[Tom's Home Page]
[Professional] [Life,
Fun, &c] [Tell Me...]
[Bookmarks] [Publications
List] <and many papers and essays>
This is a two-page "late-breaking" paper currently under submission to the CHI '98 Conference. Due to length limitations for this type of submission, I've probably tried to pack too much into this; I plan to expand this into a much longer paper, so comments are welcome. If you're interested in learning more about genre and on-line conversation see reference 1, Virtual Community as Participatory Genre.
Thomas Erickson
IBM T. J. Watson Research Center
P.O. Box 704, H1-M09
Yorktown Heights, NY 10598 USA
snowfall@acm.org, www.pliant.org/personal/Tom_Erickson
This paper presents an example of network-mediated interaction and uses it to introduce genre theory and motivate its application to HCI design and research.
Genre, conversation, conversation analysis, computer mediated communication, CSCW, virtual community.
My goal is to understand the conditions behind deep, productive, coherent interactions among large groups, so that we might design better systems for supporting them. In this paper I describe genre theory, a way of looking at mediated interaction among large groups. I begin with a simple, yet rich, example of collective interaction.
The following example is from a text-based, asynchronous conversation in "Cafe Utne", a public on-line 'salon.' The first message-beginning a new conversation in a conference area called "Fun"-explains what is happening:
Fun.64.1: BH <BH's account name> Tue, 03 Dec 1996 12:03:31 CST (12 lines)
Here's a fun game. We write limericks, each person contributing a line at a time. You'll recall from this example that limericks rhyme and scan (iambic pentameter, and all that) a certain way:
...[an example of a limerick is omitted]
Limit your contribution to one line at a time, at whichever point the limerick is at when you happen by.
Over the next six hours, five people created the following limerick: (For brevity, the individual message headers that begin each line have been replaced by the writers' 'initials.')
BH: I'll start:
An Internet surfer named Joe
WCC: Enjoying the World Wide Web flow,
CUP: Got hooked on a site
ENL: stayed there day and night
CLM: 'til his mother said, "Time to go, Joe"
Over the next 24 hours six new participants join with the other five to successfully compose two more limericks.
Here we have eleven strangers working together to achieve a coherent end. With only minimal preparation, the interaction goes smoothly and the results are as intended: well-formed limericks. And it even seems like fun.
Too bad it's such a trivial example. Or is it? There's a lot going on. Consider the making of the next limerick:
KMO: A fellow who's hair was bright orange,
WCC: [Spoilsport!]
KMO: Sorry, couldn't resist! :-)
WCC: Okay, Karl, start us again. For real.
KMO: There once was a "spoilsport" poster,
LFB: Who played games with a rogue and a boaster,
ENL: he sent him a flame,
LFB: for wrecking the game,
KMO: And then died sticking forks in a toaster!
To interpret this, readers need to know that the word "orange" has no rhyme in English, and that this fact is well known amongst the more expert English speakers who are drawn to such games. KMO has cunningly disrupted the game by making it impossible for anyone to compose a valid next line. WCC chastises KMO (using brackets to signal that she has stepped outside the limerick form), KMO apologizes, admitting the intentionality of his act, and WCC invites him to "start us again. For real."
Although this example is simple, it is not artificial. The limerick conversation has much of the richness that characterizes more complex and ordinary conversations: people are taking turns, following rules, breaking rules, using etiquette, 'going meta' to talk about the process, etc.
Why does this conversation work? (As of this writing, the conversation has been going on for about a year, with over 2300 contributions.) For we can examine many on-line conversations on the internet-in mailing lists, bulletin boards, and virtual communities-but we will find few cases in which strangers work together so smoothly, with so little preparation, to produce such a coherent result.
One reason for the success of the conversation is that the participants understood 'the rules': they knew they were making limericks, they had a sense of what limericks were for (i.e. fun), and they knew the patterns of rhyme and meter that make a limerick. As a consequence, at any point in the limerick-making conversation, it was possible to say where the conversation was with respect to its goal, and to understand what needed to be done next to move it in that direction (or to disrupt it!). In addition, participants had a very simple interaction model: one person makes one line at a time. Finally, participants assumed that their understandings of the process and content of this interaction were shared by the other participants. Thus, the disruption-and the subsequent reprimand-were immediately recognized as what they were (rather than as inept attempts to do limericks).
This case is important because it is an especially clear cut example of how shared understandings enable groups to communicate effectively and coherently. I suggest that it is useful to look at any kind of mediated interaction in terms of these shared understandings, and that an existing approach known as genre theory is well suited to this end.
Traditionally, genre theory has been used to categorize various types of writing, speech, and other media forms. It analyzes information artifacts in terms of their communicative purpose, and the regularities of form and substance which characterize them. Thus, resumes, to take a different example, have a purpose, a recognizable form (short, with highly structured content), and an expected type of content (information about the writer's qualifications).
Over the last two decades, scholars in rhetoric and literary criticism have developed a situated form of genre theory (e.g. [2,3]) which makes it particularly useful for our purposes. They suggest that what is important in understanding a genre is identifying the underlying social and technical forces which produce the regularities which characterize a genre. Thus, the regularities of the resume genre arise from a mix of social influences (e.g. resume readers are often trying to read quickly, so structure and brevity is desirable), and technical influences (e.g. desktop publishing enables the use of fonts and styles and indentation to produce highly structured text). Likewise, the form of the collective limericks resulted from social factors (e.g. an agreement to take turns), and technical factors (e.g. the properties of the Cafe interface-note that it would be difficult to do collective limericks via a mailing list!).
A final aspect of situated genre theory is that a genre is understood, enacted, and, over time, shaped, by a group of people who play various roles in the enactment of the genre-"a discourse community." A discourse community may range from a few people playing one or two roles (as with the collective limericks) to a vast, diffuse set of people with a multitude of roles (as with those involved in the creation, transmission, and consumption of resumes, and the mini-industry of classes, books, etc., that offer guidance on their preparation and use).
Genre theory lends itself to a number of applications in HCI. My interest in genre theory began while I was trying to study "virtual community." Genre theory was useful because it shifted the focus from community (e.g. the nature of relationships among community members) to the information artifact produced by the community's interactions. This shift, in drawing attention to how sociotechnical forces shaped the community's interactions, provided a number of insights about interfaces to a virtual communities (see [2]). Similarly, genre theory has been used to analyze organizational communication [4]. Genre theory also applies to what the high tech industry calls "content:" rather than treating content as an isolated entity, it suggests examining "content" in terms of how it is produced, consumed, and understood by its discourse community. Another application is to understanding web sites. Web sites often borrow genres from other media (e.g. the newsletter genre from print), but these borrowings often fail because all that has been borrowed is the look of the genre-the sociotechnical underpinnings are different.
When applied to the digital domain, genre theory raises some provocative issues for HCI researchers. Traditional genres have a marked separation between producers and consumers, and such genres tend to evolve slowly. Neither of these conditions holds in the digital world. Digital genres have the potential to be much more participatory, and to hence evolve much more rapidly. An intriguing challenge for HCI researchers and designers is to explore the implications of these changed conditions for digital genres.
Thanks to Leha Blaney (LFB), Bryan Higgins(BH), Clydie Morgan (CCM), and two other participants, for permission to use their words. Cafe Utne is at http://www.utne.com.
[Tom's Home Page]
[Professional] [Life,
Fun, &c] [Tell Me...]
[Bookmarks] [Publications
List] <and many papers and essays>