Narrative methods as supplement to field experience
CSCW workshop position paper on narrative methods
Published on: Mar 3, 2016
Transcripts - Narrative methods as supplement to field experience
Using Story Templates as a Method to Cumulate Knowledge in HCI and International Development John Thomas IBM Research PO Box 704 Yorktown Heights, NY 10598 USA firstname.lastname@example.org www.truthtable.comEffective design work is always a challenge but this challenge is multiplied whensomeone attempts to design systems that will be used in an entirely different context fromthe design team’s own context. One design situation that is particularly challenging inthis regard is to design systems for the so-called developing world. Designers in the so-called developing world, for a number of reasons ranging from altruism to greed, areevidencing increasing interest in doing just that, particularly for systems that requirehuman-computer interaction. The usual wisdom is that good HCI design relies onknowledge of the users, as well as their tasks, contexts, and goals. People in the so-calleddeveloping world may live in very different physical contexts from the design team; forinstance, users may live in situations of extreme heat, cold, or humidity. Further, theymay live in a social and political context that is unfamiliar to the designers. Thedesigner’s assumptions about infrastructure may also be mistaken; e.g., it may be quitedifficult to acquire a battery or a screwdriver. Adoption of new technology may dependon acceptance by elders, local government, or other authorities and have little to do with“consumer choice.” Not only may the tasks and goals of people in developing regions beunfamiliar to the designers, but the very way in which tasks and goals are understood andmeasured may be unfamiliar. For example, most participants at CSCW implicitly assumethat “efficiency” is a desirable characteristic of a good system. But in some parts of thedeveloping world, a “solution” that requires ten people to work (and therefore provideswork for ten people) may be more desirable than one that only requires one person.Attempting to resolve the issues implied by these quandaries suggests that design teamsmust gain considerable local knowledge in order to do a credible job. However, there aremany practical issues involved in doing this. It is expensive. It is disruptive of the livesof the designers and their families. Requiring extensive first-hand knowledge of everynew situation calls into question whether or not HCI can really be a cumulativeenterprise. If knowledge can be cumulative, that will probably require a range ofmethods of creating and sharing knowledge beyond the “traditional” scientificrepresentations of laws and mathematical equation. Other approaches that may prove tohave a more reasonable balance between capturing unique characteristics of situationsand enduring principles might include Pattern Languages, Stories, Cultural Probes, andPersonae. Here I focus on the possible use of Stories as one such method.
The best selling sports book of all time is a small book by a long-time golf teacher,Harvey Penick called "The Little Red Book." In this book, he essentially "teaches" golfby means of a series of short stories (without diagrams or photographs). Many peoplefind this book extremely useful in improving their golf game. Given that golf is a largelypsychological and psychomotor game played in a physical environment, I believe this is arather remarkable and salient testimony about the power of story to convey tacitknowledge. In a series of recent workshops on HCI and International and CommunityDevelopment, the participants decided that stories would serve as a good way to captureand share knowledge gained from fieldwork. The author has since been working on atemplate for such stories and the description of a process for composing such storiesbased on what is known about how to capture and construct effective stories.Stories "work" partly because they access and leverage the knowledge alreadyaccumulated in an adult rather than trying to construct complex cognitive structurespiecemeal from the "bottom up." Adults typically learn new “chunks” of knowledge at arate of only 1 every four seconds. On the other hand, at that rate, a 40-year adult hasalready stored 160 million chunks. This preexisting knowledge can be accessed,stimulated, and rearranged at a much higher rate than creating new knowledge.Not only do the actual words used in a story trigger these preexisting knowledgestructures. Additional knowledge structures can be stimulated, accessed, and rearrangedby association and by various kinds of inference. For example, if one reads the sentence,“Joe kicked the cat across the room” the reader infers far more based on severalcategories including specific similar experiences, hierarchical class relations, knowledgeof human behavior, knowledge of general cultural “scripts” and spatial-temporalrelations. Such inferences quickly blossom to a very large number. Examples that mightreasonably be inferred from the single sentence above might include the following.Joe is human. Joe is a primate. (Since Joe is a primate, he may feel jealousy). Joe is amammal. (Joe probably has hair). Joe is a vertebrate. (Joe wants to avoid death; feelspain; has a spine, etc.) Joe is a living being. Joe is a physical object. (Joe can only be inone place at a time). Joe is upset. Something happened to initiate Joe’s current upsetstate. Joe is indoors. The cat is not happy. The cat is a mammal. The cat is a vertebrate.The cat is a living thing. The cat is an object. The cat might be injured. Joe’s nervoussystem is at least partially intact. (He has emotions, spatio-temporal muscularcoordination, etc.)Not only can short stories (or even small fragments of stories) cause a very large numberof propositions to be inferred about the subject matter of the story; as a reader beginsreading (or listening) to a story, the reader may also begin making reasonable inferencesabout the larger context of the story, the teller of the story, and about the overall structureof the story. As examples of these types of inferences, consider the following storyfragment.
“When Joe came to me complaining of an incurable slice I asked him to go through hisnormal routine and I could see the problem, just as big and mean as a charging TexasLonghorn before he even started his backswing. I said, ‘Whoa! Hold it right there, Joe!’What can we tell from this single statement? First, what can we infer about the teller?We can reasonably infer that the teller of this story is something of a golf expert. Ofcourse, we can also infer many other things. The teller is an English-speaking humanbeing and therefore has the characteristics of human beings, of mammals, of vertebrates,of living things, and of objects in the physical world. The teller can see. He or she usesoutrageous metaphors. The teller uses complex sentence structures. Although this smallstory fragment is a narrative “about” a set of events that happened some time in the past,it communicates a tremendous amount of information about the teller of the story. This isnot atypical, but illustrates an important mechanism for the transfer of tacit knowledgewhen experts share stories with novices. Not only are they narrating events that mayprepare novices for future similar events. Perhaps just as importantly, they arecommunicating their values, perceptions, strategies, and so on. Implicitly, this storyfragment is saying, “If you want to be an expert golf instructor like me, you too shouldeventually reach the state where presumably subtle errors in the set-up should be verysalient to you.” In terms of using stories for HCI in International Development, thissuggests that more valuable knowledge can be shared to the extent that the teller of thestory is a potential user or stakeholder, rather than a visitor.A story fragment also communicates something about the context in which the narrationis taking place. For those readers who are familiar with the “script” of going to see a golfpro, a lot of non-explicit information is conveyed; for instance, we would expect theperson to pay for this lesson. We would expect them to listen to the advice and try to putit into practice and become better as a result. We expect that the golf pro is going to tryto help the person become a better golfer and fix the problem that they see and so on.Any one of these script-based inferences could be contradicted by the larger storycontext. For instance, if we learn that the student is a lifelong friend or close relative,perhaps no payment will be made. If we explicitly learn that the student is too stubbornto apply the lesson, then we don’t expect them to improve. If it turns out in the broadercontext that the instructor has some reason to hate the student, they may give themintentionally bad advice and so on. Any of the kinds of reasonable inferences that wemake about context based on our understanding of the typical script can be contradicted,but only if it is done so by the storyteller with the addition of specific information that“re-frames” the context.Human communication also tends to embed metadata about the communication withinthe communication itself. For example, in the story fragment above, we now expect thatthe storyteller is going to relate a dialogue between the teller/teacher and the student. Ifthis is the beginning of a story, we have some notion of the size and scope of the story tofollow. This is not the end of the communication. But it is also not a likely beginning fora full-length novel. There is every reason to believe the story will continue in Englishand in print. It is likely to be presented in a consistent font. Chances are, since this muchhas been in grammatical if colloquial prose, the rest of the message will continue in that
vein. The genre here is true-life narrative. We do not expect Extraterrestrials orLeprechauns to leap out and help our golfer.The communication also tells us about the subject matter at hand; what is inferred is fargreater than what is stated explicitly. Some of the tacit knowledge about golf that iscommunicated in such a short fragment includes the following. 1). One may have aproblem with one’s swing and yet not know what that problem is. 2). If the set up to agolf swing is wrong, there is going to be trouble; there is no reliable way to correct theproblem by compensating in the back-swing and downswing. 3). Golfers have a normalroutine that they go through before actually swinging. 4). A slice is something generallyundesirable in golf. 5). Some set up problems will result in a slice. 6). In golf, peopleoften use a medical ailment metaphor. Often, the “patient” first tries to treat the“disease” themselves and only when self-diagnosis and treatment fail, do they then go toa professional. 7). For the “cure” to work, it is important to prevent the “patient” fromcontinuing the incorrect behavior as soon as possible.Far from being atypical, the above cases illustrate several important features that arenormally true about stories. 1. Each story conveys knowledge, not only about “subjectmatter” but also conveys knowledge about the teller, the context, and the rest of thecommunication. 2. The amount of knowledge that is activated in the mind of the listeneror reader is far greater than the relatively small amount of information that is explicitlystated in the story. From a few explicit statements, thousands of reasonably certaininferences can be made. All told, in fact, what is explicitly presented in a story representsa very small fraction of the total number of related statements that are being manipulatedin the reader’s mind.In the example above, however, part of the reason we can infer so much is that we sharemany cultural assumptions with the storyteller. In telling stories about other cultures, thetables are often turned in that, much of the interest in the story is in the way that our owncultural biases and assumptions are brought to the fore by the contrast to those of anotherculture. For example, when I first visited Japan, I went to breakfast and managed to saygood morning and the greeters said good morning and I said good morning and they saidgood morning… but they clearly wanted “something else” to happen before I say down toorder breakfast. It turned out that the “something else” was to pay for breakfast uponentry whereupon they gave me a wooden chip to lay on the table to prove I had paid.There was no ordering; everyone had the same breakfast. This “story” is mildlyinteresting to some Americans precisely because it violates cultural assumptions. The“story” for Japanese would be quite different. It would be about how dense the visitingAmerican was and how he tried to go sit down to breakfast without even paying.Even though we believe stories are a good way to capture and share a lot of knowledgequickly and in a form that is interesting and memorable, there are still many tensionsinvolved in using stories as part of a design process for another culture. For instance,how are stories constructed? Who participates in the storytelling? Who “owns” thestory? How can we know that people from another culture are portrayed in way that isnot insulting or even damaging to them? How do we know just how representative this
story is? More broadly, when are stories a useful technique and when are PatternLanguages, for instance, or Persona a better way to cumulate knowledge?We do not have the answers to all these questions but will continue to keep them in mindover the next few years as people in the community attempt to share and create stories.We expect both the story template and the associated process to evolve.An example of a lesson learned so far from a story concerns the adoption of technologyin an African village. In that cultural context, the village elders were the respected“experts” on everything and held the power. They had much more difficulty learning touse a mobile device than young boys. However, it was possible to introduce the mobiletechnology provided that they were “owned” by the elders and the younger boys merely“used” the mobile phones in response to orders from the elders.Bibliography.Darwent, S., Incledon, F., Keller, N., Kurtz, C., Snowden, D., Thomas, J.(2002) YOR920000749US2Story-based organizational assesment and effect system.Frey, J. N. How to write a damn good novel II. New York: St. Martins Press, 1994.Lawrence, D. and Thomas, J. Social Dynamics of Storytelling: Implications for Story-Base Design. AAAIWorkshop on Narrative Intelligence, N. Falmouth, MA. November, 1999.McKee, R. Story: Substance, structure, style and the principles of screenwriting. New York: Regan Books,1997.Penick, Henry. Harvey Penicks little red book. New York: Simon and Schuster, 1992.Thomas, J. C. A design-interpretation model of natural English. International Journal of Man-MachineStudies, 10, 1978, 651-668.Thomas, J. C., Kellogg, W.A., and Erickson, T. (2001) The Knowledge Management puzzle: Human andsocial factors in knowledge management. IBM Systems Journal, 40(4), 863-884. Available on-line at http://www.research.ibm.com/journal/sj40-4.htmlThomas, J. C. (2001). An HCI Agenda for the Next Millennium: Emergent Global Intelligence. In R.Earnshaw, R. Guedj, A. van Dam, and J. Vince (Eds.), Frontiers of human-centered computing, onlinecommunities, and virtual environments. London: Springer-Verlag.Thomas, J. C. (1999) Narrative technology and the new millennium. Knowledge Management Journal,2(9), 14-17.Biography.John received a Ph.D. in psychology from the University of Michigan in 1971 andmanaged a research project in the psychology of aging for 2.5 years at Harvard MedicalSchool. He joined IBM Research in 1973 and conducted research in query languages,natural language processing, design problem solving, audio systems, and speechsynthesis. In 1986, he began the Artificial Intelligence Laboratory at NYNEX Scienceand Technology, rejoining IBM in 1998 to manage a project on the business use of storiesand storytelling. He has over 200 publications and invited presentations in HCI. In 1992,and 1993, he co-led two-day CHI workshops on cultural issues in HCI. More recently, he
co-led workshops on HCI and International and Community Development at CHI 2007,CHI 2008 and DIS 2008.