Meet the Bridgebloggers

12 downloads 234562 Views 1MB Size Report
Dec 23, 2005 - whether the blog was a sophisticated form of disinformation. Journalist ..... 29 Kristof,“Death by a Thousand Blogs”, New York Times, 5/24/2005.
Meet the Bridgebloggers Who's speaking and who's listening in the International Blogosphere

Presented at “The Power and Political Science of Blogs” University of Chicago September 16 – 17, 2005, revised December 23, 2005

Ethan Zuckerman Berkman Center for Internet and Society Harvard Law School [email protected] www.ethanzuckerman.com

Meet the Bridgebloggers When Iraqi architect “Salam Pax” began blogging in late 20021 on his site, Dear Raed, the term “weblog” and the practice of blogging - posting brief entries in reversechronological order to a webpage - were both five years old2. Salam Pax was not the first blogger outside the United States, or even the first blogger in the Middle East. He was, however, the first blogger from the region to receive widespread attention in the US and European blogosphere, and eventually, in the US and European press. In that sense, Salam Pax was one of the first major “bridgebloggers”. Salam Pax’s accounts of life in Iraq provided a first person perspective on the US-led invasion that was impossible to obtain through mainstream media outlets.At times funny, at times heartbreaking, Salam’s dispatches became enormously popular with fellow bloggers, leading to speculation about whether Salam Pax was actually an Iraqi or whether the blog was a sophisticated form of disinformation. Journalist Peter Maass was able to confirm Salam was authentic and, in fact, was his Iraqi translator.3 Salam Pax went on to publish a book derived from his blog posts and became a print4 and television5 journalist in Britain. Originally celebrated for providing perspectives and opinions outside the journalistic mainstream, Salam Pax is now established as part of the mainstream media.As a bridgeblogger, he’s achieved a coup, not just introducing online readers to perspectives outside of their ordinary reality, but ensuring that these perspectives influence offline media coverage and reach beyond the world of weblogs. The term “bridgeblog” was coined by this author and Chinese weblogger Xiao Qiang in October 2004 and popularized by Iranian blogger Hossein Derakshan in his . characterizations of Iranian weblogs as bridges, windows and cafés67 Bridgeblogs are weblogs that reach across gaps of language, culture and nationality to enable communication between individuals in different parts of the world.They are distinguished from the vast majority of blogs by their intended audience: while most blogs are targeted to friends and family, or to an audience that’s demographically similar to the author, bridgeblogs are intended to be read by an audience from a different nation, religion, culture or language than the author.A Tanzanian blogging in Kiswahili about local politics is not bridgeblogging; a Tanzanian blogging in English about Tanzanian politics, explaining the positions of the politicians mentioned and the context of the

1

http://dear_raed.blogspot.com/2002_12_01_dear_raed_archive.html#86159736, accessed 9/14/2005. Dave Winer’s Scripting News, begun in April 1997, is widely acknowledged as the first blog. Jorn Barger of Robot Wisdom is widely credited with coining the term “weblog” in December 1997. From various sources, including http://www.blockstar.com/blog/blog_timeline.html, accessed 9/14/2005. 3 http://slate.msn.com/id/2083847/, accessed 9/14/2005. 4 Salam Pax was given a biweekly column in the Guardian, where he reported on the US elections of 2004. http://www.guardian.co.uk/Iraq/blogger/0,13814,1018987,00.htm, accessed 9/14/2005. 5 Salam Pax’s film reports from Baghdad, produced by Guardian Films, have been extensively featured in European film festivals, including the Rotterdam International Film Festival. http://www.filmfestivalrotterdam.com/en/film/32898.html 6 http://www.alternet.org/mediaculture/21316/, accessed 9/14/2005. 7 http://blogs.law.harvard.edu/ethan/2004/12/10#a635, notes on Hossein’s talk at Harvard Law School, December 12, 2004. 2

issues debated, is brigeblogging8. Bridgeblogs don't need to bridge between another language and English – Jerzy Celicowski, a Polish citizen living in Hungary, uses his Polish language blog Jez Wegierski (jezwegierski.blox.pl) to bridge between Hungary and Poland. While bridgeblogs give North American and European readers access to conversations taking place in other parts of the world, they hint at a more complex and less-accessible phenomenon: the emergence of local blogospheres. In some communities, these conversations tend towards bridging – local bloggers write for both a local and global audience and encourage the interest of bloggers outside the nation or the region. In other communities, the conversation is linguistically or culturally closed, not intended for an outside audience. Some of these local blogospheres are becoming large, well connected and registering on the radar screens of Anglophone bloggers, if only because they share the tools and platforms with English-language blogs, though not common conversations. The appearance of bridgeblogging raises intriguing questions about “international blogging”, blogging outside North America and Western Europe. How many international bloggers are there? What fraction of them are bridgebloggers? What audiences are they attempting to reach? How effective are bridgebloggers in reaching North American and European audiences? To what extent are bloggers in North America and Europe interested in news and perspectives from the rest of the world? This paper offers cursory answers to a few of these questions, focusing more on anecdotal rather than quantitative data. Extensive quantitative research is necessary to offer definitive answers to these questions, and any answers are subject to invalidation by the rapidly changing and expanding blogosphere.The paper offers some future research directions and techniques for answering these questions quantitatively. An inescapable conclusion of this early research is that bridgeblogging is a small phenomenon in comparison to the massive growth of non-English blogospheres, especially the Chinese, Japanese, Portuguese and Spanish blogospheres. Given the growth of blogospheres in different languages, the act of bridging is increasingly important if weblogs are to remain a common, shared space. Quantifying the International Blogosphere Early in the web revolution, as the Internet moved from a text-based academic tool to a graphically-enhanced consumer medium, researchers attempted to quantify the growth of web by counting unique, accessible web pages.These ennumeration methods have always involved a good deal of inference and guesswork, as actually finding and counting all web pages – or all weblogs – is a problem well beyond the scale of most researchers. Complicating the estimation challenges is the fact that very little work has been done to 8

Tanzanian blogger Ndesanjo Macha manages to do both, blogging about Tanzanian issues in English on http://digitalafrica.blogspot.com/, and in Kiswahili on http://jikomboe.blogspot.com/.

enumerate the geographic distribution of weblogs. Much of the early research on weblogs has focused on highly-linked, high-traffic weblogs, most of which were (and to a lesser degree, still are) authored by English-speaking bloggers in the United States. Comparatively little work has been done to identify and count non-English blogs and blogs published by authors outside the United States. In 1998, Steve Lawrence and C. Lee Giles estimated the size of the web at 320 million indexable pages by interpolating from the overlap between search engine catalogs9. In a white paper for Cyveillance, Inc., Brian Murray offered a figure of 2.1 billion unique, indexable pages on the Internet based on extrapolation from a set of 350 million links10. As content on the web is increasingly automatically generated by assembling database elements into XHTML pages when the user requests them, it’s become very difficult to count the pages on the web. Recent figures for the size of the indexable have come primarily from search engines.Yahoo! recently claimed to index 19.2 billion web documents11, drawing a sharp response from the team at Google, which accused Yahoo! of counting duplicate and mirror pages12.As of 9/14/2005, Google claimed an index of 8.16 billion webpages, as advertised on the front page of their site – they have subsequently removed this figure from the homepage. Attempts to ennumerate the blogosphere have followed a similar pattern, as attempts to directly count blogposts have given way to assertions of index size from blog indexing sites like Technorati and Blogpulse. Using an estimation model and a sample of 10,000 blogs, Perseus estimated the total number of blogs hosted by major bloghosting companies at 31.6 million in April 2005, with projections of 53.4 million by year’s end13. Blog indexing services like Techorati and Blogpulse rely less on “spiders”14 (as search engines traditionally have) and more on “ping servers”15. Most modern weblogging platforms send a short message – a “ping” – to one or more central servers when a user authors a new post.Technorati, Blogpulse and others access the ping servers and spider the new posts that have been registered at a pingserver.Technorati claims an index of

9

Lawrence, Giles,“Searching the World Wide Web”, Science:280,April 3, 1998. Murray,“Sizing the Internet”, July 2000. From http://64.233.167.104/search?q=cache:gyUXGu3_oA8J:www.cyveillance.com/web/downloads/Sizing_the_Internet.pdf, accessed 9/14/2005. 11 http://www.ysearchblog.com/archives/000172.html, accessed 9/14/2005. 12 As reported by John Battelle, who was working closely with Google technical staff on his book about Google,The Search. http://www.ysearchblog.com/archives/000172.html, accessed 9/14/2005. 13 From “The Blogging Geyser”, http://www.perseus.com/blogsurvey/geyser.html, accessed 9/14/2005. 14 Spiders are software programs that traverse the web, moving from hyperlink to hyperlink and counting or indexing webpages encountered around the way. Most internet search engines rely on one or more spiders to build their index of webpages. 15 Ping servers like http://blo.gs or http://weblogs.com are servers that attempt to index posts made to weblogs. Rather than querying servers with spiders to find new pages, ping servers ask weblog authors to “ping” those servers with a short message each time a weblog post is added or modified. Most modern weblogging software is configured to send messages to one or more ping servers when a new post is made.While ping servers are essential tools for researchers who study blogs, the major ping servers have had major stability issues over the past year.This means it's not possible to depend on these servers having a complete, accurate set of ping messages for an arbitrary time interval, because the server might well have been down entirely or capable of recording only a fraction of pings during the interval in question. 10

23.1 million blogs16, while Blogpulse claims 19.7 million17. The similarity in size of Technorati and Blogpulse calls Perseus's estimation method into question. Since numerous blogs are created and abandoned, an estimation based on blogs created is likely to be a massive overestimate when considering “active” blogs. Furthermore, both Technorati and Blogpulse apply various filters to eliminate “Spam blogs” (sometimes called “Splogs”), blogs whose sole purpose is to link to each other and to other pages on the Internet, driving up the search engine rank of the linked pages. The disparity between the Perseus blog creation estimate and the Technorati/Blogpulse valid blog index size could be due in part to the elimination of spam blogs from the set. Despite rough agreement on the size of the blogosphere, data on the national origins of bloggers is much harder to obtain.The vast majority of blogs are hosted on large servers that host hundreds of thousands or millions of individual blogs.While many of these servers are located in the United States, users of the services could be anywhere in the world that has internet access. Counting existing blogs and using IP/Geotargetting services like IP2Location18 is, unfortunately, useless in determining the geographic origins of most bloggers. The National Institute for Technology and Liberal Education (NITLE) began a survey of blogs in May 2003, using a spidering method19. By early 2004, NITLE had spidered 1.7 million “likely” blogs and concluded that 1.1 million of them were updated within the previous month. NITLE used the textcat20 text categorization system to classify the language of each blog, matching to 77 known languages. Of the 1.1 million likely blogs, NITLE’s tools determined that 62% were in English.The next most popular languages, in decreasing order of popularity, were French, Portuguese, Farsi and Polish21. NITLE’s identification of over 75,000 blogs in Farsi received widespread attention in the media22, alerting readers to the magnitude of the Iranian blogging phenomenon. Other NITLE statistics caused more controversy – some observers of the Asian blogosphere were surprised that Russian, Chinese, Korean and Japanese blogs didn’t figure more prominently in the statistics23. It’s quite possible that the NITLE survey radically undercounted some languages by failing to spider large collections of URLs in those 16

http://technorati.com/, accessed 12/15/2005. http://blogpulse.com/, accessed 12/15/2005. 18 http://www.ip2location.com/ 19 http://www.blogcensus.net/?page=Methodology, accessed 9/14/2005. 20 http://odur.let.rug.nl/~vannoord/TextCat/ 21 The researchers behind the NITLE project pulled their site from the web in early 2005, probably due to controversy over the accuracy of their data.The 62% English figure and the ranking of secondary languages are from my notes, based on a retrieval of the NITLE census in February 2004.The rankings are echoed on David Schlossberg’s “Blogging By the Numbers”, http://www.themeat.org/rant/blogging/BloggingbythenumbersTheBigNumberPrint.html, accessed 9/14/2005. In December 2005, the engineers behind NITLE appear to have restarted their spiders. Current results indicate a total of 1.9 million blogs indexed, 1.3 million of which are categorized as being in English.The second most popular language is Catalan, with nearly 73,000 weblogs, followed by Spanish.The fact that Chinese ranks as the sixth most-frequent language identified suggests the current numbers still inaccurate. 22 See Alavi,“Freedom In Farsi Blogs”, Manchester:The Guardian, 12/20/2004. http://www.guardian.co.uk/online/weblogs/story/0,14024,1377538,00.html, accessed 9/14/2005. 23 See the comment thread on http://www.languagehat.com/archives/000942.php, accessed 9/14/2005. 17

languages. In the absence of a large survey of the languages of the blogosphere using the catalog of blogs generated by Technorati or Blogpulse and a textcat-type language identification system – a high research priority for anyone studying the international blogosphere estimates of the size of the international blogosphere have to rely on information from blogging service hosts.This information is notoriously messy – where users self-identify nationality, there’s no guarantee that they will do so accurately.And when blogging services report numbers, they have a commercial incentive to report a large numbers as possible, not removing duplicate or inactive accounts.With those caveats, some pointilistic statistics of country blogopsheres: -

-

Kirill Pankratov, in Moscow newspaper The Exile, observes that Russian bloggers are primarily using the LiveJournal platform24. LiveJournal’s statistics page lists the Russian Federation as the third most popular nation of origin for LiveJournal users, with 251,152 Russian users25. Cyworld, a South Korean internet service, reports that 11 million Koreans have a blog-like “mini-hompy” through their service. Research firm KoreanClick estimates that 90% of Koreans in their 20s have an account on Cyworld26. Skyblog.com, run by French radio station SkyRock, claims to host 2.87 million blogs, with 130 million entries and 182 million blog comments27. A study by the Japanese Ministry of Internal Affairs and Communications estimated 3.35 million bloggers in Japan, with just under 1 million blogs updated once per month or more28. New York Times columnist Nicholas Kristof claims that there are four million weblogs in China – his source is unclear29. Chinese search engine Baidu claims to index 37 million Chinese blogs, authored by 16 million bloggers.30

Rather than trying to resolve the conflicting claims listed above, researcher Matthew Hurst has been attempting to draw some conclusions about blogging behavior from a single day's worth of pingserver data. On the 28th of July, 2005, weblogs.com broadcast 816,704 pings, each representing a new or modified blogpost. Hurst looked at two sets of data in close detail – 262,704 pings representing blogs on the blogspot.com server and 178,128 pings representing blogs hosted by msn.com.The MSN blogs are especially interesting in quantifying the international blogosphere, as 82% of bloggers using the MSN Spaces service provide information about their location in a “profile” page 24

Pankratov,“Censor This!”, Moscow:The Exile, 7/1/2005. http://www.exile.ru/2005-July-01/censor_this_.html, accessed 9/14/2005 25 http://www.livejournal.com/stats.bml, accessed 12/15/2005. 26 Lee,“Souped-Up Blog Takes South Korea by Storm”, International Herald Tribune, 12/31/2004. http://www.iht.com/articles/2004/12/31/ptkorblog_ed3_.php, accessed 9/14/2005. 27 http://www.skyblog.com, accessed 9/14/2005. 28 Jacobsen,“Japanese Blogosphere Catching Up”, posted 5/19/2005. http://www.japanmediareview.com/japan/blog/archive.cfm?start=550, accessed 9/14/2005. 29 Kristof,“Death by a Thousand Blogs”, New York Times, 5/24/2005. 30 China Web Review, 12/20/05, http://web2.blogbeta.com/44.html, retrieved 12/23/05.

associated with their blogs. Of the 119,954 blogs that provide location information, the top five nations represented are as follows: China USA Taiwan Japan Brazil

39584 32.99% 11985 10.00% 10828 9.03% 7436 6.20% 5817 4.85%

Table 1: Results from Matthew Hurst’s “24 Hours in the Blogosphere”31.

It would be a mistake to extrapolate from these figures that there are three times as many Chinese bloggers as Americans. MSN’s service was introduced more than a year after major blogging services in the United States – many US early adopters had already selected a blogging platform and felt no reason to move to MSN’s platform. Other USbased blogging services are blocked by the Chinese firewall, perhaps because they've been less willing than MSN to build a service that restricts the posting of politically sensitive content. Because the MSN Spaces tool is accessible in China and has a Mandarin-language interface, it’s been disproportionately popular with Chinese and Taiwanese audiences32. MSN reported 4.5 million user accounts at the end of March, 200533 – if Hurst’s ratios hold true, that would suggest 1.5 million Chinese and 400,000 Taiwanese blogs on MSN alone34. A future direction for research might combine part of Hurst's method with the method used by the NITLE census. Using pingserver data to determine live, recently updated weblogs, a researcher could retrieve a representative sample of blogpost URLs. She could then retrieve each of these blogposts and identify the language of the post using the textcat system.The result would not be a comprehensive count of blogs in each language, but would help determine a language distribution for an average day's worth of data. Using this distribution and total blog counts from services like Techorati and Blogpulse, it should be possible to estimate the total number of blogs in each language.35 Even if realistic estimates for the number of blogs in each language were available, we would have deeply imperfect information about the national origins of bloggers. In several countries where blogging is widespread, English is the dominant language, which makes it difficult to determine if an English-language blog is from the US, Canada, the UK, Australia, New Zealand… or written by an Arabic speaker attempting to reach a wider 31

Hurst,“24 Hours in the Blogopshere”, 2005.Accepted for publication by the American Association for Artificial Intelligence, from a pre-print copy provided by the author. 32 Hurst looks at reported location information on the Blogspot server, the major competitior to MSN Spaces.While far less information is available, as only 34% of users provide location information, China doesn't even make the top 20 of countries represented on Blogspot... which makes sense as Blogspot is generally blocked by the Chinese firewall. 33 Perseus,“The Blogging Geyser”, 4/8/2005. http://www.perseus.com/blogsurvey/geyser.html, accessed 9/15/2005. 34 Using another extrapolation: 21.8% of the blogs represented in Hurst's ping set are MSN Spaces blogs. If 33% of those blogs are written by Chinese authors, 7.19% of active blogs should be Chinese-authored MSN Spaces blogs. With Technorati estimating over 23 million blogs, that's 1.65 million Chinese blogs on MSN alone. 35 This method works only if popular blogservers in other countries – Bokee.com in China, for instance – report page changes to pingservers, and specifically, use the same pingservers as US blog hosting companies.

audience. Language-based classification also fails to correctly categorize expatriate bloggers, multilingual bloggers, and other cases that may be disproportionately common in the blogosphere36. While IP to physical location services like those used by search engines to offer a localized interface to users would appear to offer a path towards determining the national origins of blogs, this solution fails because most blogs are hosted on large servers located within the US.The rich community of African bloggers on Blogspot all register as “California, USA” to IP translation services as they are all delivered on IP addresses controlled by Google. If weblog hosting sites were willing to cooperate with research efforts, it would be possible to resolve the IP addresses users utilize to post updates to a blog – this information would represent a major step in allowing researchers to understand where bloggers actually are. However, it would only be a useful study if several major hosting providers participated, as there appear to be major self-selection biases that govern who chooses to use which blogging platform. Journalists, Bloggers and Diarists Whether we believe there are 16 or 36 million blogs, it is widely accepted that there are different types of weblogs, in terms of authorship, audience and participation in the linking culture of the blogosphere.A small subset of top blogs have multiple authors, frequently pay their authors, and have online readerships that rival mainstream media sources – Techorati’s David Sifry offers a set of statistics that demonstrates that top blog BoingBoing is linked by bloggers more often than powerful media properties like USA Today37.These top blogs tend to be governed by a different ruleset than the vast majority of blogs: the authors resist being contacted directly by other bloggers; they expect stories on their blogs to be picked up by the mainstream media; they often disable comments. It’s worth thinking of these bloggers as journalists and entrepreneurs, and as distinct from the mainstream of bloggers. Moving to the other end of the spectrum from the journalist blogs linked to by thousands of other bloggers are diarists, who are rarely linked to by anyone. In some cases, bloggers explicitly don’t want to be linked to – Kevin Krim, the head of subscription services for Six Apart, which owns LiveJournal, reports that a quarter of LiveJournal posts are “friends-locked” – they’re not published to the web as a whole, but can be viewed only by individuals the user lists as a friend38.A smaller subset are “private” - they are viewable only by their authors, making them very much like realworld journals, and very different from link-heavy blogs. In other cases, a blogger may be willing to have posts linked to, but is thwarted by the volume of blogposts and comparative paucity of links.An exploration of a set of 9.8 36

Top 100 blogger Joi Ito blogs in both English and Japanese. Leading European blogger Loïc Le Meur blogs in French and English. 37 http://www.technorati.com/weblog/2005/08/39.html, accessed 9/14/2005. 38 Sullivam,“Kids, Blogs and Too Much Information”, 4/29/2005. http://www.msnbc.msn.com/id/7668788/page/2/, accessed 9/14/2005.

million Japanese blog pages by NTT researchers Ko Fuhimura,Takafumi Inoue and Masayuki Sugisaki turned up the remarkable statistic that only 1.15% of blog entries were linked to from other blog entries. Only 16.3% of blog posts linked to any URLs, and only 1.25% linked to another blog39. If the NTT results are consistent across different blogospheres, they imply that the vast majority of blogposts – and likely, the vast majority of bloggers – don’t participate in the widespread systemic linking that has been considered a fundamental characteristic of the blogosphere.These unlinked blogs are less like journalism and more like diaries – they’re intended to be read by the author alone, or by a small set of friends. Rather than documenting links and stories found online - “logging the web”, in the original sense of the term – they generally document daily life40.While some of these diaries might be useful in providing insight into a different culture, providing that insight isn’t the author’s intent. In Derakshan’s model, these blogs are windows, not bridges – you can look into them, but you may not understand what you see inside, and your peering through the window doesn’t constitute an interaction. Between the journalists and the diarists are bloggers.Again, there are dozens of ways to segment the set – by frequency of posts, traffic, number of inbound links, commercial aspirations, etc. In considering the influence of international blogs and the interconnection of local and global blogospheres, links may prove to be a useful metric. Bloggers link to other blogs to announce their interest in a topic and contribute to the conversation – a blog linked to by Kenyan bloggers is likely discussing topics germane to the Kenyan blogosphere; a Kenyan blog linked to by non-Kenyan blogs is likely talking about topics interesting to non-Kenyans, even if that topic is Kenyan politics. If this holds true, looking for highly linked blogs should be one method for identifying key players in local blogospheres (highly linked by bloggers from the same country/language/region) and key bridgeblogs (highly linked by bloggers inside and outside the same country/language/region.) Some hub figures might also be bridge figures. In other cases, a hub figure might be a focal point for a local-language blog community and another (bilingual) blogger might serve as a bridge between that hub figure and the larger blogosphere. Language and linking patterns give us two qualitative criteria to help identify bridge blogs. If a blog is regularly and extensively linked to by webloggers from two or more nations, especially nations that don't share a border or common language, there is a good chance it's a bridgeblog. If a blog is written in a language other than the dominant language of the country where the author is writing, it is also highly likely to be a bridgeblog. (We 39

Fuhimura, Inoue and Sugisaki,“The Eigenrumor Algorithm for Ranking Blogs”, presented at WWW 2005, May 10, 2005, Chiba, Japan. http://64.233.167.104/search?q=cache:6ZL_PIoGYPwJ:wwwidl.hpl.hp.com/blogworkshop2005/fujimura.pdf+eigenrumor, accessed 9/14/2005. It's quite possible that linking behavior in the Japanese blogosphere, which is widely considered less political and more personal than the US blogopshere, may be very different than linking behavior in the US. 40 Japanese philosopher and cultural commentator Hiroki Azuma believes these blogs are a form of primate “grooming behavior” - these blogs are important not because of their content, but because the process of commenting on these blogs reinforces the social ties between the members of a community. From a conversation with Mr.Azuma, December 2, 2005.

can imagine exceptions to this rule – a blog that includes technical information useful to webloggers in many nations might be internationally linked, but not serve to bridge between cultures.) Big in Japan Weblog search engine Technorati maintains a rankings page41 which lists the top 100 weblogs, as determined by total incoming links over that blog’s lifespan.This listing is more static than many other ways of ranking popular blogs, as the least popular page in the top hundred in September 2005 had 2,430 incoming links.A new blog would need thousands of new links to make this exclusive list (By contrast, Blogpulse’s top 40 page, which ranks blogs based on incoming links discovered in the past 30 days, tends to be significantly more dynamic, needing a few hundred links in a month to make the top blog lists.) Many of the blogs on Technorati’s top 100 have built up a substantial audience over the course of months or years. Examining the Technorati Top 100 list from mid-September 2005, 15 of the weblogs listed were identifiably from authors outside of the United States.

Weblog URL http://www.interney.net

Language Portuguese

http://beccary.com/ English http://spaces.msn.com/members/sucaiku/ Chinese http://riverbendblog.blogspot.com/ English http://xiaxue.blogspot.com/ English http://joi.ito.com/ English http://www.brokenkode.com/ English http://spaces.msn.com/members/jimmymsnspace/ Chinese http://gatofedorento.blogspot.com/ Portuguese http://www.beppegrillo.it/ Italian http://wahlblog.de/ German http://www.abrupto.blogspot.com/ Portuguese http://spaces.msn.com/members/princesscecicastle/ Chinese http://manabekawori.cocolog-nifty.com/blog/ Japanese http://spaces.msn.com/members/d3vmax/ English Table 2, data retrieved 9/14-15/2005 from technorati.com.

Blogger Technorati Location Rank42 20 Hong Kong 36 37 Iraq 46 Singapore 51 Japan 55 Lebanon 56 Taiwan 57 65 Italy 66 Germany 70 Portugal 72 Canada 86 98 India, UK 99

Analyzing links to these blogs generated in the past one to three days43, three patterns emerge. Several blogs are most likely widely linked because they provide services that are useful to other bloggers. Beccary (beccary.com) and BrokenKode (www.brokenkode.com) are heavily linked because links to their blogs appear on popular 41

http://technorati.com/pop/blogs/ As of 9/14/2005, 11AM EDT. 43 Technorati lists from 10 to 20 incoming links, generated in the past 3 days, on the first page of results for a blog’s “link cosmos”. My cursory analysis is based on the first page of cosmos results, which varied from 11 to 20 links. 42

weblog templates.While BrokenKode (a Lebanese graphic designer) has incoming links from blogs written in Romanian, Malay, Swedish, German and Chinese, it’s likely that there’s little dialog taking place between those bloggers – instead, they’re all using the cool template with the coffee cup. MSN spaces pages that provide graphics and information are also widely internationally linked. PrincessCeciCastle (http://spaces.msn.com/members/princesscecicastle/) , a Canadian blogger who writes in Chinese, is linked by bloggers who write in Japanese, Chinese and Thai. In all cases, the blogs that link to her are dominated by long strings of icons – PrincessCeci evidently hosts many of these icons on her blog, and bloggers aren’t shy about crediting her for sharing these images. HackMSN (http://spaces.msn.com/members/d3vmax/) provides information for customizing MSN Spaces blogs. Recent links to the site are in Japanese, Chinese (written by a 13-year old Malaysian) and in English, with authors from the US, UK and Australia. Evidently the incentive to customize a blog is sufficient to encourage bloggers to overcome linguistic barriers and link across language barriers. Sukaicu also provides information on customising MSN, but has links only in Chinese. Seven of the 15 blogs in the set are only recently linked by blogs written in the same language; Interney.net, a Portuguese blog is, bizarrely, linked to by 16 Portuguese blogs and one in Bahasa Indonesia44.These blogs are likely central figures in local, languageconstrained blog communities. It’s likely that many German bloggers read Wahlblog.de, while many Italian bloggers read bepegrillo.it – and less likely that Italians read Wahlblog.de, if only for reasons of language barriers. Four blogs in the set: Beppe Grillo (beppegrillo.it), Baghdad Burning (riverbendblog.blogspot.com), Joi Ito (joi.ito.com) and Xiaxue (xiaxue.blogspot.com) can be characterized as bridgeblogs.While Beppe Grillo addresses Italian issues and is widely linked to by Italian bloggers, it also appears in English and provides an insight into Italian politics for non-Italian speakers – as a result, it is occasionally linked to by non-Italian speakers.Written by an Iraqi woman in English about Iraqi current events, Baghdad Burning has recent links in English and French – bloggers writing in English identify themselves as coming from the US, India, Palestine and sub-Saharan Africa. Joi Ito is a bilingual, bicultural45 Japanese entrepreneur – recent links to his blog include two from Japanese language blogs, one from a Spanish language blog, and eight from Englishlanguage blog. In one case, a Japanese blog is translating Joi’s English words into Japanese. Xiaxue (Wendy Cheng) is a young Singaporean woman, known for posting glamorous photos of herself on her blog – accompanying texts are in English. Recent blogs linking to her page have been in English and Chinese; several identify themselves as blogging from Singapore or Malaysia.While the appeal of her blog doesn’t appear to extend beyond Southeast Asia, within the region, it evidently crosses cultural and language 44

The link comes from a blog related to Indonesia Blogshares, an online contest that mimics a stock market, using blog link rankings instead of share prices. In this context, the linguistic disparity is less surprising as Interney is historically one of the most popular blogs on the web. 45 Joi Ito grew up in Tokyo and Detroit and is fluent in English and Japanese.

barriers. It’s worth noting that all three apparent bridgeblogs are written in English. Most blogging platforms, link tracking tools and search engines are available in English, and many early adopters of blogging had at least some English fluency.As bilingual blogger Loïc Le Meur, puts it,“english is the TCP/IP of language46” in the blogosphere.There’s an intriguing possibility that Chinese will emerge as a linguistic rival for bridgeblogs in East Asia as bloggers in China,Taiwan, Singapore and Malaysia communicate across borders. Looking at the same Technorati Top 100 page three months later, the results are remarkably different. Xiaxue and Beppe Grillo have vaulted in the standings (she from #51 to #29, he from #66 to #12), but they're two of only four of September's international entrants in the top 100 still ranked in December. ( Chinese/Canadian princesscecicastle jumped from #86 to #31, and Japanese blogger Kaori Manabe leapt from #98 to #23.) While the other 12 international blogs listed in September have dropped from the top 100, they've been replaced by 20 new international blogs in the top 100. Weblog URL Language http://www.beppegrillo.it/ Italian, English http://spaces.msn.com/members/mmadrigal/ Spanish http://manabekawori.cocolog-nifty.com/blog/ Japanese http://spaces.msn.com/members/locker2man/ Chinese http://xiaxue.blogspot.com/ English http://spaces.msn.com/members/princesscecicastle Chinese http://spaces.msn.com/members/huangmj/ Chinese http://spaces.msn.com/members/gb2312 Chinese http://spaces.msn.com/members/silicn Chinese http://blogs.yahoo.co.jp/nosz50j Japanese http://spaces.msn.com/members/slim Chinese http://spaces.msn.com/members/MSN-SA Chinese http://spaces.msn.com/members/chattie555 Chinese http://spaces.msn.com/members/gouy2k Chinese http://blog.livedoor.jp/insidears Japanese http://spaces.msn.com/members/king19120/ Chinese http://blog.livedoor.jp/aki09041/ Japanese http://spaces.msn.com/members/lwhanz1981530/Chinese http://www.bildblog.de/ German http://spaces.msn.com/members/atiger Chinese http://spaces.msn.com/members/hcy521/ Chinese http://blog.so-net.ne.jp/atsuya-furuta Japanese http://www.microsiervos.com/ Spanish http://www.kennysia.com/ English

Blogger Location Italy Spain Japan China Singapore Canada Chinese USA Japan

Canada Australia

England Germany USA China Japan Spain Singapore

Technorati Rank Three Rank Months Prior 12 66 22 23 98 24 29 51 31 86 33 34 35 37 46 54 57 58 60 66 73 76 78 79 81 89 93 97

Table 3, data retrieved December 12, 2005, from technorati.com.

Five of the top blogs are in Japanese.Two are weblogs by Japanese celebrities – swimsuit model Kaori Manabe and baseball player Atsuya Furuta. Encouraging celebrity blogging appears to be a strategy of Japanese blogging providers like Nifty, which has named Manabe to head a committee on the promotion of weblogging47.This suggests a possible 46

Le Meur,“The world goes beyond english speaking blogs”, 8/23/2005. http://www.loiclemeur.com/english/2005/08/technorati_and_.html, accessed 9/15/2005. 47 From Mainichi Daily News, December 5, 2005, accessed at http://mdn.mainichimsn.co.jp/entertainment/news/20051205p2a00m0et040000c.html on 12/19/2005.

trend worth watching – will more Japanese celebrities start blogging? Or as blogging increases in popularity in Japan, will bloggers become famous for their online activities and make offline celebrity blogs less relevant? Will other blogospheres encourage celebrities to start blogs in the hope to encouraging local-language blogging communities. Twelve of the 24 leading non-English/non-US blogs are in Chinese.At least nine of the 12 focus solely, or in part, on ways to customize or hack the MSN Spaces service, some offering downloadable tools to assist with customization.The MSN customization phenomenon isn't limited to Chinese-language blogs – Herramientas para Blogs48 is a collection of tips and tricks in Spanish for customization of MSN spaces, offered by a 39year old IT professional in Madrid – it ranks 22nd on Technorati's list. But Chineselanguage bloggers from all over the world seem to be leading the movement – the new customization blogs are maintained by bloggers in the US, Canada,Australia, as well as China49.The linking patterns to these customization sites suggests interlinking between Chinese speakers around the world – links to “locker2man”, one of the major customization and HTML resource sites – has recent links from Taiwan, Malaysia and New Zealand as well as from China.These MSN customization sites appear to be linked to primarily from other MSN Spaces sites, suggesting this community may not just be linguistically closed, but closed to users of other popular Chinese blogging platforms. Two of the three other new blogs in the top 100 are European blogs aimed at European audiences – a German-language blog about the German media, and a Spanish-language site about the European technology industry.The third, Kennysia.com, is a Malaysian bridgeblog in the spirit of Xiaxue50, linked to extensively by other Malaysians, but also by Singaporean,American, and Australian bloggers. The disappearance of Joi Ito (now ranked #172) and Baghdad Burning (no rank listed, but slightly fewer links than Joi Ito) from the top 100 prompts two hypotheses – that bloggers are less interested in linking to these bridgebloggers than they were three months ago, or that the emergence of strong local blogospheres in other languages (Chinese, Japanese, Spanish and German) is creating more competition for the top spots. As heroes of local blogospheres emerge (like Beppe Grillo in Italy), it's possible that we will see less linking across linguistic lines as bloggers are able to link to leading bloggers who write in their own language. If this hypothesis is true, we would be expect to see more popular blogs emerge in languages that aren't widely spoken as a second language, like English. Big fish, small pond While writing in English is apparently helpful in creating a well-read bridgeblog, it’s not a sufficient condition for success. BlogStreet is a service that tracks blog popularity based 48

The title translates as “Tools for Blogs”. http://spaces.msn.com/members/mmadrigal/, accessed 12/19/2005. Location information self-reported by bloggers on their “profile” pages – there's no guarantee that this information is correct. 50 Though Sia would protest that he's a more “serious” blogger than Xiaxue and prefer a comparison to Singaporean blogger Mr. Brown (http://www.mrbrown.com). 49

on links in weblog “blogrolls” – static collections of links to other weblogs a user reads regularly, wishes to promote, or wishes to display a social connection to51. BlogStreet India provides a list of the hundred blogs most frequently listed on the blogrolls of Indian blogs. In most cases, these blogs are also maintained by Indians, or highly India-focused, though a few highly popular non-Indian blogs place in the listings52.While these blogs are highly influential within the Indian blogosphere, none register in the top blog lists on Technorati or Blogpulse: Blogstreet Technorati Technorati Technorati Blogpulse Blogpulse Links links Sites Ranking links ranking http://www.kiruba.com/ 122 749 318 3082 132 1888 http://www.lazygeek.net/ 108 870 262 4087 91 2387 http://gauravsabnis.blogspot.com/ 92 623 597 542 http://www.anitabora.com/blog/ 82 299 148 9062 46 5998 http://www.selectiveamnesia.org/ 68 400 131 10622 65 3264 http://indiauncut.blogspot.com/ 58 2245 666 1062 684 649 http://www.emergic.org/ 58 364 363 1286 http://www.yazadjal.com/ 54 300 119 11958 102 4899 http://vinodg.blogspot.com/ 50 271 136 10122 17 http://mdeii.blogspot.com/ 47 49 16 http://www.madhoo.com/ 46 335 99 15413 91 5214 http://sidin.blogspot.com/ 43 146 20 http://www.ravikiran.com/ 43 115 126 2089 http://theultimatereality.blogspot.com/ 41 117 55 35607 13 http://quatrainman.blogspot.com/ 41 82 71 4255 http://youthcurry.blogspot.com/ 39 1276 536 1506 341 http://sudhon.blogspot.com/ 35 64 5 http://indianbloggers.blogspot.com/ 34 50 6 http://varnam.org/blog/ 34 96 119 4513 http://dcubed.blogspot.com/ 34 203 175 1967

Table 4, data retrieved 12/13/2005, from india.blogstreet.com, blogpulse.com and technorati.com

The blog with the largest number of incoming links as tracked by both Techorati and Blogpulse is IndiaUncut (indiauncut.blogspot.com).This frequently updated blog features collections of links from across the blogosphere, with a strong Indian focus, and maintains an excellent blogroll of other top Indian bloggers. In other words, IndiaUncut serves as one of the major hubs of the Indian blogosphere, amplifying stories from the mainstream media to the blog community, as well as posts made on less widely read Indian blogs. The blog with the lowest (i.e., most prominent) Blogpulse rank,“Vantage Point”(gauravsabnis.blogspot.com), is an interesting case study in controversy and linking behavior. In July 2005, Indian youth magazine JAM published an exposé of the Indian Institute of Planning and Management, a graduate school that advertised widely in mainstream Indian newspapers.The article accused IIPM of misrepresenting its facilities, 51

danah boyd’s “The Biases of Links” (http://www.zephoria.org/thoughts/archives/2005/08/07/the_biases_of_links.html) provides some useful insights about the limitations determining blog popularity through blogrolls. 52 American political blog, Instapundit, ranks 17th in the Blogstreet India blogroll rankings, as accessed at http://india.blogstreet.com/top100.html on 9/14/2005.

accreditation and job placement records. Indian blogger Gaurav Sabnis linked to the JAM article and added the accusation than IIPM founder failed his graduation exams on his first attempt53. IIPM reacted strongly to the article, threatening Sabnis and several other Indian bloggers who'd linked to the piece with libel suits. (They also are alleged to have conducted a campaign to smear Indian bloggers with insulting comments and rival weblogs.) They took a step further with Sabnis, calling his employer – Lenovo – and threatening to hold a public burning of all IBM/Lenovo laptops used by IIPM unless Sabnis removed his post54. Out of concern for his employer, Sabnis resigned, and became something of a folk hero to the Indian blogopshere. Rashmi Bansal – the editor of JAM magazine – has also seen her blog Youth Curry (youthcurry.blogspot.com) benefit from the IIPM attention – it ranks third in this set of Indian blogs in terms of incoming Technorati links55. The Indian blogosphere has received major international attention around the IIPM situation, including links in prominent American blogs like Instapundit. In the aftermath of the Indian Ocean Tsunami, Indian blogs, especially the South East Asia Earthquake and Tsunami Blog (usually called “SEA-EAT” - tsunamihelp.blogspot.com) received major coverage in Western mainstream media as well as US and European blogs.Yet the most prominent of Indian blogs – IndiaUncut – ranks #1062 on Technorati56.While many of the top Indian blogs could be considered as bridgeblogs, the modest ranking of India’s best blogs on a global scale suggests at least two possible explanations: that India’s blogosphere is so small57 in comparison to US, Chinese and Japanese blogospheres that top Indian bloggers are crowded out of global rankings; or that blogs focused on Indian politics are, generally speaking, less interesting to a global readership than blogs about technology, entertainment or US politics. Skyblog URL http://justforakiss.skyblog.com/ http://mon-ptit-coeur.skyblog.com/ http://enyasims.skyblog.com/ http://blog-a-voir.skyblog.com/ http://skatastr0ophe.skyblog.com/ http://victim19.skyblog.com/ http://sachablog.skyblog.com/ http://bell-goss-2006.skyblog.com/ http://so-chic-so-class.skyblog.com/ http://americangirl7.skyblog.com/ 53

Skyblog Technorati Blogpulse Visits Sites sites 330888 8 3 247479 8 34 211698 0 0 176283 17 0 165059 0 0 158470 12 33 146411 0 0 136649 0 0 128772 1 1 125683 0 0

Sabnis,“The Fraud that is IIPM”, http://gauravsabnis.blogspot.com/2005/08/fraud-that-is-iipm.html, accessed 12/20/05. Reported by Sabnis on his blog, 10/10/2005. http://gauravsabnis.blogspot.com/2005/10/update.html, accessed 12/20/05. See also Mark Glaser,“Business School Flap a 'Breakout Moment' for Indian Blogosphere”, 10/26/2005, http://www.ojr.org/ojr/stories/051026glaser/, accessed 12/20/05. 55 I've written at more length about the IIPM situation in a blog post titled “Indian Bloggers put IIPM's Feet to the Fire”, 10/14/2005, http://www.ethanzuckerman.com/blog/?p=208. Neha Viswanathan, writing for Global Voices, provides an overview of the situation,“India: Defending Freedom of Speech”, 10/10/2005, http://cyber.law.harvard.edu/globalvoices/2005/10/10/india-defending-freedom-of-speech/. 56 IndiaUncut ranks #649 on Blogpulse, which tracks links over a shorter period of time than Technorati, and suggests that IndiaUncut's Technorati ranking is likely to rise. 57 Nivedita Mookerji, in an article titled “Blogosphere Takes off in India”,The Financial Express, 5/25/2005, asserts that there are 100,000 “India-centric” blogs. It’s unclear where this figure originates or if it is accurate. http://www.financialexpress.com/fe_full_story.php?content_id=91895, accessed 9/15/2005. 54

Table 5, data retrieved 12/20/05 from skyblog.com, technorati.com and blogpulse.com

While the BlogStreet India list tells us that A-List Indian bloggers make the “B” or “CList” of global bloggers, it’s clear that some of the bloggers featured here are linked to on a global scale.This is less clear for top 100 blogs listed by French bloghost Skyblog.net58 – five of the top ten blogs listed on Skyblog's site have no links from another blog as tracked by Technorati or Blogpulse.The best linked two of these top Skyblogs have 17 links on Technorati and Blogpulse respectively. While there may be 2.8 million French-language Skyblogs, it’s likely that most are not engaged in the global linking practice tracked by blog search engines. This does not, however, mean these blogs are without political impact.Three Skyblog users were arrested for inciting violence based on blogposts that encouraged rioters in Ile-de-France and other Paris suburbs to burn down police stations. Skyblog responded by stating they would remove any content they deemed overly inflamatory.59 A whirlwind world tour of local blogopsheres While the magnitude of some local blogospheres can be grasped by examining global blog rankings, understanding the impact of blogs in other nations requires looking closely at blogger communities and government and media reaction to those communities.A whistle-stop tour of selected national blogospheres points to the importance of local context in understanding blogging behavior, but also hints at themes that cut across national communities. Iran The early and rapid rise of weblogging in Iran has led to an awareness of the Iranian blogging community in the blogosphere as a whole, though understanding of the contents of the Iranian weblog community – “weblogestan60”, as it has been termed by some Iranian bloggers – has been limited by language and cultural barriers. In September 2001, Salman Jariri established the first Persian language blog; in November of 2001, Hossein Derakshan published weblogging instructions in Persian on the blogger.com site, encouraging fellow Iranians to begin blogging in Persian61. Less than a year later, PersianBlog began offering a blogging platform fully localized into Farsi, with free user accounts. In 2004, the NITLE weblog census reported over 75,000 weblogs in Farsi, making it the third-best represented language in the blogosphere62. It’s unclear whether the weblogs detected by NITLE were entirely from within Iran, or whether some percentage were maintained by expatriate Iranians. 58

From http://www.skyblog.com/php/detail.php/page/top100, accessed 9/14/2005. Thomas Crampton,“Blogs and text messages spread call to violence”, International Herald Tribune, 11/9/2005, http://www.iht.com/articles/2005/11/08/news/blogs.php accessed 12/20/05. 60 Doostdar, "The Vulgar Spirit of Blogging: On Language, Culture, and Power in Persian Weblogestan” American Anthropologist 106(4). http://www.doostdar.com/articles/vsob.pdf, accessed 9/14/2005. 61 I’m relying on Wikipedia’s “Iranian Weblogs” entry for a timeline of the Iranian blogosphere. http://en.wikipedia.org/wiki/Iranian_blogs, accessed 9/15/2005. 62 NITLE was almost certainly undercounting Chinese and Japanese blogs, but Iranian bloggers took a great deal of pride in their third place status. 59

Blogging in Iran rapidly became political, controversial and, sometimes, dangerous for the blogger. Reporters Sans Frontières, in a section of their “Internet Under Surveillance” Report on Iran63, speculates that independent journalists turned to the web as the government ordered the closure of more than 100 newspapers. Iranian journalist and blogger Sina Motallebi was arrested in April 2002 and held for 23 days, apparently in connection with an entry on his blog defending a fellow reporter from reformist newspaper Hayat-é-No who had been arrested some months earlier. Dozens of Iranian bloggers have been detained, generally for short time periods. Others have been imprisoned for extended stretches, like blogger Mojtaba Saminejad, who was sentenced to two years in prison for “insulting the Supreme Guide”64. While Iranian authorities have taken steps to prosecute journalists using weblogestan as a platform for government criticism, some political figures have embraced it as a way of communicating with a younger generation of constituents. In November 2003, Iranian Vice President for Legal and Parliamentary Affairs, Mohammed Ali Abtahi, started a weblog,Webnevesht, featuring photographs and posts in Farsi,Arabic and English.Abtahi surprised many observers of Iranian politics by supporting Iranian blogger Sina Motabelli with a passionate post attacking the detention of Motabelli’s father. Other Abtahi posts critiqued the closure of reformist English-language newspapers65.Abtahi has subsequently resigned from his cabinet position, but remains active in the Militant Clerics Society66 (an Iranian political party) and an active weblogger. While Iranian weblogs have hosted extended, passionate debates67, Iranian bridgebloggers have been less successful in opening these conversations to the wider blogosphere. IranFilter.com began as a project to expose Iranian blogs, translated into English, to the wider blogosphere; it recently re-launched as a site filtering news headlines and linking to mainstream media stories that address Iran – the shift was due, in part, to difficulties getting Iranians to translate relevant Farsi blog posts into English68. Hossein Derakshan’s English weblog, probably the most prominent of Iranian weblogs in English, ranks 2571 on Technorati (cumulative incoming links), 316 on Blogpulse (recent incoming links), suggesting an ongoing interest in English-speaking countries in what's being talked about in the Persian blogosphere and a desire to cross language barriers69. 63

Reporters Sans Frontières,“The Internet Under Surveillance”. http://www.rsf.org/article.php3?id_article=7268, accessed 9/15/2005. 64 Saminejad had also been accused of “insulting the prophet”, a crime punishable by death, but was acquitted of that charge. From a Reporters Sans Frontières news release, 7/11/2005, http://www.rsf.org/article.php3?id_article=12564, accessed 9/15/2005. 65 Glaser,“Iranian Bloggers Protest Government Crackdown on Reformist News Sites”, http://ojr.org/ojr/glaser/1095807595.php, 9/22/2004. accessed 9/14/2005. 66 From Wikipedia article “Mohammed Ali Abtahi”, http://en.wikipedia.org/wiki/Mohammad_Ali_Abtahi accessed 9/15/2005. 67 Notably the “vulgarity” debate, explored at length by Doostdar in her American Anthropologist paper. 68 From discussions with Hossein Derakshan between December 2004 and present. 69 A comparison of links to Derakshan's blogs in English and Persian on Blogpulse and Technorati suggest that Derakshan is becoming increasingly popular in the English-language blogosphere and less popular in the Persian blogosphere. Searches on Technorati, which track long term popularity, rank his Persian blog at 2376th, and his English blog at 2571st. Blogpulse, which tracks links in the past month, list his English blog at 316th and his Persian blog at 9712nd.This effect is due, in part, to widespread censorship of Derakshan's Persian blog in Iran and fewer restrictions on his English blog.

Despite early enthusiasm about the rise of weblogs leading to the rise of more liberal regimes in Iran, the election of arch-conservative Mahmoud Ahmadinejad has given the Iranian blogosphere a “Howard Dean” moment, a realization that opinions expressed in blogs may not be reflective of opinions in the wider populace.A June 2005 BBC article on Iranian weblogs and the upcoming election didn't mention Ahmadinejad by name, focusing on debates over whether or not to boycott the election, and blogger support for liberal candidate Mostafa Moin over frontrunner Ali Akbar Hashemi Rafsanjani70. Some Iranian bloggers would argue that Ahmadinejad's election was not representative of the popular will. Moin, the candidate most strongly supported by Iranian bloggers, accused the Guardian Council of carrying out a campaign to sway the election in Ahmadinejad's favor71, and some Iranian bloggers believe at least the first round of the election was rigged. China If Iran surprised the blogosphere with its presence in late 2003, bloggers are just now getting used to the notion that they share the Internet with millions of Chinese bloggers. While there are few believable figures for the total number of Chinese weblogs, there are a few clues that the Chinese blogosphere is huge and growing: -

-

In September 2005, 3 of the top 100 sites on Technorati were in Chinese, tying it for second place with Portuguese for the best-represented language. In December 2005, 13 were in Chinese, vastly outpacing all other languages besides English Matthew Hurst’s recent research suggests that 33% of posts to Microsoft’s popular MSN Spaces service come from Chinese blogs. The founder of Chinese blog hosting site Bokee claims 2 million blogs hosted on his service, of five million total Chinese blogs.72

One possible explanation for recent growth in the Chinese blogosphere is “Circular 17” a decree from the Chinese Education Ministry, which established new rules for use of popular online bulletin board systems hosted by Chinese universities. Prior to the decree in February 2005, bulletin board systems were used by hundreds of thousands of users, most of whom were not university students73.As the decree came into force, nonuniversity users were forced off the bulletin boards. Many may be turning to weblogs to find another space to articulate political and social concerns. 70

Mehdi Jami,“Iranian blogs take on the election”, 6/17/2005, BBC Persian. http://news.bbc.co.uk/1/hi/world/middle_east/4099380.stm, accessed 12/20/05. 71 “Iranian election 'not fraudulent'”, BBC News, 6/20/2005, http://news.bbc.co.uk/2/hi/middle_east/4110792.stm, accessed 12/20/05. 72 Jamil Anderlini,“Blog Founder Seeks Success Writ Large”, South China Morning Post, 7/12/05. http://www.asiamedia.ucla.edu/article.asp?parentid=26700, accessed 12/20/05. 73 Marquand,“Popular University Websites Restricted”, 6/4/2004. http://www.indexonline.org/en/news/articles/2005/2/china-popular-university-websites-are-restri.shtml, accessed 9/15/2005.

As political speech has moved into the sphere of weblogs, Chinese authorities are showing an interest in ensuring that such speech is not anonymous. In June 2005, Chinese authorities announced that all Chinese bloggers and website owners would be required to register their sites with authorities, using their real names – unregistered users were threatened with arrest74. In practice, the Public Security Bureau only required owners of standalone domains to register their sites – bloggers who had accounts on bloghosting sites like Blogger or MSN Spaces were not required to register with authorities75. Microsoft raised the ire of some Chinese bloggers when it became apparent that their Chinese-language MSN Spaces service prohibited the use of certain words – including the Chinese words for “democracy” and “human rights” in the title of a blog76. Users who attempted to create blogs with controversial titles received an error message that read in part “The title must not contain prohibited language, such as profanity.” Chinese users responded almost immediately by using modified censorship circumvention instructions by Bennet Hazelton77 and translating these instructions into Chinese.78 Microsoft is not alone – Chinese blogging companies routinely censor blogs as well, sometimes blanking out offensive words, sometimes blocking posts with offensive words from being posted, sometimes removing offensive posts after publication.79 While some projects exist to bridge between the world of the Chinese “bo ke”80 and the wider blogosphere, notably Xiao Qiang’s China Digital Times (chinadigitaltimes.net), the language barrier and self-censorship by Chinese authors make it difficult to get an accurate sense of attitudes and opinions within the Chinese blogosphere. Rebecca MacKinnon, whose fluency in Chinese makes it possible for her to read untranslated posts from the Chinese blogosphere, is fiercely critical of American pundits like Nick Kristof who believe the spread of the Internet will lead inexorably to a democratized China81:“In fact, the bulk of voices in Chinese cyberspace today are more anti-American, more anti-Japanese, and more inclined to go to war against Taiwan than the Chinese government is.This is made more acute by the skewed mix of information Chinese internet-users are exposed to thanks to the Great Chinese Firewall…”82 It’s possible 74

Marquand,“China Cracks Down on Web and Expats”, Christian Science Monitor, 6/10/2005. http://www.csmonitor.com/2005/0610/p01s02-woap.html, accessed 9/15/2005. 75 From an interview of Isaac Mao by Rebecca MacKinnon, 6/10/2005 http://cyber.law.harvard.edu/globalvoices/?p=221, accessed 9/15/2005. 76 MacKinnon,“Screenshots of Censorship”, 6/16/2005. http://rconversation.blogs.com/rconversation/2005/06/screenshots_of_.html, accessed 9/15/2005. 77 Haselton, MacKinnon,“How to Hack Chinese MSN Spaces to Use Banned Words”, 6/15/2005, http://cyber.law.harvard.edu/globalvoices/?p=232, accessed 9/15/2005. 78 “Portnoy”, blogpost, 6/16/2005. http://www.bigsound.org/portnoy/weblog/000813.html, accessed 9/15/2005. 79 Forthcoming research from Rebecca MacKinnon, from email correspondence with the author. 80 Qiang,“The ‘Blog’ Revolution Sweeps Across China”, New Scientist, 11/24/2004, http://www.newscientist.com/article.ns?id=dn6707, accessed 9/14/2005. 81 MacKinnon,“Kristof’s World vs. Real Life”, 5/25/2005. http://rconversation.blogs.com/rconversation/2005/05/kristofs_world_.html, accessed 9/15/2005. 82 MacKinnon,“A long post: China, internet discourse & censorship”, 12/1/2004. http://rconversation.blogs.com/rconversation/2004/12/the_china_probl.html, accessed 9/15/2005.

that, if the blogosphere had good access to the opinions and views inside the Chinese blogosphere, early communication would be at least as much about confrontation as about dialogue. East Africa Kenya has emerged as the unlikely capital of weblogging in sub-Saharan Africa. One Kenyan weblog ring, KenyaUnlimited, tracks over 100 weblogs by Kenyans within the country and in the diaspora83. Prominent Kenyan weblogs, like Mental Acrobatics (mentalacrobatics.com) have been online since at least March 2003.While there are likely more active webloggers in South Africa, many South African blogs are indistinguishable from US technology blogs; Kenyan blogs tend to be identifiably Kenyan, and more likely to be focused on local politics and culture. The Kenyan blogosphere has been characterized by a strong sense of community identity. Many Kenyan bloggers maintain extensive blogrolls, pointing to dozens of other Kenyan blogs. Most Kenyan blogs belong to Nchi Yetu Daily (daily.nchiyetu.com), an aggregator which features the most recent posts from Kenyan blogs on its front page. KenyaUnlimited provides some background materials about weblogs and ways to get started with blogging, in both English and Swahili.84 While most American bloggers may be unaware of the Kenyan weblogging scene, Kenyan newspapers are clearly paying attention.“The Thinker”, a Kenyan IT professional who blogs at “Thinker’s Room”, posted a satirical help wanted ad on his blog, recruiting new ministers for the Kenyan cabinet on February 15, 200585. On February 25, 2005, Clay Muganda, a popular columnist for Nairobi’s “The Nation”, reproduced the entirety of Thinker’s post in his weekly political column, lamely noting that the original version of the piece was available “on the internet”86. Kenyan bloggers and blog readers responded angrily, emailing and calling The Nation, and eventually received a printed apology in the paper, which agreed that Muganda’s actions constituted plagiarism87. The Kenyan blogosphere cemented its reputation for good humored political punditry with extensive coverage of the November 2005 referendum on amending the Constitution. Kenyan bloggers had having a field day with the Electoral Commission’s choice of symbols for the referendum, an orange signifying “no” and a banana signifying “yes”. Ory Okolloh, a young Kenyan lawyer who blogs as “Kenyan Pundit”, wryly observed that oranges had already sold out in a local market88.A roundup of Kenyan blogs shortly before the vote featured commentary from a dozen Kenyan blogs, including 83

http://www.kenyaunlimited.com/kbw/kbwblogs.php, accessed 9/15/2005. See http://www.kenyaunlimited.com/kublogu/, an introduction to blogging in Swahili.Accessed 12/20/05. 85 “The Thinker”,“Job Openings”, 2/15/2005, http://thinkersroom.blogspot.com/2005/02/job-openings_15.html, accessed 9/15/2005. 86 Muganda,“Clay Court: Houses of Corruption”,The Nation, 2/25/2005. http://allafrica.com/stories/200502240777.html, accessed 9/15/2005. 87 “The Thinker”,“The Clay Court Affair, Part IV”, 3/8/2005. http://thinkersroom.blogspot.com/2005/03/clay-courtaffair-part-iv.html, accessed 9/15/2005. 88 Okolloh,“Bananas (apples?) and Oranges”, 9/8/2005. http://www.kenyanpundit.com/?p=8, accessed 9/15/2005. 84

one written in Gikuyu89. Thus far, Kenyan authorities have not appeared threatened by blogs, and there are no known cases of Kenyans being arrested for online publishing activities. One reason blogs may be unthreatening to Kenyan authorities is that the vast majority of Kenyan blogs are written in English, rather than Kiswahili.While both are official languages, Kiswahili is more widely spoken, and English is generally the language of upper-class, wealthier Kenyans. Kenyan bloggers are unlikely to influence the mass of Kenyan citizens by blogging in English – most Kenyans do not or cannot access the Internet and are less comfortable reading English than Swahili. On the other hand, Kenyans may be effective in bringing their perspective into international debates by participating in the English language blogosphere – Kenyan bloggers were prominently featured in a recent Blogosphere debate over the utility of the Live8 concert series90 and surprised American and European bloggers by their vigorous criticism of the aims and goals of Live8.Alan Conor, writing for the BBC, featured several African blogs in his 7/4/05 article “Music blogs close ears to Live8”91. Many of the posts he featured received an influx of traffic from European and US commentators who otherwise might not have heard these African voices.“Thinker's” two posts on Live 892 received 66 and 39 comments respectively. Commenters identified themselves as from Kenya, Nigeria, Zimbabwe and South Africa as well as from Europe and North America. Posts by other prominent African bloggers received dozens of additional comments. The emerging blogosphere in Tanzania stands in sharp linguistic contrast to its northern neighbor. Of the blogs listed on a wiki page dedicated to documenting the Tanzanian blogosphere, 19 blogs are written in Kiswahili, five in English, and four are listed as “bilingual”, though they contain largely Kiswahili content93.While English and Kiswahili are both official languages in Tanzania as is Kenya, English is not as widely spoken, and speaking Kiswahili is seen by some Tanzanians as a source of national pride. The Tanzanian blogosphere can largely be traced back to the efforts of one man, Ndesanjo Macha, a Tanzanian journalist, activist and blogger who appears to have a personal relationship with virtually all Tanzanian bloggers. Macha currently resides in the US, where he teaches Kiswahili – his fondness for the language may have set the linguistic tone for the national blog community. Recently, Macha has started blogging in English, as well as publishing summaries of conversations taking place in the Kiswahili

89

Nish Matenjwa,“Kenya and the Constitutional Referendum”, Global Voices, 11/21/05. http://cyber.law.harvard.edu/globalvoices/2005/11/21/kenya-and-the-constitutional-referendum/, accessed 12/20/205. 90 See “Thinker’s” post “Live 8? Please!”, 6/28/2005 http://thinkersroom.blogspot.com/2005/06/live-aid-please.html, as well as this author’s roundup of African voices on Live8, http://cyber.law.harvard.edu/globalvoices/?p=263, 6/29/2005. 91 Alan Conor,“Music blogs close ears to Live8”, BBC Magazine, 7/4/05 (http://news.bbc.co.uk/1/hi/magazine/4649043.stm, accessed 12/20/05.) 92 http://www.thinkersroom.com/blog/2005/06/live-aid-please/ and http://www.thinkersroom.com/blog/2005/07/getreal-global-politics-101-live-8/, accessed 12/20/05. 93 http://cyber.law.harvard.edu:8080/globalvoices/wiki/index.php/Tanzania#Bilingual_Blogs, accessed 9/15/2005.

blogosphere94, hoping to open the closed conversations to a wider audience. Macha can be seen as a bridge blogger in another direction as well, opening Englishlanguage conversations to Kiswahili speakers. Invited to attend Pop!Tech, a technology and futurism conference held in rural Maine in October 2005, Macha responded by blogging over two dozen presentations in Kiswahili, translating complex political and technical terminology in the process95. Commenting on a blog post I made featuring Macha's efforts, Kiswahili blogger Jeff Msangi observed,“Ndesanjo is gold among us swahili speakers when the issues of keeping us update and informed.I also applaude him for wonderfully bringing to us the Tech!Pop conference live!Ndesanjo,if you are reading this here,keep it up.You are amazing and more than inspiring to us swahili speakers.”96 Jordan and the Arabic Blogosphere While Iraqi blogs like Baghdad Burning (riverbendblog.blogspot.com) and Iraq the Model (iraqthemodel.blogspot.com) are well known in the blogosphere and mainstream media, less attention is generally paid to thriving blog communities in Jordan, Bahrain,Tunisia and other Middle Eastern nations. Like the community in Kenya, the Jordanian, Bahraini and Tunisian communities maintain national blog aggregators and create a sense of blogosphere identity through cross linking.There’s a strong emphasis in these communities on face to face interaction – in all three countries, there are monthly gatherings of webloggers to discuss collaboration and common issues.There’s also a strong interest in creating an Arab blogosphere, linking across national borders and, especially, in celebrating bloggers from Palestine. Despite a strong regional identity, many bloggers from the Middle East write in English. The vast majority of the three dozen bloggers aggregated on Jordan Planet (jordanplanet.net) blog in English. Prominent Tunisian blogger Subzero Blue (subzeroblue.com) blogs in English and Arabic, but his English blog is updated more frequently. 29 of 32 blogs listed on BahrainBlogs.com are titled in English. Ahmeid Humeid, Jordanian designer and blogger (360east.com), speculates that one reason the Jordanian blogosphere writes in English is the desire to reach a worldwide audience: “As in all small countries, you feel the need to communicate with the outside world.The same thing as in Holland and all the Scandinavian countries. It’s not like in Egypt where you have 60-70 million people… We feel that to address a wider audience, to address the outside world, we have to talk in English.” Despite an audience of tens of millions of Arabic speakers, many prominent Egyptian 94

Macha,“Kiswahili Blogosphere”, 6/5/2005. http://cyber.law.harvard.edu/globalvoices/2005/06/05/kiswahiliblogosphere/, accessed 9/15/2005. 95 See Macha,“ Pop!Tech: Negroponte na kompyuta za dola 100”,10/21/05, http://jikomboe.blogspot.com/2005/10/poptech-negroponte-na-kompyuta-za-dola.html, a post about Nicholas Negroponte's $100 laptop project for an example of Macha's translation and blogging. 96 http://www.ethanzuckerman.com/blog/?p=231#comment-4162, accessed 12/20/05.

bloggers – like Big Pharaoh (bigpharaoh.blogspot.com) and 1 Pissed Arab (1pissedarab.blogspot.com) - choose to write in English as well.A counter on Big Pharaoh’s website announces that he receives the most traffic from the United States, followed by Canada, then Egypt97. Others, like Kefaya activist Alaa Al Fateh, tend to write in Arabic when organizing local events, and in English when reporting on Kefayah actions98. An awareness that Arab blogs have an international audience is apparent in the statements and actions of bloggers throughout the region. Projects like Haitham Sabbah’s NoToTerrorism.com, a collection of images from the Middle East with an anti-terrorism message, attempt to challenge perceptions of Arabs and Muslims as terrorists.As Bahraini Mahmoud Al-Yousif of Mahmoud’s Den (mahmoud.tv) states on his “about” page: “I try to dispel the image that Muslims and Arabs suffer from - mostly by our own doing I have to say - in the rest of the world. I am no missionary and don't want to be. I run several internet websites that are geared to do just that, create a better understanding that we're not all nuts hell-bent on world destruction. I hope that I will be judged that I made a small difference.99” Despite the international orientation of some Middle Eastern blogs, a number of bloggers in the region have attracted unwanted attention from legal authorities.Tunisian journalist Zouhair Yahyaoui served over 14 months in prison in conjunction with his satirical website TUNeZINE100. Bahraini bloggers have reacted with hostility towards a plan by the Ministry of Information to register web sites, allegedly to protect intellectual property created by authors101. Despite concerns about government censorship or monitoring,Al-Yousif’s Mahmoud’s Den features on its front page the messages:“394 days left to BOOT this parliament OUT!” and “I don’t need the MoI’s protection, thanks very much!”102. Given the attention paid to issues in the Middle East by global media – newspapers and bloggers alike – it's unsurprising that there is an audience in the US blogosphere for English-language commentary, especially from Iraq. Popular Iraqi blogs like "Baghdad Burning" and "Iraq the Model" are well-known within the United States, where left-wing political bloggers link to Baghdad Burning and right-wing bloggers to Iraq the Model, but it is unclear to what extent their readership extends beyond the US. By contrast, the Jordanian,Tunisian and Bahraini blogospheres cross-link extensively, maintain national blog aggregators and have frequent face to face meetings between bloggers.The sophistication of these blogospheres becomes apparent only when an event like the 97

http://bigpharaoh.blogspot.com/, accessed 9/15/2005. See Alaa’s 9/13/2005 blogpost, http://www.manalaa.net/yet_another_security_less_march.“Kefayah” translates roughly as “enough!” in Egyptian arabic and is a student-led movement for increased democratization of Egypt. 99 http://mahmood.tv/index.php/docs/82, 6/1/2003.Accessed 9/15/2005. 100 From RSF,“Internet Under Surveillance, 2004”, http://www.rsf.org/article.php3?id_article=10768, accessed 9/15/2005. 101 Al-Yousif,“State Censorship in Disguise”, http://mahmood.tv/index.php/blog/1610, accessed 9/15/2005. 102 http://mahmood.tv/index.php/blog/1, accessed 9/15/2005. 98

Amman bombings calls attention to local voices. Jordanian bloggers set up a “virtual newsroom” on Jordan Planet (jordanplanet.net) and Global Voices (globalvoicesonline.org), covering the aftermath of the bombings and the resulting antiZarqawi marches in Amman, using first person accounts, digital photographs and screenshots of television broadcasts from around the Arab world – the effort involved Jordanian bloggers in the US and Bahrain as well as in Jordan103. Reactions to an earlier bombing – the July 7, 2005 London Bombings – set the stage for a shouting match between American Christians and moderate Muslims from the Middle East.A roundup of Muslim reactions to the London bombings (universally condemning the attacks) on Global Voices was linked to by prominent US conservative blogger Glenn Reynolds (Instapundit.com) at 10:21am on 7/8/2005104. 24 minutes later, the first of several dozen comments critical of Islam and the Arab response to the bombing were posted to the thread, beginning with a comment by “Bongo”: “Middle East bloggers are hardly representative of the majority…as we saw with their embarrasing failure to correctly predict the outcome of the Iranian elections. Check out the comments in English on some of the general Arabic sites.They think it’s absolutely knee-slapping wonderful.105” Subsequent comments became even more hostile, with Bradley M. Cooke writing,“There is a sickness in the muslim culture that worships death, hates jews, fears modernism and pluralism.They cut off heads, they dance in the streets at the massacre of thousands on 9/11, they blame their societal woes on bogus Zionist conspiracies, and enshroud their women in garments that are literally and figuratively oppressive.There is nothing good about muslim culture today. Nothing.106” By the next morning, Muslim bloggers from the Middle East were responding in the thread. Mohamed of “From Cairo,With Love” (fromcairo.com) began his comment,“To all the commenters here. Just in case you are wondering, no, we are not going to renounce our religion –Islam– just because some of the ignorant comments made here about the faith, or because of some criminal terrorists who think they are Muslims who went off to kill innocent people in a heinous manner.107” Subsequent conversations on this thread and on other threads posted by Muslim authors on Global Voices turned into angry exchanges on the nature of the Koran, whether Islam was inherently violent, and whether US imperialism or European colonialism was ultimately responsible for violence in the Middle East. It would be unrealistic to suggest that more light than heat was generated during these exchanges, but the online interaction did allow two groups with radically different opinions to confront each other's preconceptions directly, as in a debate on violence in the Koran 103

See “Explosions Rock Jordanian Capital,Amman”, 9/9/2005, http://cyber.law.harvard.edu/globalvoices/2005/11/09/explosions-rock-jordanian-capital-amman/ and archive pages from Jordan Planet on 11/11 and 11/12: http://www.jordanplanet.net/archive/2005/11/11, http://www.jordanplanet.net/archive/2005/11/12, all pages accessed 12/20/05. 104 http://instapundit.com/archives/024124.php, accessed 12/20/05.While Reynolds was the first prominent US blogger to link to this thread, Jeff Jarvis linked later in the day at http://www.buzzmachine.com/index.php/2005/07/08/themorning-after/. 105 http://cyber.law.harvard.edu/globalvoices/?p=310#comment-1397, accessed 12/20/05. 106 http://cyber.law.harvard.edu/globalvoices/?p=310#comment-1418, accessed 12/20/05. 107 http://cyber.law.harvard.edu/globalvoices/?p=310#comment-1481, accessed 12/20/05.

between Rufus Lee King and Zaid Hassan108. In this sense, the interaction met some of the stated goals of some Middle Eastern bloggers, allowing them direct dialog with people who clearly had a negative and hostile view of Arabs and Muslims. Understanding community blogospheres The diversity of local blogospheres makes it unwise to make generalizations across different communities, even the four communities featured here.There are, however, a number of themes, some contradictory, that arise from an exploration of these communities. Blogs can be a printing press. In environments where journalism is controlled, blogs become an alternative platform for journalistic endeavor. Chinese bloggers, chased off bulletin board systems, appear to be turning to blogs to report and speculate on political matters. Iranian journalists, facing the closure of independent newspapers, turned to blogs as an alternative venue for criticism of government and religious authorities. Even in nations where free speech is less obviously threatened, blogs may become an alternative news venue. In South Korea, where newspapers are historically ideologically conservative, reformists rallied around OhMyNews, an online “citizen journalism” website that supported Roh Moo Hyun to election, and through his subsequent impeachment109.There's no guarantee, though, that a local blogosphere will become journalistic or political. Despite widespread net penetration and a generally tame domestic press, political blogging has yet to become a major force in Japan.And in the US, despite numerous other outlets for political opinion, one aspect of the blogosphere has become highly political and polarized. Local blogospheres can be a closed community.The emergence of large blog communities in Iran, China or Brazil doesn’t guarantee that these communities will widely interconnect with the Anglophone blog community. In many cases, discussions concern local topics, are held in local languages and are kept closed to outside eyes. While this behavior is observed in large blogospheres, it can occur in small language communities as well.A perusal of blogs from Iceland110 (generated from Blogger.com through a clever technique documented by Jordanian blogger Haitham Sabbah111) reveals that more than half of active blogs are in Icelandic, a language spoken by fewer than 300,000 people worldwide. Given that many Icelanders are also fluent in Danish or English, the choice of language appears to be a conscious decision to keep discussions closed to the linguistic community.

108

King's background is unknown, but his comments suggest he is a US-based conservative. Zaid Hassan is a Pakistani Muslim, living in the US and the UK.The textual debate begins with a series of Koranic “quotations” from King (http://cyber.law.harvard.edu/globalvoices/2005/07/10/is-it-a-muslim-problem/#comment-1661), most of which Hassan points out aren't actually from the Koran. (http://cyber.law.harvard.edu/globalvoices/2005/07/10/is-it-a-muslimproblem/#comment-1682). 109 Committee to Project Journalists,“Attacks on the Press, 2004”, http://www.cpj.org/attacks04/asia04/south_korea.html accessed 9/15/2005. 110 http://www.blogger.com/profile-find.g?t=l&loc0=IS, accessed 9/15/2005. 111 Sabbah,“Searching Blogger [dot] com by Country”, 8/19/2005, http://sabbah.biz/mt/archives/2005/08/19/searchingblogger-dot-com-by-country/, accessed 9/15/2005.

Local blogospheres can have international aspirations. Bloggers in Jordan and Bahrain are self-consciously blogging for an international audience. Bloggers in Tanzania are beginning to change their behavior to reach a wider audience. Kenyan bloggers may not acknowledge targeting an international audience, but their choice of language suggests a desire to reach across national borders. Some bloggers are becoming aware that blogs influence on search engines means that the domestic blogosphere has some power over a nation’s image online112. Blogs can be spaces for conversations that can’t take place anywhere else. Hossein Derakshan suggests that blogs became popular so rapidly in Iran because they allowed women to speak anonymously, and because they allowed an older generation to understand what the younger generation was thinking and talking about113.There are social restrictions on conversations between genders and age groups in Iran, and blogs create a new space where these conversations can take place. It’s unlikely that the conversations referenced above about Live8 involving Africans and Europeans, or about the London bombings would – or could – have taken place in the physical world. By creating an online space where speakers are unconstrained by their ability to gather physically (though still constrained by language and their ability to access the Internet), blogs allowed entirely new discussions to take place. They’re Speaking. Who’s Listening? Global Voices, a website run by Rebecca MacKinnon and this author at Harvard Law School to aggregate bridgeblogs, has as its motto:“The world is talking.Are you listening?” While this question may not be relevant for all international bloggers, it’s profoundly important for bloggers who explicitly seek to reach an international audience and challenge dominant perceptions of their nation or culture. In 1965, Johan Galtung and Mari Ruge postulated the existence of twelve newsgathering factors that govern coverage of international news in newspapers in their seminal paper, “The Structure of Foreign News”.These factors included: “...F9:The more the event concerns elite nations, the more probable that it will become a news item. F10:The more the event concerns elite people, the more probable that it will become a news item.

112

Iranian bloggers success in “googlebombing” the search terms “Arabian Gulf” and linking searches to a page that declares “The Gulf that you are looking for does not Exist.Try Persian Gulf” is a particularly funny example of blogger’s power over national image. See http://arabian-gulf.info/, http://www.google.com/search?q=arabian+gulf, and the Wikipedia entry on Googlebombing (http://en.wikipedia.org/wiki/Googlebomb) for more information. 113 Personal notes on a talk given by Hossein Derakshan, 12/10/2004. Confirmed by Rebecca MacKinnon’s blogpost about the talk: http://rconversation.blogs.com/rconversation/2004/12/is_conference_h.html, accessed 9/15/2005.

F11:The more the event can be seen in personal terms, as due to the actions of specific individuals, the more probable that it will become a news item…”114 These factors suggest that developing nations will be disproportionately undercovered by mainstream news sources, crowded out by stories about wealthier nations, wealthier groups of people and celebrities. Seeking to address these factors, UNESCO proposed a set of media reforms, outlined in a book-length proposal, Many Voices, One World115, which outlined a vision of a New World Information and Communication Order (NWICO).The vision experienced stiff resistance from delegations from the US, the UK and Singapore and went largely unimplemented. It comes as little surprise that current quantitative studies of coverage of the developing world in mainstream media reveal many of the problems Galtung and Ruge anticipated. My research on media attention116, analyzing stories that mention the names of nations, suggests that coverage of international news in newspapers and on television, is strongly correlated to national wealth, as measured by gross national income (GNI), and much less strongly correlated to factors like population.

Figure 1:Visualization of news stories available through Google News, May 5, 2005. Nations depicted in red had the most stories available; nations depicted in dark blue had the fewest117.

At first glance, patterns of media attention in weblogs appear to closely echo patterns seen in mainstream media.

114

Galtung, Ruge,“The Structure of Foreign News”, Journal of Peace Studies, 1965. MacBride et al., Many Voices, One World, Paris:UNESCO, 1980. 116 Zuckerman, "Global Attention Profiles - A Working Paper " (August 2003). Berkman Center Research Publication No. 2003-06. http://ssrn.com/abstract=487943 117 Current maps of media attention and detailed information on methodology are available at http://h2odev.law.harvard.edu/ezuckerman. 115

Figure 2:Visualization of blog posts stories indexed by Blogpulse over the 90 days preceding May 5, 2005. Coloring as above.

Blogposts appear to concentrate on countries well covered in mainstream media – China, Iraq, the USA, France – and ignore the same regions neglected by the media – Central Asia, Sub-Saharan Africa and Central America.

Figure 3: Comparison of Blogpulse 14 day sets with Google News 14-30 day sets. Nations in blue had proportionally more hits on Google. Nations in red had proportionally more hits on Blogpulse. Generated from data retrieved May 5, 2005.

Bloggers, as a whole, appear to ignore developing nations more to a greater degree than mainstream news coverage, suggesting that bloggers in developing nations may have an uphill battle in a search for recognition by their fellows in the North.A comparison of Google News data and data from Daypop paints a different picture.

Figure 4: Comparison of Daypop 14 day sets with Google News 14-30 day sets. Nations in blue had proportionally more hits on Google. Nations in red had proportionally more hits on Daypop. Generated from data retrieved May 5, 2005.

Daypop tracks approximately 40,000 top-ranked blogs, in sharp contrast to Blogpulse, which tracked approximately 10.5 million when this data was collected, and attempted to collect data from all non-spam blogs. In other words, Daypop tracks bloggers, but not diarists, while Blogpulse tracks both. It’s possible that the disparity between maps reflects a difference in attention patterns between top bloggers and the blogosphere as a whole – while bloggers as a whole might pay less attention to developing nations than mainstream newspapers, top bloggers might pay less attention to the USA and Israel, and more attention to a diverse set of nations. It is also possible that this intriguing result is an artifact generated by different methods used by Daypop and Blogpulse. Higher confidence results could be generated by comparing the results of the entire Blogpulse catalog with results from a catalog selected only from the ten thousand blogs Blogpulse sees as most popular in terms of incoming links. There is also evidence that bloggers will cover stories pertinent to the developing world if they are adequately primed by mainstream media coverage of the event. Research commissioned by AlertNet, a nonprofit arm of Reuters, indicated that coverage of the Indian Ocean Tsunami dwarfed coverage of ten other pressing humanitarian stories118 the ten “forgotten” stories put together received only 78% of the coverage the tsunami received.A rough parallel study, using data from Blogpulse and keywords designed to approximate the AlertNet searches119, revealed that coverage of the tsunami in the blogosphere dwarfed coverage of the ten “forgotten” stories to an even greater degree than in mainstream media – bloggers wrote 39% as many stories about the forgotten 118

Jones,“Tsunami coverage dwarfs ‘forgotten’ crises – research”, 3/10/2005. http://www.alertnet.org/thefacts/reliefresources/111044767025.htm, accessed 9/15/2005. 119 I have not been able to obtain methodology information from Reuters, which has made it impossible to replicate their results exactly.

stories as about the tsunami120. It’s possible that the larger gap between tsunami and “forgotten” stories in the blog data is explainable less by lack of interest by bloggers in humanitarian crises and more by their extraordinary interest in the Indian Ocean Tsunami. Searches on the New York Times for “tsunami” and “Iraq” revealed 11 times as many Iraq stories as tsunami stories over a yearlong interval. Similar searches on Blogpulse revealed a 2:1 ratio between Iraq and tsunami stories. In other words, bloggers took a disproportionate interest in the tsunami, covering it with an intensity that rivaled their interest in the ongoing US invasion of Iraq. It will be interesting to see whether blogosphere coverage of Katrina is similarly passionate. Results presented thus far seem to suggest that bloggers seeking attention for their national issues need to increase their national wealth or suffer a massive national disaster. Close analysis of what headlines bloggers choose to amplify suggests a less radical path for gaining attention: blog about topics bloggers follow. By collecting the headlines and URLs of stories published by media outlets from RSS feeds and checking in subsequent days to see whether links to these stories appear on Blogpulse, it is possible to determine what types of stories are widely amplified by weblogs and which are largely ignored121. Technology Science/Nature Front Page Middle East Health Americas Asia/Pacific Entertainment Europe UK News Business South Asia Africa Totals Means

Stories Links Unlinked Mean links % unlinked 27 165 1 6.11 3.70% 26 255 2 9.81 7.69% 359 2466 57 6.87 15.88% 60 195 16 3.25 26.67% 41 162 11 3.95 26.83% 95 288 36 3.03 37.89% 82 146 34 1.78 41.46% 74 161 32 2.18 43.24% 135 259 59 1.92 43.70% 315 377 143 1.2 45.40% 33 68 16 2.06 48.48% 119 158 63 1.33 52.94% 111 171 63 1.54 56.76% 1477 4871 533 3.3 36.09%

Table 6. Synthesis of data from July 1-7, 2005 and December 4-10, 2005. Link analysis of 4-day old BBC stories, checked in July on Technorati, in December on Blogpulse.

120

Detailed methodology and results are available on my weblog, in a post titled “’Simple statistics’ and the blogging of humanitarian disasters”, 5/5/2005. http://www.ethanzuckerman.com/blog/?p=60 121 My research on amplification of headlines was presented at the 14th WWW conference in Chiba, Japan in May 2005. Data collection is ongoing and final publication of results are pending. Information about the headliner project is available at http://h2odev.law.harvad.edu/ezuckerman/headliners. Data in the table below was calculated from July 1-7, 2005, using BBC RSS feeds and link data from Technorati, and from December 4-10, 2005 using BBC RSS feeds and link data from Blogpulse. Synthesis of the data was a simple summing of sets – no attempt to normalize the two sets was made.

Technology stories published by the BBC had a 96% chance of being linked to by one or more blogs within 4 days of publication – the average technology story was linked to by 6.1 blogposts within 4 days. Heath and Science/Nature stories are also well blogged. Front page stories, unsurprisingly, are popular – they’re both the stories editors thought were most important, and the stories that are most promoted. Of stories germane to a particular region, stories about the Middle East are the best blogged, and stories of South Asia and Africa the worst. Stories about the UK also fared poorly.The latter result is consistent across other news sources- local news and sports stories are the least blogged stories from the New York Times, for instance.A possible explanation: both the BBC and the New York Times have international readerships – local news is likely to be interesting primarily to people living in the coverage area, while other stories have a global audience. Looking closely at links to specific headlines, another pattern becomes apparent – stories that involve a US troop presence, or focus on terrorism are well blogged regardless of what region they are classified into. Bloggers interest in science and technology suggests that bridgebloggers who focus on these topics – like Joi Ito – may have increased success drawing audiences to their blogs. Other bridgebloggers are likely to generate an audience by virtue of blogging from a location of high interest to blog readers, like Baghdad Burning or Iraq the Model. Further directions and open questions Quantitative analysis of weblogs is a relatively new field and research techniques are still being refined.Analysis of international weblogs is in an even earlier state, complicated by language issues and the recent emergence of international weblog communities.Any research on the dynamics of international blogs is complicated by lack of believable data on the size of blogospheres in different languages and countries.A comprehensive, regularly updated study of the size and demographics of the international blogosphere would be a critical baseline for future studies in this area. Such a study would likely require cooperation from blogging service providers, blogging analysis companies and multilingual expertise, especially in Asian languages. As researchers gain expertise in partitioning the set of “bloggers” into subsets by language, nationality, gender and type (diarists/bloggers/journalists), it will become possible to answer questions that compare blogger behavior. Do bloggers in Japan and the US link the same way, or is the US blogosphere link-heavy? Do male and female bloggers, as danah boyd suggests, use links differently122? Do bloggers write more about the developing world than diarists? What blogospheres make efforts to connect to other cultures and which remain isolated? With data on the national origin of blogs (or at least language data), a study could be designed to help determine whether bridgeblogging is a mainstream or fringe activity. Of all blogs, what percent have ever linked to a blog in another nation? Another language? 122

boyd,“the biases of links”, 8/7/2005. http://www.zephoria.org/thoughts/archives/2005/08/07/the_biases_of_links.html, accessed 9/15/2005.

What languages and nations are more likely to link across languages and borders, and which are more isolating and isolated? Is bridging an activity we expect to see at all levels of the blogosphere or only at the elite, high-traffic blog level? There is a sense held by many bloggers – especially by English-speaking cyberenthusiast bloggers - that the world of weblogs is an extended, interconnected community. Data presented here and elsewhere on the rapid growth of non-English blogospheres suggests this phase of the global weblog movement may be coming to a close.As these nonEnglish blogospheres grow, it seems likely that they will be linguistically or culturally insulated from the existing blogosphere in the absence of bridging efforts.This, in turn, suggests that these bridging efforts, nascent as they are, are essential if the blogosphere remains an interconnected community. A bridged blogosphere or multiple separate blogospheres? Both appear to be possible futures as weblogs are adopted by an growing international population of Internet users.