This contribution critically examines various concepts related to damage functions in current use; it does not represent a comprehensive review, however. In heritage science, damage functions can be defined as functions of unacceptable change, dependent on agents of change. Unlike in other domains of science, the reference to unacceptable change implies that a value-based decision needs to be applied to the analytically determinable change, or dose-response function. Since there are a number of values associated with heritage, there can be a number of damage functions describing one and the same physical or chemical process of change. For practical reasons, it is thus useful to decouple the value function from the dose-response function, as the latter can be independently deduced (empirically or experimentally). While the behaviour of single materials and objects is often modelled deterministically, the probabilistic approach can be used to describe the degradation of complex structures, e.g. buildings, although not commonly in heritage research. There is a case to be made for more in-depth exploration of incremental and stochastic processes of degradation of heritage materials and structures, and the associated values, with a view towards developing more damage functions. Damage functions find use in modelling for predictive maintenance. However, predictions are associated with uncertainty due to the need for extrapolation, and this is also rarely analytically examined. Validation of modelling processes is possible using accurate methodologies of asset monitoring, which would enable an effective appraisal of preservation strategies. © The International Institute for Conservation of Historic and Artistic Works 2013.