A significant part of my research is driven by my interest in complexity in interaction . For almost as long as people have been doing HCI research, there has been an awareness that there are important aspects - context, the role of feedback between individuals and between individual and computer - which exceed our existing models.

This can seem a trivial observation: something like this can be said of almost any discipline at almost any time in its development. However the issue is perhaps particularly pertinent for HCI since it focuses on interaction between entities. Interaction is a notoriously difficult phenomenon to understand across many domains. Many of our best established statistical techniques assume that the events they observe are independent, and driven by factors which either do not interact, or interact in a straightforward linear way. So it is not surprising that subsequent generations of HCI researchers have returned to pick at the problem of interaction as such. Early examples came in Winograd and Flores, and Lucy Suchmann’s accounts of Situated Action, then later the ideas of Embodied Interaction. Despite the importance and value of this work, the issue is still live: by 2018 Hornbaek and Oulasvirta at least, were still of the opinion that we lacked an understanding of what interaction is, and the role it plays in how user behaviour unfolds.

HCI is not alone in struggling to understand the role of interaction. Interaction is something which most current models in behavioural sciences struggle to account for. The cognitive scientist Scott Kelso has recently argued that while we now know quite a lot about how components of behaviour work. What we really don’t know is how coordination between those components works. And this is unclear not only between organisms, or between organisms and contexts, but within individual organisms: how cognitive and physiological processes interact to give rise to behaviour. Addressing this issue seems to me to be an important question for HCI in general, but perhaps most particularly to understand especially “interactive” kinds of interaction: those which involve fast or skilful engagement: gaming, artistic and musical work, and safety-critical interactions like piloting and surgery. It also seems important for understanding interactions involving groups of users, and cases where technologies are incorporated into our bodies, or operate beneath the threshold of our reflective perception.

The good news for HCI is that in recent years researchers like Kelso and have begun to frame questions of coordination and interaction more precisely. A long list of researchers including not only Kelso, but also Mengsen Zhang, Tony Chemero, Damien Kelty-Stephen, Hanne De Jaegher, Dobromir Dotov, Anna Ciaunica, Eric Rietveld, Jelle Bruineberg, and Tom Froese have been refining methodologies and theories to account for interaction as such. There is now a great opportunity for HCI to capitalise on this work. This thought drives both my current PhD thesis on multifractal analysis of interaction behaviour, and a (workshop)[] we recently organised at ACM’s CHI conference.

Complexity -