Metaphor in User Interface Design: References

This is an annotated collection of references about the use (and mis-use) of metaphor in user interface design, which I collected while preparing a presentation for James Landay's Research Topics in Human-Computer Interaction class. The order of these items is roughly my order of preference.

George Lakoff and Mark Johnson, Metaphors We Live By, University of Chicago Press, 1980.

A great little book, very accessible, for the most part, to non-linguists (such as myself!), yet deep and insightful. Lakoff and Johnson argue that our conceptual system is fundamentally metaphoric -- the way that we understand and perceive is structured by metaphoric concepts. They state, "the way we think, what we experience, and what we do every day is very much a matter of metaphor." (If that's hard to swallow, please don't be put off -- the book offers much that does not require you to accept this thesis!)

L&J given many examples of metaphor, that we use in everyday discourse: argument is war, love is money, time is a valuable resource, happy is up/sad is down, the mind is a machine, theories are buildings, and dozens of others. They explain how metaphors form a coherent and systematic basis for underatanding, and how the particular metaphors we choose color that understanding. For example, they ask us to imagine a culture in which argument is a dance instead, and how different would be the approach to argument in that culture.

Metaphor has used and unused parts. For example, the theories are buildings metaphor uses the concepts of "framework" and "foundation," but not other aspects of a building: his theory has a problem with the plumbing illustrates the unused part of this metaphor. (This is something that both sides of the UI metaphor debate appear to have missed.) New metaphors provoke new new insight -- they give, as an example, love is a collaborative work of art, and illustrate how the entailments of this new metaphor create a new reality if you choose to live by it.

Aristotle said something like "Ordinary words tell only what is already known; it is only through metaphor that one can grasp something new."

Kim Halskov Madsen, "A Guide to Metaphorical Design," Comm. of the ACM, 37(12), December 1994, pp 57--62.

Halskov Madsen uses the "pragmatic approach" to metaphor of Carroll, Mack and Kellog (which is the only one that makes sense to me). She/he gives five examples of the use of metaphor in system design, and point out that users construct metaphors to explain system operation whether given one or not. In one example, this was used to then explicitly add the metaphor to the interface in order to support users' understanding of how the system worked.

Halskov Madsen understands the complexity of metaphor, and provides guidelines for developing and working with metaphor. One I particularly like is "Do not necessarily explicitly incorporate the metaphor in the final design." The metaphor can be used to develop new ways of interacting with the system, but sometimes it is better to throw the original metaphor away. Choose multiple metaphors -- in one example, several metaphors were used to represent a digital library, and the interaction and differences between them in fact stimulated reflection and criticla awareness. Choose metaphors with a conceptual distance between the source and target domains in order to help users see things in a new way -- few of the other writers on metaphor in UI understood this.

Gerald J. Johnson, "Of Metaphor and the Difficulty of Computer Discourse," Comm. of the ACM, 37(12), December 1994, pp 97--102.

Johnson writes a great little article in which he demonstrates the ubiquity of metaphor in computer discourse, and the difficulties associated with this. Because computer concepts are a) new, and b) abstract, we must use metaphor to understand and explain them -- examples include gateway, network, address, pointer, agents, scripts, server, client, interface, platforms, and so on. He uses a quotation from Larry Ellison to illustrate their use even in informal discourse. Because of the number of them, we mix metaphors wildly. Johnson warns that we need to be careful to keep the metaphors metaphorical, and not literal.

John M. Carroll, Robert L. Mack and Wendy A. Kellogg, "Interface Metaphors and User Interface Design", in Handbook of Human-Computer Interaction (first edition), M. Helander (ed.), Elsevier Science, 1988, pp 67--85.

Unfortunately, I hadn't read this article before recommending the readings for the metaphor in UI class. It would have been a good choice, but I had only seen the more recent article by Neale and Carroll.

CM&K point out in their introduction that metaphors, "by definition must provide imperfect mappings to their target domain," and that composite metaphors are the norm. They outline three approaches to metaphor research in users interfaces: operational, structural, and pragmatic. Operational approaches focus on measuring the effects of employing metaphoric representations. Structural approaches try to understand what metaphor is and to find ways to find the mappings between source and target domains. It asks questions such as the appropriate grain of representation and how particular metaphor properties relate to congitive aspects of their use. Pragmatic approaches to metaaphor focus on the things such as the incompleteness and mismatches of metaphors and how these affect use in complex real-world situations. In partciular, they focus on how "structural flaws" play a useful role.

CM&K then propose a method for designing with metaphors that tries to find likely metaphors that users will come up with anyway. They identify matches and mismatches in the metaphor, and attempt to assess the "goodness" of the metaphor. (They point out that mismatches can in fact be good.) They suggest the use of composite metaphors to resolve mismatches.

Dennis C. Neale and John M. Carroll, "The Role of Metaphors in User Interface Design," in Handbook of Human-Computer Interaction, second edition, M. Helander, T. K. Landauer, P. Prabhu (eds.), Elsevier Science, 1997, pp 441--462.

This paper is completely different to the paper in the first edition of the Handbook. As reading to understand metaphor, I didn't find this article useful. As a source of references to current research and a jumping-off point to a more extensive reading program, however, this paper would be invaluable.

Alan Cooper, The Myth of Metaphor, in Visual Basic Programmers Journal, June 1995.. This appears to be a short version of Chapter 5 of his book, About Face: The Essentials of User Interface Design, IDG Books, 1995.

Cooper is typically blunt in this short article. He offers three paraidms of UI development: technological, in which the UI reflects the way that the program is structured underneath; metaphorical, in which the user interface is structured in terms of metaphors and visual representations; and idiomatic, in which the user interface is constructed from a collection of complementary idioms, such as close boxes, and the way that the mouse works.

Cooper has a lot to say about how bad a UI design tool metaphor is; he is, however, focussing on structural and literal metaphors, such as "the desktop looks and acts like a real desktop" (which, of course, it doesn't). I think Cooper is right that in criticising that particular use of metaphor, and that the interaction idioms are important, but he uses "metaphor" in a very specific and narrow way.

Aaron Marcus, "Human Communications Issues in Advanced UIs," Comm. of the ACM, 36(4), April 1993, pp 101--109.

This article is a hodge-podge of semiotic mish-mash, lists of literal metaphors in user intefaces, and half-baked suggestions of new uses of "metaphor." (Sorry, but that's what I think.) Marcus suggests ridiculous "metaphors" such as visiting planets or star systems, or lassooing cattle, to select files. To him, metaphor means using, for example, a lab workbench metaphor for engineers, and a trading floor metaphor for stockbrokers.

But then again, who knows -- maybe that's what we'll get. Microsoft could build it into Office...

Steven Johnson, Interface Culture: How New Technology Transforms the Way We Create and Communicate, Harper San Francisco, 1997.

In pages 42 to 65 (first half of Chapter 2), Johnson discusses the history of the desktop and the use of metaphor. He outlines the progresssion (or deterioration...) of the desktop metaphor from a new insight (the jump from tiled windows to overlapping windows) to over-literal simulations. I find him interesting reading, although, as James Landay pointed out, he is not an authority on user interfaces, so you should make your own interpretations.

Keith J. Holyoak and Paul Thagard, Mental Leaps: Analogy in Creative Thought, MIT Press, 1995.

I bought this book because Neale and Carroll referred to it when stating that, in contemporary research, metaphors are conceived of as "cross-domain mappings." The first chapter is interesting reading, but I have got no further than that yet.

Kenneth E. Mohnkern, Affordances, Metaphor, and Interface Design, Masters Thesis, College of Fine Arts, Carnegie Mellon University.

A nicely-written short thesis, which I didn't get hold of until after the presentation.