by Jeffrey Barlow <barlowj@pacificu.edu>
Editor, Interface
Introduction:
The first part of this editorial series, "Defining the Problem," is found at http://bcis.pacificu.edu/journal/2006/04/edit.php. There we argued that current methods of filtering content which particular groups of end-users find objectionable are inefficient and inherently create conflict.
Essentially the status quo is one in which content providers simply make content accessible, and end users, or their designated authorities, whether at the individual level (for example, a Christian family attempting to screen out anti-religious content), at the institutional level (a corporation or a K-12 school blocking pornography) or at the national level (Germany attempting to prevent the dissemination of pro-Nazi information) attempt by various means to block it.
This environment inevitably creates problems. Among these is the politicization of many issues that in the absence of contested filtering might not occasion disagreement. These issues might be as simple as German laws forbidding the sale of Nazi regalia on E-bay. Or they can be much more complicated as, for example, when Chinese national authorities require that Yahoo or Google remove certain sites from their indexes under the threat of being banned from doing business in China.
We believe that a process of mutual negotiation between content providers and end-users is a preferable and possible alternative to the status quo. While such negotiations cannot offer a total solution to the problem-too many interest groups profit from pushing content that many users will find objectionable-they can do much to defuse many issues.
The major problem which first requires addressing, and the topic of this editorial, is how best to define objectionable materials. There are a number of possible approaches, of course. As indicated in our earlier piece, we believe that the most useful approach is to define materials as objectionable to specific cultural or ethnic groupings [1]. While individual differences among members of such groups will always present anomalies, nonetheless this approach can create useful data sets of values specific to large human populations.
At the Berglund Center we have been working on this particular problem for some time and as a result have some useful experience. Our first realization was that it was not useful to us to approach this issue in a positive light; that is, it is not practicable to electronically query end users as to which sorts of values they wish to see presented in their materials. We are interested initially, of course, in the largest value sets we can define. To do this in a positive manner is a hopeless task.
Why We Must Negatively Define Values:
Recently, we invited representatives of several major religions to participate in a Berglund Round Table held on September 19, 2006. Among the participants was the very thoughtful Rabbi Gary Schoenberg of the Jewish outreach project, Gesher. The Rabbi is an enthusiastic user of the Internet and believes it to be very important to both his professional activities and leisure time pursuits. He was dubious of participating in a panel the topic of which was: "According to your religious beliefs, what sorts of content should not be presented on the Internet?" He felt, quite reasonably, that Americans of late have experienced an excess of fear mongering and wondered if perhaps this was just another know-nothing attack on the Internet.
I explained that we had begun trying to proceed with the Golden Rule as stated positively in the Christian tradition, bowdlerized for our purposes as: "Push at others only the content that you yourself wish to see." This had gotten us no place, because as we argued in Part I above, on the Internet the context of information is everything. What users wish to see is, in effect, everything except that which they do not wish to see.
We found the Confucian negative statement of the Golden Rule, again altered for our narrow purposes, much more useful: "Do not push at others content that you yourself do not wish not to see." And as the Rabbi informed me, this was also the Jewish mode of stating the Golden Rule. [2]
Since it is our purpose to propose a broadly useful system for negotiating culturally appropriate content, this requires considering a wide number of culture. So our final statement might well be, "Do not push at particular sub-sets of end users content that they may well find culturally inappropriate without giving them a chance to negotiate their reception of the suspect materials with you." Fortunately, the digital implementation of this statement can be considerably shorter.
The current difficulties of a Culturally Sensitive Search
Search engines already filter, of course, with regard to subject matter. That is, if I want material as a practicing Muslim, on Islamic temples in my region, I simply run such a search. As example, here is a Google search string I ran for this editorial: <Portland, Oregon, Islamic temples> This returned 432,000+ hits. A quick glance through the first several pages shows that the querent is quickly into highly politicized territory. If I am the father in a devout Muslim family, the first thing I do not want this search string introducing into my home is an image of individuals I deem holy, and most especially not that of the Prophet Mohammed. Nor do I want my beliefs that such images are iconoclastic attacked in an adversarial fashion.
Yet if I use Google's advanced search techniques to "Search within these results" to run the string <Image Mohammed> I find that within my initial results were 63,000 sites with at least references to the image of Mohammed. [3]
If a Muslim, I am probably, in fact, not going to be happy with the number of sites discussing the controversy over the Danish cartoon picturing Mohammed as a sort of human bomb, yet the "Search within results" string <Portland, Oregon, Islamic temples image Mohammed cartoon> suggests that there may be as many as 53,000 such references in my initial 432,000 hits. I would certainly like to screen these out; I wish only to provide a list of useful local materials or places of worship for my children to do research into our faith.
We could, of course, replicate this problem using the prohibited values of any number of religions. As a quick scan of the 53,000 references suggest, many providers of these pages actively seek conflict; they want to compare the reaction of some Muslims to the cartoon's distribution to horrors attributed to some Muslims, or to compare it unfavorably to their own tolerance of Christian images, or to staunchly defend freedom of speech. [4] These are understandable reactions, but also adversarial ones. How much better if we could provide a system where providers had at least the option to signal potential devout Muslim viewers:
In return, as a devout Muslim householder, I signal my browser that I do not want to receive material with content categories 1 and 2 above. The provider and I have now negotiated a download of culturally appropriate materials. The provider has given me the opportunity to do so by coding his or her pages, and I have replied by rejecting some materials but not others. We have established a trust-based relationship by entering into negotiations.
In fact, I have twice screened the materials; once by requesting from Google a particular sub-set of pages found on the world wide web, <Portland, Oregon, Islamic temples> and then querying that list of results and simultaneously asking three questions of the pages:
How Can Such an Enormous Undertaking Possibly Proceed?
We clearly are proposing an incredible effort to create data sets classifying materials felt culturally inappropriate by a wide variety of users: potentially every human ethnic group.
The answer to this question is that such a project would have to be distributed and shared on the Internet itself, a topic to which we will return. For now, however, we turn to the question of creating appropriate value sets.
Our methodology has begun by discussions with users of the Internet who have the understandings to represent their groups' ethnic or cultural views. This process started with the Berglund Roundtable of religious believers mentioned above. A second "Culturally Appropriate Materials" Roundtable is scheduled in Wenzhou, China, in December of 2006. [5] Participants will include Chinese familiar with both traditional and modern Chinese value systems, and with the current implementation of Chinese firewalls and blocking systems.
While such real-time querying of representatives of disparate cultural groups is interesting and informative, it is not truly necessary. Any thoughtful member of a given society and culture could quickly list 10-20 common sorts of objectionable materials, as could students or scholars outside the group.
And, of course, the great preponderance of objectionable materials is objectionable to large numbers of cultural groups. Purely for heuristic purposes, let us envision this schema:
| Value | Ethnic Group | ||||
| To be coded as culturally inappropriate | To: Chinese mainlanders To Muslims | To Muslims | To Christians | To Americans | To Frenchmen |
| 0001. Content calling for the violent overthrow of the national government | 0001. Marked to be screened if the end-user so wishes | 0001. | 0001. | 0001. | 0001. |
| 00011 Content Calling for the partition of the national territory | 00011 (Tibet) (Taiwan) |
00011 Sw border regions to be returned to Mexico |
00011 Alsace-Lorraine back to Germany |
||
| 0002 Sexual Depictions of Children | 0002 | 0002 | 0002 | 0002 | 0002 |
| 00021 Depictions of a violent sexual context |
00021 | 00021 | 00021 | 00021 | 00021 |
It is clear that there are a number of potential objections to even the above simplistic classifications. Swedes, for all we know, may have given up objecting to 00021 in defense of a higher value, the total freedom of graphical information. This will require work to ascertain; but in an Internet distributed project such information can be provided rather quickly and will be hotly debated. And if the area is contested, very well; Swedish users can, in any event, simply program their browsers to ignore metadata classifications dealing with this matter, or indeed, with any or all of the categories, as can any end-user.
Why Would Content Providers Participate?
As stated above, significant groups of providers will opt out of this system, or worse, spoof their participation in it while providing objectionable materials for personal advantage or political purposes. [6] An example of the latter motivation might be a provider who believes passionately that Tibet is not part of China and wants China to quit Tibet. If such a provider does not choose to participate in this project, he or she will simply keep pushing such arguments on their web pages. In that regard, the status quo has not changed. If, however, I wish to participate in this project, I cannot deceive Chinese readers who do not want to read my arguments, or Tibetan ones for that matter. The others can simply, like the Swedes, fail to program the browser to trip the negotiated rejection.
There are, however, a variety of strong reasons for wishing to participate. One is simply self-interest. At the Berglund Center, with thousands of files dealing with China, we emphatically wish the Chinese government not to block our ISP address. The same would be true for such enormous providers of electronic materials as multi-media corporations like United Artists or Pixar. We are willing to signal the inappropriateness of some of our materials for Chinese readers in order to gain access for the remainder of our content.
Many will also wish to meet as many local laws as possible-to not provoke German authorities from blocking all our sales of war memorabilia because we provide some Nazi memorabilia. Nor do we want to remove all Nazi memorabilia simply because one or several cultural groups find it inappropriate and it is illegal in some nation states but not in all. This system facilitates a selective approach and is infinitely preferable to removing all such memorabilia.
And, of course, many content providers will wish to mark their content simply because it is culturally appropriate to do so. They do not wish to do unto others that which they do not wish done to themselves.
But, ultimately, the success of this schema must depend on its widespread adoption, like the Internet itself with its elaborate system of protocols. If either Google, Yahoo, or the largest Chinese provider, Baidu, elevated the rankings of materials coded to be culturally appropriate, soon it would be necessary to participate or risk low page rankings and diminishing one's audience.
In our next installment we will discuss the procedures for implementing this scheme.
References
[1] Once again we are defining "ethnic" here not as a racial group (A useless categorization not validated by science in our understanding, in any event.) but as a group sharing a largely common culture. In recent research in Wm. Theodore De Bary and Tu Weiming (Eds.) Confucianism and Human Rights, New York: Columbia University Press, 1998, we encountered in the chapter by Sumner B. Twiss, the term "Cultural Moral Anthropologies," (p. 35) which corresponds quite closely to our notion of culturally appropriate content, i.e., materials of which I approve or disapprove because of my cultural heritage, shared by many others in my ethnic grouping.
[2] For a marvelous site listing statements of the Rule from a wide variety of religions, go to: http://www.religioustolerance.org/reciproc.htm
[3] When I in fact run this search string in Google's image search, I find that there are probably no images of Mohammed as such, but this process has taken me a good five minutes and some sophistication to provide the assurance that I might well have wanted.
[4] I do not take the issue of the freedom of speech vis-ид-vis Internet content lightly, and will return to discuss it below.
[5] To view this event, begin at: http://bcis.pacificu.edu/roundtables/index.php
[6] We believe that spoofing can be limited to an extent by procedures to be discussed later.
Chris Pruett - Snap, Crackle, Crunch Time
Leonard D. DuBoff - What is Mediation?
Mark Szymanski - NASA: A Full Spectrum of Web-Based Support for Educators
Charles Boulet - Digital Hygiene: Email Techniques
Françoise Mengin's Cyber China. Reshaping National Identities in...
Joost Raessens's Handbook of Computer Game Studies
Negotiating Culturally Appropriate Data Transfers: Part II: Creating Culturally Sensitive...