Oceanic spreading rates are highly variable, and these variations are known to correlate to a variety of surfaceobservables, like magmatic production, heat flow or bathymetry. This correlation lead to classify ridges into fastand slow spreading ridges, but also into the more peculiar ultraslow spreading regime. Here we explore the dynamicrelationships between spreading ridges, plate tectonics and mantle flow. We first focus on the thermal signatureof the mantle, that we infer from the global S-wave seismic tomography model of Debayle and Ricard (2012).We show that the thermal structure of ridges gradually departs from the half-space cooling model for slow, andabove all ultraslow spreading ridges. We also infer that the sublithospheric mantle temperature decreases by morethan 150 degrees C from fast to ultraslow spreading regimes. Both observations overall indicate that the mantleconvection pattern is increasingly chaotic underneath slow and ultraslow spreading ridges. We suggest that this isdue to far-field tectonics at the other ends of lithospheric plates: not only it modulates the spreading rates but it alsoalters the convection regime by obstructing the circulation of plates, which in turn modifies the surface kinematicconditions for the convecting mantle. We test this hypothesis using a thermo-mechanical model that represents aconvection cell carrying a continental lithosphere atop. The continent gradually drifts away from the spreadingridge, from which the oceanic lithosphere grows and cools while the continent eventually collides at the oppositeside. In turn, this event drastically modifies the upper kinematic condition for the convecting mantle that evolvesfrom a mobile lid regime to an almost stagnant lid regime. Implications on spreading ridges are prominent: heatadvection decreases with respect to thermal conduction, which causes the oceanic lithosphere to thicken faster;the oceanic plates get compressed and destabilized by a growing number of small scale transient plumes, whichdisrupt the structure of the oceanic lithospheres, lower the heat flow and may even starve ultraslow ridges frompartial melting. It follows that the spreading rate of a modern ridge mirrors its status in the global plate tectonicsframework within a unique breakup, drift, collision scenario, within the transition from mobile to stagnant lid, andthat it is the same mechanism that build mountains at converging boundaries and control spreading rates. Oceanicridges thus can be regarded as a sensor of the resisting rather than driving forces. Both the model and the seismicstructure of the mantle underneath ridges reveal that the temperature variations are largest at shallow depths inthe upper mantle, i.e. at the critical depth where the melt supply to the above ridges can be modulated, therebyalso explaining why slow and ultraslow ridges are almost exclusively associated to cold mantle. It follows that thechemistry of oceanic ridge basalts may not strictly reveal the mantle potential temperature, but the variations in thesublithospheric temperature field.