Ciceronian Review

Just Say No Nuance

The NYT has an opinion column on the sorts of information consumer genetic testing provides people, focused this time on race.  The point seems to be, according to the title writer at  least, that White Americans will learn that they have African ancestry - the brutal history of racial slavery: "America's Brutal Racial History Is Written All Over Our Genes."  The essay has almost the same point.  Anecdotes about African-Americans meeting the descendants of the slave owners, and some statistics about the prevalence of African genes in White southerners: "About 12 percent of whites from Southern states like South Carolina and Louisiana had 1 percent or more of African ancestry."  And more generally, "close to 4 percent of the company’s customers who identified as white Americans had at least 1 percent African ancestry, consistent with an African ancestor within the last 11 generations or so" (of clients of 23andme).  Written all over our genes is a bit of an overstatement given the numbers provided.  The history is in some genes, but not all over our genes.  There are all those European immigrants who came in the latter 19th century and in the 20th century.  And those immigrating from Central and South America, and from China and Japan, and so on.  This story, and its like, seem framed by an assumption that all that matters involves people east of the Mississippi before the Civil War and their descendants.  To be fair, there are some glancing mentions of the degree of population movements over the centuries, and at least allusion to the shaky foundation of claims about German or Italian written in genes.  I wish there had been more on both.     

Earth Colors

The NYT had a story yesterday about race and sex diversity in the Earth Sciences.  The title: "Earth Sciences Has A Whiteness Problem" with a subheading: "Barely 10 percent of doctoral degrees in the geosciences go to recipients of color. The lack of diversity limits the quality of research, many scientists say."  I thought the last sentence especially interesting.  How would a different demography improve the science or the quality of the science?  Assuming Earth Sciences includes climatology subjects, I thought it might be that more attention to local history would be informative about climate patterns, for example.  I read the article because diversity affecting the quality fo the science would be interesting and not something I would come to on my own.  Actually, I read it twice because there was, literally, nothing in the article supporting the notion that the 'whiteness problem' had or was affecting the quality of the science.  Thinking I might be particularly dense on the topic or blind to it in some way I also read through all 247 comments to see if someone had identified what I missed.  No.  There was nothing about the quality of the science.  The closest was the claim that "earth science classes could be enriched by a greater focus on nonwhite and Indigenous histories and voices, given that 'Indigenous people have a unique connection to the land.'”  It is a surprising claim.  I do not think it has much to do with Earth Science really, and certainly not a showing that the quality of the science is affected.  It is an odd statement on its own.  Indigenous people have a unique connection to the land is pretty far from anything like science.  It is not remotely credible that Indigenous people have a unique connection to the land because it is not remotely credible that Indigenous people all have the same connection to the land, or any reasonably identifiable set of connections to the land.  Even as to people doing the same things with respect to the land, like farming, do not have the same set of connections to the land. Indigenous people does not mean much in this context.  No migrations, apparently, until the White people showed up?  No discussion of class in the article, except the suggestion that all White people are wealthy.  

What is going on at the NYT? 

Judge Rakoff has another provocative essay in the NYRB, discussing forensic science in the US courts.  Or should one write "forensic science"?  It covers, briefly, the absurdity of the Mayfield case, making clear just how thin the grounds for his detention were.  It was more than confirmation bias at work.  Or rather a good deal less.  But the overall topic, the treatment of science in the US courts is large and serious topic.  There is greater focus on the use in criminal cases.  It creates a more stark moral picture to have people (men, mostly) off to prison for decades based in large part on ridiculous or dishonest presentations in court.  There is no science of blood spatter or science of bitemarks, TV notwithstanding.  The misuse of evidence like DNA or fingerprints is not of the same order, but it is certain that the discussions by witnesses are not just false, but structured in ways that deeply corrupt, by which I mean more or less intentionally obfuscatory, inaccurate, and misleading.  The story is no better in civil cases.  Experts are vetted to give the right outcome as part of deciding to hire, come to court with the predicted testimony and rely on methods that are in no way science.  It is just that money is what changes hands, not freedom.  

There is plenty of blame to share.  Lawyers do what they are expected to do, paid to do, and what the rules of professional conduct direct them to do -- look to the interests of their clients.  That is the procedural structure.  It is useless to ask lawyers to act in a different way.  Still, that does not get one out from under the tent.  Lawyers can, because I know lawyers who do, try for experts who have a reliable methods, are part of properly scientific or expert social structure (in other words, are actively involved in the relevant culture and practice).  On the criminal side, the economic constraints alone make the whole conversation painful.  Criminal defense just does not have the resources to mount challenges to prosecution experts, let alone to present experts of their.  The whole process of the criminal law system is, well, dismal to contemplate.  

One link in the system could be improved: the judges.  They could start learning some of the relevant statistics and science, they could have schools on social sciences.  I do not think it will happen any time soon, but these things are possible.  And necessary, because it is judges who decide what is admitted to trial.  Far too many judges have no understanding of the materials from experts they admit.  Rakoff has good discussions of these issues.  There is one additional point I would make here -- another problem is that the courts create fake facts.  Some analysis is admitted, the judge says it is good enough, and then that is set forever.  Two examples, one has a little light and the other not.  Shaken baby syndrome and fingerprints.  For some years testimony that an infant was killed or injured because it was shaken was taken as established based on inferences from spinal and brain injuries.  But over time the medical science changed  (or more accurately, some science developed), establishing that the inference just was not substantiated, was not reliable.  Or recovered memories.  And some courts stopped accepting the testimony.  On the dark side, fingerprint testimony is accepted everywhere even though it lacks a reliable scientific foundation.  The probabilities are made up, the standards for matching are no more than "connoisseurship". 

Courts should engage their own experts more.  The standards for admissibility ought to be tightened.  I should be the next Doge.

Keeping an Eye on You

Astronomy Picture of the Day.

 

No To Local Control

There was a remarkable press release published Tuesday, noted I think by a commentator below.  The Pew Trust published the results of a survey showing a rather large portion of the American public believes that creationism should be taught in public schools (or any schools).  It is a might depressing.  Deepened by the fact that about 40% beleive in creationsim.  The gloom is enhanced by this other story from yesterday's news, showing that a remarkably large number of people -- a majority in fact -- don't understand the relevant concepts.