Please tell us what you think of the new Bulletin interactive pdf!  Feedback

Bulletin, October/November 2007


Trouble in Paradise: Conflict Management and Resolution in Social Classification Environmentsby Richard Urban

by Chris Landbeck

Chris Landbeck is a doctoral student in the College of Information, Florida State University. He can be reached by email at clandbeck<at>fsu.edu

The advent of personal computers and the Internet has also brought about social indexing groups on a scale never before seen. These folksonomies organize information with little regard for the methods used in library or information sciences. The collaborative nature of folksonomies almost guarantees conflict will arise among some members of the community. A conflict can manifest in several ways: 

  • It can be an honest disagreement between two community members about the truthfulness of a given comment or whether it is appropriate.
  • It can be an unreasonable demand on the community as a whole on the part of one member.
  • It can come from a troublemaker, looking to sow the seeds of discontent in the community for nothing more than entertainment purposes. 

If these groups and their sponsors wish their communities to persist, they must protect themselves from the potentially negative consequences of such conflicts by providing some method for resolving them.

What conflict resolution methods exist in these communities? What lessons can we take from their examples as both organizations and as collections of data? One might speculate that the advent of new rules in these communities, whether slowly emergent or suddenly imposed, will affect aspects of the organization – how it is perceived by its members, how it operates and how it gathers and organizes information. Such rules either mandate or forbid certain actions by the members of the respective societies, and therefore we might expect them to directly affect their operations. By studying rules and regulations of these groups as they affect communication among members and the information they gather, we might find out how the rules develop, what the intent of the rules is from the sponsor’s perspective and both the effectiveness and pertinence of these rules in successful social classification communities. In this article we will examine certain conflicts that arise among members of a community or between a portion of the community and the people who run it. However, copyright and other legal issues are beyond the scope of this work, as are standards for determining truth or shaping community standards for decency or other related matters.

To this end, we look at three very different organizations: wikipedia.org (an attempt at creating an encyclopedia of general knowledge), flickr.com (a collection of personal images with descriptive tags) and espgame.org (an effort to tag all the images of the Internet). These organizations were chosen because they are representative of large, ongoing, noteworthy and noted social classification efforts. Note that the content of the various collections is not being studied because the data therein are, for the purposes of this work, irrelevant. We are not looking at the content so much as we are looking at how the stuff of the community is organized. For this study, we analyzed the pages from each website speaking to the organization’s efforts to manage conflict among its users and counted them as an indication of the level of development of each conflict resolution system. The content of these pages was also analyzed to determine their nature.

Wikipedia (49 official policies: 41 governing behavior including 7 governing conflict)
Wikipedia concerns itself with building a free online encyclopedia that anyone can contribute to or comment on. The company that runs Wikipedia does not resell the data contained in the community, nor does it profit by the work of it members: Wikipedia runs on donations. Facts are established by consensus, and any and all subjects of interest are allowed in this community. It is easily the largest social classification community extant today and has the most developed system of conflict resolution. It has 49 separate official policies that speak to the operation of the community, by which is meant the guidelines that spell out how the various essential components of the community function – how articles should be written, how corrections should be made, how members should deal with one another, how citations should be made and so forth. Of these 49 policies, 41 deal with what is and what is not appropriate behavior – either while acting alone (writing or correcting articles) or when interacting with others. Of those 41 policies, seven deal exclusively with what to do when conflicts arise, whether it is between two sincere people or when dealing with some less-than-scrupulous actor within the community. This social classification environment has a linear, hierarchical progression of steps to take for dealing with disagreement and personnel designated as resources for parties that find themselves in a dispute. 

More than any of the other organizations studied here Wikipedia.org also has a highly developed vocabulary for those who “game the system” – those who work against the stated principles and intent of the community. Wikipedia is vulnerable to such behavior and, therefore, experiences this type of conflict, for which there is no easy solution. Because of the way the community is set up, one side in this conflict can, for instance, create “sock puppets,” fictional user accounts that one side uses to deluge the other with responses in the survey phase of a resolution exercise. If real people engage in this behavior (say, a boss to his employees) they are then called “meat puppets.” “Trolls” are those who are disruptive to the point of being a nuisance, while those who have been banned from adding to Wikipedia and have somehow circumvented the block are said to be “reincarnated.”

These distinctions point to a system of conflict resolution, one based on the collective experience of the community in question, which is more highly developed than most. Indeed, the progression of steps recommended to community members is clearly spelled out, is linear in nature and includes step-by-step instructions for the members to follow (see Figure 1). At all levels, these recommendations consistently point to following the editorial guidelines such as neutral point-of-view, references and leaving one’s ego at the door as the first and best way to avoid conflict. Compared to the other communities studied here, Wikipedia is easily the most complete and fully formed system of resolution. Furthermore, there is some anecdotal evidence in Wikipedia itself that suggests that the rules evolved as they were needed, that they were not the product of forethought but rather of necessity.

Figure 1. Steps in conflict resolution for Wikipedia

Figure 1

When we consider the objective of Wikipedia, we see that it invites discussion and argument as a means of building consensus on any and all contentious issues. Anyone can initiate a disagreement on any subject, at which time a resolution hierarchy should (but does not always) come into play. Wikipedia seems to be seeking what might be termed “truth” or “fact” or “reality,” recognizing only the authority of consensus, which might explain the nature of the conflicts found in that community. According to the record, the nature of the information to be collected was the first and greatest concern, and the conflict resolution methods that have developed over time reflect that continued emphasis. For Wikipedia, it appears that the chicken of “what should be done” came before the egg of “how it should be accomplished.” However, what Wikipedia says about itself is only one source of information, and we should be aware that it might have its biases.

Flickr (25 official policies: 3 governing behavior including 1 governing conflict)
Flickr’s approach to conflict resolution was built directly into the system from the outset, although we cannot be certain if this was a matter of planning or of serendipity. Flickr allows community members to post their images to the web while both the owner of the image and other members tag the images in question with free-form descriptors. In this way, Yahoo (the owner of Flickr) gathers several tags per image, providing an online image-sharing community in exchange. Conflict resolution comes into play when two members disagree on what tags should be applied to the images. In every case, the owner of the image – the person who posted the image for community tagging – is the ultimate and final arbiter of who may apply tags to their images, what tags will be allowed, who sees the images and who might comment on them in sentence or paragraph format. In the rare cases where someone gains alternate identities to circumvent being blocked and simply will not leave an image owner alone, Flickr will get involved, identifying and blocking the offender using more sophisticated methods than those available to the image owner. While conflict is possible in Flickr, it is far more difficult to sustain than in Wikipedia because, in the end, the members can pick and choose whom they associate with.

Figure 2. Owner control in Flickr

Figure 2

In Flickr, the person who posts an image is the owner of it, which gives that person the right to veto others’ descriptions of their images. In that organization, the owner reigns supreme in any conflicts that may arise. Flickr seems to place equal value on social indexing and community building, realizing a profit from the gathering of images and maintaining the community as a way to ensure the continuous flow of tags into the system. With Flickr, no information about its development from concept to reality was found on the site. One might speculate that both the nature of the data to be collected and the definition of authority were conceived and implemented at the same time, differing from Wikipedia in that the chicken and the egg arrived simultaneously. If this hypothesis is true, the founders of Flickr must have known that the pre-existing conditions of authority would necessarily affect the nature and quality of the data to be collected and have made a decision that this structure would meet the needs of both the sponsoring organization and the community members themselves.

ESPGame: (Five sections, one governing behavior and conflict)
When playing the ESP Game, two users view the same image for a set amount of time while both give their impressions of what tags best apply to the image. When players offer the same tag, points are awarded – the faster a common term is found, the more points are given. In this milieu, there is very little chance for conflict. It is not possible to communicate directly within the game, and there is no opportunity to exchange information at all as this would defeat the purpose of the game. The structure of the community is such that conflict is precluded, yet ESPGame remains a social classification community in that two people are brought together to examine the aboutness of the images in question.

At ESPGame, no conflict is possible (regardless of any disagreements), as the two people randomly paired to blindly assign mutual tags to images are both anonymous and isolated from one another. ESPGame exchanges simple entertainment for image tags. Indeed, in this arena, the website holds the data to be far more important than any potential community, even though it hopes that the community will value the game enough to keep coming back, supplying more tags for more images. While we might again speculate about the whys and wherefores of its creation, there is even less to go on here than there was for Flickr, which has a simple but complete explanation of what it will do with the data it collects and is reasonably transparent about its creation and implementation. ESPGame has none of this and indeed very little at all about its purpose. Here, it seems, there is no chicken, only the egg of data collection.

Applying These Examples in the Workplace
How might we, as information professionals, seek to use these examples as models in the workplace? What lessons can we take from those who have pioneered the realm of social classification? 

Looking at the Wikipedia model, let us suppose that there is a local historical society with ongoing efforts to catalog the landmark buildings in the area. Toward this end, they might seek to collect and electronically publish legal records about the history of ownership of a particular building, newspaper articles from the archives of local papers, construction records and oral histories from residents and former residents of the buildings being written about. The members of the community would seek to exchange ideas about where information might be found, difficulties or successes in forwarding the effort of the group and other details about what has been found and things that might yet be found. 

As this community seeks to create and maintain a record of its activities and discoveries, the local library might be asked to administer and host a wiki for them, as the wiki technology and approach would seem to be the best method for this application, since it allows community members to easily record, store, publish and update records and to discuss the accuracy or appropriateness of records within the community. But the very capabilities that are needed would allow conflict to enter into the equation – the tools that allow the work of the group to go forward can sometimes be used to pull it back. Generally, the members will take care of honest conflicts themselves – serious discussion of issues of mutual interest are the bread and butter of wikis, and amiable (or at least civil) disagreements are actively sought in such environments. However, as the administrator of such a community, the librarian might be asked to mediate conflicts between members when they reach an impasse, either as an impartial mediator or as an information expert. This librarian might also be called on to deal with graffiti, with malicious members of the wiki or with other incidental problems that come with the territory that members might not wish to deal with. It is sometimes easy to forget that wikipedia.org runs on donations and that those donations go toward the salaries of those who run it. This fact is a reminder that, while running the wiki for this notional historical society might not by any means be a fulltime job, it will need someone both to settle disputes and to attend to the workaday issues that arise from any community-based effort.

Another implementation might draw on the Flickr model. The library might be asked to establish a virtual scrapbook of a trip that members of a civic group took as part of its relief or charitable efforts. Here, the members could contribute their pictures to a central location that would allow those who went on the trip to post comments, stories and descriptors of what each picture is about. In this application, the librarian has a different set of concerns. As the administrator of the historical website, the librarian was concerned with developing records; now she is being asked to help manage the records of others. She may be asked to assume an active role, serving both as administrator of all things technical and perhaps as the point-of-contact for the scrapbook.

Adopting the Flickr approach up front – that the person contributing each photograph can accept or reject the comments on it – could provide an appropriate rule for settling disagreements. However, the librarian, as the administrator, might also be asked to serve as the de facto help desk for such an effort, answering how-to questions from less tech-savvy users. She might also be sought in her traditional role of librarian to help pin down facts about the pictures. How often have we ourselves taken a picture then forgotten where it was taken and in what way the thing in the picture is important? While the nature of the possible conflict in this social classification environment is different, it is no less a conflict that can disrupt an otherwise happy community.

Finally, in a different vein, a librarian might think it would encourage library use to set up something like the ESP Game, not for images but for, say, children’s books. She might encourage two patrons who have checked a book out to provide terms that describe the book. She might then gather together these terms, rank their frequency and put the staff in a better position to recommend (or not) books based on what others have said about them, possibly even integrating this data into an online catalog. As noted, there is little chance for conflict to arise here, as the way the system is built prevents it. But such an effort might provide a new channel for old conflicts about the appropriateness of certain materials.

Conclusions
In considering which model to adopt, three general questions should be considered:

1. What do you want your community to be able to do?

2. What kind of data should be collected?

3. What kind of time and resources can be given to the project?

We see in the examples above a wide range of possible activities that users might need or want to engage in. Wikipedia keeps a long history of article revisions and allows for a great deal of interaction, depending on the goodness of its user’s souls to keep things aboveboard and honest. Flickr has placed a great amount of somewhat authoritarian control in the hands of its image owners, again depending on the beneficence of its user base, while ESPGame has taken away the forum for conflict, stopping it in its tracks. Each has different possibilities for interaction among users in their respective communities. What level would you like to see in yours? What is needed? And how do these approaches help or hinder your users in their needs and preferences?

In a similar vein, Wikipedia deals only with full thoughts expressed in (what is hoped to be) clear, concise and complete paragraphs as its finished product; Flickr only deals with images and the associated words or phrases; and ESPGame deals only with words as tags for images. Each of these is a different kind of data to collect, and each has its own ethical and social concerns to deal with. What kind of data is necessary for your community? What kind of issues can be dealt with in advance (and how will they be dealt with)? And, most importantly, who owns the data, and is the owner different from the keeper?

The thread that runs through all of this discussion is the most burning question to ask before implementing any folksonomy: What are you willing to do to make it work for your user community? What kind of skills can you and your organization bring to the task? What kinds of hassles and conflicts are you willing to settle? And who in your organization is best suited to oversee such an endeavor? In each of the examples presented here there is a way to contact the ones who run the community, hear complaints, listen to suggestions or provide help to those in need. While there is necessarily a greater (Wikipedia) or lesser (ESPGame) need for supervision, there is nonetheless someone in charge. Who that is for your organization is essential to the success of the community.

Resources 
Bearman, D. & Trant, J. (2005). Social Terminology Enhancement through vernacular engagement. Retrieved July 10, 2006, from www.dlib.org/dlib/september05/bearman/09bearman.html

ESP Game. (2006).The ESP Game: Labeling the Web. Retrieved July 15, 2006, from www.espgame.org 

Flickr. (2006). Flickr.org. Retrieved July 20, 2006, from www.flickr.org

Kroski, E. (2005). Tag! You’re it! Retrieved July 10, 2006, from http://blocklevel.com/weblog/information_architecture/tag_youre_it/ 

Mote, N. (2006). What we can learn from folksonomy. Retrieved July 8, 2006, from http://fairuz.isi.edu/blog/index.php/2006/06/24/what-we-can-learn-
from-folksonomy-anddelicious/

Wikipedia. (2006). Main Page – Wikipedia, the free encyclopedia. Retrieved July 12, 2006, from http://en.wikipedia.org