Previous Month | RSS/XML | Current | Next Month
WEBLOG
March 31st, 2024 (Permalink)
Cry "Havoc!"
An article in a popular science magazine from a couple of years ago had the title: "Climate change is wrecking havoc on microbial diversity"1. What about microbial equity or microbial inclusion? Where's the microbial DEI officer when you need her?
"To wreak" is an uncommon verb with the uncommon meaning "to cause something to come about violently"2, which is most commonly encountered in the phrase "to wreak havoc". Though you may sometimes see other things wreaked―such as vengeance, disaster, or devastation―havoc is almost always wreaked, whereas the others may be brought about in less violent ways.
"Wreck", in contrast, is a familiar word, especially when applied to automobiles. Whereas "wreak" is only a verb, "wreck" can be both a verb and a noun3. For instance, your car might be wrecked in an accident, after which it becomes a wreck, like the ramblin' wreck from Georgia Tech4. The pronunciations of "wreak" and "wreck" differ only by the sound of the vowel, which is short in "wreck"―pronounced like "reck"―but long in "wreak"―pronounced recklessly like "reek".
Given the unfamiliarity of the word "wreak", those who only read the phrase "wreaking havoc" may mistake the word for the more familiar and similar-looking "wreck". This might explain the PopSci article title, above, as well as other occurrences of the misspelling. After all, you can't really wreck havoc, though havoc may wreck other things.
Another possible source of confusion is that the two words have similar meanings: to wreck something is to damage or destroy it, and to wreak is to bring something about in a destructive way. It's no accident that the two words are so similar in meaning and so often confused since both are derived from the same Indo-European root5.
A less common misspelling is "reeking havoc"6. Since "wreak" is pronounced like "reek", those who have only heard the phrase may think it's spelled "reeking havoc", though that would mean bad-smelling havoc.
"Wreaking havoc" is one of those hackneyed "turns of speech" that George Orwell warned about when "prose consists less and less of words chosen for the sake of their meaning, and more and more of phrases tacked together like the sections of a prefabricated hen-house.7" It's a boilerplate idiom: when I read it, I have a vague sense of what is meant, but it certainly doesn't conjure up any strong images of either havoc or wreaking.
Though it's not really a metaphor, it would be best to consign "wreaking havoc" to the same bin as Orwell's "worn-out metaphors":
Many of these are used without knowledge of their meaning…, and incompatible metaphors are frequently mixed, a sure sign that the writer is not interested in what he is saying. Some metaphors now current have been twisted out of their original meaning without those who use them even being aware of the fact. … [A] writer who stopped to think what he was saying would avoid perverting the original phrase.7
Notes:
- This is the title that shows up in the browser tab at the top of the article and in search results; the article itself, by Carla Delgado, has the following headline: "Diverse microbes are key to healthy soil. Climate change is threatening that.", Popular Science, 7/3/2022.
- "Wreak", Cambridge Dictionary, accessed: 3/30/2024.
- "Wreck", Cambridge Dictionary, accessed: 3/30/2024.
- Evan Scott Schwartz, "Team traditions: The history behind Georgia Tech's ramblin' wreck", Sports Illustrated, 11/4/2014.
- John Ayto, Dictionary of Word Origins (1991), under "wreck".
- This seems to be verified by Google's Ngram viewer, though some of these results are probably puns; see: "Wrecking havoc,reeking havoc", Google Books Ngram Viewer, accessed: 3/31/2024. Oddly, both misspellings seem to have peaked in the first decade of this millennium.
- George Orwell, "Politics and the English Language", The Orwell Foundation, 4/1946.
March 25th, 2024 (Permalink)
Who's Counting?
The first question to ask about any statistic is: Who counted it? That is, who is responsible for the statistic; who produced it? Statistics do not grow on trees. To produce a statistic, things must be counted, and people can miscount. So, any statistic is only as trustworthy as those who produced it.
The answer to this first question should prompt some follow-up questions: Were those who counted in a position to do so accurately? Did they have any conflicts of interest that would undermine their objectivity? What does their history show about their trustworthiness? Have they been sloppy or otherwise inaccurate in the past? Have they been caught lying or exaggerating? If those counting were in no position to do so accurately, objectively, or are simply untrustworthy, there's no reason to trust the count.
Case in point: No doubt people are dying in Gaza as a result of Israel's invasion, but how many? A current statistic is 32,3331, a number precise down to the ones place, which makes it appear to be an exact count rather than an estimate2.
The answer to the first question of who produced this statistic is: the Gaza Health Ministry (GHM). The GHM is the official health agency of Gaza and, since Gaza itself is controlled by Hamas, the GHM is also controlled by Hamas3. Hamas is a terrorist group that targets civilians, kills unarmed people of all ages4, takes hostages, and rapes5 and tortures its victims, including fellow Gazans6. In addition, Hamas members do not wear uniforms, hide among the civilian population of Gaza, use hospitals and schools as shelters, then exploit the inevitable civilian casualties that result for propaganda7. If Hamas is willing to use civilian deaths as propaganda, why should it not be willing to exaggerate the death toll in Gaza for the same reason?
Hamas has a strong motivation to exaggerate the death toll, and it's impossible to believe that it has any moral compunctions against doing so. Those defending the statistics put out by the GHM argue that past numbers have been in line with those gathered by the United Nations and other independent agencies8. However, we don't have to go back to earlier wars to question the GHM's death toll statistics. Remember the alleged Israeli airstrike on the al-Ahli hospital? It was GHM that almost immediately claimed that 500 people were killed in the explosion, though it later revised its statistic down a little. We later learned that almost everything about the initial reports of the explosion were wrong9: it was caused by an errant Palestinian rocket rather than an Israeli airstrike, and the rocket did not hit the hospital but exploded in a nearby parking lot10. Was the death toll also wrong?
Here's what a Human Rights Watch report on the incident had to say about that death toll:
The Ministry of Health in Gaza reported that 471 people were killed and 342 injured. Human Rights Watch was unable to corroborate the count, which is significantly higher than other estimates, displays an unusually high killed-to-injured ratio, and appears out of proportion with the damage visible on site.11
The "killed-to-injured ratio"―also referred to as the "wounded-to-killed ratio"―refers to the fact that the number of those wounded in war usually outnumbers the dead. This is common sense, but it's also borne out by history, as one historian writes: "the typical ratio of those wounded to those killed in conflict has historically hovered around the 3:1 mark.12" Due to advances in modern medicine, the ratio has tended to increase in recent decades as more casualties are able to survive their injuries. In contrast, the dead outnumber the injured in the statistics provided by the GHM.
The second question to ask about any statistic is: How was it counted? In the case of the hospital incident, the 500 number was released by the GHM only about an hour after the explosion. It seems unlikely that an actual count could have been done so quickly, so this was probably an estimate as is also suggested by the roundness of the number. However, the later downward revision of the number to 471 could have been the result of an actual count.
In contrast, independent agencies estimated between one and three hundred killed13. These, of course, are only estimates since such agencies don't have the access needed to do an actual count. In the absence of a reliable count of deaths, the best that we can do is rely on estimates from reliable sources.
There are other reasons to doubt the death toll put out by the GHM14, but the above considerations are enough to treat it with skepticism. Given that we cannot trust the GHM, what can we do? We'll probably have to wait until the war is over for anything approaching an objective count of the casualties. We'll also have to wait until Hamas has been removed from power to find out how much control it exercised over the statistics coming out of the GHM, and how much the death toll was inflated. In the meantime, all that we can do is guesstimate the death toll based on what we know; let's do so by estimating a maximum and a minimum.
First, the maximum: the GHM's statistics are not completely useless because we have every reason to believe that they're inflated. So, we can take the GHM's current statistic of around 32K as a maximum.
The minimum is only a little more difficult: Israel itself claims to have killed 12K members of Hamas15, so clearly the order of magnitude (OOM) will be at least tens of thousands. Therefore, both the minimum and the maximum OOM is tens-of-thousands, so that we can claim with some confidence that the death toll in Gaza is in the low tens-of-thousands, but that's about as precise as we can reasonably be.
Notes:
- "Health Ministry In Hamas-run Gaza Says War Death Toll At 32,333", Agence France-Presse, 3/25/2024. The Gaza Health Ministry's own website appears to be inaccessible, presumably due to the war.
- One thing to be wary of in such statistics is over-precision: due to wartime conditions, there's no way that such a number can be completely precise even if it is a count and not an estimate. What is its range of measurement error? We're not told, which is a reason to be skeptical. See: Overprecision, 8/27/2022.
- Isabel DeBre, "What is Gaza’s Ministry of Health and how does it calculate the war’s death toll?", PBS News Hour, 11/7/2023.
- "14 kids under 10, 25 people over 80: Up-to-date breakdown of Oct 7 victims we know about", The Times of Israel, 12/4/2023.
- Farnaz Fassihi & Isabel Kershner, "U.N. Team Finds Grounds to Support Reports of Sexual Violence in Hamas Attack", The New York Times, 3/5/2024.
- "‘Strangling Necks’: Abduction, Torture and Summary Killings of Palestinians by Hamas Forces During the 2014 Gaza/Israel Conflict", Amnesty International, 5/25/2015.
- Ahmed Fouad Alkhatib, "The Origin of Hamas's Human Shields Strategy in Gaza | Opinion", Newsweek, 2/27/2024.
- For instance: Jessie Yeung, Duarte Mendonca, Abeer Salman & Eyad Kourdi, "UNICEF defends accuracy of Gaza death toll as horror unfolds in ravaged enclave", CNN, 11/9/2023.
- Yascha Mounk, "How the Media Got the Hospital Explosion Wrong", The Atlantic, 10/23/2023.
- John Leicester, "French intelligence points to Palestinian rocket, not Israeli airstrike, for Gaza hospital blast", AP, 10/20/2023.
- "Gaza: Findings on October 17 al-Ahli Hospital Explosion", Human Rights Watch, 11/26/2023.
- Tanisha M. Fazal, "Nonfatal Casualties and the Changing Costs of War", International Security, 11/2014.
- Jeremy Herb, "Between 100 and 300 believed killed in Gaza hospital blast, according to preliminary US intelligence assessment", CNN, 10/19/2023.
- For instance, see: Abraham Wyner, "How the Gaza Ministry of Health Fakes Casualty Numbers", Tablet Magazine, 3/6/2024.
- Emanuel Fabian, "IDF says 12,000 Hamas fighters killed in Gaza war, double the terror group’s claim", The Times of Israel, 2/20/2024.
March 18th, 2024 (Permalink)
How to Lie with Headlines
Some recent headlines:
Trump suggests "it's going to be a bloodbath" if he loses the election1
Trump says there will be a ‘bloodbath’ if he isn’t reelected2
Trump says there will be a 'bloodbath' if he loses the election3
This is just a selection; virtually all of the usual suspects in the establishment news media followed suit. Given the similarity of these headlines, it's tempting to think that a memo went out to these different news outlets and they all just followed directions. However, I think that temptation should be resisted; instead, I expect that this is the result of group think and herd behavior: some bellwether of the flock was the first to put out such a headline, and the others quickly followed without bothering to ask where they were going.
Some of the above reports provide enough context to debunk their own headlines. For example, beneath the CNN headline, we read:
Former President Donald Trump suggested Saturday that if he were to lose the 2024 election, “it's going to be a bloodbath for the country." The remark came as Trump promised a "100% tariff" on cars made outside the US. “We're going to put a 100% tariff on every single car that comes across the line, and you're not going to be able to sell those guys if I get elected," Trump said in Vandalia, Ohio. "Now, if I don't get elected…it's going to be a bloodbath for the country."4
Here's what Trump said in context:
If you look at the United Auto Workers what they've done to their people is horrible. They want to do this all-electric nonsense where the cars don't go far, they cost too much, and they're all made in China, and the head of the United Auto Workers never probably shook hands with a Republican before. … China now is building a couple of massive plants where they're going to build the cars in Mexico, and they think that they're going to sell those cars into the United States with no tax at the border. Let me tell you something to China: … those big monster car manufacturing plants that you're building in Mexico right now, and you think you're going to not hire Americans and you're going to sell the cars to us now, we're going to put a 100% tariff on every single car that comes across the line, and you're not going to be able to sell those guys. Now if I don't get elected it's going to be a bloodbath for the whole―that's going to be the least of it―it's going to be a bloodbath for the country, that'll be the least of it, but they're not going to sell those cars.5
Trump is difficult to understand because he tends to talk in one long, rambling, run-on sentence, frequently interrupting himself in the middle of a thought to go off on some digression, sometimes returning to finish the thought and sometimes not. So, there's often some excuse for reporters misunderstanding what he says, but little excuse in this case. In context, it's clear that the "bloodbath" he was talking about was to the automobile industry in the United States, and not some kind of civil war or rioting. Out of context, in the headlines, the false impression is created that Trump is predicting, or perhaps threatening, political violence if he loses.
Some of the news outlets quoted above have now edited their headlines to make them less misleading; for instance, the ABC News headline shown above, which is a screenshot of the original, now reads:
Trump, addressing auto industry, says there will be 'bloodbath' if he loses election6
This is better, but it still doesn't make it clear that the "bloodbath" would be to the industry, rather than that he was just "addressing" the industry when he said it.
Many people read only the headlines, and even those of us who make an effort to follow the news often do not read beyond the headlines. In fact, I usually read just the headlines, perusing the underlying articles only if the headline piques my interest. The headline-only reader will be misled into thinking that Trump was threatening a bloodbath if he's not re-elected. When I first saw these headlines, I thought instead that he was probably warning that his followers would be angry enough to riot if he were to lose again. I had to actually read one of the articles to find out that he was talking about a "bloodbath" to the automotive industry. So, even in the case when the article itself includes enough context to show that the headline is misleading, the headline will still mislead many readers.
The headlines from establishment news sources now resemble those of the old tabloid newspapers, which were notorious for promising more than the story delivered.
Notes:
- Kit Maher & Alayna Treene, "Trump suggests 'it's going to be a bloodbath' if he loses the election", CNN, 3/16/2024.
- "Trump says there will be a ‘bloodbath’ if he isn’t reelected", Today, 3/17/2024.
- Emma Barnett & Jillian Frankel, "Trump says there will be a ‘bloodbath’ if he loses the election", NBC News, 3/16/2024.
- Ellipsis in the original; paragraphing suppressed.
- "Donald Trump talks about Bernie Moreno, his presidential campaign WCPO 9in Ohio", WCPO 9, 3/16/2024.
- Gabriella Abdul-Hakim, Libby Cathey & Fritz Farrow, "Trump, addressing auto industry, says there will be 'bloodbath' if he loses election", ABC News, 3/17/2024.
Recommended Reading: Leo Benedictus, "The media must stop using misleading headlines", Full Fact, 5/28/2021
March 13th, 2024 (Permalink)
Seeing is Disbelieving
- I pay as little attention as possible to the British royal family, but I am interested in fake photography, which leads to the following entry. Kate, the Princess of Wales, has had some recent health problems, including abdominal surgery, and has been out of the public eye since last Christmas. In an apparent attempt to reduce the spread of rumors about her state of health, a photograph that appears to show her with her children was released. Subsequently, the photo was retracted by the Associated Press (AP)1, Reuters2, and even the official UK news agency PA3. Finally, the princess herself issued a public apology for supposedly personally "editing" the photo4.
What exactly was wrong with the photo? I'm not an expert on Photoshop―in fact, I've never used it―so I'll simply point you to some expert discussions in the notes5. However, while the photo can certainly pass the sort of casual glance that most people are likely to give it, even the untrained eye can see some odd things on closer examination. To me, the most glaring fault is at the top of the daughter's skirt on the side closest to her mother, where it appears that a part of the skirt was copied and pasted. Also, as pointed out by the AP and others, the daughter's left sleeve is misaligned with her hand, suggesting that the position of her arm may have been moved―see the detail of the photograph, above; you can see the full photo in any of the sources listed in the notes, below.
The family has so far refused to release the original, unedited photograph, thus contributing to suspicions about why it was retouched. In the absence of a full explanation, all we're left with is speculation as to how and why the photo was manipulated. For what little it's worth, my own speculation is that Kate's face and hands were added from one or more different photos, presumably older ones. If so, I hope it was done out of vanity and not from some more sinister motive.
Perhaps the most worrisome aspect of this affair is that the experts insist that the work done on the photo was amateurish, which is why some of the changes are so glaring. This suggests that a more professional job might have passed scrutiny. In the near future, it may be impossible to tell a professionally faked photo from a "real" one; in fact, the distinction between fake and real photography may be disappearing.
- Update: Late last year, I described how the sports website Deadspin smeared a boy as a racist by means of a misleading photograph6. Deadspin has now been sold and its entire staff laid off7. How much the sudden sale had to do with the controversy is unclear, but the boy's family filed a lawsuit just last month8. I'm no more a lawyer than I am a photography expert, but the parents seem to have a strong case against Deadspin for libel9.
Notes:
- Brian Melley, "Why the AP retracted the first official photo of the Princess of Wales since her abdominal surgery", AP, 3/11/2024
- "News agencies withdraw photo of UK's Princess of Wales", Reuters, 3/11/2024
- "Kate apologises for ‘confusion’ after digitally editing family photo", PA, 3/12/2024
- Bill Chappell & Fatima Al-Kassab, "What to know about the 'confusion' over Kate Middleton's edited family photo", NPR, 3/11/2024
- See:
- Alex Abad-Santos, "Kate Middleton’s edited Mother’s Day photo, explained by an expert", Vox, 3/12/2024. The most thorough analysis of the photo. If you only have time to read one article, read this one.
- Sydney Lake & Irina Ivanova, "Kate Middleton’s Photoshopped family photo and the glaring errors that led to kill notices: ‘Think of it as a Cat. 5 cyclone’", Fortune, 3/11/2024.
- Reece Rogers, "The Kate Middleton Photo’s Most Glaring Photoshop Mistakes", Wired, 3/12/2024.
- How to Lie with Photographs, 12/9/2023
- Liam Reilly, "Deadspin’s entire staff has been laid off after the sports site was sold to a startup", CNN, 3/11/2024
- Ahjané Forbes, "Family sues Deadspin after blackface accusation at Kansas City Chiefs game", USA Today, 2/8/2024
- See: Jonathan Turley, "Deadspin Defamation: Parents of Holden Armenta Move Toward Libel Action Over Black Face Allegation", 12/6/2023
March 3rd, 2024 (Permalink)
What Red Said
Three people nicknamed Goldilocks, Brownie, and Red were at the hair salon. Each had hair of a different color: one was a blonde, one a brunette, and one a redhead, but not in that order.
One of the three said: "I just noticed that none of us have natural hair color that matches our nicknames. Isn't that odd?"
Another replied: "That's true, but the really odd thing is that all of us are getting our hair dyed a color that doesn't match our natural hair color or our nicknames."
Red added: "Well, I'm not dying my hair brown."
Assuming that what each of the three said is true, what is the natural hair color of each and the color after the hair is dyed?
Since no nickname matches the person's natural hair color, there are only two possible color distributions:
Goldilocks | Brownie | Red | |
---|---|---|---|
1. | red | blond | brown |
2. | brown | red | blond |
Goldilocks | Brownie | Red | |
---|---|---|---|
Natural: | red | blond | brown |
Dyed: | brown | red | blond |
Explanation: Since Red was not getting his hair dyed brown, and he couldn't get it dyed red―since that would match his nickname―he must be getting it dyed blond. So, his original color must be brown. Because blond is taken, Brownie must be getting her hair dyed red and her original color must be blond. Finally, Goldilocks must be getting her hair dyed brown, and is a natural redhead. How they got those nicknames is a puzzle that must remain unsolved.
March 1st, 2024 (Permalink)
Heterodoxy Vs. Heresy & the Dog Ate My Data
- Adam Rubenstein, "I Was a Heretic at The New York Times", The Atlantic, 2/26/2024
James Bennet, the [New York] Times’ editorial-page editor, and James Dao, the op-ed editor, were committed to publishing heterodox views. … The Times had hired me to provide research for columnists and to solicit and edit newsy, against-the-grain op-eds. I brushed off my discomfort about the office politics and focused on work. Our mandate was to present readers with “intelligent discussion from all shades of opinion,” as the Times’ founder, Adolph Ochs, put it in 1896. This meant publishing arguments that would challenge readers’ assumptions, and perspectives that they may not otherwise encounter in their daily news diet. …
Ochs was not, of course, calling for publishing just any opinion. An op-ed had to be smart and written in good faith, and not used to settle scores, derive personal benefit, or engineer some desired outcome. It had to be authentic. In other words, our goal was supposed to be journalistic, rather than activist. This, I learned in my two years at the Times, was not a goal that everyone shared.
Being a conservative—or at least being considered one—at the Times was a strange experience. …[T]ake the Hunter Biden laptop story: Was it truly “unsubstantiated,” as the paper kept saying? At the time, it had been substantiated, however unusually, by Rudy Giuliani. Many of my colleagues were clearly worried that lending credence to the laptop story could hurt the electoral prospects of Joe Biden and the Democrats. But starting from a place of party politics and assessing how a particular story could affect an election isn’t journalism. Nor is a vague unease with difficult subjects. “The state of Israel makes me very uncomfortable,” a colleague once told me. This was something I was used to hearing from young progressives on college campuses, but not at work.
There was a sense that publishing the occasional conservative voice made the paper look centrist. But I soon realized that the conservative voices we published tended to be ones agreeing with the liberal line. It was also clear that right-of-center submissions were treated differently. They faced a higher bar for entry, more layers of editing, and greater involvement of higher-ups. Standard practice held that when a writer submitted an essay to an editor, the editor would share that draft with colleagues via an email distribution list. Then we would all discuss it. But many of my colleagues didn’t want their name attached to op-eds advancing conservative arguments, and early-to-mid-career staffers would routinely oppose their publication. After senior leaders in the Opinion section realized that these articles were not getting a fair shake, the process evolved. Articles that were potentially “controversial” (read: conservative) were sent directly to the most senior editors on the page, to be scrutinized by the leadership rather than the whole department.
I have omitted all of the author's account of editing the controversial opinion piece by Senator Cotton, together with the repercussions of its publishing, because it is long and needs to be read as a whole.
If the Times or any other outlet aims to cover America as it is and not simply how they want it to be, they should recruit more editors and reporters with conservative backgrounds, and then support them in their work. They should hire journalists, not activists. And they should remember that heterodoxy isn’t heresy.
- Oliver Wiseman & Vinay Prasad, "We’re Not Curing Cancer Here, Guys", The Free Press, 2/22/2024
A top cancer surgeon at Columbia University is under scrutiny after one of his research papers was retracted for containing suspect data. Twenty-six other studies by Dr. Sam S. Yoon, who conducted his research at the Memorial Sloan Kettering Cancer Center, have been flagged as suspicious by a British scientific sleuth called Sholto David. David raised the alarm after spotting the same images across different articles that described wholly different experiments. He has also found duplications and manipulated data in papers published by researchers at the Dana-Farber Cancer Institute in Boston that have since been retracted.
… In search of reassurance, I called up an expert: oncologist, [University of California, San Francisco] UCSF professor, the author of more than 500 academic papers, …Vinay Prasad. Here’s an edited version of our conversation. …
Vinay, how worried should we be about the problem of fraud in cancer research?
Extremely worried. There’s something very unique about all these papers that allows people to find the fraud, and that is they report the raw data, in the form of images. Most papers, though, do not contain images. The data is all hidden. The researchers only provide a summary of the data. You have to worry how much fraud you’d find if everybody provided all the raw data. I suspect you’d find a gargantuan amount of fraud. This is merely the tip of the iceberg.
Most laymen like me assume all the data is transparent in medical research. You’re telling me that’s not how it works?
Scientific papers are like someone’s dating profile on an app. They’re picking what pictures to show you and what stories to tell you. You don’t get to see the whole library of photos on their phone. Researchers are only presenting a sliver of what they’ve actually done. And just like a dating app on your phone, everything is inaccurate.
That’s shocking, Vinay. What can we do about it?
These concerns have been brewing for a while and they are reaching a tipping point. The fact that … there’s been all this image manipulation shows that the most venerable institutions are no safeguard against malfeasance. What punishment have any of these researchers actually faced? … All of the authors of these disputed papers have, to my knowledge, faced no sanction. Their paper gets withdrawn, but they still get promoted. There’s no punishment. A few years ago, there was a proposal by the International Committee of Medical Journal Editors arguing that every paper published in the top journals should make the raw data available. That proposal was shot down because people were worried about their careers, and that other researchers would take their data and use it to make breakthroughs before them. Sharing is the solution. You should have to make all the data available whenever you publish medical research.
Not just medical research.
We start this month with another lengthy article* on the politically-motivated collapse of journalism at The New York Times.
* See: Illiberal Journalism & Tea with Terrorists, 1/1/2024
Disclaimer: I don't necessarily agree with everything in these articles, but I think they're worth reading as a whole. In abridging them, I have sometimes changed the paragraphing and rearranged the order of the excerpts in order to emphasize points.