Featured Content   Leave a comment

Featured Content: Interview with Ethan Zuckerman, Director of the Center for Civic Media at MIT

It’s a pleasure to welcome you to our informal conversation with Internet media scholar and activist Ethan Zuckerman (http://www.media.mit.edu/people/ethanz) via Skype on August 16, 2012.  Ethan is director of the Center for Civic Media at MIT and a principal research scientist at the MIT Media Lab (http://web.mit.edu/newsoffice/2011/zuckerman-civic-media.html), located in Cambridge, Massachusetts, U.S. Along with Rebecca MacKinnon, he co-founded in 2005 the international blogging community Global Voices (http://globalvoicesonline.org/), an initiative that had caught our interest and which we wanted to hear more about, among other things.


TS: Good morning, and thank you for taking time out of your busy schedule to talk with us. Basically what we would like to do is have a rather informal conversation, to talk about the history, the idea, the concept behind Globalvoicesonline and how you thought that this might be a viable solution for trying to bring more of the news that is not really reported on a regular basis from various places of the world and other languages. It seems to have started in 2004 at a conference, at the Harvard Law School?


EZ: Right. The history of it is that Rebecca MacKinnon (http://rconversation.blogs.com/about.html) and I were both visiting scholars at the Berkman Center, at Harvard University, which at that point was the Berkman Center for Internet and Society at Harvard Law School. It’s now at Harvard in general. She was just coming off over a decade with CNN. I was wrapping up a non-profit project I’d run called Geekcorps, which was doing technology volunteering in the developing world. We were both very interested in how blogs were making it possible to follow news in parts of the world that we were interested in. She was very closely following North Korea, and was often able to get some news out of North Korea from Chinese language blogs of people who’d been traveling there on business. My interests were pretty broad. I was trying to keep up with news on the African continent, particularly from Ghana, Senegal, Rwanda, countries where we’d been doing work as Geekcorps. And so we started developing lists of blogs we were paying attention to, and we realized that much of the debate about what was going on in the blogosphere, most of the conversation, was really about the U.S. tech press, so, when people talked about how blogs were changing the future, people were really talking about how it was changing technology coverage in the U.S.


EZ: And increasingly she and I got frustrated by that, and just wanted to have a more international perspective on those discussions. So we sort of inserted ourselves into this conference that Berkman was having, and sort of said ‘look, we’re going to have a conversation about blogging in a more international sense, we’re going to invite some of the people that we admire, read and admire from all over the world, and bring them together into a room.’ And it wasn’t really meant to be starting a movement or starting a project, it was really sort of meant as a conversation. But once we had that many people in the room, people very quickly said ‘well, look, you know, we have a lot of people working on the same issues here. We need to start finding some ways to work together’. And it very quickly turned into a movement.


TS: How many languages did you start with?


EZ: Well, we really only started with English. One of the things to remember is that this is 2004. Most people blogging were blogging in English, even if they were in countries where English wasn’t the primary language spoken. I’ve written, I’ve used the example several times, and I’m going to use it in the book that I’m finishing up right now, of my friend Ahmad Humeid who’s a Jordanian blogger, who continues to blog in English, as 360east (http://www.360east.com/), and, when I met him, I met him at a gathering of Jordanian bloggers, and I said ‘look, you know, you guys all speak Arabic, better than you speak English, why aren’t you blogging in Arabic?’ And his response at that point was, ‘well, you know, no one reads it.’ There are so many more English speakers online. Now that’s changed. But at that point there were a lot of people writing about their countries in English. So, remember, this is a moment at which probably the most famous blogger in the world is Salam Pax (http://en.wikipedia.org/wiki/Salam_Pax), who’s Iraqi but writing in English as a way of letting English-speaking readers understand what is going on in Iraq. So we started in English knowing full well that we would be missing some conversations but both in terms of practicality and also in terms of the people that we wanted to reach, that was the language that made sense.


TS: Has this been the case too, with people tweeting as the revolutions began in the Arab world? Were most of those started in English?


EZ: No, look, I mean we’re now really talking about ancient history, and in Internet terms that’s a very, very long time. So, I would say that it’s a very different experience now. I think there’s much, much more local language content being produced. I think fewer people are saying ‘I insist on writing in English’. I would say particularly around the question of the Arab Spring there was a very, very nice data analysis done that showed prominent Twitter accounts based on whether they were writing in English or Arabic, or both, and unsurprisingly, there’s a community of people who are writing mostly in Arabic, there’s a community of people who are writing solely in English, and then there’s a sort of bridge community between the two. And a notable small number of bridge figures who are straddling the gap between the two. The ones who speak English are going to be the ones who get the international attention. This probably explains why the world’s so badly misunderstood the Green Revolution in Iran. It was very, very easy to overestimate the amount of support that revolution had because, you know, there were so many people writing in English, and those accounts in English were being amplified again and again and again. They were not the only accounts. These were not the only relevant accounts but those were the ones that everyone was paying attention to.


TS: It would seem that what we’ve got here, and what Ethan’s referred to, is several different discourse spheres. One group, the group that’s blogging in English, has an intent to reach a different audience than those that are often blogging in their own languages or those that are the bridge groups blogging between languages. A lot of the tweets in the Egyptian revolution were for local organizing, for instance. That audience is a local audience. And there would be no call for that to be in English. So I think what we see over the last few years are different audiences, and those audiences because of the Internet technology have become available for us to eavesdrop upon, more easily than we could in the past.


EZ: Right. I do have a very old paper, it’s actually in Public Choice, called “Meet the Bridge Bloggers” (http://ethanzuckerman.com/meetthebridgebloggers/ezuckermanbridgeblog122305.html), which looks at this with data from 2005, 2006, so there’s pretty good data about this. I’m slightly more skeptical of the idea that much of the Arabic twitting in the Arab Spring was coordination. I think this is where my friend Clay Shirky sometimes gets a bit ahead of himself. Certainly there was some coordination, but there was also a lot of discussion and propaganda and sort of back and forth as well. Because Twitter’s a public channel, it’s not a great channel in which to actually organize a revolution, and a lot of what we know about what actually happened in the Arab Spring is that people often were using these channels for disinformation as much as for information. There’s a classic story of what people refer to as the January Head Fake, which is to say that activists in Egypt used Twitter and Facebook to actually give the authorities the wrong information about when they would be showing up in Tahrir [Square]. I tend to think that actually these tools are mostly powerful because they help you manipulate media.


TS: Which I guess is a form of communication of yet a fourth kind: a different audience, another audience.


EZ: Sure. I actually think that in many cases in activist communication that’s the audience they’re looking for.  I think one of the reasons that you see so much more focus on Arabic media is that Arabic media has become more independent and more influential. Trying to influence Arabic media before the rise of Al Jazeera didn’t make any sense. Now it makes an enormous amount of sense.


TS: It’s interesting, though, that this idea of Twitter as an organizing tool became one of the media themes. Although, as you’re arguing, it’s probably not one of the major purposes.


EZ: Well, so, yes. What happens all the time whenever you have a story about rapid political change and technology is that people want to draw a line between the two. I have a piece in Foreign Policy titled “The First Twitter Revolution?.” I didn’t put the title on it; in fact, the sub-title is much closer to what the article was actually about: ‘’Not so fast, Internet can take some credit for toppling Tunisia’s government but not all of it.” (http://www.foreignpolicy.com/articles/2011/01/14/the_first_twitter_revolution) This was me writing the day after the Tunisian government fell, which we’d been following very, very closely at Global Voices. We have a lot of Tunisian activists within our community, but the case that I make there is that really what happened was yes, it had a coordinating function, but not a secret coordinating function. Social media had the function of letting people document the revolution, and then sharing that information with Al Jazeera and then Jazeera was in fact broadcasting that information and that had a sort of synchronizing effect. Look, people always want to draw the simple, straight line. It allows you to— basically it’s a form of nationalism and possibly racism. You know, Arabs couldn’t possibly rebel against their dictators; it must have been the magical technology built by white people in the United States. So, you have to be very, very sensitive to it, because it’s not an uncomplicated mistake. It’s actually a deeply agenda-ed mistake that people make.


TS: You mentioned in the beginning that most of this had started out in English. When, in what year more or less did you notice that began to turn, so that there was more communication in social media in other languages?


EZ: You know; it’s different in every country. It really was very different in every country. I think what happens is if you are the first blogger in Jordan, let’s say, writing in Arabic doesn’t make any sense, because the people who read blogs tend to be the people who write blogs, and you’re the first user of it. So all your ties are going to be to people outside of Jordan, and at that point the dominant language is English. At some point when there’s a sufficient Jordanian blogger community, then it makes sense either for you to switch, or maybe what happens is if you’re comfortable writing in Arabic rather than comfortable writing in English, then it becomes a space in which you can switch into. The tipping point problem. It’s not necessarily 50-50; it might be a small percentage, at which point it makes sense to have a flood of people speaking another language in there.


EZ: So I would say that it was different from place to place. For us, we noticed that we were not able to cover the world responsibly in one language, probably as early as 2005 or 2006. What we started doing was, we started bringing on editors. The project starts in 2004. In early 2005, Rebecca and I are editing it by hand. Going through as many blogs as we can and inviting people in. By summer of 2005, we have an intern working on this; by fall 2005 we’re hiring editors. And we hire some editors with geographies in mind, which is to say we hire a sub-Saharan African editor, but other editors with a combination of a geography and a linguistic group. So, we know that our Middle East editor needs to speak Arabic, and we end up with someone who’s able to read and point to stuff in Arabic. We know that our East Asia editor is going to speak Chinese. And then as we go further and further and further, we start splitting up, having language editors as well as regional editors. So we sort of split the two, and we basically say ‘OK, you’re responsible for Chinese language wherever it is in the world, and you’re responsible for East Asia, so you might be covering Japanese news as well.


EZ: Where things really start changing is when people start producing the website in other languages. And that starts with a project in Chinese that’s put together really by volunteers. There’s a guy, Portnoy Zheng, who takes advantage of the fact that our site is available under Creative Commons and starts creating a Chinese language version of that in Taiwan. And we think ‘this is great’, and we actually reach out to Portnoy and say ‘that’s wonderful, would you like to have some space on our server? We’ll make it possible for you to do that formally.’


TS: I’m wondering, for the translation purposes, how do you deal with – because from what I’m able to tell, the content is brought in from different bloggers, and then posted, and then made available to people who would like to translate it into other languages—how do you work the issues of quality?


EZ: Well, let’s make sure first that you understand how we got to where we are. So Portnoy starts translating this. He’s reading the English site and he is himself translating it into other languages. He starts developing a team, initially of Taiwanese, and then eventually of Taiwanese and Mainland Chinese who are working together to translate the site. We offer him space on our server. We basically say ‘look, this is great, join on in, the one rule we sort of ask for is: don’t put up any original content, just translate our content’. And the theory behind that is that since it’s on our server we have editorial responsibility for it. So if we’ve read it first, we can feel pretty comfortable that we’re not flirting with libel, for instance, based on what we’re publishing. And then for the translation, we will assume that the people who are translating and editing the translations know what they’re doing. It helped, in fact, that we started with Chinese, which Rebecca speaks fluently, and could review and say ‘yes, absolutely, this is good stuff’.


EZ: What changes is we then get many, many more people coming to us and saying ‘let’s create languages, let’s create editions that are in our own language’. And what we basically say after a while is ‘look, you need to demonstrate to us that you can pull this off. We’ll give you the server space but we’ll do it on a trial basis. You have to demonstrate that you’re able to do a certain number of translations a month, otherwise, you know, we aren’t going to give you the space forever.’ And so now we have about 30 of those communities who are doing it. They generally have two things: they have their internal organizing system, they have a mailing list or some way of saying ‘you work on this, you work on that’; and then our WordPress system is very good. We’ve designed a system that basically communicates between all the different blogs that we write, so, when you translate an English language post into Japanese, for instance, the Japanese language edition builds a link back to the English post, and the English post builds a link to the Japanese edition. It’s a bunch of tech that we created.


TS: With WordPress?


EZ: We build across; we build on top of WordPress. We’ve heavily customized it to be able to do this. We run 32 different instances of WordPress. There is an instance of WordPress that is not visible to the general public, which only deals with routing messages. It says: ‘A post of this idea on this blog has just been published. Alert the other editions that this is now available for translation’, and ‘Translation has now been published. Alert this initial post that this is there.’ So, we have 31 language editions that slave to this master translation server. It’s all been custom-coded by a couple of guys on our staff. Matt Mullenweg of WordPress has worked with us on some of the code, but for the most part it’s not incorporated within the mainstream of WordPress development.


EZ: The last thing to understand about all this, before you ask the question about what it means in terms of quality— the final frontier for us was, for many, many years—content had to appear first in English, and our logic behind this was if we didn’t do that, we couldn’t have editorial control over it. What we ended up realizing over time was that so many more people wanted to publish in their original language, and this was particularly important for our francophone Africans who in many cases could work and trade email in English but really weren’t comfortable writing in English. So we then made a decision that we would allow people to write new content, new Global Voices posts, in their own languages, published on the Net language edition and then do our best to get it translated into English from there. Nowadays this means we do have situations where someone writes in French, it gets translated into English and then it gets translated into Spanish.


TS: At this point— this is when you would have partnered up with Lingua? Is that how it worked?


EZ: Well, Lingua was our name; it really was no partnering.


TS: Typically, when, let’s say Facebook or Skype has gone to using volunteer translators, there are a couple of different avenues they take to really deal with the issue of quality control. One of them, as in the case of Facebook, is that you have a very large body of very interested user translators and the volume, the large number of translators, helps compensate for quality. Skype went a different route, and they looked for people that had experience, certain motivation, and incentives. In any case there are a number of different techniques that you can use. Do you use the volume route or do you use a sort of gatekeeping-training route, or a mixture of that?


EZ: I think we probably use a different route than either of those. And I would say that the route that we’ve ended up using has been influential as well. Just in some different ways, and not everybody knows about it. What we’ve basically said is that we consider translators to be as important as authors. And that’s a very, very different route, certainly, than the Facebook route. So if you look at our translations, you’ll see that a translation lists -in the same size and same font and with the same photo- the translator, as well as the author of the post. We consider translators as important to our community as the writers. Translators are actually more numerous in our community now than writers. They have the same profile pages. They have the same mailing list access. They have the same privileges. They get invited to our conference in the same way. And so, our feeling on this is, what we want is, a small number of extremely talented, well-respected folks who are able to do this at a very professional level.


EZ: So, the people who do this for us do it out of a combination of motivations. The main motivation is interest in the material and wanting to make sure that it’s available in their language. The second interest is sort of professional training; there’s a long track record of people working in our community and then going on to highly profitable translation jobs, and then there are probably tertiary reasons, but I think those are sort of first and second. It’s not money, because the vast majority of our translators are unpaid. What’s interesting is this model of essentially saying ‘we’re going to celebrate the translators. We’re going to make them very visible.’ It’s a model that we then introduced to our friends at TED, who were experimenting with paid translation for their TED Talks. We ended up saying to them, ‘look, just try this out, just start recruiting some folks, and see whether your quality goes up’. Quality went way up. They got much higher quality from honoring their volunteer translators, which they do very much the same way we do. And they ended up deciding not to use professional translators. What they use professionals to do, actually, is transcription, which, it turns out, is hard to get people to do. But they end up using volunteers for all the translation, and it’s worked incredibly well for them.


TS: The major difference here, then, between Global Voices and something like Skype or Facebook is that typically in Skype or Facebook you’re translating the interface. You’re translating pieces, for instance small segments, but the ability to translate content, content that the translator can feel invested in, from a political, social or cultural point of view, probably also makes a huge difference. That is to say, there’s a bigger intrinsic motivation to translate — which you can then honor.


EZ: One thing we said to TED was that we think that this model of volunteer translation only works if people can decide what they want to translate. We don’t strive to translate every story. We strive to make the stories available, and let people translate what they are compelled to translate. TED has taken that, and they basically never say to a translator: ‘OK, now here’s a talk, and you’re to translate it’. Instead, you get to look at it and say ‘well, this looks interesting, I’m going to spend a bunch of time on this.’ So, I think that’s critically important. I’m not sure that I would recommend our model for translating interfaces; for one thing, interfaces need consistency. We do not use translation memories. We’ve tried introducing them to our different communities a number of times; our folks do not like them. Almost everyone in our community translates completely freehand. They tend not to use any automated tools. They tend not to use any translation memories. We’d love it if they did. We actually think that those translation memories could be really helpful, rather useful, but that’s not what people in our community want to do. I think that would probably give you fairly poor translation for interfaces. But again, it’s sort of based around this model that basically works because people are doing what they enjoy.


TS: Although of course as you mentioned, you’re losing a lot of important data that could help your process become even better.   But you’re balancing, I guess, the technology off against the needs and desires of your translator community.


EZ: As a software developer, it pains me every time. I feel like we’re wasting effort and, and every time I say ‘you really should use a translation memory’, they respond by saying ‘hell, no, actually this is what works for me and I’m going to do this because this is fun and I enjoy it.’


TS: That may change over time as translators, as that technology trickles down to individual translators. The problem of course is if many of your translators are not professionals, they haven’t been introduced to the notion that this is a standard mechanism or standard way to translate. And by professional it doesn’t mean that they’re not good translators or cannot produce quality translations, but they don’t work in a professional environment where’s there’s a more established workflow.


EZ: So, while I hear you, I think I do disagree with you to a certain extent. I actually think that this model of non-full time, non-professional –in the sense that ‘this is my main job’—translators are on the rise. And, my sense is the real shift that’s happening is that translation is going from being an extremely expensive activity that you’re doing at rates that make it possible to make a living, to an activity that’s either being done by volunteers or is being crowdsourced in one fashion or another. I don’t think that professional translation will go away, but I do think that it will not necessarily be the dominant force that it is right now. And I think that this sort of assumption that people who are participating in something like Global Voices are going to find their way into the professional translation community is not necessarily a valid one. I think what’s actually possible is that this model that we’ve brought to the table may end up becoming the dominant model over time.


TS: No, it’s not arguing that more people would eventually come into the professional translation community. It’s more about people adopting technologies, such as translation memories, that are currently restricted to the professional translation community.


EZ: Well, although, Google now has a set of tools that can give you some of those powers, even as an amateur. That’s what I’m saying, that we’ve been surprised that our translators are not particularly excited by them. Even as our translators are getting more and more professional work, they are often saying to us ‘well, that isn’t a particularly fun way to translate. So we don’t want to do that. Maybe I will have to use that in my professional work, but I don’t necessarily want to bring it into play for the Global Voices work.’


TS: That’s very interesting.


EZ: I don’t know. I’m slightly baffled by it. I’m frustrated by it. But it’s interesting. It happens again and again. I guess the only place where I’m disagreeing with you is that I have been pushing this very hard in our community for four years and I have had absolutely no progress on that.


TS: There is the question of motivation. It might have to do with your perception of how often you will need to reuse the material, of whether you believe that this is part of a larger process, or perhaps a mix of motivations that make you decide whether to adopt technology that allows reuse.


EZ: Let’s talk about some parallel problems. Knowledge management systems  are a very popular idea, right? The whole idea is ‘wouldn’t it be great if we had all our organizational knowledge in the same place, we’d have much better learning from our past experience, and so on and so forth. People have spent amazing amounts of money building them. They don’t work because people don’t use them. And so that whole field has now sort of disappeared. Now, rather than forcing people to enter stuff into knowledge management, people try to find ways of harnessing informal communication within an office whether it’s in the form of blogs or Twitter or something along those lines. You know, my history is as a coder, and I know, generally speaking, that it would be better for me to write reusable functions and document them in such a way that I don’t have to write them again. Do I? Occasionally. It’s not the easiest thing for any of us to do. I think people use translation memories mostly when they’re working in a workplace that says ‘this is mandatory, and we’re going to work this way’. I’m not convinced that people will voluntarily work that way. Until you get to projects that are at the level of organization, of, say, a well put-together open source project where people start doing things correctly, using GitHub, and so on and so forth.


TS: What would you say, what do you think your biggest challenges are going to be, coming up, say in a 5-10 year window?


EZ: Well, I would say, one of the big things that we are wrestling with is that we are able to translate the posts that we write, but we’re not necessarily able to translate the posts that we point to. So, what happens on Global Voices is we’re basically saying ‘here’s what’s out there on the Web, it’s really exciting, and you should read it’. But in many cases what we’re really saying is ‘you should read it, oh and by the way, it’s in Portuguese, and we know that you can’t read it.’ So, trying to figure out how we get to the point where we’re able to point to and contextualize content that we’re not able to translate, is a massive one. One of the places where this probably becomes the most important is video. We’re pretty convinced that online discourse is really moving into video, that video is increasingly how people are creating content. We’re really dreadful at figuring out how to transcribe it, how to translate it, how to get it out there. We’re aware of some tools. Our leading video person works for Amara which is the leading sort of subtitling, translation video platform out there. But it still doesn’t mean like we feel like we have our act together on it. On maybe more of the theoretical side of things, I think we’re really interested in figuring out how you build relationships between human translators and the machine translation systems like Google. So, we look at something like our English-Malagasy work, and we have probably the largest English-Malagasy parallel corpus out there. But we’ve yet to be able to find a way that Google is willing to take it from us. So far, Google hasn’t really expressed an interest in it. So we’re interested in trying to figure out how to make that work better.


TS: To exploit, or in lieu of the word exploit, to leverage the corpus in some way to allow your coverage of Malagasy-English communication to grow.


EZ: Exploit is fine. I mean, I think what we would say is that we understand the need for machine translation, and we want machine translation because what we want in terms of coverage can’t happen without it and it’s very, very hard to do this at scale without it. So we’d love to see people do a better job of this.


TS: So your translation function in the future, if you could get machine translation to be more a part of your attempt to reach out to these posts that currently are untranslated, means that you might have more post-editing effort on the human side.


EZ: I don’t know that we’re ever going to use machine translation in the work that we’re posting on the site. Because the truth is, our translators, when we say ‘hey, why don’t you machine translate this and then clean it up?’ say ‘no thank you, that isn’t actually very helpful’. We want machine translation because we want all content on the Internet to be accessible to all people. And our job is working on content diversity on the Internet. Being able to point to content that is in another language and have it be at least modestly readable, would be very, very helpful for us. And right now I don’t feel like we’re even modestly readable.


TS: That’s really interesting. In fact, for some time there had been work going on in different corners of the world on Haitian Creole and suddenly when we had the disaster in Haiti, a lot of that got pulled together, through collaboration from different sources in order to get that content available, make it available in different formats, in different media, in different platforms, so that that it could be used for disaster and humanitarian relief. And you’re right to want to proceed like that, especially for these different language pairs and directions, in case something like this would happen; it would be right there available for the public to use.


EZ: Right! And I think that production of these corpora and rapidly turning them into MT systems is incredibly important. So, yes, we’re interested in that. We’re interested in making sure that that’s possible. But again, what I would basically say is, for us, the trickiest thing in all of this is that we’re only ever going to be able to translate a very small fraction of the content that’s out there. Trying to figure out how we deal with the rest of the content is a very interesting one for us, and particularly content that isn’t text but that’s video.


TS: Right, it seems that one of the issues here, then, is that the volume of the content is always advancing out ahead of what your capacity is going to be. Even with machine translation. So, you’re also going to have to have some way of deciding what to translate.


EZ: Well, that’s not a big problem for us. That is what we do. We have professional editors that are sitting there and saying ‘OK, we’re reading 500 posts today; what are the five that we say ‘hey, you really ought to go read this’ and then what’s the one that we excerpt and translate?’ What would be really nice is for the four that we point to, but don’t translate, to be able to say ‘here, you’ll be able to get at least a partial reading of this in machine translation’. That discovery and filtering function is something I feel that we already do quite well. And, I would say that if we wanted to go further, we could then, you know, translate not just one, but then two, three, four, five of the stories there.


TS: But when you say that you read 500, and then you select, well, that 500 is just a small fraction of what you could be processing. Is that of concern?


EZ: I’m not sure that that’s true for us. When we’re talking about who’s blogging about contemporary politics in Ghana, I’m not sure that the supply’s as unlimited as you think it is. Remember, our job isn’t to deal with everything; our job is to deal with interesting topics, with a potentially international audience coming from different countries.


EZ: So, our job is never going to be to translate everything that’s coming out of Madagascar, it’s going to be first of all to look at Madagascar, figure out what’s potentially interesting to an international audience, and then prioritize how to translate it and contextualize it.


TS: So, you have a smaller pool of really eligible articles, of blog posts, to begin with then.


EZ: Supply is actually the issue.


TS: Interesting. Finally, how do you see, how would you characterize the online, global dialogue in general, today, in terms of multiple languages, multilingualism, or translation on the Web? What is it that you’re finding? Is there a bit of this ghettoization that you had spoken of in other pieces that you’ve done?


EZ: Well, I’m finishing up an 80,000 word book manuscript on that topic. Look, there’s an enormous amount of dialogue, worthwhile dialogue, going on around discussion of ideological isolation. For a long time, Cass Sunstein has talked about the ‘echo chamber’. Right now you have Eli Pariser talking about the filter bubble. Basically, Sunstein identifies this as being about politics and ideology. If you’re on the left, you’re only going to listen to the left, and you can filter so that you don’t accidentally have to hear the right. He goes further and basically says that not only that, but now the technology companies are doing it whether you like it or not. They’re both right, but they’re also both a little too narrow.  You know, the truth is, we’re all terribly parochial, terribly nationalistic, terribly local in our views, unless we’re working very, very hard not to be. And, while the Internet makes it conceivable that you could get those much broader views of the world, it doesn’t guarantee that you will. And so you actually have to make a very, very conscious effort to do so.


EZ: I actually worry in many cases that the fact that you can see all these different languages, that, you know, you have that sort of experience of seeing an unfamiliar language and then hitting the back button actually gives you this sort of experience of imaginary cosmopolitanism. You sort of can believe that you’re actually getting much more diverse information than you are. I do see pretty good evidence that some of the systems that we play with, like Facebook, are probably reinforcing existing local ties at the expense of building new international ties. And if I were to try to sort of engineer around this, which is what I’m trying to do in many cases, I think getting people to look for common grounds around topics is a good way to go. I think helping people see what they’re seeing and identify the things that they may not be seeing and then trying to build serendipity systems around that are all potentially helpful. But, in the end, do I see linguistic and ideological isolation continuing? Yes, yes, I do.


TS: Ethan, thank you so much for sharing your thoughts with us. We’re sure it will stimulate some interesting discussion! By the way, when is your book due?


EZ: I’m working on it today!


Some references:








Posted October 7, 2012 by Gregory Shreve

Leave a Reply