Different kinds of evidence are taken into account in sentience research. Given the potential biases of phylogenetic evidence, the analysis of behaviour has a foundational role. In this respect, sentience is associated to high degrees of behavioural flexibility – as the latter is supposed to be evidence of some kind of organismal “choice”. Several notions of behavioural flexibility can be distinguished, whereby all are vernacularly contrasted to mechanistic notions of automaticity and determination. I shall attempt to provide a taxonomy of notions of behavioural flexibility and illustrate how such notions might be deployed – whether coherently or incoherently remains an open question – to support sentience ascription claims to different lineages.