Studies of experimentally deformed rocks and small-scale natural shear zones have demonstrated that volumetrically minor phases can control strain localisation by limiting grain growth and promoting grain-size sensitive deformation mechanisms. These small-scale studies are often used to infer a critical role for minor phases in the development of plate boundaries. However, the role of minor phases in strain localisation at an actual plate boundary remains to be tested by direct observation. In order to test the hypothesis that minor phases control strain localisation at plate boundaries, we conducted microstructural analyses of peridotite samples collected along a ~1 km transect across the base of the Oman-United Arab Emirates (UAE) ophiolite. The base of the ophiolite is marked by the Semail thrust, which represents the now exhumed contact between subducted oceanic crust and the overlying mantle wedge. As such, the base of the ophiolite provides the opportunity to directly examine a former plate boundary. Our results demonstrate that the mean olivine grain size is inversely proportional to the abundance of minor phases (primarily orthopyroxene, as well as clinopyroxene, hornblende, and spinel), consistent with suppression of grain growth by grain-boundary pinning. Our results also reveal that mean olivine grain size is proportional to CPO strength (both of which generally decrease towards the metamorphic sole), suggesting that the fraction of strain produced by different deformation mechanisms varied spatially. Experimentally-derived flow laws indicate that under the inferred deformation conditions, the viscosity of olivine was grain-size sensitive. As such, grain size, and thereby the abundance of minor phases, influenced viscosity during subduction-related deformation along the base of the mantle wedge. We calculate an order of magnitude decrease in the viscosity of olivine towards the base of the ophiolite, which suggests strain was localized near the subduction interface. Our data indicate that this rheological weakening was primarily the result of more abundant minor phases near the base of the ophiolite. Our interpretations are consistent with those of previous studies on experimentally deformed rocks and smaller-scale natural shear zones that indicate minor phases can exert the primary control on strain localisation. However, our study demonstrates for the first time that minor phases can control strain localisation at the scales relevant to a major plate boundary