Performing well requires effort – which feels hard. But if motivation arises naturally and spontaneously, then why should it feel hard? I am investigating the idea that this feeling of effort might come about when an inflexible system is influenced by a flexible one. A group of relatively-isolated neurons will follow a particular rigid state trajectory, but this can be stabilised by external inputs from flexible neural systems, forming an attractor. The additional stabilising signals may constitute a cost, that counts as effortful. I will outline three ways in which neural attractor states might be corrupted by random or irrelevant input, and thus deviate from a desired path. Simple attractor models can provide testable predictions about how these disruptions impact behaviour. I will show some of our attempts to study these mechanisms using saccades and working memory.