Recursive intertextuality and the impossible task of fact-checking Glenn Greenwald

How Shannon and Weaver’s Transmission Model of Communication can save us from information chaos.

Last week, Glenn Greenwald released an article and video, both called New Proof Emerges of the Biden Family Emails: a Definitive Account of the CIA/Media/BigTech Fraud with a subheading that reads “An axis of the CIA, Big Tech and the DNC-allied wing of the corporate media spread an absolute lie in the weeks before the 2020 election. We now have definitive proof.”

This new evidence promised in the headline is presented halfway through the article. He uses the first half to set the scene, laying out his own account of the game so far, so we all have sufficient context for his imminent dunk. 

This context includes a reference to the recent “destruction by Big Tech monopolies of Parler at the behest of Democratic politicians” – a claim which he supports by linking to his own article on the topic, in which he writes, “today, if you want to download, sign up for, or use Parler, you will be unable to do so.” But that is not true. For Android users like me at least, Parler is still available. It is on my phone, as the video below shows. It’s the 18th of 36 research highlights (18/36).

You have to go to their website, rather than to the Play Store, but you are able. Greenwald has overstated the power of the tech behemoths. Indeed, the difference between what I experience and what Apple users (who I think can only download apps onto their phones if they come from the App Store) experience is a demonstration of one important choice the market still offers: The increased security of the Apple closed garden approach vs the greater liberty offered by PC and Android’s more open architecture. This is rushed past, giving his readers an inaccurate picture of the strategic landscape they inhabit, inhibiting their ability to navigate it, rather than empowering them. If they took him at his word, they would not look for Parler, and therefore would not find it. 

Correction: Amazon Web Services discontinued providing hosting services to Parler, per an announcement on January 10, 2021. It was not until February 15, 2021 that SkySilk Cloud Services began hosting Parler. Parler posts made before January 10, 2021 were not accessible when the platform came back online. On January 12, 2021, when Greenwald published the article referenced above, Parler was not accessible, for any user. As of May 2021, the Apple App Store has allowed users to download the Parler App. A record of the correction research is accessible here:

Another section of this preamble-to-new-proof reads:

But this phrase on which Greenwald places so much importance, “Russian disinformation”, cannot be attributed to James Clapper or any of the other former intelligence officers. The only time they use the word “disinformation” in their letter (to which Greenwald links) is in a quote. Here is the paragraph where it appears:

These senior figures in the intelligence community are quoting USA Today, discussing an investigation by federal authorities. What they actually say, in their own voice, is that the “arrival on the political scene of [Hunter Biden’s] emails… has all the classic earmarks of a Russian information operation” (my italics).

One journalist who Greenwald does name, with intent to shame, is “the co-queen of Russiagate frauds, Natasha Bertrand”. But unlike Greenwald, she gets the quote right, at least in the body of her article, after using the abbreviation disinfo in her headline (without quotation marks).

It seems we can safely assume this was, on Greenwald’s part, an honest mistake, since elsewhere he stresses the differences between the intelligence agents’ letter and the reporting of it, particularly the media’s failure to highlight the retired officers’ admission that they have “no evidence”.  Had he caught it, he would have pointed out this discrepancy, too. 

Whatever rating we give Greenwald’s article overall, the video is worse. 

Near the start of the video, at 01:20, he says the documents are authentic and “raise serious questions” about Joe Biden’s behaviour. But by 2:40, his confidence increases and says “these documents, reported first by the New York Post on October 14, detailed how Joe Biden used and abused his power as vice president, with the responsibility for Ukraine in his portfolio, to benefit a company: Burisma energy”. 

This second statement, like the original headline of the New York Post article in question Smoking-gun email reveals how Hunter Biden introduced Ukrainian businessman to VP dad, and its first sentence “Hunter Biden introduced his father, then-Vice President Joe Biden, to a top executive at a Ukrainian energy firm…”, is an unproven allegation. The circumstantial evidence, like the 50,000 USD per month Hunter Biden received to sit on Burisma’s board despite having no obvious qualifications, is compelling. But it’s not conclusive. Hunter may be guilty mostly of conning his colleagues by peddling influence over his father that he did not really have.

None of this is changed by the new and supposedly “definitive” proof which has emerged; passages from a recent book about the Biden family. In it, the author Ben Schreckinger of Politico cites a number of anonymous sources who in two cases were recipients of emails included in the cache, who could verify the emails they received matched verbatim those published by Biden’s accusers. The third is someone who (according to Schreckinger) could confirm that Hunter Biden had received emails, “the substance” of which matched those published, but who could not verify they were exact, verbatim, copies.

This is definitely newsworthy. And further evidence of authenticity. It fits with my immediate impression from the time that the emails were almost definitely real because, as Greenwald and others have noted, the Biden campaign failed to immediately and forcefully denounce them as forgeries and fakes.

This lack of a forceful and comprehensive denial was, pretty much, the ball game as far as the question of the emails veracity. I said so at the time and lost friends over it. But all that proves is that this person, who isn’t a member of the Biden family (but is involved in business dealings with Hunter Biden, whose role in Burisma is clearly suspect), wrote to Hunter saying “thank you for inviting me to DC and giving an opportunity to meet your father and spent [sic] some time together. ”

Assuming the author of the email isn’t lying or crazy, we still cannot tell whether they are saying thank you for a meeting that has occurred, or has been (he thinks) promised. It’s very possible there was some verbal discussion, in which access to the VP was raised and this person is hoping to lock in that commitment (real or imagined). People send emails like this often after verbal exchanges, with the goal of confirming their recollection of the details.

If new reporting about how the allegedly corrupt Biden family insist on keeping mentions of the elder Biden’s name out of written exchanges are true, that would imply this particular would-be collaborator harmed his chances of an audience by breaking this taboo. It may be that precisely because the evidence exists, the crime did not take place. Or it could all be true and President Biden could be guilty. I don’t know. 

Neither does Glenn Greenwald. Going on the current state of public record, it is impossible to know and impossible to rule out. Greenwald is smart enough to navigate these degrees of ambiguity and uncertainty, and help his audience do the same, if it suits him to do so. But instead he adds to the confusion. 

I say this as someone who agrees with the general thrust of Greenwald’s take: The response by much of the news media, and social media companies, especially Twitter, who outright blocked the transmission of the NYPost link on their platform, even in private messages, was wrong. The story had real news value. The authenticity of the emails outweighed their misrepresentation in the NYPost headline and first line, which an intelligent person can easily spot for themselves. 

He is right in the tweet thread, posted after the article and video went out, that there should be acknowledgment of this further evidence of the emails’ authenticity by the outlets which cast doubt on or denied it at the time. 

But then he goes too far.

We can ask ourselves a different question though, too. If the book had mentioned three anonymous sources who supposedly had knowledge of the emails as saying they were totally fake and the Bidens were innocent, would that be “proof”? And would it be “definitive” enough for Greenwald to retract the article he quit The Intercept over when they wouldn’t publish it?

I hardly think so. It wouldn’t make sense to. On its own, the new information changes the overall picture very little. A set of anonymous sources quoted by an author wouldn’t be enough to contradict the documentary evidence, which the Bidens have all but admitted is real. So Greenwald’s dunk is actually more like a victory lap. This would be fine, but it feels like he needs it to be more, and is trying to convince others and himself that it is. I want something better for Greenwald than this endless, obviously ungratifying, tactical scramble.

He has a rare mind, and rarer courage. He is someone who has done important work. 

Whatever Greenwald’s faults, I am sure he is doing his earnest best in an information ecosystem (and therefore political climate) that brings out the worst in everyone.

Let’s leave aside questions of intent and motive for a moment, and think in terms of nodes in a network. Consider the following diagram:

If we simplify Greenwald’s claim to the two words “they lied”, we are presented with a minimum of three tracks of investigation to consider.  Who “they” are, what “they” said, and what really happened. If we take the system one step further, we might get the following arrangement:

We still don’t know who “they” are, and knowing that might be partially defined by who said what, since inclusion in the “axis” depends on having taken a particular editorial line. Leaving this aside, let us assume that there is a coherent “they” (it could be “an axis”), and what they said was “The Post lied”. What actually happened (best I can tell) is that The Post exaggerated. So, as this diagram shows, their accusation of dishonesty on The Post’s part is itself half true, as is Greenwald’s claim that they outright lied:

In the top right of this image, you will see we can subject my determination about The Post’s actual behaviour to the same three-pronged investigation, setting off a potentially infinite set of subsequent investigations into supporting evidence for supporting claims and so on. Each node you add to the network, potentially, affects values across the whole network. 


His article is full of similar claims; That the actions taken by big tech against Parler were taken at “the behest of Democratic politicians”, or:

There are “increasingly explicit threats from elected officials in the majority party of legal and regulatory reprisals in the event that tech platforms do not censor more in accordance with their demands.

Or that such a thing as “An axis of the CIA, Big Tech and the DNC-allied wing of the corporate media” exists, and that it engaged in a “joint campaign” .

Any one of these claims, referencing as they do statements and motives by others, which will refer to statements and motives of others, which will refer to statements of motives of others, could be the start point for a new fractal tangent map.

This is life in hypertext. There is a recursive intertextuality, which we have not yet mastered. It’s something that demands more from us than just accusations. The question becomes primal. 

Who do you trust?

But what if we remove some of the emotional charge and view the problem as an issue of quality – noise versus signal – rather than a question of intent (lies versus honesty)?

In the above diagram all the claims (except mine, of course) ended up getting a 50% truth score. But a network where each node has 50% fidelity does not have an overall fidelity of 50%. It gets much worse, very fast. 

Assuming we start at 100% accuracy, and transmit in a straight line from one node to the next, with 50% fidelity each time, we go from 100% signal – 0% noise, to 50% signal – 50% noise, to 25% signal – 75% noise, and so on, as half truths are told about half truths about half truths about half truths until homeopathic levels of dilution are achieved. 

But this problem is not new. A technical version of this problem has already been encountered and overcome – otherwise large scale networks like the Internet would not be possible. There is always the potential for noise to corrupt a transmission. Even small distortions, if allowed to build up cumulatively, would overwhelm any system (like a telephone network of ever growing complexity). This problem was addressed largely by the “transmission” model of communication developed by Claude Shannon and Warren Weaver. 

They started with a simple model of transmission comprising a source, a transmitter, and a receiver. Between the transmitter and the receiver, sometimes noise will enter the system. 

To counteract this, Shannon and Weaver recommended an observer receive a copy of the information on its way to the transmitter, and from the receiver after its arrival. The information can then be compared, before sending correction data to a correction device, added at the end of the chain.

We can imagine a human-level version of the same system. The “source” becomes the “source material” including documents, interviews, field work; the transmitter is the journalist; and the receiver is the news consumer.

This is where the product my team and I have built becomes important. Stone is the software I used to track my initial (abortive) attempt to document and check the claims in Greenwald’s latest piece (which I expected to overwhelmingly agree with – and still largely do, despite my objections to his sloppy language and inexact sourcing). That means those who question my motives or my methods can review them in detail. The highlights in the player above are short webcam commentaries I added during the research process. The rest of the work is captured in a silent timelapse, without my face in the way. If you scroll up and click “see research”, that’s what you’ll see. 

Our tool creates the possibility for news consumers to become a special kind of prosumer, a research scrutineer, who examines the journalist’s engagement with the source material and compares it with the news deliverable, sending correction data (uniquely informed feedback) to the correction device – a rating system within Stone (under development) or directly to their own blog, publication, or social media account. This feedback might be “Jane Smith from the ‘Journalism Times’ has really done her homework on this one” or it could be “at 12 minutes and 34 seconds in, you made a terrible mistake”. 

But those mistakes will happen less often. The reason I caught the erroneous quotation in Greenwald’s article, for example, is not that I am so much smarter than him, but that knowing my research is being logged for later inspection, I was more careful, checking every quote against the source material provided.

This fits with my position that the real problem in the media is not bias. It’s quality. Have an opinion, fine, but get your facts right, use precise, checkable language to express it.  

We are throwing down the gauntlet to Greenwald, and Bertrand, to the alternative and mainstream media, and conservative media like the NYPost, to raise their standards, which are, by default, abysmal.

Make yourself accountable to your readership, show them all the thorough, fair, and relevant research work you are doing to provide them with reliable news. Transparency is the sword that can slice through the Gordian knot of claim and counterclaim. Pick it up.

Or are you all – all of you – afraid of what they might see if the curtain falls away and reveals an ordinary man frantically pulling levers, triggering lights and smoke, rather than the Great and Powerful Wizard of Oz?

Or are you afraid of what you might see, if, like a prizefighter, you trained in front of a mirror to examine and perfect your own technique?

If so, then it will fall to the next generation to do it better. That generation is already here

Leave a Reply