WEBLOG

May 11th, 2024 (Corrected: 5/13/2024) (Permalink)

How to Lie with Photographs, Part 2

"Let the jury consider their verdict," the King said….
"No, no!" said the Queen. "Sentence first―verdict afterwards."
"Stuff and nonsense!" said Alice loudly. "The idea of having the sentence first!"¹

The best evidence for factual claims is usually that of your own senses, but you probably won't be able to check most claims you read in the media so directly. The second best evidence is likely to be video or still photographs, but both can be manipulated in ways that may not be detectable by your own senses and, instead, require technical expertise to expose. For this reason, just because a factual claim is based on videos or photographs is no reason to accept it uncritically.

It used to be the case that you needed a darkroom and special equipment to fake photographs, but with the advent of digital photography and programs such as Photoshop, anyone with a computer can do so―even princesses². Despite these developments―pun noted―many people still look at photos as though they never mislead and, as a result, they accept such evidence without critical examination.

Though it sometimes takes specialized knowledge to detect fake photography, not every misleading photo involves such trickery. As we saw in the first part of this series³, sometimes it's not the photo itself that lies but its caption, or the text of an article to which the photo is attached.

Dino Brugioni, in his book Photo Fakery, classifies fake photography into four types: "deletion of details", "insertion of details", "photomontage"⁴, and "false captioning"⁵. About the latter, he writes: "The falsely captioned photo differs from other groups of fake photos in that, although the photography has not been altered, the context of what the photograph purportedly conveys is simply falsified.⁶"

In order for Brugioni's four categories to cover all types of misleading photograph, the name "false captioning" must not be taken literally. Rather, any photograph that is misleadingly described, whether in a caption or accompanying article, is an example. In particular, staged photos which are represented as portraying something other than what they actually portray, should fall in this category.

The infamous Cottingley fairy photographs⁷ are examples of this last type. The two young cousins^* who took the photos lacked both the knowledge and access to the tools required to engage in any sophisticated fakery. Instead, they simply copied drawings from a children's book, cut them out, and propped them up in the weeds with hatpins. The problem was not with the photos themselves but with gullible adults, such as Arthur Conan Doyle, who wrote books claiming that they showed real fairies⁸.

Let's look at a more recent example of false captioning. I won't display any of the photos or video discussed, since they are probably owned by the photographer, and you can see them on many of the articles linked in the Notes, below.

On September 20, 2021, the El Paso Times published a report which began:

A mounted U.S. Border Patrol agent shouted commands in a tense encounter with Haitian migrants wading through the Rio Grande near Del Rio, Texas. As the Haitians tried to climb onto the U.S. side of the river Sunday afternoon, the agent shouted: "Let's go! Get out now! Back to Mexico!" The agent swung his whip menacingly, charging his horse toward the men in the river who were trying to return to an encampment under the international bridge in Del Rio after buying food and water in Ciudad Acuña, Mexico. One migrant fell as he tried to dodge, others shielded their heads with their hands.⁹

The article was illustrated by a couple of photos showing Border Patrol agents (BPAs) on horseback, and in the one at the top of the report a BPA appears to be holding up a long, thin cord or strap. If you glance at the photo after reading the above description of a BPA swinging a whip, you might interpret it as showing a whip at the top of its swing.

However, the next day the highlighted sentence was changed to: "The agent menacingly swung his reins like a whip, charging his horse toward the men in the river who were trying to return to an encampment under the international bridge in Del Rio after buying food and water in Ciudad Acuña, Mexico."¹⁰. Also, a "Clarification" was added at the top of the story:

Our reporting team witnessed at least one agent on horseback swing his reins like a whip. We have updated the story to clarify that fact since it was not an actual whip.

In addition to the still photos, there is some video of a BPA on horseback twirling one of this reins, though not appearing to make contact with anyone¹¹. Notice that even the uncorrected report quoted above does not claim that the BPA who supposedly swung a whip "menacingly" actually used it to whip anyone.

Nonetheless, this tiny spark was enough to start a firestorm of criticism of the Border Patrol over the next few days. The condemnation went all the way to the top, namely, President Biden, who was quoted as saying:

It's horrible what you saw. To see people like they did, with horses, running them over, people being strapped, it's outrageous…. I promise you: those people will pay… There is an investigation underway right now and there will be consequences.¹²

Despite referring to an investigation, Biden apparently had already made up his mind that the BPAs should be punished. Verdict first, investigation afterwards. Does the presumption of innocence not apply to BPAs, or is the president exempt from applying it?

"Strapped" is an interesting choice of words because "to strap" as a transitive verb usually means to tie something down, as with straps, but there's a less common meaning, namely, to beat with a strap¹³. Why not say "whipped" rather than the rare "strapped"? Had Biden been informed that no whip was involved? Had he seen the photos or videos? He talks as though he had, pronouncing the sight "horrible" and "outrageous", but did he think that he had seen people beaten with straps in the photos or video?

A few days after Biden's remarks, the outrage continued with an opinion piece in The Seattle Times:

In recent days, images of Border Patrol agents on horseback whipping Haitian asylum seekers at the Texas-Mexico border have reminded Americans of the racist origins―and current practices―of our nation's immigration enforcement agencies. … Title 42…is the Trump-era policy of expelling asylum-seekers, ostensibly on public health grounds; the whip-wielding Border Patrol agents were enforcing this rule when they forced Haitian refugees back into Mexico.¹⁴

Within a couple of days, "Whip-wielding" was deleted and "whipping" changed to "corralling" ¹⁵.

As for the investigation mentioned by Biden, Department of Homeland Security Secretary Alejandro Mayorkas claimed that it would "be completed in days, and not weeks", but six weeks later NPR was fretting that it was still ongoing¹⁶. In reality, the investigation was not completed until almost a year later¹⁷, concluding:

There is no evidence that BPAs involved in this incident struck, intentionally or otherwise, any migrant with their reins. The horses involved in this incident were equipped with split reins which can be twirled by the rider to guide the horse's movements. One BPA involved in this incident also reported twirling these split reins as a distancing tactic.¹⁸

By the time the report was released, the outrage had died down, the incident was mostly forgotten, and few cared what had really happened.

As with the Cottingley photographs, the Border Patrol photos and video were not fake, they just didn't show what some people―including the President of the United States―claimed they did. There were no fairies at the bottom of the garden in the Cottingley photos, and there were no "whips" in the hands of BPAs. In both cases, you don't need any advanced knowledge of photography to detect the false claims, you just have to look closely.

Correction: I originally referred to the two girls who took the Cottingley fairy photos as "sisters"; they were cousins. See: "Cottingley Fairies: How Sherlock Holmes's creator was fooled by hoax", BBC, 12/4/2020.

Notes:

Lewis Carroll, Alice's Adventures in Wonderland (1865), chapter 12: "Alice's Evidence".
See: Seeing is Disbelieving, 3/13/2024.
For Part I, see: How to Lie with Photographs, 12/9/2023.
This is what I initially thought National Enquirer's Oswald and Cruz photo would be, but it turned out to be false captioning; see: Mashed or Matched?, 4/27/2024.
Dino A. Brugioni, Photo Fakery: The History and Techniques of Photographic Deception and Manipulation (1999), chapter 2. See, also, by the same author: "Spotting Photo Fakery", CIA Historical Review Program, 9/22/1993. This shorter and earlier work also divides fake photos into four types, but the first three types differ from those in the later book, though the last type is also "false captioning".
Ibid., p. 22.
See: Fairy Tale, 2/6/2013.
Arthur Conan Doyle, The Coming of the Fairies (1922).
Martha Pskowski, "Haitian migrants face tough choices in Del Rio amid crackdown at Texas-Mexico border", El Paso Times, 9/20/2021. This is the Internet Archive's copy of the original report which has since been corrected. Paragraphing suppressed.
Martha Pskowski, "Haitian migrants face tough choices in Del Rio amid crackdown at Texas-Mexico border", El Paso Times, 9/21/2021. This is the earliest snapshot of the corrected report.
"White House condemns whip use on Haitian migrants", Reuters, 9/21/2021. The title of this video is still uncorrected despite the fact that the description beneath it reads: "The White House criticized the use of horse reins to threaten Haitian migrants after images circulated of a U.S. border guard on horseback charging at migrants near a riverside camp in Texas". So, Reuters appears to have known that there was no whip.
Kevin Liptak & Kate Sullivan, "Biden and Harris harshly condemn horseback wrangling depicted in images from US-Mexico border", CNN, 9/24/2021. Second ellipsis in the original; paragraphing suppressed.
"Strap", Collins Dictionary, accessed: 5/10/2024.
Angelina Snodgrass Godoy, "ICE arrests in Washington tell stories of suffering that goes unseen", The Seattle Times, 9/28/2021. This is the Internet Archive Wayback Machine's archived copy of the uncorrected op-ed article.
Angelina Snodgrass Godoy, "ICE arrests in Washington tell stories of suffering that goes unseen", The Seattle Times, 9/30/2021. This is the corrected version of the op-ed article.
Joel Rose, "The inquiry into border agents on horseback continues. Critics see a 'broken' system", NPR, 11/6/2021.
Carolina Cuellar, "CBP report finds 'no evidence' of border agent whipping Haitian migrants, despite viral photo", NPR, 7/8/2022.
"Report of Investigation", Department of Homeland Security, 5/2023, p. 5.

May 3rd, 2024 (Permalink)

Crack the Combination VII^*

The combination of a lock is three digits long and each digit is unique, that is, each occurs only once in the combination. The following are some incorrect combinations.

807: No digits are correct.
592: Two digits are correct but both are in the wrong position.
012: One digit is correct and in the right position.
758: One digit is correct but in the wrong position.

Can you determine the correct combination from the above clues?

^*Previous "Crack the Combination" puzzles: I, II, III, IV, V, VI.

May 1st, 2024 (Permalink)

Ich Bin ein Berliner & the "Disinformation" Industry

Uri Berliner, "I've Been at NPR for 25 Years. Here's How We Lost America's Trust.", The Free Press, 4/9/2024
In case you don't know, "NPR" stands for "National Public Radio", a government-supported radio network¹.

It's true NPR has always had a liberal bent, but during most of my tenure here, an open-minded, curious culture prevailed. We were nerdy, but not knee-jerk, activist, or scolding.
In recent years, however, that has changed. Today, those who listen to NPR or read its coverage online find something different: the distilled worldview of a very small segment of the U.S. population.
If you are conservative, you will read this and say, duh, it's always been this way. But it hasn't.

Maybe not, but it's been that way as long as I can remember, which is decades. Perhaps it's worse now, but so are all the other establishment news media outlets.
…Back in 2011, although NPR's audience tilted a bit to the left, it still bore a resemblance to America at large. Twenty-six percent of listeners described themselves as conservative, 23 percent as middle of the road, and 37 percent as liberal.
By 2023, the picture was completely different: only 11 percent described themselves as very or somewhat conservative, 21 percent as middle of the road, and 67 percent of listeners said they were very or somewhat liberal. We weren't just losing conservatives; we were also losing moderates and traditional liberals.
An open-minded spirit no longer exists within NPR, and now, predictably, we don't have an audience that reflects America. That wouldn't be a problem for an openly polemical news outlet serving a niche audience. But for NPR, which purports to consider all things, it's devastating both for its journalism and its business model.

At this point, I've omitted three examples Berliner gives of NPR's misreporting of major news stories in recent years: "Russiagate", Hunter Biden's laptop, and the lab leak hypothesis of the origin of COVID-19. About these stories, he goes on to say:
It is one thing to swing and miss on a major story. Unfortunately, it happens. You follow the wrong leads, you get misled by sources you trusted, you're emotionally invested in a narrative, and bits of circumstantial evidence never add up. It's bad to blow a big story.
What's worse is to pretend it never happened, to move on with no mea culpas, no self-reflection. Especially when you expect high standards of transparency from public figures and institutions, but don't practice those standards yourself. That's what shatters trust and engenders cynicism about the media. … But to truly understand how independent journalism suffered at NPR, you need to step inside the organization.
You need to start with former CEO John Lansing. … He declared that diversity―on our staff and in our audience―was the overriding mission, the "North Star" of the organization. Phrases like "that's part of the North Star" became part of meetings and more casual conversation.
Race and identity became paramount in nearly every aspect of the workplace. Journalists were required to ask everyone we interviewed their race, gender, and ethnicity (among other questions), and had to enter it in a centralized tracking system. We were given unconscious bias training sessions. A growing DEI staff offered regular meetings imploring us to "start talking about race." Monthly dialogues were offered for "women of color" and "men of color." Nonbinary people of color were included, too.
These initiatives, bolstered by a $1 million grant from the NPR Foundation, came from management, from the top down. Crucially, they were in sync culturally with what was happening at the grassroots―among producers, reporters, and other staffers. Most visible was a burgeoning number of employee resource (or affinity) groups based on identity. … All this reflected a broader movement in the culture of people clustering together based on ideology or a characteristic of birth. If, as NPR's internal website suggested, the groups were simply a "great way to meet like-minded colleagues" and "help new employees feel included," it would have been one thing. But the role and standing of affinity groups, including those outside NPR, were more than that. …
…[W]hat's notable is the extent to which people at every level of NPR have comfortably coalesced around the progressive worldview. And this, I believe, is the most damaging development at NPR: the absence of viewpoint diversity.
There's an unspoken consensus about the stories we should pursue and how they should be framed. It's frictionless―one story after another about instances of supposed racism, transphobia, signs of the climate apocalypse, Israel doing something bad, and the dire threat of Republican policies. It's almost like an assembly line.
The mindset prevails in choices about language. In a document called NPR Transgender Coverage Guidance―disseminated by news management―we're asked to avoid the term biological sex. (The editorial guidance was prepared with the help of a former staffer of the National Center for Transgender Equality.) The mindset animates bizarre stories―on how The Beatles and bird names are racially problematic, and others that are alarmingly divisive; justifying looting, with claims that fears about crime are racist; and suggesting that Asian Americans who oppose affirmative action have been manipulated by white conservatives.
More recently, we have approached the Israel-Hamas war and its spillover onto streets and campuses through the "intersectional" lens that has jumped from the faculty lounge to newsrooms. Oppressor versus oppressed. That's meant highlighting the suffering of Palestinians at almost every turn while downplaying the atrocities of October 7, overlooking how Hamas intentionally puts Palestinian civilians in peril, and giving little weight to the explosion of antisemitic hate around the world. …
For years, I have been persistent. When I believe our coverage has gone off the rails, I have written regular emails to top news leaders, sometimes even having one-on-one sessions with them. … Throughout these exchanges, no one has ever trashed me. That's not the NPR way. People are polite. But nothing changes. So I've become a visible wrong-thinker at a place I love. It's uncomfortable, sometimes heartbreaking. … But what's indisputable is that no one in a C-suite or upper management position has chosen to deal with the lack of viewpoint diversity at NPR and how that affects our journalism.
Which is a shame. Because for all the emphasis on our North Star, NPR's news audience in recent years has become less diverse, not more so. … Despite all the resources we'd devoted to building up our news audience among blacks and Hispanics, the numbers have barely budged. In 2023, according to our demographic research, 6 percent of our news audience was black, far short of the overall U.S. adult population, which is 14.4 percent black. And Hispanics were only 7 percent, compared to the overall Hispanic adult population, around 19 percent. Our news audience doesn't come close to reflecting America. It's overwhelmingly white and progressive, and clustered around coastal cities and college towns.
These are perilous times for news organizations. Last year, NPR laid off or bought out 10 percent of its staff and canceled four podcasts following a slump in advertising revenue. Our radio audience is dwindling and our podcast downloads are down from 2020. The digital stories on our website rarely have national impact. They aren't conversation starters. Our competitive advantage in audio―where for years NPR had no peer―is vanishing. There are plenty of informative and entertaining podcasts to choose from.
Even within our diminished audience, there's evidence of trouble at the most basic level: trust.
In February, our audience insights team sent an email proudly announcing that we had a higher trustworthy score than CNN or The New York Times. But the research from Harris Poll is hardly reassuring. It found that "3-in-10 audience members familiar with NPR said they associate NPR with the characteristic 'trustworthy.'?" Only in a world where media credibility has completely imploded would a 3-in-10 trustworthy score be something to boast about.
With declining ratings, sorry levels of trust, and an audience that has become less diverse over time, the trajectory for NPR is not promising. Two paths seem clear. We can keep doing what we're doing, hoping it will all work out. Or we could start over, with the basic building blocks of journalism. We could face up to where we've gone wrong. News organizations don't go in for that kind of reckoning. But there's a good reason for NPR to be the first: we're the ones with the word public in our name.

Actually, this is a reason why NPR probably won't change. As a government-supported news outlet, it's insulated from the market; whereas, a for-profit corporation would be more likely to change in the face of a declining audience.
Despite our missteps at NPR, defunding isn't the answer. As the country becomes more fractured, there's still a need for a public institution where stories are told and viewpoints exchanged in good faith. Defunding, as a rebuke from Congress, wouldn't change the journalism at NPR. That needs to come from within.

It should come from within, but this article is itself evidence that it's unlikely to do so, and a financial "rebuke" from government may be the only way to force it to change. Given that Berliner himself has just documented how NPR is not now "a public institution where stories are told and viewpoints exchanged in good faith", why should taxpayers keep paying for it?
Unsurprisingly, a week after publishing this article, Berliner resigned from NPR², which is unfortunate because now NPR has one less actual journalist, and one less voice for intellectual diversity. I don't blame Berliner for resigning―I don't know the details, of course, but the internal backlash against his article probably made it unpleasant for him to continue―but there's even less chance now of any kind of reform coming from within.
Disinformation is a real problem, but the word "disinformation" has been hijacked by censors. "Disinformation" originally referred to misinformation that was intentionally and knowingly spread, usually by governments; now, the word often refers to information that the government doesn't want spread, even if true. Check out the following video on the way that "disinformation" is used as an excuse for censorship: