Blogging Back then: Annotative journalism in I.F. Stone’s Weekly and Talking Points Memo

Introduction

Online journalism in general, and blogging in particular, have invited frequent comparisons to earlier eras of journalism: to ancien regime France, to pre-revolutionary pamphleteering, to the party press of the 19th century (e.g. Barlow, 2007; Darnton, 2010). This article proposes a very particular comparison, to a mid-century, muckraking newsletter, in order to highlight the material, textual practices of blogging as a form of newswork – one that defies customary distinctions between factual reporting and opinion or commentary. I call this annotative journalism, defined simply as journalism that proceeds mainly through the critical analysis of published texts, where those may be news accounts, official documents, or other material, publicly available texts. ¹

Annotative journalism is not simply investigative reporting, but rather a style of newswork that unearths new facts by publicly dissecting and comparing news accounts and other evidence through the lens of a wider media and political critique. This larger critical-political framework obviates the narrative coherence of conventional American news, so central to classic studies of news and newswork (e.g. Lule, 2001; Tuchman, 1978; Zelizer, 1992). The result is an ideological journalism which is nevertheless more transparent in important ways than its objective counterpart. It reveals reporting routines to the reader and builds explicit arguments both from and about the work of other journalists, making visible the intertextual nature of news production. The annotative journalism developed by a network of blogger–journalists over the last decade finds a remarkable precedent in the work of I.F. Stone, and especially in I.F. Stone’s Weekly, the newsletter he produced from January 1953 through the end of 1971.

The muckraking tradition has often been seen as an antecedent of blogging (e.g. Barlow, 2007; Cohen, 2008; Rosenberg, 2009). The obvious parallel is that like the best-known blogger–journalists, Progressive-Era muckrakers did not strive to be dispassionate presenters of fact. Consider Lincoln Steffens’ (1904) frank disavowal of journalistic detachment, offered in the introduction to a collection of his essays on urban blight and corruption. ‘This is all very unscientific, but then, I am not a scientist. I am a journalist’, Steffens (1904) wrote, continuing,

I did not gather with indifference all the facts and arrange them patiently for permanent preservation and laboratory analysis. I did not want to preserve, I wanted to destroy the facts. My purpose was … to see if the shameful facts, spread out in all their shame, would not burn through our civic shamelessness and set fire to American pride. That was the journalism of it. I wanted to move and to convince. (p. 12)

But beyond that political parallel lies an important material one: the distinctive set of textual practices invited by a journalism that seeks ‘to move and convince’. Informed by the Progressive faith in science and reason, turn-of-the-century muckrakers practiced an evidence-driven, argument-building brand of reporting (Swados, 1962; Weinberg, 2001). They assembled facts and gathered documents on an unprecedented scale from diverse textual sources, including news accounts – data that gained coherence in the broader project of progressive political and economic reform. ² As Guttenplan (2009) writes, ‘it was the Populist critique of the economy that gave the facts so painstakingly assembled by the muckrakers their significance’ (p. 53). Stone learned his craft, in the 1930s, in newspapers which were the inheritors of that tradition. In his own weekly newsletter, free from the format and genre constraints of newspaper journalism, Stone married the muckrakers’ document-driven methods to an annotative style of presentation. What the following analysis of the Weekly will show is how a critical politics affords uncommon latitude in working with texts. Stone’s approach yielded a precursor of the transparent, fragmentary, and relentlessly intertextual style of news which has come into flower on the Internet.

Method and literature

This article illustrates annotative journalism through close study of two cases separated by a half-century. Both stand out in hindsight as major investigative coups, though neither relied on the tools we associate with that kind of journalism: anonymous sources, secret documents, ‘shoe-leather’ reporting, and so on. Instead, each investigation drew mainly on published reports in other news outlets. The first case is Stone’s exposure of a government conspiracy to misrepresent the results of a 1957 nuclear test in order to discredit opponents of such testing. (Another story Stone worked on, the 1964 Gulf of Tonkin incidents, will be reviewed briefly.) The second major case is the award-winning work done by the blog Talking Points Memo (TPM) to uncover a ‘purge’ of federal prosecutors at the Justice Department in 2007.

These two cases stand out, it has to be acknowledged. The news analysis and media criticism on news-related blogs do not usually lead to major investigative breakthroughs. Similarly, I.F. Stone’s newsletter (modeled on George Seldes’ In Fact of the 1940s, ‘an Antidote for Falsehood in the Daily Press’) mostly reacted to headlines and rarely broke news. At the same time, some of the reporting work described below would be familiar to any investigative journalist. This is why annotative journalism, in its complex, ecological relationship to traditional news, deserves closer attention. These cases highlight annotative techniques which have a long history outside of the journalistic mainstream and which are increasingly basic to the vocabulary of media production today – on blogs, in professional newsrooms, and across a wider news landscape that includes venues from Fox News to The Daily Show and Democracy Now!.

This raises an important point: Not all blogs are the same. The boundaries of ‘blogging’, always contested, have only become murkier as traditional news organizations embrace the genre and as prominent blogs have professionalized and been incorporated into elite media-political networks. (This is one good reason for scholars to focus instead on categories of practice, such as annotative journalism.) This study takes TPM as emblematic of that class of news-and-politics blogs which has most interested journalism studies, for their practice as well as their critique of journalism. The site has been involved in several of blogging’s signature ‘scoops’ over the last decade and figures prominently in academic literature.

The defining feature of this class of blogs for scholars has been their oppositional stance toward traditional news outlets, yielding a news discourse organized ‘around the idea of challenging mainstream journalism’ (Matheson, 2004: 452; Park, 2009). Many studies read this antagonism as an ideological challenge to journalism’s always-tenuous professional project (Lewis, 2012; Schudson and Anderson, 2009). In the same vein, scholars have seen traditional news outlets responding with efforts to ‘domesticate’ the new medium in a way that preserves professional norms and status (Domingo et al., 2008; Robinson, 2006; Singer, 2005).

Blogging’s critique of journalism has invited an ecological approach to understanding its role in a wider news environment. Any number of studies point out that blogs engage in little primary reporting and instead rely on the work of other (especially elite) news outlets (e.g. Reese et al., 2007; Reich, 2008). Blogs have been assigned a downstream role in established models of political communication, engaging in ‘second-level agenda-setting’ which tells readers how to interpret the issues which journalists report on (Meraz, 2009; Murley and Roberts, 2006). They may operate not as gatekeepers but instead as ‘gatewatchers’ who publicize and prioritize stories from the body of news produced every day (Bruns, 2005: 17–23). At the same time, top bloggers can set the wider news agenda because ‘opinion-makers within the media’ take them seriously (Drezner and Farrell, 2008: 29–30). Anderson (2010) complicates the distinction between reporters, bloggers, and activists with his notion of ‘fact entrepreneurs’ who promote and shape a developing story in a local news ecosystem.

Annotative journalism as conceived here fits squarely within this broader ecological understanding of newswork. In an early report on ‘participatory’ media, Bowman and Willis (2003) identified ‘annotative reporting’ that supplements a news account with a ‘point of view, angle or piece of information … missing from coverage in the mainstream media’ (pp. 34–35). However, annotation can do more than ‘supplement’ the news; as practiced over time, in the context of an unfolding story, annotative techniques may yield vital new information even in the absence of original reporting. As I argue elsewhere, news-related blogs have advanced major stories through a kind of ‘distributed news analysis’: Collectively they can act ‘as an engine for distilling and dissecting news accounts, testing them against one another and against established facts to solidify … the real, factual context for future news accounts’ (Graves, 2007).

This study seeks to draw our attention to the material, textual practice of annotation as a kind of newswork. An annotation acts upon another text. It opens up a critical distance between two or more texts and constructs meaning by exploiting that distance. This textual awareness affords a particular economy of communication, visible in various annotative genres which have emerged historically. Thus, for instance, in medieval florilegia – collections of excerpts meant to serve as guides to a work, an author, or a topic – the mere fact of a passage’s selection, or its position within an anthology, conveys valuable information. Moss (1996) writes that florilegia ‘compose a signifying universe which is wholly literary, in which texts illuminate texts in a self-sufficient environment where dialectical inference and extra textual reference are only minimally necessary’ (p. 106). In the same way, annotation may structure an argument from news texts with the simple juxtaposition of clashing accounts or by adding an ironic headline to a copied passage – a favorite tactic of bloggers (and of programs like The Daily Show) also applied in Stone’s Weekly.

Annotation runs counter to the narrative logic of conventional reporting. Emile Benveniste writes that ‘the objectivity of narrative is defined by the absence of all reference to the narrator … Here no one speaks. The events seem to tell themselves’ (quoted in White, 1980: 7). This sense of narrativity underlies Tuchman’s (1978) notion of the self-validating ‘webs of facticity’ produced in objective news accounts. In contrast, annotative journalism is critical, evaluating and assessing news texts within a larger normative framework. It thus challenges traditional journalism’s constantly reinforced distinction between original reporting and opinion or comment. The cases reviewed below will illustrate how annotative reporting, by juxtaposing and dissecting news texts, may unearth new facts, drive a story forward, and even yield decisive scoops. Both cases highlight an intertextuality that is deeply unsettling to objective news accounts in which the facts seem to speak for themselves.

I.F. Stone and the nuclear testing controversy

Two episodes from Stone’s career will illustrate how the Weekly married document-driven muckraking to the newsletter format in a way that anticipated the annotative journalism practiced online today. The first, which Stone would call ‘the biggest scoop I ever got’ and the best illustration of his reporting style, was the Weekly’s unmasking of an official campaign to discredit the nuclear test-ban movement – and to forestall a resumption of US–Soviet talks over a ban – with misleading seismic data. (Accounts of this episode are in Alterman, 1988; Bruck, 1973; Guttenplan, 2009; Patner, 1988.) Opponents of banning nuclear tests, led by Dr Edward Teller, had argued that the Soviets would be able to cheat by testing weapons in secret, underground. On 6 March 1958, the Atomic Energy Commission (AEC) gave ammunition to those arguments by publishing the results of the first US underground nuclear test, carried out in Nevada the previous September. The AEC report claimed seismologists had not been able to detect the blast beyond a radius of about 250 miles – far less than the roughly 600 miles between ‘seismic listening posts’ that Moscow had tentatively agreed to. In a matter of days, however, Stone’s reporting would force the AEC to retract its claim.

The most remarkable feature of this reporting coup is that Stone accomplished it using only public documents and news reports. National reporters had been invited to Nevada to cover the underground test in September 1957. The New York Times reported from Las Vegas that the test ‘seemed to have conformed with predictions of A. E. C. scientists that the explosion would not be detectable more than a few hundred miles away’ (Hill, 1957). As Stone would later explain, however, his edition of the Times included a tiny ‘shirt-tail’ reporting a claim that the blast had been detected in Toronto; later editions had similar bulletins from Rome and Tokyo (see Figure 1). The Times did not acknowledge the contradiction or follow up in the months to come. But it caught Stone’s eye. Lacking even the resources to cable those cities for more information, Stone filed the clippings away (Guttenplan, 2009: 337, 442–445).

OPEN IN VIEWER

Figure 1.

I.F. Stone clipped ‘shirt-tails’ from the Times indicating the 1957 US nuclear test was detected far more widely than officials claimed.

When the AEC report on the underground test finally came out 6 months later, national reporters enjoyed a tour of the blast site and reported on the peacetime nuclear applications to be yielded by this safe new testing regime (e.g. Herbert, 1958). Stone, however, saw the report as an obvious effort to bolster the case against disarmament. The cover story of the 10 March Weekly took aim at Dr Teller’s ‘hint-and-run’ campaign against nuclear disarmament (Stone, 1958a). To undercut pro-nuclear arguments, Stone relied on an annotative technique he used often in the Weekly: quote boxes, freestanding textual excerpts offered with no editorial comment beyond a provocative title or a jarring juxtaposition. (The first issues of the Weekly, from 1953, featured a straightforward layout with no text boxes or lengthy excerpts. By the late 1960s, the newsletter included at least one quote box, and usually two, on every page.)

Thus, a quote box on the front page, titled ‘Dr. Teller’s Point of View’, united the scientist’s claim that because of cheating ‘disarmament is a lost cause’, from a recent appearance on NBC’s Meet the Press, to a more disturbing argument he had advanced in a 1957 magazine article: ‘We must overcome the popular notion that nuclear weapons are more immoral than conventional weapons’. The subtext was clear: However pragmatic he sounds now, Teller’s opposition to a test ban is ideological. A quote box on the last page asked, ‘Is Fallout as Negligible as Dr. Teller Says?’ over a quote from the scientist likening nuclear fallout to smoking a cigarette every 2 months – followed by a contradictory passage from the Bulletin of the Atomic Scientists (see Figure 2). Fully half of the four-page newsletter was excerpted from the Congressional Record; Stone carefully selected passages of recent testimony from a prominent test-ban advocate, using boxed quotes and boldface subheads to guide the reader and reinforce the argument. Stone’s own words fill just a third of the issue, though his voice is everywhere.

OPEN IN VIEWER

Figure 2.

The Weekly used boxed quotes with ironic headlines or jarring juxtapositions to make editorial points.

That 10 March Weekly also noted the New York Times ‘shirt-tails’, from September, that seemed to contradict the new AEC report, though Stone still had not confirmed the bulletins. As the issue went to press, he lodged a request for information with the Coast and Geodetic Survey in the Commerce Department, which, as he later reported, ‘seemed to be unaware of the AEC release’ (Stone, 1958b). Within days, the government scientists gave Stone a list of 19 seismic stations across the United States and Canada that had recorded the nuclear test. He confronted the AEC with the new information. By the time the 17 March Weekly came out, the AEC had issued a ‘note to editors and correspondents’ amending its earlier report to say ‘earth waves’ from the blast had in fact been detected more than 2000 miles away. Stone’s (1958b) cover story provided a blow-by-blow account of his own reporting under the headline, ‘Why the AEC Retracted that Falsehood on Nuclear Testing’.

Stone thus produced a remarkable piece of investigative journalism about a vital area of national security policy without any inside or anonymous sources. However, it is crucial to understand that Stone supported nuclear disarmament and had often used his newsletter to argue for the cause. In July 1957, Stone dedicated an entire issue of the Weekly to reprinting official statements from a meeting of top scientists convened by Bertrand Russell to discuss the nuclear threat. Stone inserted provocative subheads – ‘War Would Leave No Country Unscathed’ – to organize the scientists’ message. (He also faithfully highlighted conclusions that complicated the anti-nuclear argument, for example, ‘Medical X-Ray Worse Than Fallout’.) ‘The Scientists Warn Mankind’, the cover bellowed, followed by a brief editor’s note that took direct aim at establishment journalism:

As we went to press on July 18 not a single newspaper in the United States had been sufficiently interested to publish the text of the warning issued by twenty world famous scientists of the Soviet and Western blocks after a historic meeting in the little Nova Scotian fishing village of Pugwash the week before. (Stone, 1957a)

Stone was also already suspicious of the AEC and its ‘Madison Avenue Techniques’, in the words of an October 1957 headline in the Weekly. That article used Vance Packard’s The Hidden Persuaders, then newly released, to frame an analysis of government efforts to downplay nuclear risks – most dramatically in choosing the name ‘Project Sunshine’ for a study of nuclear fallout. ‘It is as if from the very start the intent was to make us assume that the radioactivity let loose by nuclear testing was something like sunshine and natural radiation’, Stone (1957b) wrote. The story went on to dissect Senate questioning of a top AEC scientist, annotating the exchanges with notes on the motives of the lawmakers. The result reads like a pre-digital blog post, with boldface doing the work of indentation to move between textual registers (see Figure 3).

OPEN IN VIEWER

Figure 3.

Stone often annotated published texts. Here, his editorial remarks guide readers through a congressional hearing about ‘clean’ nuclear weapons.

In this way, Stone’s opposition to nuclear testing and his critical view of the news media informed the annotative, document-driven approach that yielded his big scoop. He filed away the suspicious ‘shirt-tails’ because he put little faith in either the AEC or the New York Times. Stone’s progressivism provided the analytical context in which a discrepancy between documents becomes a ‘story’ worth reporting. This approach demands that documents be treated explicitly as such – as texts to be labeled and analyzed. Throughout these issues of the Weekly, Stone referred directly to news articles, government reports, congressional transcripts, and even press releases, rather than weaving their contents into a narrative whose material seams and sources are hidden from the reader.

The coherence of this annotative approach assumed as much about the Weekly’s readers as about its author. A shared critique of the political and media establishment permitted reporting and presentation to be fragmentary and unfinished. Stone did not have to wait until he had the whole story. He could present inconclusive but suggestive bits of evidence – like his first mention of the ‘shirt-tails’, or in the same issue, the point that an above-ground Soviet test had been detected and thoroughly analyzed by the United States in a matter of hours (Stone, 1958a) – because his readers shared an interpretive framework and understood where the story might lead. Likewise, a few sentences from either friend or enemy could stand alone in a quote box, with only an ironic headline to add context, because annotator and reader would see it the same way.

It is worth briefly reviewing a case that shows the same annotative techniques at work in a much larger controversy. Stone played a decisive role in uncovering the truth behind the Gulf of Tonkin incidents of 2 and 4 August 1964, and more generally behind the escalation of US involvement in Vietnam. His reports quickly cast doubt on the circumstances of the first Tonkin incident, suggesting that it may have been provoked, and on whether the second incident had occurred at all. In this case, however, no official retraction was forthcoming; although Stone pursued the issue doggedly in the Weekly and elsewhere, his suspicions would not be confirmed until the release of the Pentagon Papers in 1971. (In the 2003 documentary ‘The Fog of War’, Robert McNamara admitted that the second incident never took place.) Still, Stone had the contours of the story within weeks of the incidents – again, relying entirely on public documents and news reports. As one historian has observed,

It was one of the most remarkable accomplishments in history of investigative journalism, I think you could say, given his physical condition. He was practically stone deaf at this point in his life, so he couldn’t go to cocktail parties, he couldn’t chat it up with inside dopesters. He could only look in the public record. But, at the same time, he did have a larger critique in mind, and that is that the Vietnam War was sparked by anti-colonial nationalism and not by Moscow. (Jackson Lears interviewed by On the Media, 2009)

As in the nuclear testing stories, that larger critique made Stone suspicious of official accounts and sensitive to contradictory data. One source of such data was the overseas press. In the 10 August Weekly, which went to press just 2 days after the first Tonkin incident, Stone (1964a) cited North Vietnamese radio reports suggesting it was a response to shelling by US and South Vietnamese warships. The 24 August issue featured a boxed quote from Le Monde, reporting on the secret history of US operations against North Vietnam (Stone, 1964b). The front page of the 28 September Weekly excerpted a skeptical analysis of the Tonkin incidents from the Peking Review, an organ of the Chinese Communist party (Stone, 1964c).

However, Stone’s main source of information was again the elite US press. The first issue after the incidents cited three New York Times reports and one from the Washington Post to make a point the papers themselves had not: that the Pentagon had been ‘carrying on war behind our backs’ (Stone, 1964a). As in the test-ban controversy, Stone’s annotative analysis was framed by a deep skepticism about the elite news media and its cozy relationship with official Washington. The 24 August edition opened by flatly asserting that the ‘American government and the American press have kept the full truth about the Tonkin Bay incidents from the American public’ (Stone, 1964b). Building his case from congressional testimony and various press accounts, he concluded,

The process of brain-washing the public starts with off-the-record briefings for newspapermen in which all sorts of far-fetched theories are suggested to explain why the tiny North Vietnamese navy would be mad enough to venture an attack on the Seventh fleet, one of the world’s most powerful. Everything is discussed except the possibility that the attack might have been provoked. (Stone, 1964b)

The piece cast doubt on whether the second confrontation had occurred at all:

It is strange that though we claim three boats sunk, we picked up no flotsam and jetsam as proof from the wreckage. Nor have any pictures been provided. Whatever the true story, the second incident seems to have triggered a long planned attack of our own.

Once again, the Weekly’s annotative layout reinforced its critical stance. ‘Prize Explanation’, announced one quote box, over Sen. McCarthy speculating on CBS’s Face the Nation that the North Vietnamese attacked the US Navy because ‘they were bored’ (Stone, 1964b). In the 28 September Weekly, Stone contrasted the language in a New York Times report with a more skeptical Associated Press (AP) account, calling the former an example of ‘phony news stories’ that advance the cause of war. He took particular exception to the use of anonymous sources to advance the White House line (Stone, 1964c).

Several years later, in a series of articles in the New York Review of Books (NYRB), Stone would build a decisive case that the United States had begun a major escalation in Vietnam well before the Tonkin incidents. He could not take the same liberties of style and format in the NYRB, but Stone’s reporting in these articles reads like forensic document analysis, parsing phrases and comparing dates to find the gaps in official accounts (e.g. Stone, 1968). Once again, news stories, press releases, and congressional testimony were not just sources of information but texts to be deconstructed. Any good reporter reads between the lines. But Stone’s politics allowed, and his annotative methods demanded, that he make those readings explicit for readers.

TPM and the ‘running massacre’ of federal prosecutors

Perhaps the best illustration of annotative journalism online is the work of the news blog TPM, launched in 2000 by Joshua Marshall. TPM has been compared in its idiosyncrasy to I.F. Stone’s Weekly, and the similarities are unmistakable. Marshall’s handful of major reporting successes have all come by focusing on stories or on angles which, as his posts often point out, are being neglected in mainstream coverage. When one of these stories does break into national headlines, due in part to TPM’s persistent focus, the blog appears in hindsight to have been ahead of traditional news outlets. (The Trent Lott affair of 2002 is a much-studied example; see Glenn, 2007; Rosenberg, 2009; Scott, 2004.) And, like Stone, Marshall has been open about his progressive political views as well as his criticism of conventional journalism. TPM found its voice as part of a network of progressive blogs focusing on scandal and corruption in the White House of George W. Bush.

More important here, TPM also bears a strong material resemblance to the Weekly. Like other news-related blogs, TPM excerpts heavily from published sources. Quotes from public figures are almost always borrowed from a newspaper or broadcast outlet. Marshall has explained that he tries not to draw a bright line between original reporting and this kind of news ‘aggregation’. The bloggers at TPM also develop their evidence over a period of weeks or even months, in full view of (and with assistance from) their readers, rather than amassing it for a single, airtight exposé. ‘We have kind of broken free of the model of discrete articles that have a beginning and end’, Marshall has said. ‘Instead there are an ongoing series of dispatches’ (Cohen, 2008). The effect has been likened to reading an investigative reporter’s private notebook (Apple, 2007).

Journalistic encomiums to TPM have recognized its challenge to the traditional division of labor in newswork, in mixing ‘liberal opinion with original reporting’ (Apple, 2007) and in ‘synthesizing the work of other news outlets with original reporting and tips from a highly connected readership’ (Cohen, 2008). What has been difficult to recognize is that the site’s reporting successes come not in spite but because of a wider political critique that frames its annotative journalism. This was clearly illustrated in TPM’s most celebrated work, covering a scandal that enveloped the Justice Department in 2007. That work earned a Polk Award for Marshall and his staff, the first time a blog had received the honor. The award citation noted that TPM had ‘connected the dots and found a pattern of federal prosecutors being fired for failing to do the Bush administration’s bidding’, and that its ‘tenacious investigative reporting sparked interest by the traditional news media and led to the resignation of Attorney General Albert Gonzales’ (cited in Cohen, 2008; see also McDermott, 2007; McLeary, 2007).

TPM’s bloggers ‘connected the dots’, as usual, by reading the news – in this case, local press accounts of federal prosecutors being dismissed in early 2007. ‘What’s the White House Doing to Prosecutors?’ asked the headline of the first post to establish a suspicious pattern, in mid-January, linking to local reports of seven firings in six states (Rood, 2007). Soon TPM was offering line-by-line analysis of reported speech, official statements, legal documents, and other texts that emerged in the widening scandal. Typical of TPM’s reporting was its campaign, in early March, to identify a pair of Republican lawmakers who had pressured a federal prosecutor in New Mexico to announce indictments against state Democrats. TPM meticulously compared evidence from multiple news outlets – one brief post cites McClatchy, radio station KQRE, National Public Radio (NPR), the AP, the Washington Post, and the Seattle Times – in order to highlight discrepancies and assemble the most comprehensive picture of the incident. On the basis of these comparisons, the site could identify errors in news reports, endorse those that were on the right track, and establish what would be at stake in upcoming congressional hearings (e.g. Kiel, 2007a, 2007b).

TPM also maintained a master timeline of the affair which indexed all of the relevant documents and events, reaching back to the start of the Bush Administration (TPM, 2007). This remarkable document grew to more than 11 pages spanning hundreds of bulleted entries, linking to scores of news items and other sources. Called simply the ‘TPM Canned US Attorney Scandal Timeline’, it offered a guide to journalists and others interested in the story. But it also acted as an indictment – an annotated argument that the events deserved the label TPM had given them.

Relying on news accounts as evidence demanded critique of those accounts. It meant assessing them for accuracy and completeness and also reading between the lines to determine who their sources were and what interests they served. On TPM (as on other news-related blogs), criticism of the mainstream media (‘MSM’) offers a ready framing device for analysis that runs counter to conventional news narratives. An irony-laden post by Marshall (2007) in March 2007 opened this way alongside an excerpt from the Washington Post and a headshot of its reporter:

So there you have it: the White House’s side of the canned US attorney story provided by the Post’s John Solomon. … It turns out the whole thing is just one of those unfortunate misunderstandings the Bush White House now and again finds itself in.

More than a scathing review, though, the critique offered the clearest formulation Marshall had yet given of why the ‘canned US attorney story’ mattered – a scaffold for his argument that the firings amounted to a political purge. He analyzed anonymous leaks to the Post for what they revealed about the Administration’s strategy for handling the crisis: ‘when a White House tries to get out ahead of a story like this it’s key to note the admissions of salient facts that come along with the larger bamboozlement’ (Marshall, 2007). Like Stone, Marshall offered readers an account of what insiders, including reporters, were saying or thinking in private. As he had explained several years earlier, ‘If all the journalists in Washington kind of know something and no-one’s talking about it, I’m enough of a populist to think more people should know that, let’s get it out there’ (Marshall, 2003).

The scandal came to a head in March of 2007, as TPM’s persistent attention drew other reporters to the story. Congressional hearings and a trove of Justice Department emails added fuel to the fire. Several glowing profiles that month focused on the site’s role in driving coverage of the affair. In an interview, Marshall resorted to ecological language to explain his site’s impact: ‘This is sort of the nature of our role in the journalistic ecosystem … Once a story catches fire, the big players are going to start getting the big scoops’ (quoted in McLeary, 2007; see also Niles, 2007).

But that understates the resistance TPM’s narrative encountered at first from its better-established peers. Two months earlier, when the site was all but alone in covering the ‘running massacre’ of federal prosecutors, Time magazine’s Washington bureau chief took issue with that framing. ‘It’s all very suspicious-sounding’, Jay Carney wrote on Swampland, the magazine’s political blog. ‘Of course! It all makes perfect conspiratorial sense! Except for one thing: in this case some liberals are seeing broad partisan conspiracies where none likely exist’ (Carney, 2007b). Later the reporter reversed course, declaring that Marshall ‘and everyone else out there whose instincts told them there was something deeply wrong and even sinister about the firings’ had been right. He explained why he had believed there was less to the affair than met the eye:

When this story first surfaced, I thought the Bush White House and Justice Department were guilty of poorly executed acts of crass political patronage. I called some Democrats on the Hill; they were ‘concerned’, but this was not a priority. The blogosphere was the engine on this story, pulling the Hill and the MSM along. As the document dump proves, what happened was much worse than I’d first thought. I was wrong. (Carney, 2007c)

Critics read the fact that Carney’s reporting amounted to calling ‘some Democrats on the Hill’ as further proof of mainstream journalism’s subservience to political interests. Just as illuminating, however, is the un-self-conscious way the reporter, describing his own journalism, applied the evidentiary language of ‘facts’ and ‘proof’ to the prosaic reality of sourcing a piece of political news. Carney had objected earlier that TPM’s analysis ‘was purely speculative. Suspicions aren’t facts’ (Carney, 2007a). But he didn’t say what kind of facts his own calls to Capitol Hill might have turned up – a Democrat willing to supply evidence that the affair was a conspiracy, or one willing to supply a quote calling it a conspiracy? Everyday news practices can elide the distinction between facts in a statement and the fact of a statement.

Similarly, Carney didn’t specify what evidence in the ‘document dump’ now made it objectively factual to speak of a broad conspiracy. What is clear is that by mid-March, it was becoming uncontroversial to use the language of scandal and conspiracy. This was due not only to new evidence, but also to changing political circumstances: the fact of growing media attention, the fact of the congressional hearings, the fact that Justice Department officials began to resign, the fact that President Bush distanced himself from Attorney General Gonzales, and the fact that political leaders, including some Republicans, were calling for Gonzales to resign.

In one sense, then, Marshall won a prestigious reporting award for the triumph of an argument about how the affair should be read, an argument subsequently borne out by events which TPM itself helped to set in motion. The site’s reporting through the scandal consisted mainly in gathering public texts and arraying them in damning fashion; even today, the clearest evidence of a ‘conspiracy’ or ‘purge’ at the Justice Department remains the simple pattern of sudden firings TPM identified at the outset. Of course, that a consensus would emerge so quickly around TPM’s version of events was not inevitable. I.F. Stone, moved by similar distrust of the White House, had sketched the outlines of a grave conspiracy to lead the nation into war within weeks of the Gulf of Tonkin incidents. But it would be decades before his account was completely ratified.

Discussion: Objectivity and intertextuality

The decades of the Weekly’s publication coincide with what Hallin has called the ‘high modern’ period in American journalism. Dominated by the Cold War political consensus, this was an era ‘when the historically troubled role of the journalist seemed fully rationalized, when it seemed possible for the journalist to be powerful and prosperous and at the same time independent, disinterested, public-spirited, and trusted and beloved by everyone’ (Hallin, 1992: 16). The doctrine of journalistic objectivity which took shape after World War I was, by the 1950s and 1960s, deeply entrenched (Schudson, 1978). Stone’s politics made the Weekly an outlier, a bridge of sorts between Progressive-Era muckrakers and the advocacy journalism and investigative reporting that would thrive again in the 1970s. The Weekly was part of the long parallel tradition we now call ‘alternative’ journalism, so named precisely for rejecting the profession’s reigning orthodoxy (Atton, 2002; Schudson and Anderson, 2009).

But to focus on advocacy or partisanship is to miss half of the story. A faithful mid-century reader of the New York Times or the Washington Post who came across the Weekly would have been struck by any number of jarring textual practices, practices common on news-related blogs today. Newspapers occasionally excerpt official documents or other print sources; the Weekly did this every issue and often at great length. (In this respect, Stone’s newsletter might have been more familiar to readers in previous centuries.) ³ Professional reporters almost never cite one another’s work; the Weekly was filled with direct references to other news outlets. News reports often fail to specify the documentary source of a claim or a quote, especially when that is a press release; Stone invariably gave his readers this material context. Most striking is what’s missing from the Weekly: human sources. Stone rarely quoted from personal interviews and did not rely on anonymous sources.

The muckraking newsletter and the pioneering blog both exhibit a promiscuous intertextuality that objective reporting generally abhors. Hallin (1992) writes of the ‘wholeness and seamlessness’ that characterized the high-modern self-understanding of professional journalists (p. 14). The same adjectives apply to stories produced under the objectivity norm, which guides reporters to obscure not only their politics but also their reporting practices and their position in a political economy of information. The textual isolation of conventional news is so common that its strangeness eludes notice: the fact that news organizations, so interdependent in their daily work (Reinemann, 2004; Reese and Danielian, 1989) and so overlapping in the texts they produce, are nevertheless so reluctant to acknowledge one another within those texts. Some of the frankest statements of the routine copying and cue-taking in the news business have come in court. For instance, an amicus brief by Google Inc. and Twitter, Inc. (2010) declares that ‘for decades, television and radio news stations have broadcast information obtained from newspapers. And newspapers and Internet news organizations learn and write about events originally reported on television’ (or see the 1930s ‘press-radio war’; Jackaway, 1994).

It is important not to overstate the case. Newspapers do sometimes carry lengthy excerpts from reports or speeches, in extraordinary instances – the Pentagon Papers or the WikiLeaks diplomatic cables – as well as more routine ones. Zelizer (1995) has remarked on the latitude reporters enjoy in choosing between ‘text’ and ‘talk’. But textual excerpts are vastly outnumbered by reported speech in conventional news accounts, and they are often deployed in the same fashion as reported speech: to buttress the narrative of the story (e.g. the key findings of a report) rather than to sustain a critical analysis. Journalists increase their own authority, Zelizer argues, by using quotes to emphasize proximity to events or to powerful individuals and ‘to make claims without the accompanying responsibility’ (1995: 35). Elsewhere, she has written that ‘reporters use quoting practices to create the illusion of a whole’, in a way meant not to clarify discourse but to ‘blur its spatial and temporal parameters’ (Zelizer, 1989: 372–373).

In other words, traditional journalistic quoting practices are narratival rather than critical. The use of irony offers a revealing lens. In the cases above Stone and Marshall both employed irony liberally in critiquing textual excerpts or arraying them against each other. Conventional reporting by contrast adopts the ironic voice warily and only when moral circumstances permit – for instance, Ettema and Glasser argue, to subvert the claims of an official who has been revealed to be corrupt. In such cases, an ironic juxtaposition ‘transfigures the conventions of journalistic objectivity so that the very textual devices intended to assure the differentiation of fact and value become the means to express their fundamental unity’ (Ettema and Glasser, 1994: 5). The investigative reports that permit such ‘condemnation’ cut against the grain of conventional objectivity in the news, invoking an evidentiary rather than detached and neutral basis for the reporter’s authority.

More than one factor accounts for the aversion to intertextuality in traditional journalism. One durable explanation is competitive pressure, both professional and commercial, that makes reporters reluctant to credit other news outlets or to send audiences their way. But the counterexample of annotative journalism underscores how intertextuality also violates objective reporting’s standard of internal completeness and coherence (touched on in Gans, 2004 [1979]: 162, 172). It reminds us that other versions of the story exist, and in this way, it draws attention to the behind-the-scenes work of story construction. ⁴ Intertextuality invites scrutiny of the choices different reporters make.

A large tradition in journalism research has focused on techniques of story construction designed to efface the reporter’s role and make it seem, in Tuchman’s (1972) phrase, as if ‘the facts speak for themselves’. The reporting studied here also seeks to make the facts speak, but in a way that does not obscure the journalist’s role in giving them voice. Annotative journalism lacks the seamless narrative coherence of news reports carefully grounded in a place and time – the ‘dateline’ – but not in the web of documents and sources from which they are built. Its intertextual style of newswork breaks down the cardinal distinction between original reporting and opinion, yielding news reports grounded in a political and media critique. In this way, it highlights the interlinked material and ideological dimensions of journalism – the relationship between political commitments (including the commitment to ‘neutrality’) and reporting strategies, textual practice, and the affordances of a communications medium.