Features

XXII.6 November - December 2015
Page: 37
Digital Citation

Filling the big hole in HCI research


Authors:
Alan Blackwell

Recent issues of Interactions have laid the groundwork for a debate about the disciplinary status of HCI. This was initially perceived as a “big hole” in HCI research—the concern that HCI does not seem to have a solid intellectual or methodological core [1]. That article by Vassilis Kostakos drew on his work with Liu and others, whose bibliometric analysis of keywords in CHI publications over the past 20 years showed that the field seemed to follow technical fashions rather than long-term research themes.

ins01.gif

Quite reasonably, Kostakos and his colleagues observed that it is hard to maintain an academic discipline if there are no persistent ideas or methods by which the discipline can be defined. In responding to their Interactions article, Stuart Reeves [2] recommends that we should think of HCI not as a discipline but rather as an interdiscipline, an argument drawing on previous suggestions by Yvonne Rogers, and on a paper I had presented at this year’s CHI conference with the title “HCI as an Interdiscipline” [3]. My own paper had also been written in response to the work of Liu and Kostakos. This article summarizes it, and also proposes a way forward, thus filling the hole.

However, I should start by noting that the nature and significance of the hole is open to question. Drawing attention to a hole at the center of HCI research is like drawing attention to the hole at the center of a horse and carriage. It is true that the space between the horse and carriage is mostly empty, but this is an important feature of the whole apparatus—filling it up would miss the point of having a separate horse and carriage in the first place! I hope to explain why there is little need to worry about the hole in HCI research, for related reasons. In my slightly flippant analogy, the horse of HCI is insights into human behavior, and the carriage is technological innovation. The space between is a different kind of thing—neither a horse nor a carriage, but essential to the overall enterprise.

ins02.gif

My argument is supported by evidence from two large-scale case studies, reported in more detail in my recent CHI paper [3]. The first is a systematic analysis of 180 collaborative projects in my Crucible network for research in interdisciplinary design, carried out over 15 years with more than 450 collaborators across the arts, humanities, and social sciences. These projects were initiated without asking whether they counted as HCI—from any definition of the term. Much of this work has not been published in HCI, and never will be. For example, although several eminent anthropologists have been involved in this work, they would receive little recognition among their peers for HCI publications. Because anthropology does not have an extensive culture of co-authorship, I do not work with them to translate or interpret their findings. Nevertheless, despite the resulting distance from mainstream HCI, I argue that analyzing this large corpus of work can help us understand the motor component of my horse and carriage analogy: insights into human behavior.

The second case study addresses the carriage end of the analogy: using research to achieve technological innovation. This work was originally commissioned by Nesta, a national agency concerned with creative innovation in the U.K. My team had been asked to explore the relationship between interdisciplinarity and innovation, a study we undertook through a snowball sample of leading interdisciplinary innovators, followed by intensive reflective workshops with a selection of those leaders, contextual interviews, and in-depth development of grounded theory to describe the phenomena we observed.

These studies have also provided an opportunity to reflect on my own use of theories and methods, and the ways in which I draw on other disciplines to do HCI work. Originally trained as an engineer and with a master’s in AI, I had spent many years designing user interfaces and understanding and evaluating user requirements before I learned of academic HCI. This previous life taught me that the word discipline has quite different meanings in academic research and in business. My commercial design projects always involved multidisciplinary teams—people trained with skills in software, electronics, packaging, marketing, finance, and others. It was clear what responsibilities each would take, and there was no anxiety about the combination being interdisciplinary. In academia, the problems arise when innovation happens; when a new theory is constructed, who gets to own it, and in what discipline should it be taught?

The possibility that HCI might be not so much a discipline as a way of talking between applied practitioners had already been rather eloquently stated before the start of the 20-year period studied by Kostakos and colleagues. In his contribution to the classic collection The Art of Human-Computer Interface Design, Scott Kim offered advice on the value, conduct, and need for interdisciplinary cooperation in HCI [4]. This book was targeted to an audience of practitioners rather than scientists, who in the early years of the GUI were addressing the challenge of how graphic artists and designers could work effectively alongside programmers. Kim observed that working effectively in this context meant that it is necessary to “stick your nose into other people’s business” because different disciplines have different priorities. This requires the skills of an ambassador, as well as organizational structures that support and maintain such skills. In much the same way as Stuart Reeves’s recent Interactions article, Kim concluded that discipline is the wrong way to think about the value inherent in a field like HCI. Rather than calling it a specialty, he suggests that an interdisciplinary field might be called a generality.

We do have more sophisticated ways to describe and analyze these dynamics through the sociology of knowledge. Peter Galison’s metaphor of the trading zone can be used to describe the languages through which engineers and technology designers are able to achieve productive exchanges with researchers offering insight into user experiences and behavior. Pragmatic advice for the day-to-day operation of HCI initiatives can be taken from Susan Leigh Star and James Griesemer’s description of the boundary object as a fluid construct able to accommodate different conceptual interpretations within different knowledge systems. Organizational sociologists have drawn attention to the bounded communities that act somewhat like “disciplinary” centers within large organizations, but also rely on mavericks and brokers to achieve effective work across those boundaries.

These are useful ways in which to understand the things that happen in an interdiscipline, but it is perhaps more important to ask why interdisciplines occur in the first place, and why they are a necessary feature of both professional practice and public policy. Donald Stokes describes the value of publicly funded research in Pasteur’s Quadrant, where use-inspired basic research is driven both by practical needs and a quest for fundamental understanding. This aligns nicely with Donald Schön’s characterization of the reflective practitioner, which is often cited as a model for HCI as an intellectually engaged professional endeavor.

It was with this background in mind that I approached the analysis of 180 projects from the Crucible network, an organization established to encourage collaboration between technologists and researchers in the arts, humanities, and social sciences. Crucible had been a strategic response to two specific opportunities at the University of Cambridge. The first was that the 50 percent of Cambridge undergraduates in humanities disciplines represent an untapped resource for the large Cambridge technology sector. The second was that it is so easy to conduct interdisciplinary research in a university, because each college brings together students and faculty across many disciplines.


Kim concluded that discipline is the wrong way to think about the value inherent in a field like HCI. Rather than calling it a specialty, he suggests that an interdisciplinary field might be called a generality.


Using the historical corpus of Crucible projects, I set out to test whether there had been any convergence on a set of disciplinary concerns (either theoretical or methodological) of the kind that Kostakos and his colleagues hoped might arise in HCI. I therefore classified each of the 180 Crucible projects according to which CHI review subcommittee might be willing to publish it. Of course, the projects in this sample were not all HCI projects—they included technologies in architecture, engineering, biotech, and others—so the classification required analogies between the kinds of interactivity associated with digital technologies, and “interactivity” in these other contexts. The results, as seen in Figure 1, showed that there seems to be no predetermined natural mapping between disciplinary concerns and the ways in which the CHI community structures itself. More detail of this analysis, and of the lessons it offers with regard to the role of specific CHI communities, can be found in the CHI paper [3].

In order to explore how this situation might arise, the second case study, based on our national survey of interdisciplinary innovation, offers some helpful observations. Both public funding and commercial sponsorship impose an expectation on the CHI community to be a generator of innovation, which Erik Stolterman and Mikael Wiberg observe to be a key criterion for concept-driven research [5].

Our key finding was that the relationship between interdisciplinarity and innovation has to do with unexpectedness. Disciplinary research sets up a framework that defines and thus constrains the outcome. Innovative research, on the other hand, has results that were not expected. More surprising is that it may not even be possible to describe the nature of the research question. Well-posed research questions can be well-posed only in terms established by existing bodies of knowledge. Our informants in this project often drew attention to the ways in which their interdisciplinary research answered questions that did not even exist at the outset. Rather than arising through crude metaphors of “cross-fertilization” or “breaking down silos,” innovation arises where disciplines challenge one another’s definitions of knowledge.

This challenge is achieved, practically universally, through collaboration—between members of a team who hold different understandings of the world but work together to reach a new one. Any partner must be able to challenge disciplinary assumptions, as observed by Scott Kim. Once again, a common metaphor for interdisciplinary work is insufficient—it is not simply the case that disciplinary specialists “speak different languages” and need a “translator” to be able to understand each other. On the contrary, skilled researchers are often skilled communicators, easily able to share their understanding in a register appropriate to a new audience. The challenge, as we identified it, is that different disciplines are actually setting out to achieve different things. They talk at cross-purposes because they do not appreciate how different the intentions of their collaborators might be.


Rather than arising through crude metaphors of “cross-fertilization” or “breaking down silos,” innovation arises where disciplines challenge one another’s definitions of knowledge.


These insights into the nature of interdisciplinary innovation highlighted the need for time and resources in which members of the team can accept that they are wrong—or at least in need of a new understanding. Humility is invaluable, but it is never easy for skilled professionals to accept challenges to their understanding. Since professional skill is acquired within a discipline, innovation from elsewhere is always challenging. The response often occurs over a period of years rather than months, during which it may turn out that the original objectives of a project must be abandoned in light of the understanding gained.

Although these processes are collaborative, we found that the role of the leader was critical. The leader must articulate a vision that persuades patrons and publics to commit the necessary resources, and also recruit team members to a shared moral purpose that overrides the commitments of their disciplinary training. More surprising, the leader must then be able to re-orient the whole enterprise when unanticipated opportunities arise. We found that certain professions—theater producers, product designers, journalists—appear to be particularly effective in nurturing these pragmatically collaborative leadership techniques. In an unavoidably risky process, management of risk is also a high priority. This cannot take the form of removing risk, which would render the innovation process meaningless. Instead, necessary risk must be managed as in a financial portfolio, but here the portfolio is an intellectual one, with a good mix of methods, theories, and discursive or practical styles.

To return to the horse-and-carriage analogy presented earlier, readers will no doubt have observed that separate horses and carriages are not as common as they once were. Surely we should just combine the horse and carriage into an automobile, in which case the hole disappears, and the problem goes away? In the terms of my argument, this would be like saying that humanities and computer science should be merged into a single super-discipline, with no internal holes, and the new community sharing far greater resources for the good work of HCI.

A more likely outcome is that we will remain an interdiscipline, a kind of parasite, dependent for our existence on the other disciplines that we couple together in the interests of innovation. If this is the case, then we must remain not only diplomats (as Kim observes) but irritants. Articles such as this one are potentially annoying distractions from HCI business as usual, but I believe that CHI itself is a perennial irritant within its host discipline of ACM-defined computer science. By drawing attention to social obligations, intellectual flaws, playful creativity, and alternative kinds of knowledge, HCI does not make computer science a more comfortable place. If we continue to pay attention to this core mission, we are unlikely to become a discipline. However, if we can be an effective interdiscipline, we will fill a very important hole.

References

1. Kostakos, V. The big hole in HCI research. Interactions 22, 2 (Mar.–Apr. 2015), 48–51.

2. Reeves, S. Locating the ‘big hole’ in HCI research. Interactions 22, 4 (Jul.–Aug. 2015), 53–56.

3. Blackwell, A.F. HCI as an inter-discipline. Proc. of the 33rd Annual ACM Conference Extended Abstracts on Human Factors in Computing Systems. ACM, New York, 2015, 503–516.

4. Kim, S. Interdisciplinary cooperation. In The Art of Human-Computer Interface Design. B. Laurel, ed. Addison-Wesley, 1990, 31–44.

5. Stolterman, E. and Wiberg, M. Concept-driven interaction design research. Human–Computer Interaction 25, 2 (2010), 95–118.

Author

Alan Blackwell is professor of interdisciplinary design at the University of Cambridge Computer Laboratory. While practicing as a professional engineer, he has studied comparative religion, computer science, music performance, and cognitive psychology. His current research interests are in visualization, digital music, and critical technical practice. afb21@cam.ac.uk

Figures

F1Figure 1. Mapping of 180 collaborative interdisciplinary projects to CHI topical review subcommittees.

Copyright held by author. Publication rights licensed to ACM.

The Digital Library is published by the Association for Computing Machinery. Copyright © 2015 ACM, Inc.

Post Comment


No Comments Found