Ethan Zuckerman’s work toward a Serendipity Engine

Ethan Zuck­er­man spoke at the New School last Wednes­day to a turn out of about 10 peo­ple (where was every­one else from DT?).  Zuck­er­man is a fel­low at the Berk­man Cen­ter for Inter­net and Soci­ety at Har­vard.  His career has been marked by pio­neer­ing projects focused on par­tic­i­pa­tion, plu­ral­ity, and atten­tion on the web.  One of his most acclaimed projects is Global Voices, a citizen’s media project that aggre­gates, trans­lates, inter­prets, and sum­ma­rizes blog­ger activ­ity glob­ally to high­light con­ver­sa­tions hap­pen­ing in grass­roots media and pro­vide win­dows into media spaces often ignored or mis­un­der­stood my main­stream news.

I’ve been putting off writ­ing about the talk because I have con­sid­er­able research to do in order to artic­u­late Zuckerman’s ideas accu­rately.  I’d rather do my best now while it’s fresh, so for­give me if there are gaps or unclear concepts.

Zuck­er­man spoke with us about his cur­rent project to develop dig­i­tal archi­tec­tures for serendip­ity.  He opened with the exam­ple of the band Jour­ney find­ing a replace­ment for lead singer Steve Perry by watch­ing youtube videos of cover bands.  They found Arnel Pineda, a Fil­ipino singer, flew him to the US, and toured the band with him.  It seems that the inter­net is built for exactly this kind of serendip­i­tous, global human inter­ac­tion.  At the same time, dig­i­tal media can actively facil­i­tate what Zuck­er­man and other iden­tify as homophily, the ten­dency of peo­ple to hang out with peo­ple like them.  The issue with homophily in media is that when the infor­ma­tion we get about the world is largely made by peo­ple like us and about peo­ple like us, we only hear the infor­ma­tion we want or already know.  To illus­trate this, Zuck­er­man out­lined to phe­nom­e­non of Nige­ri­ans attempt­ing to scam peo­ple in the devel­oped world through email.  The result is that some peo­ple are actively try­ing to block any com­mu­ni­ca­tions from Nige­ria, be that email or inter­net pub­lish­ing.  In this case, the result­ing media archi­tec­tures pro­mote xeno­pho­bia and clas­sism. (Please see Zuckerman’s detailed post on homophily as he presents the issue much more ele­gantly than I could.  His inquiries are worth read­ing at length.)

A pos­si­ble anti­dote to homophily is serendip­ity, find­ing exactly what you need with­out know­ing what you’re look­ing for.  One place where Zuck­er­man sees serendip­ity hap­pen­ing is on the front of the New York Times.  The front page presents a col­lec­tion of sto­ries cho­sen by the edi­tors that are meant to be of gen­eral inter­est and import to the Times’ read­ers.  Some of these sto­ries are impor­tant in ways that the reader can not antic­i­pate, so it’s the edi­tors job to choose and present them in a way that will pull read­ers in.  In com­par­ing the front of the print sec­tion of the Times and the dig­i­tal ver­sion, the dig­i­tal ver­sion presents con­sid­er­ably more head­lines and links.  The pos­si­bil­ity that a reader will engage with a story is depen­dent upon less infor­ma­tion about each story, mak­ing it eas­ier for a reader to ignore sto­ries that are not of the most obvi­ous inter­est to them.  Thus, more NYTimes.com read­ers will arrive more quickly at arti­cles about local elec­tions and home dec­o­rat­ing, miss­ing the news about pol­i­tics in the devel­op­ing world or global food secu­rity that might smack the New York Times reader in the face.  If you are fol­low­ing this logic, you can see how the ear­lier exam­ple of Jour­ney find­ing their new lead singer is actu­ally an exam­ple of inter­net homophily.  Jour­ney knew what they were look­ing for.  The inter­net just pro­vided more oppor­tu­ni­ties to find it.

Zuck­er­man sees libraries func­tion­ing sim­i­larly to the New York Times, where upon search­ing for a book you can’t help but see the books grouped around it, often find­ing rel­e­vant and unan­tic­i­pated infor­ma­tion.   Libraries and news­pa­pers are prece­dents of infor­ma­tion archi­tec­tures for serendip­ity that Zuck­er­man hopes to emu­late com­pu­ta­tion­ally on the web in the effort to off­set exist­ing dig­i­tal archi­tec­tures for homophily and predictability.

I asked Zuck­er­man about the roll of nar­ra­tive or story telling in serendip­ity.  He pointed to what David Wein­berger calls the Ninja Gap.  Nige­ria and Japan have roughly the same pop­u­la­tion and dra­mat­i­cally dis­pro­por­tion­ate media cov­er­age.  We as Amer­i­cans gen­er­ally have an idea of Japan, even if that idea is lim­ited to nin­jas.  This is the con­text we bring to sto­ries about Japan.  Nige­ria on the other hand has no ninja equiv­a­lent in the Amer­i­can pop­u­lar imag­i­na­tion.  No con­text, no inter­est, no atten­tion.  Story telling is impor­tant because it cre­ates the con­text that will moti­vate some­one to pay atten­tion long enough for serendip­ity to happen.

I also asked about how to account for the fact that the best serendip­ity tends to come from friends because of their inti­mate and sus­tained rela­tion­ship with us.  They can rec­og­nize things that might be import­nat to us through their intu­ition, that quiet feel­ing of infor­ma­tion becom­ing mean­ing before we’ve coded the mean­ing into language.

There is a lot of social soft­ware that is reach­ing toward serendip­ity.  Insti­tu­tions like Google are on the fore­front of with projects that antic­i­pate flu trends.  This can be seen as serendip­i­tous because it presents us with use­ful infor­ma­tion, indi­ca­tion of an oncom­ing epi­demic, that we might not have known to look for.  Zuckerman’s serendip­ity project is dri­ven not by an inter­est in social con­trol (as sug­gested by mod­ern his­to­ries of the epi­demi­ol­ogy orig­i­nat­ing in cholera out­breaks) or cre­at­ing incen­tives for pop­u­la­tions of sim­i­lar procilvi­ties to gather in one place for mar­ket­ing (as face­book does) but by an inter­est in chal­leng­ing clas­sims and xena­pho­bia on a global scale.

This is excit­ing.  I’m glad I went to the talk as Zuck­er­man is deal­ing with many issues that are near and dear to me, namely inter­cul­tural under­stand­ing and slow­ing down enough to lis­ten to each other.  Zucher­man helped com­pli­cate my assump­tions about the rela­tion­ship between the increas­ing speed of new media and atten­tion economies.  Thanks Dave Car­roll for sug­gest­ing I go to the talk. I look for­ward to see­ing how the serendip­ity project develops.