The difference between various models of dark energy or modifications of
gravity becomes apparent at the level of growth of large-scale structure in
the universe. In addition to measuring the background expansion, we are now
beginning to probe this aspect. The problem that is not particularly
understood currently is the impact of the assumption of a particular model
of dark energy on the interpretation of observations.
I will show that we in principle cannot disentangle the contribution of dark
matter from that of dark energy to neither the evolution of the cosmological
background nor that of large-scale structure unless a model is assumed in
the first place. This degeneracy limits severely the measurements that we
can obtain. I will discuss the total set of model-independent observables
that we can obtain from the sky and will discuss how we could in principle
rule out various classes of generalised Galileons as models of cosmic
acceleration in their entirety.