Quantum contextuality is the impossibility of assigning outcome values independent of measurement context and is treated as a fundamental axiom of quantum theory (Cabello et al. 2010) arXiv:1010.2163. Yet circuits that stay inside the Clifford/stabilizer framework (e.g. those that create GHZ states) remain classically efficiently simulatable, so contextuality by itself yields no computational advantage. The missing ingredient is a magic state: any pure state outside the stabilizer set necessarily has Wigner-function negativity and therefore exhibits contextuality (Gross 2006). Injecting such a state (or an equivalent non-Clifford gate) lifts a Clifford circuit to full universality. I
Howard proved an operational equivalence between the onset of contextuality and the ability to distil magic states, establishing contextuality as necessary and (in this distillation model) sufficient for quantum speed-up (Howard et al. 2014)Contextuality supplies the ‘magic’ for quantum computation
Thus, every magic state is contextual, whereas contextuality guarantees an advantage only when it is present in a consumable resource. This explains why the GHZ state is contextual but gives only advantage when magic is included.
GHZ non negative:
In the phase-space formalisms used in the literature (Gross 2006 for odd-prime qudits; Kocia & Love 2017 for qubits) every pure stabiliser state is represented by a uniform, non-negative quasiprobability distribution that is supported on an affine, maximally isotropic subspace (“a line”) of the discrete phase space.For a single qubit the phase space has points; a stabiliser state (say) is non-zero on the two points of one vertical line, with value on each.
In general, an qubit stabiliser state lives on points one line per qubit tensor-factor, each carrying weight. Because every Clifford operator acts as a permutation of phase-space points. it can only reshuffle the values of ; it can never change the set of values . The GHZ-preparation circuit is a sequence of Cliffords that maps the line supporting onto another line, so inherits exactly the same non-negative weights (1/8 on eight points, 0 elsewhere).
This illustrates two distinct facts:
All stabiliser states (including GHZ) are non-negative in these Wigner representations (Kocia & Love 2017)https://arxiv.org/abs/1705.08869
Covariance of the Wigner function under Clifford gates ensures that Clifford circuits cannot create negativity (Gross 2006, §III.B)(Physical Review). https://journals.aps.org/prx/abstract/10.1103/PhysRevX.5.021003?