Crowdsourcing Translation - An Experiment

Extracted text from Draft_-Crowdsourcing_Bible_Translation_-An_experiment__Gravelle.pdf:

Crowdsourcing as a Method of Bible Translation Gilles Gravelle 1

Rough Unedited Draft

1. Crowdsourcing

Crowdsourcing Bible translation is a disruption of how groups have worked in the past. Disruptive thinking, from time to time, is not a bad thing. As one author describes it:

A theologian recently described a community as an aggregate of individuals. 2 The term suggests a collection of diverse people brought together to form a whole. The notion of crowdsourcing as described by various authors suggests the same: 3 Human beings dont have complete information. They have private, limited information. Groups that are too much alike find it harder to keep learning, because each member is bringing less and less new information to the table.

Disruptive means, innovative ideas that disrupt the status quo when the status quo is no longer effective in achieving mission and impact. (Vijay Vaitheeswarn, The Economist Magazine). 1

It may be valuable information and it may be accurate (or it may be useless and false), but it is always partial. That doesnt mean that intelligence is irrelevant. But it does mean that, on the group level, intelligence alone is not enough, because intelligence alone cannot guarantee you different perspectives on a problem. Its not necessarily about size as it is about scope. This does not mean that if you assemble a group of diverse but thoroughly uninformed people, their collective wisdom will be smarter than an experts.

Rather, if that same group, though, has a means of aggregating all those different informed opinions, the groups collective solution may well be smarter than even the smartest persons solution. Decentralization as a way to aggregate collective wisdom is needed. Decentralization offers diversity.

1

Director of Research and Innovation, The Seed Company, Arlington, TX.

1

The Results of Crowdsourcing

Diversity helps because it adds perspectives that would otherwise be absent and because it takes away, or at least weakens, some of the destructive characteristics of small group decision making.

Diversity means adding people (youth, women, illiterate) whose opinions may not normally be sought after.

Open Source

More wisdom, knowledge, perspective, skills, experience, and insights are aggregated. The aggregate is greater than any individual part of the crowd. The crowd informs, learns from one another, and the collective intelligence of the crowd grows. The highly iterative and cyclical process continues to make improved decisions

Related to crowdsourcing is open sourcing. It is a means to insure a wide variety of aggregated intelligence by opening the creative process to anyone who is interested in contributing. Examples of open source collaboration are the Linux computer source code and Wikipedia. Self-monitoring When it comes to quality control, self-monitoring for accuracy seems to be a natural occurrence with open source and crowdsourcing collaboration. When people work together on something they highly value, they develop a greater sense of ownership. Group Creativity for Innovation Paulus & Nijstad (2003) claim most organizations and much of the scientific process now rely on the work of teams with diverse skills and knowledge. Other recent books on creativity confirm Paulus & Nijstads claims. This means open innovation rather than closed innovation. It also means work is done through group collaboration rather than hierarchy. This suggests that the end-user of a localization effort is an active participant rather than a passive consumer. 4 Group diversity is one of the most important factors in group collaboration for innovation. More diversity brings new information. Information diversity can lead to higher quality decisions, more so than an individual might make relying solely on their particular area of expertise.

2

Diversity of team membership can produce higher-quality team decision making. It creates synergy, where one members ideas provide new information that stimulates more creative thinking by others in the group. Therefore, open sharing is needed: Group members must be willing to share their unique information for the purpose of more innovative group decisions. Groups that have only old members over time tend to think alike, so adding to or changing group membership is necessary.

Why this age is ripe for open collaboration and crowdsourcing of translation

In 2010 the global internet-connected population will cross two billion people, and mobile phone accounts already number over three billion. Since there are something like 4.5 billion adults worldwide (roughly 30 percent of the global population is under fifteen), we live, for the first time in history, in a world where being part of a globally interconnected group is the normal case for most citizens. When you aggregate a lot of something, it behaves in new ways, and our new communications tools are aggregating our individual ability to create and share, at unprecedented levels of more. 5

Clay Shirky (2010) writes, The social uses of our new media tools have been a big surprise, in part because the possibility of these uses wasnt implicit in the tools themselves. When we use a network, the most important asset we get is access to one another.

Crowdsourcing as a Localization Model

Localization of translation, community development, and resource development takes into consideration end-user relevance, accessibility, and open communal learning. Given the current mission shift (indigenization, globalization) a localization model, as Simon Crisp describes it, for translation: extends local ownership and control of translation to the community and local empowerment. This doesnt mean the local community can determine it to mean anything they want it to mean. Community acceptance of the translation, though, is likely in many cases to be directly related to the level of community involvement in its production. Next, and as a consequence of this locally rooted and appropriate capacity building, we could expect to see the key functions of text checking and quality control,

3

traditionally associated with external translation consultancy, delivered in future by members of the host community rather than by outside experts. A component of the local empowermentis the investment of substantial resources in widening access to technology of all kinds rather than keeping it in the hands of the few.

As Shirky argues, the proliferation of smart phone and PC technology (the age of the PC tablet) along with global internet penetration, the goals of crowdsourcing, described in section 3, are now achievable, and at significantly reduced cost. Application: The Goda 2 Crowdsourcing Translation Crowdsourcing Experiment

The following section makes claims about social networking and crowdsourcing theory and compares those claims with what actually happened in a Bible translation crowdsourcing experiment with the Goda community of Northeast India. This experiment is not necessarily a radically new way to do Bible translation work. Rather, its a way to recombine important parts of the work to achieve better results. In this case, Bible translation carried out as an open communal process. The project data below reflect two months of crowdsourcing activity. People want to do something to make the world a better place. They will help when they are invited to. o 1,323 people have responded by contributing their time to the translation o Over 100,000 votes have been cast dealing with various content topics o 78 users have drafted o People from seven different regions where Goda people reside have participated

Access to cheap, flexible tools removes many of the barriers to trying new things. o Netbooks and desktop PCs through internet connections have been utilized. SMS messaging has been used to communicate with the crowd via cell phones. Decentralization allows for diversity to happen o 10 translation hubs and 3 additional mobile stations have attracted diverse members of the community to work together. They say that more stations are needed. 2

A Pseudonym

4

A decentralized system can only produce genuinely intelligent results if theres a means of aggregating the information of everyone in the system. o The group collaboration Website, designed for the Goda project, allowed leaders and users to see the aggregated results of their work. They volunteer their time for intrinsic motivation o There are no paid staff roles in the actual translation project. In spite of obstacles adversity, many still persist in being involved.

Intrinsic motivation is a catch-all label, grouping together several reasons one might be motivated by the reward that an activity creates in and of itself. o People say they enjoy working with their language and are growing in their understanding of that language as well as the regional language, learning together, and understanding the Bible they are translating better. They also communicate joy in being able to contribute to development of their language and in making the Bible available to their people.

Rather than functioning as a generic volunteer corps that does whatever is needed, people have an incentive to take on specific jobs. o Of 1,300+ participants, less than ten are receiving compensation. Crowd leaders organize; invite participation, supervise use of netbooks and internet, manage internet access. Technology enable these behaviors, it doesnt cause them. They will use it even if the tool is very different from what existed before, provided it lets them do things they want to do. o Even people with low to no exposure to the Web technology see how it enables their participation and they are enthusiastically utilizing it. Even with unpredictable internet connections, participants are eager to work.

Most collaborative circles consist of a core group who interact frequently and a peripheral extended group who vary in their degree of involvement. o Of the 558 users voting, over 100 have cast over 100 votes. Five voters were promoted to drafters after logging 800 votes.

As the groups ability to learn and work together gets stronger, it attracts more participants. o Eighteen to twenty nine-year olds are most responsive; however, older people and woman have also joined the groups and are participating through a communal discussion process.

5

As a general rule, it is more important to try something new, and work on the problems as they arise, than to figure out a way to do something new without having any problems. o Less internet cafes than thought; Christians reluctant to use Internet Cafes. Adjustment: supplied with netbooks and USB cellular data modems used instead at schools and church locations which became popular hubs for participation. o Many Web interface changes based on user feedback o Made possible for other contributors to take ownership of work left orphaned by others. Users who freely reveal what they have done often find that others then improve or suggest improvements to the innovation, to mutual benefit. o People tend to participate in groups and to discuss their responses before entering them on the computer. The younger people are showing their work to the elders who cant read, but speak Goda. The elders helped to improve their work, which the younger people then typed into the Web-based translation text.

Increases in community size, decreases in cost of sharing, and increases in clarity all make knowledge more combinable, and in groups where these characteristics grow, combinability will grow. o First month, five hundred participants; and then three hundred per month for next few months. Numbers are greater because of shared log-in. If an open source projects community members are committed to improvement, then they will commit to the hard work of identifying and making those improvements. o 1,374 total hours spent voting on scripture; 225 hours drafting, 11 hours per drafter in 10 days. In total 3,000 users, 714 hours per month.

Impact from crowd-sourcing The Goda Translation is notable in terms of localization of Scripture. During the course of the project, pastors reported church growth after people who joined the translation effort began valuing Scripture in their language. Their interest in knowing more Scripture caused them to remain engaged in the project. This includes people from two other major religions in the area who participated in the translation work. Another person reported how evaluating translation passages as a group has drawn them closer together as a church. They now want to conduct services in Goda. Conclusion

6

How can translation, community development, resource development, and curriculum development benefit from having more of the affected community participate in the creation and distribution of those materials? The result is materials that are more fully localized because of diversity. In this case, the community members viewed the work as relevant and impacting, and importantly, the fruit of their own work, rather than that of a small group of outside experts. Sources 1

Crowdsourcing is not so much a methodology as it is an idea. Indeed, open source collaboration requires that only the most basic parameters, structures, and policies are put in place before a collaborative effort begins. If the collaboration is overdeveloped from the start by a small group of experts, then people will see it as hierarchical, highly developed, and inflexible, thus not conducive to change or improvement, accept for incremental adjustments. The idea behind crowdsourcing and open source collaboration is to allow the aggregated wisdom of individuals to develop the processes and methods as they work.

Quote at The Economist Conference: Innovation and Entrepreneurship for a Disruptive World. UC Berkeley, March 23-24. Joseph Crockett, Eastern Pennsylvania University. Spiritual Metrics Conference, Oct. 21-22, 2011. Seth Godin. 2008. Tribes, We Need You To Lead Us. Portfolio. Steven Johnson. 2010. Where Good Ideas Come from: The Natural History of Innovation, Penguin Group. Paul B. Paulus & Bernard A. Nijstad (Eds), 20032. Group Creativity: Innovation Through Collaboration, Oxford University Press. Clay Shirky. 2010. Cognitive Surplus; Creativity and Generosity in a Connected Age, The Penguin Press. James Surowiecki. 2005. The Wisdom of Crowds, Anchor Books. Eric Von Hippel. 2005. Democratizing Innovation, MIT Press. Everett M. Rogers and Everett Rogers. 2003. Diffusion of Innovations, 5th Edition

2 3

4 5

Andrew Pym. 2005. The Moving Text: localization, translation, and distribution, p. 47. John Benjamins. Clay Shirky. 2010. Cognitive Surplus; Creativity and Generosity in a Connected Age, [Kindle Edition]. The Penguin Press

7