Wednesday, June 04, 2008

Orders of Magnitude

I have a friend, cosmologist Ned Wright, who studies the universe. He teaches a graduate course in astrophysics at UCLA called "Order of Magnitude." As I understand it, the point of the class is to have the student who is assigned a problem not reach an actual definitive numerical solution, as laypeople might expect experts to do, but merely approach the correct "order of magnitude" for the situation involved. After all, when you are studying the universe, knowing the right order of magnitude can be the critical piece of information.

This kind of thinking may be more common in the sciences than in the humanities, but today's expansive talk about "Cultural Analytics" by Lev Manovich was designed to address these kinds of questions of scale as they apply to the millions of objects of study of interest to those in the humanities. Like Franco Moretti, Alan Liu, Todd Presner, and other faculty on California campuses, Manovich has become associated with ambitious plans for data visualization that involve multiple times, places, texts, and artifacts. Although he didn't get to it in his disquisition on "crazy ideas," he has also been affiliated with the Blue Brain project in Lausanne, which has been working on elaborate brain simulations that give researchers over thirty million synapses with which to experiment.

Manovich established his ethos with the largely undergraduate audience by citing his own social networking credentials, which he said included a presence on Facebook, MySpace, and Flickr. Laptop-bearing students were soon checking out his various online personae as he talked. He frequently situated social media in relationship to discursive exchanges, such as the "conversation around" "production, reception, distribution, remixing, uploading, and downloading" and the need for an "intelligent discussion about cultural production."

Although he's a major figure in the field, in the past I've criticized Manovich for discounting the importance of the rhetorical in software studies, particularly in one regrettable passage on page seventy-seven in the otherwise excellent The Language of New Media. But Manovich has clearly come a long way on the subject since the book's initial publication in 2001, as evidenced by his inclusion in the Software Studies group of Ian Bogost, known for his theories of "procedural rhetoric," and Michael Mateas, who recently gave a talk about the relationship of "code machines and rhetorical machines."

Manovich sketched out what he had seen of a "culture mediated by software" since the end of the Cold War in 1990. Along with the rise of user-generated content, he drew attention to a
"shift in production tools" as designers and editors across the globe embraced common proprietary software packages such as Final Cut Pro, Flash, After Effects, Maya and 3D Studio MAX.

Rather than be "interested in particular masterpieces" of our era, Manovich argued that we should be interpreting the emergent software techniques "as a kind of language" like "Chinese or English in the 10th century." He described himself as intellectually engaged with such "global cultural flows and patterns" as a way to understand what affects our imagination. Scholars, he insisted, should be doing more than just "using intuitions," which -- in the case of more traditional film faculty -- might be based on nothing more than arbitrary "subscriptions to Netflicks" from which one can "try to make some theories." He reminded his audience that "none of us have the physical ability to see" more than ".0001 percent of what is produced." Thus, Manovich asked, "How can we have an intelligent conversation about contemporary cultural production?"

(For another perspective on the problem of the mathematical sublime in digital research, see this talk by Jeffrey Bardzell, which argues that there are limits to the relevance of number-crunching in interpretive fields.)

To solve the problem, Manovich said that he had coined the term "cultural analytics" three years ago. Unlike conventionally "small data sets" used in the Humanities, such as "classical Hollywood cinema" and "Italian Renaissance," he suggested that we should be looking across genres, cultures, traditions, disciplines, and types of subject matter with very large samples in mind. He argued that the Humanities should follow the lead of "linguistics, economics, life sciences, and genetics" and be working closely with "huge amounts of data."

During his talk, he launched into a series of examples, illustrations, anecdotes, hypotheses, mock-ups, and prototypes, many of which were derived from his enviable jet-setting lifestyle as an arbiter of academic taste. (Unfortunately, not all of us can get away with wearing black leather to work, although -- as this blurry image shows -- Manovich apparently does.) For example, he asserted that a country like Japan "had culture as its most important export." And he showed a slide of a map of fashion weeks that already had covered the globe in 2005 to illustrate the concept of "global culture space."

In Manovich's vision of the world, it seems that "everybody has access to the same cultural software." Moreover, the mobile students that he meets across the globe might be "working in any other city" when they grow up. Of course, Manuel Castells has argued that there are still certain kinds of more subtle "digital divides" with which to contend, particularly between "the interacting" and "the interacted upon." As someone who studies digital rhetoric, I have noticed that 3D models, animations, simulations, and games have significant rhetorical force as means of persuasion in the public sphere. There are many genres that incorporate these cultural imaginaries.

For example, the software package Maya may be used to create digital film effects, buttress a scholarly argument in archeology or evolutionary biology, provide an illustration of the testimony of an expert witness in a courtroom, reenact a violent altercation in news coverage, elaborate a proposal for urban development, churn out gimmicky graphics in advertising and corporate promotion, provide hyperreal imagery for satire or political rhetoric, and build CGI environments for spectacle, deliberation, and pedagogy. Yet I've watched relatively sophisticated designers and computer users open Maya on their desktops as novices, and for them it might not be easy as autodidacts to "drive it around" for the first time without crashing into a few things while they are at it. Thus, even less computer-literate citizens are unlikely to be able to assess how this software creates artifacts of visual culture that may be constrained (or enabled) by certain features. The way that proprietary software may manifest certain architectures of control is worth attending to, as is the fact that young designers in the United States -- even those on U.C. campuses -- often don't have access to machines with these programs because such training is relegated to vocational institutions.

Manovich opined that the "humanities is the one area that isn’t globalized," and he noted that professors of English rarely come from India and China unlike the faculty in other disciplines. Speaking metaphorically, he talked about this circulation of information as being "like a new ecology" in which "cultural ideas spread like viruses." In contrast, he noted, "When I was growing up in the Soviet Union," one needed "access to a special library" even to read the works of Freud.

Manovich observed that "social activities leave digital traces" and that while a cellular telephone provider like Verizon may "only keep this information so they can bill you," many other kinds of data about communicative actions are being tracked for many possible purposes. Because "all this information is digital," it is possible to see the "digital shadow" that constitutes having a "public persona" with a "cultural identity." As he also noted, artists can use similar software to provide an alternative to corporate-controlled tools for data visualizations. (For more on this concept from another source, see "disruptive technologist" Virgil Griffith's work on "amateur data mining.") To illustrate Manovich's point about how these digital traces can be assembled into a "trend history," he showed a sample result screen from Blogpulse live that included "diaries" and "memes" along with "sports" and "politics" as a categories with which to understand our current fluctuating economy of attention or "mindshare."

Manovich also argued that one could no longer visit a few cultural capitals to know what might be happening in painting or architecture, because of the presence of young artists and designers who are "hungry for knowledge" in places around the world, such as Singapore and Mexico City. As opposed to the argument in Thomas Friedman's popular 2000 book, The World is Flat, Manovich is making what he called an "Inverted World" argument in which speaking English, using the same software, having access to the Internet, and being able to fly on discount airlines (which he admitted may be an era that is ending) means that "the margins are taking over the center." He noted that the U.S. is now far behind many other countries in its broadband access, and that whole countries -- such as Estonia and Singapore -- have free wireless, which he attributed to the fact that there are "thirty-five and forty-year-old decision makers" rather than "elite cultural institutions run by seventy-year-olds."

So "how do you tell the stories," Manovich asked, when there are no longer "a small number of capitals and schools"? He continued this line of inquiry by asking the following questions:

Can we create quantitative measures of cultural innovations? Can we have real-time detailed interactive maps of global cultural production, consumption, remix, and collaborations?
Can we visualize flows of cultural ideas, images, and trends?
Can we visually represent how cultural and lifestyle preferences – whether for music, forms, designs, or products -- gradually change over time?

After lobbing out these questions, he asked if it were possible to use interactive graphics to "predict future development" as well. As he said, the humanities tends to be occupied with "interpreting the past," which "doesn’t make for a dynamic field." He argued that humanities scholars could be "prototyping the future" in ways that would be much more ambitious than the kind of marketing model currently in use to sell commercial products such as Lexus cars. According to Manovich, "the joint availability of large cultural data sets" that is growing exponentially and being facilitated by the digitization efforts of museums and libraries makes it possible to treat "culture as data" and study "social media that can be mined and visualized."

To introduce the work that he was doing with the HIPerWall, he emphasized the concept of "situational awareness." He pointed out the irony that there were "innovative solutions for command centers" for security and for the military that capitalized on new computer interfaces, and yet history lectures were little more than a matter of showing "two slides and then two more slides." He argued that this technology could be used for more than visualizing obvious needs like "global patterns of global warming or economics" by being applied to a kind of "real time traffic display" for cultural production.

He pointed to the work of researchers working with anime music videos, commonly knowns as AMVs, and suggested that computer vision software would be able to manage the hundreds of thousands of videos and identify ones that are "statistically improbable" through descriptions and markers. The data aggregated could include fan-creators, places of origin, video lengths, years of composition, and even original authors, although this marker of initial intellectual property ownership also raised the issue of potential copyright policing by multinational media conglomerates, which could prove to be repressive to the very cultural production being studied.

Manovich indicated that the model of history promulgated by important thinkers such as Foucault and Kuhn frequently presented cultural production over time in terms of breaks and radical reconfigurations. Inspired by the example of calculus, he also argued that it was important to be able to visualize continuous functions as well, such as how "hip hop becomes techno."

In his HIPerWall slides, Manovich was careful to resist overemphasizing either high or low culture. For example, in one visualization, the object of study seen in close-up was a twin photograph of similar postmodern dual skyscrapers in different cities; in another image he showed contrasting "vernacular architecture" from Mexico city and paid tribute to the cultural contributions of favelas and shantytowns. Unlike Jane Jacobs, who dismissed "pervert parks," such as the one in Philadelphia, Manovich indicated that he wanted to include cultural production that leaves "no glamorous traces in public squares," which may include the work of "sex services" and "servants," as well as social service agencies. Further on in his lecture, he discussed how could be mined for "books about the Iraq war" to show "different genres" and "connections between different books."

At one point, Manovich also showed a citation matrix derived from the Web of Science in which different colors represented different disciplines. As he wryly commented, "people in the humanities don’t talk to anyone." He also displayed how this distribution could be visualized in terms of different countries. From this example, he suggested a kind of "art project" that looked at "influences" rather than producing a paper with clear citations. What he suggested was what he called "relationship maps" that showed a "response to particular aesthetic or cultural teachers," which could encompass oppositional patterns, images, metaphors, camera moves, and types of characters in computer games. As Manovich implied, such influence maps might free humanities scholars from the simple-minded periodicity that goes from "early" manifestations of a given phenomena to the cultural product in "decline."

He concluded by looking at his own undergraduate teaching and the ways that students in one class can work on visualizing data sets created by students in a previous class. To build from more simple stages to an eventual “global educational platform,” Manovich used freely available tools at Many Eyes for starters. He also showed how students mapped "connections between subcultures" that visualized how "people connected through cultural brands" that could be sliced according to "casual" or "trendy" orientations.

As an academic married to a designer, I think that I know what Manovich means when he talks about the million or so people who serve as "cultural producers" on this planet, since even my teenager creates street art and mixes beats in ways that have involved mentoring from professionals at Cal Arts and the Grammy Foundation. By having some sense of the connections between members of this distributed population, however, I wonder how much this kind of first-stage mapping can reveal of the often substantive covert connections between persons and even between their ideas.

In other words, will only collaboration be visualized in Manovich's cultural analytics? What about competition, which also serves as an important creative engine, particularly in battling niche technologies? Even seemingly counterproductive dynamics like pecking orders and petty rivalries are important in artistic and activist communities and large bureaucratic organizations alike. Furthermore, should Manovich's Web of Science visualization eventually include less obviously institutional forms of relatedness and tag individuals who are spouses, ex-lovers, childhood friends, and classmates to help novices with navigating the social landscape of public intellectuals?

Finally, what about software that doesn't just represent how social actors are related but fundamentally determines the nature of their relationship? Right now we might not think much about how the algorithms of social bookmarking and social networking sites may connect cultural producers, but in the future this might actually be important. For example, what is my kinship to the conference chair and editor of scholarly publications that I happened to have dated at sixteen during the punk rock era with whom I reestablished contact after the algorithm suggested that I should be interested in his work on cyberspace? In many ways, the latter connection of the marketing logic of the online bookseller is much more important to us in our present collegial (and platonic) relationship than the former tie. And what about Alice Robison to whom I am closely related in Facebook genealogy, although we have never met in person? See my blog posting and her blog response to see the nature of our kinship and how the interface design of the social networking site brought us together.

This talk also concluded Peter Krapp's Software Culture lecture series at U.C. Irvine, which was covered here, here, here, here, and here. Regular readers in the lay population can now look forward to summer and three months without conferences or talks during which I can provide more coverage of public displays of affection, taxpayer-funded cartoon characters, awards for really bad government websites, and even information about the erstwhile subject of this blog: institutions as digital media makers, such as universities and government agencies, who must simultaneously serve as regulators and content-creators.

Labels: , , ,


Blogger Lupton said...

As someone who works on the "limited data set" called Shakespearean drama, yet who yearns to travel in the wider terrains of social media, I find Lovink's call for a greater order of magnitude in humanistic inquiry both inspiring and daunting. My generation won't be the one to do this new work, which will probably replace the word processor with the cell phone as the favored research instrument. But I hope I can follow the show!

12:08 PM  

Post a Comment

<< Home