Context. The Galactic chemical evolution of sulphur is still under debate. At low metallicities some studies find no correlation between [S/Fe] and [Fe/H], which is typical for a-elements, while others find [S/Fe] increasing towards lower metallicities, and still others find a combination of the two. Each scenario has different implications for the Galactic chemical evolution of sulphur. Aims: The aim of this study is to contribute to the discussion on the Galactic chemical evolution of sulphur by deriving sulphur abundances from non-local thermodynamic equilibrium (LTE) insensitive spectral diagnostics in disk and halo stars with homogeneously determined stellar parameters. Methods: We derived effective temperatures from photometric colours, surface gravities from stellar isochrones and Bayesian estimation, and metallicities and sulphur abundances from spectrum synthesis. We derived sulphur abundances from the [S i] ?1082 nm line in 39 mostly cool and metal-poor giants using 1D LTE MARCS model atmospheres to model our high-resolution near-infrared spectra obtained with the VLT, NOT, and Gemini South telescopes. Results: We derive homogeneous stellar parameters for 29 of the 39 stars. Our results argue for a chemical evolution of sulphur that is typical for a-elements, contrary to some previous studies that have found high sulphur abundances ([S/Fe] ? 0.6) for stars with -2.5 < [Fe/H] < -1. Our abundances are systematically higher by about 0.1 dex than those of other studies that arrived at similar conclusions using other sulphur diagnostics. Conclusions: We find the [S i] line to be a valuable diagnostic of sulphur abundances in cool giants down to [Fe/H] ? -2.3. We argue that a homogeneous determination of stellar parameters is necessary, since the derived abundances are sensitive to them. Our results ([S/Fe]) agree reasonably well with predictions of contemporary models of Galactic chemical evolution. In these models sulphur is predominantly created in massive stars by oxygen burning and is ejected into the interstellar medium during Type II supernovae explosions. Systematic differences with previous studies most likely fall within modelling uncertainties. Based partly on observations obtained at the Gemini Observatory, which is operated by the AURA, Inc., under a cooperative agreement with the NSF on behalf of the Gemini partnership: the NSF (US), the PPARC (UK), the NRC (Canada), CONICYT (Chile), the ARC, CNPq (Brazil), and CONICET (Argentina).Based partly on observations collected at the European Southern Observatory, Chile (ESO program 080.D-0675).Figures 3 and 4 are available in electronic form at https://urldefense.proofpoint.com/v1/url?u=http://www.aand a.org/&k=diZKtJPqj4jWksRIF4bjkw%3D%3D%0A&r=UpUi4hY04iGuRFIqvI40bIbHeCQwTKuVYBJx wGoVKtk%3D%0A&m=iqcfCKSx8rh%2FpmcK41TCrt4Ikno4QmZlEAZdLNaaujk%3D%0A&s=72eb8 5ad241058a299a6483d6e0534217d09413765b2813d578bde6ab210dccc
