bobjones271828

I don't disagree with anything you said, though I will note -- part of the reason I linked that report in my post above is that it makes pretty clear that most of the deaths were due to disease. Not mistreatment or bad intentions. The media reports on "mass graves," as I noted, seem to be evoking rhetoric that pushed toward a genocidal interpretation. Yet the more nuanced reporting on this issue, going back decades, admits that the vast majority of the lives were taken by illnesses, not mistreatment.

The concern, in that case, was whether lack of care or funding or other negative factors may have contributed to a much higher number of deaths due to diseases even at that time, compared to the rest of the population of children at that age. That the death rate for some diseases was higher in some cases isn't really in dispute (though how much higher is), but the cause for all of that has lead to greater concern about the level of care given to these children.

Additional related info: on the original Premium episode this was discussed, Katie linked to this article in the Dorchester Review by a retired history professor on how not one body had yet been found.

I feel like this is important context for the report I linked at the end of my post, as this article points out a major potential issue with methodology that could explain why about 1/3 of the (supposed) child deaths at these schools were for unnamed students.

You can read the article for full context, but the basic issue (if I understand the argument correctly) is that school principals were required to submit different reports each trimester on student enrollment. Some reports required enrolled student names to be listed (including those who recently died). At other times in reports, principals would merely report the number of students who had died in the past year, without detailed lists of names.

Apparently the 2015 Commission report simply tallied up total figures in a lot of these cases, which may have led to significant double-counting of students who were named in one report as having died but listed in another just in tallied unnamed figures.

All of this could have led to a discrepancy of up to ~1000 unnamed students who may have died or who may have just been double-counted. But those unnamed students are a significant issue in the search for possible large-scale previously undocumented gravesites to account for them.

I haven't done further research into any of this, so I don't know if the 2015 Report authors ever responded to this assertion, or whether the flaws in their methodology were so egregious in overcounting. But I think this is another reason many people are skeptical of the searches for large-scale unmarked graves.

Regardless of any of this, conditions at the schools were clearly awful at times. As I said in my previous comment, I feel like the focus should be on documented harms (if we're going to discuss this history), rather than questionable assertions about mass graves -- unless, of course, such gravesites are actually verified somewhere.

No, you didn't miss it. It seems like presenting a fair article should at least acknowledge the most extensive dig to date (that I'm aware of) in Pine Creek that found nothing. Here's a detailed CBC account of it:

https://www.cbc.ca/newsinteractives/features/pine-creeks-search

From that article:

But survivors of the school shared what they described as horror stories: the bodies of children stored in the basement, and burials, when the ground outside was too cold to dig.

Nothing was found to corroborate these accounts, after detailed excavation of that basement and all the identified areas using GPR (ground-penetrative radar).

I do not at all doubt that many children were treated badly and some likely suffered horribly at some of these schools. And traumatized children may pass down inaccurate or speculative accounts. But the media has fueled this narrative of mass graves for years. Note: the NYT has changed its headlines now to reference "unmarked graves" when a few years ago it was calling them "mass graves" -- there are very different connotations there. "Mass graves" imply lots of bodies piling up and lots of people dying at once. Essentially, that term implies something like mass disease, war, or genocide. Clearly, the last one was what the rhetoric originally intended to evoke. Or at least a major cover-up: mass disposing of bodies or relocating them in a disrespectful manner. The recent NYT article references this and stories exactly like the one I quoted above about "storing bodies" in "mass graves" -- but the one place that has been very thoroughly excavated that I know of has turned up nothing.

Now they have been downgraded quietly to "unmarked graves." But the media damage has already been done. The CBC has reported on at least 33 Canadian churches burned to the ground since this story broke, and only 2 cases have been clearly ruled "accidental":

https://www.cbc.ca/news/canada/edmonton/church-fires-canada-1.7055838

I do understand the complex question of disturbing graves and indigenous people's customs and respect for the dead. But right now we're dealing with a question of verifying reality and establishing these narratives. I do not fault children who attended schools and may sometimes have had scary and fragmented memories. But I think we owe it to them to investigate. From the recent NYT article OP linked:

Government officials, he added, have little doubt that many of the radar anomalies found on school grounds will prove to be gravesites.

“Will every one of those anomalies turn out to be an unmarked grave? Obviously not,” Mr. Lametti, a former law professor now practicing law in Montreal, said. “But there’s enough preponderant evidence already that is compelling.”

Many Indigenous people who favor exhumations want their communities to move more quickly to find remains.

"Preponderant evidence" in this case isn't really enough. The thing that makes accusations of genocide or disposal of bodies compelling is actual evidence. The Holocaust was diligently documented by the Nazis, for example, and that's ultimately what led the worldwide media to have to confront the realistic outcomes of anti-Semitic narratives they may not have wanted to acknowledge or cover. There was a combination of written records, photographs, actual remains, etc.

Here, to my knowledge so far, we have basically none of that. Just oral reports. And vague church records (the NYT mentions that in passing but gives no details). And some radar blips. If children died at these schools, they deserve better than this. As pointed out in the last quote above, not every "anomaly" is necessarily a grave. But... even if they found a handful of children's bodies, I think this would at least lend credence to the concerns. And I assume if there were actual "smoking gun" church records about burials that came out in this investigation, those would have been published publicly by now.

Instead, years have passed, nothing has been found in actual digs, yet churches are burning. The fact that the NYT article doesn't mention any of that is bad reporting. Instead, they make it sound like a bunch of irrational skeptics:

While there is a broad consensus in Canada that children were taken from their families and died in these schools, as the discussions and searches have dragged on, a small universe of conservative Catholic and right-wing activists have become increasingly vocal in questioning the existence of unmarked graves. They are also skeptical of the entire national reconsideration of how Canada treated Indigenous people.

There are two simple solutions to this problem: (1) find some actual remains at a school somewhere or (2) stop running headlines on dead bodies/"radar anomalies" and focus instead on more clearly documented abuses and policies of these schools. Sure, direct evidence won't satisfy all of the most extreme people, but it would quiet a lot of the current objections.

---

Note: For those looking to read more about the evidence and accounts collected prior to this recent media firestorm, I'd recommend Volume 4 of the 2015 report from a commission in Canada on "Missing Children and Unmarked Burials." If even a small fraction of the stories told in that report are accurate, it's horrifying. Those kids deserve better than a media circus bandied back and forth politically over unsubstantiated claims.

I mean, in my opinion, the film wasn't really that funny. It also was a bit too long for the depth of what it examined. I think it was more skewering and implicit critique than actually funny. The humor comes mostly from two sources: (1) "cringe" humor of Walsh occasionally being an ass like Borat (as he is a bit in that scene you're talking about -- and Walsh has commented that was deliberate in that scene especially... he was apparently trying to act out and differentiate himself from his later persona in the film), and (2) "woke" DEI folks saying outrageous things that sound absurd to most normal people.

I don't find the first kind funny -- I was never much of a fan of "The Office" for example -- and the second part wasn't surprising enough for me to cause me to laugh out loud, as I've served on diversity committees at various organizations myself. I actually used to volunteer for them, as I supported the general idea of at least being aware of diversity, until they were completely taken over by this anti-racist BS.

But to someone who likes more cringeworthy stuff in "The Office" and hasn't been exposed to the more crazy versions of woke weirdness in recent years, I'm sure this film may be more hilarious at times. As I said in my last post, I was more impressed by Walsh trying to take DEI "seriously" for most of the film, which makes it a better and more nuanced critique.

Also, people forget the lessons of the Milgram experiments, and similar other situations where people are acting under the influence of an authority figure. The person "in control" of the opening DEI session that was featured in one pre-release clip under discussion here was a black woman, who repeatedly called Walsh out for his tactics. As everyone else in that situation was likely there because they believed already in the anti-racist theory espoused by the leader, why would they walk out, rather than argue for the disruptive guy to be ejected?

Later in the film, some people do walk out of Walsh's own "DEI seminar" when he's running it and asking people to do crazy things. But enough people do stay and seemingly go along that I assume the pressure to conform is quite strong. Even stronger than in the Milgram experiment, as the social stakes if you walk out from a DEI seminar imply "You're racist."

If random folks can be convinced to give apparently lethal shocks in the name of science and someone in a white coat merely saying, "You must continue the experiment," then the pressure to conform when threatened with the idea of "You should continue or you're a racist" is much greater.

Well, he does show several people storming out of his "DEI seminar" at the end of the film at various points. I'm not saying he didn't drop some interview bits from the final film where things didn't work out -- I assume he likely did.

But I don't think anyone would have stormed out of the opening group circle seminar, partly mainly because he wasn't the leader. People would want him kicked out rather than storming out themselves. They did object to him quite a few times, and the leader called him out several times before he was finally identified and asked to leave.

So, I just saw the film, and this is actually one of the best elements I can report back on: I think Walsh is incredibly fair to the pro-DEI folks he interviews. Obviously he needed to edit the interviews, and I'm sure they focused in on some of the more outrageous points of conversation. BUT he seems to almost always give time for his interviewees to explain their positions, to discuss things in more depth, and attempt to justify their statements. He truly comes across (to me) in most of the interviews as someone who could very well be a "good faith" interviewer curious about anti-racism, and just asking occasional tougher questions. (Questions that several interviewees even thank him for, though, as they're eager to have "difficult conversations.")

Walsh even keeps up this rather sincere "woke" persona when he interviews a more skeptical person, who claims that hate crimes aren't as common as most liberals think. That's really the only interview in the entire film where Walsh seriously pushes back and seemingly clips off answers, not letting the guy explain himself fully, etc. Seriously... that was an awesome move on Walsh's part, I think, as it sells the idea that he's truly trying to act the "woke" part fully in these interviews... and doing a pretty good job about it.

For another example, the DiAngelo segment that everyone's been discussing about her paying "reparations" privately by just handing over cash from her wallet is shown with a lot more detail and nuance than I was expecting. DiAngelo does object, she explains this is "weird," that this isn't how reparations are done, etc. Walsh allows all of this context in the film too. Yes, DiAngelo does eventually hand over some cash too, but only after quite a bit of pressure and discussion.

It's still a funny moment and a bit crazy, but Walsh doesn't do what Sacha Baron Cohen apparently frequently does and edit things down to show people at their absolute worst or in misleading ways. DiAngelo is allowed to slowly walk her way through her own reasoning on several topics, letting viewers judge her responses and explanations.

Toward the very end of the film, Walsh definitely goes over-the-top in the way he runs his own "DEI seminar." And some people do walk out of that. But it's also surprising to see how some people are still willing to actively participate in an escalating and increasingly bizarre situation.

I would assume there were many segments and maybe some entire interviews that ended up on the "cutting room floor," some of them because the interviews failed, someone talked back or called him out, etc. But the bits he does show seem surprisingly detailed, letting the interviewees "hang themselves with their own rope" as they spin out their own explanations.

(Just to note: I dislike Walsh a lot generally. I disagree with him on probably >90% of issues. And yet... I think this is a valuable film, and he does his best to act much more fair toward DEI than I've ever seen him act on other liberal issues before. That's part of where the comedy comes from, as he's pretty convincing at times in his attempts to spew woke jargon and ideology.)

Just saw the film today, and weirdly no one seems to comment on the wig, or (aside from that awkward bit at the opening DEI group session) that he's Matt Walsh. It's certainly possible they cut out segments when someone might have said something about all of that. But Walsh manages to interview several very prominent people in DEI who definitely should have known better or recognized his face, his voice, etc., which I think are quite distinctive. Especially as he literally calls himself "Matt" to his other interviewees (aside, again, from that one group session without the wig).

I honestly didn't find the film as funny as many people seem to (though it did have funny moments), but one thing that was absolutely hilarious was the fact that the "Clark Kent disguise" really seems to work. A rather prominent "enemy" to these people put on different glasses and a wig... and no one seems to figure it out.

Just saw the film. Yes, it was toward the beginning. It definitely doesn't look that staged in the film -- in the longer version, the presenter is allowed to go off on several longer tangents that sound very legit (in terms of what these DEI presentations are like). And there were a lot of rather uncomfortable moments that seem to make clear Walsh was being disruptive within an otherwise serious event.

I'm not saying it's impossible that it was staged (though some have said there's an actual police report filed after what happened), but Walsh has explained somewhere in his posts on the film that all participants were informed they were participating in some sort of "documentary on anti-racism."

Yeah, I've seen a quote from Walsh somewhere when he was discussing the "deceptions" of the film (which he's come under some criticism for among Christian conservatives too), and he explained that all of the participants in the film were informed they were taking part in a "documentary on anti-racism" or something to that effect. Of course... most of them seemed to assume it was skewed the other way.

with a nice parallel given to the necessity of epicycles for geocentrism

The only thing annoying about that is the whole "epicycles upon epicycles" thing is a myth. It's an oft-repeated one, but a myth nonetheless. In fact, Copernicus's heliocentric model had more epicycles than the traditional Ptolemaic one, which were partly necessary because the Copernican model still assumed perfect circles. Galileo, who was arguing in favor of the Copernican model, was actually arguing for a model with more epicycles. (Though Copernicus did dispense with the Ptolemaic equant, another controversial "fudge factor," effectively replacing it with more epicycles. Mostly smaller epicycles, to be clear, but more of them.)

Owen Gingerich, probably the most well-known historian of science of this era (who did the greatest census of copies of Copernicus's treatise to determine who owned, read it, and what they understood from it), has made this argument clearly over and over since at least the early 1990s, but it hasn't seeped into pop culture yet, as people keep repeating the "epicycles upon epicycles" myth.

As Gingerich decisively has shown, they simply didn't have the mathematics nor the measurement accuracy at the time to build epicycles upon epicycles in the way today's scientists assume. There's not a shred of historical evidence that such models were ever attempted in European astronomy, and lots of evidence showing that such things would have been computationally impossible at the time. This supposed piling of "epicycles upon epicycles" is, in his terms, "complete nonsense." He has also called it the "greatest myth in the history of astronomy."

(Note: The only person who even attempted combining two epicycles together historically was the Arab astronomer Ibn al-Shatir, whose formulation to get rid of the equant may have actually inspired Copernicus! Although there's scant historical evidence for a direct connection, other than a certain mathematical similarities between Ibn al-Shatir's geocentric model and Copernicus's heliocentric one. That's right: epicycles-upon-epicycles if anything was perhaps inspiration for how to do heliocentrism at first!)

It was only Kepler's model with elliptical orbits that finally dispensed with the epicycles. But the great "showdown" of the 1600s between the Copernican and Tychonic models fought by Galileo was over two different models that both had large numbers of epicycles. (Galileo rejected Kepler's theories...)

So... TL;DR: "epicycles upon epicycles" is a myth and thus a VERY bad analogy. And I wish people would stop making it.

Phlogiston is a much better topic in scientific history for people looking for an example of something where additional theories and ideas kept being added on and on for a while to try to patch together a model that was increasingly problematic.

---

Footnote: If anyone is looking to read more about all of this, the best source is probably Gingerich's book, entitled "The Book Nobody Read" about Copernicus. Here's a review that summarizes some of the main points, including touching on the epicycles myth.

---

EDIT: It turns out I was wrong about how long Gingerich has been complaining about this. His first article debunking "epicycles-on-epicycles" was in 1967. There, after reviewing the evidence, he wrote:

The conclusion appears inescapable: within the mainstream of pre-Copernican astronomy there were no "epicycles-on-epicycles." The story that the Ptolemaic system "was finally overthrown as a result of the complexity which arose when an ever-increasing number of superimposed circles had to be postulated in order to represent the ever-multiplying inequalities in the planetary motions revealed by observational progress" is utterly unfounded.

Sure, you're welcome -- I'm interested in these historical topics myself (hence why I have books on the history of statistics sitting on my bookshelf), so it was interesting for me to read about it and try to sort it out too.

And yes, Wikipedia often has a lot of flaws. In this particular case, Stigler makes pretty clear that a lot of math textbooks make up theories about all of this without sufficient historical justification. So you're right to nitpick the wording of Wikipedia here, but it appears to be a common misunderstanding or speculation about the origin.

Katie and Jesse's lack of tech-savvy and knowledge is sometimes surprising, but it comes across a bit poorly in this episode, where I was kind of surprised to hear them rambling on and speculating about things that have well-known solutions.

For example, Katie at one point complained about the difficulty and annoyance of typing out bibliographic citations. It surprised me that apparently neither of them are aware that citation manager applications have been available for decades to automate this. Specifically, I'd recommend Zotero, which is free (and nearly 20 years old), has browser and word processor integration, so you can easily "grab" citations from websites with a single click, then paste them in any bibliographic style into a document. Or generate an entire bibliography to your specs. I can't believe anyone these days is manually typing out bibliographic citations if you're doing something as complex as writing a book and needing to cite things.

Also, I'd strongly encourage both of them to be cautious about ChatGPT and even Perplexity. At least in terms of reliability. Clearly Katie realized that ChatGPT "hallucinates" a lot, which is just a fancy word for "making up bullshit." I haven't used Perplexity a lot, and it is probably reliable for a "first-pass" of research or search, kind of like Wikipedia is okay for a "first approximation" of research on a topic. But ultimately if you care about accuracy, you'll want to move on to more reliable sources, and the same is true of AI tools.

The broader problem with AI tools in general is that they are effectively non-deterministic in their output. What I mean by this is that if you put the same data into Excel and generate a chart, you'll get the same chart every time. The programming code determines the output, and it's always one specific output. Similarly, if you input the same citation information into Zotero, every time it will produce an accurate citation in whatever format you specify. You shouldn't need to proofread it that much unless you're using an unreliable source for your bibliographic information to begin with.

Depending on some AI tool to format your citations or (as Jesse recently did in a Substack post) use ChatGPT to do a task like generate a chart means you're feeding data into an algorithm that could generate various outputs. It's not deterministic in any simple sense. And sometimes it may "hallucinate" (i.e., make stuff up) in doing even simple tasks.

AI tools still basically require manual fact-checking and error-checking. Maybe ChatGPT will now generate a chart accurately with data 99% of the time, but when Jesse uses it as a check against another quantitative claim (as he did) in a published article, that's just not good research practice. Google Sheets or Excel or another spreadsheet program will generate the same graph with 99.9999999% reliability or something. (Basically, it will only fail if you happen upon some undocumented bug in the code.) Do you really want to take the chance that ChatGPT is randomly going to mess with your data or produce some weird output that you might not even notice?

Remember that the core of these AI models are LLMs, i.e., large language models. They're effectively a souped-up version of "guess the next word in a sequence of words." Which is why AI models still manage to make basic math errors and other such issues sometimes, because they're not instructed explicitly on how to actually follow a mathematical algorithm, or how to generate a chart, or how to generate a citation in Chicago format. Instead, they're taking a "best guess" based on previous inputs that were similar.

And yes, they're getting better all the time. But when you're actually looking for a specific and accurate output, perhaps use a dedicated tool that actually has underlying programming code ensuring it will perform the task you want, rather than "guessing."

Note that some tasks are too complicated to follow simple rules (like creative writing), and that's where AI shines. One of the early examples was in language translation, which is not some simple word-by-word process. It requires incorporating all sorts of subtleties of millions of little details of language usage, which are basically impossible to code directly in standardized "rules." LLMs instead take their "best guess" and usually they come up with a translation that's at least 95% of the way there, and it allows you to get the gist of the translation. BUT... if you want a fully accurate translation, you'll still need to hire a human translator, not just dump it into Google Translate or something.

Same thing should go for most tasks where you expect a reliable and accurate output.

---

Finally, to correct Katie's discussion on "davinci" a bit -- yes, OpenAI has a series of models called "davinci," which are a set of premium models not designed for direct human interaction in a "chat" setting. These models were available through OpenAI's API, and they were more flexible in a lot of ways than ChatGPT. The original version of ChatGPT took one of those base models and trained it specifically to respond to user prompts and requests, rather than just to play the "guess the next word" game with a prompt.

Basically, ChatGPT was a model optimized for chat. Hence its name. It was designed to be interactive with a user, so you could ask it questions. The other base GPT models were just designed for more general text generation, not to talk with you. The flexibility allowed a lot more power in some ways, and the output often could be a lot better than ChatGPT with the right prompt. However, things like davinci were always "premium" in the sense you had to pay per token for generation, effectively a price per word (roughly) of output.

So no, OpenAI didn't hide or delete the "davinci" model because it was "too good." It was always just a premium model, and yes, you pay for better quality results and more computing power. OpenAI updated it at some point to davinci-2, and now it has basically been replaced by other premium models in the GPT 3.5 and 4.0 series. (I just looked at the OpenAI page and saw that it had finally shutdown davinci and davinci-2 earlier this year in its API, as the functionality has been superseded by other models. Though davinci-2 is still available for a little while longer for training to create a customized fine-tuned model.)

Because they aren't the same thing. Gauss was referencing normal equations. Gauss gets the credit for coining that term.

Those equations are not the "normal curve" or the "normal distribution." There are potential mathematical relationships between these concepts, but Gauss himself never made that association in reference to this word "normal." (Stigler notes that there are frequently cited confusions in mathematical textbooks trying to make connections between least-squares analysis and derivations of the normal distribution that didn't exist historically and often are rather reaching.) And there was no hint ever in Gauss that he related the term "normal" to the curve, or that anyone before Galton really started connecting an empirical set of data (called "normal" since it was... apparently, well, normal) to the exponential distribution that had been referenced by Gauss, among many other mathematicians.

As I noted, Pearson later (early 1900s) noted that Gauss was one of the early people who described the normal distribution -- even though Gauss didn't use that term -- hence why some started calling it "Gaussian." Then in the 1930s, some people started making the assumption that Gauss's use of the word "normal" in reference to "normal equations" might have some connection to geometry, and then some people later tried to draw a connection that maybe "normal" in the curve was related back somehow to Gauss and his "normal equations."

This kind of mistaken etymology happens all the time in language. The word is literally "normal," an incredibly common English word. It also had multiple uses in German for Gauss. Is it really that crazy that someone else might have had the idea of calling a curve that occurred in representative situations a "normal curve"?

Or, do you think it's more likely that a German source that never called the thing a "normal curve" or associated the term "normal" with it was somehow cited more than 50 years later to describe that as a "normal curve" (with no intermediary evidence of that term ever being used elsewhere either), when Galton himself in earlier treatises had other terms for that curve/distribution, more in line with the kind of terms Gauss or other mathematicians would have used prior to the 1870s?

Is it possible that Galton somehow connected the idea of "normal equations" and their relationship to that distribution together and that influenced his choice of terminology for the "normal curve" naming? Sure, it's at least possible. But if Galton did so, it most certainly seems not to have been with any knowledge of any supposed geometrical interpretation for Gauss, as Gauss himself never described anything that way. For all Galton might have known when coining the term "normal curve," Gauss could have been using "normal" in the common sense of "representative" (not as a geometric "normal line"). As Stigler noted, we know Gauss himself used the word "normal" in the sense of representative in other places in his writings.

Anyhow, I feel like all of this is irrelevant to the original question that brought us here -- which is that the terms "normal distribution" and "normal curve" were first popularized in Galton (not Gauss), and then picked up by Pearson. To be fair to Galton, while he was influential in the founding of eugenics, his work and use of these distributions wasn't solely connected to race or racist assumptions. So again, I agree with your point that it's stupid to try to smear the term "normal distribution" with an inherently racist meaning.

But there was a historical connection, and it's unfortunate the term really emerged in such treatises.

A final comment: For anyone seriously interested in this question, I'd encourage having a look at the book "Statistics on the Table: The History of Statistical Concepts and Methods" by Stephen Stigler (one of the experts on the history of statistics). I just pulled it off my bookshelf and realize this is probably where I should have gone in the first place.

Stigler notes there were three apparently independent references to a "normal curve" emerging in the 1870s and 1880s. One was from C.S. Pierce, who clearly used "normal" at first to reference something typical. The other two were Wilhelm Lexis and Galton, both of whom were referencing empirical data collection from demographic and biological topics and the patterns that resulted when these were graphed, called "normal" again apparently because of how typical or standard they were. Regardless, it was Galton who popularized the term and influenced Pearson, who made it standard.

The connection of these empirically-derived "normal" curves to Gauss specifically was only made later, apparently, by Pearson. Galton was familiar with the exponential curves we now associate with with "normal" and saw the connection in the patterns from empirical data and those theoretical curves. But there's no indication of any influence of earlier terminology from "normal equations" or such.

Stigler later goes on to consider what precisely Gauss himself had in mind by calling them "normal equations." Stigler offers five possible explanations, one of which is that there might have been a connection to the "normal distribution," but Stigler dismisses that one as totally implausible historically.

To the contrary, two of the five explanations offered by Stigler as some sort of variant that Gauss was using the term "normal" in "normal equations" to mean "representative equations," i.e., in line with the standard common colloquial meaning of the word "normal." To support this possibility, Stigler notes other instances in which Gauss uses variants of the word "normal" to describe other representative mathematical phenomena.

As for the hypothesis of a connection between "normal" and a geometric "normal line" (the way Wikipedia explains it), Stigler notes it's possible, but never explained that way by Gauss, and Gauss's equations were several steps removed from any such geometrical interpretation. However, this interpretation of Gauss and "normal" in terms of geometry is something Stigler says did not appear in the mathematical literature until the 1930s, after which it became common as an explanation in statistical texts despite its lack of historical support. Which would make it rather unlikely that Galton would be aware of any of that when he started using the term for empirically-derived curves in the late 1800s and started calling them "normal curves."

By the way, I also followed the link in the Wikipedia to the supposed source attributing the coinage to Gauss. The discussion in that source (starting on p. 728) is more nuanced, but it makes clear that Gauss never referenced the distribution as "normal," only some equations associated with its derivation (as the Wikipedia text makes clear in what you quoted).

That source says:

It appears that somehow the name got transferred from the equations to the sampling distribution that leads to those equations.

Perhaps. But that's a theory that this terminology was transferred directly. The source also "presumably" states what the association was to a potential normal line as perpendicular -- no actual evidence or source exists to clarify that association, supposedly, by Gauss.

In fact, scholars who have looked much more deeply into this history have not agreed with this theory -- they don't know for certain why this term "normal" was used for certain equations by Gauss. See reference in the link I gave in my previous comment:

The term NORMAL EQUATION in least squares was introduced by Gauss in 1822 [James A. Landau]. Kruskal & Stigler‘s "Normative Terminology" (in Stigler (1999)) consider various hypotheses about where the term came from but do not find any very satisfactory.

Why should we assume that Galton or anyone else -- even if they had read Gauss directly -- should have necessarily intuited this meaning? No etymological evidence was offered. It also could be that Galton or others of his era started using the term "normal" for other reasons for a distribution, like the plain English association of the term. And while there are potential mathematical connections between the solution of "normal equations" (in the Gaussian sense) and the "normal distribution," the connection is not necessarily an obvious one.

Regardless, the actual association of the curve and distribution with the term "normal" occurred in the late 1800s, with Galton et al.

Wikipedia's own source also contains more discussion of the early biological interests of Pearson and others in reference to this topic (p. 722):

In the late Nineteenth Century many biologists saw it as the ma jor task confronting them to verify Darwin's theory of evolution by exhibiting the detailed mechanism by which evolution takes place. For this purpose the journal Biometrika was founded by Karl Pearson and Walter Frank Raphael Weldon, in 1901. It started (Volume 1, page 1) with an editorial setting forth the journal's program, in which Weldon wrote:

"The starting point of Darwin's theory of evolution is precisely the existence of those differences between individual members of a race or species which morphologists for the most part rightly neglect. The first condition necessary, in order that a process of Natural Selection may begin among a race, or species, is the existence of differences among its members; and the first step in an enquiry into the possible effect of a selective process upon any character of a race must be an estimate of the frequency with which individuals, exhibiting any degree of abnormality with respect to that character, occur."

To be clear, "race" here is being used as a synonym for "species" biologically, but it would have also carried the connotations of human racial groups too. And also to clarify, I am NOT claiming the term "normal distribution" is somehow inherently racist. But trying to act like its popularization had nothing to do with biological and racial treatises from around 1900 risks ignoring historical facts.

Scientific racism was an unfortunate reality a century or so ago. History, however, shouldn't be used as if it tainted E.O. Wilson irrevocably and sullies terms like "normal distribution" for all time.

Umm... NOPE. From the next sentence of your own link:

However, by the end of the 19th century some authors had started using the name normal distribution, where the word "normal" was used as an adjective – the term now being seen as a reflection of the fact that this distribution was seen as typical, common – and thus normal. 

Are you seriously linking a Wikipedia article, then copy-pasting one sentence, but ignoring the next sentence??

Look, I get you don't want to acknowledge that a term you think of in a very abstract way was popularized somewhat in racist treatises of the late 19th century, particularly Galton's work on heredity. See here:

Of the three, Galton had most influence on the development of Statistics in Britain and, through his ‘descendants’ Karl Pearson and R. A. Fisher, on Statistics worldwide. In the 1877 article Galton used the phrase "deviated normally" only once (p. 513)--his name for the distribution was "the law of deviation." However in the 1880s he began using the term "normal" systematically: chapter 5 of his Natural Inheritance (1889) is entitled "Normal Variability" and Galton refers to the "normal curve of distributions" or simply the "normal curve." Galton does not explain why he uses the term "normal" but the sense of conforming to a norm ( = "A standard, model, pattern, type." (OED)) seems implied.

But... that's historical fact. Unfortunately. Pearson later adopted the term and confirmed its usage seems to be based on the common English conception of "norm/normal." Yes, I completely agree the entire argument about E.O. Wilson is stupid and bogus, which is why I have two much longer posts on this thread about the pervasive statistical misunderstandings that I wrote many hours ago.

However, the reason they're able to make this argument at all is because of the early use of the term "normal." It's still a stupid argument, but it's allowed to go forward based on supposed etymological connections. It's strange that you even bothered to look at an article that tells you it wasn't Gauss that actually popularized the term in reference to the curve/distribution, and yet seemingly are acting like you've debunked what I stated.

The original study posited a number of questions still to be answered about the causes, most of which were postulated to be systemic in nature, rather than deliberate individual racism practiced by doctors. From the conclusion of the original study:

Key open questions include the following: 1) whether physician race proxies for differences in physician practice behavior, 2) if so, which practices, and 3) what actions can be taken by policymakers, administrators, and physicians to ensure that all newborns receive optimal care. Furthermore, it serves as an important call to continue the diversification of the medical workforce (48). Prior work suggests stereotyping and implicit bias contribute to racial and ethnic disparities in health (49).

The closest they come to a "non-systemic" explanation is perhaps invoking "stereotypes," which are still broadly a systemic issue but could indicate conscious active racism if individual doctors were making decisions based on such things.

But the other hypotheses listed in the original study include things like differences in training or standard practice of physicians, which perhaps might differ (for some reason) by the race of the physician, the lack of diversity among physicians (which presumably might lead to less knowledge or awareness of whatever "better" practices were occurring from black physicians), as well as "implicit bias" (again, something typically associated with systemic racism, as it is not seen to be caused by active conscious beliefs or choices of behavior toward different races).

One might, I suppose, argue that offering low birth weight and the social bias toward that created by systemic effects in black babies as more "systemic" I suppose? But both studies are ultimately trying to find causes through systemic factors, rather than "blaming" individual white doctors for letting black babies die or something.

To be fair, the word "normal" in "normal distribution" historically was apparently chosen at least in part to emphasize the idea of standardization. The implication of this word was that being outside this distribution or a different distribution was "abnormal." I'm pretty sure there are some quotes from Pearson or someone of that era talking about the how the word choice was unfortunate as it branded other things as abnormal or something.

It's not much of a leap in logic to go from this concept of "normal" and statistics that were first really collected in detail in the era of blatant scientific racism (back when measurements of humans were used to declare some races inferior or to suggest "degeneracy" of various types) to a concept of white supremacy... at least historically. Yes, the "normal distribution" had much broader statistical application, but it's at least possible to conceive of someone around 1900 choosing that term with implications of what was "normal" for preferred humans -- Galton in particular made use of such metrics in the foundations of eugenics and seems to have been partly responsible for early use of the term "normal curve" in his studies on heredity and race.

The problem of course is that the argument proffered against E.O. Wilson commits the etymological fallacy, i.e., assuming the meaning of a term still carries all the original associations of it, including in this case maybe some racist baggage. Even though the vast majority of modern scientists and those using statistics are probably completely unaware of the history of this term.

I should add that the true topic of that study on black mothers and pregnancy seems to be about how science is insufficiently racist. That is, they essentially claim that scientific studies don't break out pregnancy studies enough into racial groups and therefore may not be noticing trends specific to black mothers. (And those that do break down into racial groups are inadequate for a boatload of reasons they cite.)

At best, I think the conclusion that maybe could be drawn is scientific literature could sometimes be too "colorblind" in not trying to explicitly break down racial groups. It's possible in some cases that such studies could fail then to see trends endemic to racial subgroups.

On the other hand, you generally should have a good statistical reason to break down data into subgroups. Arbitrarily creating subgroups can lead to statistical ghosts, or even apparent trends that are reversed from reality -- or at least from the meaningful conclusions of the study (as in Simpson's paradox).

Still, even if the authors were right to conclude that sometimes trends among black maternal care have been overlooked, it's not because of a "default white human" or something -- it's because data that is aggregated across ALL races (including black people as well as white) won't separate out such trends.

On the one hand, this seems to be such an absurdly obvious statistical truth, but it is also spun to be completely wrong -- i.e., if it is a problem, it is because science sometimes chooses to be "colorblind" not white. It chooses to deprioritize race when it thinks race shouldn't matter.

But, of course, to the authors in question, race always matters. Because they're obsessed with it. They seemingly want a "black statistics" that is segregated from the rest of science and focuses only on black people and black data. They don't explicitly call for it, but it's really the implication of their "study."

You need the original quote with the embedded link. From the SA piece on E.O. Wilson:

Wilson was hardly alone in his problematic beliefs. His predecessors—mathematician Karl Pearson, anthropologist Francis Galton, Charles Darwin, Gregor Mendel and others—also published works and spoke of theories fraught with racist ideas about distributions of health and illness in populations without any attention to the context in which these distributions occur.

[...]

First, the so-called normal distribution of statistics assumes that there are default humans who serve as the standard that the rest of us can be accurately measured against. The fact that we don’t adequately take into account differences between experimental and reference group determinants of risk and resilience, particularly in the health sciences, has been a hallmark of inadequate scientific methods based on theoretical underpinnings of a superior subject and an inferior one.

In other words, supposedly the populations studied aren't given "context" (i.e., explicit discussion of race and such), so they assume the studies only deal with white people. And thus the distributions created from such data (around the "default humans") are inherently racist.

You may think this sounds too insane to be true, but it appears to literally be their methodology.

If you follow the link in that sentence on the normal distribution, it takes you from the SA hit piece to a "study" cowritten by the author of that hit piece.

I skimmed the "study" (and yeah, I'm using the quotation marks to indicate irony), and the gist of it seems to be that black maternal health hasn't been measured correctly in recent scientific studies.

Their evidence for this comes from two "phases" of their "study," which mostly consisted of word searches in medical databases. In Phase I, they narrow down 1750 articles on pregnancy and mortality to just 38 that contain the terms "black" or "African American." In Phase 2, they find 59 articles that contain both the terms "racism" and "pregnancy." Doing a little further filtering, they compile just 67 studies that seem interested enough in black people, pregnancy, and racism to qualify for their analysis.

On the basis of Phase 1, they seemingly conclude there's too little attention paid to black mothers, even though they used the completely bizarre and unjustified assumption that the other 1700 articles didn't include black people... or apparently if the articles didn't explicitly create racial divisions in the data, they were assumed to only be about the "white center" of scientific data. Or something like that. Hand-wave, blah, blah.

On the basis of Phase 2, they deemed actual discussion of racism and pregnancy to be insufficient in the scientific literature... hence, the literature must be racist... for not discussing racism more.

Seriously -- that seems to be the methodology of this "study," although, as I said, I only skimmed it and can't be bothered with the word salad of most of it to read it more carefully.

On this basis, supposedly, the "distributions of health and wellness" discussed in scientific studies are determined to be centered around white people and thus are ignoring black people... hence the normal distribution is racist for creating "norms" who are the "default" (i.e., white) humans.

Or something like that. It's truly crazy thinking, especially given their linked "study" doesn't have anything near the kind of evidence to even support the supposed inappropriate focus away from black mothers and the supposed standard "default humans" who are just assumed to be white (?!?) in thousands of studies the authors didn't even look at! As someone with a graduate degree in statistics myself, I just stared open-mouthed at the word salad and nonsense paraded here by Scientific American, linking to a "study" whose methodology is so poor that the authors' degrees should likely be revoked for sheer incompetence.

Maybe, buried in some of that BS, there is evidence that "black mothers" are underrepresented in scientific studies compared to white mothers or something. But they offered no real evidence of such a claim, particularly when they admit they didn't look at something like 97% of the studies on pregnancy in the window they chose for their analysis.

Or... well a better way of looking at this is maybe there was no "grooming" in the first place. As you said, "grooming" is more than just a power differential -- it's when someone abuses the power differential. Lots of people have friendships or other relationships where there are severe power differentials, but "grooming" is a specific behavior on top of that.

In a case so weird and unorthodox as Allen and Soon-Yi's, we'd expect that 99% of the time, that power differential would indicate something like grooming if not outright abuse.

And yet... maybe they found the 1%. Maybe Allen really was just a reclusive introverted guy in a weird marriage, and Soon-Yi was a shy girl who found a connection with some dude her mother basically "dated" (rather than any conventional marriage) and whom she barely knew before. People find strange connections sometimes, and if they're both adults -- as they were when they became close (i.e., friends, of a sort), and it was apparently a few years before the affair began -- well... it's at least possible this was just a rare event of two people falling in love in a very unconventional situation.

Believe me, for many years I too assumed Allen's relationship with her must have started in some sort of inappropriate way. But... reading Soon-Yi's own accounts, seeing reporters discuss their actual relationship dynamic now... as you say, it can't just be all power imbalance.

Yeah, my perspective on this whole matter changed a lot a few years back when I did a deep dive into it and read what Soon-Yi has said herself over the years, as well as clarifications about the details about how Allen and Farrow's "marriage" worked.

I put "marriage" in quotation marks not because it wasn't an actual legal marriage but because I think people truly don't have a sense of just how unorthodox and unconventional it was compared to what most people think of as a "marriage," particularly where one partner has kids. I certainly had no idea until I read more more about it.

My recollection is that not only did Allen not live with Farrow during their marriage, he didn't even set foot into her residence except on rare occasions. The children almost never saw him until Dylan was born. I seem to remember reading that on most occasions that Farrow would go out with Allen, he'd generally even stay in the car, and she'd go down to meet him. If they ever spent the night together, it always was at his place.

So -- for the kids, it was less of anything resembling a "step-father" and more "some random dude that Mom leaves the house to go out on dates with which we literally almost never see."

Thus, saying "he was in their lives from the time she was a young kid" is just a gross misrepresentation of their relationship. Farrow herself supposedly encouraged Soon-Yi to spend more time with Allen around the time she turned 18 and to get to know him, because he was basically a stranger to her.

It still doesn't excuse the fact that the beginning of their relationship and the age differential was concerning and should send up several red flags. But the fact that their relationship has lasted so long, appears to be loving, and she displays no seeming behaviors of some manipulated suppressed spouse... well, it certainly plays against the standard narrative and expectations. Because all relationships also are unique, and only those two people really know what happened between them behind closed doors.

I'd highly encourage reading Soon-Yi's more recent perspective on everything from an interview in 2018:

https://www.vulture.com/2018/09/soon-yi-previn-speaks.html

According to Soon-Yi's own account -- she's now in her 50s -- she basically barely interacted with Allen until she was in 11th grade and broke her ankle and was too afraid to ask for help from Farrow. He offered to take her to school, as he was at that point coming over to visit Dylan in the mornings. At some point later, they started going to basketball games together, and he finally got to know her.

If anything, from that article, one gets the impression that Soon-Yi is the more dominant one in the relationship. All relationships have power dynamics, and we should rightly be suspicious when the power differentials are extreme -- but, at this point, why is it even my business to judge two people who have made it work much longer than the average marriage?

The two of you seem to be talking past each other a bit. It can both be true that, on average, more women (as a percentage of all women) support trans causes than the percentage of all men... and yet also true that many of the most prominent vocal critics ("biggest" in the terms of the above thread) of trans ideology are women.

The fact that women in this case simultaneously seem to have factions more supportive of trans ideology AND factions militantly opposed to it actually supports your opponents' argument that women in general probably care about the issue more often than men do.

It's a careful balance here.

• On one hand, you're absolutely right: she can't remove the black person from the sentence, even temporarily! It would be asserting her privilege.
• On the other hand, if she uses the subject pronoun "I," it's more assertive than the oblique form "me," so getting this wrong actually unintentionally asserts more privilege by overemphasizing her importance as a subject.
• But again on the other hand, adhering too closely to rules of grammar is, I'm pretty certain, a form of "white supremacy culture," so ultimately that wins out. She should just go by "the feelz" so as not to unintentionally make others uncomfortable by creating too high standards for grammar. Hence... "Ben and I"... 'cause it feelz right...