're actually allowed before you decide a goal is out of reach.
A conserved quantity is a wall no search can climb. When something every allowed move preserves separates you from the goal, more attempts only re-confirm the barrier. Look for the invariant before you look for stamina.
Structural impossibility dissolves the instant the toolset changes — grant one move and the barred becomes trivial. If a goal is truly out of reach, spend your energy acquiring a new permission, not repeating an old one.
In the wild
Trisecting a general angle, doubling the cube, and squaring the circle are all impossible with compass and straightedge — proven, not merely unsolved. The permitted constructions can't reach the numbers those goals require.
The general quintic has no solution by radicals — Abel and Galois showed its symmetry group isn't "solvable," an invariant no formula in roots can overcome. Some equations are unreachable by the tools of √ and ∛.
The halting problem and other undecidable questions are the same shape in another key: no algorithm, however clever or long-running, can decide them. The barrier is structural, not a matter of faster machines.
The mapping, exactly
| Mathematics | Life |
|---|---|
| the permitted moves | The means you actually have — the rules you must obey, the resources handed to you, the moves your position permits. |
| the reachable set | Everything those means can ever produce, in any order, however long you keep at it. |
| a conserved invariant | The structural fact that bars the goal — a parity or symmetry every permitted move must preserve, whatever you do. |
| target outside the set | A goal impossible in principle for these tools — not far, not hard, but on the wrong side of an invariant. |
| breadth-first searching | Trying harder, longer, more exhaustively — effort, which sweeps the reachable set but never crosses an invariant. |
| granting an extra move | Changing what's permitted — a new tool, credential, or rule that breaks the invariant and makes the barred reachable. |
The honest model
Be explicit: this widget demonstrates reachability on a lattice, which is an honest stand-in for Galois' criterion, not the criterion itself. From the origin, a real breadth-first search applies only the permitted move-vectors and shades every cell it reaches (computed on a grid larger than the one you see, so the verdict for visible cells is exact). A cell's colour is decided by whether a genuine chain of moves lands on it — never assigned in advance.
The proof that the target can't be reached is a conserved invariant: every diagonal move changes x + y by ±2 or 0, so (x + y) mod 2 is fixed for all time, on a board of any size — the flood you see is just a slice of an infinite fact. The full Galois story is deeper: a polynomial is solvable by radicals exactly when its Galois group is a solvable group, and constructibility asks whether a number lies in a tower of degree-2 field extensions. Those invariants are subtler than a parity — but the skeleton is the same: permitted operations generate a closed set, and an invariant proves some targets lie outside it forever.
Where the metaphor tears
My toy makes the wall obvious — a single, visible invariant does all the work. Galois' real content lives where no simple conserved quantity is available, and the impossibility hinges on the fine structure of a symmetry group being unsolvable. The lattice makes barred goals look more legible than they usually are; real structural limits can be genuinely hard to even recognise.
Every claim here is with these moves. Change what's permitted and the impossibility can vanish — which is exactly why "grant a tool" unlocks the board. In life the permitted set is rarely as fixed as compass-and-straightedge; treating a contingent limit as an eternal law is its own error, the mirror image of trying too hard.
Some goals are barred by structure — but "Galois-unsolvable" can also become a comfortable excuse for not trying at things that are merely hard. The whole discipline is telling the two apart: a real invariant you can name and prove, versus a wall you've assumed. Mislabel a difficult climb as impossible and you quit a door that was only stiff.