In gauge theory, tree-level scattering amplitudes with a single negative-helicity gluon vanish. This is not approximate — it is exact, a consequence of supersymmetric Ward identities that constrain the structure of amplitudes even in non-supersymmetric theories. The result is textbook, proven in the 1980s, and foundational to the modern amplitudes program.
Note (2026-11-13; read against the paper's full text): the essay says gravity breaks from the gauge theory pattern in this corner of kinematics. The paper says the opposite about the relation: it presents the nonzero single-minus gravity amplitudes as extending to gravity a similar result for Yang-Mills theory, describes its recursion as the gravitational analogue of the ordered single-minus recursion in gauge theory, and notes that a similar formula exists for gluons. The single-minus amplitudes are nonvanishing for half-collinear configurations in Klein space or for complexified momenta; in a restricted decay region they reduce to an (n minus 2)-fold product of soft factors, and with suitable analyticity assumptions they are generated by a recursive L w(1+infinity) Ward identity seeded by the three-graviton amplitude. The essay's explanation that spin-2 soft factors prevent the gauge-theory factorization, and its conclusion that 'gravity is not gauge theory squared' here, are not statements I found in the paper.
Gravity is supposed to follow the same pattern. Graviton scattering amplitudes with a single negative-helicity graviton should vanish at tree level, by gravitational analogy with the gauge theory result. This has been assumed for decades.
Guevara, Lupsasca, Skinner, Strominger, and Weil (arXiv:2603.04330) show the gravitational amplitudes are nonzero. Not in a loophole — in a specific kinematic configuration (“half-collinear”) that exists in Klein space (signature 2,2 rather than the physical 3,1). The amplitudes have a closed-form expression as a product of soft factors, governed by a W_{1+∞} Ward identity — an infinite-dimensional symmetry algebra that organizes the soft graviton expansion.
The gauge theory result still holds. The gravitational result breaks from it. The divergence traces to a structural difference: gauge theory amplitudes factorize in a way that forces the single-minus configuration to vanish, but gravitational amplitudes have additional soft factors from the spin-2 structure of the graviton that prevent the same factorization. Gravity is not gauge theory squared in this corner of kinematics.
The result is exact and has a clean physical interpretation: the nonvanishing amplitudes describe the collinear emission of gravitons in configurations where the gravitational soft theorem acquires corrections that gauge theory soft theorems do not. The corrections are organized by the W_{1+∞} algebra, which acts as a symmetry of the celestial sphere — the sphere at infinity where gravitational radiation is observed.
The structural lesson: analogies between gauge theory and gravity, which have been extraordinarily productive (BCJ duality, double copy, KLT relations), have limits. The limit here is specific and informative — it tells you exactly where the “gravity = gauge theory squared” picture breaks down and what additional structure gravity contains.
Guevara, Lupsasca, Skinner, Strominger, & Weil, "Single-minus graviton tree amplitudes are nonzero," arXiv:2603.04330 (2026).