--- schema_version: "library-reader-v1" source_type: "original_book" title: "If...Then Algorithmic Power and Politics" author: null edition: null publisher: null year: null isbn: "2017054909" library_document_id: "1d3bd28e-c1ba-4af2-8580-60b32ef26053" source_sha256: "9d6111c11832b55e134d36eb9ec970d190a989ee71c78f6d87e076c351752a78" pdf_physical_page_count: 211 pdf_physical_page_index_base: 0 range_start: 180 range_count: 10 html_url: "https://workshop.closeclaw.top/api/ai/read/1d3bd28e-c1ba-4af2-8580-60b32ef26053?start=180&limit=10" markdown_url: "https://workshop.closeclaw.top/api/ai/read/1d3bd28e-c1ba-4af2-8580-60b32ef26053.md?start=180&limit=10" previous_url: "https://workshop.closeclaw.top/api/ai/read/1d3bd28e-c1ba-4af2-8580-60b32ef26053.md?start=170&limit=10" next_url: "https://workshop.closeclaw.top/api/ai/read/1d3bd28e-c1ba-4af2-8580-60b32ef26053.md?start=190&limit=10" --- # If...Then Algorithmic Power and Politics > 页码规则：PDF_PHYSICAL_PAGE_INDEX 是从 0 开始的权威文件物理页索引；PRINTED_PAGE 是页面中识别出的印刷页码，可能为空，二者不得混用。 ## PDF_PHYSICAL_PAGE_INDEX=180 | PRINTED_PAGE=170 10. Film, art and other cultural mediations also lend themselves to the analysis of unknown and hidden aspects of technoculture, an approach that is often used in humanities-oriented media studies that seek to address the experiential dimensions of technologies. See, for example, Hansen (2004) on the phenomenology of new media or, more recently, Hansen (2012) on the micro perception of code and ubiquitous computing. Recently, Lisa Parks and others have also called for a critical study of media infrastructure that focuses on making the materiality of infrastructure more visible by paying atention to images, art and film (Parks, 2007; Parks & Starosielski, 2015; Sandvig, 2013). For an understanding of computers through art and literature, see, for example, Hayles (2010). 11. Depicting algorithms has always been a representational challenge whether it is flowcharts, diagrams, visualizations, introductory computer science textbooks, or commercial depictions (Sandvig, 2015). See also A. Galloway (2011) on the limits of representation in data science. 12. For more on the role of popular imagery as representational practice, see Kirby (2011). 13. Both the names and the exact wording of the tweets have been slightly changed to protect the privacy of the participants in the study reported on in this chapter. 14. Although the term “ordinary people” is both contested and highly ambiguous, the prefix “ordinary” is simply used to denote the aim of the study, which was to try and talk to people who are neither computer specialists nor social media marketers but, on the face of it, have no specialist knowledge of or any obvious vested interest in algorithms. 15. I conducted similar searches for the following platforms and services: Twiter, Instagram, YouTube, Amazon, Google, OkCupid, Tinder, and Spotify. While users tweet about algorithms in connection with all of these platforms, most of the searches have been conducted on Facebook and algorithms, Netflix and algorithms, and Twiter and algorithms because they were the platforms that seemed to solicit the most consistent streams of user responses. The main period of data collection regarding the Twiter searches took place during a ninemonth period stretching from October 2014 through June 2015. 16. All 25 participants are pseudonymized, whereas their real age, country of residence, and occupation are disclosed. Australia: Steven (24, graphic designer). Canada: Jolene (22, fashion blogger), Nora (20, student), Larry (23, works in television), Anthony (64, art professor), Richard (41, manual laborer), Alex (age unknown, occupation unknown). Norway: Sarah (33, biologist). Philippines: Louis (20s, former student, current occupation unknown). United Kingdom: Jacob (38, on leave from a university degree). United States: Amber (25, student), Kayla (23, student), Michael (21, Musician), Rachel (24, journalist), Jessa (20s, journalist), Lucas (25, quality assurance engineer), Shannon (45, career counselor), Lena (20s, graduate student), Chris (20, student), Albert (42, works in advertising), Kate (36, former school teacher), Nancy (age unknown, public policy associate), Caitlyn (30s, teacher), Robyn (age unknown, graphic designer). Unknown location, age, and occupation: Tom, John. 17. The participants were recruited from the personal Facebook network of Nora, one of the participants in the Twiter study (see note 16). It was Nora’s own interest in the subject mater and ofer to distribute a request for interview participants via her own personal Facebook account that helped me recruit new participants. Together, we designed the recruitment request for participants but phrased in the way Nora would normally speak. Nora’s status update asked people to contact her privately if they “would be willing to talk to a researcher about your experience with social media, and how it shapes the type of information that you get.” Out of the 12 people expressing an interest, 10 ended up engaging in email conversations and face-to-face interviews over Skype during May and June 2016. Once again, the names provided are pseudonyms to protect the privacy of the participants. Due to Nora’s physical locality in a university town in Canada, many of the participants recruited come from a similar geographical and socioeconomic background. They are all in their early twenties, Canadian, engaged in college education, and relatively active on social media (however, none are involved in computer science or similar subjects that would entail knowledge of algorithms). 170 NOTES TO PAGES 97–100 ## PDF_PHYSICAL_PAGE_INDEX=181 | PRINTED_PAGE=171 NOTES TO PAGES 101–114 171 ## PDF_PHYSICAL_PAGE_INDEX=182 | PRINTED_PAGE=172 55. I am well aware that talking about the illusionary and the real in the context of the imaginary may create associations with the psychoanalytical theories of Jacques Lacan. However, I am using these terms in a much more pragmatic, everyday sense. For more on Lacan’s notion of the imaginary, the symbolic and the real, see Lacan & Fink (2002). 56. Based on multiple emails exchanged on November 5, 2014. Chapter 6 1. The term “fake news” should be understood as encompassing many related phenomena, which held together refer to various forms of mis- and disinformation, ranging from deliberate misleading content, over parody, through to news that is ideologically opposed (see Tambini, 2017). 2. Although the material turn has long been underway in the social sciences and humanities, I am well aware that many other (often neighboring) fields and areas of studies have not yet fully acknowledged the power of things or adopted a material sensibility. Within the field of journalism studies, the assertion that “materiality maters” is still a relatively novel claim. Although journalism scholars have talked about the role of technology and computers in news-making for decades, “a renewed interest for materiality has emerged in the last few years” (De Maeyer, 2016: 461). For more on the material turn in journalism studies and the role of nonhumans in journalistic practice and products, see, for example, Anderson (2013), De Mayer (2016), Steensen (2016). 3. Scholars have used a variety of diferent but interrelated concepts to describe this movement toward the algorithmic processing of digital and digitized data, including computational journalism (Diakopoulos, 2015; Karlsen & Stavelin, 2014), data journalism (Fink & Anderson, 2015), data-driven journalism (Parasie, 2015), robot journalism (Carlson, 2015), journalism as programming (Parasie & Dagiral, 2013), and algorithmic journalism (Dörr, 2016). The precise label chosen to describe the intersection between journalism and computing depends, in part, on the specific area of work that is being emphasized (Karlsen & Stavelin, 2014). For example, the terms data journalism and data-driven journalism are most often used to describe an emerging form of storytelling in which journalists combine data analysis with new visualization techniques (Appelgren & Nygren, 2014). Broadly conceived, computational journalism refers to “finding, telling, and disseminating news stories with, by, or about algorithms” (Diakopoulos & Koliska, 2016: 2), and serves as an umbrella term for all kinds of “algorithmic, social scientific, and mathematical forms of newswork” (Anderson, 2013: 1005). 4. Precision journalism was conceived of as “the application of social and behavioral science research methods to the practice of journalism” (Meyer, 2002: 2). Computers and software, Meyer foresaw, would be essential for a new kind of journalism. 5. Although Jefrey Alexander writes “information will be free,” I assume he means “information wants to be free” as this is the iconic phrase atributed to Stewart Brand. Interestingly the first part of the mantra is usually overlooked: “On the one hand information wants to be expensive, because it’s so valuable. The right information in the right place just changes your life. On the other hand, information wants to be free, because the cost of geting it out is geting lower and lower all the time. So you have these two fighting against each other” (Doctorow, 2010). 6. The Scandinavian news media context is characterized by what Hallin and Mancini (2004) describe as the “democratic corporatist” model. According to this model, media are marked by a unique “coexistence” that brings together “the polarized pluralist tendency toward more partisan, opinion-oriented journalism, and the liberal tendency toward a more commercialized, news-driven journalism—often assumed to be ‘naturally’ opposed” (Benson et al., 2012: 22). For more on the Scandinavian context in general, see Syvertsen et al. (2014). 172 NOTES TO PAGES 114–124 ## PDF_PHYSICAL_PAGE_INDEX=183 | PRINTED_PAGE=173 7. The following interviews were conducted for chapter  6. Interviewees disguised by pseudonyms requested anonymity. The gender of anonymous interviewees should not be inferred from their pseudonyms. 1. Hans Martin Cramer, product developer at Aftenposten, March 2014, August 2016. 2. Stig Jakobsen, editor-in-chief at iTromsø, May 2016. 3. Jens Nicolaisen, digital director at Jyllandsposten, February 2015. 4. Ingeborg Volan, director of innovation and newsroom development at Adresseavisen, August 2016. 5. Anders Emil Møller, former director of digital development at Politiken, March 2014. 6. Johan Koppel Stræde, director of digital development at Politiken, September 2016. 7. Espen Sundve, vice president of product management at Schibsted, May 2016. 8. Markus Gustafsson, editor-in-chief at Omni, April 2014. 9. Markus Gustafsson, editor-in-chief at Omni, March 2015. 10. Henric Englund, software engineer at Omni, March 2015. 11. “Clare,” editor at Omni, March 2015. 12. “Frank,” editor at Omni, March 2015. 13. ”Emma,” editor at Omni, March 2015. 14. Catarina Smedshammar, UX designer at Omni, March 2015. 15. Alexander Nordström, technical lead at Omni, March 2015. 16. Linda Zoumas, product manager at Schibsted Media group, March 2015 17. Christian Gillinger, head of social media at Sveriges Radio, May 2016. 18. Ola Henriksson, web editor and project manager at Svenska Dagbladet, June 2016. 19. Jørgen Frøland, project manager for personalization at Polaris Media, August 2016. 20. Geir Larsen, former head of editorial development at VG, March 2014. 8. The informants were selected for their status as experts in the domain of digital and computational journalism, broadly defined. The interviews contained questions about the use of algorithms, data mining and other computational techniques within the respective news organizations, issues and challenges arising from the use of algorithms to support and shape the news, participants’ views about the current state of journalism and how they see algorithms as part of it, including the potential for automating and personalizing journalistic practices and products. The interviews averaged at 60 minutes with the shortest being 40 minutes and the longest being 110 minutes. They were taped and subsequently transcribed. The interviews were examined using open coding techniques associated with grounded theory (Corbin & Strauss, 2008). The interview transcripts were subsequently read and analyzed for the specific preoccupations and concerns that the actors were speaking to. 9. For Kristin Asdal and Ingunn Moser this is a move that implies a sensibility toward “contexting,” by which they “mean that contexts are being made together with the objects, texts, and issues at stake” (2012: 303). 10. Aftenposten is among the 29 titles owned by the Norwegian media house Schibsted. Schibsted Media Platform is Schibsted’s new efort to assemble the technology and product teams of all of its titles into a single global technology and product organization. 11. Equipped with the data infrastructure and algorithms delivered by the Norwegian big data analytics company Cxense (of which Polaris Media, the company that owns iTromsø, is one of the company shareholders), iTromsø was the first media outlet in Norway to fully automate its mobile news feed. 12. According to the Columbia Journalism Review, structured journalism is an umbrella term that refers to thinking of journalism as bits and pieces of information that can be mixed and matched in infinite ways (Gourarie, 2015). 13. As I did not explicitly ask for permission to publish their names, all names of the editors I followed and interviewed during my fieldwork at Omni are anonymized to protect their privacy. NOTES TO PAGES 125–146 173 ## PDF_PHYSICAL_PAGE_INDEX=184 | PRINTED_PAGE=174 14. “Event thinking” has emerged as an influential philosophical avenue for contesting the concepts of facts and substance—particularly, as it relates to epistemology and the philosophy of science. Theorists such as Bruno Latour, Isabelle Stengers, Alfred North Whitehead, and Gilles Deleuze have all contributed to bringing the notion of the event to the forefront of process-relational theory. Diferences among diferent theorists aside, the notion of event can be understood as “part of an anti-reductionist project that seeks to describe the relations between actual things, bodies and happenings” (Fraser, 2006: 129). Latour and Stengers both use the notion of event as a way of thinking about science—not in terms of discovery but as an invention of sorts. In Pandora’s Hope (1999: 305), Latour says he borrows the notion of event from Whitehead as a way to avoid giving all scientific credit to human discoverers, at the expense of objects that typically remain immobile. Similarly, Stengers (2000) conceptualized science as event in the sense that it invents new relations between facts and fiction. In problematizing accepted boundaries, science becomes political in Stengers’ account. An entirely diferent way of conceptualizing the event can be found in Alain Badiou’s thinking, which, in many ways, is exactly the opposite of how process-relational philosophers conceptualize the event. For Badiou (2005), events are not premised on contingency but, rather, on a radical break with the past. For more on the notion of event in recent philosophical theory, see Fraser (2006) and Mackenzie (2005). 15. By “diferent,” I do not necessarily mean “all new” or “radically disparate” from how things were before. My intention is not to talk about before and after at all, but to examine how boundaries and relations are enacted in practice. As I hope should be clear by now, I do not mean to suggest that the algorithm causes all of this to happen on its own or to anthropomorphize the algorithm. To remind readers, any distinction is analytical only, and made with the recognition that these entities necessarily entail each other in practice. 16. The notion of the pharmakon I take from Bernhard Stiegler’s (2012; 2013) work, where he uses this Greek concept to talk about various technologies as both poison and cure. Chapter 7 1. Wendy Chun (2016) argues that it is through habits that new media become embedded in our lives. It is not the push to the future characteristic of Big Data that structures lives in the age of smartphones and “new” media, but rather habitual forms of updating, streaming, sharing etc. 2. See also Gillespie (2017); Grimmelmann, J. (2008). 3. A widely used example of platform intervention is how Google responded diferently in two cases of “Google bombing,” the practice of manipulating Google’s algorithms to get a piece of content to the top of the search results for a given topic. While Google chose to intervene and to manually “clean up” its search results when a group of people started a racist campaign against Michelle Obama in 2009 (to make search results for her name yield images of a monkey), the company did not intervene when in the wake of the 2011 terrorist atacks in Norway a campaign was launched urging people to upload pictures of dog droppings, tagging them with the terrorist’s name. 4. Claims about heterogeneity and relationality do not automatically imply symmetry or equality. Despite the long-time insistence on symmetries within studies of science and technology, realities are seen as enacted and made through practices. As Pickering suggests with regards to humans and nonhumans, “semiotically, these things can be made equivalent; in practice they are not” (Pickering, 1995: 15). How agency and power manifest in particular practices is therefore an empirical question. 5. Here I am borrowing from Ganaele Langlois, who writes about meaning machines and how they “not tightly control what users say, but rather how they come to speak” (2013: 103). 6. In Karen Barad (2007) and Donna Haraway’s (2004) terms, we might also say that the “response-ability” is shared. Accordingly, reponsibility and accountability is not so much about taking charge, as it is about the ability to respond to others. 174 NOTES TO PAGES 146–159 ## PDF_PHYSICAL_PAGE_INDEX=185 | PRINTED_PAGE=175 Bibliography 175 ## PDF_PHYSICAL_PAGE_INDEX=186 | PRINTED_PAGE=176 176 Bi b liography ## PDF_PHYSICAL_PAGE_INDEX=187 | PRINTED_PAGE=177 Bi b liography 177 ## PDF_PHYSICAL_PAGE_INDEX=188 | PRINTED_PAGE=178 178 Bi b liography ## PDF_PHYSICAL_PAGE_INDEX=189 | PRINTED_PAGE=179 Bi b liography 179 下一段：[继续读取](https://workshop.closeclaw.top/api/ai/read/1d3bd28e-c1ba-4af2-8580-60b32ef26053.md?start=190&limit=10)