The asteroseimsology of individual solar-like oscillators generally allows more precise inferences about many stellar model parameters than with only non-seismic constraints. Some of these parameters (e.g. the mass and age) are fundamental physical properties of the star but some (e.g. the mixing-length parameter for near-surface convection) represent uncertainty in the physics of our stellar models. Most asteroseismic modelling takes these parameters as free or fixes them to some value, motivated by other constraints. I will show how one can introduce correlations between the parameter values for different stars to quantify our expectation that, even if we don’t know exactly what values these non-fundamental parameters take, we don’t expect them to differ much between similar stars. e.g. if two stars have surface gravities and effective temperatures that differ by 0.05 dex and 50 K, respectively, they probably have similar mixing-length parameters, too. I demonstrate this concept using a binary star system observed in Kepler, in which I also experiment with how close we expect the stars’ ages and compositions to be. In principle, the method is not limited to binary stars and, the more efficient the model-fitting for a single star, the more stars can be included in a simultaneous fit that constrains how the non-fundamental parameters vary across a (sub)sample of solar-like oscillators, like the sample that PLATO will study.