A New Way to Think About Press Freedom: Networked

A NEW WAY TO THINK ABOUT PRESS FREEDOM:

NETWORKED JOURNALISM AND A PUBLIC RIGHT TO HEAR

IN AN AGE OF ―NEWSWARE‖

A DISSERTATION

SUBMITTED TO THE DEPARTMENT OF COMMUNICATION

AND THE COMMITTEE OF GRADUATE STUDIES

OF STANFORD UNIVERSITY

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

Mike Ananny

March 2011

This work is licensed under a Creative Commons Attribution- Noncommercial 3.0 United States License. http://creativecommons.org/licenses/by-nc/3.0/us/

This dissertation is online at: http://purl.stanford.edu/qw962by3920

ii I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Theodore Glasser, Primary Adviser

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Jeremy Bailenson

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Sarah Jain

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Frederick Turner

Approved for the Stanford University Committee on Graduate Studies. Patricia J. Gumport, Vice Provost Graduate Education

This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file in University Archives.

iii (This page intentionally left blank.)

(This page intentionally left blank.)

iii

(This page intentionally left blank.)

ABSTRACT

This dissertation presents a new way to think about press freedom in the context of online, networked news production.

Essentially, if individual freedom means something other than just being left alone, and if press freedom is anything other than an anachronism from a time when only a privileged few could print or broadcast, then there is a democratic reason to defend press freedom and networked press dynamics to be discovered. To be free, people still need relationships with others and a right to hear – conditions that are, ideally, guaranteed in part by an autonomous press. The need for freedom remains, but the press is negotiating its autonomy in new ways: distancing itself from and depending upon publics, markets and states through online networked infrastructures I call ―newsware‖.

I begin by unpacking the idea of autonomy as a general philosophical and democratic concept. I trace autonomy through a number of democratic rationales and situate it within an institutional understanding of the democratic press and an affirmative interpretation of the U.S. Constitution‘s First Amendment. I then use

Bourdieu‘s Field Theory and New Institutionalism to show how the press can be understood as a field and use this model to trace how the mainstream press has historically negotiated its autonomy. Focusing on contemporary dynamics of online news production, I identify a new type of infrastructure called ―newsware‖ through which the press negotiates its autonomy today.

My empirical investigation focuses on one type of newsware: news organizations‘ application programming interfaces (APIs) that give publics access to their content. I present what I believe to be the first integrated account of three leading news organizations‘ APIs (The New York Times, The Guardian, and National

Public Radio) and identify three ways in which they use them to negotiate distance from and dependences upon software programmers and internet users.

I conclude by claiming that this new way of thinking about press autonomy—as a set of negotiated separations and dependencies among distributed actors connected through newsware infrastructures—better lets us both define and defend the press as an ideal networked institution that ensures a public right to hear.

ACKNOWLEDGEMENTS

I have many people and organizations to thank for their guidance, support, friendship and love.

My advisor, Ted Glasser, patiently and generously taught me how to read literature on the press, journalism, and democratic theory. Through his own writing and our conversations, he showed me how to craft complex and critical research projects by asking seemingly simple questions.

My dissertation committee was tremendously helpful, providing critiques that will continue to guide my work. Since coming to Stanford, Fred Turner has taken me on a rich and lively tour of ideas running through social and historical studies of networked technology, always pushing me to see people and technology in conversation with each other. Jeremy Bailenson, at a critical point in this work‘s development, asked hard questions of my empirical study, pressing me to frame my project in a more interesting, realistic and playful manner. Lochlann Jain gave feedback on both my dissertation proposal and final document, rightly urging me to think more critically about what kind of publics and counterpublics my study of the press assumes. Barbara van Schewick was a generous and supportive committee chair; but before assuming this official role, she introduced me to new ways of thinking about the internet in terms of legal and technical architectures. I‘m extraordinarily grateful for our conversations and her wise advice over the last few years.

vii

I‘m thankful for the financial and administrative support of the Stanford

Communication Department: Susie, Barbara, Mark D, Mark S and Katrin patiently and professionally guided me through the university‘s various logistical mazes, with

Susie being especially competent and generous throughout.

Relatively early in my time at Stanford I was extremely fortunate to meet

Woody Powell. Not only did he introduce me to the complex and fascinating field of institutional sociology, but he has been a consistently constructive mentor. I‘m grateful to him, Rob Reich, Deb Meyerson, Kim Meredith, and all the workshop participants for creating a wonderful home for me at Stanford‘s Center on

Philanthropy and Civil Society (PACS) –and for the PACS fellowship that supported a year of my time at Stanford. I am similarly grateful to Roy Pea in the Education

School and Claudia Engel in the Anthropology Department for financially supporting my work and introducing me to new ideas that have helped to shape my work.

The Pierre Elliott Trudeau Foundation not only provided me with financial and travel support that made my work more dynamic and interdisciplinary than it otherwise would have been – the Foundation also linked me to Canadians and

Canadian ideas that hold special places in my mind and heart. Special thanks to

Foundation President PG Forest for being a leader of people and ideas who is simultaneously insightful, principled, caring and irreverent. I look forward to continuing to learn from him. Thanks, too, to my official Trudeau Mentor, Ray

Speaker, and my unofficial mentor, Ray‘s wife Ingrid, for their generous friendship and perspective on political populism that I continue to ponder and value. The

Foundation staff—especially Josée St. Martin, Dr. Bettina Cenerelli—expertly created

viii events and relationships that helped me learn, debate, and contribute in new ways. It was a tremendous honoUr to be a Trudeau Scholar, especially alongside people I now count as good friends: Kate Hennessy, Lisa Helps, Taylor Owen, Jill Boyd, Myles

Leslie, Kate Parizeau, Leah Levac, Chris Tenove, Lindsey Richardson, Sonali

Thakkar, Lisa Freeman, Jason Luckerhoff, Alex Aylett, Simon Collard-Wexler, Josh

Lambier, Scott Naysmith, Elaine Craig, Fiona Kelly, Jay Batongbacal, Emily Paddon,

Bob Huish, Meredith Schwartz, Sam Spiegel.

I was also fortunate to have come to Stanford with a set of colleagues and mentors from my time at the MIT Media Lab and Media Lab Europe who continued to be generous and insightful friends, colleagues and mentors throughout my PhD. I‘m especially grateful to: Carol Strohecker for providing me with the best ―pre-doc‖ I could have imagined; Bakhtiar Mikhak for his gentle wisdom and razor-sharp thinking; and Lis Sylvan for her easy friendship that gave me new and healthy perspectives on life and work. Carol, Bakhtiar and Lis are all friends and mentors – and developmentalists in the very best senses of the word.

I‘m also grateful to have had a set of friends—in San Francisco, at Stanford, and beyond—who made life within and outside of the PhD very enjoyable. This list probably accidentally omits some people who are most certainly not forgotten: Abby

Reisman, Adam Smith, Ben Adida, Bob Monti, Brent Bannon, Brent Kawahara,

Carina Wendel, Dan Chak, Dan Kreiss, danah boyd, Destiny Lopez, Gian Pangaro,

Herve Gomez, Janet Go, Jeff Gubitosi, Jenny Scroggs, Jessie Dorfman, John Pien,

Josh Spanogle, Kat Murray, Kim Gomez, Leila Takayama, Lise Marken, Lori

Gauthier, MaryBeth Lorence, Matt Vespa, Michelle Hlubinka, Megan Tompkins-

Strange, Mike Hooper, Ray Bacerdo, Rhonda Fernandes, Scott Cataffa, Seeta Peña

Gangadharan, Stacy Abder, Tim Bates, Tom Coates, Tricia Bromley Martin, Vince

Fecteau, Wendy Ju.

Beyond these friends, I‘m also grateful to the GLBT National Help Center for letting me volunteer, giving me a break from academia, and reminding me every week how difficult and important it is to listen carefully to everyone‘s story. I‘m also grateful to Dr. Lisa Covey, Anju Guirani, Michelle Wallace, Hannah Sowd, and Mary

Jenn for literally helping me get on my feet again after the Great Sneeze.

I also want to thank my Ananny and O‘Connor families for all their visits, emails, phone calls and support through what often seemed like a never-ending PhD – especially Uncle Frank who first showed me what good journalism is and why it matters long before I ever thought I would study it.

I‘m also grateful to a newer but no less caring and supportive family, the

Lawsons. More times than I can remember, Ann and George generously provided me with a fantastic meal, a welcoming home, and words of encouragement – intermixed with sincere friendship, easy laughter, and logistical irreverence. Many, many thanks to them both. And my thanks to Katie and Virginia for their similarly generous long- distance support. (You might not yet know it, Lucy, but you‘ve won the family lottery.)

My thanks, too, to two other people who have provided me with long-distance love and support over the years: Dad and Irene. I‘m grateful for all you‘ve done to

x help me get here; more good years are ahead. And to Mom, whose long-distance love and support continues to be very real.

Almost finally: John! For my entire life you‘ve provided me with love, support, fun, encouragement and understanding. I could not have asked for a better brother and am grateful every day for how generously and reliably you‘ve helped me through every phase of life so far.

And, finally: Wells. Wow. This dissertation is dedicated to you because I think you often had more confidence than I did that it would ever be finished. I‘m profoundly grateful for your love, tenacity, generosity, good humour, and consistently kind spirit. I do have the words to say how much I appreciate everything you do, and everything you are. :-) I‘m very much looking forward to our post-PhD time together.

TABLE OF CONTENTS

CHAPTER ONE: Introduction ...... 1

Introduction ...... 1

Summary of Chapters ...... 8 Chapter 2: Press Autonomy in Legal Theory and Supreme Court Jurisprudence ... 8 Chapter 3: Locating the Press and Tracing Its Socio-Professional Autonomy ...... 11 Chapter 4: Negotiating Press Autonomy in Contemporary Online Journalism ..... 14 Chapter 5: Application Programming Interfaces as Sites for Press Autonomy ...... 16 Chapter 6: Conclusion ...... 27

CHAPTER TWO: Press Autonomy in Legal Theory and Supreme Court Jurisprudence ...... 29

1. Introduction ...... 29

2. The Idea of Autonomy in Democratic Systems ...... 30

3. Rationales for Protecting Free Speech ...... 41 3.A. The Argument from Truth ...... 43 3.B. The Argument from Democracy ...... 46

4. The Press Clause, Press Autonomy and the Press as Instruments of Free Speech ...... 54 4.A. The Institutional Press ...... 57 4.B. Supreme Court Press Clause Decisions ...... 61

5. Contemporary Speech Infrastructure, The First Amendment and The Press...... 72

6. Conclusion ...... 78

CHAPTER THREE: Locating the Press and Tracing Its Socio-Professional Autonomy ...... 81

1. Introduction ...... 81 xii

2. Grounding Press Autonomy: Field Theory, New Institutionalism and the U.S. Press ...... 82 2.A. Introduction ...... 82 2.B. Structure-Agency Dialectics and the Development of Field Theory ...... 85 2.C. Bourdieu’s Field Theory ...... 90 2.D. Bourdieu’s Journalistic Field ...... 98 2.E. Bourdieu’s Journalistic Field & New Institutionalist Journalism ...... 106

3. Toward a Sociological Model of Press Autonomy ...... 114

4. Using the Model: Tracing Press Autonomy through Sites of Dependence ...... 117 4.A. Introduction ...... 117 4.B. Autonomy through Professional Ideals of Objectivity ...... 119 4.C. Autonomy through Ritualized Organizational Routines ...... 126 4.D. Autonomy through Strategic Public Distancing ...... 134

5. Conclusion ...... 141

CHAPTER FOUR: Negotiating Press Autonomy in Contemporary Online Journalism ...... 144

1. Introduction ...... 144

2. Autonomy & The Contemporary, Online Press ...... 146 2.A. Press Economics ...... 147 2.B. Reader Participation ...... 155 2.C. Professional Practices ...... 170 2.D. Intellectual Property and Organizational Systems ...... 189 2.E. Distributed Human-Machine Intelligence ...... 203

3. Press Autonomy & Newsware ...... 216

4. Conclusion ...... 222

xiii

CHAPTER FIVE: Application Programming Interfaces as Sites of Press Autonomy ...... 223

1. Introduction ...... 223

2. What is an Application Programming Interface (API)? What is a News API? And Why Study News APIs to Understand Press Autonomy? ...... 224

3. How News APIs Work: A Descriptive Study of Three Leading APIs ...... 231 3.A. Experimenting with a News API ...... 236 3.B. How do news organizations describe their APIs, why do they say they offer them, and who do they think will use them? ...... 246 3.C. How long have the APIs been in existence, how do people gain access to them, and how is their usage monitored or restricted? ...... 255 3.D. How many people have used the APIs? How much data and how many stories have been served through them, and how many systems have been built with them? ...... 265 3.E. What help or support exists for those who want to learn, use, or are currently using, the news APIs? ...... 267 3.F. What data is available through the APIs? How is API content indexed and served? ...... 273 3.G. What rights do programmers have when working with news APIs, and to the systems they build with them? ...... 292 3.H. What assumptions do APIs make about the timeliness of data or news stories? ...... 295 3.I. Are API users allowed to earn revenue from systems they create? ...... 298

4. Discussion and Conclusion: News APIs as Sites of Negotiated Press Autonomy ...... 304

CHAPTER SIX: Conclusion ...... 310

DOCUMENTS CITED IN CHAPTER FIVE ...... 326

REFERENCES ...... 332

xiv

CHAPTER ONE:

Introduction

This dissertation is a study of contemporary press autonomy1. I argue that the press is a distributed entity: a set of actors, technologies, norms and ideals that exist in relation to each other, and that the scope and nature of these relationships are shaped and regulated by a historically and socially constructed field of forces. The contemporary press thus exists in no single profession, set of organizations, or material forms but, rather, among a collection of social and technological influences. These influences emerged from a history of news production that, in turn, shapes today‘s networked press and how we encounter news online – giving rise to a new set of tools and practices that both reproduce and challenge historical patterns of news production.

If the institutional nature of the press is changing, then so is its identity as a distinct, independent field with autonomy. That is, if the press is distributed then so is its autonomy. At first this seems counterintuitive: how can something that is distributed be autonomous? Answering this question involves two steps. First, we

1 Although this dissertation is meant to be generally applicable beyond any single nation, its theoretical framework relies mostly on accounts and models of the U.S. press. There are two exceptions. First, I rely upon several studies of journalists and readers that investigated how, for example, journalists at the BBC interact with user-generated content, or how readers of Western European online papers are regarded by editorial staff. This is because these studies offer valuable insights that do not seem to be irrelevant to the U.S. context and a general study of press autonomy. Second, my empirical study includes an investigation of The Guardian‘s (a UK-based news organization) application programming interface. The Guardian‘s system was included in my sample because it is widely regarded by professional journalism and software engineering communities as having one of the most sophisticated and mature systems for giving publics access to its data. More detail on this second point is offered in the Chapter Five.

need to revisit the very idea of autonomy, re-examining its conceptual and philosophical underpinnings for ways of thinking that might help us understand what autonomy means in networked societies. Second, equipped with this more nuanced and complex appreciation of autonomy, we can look for evidence of how the press— as a distributed entity—exhibits and negotiates its institutional autonomy. Who does it both rely upon and distinguish itself from? How does it do this and why? And what do such negotiations in autonomy tell us about how the press is, or could be, functioning in a networked society? The core idea in this dissertation is that the press lives not as a single configuration of these forces and influences, but as a negotiated and dynamic process in which actors configure themselves and their relationships, periodically distinguishing themselves from some, while simultaneously depending upon others. This dissertation frames and unpacks these dynamics.

Grounded in three existing literatures—legal theories of autonomy and U.S.

Supreme Court jurisprudence on the First Amendment‘s press clause, sociological and field theories of the press, and science and technology literature on the social meanings of technical infrastructures—the project problematizes the idea of press autonomy in the terms of contemporary online networked journalism. It works across these literatures to identify a class of networked information infrastructures called

―newsware‖ that mediate interactions among publics, journalists and technology designers.

Infrastructure—what Bowker & Star (1999: 47) call ―scaffolding in the conduct of modern life‖—of this kind essentially sets the conditions under which a press might be considered autonomous. These are the material and symbolic

2 conditions (i.e., the tools, forms and practices of news production and their social and cultural meanings) that different actors might use to understand the press.

Professional news organizations might want to understand how and where to direct their resources, investing only in those individuals, projects and associations in which they can see their own ideal of an autonomous press. Publics may wish to become better critics of journalism, perhaps choosing to read and trust news sources that can clearly articulate how they are autonomous and why that configuration of autonomy is most appropriate for its reporting. Advertisers might want to invest only in those news organizations, bloggers or technical infrastructures that can describe and demonstrate their autonomy. Individuals who are new to reporting or producing news—whether new hires at official news organizations or independent, self-styled bloggers trying to produce original stories and commentary that attracts an online following—may benefit from seeing what an autonomous looks like and how it behaves, so that they might see and articulate the kind of autonomy—a system of separations and dependencies—they wish to enact in their own work

And courts may need to understand more about how and when a press is autonomous if they are to understand what the First Amendment‘s press clause means today amongst the array of influences on the press. This might make possible what

Schauer (2005) calls an institutional understanding of the First Amendment. By appreciating the ―important historical, structural, economic, and cultural differences among the various channels and institutions of communication,‖ (1271) Schauer suggests that courts may avoid the risk of issuing decisions that ―institutionally compress‖ (1272) the First Amendment:

A Supreme Court unwilling to distinguish among the lone pamphleteer, the blogger, and the full-time reporter for the New York Times is far less likely to grant special privileges to pamphleteers and bloggers than it is, as it has, to grant privileges to no one … The Supreme Court in Branzburg v. Hayes [the decision denying reporters a federal shield from giving forced testimony] was perhaps right to worry about the lines between journalists and other information gatherers and information disseminators, and contemporary bloggers and others are perhaps right to be worried that such lines would be drawn to their disadvantage. Yet such a pristine approach to potentially messy lines has never characterized either the common law, constitutional adjudication, or the development of First Amendment doctrine. (Schauer, 2005: 1272-1278)

It is by understanding the distributed nature of the press, and how its autonomy can be influenced by networked information infrastructures that we might better understand the types of press autonomy that are envisioned and desired by journalists, publics and courts. Such articulations can help focus normative discussions—beyond the scope of this dissertation—of what people think press autonomy should be.

Indeed, if the press—as an autonomous, professional institutional entity with traditions, norms and products different from those of individuals with free speech rights—is to exist, we need to ask seriously whether digital technologies have made the First Amendment‘s ‗speech‘ and ‗press‘ clauses practically redundant. If the press is viewed as an institution housed within particular organizations or static definitions of journalism, then it the press clause and the idea of press freedom of the press is an anachronism. With new technologies that make it possible for anyone to publish an opinion online with the potential to reach mass audience, the press clause is simply a holdover from a time when citizens needed the help of a powerful few who owned printing presses if their perspectives were to be represented and heard by others. If, though, the press clause represents something else distinct from the speech clause, then we need to separate the material and economic practices of speech (it is virtually costless to put an opinion online) from the normative and democratic value of hearing 4

(markets for speech do not ensure that citizens get to know everything they need to know).

This tension between speaking and hearing is at the heart of the First

Amendment‘s press clause and the meaning of press autonomy today. The press clause is essentially a structural provision (Fiss, 1986, 1996) concerned with the public‘s right to listen to a diverse set of ideas – i.e., motivated not by the goal that

―everyone shall speak, but that everything worth saying shall be said‖ (Meiklejohn,

1948: 25). The more fundamental idea underlying this interpretation of the First

Amendment is that liberty has both negative and positive aspects: someone‘s right to be ―left to do or be what he is able to do or be, without interference by other persons‖ fundamentally depends upon not an idealized, isolated self but also on ―source[s] of control or interference that can determine someone to do, or be, this rather than that.‖

(Berlin, 1969: 121-122)

Lichtenberg nicely summarizes this idea as such: ―A person cannot think freely if he cannot speak; and he cannot think freely if others cannot speak, for it is in hearing the thoughts of others and being able to communicate with them that we develop our thoughts. Thus autonomy requires freedom to speak as well as freedom to hear.‖ (1987: 335-336)

Liberty depends upon not only the absence of constraints on individual natures, but on the presence of opportunities to construct an autonomous self, opportunities that require coordination, regulation and interpretation by others. Or, as Meiklejohn asserts, when the first amendment‘s speech and press clauses are taken together they

5 protect not just ―an absence of regulation [but] the presence of self-government.‖

(1961: 252)

Individuals today have tremendous opportunities to exercise their speech rights using new technologies—online forums, social networks, blogs, microblogs—but the more anyone speaks, the more necessary it becomes for everyone to rely upon a different set of technologies—search engines, filtering algorithms, reputation systems and categorization schemes—to hear. These listening technologies—with their abilities to index, interpret, privilege and amplify voices—perform one of the functions an ideal press performs, one that ultimately influences the kind of self- realization and positive liberty citizens might enjoy in today‘s large-scale democracies.

It is this fundamental sense of autonomy—a two-featured concept developed further in Chapter Two that entails both your freedom from unreasonable restrictions on your speech rights and freedom to hear what you reasonably need to know to realize your civic potential—that guides this dissertation‘s ideal of institutional press autonomy, and its search for press autonomy in newsware infrastructure.

I thus mean ―press autonomy‖ in a specific sense: as a structured system of dynamic and institutionally embedded dependencies among networked actors, norms, regulations and technologies that, together and in their ideal forms, enable both individuals’ freedoms to speak and publics’ rights to hear.

Furthermore, this autonomy is made possible through the design and use of networked information technology infrastructures called newsware. Newsware is a set of networked technologies, algorithms, interfaces, practices and norms that mediate

6 between online presses and publics. It is the infrastructure of online news production, the shared, embedded and largely invisible set of material and ideological conditions and logics governing press-public interactions online.

The third part describes each chapter of the dissertation, giving a short summary of each. Included in this section is an outline of the dissertation‘s empirical study: a comparative content-analysis of three different news organizations‘ (The New

York Times, The Guardian, and National Public Radio) application programming interfaces (APIs). These interfaces—libraries of computer code that selectively let individuals with programming skills access and repackage the content of news organizations to create news feeds—are an instance of newsware, a new site in which press autonomy is being worked out.

My aim is to provide an interdisciplinary account of press autonomy, drawing upon legal, socio-technical, and socio-professional literatures, in order to make a contribution to the scholarship on contemporary understandings of the press and journalistic practice.

These three products—an updated notion of press autonomy, the sensitized concept of newsware infrastructure and an account of how APIs encode press autonomy—are this dissertation‘s core analytical contributions. I develop a networked notion of press autonomy grounded in the idea that the press only deserves to be autonomous insofar as its actors and systems of associations guaranteed a public right to hear. While ―the press‖ may come in various forms and configurations—I make no nostalgic defense here of traditional, organizational journalism—its modern autonomy

7 can only be legitimately derived from how well, as a network, it ensures that publics hear what they need to hear in order to be citizens.

Summary of Chapters

Below I summarize each chapter in the dissertation, giving a brief overview of each chapter‘s argument and relating it to the dissertation‘s overarching study.

Chapter 2: Press Autonomy in Legal Theory and Supreme Court

Jurisprudence

This chapter addresses three questions: what does ―autonomy‖ mean most broadly and why is it defined as a concept that involves both separation from and dependence upon other actors; how has the U.S. Supreme Court historically understood press autonomy and how does a dualistic understanding of autonomy underpin an affirmative interpretation of the First Amendment; and what legal models and theories exist for understanding online press autonomy, specifically in terms of how the law understands networked infrastructures?

To frame the dualistic core of this dissertation‘s understanding of press autonomy—as institutionally situated dependencies and separations—this chapter begins with by explicating freedom as a multi-faceted concept. In extending Mill‘s liberalism—in which freedom is largely viewed as self-centered concept in which individuals are protected from unwarranted exercises of power (Held, 2006: 80)—

Berlin argued that freedom involves both freedom from control and freedom to thrive.

Someone‘s right to be ―left to do or be what he is able to do or be, without interference

8 by other persons‖ fundamentally depends upon not an idealized, isolated self but also on ―source[s] of control or interference that can determine someone to do, or be, this rather than that.‖ (Berlin, 1969: 121-122) Freedom depends upon not only the absence of constraints on my individual nature, but on the presence of opportunities to construct an autonomous self and make informed decisions (Benn, 1988; Dworkin,

1988) – opportunities that require coordination, action, involvement, and interpretations of others. Relying on others who supplemented and extended Berlin‘s concept of dualistic freedom (e.g., Fallon‘s [1994] description of the internal- and external-facing features of autonomy, and Simhony‘s [1993] and MacCallum‘s [1967] further explorations of negative and positive descriptions of freedom), this chapter describes the starting point for the dissertation‘s theoretical framework. I use the term autonomy—instead of ―freedom‖, ―independence‖ or ―liberty‖ that often involve distinction from influences—to denote a dualistic, relational concept with positive and negative, external and internal elements in which there is both separation from and dependency upon different actors and forces.

Using this sense of autonomy, I then relate it in to ideal aims of press autonomy. What can press autonomy mean, how is it rationalized and defended, and how has press autonomy been constitutionally protected in interpretations of the U.S.

Constitution‘s First Amendment?

Essentially and ideally, the First Amendment‘s press and speech clauses work together to create an environment in which individuals have freedom to speak and publics have freedom to hear (Emerson, 1966; 1981). This interpretation sees the press in structural terms—as an institutional protector of a right to listen to diverse,

9 quality speech—precisely because, following Berlin‘s formulation of autonomy, an individual depends upon not only independence from power that might prevent her from speaking, but upon the knowledge and perspectives of others that can only be encountered if people speak and are heard. As Meiklejohn asserts, when the first amendment‘s speech and press clauses are taken together they protect not just ―an absence of regulation [but] the presence of self-government.‖ (1961: 252) The press essentially is akin to schools, museums, libraries and other public-oriented institutions that, although patronized by individuals, exist to fulfill collective, public goals.

Equipped with a dualistic theory of autonomy and an affirmative interpretation of the press clause, this chapter then turns to a brief examination of Supreme Court of the U.S. (SCOTUS) rulings on the press clause. Primarily relying upon Bollinger‘s

(1991, 2010) typology of SCOTUS judgments in relation to press clause principles, and Bezanson‘s (1999) essay on the developing law of editorial judgment (in which he offers both product- and process-oriented accounts of how courts evaluate and privilege editorial decision-making), this section reviews legal doctrine on press autonomy and how the court has understood the rights and responsibilities of reporters. The aim here is to move from theories of autonomy and ideals of an affirmative press clause to an appreciation of how and why such models have been accepted, rejected or reframed by SCOTUS decisions.

Finally, to tie together general theories of autonomy with theories of the press clause and SCOTUS interpretations thereof, this chapter ends with a review of contemporary legal theory focused on how it understands infrastructural aspects of online speech. Specifically, if the press clause is to be understood in structural terms

10 as helping to create an environment for both speaking and listening, it is worth understanding how legal theory understands the digital, networked conditions of speaking and hearing. This section focuses on understanding how infrastructures facilitate the filtering and rejection of speech and, thus, a right to hear (Garry, 2004,

2005); how digital intermediaries like Internet and Application Service Providers control access to speech and apparatuses for speaking (Kreimer, 2006; Tushnet, 2008); how computer code enacts legal regimes (Kesan & Shah, 2003, 2005; Lessig, 1999); and how infrastructures automate the selection of speech from one communication environment for dissemination to, and transformation within, another (Bezanson,

2003).

Taken together, there is a growing body of legal research concerned not only with enabling the internet‘s populist or democratic potential (Balkin, 2004, 2008), but also with understanding and defending the structural conditions under which online infrastructures publish, filter and contextualize speech. Such legal theories can inform the concept of ―newsware‖ infrastructure, and a networked understanding of press autonomy grounded in a dualistic notion of freedom, an affirmative interpretation of the First Amendment, and a structural understanding of the press‘s democratic responsibilities.

Chapter 3: Locating the Press and Tracing Its Socio-Professional Autonomy

This chapter essentially addresses three questions: what does it mean to understand a domain of activity as a ―field‖; what is a ―journalistic field‖ and what special properties or features define it; and, revisiting sociological understandings of

11 professional journalism, how can we reinterpret the history of journalism as a struggle for press autonomy – as a series of dependencies and separations?

Chapter 3 traces the idea of Field Theory with a review of how Cassirer

(1923/1953, 1942/1960), Lewin (1951) and Bourdieu (1985, 1990, 1993) variously define the dynamics of a field. Each articulates how social science and human associations can best be understood as tensions between structure and agency in which dynamic actors are engaged in constant struggles for position relative to others, in conflicts over what constitutes identity and legitimacy, and in contests for power to influence other forces. These interactions are underpinned by values and normative positions about how fields ought to work, fears and expectations about the future and guilt for the past, and inertial tendencies to maintain positions and patterns over time.

These identities, interactions and values, Boudieu argues in his studies of

French literary and artistic fields, both constitute and are embedded within a ―field of struggles‖ or ―space of possibles‖ (1993: 30) in which symbolic goods are produced and traded. Such social systems of creation and exchange are mediated both by

―agents of consecration‖ (1993: 121) who confer legitimacy on producers of small- scale, restricted goods who take risks and develop new forms of art (imagine an avant- garde painter who depends upon patrons for support and critics for legitimacy), and by large-scale markets of production that signal successful production through populist expressions of approval and economic earnings (imagine fiction writers under commercial contract to regularly produce books with essentially the same literary structures, themes, plots and characters).

A group of scholars working under a ―new institutionalism‖ paradigm and including both sociologists (e.g., DiMaggio & Powell [1991], Scott [2001]) and journalism scholars (e.g., Benson & Neveu [2005], Cook [1998, 2000], Sparrow

[1999, 2006]) extend Field Theory and apply it to the study of institutions, including the press. This approach looks beyond particular organizations or individuals, an approach taken by media sociologists who generated valuable insights into journalism practices, but who tended to bracket their studies within a particular press or set of news organizations (e.g., Fishman [1980], Gans [1979], Molotch & Lester [1974],

Tuchman [1972, 1973]). Instead, it sees institutions as systems in which power, sanction, legitimation, reproduction and taken-for-granted beliefs (DiMaggio &

Powell, 1991) are distributed across a variety of actors, organizations, professions, and social contexts that include normative elements (e.g., ethics of press professionals), regulative features (e.g., legal theories and jurisprudence governing press freedom) and cognitive-cultural phenomena (e.g., what people expect of the news and how they assume news work is performed).

This new institutionalist framing of Field Theory and the press is then used to generate a sociological model in which the press‘s institutional autonomy—the extent to which it can define and defend its dependencies and separations—is described in terms of five types of relationships including: field interactions, capital competition, negotiated production, transitioning actors, and public work. Each relationship is a sociological site in which the press works out its autonomy.

Finally, equipped with this history of Field Theory, its new institutionalist update and application to the study of journalism, this chapter revisits the sociological

13 history of the U.S. press to recast it as story of negotiated autonomy. Specifically, the chapter suggests that the press‘s ideal of objectivity, its ritualized organizational routines and its strategic distance from publics can be understood as exercises in establishing distance from, and reliance upon, a set of actors and ideas.

The purpose of this chapter is to show how a networked model of press autonomy connects to old and new theories of fields, and how it might be read into existing sociological histories of the press.

Chapter 4: Negotiating Press Autonomy in Contemporary Online Journalism

This chapter tackles three questions: what are the socio-technical features of contemporary networked news production; how do these features relate to studies of automaticity and infrastructure in Science, Technology & Society literature; and how can online press autonomy be understood in terms of ―newsware‖ infrastructure distributed among visible, automated processes and human decision-making? The aim of this chapter is to prepare the concept of ―newsware‖ for Chapter 5‘s comparative case study of one particular kind of newsware, the application programming interfaces of three mainstream news organizations.

Specifically, this chapter starts from the premise that networked news production is fundamentally a kind of relational work grounded in network logics.

Such logics see the news emerging among interdependent actors with complementary roles, governed more by norms of reciprocity and reputational concerns that span different partnerships and rule systems than any single definition of what the press is, or what journalists are (Powell, 1990; Turner, 2005).

Networked, online news production is a relatively new field of practice and research, but with a growing number and diversity of studies (Mitchelstein &

Boczkowski, 2009). There are ethnographies tracing the development and use of new technologies in newsrooms, and co-evolution of journalistic and technological logics

(e.g., Boczkowski, 2004a, 2004b, 2004c, 2009; Cottle, 2007; Kleinberg, 2005; Royal,

2010). There are studies of how professional journalists and participating publics interact and co-produce news in online environments (e.g., Braun & Gillespie, 2010;

Domingo et al, 2008; Hermida, 2009a; Hermida & Thurman, 2008; Lewis et al, 2010;

Lowrey & Mackay, 2008; Muthukumaraswamy, 2010; Neuberger & Nuernbergk,

2010; Reich, 2008; Singer, 2010; Thurman, 2008; Williams et al, 2010), how such interactions are changing the professionalism, culture and responsibilities of journalism (e.g., Bardoel, 1996; Bardoel & D‘Haenens, 2004; Beckett & Mansell,

2008; Deuze, 2007; Kunelius, 2006; Kunelius & Ruusunoksa, 2008; Robinson, 2007;

Singer, 2006, 2007), and how, despite there being a greater number of online news outlets, there is less diversity available to online news readers (e.g., Boczkowski & deSantos, 2007; Haas, 2005; Hindman, 2008).

Equipped with a review of patterns in networked news research, this chapter then argues that the networked press‘s autonomy might best be understood in terms of

―newsware‖ infrastructure: technologies and practices that structure its systems of dependencies and separations. Specifically, this chapter operationalizes ―newsware‖ infrastructure in terms of two features: accessibility and automation. That is, what aspects of networked news infrastructure can be seen and controlled, and by whom?

How much of the system of expression defined by online news can be seen and

15 controlled by human actors, and how much of it is embedded within structures and algorithms that are readable primarily by machines or software designers?

While newsware could be operationalized in other ways for other studies— e.g., tracing flows of financial capital, professional reputation, or inter-organizational influence—I focus here on accessibility and automation because, together, they speak to a contemporary autonomous press‘s structural potential. Essentially, the ability of the online press—however it might be institutionally defined and configured—to enable both publics‘ rights to hear and individuals‘ rights to speak depends upon how well it and others understand and can access the means by which it produces news.

Only a press that can see, appreciate and thoughtfully reconfigure its infrastructure is capable of achieving its structural goals.

The primary aim of this chapter—its review of networked news practices, its infrastructural definition of newsware and its reliance on social informatics—is to set up the next chapter‘s comparative study of one type of newsware: the application programming interfaces (APIs) used by news organizations to give people outside of their news organization access to their online content.

Chapter 5: Application Programming Interfaces as Sites for Press Autonomy

This chapter is an empirical investigation of newsware, a study of three mainstream news organizations‘ APIs.

Data show that the websites of mainstream, ―legacy‖ media—most of whom existed before the internet—still draw the majority (67 percent) of online news traffic

(Pew, 2010) and are still consistently the most highly ranked by search engines

(Hindman, 2008). Although there always new news websites and blogs that may behave like the mainstream media, the websites of traditional news organizations account for a significant portion of internet news reading. It seems particularly important, then, to study their use of newsware infrastructure.

Three mainstream news organizations in of particular interest—The New York

Times, The Guardian and National Public Radio—not only because of their considerable readership and histories, but also because they are leaders in developing public-facing newsware infrastructures.

Each organization is currently engaged in a considerable amount of activity designing and attracting users for their application programming interfaces (APIs).

These APIs are essentially publicly available software toolkits with computer code, documentation and terms of service that detail how people outside of these news organizations can access and republish content created within them. The process for using these APIs is generally as follows:

1. Visit the news organization‘s API website and apply for an account and ―key‖

that unlocks the API‘s features.

2. When approved with a key, download the API itself, a collection of computer

codes that are essentially gateways into the parts of the news organization‘s

content management system that it has made accessible through the API.

3. Create your own website that uses the API to access and download the news

organization‘s content; repackage and republish this content on your own site,

according to the API‘s terms of service.

Such APIs let anyone with the programming skills create their own news websites or

―widgets‖ (smaller web-based programs that can be embedded within websites).

Each API is slightly different. The National Public Radio API (NAPI), for example, launched in 2008, essentially offers the ability to query its database of stories, stations and transcripts through lists it maintains of topics, programs, blogs, biographies, artists and series. Programmers could, for example, query NAPI for all stories about Afghanistan that originated from WHYY (Philadelphia) and WBUZ

(Chicago) and be given the transcript of an interview Terry Gross conducted on ―Fresh

Air‖ and a pointer to a ―This American Life‖ podcast. In April, 2010, NPR claimed that, in the last 6 months, it had delivered almost 5 billion stories through its API

(Melanson, 2010).

Also launched in 2008, the New York Times‘ API ―Developer Network‖ (DN) is an extensive collection of tools for accessing and republishing Times content. It states that its APIs helps the Times to ―learn more about what our readers want and gain insight into how news and information can be reimagined‖ as well as to ―fulfill the newspaper's journalistic mission by putting more information in the hands of the public. [The APIs also] expand that mission by giving users the ability to find and tell their own stories.‖ (New York Times, 2010)

The Times API is actually a set of 13 sub-APIs including everything from the

Community API (a code set to download comments left by NYTimes.com users) to

The Most Popular API (a code set to access the links and blog posts that are most frequently emailed, shared and viewed by NYTimes.com readers). The Article Search

API gives users access to the headlines, abstracts and multimedia content of any Times

18 story published since 1981. In addition to these APIs, the DN includes a set of tools that, for example, let users access API features without writing code and create their own visualizations of Times data.

The Times does not offer statistics on how often its API is used, but lists 32 applications created using the DN, everything from ―CEO Politics‖, an application that uses the Times Campaign Finance API and Google Maps to display the geographic origins of corporate contributions to the 2008 presidential campaigns; to ―Reading

Radar‖, an application that uses the Times Best Sellers API to create lists of NYT best- selling books with links to purchase at Amazon.com.

The Guardian API, called ―Open Platform‖ (GOP), is perhaps the most extensive news API of the three. The GOP has two key components: a ―Data Store‖ and ―Content API‖. Like the DN and NAPI, the GOP serves content already published by The Guardian (over 1 million articles published since 1999 are available). However, its Data Store also includes an extensive collection of data sets from outside The Guardian. It includes everything from private industry figures (e.g., it includes data that AIG publicly released describing it and other banks‘ receipt of

U.S. government subsidies) to non-governmental studies (e.g., UNESCO figures comparing education, life expectancy and income between men and women worldwide) to a searchable index of government datasets from around the world (e.g., real-time data on London traffic, Vancouver‘s database describing city-owned properties, and New York City‘s inventory of privately owned public spaces). The

GOP does not simply link to these data sets, it reformats many of them in ways that

19 are public accessible (e.g., Google Docs), vets them for accuracy, and makes them readable to the second part of the GOP system, its Content API.

Like the NAPI and DN, the GOP Content API is essentially a software toolkit for making queries of GOP indexed data and stories. It lets users make queries of

Data Store content using a variety of index categories. For example, it lets GOP users make queries using standard categories like story title, website section, byline, and date; but it also offers other descriptors like whether the story is an ―editor‘s pick‖, has content judged by editors to be ―related‖, or whether the content is free, requires a subscription, or is only accessible through a higher level of approval that serves content without advertisements, in higher-quality data formats (e.g., greater resolution in photos and videos), and without any volume restrictions on how much content can be served. The GOP documentation states that all of these categories are manually maintained and applied by its editorial staff. As a sample, Figure 5 (below) shows a story that appears on The Guardian’s main site guardian.co.uk (top) and the corresponding API architecture underlying its data (bottom).

Figure 1: A sample story from The Guardian (top) with the same story‘s representation in the Guardian Open Platform API (bottom)

Anyone can access these indices and search for stories, but (like the DN and

NAPI) accessing Guardian data—but not the governmental, NGO or private industry data The Guardian serves—requires registering with The Guardian for an API ―key‖ that lets it track how an API programmer is accessing the GOP.

The point here is that The Guardian claims that it offers API users access to its

―complete information architecture‖ (both stories it generates and public data sets) but such access is mediated and regulated by a system that affords and constrains what such access means.

This chapter is a study of how autonomy is encoded: within these three different API systems; in trade- and popular-press coverage describing their design and use; and in a selection of online systems built using these APIs.

I follow a content analysis approach in which I begin with the overarching concept of autonomy and the two features of infrastructure identified in Chapter 4— accessibility and automation—but purposefully keep the operationalizations of these concepts sufficiently flexible in order to respond to categories I observe in my empirical material. Although this is by no means a full ethnography, I intend to follow the spirit of Marcus‘s (1995) ―multi-sited ethnography‖ by attempting to ―follow autonomy‖—closest to Marcus‘s method of ―follow[ing] the plot, story, or allegory‖

(1995: 109)—and triangulate among three types of data:

API material. Each API (NAPI, DN, GOP) publishes its code, documentation

describing its use, and user forums in which coding problems and design

decisions are publicly discussed. These codes, documents and forums are

accessible to those who have registered with the news organizations.

Meta-API material. Each API has a set of intellectual property agreements,

frequently-asked-questions and marketing materials that describe and

contextualize its APIs. Additionally, there is a growing set of popular and

trade-press stories about the APIs that contextualize and critique them. I have

been building a database of these stories indexed by LexisNexis and Google.

API-built systems. Each API has produced a set of systems, designed by third-

parties external to the news organizations. These systems are publicly

accessible and indexed both at each news organization‘s site, and in popular

and trade-press articles.

My analysis of these materials has two purposes: to identify categories into which the different materials fit, and to trace among instances within these categories moments relevant to the idea of press autonomy as developed in Chapter 4. Specifically, I follow a content analysis technique called ―categorical aggregation‖ in which the goal is to ―seek a collection of [categorical] instances from data‖ (Cresswell, 1998 : 154) in order to form larger, issue-relevant meanings that connect the data to the research framework (Holsti, 1969, cited in Jones [1996: 128-129]).

Sample questions I ask as I code the data include:

- Within APIs, how is news content (stories, themes, people, events) coded?

Where are these codes defined, how are they updated, and who has power to

change them?

- What are an API‘s technical boundaries? For example, what other libraries or

coding languages does it rely on? What formats can queries be made in and

what data formats are returned? Are any other organizations or organizational

data streams explicitly included in the APIs and, if so, who are these

organizations?

- What distinctions, if any, does an API make with respect to genres of

journalism? For example, does the API code categories using traditional labels

like ―investigative reporting‖, ―opinion/editorial content‖ or ―business news‖

or are there new categories used to describe its content?

- Related to the idea of how objectivity may appear in the design and use of

news APIs, are distinctions made between descriptive reporting and

interpretive reporting, e.g., when are datasets visualized, when are they

interpreted, and when is there a combination of both? Do systems suggest the

desired or acceptable roles it envisions readers to playing? E.g., when do these

systems offer interfaces that allow readers to interact with datasets, and when

do they offer data interpretations with particular perspectives or arguments that

audiences are meant to receive but not change?

- What materials are made available through an API? Do API users have access

to full stories, story segments, reporter notes, background materials? Does the

API offer access to other data sources created or maintained by other

organizations (e.g., government, private sector, or NGO data)?

- Are the APIs themselves open-source and examinable by users? If so, have

changes to the APIs been made and, if so, by whom? Have any related APIs

spawned from the news organizations‘ APIs and, if so, by whom?

- What reasons do news organizations give in their documentation for why they

are offering the APIs? Do they offer arguments about the type and extent of

material they are making available? Are these reasons and arguments different

from those that appear in the trade- or popular-press coverage?

- What assumptions, if any, appear within API code, documentation, or news

stories about who is using or will use these APIs? What skills and

relationships, if any, are assumed by the API designers and what support, if

any, is offered to those lacking those skills and relationships?

- What assumptions, if any, are made about those who will use the systems built

with the APIs? How is the audience for these systems described, either by the

designers of APIs or the creators of systems that use the APIs? Are people

within the news organizations permitted to use the APIs and are any

distinctions made between their use and other people‘s use?

- What kind of intellectual property agreements must an API user enter into?

How are these terms of usage different from the content terms on which the

news organization itself operates? Do these agreements make mention of any

press-specific laws, e.g., regarding libel or shield protection? What kind of

risks and responsibilities are embedded within these agreements and what, if

any, opportunities are there to contest them? Are there any examples of API

users breaking these agreements and, if so, what were the consequences?

- Where do data reside in news organizations‘ APIs? That is, different from

legal claims to content ownership, who technically owns and controls content?

Do API users download copies of content from the organizations‘ sites, access

streamed content that remain on news organizations‘ sites, or some

combination of both approaches? Essentially, can API users create systems

that can exist as stand-alone experiences independent of news organizations, or

is a relationship through data always maintained?

- What assumptions does the API make about news timeliness? That is, do API

users have access to content at the same time as the news organizations? How

often are API users able to publish content updates?

- What information, if any, is collected about the users of APIs by news

organizations, or about people who use the systems created by API

programmers? For example, do APIs offer any channels (e.g., browser

cookies, user accounts) by which news organizations learn about API

programmers or online news readers? How does the API allow access to this

information and what restrictions, if any, are there on what kind of information

APIs can collect?

- Are there cases of API privileges being revoked?

- Do news organizations allow or require APIs to gather revenue?

The goal in answering these questions is to construct an empirical account of how the design and use of APIs—as an example of newsware infrastructure—encode autonomy.

I should acknowledge three points here. First, the story of autonomy I tell using these data is arguably from the perspective of these news organizations. It is their APIs that form the core of the materials I will analyze. Given the theoretical framework of press autonomy developed throughout the dissertation—the idea of a distributed, networked autonomy—it is important to note that my empirical study is

26 largely focused around these news organizations. Even if their perspectives are critiqued in press stories on the APIs or through subversive usages of the APIs, it is still their perspective that is the main starting point.

Second, this analysis is not informed by interviews with the designers or users of these APIs. A more empirical investigation would include on-site observation of how these APIs are created and changed by news organizations, as well as interviews with API users about how and why they made certain design decisions over others. In that kind of study, my challenge would be to trace press autonomy among not only published content but among the statements, behaviors and rationalizations of various actors – and then to analyze these alongside the other content studied here.

Third, I should self-reflexively state my own background as a programmer and system designer. I studied Computer Science, worked as a design researcher at the

MIT Media Lab and Media Lab Europe and have built several software and hardware systems using commercial off-the-shelf materials similar to the APIs studied here. My observations of the API code, my readings of the press coverage on them, and my analyses of the systems built with them are arguably impacted by my own opinions on, and experiences with, programming, system design and its cultures.

Chapter 6: Conclusion

This chapter summarizes the dissertation‘s arguments in light of Chapter 5‘s empirical findings. Specifically, it revisits Chapter Two‘s model of autonomy, reviews Chapter

Three‘s historical tracing of press autonomy negotiations and summarizes Chapter

Four‘s study of contemporary press autonomy and newsware infrastructure. I discuss implications and limitations of the study and future work.

CHAPTER TWO:

Press Autonomy in Legal Theory and Supreme Court Jurisprudence

1. Introduction

This chapter asks four questions intended to ground this dissertation‘s study of autonomy in legal theories of press freedom, and jurisprudence on the First

Amendment‘s press clause.

First, why is autonomy framed in terms of separation from, and dependence upon, other actors? Second, how is this concept of autonomy related to the idea of a free press and its ideal institutional characteristics? Third, focusing on Supreme Court of the U.S. decisions, how has the press clause been interpreted, and what rights and responsibilities have these interpretations conferred upon the press? And finally, using this ideal concept of autonomy and this understanding of press clause jurisprudence, what are some key ideas in legal theory guiding contemporary models of press freedom in light of distributed journalism?

By the end of this chapter, the reader should have a good understanding of: how this dissertation understands autonomy as a general democratic concept; how this concept is intertwined with legal theories underpinning a free press; how the Supreme

Court has struggled with and decided questions of press freedom; and key ideas guiding contemporary models of press freedom, specifically those concerned with how to understand the press‘s socio-technical infrastructure in relation to its autonomy.

2. The Idea of Autonomy in Democratic Systems

What is meant by autonomy and why be concerned with it at all? As Feinberg (1986)

(cited by Christman, 1988: 109-110) argues, the word has usually been accompanied by an implicit modifier: ―personal.‖

―Personal autonomy‖ variously means ―the ‗capacity‘ to govern oneself, the

‗actual condition‘ of self-government, an ideal of virtue derived from that conception, or the ‗sovereign authority‘ to govern oneself.‖ The term has generally been taken to mean: a future-oriented concept describing what people are able to imagine and realize; a way to critique someone‘s current circumstances and evaluate present conditions; a normative, theoretical aspiration independent of any particular set of conditions; and the fundamental basis for a right to act or behave as one wishes.

Others state that autonomy can mean the ability to: direct your own actions independent of others; change your desires in the face of circumstances and feasible options that you cannot influence (Meyer, 1987); make choices and compromises with full knowledge of various consequences (Benn, 1988; Dworkin, 1988) and understand whether these decisions have been made under the influence of others and free of coercion (Arneson, 1985; Neely, 1974). Finally, and most broadly, autonomy can mean distinguishing between ―global‖ and ―local‖ levels of control: knowing the difference between factors that are beyond one‘s immediate control (e.g., legal or military force, broad cultural trends and social traditions) and others that are more responsive to individual differences and preferences (e.g., personal conversations with members of the same community, or negotiations amongst a few known people who share resources and are commonly affected by outcomes) (Dworkin, 1981).

Essentially, autonomy means knowing what you can influence and change – and how difficult it might be to realize your preferences.

Most fundamentally, though, autonomy is a basic concept ―at the center of the modern democratic project, a principle which can provide an anchor point for both conceiving and building a new and more robust account of democracy.‖ (Held, 2006:

260) This anchor, Held claims, is a kind of organizing principle – an umbrella of aspirations and goals under which other democratic values fit. Held (2006: 263) defines these aspirations as:

1. the creation of the best circumstances for all humans to develop their nature and express their diverse qualities (involving an assumption of respect for individuals‘ diverse capacities and their ability to learn and enhance their potentialities); 2. protection from the arbitrary use of political authority and coercive power (involving an assumption of respect for privacy in all matters which are not the basis of potential and demonstratable ‗harm‘ to others); 3. involvement of citizens in the determination of the conditions of their association (involving an assumption of respect for the capacity of individuals to come to reasoned standpoints); 4. expansion of economic opportunity to maximize the availability of resources (involving an assumption that when individuals are free from the burdens of unmet physical need they are best able to realize their ends).

There are several key ideas to extract from Held‘s statement of the autonomy principle, ideas which guide this dissertation‘s understanding of autonomy.

First, Held‘s principle accepts the idea that autonomy depends on a set of

―circumstances,‖ a system of conditions that exists in the world, not within a single individual. Yes, individuals make themselves and decide who they are – but only within the messy and complicated constraints and empowerment of these conditions.

This is related to, but conceptually distinct from, what Dahl (1989: 100) characterized as the ―presumption of personal autonomy‖ which assumes that in ―the absence of a compelling showing to the contrary everyone should be assumed to be the best judge 31 of his or her own good or interests.‖ Held sees social conditions as dualistic constraints that both limit and enable individuals while Dahl describes autonomy as an ideal that exists, in its purest form, absent from social relations.

This presumption is consistent with the liberal political tradition (Mill,

1859/1974: 81) in which equality is considered to be a private and individual matter focused on removing obstacles to personal realization:

The only freedom which deserves the name, is that of pursuing our own good in our own way, so long as we do not attempt to deprive others of theirs, or impede their efforts to obtain it. Each is the proper guardian of his own health, whether bodily, or mental and spiritual. Mankind are greater gainers by suffering each other to live as seems good to themselves, than by compelling each to live as seems good to the rest.

Mill acknowledges that people live within associations, but his model of liberty is focused on ensuring that these relationships do not interfere with an individual‘s opportunity to make for themselves lives they see as desirable. Indeed, the liberal, pluralist tradition is principally concerned with ensuring that people have equal opportunities, seeing community and social associations as either instrumental to the achievement of personal advancement or as sentimental tangents to the core project of personal liberty (Christians, et al. 2009: 96-99). Whereas Mill and Dahl see autonomy as an ideal individuals realize in the absence of state interference, Held argues that political autonomy is a state of being distinct from some kind of pure, ideal individual independence (a fundamentally unrealizable ideal). Rather, autonomy emerges from and depends upon a set of social, political, economic and cultural circumstances. It is produced by a system of relationships. The core practical challenge to individuals is to decide amongst themselves what kind of constraints and conditions best enable a

32 kind of dualistic individual and collective autonomy that requires associating with each other.

Second, while Held acknowledges that autonomy requires protection from state force and unreasonable interference, he importantly qualifies this statement by saying that autonomy depends upon the absence of ―the arbitrary use of political authority.‖ That is, there is some acceptable set of political uses of power that fit within the principle of autonomy – but they are only legitimate if they respect a distinction between private matters and public concerns. This principle of autonomy argues neither for the absence of the state—or a thin version of governmental influence that leaves citizens alone—nor for a state that casually exerts control over citizens‘ private matters.

Third, this principle of autonomy assumes that citizens have the capacity ―to reason self-consciously, to be self-reflective and to be self-determining. [This capacity] involves the ability to deliberate, judge, choose and act upon different possible courses of action in private as well as public life.‖ (Held, 2006: 263) Note a distinction Held makes here: he does not claim that citizens possess, a priori, the ability to be self-conscious, reflective and determining; rather, he asserts that each person must have the potential to be so. As with his emphases on the circumstances of human development and the contingent and contestable use of political power, Held frames individual autonomy as a phenomenon that emerges from a set of relationships.

These relationships have a recursive and self-referential nature: they must help people develop the very capacities they need to consciously enter into and manage the kind of associations that political autonomy requires.

Finally, Held sees autonomy as dependent upon a set of resources: material and symbolic goods that citizens must have access to and be able to use if they are to realize their potentials. Essentially, someone‘s autonomy depends not only upon having resources to execute pre-existing preferences (e.g., money or tools to broadcast their opinions or implement their designs) but exposure to new perspectives that come from outside themselves. Autonomy means having the power to assert a preference and using that power to encounter new perspectives that will, in turn, make future preferences richer and more multi-faceted.

It is only by acknowledging the importance of these external factors— essentially, being secure enough in yourself to learn things that might challenge that security—that people might move closer to a self-reflective ideal of autonomy in which they can appreciate, trace and be in conversation with the influences around them. As Benn (1988: 124-125) argues:

And at the highest level one might say that a person, though fully competent and subject to no external pressures, was not causally independent if he was dominated by his own prejudices blinkered by unexamined ideology, or a slave to convention. One says of such a person that he is heternomous; one looks for the causes of his decisions in the opinions and believes of other people which his own merely reflect.

That is, any ideal of autonomy that fails to acknowledge the relational and socially constructed nature of individual choices actually creates an illusion of autonomy.

Someone may seem independent and free of external influences, but he is actually simply receiving and uncritically recreating a system of values and influences he cannot see2. (For example, this type of illusory autonomy may appear in a person‘s

2 This idea of unseen influences and hegemonic force has a long and complex history; a complete review of which is beyond the scope of this dissertation. However, it is worth noting that this idea of autonomy as a negotiation between individual agency and structures of influence and domination is more fully explored in such works as: Arendt‘s ―The Human Condition‖ (1958/1998) in which she 34 supposedly rational choice of a product or service, without understanding that their choice has likely been influenced by a system of media advertising, a company‘s strategic positioning in relation to competitors, pre-existing brand loyalties, or assumptions about what people like them usually buy.)

The underlying ideal here is that autonomy means having both ―negative‖ and

―positive‖ freedoms. That is, citizens have both the right to be ―left to do or be what he is able to do or be, without interference by other persons‖ but such a right is inseparable from ―source[s] of control or interference that can determine someone to do, or be, this rather than that.‖ (Berlin, 1969: 121-122) Although this distinction is a helpful starting point for understanding the dualistic nature of freedom—as something that involves both the individual and her surrounding environment—it is too simplistic.

Notably, MacCallum (1967: 314) argues that freedom actually emerges from a triad of relations: ―freedom is always of something (an agent or agents), from something, to do, not do, become or not become something.‖ That is, the situated nature of the actor—her identity and position within a network of others—is critically important for understanding, practically, what it means to be free from control or free to act. Identities that appear to be properties of individuals are, in fact, relationally defined and thus intricately define what kind of freedom is pragmatically possible.

argues that an ―active life‖ requires seeing oneself as never fully in control of one‘s personality because it is always subject to constructs beyond one‘s control (the best we can do is promise ethical behavior and forgive lapses thereof); Foucault‘s ―Discipline and Punish‖ (1979) and ―Governmentality‖ (1991) in which he traces a series of social control techniques that reconfigure citizens‘ behaviors and identities in ways that best suit the aims of the state and coercive power; or Gramsci‘s ―Prison Letters‖ (1988) in which he argues that an individual‘s freedom (particularly for workers who have historically had little access to many different forms of power) depends not only upon securing economic or political capital for one‘s own personal use, but on critically understanding and mastering often unseen cultural influences and social systems that control the very ideas that an individual might be able to imagine. 35

This view challenges the assumption (made by Held and others) that citizens‘ capacities are ideally and evenly distributed. At any given period of time—it is important to remember that distinctions between what is possible or impossible are always historically contingent—a particular agent may be constrained characteristics that are thought to be ―natural,‖ or so ingrained as to be immutable3. For example, at various points in history, different cultures have considered people of different ethnicities, genders, physical abilities, sexual orientations, or family backgrounds to be—by virtue of inherent, existential features—inferior and incapable of full citizenship. MacCallum‘s point is not that autonomy should be structured using these features but, rather, to reassert that, in any discussion of freedom, the characteristics of agents cannot be forgotten or lost amidst accounts that abstractly distinguish between positive and negative freedoms. Especially when articulating ideals of autonomy, we must be sure not to state explicitly—for any particular time in history—which people exactly it is that we have in mind as free, and how these people are practically able to execute their freedoms from influence or freedoms to act (MacCallum, 1967: 315-

319). MacCallum‘s clarification helps us work within the system of positive and negative liberty to ask not just ―is someone free?‖ but ―when are particular people free?‖

Another critique of the positive-negative theory of freedom comes from Fallon

(1994) who argues that autonomy might better be understood as descriptive and

3 See Friedman (1986) and Christman (1987) for good discussions of the idea of a ―true self‖ and whether people are ever able to imagine and act to realize such selves. Friedman argues that, despite people‘s impressions that they act in relation to a single and consistent set of values, their self-visions are always subject to forces beyond their control that temper the idea a ―true self.‖ Instead, a ―true self‖ might better be thought of as a dynamic way of responding and acting in relation to always changing circumstances. 36 ascriptive concepts. As a description, autonomy refers to people‘s ―actual condition and signifies the extent to which they are meaningfully ‗self-governed‘ in a universe shaped by causal forces.‖ (877) Like MacCallum, Fallon is concerned with pragmatically evaluating individuals‘ actual capacities in particular sets of circumstances because, he says, to do otherwise is to ignore how different individual and environmental conditions pragmatically afford and constrain autonomy4.

Autonomy is not dualistic concept—a binary ideal—but, rather, is a ―matter of degree‖ (877) that depends in part upon how sophisticatedly someone understands their position in relation to others, and the influences upon him. Similarly,

―paternalism can sometimes be defended as a means of preserving or promoting autonomy.‖ (877) For example, limits on fast food advertising may be needed to curb the addictions of people who do not understand that—although they may feel like they are making independent eating decisions—their behaviors are subject to the complex manipulations of advertisers whose media messages are more powerful than personal willpower. Essentially, autonomy may depend upon a system of tradeoffs that may entail impinging upon some personal freedoms in order to secure others.

In contrast to descriptive autonomy, Fallon defines ascriptive autonomy as an ideal, ―the autonomy we ascribe to ourselves and others as the foundation of a right to make self-regarding decisions … a moral entailment of personhood.‖ In the same spirit as MacCallum, Fallon makes this distinction to draw attention to the differences

4 In an essay reflecting upon Berlin‘s notion of positive and negative freedoms in relation to the writings of T.H. Green, Simhony (1993) argues autonomy is better described as struggle between internal-facing and external-facing capacities. That is, autonomy depends upon abilities that reside within the single individual (things she is able or not able to do at any moment) and abilities that exist within the social environment surrounding the individual (both things that collectives are able to do and things that collectives allow individuals to do, allowances made through coercive force or cultural signaling). 37 between idealized and actual autonomy. He does this not to argue against the kind of ideal autonomy Held describes but, rather, to provide a systematic way of understanding how and why autonomy is limited, and to appreciate how compromises of autonomy might be justified and critiqued. In making this distinction, Fallon aims to do two things. First, he is trying to rescue and preserve Kant‘s original ideal of formal autonomy as a ―freedom of the will from causal determinism‖ (Fallon, 1994:

878) and a freedom to use reason (as opposed to faith or other types of logics) when making decisions (Kant, 1785/2002). Second, he aims to re-articulate this ideal in a way that can help people critique their actual freedom in contemporary circumstances.

Common to these critiques of positive liberty—MacCallum‘s call to pay more attention to the subject of autonomy and Fallon‘s distinction between pragmatic and ideal forms of autonomy—and consistent with liberal responses to positive liberty is an idea that social forces constrain or limit individual self-government (see Christman

[1991] for a good review of liberalism and positive freedom). However, in addition to constraining or controlling individuals, social influences can also have empowering effects on people and are indeed integral to realizing autonomy:

To be autonomous one must have reasons for acting and be capable of second thoughts in the light of new reasons … And for reasons one must have a system of beliefs from which action commitments derive and into which new evidence can be assimilated, yielding new commitments. How could anyone come by these bits of basic equipment except by learning them in the first instance from parents, teachers, friends, and colleagues? Someone who had escaped such a socialization process would not be free, unconstrained, able to make anything of himself that he chose; he would be able to make nothing of himself, being hardly a person at all. Within this conception of a socialized individual, however, there is still room to distinguish as autonomous a person who is committed to a critical, creative and conscious search for coherence within his system of beliefs ... The resources on which he will rely for this critical exploration must lie, necessarily, within the culture itself, supplemented, perhaps, by those elements of alien cultures with which he has become acquainted. (Benn, 1988: 179)

This is a critical point: autonomy requires both facility within a culture— understanding how to craft consistent yet contestable beliefs with those you encounter most—and facility among cultures—encountering and incorporating new perspectives that you are unlikely to learn from those within your usual community. Autonomy means knowing who you are and exploring who you could become.

This issue of genuine choice is central to Benn‘s theory of freedom. He distinguishes between choices and decisions, arguing that people often mistake observable decision-making (or constraints on actions) for the exercise of true choice

(among a set of options whose costs and benefits are fully appreciated):

Though an agent does willingly what he has no option but to do, the fact remains that an external agency has interfered with his conditions of action and has made him unfree to do anything different; he may not resent having to do it, and because he would have done ‗of his own free will‘ what he is required to do, one might be inclined to say that he does it freely. But if he is not free to do anything different, what he does cannot in fact be the thing he chooses to do, for he has no choice. It is at best what he would have chosen, had the choice been available to him. To abridge an agent‘s freedom is to abridge the range of possible choices, not necessarily to act so that he does something that he might otherwise not have done. To say that he willingly or freely acts in the required manner is to say that he acts without resenting the requirement. But whether one is actually free is a function of what options one might have had but for someone else‘s intervention, or neglect to intervene; it is not a function of the absence of feelings of resentment, of grievance and frustration. (Benn, 1988: 145)

Benn encourages us, then, not to concentrate on whether people feel like they are free in their decision-making (relying on self-reported satisfaction) but instead to more closely examine the conditions under which they make decisions and ask how meaningful the choices are that they consider while deciding. Benn‘s focus is rightly not only on ―objective choice conditions‖—constraints rooted in a person‘s physical realities that ―exist independently of his beliefs about them‖ (153)—but on the often

39 hard-to-see ―subjective choice conditions,‖ factors that influence the beliefs on which someone chooses to act.

A person can be misled by false information about the resources at his disposal into believing that he is unable to do what he can in fact do; or he may be wrongly informed about the environment which provides and constrains his opportunities … With his beliefs and preferences so managed, his actions can be controlled with no alteration to his objective choice conditions, effectively depriving him of the freedom to do whatever he nevertheless has the power or capacity to do. (Benn, 1988: 154)

Seen as such, decision-making is not a deterministic activity over which individuals have complete control. It instead becomes a probabilistic phenomenon in which someone‘s autonomy depends upon the likelihood of encountering and appreciating a truly diverse set of choices. It depends upon social conditions (systems of associations and mutual dependencies) and cultural logics (meaning-making that takes symbolic forms) that produce options that can be articulated and seen. Simply put, my personal autonomy depends upon other people providing me with a truly diverse set of choices that I might select amongst and thoughtfully decide to act upon5.

In sum, autonomy sits at an intersection between the individual and the collective, the private and the social. We might see and judge personal autonomy according to the actions of individuals but, more fully considered, these behaviors emerge from ―conditions of enactment, that is, institutional and organizational requirements‖ (260) that enable individual autonomy. The circumstances, protections, associations, and resources required for autonomy might be recognized in individual action but they are made possible through social inter-actions. Pragmatically, autonomy thus becomes a problem of institutional design: understanding how the

5 This idea is further explicated in Chapter Three of this dissertation, but it is closely related to what Bourdieu calls a ―space of possibles‖ (1993: 30) that delineates a field of activity‘s scope while simultaneously pointing to the creative potentials within it. 40 democratic ideals enabling Held‘s principle of autonomy are instantiated in, and challenged by, the organizational arrangements that govern social relations.

It is with this understanding, then, that we might turn to the design of one particular institution: the press. Essentially, if personal autonomy requires socially constructed and collectively managed circumstances, protections, associations, and resources, we need to consider how, in its ideal form, the institutional press might support the democratic autonomy of individuals. An autonomous press is a kind of interstitial, institutional glue that works across multiple levels: it enables an individual‘s autonomy by ensuring that she hears everything she needs to realize her personal autonomy; it understands and depends upon social relations to ensure that it presents individuals with truly diverse options; and it uses law, culture and technology to distinguish and protect itself sufficiently well from other types of institutions so that it can do the work it needs to do to help democracies realize autonomy.

3. Rationales for Protecting Free Speech

If we accept that autonomy—both as an ideal and as a pragmatic lens for critique—is a critical element of democratic life encompassing both a freedom to be free from unreasonable control and a freedom to associate with others, then we must ask which institutions and organizational conditions enable us to be autonomous, ―socialized individuals.‖ (Benn, 1988: 179)

The press is not the only public-facing institution that enables such dualistic autonomy and socialization. Although each has a particular history as an organization serving democratic functions: public schools and educational institutions have been

41 ideally envisioned to ―provide every child with an opportunity to choose freely and rationally among the widest range of lives‖ (Gutmann, 1987: 34); U.S. museums have historically been seen as sites for the display and critical discussion of objects not normally accessible to private individuals (DiMaggio, 1991); and public libraries were originally established to give people means to experience new knowledge in both public environments and private settings (e.g., consider both libraries‘ education and lending programs) (Kerslake & Kinnell, 1998).

The press‘s identity as a public-facing institution is similarly grounded in the idea that individual autonomy requires resources beyond the power of any single individual; but the press‘s relationship to the idea of positive freedom and dualistic autonomy is predicated on its ability to act as an institution that supports free speech.

The relationship between speech and freedom is a complex one, but it is fundamentally based on the idea that speech is an ―other-regarding act.‖ (Schauer,

1982) Simply put, speech affects others and can thus be defended and regulated on what Scanlon (1972: 204-205) calls ―consequentialist‖ grounds in which acts of speech are pragmatically weighed for their ability to produce both good and bad outcomes. Theories of speech regulation that encourage some speech conditions over others, that punish some speakers but not others ideally aim to maximize the benefits of speech and minimize its harms. These effects can be short-term and personal—

―saying or printing something untrue (or true) about another person may damage his reputation, humiliate him, invade his privacy, offend him, or cause emotional distress‖—or long-term and public—―the disclosure of military secrets, or the spread

42 of lies (or truth) about government may impair the efficiency of the machinery of the state‖ (Schauer, 1982: 10).

In a democratic system, this ―machinery‖ is critical to realizing a type of self- government in which individuals knowingly and freely agree and submit to constraints on their freedom that are imposed and administered by the state. As Haiman (1981: 6) describes it, ―[s]ocial order is a means to maximizing individual liberty and security‖ but, for this order to function properly, it requires people to engage in ―symbolic behavior‖ in which they express themselves, debate ideas, agree to resolutions, or maintain dissent.

Knowing that speech has effects—that it matters both to individuals and the degree of personal autonomy they derive from collective experiences—there are two main frameworks from which to consider the relationship between free speech and personal autonomy. Essentially, these are principal reasons why freedom of speech is critical to democratic societies and why speech deserves special consideration, an

―argument from truth‖ and an ―argument from democracy.‖

3.A. The Argument from Truth

The first is often called the ―argument from truth.‖ This argument is grounded in

Mill‘s assertion that knowing the truth requires the expression of others. He argued that the ―peculiar evil of silencing the expression of an opinion‖ (Mill, 1859/1974: 76) harms not only those who hold that view but also those who disagree with it: ―if the opinion is right, [individuals] are deprived of the opportunity of exchanging error for truth: if wrong, they lose, what is almost as great a benefit, the clearer perception and

43 livelier impression of truth, produced by its collision with error.‖ (Mill, 1859/1974:

76) Furthermore, the truly autonomous individual must be free even to experience harm that might result from encountering false statements; otherwise, he would ―have to concede to the state the right to decide that certain views were false and, once it had so decided, to prevent him from hearing them.‖ (Scanlon, 1972: 217)

Essentially, if truths are to be discovered and agreed upon they need environments in which conversation, debate and claim-making can proceed unfettered.

This idea of has also appeared in the rationales offered by U.S. Supreme Court justices for laissez-faire approaches to speech regulation and the importance of lightly regulated speech to the goal of ongoing social progress. Consider, for example,

Justice Holmes‘ assertion that ―the best test of truth is the power of the thought to get itself accepted in the competition of the market‖ and Justice Frankfurter‘s opinion that

―the history of civilization is in considerable measure the displacement of error which once held sway as official truth by beliefs which in turn have yielded to other truths.‖

(Schauer, 1982: 15)

This market-based theory of free speech is a powerful argument but, as Baker (1989:

6-15) shows, it suffers from three principal weaknesses.

First, adopting the theoretical perspectives of symbolic interactionism and social constructionism (e.g., see Blumer, [1969], Goffman [1959, Mead [1934/1967]), he argues that the marketplace model assumes that all truths are objective, discoverable realities but it fails to explain why some truths are better than others, truths that might be produced under some other rules of conversation, set of participants, or places in time.

Second, he claims that the classic theory assumes that ―people‘s rational faculties … enable them to sort through the form and frequency of message presentation to evaluate the core truth in the messages.‖ Essentially, he says that the marketplace model assumes that that people are somehow capable of stepping outside themselves and their own frames of understanding and then see others people‘s messages as independent of their identities and social positions.

Third, Baker claims that the value of such a marketplace, even if it could exist, is unclear because people may not always want to discover a particular truth. They may be guided by other ―irrational‖ priorities or desires to reach solutions that are

―satisficing‖ or ―good enough‖ (Kahneman, 2003; Simon, 1978, 1983) for their particular circumstances. Essentially, Baker‘s critique undercuts the assumption that a marketplace of speech—a lightly regulated space in which the state is mostly silent, taking little or no action to structure the conditions under which individuals encounter new ideas—is the desired ideal, opening up the idea that some other argument is required for a more fuller account of freedom of speech.

Further to these critiques, it seems that market-based approaches to speech do not set a deadline for when ―truth‖ might be achieved. That is, it may take a short or long time for truth to emerge; depending on the issue, people may needlessly suffer or make errors in judgment with significant consequences if they do not receive high quality information in a timely manner. (For example, consider the need, during election cycles, for there to be easy access to quality information if citizens are to cast their votes knowledgeably.) Finally, it is worth considering the U.S. Supreme Court‘s recent ruling in Citizens United v Federal Election Commission (2009: 5) in which the

45 court stated that ―[a]ll speakers, including individuals and the media, use money amassed from the economic marketplace to fund their speech, and the First

Amendment protects the resulting speech.‖ While this is true, the court‘s finding ignores the fact that some amounts of money make it more or less likely that a message will be heard in the marketplace of speech. That is, by finding that ―First

Amendment protections do not depend on the speaker‘s ‗financial ability to engage in public discussion,‘‖ (5) the Court simultaneously accepts a marketplace model of speech and makes no provisions for the fact that those with considerable resources to make themselves heard (e.g., corporations that have amassed large amounts of money from an economic marketplace) may drown out the speech of those with fewer resources. The result is an elide between the marketplace as metaphor (representing an ideal of communication in the search for truth) and the marketplace as a structural gatekeeper (harming the pragmatic creation of conditions under which equitable communication might take place and, Baker‘s critiques aside, truth might emerge).

3.B. The Argument from Democracy

The second type of rationale for defining and defending freedom of speech is often called the ―argument from democracy.‖ (Schauer, 1982: 35-45) This is primarily a political argument that free speech is necessary for providing the electorate with the

―information it needs to exercise its sovereign power and to engage in the deliberative process requisite to the intelligent use of that power,‖ and that a ―freedom to criticize‖ makes it possible to hold government officials and other powerful agents accountable for their actions (36). If the Constitution‘s main function is to delimit the state‘s

46 power over self-organizing individuals—stating how and when the state may constrain individuals‘ personal freedoms—then a critical prerequisite is that citizens have opportunities to knowledgeably and fully participate in the state‘s functioning. As

Meiklejohn (1948: 88-89) argues, this ideal of self-government (of citizen-determined limits on their freedoms) requires a set of communicative processes that depend upon freedom of speech:

The First Amendment is not, primarily, a device for the winning of new truth, though that is very important. It is a device for the sharing of whatever truth has been won. Its purpose is to give to every voting member of the body politic the fullest possible participation in the understanding of those problems with which the citizens of a self-governing society must deal. When a free man is voting, it is not enough that the truth is known by someone else, by some scholar or administrator or legislator. The voters must have it, all of them. The primary purpose of the First Amendment is, then, that all the citizens shall, so far as possible, understand the issues which bear upon common life. That is why no idea, no opinion, no doubt, no belief, no counterbelief, no relevant information, may be kept from them.

This argument from democracy is thus related to the argument from truth, but with a significant distinction. It motivates freedom of speech not from a perspective that values the apolitical exchange of information or seeking of truth but, rather, it attaches public aims to the importance of free speech because it views unfettered speech as essential to the good-faith agreements that citizens make with each other—through the state—about how and why to manage public goods and constrain each other‘s freedoms. It is thus important to note the First Amendment guarantees not a freedom to speak but, rather, a freedom of speech. Or, as Meiklejohn (1948: 25) famously wrote, the First Amendment‘s ―point of ultimate interest is not the words of the speakers, but the minds of the hearers … what is essential is not that everyone shall speak, but that everything worth saying shall be said.‖

What this rationale and Meiklejohn‘s words are essentially focused on is ensuring the social and institutional conditions that enable ―self-government‖. In line with Schauer‘s (1982) observation that speech has effects and Haiman‘s (1981) argument that public expression is the principal way in which social order, individual liberty and security are maintained, Meiklejohn‘s argues that democratic self- government—the system in which governments ―derive their just powers from the consent of the governed‖ (1948: 3)—is essentially a communication concept in which citizens‘ give their common consent to be governed. Although people certainly perform individual acts of communication that make informed citizenship possible— e.g., reading newspapers, voting in elections, writing letters to representatives, arguing ideas with neighbors—the fundamental, underlying idea of self-government is a collective one that makes limits on individuals legitimate.

This legitimacy derives from a process of communication, a making and remaking of a social compact in which ―the body politic, organized as a nation, must recognize its own limitations of wisdom and of temper and of circumstance, and must, therefore, make adequate provision for self-criticism and self-restraint.‖ (1948: 12-13)

Put slightly differently by another free speech scholar, Owen Fiss:

The purpose of free speech is not individual self-actualization, but rather the preservation of democracy, and the right of a people, as a people, to decide what kind of life it wishes to live. Autonomy is protected not because of its intrinsic value, as a Kantian might insist, but rather as a means or instrument of collective self-determination. We allow people to speak so others can vote. Speech allows people to vote intelligently and freely, aware of all the options and in possession of all the relevant information ... The critical assumption in this theory is that the protection of autonomy will produce a public debate that will be ‗uninhibited, robust, and wide-open.‘‖ (1986: 1409-1410; emphasis added)

This is essentially the communicative basis of self-government and the link between the importance of free speech and the maintenance of democratic society.

But such a concept of self-government, though, is open to a fundamental criticism that focuses on what theory of autonomy, exactly, best enables the kind of self-government Meiklejohn envisions. In his article titled ―Meiklejohn‘s Mistake,‖

Robert Post (1993) argues that Meiklejohn‘s ―collectivist‖ vision of free speech is fundamentally misguided because it assumes an end without specifying how it might be achieved. Meiklejohn‘s ideal, Post argues, adopts as its discourse model a

―‗traditional American town meeting‘ [that is] ‗not a Hyde Park‘ [or a] scene of

‗unregulated talkativeness.‖ This town meeting model presumes the existence of a common agenda, set of goals, and subservience to the meeting leaders; it fails to say exactly how such an agenda might arise, who would be responsible for deciding whether everything worth being said had been said, and what might become of citizens who either cannot or will not participate in what Post calls the ―managerial‖ structure of a town hall‘s authority. Post is right to call out an element of circularity in

Meiklejohn‘s reasoning: Meiklejohn‘s ideal of self-government relies on the existence of a system of free expression in which there is some kind of shared communication, but it fails to articulate exactly how this communication arises in the first place.

Unless we grapple seriously with this circularity, Post argues, we risk prematurely accepting theories of free speech in which we ―create organizations of heteronomy

[that] we shall all, sooner or later, be condemned to inhabit … We shall become the subjects of a power not our own.‖ (Post, 1993: 1136) Instead, Post argues that we reject overt attempts by the state to manage public discourse on our behalf because, by

49 doing so, we relieve ourselves of the individual power to influence the conditions of public discourse and, Post argues, the chance to realize the very ideal of self- government Meiklejohn envisioned6.

Lichtenberg (1987) also critiques Meiklejohn‘s ideal but does so from a slightly different perspective than Post, one that emphasizes the need for equality among individual speakers, rather than their independence from state control. She offers three reasons why any system of democratic free speech must work to ensure equality among speakers: ―there is no way of telling in advance where a good idea will come from‖ (systematically and structurally excluding some speakers will prevent quality perspectives from entering into public discourse); ―valuable contributions to arriving at the truth come in many forms, speaking the truth being only one of them‖

(learning about others and becoming knowledgeable about society involves in many other ways than simply through the exchange of factually truthful or false statements); and ―much of the value of a person‘s contribution to the ‗marketplace of ideas‘ is its role in stimulating others to defend or reformulate or refute‖ (simply communicating with someone and presenting yourself to them for their reception is a valuable exercise in realizing differences, independent of the quality of the original ideas) (Lichtenberg,

1987: 338). Lichtenberg asks the designers and regulators of systems for public

6 In an important response to Post‘s critique of Meiklejohn, Fiss (1995) cautions against Post‘s uncritical embrace of civil society‘s ability to self-regulate speech and his general rejection of state participation in the public sphere. Agendas, Fiss argues, can come not only from town hall managers but, more insidiously, from largely unseen cultural and economic forces that can set and control topics of discussion free of any requirements to be transparent or inclusive. For example, through advertising, campaign sponsorship or direct control of media companies, private corporations can have significant power to set public discourse agendas. Fiss instead envisions the state serving a parliamentarian-like role that lightly administers ―time, place and manner‖ restrictions on speech, and creates a set of incentives and disincentives to encourage equitable participation by a diverse range of speakers (guarding against the ―heckler‘s veto‖) (Fiss, 1995: 86). Fiss‘s position here is consistent with his earlier argument (1986: 1412) that the most powerful regulator of free speech, pragmatically, is not necessarily the courts or the state, but the ―forces that dominate the social structure.‖ 50 expression in general, and the press in particular, to see themselves not only as facilitators of free speech and self-government but as gatekeepers for particular kinds of self-government. She argues that the press, the state and corporations—as public- facing collectives—only enjoy free speech rights and privileges insofar as their actions serve to increase the diversity and support the equality of speech within the public sphere. That is, collective actors have structural responsibilities to facilitate free speech and ―equalize the chances of all participants or points of view to speak.‖

(351ff)

Post‘s and Lichtenberg‘s critiques help to distinguish between a normative ideal and a set of empirical conditions. That is, we can still accept Meiklejohn‘s primary theoretical aim—a system in which the consent to be governed emerges from processes of communication in which citizens knowingly and freely debate the limits they place on themselves and each other—while accepting Post‘s plea to keep dynamic and debatable the conditions of self-expression. Free speech is not an end in itself (a static state of affairs in which expression is managed by any central authority) nor is it adherence to any particular ideology (e.g., one in which the individual‘s freedom to speak is privileged over a collective right to enlightened self- determination) but, rather, as Emerson puts it, free speech can best be thought of as a system of freedom of expression. Such a system entails

the right to form and hold beliefs and opinions on any subject, and to communicate ideas, opinions, and information through any medium … the right to remain silent … the right to hear the views of others and to listen to their version of the facts [and] the right to assemble and to form associations, that is, to combine with others in joint expression. (Emerson, 1970: 3)

Essentially, Emerson links Meiklejohn‘s ideal with Post‘s critique by showing how the state might protect the rights of publics to hear speech not through censorship or managerial control, but by adopting an ―affirmative role‖ by, for example, ―furnishing facilities, eliminating distortions in the media of communication, or making information available.‖ (Emerson, 1970: 4)

In his later work Emerson (1981) elaborates on these state roles and distinguishes between two types of expressive government activity7. The first involves the promotion of the system of freedom of expression: e.g., granting subsidies to all electoral candidates without preference, building a cultural center for general use by any community organization, regulating airwaves to ensure the predictable behavior of public spectra carrying messages, protecting citizens‘ rights to speech on streets, parks, open spaces and even privately owned land that appeared and behaved like public spaces. The second entails government participation in the system of freedom of expression: e.g., when a government official issues information, a state agency makes a report, or a representative delivers a public speech. The only circumstance in which the government might legitimately exercise what Post might call its ―managerial‖ powers occurs when the state is promoting the overall system of freedom of expression, making possible ―greater opportunity for expression, increased diversity, or similar improvements in the system.‖ (Emerson, 1981: 799) Furthermore, under Emerson‘s system, the government ―may support expression by selecting a general area, or a broad subject-matter, as the object of its intervention, but may not

7 See Bezanson & Buss (2001) for an extensive review of scenarios in which the state speaks and legal judgments thereof. They go further than Emerson‘s focus on defending and constraining government speech to argue that, in democratic societies that value two-way communication between the state and the citizenry, the government has an obligation to participate in speech systems. 52 control expression within that area ... In terms of negative interference, the government must keep its hands off all expression, regardless of the area or subject matter.‖ (803-804) The government, Emerson argues, should always be expressly prohibited from: holding an audience captive for communication; communicating covertly or without disclosing itself as the state; mobilizing citizens through grassroots efforts that pit one branch of government against another; and promoting in even an implicitly partisan manner one religion or political candidate over another, especially within institutions like schools and museums designed to educate citizens (835-848).

Essentially, what Emerson‘s systems aims to do is two-fold: first, to acknowledge the state‘s legitimate and constitutionally mandated role in facilitating freedom of expression (market solutions alone will not produce the kind of communication citizens require for knowledgeable self-government); and second, to examine closely and limit those circumstances in which the state may impact the structure of, or participate in, the system of free expression.

His system sees the First Amendment‘s speech clause not as a constitutional license for engaging in unfettered expression in search of theoretical truths under any set of conditions markets allow, nor as a right for the state to create and control whatever arrangements it desires under the auspices of facilitating its vision of self- government. Rather, it both articulates an ideal of speech and specifies conditions under which that ideal might be worked out and worked toward8. What Emerson‘s

8 Schauer (1995: 20) makes a relevant distinction here when he argues that a mature system of free expression is not one focused on hard-and-fast concepts like the ―truth‖ and ―falsity‖ of a proposal but, rather, ―where social problems are concerned, the question is more likely to be [the] soundness or unsoundness‖ of an idea. That is, he claims (albeit without evidence) that most of the issues citizens struggle with in modern, complex democracies involve making judgments based on incomplete information and biased opinions where ―truths‖ might never be known and falsehoods never proven. It 53 system suggests is that the First Amendment‘s speech and press clause might work together to create a set of circumstances yielding freedom of expression. It is in this understanding of the press clause—as an institutional companion to, and enabler of, free speech—that we might best understand when and why the press as enjoyed and asserted its autonomy, and how such autonomy has been interpreted and limited by the

Supreme Court (in both its members‘ extra-judicial writings and its bench opinions).

4. The Press Clause, Press Autonomy and the Press as Instruments of Free

Speech

Equipped with these rationales for valuing free speech—the argument from truth and the argument from democracy, and criticisms thereof—we might now ask what role the press might play as an institutional actor in the system of free expression. That is, what are differences between the First Amendment‘s speech and press clauses, when has the Supreme Court recognized the press, and what might this mean for understanding press autonomy?

For most of U.S. constitutional history, the speech and press clauses were used interchangeably and largely without distinction (Anderson, 1983, 2002). In fact, it was only in the 1920s that the Supreme Court began to consider the press as a potentially distinct institutional actor with democratic functions. Specifically, the

Court heard but, largely rejected, the press‘s arguments that it should be free to criticize the government and promote dissenting opinions (issuing decisions that largely upheld the Espionage Act of 1917 and the Sedition Act of 1918). It was not

is thus important to design a system of free expression that moves beyond Millian notions of truth and error, no matter how appealing such an ideal might seem. 54 until Near v. Minnesota (1931) that the Court overturned a Minneapolis state injunction prohibiting the press from publishing articles critical of police officers, finding that the state‘s prior restraint of the newspaper was unconstitutional (Garry,

1994: 15). This was an early case in what would become a series of Supreme Court cases in the 1960s and 1970s focused on deciding exactly what, if any, distinct rights or privileges the press enjoys.

Parallel to these court cases was a thread of developing legal theory arguing that the First Amendment could best be understood as a structural provision concerned with the conditions under which citizens express and hear speech. The core idea is that the press‘s autonomy is not to be defended on its own as a constitutional privilege guaranteed by the press clause but, rather, that the press and speech clauses work together to create a system of expression that is only worth privileging insofar as it supports democratic aims. As Fiss (1986: 1411) explains, the

key to fulfilling the ultimate purposes of the first amendment is not [press] autonomy, which as a most uncertain or double-edged relationship to public debate, but rather the actual effect of the broadcast: On the whole does it enrich the public debate? Speech is protected when (and only when) it does, and precisely because it does, not because it is an exercise of autonomy. In fact, autonomy adds nothing and if need be, might have to be sacrificed, to make certain that the public debate is sufficiently rich to permit true collective self-determination.

Essentially, deciding whether the press clause represents a significant and meaningful concept distinct from, but complementary to, the speech clause requires accepting the democratic value of the dualistic autonomy described here. That is, if you accept idea that self-government entails not only freedom from interference by others (letting you pursue your own interests, speech and actions on the assumption that you are largely independent from others) but also that freedom depends upon what others contribute

55 to the pursuit of truth and democratic government (enabling you to encounter information, perspectives and proposals that you could not have discovered or created on your own), then the press clause represents a normative ideal. The ―press‖— however it might be defined at any moment in history—is an institutional exemplar of a free speech system in which citizens have both the right to speak and the right to hear.

The central problem of understanding the press clause and press autonomy then shifts to a pragmatic one that asks: to what extent does the press institutionally facilitate the ideal of dualistic autonomy? Press autonomy might best be defended not through trying to reconstruct constitutional history—attempting to discern what the framers of the constitution may have meant by ―the press‖9—but, rather, by examining the contemporary conditions and circumstances under which the press performs its ideal function of enabling dualistic autonomy10. The remainder of this section focuses on explicating an institutional theory of the press clause, and a discussion of law of editorial judgment as a process- and product-oriented idea that considers both how the press does its work and what the press contributes to democratic life.

9 See Scalia (1997) for a general theory of constitutional interpretation in which the goal is to determine the original intent of the constitution and what its text meant at the time it and its amendments were ratified. See Levy (1960, 1985) for controversial histories of the press clause, including a claim that the framers only intended to prevent prior restraint of the press; and Anderson (1983) for a good overview of debates in press clause history. 10 In contrast to Scalia‘s (1997) originalist approach, Justice Breyer describes a ―pragmatic‖ approach to constitutional interpretation in which he says that judges use five elements in interpreting the constitution—statue language, jurisprudential history, legal traditions, case law precedents, constitutional purpose, and pragmatic consequences—but that, ideally, they should ground their decisions in understanding the overall purpose of the constitution and interpret that purpose in contemporary terms (Breyer, 2005). 56

4.A. The Institutional Press

The idea of an institutional press has never been explicitly stated in U.S. Supreme

Court jurisprudence. Indeed, the court has historically been reluctant to distinguish journalists or news organization owners from other citizens or give them special privileges, principally because of a fundamental notion that all constitutional rights are enjoyed by all citizens. To justify special treatment for the press by using the press clause would be establish a two-tiered constitution that would then require some kind of constitutional license that privileged journalists over others. The Court‘s has historically protect speech with little regard to where it came from. When it has limited speech rights, it has done so on contextual bases, not on identity bases that gave some individuals rights but not others.

The Court has historically (and usually implicitly) used two broad, intertwined logics to judge and regulate speech. First, it has considered whether the speech in question is considered ―low-value‖ (e.g., types of pornography, sexual and racial harassment, threats) or ―high-value‖ (e.g., political and deliberative discussions on current issues or entertainment considered to be of public value). Second, the Court has evaluated whether the state‘s restrictions on speech are: content-neutral (in which the ―content of the expression is utterly irrelevant to whether the speech is restricted,‖ as in a decision to ban all speech on billboards regardless of who purchased the space or what they said); content-based (limitations that consider the type or category of speech, e.g., banning all political speech in a particular area); or viewpoint-based

(restrictions that actively take sides in a debate, limiting the speech of those who disagree with the government‘s position). The Court rarely accepts viewpoint-based

57 restrictions, periodically approves content-based restrictions, and has been most accepting of ―time, manner, and place‖ speech restrictions that are blind to who the speaker is or what she is saying (Sunstein, 1994: 1-23; 167-208).

Within these logics, the Court has readily differentiated among types of speech, distinguishing ―incitement from advocacy, commercial speech from noncommercial speech, obscenity from indecency, public interest speech from personal interest speech, public forums from nonpublic forums and from ‗designated‘ public forums‖

(Schauer, 2005: 1263) however, with the exception of the broadcasting industry

(considered below in the review of Supreme Court case law), it has historically not explicitly considered a speaker‘s institutional identity. It has traditionally refused to organize its ruling along organizational or institutional lines, avoiding decisions, for example, in which it defines the press or journalists. Schauer identifies three reasons for this reluctance. First, justices aim to make decisions according to legal distinctions rather than social theories: ―What distinguishes categories like viewpoint discrimination, content regulation, public forum, and prior restraint from categories like universities, libraries, elections, and the press is that the former exist in the First

Amendment but the latter exist in the world.‖ (Schauer, 2005: 1265) Second, courts are traditionally conservative entities focused on with maintaining stability and avoiding radical change movements. And third, there is a concern that if Courts consider contemporary circumstances and conditions too closely, the First

Amendment may become a weaker constitutional provision unable to be useful across different eras and contexts.

By implicitly accepting these conservative rationales and refusing to make distinctions among institutions, the Supreme Court may actually harm free speech. If the result of institutionally blind decisions means less protection for the kind of speech that is essential for self-government (both individuals‘ rights to speak and publics‘ rights to hear), then there needs to be a way of letting institutional differences enter into First Amendment decisions in general, and press clause cases specifically. As

Schauer (2005: 1272) argues, there is a danger that institutional blindness may result in what he calls ―institutional compression‖ in which actors are artificially separated from the contexts, norms, principles and incentives of their circumstances. The First

Amendment will become less powerful because of the Court‘s reluctance to enter into the messiness of institutional action: A ―Supreme Court [that is] unwilling to distinguish among the lone pamphleteer, the blogger, and the full-time reporter for the

New York Times is far less likely to grant special privileges to pamphleteers and bloggers than it is, as it has, to grant privileges to no one.‖

What is required is a kind of institutional middle ground in which the Court can anchor its decisions in principles that affirm a broad interpretation of the First

Amendment (making room for consistency with unanticipated contexts) while staying relevant and timely (providing guidance to citizens and governments about what conditions of speech might be more or less encouraged and protected). Schauer

(2005: 1275) sketches a two-tiered system in which the Court might consider the institutional nature of speech:

We first locate some value that the First Amendment treats, or should treat, as particularly important. Then we investigate whether that value is situated significantly within and thus disproportionately served by some existing social institution whose identity and boundaries are at least moderately identifiable. If so, then we might develop a kind of second-order test. If there

is a reporter‘s privilege, for example, we might ask not whether this exercise of the privilege serves primary First Amendment purposes, but instead simply whether the person claiming the privilege is a reporter. Obviously, defining the category of people who receive the privilege will be based both on the reasons for having the privilege and the reasons for locating it in a particular institution, but the case-by-case inquiry will largely consist of applying the rule, rather than applying the reasons lying behind the rule directly to individual cases.

This model is correct to ground this institutionally sensitive model in First

Amendment values, but his focus on defining actors might better be replaced with an emphasis on articulating actions—e.g., although there is no federal shield law, the

Court might better consider how to connect First Amendment values to ―reporting‖ rather than ―reporters‖. Such an action-oriented grounding of press clause questions might prevent the court from becoming entangled in the subtleties of ever-changing technologies—e.g., are bloggers and Twitter users both journalists and, if so, how does blogging differ from tweeting?—and allow it instead to focus on the practices or standards of reporting that are likely to produce the kind of speech envisioned by an affirmative interpretation of the press clause.

The aim here is not to create a First Amendment that is unpredictable and too dynamic for anyone person or institution to be assured of its protections. An unstable

First Amendment is as useless as an antiquated one. Rather, the goal here is to articulate a way in which the Court might pay attention to the ―contingent, empirical, institutional‖ terms under which contemporary speech is produced, and offer an institutional model of the press clause. As Justice Hughes wrote in a 1938 opinion,

―[t]he liberty of the press is not confined to newspapers and periodicals. It necessarily embraces pamphlets and leaflets ... the press in its historic connotation comprehends every sort of publication which affords a vehicle of information and opinion.‖ (cited

60 in Horowitz, 2006: 45) The question now is two-fold: how can the Court continue to recognize the multiple ways in which speech is produced and disseminated; and how might the Court depart from considering particular forms of publication (blogs versus newspapers versus pamphlets) and instead establish standards for evaluating the value of practices of publication? Answers to these questions might be found in two locations: a review of select Supreme Court cases that have—largely without explicitly invoking the press clause—identified what it considers to be valuable roles for the press to play in democratic life; and in what Bezanson (1999) calls a

―developing law of editorial judgment.‖

4.B. Supreme Court Press Clause Decisions

While the Supreme Court has never broadly or formally recognized the press as a distinct, identifiable entity with special constitutional privileges or protections or used the press clause exclusively as the basis for a decision (Anderson, 1983; Bollinger,

2010), it has, at various times and for different reasons, assumed that something called

―the press‖ exists and that its existence serves purposes protected by the Constitution.

Although not part of a Supreme Court decision, the most famous opinion offered by a justice on the press‘s identity has come from Justice Potter Stewart. In a

1974 address at Yale Law School, in the wake of the Watergate scandal and a general celebration of the press‘s investigative, watchdog role, Stewart (1975: 60) argued that the press and speech clauses are meaningfully distinct because ―the Free Press guarantee is, in essence, a structural provision of the Constitution.‖ Without such an institutional guarantee, Stewart argued, the press could fall prey to one of two

61 powerful forces: market forces that purport to offer a ―neutral forum for debate‖ (60) but that actually exclude voices who lack the financial capital to purchase media space and the public‘s attention; and government forces that could require newspapers, television networks and magazines to ―promote contemporary government policy or current notions of social justice.‖ (63) Although Stewart could not predict the changes in the nature of communications media (e.g., most notably cable networks and the

Internet), his view of an institutional press is still relevant today. Although he expressed his opinions in a speech and not a ruling, Stewart‘s role as a Supreme Court justice11—making arguments in relation to his contemporaries and precedents— suggests more closely examining how Supreme Court decisions may implicitly signal a constitutionally recognized institutional role for the press.

These press clause decisions essentially all grapple with a central concern. If the Court recognizes press clause privileges distinct from those guaranteed by the speech clause, begs a natural follow-up question: namely, what then constitutes ―the press‖ and why does the constitution offer special rights to members of that profession? How could the press be defined and (even if such a definition could be agreed upon), why does this profession enjoy special rights not available to other residents of the United States? The Court runs the risk of creating a two-tiered First

Amendment in which not all clauses apply equally to all individuals12. It was this

11 Nine years before Justice Potter‘s more famous speech, another justice, William Brennan (1965), gave somewhat similar public remarks at Brown University as part of the Alexander Meiklejohn Lecture Series. In his speech, he seemed to agree with Meiklejohn‘s argument that the state is responsible for providing to publics conditions that enable individual self-government and, in a discussion of the then recent New York Time v Sullivan case, he stated that so long as a free press provides speech of ―redeeming social value,‖ it plays a valuable constitutional role in securing individual self-government. 12 In the same public remarks in which he argued for a structural understanding of the press, Justice Stewart (1975: 60) answered this question directly by stating that ―the Free Press Clause extends 62 exact set of questions that Chief Justice Burger had in mind when wrote ―the very task of including some entities within the ‗institutional press‘ while excluding others … is reminiscent of the abhorred licensing system,‖ which ―the First Amendment was intended to ban.‖ (cited in Bollinger [2010: 10)

The modern Court‘s case law are concisely and thoughtfully summarized by

Bollinger (1991, 2010), and it is his typology that I adopt here. He identifies three broad logics in Supreme Court press clause jurisprudence: protecting publics against censorship; no right of access to information; and improving the press through regulation. By offering this typology, I do not mean to argue that the Court defines the press in these terms, that this is how it sees the press clause (indeed, the press clause is rarely explicitly mentioned in Court decisions), or that I think this is how the

Court should define the press. Rather, these logics—culled by Bollinger from a body of complex decisions—suggest that this is how the Court has understood the idea of an institutional press.

These cases hint at how it has described and protected the general concept of a press that serves a democratic function, without carving out specific privileges for a particular definition of the press. I briefly summarize each of Bollinger‘s principles below, integrating, as he does, Supreme Court cases that illustrate these principles.

protection to an institution. The publishing business is, in short, the only organized private business that is given explicit constitutional protection.‖ Stewart, though, elides between two important concepts: the press as an institution and the press as a publishing business. The idea of an institution is explored in greater detail in Chapter Three of this dissertation but, for now, it is simply worth noting that a business is not the same as an institution: a business has a narrow focus on the production and distribution of commodified goods, while an institution might best be thought of as a system of associations in which power, sanction, legitimation, reproduction and taken-for-granted beliefs (DiMaggio & Powell, 1991) are negotiated among a variety of actors, organizations, professions, and social contexts. 63

4.B.(i). Protecting Publics Against Censorship

There is a set of Supreme Court cases largely focused on ensuring that speakers are not subject to unreasonable prior constraint on their expressions. In New York Times

Co v. Sullivan (1964), the Court held—in an opinion written by Justice Brennan a year before he made the speech ostensibly supporting Meiklejohn‘s view of the First

Amendment—that even though Sullivan‘s personal reputation may have been damaged by an error-filled and critical advertisement run by Times, Sullivan was a public official who had voluntarily entered the public sphere and, as such, the public had an overriding interest in hearing such criticisms. Unless an individual or news organization had ―spoken defamatory falsehoods knowingly or in reckless disregard for the truth‖ (Bollinger, 1991:8) against public officials—and even unelected public figures, as the Gertz v. Robert Welch Inc. (1974) ruling later found—people with public identities had less claim to libel protection than other citizens. The core idea, the Court ruled, is that false statements ―must be protected if the freedoms of expression are to have the breathing space that they need to survive‖ (Bollinger, 2010:

16) and that the press is ―entitled to act on the assumption that public officials and public figures have voluntarily exposed themselves to increased risk of injury from defamatory falsehoods concerning them.‖ (18) Essentially, the Court found the press—acting in its role as a producer and distributor of potentially valuable public information—could act with less fear of being found libelous in its publishing.

The Court adopted a similar view of the press as a provider of valuable public information when it ruled in Cox Broadcasting Corp v. Cohen (1975) that as long as

―newspaper lawfully obtains truthful information about a matter of public significance

64 then state officials may not constitutionally punish publication of the information, absent a need to further a state interest of the highest order.‖ Two elements of this ruling are important to note: first that the court granted a newspaper this privilege of publication (not any other organization, type of media, or individual); second, the

Court recognized that there are issues of public significance, separate from state interests. The Court was essentially using similar logic as it had used in the Pentagon

Papers case (New York Times Co. v. United States, 1971) when it found that the federal government had failed to convince the Court that it should prevent the New

York Times and Washington Post from publishing a classified Department of Defense report. The government had not met the rigorous constitutional test for restraining expression, finding that the public‘s interest in knowing the newspapers‘ findings overruled the government‘s claims that state secrets would be revealed.

Finally, in Nebraska Press Association v. Stuart (1976), the Court found unconstitutional a lower judge‘s ban on press reports of a court case. Sympathizing with the judge‘s concern that media scrutiny would damage the defendant‘s right to a fair trial—while simultaneously admonishing him for not exploring alternative arrangements for protecting the defendant—the Court essentially ruled that part of what might make a trial fair is public scrutiny. That is, if a court, on the public‘s behalf, is trying an individual for a crime, the public—in this case, through the actions of the press—has a protected interest in learning about how the judge is administering justice. While this ruling, like others, identifies no explicit press right (i.e., it does not derive its finding from the press clause), it effectively finds that the press can act to ensure public oversight of and participation in the administration of justice. The

Court, then, implicitly sees the press not as a quasi-branch of government13, but as an institution in its own right, grounded ideally in a logic of public oversight – invoking what Blasi (1977) calls the ―checking value‖ of the First Amendment.

In this body of cases, the Court seems to give the press an institutional role, one that protects it from being censored precisely because it is presumed to be acting on behalf of the public, giving individuals the information they need to have to be self- governing citizens.

4.B.(ii). No Right of Access to Information

Although the Court seemed to adopt the principle that the press provides valuable information to citizens, it has largely refused to grant the press any special privileges as it works to seek out that information. That is, the press enjoys no special rights as it goes about its reporting work.

Although there are state laws that shield reporters from being compelled to testify in court and reveal their sources, in Branzburg v. Hayes (1972) (actually a collection of cases anchored around this particular decision) the Supreme Court ruled that reporters have no federal ―constitutional right not to reveal the identities of their sources in grand jury or other criminal proceedings.‖

Reporters argued that they needed a ―shield‖ from compelled testimony to prevent two kinds of speech chilling, and that motivations for testifying were more often politically motivated and had little to do with the facts of a case. First, they claimed, ―news reporting consists of a succession of events‖ (Anderson, 1973: 553)

13 Although press rights may be implicitly asserted by invoking the public as its client and patron, see Chapter Three of this dissertation (especially section 4D) for a discussion of how the press asserts its autonomy, in part, by strategically distancing itself from publics. 66 that are often sensitive and confidential; reporters would self-censor this process and produce less powerful stories if they knew that they could be called upon to testify about any aspect of their work at any time. Second, if sources knew that their conversations with reporters were potentially matters of public record they would be far less willing to revealing sensitive information that might be of critical interest to the public (Murasky, 1974: 851-862). Finally, the reporters observed that they were most often called to testify in cases when the government wanted to know the identities of suspected criminals (effectively deputizing reporters), when a reporter‘s work was relevant to a civil case (forcing them to take a litigant‘s side), when the case in question is more focused on a third party (drawing connections beyond those of the reporter‘s story), or when the reporter‘s work relates to a governmental matter

(making the reporter an unwilling member of legislative committee) (Guest &

Stanzler, 1969). The press needed a testimonial shield—a ―defensive‖ right that Baker

(1989: 246) likens to freedom from searches and seizures—to prevent its reporting from being muted, its sources‘ speech from being chilled, and reporters‘ work from being implicated in a set of activities that had little to do with their professional goals and reputations.

Two years later, in Pell v. Procunier (1974), the Court similarly found that reporters had no right of access to prisoners who journalists thought may be being mistreated: ―It is one thing to say that … the government cannot restrain the publication of news [but] [i]t is quite another thing to suggest that the Constitution imposes upon government the affirmative duty to make available to journalists sources of information not available to members of the public generally.‖ In both cases, the

Court essentially ruled that the state‘s interest in protecting the public through the criminal justice system overrides the press‘s interest in protecting the integrity of its reporting practices, regardless of whether the press‘s root interest is to inform the public, or what the state‘s intention is in compelling the reporter‘s testimony14.

4.B.(iii). Improving the Press Through Regulation

Finally, there is a series of decisions in which the Court has, under the auspices of encouraging the kind of system of expression deemed beneficial for democratic societies, found constitutional regulations that attempt to structure the press. As

Bollinger (2010: 29) notes, the trend began with the Communications Act (1934) that allowed the government to ―take and to keep control over channels of ‗radio transmission‘; and to provide for the use of such channels, but not the ownership thereof, by persons for limited periods of time.‖ Essentially, the state asserted its right to manage the technical systems of mass communication, recognizing that it needed to, on the public‘s behalf, ensure that the means of speaking and hearing opinions were structured in the public‘s interest.

The underlying logic of this act—a ―Fairness Doctrine‖ that aimed to provide equal speaking time and reasonable access to airwaves for public office candidates— was first put to the test in the landmark case Red Lion Broadcasting Co. v. FCC

(1969). In this case, the Court affirmed the Fairness Doctrine finding that, because radio and television spectrum is limited and of great public value, the government has

14 As Bollinger (2010: 27-28) notes, though, there is one case that seems to be an exception to cases that limit the press‘s right to perform reporting practices it deems to be of public value. In Richmond Newspapers Inc. v. Virginia (1980) the Court ruled that the public and the press have a right to attend criminal trials because such access may create ―greater public confidence in the criminal justice system and better decisions in trials.‖ 68 a constitutional interest in protecting its allocation and administration on the public‘s behalf. Most critically for advocates of an affirmative interpretation of the First

Amendment, the Court grounded its ruling in an idea that there is a public right to hear that overrides a broadcaster‘s private right to expression: ―what is ‗crucial‘ in this environment ‗is the right of the public to receive suitable access to social, political,

[a]esthetic, moral, and other ideas and experiences.‖

The Court‘s affirmative interpretation of the First Amendment, though, was short-lived and, it emerged, only applicable to media that had limited spectra. Five years later, in Miami Herald Publishing Co. v. Tornillo (1974), the Court found that no ―right of reply‖ existed in newspapers and that editors and newspaper owners were free to decide what would be printed in their publications. These two cases—Red Lion and Miami Herald—are a curious pairing. While Red Lion ostensibly identifies the press as playing an institutional role (guaranteeing the public‘s right to hear), this responsibility seems not to be grounded in an affirmative interpretation of the First

Amendment or the state‘s responsibility to ensure the conditions of public hearing.

The court recognized no special role for journalism as a profession that might enact publicly valuable principles that span forms of media, and instead focused on the state‘s right to regulate media based on technological considerations.

The Court confirmed this logic when it considered the structural properties of the cable industry in Turner Broadcasting System, Inc. v. FCC (1994), and whether cable providers could be compelled to carry channels. The Court found that because the technology offered virtually limitless bandwidth—there existed no technical reason to regulate spectrum—there was no rationale for government regulation of

69 cable the same way Red Lion allowed for broadcast regulation. The only potential weakness of the cable industry, Justice Kennedy offered, was its potential to create economic monopolies that might interfere with the functioning of free speech markets.

The risk of dysfunctional markets was considered an insufficient reason to regulate, essentially ending any hope that Red Lion‘s affirmative interpretation of the First

Amendment might survive in other mass communication media technologies

(Bollinger, 2010: 38). (The Court further confirmed this logic in Reno v. American

Civil Liberties Union [1997] when it struck down parts of the Communications

Decency Act [1996] that attempted to ban ―indecent‖ material on the Internet. It ruled that since the Internet ―provides relatively unlimited, low-cost capacity for communication of all kinds,‖ no spectrum scarcity rationale applied and, since the medium was not considered ―invasive,‖ people were free not to expose themselves to

Internet content. It found First Amendment rationale to regulate Internet content15.)

Another element of the press‘s relationship to state regulation is the extent to which the government funds public broadcasters, and whether such funding carries with it the ability to influence content. In one notable case, the Court found unconstitutional the part of the Public Broadcasting Act (1967) (the legislation founding the Corporation for Public Broadcasting) that prohibited publicly funded

15 Although it is true that technological scarcity makes little sense in relation to the Internet, it might be argued, especially in light of recent research on the deleterious effects of multi-tasking (e.g., Ophir et al. [2009]) that there is a scarcity of public attention. That is, it is not technological spectrum that needs to be managed but, perhaps, social attention to issues considered to be of public concern. I am not arguing that people be forced to read, watch or attend to media. Rather, there may be a rationale for considering the socio-technical conditions that make it more or less likely that people will encounter certain perspectives over others. That is, even when they ostensibly choose for themselves what media to attend to, their media preferences are influenced by forces beyond their control (e.g., search engine rankings and other algorithmically structured mass behaviors) that, practically, make it more or less easy to encounter the kind of diverse perspectives required for self-government. See Chapter 4 of this dissertation for a more detailed discussion of the role of infrastructure in mediating people‘s consumption of news content. 70 stations from editorializing. The Court ruled in FCC v. League of Women Voters

(1984) that broadcasters—regardless of their funding sources—could not be prohibited from ―speaking out on public issues‖; to do so would be a ―substantial abridgement of important journalistic freedoms which the First Amendment protects.‖ Again, we see the Court—without explicitly invoking the press clause—implying that the press‘s freedom to function is intimately linked with the public‘s right to hear speech, and that these freedoms and rights are not just the domain of individual decision-making but of institutional structures.

Most broadly, Bezanson (1999) argues that court decisions (at both federal and state levels) about journalism and the press might best be thought of not as a specific cannon that defines what ―the press‖ is or what the press clause means but as a

―developing law of editorial judgment.‖ Through an extensive review of decisions involving the press and its claims to editorial privileges, Bezanson identifies a set of considerations that Courts seem to use in structuring their opinions: the subjective intent of editors (e.g., when deciding whether news organizations are guilty of libel); the objective result or effect of a particular decision to publish (e.g., favoring news organizations that can demonstrate a controversial story‘s ―newsworthiness‖ in terms of the public‘s interest in its information or genre of reporting, and not business concerns); the fact versus opinion nature of a story (e.g., privileging decisions to publish stories containing verifiable, truth-oriented reporting over subjective, opinion- based commentary); the truth versus falsity feature of a story (e.g., favoring stories that later prove to be true versus those that are found to have factual errors); the process of how journalists prepared a story (e.g., looking positively on publishing

71 decisions that appeared to involve a rigorous editorial process that considered the controversies in question). Bezanson offers hints about how Courts may understand the press‘s function in the absence of their close consideration of the press clause and their refusal to clearly define and delimit the press. His typology emphasizes that the

Court seems to consider both the purpose of journalism—its function in society and the effects it can have—and the process of journalism—how its professionals conduct themselves, grapple with controversial decisions, and publish content.

5. Contemporary Speech Infrastructure, The First Amendment and The Press

Given this history of the Court‘s engagement with press freedom, Schauer‘s call for a more institutional understanding of the press, and the increasingly distributed and multi-sited nature of online news production16, what contemporary legal concepts might be used to understand press autonomy? That is, what ideas exist in legal theory today that might help us understand press autonomy in terms of online speech and a diversity of networked actors that include, for example, mainstream professional journalists, unaffiliated self-regulating bloggers, indexing technologies that sort and rank speech, publics that participate within a variety of news forums. What broad legal principles might support a modern, ―audience-centered theory of free speech‖

(Tushnet, 2008: 101) in which a dualistic sense of individual autonomy—an ability to be both free from unreasonable restrictions on speech and embedded within rich public discourses that can entail personal insight and learning—underpin an

16 See Chapter Three of this dissertation for a history of social production of news, and Chapter Four of this dissertation for a more in-depth discussion of the current conditions of online news production and its distributed, networked nature. 72 interpretation of the First Amendment in which the press‘s right to gather and report news is considered integral to publics‘ rights to hear?

These legal principles are intertwined with how the law and courts understand technologies of online speech. There are two core legal concepts critical for understanding press freedom in online speech environments: an intermediary, and a selection judgment.

Most clearly articulated in section 230 of the U.S. Communications Decency

Act (1996), an intermediary can be defined as an ―interactive computer service‖ that shall not ―be treated as the publisher or speaker of any information provided by another information content provider‖ and that shall not ―be held liable on account of any action voluntarily taken in good faith to restrict access to or availability of material that the provider or user considers to be obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable, whether or not such material is constitutionally protected.‖ That is, Yahoo! may not be held responsible as the speaker if a user posts defamatory speech on one of their sites and they are within their rights to removes such material if they feel that it is unacceptable, even if such removal violates a user‘s constitutional rights. Essentially, it is not intermediaries who are responsible for ensuring individuals‘ speech rights – they are allowed to use their own judgment in determining the content that appears on their sites.

This dualistic role of intermediaries—as a provider of others‘ content and as a judge of that content‘s value—makes them seem a lot like news organizations: they carry others‘ speech, they evaluate that speech in relation to standards (e.g., through the use of pre-hoc filtering technologies that prevent speech from being published or

73 post-hoc decisions that remove speech from sites), and they act as forums or virtual street corners where others come to hear speech. Missing from this characterization, though, is any sense of stated editorial judgment—a core feature of journalistic traditions—in which reporters (ideally at least) operate under a set of socially and professionally regulated norms that heavily influence their decisions to publish. For online intermediaries, there is what Tushnet (2008) calls ―power without responsibility‖ as ―interactive computer services‖ like Yahoo!, Google, Technorati,

Facebook and Twitter have the ability to make decisions about what content to publish, but without the normative constraints that journalists place on themselves, or the constitutional restrictions the Court puts on the state (recall Sunstein‘s [1994: 1-23;

167-208] discussion of time-manner-place regulation regime). This idea of an intermediary is key to understanding what kind of future might lie ahead for constitutionally protected press freedom as, increasingly, the work of journalism becomes integrated with online information systems with the power to both carry and suppress speech.

What, exactly, though are the nature of the decisions these intermediaries make, and how might the decisions of information service providers—many of which are automated within code and algorithms that, once designed and executed, can process and restrict speech almost instantaneously—be different from those of journalism? Essentially, although Lessig (1999) argues that code itself is, or should be, considered speech, what should we make of code that operates on others‘ speech – that makes some speech more or less likely to be heard, acting not as one voice among

74 others within an equitable public forum, but as source of rules for how that public forum practically behaves and what people effectively hear?

There is some help here in the second core concept likely to influence the press‘s autonomy in online speech environments: the selection judgment. As defined by Bezanson (2003: 983), selection judgments ―involve the appropriation or selection of speech originally created elsewhere (by another) and the secondary deployment of that material in another context by a person or entity different from the original creator.‖ For example, consider reporters who take the text of a presidential speech, an interview with a source, or an industry report and selectively use, with citation, that text as part of a story. Bezanson argues that such selection judgments be considered protected speech under the First Amendment because they are instances in which ―(1) a person makes a speech selection for purposes of expression; (2) the selected speech is adopted as an expression of the speech selector‘s own ideas or opinions; and (3) the selected expression is transformed in meaning or significance by the act of selection.‖

(2003: 983) Essentially, if the First Amendment‘s primary goal is to protect individuals wanting to create speech and publics wanting to hear ideas, then First

Amendment must make a distinction between the intentional distribution and transformation of ideas for an audience, versus accidental or non-transformative speaking to no one in particular.

Although Bezanson‘s analysis suffers from a lack of clarity—e.g., he never really offers compelling instances of accidental speech and his examples of speakers without audiences are open to much debate—his core concept of selection judgment is important in the context of online speech for several reasons. First, is the software

75 programmer who creates an algorithm for filtering speech—e.g., Google offers a default ―safe‖ setting that few users change that filters out search results deemed to be offensive without explaining what it considers offensive to mean or when its definition might change—engaging in an intentional selection judgment when she writes the code or when the code executes? Who is doing the selection, the person or the machine? If it is the person, then Google looks a lot like a newspaper and the programmer is an editor making a purposeful selection judgment of the kind that

Bezanson would give constitutional protection. If, however, it is the machine who is making the selection judgment—and doing so without regard to the unpredictable power of social context to create meaningful exceptions to selection rules—then such an edit would fail Bezanson‘s test, since the machine itself can have no sentient intentionality. If we are to apply Bezanson‘s notion of selection judgment to the intermediaries protected by the Communications Decency Act, we must make some determination about when automated selections are meaningful and distinct, what aspects of such judgments require human intervention, and whether code that acts on other speech should itself be considered speech.

Second, despite Bezanson‘s claim that his selection judgment model is meant to improve the quality of constitutionally protected speech, his test may actually undermine a public‘s right to hear. That is, the best system for protecting the public‘s right to hear might be one that enables publics themselves to determine the difference between purposeful speaking to for audiences and accidental speaking to no one in particular. For example, in an online news context, must journalists have a model of their audiences in mind for their speech to be protected? What of evidence that

76 journalists often write for each other, their competitors, their friends and family, or that they often shy away from imagining their readers?17 Imagine a diary-like blog whose log files show that no one other than the writer has ever visited – is this speech only worthy of constitutional protection once someone encounters the blog? What if the blogger had never intended anyone to find it and actually does not want any readers? Essentially, there may be instances in which the audience for a message is unknown at the time and only later discovered: consider the blogger who does not know whether others might find her ideas are valuable but publishes online to see if a search engine indexes her and readers find the blog. Denying the public the right to hear that kind of speech—or diminishing its constitutional value in relation to more purposeful speaking—may effectively harm the public‘s right to use systems for listening to discover and benefit from ideas in which speakers did not think anyone would be interested. For example, strictly applying Bezason‘s test may have a kind of

―meta-chilling effect‖ by signaling to speakers that they should not engage in producing speech or experimenting with new forms of speaking for which there currently exist no audiences (e.g., in addition to the blogger who is later discovered, consider the investigative journalist who wants to pursue a topic for which there seems to be no public audience but about which the public would benefit from hearing).

The underlying point behind this discussion of intermediaries and selection judgments is the idea that First Amendment law‘s notion of speech and its value to publics may need to be rethought. In online contexts, information is increasingly mediated by actors with the power to structure the speech that is available in

17 See Chapter Three of this dissertation for more discussion of studies of who journalists imagine as their audiences. 77 seemingly public forums automatically, making it more or less likely that some ideas are ever heard by many people. But these same actors may lack the professional norms that—however imperfect or unevenly applied—guide journalists and be unsure of when or why to make some selection judgments over others. I offer no remedy for these tensions but, simply aim to point out that the ideas of intermediaries, selectivity and automaticity will be key to understanding what kind of speech can and should be protected in future First Amendment jurisprudence and legal theory.

6. Conclusion

The purpose of this chapter has been four-fold. First, I aimed to ground the idea of autonomy in a core, dualistic understanding of democratic liberty, one that values both a freedom from unwarranted restrictions on speech and a likelihood of engaging in the kind of expressive social relations that make possible individual enlightenment.

Essentially, individual autonomy certainly requires separation from forces of power and control that aim to limit personal freedom and facility and engagement with people and cultures that can provide the new information and perspectives individuals are unlikely to generate in isolation.

Second, I discuss how this autonomy requires a particular kind of system of speech. Essentially, for individuals to be self-governing—to know what it is they need to know to make informed and thoughtful decisions about how to live their own lives and interact with others—they need to be able to discover truth and they need to know that their democracy entails being in communion with others who also enjoy free exposure to quality speech.

Third, using this dualistic sense of autonomy and theory of free speech, I discuss the kind of institutional and organizational conditions that might enable the production and protection of free speech that liberty requires. In its ideal role, the press is a public-serving institution that makes it possible for individuals to hear ideas and encounter perspectives that, left alone, no single person could reasonably discover for themselves. In this way, the press is one of many institutions that ideally enables the dualistic notion of autonomy: offering, through a system of professionally regulated journalistic norms, it is an instrument of free speech. It is a way for people both to see their own perspectives represented in public discourse and to encounter new ideas of those they could not talk with directly.

Finally, I discuss in conceptual terms using Supreme Court jurisprudence and an emerging body of legal literature, the kind of institutional press that might enjoy

First Amendment protection. Despite the Court‘s lack of rulings that clearly and squarely address the contemporary meaning of the press clause—it instead has carved out a tenuous, implicit identity for the press through various interpretations of the speech clause—there is, or could be, an more distributed, modern understanding of the press that preserves the ideal functions of the press, while worrying less about who qualifies as a journalist or what defines a news organization. I conclude this section with some thoughts on how contemporary, online news production will likely need to grapple with the role of information intermediaries, and the exact nature of selection judgments.

Unlike the next chapter in which I locate the press amongst a set of socio- historical forces and Chapter 4 in which I discuss the contemporary pressures on

79 online news production, the aim of this chapter has been to take a normative stance on what the press—in its broadest and most conceptual sense—should be about.

Essentially, our individual autonomies—our rights to be left alone and our rights to be enlightened by others—depend upon us—and the publics to which we belong—being able to speak and hear. To the extent that the press enables such individual autonomy, so should it be autonomous. That is, I am purposefully employing two senses of autonomy: an individual one rooted in a right to democratic self-governance that requires both separation from and dependency upon others; and an institutional one in which the press earns its rights to be distinct from and reliant upon a variety of actors

(including state, markets, consumers, publics, and technologists) in so far as it works to create the kind of system of free expression that individuals need to realize their autonomy. If asked what I mean by the press as an ideal—distinct from what I mean by the press as a sociological or technological construct that I discuss in the following two chapters—I answer that the press should be autonomous in so far as it enables an affirmative interpretation of the First Amendment in which individuals enjoy the right to speak, and publics have the freedom to listen.

CHAPTER THREE:

Locating the Press and Tracing Its Socio-Professional Autonomy

1. Introduction

The aim of this chapter is to locate the press so that the dynamics governing its autonomy might be better understood. I say ―locate‖—instead of define, defend, or bound—the press to emphasize that the goal here is to situate the press amongst a set of influences and historical forces that have shaped its practices and the meanings of its products. More specifically, this chapter locates the press by tracing claims associated with its autonomy to argue that we might best understand what the press is—or should be—by appreciating when and why various actors (researchers, producers and consumers of news) have asserted or advocated for the press‘s independence.

The aim here is to offer a framework, grounded in a sociological study of fields, that sees the U.S. press as an inherently difficult institution to pin down or understand in terms of any single tradition, actor, organization, technology or norm. It is a profession that requires no license; a profit-making private enterprise awkwardly restrained by a public mission; a democratic institution that relies upon, shuns and investigates state influences; a way of both entertaining people—fulfilling desires, reinforcing stereotypes, and perpetuating ignorances—and enlightening them—telling them things they didn‘t know they needed to know, and offering opportunities to create communities across distances too great to experience face-to-face.

My aim here is to show how the press is a field, and how the autonomy of this field is actually a set of ongoing, strategic negotiations among a number of different actors. Essentially, the press‘s autonomy is not a static or binary concept but is instead constantly remade in ways that free and empower journalists (giving them sociologically privileged positions), while simultaneously constraining them (by putting them in constant contact with a set of actors and forces that limit and contextualize their work).

I start by tracing the idea of a field through its foundations in the work of Ernst

Cassirer, how these core concepts were adopted and adapted in Pierre Bourdieu‘s

Field Theory, how this theory was applied to the studies of the press – and how these applications proved limited in light of the complexities highlighted by New

Instititutionalist ways of understanding fields. From this literature I extract a model of press autonomy and apply it to a brief history of three elements of US journalism— professional identity, organizational practices, and public interactions—showing how press autonomy is not an single ideal to be attained but, rather, a guiding ethos that takes different forms in various aspects of journalistic practices.

2. Grounding Press Autonomy: Field Theory, New Institutionalism and the U.S. Press

2.A. Introduction

The idea of whether journalism is a field – or whether it should be considered one – has its roots in one of the most fundamental questions concerning the sociology of media: how individuals and groups use and adapt systems of symbol representation to create and sustain communities.

The idea of a journalistic field can be traced to two theorists who never researched journalism at all. Ernst Cassirer‘s philosophy of culture and Kurt Lewin‘s field theory of social psychology are the intellectual roots of later theories that describe journalism as a field. Both theories share a fundamental view that social systems are essentially ongoing struggles between subjective agency (how individuals interpret, imagine and assert their free will) and objective structures (how organizations systematize, regulate, afford and constrain). This dialectic forms the core of Bourdieu‘s Field Theory, which grew out of, in part, his critical readings of

Cassirer and Lewin and their subsequent incorporation into his historical accounts of how European literary and artistic fields emerged and gained autonomy. Bourdieu‘s work consists of both empirical accounts of different histories of cultural production and a general thesis on how individual actors compete for symbolic and economic capital within fields of forces that both restrict and suggest their individual practices and interpretations. Most relevant to the sociology of media, Bourdieu and others have applied his Field Theory to the study of journalism, arguing that a distinct journalistic field exists in which its autonomy is secured by privileging consecrated, restricted production of symbols and representations over the large-scale, market- oriented manufacture of cultural goods. Such a hierarchy, Bourdieu argues, assures that journalists are insulated from the public‘s attraction to spectacle and are able to benefit from the long-term high-risk cultural practices that eventually produce enlightened citizens.

In a complementary but distinct effort, Cook and Sparrow describe a ―new institutionalist‖ approach to the study of journalism. Grounded in the broader

83 sociological project of new institutionalism, Cook argues that any notions of autonomy – either of the field of journalism or the production of objective accounts by journalists – essentially mask journalism‘s identity as a fundamentally political institution. The news, he shows, emerges as a co-construction, negotiated among journalists who enjoy a variety of government subsidies and government officials who need journalists‘ public validations.

Framed as such, the press is a field that lives not within a single organization, profession, set of practices or cultural expectations and the image of an autonomous journalist is not as an artist ―who lives (figuratively) outside of society, beyond normal conventions, and who is therefore better able to see and expose its shortcomings.‖

(Bollinger, 1991: 55) Rather, the press is a system of forces in which in a multiplicity of actors—professional reporters and editors, self-styled bloggers and content creators who claim to act like professional journalists, sources who appear in stories, publics that both consume news and participate in its production, private industry funders, government elites and decision-makers, news organization leaders, state regulators and judges, and technological innovators—negotiate positions in relation to each other and a set of values or ideals. This chapter explains the dynamics of this system of forces, uses these dynamics to propose a model of institutional press autonomy, and reinterprets the U.S. press‘s autonomy as a sociological negotiation among actors and ideals working within a field of forces that defines the press.

2.B. Structure-Agency Dialectics and the Development of Field Theory

Understanding contemporary journalism requires appreciating both the practice of journalism and the conditions under which it is practiced. The actions and assumptions of individual journalists, their news organizations, the commercial market, advertisers and audiences all reinforce and reflect the particular ideals and practices of journalism at any given time. This means that the study of journalism is essentially a mezzo-level—in between micro, individual factors and broader, macro phenomena—one works across both individual practices and institutionalized norms.

Such investigations are fundamentally studies of structure and agency: how the expectations, patterns and categories that guide how people, groups and cultures behave are both reproduced and changed by the languages, interpretations, and values they bring to their personal and collective practices. For example, consider an individual journalist who may be motivated to be a crime reporter because of family member‘s experience with a murder or theft, but who must channel or reconfigure these passions into traditions of objectivity and professionalism demanded by her news organization‘s code of ethics, her colleagues‘ expectations, and her desire to win reporting awards. As she reports—frames issues, interviews people, writes stories, works with editors—she must reconcile her personal perspectives (which may be motivating and inspirational) with a system of professional and cultural values that expects crime reporters to behave a certain way. She is operating at a mezzo-level that requires constant negotiation and re-orientation to her own passions and the norms of those around her.

Most fundamentally, mezzo-level research is the study of how a piece relates to a whole, how an individual is both a product of – and an agent defining – the social systems to which she belongs. This basic dialectic between structure and agency underpins several theories that see communication and organization as socially- defined structures. For example, Mead‘s observation that language consists not of

―inner meanings to be expressed, but in its larger context of co-operation in the group taking place by means of signals and gestures‖ (Mead, 1934/1967: 6) led to Symbolic

Interactionism‘s rejection of the notion that people simply use language but, rather, that it is a material with which people assemble interpretations and create sharable, context-dependent understandings (Blumer, 1969: 5). Structuration Theory (Giddens,

1984) makes a similar point when it argues that social structures are conceptual tools with which individuals think, that they must always be considered ―the medium and the outcome of the practices [that] constitute social systems‖ (Sewell, 1992: 4)18.

Understanding Bourdieu‘s Field Theory and how its underlying dialectics shape a distinct approach to researching social phenomena requires appreciating two principal philosophies out of which contemporary Field Theory emerged, namely the

―philosophy of culture‖ articulated by Enrst Cassirer and the theory of social- psychological fields developed by Kurt Lewin – both of which focus on understanding the meaning of an instance in a pattern, and a particular organism to an ecological whole.

18 Note, too, another related and relevant body of literature: Actor-Network Theory. It extends Symbolic Interactionism and Structuration Theory to argue that sociology should not see individuals, groups or organizations as its primary units of analyses but, rather, should focus on ―tracing of associations‖ among connected ―things,‖ not just people or groups, that may not appear social but that have socially mediated relationships (Latour, 2005: 5). 86

A founder of modern social science, Cassirer was generally concerned with creating a ―science of culture‖ that might explain how humans both interpret and structure their environments – how they ―give form to experience‖ (Cassirer,

1942/1960: 22; Cassirer, 1923/1953). He challenged what he saw as a ―crisis in the concept of evolution upon which the entire naturalistic philosophy of culture is built‖

(Cassirer, 1942/1960: 34) and argued that physical models of causality were being over-applied to the realm of culture. He criticized the idea that ―all events in the human world are subject to the laws of nature and that they are the outgrowth of fixed natural conditions‖ (16), distinguishing between physical phenomena that are largely beyond the human influence and social phenomena, the nature and meanings of which are constantly changing due to human intervention and interpretation.

He did not reject causality, determinism or logical proof as epistemologies that might have value in the social sciences, but instead firmly placed their strict application in the realm of objective, scientific investigations of physical phenomena.

Cultures and customs already constrain and create hegemonies, he argued, such that an individual is ―determined and limited from his first movements by something over which he has no power‖ (42). A philosophy of human behavior that privileges causal reasoning or universal objectivity might result in a ―tragedy of culture‖ – a consolidation of life in which objectivity rules the ego, not allowing it to ―go beyond its own sphere and enter one foreign to it‖ (189). Cassirer‘s primary concern was that an overtly powerful naturalist philosophy might prevent the individual from creating an ―aesthetically liberated life‖ (Cassirer, 1946: 98) and instead bound her within an environment she cannot change.

Cassirer‘s own philosophy of culture instead asserts what is uniquely human about social phenomena and calls for models of society that put imagination and agency on equal footing with objective facts:

[T]he titanic creative power of the individual is in conflict with the forces with have as their goal the preservation and, in a sense, the immortalization of the status quo. Creativity is forever in conflict with tradition. But here, too, it is wrong to paint the conflict in black and white – as if one side had a complete monopoly on values and the other side a total absence of values. Tendencies toward perseveration are no less significant and just as indispensable as those which seek renewal; for renewal can only come to flower through being preserved, and perseveration is possible only through self-renewal (Cassirer, 1942/1960: 197)

Juxtaposing creativity with tradition is a powerful way of understanding the dialectics underlying Field Theory (individual-group, structure-agency, objectivity-subjectivity) because it locates them in practices that produce symbolic forms. These forms

(language, myth, art, religion) are the bases of human communication and negotiation since it is within them that we might see the evolution of shared systems of meaning and the foundation of community. Cassirer‘s philosophy lets us moves beyond essentialist debates that largely ignore the socially constructed and symbolically mediated nature of community – trying to define what free will or social constraints are – to discussions of how to create the philosophical conditions in which the dialectics underpinning Field Theory might endure and change. Meaning is not defined at the level of an individual, but through a set of social relationships that signal significances through symbolic forms that are constantly remade and reinforced by these same relationships.

A founder of social psychology, Lewin‘s Field Theory (1951) emerged from a desire to understand the nature of psychological phenomena, not simply their effects

88 on individuals. Framing his work as a study of social concepts that take psychological form, he departed from dominant forms of psychology at the time that tried to catalogue perceptual phenomena. He instead saw individual psychology as a series of

―dynamic and nondynamic constructs‖ (Lewin, 1951: 39) that, together, define an individual‘s experience: position (an person‘s location relative to a group, occupation, activity, including how near or far he might feel or physically be from them); locomotion (the pattern of positions over time); force (the tendency for locomotion); goal (a field that distributes forces in space); conflict (overlapping force fields); fear

(expectations about the present, hope for the future, and guilt about the past); power

(the likelihood of inducing forces); values (like force fields but with positive and negative valences).

For Lewin, Field Theory was a way to integrate among such seemingly subjective factors, a ―method of analyzing causal relations and of building scientific constructs‖ (Lewin, 1951: 45) so that he might move beyond the analytical traditions that dominated psychology at the time: anamnesis, or studying past behavior; and

―tests of the present‖ (1951: 49) that focused on diagnosing a pre-existing state.

Missing from psychology was the recognition that the

individual sees not only his present situation; he has certain expectations, wishes, fears, daydreams for his future. His views about hs own past and that of the rest of the physical and social world are often incorrect but nevertheless constitute, in his life space, the ―reality-level‖ of the past (Lewin, 1951: 53).

Critical to Lewin‘s conception of fields is the idea that the past and the future are simultaneous parts of an individual‘s or group‘s perception at any given time. That is,

Field Theory aims not to describe the past or predict the future but, rather, to explain a particular time as a product of both past influences and future expectations.

In addition to providing a set of analytical constructs and a model of psychological time, Lewin‘s Field Theory presages the structure-agency dialectic by arguing that understanding our personal ―life space[s]‖ (1951:58) – the behaviors available to us at a particular time – requires seeing our environments as a set ―of coexisting facts which are conceived of as mutually interdependent‖ and that objectively define the variety of ―possible life spaces‖ (1951: 240, emphasis added) available to us at any given time. Essentially, Lewin argues that we are agents with free will – but that our perceptions and choices at any given moment always depend upon how we relate to a set of enduring facts that structure our environment.

2.C. Bourdieu’s Field Theory

Motivated by Cassirer‘s philosophy of culture and Lewin‘s social psychology of fields, Bourdieu constructed a new theory meant to bridge the ―most fundamental, and the most ruinous‖ divide in social science – between subjectivism and objectivism

(Bourdieu, 1990: 25). His principal aim was to understand how individuals make, replicate, act within and adapt the conditions that afford and constrain their behaviors, focusing on the dynamics of cultural production. For example, cultural products emerge from social systems of creation and exchange that are mediated both by

90 commercial contract to regularly produce books with essentially the same literary structures, themes, plots and characters). Cultural production is never an isolated phenomenon in which a single individual working in isolation can have an idea, give that idea form, and distributes the form among waiting recipients who will simply receive the intended idea without alteration or reinterpretation.

Such a theory, he argued, must simultaneously reject a notion that have traditionally guided social science: that objective observers can assume ―sovereign viewpoints‖ (1990: 27) from which they can operationalize human behavior and claim to understand the deep structures of social ecologies.

Instead of first-person subjectivity or omniscient objectivity, Bourdieu‘s Field

Theory calls for a relational epistemology in which a viewpoint (consisting of both a position and interpretation of others‘) is the principal unit of analysis and the ―work of symbolic representation‖ (Bourdieu, 1985: 727) is the process through which individuals imagine and realize different perspectives. It is this ongoing balance between dispositions and structures that Bourdieu captures in his concept of habitus as

―systems of durable, transposable dispositions … structuring structures … which generate and organize practices and representations that can be objectively adapted to their outcomes without presupposing a conscious aiming at ends or an express mastery of the operations necessary in order to attain them.‖ (1990: 53)

These structures – codified in social systems, emerging from historical processes, but still subject to adaptation – are organizing principles that guide people‘s expectations of themselves and others. Habitus becomes a kind of practice-rendered language because it suggests behaviors while constraining their possible meanings: it

91 gives people freedom and independence to express ―thoughts, perceptions and actions‖

(1990: 55) – but only within a particular set of conditions. Like Lewin, Bourdieu does not intend habitus to be a way to predict future conditions, explain past behavior, or proscribe correct practices. Rather, habitus are evidence for how forces, at a particular time and in a specific set of conditions, structure agents‘ actions.

These forces, in turn, constitute a field – a ―field of struggles‖ or a ―space of possibles‖ (Bourdieu, 1993: 30) – in which a ―structure of objective relations [allow] us to account for the concrete form of … interactions‖ (182).

Bourdieu defines a field most generally as ―a field of forces within which agents occupy positions that statistically determine the positions they take with respect to the field, these position-takings being aimed at either conserving or transforming the structure of relations of forces that is constitutive of the field.‖ (Bourdieu, 2005:

30) The literary field in particular is defined as ―an independent social universe with its own laws of functioning, its specific relations of force, its dominants and its dominated … where, in accordance with its particular laws, there accumulates a particular form of capital … [and] entirely specific struggles.‖ (1993: 163-4)

(Bourdieu does not mean ―laws‖ in a narrow sense focused on formal, state- authored regulations but, rather, as a more general term to describe a set of actions and dispositions that socially constructed norms and values considered to be more or less permissible. Similarly, ―capital‖ does not only mean currency or commodities but socially created resources like reputation, class, or family lineage that can be recognized both in formal systems [e.g., consider military ranks, university degrees, and caste systems that create clear hierarchies] and in more informal, implicit ways

[e.g., consider how a worker‘s perceived value in an interview setting may be a function not only of her employment history and letters of reference, but also her dress, accent, demeanor, and the reputations of others who may resemble her work history or educational background].)

There are several features of these definitions to extract to understand

Bourdieu‘s Field Theory in relation to Cook‘s New Institutional model of journalism.

Specifically, a principal concern of Bourdieu is how autonomous a particular field

(literary, artistic, religious, journalistic) is at any given time, how this autonomy can be traced through individual subjectivities and structural conditions, and how agents‘ habitus and position-takings are encroached upon by extra-field forces (influences from actors and values that are considered to be, either explicitly or implicitly, beyond the scope of a particular field) and new entrants (actors from another field who are attempting to join a field, and whose value and position within the new field must be negotiated).

In tracing the emergence of French literary and artistic fields in the second half of the nineteenth century, Bourdieu identifies a series of moves in which writers and artists gained increasing autonomy from those who traditionally supported their work: the financier, commissioning patron, or official state-sponsored ―protector of the arts‖

(Bourdieu, 1996a: 49). Such direct support and domination was broken only when writers and artists began to develop two alternative modes of support: markets and

―durable links‖. Bourdieu argues that the sanctions and constraints of markets recast artistic works as ―symbolic goods‖ (the success of which can be measured through sales figures and revenue streams) while ―durable links‖ signal entirely different

―apparatuses of consecration‖ (salons, universities, elite journals) based more on shared affinities and lifestyles among a small set of elite cultural producers. A field‘s autonomy, Bourdieu argues, exists in its ability to sustain these two venues for the production and evaluation of symbolic goods, each driven by different logics. One domain uses large-scale, commercial logics to produce low-risk symbolic commodities over short production cycles, creating goods with pre-established forms already accessible to ―mass publics‖ with well-understood tastes (Bourdieu, 1993:

129; 1996a: 142). In the restricted sphere of production, artists use longer, higher-risk production cycles to create experimental ―pure art,‖ competing for the ―cultural consecration‖ of other producers, competitors and ―privileged clients‖ (Bourdieu,

1993: 115).

Although cultural production indeed involves individual agents working within habitus to interpret, create and imagine new possibilities, in both systems, position and power is earned through a set of objective relations: markets signal success through popular appeal and economic revenue, while ―agents of consecration‖ (Boudieu, 1993:

121) confer legitimacy through restricted evaluation criteria, largely comprehensible only to a small set of powerful critics ensconced in stable roles.

For Bourdieu, a field‘s autonomy is most clearly seen in the relationships between these two sub-fields: restricted and large-scale production. If critically- oriented, small-scale artists and writers can create new works independent of the demands that structure market-oriented, mass-scale production – more technically, if the field‘s ―internal hiearchization‖ dominates ―external hierarchization‖ (Bourdieu,

1996a: 217) – then a distinct quality of cultural product might emerge that does not

94 depend on the audience or public. Bourdieu avoids arguing simplistically that restricted production results in ―better‖ or ―higher-quality‖ works than large-scale production but in privileging the creation of novel, high-risk, experimental, ―pure‖ art in the sub-field of restricted production, he gives the ―agents of consecration‖ – not the

―mass public‖ – responsibility and power to change a field.

For Bourdieu, these questions of autonomy and change are linked and tied to the problem of how to define a field‘s boundaries. That is, to understand how a field emerges as a distinct area of struggle means understanding how habitus emerge, how sub-field hierarchies shift, how new actors enter a field, and how a field‘s autonomy is infringed upon by the influence of different forms of capital. Bourdieu highlights several mechanisms for such change. A field‘s internal struggles, its habitus, its networks of consecration and capital exchange may be subject to factional rifts among social classes not directly tied to the field in question; e.g., the 1848 French uprising briefly encouraged artists to produce ―social art‖ (Bourdieu, 1993: 58) before they returned, post-uprising, to relating to more established legitimacy from markets and fellow producers. Or younger newcomers may enter the field, bringing with them

―dispositions and position-takings which clash with the prevailing norms of production and the expectations of the field‖ (1993: 57). Such new entrants bring with them a youth and sense of time that, Bourdieu argues, is a de facto feature of change.

Both mechanisms for change still locate newness and innovation mostly within the sub-field of restricted cultural production in which the ―mass public‖ plays one of two roles. It is either an audience that consumes large-scale production, the symbolic nature and meanings of which originate from ―consecrated authors dominating the

95 field of production [who] also tend to make gradual inroads into the market, becoming more and more readable and acceptable [in a] more or less lengthy process of familiarization.‖ (Bourdieu, 1996a: 159)

As the mass public adopts these tastes (―understood as systems of preferences concretely manifested in the choices of consumption‖ [Bourdieu, 1996a: 160]) and impact markets and revenues, they signal to restricted producers what might eventually constitute successful large-scale cultural production. The mass public might also, though, induce temporary change in artists and fields of cultural production when they participate in revolutions and political movements that restructure fundamental social relations and power structures.

Importantly critical for the understanding of the journalistic field, Bourdieu does not seem to see lasting field change originating in a public sphere through public activity but instead gives more change-making power to the ―struggle between the established figures and the younger challengers‖ (Bourdieu, 1993: 60). (For instance,

Boudieu‘s use of the term ―public sphere‖ has little of the sophistication of

Habermas‘s [1989] ideal public sphere—designed to foster truth-seeking through rationality, the bracketing of individuals‘ identities and a focus on matters of common interest—or Young‘s [2000] ―decentered‖ public sphere in which diverse and mutually accountable participants engage directly with each other‘s differences.)

These restricted-production actors are engaged in a different kind of struggle for change – one for ―cotemporaneity‖ (Bourdieu, 1996a: 158) in which the power to define the field‘s timeframe carries with it the power to categorize actors either as forward-looking and avant-garde or as part of the conservative rearguard.

This kind of struggle brackets out the public, though, because, within the field of restricted production, no matter what an actor‘s chronological age, there is a fundamentally shared system of artistic (and ―journalistic‖) values. These values constitute what Bourdieu calls the illusio of the ―aesthetic disposition‖ (Bourdieu,

1993: 257): ―Participation in the interests which are constitutive of membership of the field (which presupposes them and produces them by its very functioning) implies the acceptance of a set of presuppositions and postulates which, being the incontestable condition of discussions, are by definition sheltered from debate.‖ (Bourdieu, 1996a:

167)

Bourdieu‘s Field Theory develops a number of concepts essential for describing any system of structured, social relations: habitus as a system of structuring yet adaptable principles for organizing practices and expectations; field as a site of struggle for capital and power in which individual agents occupy positions in relation to a space of possible positions; autonomy as a way of understanding a mature field‘s ability to shape the conditions under which it engages in restricted and large-scale production and field-change; and illusio as a set of shared principles that signal the value-laden preconditions for participation.

Equipped with these concepts, we can use Field Theory and its extensions to journalism by Bourdieu (1998, 2005) and others (Benson, 1999, 2004, 2006; Benson

& Neveu, 2005) in conjunction with Cook‘s New Institutionalism (1998, 2000, 2006) to construct a sociological model of press autonomy and trace it among different logics of journalistic practice, specifically, professional ideals of objectivity, ritualized organizational routines and strategic public distancing.

2.D. Bourdieu’s Journalistic Field

Bourdieu‘s influence on journalism studies is primarily through a short book, On

Television, a companion essay, and a series of secondary analyses (Benson, 1999;

Benson, 2004; Benson & Neveu, 2005; Neveu, 2007) in which journalism and journalists are critiqued from the perspective of Field Theory.

Bourdieu‘s main thesis in On Television (1996b) is that, as journalism has become more dependent on large-scale commercial production logics it has become less autonomous as a field. First, he claims that the medium of television is driven by a kind of ―structural corruption‖ (1996b: 17) that rewards people who produce spectacles without a clear idea of what they want to say to an audience, or why an audience should listen to them. They are forced, by television‘s time constraints, to become ―‗fast-thinkers,‘ specialists in throw-away thinking‖ (1996b: 35) who ―say banal things‖ (1996b: 18) and do a kind of ―symbolic violence‖ (1996b: 17) as they comment simplistically on complex issues. Although Bourdieu acknowledges the dialectical nature of such production systems—saying that journalists and sources are as much manipulated by the conditions of television as they are agents of manipulation—he argues that individual journalists most clearly demonstrate the field‘s impoverished, non-autonomous state when they engage in ―relations of competition (relentless and pitiless, even to the point of absurdity) and relations of collusion, derived from objective common interests.‖ (1996b: 36) Journalists, he says, are more motivated to report issues that are ―exceptional for them‖ as they engage in deadline-driven ―relentless, self-interested search[es] for the extra-ordinary.‖ (1996b:

20) And implicitly invoking communication research‘s Cultivation Theory (Gerbner,

1970; Gerbner, Gross, et al., 2002), Bourdieu argues that such journalists—even those

―acting in all good faith and in complete innocence‖—can create ―reality effects‖ that

―unleash strong, often negative feelings, such as racism, chauvinism, the fear-hatred of the foreigner, or xenophobia.‖ (1996b: 21)

The important point to note here is that, although Bourdieu‘s Field Theory is a rich and useful account of literary and artistic fields, his application of Field Theory to journalism and the press is seriously lacking. He presumes that they uniformly and uncritically adopt the logics of large-scale commercial production and traces trends in what he claims is a general homogeneity and decline: newsmagazine headlines are increasingly similar, the same commentators reappear, opinion poll results are ritually reported, and television news broadcasts might only differ in the order of the stories presented (Bourdieu, 1996b). Without supporting his claim, Bourdieu argues that the journalistic field is at a structural disadvantage compared to other domains of cultural production because it is ―much more dependent on external forces … depend[ing] very directly on demand [and] the decrees of the market and the opinion poll.‖ (1996b:

53) His seemingly complete rejection of journalism‘s legitimacy as a field is curious in the absence of a either a comparison to other media fields—e.g., consider market influences on reality television, cable talk shows, or talk radio—or a more complete dynamic of differences and dynamics within journalism‘s field (e.g., consider differences between niche and mass-market presses or daily beat journalism versus long-form investigative reporting).

He also presumes that journalism once was autonomous, as he asserts without empirical evidence that its autonomy as a field has dwindled. Bourdieu argues that as

99 journalistic habitus and the distribution of symbolic and economic capital become increasingly similar across field participants, that journalistic ―competition‖ only shallowly masks the increasing uniformity among its products. The journalistic field and its actors are, he claims, ―increasingly subject to the constraints of the economy and of politics‖ (Bourdieu, 2005: 41) as are the organizations and publications comprising the field. In Field Theory terms, if journalists are to retain their positions – and have the chance to assume other positions within the field – they, consciously or not, must adopt the habitus and illusio rewarded by the market-oriented logics of large-scale production.

It is in these rants and laments—the tone of Bourdieu‘s analysis barely conceals a normative plea to reform journalism—that we see some significant shortcomings of his application of Field Theory to journalism (e.g., see Marlière [2000] and

Schudson [2005] for good critiques). The most obvious is a kind of circular logic as he a priori assumes the homogenization of journalism and journalistic practices while simultaneously arguing their uniformity. He does not, for example, distinguish among different kinds of journalists, e.g., beat reporters, feature writers, investigative journalists, editorial commentators, and he moves quickly among critiques of television, newspapers and magazines without exploring the professional, commercial and symbolic differences among them. Indeed, without any definition or explication of the diversity that exists with journalism (albeit it in ways not explicitly appearing in his original Field Theory), his critiques of the field are difficult to engage with fully.

Bourdieu also only briefly touches on the public or democratic features of journalism (at least in its ideal form) as central elements distinguishing it from other

100 fields of cultural production. He talks about journalism‘s ―de facto monopoly on the large-scale informational instruments of production and diffusion of information

[through which] they control the access of ordinary citizens … to the ‗public space‘‖ and its control of ―public existence, one‘s ability to be recognized as a public figure‖

(Bourdieu, 1996b: 46) – but he does not develop these public dimensions of journalism further. In fact, Bourdieu seems to argue that journalism should be anti- public, suggesting that the journalistic field should have a higher ―entry fee‖ and that some of its actors have a ―duty to get out.‖ (Bourdieu, 1996b: 65) In refuting anticipated charges of elitism, Bourdieu claims that he is defending ―the social conditions of production‖ (Bourdieu, 2005: 46) that might make possible enlightening cultural products that could not emerge from large-scale, mass-oriented production regimes.

This equating of the public as consumers of large-scale cultural production is perhaps at the heart of a critical and under-articulated problem with Bourdieu‘s application of Field Theory to journalism. Specifically, in importing his generic notions of autonomy from his histories of other fields of cultural production directly into journalism, he collapses the public onto the market in a way that prevents one from seeing anything other than journalism‘s decline as an autonomous field.

In Rules of Art and The Field of Cultural Production, Bourdieu argues that a field is most autonomous not only if maintains a separation between two poles of production – restricted and large-scale – but if the restricted producers are symbolically privileged over those who produce for the market (Bourdieu, 1996a:

217). Essentially, if the critical interpretations and meaning-making of elite producers

101

(insulated from markets and given more time to develop their ideas) are considered superior to the quick productions targeting mass publics, then, over time he claims, the products of restricted production will lead a field‘s change, ―making gradual inroads into the market.‖ (Bourdieu, 1996a: 159) Such claims mostly rest on his re-readings of histories of French painters and their patrons, and his analysis of how critically recognized authors became best-selling authors (and vice versa) in 1970s France.

It is unclear what this means for journalism since Bourdieu‘s conception of symbolic quality—and his rationale for protecting elite symbol-producers—seems to depend upon uncritically accepting that separation from the masses in both space and time produces higher quality, more forward-looking, risk-taking culture. If, however, we view journalism not simply as another field of cultural production but, instead, as a particular domain of symbolic activity that—in its ideal form—is oriented towards publics and not only markets, then Bourdieu‘s call for journalistic autonomy is less defensible. The claim here is that journalism is essentially different from other fields of cultural production because of its public nature and, thus, requires a different understanding of autonomy. For example, journalism is oriented around a number of central tensions: whether the press should primarily serve a civic information and analytical ideal (e.g., equipping people with data and analyses they need to know in order to vote knowledgeably), a consumer-driven market ideal (e.g., providing people with news that they find pleasing or entertaining), a community-oriented agenda (e.g., in which it assembles and facilitates the kind of rational, cross-cutting conversations among diverse people envisioned, for example, by Habermas [1989], Young [2000] and Fraser [1990]), a watchdog ideal in which the press monitors elites‘ and elected

102 officials‘ uses of corporate and governmental power on citizens‘ behalf, or a mobilizing function in which they organize citizens into groups capable of challenging power (Schudson, 2008b: 11-26). Essentially, journalism cannot easily separate itself from commercial, governmental or civil society influences and activities without relegating itself to a narrow role that ignores its multi-faceted nature.

Differently put, the press can normatively serve a number of different roles that become impossible if its autonomy is defined narrowly as separation from markets, states or civil society. Christians, et al. (2009) define four roles for the press. First, a monitorial role in which the press performs a number of information practices

(everything from keeping and publishing agendas of public events to maintaining presences in official forums and warning publics of both short- and long-term issues of concern) in the service of citizens who need and demand information relevant to their local circumstances (139-157). Second, a facilitative role in which the press creates and manages forums for public deliberation, helps civil society groups (neighborhood organizations, churches and other non-governmental and non-state associations), and influences cultural environments in which ―autonomy and rights are meaningful.‖

(166) Third, a radical role in which the press ―insists on the absolute equality and freedom of all members of a democratic society in a completely uncompromising way‖ (179) so that social power can be redistributed ―from the privileged (typically few) to the underprivileged (typically many)‖ (181). This is a ―revolutionary ideology‖ in which journalism is ―an instrument for challenging and changing political and economic systems.‖ (181-82) Finally, the press can serve a collaborative role in which the press and the state engage in partnerships ―built on mutual trust and a shared

103 commitment to mutually agreeable means and ends.‖ (198) Collaboration can be understood as ―compliance‖ (a weak form in which the press is coerced into behaviors, apathetically disregards its interests, or uncritically follows traditions),

―acquiescence‖ (a slightly stronger form in which the press begrudgingly recognizes the pragmatic benefits of cooperation with the state and cooperates for strategic reasons), and ―acceptance‖ (a strong form of collaboration in which journalists‘ own judgments lead them to believe that they are right to and normatively should cooperate with the state) (198-200).

It is worth reviewing each role here in detail because each role has different implications for press accountability. That is, the roles that the press enters into— monitorial, facilitative, radical, collaborative—entail being responsible to different actors and norms. It is too simplistic for Bourdieu to argue that press autonomy is an exercise in securing independence from fields that are perceived to contaminate or interfere with journalism‘s one true mission. Such a view of autonomy limits the press‘s potential roles, and citizens‘ chances of benefitting from them. And in contrast to Bourdieu‘s claim that field change is achieved by an autonomy that comes from separating the commercial, market-place sphere of activity from critical, restricted production, Schudson argues that

[w]hat keeps journalism alive, changing, and growing is the public nature of journalists‘ work, the nonautonomous environment of their work, the fact that they are daily or weekly exposed to disappointment and criticism of their sources (in the political field) and their public (whose disapproval may be demonstrated economically as readers cancel subscriptions or viewers change channels). Vulnerability to the audience (the market) keeps journalists nimble in one direction, vulnerability to sources (the government) in another … [A]bsent these powerful outside pressures, journalism can wind up communicating only to itself and for itself (Schudson, 2005: 219; emphasis added)

104

For Bourdieu, then, the influences of other cultural fields of production are sources of contamination, mechanisms by which journalism‘s autonomy erodes and disappears – evidence that democracy is suffering. But, for Schudson and others who see journalism as a fundamentally public domain, such influences are one source of accountability – ways of increasing the likelihood that journalism remains responsive to and engaged with hybrid public/market/civil society forces that make journalism‘s work purposefully responsive to multiple audiences.

Although Bourdieu‘s own application of Field Theory to journalism needs further development, the fundamental concepts of Field Theory offer powerful ways to understand the dynamics of the journalistic field. Such a formulation lets us see journalism not as the exclusive domain of any particular organization or definition of

―journalist‖ but as an arena of struggle for economic and cultural capital where

―economic capital is expressed via circulation, or advertising revenue, or audience ratings, whereas the ‗specific‘ cultural capital of the field takes the form of intelligent commentary, in-depth reporting, and the like – the kind of journalistic practices rewarded each year by the US Pulitzer Prizes.‖ (Benson & Neveu, 2005: 4)

But even recasting journalism within a broader field of power does not help us address the question of how journalism relates—or should relate—to the public sphere or to the state. As Benson & Neveu acknowledge (2005: 9), Bourdieu‘s Field Theory maintains a ―more ambivalent or at least indirect relation‖ to democratic normative theory. He is more concerned with the social conditions by which ―specialized knowledge‖ and ―enlightened citizens‖ are produced – the environment best created

105 by an autonomous sub-field inhabited by elites engaged in restricted production for other journalists.

If Bourdieu‘s Field Theory views journalism as a field struggling for autonomy from the state and other cultural fields, Cook‘s ―New Institutionalist‖ approach to journalism explicitly tackles the non-autonomous nature of the field. He argues that journalism has, in fact, always been a political and social institution in which news emerges from historically blurred boundaries between the press and government.

2.E. Bourdieu’s Journalistic Field & New Institutionalist Journalism

The work of Bourdieu and Cook have much in common. They both see journalism as an institutional or field-level phenomenon with norms, structures and practices that shape—and are shaped by—the subjective actions and interpretations of individual journalists. Likewise, they both understand the journalistic field as one that is in constant transition as new agents—e.g., young reporters who self-identify as journalists, but also others like public relations officers and industry spokespeople who may periodically attempt to be seen as journalists for strategically reasons—enter the field and compete for limited roles, and scarce economic and symbolic capital.

And they are both concerned with questions of journalistic autonomy, understanding the field to be essentially one in which symbolic goods are developed and shared by organizations that need to maintain relationships with other fields, markets and publics.

Despite these commonalities, though, there are significant differences between the two models that cluster into two broad areas of change and autonomy.

106

New institutionalism tends to see the journalistic field changing through its interaction with other fields, primarily the political field. In his account of how news is negotiated among and co-constructed by journalists, government officials and politicians, Cook describes a set of complementary and tightly coupled practices maintained by both journalists and government officials. In this sense, there is no single journalistic field in Bourdieu‘s sense but, rather, a hybrid area of struggle and cooperation in which journalists and state actors both compete and help to create representations that reflect their respective perspectives.

In Bourdieu‘s model, field change is more of a time-based one reflected in tensions between younger and older producers and in the relationships between the two sub-fields of production (large-scale and restricted). In the first kind of change, newcomers usurp experienced producers; new entrants are bound to continually banish to the past … those consecrated producers against whom they measure themselves and, consequently, their products and the taste of those who remain attached to them.

(Bourdieu, 1996a: 157-158)

The struggle between these newer and older producers is essentially a struggle to define themselves in terms of time – identifying themselves and each other as experienced and conservative, or naïve and avant-garde. In journalism, this kind of change might present itself within a single publication as younger reporters encroaching on the issues or beats usually covered by more seasoned journalists; or as the popularity of new forms of journalism (e.g., standards of citation that quotes from anonymous sources, or expectations that journalists respond to readers‘ posts in online forums) over traditional practices; or in the crowding-out of traditional news outlets by

107 newer publications (e.g., blogs or Wikipedia-like sites garnering more attention or revenue than newspapers or newspaper websites).

In the second kind of change, the cultural work of elite, consecrated producers gradually enter into the domain of large-scale production (e.g., imagine investigative reporters like those at ProPublica who may work for many months or years, with the funding and organizational support of a foundation or news organization, to break a detailed story that is serialized in a mass-scale publication like the Wall Street Journal or USA Today, or highlighted in a television magazine program like 60 Minutes or

20/20). This kind of symbolic flow lets cultural works that take longer to produce

(and that signal greater risk) join the kind of quickly produced, largely homogenous products usually offered to mass-scale audiences. In a journalistic sense, this kind of development might be seen in the publication of investigative stories in serial form in popular press outlets; or by feature reporters writing about the structural conditions and patterns of events and organizations and perhaps suggesting to daily beat reporters the kinds of questions they might ask.

The second point of departure between Bourdieu‘s journalistic field theory and

Cook‘s new institutionalist approach concerns their views of autonomy and, by extension, their conceptions of the public. Bourdieu sees the state as one player in the field of power with its most notable direct role being to impose its visions of legitimacy on the ―production of producers‖ through its sanction of particular academies and subsidy of particular commissions (Bourdieu, 1993: 250-251). But beyond these general interventions, Bourdieu makes little comment on the relationship between the ―political field‖ and the ―journalistic field‖ except to argue that the most

108 autonomous, ―purest‖ journalism exists independent of ―state power, political power, and economic power.‖ (Bourdieu, 2005: 41) Presumably, for Bourdieu, press autonomy is achieved by insulating itself against all such powers, even if the public‘s perspectives might be represented in the actions of the state, politics or commerce.

For Boudieu, any attempt to represent the community’s agenda is anathema to the very idea of an autonomous press since his theory sees such reform as nothing other than pulling the journalistic field further away from where its symbolic authority should lie

– among a small set of elite, symbolic workers engaged in restricted production

(rewarded by agents of consecration) that may eventually be accessible to large audiences.

Cook, though, sees the state as a central influence on the press‘s institutional identity. He goes beyond simply calling the news media a political institution— tracing their co-constructed histories, processes and products—and calls for a kind of journalistic autonomy motivated by concern for the public. He expresses no particular faith in journalism‘s ability to report the community‘s agenda or specific concern with the influence of large-scale production logics on the symbolic quality of the news.

Rather, he questions the press‘s ability to be a competent partner in government and critiques its institutional ability to be a trusted, public contemporary of the state. He observes that there is little historical motivation for public-oriented journalism because

―the public has never been, for better or worse, pivotal to the ability of the news media to act as an intermediary political institution.‖ (Cook, 2000: 193) Cook suspects that journalists are ―poorly equipped to assist in governance,‖ notes that no one elected reporters or can hold them accountable as representatives in government, and calls for

109 media policies ―to ensure that the news we receive gets us toward the politics and toward the democracy we want‖ (Cook, 1998: 4)

In his comparison of Bourdieu and Cook, Benson (2006) essentially agrees that

Cook‘s inclusion of the state helps fill a major gap in Bourdieu‘s journalistic field, and suggests that Bourdieu‘s model of strict autonomy deserves an institutional critique that extends beyond rational-actor economics or studies of the press that bracket news production within news organizations or professional journalism. Such a critique underpins a ―new institutional‖ approach to studying journalism (e.g., Cook, 1998,

2000, 2006; Sparrow, 1999, 2006).

This approach is grounded in a broader ―new institutionalist‖ understanding of organizations, practices and norms that looks beyond particular groups, locations, individuals or professions. It instead attempts to integrate across multiple sites of analysis, aiming to understand in an integrated fashion the regulative, normative and cultural-cognitive (Scott, 2001: 51-58) features of institutions: that is, the system of formal rules and state-enforced laws that bound and regulate them (e.g., in the U.S. press, consider First Amendment and libel law); the values and priorities embedded within them and their professions (e.g., consider the largely unquestioned ideals of objectivity and public service implicitly adopted by many journalists); and the social expectations and psychological processes governing how members of institutions relate to each other and publics (e.g., consider the public‘s and journalists‘ assumptions about constitutes a news story, what a credible sources is, or when news should be delivered and by whom).

110

A new institutionalist framework is concerned with systems in which power, sanction, legitimation, reproduction and taken-for-granted beliefs (DiMaggio &

Powell, 1991) are negotiated among a variety of actors, organizations, professions, and social contexts. In essence, ―whereas the old institutionalism viewed organizations as organic wholes, the new institutionalism treats them as loosely coupled arrays of standardized elements‖ (DiMaggio & Powell, 1991: 14). This struggle to work simultaneously with both individuals‘ interpretations and organizations‘ structures makes it a complementary approach to Bourdieu‘s Field Theory: ―much of

[Bourdieu‘s Field Theory] dovetails with and may contribute to a broadening and deepening of the institutional tradition.‖ (DiMaggio & Powell, 1991: 26)

Applying this ―new institutionalist‖ approach to journalism, Cook in particular describes a cross-organizational way of understanding the news media that puts more emphasis on the norms, practices and products shared by the journalistic and political actors than the ―media effects‖ thinking that has traditionally dominated communication research‘s studies of news media. (Cook, 1998: 4; Sparrow, 1999:7) It is not surprising that Field Theory and ―new institutionalism‖ are ideal and complementary research tools for studying journalism. Both approaches are fundamentally concerned with creating models that explain the persistence of both subjective and objective elements within social systems. Such a tension is especially central to journalism since ―journalists work hard to maximize their autonomy‖ (Cook,

1998: 7) as ―professional communicators‖ who convert the messages, ideas and purposes of sources into symbolic strategies designed to inform or persuade audiences

(Carey, 1969/1997: 132-133). But they do so while simultaneously ―work[ing] hard to

111 present a news account that seems largely beyond their individual control‖ (Cook,

1998: 7) and dependent upon unspoken and largely unexamined assumptions that equate ―news judgments (what‘s important) [with] moral judgments (what‘s right)‖

Ettema & Glasser (1998: 8). Ideally, Schudson says, such a tension between structure and agency is resolved socially, ripening into a ―mature subjectivity … aged by encounters with, and [a] regard for, the facts of the world.‖ (Schudson, 1978: 192)

Cook (1998) further develops this tension within an institutional, inter- organizational context and shows how the journalistic field in the United States has never been autonomous from extra-field forces in the way Bourdieu would prefer.

Cook challenges the idea that news media are a collection of independent organizations, instead describes them as an institution: ―[t]aken-for-granted social patterns of behavior valued in and of themselves [that] encompass procedures, routines, and assumptions, which extend over space and endure over time, in order to preside over a societal sector.‖ (Cook, 1998: 84) He then uses three kinds of evidence to show how news media are not an independent, neutral institution that reports on government – but is instead a social and political institution that reports with the support of government.

First, he argues that the news media have traditionally relied upon a variety of government subsidies: early newspapers were sponsored by powerful political figures; the Postal Act of 1792 made it cheaper to mail newspapers than other correspondence;

Congress funded Samuel Morse‘s early telegraph experiments that eventually supported news wire services; government information offices gave updates on state activities in forms that could be easily digested and reprinted by the news media; and

112 even the seemingly restrictive government regulation of radio and television spectra helped the press standardize its use of the new media and come to expect regular revenue streams.

Second, Cook shows that news organizations—even those with different owners, using different media, and targeting different audiences—have largely converged into a single institution with ―shared processes and predictable products‖

(Cook, 1998: 3). For example, in order to meet strict expectations of timeliness, reporters ―routinize the unexpected‖ (Tuchman, 1972) by following particular beats

(e.g., the White House, Congress, police departments) that mimic the organizational structures of the state and define the divisions of labor within newsrooms. To ensure that excessive competition for news does not isolate reporters from their colleagues or make it too difficult to sell and defend their stories with editors, beat reporters often coordinate and homogenize their coverage, consciously or unconsciously producing leads that echo their fellow reporters‘ interpretations (Cook, 1998: 78-79)19.

Third, Cook demonstrates how public officials and government sources in particular ―dictate conditions and rules of access and designate certain events and issues as important by providing an arena‖ within which reporters are expected to work. By regulating press credentials (signaling certain events to be worthy of selection and exclusion) and providing quotes that easily fit within reporters‘ norms of

19 It is important to note here that Cook‘s account is focused on what Entman (2005) describes as ―traditional journalism‖ conducted by reporters who adhere to five key standards in the production of news: accuracy and factual claim-making, balance and separation of personal views, allowing news judgments to override profit considerations, focusing on checking sources of power, separation of news and editorial staffs. Entman contrasts this type of reporting with advocacy journalism that has a purposeful change-making agenda; tabloid journalism designed for profit, amusement and policy influence; and entertainment targeted to meet the desires of audiences. These categories are not mutually exclusive—and, indeed, are implicitly critiqued in the more sophisticated arguments of Christians, et al. (2009)—but they offer a grounding to better understand the kind of journalism Cook focuses on. 113 newsworthiness, government officials have significant power to suggest what is both interesting and important about state business. Even bureaucratic credentials signal to reporters the quality of information: a more highly placed official is likely to provide perspectives that are more authoritative and, perhaps, trustworthy than another lower- ranked source.

In arguing that the news media constitute a political and social institution,

Cook describes, in institutional terms, the recurring tension between structure and agency, subjectivity and objectivity underlying social phenomena in general and journalism in particular. That is, while journalists ultimately have the ability to make independent judgments about the quality or newsworthiness of information, events or individuals and to produce stories that reflect their judgments, they are understandably influenced by the state‘s own systems of legitimation that are commonly understood to signify the quality, authority and hierarchy of people and information. As such, the news media are not independent producers of news but, rather, one participant of many in a socially situated ―negotiation of newsworthiness.‖ (Cook, 1989: 169)

3. Toward a Sociological Model of Press Autonomy

Equipped with these understandings of how journalism can operate amongst different arenas of activity, it becomes possible to articulate a model of press autonomy that spans Field Theory‘s notions of structure-agency tension and New Institutionalism‘s arguments for socially situated relational power. The table below spans these concepts to articulate not a definition of press autonomy, per se, but a set of five mechanisms and influences through which journalistic independence is asserted and justified:

114

Table 1: Elements of Press Autonomy: Tracing autonomy through configured dependencies.

Field Most generally, the press‘s autonomy depends upon how it relates to other Interactions fields including, but not limited to: state legislation and regulatory agencies; courts; literary and artistic productions and tastes; technological innovations; and managers of financial capital managers. Any discussion of press autonomy exists in relation to these fields.

Capital The press‘s autonomy depends upon its ability to garner both economic Competition capital (resources derived from the market, state subsidies, private funders) and symbolic capital (signals of legitimation from critical acclaim, circulation numbers, awards, recognition from elite decision-makers). These capitals make possible the press‘s illusio or motivating normative stances – the public- oriented, risk-taking, truth-telling values that journalists and press supporters claim distinguish the press from the systems that create these same capitals.

Negotiated The press‘s autonomy depends upon a negotiation between two sub-fields of Production production, each of which provides both economic and symbolic capital: restricted activity that depends upon a small number of patrons with the power to fund and critically legitimate presses oriented higher-risk, experimental journalism; and large-scale information production that relies on consumer- oriented markets to finance a press, signal its populist value, and make claims to representing the interests of mass publics.

Transitioning As much as the press‘s autonomy depends upon flows of capital and Actors intersecting production regimes of different fields, its independence also relies upon interactions between new and old entrants. Younger actors—who may be defined in terms of generational age, duration within a field, or attitudes toward meanings of change and time—may challenge a field‘s existing, dominant norms to simply usurper older, more powerful actors experienced with existing production regimes. Generic calls for newness, change,

115

experimentation and differentiation become ways for newer actors to gain entry into a field and perform broader generational identities.

Public Work The press‘s autonomy depends upon how it performs different aspects of publicness and defines its illusio in terms that implicate publics: how it conceptualizes its interactions with publics and how it defines informational public goods; how visible its activities are and what rationales it gives for private action; what mechanisms and practices it creates and professionally values for interacting with publics; how its conceptions of publicness challenge or reinforce other fields‘ conceptions of publicness (most notably, the state‘s, the market‘s and civil society‘s). How the press performs public work can be a way of visibly differentiating itself from other arenas of activity, earning legitimacy among publics by managing spheres of activity distinct from states and markets.

Note that there is no explicit category for ―technology‖ here since we might best see technological changes as both evidence and drivers of change across these different elements of autonomy.

How has this model appeared in the US press‘s history, and how might we historically situate any contemporary understanding of press autonomy? What actors, resources, practices and ideals have attempted to configure the ―right‖ set of dependencies that lead to press independence? As an introduction to the next chapter that deals more directly with technologies and practices for constructing autonomy in online networked environments, the final section of this chapter uses Field Theory and

New Institutionalist frameworks to trace where and how press freedom has traditionally been negotiated in U.S. journalism.

116

4. Using the Model: Tracing Press Autonomy through Sites of Dependence

4.A. Introduction

A journalist can be described in many different ways. Some descriptions focus on mechanistic and uncritical descriptions of actions. Journalists are seen as

―systematically [keeping] a public record of events in a given time frame‖ (Zelizer,

2005: 66) at locations where people expect news to happen (e.g., during natural disasters and crimes, in city halls and police stations, or anywhere large numbers of people are gathered) in ways that are accurate, fair and balanced (Bennett, 2007: 189).

Journalists are seen as trusted information source serving a public mission: ―sense makers‖ with obligations ―to verify what information is reliable and then order it so people can grasp it efficiently‖ (Kovach & Rosensteil, 2001: 19) and, most fundamentally, to create the news that constructs, maintains and repairs our shared knowledge and ―common sense‖ (Campbell, 1991: xxii; 9). Still others view them more strategically, as skilled communicators who can subtly engage in ―multilingual, code switching from neutral interpreters to guardians of social consensus and back again without missing a beat‖ (Schudson, 2000: 193), often subtly if unconsciously advocating for particular positions that perpetuate social status quos, systematically excluding from their reporting the actions of people and ideas that do not fit within established norms and understandings of the world (Gitlin, 1980).

Being a journalist means negotiating among different identities: they are both individuals with personal backgrounds and perspectives that inform how they see and report on the world (an embodiment of the kind of subjectivity Cassirer and Lewin

117 engage with); they are citizens who can only ultimately claim the same rights and privileges afforded under a First Amendment that is under constant reinterpretation; and they are (official or otherwise) members of a loosely self-regulating professional community (albeit not through formal mechanisms like doctors and lawyers whose licenses and bar admissions are legally regulated) from whom the public expects certain behaviors and ideals (Schudson, 2000); and, whether as staff or stringers, they are employees whose work is judged in relation to a set of organizational, economic and political forces often beyond their direct comprehension or control. Whether they are Pulitzer-winning reporters employed by traditional news organizations, self- employed stringers moving among associations, or unaffiliated bloggers producing reports or commentary without professional or educational credentials, journalists or are essentially always acting in relation to some set of traditions, institutions and ideals that exist beyond their own personal histories and perceptions. That is, mainstream journalists may have professional reputations to maintain, stringers may have reporting contracts to sustain, and bloggers may feel beholden to shifting ethical expectations and information practices in order to gain visibility on the internet – but each type of actor exists within a field of ever changing forces in which they must find and maintain positions.

Although they have long histories as social constructs, economic activities, public and cultural artifacts, sites of technological invention, ideological positions, and means of mass communication (e.g., for surveys see Briggs & Burke [2005], Carey

[1974/1997] Eisenstein, [2005:102-120], Mindich [1998], Schudson [1978, 1998],

Star [2004], Stephens [1988]), the sociological roots of an autonomous (potential,

118 actual or asserted) modern, professionalized U.S. press might best be understood in relation to a series of institutional process that began shortly after the First World War in which the press asserted its autonomy. One history of the modern press‘s autonomy might thus be told using the elements explicated above, traced across three broad conceptual areas: professional identity; organizational routine; and public interaction.

Each concept serves as both a domain and evidence of the press‘s negotiations—i.e., where the press developed and what the press developed into—as it negotiated its autonomy.

4.B. Autonomy through Professional Ideals of Objectivity

Before the 1920s, as Schudson (1978), Kaplan (2002) and Mindich (1998) all argue, journalists (and most other social practitioners) saw little difference between facts and values. The generally held assumption was that the world—in both physical and social forms—could be described in terms of facts that were ―not human statements about the world but aspects of the world itself‖ (Schudson, 1978: 6). Readers and reporters were not seen as an active, socially mediated actor creating understandings and biasing accounts but, rather, as naïve empiricists, discovering realities assumed to be independent of human construction.

Although the history of journalistic objectivity has no exact beginning—

Mindich (1998:11) claims that it ―was certainly an issue at least since 1690‖ while

Schudson (1978:4) argues that ―before the 1830s, objectivity was not an issue‖—the press‘s ideal of objectivity emerged from four factors that lead to a ―scientization‖

(Hallin, 1985) of contemporary US journalism.

119

First, in the aftermath of the First World War, as Lippmann argues

(1922/1997), the perspectives and accounts that journalists received of specific battles—and the war‘s underlying nationalistic rationale—from official government sources were not truthful descriptions of actual events but, often, were stories constructed to persuade publics and elites to hold particular beliefs or attitudes. This use of widespread propaganda suggested to journalists that the democratic institutions they assumed to be acting in the public interest were actually meaning-making organizations interested in advancing their own priorities and values – independent of whatever the ―facts‖ might be.

Writing in 1922 as an already accomplished member of elite intellectual society in the United States, Walter Lippmann (1922/1997) worried that the press simply created ―pseudo-environments‖ (10) that, when inserted between ―man and his environment,‖ created ―pictures which are acted upon by groups of people‖ that, in turn, become ―Public Opinion with capital letters.‖ (18) He essentially argued that the press of the day was simply an instrument of stereotyping (his newly coined term): at best, sloppily assembled half-truths that required more explanatory work and, at worst, purposeful manipulations that required real thought and power to overcome. He argued that the entire system of electoral democracy in which citizens choose their representatives was unworkable without ―an independent, expert organization…making the unseen facts intelligible to those who have to make the decisions.‖ Public opinions, he argued, ―must be organized for the press if they are to be sound, not by the press as is the case today‖ (Lippmann, 1922/1997: 19, emphasis added).

120

Lippmann was calling for a kind of press expertise—and relationship to experts—in which journalists acted on behalf of citizens who would remain largely removed from the press‘s day-to-day activities but who would benefit from the press‘s rational and professionalized oversight of public administrators, and act ultimately as the press‘s critically consuming clients. Although seemingly elitist and anathema

(Carey, 1987) to the kind of critical, populist, communicating democracy envisioned by, among others, Dewey (1920/1963, 1927/1954, 1938/1997), Lippmann actually realistically and optimistically envisioned the press as an instrument of democracy, one of many public bodies that could critically execute and manage some of the investigative and communicative aspects of complex modern democracy on the public‘s behalf, much like the Census Bureau and the Department of Agriculture each has its own area of expertise (Schudson, 2008a). Lippmann expanded on this positive conception of journalists as ―insiders‖ who understood how political systems worked when he wrote in the Phantom Public that

only the insider can make decisions, not because he is inherently a better man but because he is so placed that he can understand and act. The outsider is necessarily ignorant, usually irrelevant and often meddlesome, because he is trying to navigate the ship from dry land… [C]ompetence exists only in relation to function; […] men are not good, but good for something. (Lippmann, 1925/1993: 140)

In this way, Lippmann‘s overarching goal of democratic governance amidst modern complexity was entirely consistent with Dewey‘s review of Lippmann‘s Phantom

Public in which the former wrote that ―[t]he modern state is so large that decisions made and the executions initiated are necessarily remote from the mass of citizens; modern society is not only not visible, but it is not intelligible continuously and as a whole.‖ (Dewey, 1925/2008: 214-215) Dewey and Lippmann both envisioned a

121 public that would ―intervene occasionally‖ (Dewey, 1925/2008: 216) and a professional press that would equip it to do so.

But understanding exactly who would organize public opinion ―for the press,‖ as Lippmann put it, leads to the second element in the process of journalism‘s scientization. Essentially, organizations and politicians began to recognize the value of public relations and the benefit of employing individuals to advocate particular positions in public forums. These public relations professionals were not disinterested reporters or even accountable government officials. Rather, they represented a new kind of intermediary inserted between the elites and the public with the goal of communicating and advancing the perspectives of their bosses. Schudson (2003: 83) reports that in the early 1920s ―50 or 60 percent of stories, even in the venerable New

York Times, were inspired by press agents and that the new Pulitzer School of

Journalism at Columbia University in New York City was churning out more graduates for the public-relations industry than for the newspaper business.‖

Journalists began to wonder whether the public would be able to tell the difference between their reporting and the work of public relations specialists.

Journalists worried about the impact of these new public relations ―parajournalists‖

(Schudson, 2003: 83) not just on their ability to report the news in the service of the kind of democracy Lippmann aimed to protect, but also their ability to distinguish themselves from the emerging field of public relations in order to retain their identity and the unique and economically viable quality of their product, newspapers.

Although journalism is a profession that requires no license or degree and is ultimately dependent upon appeals to the First Amendment shared by all residents of

122 the United States, as journalism schools, codes of conduct and professional associations appeared—American newspaper editors formed their first professional association in 1922, declaring that ―news reports should be free from opinion or bias of any kind‖ (cited in Schudson, 2003: 82)—journalists began the early stages of professionalizing, of articulating a still-contested professional ideal: creating a teachable body of knowledge that takes an academic form; making moral commitments of public service; asserting independence and differentiating their labor from other professions; and associating professional identity with economic earning

(Abbott, 1988; Hatch, 1988; Larson, 1977; Schudson, 1988)20.

Third, in the 1920s social scientists were beginning to refine polling techniques to the point where ―sampling referenda‖ were viewed as objective representations of what the public thought on a particular topic. Pollsters were becoming a kind of intermediary between the governing elite and the mass public, challenging the very notion that journalists or politicians could understand and represent the beliefs and intentions of the citizenry. In 1939, Gallup (1939: 15) wrote that ―through the process of sampling referendum, the people, having heard the debate on both sides of every issue, can express their will. After one hundred and fifty years, we return to the town meeting. This time the whole nation is within the doors.‖

Pollsters argued that they had the objective, scientific techniques that could recreate, on mass scales, the kind of intimate conversations that had traditionally defined deliberative, participatory democracy. And polls let journalists transform ―the subjectivity of opinion into the objectivity of fact,‖ and avoid taking any responsibility

20 Also see Schudson & Anderson (2008) for a good overview of how journalists have situated their claims of authority and professionalism in terms of objectivity. 123 for the quality of public opinion. Such ritualized use of polling by journalists tapped into two fundamental features of modern democracy: a liberal notion that individual expression (aggregated in statistically defensible forms like polls) is the purest form of power in a democratic society made up of rational citizens (Salmon & Glasser, 1995:

444); and the idea that modern democracies are of a scale and complexity that require systems of public opinion that use ―quantification and statistics … to routinize processes of observation.‖ (Herbst, 1995: 12) Even if these processes of aggregation and rationality bear little resemblance to how citizens themselves think about public opinion when given the chance to discuss their experiences and perspectives (e.g.,

Gamson [1992], Herbst [1993], Igo [2007]), the press‘s ritualized use of polling lets them recast the public as ―a statistical artifact‖ (Carey, 1995: 392) that ―makes the public a demographic segment or data set rather than a realm of action.‖ (Peters, 1995:

20)

Opinion polls, public assumptions that they represent reality, and largely inaccessible methodologies that rely on complex statistical techniques let the press distance itself from publics, reporting on (as opposed to with or for) publics and reasserting a practice of objectivity that underpin its claims to professional autonomy.

Fourth, physical sciences were developing increasingly powerful techniques to describe the world, putting pressure on social scientists—and journalists as professional practitioners of social science—to demonstrate that they, too, possessed robust analytic tools of observation and explanation. In his famous critique of professional journalism and its ability to represent public opinion Lippmann

(1922/1997) argues that even though the press has ―come to be regarded as an organ of

124 direct democracy … with the function often attributed to the initiative, referendum, and recall‖ (Lippmann, 1922/1997: 229), the journalist has no analytical tools analogous to the physicist‘s: ―The control exercised over [the journalist] by the opinions of his employers and his readers, is not the control of truth by prejudice, but of one opinion by another opinion that it is not demonstratably less true.‖ (Lippmann,

1922/1997: 227)

Lippmann wanted journalists to be able to derive facts about the social world in the same way that, he believed, physicists were able to discover truths about the physical world. In making such a critique, however, Lippmann did more than decry the professional practices of journalists: he made an epistemological assumption, equating the social worlds that journalists attempt to explain with the material worlds that physicists aim to predict.

Taken together, these four influences led to the ―scientization‖ of journalism: journalists realizing the socially constructed nature of reality and the power of propaganda after the devastation—and information impoverishment—of the First

World War; the growth of public relations as a profession and democratic intermediary that some reporters worried the public might find indistinguishable from journalism; the development of public opinion polls as ―objective‖ sources that journalists could use to cover democratic life; and the pressure on journalists to develop robust methods for explaining social worlds analogous to the analytical tools physical sciences were using to so successfully predict behavior of physical worlds.

Journalism‘s response to these pressures was to develop objectivity as a professional ideal. Objectivity in this sense means that ―a person‘s statements about

125 the world can be trusted if they are submitted to established rules deemed legitimate by a professional community. Facts here are not aspects of the world, but consensually validated statements about it.‖ (Schudson, 1978: 7) This definition properly situates objectivity within a social process and not an ontological certainty, rightly suggests a difference between objectivity and neutrality and a need to appear neutral:

The objective investigator may start out neutral (more likely she is simply good at keeping her prior beliefs from distorting her inquiry) but she does not necessarily end up neutral … As a journalistic virtue, then, objectivity requires not let their preconceptions cloud their vision. It does not mean they see nothing, or that their findings may not be significant and controversial ... The safest way to seem objective, then, may be to look neutral. (Lichtenberg, 2000: 251-252, emphasis in original)

For journalists, objectivity thus became an overarching professional philosophy, an epistemology that guided how they could learn about the world in ways that would be useful to—and respected by—the American masses and elites that increasingly trusted only those perspectives constructed with scientific methods. Understanding the exact forms that objectivity—the ways in which the ―consensually validated statements‖ about the world are agreed upon—is an organizational issue, and a further conceptual site in which the press‘s autonomy is negotiated.

4.C. Autonomy through Ritualized Organizational Routines

This belief in objectivity and this desire for ―scientific‖ journalism manifested itself in journalistic practice through organizational routines and rituals that, implicitly or explicitly, attempted to standardize and codify journalistic practices so they might be defensible against other fields (e.g., in different ways as detailed above, physical sciences and public relations) and help the press assert autonomy. Explicitly or

126 otherwise, U.S. news organizations tried to situate journalism‘s epistemological and professional struggles—how it dealt with objectivity, interpretation, analysis and public relations—within organizational practices and priorities that defined journalistic judgment, made people and events easier to process, and that ultimately ritualized and standardized the production of news.

A host of studies show that journalists craft accounts of the world according to largely unspoken heuristics about what constitutes news (Gans, 1979; Molotch &

Lester, 1974), what publishers expect the news to be (see Breed [1955] for the foundational study and Chomsky [2006] for a more contemporary account) and remarkable similarity among organizations‘ judgments about what news is (Ryfe,

2006). Although reporters may claim considerable autonomy in their individual reporting (Gans, 1979: 101), such independence rarely extends beyond narrowly

―operational‖ decisions (relatively minor decisions within the boundaries of a story on, for example, who to interview, which quotes to use) to ―allocative‖ issues (more fundamental decisions about the conditions under which the news organization operates, for example, what sections a newspaper should have or who should be hired or fired) (Murdock, 1977). Moreover, Gans (1979: 101) found in his study of news organizations, that although journalists frequently claimed the right to make their own news judgments free of any hierarchical instruction, ―the suggestions of powerful superiors are, in fact, thinly veiled orders.‖

News organizations work hard to ensure that the events, locations and sources they use within their reporting fit well within the rhythms—the beats—they use to coordinate the production of news (Gans, 1979; Tuchman, 1973). With the possible

127 exception of long-term, high-risk investigative reporting, it becomes difficult if not impossible for a journalist, editor or publisher to convince their colleagues or publics that a story can exist outside a beat, or any of the other criteria that have traditionally defined enterprise news production: conflict, timeliness, proximity, prominence and the unusual (Stephens, 1988: 32-33). Stories appear in newspapers because, essentially, they could be easily understood ―in the context of what has gone before and anticipated in the future‖ (Molotch & Lester, 1974: 101): journalists were more likely to cover routines, accidents, scandals and even serendipities that more easily cohered with their deadlines, their editors‘ expectations, and what their readers had come to understand as the news.

One source of evidence is the increasingly common use of ―strategic rituals‖ of objective reporting that developed in the years following the 1920s that are still largely present in contemporary newsrooms (Bennett, 1996; Ryfe, 2006). Reporters used these ―rituals of objectivity‖ (Tuchman, 1972, 1973, 1978) and reliable relationships to bureaucracies and predictable sources (Fishman, 1980; Sigal, 1973, 1986) to ―orient man and society in an actual world.‖ (Park, 1940: 669) News was understood to be an informational product with social meaning and public value created by organized professionals for the consumption of audiences with well-understood preferences and expectations. Tuchman (1978) argues that journalists use specific practices to protect themselves against claims of biased reporting, specifically: conflicting truth claims

(constructing stories by juxtaposing not their own interpretations but the statements of socially sanctioned experts); attribution (distancing themselves from the ideas of others by overtly marking text as originating from an external, named source);

128 additional facts (overwhelming the reader with information to buttress themselves against claims that their stories lack research); the inverted pyramid (organizing a story in the professionally accepted form in which the most important information comes at the beginning of the article).

In a kind of update to Tuchman‘s study, Bennett (1996) found that most contemporary journalists must still follow certain implicit rules of production, uncritically accepted practices that: ―base stories on official sources; index views according to the magnitude and content of conflicts between these sources; follow the trail of power; narrate stories according to the prevailing customs of the political culture; and the emergence of credible dissenting icons may license journalists to move beyond official sources in their coverage‖ (Ryfe, 2006: 204-205 citing Bennett,

1996). Adherence to such rules, Bennett further found, means that much of the mainstream press‘s function is simply to ―index‖ (Bennett, 1991) existing debates, tying ―story frames to the range of sources and viewpoints within official decision circles, reflecting levels of official conflict and consensus.‖ (Bennett, Lawrence et al,

2006: 468) Essentially, if elite sources are not debating an issue, it is virtually impossible for news organizations to comment on it. To do so would be to acknowledge press interests that are somehow separate from those of elite decision- makers and publics.

Even when publics and social movements emerge to challenge elite consensus, the press has difficulty acknowledging the legitimacy of these activities. As further evidence of ritualized reporting, Gitlin described how the press‘s reporting implicitly adheres to uniform, hegemonically constructed ―media frames‖:

129

persistent patterns of cognition, interpretation, and presentation, of selection, emphasis, and exclusion by which symbol-handlers routinely organize discourse, whether verbal or visual. Frames enable journalists to process large amounts of information quickly and routinely: to recognize it as information, to assign it to cognitive cateogories, and to package it for efficient relay to their audiences. (Gitlin, 1980: 7)

Gitlin saw evidence of such frames in how mainstream media reporters and oppositional social movements interacted during the 1960s: political groups learned that, in order to gain the kind of power that comes with national media attention, they needed to submit their own goals, self-narratives and events ―to the implicit rules of newsmaking‖ and ―journalistic notions … of what a ‗story‘ is, what an ‗event‘ is, what a ‗protest‘ is.‖ (Gitlin, 1980:3) Such submissions to media frames, Gitlin argues, is a kind of press hegemony in which ―one picture of the world is systematically preferred over others.‖ (257) Such techniques let journalists distance themselves from stories, claiming that the stories they write are not their own accounts of the world but, rather, accounts of other people’s accounts of the world.

This distinction between a journalist‘s story and a source‘s account highlights the distinction between the ―professional communicator‖ and the ―individual interpreter.‖ In Carey‘s (1997) use of the terms, the rise of the professional communicator paralleled the rise of science in the 1920s. The professional communicator distances herself from her data and uses techniques of objective reporting (like those identified by Tuchman) to claim that a story‘s validity arises not from the journalist‘s own interpretation but from the veracity of the information provided to her by socially sanctioned sources. The job of the professional communicator is to seek out or respond to these sources (e.g. institutions like the

White House, police stations and government agencies that have been pre-sanctioned

130 as socially reputable) and to use openly accessible forms of language— ―elaborated codes‖ (Bernstein, 1962)—that allow non-specialists to understand a story‘s meaning.

This distinction between professional communication and individual interpretation mirrors a distinction Schudson makes between the ―information‖ and

―story‖ ideals of journalism (Schudson, 1978: 88-120). In the information ideal, newspapers are unique genres of literature that provide readers with accounts that are understandable in themselves, that require no special intelligence or skill to discern and that—rightly or wrongly—are associated with ―fairness, objectivity and scrupulous dispassion‖ (Schudson, 1978: 90). In the story ideal, newspapers are seen as sources of entertainment that create ―satisfying aesthetic experiences which help

[readers] to interpret their own lives and to relate them to the nation, town, or class to which they belong.‖ (Schudson, 1978: 89)

(In terms of broader communication theories, the information ideal might best be thought if in terms of a transmission model in which messages are ―transmitted and distributed in space for the control of distance and people‖ [Carey, 1989: 15] while the story ideal might be considered closer to a ritual model in which communication functions to maintain society and represent shared beliefs [18].)

The essential features of the professional communicator are, in a sense, to perform Tuchman‘s strategic rituals of objectivity in her professionalism, to adhere to

Bernstein‘s elaborated codes in her writing, and to embody Schudson‘s information ideal in her reporting21.

21 The value of such reporting traditions, steeped in a professionalized goal of objectivity, is debatable and not universal among reporters. Hallin & Mancini (2004: 226) report that norms of objectivity can make reporters less autonomous within the global profession: European journalists ―would say that the 131

There are ways, though, for a professional communicator to periodically behave like an individual interpreter—to enact the ―story ideal‖—highlighting tensions within journalism between the ideals of professional objectivity and the practices of individual storytelling. There are two senses in which tensions between the standards of professional journalism and journalists‘ individual interests and backgrounds manifest themselves: in the language reporters use to subvert norms of objective reporting, and in the policies their newsrooms use to promote workforce diversity.

Glasser and Ettema (1993), for example, describe how journalists can use irony to ―quietly and discretely‖ pass judgment on the ―facts‖ they report. Their analysis of journalist Lou Cannon‘s 1985 story about President Reagan‘s statement that the South

African apartheid government had ―eliminated‖ segregation demonstrates that skilled journalists have some freedom—albeit by using ―restricted‖ rather than ―elaborated‖ codes (Bernstein, 1962)—to infuse professional communication with personal interpretation. Journalists, for example, may use selective quotes that report a source‘s words verbatim but that simultaneously embed them in framing language that, to the observant reader, reverses their meaning. Irony can thus become a language game, freeing the reporter from the usual professional norms and standards of objectivity and inviting the reader to reinterpret the newspaper‘s information genre as a more subjective space for storytelling.

Similarly, Lipari (1996) shows how journalists can skillfully use adverbs to strategically legitimate certain claims, guiding readers into particular readings that

Americans were ‗unprofessional‘ because they were so constrained by the routines of balance and ‗objectivity‘ that they didn‘t exercise independent judgment.‖ 132 contradict the supposedly objective, disinterested tone of newspaper stories. For example, she notes how journalists can use the word ―obviously‖ to connote an empirical claim (inviting readers into shared understandings based on presumed common experiences), an inferential claim (assuring readers that the claim is logical and causally follows), or a metalinguistic claim (linking attributed facts with unattributed inferences readers are presumed to share).

Although professionally sanctioned practices of objectivity may seem to bind reporters into the information ideal, skillful journalists can sometimes connect with attentive readers in ways that subvert the genres and norms of objective reporting.

Such insights into journalism as an organizationally situated phenomenon suggested that the press—and the news it produces—is a socially constructed entity with shared routines, rituals and norms that cannot be fully explained by the rational actions of reporters and organizations, or critical information-seeking audiences. That is, the news is made—and made meaningful—by interdependent, organizationally situated entities: professional reporters and their sources; schools and companies that train and employ reporters; markets and investors who value news organizations; publishers who set organizational directions and control operational resources; consumers who buy news and citizens who vote based on news; legislatures and courts that regulate reporters and organizations.

What kind of press autonomy emerges amongst these organizationally situated practices and norms? Although journalists may have considerable control over individual decisions made while reporting a story—free from close editorial scrutiny during an interview or while writing—and may benefit considerably from their

133 organizations‘ resources—free to engage in reporting that requires time, money and access for its success—journalists‘ autonomy is largely situated within organizational practices and norms that are often left largely unexamined. The news they thus produce emerges from structures—media frames, strategic rituals, the professional communicator stereotype, and the information ideal—that rely upon professionally rooted organizational forms for performance and replication.

4.D. Autonomy through Strategic Public Distancing

As much as the press defines and asserts its autonomy through professional cultures of objectivity and organizational rituals of news-making, the press must also conceive of its independence in relation to publics that consume and rely upon news. Sometimes journalists have assumed that the public is simply interested in whatever journalists are interested in, and are content to write ―for their superiors and for themselves‖

(Gans, 1979:230). Or, as Darnton (1975: 182-88) chronicles, reporters often also write for a personal reference group (―the inner circle of a reporter‘s public‖): not only reporters and editors at his own paper, but also who he considers to be competing reporters at other papers, friends and family members who regularly read their stories, reporters who are no longer practicing but are still present in his memory, and sources and subjects who regularly appear in their stories.

At other times journalists have a ―fear of the audience‖: the responsibility of producing news for so many people and a concern that ―if audience wants were considered, journalistic news judgment would go by the wayside‖ (Gans, 1979: 235), impacting both press independence and news quality. Although they may seek

134 informal feedback from friendly non-journalists—the ―known audience‖ of friends, neighbors and family members (236)—journalists are largely dismissive of readers‘ letters, seeing the readers who write letters as constituting an ―idiom of insanity‖

(Wahl-Jorgensen, 2002) that should be discounted for its own good. In order to justify their distance from publics and retain the freedom and power that come with being

―self-appointed and unaccountable audience representatives,‖ (Gans, 1979: 238) the press needs to bracket publics—both conceptually and practically—as audiences: ―a receptacle to be informed by experts and an excuse for the practice of publicity.‖

(Carey, 1995: 391-392) As long as it has the power to describe the public on its terms, the press can set up a strategically circular set of rituals in which the press decides the public‘s interests, reports on them and then selectively publishes public comments on the news that reassert the very need for an independent, caretaking press that ―justifies itself in the public‘s name but in which the public plays no [significant] role‖ (Carey,

1995: 391). Largely absent from these press-controlled conceptions of the public is a sophisticated self-reflection on what role the press ought to play in public life.

Two sets of quotations represent a fundamental tension between conceptions of press autonomy and publics. The first is by former Federal Communications

Commission chair Mark Flower who said that ―the public interest is that which interests the public‖ (cited in Hallin, 2000: 234) and that the state ―should rely on the broadcasters‘ ability to determine the wants of their audiences through the normal mechanisms of the marketplace‖ (Fowler & Brenner, 1982: 210). The second is the call made by the Hutchins Commission on Freedom of the Press to ―free the press from the influences which now prevent it from supplying the communication of news

135 and ideas needed by the kind of society we have and the kind of society we desire.‖

(Commission on Freedom of the Press, 1947: 79) Essentially, in claiming its autonomy, the press must decide how it is going to envision and interact with publics.

The tension captured in these quotes is a difference between two ways the press can view its readers. In the first, the press can see its job as informational and service-oriented: to sense, respond to, and meet public desires for information within a largely market-oriented system in which ―consumers‘ desires drive news coverage.‖

(Hamilton, 2006: 7; see also Baker, 2001) Market-oriented journalism tends to have six characteristics: consumers define news‘ quality and value; products judged to be quality will drive other products from the market; self-correcting markets will produce the information consumers lack; producers are under constant pressure to create the appearance of new, improved goods that situate them in relation to other producers; scarce resources will always go to producers consumers consider to be more valuable; and consumers are always free to choose among products without coercion.

(McManus, 1994: 4-5) In this model—most closely associated with a ―competitive‖ model of democracy ―through which individuals, acting alone or in concert with others, build support for what they want‖ (Christians, et al., 2009: 94)—the public is seen as an aggregation of individual citizens with their personal preferences. The public becomes a collection of individual information-oriented clients for whom the press produces commodities – but it only does so in proportion to citizens‘ desires for them.

News organizations argue that such market-oriented motivations give them an important kind of independence: because they are beholden to no set of identifiable

136 interests that might interfere with their editorial judgments, they can exist in relation to value-neutral markets whose patterns emerge from the independent actions of individuals with preferences. Press autonomy is intact because it is only responsible to market forces that lack ideologies22.

But there are at least two ways in which market characterizations of the publics actually hinder journalist autonomy. First, markets tend to reward choice among goods, but not necessarily diversity among products. It is easier to navigate a marketplace and earn market share if the set of competitors is small and known, and if competition exists around only a few, discrete, easily measurable set of elements. If products differ too much, it becomes hard to create a language of competition within which choices can be compared. Second, for journalists to survive within such markets they must learn which factors consumers can differentiate amongst and compete on these bases. This creates a ―scoop and shun‖ phenomenon in which journalists are, for example, more likely to focus on producing stories that are more timely—it is easy to recognize which news organization published a story first, regardless of how a story‘s urgency relates to its importance—or stories that ignore what their competitors may have reported—it is likewise more difficult to distinguish and defend exactly how your reporting on a particular topic is different from another journalist‘s and much easier to simply write a different story. Lost in this focus on autonomy through marketplace competition is any concept of what publics— conceptually independent of consumers or audiences—may need or want from their news organizations (Glasser & Gunther, 2005).

22 For an historical perspective on this idea, see Schudson (1978). 137

In the second view, the press‘s job is to sustain the conditions and relationships necessary for civic life at the scale modern democracies operate. Essentially, the press‘s economic model, its conceptions of citizenship and its public interactions should always be service of a republican ideal of democracy characterized by ―the subordination of egoistic concerns to the public good, and by the subsequent opportunity this creates for the expansion of welfare, individual and collective.‖ (Held,

2006: 43) This welfare can be worked out in many different kinds of public spheres, including:

 a truth-seeking and, in principle, inclusive one in which differences are

bracketed, rationality is paramount, discussion focuses on matters of

objectively common interest and where the ―media are used to create occasions

for consumers to identify with the public positions or personas of others‖

(Calhoun, 1992: 13-26; Habermas, 1989);

 a ―weak‖ one in which ―deliberative practice consists exclusively in opinion

formation‖ or a ―strong‖ one ―whose discourse encompasses both opinion

formation and decision making‖ (Fraser, 1992: 134, emphasis added);

 a ―decentered‖ sphere or ―sphere of publics‖ (Haas, 2004) that ―includes

differently situated voices that speak across their difference and are

accountable to one another‖ (Young, 2000: 107) and not a ―comfortable place

of conversation among those who share language, assumptions, and ways of

looking at issues‖ (111);

 or ―counterpublics‖ that are ―by definition, formed by their conflict with the

norms and contexts of their cultural environment‖ (Warner, 2005: 62) and that

138

have ―at some level, conscious or not, an awareness of [their] subordinate

status.‖ (56)

Regardless of what kind of public sphere such a model of the press assumes or advocates for, this second view sees the press‘s primary responsibility as social in that it that is somehow accountable to, or in negotiation with, publics, and not simply statistical aggregates of citizens derived from polls or market studies, or the caricatures of citizens that seem to pervade journalists‘ understandings of their readers. Essentially, a public-oriented conception of press-public gives the press a different kind of freedom steeped in editorial judgments that allow journalists to, for example, take stances on what truths and public goods should attract public attentions

(Marvin & Meyer, 2005). To take such positions—even within the kind of ―mature subjectivity‖ (Schudson, 1978: 192) and ―neutrality‖ (Lichtenberg, 2000) advocated for by those who do not reject the ideal of journalistic objectivity out of hand—is to depart from the kind of independence that comes from market-oriented journalism, recasting it instead within a judgment-focused that press that is transparent and accountable in its work.

In practice, newspapers and television stations have translated these ideals into a variety of mechanisms: sponsoring open public forums for semi-structured debate on community issues; commissioning polls before and during election times to gauge citizens‘ opinions on issues; conducting focus groups and forming citizen panels to discover community issues that may be absent from press coverage; soliciting questions from readers that journalists can then ask political candidates; devoting reporters‘ time and resources to issues developed through citizen contact; and giving

139 citizens information and opportunities to help them become involved in political processes (Glasser & Lee, 2002: 216).

To ensure that these community engagements impact the ethos and contents of news publications means changing newsroom cultures. Glasser & Lee (2002: 211) describe a set of principles guiding newsrooms that practiced ―public journalism‖: reporters should look for sources with moderate views, not extreme opinions; coverage should be framed in terms of people‘s daily experiences instead of officials‘ policy priorities; stories should use emotions to explain how people arrive at positions instead of merely to add ―color‖; reporter should uncover the values that people bring to issues, seeing them not just as sources of conflict; a citizen‘s knowledge should be valued alongside an expert‘s; and stories should explain why citizens should care about an issue, and reporters should offer people opportunities to take action and design solutions.

In addition to these changes in reporting styles and newsroom practices, publications following public journalism principles might make organizational changes. A common restructuring involves replacing the traditional beat system—in which reporters tend to rely on a limited set of bureaucratically reliable official sources (Fishman, 1980; Sigal, 1973, 1986)—with a ―reporting circle‖ in which reporters from different beats collaborate on a variety of broad themes (e.g., individual quality of life, city values and governance, personal leisure) that correspond to community priorities instead of newspaper sections (Johnson, 1998).

A critical question emerging from such optimism in citizens‘ abilities, and faith in (and nostalgia for) community presses is what kind of autonomy, if any, is left

140 for the press as an institution and journalists as individuals? That is, if autonomy is demonstrated by—and protected by claims of—journalistic expertise and authority then public journalism may be seen as an erosion of the kind of professional responsibility and independence journalists have historically claimed. By ―denying the press the authority to set its own agenda, public journalism substitutes the community‘s judgment, however defined, for the judgment of journalists‖ and, in doing so, ―deprives the press of an opportunity—and diminishes the importance of its obligation—to set forth, clearly and convincingly, its politics.‖ (Glasser, 2000: 684-

685)

Throughout these phases of US journalism—an early embrace of professionalized objectivity, a subsequent coding of objectivity in organizational routines, and a relocating of objectivity within public-oriented mechanisms—the true dynamics constituting the press‘s autonomy have been largely hidden: journalists have been free to ―systematically frame the news to be compatible with the main institutional arrangements of the society … thus sustain[ing] the dominant frames through the banal, everyday momentum of their routines. Their autonomy keeps them within the boundaries of the hegemonic system.‖ (Gitlin, 1980: 269)

5. Conclusion

The aim of this chapter has been to locate the press as a field of activity that negotiates its autonomy through a complex set of social interactions. This chapter has had three main contributions. First, I aimed to ground the idea of press autonomy in Field

Theory and New Institutionalist approaches to journalism to demonstrate how fields

141 essentially constitute themselves through negotiations between subjective agency

(how individuals interpret and describe their worlds) and objective structures (how organizations systematize, afford and constrain behaviors and truth-telling). We can best understand the conceptual underpinnings of the contemporary US press and its assertions of autonomy if we know that such concepts and claims are grounded in communication sociology understandings of how actors shape and change fields of practice.

Second, I summarized and extended this comparative review into a model of press autonomy that draws upon Bourdieu‘s Field Theory and New Institutionalist sociology. I defined press autonomy as a set of dependencies among five economic, sociological and public concepts: interactions with other fields; competition for economic and symbolic capital; negotiated production among small- and large-scale activities; transitioning actors vying for field positions; and values that implicate publics. My aim in this model was not to define press autonomy as a static or ideal concept that might be attained but, rather, to see it as an active contest amongst elements of the journalistic field.

Third, I then used this conceptualization of press autonomy to review how practices and debates in which the press has asserted its autonomy: through its construction of a professional identity grounded in performed objectivity, essentially a way of achieving autonomy by establishing the ―scientistic‖ conditions on which journalists would engage the world and tell truth; through its use of organizational routines to systematize and ritualize the construction of news, fundamentally a way of achieving autonomy from the people, events and issues journalists aimed to cover;

142 and, finally, through its design and use of public interactions methods in which journalists simultaneously distanced themselves from, and embedded themselves within, their publics.

Although not an exhaustive history of American journalism, this review is an exercise in applying the model of autonomy derived above to historical phases of the

US press. The press‘s early differentiation from the public relations profession and its later experiments with public journalism might be seen as attempts to articulate illusio or public-oriented norms that spoke to broader cultural trends. The press‘s organizational rituals of reporting helped standardize its practices and relations to other fields so that it could secure economic capital (e.g., news organizations that created predictable products for segmented audiences could rely upon stable streams of advertising revenue) and symbolic capital (e.g., news organizations that organized their work in relation to Pulitzer categories, the priorities of New York Times, or impacting elite decision-making readers could claim a critical legitimacy and relevance).

Today, while all these mechanisms of autonomy—field interactions, capital competition, negotiated production, transitioning actors, and public work—are still relevant to journalism and the press‘s production of news, they are taking new forms.

In particular, the next chapter explicates how a new kind of press autonomy— grounded in the historical context of the story told here—is being negotiated at technosocial boundaries inhabited by tools, practices and emerging norms that aim to mediate press-public relations: a set of networked algorithms, interfaces, uses and expectations I call ―newsware.‖

143

CHAPTER FOUR:

Negotiating Press Autonomy in Contemporary Online Journalism

1. Introduction

The chapter aims to address this question: how can the mainstream press‘s autonomy be understood in terms of contemporary, online environments? Specifically, how does the online press strategically negotiate its distance from publics, states, and markets— e.g., asserting independence from participants, government information sources or financial models—while simultaneously maintaining a reliance upon these same actors – e.g., interweaving its production processes with online systems for public participation in the news publishing, monetary valuation of digital content, and use of legal doctrines to protect its intellectual property in online environments?

In contrast to earlier chapters that examined the philosophical underpinnings of the idea of autonomy, legal and jurisprudential understandings of press autonomy, or historical, field-theoretic negotiations of press autonomy, this chapter attempts to situate the press‘s autonomy in relation to a new set of emerging relationships among distributed actors that share, to certain extents, the same broad news infrastructures— what I call ―newsware‖—for journalistic work and news production.

This chapter has three main parts. The first part—the bulk of the chapter, with five subsections—situates the dissertation in terms of contemporary dynamics in the

U.S. press. Since much of this project is essentially a response to changing contemporary circumstances, the goal of this section is to ground and shape the project in terms of specific conditions in which the press is operating. It is essentially answers

144 the questions: why is this idea of networked press autonomy timely and relevant, how is press autonomy being negotiated in the online press today, and what are the elements of newsware?

This part is organized around a discussion of five features of the contemporary

American press: press economics, reader behaviors, professional practices, intellectual property and organizational systems, distributed human-machine intelligence.

Although this section offers considerable detail on various elements of online press dynamics, I am careful to focus on patterns among them and not to become too narrowly focused on the particular circumstances of any one element of the system of autonomy. I weave throughout this part not only a description and contextualization of this system, but why they matter to the dissertation‘s larger argument that press autonomy is a networked concept that exists in relation to a set of information technologies and practices.

The second part of this document reflects upon the first section to more completely define ―newsware‖ as a set of institutionally embedded technologies and practices—an infrastructure—through which press autonomy is negotiated. These socio-technological systems work to structure relationships among online readers, journalists and technology designers and, thus, define the terms on which online press autonomy is negotiated. I argue that understanding the contemporary nature of networked press autonomy requires a nuanced understanding of how and when journalistic practices intersect with—are reflected in and shaped by—networked information technologies. It is in the practices and norms, the ―artful integrations,‖

(Suchman, 1994: 34-35) that exist in and around information technologies that we

145 might distinguish the press from others, better understanding how and when it knowingly or unknowingly asserts its autonomy, and point to normative arguments for modern press autonomy.

2. Autonomy & The Contemporary, Online Press

A review of the last three years (2008-2010) of the Pew Project in Excellence in

Journalism ―State of the News Media‖ reports shows considerable change in how the news media are funded and owned, how journalists perceive their roles, and how audiences and readers are behaving. In addition to these trends, there are changes in how news media are determining and signaling the value of their news products, how news organizations are relating to each other and how networked technologies are both shaping and reflecting new distributions of journalistic practice between automated, machine processes and human, editorial judgments.

It is important to note two distinct scales of description here. Some statistics presented describe industry-wide phenomena that show broad trends in how news is commodified, produced and consumed. Other data discussed here are more like examples, seemingly anecdotal or idiosyncratic in nature, but with characteristic details that present particular instances of change in detail. Taken together, these descriptive statistics and behavioral examples are meant to tell a story of how online news in the U.S. is produced today. This story is organized into five interrelated categories: press economics, reader behaviors, professional practices, organizational systems, distributed human-machine intelligence. The aim here is not only to describe these categories but to highlight within them how their social, legal and professional

146 features relate to an understanding of press autonomy that is distributed among actors and negotiated through newsware.

2.A. Press Economics

The funding model that has historically underpinned mainstream print journalism has changed dramatically in the last few years. In his summary of National Newspaper

Association and Pew Research data, Varian (2010) summarizes four descriptive statistics that, together, demonstrate the main trends in the news industry‘s economic decline:

i) newspaper advertising revenue has historically followed trends in GDP growth

and decreased during recessions but, since approximately 2000, has departed

from this trend, dropping precipitously in relation to GDP (Figure 1);

ii) all types of newspaper advertising revenue are decreasing with the only local

retail advertising staying essentially flat (Figure 2);

iii) newspaper circulation per household is decreasing significantly (Figure 3);

iv) more Americans are increasingly getting national and international news online

and not from print newspapers (Figure 4).

147

Figure 1: Newspaper Advertising Revenue & GDP

(National Newspaper Association, as cited in Varian [2010: 6])

Figure 2: Newspaper Advertising Revenue (in USD$ millions) by Type of Source

(National Newspaper Association, as cited in Varian [2010: 7])

148

Figure 3: National Newspaper Circulation Per Household

(National Newspaper Association, as cited in Varian [2010: 11])

Figure 4: Yearly responses to question ―where do you get most of your national and international news?

(National Newspaper Association, as cited in Varian [2010: 12])

Together, these statistics suggest that newspapers‘ revenue declines are not simply a symptom of the more general economic recession; that newspapers‘ main advertising sources are decreasing across almost all categories and that online advertising is not sufficient to make up for lost print revenue; that newspapers‘ circulation numbers are

149 continuing a multi-year decline; and that people are increasingly seeking news from online environments, not newspapers.

These data suggest that the news industry‘s economic changes are primarily a tension between a newspaper‘s material form (as a physical object with limited space that is expensive to produce and distribute) and its ability to connect advertisers with audiences they wish to reach. That is, there are now cheaper, faster and easier ways for advertisers to reach readers. Online advertisements may represent a new category of earnings for print newspapers‘ online editions, but their revenues are dwarfed by the newspaper‘s historical earnings from print advertisements, classifieds and circulation. And since consumers have largely grown accustomed to online news being free (Fallows, 2010; Starr, 2009) (aside from relatively niche finance and international oriented publications like the Wall Street Journal and The Economist), newspapers have a double challenge that impacts both advertising and circulation:

Newspapers have been able to make money from their print editions at both ends: by charging advertisers for eyeballs, and by charging the eyeballs, too. But online there are other news sources such as sites run by TV and radio stations, which have never charged their viewers or listeners. So, for newspapers, there goes circulation as well as advertising income. (Starr, 2009: np)

As most newspapers lose their value to advertisers (small community newspapers with the ability to provide locally targeted print advertisements seem to be a current exception), they also lose the main revenue stream by which their public interest, investigative reporting has historically been subsidized (Starr, 2009). As a recent

Federal Trade Commission report on the ―reinvention of journalism‖ put it, consumers have lost the ―ancillary benefit‖ of ―receiv[ing] news about a wide variety of topics,

150 including important public affairs‖ (Federal Trade Commission, 2010a: 2) that came with communing around newspaper advertisements.

There is a great deal of interest in restructuring news business models, especially creating ways to produce the kind of news that seems to have lost its advertising patrons. One effort exists within technology organizations themselves, among online indexers who argue that they need high-quality professionally produced news to index so that they can create advertising markets, charging more for advertisers who wish to appear beside content that consumers consider to be high-quality and valuable. Google assumes that if the information it indexes is ―uninteresting, inaccurate or untimely, people will not want to search for it,‖ (Fallows, 2010: np). It has therefore begun to help news organizations with online

distribution, engagement, and monetization. That is: getting news to more people, and more people to news-oriented sites; making the presentation of news more interesting, varied, and involving; and converting these larger and more strongly committed audiences into revenue, through both subscription fees and ads. (Fallows, 2010: np)

While news organizations may indeed benefit from such help, Google‘s mission to redefine news economics begs questions that have traditionally been the purview of journalists. For example, how does Google define what ―news-oriented sites‖ are?

How are such sites judged to be ―more interesting, varied, and involving‖? And how will volume-oriented metrics—focused on driving traffic to news sites—that can count how many people read particular stories mesh with other measures of journalistic quality, e.g., professional awards like Pulitzer Prizes, bylines of particular reporters, or the brands of news organizations like The New York Times or the Wall Street Journal that have cultivated reputations as high-quality publications? While Google News

151 recently experimented with letting editors manually place stories (Garber, 2010a), running an experiment called ―Living Stories‖ (now available as open source code) that tweaked its index to prioritize content from The New York Times and The

Washington Post (Fallows, 2010; Google, 2010a), it mostly ranks content using a proprietary algorithm. It states that the ―selection and placement of stories on this page [are] determined automatically by a computer program‖ (Google, 2009a) that selectively indexes a ―news site or blog‖ according to an unpublished standard

(Google, 2009b).

The press is also developing new funding through relationships with philanthropists and foundations. While nonprofit news organizations have long existed in the U.S. and elsewhere (Overholser, 2006), news organizations only relatively recently began to rely on philanthropies directly to support news production.

Foundations are subsidizing existing newspapers (e.g., the Scott Trust‘s support of the

Guardian Media Group), underwriting new journalistic ventures (e.g., the Sandler

Foundation‘s funding of ProPublica.com, which conducts and then shares content with mainstream news organizations), supporting new media organizations (e.g., the

Atlantic Philanthropies‘ co-founding of the Huffington Post‘s new Investigative Fund to support original investigative journalism), and even creating entirely new news services narrowly focused on topics that interest foundations (e.g., the Kaiser Family

Foundation‘s independent and free news service that reports exclusively on the U.S. healthcare system).

Such examples suggest that philanthropic foundations are not simply sources of funding that might help traditional news organizations weather this period of

152 economic change. Rather, they may become active participants in defining what kind of news is financially sustainable. In considering the press‘s autonomy, it is worth asking what new relationships or dependencies emerge from such funding models, and how foundations‘ different missions, ideologies, and issue-specific focuses may fit with historical ideals of the professional press.23

In addition to industry and philanthropic efforts to influence press economics, the

Federal Trade Commission (FTC) is beginning to explore public policy initiatives to support the production of professional quality news. A recent FTC report (FTC,

2010a) suggests a wide range of policies designed to spur production of high-quality news: everything from increased direct government subsidies for the press (a long- standing but little acknowledged part of the U.S. media system [Cook, 1998]), to more indirect tax regimes on the devices and spectra that deliver the news and new IRS tax statuses for news organizations. Other such proposals include:

i) Federal ―hot news‖ legislation to protect, for a short period of time, news

producers from unauthorized borrowing of facts—distinct from copyright

protection for news producers‘ expressions of facts—by news aggregators who

repackage original reporting without permission, attribution or payment.

ii) An amended Copyright Act ―fair use‖ doctrine to prevent the ―routine copying

of original content … by a search engine in order to conduct a search

(caching)‖

iii) Statutory licensing fees, administered by internet service providers, that let

newspapers charge those who use their content, much like the Copyright Act

23 See Guensburg (2008), Kosterlitz (2008) and Lewis (2007) for good general overviews of issues in nonprofit news production; and Bollinger (2010) and McChesney & Nichols (2010) for proposals of publicly funded U.S. news media. 153

currently regulates those who ―make and distribute phonorecords‖ and

―jukebox operators‖ (FTC, 2010a: 12)

Embedded in each of these proposals is an assumption about what constitutes the press, and press autonomy. That is, in the absence of a First Amendment challenge, such subsidies proposals would de facto separate the press from other industries and activities, begging questions about how to define a news organization or a news aggregator.

These proposals are one example of how the state, in an effort to support a type of journalism that appears to be losing its financial footing in online environments, is implicitly defining the press through subsidies that distinguish among news organizations. Such proposals have been criticized not only by the information technology companies who aggregate content and stand to lose from such restrictions

(Google, 2010b), but from news producers themselves who note that they often involve rely upon the very aggregators and search engines that such policy recommendations are meant to regulate (FTC, 2010a). Such regulations, they argue, might make it harder, not easier, for them to produce high-quality journalism because they may lose the newfound underwriting of information indexers who need to rank quality content.

To be sure, the press‘s economic viability likely requires some involvement from information technology companies, but it is important to distinguish between systems that evaluate and commodify large amounts of content in economic terms, from those that signal journalistic quality through other means (like professional prizes, organizational reputations, and the ability to influence elite decision-makers).

154

The online press‘s autonomy is, in part, determined by how it negotiates between such systems of valuation: how it depends upon funding from industries, foundations and governments, and how it critically questions how such relationships impact its own norms and goals and separates itself from such actors.

2.B. Reader Participation

To understand how and why press autonomy is changing means appreciating not only its economic challenges, but also how relationships between mainstream news production (traditionally the function of journalists who created news) and consumption (historically the responsibility of audiences who read news) are changing. That is, at the exact moment that the internet has made it cheaper and easier for advertisers to bypass newspapers in their search for interested readers, it has also demonstrated new ways for readers to interact with the news, highlighting the mainstream news industry‘s historically top-down, broadcast logic in which a few stories are told to many different readers.

As much as the internet has challenged the press‘s economics, it is also putting into question how readers can and should interact with journalists and the news. What clues do such relationships offer to the networked nature of press autonomy?

The very distinction between information production and consumption—itself the product of broadcast media logics in which a few messages were delivered to large and mostly undifferentiated masses—has, in online contexts, been challenged by those who describe internet users as ―prosumers‖ (Toffler, 1980)—individuals who both produce and consume content—who take part in a ―participatory culture.‖ Jenkins

(1992) defines this culture as having:

155

―1. relatively low barriers to artistic expression and civic engagement, 2. strong support for creating and sharing creations with others, 3. some type of informal mentorship whereby what is known by the most experienced is passed along to novices, 4. members who believe that their contributions matter, and 5. members who feel some degree of social connection with one another.‖ (Jenkins, 2009: 7)

In one sense, such prosumption is not new; non-specialists have always helped to create content for mass audiences. Consider what Downie & Schudson (2009) call the

―adjunct reporting‖ of census-takers, Center for Disease Control health workers, and

Audubon Society scientific bird watchers who have historically gathered information for review and publication by centralized authorities. Or note the history of letters

(both published and unpublished) written by non-journalists to editors critiquing news stories or expressing opinions on matters they perceive to be of public concern (Wahl-

Jorgensen [2001, 2002]). Or consider interviews by journalists of non-journalists who then become central sources and authorities within news stories (Darnton, 1975: 182-

85; Gans, 1979; Hess, 1981). Non-specialists have always had some presence in the production of news.

Recently, though, public participation in the production of news has taken new forms online, including but not limited to:

i) Systems for publishing user-generated content in mainstream news

websites, including: journalism staff blogs; reader comments on journalist-

produced stories; edited and curated content produced by readers (e.g.,

CNN‘s iReport, Fox News‘s uReport or BBC News‘s User-Generated

Content Hub); lists like ―most read‖, ―most emailed‖ and ―most blogged‖

offered by most online news sites that are automatically ranked according

to how often users view and share stories.

156 ii) Systems for judging for value of user-generated content, including:

systems for ranking comments (e.g., Huffington Post‘s system that rewards

readers who judge comments the same way the site‘s official moderators

do with ―Moderator‖, ―Networker‖ and ―Superuser‖ award badges [Tenore,

2010a]); systems for relating a news organization‘s stories to blog content

from outside the news organization (e.g., CBC News‘s ―Blogwatch‖ that

links its stories with blog posts that are highly ranked by blog indexer

Technorati). iii) Systems for generating story ideas based on reader behaviors, e.g., the

travel section of USA Today that uses Demand Media‘s service to generate

content that its algorithms think will be highly ranked by Google. Or

consider AOL‘s website Seed.com, edited by former New York Times

columnist Saul Hansell, that organizes independent journalists to work on

stories that search engines judge to be most popular with audiences. Said

Lewis Dvorkin, founder of site True/Slant, such systems make journalists

think ―wow, this is the audience, and now all of a sudden I have to respond

to the audience because this is what they‘re interested in.‖ (Rice, 2010: np) iv) Systems for publishing professional content according to reader

behaviors, e.g., the Huffington Post‘s use of comparative, real-time testing

to see how well one story headline garners clicks over another. After

approximately ―five minutes … enough time for such a high-traffic site‖

(Seward, 2009a: np) the headline that attracted the most clicks becomes the

one that all readers see. Or consider Google News‘s use of a ―personalized

157

news recommendation system‖ to organize headlines according to

individual users‘ past click behaviors, attracting more frequent visits to the

websites the system anticipates as being relevant (Liu, Dolan & Pedersen,

2010). Journalists increasingly and ritualistically consult data about user

behaviors as they and their editors judge, sometimes hour-by-hour, the

traffic and demographics of particular stories (Domingo, 2008; Leibovich,

2010; MacGregor, 2007) and shift their reporting to follow the traffic

(Peters, 2010).

v) Systems for funding reporting, e.g., those used by Spot.us in which

readers collectively decide which news stories will be written as they

pledge money or spend ―spot credit‖ (earned by taking surveys from

Spot.us advertisers). Or consider earlier, less formal arrangements in

which independent reporters solicited their blog readers for funding to

engage in expensive foreign reporting trips (Rosen, 2003a).

vi) Systems to fact-check reporting of mainstream news organizations, e.g.,

Newstrust.net‘s Truthsquad project (funded by the Omidyar Network

foundation and the Macarthur Foundation) in which its editors select

―dubious quotes‖ from politicians and pundits—many of which already

appear in mainstream news publications—and invite visitors to the site to

vote on whether the statements are true.

Such systems and practices are not without their critics. Most conceptually,

Livingstone (2008) argues that terms like ―user‖ and ―user-generated content‖ fail because they:

158

 see audiences as recipients and consumers of content rather than discerning

readers who likely need help to critically evaluate media;

 off-load meaning-making to a presumably critical reader, relieving journalists

of their responsibility to produce unbiased, coherent and complex messages

that might be meaningfully understood by broad audiences;

 and lack ―any direct relation to communication in particular … [impl]ying an

instrumental individualism rather than a collective, even public, status.‖

(Livingstone, 2008: 52)

Such criticisms speak to the press‘s ideal role as a convener of conversations (Carey,

1987: 14) among a wide variety of people who are actually able to participate, not simply presumed to be participating.

Similarly, in his study of the history of the idea of ―mass man,‖ (inspired in part by Blumer‘s [1939, 1948] early discussions of crowds, publics and masses)

Butsch (2008) argues that widespread use of terms like ―user‖ and ―user-generated content‖ collapse across important historical concepts like audiences, citizens, crowds, and publics. Crowds, he argues, can be viewed as

gatherings of people physically together and sharing a common activity. They are contrasted to a mass that is a dispersed population. Both of these are distinguished from publics that exhibit a dimension of debate or discussion absent in a crowd or mass … Publics in the political sense have been conceptualized as bound up with media. Unlike crowds, publics are dispersed and therefore necessitate means of communication … The rowdiness of crowds constitutes forcible action rather than reasoned agreement. Publics presume a society of equals where various parties can reason with each other and achieve a consensus or settlement without resort to force. A public sphere is premised on the existence of common ground not only physically but also socially and politically. (Butsch, 2008: 8-12)

159

The point here is not to review in detail the histories of these different terms and their relationships to public opinion24 but, rather, to argue that uncritically accepting terms such as ―user-generated content‖ or ―audience participation‖ in online news production elides important conceptual differences in how and why groups gather and mediate their communication.

More pragmatically, it appears that users and user-generated content might have little real impact on the professional press. In a survey of twelve UK national newspapers, and interviews with their staff, Hermida & Thurman (2008) found that, although people have more ways to contribute content online—variously called

―Polls‖, ―Messageboards‖, ―Have your says‖, ―Comments on stories‖, ―Q&As‖,

―Blogs‖, ―Reader blogs‖, ―Your media‖, and ―Your story‖—there is significant internal resistance among editorial staffs to public participation in the production of news. Some fear their work being marginalized by user-generated content; others view such publicly produced content as low-quality—―‗extremely dull‘, ‗mediocre‘ or of ‗very marginal interest‘‖—while others see user-generated content as a complement to professional reporting only in the few cases when audiences are ―very knowledgeable about certain areas.‖ (349) Essentially, Hermida & Thurman find that journalists persist in their roles as ―gate-keepers,‖ devaluing user-generated content because of a desire to protect their organizations‘ brands, control the style and scope of news stories and conversations around them, and avoid the costly labor required to monitor user-generated content and answer readers‘ questions (350-351).

24 For good overviews and critical histories of publics and mass configurations of public opinion, see Calhoun (1992), Hardt (2004), Herbst (1994: 1-32; 1995), Peters (1995, 1999). See also Chapter 3 of this dissertation for a discussion of how presses negotiate autonomy in relation to publics. 160

In his study of nine British news websites, Thurman (2008) found similar resistance to integrating user-generated and professional content. Editors were concerned with audiences‘ abilities to meet ―the standards of professionally produced output‖ (144), fearful of legal action ―resulting from libelous comments posted on unmoderated user forums‖ (150), concerned about duplication among readers‘ submissions, and discouraged by the overtly personal tone of many web comments.

Instead of attributing such resistance to any broad, industry-wide patterns, though,

Thurman attributes such opinions to the local conditions of news production, finding that editors highlighted their organizations‘ particular challenges without necessarily making any fundamental critiques of the idea of public participation in news production25.

Valuation of user-generated content by professional mainstream journalists appears similar in the United States. The Pew Project for Excellence in Journalism

―State of the Media‖ report (2009) surveyed online journalists and found that 43 percent of respondents said that user-generated content was either the least or second- to-least most important part of an online news site. Despite its diminished importance in the eyes of journalists themselves, many U.S. news organizations struggle with how to involve readers in the practices of news production, in particular as commentators

25 Such findings are confirmed by other studies. In their study of 16 leading newspapers from 10 Western democracies, Vujonic et al. (2010) found that journalists and editors viewed user-generated content as successful only when it built the organization‘s brand loyalty, increased traffic to the website, or distinguished the organization from a competitor. Similarly, Domingo et al. (2008) studied 16 online newspapers in Europe and the U.S. and observed that editors limited user-generated forums: they were designed as ways for readers to debate current events within the story frames provided by the papers, not as mechanisms to involve citizens in other stages of news production. And in his analysis of internal BBC documents and interviews with six BBC journalists about blogging within the BBC, Hermida (2009a) concludes that although journalist blogs are seen as fast and helpful ways of receiving feedback from readers and fit within the organizations public-oriented mission, blog conversations are largely one-way and opaque. Journalists rarely respond to blog readers‘ comments and the BBC reveals little about how it moderates blog comments. 161 on news stories. The Washington Post and The New York Times, for example, require that readers first register with the websites, ―providing some information about themselves that is not shown onscreen,‖ before providing comments. Such disclosure is thought to decrease the likelihood of users contributing offensive or inappropriate content, but both papers still review each comment before it is published.

The Wall Street Journal now offers readers the option of seeing only those comments posted by readers who have purchased subscriptions, under the assumption that paying readers will post more responsibly. The Huffington Post is in the process of instituting a design with a similar aim in which commentators who are better known and trusted among its readers are displayed more prominently in lists of comments

(Pérez-Peña, 2010). Smaller papers are not immune to the issue. The central Illinois

Pantagraph recently discontinued comments on local stories for two days—instituting a ―cooling-off period‖—after it found that most reader comments were ―offensive and devoid of civility‖ and against the paper‘s goal of fostering a ―spirit of community involvement and conversation.‖ (Pantagraph, 2010) Similarly, the Voice of San Diego announced that it was discontinuing anonymous comments altogether after finding that ―many of the conversations that take place underneath the articles on news websites devolve into name-calling, racist or sexist remarks, and other vulgarities.‖

(Readers are still able to submit anonymous questions and concerns directly and privately to the editorial staff.) (Donohue, 2009)

Perhaps unsurprisingly, some news organizations are experimenting with outsourcing comment moderation in an attempt to distinguish between news work and other kinds of work. For example, comment threads on NPR.org are now moderated

162 in part by a ―group of professional moderators … a Canadian-based company called

ICUC Moderation Services‖ so that NPR ―news staff can now concentrate on doing what we've always intended - to use the comments for fostering intelligent dialog, finding potential sources, fleshing out story ideas and like.‖ (NPR, 2010: np) Such outsourcing may indeed help NPR better use scarce personnel resources, but it may also further distancing reporters and editors from participating publics. To understand the difference between NPR and ICUC Moderation Services, it is also worth asking what kind of reporting metrics NPR receives regarding how comments were moderated, what kind of decisions ICUC makes on its own, and how NPR learns to be aware of what kind of comments people are submitting but its staff are not seeing.

And even though it officially encourages its reporters to engage publicly with readers who commented on their stories (Thornton, 2009), the Cleveland Plain Dealer recently decided to reconsider whether it should allow comments at all after being sued by a judge who argued that the paper violated her privacy by printing disparaging comments from a false email address (Pérez-Peña, 2010).

There are a few tensions inherent in this debate over how and when to publish online comments. Some papers cite their public responsibilities, seeing their ability to foster healthy commenting forums as evidence of their roles as facilitators of quality public debate. Others make economic arguments noting that comment pages bring in little revenue because advertisers are unwilling to appear next to unpredictable reader feedback; others flip this, arguing that comments are dramatically increasing in popularity—The Washington Post reported a 33 percent increase in comments last year (Alexander, 2010)—and that, properly moderated, comment pages are a healthy

163 source of advertising revenue. Some journalists argue against commenting systems saying that they are unable to do the kind of professional reporting they would like when such systems are turned on because many sources ask first whether the reporter‘s story will allow online commenting; they are unwilling to appear in stories that are open to public comments for fear of being harshly criticized by unruly readers not properly monitored by the news organization (Alexander, 2010; Davenport, 2010;

Pérez-Peña, 2010).

Amidst this debate remains the critical question of how exactly to design, fund and staff online news forums that support quality public participation. Achieving this kind of ―quality‖—a vague and underspecified concept at most online newspapers— often means: costly, close monitoring of readers‘ comments by professional staff who may not value public participation in the production news26; relying on ―crowd- sourced‖ determinations of quality that can be susceptible to manipulation by self- organized sub-groups with vested interests in achieving certain outcomes; or using

26 Online comments under news stories are akin to the editorial letters. Although editorial letters have, at various times in the past, been seen as valuable and growing parts of the newspaper (Hynds, 1994) or viewed them as proxies of public opinion (Sigelman & Walkosz, 1992), they have also been understood to be poor representations of how diverse public opinions can be (Grey & Brown, 1970) and more often than not were simply reflections of what editors perceive to be important topics (Renfro, 1979). They now seem to be viewed with little value by editorial page staff and are more tightly edited and regulated than online comments. In her study of newspaper editorial page, Wahl-Jorgensen‘s (2001, 2002) found that editors preferred letters that were sensational or emotionally charged while simultaneously viewing their writers as ―insane or ‗crazy,‘‖ (2002: 183) and using such perceived instabilities as evidence of letter-writers‘ inability to create meaningful public spheres. (For short surveys of how letter-writers have been viewed at different stages in modern U.S. press history, see Forsythe [1950], Tarrant [1957], Vacin [1965] and Volgy et al [1977].) Nielsen (2010) similarly found that editors only printed letters they judged to be novel, personally evocative, publicly resonant, timely, written by individuals (not groups), fair and objective in the traditional journalistic sense of telling a story with two sides, and likely to incite further public debate (even at the expense of printing letters that editors suspect contain false statements). (Although note that editors are also cautious not to print ―attack letters that might result in libel suits‖ or letters that are ―openly racist, sexist, or homophobic.‖ [Wahl-Jorgensen, 2004: 89]) Across all these studies, the letters section generally seems to be a place for a small number of people to participate in a limited form of public expression that news organizations use to provide the illusion of broad public debate. Most critically, editorial letters can be seen as strategically selected examples of private opinion marketed as evidence of public discussion (Wahl-Jorgensen, 1999; Raeymaeckers, 2005). 164 tiered, ranking systems that reward certain commentators over others, creating reputation economies and ―echo chambers‖ (Sunstein, 2001) in which it can be difficult to hear newcomers, or new and diverse ideas.

These tensions—among the opportunities for and challenges to publishing user-generated content—point to a deeper intersection between two ideologies underpinning the design and use of networked information technologies for public participation in the production of news. One ideology, best summarized in Jenkins‘

(2009) model of participatory culture, emphasizes the civic benefits that can be derived from widespread information sharing, online mentoring, and feelings of consequential action and social belonging that might come from participating in online news environments. The other, best captured in Schudson‘s (2008b) ideal model of journalism, also focuses on the civic functions of the press but frames them in terms of the professional production of information and forums that can, in turn, inform citizens, foster empathy, develop communities, and mobilize individuals around causes of public importance.

Table 1: Jenkins Participatory Culture vs Schudson’s Ideal Functions of Journalism

Jenkins’ (2009: 7) Schudson’s (2008b: 12) Five Elements of Participatory Six Ideal Functions of Journalism Culture

1. Relatively low barriers to artistic 1. Information: the news media can provide fair and expression and civic engagement. full information so citizens can make sound political choices. 2. Strong support for creating and sharing creations with others. 2. Investigation: the news media can investigate

165

concentrated sources of power, particularly 3. Some type of informal mentorship government power. whereby what is known by the most experienced is passed along to novices 3. Analysis: the news media can provide coherent frameworks of interpretation to help citizens 4. Members who believe that their comprehend a complex world. contributions matter. 4. Social empathy: journalism can tell people about 5. Members who feel some degree of others in their society and their world so that they social connection with one another. can appreciate the viewpoints and lives of other people, especially those less advantaged than themselves.

5. Public forum: journalism can provide a forum for dialogue among citizens and serve as a common carrier for the perspectives of varied groups in society.

6. Mobilization: the news media can serve as advocates for participatory political programs and perspectives and mobilize people to act in support of these programs.

Press autonomy in the context of online participatory cultures of prosumption means negotiating between these two models, deciding when and why the press should distance itself from publics and retreat to more traditional models of news production, or when it should help to facilitate public participation in the news, performing a role that is possibly valuable for the press‘s civic ideals and potentially lucrative to the news organization.

166

It is unclear exactly what kind of technological design or operational model might result in high-quality, high-yielding user-generated content news pages but, fundamentally, it means negotiating between a model of participatory culture and an ideal set of press functions. It also means noting how Jenkins highlights the power that comes from individual expression and communal association—using ―members‖ as his unit of analysis—while Schudson emphasizes the informational requirements and power dynamics of civic life—using ―citizens‖ as his object of study.

Also note that, while Schudson mentions that the press would ideally garner

―social empathy‖ for ―other people‖ who are ―less advantaged‖ and sees the press as a

―common carrier for the perspectives of varied groups,‖ Jenkins is largely silent on how diverse, inclusive (Young, 2000) or dissenting (Sunstein, 2003) counterpublics

(Fraser, 1990; Warner, 2005) might be recognized or encouraged within participatory cultures. Jenkins‘ participatory culture has little of the sophistication of Habermas‘s

(1989) ideal public sphere—designed to foster truth-seeking through rationality, the bracketing of individuals‘ identities and a focus on matters of common interest—or

Young‘s (2000) ―decentered‖ public sphere in which diverse and mutually accountable participants engage directly with each other‘s differences.

Jenkins talks about membership being limited by barriers to expression without acknowledging other kinds of power that might exclude entrants, while Schudson presumably means understands ―citizen‖ with nation-state boundaries and not as a more general, global role. Schudson sees mobilized citizens as a potential product of journalism while Jenkins seems to see mobilization a central activity of participatory

167 cultures, albeit without commenting on why members might mobilize and what they might be mobilizing against.

In the extreme, Schudson‘s citizens are susceptible to being mobilized externally by the press‘s goals, while Jenkins‘ members may be limited to mobilization derived from internal self-referential aims or those that come from self- selected communities. In the context of press autonomy, though, it is worth recalling

(from the review of empirical studies of how news sites treat user-generated content) that journalists and editors largely exclude non-journalists from core aspects of news production, making it difficult for non-professionals to be mentored (as Jenkins calls for) or engage with the ideas of diversity, quality and empathy that Schudson ideally asks journalists to prioritize. Negotiating the exact nature terms of a hybrid professional/citizen journalistic ―community of practice‖ (Wenger, 1998)—and the terms of ―legitimate peripheral participation‖ (Lave & Wenger, 1991) under which citizens might be mentored by professional journalists—is a critical aspect of any new model of press autonomy in which journalists are simultaneously distinct from and reliant on publics.

Finally, it is worth noting that, in terms of the press‘s public responsibility (at least in its ideal form) such definitions of participatory culture are somewhat apolitical. They offer little guidance about how a broadly participatory news culture would manage public communication infrastructures, negotiate between majority and minority viewpoints, or decide when to restrict citizen-members‘ creations and sharing. Essentially, Jenkins‘ definition foregrounds questions of visibility and collaboration while remaining largely silent on deeper questions of public decision-

168 making or deliberation among diverse groups whose membership derives not from interest or creativity, but a fundamental notion of constituency associated with being a citizen within a political unit.

Jenkins‘ ideal of participatory culture has at its core a sense of shared experience, but it is missing a notion of shared consequences. Dewey argued that a key feature of a public is its ability to collectively manage common fates and pragmatically appreciate outcomes that affect community constituents. He wrote that the public‘s essential problem is ―that of perceiving in a discriminating and thorough way the consequences of human action (including negligence and inaction) and of instituting measures and means of caring for these consequences.‖ (Dewey,

1927/1954: 21) Jenkins‘ participatory culture does not focus on political outcomes or public goods. He states only that members should ―believe‖ that their contributions matter and ―feel‖ some degree of social connection.

An autonomous press—practicing the kind of ideal journalism envisioned by

Schudson but within the kind of participatory cultures Jenkins (1992) has chronicled in various online communities—must be both distinct from publics and inclusive of them. Essentially, if the press is to enact a public mission in which its readers— citizens—encounter and empathize with information and people they might not purposefully seek out, it must be critical of self-organized systems susceptible to echo chamber effects (Sunstein, 2001) that foster feelings of participation but that produce little in the way of new ideas. Conversely, if the press continually characterizes user- generated content as dull, or mediocre (e.g., doing little across to foster within its readership the kind of critical skills necessary to participate in quality news

169 production), it risks marginalizing itself among new quasi-journalistic online forums that provide information people want, but not necessarily perspectives that citizens need.

2.C. Professional Practices

Contemporary press autonomy is negotiated not only as the economics of news production change and as readers contribute content, but also within professional practices of mainstream journalists. Three aspects of journalistic practice stand out in which press autonomy is being negotiated in terms of individuals‘ professional behaviors: the attitudes and behaviors of individual journalists in online news production; mainstream news organizations‘ policies and guidelines around social media use; and the emergence of seemingly new official roles and responsibilities for journalists at online news organizations27.

2.C.(i). Journalist Behaviors and Attitudes

In 2009, The Pew Project for Excellence in Journalism ―State of the News Media‖ report (Pew, 2009) surveyed journalists‘ understanding of their profession‘s

―fundamental values‖ in light of new online news environments. The survey revealed two trends relevant to this discussion or press autonomy. The first centers on the notion of speed: 45 percent of respondents said they feel pressure to produce reports faster and that they had less time for fact-checking, reflection or verification with other sources; 25 percent agreed with the stronger statement that immediacy in news is

27 Chapter 3 of this dissertation goes into greater detail about professionalism as a concept and how journalists understand and signal their professional roles. 170 now ―more important‖ than accuracy. This finding agrees with the findings of PR

Newswire‘s (2010) survey—albeit a methodologically weak one that groups together the responses of ―traditional‖ journalists, ―non-traditional‖ journalists‖ and public relations professionals without defining these categories—in which 20 percent of respondents (up from 5 percent in 2009) said that their professional success is measured in terms of how well they ―break news and chronicle events as they happen.‖ (2010: np)

Journalists and news organizations have traditionally valued speed in reporting; this is not a new phenomenon. However, the nature of journalistic time seems to be changing, as Carr (2010: np) describes when he says that news today is temporally organized not into pages that arrive with a predictable frequency (e.g., the daily newspaper, the weekly magazine) but into a ―stream‖ of ―fast-moving, ever- shifting flow of bite-sized updates and messages‖. Indeed, one study of temporal relationships between mainstream media websites and blogs found that it took an average of 2.5 hours between a meme‘s peak appearance in mainstream news sites and the same meme‘s peak appearance in blogs (Leskovec et al, 2009). Mainstream journalists may have a head-start on bloggers for most stories, but to retain this lead they must—very quickly—either replicate a new version of the story for their own organizations or deliver an even newer story that simultaneously distinguishes them from their competitors and attracts their (and, later, bloggers‘) attention. These dynamics are similar to the ―scoop and shun‖ phenomenon in traditional print journalism in which the

pressure to be fast and first shifts the journalist‘s attention away from judgments about what the public needs or wants and toward a calculation of what the competition might do, an intramural contest of some significance to

171

newsroom insiders but of little or no interest to most readers, viewers, and listeners. Shunning the competition is the other side of the coin. Competition in this case results in a very peculiar form of self-censorship: a refusal to cover a story only because it appeared first in another publication. (Glasser & Gunther, 2005: 396)

The core of this phenomenon is likely still true in online news environments, but its character may differ because of the qualitatively different nature of time in online news cycles. Specifically, given that online news is essentially consumed at the moment it is published—most news organizations‘ content is, for example, immediately indexed by search engines and advertisers—there is often virtually no lag time between a decision to publish and an opportunity to receive online reactions to the publication. There are rhythms to the internet—time of day matters in both the production and consumption of news stories (Leskovec, 2009; Varian, 2009)—but such rhythms are global, spread across producers and consumers in different time zones. There is no morning or evening print run to work toward, no six o‘clock newscast to convene around, and no regularly scheduled programming to interrupt.

Journalists‘ may indeed continue to engage in scoop-and-shun behavior, privileging their perception of competitors timeliness over considerations of what information publics might need and when it might need it. But their decisions may also become of great interest to algorithms and advertisers who model—and assign economic value to—readers‘, viewers‘, and listeners‘ consumption patterns. It may still matter whether a journalist‘s competitor is scooped or shunned, but it may be more beneficial—in terms of readership and advertising revenue—to have a particular version of a story appear at a particular time within the online news cycle, one that attracts search engines at a different arc in time.

172

Although the dynamics of contemporary, accelerated news cycles are understudied and not well understood, press autonomy may have a temporal component to it: a journalist may be more or less free of influence from competitors depending on why and how well she selectively and strategically accesses the online news stream, responds to it at particular points, and separates herself from it in order to prepare a story that is different from what is being published at a moment in time.

More institutionally, the press‘s autonomy might be defined by how thoughtfully it orients publics in time, distinguishing between information flows that appear urgent and news-like but that are actually more about providing ―nonstop information without interpretation, and nonstop interpretation without information.‖ (Katz, 1992: 8)

The second trend in individual journalists‘ professional practices concerns the long-standing issue of viewpoint, or whether individual reporters should state their perspectives on the news they cover. Half of online journalists surveyed (Pew, 2009) said it was a ―good thing‖ that online news outlets have ideological points of view and slightly less (45 percent) felt that, in online environments, it is ―becoming more acceptable to report news with an obvious voice or bias.‖ (Although there are again methodological problems with comparing survey data sets—Pew is unclear about the difference between ―online journalists‖ and members of the ―mainstream press‖—this finding is in contrast to Pew‘s 2007 survey in which 75 percent of the national, mainstream press said it was a ―bad thing if some daily news organizations have a decidedly ideological point of view.‖)

Again, the tensions in such broad survey-level results are supported by commentary among journalists and contemporary conversations about the value of

173 objectivity in online reporting,28 and whether the growing use of social media tools by journalists means that ―transparency is the new objectivity‖ (Weinberger, 2009;

Ingram, 2009). Essentially, what this phrase captures is the idea that, instead of relying upon the reputations and brands of their organizations, individual journalists should now earn readers‘ trust by demonstrating how they do their reporting. They should, e.g., reveal their stories in draft form, state personal biases and potential conflicts of interest, provide links to background materials, acknowledge when they are using competitors‘ reporting, which sources influenced their reporting and, most broadly, sustain a meta-conversation about their reporting that helps readers understand why their stories take the forms they do (Phillips, 2010).

This kind of trust, seems to be built not only by news organizations or reporters who disclaim being influenced by anything other than a search for truth; it is also created by readers who presumably weigh a story‘s and reporter‘s meta-data alongside its reporting. It envisions trust as a contract co-constructed by self-revealing journalists and critical audiences. And, although the idea of professional reputation is not new, advocates of online transparency recommend that both journalists with established records of reporting (Glaser, 2009) and newer journalists and students

(Hermida, 2009b) cultivate and demonstrate not just their reporting skills but also personal brands—in the form of personal websites, blogs, social media feeds, link

28 Although often dramatically pitted against a supposedly uniformly shared and outdated notion of journalistic objectivity, such oversimplifications should be viewed with skepticism. (Perhaps relying upon Daston‘s [1992] critical use of the same term to describe the scientist‘s perspective in the supposed absence of social relations, Rosen (2003b) says that journalists routinely claim to have a ―view from nowhere‖ to mask their biases.) Journalistic objectivity is not a single, homogenous concept that lives within a particular ideal person or text but, rather, can be better viewed as a locally contingent practice, a kind of ―mature subjectivity‖ (Schudson, 1978:92), ―neutrality‖ (Lichtenburg, 2000) or objective method, rather than within a single person (Kovach & Rosenstiel, 2001). The relationship between objectivity and press autonomy is a large and complex topic covered in greater detail in Chapter 3 of this dissertation. 174 archives, un-commissioned reporting, commentaries—that signal not just their professional histories but their personal values and opinions to both news organizations and audiences. Essentially, earning a position of respect within the online press depends not only on journalists establishing reputations with news organizations or a body of published work – it also now requires journalists to maintain a personal brand, housed in social media, that complements and is linked to their professional work.

2.C.(ii). Social Media Policies

Another way in which the press‘s autonomy appears in professional practices is the degree to which journalists are encouraged or dissuaded from engaging with audiences through online networked social media sites like Facebook, Twitter, MySpace, and

LinkedIn. How do news organizations‘ social media policies suggest or prohibit staff from engaging with publics through these other media, and what do such policies reveal about how presses negotiate their autonomy through sanctioned employee activities?

While there are few good industry-wide data about the use of social media technologies by the mainstream press, several news organizations seem to be encouraging its use amongst their staffs. For example, Sky News installed the Twitter application TweetDeck on all news desks and began an ―education drive to get [all journalists] up to speed‖ using the tool (Betancourt, 2010: np). And, when asked whether it was important for BBC journalists to use social media tools like Facebook and Twitter in their reporting, the BBC‘s Director of Global News stated:

175

I'm afraid you're not doing your job if you can‘t do those things. It's not discretionary … If you don‘t like it, if you think that level of change or that different way of working isn‘t right for me, then go and do something else, because it's going to happen. You‘re not going to be able to stop it. (Bunz, 2010: np)

One way to understand the meaning of social media tools in relation to reporting is to understand how news organizations regulate their use. I conducted a close, critical reading of the publicly available social media policies and guidelines of The

Washington Post (Kramer, 2009), NPR (2009), Associated Press (Wired, 2009), and

The New York Times (PoynterOnline, 2009) and extracted five principles that seem to govern these news organizations‘ approach to social media. The aim here is to better understand how mainstream news organizations—officially at least—would like their reporters to both depend upon and distinguish themselves from social media29.

1. Social media is a boon for information-gathering and a minefield for

professional journalism. All five organizations explicitly mentioned the

potential value of social media for news gathering—their ability to ―speed

research and extend a reporter‘s contacts‖ (NPR), deliver ―citizen journalism

material‖ (AP) or do ―triangulation on difficult-to-research subjects‖ (New

York Times)—but each in turn also cautioned that such tools should only be

used in sanctioned ways to ensure that the organization can ―protect its

reputation‖ (AP) and ―professional integrity‖ (Washington Post) against the

―virtually unlimited‖ ―possibilities of digital distortion‖ (New York Times).

29 It should be noted here that the definition of ―social media‖ used here is the same one that these organizations‘ policies implicitly adopt, but never explicitly state. They all listed examples of technologies the policies are designed to cover (Facebook, Twitter, LinkedIn, MySpace) but none stated a definition of social media that might accommodate future technologies. The lack of a definition makes the social media policies analyzed here contingent upon the unstated assumptions guiding these organizations‘ understandings of social media. 176

2. Journalists are almost always journalists and they are always publishing.

While some organizations acknowledge their reporter‘s multiple identities—

NPR understands that ―regardless of how careful you are in trying to keep

them separate, in your online activity, your professional life and your personal

life overlap‖ and the AP only asks that reporters identify themselves as AP

employees in online networks when using them for work—the Washington

Post states that reporters ―must remember that Washington Post journalists are

always Washington Post journalists … [that] all Washington Post journalists

relinquish some of their personal privileges as private citizens‖ and that the

policy makes no distinction between personal and professional use of online

social networks. The New York Times and Washington Post caution their

reporters that there is no online sanctuary from reporting—―‗personal blogs‘

and ‗tweets‘ represent you to the outside world just as much as an 800-word

article does‖ (NYT) and ―any content associated with [reporters] in an online

social network is, for practical purposes, the equivalent of what appears

beneath their bylines in the newspaper or on our website‖—and the Times

further reminds its staff always to stay within their organizational roles in their

online behaviors: ―don‘t editorialize, for instance, if you work for the News

Department.‖

3. Nothing is private and shield laws won’t protect you. All organizations‘

policies emphasized that, even though social network sites offer a variety of

177

privacy settings and restrictions on viewing content, they are not to be trusted.

NPR tells its employees to ―use the highest level of privacy tools available to

control access to your personal activity when appropriate, but don‘t let that

make you complacent … [E]verything you write or receive on a social media

site is public [and] can be easily circulated beyond your intended audience …

[L]aw enforcement officials may be able to obtain by subpoena anything you

post or gather … without your consent – or perhaps even your knowledge.30‖

The Washington Post similarly distinguishes social network activity from other

kinds of background activity when it tells reporters that what ―you do on social

networks should be presumed to be publicly available to anyone, even if you

have created a private account.‖ The AP adds that even personally shared

information with no apparent link to a reporter‘s professional role is potentially

public ―even if staffers restrict their pages to viewing only by friends. It‘s not

just like uttering a comment over a beer with your friends.‖

4. Objectivity in social media means being transparent about who you are and

what you’re doing, but not what you think or who you really know. All

organizations‘ policies seem to prevent reporters from engaging in anonymous

or open-ended observation in online environments. They instruct journalists

both to identify their professions and organizational affiliations, and to reveal

their intentions. The Washington Post asks staff using social media to be

30 Curiously, while NPR assumes that all of its reporters‘ online activities are essentially public, it maintains a distinction between on- and off-the-record communication for other people‘s online actions. Its policy instructs employees who participate in any online forum to ―ask first if the forum is on or off the record before distributing information or content about it.‖

178

―transparent in [their] intentions‖ and NPR tells its reporters: ―[j]ust as you would do if you were working offline, always explain to anyone who provides you information online how you intend to use the information you are gathering.‖ But the transparency seems to stops there. All organizations prohibit journalists from using online social networks to take public stances on issues, including any appearance of bias that may emerge from joining an online group. The Times specifically tells its staff to ―leave blank the section

[on your profile] that asks about your political views‖ while the Post warns against any general but undefined ―pattern of [social media] use‖ that may suggest bias. The Post further warns reporters not to reveal the kind of

―internal newsroom issues such as sourcing, reporting of stories, decisions to publish or not to publish‖ that advocates of objectivity-through-transparency call for. Finally, although the news organizations seem unclear about the meaning of online ―friends‖—the Times states that ―being a ‗friend‘ of someone on Facebook is almost meaningless‖—all organizations see online friending as a potential source of bias and recommend the following: ―if you

‗friend‘ or join a group representing one side of an issue, do so for a group representing the competing viewpoint‖ (NPR) because ―[i]t would not have looked good in the presidential election campaign for a national political reporter to agree to be a ‗friend‘ of Barack Obama without first making sure to be a ‗friend‘ of John McCain, too.‖

179

5. News organizations still own their content. Finally, most news organizations

explicitly warned reporters not to use social media to scoop their employers:

―We‘re still the AP. Don‘t report things or break news that we haven‘t

published, no matter the format, and that includes retweeting unconfirmed

information not fit for AP's wires.‖ NPR tells reporters that they are not to

―repost NPR copyrighted material to social networks without prior permission‖

and instead channels them to the application programming interfaces (APIs)

and ―widgets‖ NPR makes available for sanctioned distribution of its content31.

Essentially, although reporters are cautioned that their professional and public

identities are essentially identical, when using social media sites, they are

given no special professional privileges different from the general public to use

the content they generate in their reporting roles.

Taken together, these policies describe a set of tensions in mainstream news organizations‘ official use of social media that can be understood as tensions in press autonomy. Specifically, such guidelines are essentially about determining a journalist‘s ―proper‖ distance from online publics and non-journalists. They see social media as opportunities for news organizations to increase their informational efficiencies—to speed up reporting, reach new sources or verify information—but not to question underlying assumptions about how and why to form relationships with those considered outside of the profession. Essentially, the guidelines recast traditional journalistic objectivity in terms of social media, updating it for the material conditions of online news production without reexamining its underlying logics.

31 See Chapter 5 for a more detailed discussion of the NPR API. 180

Recall the idea that autonomy is understood here as a system of dependencies and negotiated distances. Social media can become close and integral to journalists‘ work and their relationships to audiences (e.g., consider Hermida‘s proposal for

Twitter-fueled ―ambient journalism‖ in which both citizens and reporters constantly

―maintain a mental model of news and events around them‖ [Hermida, 2010: 297]).

But social media can also distance journalists from their audiences by simply recasting old objectivity in new material. (They must, for example, keep personal views hidden in online social networks, and be sure to make friends with both side of an issue.)

Perhaps of greater concern, though, is how stridently the guidelines warn reporters that all social media activity is always essentially public. Such online networks are not seen as professionally private spaces for recording or debating information—compare this with reporters‘ notebooks, editorial meetings or background sourcing—in which journalists might casually experiment with story ideas and angles without fear of their work being exposed prematurely. If social media activities are seen by news organizations as unfailingly public activities and core to a news organization‘s relationships with online publics, it becomes difficult to separate—conceptually and in practice—the production, representation and consumption of news. The fewer private spaces there are that are distinct from and inaccessible to publics—places where journalists can conduct at least some online work, testing ideas with colleagues, following leads that prove erroneous, pursuing stories that might be unpopular to masses or elites, offering sources anonymity, or repairing mistakes—the greater the risk there is that news may become safe and sanitized. Essentially, journalists and publics may learn to co-produce news that can

181 successfully withstand constant public scrutiny, but that may ultimately fail to equip citizens with the critical or risky information they need to be self-governing32.

2.C.(iii). Official Roles

Finally, another way to understand the press‘s contemporary professional practices— and their impact on press autonomy—is to look for evidence of new official roles and responsibilities emerging in relation to online news production and the public‘s participation therein.

Several national news organizations have proposed new official roles in response to the growing popularity of online social media tools like Facebook, Twitter and MySpace but, before examining them in detail, it is worth first putting them in a context of other newsroom roles that have been designed to mediate press-public

32 This discussion of what Bourdieu (1996a) refers to as small-scale, private restricted production is more clearly expanded upon in Chapter 3. Also, this point is conceptually similar to studies of social processes by which objects are visibly constructed, the value of observing production, and the varied locations of making. For example, Lessig (2009) cautions against ―open government‖ initiatives that advocate for unfettered transparency—seemingly for its own sake—in how government works; he worries that rampant openness may lead to casual, ill-informed critiques of complex public systems that fail to recognize the value of work that is done in private settings for public interests. Similarly, in his dramaturlogical studies of ―impression management,‖ Goffman (1959) distinguishes between ―frontstage‖ and ―backstage‖ work: the former consists of public performances in which people reveal practiced and strategically constructed aspects of self as they embody known roles in front of expecting audiences; the latter involves more private, preparatory work rehearsing and experimenting with features of self-identity that complement their public-facing roles. Becker (1982) makes a similar point in his study of artists when he notes how they protect distinctions between finished and unfinished works, work studios and showing spaces. This is because an artist‘s reputation depends, in part, on her ability to demonstrate to colleagues, audiences and critics—her ―art world‖—that she understands the complementary yet distinct phases of art production, from having an initial idea to displaying a finished product. And note that learning sciences researchers (e.g., Turkle & Papert [1992], Resnick et al [2000]) increasingly advocate for systems that make transparent and manipulateable aspects of complex mathematical and engineering systems. Such features, they argue, can make an individual‘s understandings of a system concrete, rendering it in visible and dynamic forms that offer both publicly accessible records of learning, as well as starting points for collective debugging and revision. Transparency makes it possible to see how individuals or groups learn differently, how understandings change over time, and how certain system features may favor particular forms of participation or types of participants over others. (See Gilligan [1982], Keller [1983] and Suchman [1995] for feminist intellectual roots of contemporary learning science‘s understandings of epistemological pluralism and the value of transparent system design.) 182 interactions. The first and most obvious is the editorial-page staff. They have historically been charged with reading, editing and publishing letters to the editor, acting as gatekeepers for a narrow form of public participation in the news. They have generally thought little of such participation, viewing ―contributors to the section—the members of the letter-writing public—[as] insane or ‗crazy.‘‖ (Wahl-Jorgensen, 2002:

185)

During the public journalism movement in the U.S. in the 1990s, news organizations also experimented with a variety of official organizational forms and roles designed to bring their staff closer to their readers. For example, Willey (1998) describes a set of outreach techniques newspapers used to better understand which public concerns should guide reporting: focus groups, resource panels (panels of experts), community conversations (among invited community members), pseudo focus groups (among people who responded to ads soliciting their participation), neighborhood conversations (pseudo-focus groups within particular geographic districts), and living room conversations (focus groups conducted in participants‘ homes). And, internally, Johnson describes the reorganization of some papers‘ journalists into ―reporting circles‖ (Johnson, 1998) who covered perceived community themes—e.g., quality of life, city life and governance, leisure—instead of traditional reporting beats like city hall and police stations.

Finally, it is worth considering some of these roles and organizational changes in the context of the ―ombudsman‖ or ―reader representative‖33 position that news organizations have often used and to connect with, and distance themselves from, their

33 While some papers use the more gender-neutral ―reader representative‖, the term ―ombudsman‖ remains accurate and will be used here. Most people in this role have been male and the national group is still called the ―Organization of Newspaper Ombudsmen.‖ 183 reading publics. First begun in the U.S. in the mid 1960s as part of an attempt to establish credibility with readers and present a more public-oriented view of the press, the ombudsman was ―typically a senior editor equipped with the authority to investigate complaints and get answers for readers.‖ (Nemeth, 2003: 2) Although some papers have experimented with adding ombudsmen and changing their roles, it is still a relatively new and rare position in contemporary U.S. pressrooms (McKenna,

1993; Meyers, 2000). (The Organization for Newspaper Ombudsmen [2010] lists only

37 news organizations—worldwide—that have ombudsmen or readers‘ representatives.)

Most ideally envisioned, the ombudsman has the ability to sanction journalists, establish press credibility in public eyes, and maintain a sufficient distance from journalists, editors and owners to sustain ongoing critique of a particular news organization and, more broadly, the journalistic profession (Meyers, 2000). The ideal ombudsman, Cline (2008) argues (through a critical comparison of Daniel Okrent‘s ombudsmanship at The New York Times with Michael Getler‘s at the Washington

Post), is not an ―omniscient journalist‖ who simply corrects journalists about what should have been done, or lectures readers on how newspapers ―really‖ works.

Rather, the ideal ombudsman is a self-reflective, ―privileged reader‖ whose first mission is to critique the paper‘s integrity by critically examining its reporting practices as an experienced outsider.

Supporters of the role describe it as an instrument of accountability that can make relationships between a paper and its readers more responsive and less adversarial. For example, managers at newspapers without ombudsmen tend to

184 resolve reader complaints privately without committing themselves to making systemic changes in practices (McKinzie, 1994) while readers of newspapers with ombudsmen are less likely to lodge formal or legal complaints against the news organization when compared with papers without ombudsmen (McKenna, 1993;

Pritchard, 1993).

More commonly, though, it seems that the ombudsman role has been less than effective, more symbolic than substantive. A national survey comparing the reporting of journalists who worked at papers with ombudsmen versus those at papers without found that a reporter‘s age—not the presence of an ombudsman—was a better predictor of whether he would engage in what the survey described as ―controversial newsgathering techniques.‖34 (Pritchard, 1993: 77) More significant than this ―null effect‖ finding in which the presence of an ombudsman had no impact on reporters‘ behaviors is the finding that, when papers do have ombudsmen, journalists tended not to share complaints they received from readers and were more likely to simply ignore or hide negative feedback (Bezanson et al., 1987). This finding supports McKenna‘s

(1993) argument that ombudsmen actually isolate reporters from readers because they outsource the sense of public responsibility that, he argues, should be located within an individual reporter‘s personal work ethic. Finally, and perhaps most damaging to the role‘s ideal, Ettema & Glasser (1987) found that ombudsmen‘s own understandings of their roles were so inflected with persuasive, public relations logics—stating, for example, that the ―single most important aspect of their job‖ is to

34 The survey questioned respondents about their opinion of four reporting practices thought to be controversial: using an individual‘s personal documents without their permission; repeatedly contacting sources who stated their unwillingness to talk to a reporter; using confidential business documents or classified government documents; becoming an employee within a firm or organization in order to gain information (Pritchard, 1993: 82). 185

―give readers a sense that the newspaper cares about them‖ (7)—that they were more likely to represent the paper’s interests to the readers, rather than be critical liaisons or cultural bridges between the news organization and its publics. This work agrees with

Nemeth & Sanders‘ (1999) content analysis of a year of ombudsmen‘s columns across member organizations of the Organization of Newspaper Ombudsmen. They found that approximately 85 percent of columns consisted of public relations-oriented behavior and language aimed at convincing readers that a paper had acted correctly and discouraging public debate.

It is with these organizational structures—letters page editors, public journalism experiments, newspaper ombudsmen—and ideals in mind that we might critically evaluate newer roles, networked technologies and information practices that news organizations are using to mediate between a paper‘s staff and its reading publics. What assumptions about public capabilities, professionalism and ―proper press-public distance‖ are embedded in these positions and programs? How are these assumptions encoded in policies, job descriptions, practices and technologies? How are both public participation in the news and professional oversight of journalism bracketed in these structures and why? And how does the ability to economically and demographically value public participation in the production of news impact news organizations‘ decisions to implement new roles and structures? E.g., does an advertiser‘s willingness to appear on web pages of readers‘ letters, user-generated content, or ombudsmen columns impact how the news organization experiments with roles and structures for public-press interactions?

186

It is still early to answer such questions completely, even though several major mainstream news organizations are creating new roles and structures that depend upon networked social media, and that aim to change existing journalistic practices. The

New York Times, for example, appointed a ―Social Media Editor‖ in May 2009—to work ―full-time on expanding the use of social media networks and publishing platforms‖ and to ―guide [Times staff] on how to more effectively engage a larger share of the audience on sites like Twitter, Facebook, Youtube, Flickr, Digg, and beyond‖ (Seward, 2009b)—but then discontinued the role a little over a year later saying that ―social media can‘t belong to one person; it needs to be part of everybody‘s job … it has to be integrated into the existing editorial processes and production processes.‖ (Tenore, 2010b)

The Voice of San Diego similarly recently advertised for an ―engagement editor‖ to ―revolutionize how it presents its content and engages the San Diego community‖ (Garber, 2010b)35. And although the Los Angeles Times continues to employ a ―readers‘ representative‖ in the tradition of ombudsmen (recently promoting its national copy desk editor to the role), it also recently hired a ―Senior Producer for

Social and Emerging Media‖ tasked with drawing readers into the Times website.

This producer is to use ―sites and services outside of latimes.com [to] engage new readers, spread the word about some of [the Times’] best work, and do a better job of listening to the larger conversation on the Web.‖ (LA Times, 2008: np) The producer is tasked with using ―Digg, Twitter, Facebook, Flickr, You Tube and other sites that provide opportunities for us to share what we do and better connect with readers …

35 The role seems to be a response to unstated inadequacies with the opinion editor tradition, as Voice of San Diego CEO Scott Lewis suggested in announcing the position: ―[i]magine if there were an opinion editor who had never heard of what an opinion editor was in a newspaper.‖ (Garber, 2010b) 187

[and] attract readers to [Times] stories, photos, video, database projects, etc.‖ (LA

Times, 2008)

Although a wider and more systematic comparison of such roles is required, note what these roles emphasize. They are largely tasked with bringing social media users to the news organizations‘ sites, delivering content,

―engaging‖ readers, and attracting the attention of third-party applications like search engines and online social network sites that are now integral to building readership. With their emphases on the power of social media to share content easily among people who feel like they are making social connections, such roles seem to embrace parts of Jenkins‘ (2009) model of ―participatory culture‖ while largely remaining silent on other aspects. For example, none of the roles talks about providing ―informal mentorship‖ in online news production or helping non-professionals learn how to participate in news production in a way that the organization might value. And such roles seem very distant from

Schudson‘s (2008b) functions of ideal journalism, saying nothing, for example, about what kind of analysis, empathy, mobilization, diversity, or reliability such social media-fueled press-public interactions might provide.

In considering the press‘s autonomy—its system of institutionally distributed relationships and dependencies—it thus becomes important to consider not only the institution‘s relationships to economic forces and readers actors but also to the professional practices guiding its change. It is argued here that three sites of professional practice in which we might see press autonomy negotiated are in journalists‘ individual behaviors in networked environments, news organizations‘

188 policies regulating staff use of social media, and new official roles and structures designed to mediate press-public interactions.

2.D. Intellectual Property and Organizational Systems

In addition to funding models, reader participation and professional practices, another domain in which the press‘s autonomy is both defined by, and reflected within, online news production is the intellectual property and inter-organizational relationships governing the distribution of news content. The contemporary press‘s autonomy—its dynamic, networked system of dependencies and distances—might be understood, in part, by examining how news organizations try to control circulation of their products through both intellectual property ownership claims and strategic relationships.

One of journalism‘s basic challenges lies in a key feature of its primary product. News, historically in the United States at least, is mostly a privately produced public good (Hamilton, 2006). That is, it is non-rivalrous (one person‘s consumption of the good does not prevent another person from using it) and non- excludable (realistically, no person can be barred from enjoying it). Such public goods, though, are open to abuses by ―free riders‖ who use more than what is considered to be their ―fair share‖ of a public good, or who fail to make what is considered to be a fair contribution to the production of a public good. Such people are problematic when they prevent a public good from being produced in sufficient amounts, or when public goods are not produced at all. Since free riding is especially easy online in online news environments that mostly offer news stories for free, it is

189 often difficult to ensure that producers of news are compensated appropriately.

(Federal Trade Commission, 2010a: 4)

For much of the press‘s modern history, this arrangement worked—i.e., consumers enjoyed relatively inexpensive high-quality news—because the news was bundled with advertisements that subsidized its production, and because it was difficult to share free newspapers easily, mitigating free-riding as a major problem.

News production, labor and circulation costs were relatively high, but these costs were covered because newspapers had a virtual monopoly on connecting advertisers and readers, who then became consumers of the advertised products and services (Fallows,

2010; Hamilton, 2006). It mattered little that there was a small direct market for the news because advertising subsidized its production and distribution. News was a relatively stable public good with what economists call ―positive externalities‖: theoretically at least, even free-riders who read newspapers without paying for them were more likely to be informed citizens who could make good public decisions, and active consumers who purchased more goods (Hamilton, 2006).

Although the basic economic arrangements of the news industry were reasonably stable, the organizational relationships that challenged the press‘s autonomy were usually cast in terms of three types of economic tensions:

- within news organizations, between the journalists and editors who produced

the news, and the business staff and owners who funded it36;

36 Although it is methodologically difficult to trace the contemporary influence of owners and business interests on reporters and editors—e.g., Gans (1979) found no evidence of it in his newsroom—several studies suggest that journalists and editors are not independent of commercial considerations in their decision-making. Chomsky‘s (2006) examination of historical records at the New York Times chronicles clear influence of owner-publisher Arthur Sulzberger over editor Turner Catledge; McManus‘s (1997) meta-review of surveys evaluating influences on journalists‘ ethical decision- making found that the influence (perceived or otherwise) of news organization owners trumped the 190

- among news organizations, as consolidations in industry ownership

concentrated news-making power in fewer, larger conglomerations, leading to

greater choice but less diversity for news consumers (Bagdikian, 2004;

McChesney, 2000);

- and against news organizations, when news organizations accepted various

forms of public financial support without providing commensurate public

access to their publications (e.g., Barron, 1967, 2003)

These influences likely still exist but, as news and advertising have become unbundled, private presses have lost much of the revenue they need to produce news in a way that made news behave like a public good. And, simultaneously, in contrast to physical newspapers that are difficult to share quickly among many people, digital forms of news can be copied and disseminated easily and at virtually no cost by people who may contribute little or nothing to their production. Furthermore, such modern free-riders can profit from news they did not produce by copying press content, attracting search engine attention, and earning revenue from juxtaposed advertisements. The internet has thus exposed the ―real‖ economic costs of news while, at the same time, created a new class of free-riders who can, news organizations argue, create unfair news markets that offer few incentives for news organizations to produce original news.

Mainstream news organizations have formulated different responses to this situation by creating new ways of managing the production and dissemination of their stories. Each type of response has implications on the idea of press autonomy in influence of colleagues or publics; and although Voakes (1997) found that, in some limited circumstances, journalists used independent lines of reasoning to make ethical decisions, small-group and organizational factors were far better predictors of how a journalist would behave. 191 networked news production. Specifically, news organizations are crafting new intellectual property claims and technical systems to defend them; and new organizational agreements and publishing technologies to govern production and distribution of their content.

For example, at a March 2010 Federal Trade Commission workshop on the question ―How Will Journalism Survive the Internet Age?‖ the Associated Press‘s

(AP) Associate General Counsel for intellectual property Laura Malone argued for private ownership of news stories and for copyright regimes protecting owners‘ rights:

[W]e, the content owners, we, the copyright owners, get to set the parameters by which people can republish our stuff. If people want to build sites based on the news that is published by any of the news organizations, that‘s great. We‘ll give them a license. Licenses are not difficult to get, and they‘re not terribly expensive … It‘s just that people, because they can do it for free, are doing it for free and assuming that that‘s alright and then raising a stink if news organizations bring a copyright-infringement claim (Federal Trade Commission, 2010b: 67)

Later in the same session, Malone used the ―hot news‖ doctrine to extend the AP‘s claim from a right to control distribution of its news stories to a more fundamental right to control, for a limited period of time, the informational material underlying a news story. This doctrine first appeared in legal theory following a 1918 decision

(International News Service v. Associated Press) in which the Justice Pitney of the

U.S. Supreme Court, writing for the majority, granted the AP a ―quasi-property‖ right, for a limited period of time, to the news it produces. It prevented competing news organizations—but, importantly, not the general public—from having access to this news (Epstein, 1992; Anderson, 2010). There are three important points to the hot

192 news doctrine (and the AP‘s invocation of it) that are relevant to this dissertation‘s discussion of press autonomy37.

First, the doctrine assumes the existence of news organizations that are distinct and visibly separate from publics. While not offering a definition of the press, it only affords short-term property rights to ―news‖ organizations. It is essentially concerned with commercial relationships among news organizations and the creation of fair, competitive conditions that will lead to the production of news for the public‘s benefit.

By assuming a recognizable distinction between news organizations and the public, this ruling assumes the existence of something identifiable called the ―press‖ that has property rights claims not afforded to the ―non-press‖. For this ruling to make sense in contemporary, networked news production that involves a wide variety of actors— with unclear and contestable identities and responsibilities—there would have to be some understanding of what an autonomous press looks like and how it is different from other entities. In its use of the hot news doctrine, the AP, though, offers no definition of the press or an argument for how such a doctrine would ensure its continued autonomy.

Second, recalling the discussion in section 2C of the changing role of speed in news production, it is unclear what ―limited period of time‖ means in contemporary journalism, and exactly how long this property right would extend. Indexing algorithms and online content aggregators that rank news websites can act virtually instantaneously. Should the behavior and speed of such code—not only professional journalists—be considered by the ―hot news‖ doctrine? Answering this question

37 For a good overview of the hot news doctrine and its potential impact on First Amendment law and news aggregators, see Bayard (2010). 193 requires a definition of press autonomy that is not static but that, in some way, accounts for news cycles and press times. In the context of the ―hot news‖ doctrine and legal distinctions at least, the press may only be autonomous—with property rights distinct from the general public—for a limited period of time. Determining this period‘s duration is a problem that crosses legal, social, professional and technical understandings of press autonomy since it touches upon how the law, journalists and computer algorithms all define ―limited period of time.‖

Third, this doctrine introduces the idea that news itself is ―quasi-property‖ with associated property rights, as distinct from copyrights that can be claimed on news stories. Writing for the majority in International News Service v. Associated Press,

Justice Pitney walks a fine line between recognizing that the news consists of both events that exist in the world (naturally occurring phenomena that occur in public for all to see) and renderings of these events that require material investments (a journalist‘s labor and the economic resources required to sustain her reporting). He gave the AP a short-term property right to protect it from competing news organizations who might appropriate the time, labor, skill and money the AP had invested in creating a news story, but he allowed the purchasers of newspapers to

―spread knowledge of its contents gratuitously.‖ (Epstein, 1992: 122) In relation to press autonomy, there appears to be a third distinction (in addition to separations between journalists and publics, hot and cold news) embedded in Pitney‘s ruling: he distinguishes between first-order material investments by the press in the production of something called ―news‖ and second-order conversations around and distribution of the news the press produces. Do the intellectual property claims of an autonomous

194 press—however it might be configured—extend to the use of news, not just its production?38

The AP is pursuing its version of a protected press not only through legal claims but also through the design of technical systems meant to give these claims material form (Associated Press, 2009a). Described in a confidential, internal document acquired by Harvard‘s Nieman Journalism Lab (Seward, 2009c), the AP is creating a 3-element system called ―Protect-Point-Pay‖ (also called ―AP3P‖) to

―address the urgent need to regain control of news content online.‖ (Associated Press,

2009b: np)

The AP3P system involves wrapping all digital content the AP produces in a proprietary ―news microformat‖ that contains the content‘s terms of use and a ―digital beacon‖ to track its online use and distribution. A ―news registry‖ of such tagged content will ―send reports back to the core database each time [a news] item is clicked on by an end user‖ with the ―IP address of the content viewer, the referring Web server and the time of use.‖ And ―since it is safe to assume that some users will intentionally or inadvertently remove the beacon … [a] ‗passive‘ tracking service will crawl the Web searching for [untagged] AP content [and] unauthorized uses will be pursued.‖ (Associated Press, 2009b: np) The AP3P system also distinguishes between two types of AP news content: ―utility‖ content and ―unique‖ content. ―Utility‖

38 Although now changed, in 2002 NPR‘s terms of use stated that ―linking to or framing of any material on this site without the prior written consent of NPR is prohibited‖ (Manjoo, 2002) and the Associated Press recently reached an agreement with Google News regarding how Google can access and republish AP stories (Liedtke, 2010). In slightly different ways, NPR and the AP essentially made a legal claim that it had the right to control not only how its content could be distributed, but how its content could even be described and accessed.

195 content is a ―type and amount of news that is quickly and easily available from other sources.‖ The system is designed to encourage this news to be shared and republished but, recalling the AP‘s hot news doctrine claims to news property rights, only among

AP subscribers. Such news is considered to be of little value and easily reproduced.

―Unique‖ content, though, consists of news that involves more in-depth analyses, history or context of a news story. The system is designed to let this content appear only on the AP‘s own website and to prevent it from being shared even among AP subscribers and definitely among non-subscribers.

The AP3P system appears to be essentially a technical defense, rendered in computer code, of the ―hot news‖ doctrine. Its internal documents describe the system using language similar to that used in Justice Pitney‘s 1918 decision and the subsequent relevant case-law. Specifically, AP claims that the AP3P system is designed to ―reduce misappropriation‖ of news and to ―limit or prevent redistribution of the kinds of information AP provides uniquely to ensure that hypersyndication does not drive down its value.‖ (Associated Press, 2009b: np, emphasis in original) This system is significant not only because of its connection to the AP‘s legal arguments, but also because of the assumptions its design contains. Principally, it assumes that the press essentially creates private goods, not public goods; that these goods must be protected from free-riders; and that information that is produced quickly and without context should enjoy fewer technical and legal protections than news crafted more carefully, over a longer period of time, containing more analysis.

196

As the AP‘s general counsel Srinandan Kasai explains, there is a great deal of economic benefit that can be gained from protecting and controlling the publication of unique content:

[T]he reason to do that is, then you have a bunch of links that point to a particular piece of content. You have better search outcomes. You have better exploitation of the link value, if you will, to that piece of content. So there‘s an ability to think of that piece of content differently because you‘re trying to maximize traffic as compared to other content where the benefit is really about getting the initial engagement. (Seward, 2009d)39

If such logics and systems expand beyond the AP, the general press‘s autonomy might rest not only on its ability to fund itself, empower its journalists‘ professional decision-making, or negotiate the terms of user involvement in news production – but to control the very terms on which news circulates, the types of opportunities that might exist for public learning, and the different depths of insight available on any given issue. Cast in these terms, the AP sees the news not in terms of Jenkins‘s

―participatory cultures‖ (centered on sharing, artistic expression, civic engagement and informal mentorship) or in light of Schudson‘s ―ideal journalism‖ (focused on empathy, diversity, mobilization and political decision-making) but, primarily, as the controlled production of private goods with economic value.

The aim here is not to resolve tensions in the hot news doctrine but, rather, to use it, the AP‘s property rights claims, and its design of technical systems encoding these claims to better understand one type of techno-legal shape a contemporary news organization‘s autonomy is taking.

39 Such statements are not unique to Associated Press leadership. The MinnPost‘s CEO and Editor Joel Kramer (former editor and publisher of the Star Tribune) recently said that his paper ―cares more about repeat readers than stray visitors‖ and is focusing on stories that attract the sustained and recurring interest of a reliable set of readers. Critical to this strategy, he claims, is producing stories that attract the attention of search engines and Facebook, the top two means by which MinnPost readers arrive at the website (Graber, 2010c). 197

Another type of response by news organizations to changes in press economics is the creation of new types of organizational partnerships and collaborative publishing technologies. Consider, for example, philanthropy-funded ProPublica40: a mainstream news organization that recently won its first Pulitzer Prize and that was founded by, among others, the former managing editor of The Wall Street Journal. It partners with 38 different news organizations to place its investigative news stories in their publications as well as its own website (ProPublica, 2010a). It also encourages anyone to republish its stories, under a broad but defined set of terms that allow anyone to ―steal our stories.‖ The terms:

- demand byline credit for both the ProPublica reporter and organization;

- prevent substantive editing of its stories beyond small stylistic changes;

- require linking directly to the ProPublica website and including a tag that lets

ProPublica track republications;

- allow placing advertisements next to its stories, but ban any advertisements

directly related to a story‘s content;

- prohibit ―automatically‖ republishing stories in any kind of bulk fashion.

Note here the mix of strategies: legal contracts rooted in some journalistic standards while still cognizant of the technological and economic details of online news distribution. ProPublica asserts its control over stories while simultaneously encouraging their distribution.

By partnering with everyone from the New York Times to Salon.com as well as encouraging anyone to reprint their stories, ProPublica increases the visibility of their

40 Although ProPublica‘s funding is primarily drawn from the Sandler Family Foundation, in January 2011 it began placing advertising on its site (Garber, 2011). 198 stories and reporters while simultaneously benefiting—and inviting risk—from its associations with official news organizations and unknown strangers. Thus,

ProPublica‘s autonomy—its separation from and reliance upon other actors‘ economic and reputational resources—depends upon forming and maintaining partnerships that let it strategically signal independence and association. Some of these relationships are on-going and broadly defined while others are negotiated individually: marketing strategies, publishing partnerships and funding arrangements are all crafted for particular issues, reporters and publication channels. This makes

ProPublica‘s autonomy a dynamic entity embedded within networked work practices and consisting of a continually changing set of relationships with both formal organizations and unaffiliated individuals.

In addition to publication partnerships, some online news organizations are coming together through the use of collaborative publishing technologies. That is, press autonomy is negotiated not only through intellectual property arrangements, strategic partnerships or distribution agreements, but through the shared use of networked technologies. The terms of each organization‘s autonomy—how it distinguishes itself from and depends upon other actors—are embedded in computer codes and programming practices.

For example, Publish2‘s NewsExchange system (Publish2, 2010a) is an

Associated Press-like wire service in which individual news organizations share news feeds to other news organizations within a closed network managed by Publish2

(Garber, 2010d). For example, if a news organization, self-identified journalist or blogger meets NewsExchange‘s editorial standards—the first of which is to

199

―[m]aintain editorial independence from governmental, commercial, or special interests, and … not engage in public relations, marketing and lobbying‖ (Publish2,

2010b)—it can install NewsExchange‘s content management system, begin sharing its content (not just stories but also reporter notes, background research, and other unedited material), and gain access to the materials shared by other news organizations that have been similarly vetted by Publish2.

Members can republish on their own sites any material shared within the network, either by reviewing the material first on a story-by-story basis or by automatically publishing material that matches keywords thought to be of interest to members‘ reading publics. They set the terms under which their materials are shared, control which members are allowed access to their material, and decide the fees (if any) they will charge for access to either their entire news feed or individual stories.

In many respects, the system is like a distributed, member-controlled Associated Press wire service in which self-selected partners who meet common editorial standards share news content within a private network.

Two features of NewsExchange stand out in relation to the online press autonomy. First, the editorial standards it requires all members to meet clearly define the system as a tool for practicing journalism, not for information sharing or general publishing. The price of admission to the network is a commitment to traditional tenets of journalism that aim to distinguish it from other online information collectives

(cf. email listserves and online social networks). NewsExchange separates itself from other actors through the use of traditional editorial standards.

200

Second, once members have demonstrated their journalistic commitments to

NewsExchange managers, they enter into a new set of dependencies in which editorial decision-making can be replaced or supplemented with editorial algorithms. When designing their news feeds—choosing NewsExchange‘s content to publish on their sites—members essentially program their publishing. They choose the organizations or individuals they wish to follow, the keywords they are interested in, the distribution terms they are willing to abide by, and the costs they are willing to pay. In contrast to

ProPublica‘s prohibition on automatic publishing, NewsExchange allows (and seems to encourage) news organizations to increase their web traffic by automatically publishing content. Members publish through a kind of journalistic programming, writing code that selects materials according to rules and envisioned, desired outcomes.

If these programs function as envisioned, small news organizations, independent journalists and lone bloggers—many of whom may have few resources and small audiences—are inexpensively and easily able to add presumably high-quality journalistic content to their sites, generating much-desired web traffic within budget restrictions. However, such a system may also result in:

- traffic-optimized publishing in which, much like online advertising systems,

content is selected for its likelihood of attracting visitors to a member‘s site;

- content-optimized publishing in which there exists an ideal distribution of

resources that will, over time and using stochastic processes of analysis, ensure

that the highest traffic is generated with the lowest cost (e.g., a member may

publish a certain number of free or low-cost stories with generous terms of

201

distribution, and another number of higher-cost stories with more exclusive

terms)

- post-hoc editing in which members periodically review their automatically

generated feeds and tweak their news feed algorithms, but only after they have

published NewsExchange content that may or may not have been manually

selected.

These may seem like overly negative characterizations of NewsExchange‘s potential.

To be sure, NewsExchange and systems like it have a great deal to offer smaller news organizations and independent publishers genuinely looking to create high-quality publications. But note, too, that each of the limitations described above is a feature of economic markets and information exchanges: systems like NewsExchange make it easy for publishers to program algorithms to allocate resources and track consumer desires optimally with the least amount of effort. Such systems may allow more entrants into the market and offer greater efficiencies, but they do not necessarily create news experiences that ensure that citizens learn what they need to know, in addition to knowing what online systems predict they want to know.

In one respect, the NewsExchange system—and similar systems like

Facebook+Media and NewsCloud—has all the hallmarks of an autonomous press system: its members commit to being editorially independent, original, fact-based and free of conflicts of interest (Publish2, 2010b). But this autonomy begins to change when we look at the system of news production that NewsExchange affords and the broader context in which it operates. A close reading of the NewsExchange system suggests that at least its system for networked news publishing is influenced by logics

202 of automaticity, optimization, and market exchange that may influence the conditions under which seemingly autonomous editorial judgments are made.

2.E. Distributed Human-Machine Intelligence

Related to the AP3P and NewsExchange systems is a collection of online news technologies and practices in which press autonomy is negotiated not only in relation to industry economics, reader participation, professional practices and organizational systems, but amidst a set of tools that purposefully aim to distribute news-making intelligence among humans and machines. Such news production can be seen in terms of two broad sets of information-producing activities: automated tasks performed by algorithms and robots that sense aspects of their environments and enact programmed responses; and other, more nuanced tasks in which journalists and editors use their judgment to make decisions about what the news is or should be in a particular context. In this way, news work is dispersed across machines and humans, and press autonomy—from the perspective of humans—becomes a function of how news professionals separate themselves from and depend upon machine capabilities.

This idea of distributed human-machine intelligence—in which perception, intelligence, and decision-making are distributed ―across minds, persons, and the symbolic and physical environments, both natural and artificial‖ (Pea, 1993: 47)—has a modern history that can be divided into two broad periods. The first period concentrated on how humans and machines might come together to create more efficient, reliable production of material goods. Grouped under the term ―scientific management,‖ this was a set of techniques designed to rationalize human labor in

203 terms of production systems – to couple human and machine capabilities in the most efficient ways possible. Such systems—ideally created to maximize both an owner‘s profits and a worker‘s wages—were created and managed by planning and engineering departments who kept detailed records of how labor might be optimized across people and machines. Essentially, this optimization involved designing machines and training people to create a single, synchronized, efficient production unit

(Chandler, 1977; Taylor, 1911/2005).

In the same vein, Shannon-Weaver information theory (Weaver & Shannon,

1963) aimed to formalize systems for signal exchange between machines in order to decrease communication errors and increase the quality of data transmission. The overarching goal was to create a standardized language of transmission that would ensure a communication channel‘s reliability. The message that was received would be an exact copy of the message sent. In practice, this meant designing not only a technical architecture capable of consistent and dependable performance, but operators—human senders and receivers—who could recast their communication into the forms that architectures expected (e.g., consider telegraph operators and ordinary people who were rewarded for speaking with ―brevity and efficiency‖ [Marvin, 1990:

21] on early, fragile telegraphy systems that would frequently fail).

In Norbert Wiener‘s (1950) theories of cybernetics, communication systems were understood not only as ways of reliably transmitting information, but as feedback systems. Machines could be controlled through meta-systems that were aware of their environments and human operators‘ capabilities so that they could monitor not only whether the same message was sent and received, but whether the right message was

204 sent and received. In contrast to the Shannon-Weaver model that equated reliable data delivery with successful communication, cybernetics was concerned with creating human-machine couplings that increased the likelihood of contextually appropriate meanings being exchanged.

Especially in the context of complex, post-Second World War weapons technologies capable of instantaneous global effects (nuclear missile systems that required human operators to make life-and-death judgments about distant contexts by trusting data streams and deterministic procedures), Weiner saw a need to manage the probabilistic nature of communication. The probability of errant communication was reduced if people and technologies were understood as a joint system, a

―servo/organism‖ in which humans and machines would give each other constant feedback about what they were sensing and perceiving so that misunderstandings could be repaired before either made a serious error (Edwards, 1996: 180-187)41.

Autonomy was a property of human-machine couplings, not an issue of whether an individual was free or informed enough to make decisions.

Cybernetics was arguably the conceptual precursor to a second period in the modern history of distributed intelligence that focused more on the meaning and contextualization of human-machine systems. This phase saw designers‘ intentions and assumptions as integral to a technology‘s meaning because, it was argued, they suggested or afforded ways in which technologies were to be used and appreciated

41 It is important to note that such thinking about humans and machines as complementary systems were developing at the same time—and in conversation with—other fields that were increasingly seeing societies in terms of ecologies and probabilities. For example, at the Macy Foundation conferences ―biologists, physicists, and mathematicians, including cyberneticisians such as Arturo Rosenblueth and Warren McCulloch, psychiatrists such as Ross Ashby, and sociologists and anthropologists such as Gregory Bateson and Margaret Mead … helped refine a number of cybernetic concepts, including the relationship between a system and its observers and the nature of feedback.‖ (Turner, 2006: 26-27) 205

(Kling, 2000, 2007; Sawyer & Tapia, 2007; Suchman, 1996; 2007). This phase identified a new level of analysis in the social and technical design of infrastructures.

An exchange between Suchman (1994) and Winograd (1994) illustrates a central tension of this era. In her landmark article ―Do Categories Have Politics?‖

Suchman identified a class of ―coordination technologies‖ that ―provide canonical frameworks for the representation and control of everyday communication practices.‖

(1994: 178) Her concern was that such technologies were based on a set of uncritiqued and largely invisible theories that aimed to ―discipline and control‖ people‘s actions. Essentially, when a person interacted with a piece of software they were confined within the set of behaviors that the interface—and its underlying model of the person—allowed. The system implicitly signaled some behaviors as acceptable and others as impermissible, without the ―user‘s‖ knowledge or consent42. Suchman essentially argued that the person in a human-machine coupling is far less autonomous than either the machine or its designers because anyone who enters into a relationship with a machine unknowingly surrenders part of her agencies, becoming an instrument of a system in ways that are beyond her comprehension or ability to influence.

Since it was his system that Suchman used to illustrate her argument,

Winograd (1994) responded to her article. Although he largely dismisses her core conceptual argument as ―sociopolitical drama,‖ (191) he accepts the idea that no system can ―fully capture the richness of mental life or social interaction.‖ (192) He argues, though, that standardization and coordinated action are essential for any set of

42 As with cybernetics, Suchman‘s critique was heavily influenced by ideas that were seemingly outside of strictly technological fields. E.g., she imports Foucault‘s (1979) concept of ―disciplinary practice‖ into her argument to warn against systems in which people‘s bodies become docile and subject to the wishes of controlling operators. 206 distributed activities that aim to accomplish a goal. His solution to Suchman‘s charge of disciplinary control is to invite people into the creation of computational systems, arguing that ―design succeeds when it is grounded in the context and experience of those who live in the situation.‖ (195) Winograd essentially shifts the problem of autonomy from one of human-machine interaction to one of system design: people will be free when they participate in the creation and ongoing revision of the systems they use.

This idea is particularly important in understanding networked press autonomy because it suggests the need not only to examine how professionals and publics use online news systems (like the Associated Press’s AP3P or Publish2‘s NewsExchange) but the extent to which they are involved in the conceptualization and construction of these systems. Without looking at the design processes that lead to online news systems, we may mistake individuals‘ easy, orderly and uncontroversial use of systems with their informed consent to, and knowing acceptance of, the categories that structure their behaviors.

Understanding online press autonomy thus becomes, in part, a study of infrastructure and the conditions under which networked presses and publics behave.

Infrastructure is fundamentally a relational and networked concept. It is a system that runs ―underneath‖ (Star & Bowker, 2006: 151) to connect, regulate and enable action, and is defined in the eyes of beholders (one person‘s infrastructure is another person‘s application). More formally, Star & Ruhleder (1996:113) define infrastructure as:

- embedded within ―other structures, social arrangements and technologies‖;

- transparent to use without requiring reinvention or assembly for each task;

207

- scoped beyond a single event or site of practice;

- learned as a member of a group that agrees upon its taken-for-grantedness;

- shaping and shaped by conventions within communities of practice;

- embodied in standards that define the nature of acceptable use and replication;

- built on an installed base of other infrastructures with similar characteristics;

- visible only when it breaks down.

Such a definition provides a rich starting point for thinking about online news systems as human-machine infrastructures in which press autonomy is negotiated through the design, definition and use of infrastructure. Who builds the infrastructures—the assumed, embedded and translucent rules, relationships and technologies—that create, deliver and disseminate news in online environments? What systems and practices are considered applications to some, and infrastructures to others? Where do we see

―boundary infrastructures‖ (Bowker & Star, 1999: 287) that reveal and maintain separations among communities of practice, e.g., casting software programmers as engineers who create the foundations of online news infrastructure, journalists and editors as end users of news systems, and readers as consumers of infrastructure outputs? When do such separations blur and what role does infrastructure play in this blurring? And—following Winner (1986), Bowker & Star (1999) and Friedman et al.

(2006) who, respectively, chronicle the politicization of built urban environments, information classification schemes, and computer-supported cooperative work systems—what political and cultural assumptions are embedded within news infrastructures that might influence whether publics think that an autonomous press should exist at all, and affect how courts interpret the First Amendment‘s press clause?

208

As Bowker and Star (1999:4) found in their historical study of classification systems, infrastructures can define the ―moral and aesthetic choices that in turn craft people‘s identities, aspirations, and dignity.‖

How do such infrastructures appear in online news production? In line with

Star & Ruhleder‘s observation that infrastructures grow out of other ―installed base[s]‖

(1996: 113), online news systems emerge from a kind of infrastructural bricolage in which actors purposefully draw upon and reuse the systems and rationales that have underpinned previous information systems. Indeed, such ―installed bases‖ have always been present in the production and distribution of news. On a field level, Cook

(1998) details a long history of press reliance on state-sponsored and state-regulated infrastructure: the Postal Act of 1792 established a system of subsidies that let newspapers be mailed cheaply; Congress exempted newspapers from minimum wage and child labor laws, effectively subsidizing their labor costs; the Federal

Communications Commission was established to rationalize and regulate radio spectra, creating a predictable technological environment within which radio news could develop; government officials and agencies provide extensive public relations services to news organizations, providing a steady stream of information to reporters; and the Department of Defense sponsored the initial work that led to the development of the internet‘s core infrastructure.

On the more micro level of journalistic practice, technologies and taken-for- granted practices have always structured and enabled news work. Dooley (2007) describes a number of them: increasingly fast printing presses that let news organizations print multiple editions per day; telephone and telegraphy systems that let

209 far-flung reporters call in stories to centralized editors; the inverted pyramid style of writing that survived long after it had first been invented to ensure that a field reporter could communicate a story‘s lede before a telegraph line failed; increasingly portable and high-quality photographic and sound equipment that gave audiences rich experiences and let news organizations with greater technical and financial resources claim advantages over poorer competitors.

Similarly, in their history of the ―forms of news,‖ Barnhurst & Nerone (2001) describe how newspapers, throughout history, constantly emerged from interplays between physical media and professional practices that, together, defined their outlooks. Their claim is that a newspaper is not only a reflection of journalists‘ relationships with sources, professional practices of the day, and the economics of print production. Rather, a newspaper is a record of the ideas and people that could— in a particular set of material and ideological conditions—seen and be re-presented to journalists.

The material forms of newspapers—e.g., what appeared on the front page as opposed to internal pages, which stories had accompanying visuals, how advertisements were distinguished from editorial content, whether stories were signed or unsigned, the appearance and disappearance of letters from readers—offer evidence of how journalists knew their worlds, how their technologies enabled them to see the world, and what kind of newspaper-supported civic cultures were possible (Nerone &

Barnhust, 2003). Only concepts and sources that ―fit‖ with a news organization‘s culture and technological practices could be rendered in print. All others were essentially invisible to journalists and, therefore, their reading publics.

210

Today, such visibility is defined by a new set of techno-social infrastructures with a great deal of influence over what news organizations can see, and what citizens can read. One such infrastructure is search engines. As noted above (e.g., Federal

Trade Commission, 2010a; Graber, 2010c; Rice, 2010) search engines play an increasingly important role in attracting readers to news organizations‘ websites.

They are among the most popular sites on the internet and, although there are many different search engines available, 44 percent of internet searchers rely on a single search engine and consider themselves satisfied with its results (Fallows, 2005). But search engines are not neutral indexers of web content (Hargittai, 2007). Indeed, they have been shown to ―systematically exclude (in some cases by design and in some, accidentally) certain sites and certain types of sites in favor of others‖ (Introna &

Nissenbaum, 2000a: 169). They tend to: privilege sites that are commercial in nature with particular domain names and many back-links over smaller, less professionalized sites (Introna & Nissenbaum, 2000b); favor U.S. sites over comparable sites in other countries (Vaughan & Zhang, 2007); and be susceptible to manipulation by those with the technical ability to reverse-engineer their sorting and indexing algorithms (Bar-

Ilan, 2007).

Another kind of techno-social infrastructure under which the press finds itself operating is a collection of system in which editorial decision-making and journalistic judgments are distributed among system designers, operators and press professionals.

Although it may seem fantastical and science-fiction like, consider a ―journalist robot‖ recently invented by scientists at Tokyo University (Matsumo, Nakayama et al, 2007).

It is a machine that seeks out—according to perceptual rules given to it by its

211 designers—newsworthy features of its physical environment. It interviews people, generates news stories describing its findings, and publishes them to the web.

Although certainly experimental and not in widespread use, such a robot is not unique in the history of networked journalism; the project extends an earlier robot designed by MIT professor Chris Csikszentmihalyi called the Afghan Explorer. The Afghan

Explorer was a non-autonomous proxy for journalists unable to report from dangerous

Afghan locations: as envisioned, reporters would use its wheels as feet, its cameras as eyes and its speaker as a mouth to seek out and interview local Afghanis (Wakefield,

2002).

Or, in a similar, software-oriented project, consider Northwestern University‘s

―Stats Monkey‖ project. Developed by both journalists and computer scientists

(Gordon, 2009a), it uses ―information commonly available online about many

[baseball] games‖ and a ―library of narrative arcs that describe the main dynamics of baseball games‖ to ―automatically [generate] the text of a story about that game that captures the overall dynamic of the game and highlights the key plays of and key players.‖ In an effort to replicate a kind of objectivity reporters often claim from telling ―both sides‖ of a story, the Stats Monkey system generate stories ―from the point of view of either team.‖ (Intelligent Information Laboratory, nd) Although criticized by some journalists who questioned whether stories should have anyone‘s perspective, whether the system is extensible to other, less-structured news domains, or whether the system buries the lead (Carr, 2009), others noted that the system might be a boon for sports news efficiency and inclusion. It could ―instantly write a game story as soon as the last out is made, freeing a reporter to go down to the field or the

212 locker room to do interviews‖ or ―create stories about games—for instance, college baseball—that are not routinely covered by professional journalists.‖ (Gordon, 2009b)

Understood as such, automated journalism essentially solves a labor problem, creating more reports on a greater number of issues, for less money and with fewer people.

Moving away from the idea of robotic journalists, proxy reporters, or automatically generated stories, consider systems that systems that automatically processes other news stories and guide human decision-making. Reuter‘s NewsScope program is

a machine-readable news service designed for financial institutions that make their money from automated, event-driven, trading. Triggered by signals detected by algorithms within vast mountains of real-time data, trading of this kind now accounts for a significant proportion of turnover in the world‘s financial centres. Reuters‘ algorithms parse news stories. Then they assign ‗sentiment scores‘ to words and phrases. The company argues that its systems are able to do this ‗faster and more consistently than human operators.‘ Millisecond by millisecond, the aim is to calculate ‗prevailing sentiment‘ surrounding specific companies, sectors, indices and markets. Untouched by human hand, these measurements of sentiment feed into the pools of raw data that trigger trading strategies … The latest iteration of NewsScope ‗scans and automatically extracts critical pieces of information‘ from US corporate press releases, eliminating the ‗manual processes‘ that have traditionally kept so many financial journalists in gainful employment. (Kirwin, 2009)

In this model, automated journalism solves a meaning-making and information overload problem. Too many news stories are being generated for readers to appreciate broader patterns or critique trends so meta-news systems are required to parse and automatically analyze stories. Most critically, these analyses need to be rendered in formats understood by other machines, e.g., ones making financial decisions.

Such a system is not unique to niche financial reporting. The Blogosphere

Ecosystem (2010) and Technorati (2010) perform similar functions by analyzing, indexing and ranking blogs. (Since their algorithms are not public, it is unclear 213 exactly how they distinguish between blogs and other websites, or exactly how their rankings are generated). Google News follows a similar approach when it uses a proprietary algorithm to determine the ―selection and placement of stories … automatically by a computer program.‖ (Google, 2009a) (Although, note that in June

2010 Google News announced a new feature called ―Editors‘ Picks‖ in which ―real live human news editors at partner news organizations‖—including The Washington

Post, Newsday, Reuters and Slate—select news stories that are placed alongside those chosen by Google‘s algorithm [Garber, 2010a, emphasis in original].)

There is some evidence (Farrell & Drezner, 2008; Lowrey & Mackay, 2008; Peters,

2010) that reporters sometimes use such indices when preparing stories, to search for blog sources or gather quotes for particular stories and other studies showing that news media agendas are actually co-constructed by both mainstream journalists and highly- ranked blogs (Wallsten, 2007). Essentially journalists seem to use rankings like

Technorati‘s or the Blogosophere Ecoystem‘s as, at best, starting points and references for stories and, at worst, as proxies for their own editorial judgments about sources or story themes.

This kind of sourcing is not new. Press researchers (Fishman, 1980; Sigal,

1973, 1986) have consistently shown how reporters working under tight deadlines often rely on bureaucratically reliable sources for story contexts and quotes. These sources are individuals and institutions that are trusted—because of their past behavior, organizational roles or ideological identities—to give predictable opinions.

What is new here is the nature and location of a source‘s trust and reliability: within privately created and administered algorithms that automatically analyze and rank

214 potential sources according to designs and rules to which journalists and publics generally do not have access.

Essentially, this kind of system is akin to the online ―echo chambers‖ Sunstein (2001) predicted, but instead of emerging from individuals’ choices—to filter information and hear self-confirming opinions—these are institutional echo chambers. Such chambers are created when users adopt hierarchies presented to them by systems that ranking existing news stories (Carlson, 2007) and when reporters use information infrastructures they cannot critique to create stories, giving an appearance of professional sourcing and public participation when, in fact, their sourcing and publics are those presented to them by their information infrastructure. I do not claim that online reporters mostly, or even regularly, use blog indices to find sources. Rather, such systems and their documented use by reporters is another element of a set of networked dependencies and separations out of which the press might fashion its institutional autonomy.

These experiments in automated reporting—robots that variously embody or replace journalists, algorithms that automatically generate news stories or extract meaning in existing reports, and indexes that rely on a mix of machines, humans and crowds to organize and judge the value of news—are only part of a larger landscape of increasingly automated, networked practices that distribute news work among humans and machines.

An online press that heavily relies upon search engines, algorithms and robots—to conduct online research, respond to audiences or find sources—risks letting its autonomy be defined, in part, by the kind of assumptions search engines make in

215 their indexing and the kind of information-seeking behaviors users presume to be adequate. I do not mean to suggest that editorial judgments are always being replaced with algorithms or automatic processes, nor do I think that search engines purposefully exclude information or preference some actors over others. However, if the press is to define and defend its autonomy amidst a host of other influences and actors—each operating under their own distinct but overlapping systems of norms—its practitioners, critics and champions need to understand the technological tensions under which such autonomy operates. Explicitly or otherwise, these technologies act as infrastructures that influence how the press relies upon and separates itself from other sites of power—legislatures, courts, professions, populist movements, advertisers—that have historically set the conditions under which publics hear.

3. Press Autonomy & Newsware

The aim is not to continue to review experiments in online journalism and their various tensions, nor is it argue that the examples offered here are representative of a rich diversity of networked news models. They are too numerous, too wide-ranging, and too dynamic to be captured in a single typology.

Rather, the point thus far has been two-fold. First, to show a variety of ways in which the online press is undergoing transition in its economics, relationships to readers, professional practices, organizational systems, and distribution of intelligence among machines and humans. These are the primary conceptual and empirical sites in which press autonomy appears. It is essentially an answer to the question: in this

216 particular historical moment, what changes in the press are occurring that impact the material and symbolic conditions under which its autonomy is negotiated?

Second, throughout the discussions of these changes, I have tried to weave a particular framing of contemporary press autonomy that underpins this dissertation‘s theoretical framework. That is, the online press‘s autonomy is, and will likely continue to be, fundamentally a problem of relational infrastructures that take form in networked information technologies and practices. The press‘s autonomy lives not in any legal or professional definition—bracketing what the press clause means or what a journalist is—but, rather, in a series of ongoing techno-social negotiations in which press autonomy is crafted out of both conceptual and material dependencies and separations. This is the primary theoretical contribution of this dissertation: a new model of press autonomy that considers the design and use of networked information technologies alongside legal and professional considerations of reporting practices and journalistic ideals, precisely because it is in the creation and articulation of these technologies that the press‘s autonomy can be defined and defended.

Together these two points lead to another contribution this dissertation aims to make. In developing this model of contemporary online press autonomy, I propose a new term ―newsware‖ to describe the techno-social design and use of networked information technology infrastructures upon which press autonomy rests. That is, newsware is both a set of technologies—e.g., the user-generated content systems that the BBC or CNN might produce, the rules guiding a remote robot reporter, the content-tracking system the AP aims to use to monitor the online spread of its stories, or the algorithms that Technorati and Google use to judge blogs and news stories—

217 and practices—e.g., how professional journalists judge the quality of content they did not produce, how sources are found and evaluated online, how non-journalists come together to critique and challenge press authorities, and how system designers decide which aspects of reportorial and editorial judgment to automate.

These technologies and practices are laden, implicitly or otherwise, with values about what journalism is or should be in online networked environments. To uncover and critique these values requires a conceptual framework—grounded in explications of what ‗autonomy‘ means for both individuals and organizations—that might help designers decide what to create, journalists choose how to report, and readers select what believe or when to participate in news systems. And although the Supreme

Court has yet to grapple with many of the core tensions described in this introduction—Chapter 2 includes a more complete review of press-clause jurisprudence—I suspect that a concept like ―newsware‖ will be required for courts, journalism advocates and challengers, and publics to articulate and regulate what press freedom should mean going forward43.

Whenever the press outsources to technologies the means by which it investigates power, helps people encounter new information, or facilitate community conversations, it delegates part of its structural, first amendment role in the ―system of freedom of expression‖ (Emerson, 1970) to algorithms and technological practices.

Perhaps the very idea of—and necessity for—a ‗press‘ fades as technologies that index public expression subsume the press‘s ideal editorial functions. Or perhaps the privileges that the state has historically given to the press—e.g., exemptions from

43 For example, courts in Florida and Texas have ruled that the identities of people who post comments on online news stories are protected by press shield laws—implying that the press includes not just journalists‘ stories but discussion thereof—while courts in Illinois and Nevada have rejected such logic. 218 labor laws, shield laws that excuse journalists from giving court testimonies, public subsidies, tax exemptions, preferential postal rates, access to restricted news conferences (Cook, 1998)—should be given to any fourth-estate-like actor (individual, organization, technology) that can demonstrate and normatively defend its own type of autonomy. Or perhaps the mainstream press should retain its privileges but the state and the courts should retain their power to define the press, structuring and defining its freedom as they see fit (Baker, 1989: 271).

Essentially, I argue here that since the work of journalism is now distributed among a set of organizations, technologies, people, groups, norms, laws that are increasingly networked and sometimes difficult to see, the idea of press autonomy needs updating. Some significant parts of the press‘s core function and identity— recall the ideals articulated by Schudson—now live not within any one profession, economic model or in-house set of tools but, rather, as a set of relationships that rely upon code: user models, input assumptions, data structures, algorithms, error-handling and standardized outputs. These codes—how they function, are created, talked about, and used—are a set of ―listening technologies‖ that, along with professional norms and legal statutes, constitute an infrastructure that defines the conditions under which publics can hear. They are a new way to re-evaluate the First Amendment‘s press clause‘s affirmative interpretation and the material conditions of the press‘s structural functions – to revisit Meiklejohn‘s assertion that First Amendment‘s ―point of ultimate interest is not the words of the speakers, but the minds of the hearers.‖ (Meiklejohn‘s

1948: 25) What does it mean to value the press‘s structural, listening function in a

219 context where its autonomy is distributed across sites of network technology design and use?

Most broadly, this dissertation might be used to help answer existential questions asked of and about the press. Too often today it is suggested that the professional press—and its constitutional protections in the press clause—are essentially anachronisms, remnants of a time when individuals lacked the means to publish to broad audiences. Now that, theoretically at least, anyone has the ability to speak online (e.g., creating a blog, Twitter feed or Facebook page), so too is it often assumed that everyone has the potential power to attract and hold an audience.

Too little is understood about dynamics among online speakers and audiences, but the central struggles can best be characterized as a tension between two models.

One, most often celebrated by defenders of traditional journalism, emphasizes the need for institutions that—even at the expense of public participation and transparency—reliably produce and vet the kind of quality information citizens need to be self-governing citizens. It shows how populist online information cultures often replicate offline inequalities and divisions, recreating separations among socio- economic groups and ethnicities that inclusive public spheres are ideally meant to overcome (Hargittai, 2008; Hargittai & Walejko, 2008; Hindman, 2008; Nakamura,

2007; Zillen & Hargittai, 2009).

The other model, usually offered by celebrants of online information tools and cultures, argues that reliable, quality information will emerge from networked, commons-based arrangements in which individuals self-associate and organize to create and critique data and opinions. Mainstream media institutions are criticized for

220 taking in too few perspectives, concentrating power in too few commercial owners, and focusing on creating large audiences that can deliver much-needed advertising revenue (Benkler, 2006: 196-211). In their ideal form, networked information tools enable a new kind of collective action that ―challenges existing institutions by eroding

[their] monopoly on large-scale coordination‖ (Shirky, 2008b: 143) and that creates a new set of ―digital free speech values‖ that privilege ―interactivity, mass participation, and the ability to modify and transform culture.‖ (Balkin, 2004: 1) Although there is emerging evidence that online news environments involve less segregation of people with different ideologies and viewpoints than do traditional offline news media or even face-to-face interactions among neighbors, co-workers or family members

(Gentzkow & Shapiro, 2010), advocates of this model tend to use anecdotal evidence to argue that self-organized news systems will produce desirable outcomes. Without specifying exactly how this self-organization functions or what defines desirable, such celebrants tend to equate what is technologically possible with what is socially, economically and cultural plausible – and desired.

A detailed critique of these two models is beyond the scope of this dissertation, but they are presented here to demonstrate the kinds of environments in which the press‘s existence and autonomy is and will likely be debated. Understanding contemporary press autonomy does not have to be an exercise in defending the First

Amendment‘s speech clause against its press clause, pitting the protection of individual expression against the value of institutional listening. Rather, by re-casting press autonomy in terms of contemporary actors and conditions, we might see that the

221

―press‖—as a democratic institution with constitutional protections—can actually be constructed out of a particular configuration of actors and social forces.

4. Conclusion

Perhaps we should defend a press not when it is located in a particular tool, person or organization—e.g., a technology that enables mass publishing, an individual trained in journalistic methods, a company with the resources to fund reporting—but, rather, when a state of normatively defensible press autonomy exists. This autonomy is a dynamic, multifaceted phenomenon that occurs when a set of dependencies and separations, together, constitute the values that a protected press is meant to embody.

As a link to the next chapter, it is worth reiterating that the overall goal of this dissertation is threefold: to provide a theoretical and conceptual framework for understanding press autonomy; to argue that this framework exists in a set of a set of networked, socio-technical tools and practices I call ―newsware‖; and to evaluate this framework in relation to the design, discourse around, and use of one particular kind of newsware: mainstream news organizations‘ application programming interfaces

(APIs), essentially code used between professional journalists and programming publics to mediate the production of online news. The next chapter describes what an

API is, what a news API is, and how studying news APIs can help us better understand online press autonomy as a negotiated concept among multiple actors. It describes three news APIs and identifies how each uses its APIs to negotiate its autonomy as a professional news organization.

222

CHAPTER FIVE:

Application Programming Interfaces as Sites of Press Autonomy

I’ve been working in the newsroom for the last few months. On the surface, there’s not much that distinguishes the modern journalist from the modern software developer. Gone are the days of grizzled hacks grousing into phones and banging out copy five minutes before deadline on dented Underwoods. Modern journalists work against increasingly continuous deadlines and are as likely to collaborate via e-mail and instant message as by phone. And they bang out their stories on computers just like mine in cubicles just like mine.

But there is one thing that distinguishes me from the correspondents around me. Their primary audience is humans; my primary audience is computer programs (hello there 00:1b:63:a1:34:d4, you‘re my favorite reader). Increasingly, however, our audiences will converge. As online readers face more and more information overload, intelligent computer programs will become the initial readers, filters, sorters, and social planners of all the content those journalists produce (and the bloggers blog and the critics criticize).

-- Jacob Harris, Senior Software Architect, New York Times [http://open.blogs.nytimes.com/2007/12/05/the-magical-minimalism-of- microformats/]

1. Introduction

The story of U.S. press autonomy is complex and interdisciplinary. I have so far attempted to tell it from a number of different perspectives, specifically focusing on:

- conceptual underpinnings of autonomy as an ideal and its legal and

jurisprudential application to the U.S. press;

- the history of the U.S. press as a field by focusing on different ways it has

configured itself, its separations and its dependencies;

- the state of the contemporary online press by describing how, through a critical

reading of its newsware infrastructure, it distances itself from and depends

223

upon market-based funding models, participating readers, traditional

professional journalistic practices, intellectual property regimes and

organizational systems, and hybrid human-machine techniques for distributing

work.

This chapter aims to build upon these understandings and focus on a particular kind of newsware infrastructure, the news application programming interface (NAPI). This chapter has three broad sections. The first articulates what an application programming interface (API) is, what a news API is, and why news APIs are a critical site in which to understand contemporary press autonomy. The second describes three leading news APIs, their motivations, structures and functions through a two-part technique: by analyzing existing public documents that detail and discuss these APIs; and by experimenting directly with one news organization‘s API, describing the experience of using it to build a simple news feed. The third part of this chapter uses this investigation of NAPIs as evidence of a 3-element typology of press autonomy in which news organizations offering NAPIs are able to negotiate their field-level independences and dependencies through regulatory language, system design, and network convening.

2. What is an Application Programming Interface (API)? What is a News API?

And Why Study News APIs to Understand Press Autonomy?

Before defining an application programming interface (API), it may help to start with an analogy: a post office is a kind of API. For example, if you bring a parcel of a particular size and weight to an office during its open hours, stand in (a likely very

224 long) line and pay a certain amount of money, the post office will ensure that your package arrives at a specific location at a given time. While few people understand exactly how a post office works internally, people generally understand how to behave in relation to it. Things like zip codes, classes of stamps, mail boxes, and delivery times constitute a standardized, learned, shared, and periodically revised language that you and the post office speak. It is this language that enables you to enter into a contract with the post office: if you follow its rules, you can, for the most part, depend upon its services.

Most broadly, an API is essentially a way for two computer processes to communicate with each other. It is fundamentally an interface, a set of rules by which one software program can access the resources of another software program. More formally and technically, an API is a kind of ―middleware‖ that:

offers general services that support distributed execution of applications … [I]t is software positioned between the operating system and the application. Viewed abstractly, middleware can be envisaged as a ―tablecloth‖ that spreads itself over a heterogeneous network, concealing the complexity of the underlying technology from the application being run on it … [An API] defines how an application can access the functionality of the middleware. (Puder, Romer & Pilhofer, 2006: 21)

To computer scientists, then, an API is a type of abstraction that makes system design and maintenance easier, more reliable and less costly. It is technique by which programmers can: hide and make inaccessible certain data or details of a system, while simultaneously offering approved and easy ways of accessing other parts of a system; reliably plan and create new hybrid systems that combine multiple data sources; design programs in a modular, distributed fashion by making it easy to isolate and test functions; speak a common language and organize their work, letting multiple

225 programmers know what kind of software can and cannot be designed; reuse and refine code, making it less likely that work will be duplicated.

But APIs also make it possible to centralize system control. For example: an error within API code may adversely impact multiple systems that use and rely upon the API code, making second-order systems vulnerable to weaknesses in API code; a

―chilling effect‖ on certain kinds of second-order systems design that are either not possible under the API‘s technical restrictions or that are simply more difficult to program; a change to an API‘s technical functioning or its terms of service may interrupt a second-order system that had come to rely upon the API offering a particular level of access or functionality.

Most fundamentally, a news API is system of rules by which a distributed set of actors can build new online applications and services using data that is produced, vetted and distributed by news organizations. It is essentially a structured and controlled way for people outside of news organizations to access information created or evaluated within news organizations; simultaneously, it is a means by which news organizations can share information beyond their normal distribution or syndication networks, establishing themselves not only as people who tell particular stories to known audiences but also as professionals who certify data and encourage others to create work with that data.

For example, consider the following applications that programmers have created with the three news APIs studied in this chapter.

226

Guardian API: “Nukeometer” Created by: Adam Charnok [nukeometer.com]

This application was created using the jQuery library to combine data from the Google Maps API with data published in the Guardian article ―The World in Active Nuclear Weapons,‖ [118] an article that linked to data published by the Bulletin of Atomic Scientists [119] and vetted by the Guardian [120].

Users input their city and country and the application calculates how many nuclear weapons can reach that location. (The author asks that users ―please, please, please do not make any important decisions based on this information.‖)

NPR API: “Reverbiage World Widget” Created by: Geoff Gaudreault [www.reverbiage.com/widgets]

This application indexes all NPR stories according to their location on the globe, letting users access story text and audio by manipulating a 3-D globe and accessing stories from a particular region.

227

NYT API: “CEO Politics” Created by: Jeff Larson [www.ceocampaigncontributions.info]

This application uses the NYT‘s Campaign Finance API and the Google Maps API to display and compare the campaign donations made by companies,

CEOs and industries, sorted by regions of the U.S.

In addition to the information-focused aspects of news APIs (e.g., giving publics access to data outside of the traditional formats offered by news organizations and letting news organizations become information clearinghouses and repositories), of great interest to this dissertation—and to the goal of understanding press autonomy in contemporary, distributed news systems—is the idea that news APIs are a particular kind of site. They are a place in which a new set of actors comes together: reporters

(through the stories they produce that become API data), editors and managerial staff

(through the information policies and terms of service they create and maintain), in- house API software engineers (through the technical design decisions they make about how data access will be implemented), external programmers (through the systems they design that access the news APIs), external information providers (organizations and individuals that provide data to news organizations that are then accessed through the API) and readers (those who use the systems created by external programmers). If press autonomy is, as early chapters argued, a distributed and multi-sited phenomenon

228 in which the press defines itself and its independence through a variety of relationships and dependencies, then APIs offer a new and critical location where such autonomy is negotiated.

APIs are, in a sense, new tools of news syndication. But, in contrast to syndication systems of the past that focused on distributing a particular piece of content to broad audiences, news APIs are infrastructures—tools, standards, norms, assumptions, and shared practices—through which news is not simply published but, rather: constructed (assembled from information streams according to programs that act on data), interpreted (rendered in human-computer interfaces and affordances that suggest how readers should understand data – stories of a kind) and made visible

(published on a variety of websites that rely upon a news organization‘s own site) by multiple actors all working around a shared set of data made accessible through the rules set by the news organization.

Some news APIs exist within news organizations (e.g., New York Times, The

Guardian, NPR and the now defunct BBC ―Backstage‖), designed to give public access to internal content; while others are more generic and used by many different news organizations to share content among themselves and with broader publics (e.g.,

Publish2, DayLife, Yahoo! Developer). And just in the past two years, large data sets that were ostensibly already public (but that were in file formats and bureaucratic processes difficult to access) have become more accessible and easier to use: e.g., the

Obama Administration‘s broad data.gov initiative, the UK government‘s release of the

―Coins database‖ on public expenditures, or ScraperWiki‘s project that crawls across

UK government websites to convert website information to standardized data, making

229 it accessible through an API. When public data has become more accessible, it has been easier for news organizations to do three things:

- conduct original, professional reporting on the data, uncovering stories that

were more difficult to tell with less accessible data. For example, consider

ProPublica‘s ―Dollars for Doctors‖ project that culled across and organized

large amounts of data that had been made available by drug companies to

investigate patterns among their payments to physicians [121];

- make standardized and verified data accessible to broader publics who

encouraged and given advice about how to become ―data journalists‖ [122] by

making data available through news APIs. For example, consider the

Guardian‘s ―Data Store‖ project (that obtained approximately 6500 data sets

from governmental, non-governmental and private sources, vetted each set,

organized them by topic and UN Millennium Development Goals, and

published them on its website);

- organize hybrid professional-amateur collaborations in which journalists work

with individuals not affiliated with news organizations to make sense of large

amounts of data that likely could not be analyzed by the news organizations

alone, or by individual citizens without the guidance and resources of a

professional news organization. For example, consider the Guardian‘s

―Investigate your MP‘s expenses‖ project in which it organized and provided

API data resources to approximately 20,000 volunteers for them to analyze

over 450,000 public expense reports, subsequently converting these analyses to

professionally prepared news stories [123].

230

Essentially, news APIs constitute a kind of meeting place where particular types of news are published that are constitutive of—and heavily dependent upon—the decisions of news organizations‘ journalists, programmers and editors (e.g., what data sources they choose to pursue, vet, organize and make available) and, simultaneously, the decisions of publics and readers who now have the potential not only to access or read news, but to work and program with the news (e.g., choosing to work with some parts of a news API over others, accessing certain vetted data over others, designing and building particular applications that use news APIs versus others). News APIs can thus offer a window into a distributed, news-focused community of practice

(Wenger, 1998) consisting of professionals reporters, organizational managers, software engineers, citizen journalists, readers all focused around a kind of news bricolage (Lévi-Strauss, 1966) or thoughtful tinkering with found news objects that is mediated by the norms and rules of both news and API programming.

A first step in understanding this tinkering—this distributed community, these overlapping identities, and how the norms and practices of news APIs might help structure contemporary press autonomy—is simply to describe what news API are and how they function. To do this, the next part of this chapter chronicles three leading news APIs: the New York Times‘ ―Developer Network‖, The Guardian‘s ―Open

Platform‖ and the NPR API.

3. How News APIs Work: A Descriptive Study of Three Leading APIs

The goal of this empirical section is to describe three leading news organizations‘

APIs. I began by building a data set of approximately 300 publicly available

231 documents [see the end of the dissertation for a list of ―Referenced Documents‖ with numbers corresponding to in-text citations], including:

- terms of service agreements for each API;

- entries on API blog sites that are maintained by each news organization;

- secondary blogs and news stories about each API;

- public presentations given by news organizations and independent developers,

including slides, videos and transcripts of public talks;

- code documentation for each API;

- user forums and support groups in which both developers and news

organization staff ask and answer questions about the APIs.

I was motivated by Star‘s directive that, when reading infrastructures, researchers should attempt to identify a system‘s ―master narrative,‖ the voice that ―does not problematize diversity‖ and that ―speaks unconsciously from the presumed center of things‖ (1999: 384). In this light, I noticed moments when the documents‘ language and word choices seemed to matter; e.g., when a ―we‖ was presumed, when a passive voice was used (―the data show‖), when articles were signed or unsigned, and when authors spoke on behalf of, or diverged from, an institutional voice (―We engineers built the API to …‖ versus ―At the New York Times, our API …‖). I did not conduct a content analysis that quantified such word choices or voice shifts but, rather, noted them as I read the documents, inductively revising and answering my empirical questions using a ―categorical aggregation‖ technique in which the goal is to ―seek a collection of [categorical] instances from data‖ (Creswell, 1998 : 154) in order to form larger, issue-relevant meanings that connect the data to the research framework

232

(Holsti, 1969, cited in Jones [1996: 128-129]). The questions that guided my reading of the documents and my descriptions of the APIs were:

- Within APIs, how is news content (stories, themes, people, events) coded?

Where are these codes defined, how are they updated, and who has power to

change them?

- What are an API‘s technical boundaries? For example, what other libraries or

coding languages does it rely on? What formats can queries be made in and

what data formats are returned? Are any other organizations or organizational

data streams explicitly included in the APIs and, if so, who are these

organizations?

- What distinctions, if any, does an API make with respect to genres of

journalism? For example, does the API code categories using traditional labels

like ―investigative reporting‖, ―opinion/editorial content‖ or ―business news‖

or are there new categories used to describe its content?

- Related to the idea of how objectivity may appear in the design and use of

news APIs, are distinctions made between descriptive reporting and

interpretive reporting, e.g., when are datasets visualized, when are they

interpreted, and when is there a combination of both? Do systems suggest the

desired or acceptable roles it envisions readers to playing? E.g., when do these

systems offer interfaces that allow readers to interact with datasets, and when

do they offer data interpretations with particular perspectives or arguments that

audiences are meant to receive but not change?

233