Despite having decisively won the presidential election by the only measure that counts, the Electoral College, Donald Trump recently decided to call the legitimacy of the entire process into question. “In addition to winning the Electoral College in a landslide, I won the popular vote if you deduct the millions of people who voted illegally,” Trump tweeted.

There was instant widespread condemnation of Trump. The New York Times ran a headline declaring that Trump’s claim had “no evidence.” ABC News declared it “baseless,” NPR went with “unfounded.” Politico called it a “fringe conspiracy theory.” Those news outlets whose headlines about the tweet did not contain the word “false” were criticized for failing their responsibility to exercise journalistic scrutiny.

The Washington Post swiftly sicced its top fact-checker on Trump. Glenn Kessler denounced Trump’s “bogus claim.” Kessler gave Trump a lecture on the importance of credibility, writing that since Trump was now “on the verge of becoming president, he needs to be more careful about making wild allegations with little basis in fact, especially if the claim emerged from a handful of tweets and conspiracy-minded websites.” Should Trump persist in wildly distorting the truth, he “will quickly find that such statements will undermine his authority on other matters.”

The media demanded to know where Trump had come up with such a ridiculous notion. The day after the tweet, Trump spokesman Jason Miller was asked by NPR whether there was any evidence to support the idea that millions of people had voted illegally. But surprisingly enough, Miller did have a source: The Washington Post.

In 2014, under the headline “Could non-citizens decide the November election?” the Post had run a piece from two social scientists, Jesse Richman and David Earnest, suggesting that illegal voting by non-citizens could be regularly occurring, and could even be prevalent enough to tip elections. As they wrote:

How many non-citizens participate in U.S. elections? More than 14 percent of non-citizens in both the 2008 and 2010 samples indicated that they were registered to vote. Furthermore, some of these non-citizens voted. Our best guess, based upon extrapolations from the portion of the sample with a verified vote, is that 6.4 percent of non-citizens voted in 2008 and 2.2 percent of non-citizens voted in 2010.

Richman and Earnest’s thesis was extremely controversial, and was so heavily criticized that the Post ultimately published a note preceding the article, pointing out that many objections to the work had been made. But the Post never actually retracted or withdrew the piece. It was ironic, then, that when Trump tweeted about millions of illegal voters, the Washington Post’s fact-checker chastised him for relying on “conspiracy-minded websites.” After all, the conspiracy-minded website in question was the Post itself.

After Trump’s spokesman pointed out that the tweet was consistent with assertions from the Washington Post’s own website, the newspaper’s fact-checking department became extremely defensive. They awarded Miller’s statement an additional “four Pinnochios.” Without actually linking to the Post’s original article about voting by non-citizens, fact-checker Michelle Yee Hee Lee tried to claim that the study wasn’t really in the Washington Post. Instead, she said, it: “was published two years ago in the Monkey Cage, a political-science blog hosted by The Washington Post. (Note to Trump’s staff members: This means you can’t say The Washington Post reported this information; you have to cite the Monkey Cage blog.)”

It was an embarrassing defense. The writers had explicitly said that a reasonable extrapolation from existing data was that 6.4 percent of non-citizens voted in the 2008 election. They had said so in an article that appeared on the Washington Post’s website, displayed in exactly the same manner as every single other piece of reportage. And the Post had never taken the article down or retracted the claim, and had only noted that the piece was highly controversial. Yet instead of apologizing for the Post’s role in spreading a dubious claim, Lee relied on ridiculous distinctions. She insisted that the Post had “hosted” rather than “published” the article. She attempted to enforce a made-up rule, that people aren’t allowed to cite the article as coming from the Post, but must instead cite it as coming from something called the “Monkey Cage,” which sounds far less credible. Yet on the article page itself, there is no such disclaimer to indicate a distinction between non-Post-endorsed “blog posts” and actual Post writing, and the words “Monkey Cage” appear in tiny letters beneath the ordinary full-sized Washington Post logo. There is nothing to make ordinary readers aware that the Post is not responsible for any claims made in these corners of its website.

This is not to say that Trump’s claim of massive voter fraud is correct. It is false, or at least totally unsubstantiated. We don’t have any reason to conclude that millions of people voted illegally. The original study that appeared in the Post was criticized for good reason. Attempts to conclude that millions of people voted illegally voted rest on shaky extrapolations, rather than actual positive proof. But it’s noteworthy that the Washington Post so blithely joined the chorus of those treating Trump’s claim as self-evidently bizarre and deranged, while refusing to acknowledge they had themselves helped to give legitimacy to the idea. Of course, it’s understandable that the paper would be reluctant to make such a concession. While it doesn’t make Trump any less wrong, it does undermine the idea that Trump is entirely reliant on conspiratorial and discredited sources—unless such sources include the Washington Post. But however embarrassing it may be to admit, the imperatives of professional integrity require one to concede that Trump wasn’t just making things up out of whole cloth.

The voter fraud story is indicative of a much wider problem with U.S. political media: its attempts to point out Trump’s falsehoods are consistently undermined by the media’s own lack of credibility on matters of fact. Especially with the rise of “fact-checking” websites, whose analysis is frequently shoddy and dubious, the political media contribute to the exact kind of “post-truth” atmosphere that journalists criticize Trump for furthering.

An interesting and illuminating example of this can be found in the controversy over so-called “fake news.” A few weeks after the election, a series of critics lamented the role of “fake” stories during the election cycle. A study by BuzzFeed reported that “the top-performing fake election news stories on Facebook generated more engagement than the top stories from major news outlets.” A number of commentators saw this as a bad sign for the future democratic governance. Andrew Smith of The Guardian suggested that the proliferation of false stories on social media was eroding the very foundations of reality. In the New York Times, Nicholas Kristof solemnly concluded that “fake news is gaining ground, empowering nuts and undermining our democracy.”

One of the most ominous and sinister warnings about the threat of fake news was found in (again) The Washington Post. In late November, the Post’s Craig Timberg produced a detailed report alleging that much of the “fake news” on the internet was, in fact, a carefully-crafted Russian propaganda effort designed to erode Western governments through the spread of damaging disinformation. The Post cited a “nonpartisan group of researchers” known as “PropOrNot,” who had “identifie[d] more than 200 websites as routine peddlers of Russian propaganda during the election season, with combined audiences of at least 15 million Americans.” Many news stories on the internet, the Post suggested, were not news at all, but lies propagated by Russia in order to further its own state interests. The Post concluded that while there “is no way to know whether the Russian campaign proved decisive in electing Trump… researchers portray it as part of a broadly effective strategy of sowing distrust in U.S. democracy and its leaders.”

The report landed like a bombshell. It was soon the most-read piece on the Post’s website, was covered by NPR, and was being promoted by prominent journalists and commentators as a crucial investigation. But subsequent scrutiny of the Post’s reportage revealed that its evidence for a Russian conspiracy was thin. PropOrNot’s “list” of “Russian propaganda” websites targeted a number of totally innocuous independent media outlets, including left-wing outlet Truthdig and popular financial blog Naked Capitalism. It turned out that to be classified as a “Russian propaganda outlet,” one needn’t actually be associated with Vladimir Putin or the Russian government. For the purposes of making the PropOrNot blacklist, it was sufficient that a media organization be “useful” to the Russian state. By that expansive criterion, plenty of ordinary political criticism and analysis (such as that found on Truthdig) could be classified as “propaganda.” After all, anything critical of the U.S. government could be considered helpful to the Russian government. The Post’s allegations therefore rested on a dangerous premise: the idea that if one can’t prove one isn’t helping the Russian government, then one is helping the Russian government.

Furthermore, the PropOrNot organization itself was highly mysterious and of dubious reliability. Its Twitter feed regularly accused its critics of being “fascists” and “Putinists.” All of its “researchers” were anonymous, and it was unclear what credentials or expertise they had, or who they themselves might be funded by. Thus The Washington Post tarred a series of legitimate independent media outlets as tools of the Russian state, based on the word of an unknown anonymous source.

The Post quickly received intensive criticism over the report. The Nation said it had “smeared working journalists as agents of the Kremlin” by offering up a “McCarthyite blacklist.” Adrian Chen of The New Yorker called it “propaganda.” Glenn Greewnald and Ben Norton of The Intercept said the Post had offered “obviously reckless and unproven allegations… fundamentally shaped by shoddy, slothful journalistic tactics.” In Rolling Stone, Matt Taibbi called it an “astonishingly lazy report” and said that “most high school papers wouldn’t touch sources like these.” (Though Rolling Stone may not have been the optimal venue for launching a denunciation of substandard reporting using unreliable sources.) Yet confronted with evidence that it may have reported a story riddled with falsehoods, the Post (again) refused to admit error. “I’m sorry, I can’t comment about stories I’ve written for the Post,” replied reporter Craig Timberg. (“Can’t,” as is so often the case, meaning “won’t.”)

The irony here was that in writing about the spread of so-called “fake news,” the Post had itself produced a classic example of fake news. After all, wasn’t this entirely the sort of story about which journalists were panicking? A poorly-sourced series of outlandish allegations, that brought harm to people’s reputations without actually providing proof of wrongdoing?

The Post’s catastrophically bad reporting on “fake news” illustrated an unfortunate tendency of the American political press. When it comes to news about Russia or Vladimir Putin, all the usual standards of skepticism and caution (as one might apply to claims made by Donald Trump) seem to disappear. In October, Franklin Foer of Slate wrote a story alleging that a Trump Organization computer server was sending secret communications to Russia. (Amusingly, it turned out that the server was routinely sending the Russians spam promotional flyers advertising Trump hotels.) Mother Jones published quotes from an anonymous former intelligence official, claiming Donald Trump was a secret Russian agent. After Hillary Clinton’s loss, Paul Krugman became especially paranoid and unhinged, tweeting that James Comey and Vladimir Putin had “installed” Trump as president, and declaring that the FBI was essentially in “alliance” with Putin.

Such language almost seems a throwback to the 1950s, likewise a time when sinister Russian conspirators lurked around every corner and beneath every bed. Most of the “Russians are coming” stories were thinly sourced or based on unsupported quotes from anonymous government insiders. Consider this one from (…again) The Washington Post entitled “If you’re even asking if Russia hacked the election, Russia got what it wanted.” The writer argued that the Russian government had a conscious strategy to disrupt Americans’ faith in their systems of governance, and that:

…[the] strategy manifested itself in the Russians’ strongly alleged involvement in promoting “fake news” and disseminating hacked emails stolen from the Democratic National Committee. These emails hurt Hillary Clinton’s campaign and weakened Americans’ trust in the Democratic primary.

Note the key phrase: strongly alleged. When it comes to Russian meddling, it doesn’t matter whether the proof is strong. It matters whether the allegation is strong. Once we have a strong allegation that Russia is doing something nefarious, we can treat it as fact. The press’s treatment of Trump/Putin stories was little short of deranged. One can ponder how much of this was driven by loyalty to Hillary Clinton among certain journalists, versus how much was the sensationalistic pursuit of eye-catching stories. Either way, whenever the subject of Russia comes up, the press has a tendency to blow even the flimsiest rumor into the stuff of airport espionage thrillers. “Vladimir Putin has a plan for destroying the West—and that plan looks a lot like Donald Trump,” wrote Foer in Slate. Headlines like “The secret to Trump: He’s really a Russian oligarch” and “The Kremlin’s candidate” abounded.

But the mainstream media’s looseness with facts goes well beyond stories about Russia and Trump. It’s also furthered by “explainer” websites like Vox.com, which blur the distinction between (liberal) commentary and neutral empirical analysis. Particularly pernicious is the rise of “fact-checking” websites, which are ostensibly dedicated to promoting objective truth over eye-of-the-beholder lies, but which often simply serve as mouthpieces for centrist orthodoxies, thereby further delegitimizing the entire notion of “fact” itself. As Current Affairs has previously argued at length, websites like PolitiFact frequently disguise opinion and/or bullshit as neutral, data-based inquiry.

This happens in a couple of ways. First, such websites frequently produce meaningless statistics, such as trying to measure the percentage of a candidate’s statements that are false. PolitiFact constantly spreads its statistics about how X percent of Trump or Clinton’s statements are rated false, declining to mention the fact that this statistic is empty of any content, since the statements that are evaluated haven’t been randomly selected. The centrist biases of fact-checkers also affect their decisionmaking. Fact-checkers have, for example, insisted that it was wrong to say Hillary Clinton wanted to get rid of the 2nd Amendment. But this isn’t a “factual” dispute at all. It depends on one’s interpretation of the 2nd Amendment’s essential meaning, something that varies based on one’s personal political values.

Efforts to soften critiques of the Clintons were persistent features of fact-checks during the election. For example, fact-checkers have insisted that a factory in Haiti that the Clintons helped build was not a sweatshop, despite the fact that wages in Haitian factories are under a dollar per hour and workers have complained regularly of exploitative and abusive treatment. Conservative writer Sean Davis similarly encountered the topsy-turvy world of Clinton Foundation “fact checking.” When Davis wrote an article about the small percentage of its funding the Clinton Foundation spends on charitable grants (as opposed to its own in-house programming), PunditFact argued that the claim, “while technically true” was nevertheless “mostly false.” Davis was understandably puzzled by the idea that something could be rated false despite “technically” being true.

But this happens frequently on fact-checking websites. Fact-checkers claim that while claims may literally be true, they are nevertheless false for giving “misleading” impressions or missing crucial context. For example, when Carly Fiorina claimed that she had gone from being a secretary to being a CEO, her claim was given “Three Pinnochios” by The Washington Post, even though Fiorina had indeed (by the Post’s own admission) been a secretary before she was a CEO. The Post reasoned that while Fiorina was literally telling the truth, her statement was nevertheless false since she had advantages in life that other secretaries did not have.

The fact-checkers might think that by going beyond the literal meaning of statements, and evaluating the impressions they leave, they are in fact doing a greater service to truth and reality. In fact, they are opening the door to a far more subjective kind of work, because evaluating perceptions requires a lot more interpretation than evaluating the basic truth or falsity of a statement. It thereby creates far more room for bias and error to work their way into the analysis.

A good example of the perils of fact-checking is seen in Donald Trump’s claims over birds and wind turbines. Trump doesn’t like wind turbines, and frequently rails against them on Twitter and in speeches. One of his favorite points to make is that wind turbines kill birds, specifically eagles. At one point, Trump said the following:

“There are places for wind but if you go to various places in California, wind is killing all of the eagles… You know if you shoot an eagle, if you kill an eagle, they want to put you in jail for five years. And yet the windmills are killing hundreds and hundreds of eagles. … They’re killing them by the hundreds.”

This invited a vigorous fact-check from PolitiFact, who rated the claim “Mostly False” and said that Trump was “inflating” wind turbine deaths. Yet wind turbines do kill over 100 eagles per year in California, as PolitiFact admitted. Furthermore, eagle deaths from turbines are such a serious concern to animal welfare advocates. Save The Eagles International has reported “millions” of wind turbine deaths and the Audubon Society has warned that wind turbines, while good for the environment, come with “hundreds of thousands” of unnecessary bird deaths.

Here we see how bias can affect fact-checks. Trump was clearly correct that wind turbines are a serious threat to birds, including endangered birds. Rating him “mostly false” depends on giving the least charitable possible interpretation to his words, suggesting that he meant hundreds were dying within California per year (which he did not say). And since it’s actually about 116 eagles within California per year, this would be a slight exaggeration. But note: Trump’s underlying point is still clearly valid. Wind turbines kill lots of birds. The Audubon Society is concerned. Trump isn’t making this issue up, it exists and it’s serious, and his sources are perfectly sound. The context and implications of Trump’s remarks make them true, even if his statistic is marginally off. But while context matters if it can help prove Carly Fiorina’s point is invalid, it doesn’t matter if it can help prove Trump’s point is valid.

This is a story about glass houses and stones: in order to convince people not to believe in disreputable sources, you must first give them reason to believe that you yourself are reputable.

It’s clear why the fact-checkers wouldn’t want to admit Trump’s point about birds and wind turbines is a good one. First, it sounds ridiculous, even though it happens to be true. Second, it’s Trump, and sober-minded Democratic centrists don’t like admitting that Trump is right about anything. Third, it unsettles Democratic centrist political convictions, because it seems to undermine the case for green energy. (It actually doesn’t. One can argue that wind turbines are worth the cost in bird-lives. Or one can argue that wind turbines should both exist and be made safer, as the Audubon Society does. There is no reason to be afraid of the facts.) But by refusing to admit that Trump is ever right, or at least has something resembling a point, fact-checkers render themselves untrustworthy.

When recently asked about Trump’s claims of voter fraud, Trump surrogate Scottie Hughes gave a statement about the nature of truth that shocked many people:

One thing that has been interesting this entire campaign season to watch, is that people that say facts are facts—they’re not really facts. Everybody has a way—it’s kind of like looking at ratings, or looking at a glass of half-full water. Everybody has a way of interpreting them to be the truth, or not truth. There’s no such thing, unfortunately, anymore as facts. And so Mr. Trump’s tweet, amongst a certain crowd—a large part of the population—are truth.

Politico reporter Glenn Thrush called Hughes’ remark “absolutely outrageous.” After all, Hughes was suggesting that there was no such thing as objective reality, that Trump’s claims of voter fraud were just as legitimate as the claims of those who had “reason” and “evidence” on their side.

Indeed, Hughes’ remark is somewhat terrifying. (Although it doesn’t sound particularly conservative. In fact, it rather resembles a mainstream liberal belief: that ideas of “truth” and “facts” are matters of interpretation, shaped by our personal identities rather than any “objective” reality. Hughes almost sounds as if she has been reading Foucault, and is on the verge of concluding that truth is little more than a series of differing narratives reflecting existing power relations.) If Hughes’ perspective were taken to its logical extreme, it would mean that every form of bigotry and error was just as legitimate as its opposite. Such a world is nightmarish.

But before getting too sanctimonious, journalists should question their own role in giving this perspective a boost. The garbage churned out regularly by CNN and Slate may be better than Trump’s tweets, but it is not that much better. And by failing to show humility about their own ability to generate truth, and themselves being highly detached from the real world, talking-head pundits and biased “data-based” journalists may be helping to create the “post-truth” environment, by robbing words like “true,” “false,” and “fact” of their meaning. By conjuring phony statistics (like “percentage of false statements”) and treating highly subjective and interpretive judgments as if they are Just The Facts, the press steadily erodes the credibility it will need in order to effectively hold Trump accountable. Kellyanne Conway is correct to point out that the single biggest piece of “fake news” was the story that Trump couldn’t win. It’s very difficult for places like, say, BuzzFeed to hold forth on the necessity of accuracy in journalism, when BuzzFeed itself had reported that Trump “plainly has no interest in actually running for office.” Trump has actually established some formidable credentials as a truth-teller against his critics in the press. After all, they were the ones telling him that his confidence of victory was a delusion.

In fact, BuzzFeed even published a lengthy profile mocking Trump-supporting commentator Bill Mitchell for being “post-truth” and “post-math.” To BuzzFeed, Mitchell was laughably divorced from reality for his belief that “enthusiasm” was a far more reliable predictor of electoral success than polls. Mitchell was treated with open contempt by data obsessives like Nate Silver, for his failure to understand “basic math.” But Mitchell turned out to be right. This raises an important question: if Trump and his supporters were labeled “post-truth” or “anti-facts” for the act of ignoring polls, but they turned out to be correct, then why should allegations of being “post-truth” or “anti-facts” be taken seriously? By using these phrases with overconfident abandon against Trump supporters, even when they don’t necessarily apply, members of the press diminish the currency of words like “truth.”

None of this is to suggest that the mainstream media is somehow “just as bad” as fake news from conspiracy theory websites. What’s reported in The New York Times frequently does bear a general resemblance to the truth. (Though not always, and one should never forget the Times’ uncritical repetition of government claims about weapons of mass destruction in the lead-up to the Iraq War.) The point is, rather, that even a single falsehood or misrepresentation can permanently destroy one’s credibility, and being trustworthy requires always being honest and self-critical. If phrases like “post-truth” are used cavalierly, they can become insignificant. If “fact-checks” are not really fact-checks, but are centrist opinion pieces, the word “fact” comes to connote “the highly contentious views of people who call themselves fact-checkers” rather than anything about reality or the world as it actually exists.

Those who say Donald Trump dwells in a “post-truth” realm are not wrong. He lies a hell of a lot, and misrepresents a hell of a lot more. But in order for the “post-truth” charge to be taken seriously, one must be careful and reliable in calling out “lies.” And one must be serious in understanding why people become conspiracy theorists in the first place. If the press is unaccountable, condescending, and secretive, it won’t be believed, even if it’s right. (Similarly, one of the reasons that so many wild conspiracy theories develop around Hillary Clinton is that—as even her supporters admit—she is extremely secretive. As a purely practical matter, if you act like you’ve got something to hide, people will assume you do. And they’re not irrational to make that inference.) If people are heading for fake news, then it is urgently necessary to figure out how to get them back. One won’t do that by continuing to do the same thing, such as continuing to spew biased and speculative punditry. This is a story about glass houses and stones: in order to convince people not to believe in disreputable sources, you must first give them reason to believe that you yourself are reputable.

For progressives, having a reliable and trustworthy media means not being afraid of uncomfortable truths. If wind turbines kill a bunch of eagles, let’s have the guts to admit it. If the Clintons are actually pretty noxious, let’s be perfectly honest about their failings. If Trump is right about something, then he’s right. And if he is wrong about something, but he read it in the Washington Post, then let’s admit that this reflects worse on the Washington Post than on Trump. The truth is a precious thing, and it should never, ever be distorted for partisan reasons. Being credible means being self-critical, and trying to build a press that people can depend on to help them sort truth from lies.

Having a media people can actually trust should be a fundamental goal of Trump opponents. Currently, people don’t trust the mainstream media. And the first thing the media must do is acknowledge that part of that mistrust is entirely rational and reasonable. After that, building true credibility will at the very least require a major rethink of how ordinary political media do business. They will have to interrogate their assumptions more, defend or revise their work in response to criticisms, and get serious about truth, fairness, and accountability. They will need to abandon the assumption, commonly held, that if people on “both sides” are mad at you, you must be doing your job well. And they will need to be extremely cautious in their factual assertions. If I go around asserting that Trump’s attitude toward polls is “post-truth,” then report that Trump is a possible Russian spy, I will have few grounds to complain when Trump’s supporters decide to get their news from alt-right conspiracy websites instead.

Yet it is telling that after the election, the people who were most wrong during the campaign are still producing voluminous commentary. No outlet that wanted to regain trust and build audiences would be keeping such people on its staff. But “pundit tenure” is powerful. Thus is also likely that the quest for credible media will necessitate the creation of new media. CNN and The Washington Post have never shown a particularly encouraging capacity for introspection and self-improvement, and it’s unlikely that they’re contemplating major internal overhauls in their mission and accountability practices. Their institutional imperatives consist, after all, largely of seeking views and clicks. For them, the 2016 election was a success rather than a failure. A lot of people, after all, tuned in. Why should they do things any differently? Thus it would be useful to have fresh, truly independent outlets, ones that disclose their biases, are transparent in their methods, and are constantly trying to improve themselves rather than simply pursuing the same useless sensationalism and empty horse-race punditry. If one’s only options are Breitbart on the one hand, and The Washington Post on the other, readers lose no matter what.

Credibility is extremely difficult to achieve, and extremely easy to destroy. At the moment, the press doesn’t have it. They need to acknowledge that they don’t have it. They need to figure out why they don’t have it. And then they need to begin the long, agonizing, humbling process of trying to get it. The only way to counter fake news is with real news. Not fake real news, or news that merely looks like news but is actually opinion or allegation. Actual real news. Substantive and serious reporting. A commitment to avoiding innuendo and anonymous sources. Transparency and a willingness to atone for mistakes.

Every one of the three major candidates in this election (Trump, Clinton, and Sanders) was hounded by fake or exaggerated news stories. Trump was accused of being a secret Russian agent. Clinton’s email scandal was blown out of all reasonable proportion. And Bernie Sanders was hounded by malicious and unrepresentative stereotypes about “BernieBros.” Yet none of these stories were from fringe blogs and conspiracy sites. They were all produced by the mainstream press, which gave this nonsense primacy over stories about climate change, nuclear proliferation, Syria, health care, poverty, and every other conceivable issue of consequence.

Concerns about fake news are justified. But instead of begging our Silicon Valley overlords to crack down on the free sharing of information, we might start by building a mainstream press that has credibility of its own.

This article is adapted from the forthcoming book Trump: Anatomy of a Monstrosity, now available for pre-order. Ships January 20th.