A key question for volcanic hazard assessment is the extent to which information can be exchanged between vol- canoes. This question is particularly pertinent to hazard forecasting for dome-building volcanoes, where effusive activity may persist for years to decades, and may be punctuated by periods of repose, and sudden explosive ac- tivity. Here we review historical eruptive activity of fifteen lava dome-building volcanoes over the past two cen- turies, with the goal of creating a hierarchy of exchangeable (i.e., similar) behaviours. Eruptive behaviour is classified using empirical observations that include patterns of SO2 flux, eruption style, and magma composition. We identify two eruptive regimes: (i) an episodic regime where eruptions are much shorter than intervening pe- riods of repose, and degassing is temporally correlated with lava effusion; and (ii) a persistent regime where erup- tions are comparable in length to periods of repose and gas emissions do not correlate with eruption rates. A corollary to these two eruptive regimes is that there are also two different types of repose: (i) inter-eruptive re- pose separates episodic eruptions, and is characterised by negligible gas emissions and (ii) intra-eruptive repose is observed in persistently active volcanoes, and is characterised by continuous gas emissions. We suggest that these different patterns of can be used to infer vertical connectivity within mush-dominated magmatic systems. We also note that our recognition of two different types of repose raises questions about traditional definitions of historical volcanism as a point process. This is important, because the ontology of eruptive activity (that is, the definition of volcanic activity in time) influences both analysis of volcanic data and, by extension, interpretations of magmatic processes. Our analysis suggests that one identifying exchangeable traits or behaviours provides a starting point for developing robust ontologies of volcanic activity. Moreover, by linking eruptive regimes to con- ceptual models of magmatic processes, we illustrate a path towards developing a conceptual framework not only for comparing data between different volcanoes but also for improving forecasts of eruptive activity.