CIRC09 – Mapping, Circumventing, Translating, Sharing

I’ve written in the past about my friend and colleague John Kelly’s excellent work visualising connections in different blogospheres. His best known research is on the Persian-language blogosphere, where his analysis of linking behavior showed clusters around liberal and conservative politics, but also around poetry. Subsequent analyses have seen clustering around different factors. Russian blogs appear to cluster around platforms – Livejournal users link primarily to other Livejournal users, and so the Russian “blogosphere” is a mess of disconnected communities. The Arabic blogosphere clusters based on location, rather than on interest – Egyptians tend to link to Egyptians, Saudis to Saudis.

The Chinese internet, Kelly tells us, has a complex and hybrid form. It has aspects of clustering via platform, but there are also “trading zones”, where people group by interest and mix content across platforms. He’s looking at techniques of “attentive clustering”, joining people together based on sites they’re paying attention to, rather than on direct links to one another. The research is in an early state, but it looks like Kelly’s techniques will be able to release some interesting information.

Roger Dingledine of Tor offers some insights into his unique and exciting platform for censorship circumvention and anonymity. He reminds us that it’s free software – you’re encouraged to build your own Tor network, though you might have a hard time replicating the 1500 active relays and 200,000 users he’s got on his network. Tor has the most users in China, followed by the US and Germany.

Tor is now a “real live 501c3” non-profit organization, and it’s been funded by an amazing variety of organizations: the US Department of Defense, the Electronic Frontier Foundation, Voice of America, Human Rights Watch and Google. Speaking to all these funders requires using different language. “When I talk to Walmart, I talk about communications security. Talking to my family, I mention privacy tools. To the military, it’s ‘traffic-analysis resistant communications networks’. It’s the same tool, but I phrase it in terms of the characteristics people care about.”

All these users, Roger reminds us, are needed to keep the network robust and anonymous. Good cryptography isn’t sufficient to provide anonymity – you need to disguise who’s talking to whom, which means Tor benefits from being a network used by privacy freaks, online gambling fans and human rights activists. “Nobody tries to break crypto anymore – they just do social network analysis, find the hub, then break into your house.” Tor helps with one aspect of this problem – it disguises a great deal of communication between people who could otherwise be linked via traffic analysis. On the other hand, Roger remembers a training he and I gave a few years back, where our clients explained were being surveiled both electronically and in the physical world, with parabolic microphones intercepting conversations. Online security can only take you so far.

Roger notes that groups like Tor can help control the pace of the censorship and circumvention arms race. The more publicity tools get, the more likely they are to get blocked – Roger’s very interested in building a tool that’s useful for Chinese internet users, but not aiming at overthrowing or somehow overcoming the Chinese government, because that’s almost guaranteed to make the tool a target for blocking and censorship.

Zhang Lei, the founder of Chinese translation community Yeeyan, starts his talk with a story about his last name. While Zhang is the world’s most popular last name, it’s generally considered exotic in the US, and most Americans can’t pronounce it correctly (“Jong”, not “Zang” or “Zeng”.) He sees this as an illustration of the difficulties people have in understanding one another when separated by barriers of language.

While 18-20% of world’s internet users are Chinese, it’s unlikely that Chinese is as well represented linguistically on the net. Zhang points out that there’s really no accurate data on what languages are represented online – he references an old and probably bad cite on Wikipedia that suggests that 80% of web content is in English, followed by German and Japanese. If this is true, there’s a massive imbalance between users and the content available to them. A simple experiment confirms this suspicion. A search for “breast cancer” on Google reveals 38 million pages – a search for the Chinese equivalent yields only 6 million, and the quality of content is much lower.

Machine translation isn’t a satisfactory solution. A simple paragraph of text, translated from English into Chinese via cutting edge technology, yield about one third readable text, two thirds gibberish. There’s a ton of content that would be worth translating from English to Chinese, and we’re not going to be able to do it automatically.

Zhang’s project, Yeeyan.com, wants to be “wikipedia for translation”. His community involves 8,000 volunteer translators, who’ve created 40,000 translations. The community includes 80,000 participants, who are able to comment on or improve translations. Perhaps the most exciting new project is a collaboration with The Guardian, to translate the newspaper into Chinese on a regular basis, producing an official, sanctioned edition – this is an interesting contrast to ECOTeam, which translates The Economist via an informal understanding with the publisher.

The motivation for Zhang’s project is to build understanding across gaps of language. He explains that terms can mean something very different, even in translation: “The term ‘conservative’ in relation to economic policy means ‘anti-freemarket, pro-government control’ – the opposite of what it means in the US.” These misunderstandings get in the way of dialog and understanding. In 2008, we saw major understanding gaps built on language gaps, centered on Tibet and Chinese nationalism. “I can’t solve these problems, but I can translate,” Zhang tells us. “Translation is the first step and a must to bridge the divide.”

Isaac Mao has been blogging since 2002, and he’d be the first to tell you that blogging has changed how he sees the world. His work now is on developing a theory called “shareism”, based on the idea that humans are inclined to share with one another, but that cultural barriers have emerged to restrict sharing, and that losses and absences in our society arise, in part, from our failure to share. Isaac sees the hierarchical system of Chinese society, and several thousands of years of history of top-down control, as providing an especially challenging environment for shareism.

Chinese people, he believes, are being separated into two groups – those who are connected and those who are disconnected. Bloggers spend a lot of time sharing, subscribing to other bloggers, and connecting with one another. They have more authentic relationships to one another, he believes, based on their willingness to share and connect. The unconnected are influenced primarily by mainstream media – the connected can influence each other, can access information that’s hidden from the unconnected and circumvent censorship. Ideally, they’ll connect via social media, access important information, and share information with the unconnected people, empowering them. “This could be the hope and the future of the Chinese community.”

It’s not reasonable to posit the elimination of China’s hierarchical systems – it needs to be replaced with something, and Isaac believes the sorts of connection he’s talking about could offer that necessary structure. He sees this change already happening in small ways – communities that have access to alternative media stop being as dependent on highly controlled mainstream media. As attention switches to these new spaces, business and political leaders need to pay attention to these new spaces, as do foreign journalists. He notes that journalists covering China are now paying close attention to bloggers, not just to established media sources.