This paper tests whether gravitational waves from the neutron star merger GW170817 contained an extra kind of ripple that is not predicted by Einstein’s General Relativity. The authors add a so-called scalar “breathing” polarization to the usual two tensor polarizations and fit this extended signal model to the data. They also include, for the first time in this kind of test, a measurement of the polarization angle from the gamma-ray burst afterglow. With that extra information they find a mild (about two sigma) hint for a scalar contribution when they look at the dominant quadrupole part of the signal, but no preference for a scalar when they look at the dipole part.

The team used a model‑agnostic framework called the parameterized post-Einsteinian (PPE) formalism. PPE lets one add simple power‑law corrections to the standard gravitational‑wave waveform so that many possible deviations from General Relativity can be tested without committing to a single alternative theory. In this work they extended the waveform by a scalar breathing mode and by modifications to the tensor modes, introducing three extra non‑GR parameters. They then performed Bayesian inference — a statistical method that updates what is known about model parameters in light of data — on the pre‑merger part of GW170817. The inference was done for two angular harmonics (quadrupole ℓ=|m|=2 and dipole ℓ=|m|=1) and with two different assumptions for how the extra signal evolves with frequency.

Electromagnetic observations of GW170817 played a key role. The radio imaging of the gamma‑ray burst afterglow, obtained with very long baseline interferometry, provided constraints on the source orientation and on the polarization angle. The kilonova and other electromagnetic data gave sky location and distance. The authors folded these electromagnetic constraints into their analysis as prior information. Because the polarization content of a gravitational wave depends on the source orientation, knowing the orientation tightens the allowed non‑GR signal shapes.