Ambiguity as a Resource for Design

Gaver, B., Beaver, J. and Benford, S. Ambiguity as a resource for design. in Bellotti, V., Erickson, T., et al. eds. Proceedings of CHI 2003, Ft. Lauderdale, Florida, ACM Press, New York, NY, 2003.

In HCI ambiguity is often perceived as the nemesis of usefulness and usability. Instead of regarding ambiguity as a problem, the authors suggest that it can be sees as an opportunity in the context of emerging application for everyday life (alternatives to task-oriented forms of ubiquitous computing). Indeed, ambiguity can make a virtue out of technical limitations by providing the grounds for peoples interpretations to supplement them.

Ambiguity is an attribute of the interpretation of fuzziness or inconsistency. Things themselves are not inherently ambiguous. However, they may give rise to multiple interpretations depending on their precision, consistency and accuracy on the one hand, and the identity, motivations, and expectations of an interpreter on the other.

The authors distinguish three principal kinds of ambiguity: ambiguity of information, of context and of relationship. Ambiguity of information is of prior interest with my focus on spatial uncertainty. Gaver et al. provide an example of Bystander, a mixed reality game in which mobile player’s location is tracked using GPS data, which is prone to errors:

Thus the issue is not ‘What does this display mean?’ but ‘Do I trust it?’ – that is, how does the display correspond with reality?

The traditional response to ambiguity of information in interactive systems like Bystander is to improve the technology, use statistical methods to set certainty thresholds, or ignore it and hope for the best. Bystander, in contrast, passes the ambiguity directly to players in the form of fuzzy avatars that hint at locations without specifying them. Rather than seeing uncertain GPS information (in this instance) as a flaw, Bystander treats this ambiguous information as a challenge to users, forcing them to join their knowledge of people and cities to the clues offered by the system to play the game.

There a different tactics to use ambiguity and compel people to join in the work of making sense of a system and its context:

• Use of imprecise representation to emphasize uncertainty: display information that is physically or conceptually blurred.
• Over-interpret data to encourage speculation: draw attention to possible truths rather than simple untruths
• Cast doubt on sources to provoke independent assessment: Over-interpretation and inconsistency are special cases of increasing ambiguity by casting doubt on sources of information.
• Expose inconsistencies to create a space of interpretation: the juxtaposition of incompatible elements require viewers to build their own meanings form the display.

Exposing inconsistency between sources of information (e.g. external sensors) can also be a powerful alternative to trying to resolve or hide it. For instance, just as the uncertainties of GPS tracking in Bystander are shown to participants, so are inconsistencies revealed between GPS and spoken information. By crafting the inherent limitations of the technology into the user experience, they become a means to heighten narrative intrigue by opening a space of possibilities that participants must navigate.

Relation to my thesis: The traditional response to ambiguity of information is to improve the technology. Enhancing ambiguity of information is a design tactic is a step beyond seamful design, because it challenges the users to participate in making meaning. It focuses on creating or reflecting uncertainties about information that are noticeable to people. In my context, I try to understand how uses can cope with inaccurate sensors, inexact mappings and low-resolution displays. Encouraging them to supplement uncertainties with their own interpretations and beliefs raises questions on in what context and how much workload can we transfer to the users. How much transfer can be imposed and how much can be suggested? Nevertheless, I share the approach of using ambiguity to leverage immersion in the task while keeping a certain independence (doubt) to the system.