Other people's ideas are like new clothes: You can try them on and wear them for a while. Sometimes they just don't work, so you bury them in the bottom drawer, where they are forgotten until you happen across them while rummaging for something else. Other times, they do work. They fit, they change the way you feel, and they change your whole outlook. Those clothes become part of your daily wardrobe, always on hand when you need them. Those are the good ideas.
The ideas in Lucy Suchman's book, Plans and Situated Actions, first published in 1987, became part of my daily wardrobe. They offered me a whole new outlook on human-computer interaction. Suddenly, all of my taken-for-granted tools of the tradetasks, goals, planscame up for critical questioning. Many people in HCI, myself included, didn't quite get her argument, but that didn't matter. The ideas were so rich and deep that people were able to make sense of them from their own disciplines and appropriate or misappropriate them in different waysin other words, to just wear them for a while to see how they fit. In 2007, the second edition of Plans and Situated Actions was published, but it didn't grab my attention for a while because I'd heard that it was basically the same book as was printed in 1987 with a few additional chapters. When I finally got around to having a look at those extra chapters, I realized that, once again, Suchman had popped up with some new clothes for me to try on.
HCI research has always had dual targets: understanding the phenomenon of people using computers and understanding the process of design. Both of these are interesting in their own right and are essential for helping designers to design more effectively in a user-centered way. Both targets presuppose clear separations between user, computer, and designer. But in recent years, we have seen that from Second Life to teenagers hanging out on Facebook, from hacker culture to open source, from mobile to ambient technology, these previously clear separations have become harder to maintain. Even the terms user, computer, and designer have begun to seem tired and ill fitting somehow. For me, Suchman's latest edition, renamed Human-Machine Reconfigurations, offers a set of ideas that do not seek to resolve these problems so much as to dissolve them . They do so by offering a very different way of looking at the fundamental issues of agency and interaction.
A presupposition in much HCI work is that the difference between the user and the computer in interaction involves agency and control. The human is the agent with goals, plans, and intentions who acts through the passive, albeit responsive, computer-as-tool by means of messages translated across the interface. The aim of HCI, then, is to narrow the distance between human and machine to make the interface more transparent. Suchman offers an alternative to presupposing agency as a property of either humans or machines and thus the interface as an a priori and self-evident boundary. Instead, she argues that agency can be seen as something that emerges from, or is constituted in, the relation between humans and machines that is enacted in particular ways in particular settings. Suchman's analysis is informed by the discipline of social studies of technology (STS) and, inspired by Barad , she repositions the concept of interaction in this way:
Whereas the construct of interaction suggests two entities, given in advance, that come together and engage in some kind of exchange, intra-action underscores the sense in which subjects and objects emerge through their encounters with each other .
To put this a different waya way that introduces another key concept for Suchmana person's (or thing's) capacity to act (their agency) is reconfigured when it comes into contact with another thing or person. It is only in coming together into some relation that the entities have separable identities, meanings, and agencies. Furthermore, she argues, the boundary between those entities is a shifting one.
To illustrate the first idea that agency emerges in the intra-action between people and technology, Suchman reminds us of Bruno Latour's account of the gunman. In one way this example is quite banal, but in another, it is quite profound (for a detailed analysis, see ). When a man uses a gun to shoot someone, is it the man or the gun that shot the person? Clearly the gun could not shoot the person unless the man was there to pull the trigger, but equally clearly, the man could not have shot the person if the gun had not been present. Thus, the agency of the shooting emerges in the specifics of the intra-action between man and gun. Furthermore, since the man could not have shot the person without the gun, then perhaps the person supplying the gun to the man is implicated? Perhaps the people who made the gun? Perhaps the people who designed the machines to make the gun? The point here is not to make a reductio ad absurdum argument; rather, the point is to appreciate how an act emerges from a network of people and technologies that Suchman calls sociomaterial practices. This network extends in space and time well beyond the moment of encounter between person and tool, back into practices of designers and manufacturers and back into a history of previous interactions and encounters between person and tool.
The second idea, that the boundaries of agency are not fixed but open to change, is illustrated by examples (again borrowed from STS) of medical equipment such as life support machines. When a person is attached to a life support machine, the person's agencytheir capacity for actionis radically reconfigured over time such that the responsibility for even basic life functions, such as breathing, blood circulation, movement, and consciousness is shifted not just to a machine, but to a network of people, practices, and institutions.
Suchman uses these ideas as starting points to critically reflect on user-centered design as a practice. She argues it is inevitable that any analysis of complex networks "cuts" the network in a certain way, making certain actants visible "on stage" and leaving others invisible, left "waiting in the wings" but "bracketed out" for current purposes. In HCI, that often means foregrounding user and computer while bracketing out designers, engineers, programmers, and marketers, all of whom play important parts in shaping meaning in use.
One way of dealing with the problem is to consider these invisible actors as having left traces of their agency embedded in the material artifact in some way. Hence we talk of system properties or design values or assumptions embedded in the artifact, which help keep those hidden parts of the network visible for consideration. But, Suchman argues, this embedding is often less deterministic, less rational, and more open and ambiguous than simplistic analyses would have us believe. To understand that indeterminacy, we have to make more of the network visible. So, for example, usability can be understood as a collection of properties embedded in the user interface that are the result of invisible designers making assumptions about the user's needs. But Suchman reminds us (quoting ) that in the real world, those professionally involved in the development of complex technologies don't always have a clear conception of "the user," let alone a multiplicity of individuals who make up the "user population." Developers imagine the system as a complex machine built from data structures and processes. They are imaginatively "inside the system," and thus "the very concept of a user interface can be difficult to grasp or take seriously" .
But this is not just a lack of empathic imagination on the part of developers. Suchman suggests that sometimes organizations can make it difficult for designers, developers, and programmers to actually get to users. They often consider communities of users an "unruly and potentially disruptive" influence on their properly controlled and managed development processes. User-centered design has responded to these problems by providing tamer proxies, such as scenarios, personas, and user profiles to stand in for those unruly users. These are supposed to help developers connect with the real world while allowing the organization to stay in control. But in reality, they are often too far away from those unruly subjects and distal sites to be stable proxies. Suchman concludes,
...there is no stable designer/user "point of view" nor are imaginaries for the user or settings of use inscribed in anything like a complete and coherent form in the object... The "user" is, in other words, more vaguely figured, the object more deeply ambiguous .
One of the concerns Suchman has with this reframing of HCI in terms of actor networks and socio-material practices is the risk of losing the ability to differentiate between human and machine. Her challenge is to preserve the ideas of emerging agency and shifting boundaries while at the same time maintaining a difference between human and machine so that responsibility, accountability, and creativity can be reclaimed for the human:
How might we conceptualize the granting of agency that at once recognizes the accountability of human actors, while recognizing their inseparability from the sociomaterial networks through which they are constituted? 
But what Suchman's and others' research has shown over the years is that it is in the nature of technologies (whether they be computers, printing presses, or operating procedures) to prescribe human activity, to configure the user in certain ways, and to recommend or even enforce certain courses of action on the user. But at the same time, it is in the nature of people to find creative, artful ways of working with, working around, or subverting these attempted configurations. In so doing they are creatively appropriating the technology to new uses and new meaningseven by simply getting it to work in a new setting.
With this idea of humanness as creativity in place, Suchman turns her attention in the closing chapters of the book to an exploration of digital participatory art as an alternative creative practice through which to explore emergence and human-machine difference. She describes, for example, the participatory installations "Mother and Child"  and "TGarden"  as interesting contrasts to HCI/AI projects that try to pack ever-more-sophisticated ambient intelligence into systems while providing no more access to the context of its human users than the Xerox machine Suchman studied for the first edition of her book did. In contrast to the Xerox, "TGarden" actively eschews trying to understand the intentions of its users, instead focusing on increasing the bandwidth of the sensorial window that the machine has on its human actant. As Sha Xin Wei says:
The TGarden software tracks gesture rather than recognizes gesture, because at no place in the software is there a "model" that codes the gesture... the software does not infer what the player means by her gesture, it merely tracks the gesture and continuously synthesizes responses. So what we have done is to set aside entirely the problem of inferring human intent from behavior, or more generally from observables. Yet by providing and even thickening the sensuous response, we make fertile the substrate for agency... .
If the old project around human-computer interaction posited agency as the hallmark of human-ness, and the design challenge as emulating that in machines, then what these new forms of participatory digital art suggest is that the new project around pervasive technology might be to offer a relational, performative account that configures both human and computer in intra-action. So, for Suchman, the design question is "how to configure assemblages in such a way that we can intra-act responsibly and generatively with and through them" . She elaborates on this idea:
More than a conversation at the interface, it is creative assemblages like these that explore and elaborate the particular dynamic capacities that digital media afford and the ways that through them humans and machines can perform interesting new effects. Not only do these experiments promise innovations in our thinking about machines, but they open up as well the equally exciting prospect of alternate conceptualizations of what it means to be human. The person figured here is not an autonomous rational actor but an unfolding, shifting biography of culturally and materially specific experiences, relations and possibilities, inflected by each next encounterincluding the most normative and familiarin uniquely particular ways .
In this latest edition, Suchman has offered me new ideas to try on and live with for a while. She has offered a new way of looking at familiar phenomena, and also some signposts to a new literature that has stimulated my thinking and increased my Amazon bill. It will take time for me to decide whether these ideas will end up in my wardrobe or my bottom drawer. For me, the answer to that question goes beyond Suchman's book to a question about how one might use the findings of STS to inform practical design. But what really makes Suchman's ideas good for me is not her answers but her questions. Questions like "What does it mean to be human?" seem to be more important than ever in today's world of fast-paced technological progressprogress that is so fast that even researchers and practitioners, let alone ordinary citizens, seldom have the time to stop and think about the answers, and what they imply for how, who, or what, we should be.
3. See , p. 267.
5. Agre, p. conceptions of the user in computer systems design. In The Social and Interactional Dimensions of Human-computer Interaction. p. Thomas, ed. Cambridge University press, New York, 1995, 67-106.
7. See , p. 192-193.
8. See , p. 270.
9. Tikka, H. Mother and child, 2000; http://mlab.taik.fi/~htikka/projects/motherchild.html
10. Wei, S.X. Resistance is fertile: Gesture and agency in the field of responsive media Configurations 10 (2003), 439-472; http://topologicalmedialab.net/xinwei/papers/texts/Configurations/Resistance_Is_Fertile.pdf
12. See , p. 285.
13. See , p. 281.
Peter Wright is professor of social computing in the Digital Interaction Group based in the culture Lab at Newcastle University. He has over 20 years of experience as a human-centered design researcher and is best known in the HCI community for his research into theory and methods for experience-centered design. His current projects focus on bringing experience-centered design principles to health-related services and technologies.
©2011 ACM 1072-5220/11/0900 $10.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2011 ACM, Inc.