Moving Learning Systems Toward "Embodied Technologies"

Paul Dourish, in his book Where The Action Is describes the evolution of tactile and social computing toward his notion of embodied interaction, which connects phenomenology - a philosophical point of view put forth by a number of folks in the past 150 years - to interaction design. Dourish details the three aspects of meaning - Ontology, Intersubjectivity and Intentionality - as the focus of his explication of embodied interaction.

Embodied Interaction begins to illuminate not just how we act on technology, but how we act through it according to Dourish. This theory is in contrast to the dominant theoretical basis of HCI research prior to Bonnie Nardi's "Context and Consciousness" in 1996, which mostly included mechanisms for analyzing human computer interaction as a set of as disembodied, rational problems. What Dourish proposes is engaging not just the interaction between 'human and computer', but understanding how such interaction is part of evolving phenomena. How people appropriate (apologies to Don Norman for using that term in yet another manner he probably did not intend) tools for use becomes an essential component of their use.

What does this mean and how can it be applied to understanding and examining software that's in use? It means we must examine how users apply tools in a way that results in changes to the user, the tool and the context. This is not, "can the user open a word processor and create a letter?". This question is "how does the user act through the technology, and what specific thinking about the tool is required to access the intended features?". Intent is the intent of the designer. To the extent the tool matches metaphors or mental models the user is familiar with, less explicit thinking about the tool will be required. This is neither good or bad, but it is a measure of how embodied interaction with a particular tool is likely to evolve. If a tool matches an existing mental model, it'll be adopted according to that model. If a tool isn't easily placed in a users current way of thinking about tools for a particular task, the user is probably more likely to apply the tool in a manner never intended by the designer.

This cuts both ways. An innovative tool that the user is able to think about in a way they are familiar with might never get adopted, or will get adopted "poorly and inappriately" because it breaks that users mental model. A classic example is the use of desktop databases by avid spreadsheet users in the 1980's. Anyone fortunate to be around at that time will remember running across the accountant, engineer or other skilled professional who moved their spreadsheet lists into a database and proceeded to manipulate the tables in the same manner they manipulated cells in the spreadsheet; Often complaining about the lack of features. Mental model = "Spreadsheet"; Application = "Database". User experience = Deceptively similar at the surface.

Embodied interaction is not an artifact of the user, nor is it an artifact of the tool. Rather, it is found in the marriage of the two. Applying it to the evaluation of a particular tool, such as Sakai, means that we must keep the spreadsheet-database metaphor in mind.

With this spirit and questions like this in mind, I undertook to assess Sakai, an open source course management system. I turned a keen eye toward the extent to which Sakai demonstrates embodied interaction in practice, and the extent to which its design is lacking. Since Sakai is a collection of tools, I focused on the Chat and Discussion sections, which are among the most commonly used.

Chat in Sakai
Chat is a synchronous tool that is used for immediate communication between two people. Archived chat, like the system in Sakai, is helpful because chat topics are maintained in a searchable history, contextualized by Sakai course. This feature is useful.

Sakai's chat falls short of the embodied ideal, which I choose to assess using the question, "Does the user act through this technology or is explicit thinking about the tool required to access the intended features"? There are three specific deficits that Sakai Chat demonstrates in this area:

1. When you see users who are also online, the embodied way to "act through the technology" would be to click on the user, and invite them to a private chat which could be catalouged and searchable. Sakai shows the users who are present, but denies you the ability to "use" this information. It is entombed interaction, which commonly leads to frustration, as reported by several Sakai users
2. When actually using the chat features, the application behaves slowly, so instead of "acting through" the technology, here again the user is having to think, and experiences frustration.
3. Finding the chat function is actually quite difficult to begin with. It's listed in the context of a specific course, and requires a number of clicks. For something that is generally intended for synchronous interaction, this is almost absurd, and certainly nothing approximating Dourish's embodied ideal. Though this functionality is visible in the course menu tab, nine out of nine students interviewed during the past two semesters at the University of Missouri regarding their experiences collaborating in Sakai have indicated that the chat function is hard to find, engage and use.
   

Discussion Board in Sakai
The JForum discussion board is used in the implementation of Sakai that the University of Missouri - Columbia uses. This particular discussion board is more well suited to it's purpose and embodied interaction than the chat function.

For example, to answer the question "Does the user act through this technology or is explicit thinking about the tool required access the intended features"? In the case of the discussion board, users expect to be able to add a thread and respond to a thread. The discussion board tool does this well, and observation of Sakai classes has proven that users of the system are able to, in some ways, act through the technology by communicating fluidly, cogently and adapt the tools to the immediate purposes of their communication. In the case of the discussion board, this is possible.

Discussion
One consideration when exploring the extent to which embodied interaction is made possible by a given toolset is the prevalence of previous mental models for that type of toolset. In the Sakai example, there are a number of existing synchronous chat models, and Sakai ignores them all to it's own detriment. In the case of discussion boards, there is also an established model of tool use, and Sakai's conforms well to that standard. In this way, even if the technology is not fluid or intuitive from the outset, because it mimics tool models for that type of interaction which the user is familiar with, the discussion module fares better than chat.

How Embodied Interaction Could be Better Reflected in Sakai
Like every learning management system I've ever used, Sakai applies tools that were first conceived in the early days of computer bulletin boards and compuserve, wraps them in a traditional classroom context and deploys the combined meta-metaphor for online learning. Embodied interaction, to fulfill Dourish's vision, will need to re-imagine what a learning management system is. This will mean, I think, acting to combine research in electronic community formation and electronic tutors as well as what we know about the social nature of learning to develop tools that account for these important, human characteristics in a tool.

Embodied interaction is important to apply to the design of future versions of any learning management system. Designers must be careful to not be deceived by tools they are familiar with, and apply them wholesale in different contexts. Designers must beware of how they apply or don't apply metaphors of the physical world (classrooms, assignments, etc.) to the design of learning management systems. Most importantly, designers must send explicit signals (signs, semiotics, visual & intuitive guides) to users so that users are able to discern the "best way" to apply a particular tool to the task at hand. Finally, efforts to understand how users actually use Sakai will provide an invaluable feedback loop through which new designs will emerge.

Key References
Dourish, Paul. Where The Action Is