Entanglement and Bell states

Two bits can be correlated the boring way: write the same value into both, seal them in envelopes, mail them apart. Open one, you know the other. A Bell pair is correlated in a way that has no such story. The two qubits are not secretly carrying preset values, yet they agree perfectly the instant you look. And despite decades of hype, that agreement cannot carry a single bit of message. Both of those claims need care, so we build them slowly.

The myth to drop first

The most stubborn misreading of entanglement is that measuring one qubit instantly transmits something to the other, a spooky telephone across any distance. Test the intuition before the math.

See the correlation, then try to break it

The widget builds a Bell pair with an H and a CNOT, then samples it. Watch the joint outcomes lock together while each qubit alone stays a fair coin. Then act only on the first qubit and confirm the second qubit's own statistics never move.

Why the Bell state cannot be split

We built $\frac{1}{\sqrt{2}}(|00⟩+|11⟩)$ from $|+⟩|0⟩$ last lesson. A state is entangled when it cannot be written as one qubit's state $\otimes$ another's. Suppose, for contradiction, that it could:

Matching terms forces $ac\ne 0$ (there is an $|00⟩$) and $bd\ne 0$ (there is an $|11⟩$), but also $ad=0$ and $bc=0$ (no $|01⟩$ or $|10⟩$). From $ad=0$, either $a=0$ or $d=0$, and either one kills $ac$ or $bd$. Contradiction. No product of two single-qubit states reproduces this vector.^[1] The two qubits genuinely have no separate states of their own, which is the mathematical content of entanglement, derived rather than asserted.

Where the no-signaling wall comes from

Here is why the perfect correlation still sends nothing. Consider only Bob, holding the second qubit. Before he compares with Alice, his outcome is $0$ half the time and $1$ half the time, and this is true whether or not Alice has measured, and whatever local gate Alice applied first (the widget let you check X and Z). His local statistics are frozen at 50/50. A message needs a detectable change on the receiving end, and there is none. Only when Alice and Bob bring their two lists of outcomes together, over an ordinary channel no faster than light, do the correlations appear. Entanglement is a shared correlation you can only cash in by comparing notes, never a channel.

What entanglement is not is a hidden set of preset values. Bell's theorem (which we name rather than prove here) shows no assignment of secret local values can reproduce the full pattern of quantum correlations across different measurement choices. The correlations are stronger than any envelope story allows, and experiments have confirmed the quantum prediction.^[2] This is the resource that powers teleportation and quantum key distribution, and it is why the no-cloning theorem has teeth.

from qiskit import QuantumCircuit
from qiskit.quantum_info import Statevector

qc = QuantumCircuit(2)
qc.h(0)
qc.cx(0, 1)                     # Bell state (|00> + |11>)/sqrt(2)

probs = Statevector(qc).probabilities_dict()
# {'00': 0.5, '11': 0.5}       joint: only matching outcomes

marginal_q1 = Statevector(qc).probabilities([1])
# [0.5, 0.5]                    qubit 1 alone: a fair coin, carries no message