the second hundred · metaphor 209
Some things can be given and yet never truly duplicated — a trust, a first time, a confidence placed in you. You can pass them along, at a cost, but there is no making a second identical copy that leaves the original untouched.
Think of the trust someone hands you. You can carry it forward — vouch for a third person, extend it onward — but you cannot photocopy it: run it off twice and hold two intact originals. In the act of transferring it you spend it; the thing that arrives is subtly changed, and what you kept is not the same as what you had. A first experience is like this too. You can recount it, even give someone their own version — but you cannot mint a perfect duplicate of the unrepeatable original.
Quantum mechanics turns this intuition into a theorem. An unknown quantum state cannot be cloned — not by any machine, however clever. The proof is almost embarrassingly short: cloning would have to be a linear operation, and linearity, applied to a state that is a blend of possibilities, produces something entangled rather than two clean copies. Below is one honest copier you can watch fail, exactly where the theorem says it must.
The idea
Suppose a machine could copy. Feed it |0⟩ and a blank, out come two |0⟩s; feed it |1⟩, out come two |1⟩s. So far, easy — this is just classical copying, and the instrument shows it working at fidelity 1. The trouble starts with a state that is both at once: |ψ⟩ = c|0⟩ + s|1⟩. Quantum operations are linear, which means the machine must act on the blend by acting on each piece and adding the results — it cannot peek at c and s first.
Do the arithmetic and you get c|00⟩ + s|11⟩ — the two qubits chained together, entangled. But two genuine copies would be (c|0⟩+s|1⟩)⊗(c|0⟩+s|1⟩), which has cross-terms |01⟩ and |10⟩ that the linear output simply lacks. The two are not equal unless c or s is zero — that is, unless the state was classical to begin with. There is no adjustment, no cleverer machine: the very step that makes copying copying — being linear — is the step that breaks it for anything genuinely in superposition.
What to try
Start at |0⟩. On the left, the original arrow and its two copies point the same way; on the right, the output bars match the "two true copies" ghosts exactly; fidelity reads 1.000. The machine works. Now drag the state toward |+⟩, an even superposition. Nothing about the machine changes — but the original arrow tilts sideways while the two copy-arrows stay pinned to the vertical axis. That sideways lean is the quantum part of the state, and the copies have lost it: they are flat, classical shadows, not duplicates.
Read the fidelity as you go. It slides from 1 down to 0.5 at the exact halfway superposition — the copier's worst case — then climbs back to 1 as you reach |1⟩, the machine's other native state. The ledger spells out why, line by line, in live numbers: the output is forced to be c|00⟩ + s|11⟩, while two real copies need those missing cross-terms. The failure isn't noise or imperfection; it is exact, and it is worst precisely where the state is most quantum.
The mapping
The states that resist copying are the ones in superposition — held delicately in more than one possibility at once, their whole meaning in the blend. That is a fair picture of the things we most want to preserve and cannot: a trust that lives in a particular relationship; a first experience whose value is its unrepeatability; a confidence someone places in you, made of everything unspoken around it. You can transfer such a thing — vouch onward, tell the story, pass the torch — but the transfer is not a copy. The original is spent or altered in the giving, and what arrives is a flattened version, true in its populations but stripped of the coherence that made it that exact thing.
Notice what does copy cleanly: the classical states, |0⟩ and |1⟩. Facts, credentials, a phone number — the parts of a person that are already collapsed to a definite value — duplicate freely, which is why we can broadcast them without loss. It is only the superposed things, the ones held in tension and not yet resolved, that carry the no-cloning stamp. The more a thing's worth lives in that unresolved tension, the less of it survives being copied — and the metaphor's edge is that this is not a limit of our technology but, in the quantum case, a law.
Read as life lessons
You can hand a thing on and still not have made a copy. Passing forward moves it, and often spends it; two intact originals is the thing that can't be had. Giving and copying are different verbs.
Definite facts duplicate freely; it's the things held in superposition — a trust, a first time — that flatten when you try. What copies cleanly is exactly what mattered least.
The copier fails not from clumsiness but because it is linear — the very property that lets it copy at all. Some impossibilities aren't gaps in skill; they're baked into what the operation is.
In the wild
Quantum key distribution is secure because of this theorem: an eavesdropper can't copy the qubits in transit to read them later, and any attempt to measure leaves a detectable mark.
Proposed banknotes carrying unknown quantum states can't be counterfeited — no forger can duplicate the state — while the bank, knowing the recipe, can still verify them.
You can't back up a qubit the way you back up a file. Error correction has to protect quantum information without ever copying it — a constraint that shapes all of quantum computing.
The mapping, exactly
| Mathematics | Life |
|---|---|
| an unknown state |ψ⟩ | A trust, a first experience, a confidence — a thing whose value lives in its particular, unrepeated form. |
| a basis state |0⟩, |1⟩ | A plain fact or credential — already definite, and so freely broadcast without loss. |
| superposition | The delicate part held in unresolved tension, its meaning in the blend rather than any single value. |
| the linear copier | Any attempt to duplicate the thing wholesale — to run off a second intact original. |
| the entangled output | What you actually get: not two copies but a flattened, altered version, its coherence gone. |
| fidelity → 0.5 | How far short the "copy" falls — worst exactly where the thing was most itself. |
The honest model
The instrument holds a real single-qubit state |ψ⟩ = cos(θ/2)|0⟩ + sin(θ/2)|1⟩ and a fixed copier U pinned by two rules: U|0⟩|0⟩ = |00⟩ and U|1⟩|0⟩ = |11⟩. Because a valid quantum operation must be linear, its action on a superposition is forced — there is no freedom left. The page computes that forced output, cos(θ/2)|00⟩ + sin(θ/2)|11⟩, and compares it to the ideal |ψ⟩⊗|ψ⟩. The joint fidelity is the genuine squared overlap |⟨ψψ|out⟩|² = (cos³(θ/2)+sin³(θ/2))²; the single-copy fidelity is ⟨ψ|ρ|ψ⟩ = cos⁴(θ/2)+sin⁴(θ/2), where ρ is one output qubit after tracing out the other. Every arrow, bar, and number is read straight off these formulas.
What this is and isn't. This is an illustration, not the full proof. It shows one particular copier failing; the theorem's real force is that no copier can do better than fail on some state — a two-line argument (assume a universal cloner, apply it to two non-orthogonal states, and the inner products can't match). The best any real "cloner" can manage is an approximate copy capped at fidelity 5/6 for an unknown qubit — still not a duplicate. Our toy's worst case, 0.5, is honest for this specific machine; the point it makes — that superpositions resist copying, by linearity — is the real one.
Where the metaphor tears
No-cloning is a strictly quantum statement. Ordinary information — a fact, a photo, a document — can be duplicated perfectly and endlessly, which is the whole basis of the digital world. The metaphor only bites for the genuinely superposed, delicate things; reach for it to explain why you can't copy a spreadsheet and it simply doesn't apply.
The theorem forbids duplication, not transfer. Quantum teleportation moves a state perfectly to a new location — but it destroys the original in the act, leaving exactly one copy, never two. So "you can give trust but not duplicate it" is the right reading; "you can't pass it on at all" is not.
Human trust, unlike a quantum state, is fuzzy, renegotiable, and often improves in the retelling; nothing enforces a fidelity ceiling on a friendship. The correspondence is a lens for one real feature — that some valuable things are spent or altered by copying rather than freely reproduced — not a claim that relationships obey the Schrödinger equation.