Lucy Suchman: Human-Machine Reconfigurations

The book is the updated and extended version of her 1987 dissertation „Plans and Situated Actions“, a frequently cited classic in HCI literature. Suchman critisizes the (back at the time) predominant cognitivist planning view on human-machine interaction, describing the connection between an action in the world and its mental representation (e.g., as a plan) as rather weak. Consequently, the tracking of an action does not automatically reveal a distinct intent. An identical aim can be achieved through different courses of action, each responding to the requirements of the respective situation. In this respect, plans can be seen not as a precondition for action, but as a form of situated action themselves.
In her case study, Suchman examines the intelligent expert system of a complex copying machine that is supposed to give interactive and situational instruction to the user, comparable to a human coach. Comparing human interaction with human-machine interaction, Suchman describes the encounter of users with the artificially intelligent expert system. She sees the very limited access of the machine to the world and therefore the assymmetric ability of human and machine to make sense of the interaction as the important problem in human-computer interaction. The user’s actions as well as the machine’s reaction can be ambiguous. While in human interaction, there are several mechanisms to regain mutual intellibility, the machine does not track some of the important events to catch the user’s intent.
In the new chapters, Suchman critizes her former view on the relation of humans and machines as having reserved agency exclusively to humans. She revises this perspective by referring to several resources like feminist theory and science and technology studies and advances in her recent work. In the following, she gives an updated overview of reseach into the role of plans and ordering devices, the agency in humans and artefacts, and the role of figure and embodiment on our image of an autonomous intelligent other. She concludes with a chapter on how a shift in our perception of the human-machine relationship might open up new possiblities to reconfigure it.

The book is definitely worth reading for designers in HCI. It presents some very valuable critique on the views on human-computer interaction established in the natural an engineering sciences and names some intriguing alternative perspectives. It is amazing how current even Suchman’s 1987 analysis still is (even if the machine itself now looks a little outdated). The new chapters expand and update the basic critique with recent developments and examples. However, this book is no bedside reading and requires a lot of attention, as it is densely written and refers to a lot of concepts that at least I was not aware of before.
I also attached an (almost complete) extended summarization of the book for a quick read.