Selected Papers of AoIR 2016: The 17th Annual Conference of the Association of Internet Researchers Berlin, Germany / 5-8 October 2016 Suggested  Citation  (APA):  Rossi,  L.,  Neumayer,  C.,  Vulpius,  J.  (2016,  October  5-­8).  Images  Of  Protest  In   Contested  Social  Media:  Production,  Propagation  And  Narratives.  Paper  presented  at  AoIR  2016:  The   17th  Annual  Conference  of  the  Association  of  Internet  Researchers.  Berlin,  Germany:  AoIR.  Retrieved   from  http://spir.aoir.org.   IMAGES  OF  PROTEST  IN  CONTESTED  SOCIAL  MEDIA:   PRODUCTION,  PROPAGATION  AND  NARRATIVES   Luca  Rossi   IT  University  of  Copenhagen     Christina  Neumayer   IT  University  of  Copenhagen     Julie  Vulpius   IT  University  of  Copenhagen     Images  play  a  central  role  in  activists’  social  media  communication  from  protest  events   (Andén-­Papadopoulos  2014;;  Mortensen  2013;;  Poell  and  van  Dijck  2015).  The   documentation  of  protest  events  through  images  and  their  propagation  on  social  media   has  been  conceptualized  as  eye-­witnessing,  counter-­surveillance  practices,  or   producing  a  visual  alternative  to  the  mainstream  framing  of  protest  events.  Activists’   production  and  propagation  of  images  through  social  media,  nevertheless,  also  raises   critical  questions  about  the  inherent  logics  of  corporate  social  media  shaping  activist   communication  and  privileging  violent  narratives  to  produce  visibility  over  other   collective  action  frames  (Dijck  2013;;  Fuchs  2012;;  Milan  2015;;  Poell  and  van  Dijck   2015).  Despite  the  widely  recognized  centrality  of  images  in  activists’  communication   empirical  studies  based  on  actual  social  media  data  are  rare.  This  article  addresses  this   challenge  by  analyzing  images  (photos  and  videos)  produced  and  propagated  in  the   Blockupy  Frankfurt  protests  against  the  opening  of  the  European  Central  Bank  (ECB)   headquarters  in  Frankfurt  am  Main,  Germany,  on  March  18,  2015.     This  research  is  the  second  phase  of  an  analysis  of  Twitter  data  collected  during  the   Blockupy  Frankfurt  action  using  event-­specific  hashtags  (#Blockupy,  #Destroika,   #NoTroika)  (Authors,  2016).  Through  a  social  network  analysis  of  Twitter   communication,  the  first  phase  focused  on  the  main  actors  involved  in  the   communication  processes  and  the  identification  of  groups  of  users  behind  the   production  and  dissemination  of  Twitter  messages.  This  research  produced  two  mayor   results:  it  confirmed  the  relevance  of  visual  content  in  Twitter  propagation  dynamics   (Suh  et  al.  2010)  and  it  observed  the  central  role  of  the  official  Frankfurt  am  Main  police   account.       The  centrality  of  the  police  account  shows  that  police  and  other  law  enforcement   agencies  have  adopted  social  media  logics  to  communicate  their  perspective  on  protest   events.  Social  media  appears  to  be  contested  spaces  where  multiple  actors  (in  this   research  activists,  police,  and  mass  media)  propose  diverse  conflictual  narratives.  In   these  narratives  violence  plays  a  central  role  (Cammaerts  2012;;  Juris  2005)  as  a   radicalized  tactic  to  produce  visibility  and  to  antagonize  the  opponent,  a  central  aspect   of  newsworthiness,  as  well  as  (preventing  it)  constitutes  the  raison  d’être  for  police’s   physical  and  digital  presence.       This  article  explores  how  different  conflictual  narratives  produced  by  different  actors   emerge  during  the  Bockupy  Frankfurt  protests  in  visual  communication  on  Twitter:   Which  different  actors  produce  and  propagate  different  visual  representations  of  the   events?  What  are  the  visual  narratives  emerging  on  Twitter  during  the  #Blockupy   actions  against  the  opening  of  the  ECB?     Methodology     To  answer  these  questions  this  research  followed  a  mixed  methods  approach  based  on   both  quantitative  and  qualitative  analysis  of  the  visual  content  shared  through  Twitter   messages  during  the  event.  We  first  ranked  the  tweets  containing  images  and  manually   coded  the  most  retweeted  1%  (N=119)  of  the  data-­set.  The  manual  coding  aimed  at   identifying  the  type  of  users  (activists,  media/journalists,  politicians,  police,  others)  who   produced  the  tweet  and  the  presence  of  violence  in  the  visual  content.  We  further   differentiate  between  two  types  of  violence:  Explicit  violence,  the  direct  execution  of   physical  violence;;  and  latent  violence,  not  directly  executed  but  latent  expressions  of   authority  and/or  destruction  (Fishman  and  Marvin  2006).  The  data  have  then  been   analyzed  through  cross  tabulation  (table  1)  to  verify  the  relationships  between  the  type   of  user  and  the  type  of  violence.  In  parallel  we  performed  a  narrative  analysis  of  the   images.  Combining  these  two  approaches  we  were  able  to  observe  how  different  actors   produce  specific  types  of  visual  content  and  to  identify  the  conflicting  narratives   produced  by  police,  activists,  and  journalists/media.     Results  and  discussion     Table  1  shows  the  relations  between  the  types  of  users  and  the  types  of  violence   present  in  the  images  shared  by  those  users.  Activists  mainly  shared  non-­violent   images  (46.5%  versus  18.6%  of  images  containing  explicit  violence  and  14%  containing   latent  violence).  Media,  on  the  opposite,  largely  used  images  containing  explicit  violence   (56.8%),  but  only  sharing  a  minor  number  for  images  with  no  violence  (18.2%)  or  latent   violence  (11.4%).  Police  shared  mainly  images  containing  latent  violence  (53.8%)  and   explicit  violence  (30.8%).  These  values,  supported  by  a  strong  statistical  association   (Cramer’s  V  .309),  show  how  the  major  actors  behind  the  Twitter  communication  had   very  diverse  strategies  influencing  which  images  were  shared.  In  the  shared   communicative  space  of  Twitter  hashtags  these  diverse  strategies  produced  different   emerging  visual  narratives.         Table  1:  Crosstabulation  of  user  types  and  represented  violence  (X2  45.5,  DF  20,  p<0.001,  Cramer's  V  .309)     No  Violence   Latent   violence     Explicit   violence   Mixed   N.A.   Activist   46.5%   18.6%   14%   11.6%   9.3%   Media   18.2%   56.8%   11.4%   6.8%   6.8%   Police   15.4%   30.8%   53.8%   0%   0%   Politicians   66.7%   33.3%   0%   0%   0%   Others   8.3%   41,7%   16.7%   0%   33.3%   N.A.   0%   50%   25%   0%   25%     The  emerging  narratives  show  further  nuances  based  on  the  users  who  originally   posted  the  photos  or  videos.  The  most  visible  images  and  videos  produced  and   propagated  by  police  all  include  either  explicit  or  latent  violence  such  as  images  of  riots   initiated  by  activists  by  setting  police  cars  on  fire.  The  narratives  created  by  activists  are   more  diverse:  Colorful  and  artistic  protests,  eye-­witnessing  of  physical  and  symbolic   police  violence  as  well  as  glorification  of  violent  action  initiated  by  rioting  activists.   Despite  this  diversity  of  visual  narratives,  mass  media  and  journalists  mainly  construct  a   purely  violent  or  at  most  a  mixed  narrative  of  the  events  that  also  includes  images  of   mass  protest.       Based  on  these  results  and  a  continued  analysis  of  images  of  the  Blockupy  Frankfurt   actions,  we  can  preliminarily  conclude  that  on  social  media,  images  of  riots,  peaceful   protests,  artistic  action,  as  well  as  police  and  news  media  struggle  for  public  visibility.   Different  actors  create  parallel  narratives  representing  a  positive  image  of  themselves   by  antagonizing  the  other  through  visual  narratives  of  physical  and  latent  violence  but   also  (in  the  case  of  the  activists)  by  a  non-­violent  representation  of  their  own  actions.   Despite  these  non-­violent  narratives  we  can  nevertheless  observe  a  dominance  of   physical  and  latent  violence  in  reporting  about  protest  events  through  visual  content  on   Twitter.     References     Andén-­Papadopoulos,  K.  (2014).  Citizen  Camera-­Witnessing.  Crisis  testimony  in  the   age  of  mediated  mass  self-­communication.  New  Media  and  Society  16(5),  117-­133.     Cammaerts,  B.  (2012).  Protest  logics  and  the  mediation  opportunity  structure.  European   Journal  of  Communication,  27(2),  117–134.     Dijck,  J.  van.  (2013).  The  culture  of  connectivity:  a  critical  history  of  social  media.   Oxford  ;;  New  York:  Oxford  University  Press.     Fishman,  J.M.,  &  Marvin,  C.  (2006).  Portrayals  of  Ciolence  and  Group  Difference  in   Newspaper  Photographs:  Nationalism  and  Media.  Journal  of  Communication  53(1),  32-­ 44.     Fuchs,  C.  (2012).  Social  media,  riots,  and  revolutions.  Capital  &  Class,  36(3),  383–391.     Juris,  J.  S.  (2005).  Violence  Performed  and  Imagined:  Militant  Action,  the  Black  Bloc   and  the  Mass  Media  in  Genoa.  Critique  of  Anthropology,  25,  413–432.     Mortensen,  M.  (2013).  Conflictual  Media  Events,  Eyewitness  Images,  and  the  Boston   Marathon  Bombing.  Journalism  Practice  9(4),  536-­551.     Milan,  S.  (2015).  From  social  movements  to  cloud  protesting:  the  evolution  of  collective   identity.  Information,  Communication  &  Society,  1–14.     Poell,  T.,  &  van  Dijck,  J.  (2015).  Social  Media  and  Activist  Communication.  In  C.  Atton   (Ed.),  The  Routledge  Companion  to  Alternative  and  Community  Media  (pp.  527–237).   London:  Routledge.     Suh,  B.,  Hong,  L.,  Pirolli,  P.,  &  Chi,  E.  H.  (2010).  Want  to  be  retweeted?  large  scale   analytics  on  factors  impacting  retweet  in  twitter  network.  In  Social  computing   (socialcom),  2010  ieee  second  international  conference  on  (pp.  177-­184).  IEEE.