Tuesday, December 19, 2017

Why a 'gap' in a field is not an argument for research

One quite common argument used in research papers is made by showing that there is a 'gap' in the field that no one or few have studied or researched. This argument is flawed in many ways. It is built on some assumptions that do not make sense. For instance, it is built on the assumption that the whole 'field' (whatever that means) need to be equally well researched. It also assumes that areas that have been researched do not require the same attention as other areas.

We know from the history of science that a field is never researched completely or finished. In the decades before Einstein, there was a growing sense in physics that the field was done, that the world of physics was more or less completely understood. So, if Einstein had followed the advice of only study gaps, his revolutionary theory would probably not have been developed. Instead, he studied the area of physics that was perhaps most developed, most complete, and the most popular. There was no gap for him to approach. His ideas revolutionized physics.

In my own field, HCI, the idea of studying gaps is extraordinarily strong. It seems as if the idea is that only by studying something less researched there is a chance to make a contribution to the field. This has led to a field that mainly develops its knowledge horizontally, that is, by adding new aspects or phenomena to the repertoire of study. We see much less of vertical knowledge production, that is,  in-depth studies of areas where we already have substantial knowledge.

There have been some attempts in the field to advocate for more vertically oriented knowledge production (such as Repli-CHI) but in general, the aspiration for the new and the novel in combination with the idea of the 'gap' is apparently too strong. For a dynamic field like HCI this is unfortunate and may not, in the long run, lead to a foundation of knowledge that is stable and sustainable and that can deliver increasingly deeper insights about the relation between humans and machines.

Friday, December 08, 2017

Interactive species: GOFIs, Things and Beings

At the end of our book "Things That Keep Us Busy--the elements of interaction" (Janlert and Stolterman, 2017, MIT Press) we spend some time speculating about the future of interaction. One of the ideas we present is to consider three forms of interactive 'species'. We claim that even if our ideas are speculations, they are not pure fantasy, actually, we argue that they are logical consequences of the examination of the nature of interaction that we engaged with throughout the book. So, here are a few pages that present some of these ideas (page 198-202).


"11.1 Things and Beings

Any attempt to imagine what may lie ahead easily becomes science fiction or pure fantasy. Not least when it comes to interactivity—a popular topic of futuristic portrayals in science fiction movies. It is exciting to imagine futuristic scenarios where the methods and patterns of interaction have completely changed due to some unknown technology. It is tempting to imagine future forms of interactivity that relieve us of all the complications and issues we have discussed throughout this book—without considering how realistic they may be. Even though we believe that science fiction can stimulate technological development and invention, and indeed has in many ways influenced our field, we will refrain from wishful futuristic thinking. Instead of imagining radically new forms of interactivity without connection to our present situation, we are rather trying to extend what we already have observed in existing and developing technology.

Let us also make it clear that we do not in any way foresee the demise of “traditional” interfaces and, let us call them, GOFIs, Good Old-Fashioned Interfaced artifacts and systems with dedicated interaction areas occupy- ing a limited part of their surface. GOFIs are likely to continue to be used even as faceless interaction and other nontraditional solutions become common. One reason is that a clearly located user interface generally has lower risk of accidental interaction and may simplify handling of the artifact as a physical object. By keeping the object’s “smart parts” separated from the “dumb parts,” users can have better control and confidence regarding when and which operations are actually performed. Another reason is that a clearly defined and visually recognizable interface enables users to quickly see similarities with other interfaces and draw on earlier experiences.
In the present situation we can discern the emergence of two novel species of interactive artifacts and systems, different from the ordinary GOFIs: one development toward “Things,” another toward “Beings” (we will spell them with capital initials to distinguish them from things and beings in general).

The Things line of development means abandoning the traditional dedicated, surface-bound user interface for artifacts that can be interacted with in a fashion similar to how traditional small- or medium-sized “dumb” interfaceless things, natural or artificial, are handled—that is, by being moved, squeezed, thrown, shaken, folded, twisted, rubbed, bent, and so forth, just like a chair, a pillow, or a piece of paper can be interacted with in many ways without any designated area for interaction. Some of these Things may be in more or less magical and mystical rapport with other Things. Lately, we have seen a lot of effort in HCI research (embodied inter- action, tangible interaction, ubiquitous computing, Internet of Things, and more) that can be viewed as work in this direction.

While the role model for Things is ordinary, nondigital things, the point is of course that Things have nonordinary and perhaps extraordinary properties and qualities. Given the present state of technological development, we can for instance reasonably expect to see objects entirely covered by some relatively cheap touch-sensitive display layer on top of some smart shape-changing material, equipped with various kinds of sensors and micromotors, enabling the object to change its shape and physical configuration, color, and pattern in a controlled manner under the impression of external forces and sensations—and still without adding a traditional interface. Furthermore, today wireless access to and delivery of local and remote information are already very much taken for granted. Yet, in the end we do not think there will remain a sharp dividing line between things and Things; as Things become common their once extraordinary and marvelous properties will come to seem more ordinary. What initially will set Things apart are above all their expressive and impressive abilities, particularly their dynamic, live impressions and expressions, and their ability to offer interactions that (to begin with) challenge everyday experience in unexpected and interesting ways (e.g., when you push, instead of yielding, the Thing might move in the opposite direction, contrary to the applied force). But these expressive-impressive abilities do not take the form of a developed symbolic language as we have been accustomed to in the interfaces of GOFIs, and as interactants their agency is weak—it still makes sense to think of them as “things.”

The introduction of Things into everyday life means that we will encounter new forms of interaction, and interactivity will appear in many places where none was expected before. Instead of turning on the room light with a fixed wall switch, in the future you might turn it on by tapping or stroking the wall anywhere in certain ways; you might unlock or lock the door by pressing your palm against it; you might adjust the height of the tabletop by nudging it with three fingers in the desired direction—and so on, and on. A chair might groan when you sit down if you are overweight, whine if you jump up and down on its seat, and lock its wheels (if it has wheels) if you step up on it. What used to be ordinary things may be equipped with interactive and expressive abilities that are related to their physical and tangible qualities, their materials, shapes, and forms. We will see Things that can change their appearance in ways that serve functional purposes as well as deliver expressions even at a quite nuanced and subtle level. We may encounter a chair that expresses sadness through some slightly drooping shape change, perhaps because it is in disrepair or because it commiserates with us. We may be able to interact with Things by expressing emotions through our manner of physically handling them, by our posture or intonation, for instance. There may be nothing very remarkable in any single example, but when Things are everywhere they will transform everyday life.

The other line of development we call Beings. In contrast to Things, Beings have stronger agency and may also have elaborate and sophisticated language-like methods of expressing themselves and be impressed by their users symbolically. Some may even have GOFI-like interaction areas (some robots come with an integrated screen for displaying texts and images, for example). While a Thing is basically dumb and has only a limited and fixed repertoire of behavioral patterns, a Being is smarter and can have a richly varied, adaptable behavior, capable of development. Simple Beings you might shoo at or pat; more advanced Beings you might strike up a conversation with. In any case, to interact with Beings should be more like getting along with your dog or cat than dealing with your furniture.1 But just as with Things and things, there may be no sharp line separating Beings from Things. Rather, we imagine an unbroken chain of entities at different “levels of existence,” stretching from things over Things to Beings (again reminiscent of the old idea of the great chain of being; see our earlier comment at the end of section 7.1). And why not as in earlier times indulge ourselves by chauvinistically putting humans on top of Beings, as a kind of superBeings (but let us stop there and go no further).

The dream of infusing intelligence into the things around us goes far back, but it has always led to mixed feelings. To be surrounded by “beings” that you can relate to in a supposedly more “natural” or “human” way, some see as desirable, others as a nightmare. This is also a favorite theme in many science fiction narratives. The omnipresent Being called Hal in the movie 2001: A Space Odyssey evolves from being the perfect servant to a mortal enemy with dubious moral instincts. The prospect of future artificial “superintelligence” and the potential danger for humankind it poses, as analyzed and discussed by Nick Bostrom (2014), should certainly be taken seriously.

Even though we are not yet living in a world of superintelligent Beings, a number of perhaps minor but still significant steps have already been taken on the road to populate our environments with Beings. There are numerous examples of toys, social robots, and certain everyday objects like cars and homes that already behave like Beings. They respond to commands, they perform actions based on our desires, they engage in conversations, they can proactively suggest your next activity. They may still be somewhat experimental and not very broadly used, but a rapid spread of more advanced Beings in everyday life does not seem unrealistic.

Some Beings will be able to remember us, they will know who we are, understand our needs, understand what we are doing, and they may even persuade or force us to behave and do things we are not eager to do, like diet, study, or drive under the speed limit. They may become our partners, our allies, our superegos, or “parents.” In some cases, we will interact with them through traditional surfaces or gestures, but in many cases language will be the primary mode of interaction, in some cases developed into what could be called conversations. To what extent our average future Being will be an eloquent conversation partner is not clear. We can already converse (although primitively) with our car about where we want to go, how we want to be entertained, and with whom we want to communicate. This interaction may increase our perception of the car as having a character (as we examined earlier). Whether we will experience this Being, the car, as a servant or boss (or something else) is another issue. When and where conversational interaction will be of any use is still an unknown and will probably continue to be difficult to predict, partially because it is to a large extent a consequence of what is culturally and socially accepted behavior. We have seen how the use of mobile phones in public have evolved since its initial days, not because the interaction has changed but as a result of changing social norms.

Things and Beings have in common that interaction has ceased to be a matter of having detailed knowledge about precise operations and their effects (with or without a designated interface) and instead becomes a matter of understanding and interpreting primitive reactions, expressions, and behavior patterns (Things), or objectives, needs, intentions, and plans (Beings)—and of behaving in a corresponding fashion in relation to the artifacts and systems. Living with such Things and Beings, we are undoubtedly getting closer to the animism of Toontown—even though not necessarily to its frenzy."

Tuesday, December 05, 2017

Practical (design) reasoning explained (Martha Nussbaum)

After quite many years I am re-reading an essay by Martha Nussbaum. The title is "The Discernment of Perception: An Aristotelian Conception of Private and Public Rationality" (to be found in the book "Love's Knowledge--essays on philosophy and literature" published in 1990). This essay helped me a lot when it was first published and it has influenced my thinking over the years in so many ways. It is therefore great to re-read it carefully now, many years later and realize that it is even better now.

Even though the title of this essay may scare some people with its complexity and reference to Aristotle, the essay is in my view one of the best texts ever written about practical reasoning and judgment. It is an essay that resonates perfectly with anyone who is reflecting on design practice and how designers reason, think and make judgments.

Nussman discusses why practical reasoning is not possible to understand with some simplistic (scientific)  form of logic. She builds her argumentation on the writings of Aristotle and his "attack on scientific conceptions of rationality". She summarizes her intention at the beginning of the essay by stating:

"I shall suggest that Aristotle's attack has three distinct claims, closely interwoven. These are: an attack on the claim that all valuable things are commensurable; an argument for the priority of  particular judgments to universals; and a defense of the emotions and the imagination as essential to rational choice."

Nussman then goes through these claims and explains how they lead to a definition of practical reasoning that is distinct, understandable and useful. This understanding of practical reasoning fits extraordinary well with the reality that designers face, when dealing with overwhelming but insufficient information, in their dealing with particulars and not universals, and having to rely on imagination and accept being influenced by emotions.

Just read it!!

------------------------------- Addition --------------------------

Ok, today I found my notebooks from earlier years and randomly pick one up, and randomly open up a page. At the top of the page I had written: "Good idea for an article, based on the notions of private versus public rationality by Martha Nussbaum."

Then a note (translated from Swedish): "No one has pushed this [rationality] far enough, not Churchman, not SSM [Soft Systems Mthodology]. Everyone is trying to start with how the world is, while Nussbaum starts with how people are. An article idea: How to manage systems design: the conflict between private and public rationality."

So why did I read Nussbaum yesterday and why did I happen to see that page today? Synchronicity...