Who’s Eddy? He (or she!) is a weather formation. Which is being used as a metaphor. See here for one paper looking at this idea – with hopefully more to come.
Most current approaches to studying how attention and affect are regulated in everyday settings assume, implicitly, that these processes should be thought of as fundamentally rational processes. For example, most approaches to studying the self-regulation of arousal emphasise processes such as allostasis. At its heart, this is fundamentally simple and logical: if something exogenous to me acts in such a way as to increase my arousal, then I alter my behaviours in such as a way as to decrease my arousal, and vice versa.
In 20 or 30 years I think that we’re going to appreciate better that things are much more complicated than that. First, there isn’t actually that much evidence that allostatic mechanisms operate at all, at least during early development. (This is evidence from simplified experimental paradigms, but I talk about the limitations of those here.) Second, there is evidence for the opposite types of process might exist – that both high and low arousal states might possibly become ‘sticky’ in some way – they take on a self-sustaining characteristic which ought not to be expected based purely on allostatic models. Third, there is the question of how we tell apart dysregulatory processes (where someone doesn’t want to be in a state but they lose control of their behaviours so that they end up being in that state anyway – as in panic disorder, for example) from self-regulatory processes (where someone wants to be in a high arousal state, and so effortfully upregulates arousal into that state – as in stimming in Autism, for example). Almost all research implicitly assumes that it’s always the former – that the ideal state is an affect-free state, and that self-regulation is always a process in which excessive arousal is down-regulated – but this just doesn’t feel right to me.
My gut feel says that modelling approaches – that try to forward-predict changes in arousal, using dynamic models based on non-linear dynamics such as attractor basins – might be useful here. But honestly this is a guess at the moment, as this is an area that really confuses me, the more I think about it! So do get in contact if you’ve got any ideas…