New APPS: Art, Politics, Philosophy, Science

No, Really: Justice Roberts cares about law

UPDATE: With a note on the Roberts concurrence at the end

Justice Roberts sided with the Court's liberals today in a (somewhat surprising, and really important) 5-4 decision by Justice Breyer striking down a Texas abortion law nearly identical to one the Court struck down in 2016.  Justice Roberts is not a fan of abortion.  But he is a fan of the law!  Today, he basically joined in a smackdown of the ultra-conservative 5^th Circuit, which forgot that it was an appellate court, and not a trial court.  Justice Breyer writes:

“The Court of Appeals agreed with the District Court’s interpretation of the standards we have said apply to regulations  on  abortion.    It  thought,  however,  that  the  District  Court  was  mistaken  on  the  facts.    We  disagree.   We  have  examined the extensive record carefully and conclude that it supports the District Court’s findings of fact.  Those findings mirror those made in Whole Woman’s Health [the 2016 case] in every relevant  respect  and  require  the  same  result.  We  consequently hold that the Louisiana statute is unconstitutional.” (op. slip, 3)

Yes, but before we get there, let’s remember that the job of the Court of Appeals is not actually to disagree about facts when it doesn’t like a precedent.  Breyer begins with a review session on what Court is supposed to do what:

Title VII Protects LGBTQ Individuals!

UPDATE: First, I'm being loose with terminology here - "originalism" specifically refers to a theory of Constitutional interpretation; what Gorsuch et al are advocating is "textualism" (for statutory construction).  The distinction between public meaning and expected application is important in the originalism debate - but I think it's clearly at work in the debate here.  Second, Andrew Koppelman's commentary here is worth reading on the different ways of understanding the plain meaning of a text.

Up until today, if your boss wanted to fire you for begin gay or trans, you had no recourse in federal law – so unless your state happened to protect you, you had no recourse at all.  Today, the Supreme Court issued a landmark ruling in Bostock v. Clayton County that not only hands the LBGTQ community a huge, huge set of legal protections, it does so in a straightforward way.  And the opinion was authored by Justice Gorsuch and joined by Roberts.

Gorsuch delivers a seminar on originalism.  Does Title VII prohibit discrimination against trans and homosexual individuals? Well, let’s see what it says.  It says that it prohibits discrimination because of sex.  What does “sex” mean here?  Well, according to the public meaning (more on this in a sec.) of the term when the law was enacted, it means “status as either male or female [as] determined by reproductive biology” (op. slip, 5).  Ok.  You can’t discriminate “because of sex.”  What does “because of” mean?  It establishes a but-for causality.  Gorsuch goes on to explain that an event can have multiple but-for causes, which means that sex need only be a necessary part of the decision.  As he puts it, “when it comes to Title VII, the adoption of the traditional but-for causation standard means a defendant cannot avoid liability just by citing some other factor that contributed to its challenged employment decision. So long as the plaintiff’s sex was one but-for cause of that decision, that is enough to trigger the law.” (6).  Gorsuch notes that this is sweeping, and “Congress could have taken a more parsimonious approach” (6).  But they didn’t.  He proceeds to similarly characterize discrimination.

On the intersection of masks and racist policing

Black men have to decide whether the risk of being harassed and profiled by police for wearing a mask is greater than the risk of contracting Covid for not wearing one.

Covid-19 Came to the U.S. later than we thought

Research into the spread of Covid-19 continues, with an important new preprint by Michael Woroby et al up today (tl;dr see the writeup in Stat News).  The standard narrative about the arrival of Covid-19 in the U.S. is that a patient arrived in Seattle, WA from Wuhan on January 15^th.  He felt ill, and aware of CDC messaging about a new virus circulating in Hubei, sought medical help.  On Jan. 19, he became the first confirmed case of Covid in the U.S. Then, a few weeks later on Feb. 24, another patient turned up with what appeared to be community spread.  Containment had failed.  Per Woroby:

“On February 29th, 2020, a SARS-CoV-2 genome was reported from a second Washington State patient, ‘WA2’, whose virus had been sampled on February 24th as part of a community surveillance study of respiratory viruses. The report’s authors calculated a high probability that WA2 was a direct descendent of WA1 [the first patient, from January], coming to the surprising conclusion that there had by that point already been six weeks of cryptic circulation of the virus in Washington State. The finding, described in a lengthy Twitter thread on February 29th, fundamentally altered the picture of the SARS-CoV-2 situation in the US, and seemed to show how the power of genomic epidemiology could be harnessed to uncover hidden epidemic dynamics and inform policy making in real time” (3, internal citations omitted). 

Then there was more genetic sequencing, and it turned up a strange anomaly: the cases that appeared to stem from WA2 had mutated from WA1 in two places, and there was neither evidence of a transitional virus strain between them or of other infections in Washington whose genetic sequencing matched WA1.  Woroby et al then ran a computer simulation – 1000 times! – of the epidemic, seeding it with WA1.  The result was a surprise: “when we seeded the Washington outbreak simulations with WA1 on January 15th, 2020, we failed to observe a single simulated epidemic that has the characteristics of the real phylogeny” (7).  In other words, the narrative about the early spread in Washington is almost certainly wrong.  Patient WA1 was not the source, and the virus was not spreading covertly for weeks before emerging again in WA2.  Rather, the spread around WA1 was contained, and WA2 represented a new infection, and was either the source of the actual epidemic, or near it.  Specifically, the virus arrived a second time from Hubei, around Feb. 13 (95% probability between 2/7 and 2/19) (9).

Testing Nonsense Redux

As most folks know by now, there's two kinds of Covid tests.  One of them tests for whether you have the disease now.  The other tests for the presence of antibodies in your blood, indicating that you have had the disease at some point.  You might think that an outfit like, say, the CDC, would distinguish between them in its count of Covid-19 testing.  Alas, you'd be wrong, as The Atlantic reported today.  Why would the CDC (and Georgia, Texas, and some other states) do this?  Well, it lets you (a) claim to be doing a lot more testing for current infections than you currently are (because the addition of antibody tests to the viral tests drives up the total number of "tests"), and (b) will drive down the percentage of positive test results (because a small number of people have the antibodies, generally a much smaller number than test positive in most places in the U.S., especially those that conduct relatively few viral tests).

And if you do massive numbers of tests with low percent positive, then those annoying scientists say you can reopen sooner!  It puts you on the good green side of charts like the one above (source), not the bad yellow one.

Trump has been in office too long not to assume that this is the work of Trumpian political operatives and not CDC scientists, at least until proven otherwise.  Trump himself is obviously too stupid and incurious to know the difference between the testing types, but like Henry II, he makes it clear enough what he would like to see happen.  And that's lots of tests, lots of opening, and lots of loyal servants who translate his impulses into "reality."

More on Covid Transmission

One of the widely-discussed metrics for understanding Covid-19 transmission is its R0 number - the average number of infections that a given person causes.  An R0 of 3, for example, means that each person infects an average of 3 others.  In order to stop the disease from eventually spiraling out of control you need to keep the R0 number at or below 1.  But of course R0 represents an average - and a recent write-up in Science elaborates on the emerging evidence that a relatively small number of cases account for most of the spread.  Understanding why that is could of course go a long way toward making it possible to restore some sense of normalcy, even if figuring this out is incredibly difficult.

One thing that does appear to be the case is that indoor transmission is more likely than outdoor, and that mixing large numbers of people together in small spaces is a bad idea.  That's why Michael Sorrell, president of Paul Quinn College, writes in the CHE that the desire to have in-person college in the fall is nonetheless a bad idea.

On the other hand, a group of German medical organizations just issued a statement advocating the immediate and full return of K-12 schools.  There is some nuance - older kids can do some social distancing, and there is an assumption of testing and contact-tracing - but they summarize quite a bit of research saying that there is little evidence that outbreaks occur in groups of kids, and some evidence that kids don't readily transmit the disease.  We are all about to owe France and Germany, which are both competent at testing and are gradually reopening schools, a debt for better understanding how Covid transmits among kids.

In the meantime, the dishes here keep piling up and the racist sociopath pretending to be President is deporting children who aren't sick cuz the #MexicanCaravanWuhanOBAMAChinaWTOFAKENEWS is obviously the reason 93,000+ Americans have died!  But that's ok, because we have done lots of tests, very good tests, and "per capita" is one of those phrases that libtards use to confuse you with their elitism.

On understanding what "Covid cases" means

With reference to malaria tracking, I've tried to suggest some of the reasons we don't really know what "Covid cases" means, either insofar as that is measured by positive tests (because we don't know how many more cases there are beyond the tested ones, so tested cases is at best an rough guide to Covid incidence) or symptoms (a task that gets harder by the day, as Covid continues to present in new and strange ways).  Here's a nice piece that starts with how they deal with malaria incidence, and works through a lot of the problems in translating from test results to meaningful information about Covid.

Getting a more nuanced understanding of Covid Transmission

It’s fairly clear that one of the keys to living with Covid-19 is understanding the dynamics of transmission: absent something more nuanced than what we have, “stay 6 feet away from everyone at all times!” becomes the only public health advice that can be given.  Getting past the initial maximin strategy requires better data on everything.  There’s an interesting Twitter thread by Dr. Muge Cevik (St. Andrews) that collates some of the recent studies about transmission within and through social groups.

 

A lot of discussion recently about transmission dynamics, most of which are extrapolated from viral loads & estimates. What does contact tracing/community testing data tell us about actual probability of #COVID19 transmission(infection rate), high risk environments/age?
[thread]

— Dr Muge Cevik (@mugecevik) May 4, 2020

It’s not a formal meta-analysis or anything like that but there’s emerging confirmation for the idea that most transmission depends on sustained, close contact, and that children do not appear to be a primary vector for the disease (this latter one is enormously important, because nobody knows for sure what should be done about schools).  She summarizes at the end (the paragraph breaks are mine):

“While the infectious inoculum required for infection is unknown, these studies indicate that close & prolonged contact is required for #COVID19 transmission. The risk is highest in enclosed environments; household, long-term care facilities and public transport. High infection rates seen in household, friend & family gatherings, transport suggest that closed contacts in congregation is likely the key driver of productive transmission. Casual, short interactions are not the main driver of the epidemic though keep social distancing! Increased rates of infection seen in enclosed & connected environments is in keeping with high infection rates seen in megacities, deprived areas, shelters. A recent preprint demonstrates that #COVID19 epidemic intensity is strongly shaped by crowding.

“Although limited, these studies so far indicate that susceptibility to infection increases with age (highest >60y) and growing evidence suggests children are less susceptible, are infrequently responsible for household transmission, are not the main drivers of this epidemic.”

“Finally, these studies indicate that most transmission is caused by close contact with a symptomatic case, highest risk within first 5d of symptoms. To note: this preprint suggests that most infections are not asymptomatic during infection.  In conclusion, contact tracing data is crucial to understand real transmission dynamics. Cautionary note: This data & interpretation is based on the available evidence as of May 4th. Our understanding might change based on community testing/lifting lockdown measures.”

Meat is Murder

As Daniele Lorenzini reminds us, the coronavirus pandemic exposes nothing if not the differential precarity of our biopolitics.  Sure, biopolitics is about promoting life, but it’s also about deciding that some people can die in order that others may live.  The most obvious candidates are “essential”  workers in various parts of the supply chain who form what Marx calls the "disposable industrial reserve army" of capitalism.

In case there was any doubt about this, Trump is apparently about to invoke the Defense Production Act to force meat packing and processing plants to stay open.  This is the same Donald Trump who has routinely failed to use the same DPA to get medical supplies to hospitals and caregivers. 

But what about when it’s a question of protecting a bunch of mostly Latinx folks with a median annual wage of under $28,000 and whose working conditions look like this:

“The demographics in the meatpacking industry may be multicultural in the 21st century, but workers share many of the same characteristics. They are uneducated and work in extremely unsafe work conditions. Nearly every worker in meatpacking plants has injuries. Cutting stations are located close to each other, automated lines move too quickly for workers to keep up, and workers must put in long shifts or fear losing their jobs. Men and women must wear goggles, hardhats, stainless steel mesh gloves, rubber aprons and chaps. Still, every worker from the line cutter to the cleaning crew is in danger of suffering amputations, body part crushing, burns, punctures and other traumatic injuries. Though the demographics have opened up to include all segments of society, poor safety conditions in the plants continue to plague workers. According to Human Rights Watch, the poor labor market has caused safety concerns to backslide in the United States. Collective bargaining eroded, and by 2005, the injury rate was twice as high in the meatpacking industry as any other industry in the country.”

Trump wants you to know: sustaining Americans’ egregious overconsumption of cheap meat, which involves the brutal exploitation of vulnerable workers in the best of times, is more important than PPE’s for healthcare workers!  Oh, wait, Trump says (per the Bloomberg piece) that “the government will provide additional protective gear for employees as well as guidance.”

I wonder if that will arrive before, or after, the additional protective gear for hospitals?  But let no one accuse Trump of inconsistency!  He’s using the DPA to ensure that vulnerable workers, many or even most non-white and/or female, are at high risk for Covid.

Who Owns the Law? SCOTUS Speaks

Recall that before Covid (so about 300 years ago), there was an interesting copyright case percolating through the federal courts.  The question concerned the Official Georgia Code Annotated (OGCA), which contains the text of the Georgia Code as well as various annotations.  There were two potentially conflicting principles at work.  On the one hand, the law is public domain.  On the other hand, annotations and supplemental materials by third parties are often copyrightable.  Georgia managed to produce a hybrid system: the legislature established a code commission, which outsourced most of the annotations work to Matthew Bender Corp, which was granted an exclusive license to sell it.  At the same time, the legislature every year officially adopted the GCA, and it was the authoritative source for the Georgia Code in everything from legislative proceedings to cases to public reference.  It was even published with the state seal attached.

The 11^th Circuit ruled that OGCA was not copyrightable because, even if it wasn’t quite the same thing as the statutory text, it nonetheless is “an exercise of sovereign power” (3) and “sufficiently law-like so as to be properly regarded as a sovereign work” (4).  Today, in an opinion by Justice Roberts, the Supreme Court agreed, though for somewhat different reasons.  The SCOTUS opinion basically argues that the relevant question is whether the “author” of something is a judge or legislator; answers that the OGCA is reasonably the work of a legislator, and thus uncopyrightable.

Discouraging News (probably) on Remdesivir

Remdesivir has been one of the most closely watched potential therapies for COVID-19, and a couple of early cohort and observational studies have been encouraging.  But apparently the first results from a randomized controlled trial in China indicate that it did not make a statistically significant difference.  At least, it looks that way: results were accidentally posted on the WHO website and then removed, with a promise of an official publication to come later.  It's all a bit murky, and there's a number of other trials ongoing, but this is not the sort of result you hope to see, even accidentally.

What does it mean to model “social distancing?”

By now it should be apparent just how little we know about the coronavirus pandemic, from how to treat it to basic facts about what the “number of COVID cases” means. Even “deaths due to COVID” turns out to be difficult: both New York and UK have revised their numbers up to accommodate likely cases that hadn’t been counted, and there’s a brewing political battle over how to count them. This ignorance also affects pandemic modeling; the aspect I want to look at here is over what “social distancing” means.

COVID Models are for guiding action: we want to look at them and know when we can leave the house again (or, maybe more importantly, when the kids can leave the house again!). In that sense, models are inherently political, in several ways. First, the model doesn’t tell you what to do; the decision about what to do can be informed by the model, but it requires an entire apparatus of priorities, intuitions, and whatever else goes into a decision about what to do. Part of this is a second-order decision strategy about how to process uncertainty. I’ve suggested before that we’ve been operating on maximin; whether I’m right isn’t important here except to underline that the model is insufficient. Second, the construction of the model is going to embody a number of social judgments. I’m not being Latour here and arguing that whatever comes out of laboratories or complex instruments is broadly political (though he’s right). What I do want to say is that to use a model, you have to answer some antecedent questions what your use of the model is trying to achieve, and how society operates.

U.S. COVID Testing: Still Crap

President Nero wants you to know that the U.S. has conducted a bigly number of coronavirus tests, higher than he can count, and maybe even more tests than there were people at his inauguration!  Anyway, the U.S. is still terrible at COVID testing, as the following chart from Vox reminds us:

As the accompanying article points out, this is completely mediocre.  It is a little better than before, but it nonetheless underscores that we are not ready to go back to a new-normal.  All the actual plans I’ve seen begin with the presumption of widespread testing.  The U.S. is absolutely nowhere close to that.

Cellphone Tracking for COVID Surveillance

Cellphone tracking - whether through geolocation or something like detecting the proximity of bluetooth devices - has been getting a lot of attention for its potential to improve COVID surveillance.  Given that there are estimates that a workforce of upwards of 100,000 people would be necessary to get a good contact-tracing regime going in the U.S., any automated way to reduce this would be welcome.  Apple and Google have announced that they are developing an App for that.  It would enable a user to know if they had been within six feet or so of someone who tested positive for COVID.  Assume for the moment that enough people get tested that such a strategy would generate meaningful data, what are its limitations?  Here are 5 concerns.  Not surprisingly, they echo concerns about other uses of cellphone tracking: not just privacy invasion, but whether they actually generate a lot of false positives (the person in the apartment next to yours might well be within 6 feet of you), whether they promote discrimination (neighborhood hotspots, etc.), and so on.

The EFF has some useful principles here.

Modeling the Future under Coronavirus

UPDATE: Here's a nice piece that talks about the complexities of reducing restrictions, framing the overall need in terms of keeping R0 (the number of new infections a given case leads to) from rising much above 1.

A new article in Science models our future under the new Coronavirus regime.  It is not pretty.  A few takeaways, followed by a thought on social distancing (the whole study is worth a read, because I’m cherry-picking here, and I’m not indicating anything about the modeling process, only some of the implications):

• Social distancing works, BUT when you stop it, COVID will come back. The relation between the two is complex, because effective social-distancing reduces population immunity: “We evaluated the impact of one-time social distancing efforts of varying effectiveness and duration on the peak and timing of the epidemic with and without seasonal forcing. When transmission was not subject to seasonal forcing, one-time social distancing measures reduced the epidemic peak size. Under all scenarios, there was a resurgence of infection when the simulated social distancing measures were lifted. However, longer and more stringent temporary social distancing did not always correlate with greater reductions in epidemic peak size. In the case of a 20-week period of social distancing with 60% reduction in R0, for example, the resurgence peak size was nearly the same as the peak size of the uncontrolled epidemic: the social distancing was so effective that virtually no population immunity was built. The greatest reductions in peak size come from social distancing intensity and duration that divide cases approximately equally between peaks
• You’d better hope COVID is not seasonal: “For simulations with seasonal forcing, the post-intervention resurgent peak could exceed the size of the unconstrained epidemic, both in terms of peak prevalence and in terms of total number infected. Strong social distancing maintained a high proportion of susceptible individuals in the population, leading to an intense epidemic when R0 rises in the late autumn and winter. None of the one-time interventions was effective in maintaining the prevalence of critical cases below the critical care capacity.” And: “One-time social distancing efforts may push the SARS-CoV-2 epidemic peak into the autumn, potentially exacerbating the load on critical care resources if there is increased wintertime transmissibility”

 

Republicans still hate voting

Why would you let a pandemic get in the way of voter suppression?  Much better to use it as a tactic of voter suppression.  At least, I can't see any other other coherent way to read the Wisconsin GOP and SCOTUS refusal to either delay Wisconsin's vote today, or allow absentee ballots to be delayed.  Wisconsin GOPs shenanigans to keep itself in power are well-known; we shouldn't forget that the Roberts Court is systematically hostile to voting rights.  The Onion has the right tweet, and reminds us that Wisconsin has 77 Covid deaths today.  Check that space in 2-3 weeks.

Unless it's that Republicans hate democracy so much they don't even care who wins, as long as fewer people vote.

Plus ça change...

More on Late Capitalism and COVID

By Gordon Hull

I argued a few days ago that late capitalism, with its fetishization of efficiency, leaves us unprepared for a pandemic because of vulnerabilities in the supply chain. In a recent blogpost, Frank Pasquale adds some healthcare-specific texture to the point, noting how our healthcare system is almost designed to fail in a pandemic. He cites one of his own papers from 2014:

“The reduction in hospital facilities and other resources, although “efficient” in normal times, may prove disastrous if there is an epidemic. For example, one national-preparedness plan for pandemic flu estimated that, in a worst-case scenario, the United States would be short over 600,000 ventilators. “To some experts, the ventilator shortage is the most glaring example of the country’s lack of readiness for a pandemic,” one journalist noted. The lack of “surge capacity” throughout the health care industry is a major infrastructural shortcoming, likely to cause tremendous, avoidable suffering if a pandemic emerges” (179).

The quotes are from… 2006 and 2007, and refer to warnings coming after the SARS epidemic. In other words, we’ve been as unprepared as possible for 14 years, despite a near-miss epidemic and constant warnings from epidemiologists. So Trump is an idiot and an imposter, and his son-in-law supply czar is a feckless idiot who understands nothing about supply, but, as Pasquale underscores, there is another, longer timeline to our pandemic preparation failure.

In Pasquale’s paper, he notes that part of the problem is how we frame healthcare in terms of aggregate costs (and the need to keep costs down), a construction that makes it impossible to notice that some things are over-funded and others under-funded. In particular, not only can mantras of cost-cutting shield wasteful allocations of resources like those to hedge-fund managers from scrutiny, but it can also hide the fact that some aspects of healthcare (let’s see, hmm. Pandemic prevention!) are radically under-funded. Worse, there is absolutely no way to guarantee that money saved here will actually be reallocated to something more socially useful:

“If the health care cost-cutters had a plan for reallocating excess health sector spending to pay for care that is now undercompensated or absent, they would merit the influence they have now achieved. But in reality, money freed up by cost-cutting is much more likely to be retained as profit or claimed by capital and rentiers in some other way” (191)

We see at least three versions of these problems playing out now.

COVID and Geolocation: Google edition

Geolocation data is getting increasing attention as a way of tracking social distancing in particular.  Google has just released a bunch of its geolocation data, which tracks changes in trips to retail, parks and other places.

In the meantime, a new paper in Science says that a good contact-tracing App, if sufficiently robust and adequately deployed, could avoid the need for lock-downs. 

Of related interest, Zeynep Tufekci has a smart piece in The Atlantic, pointing out that disease modeling isn't useful so much for producing truth or knowledge, but as a guide for how to avoid worst outcomes.  This seems absolutely right to me, and is in line the way health policy folks are pursuing what I've called a maximin strategy.

Late Capitalism Makes COVID-19 Worse

By Gordon Hull

I know there’s a lot of ways to develop that thesis!  Let me focus on one: the fetishization of “efficiency,” and its corollary, just-in-time supply chains.  In a recent piece in The Atlantic, Helen Lewis argues that a lot of the disruption in consumer goods (toilet paper, etc.) is initially attributable not to hoarding, but tiny, unanticipated fluctuations in demand.  Stores don’t carry extra product in the back any more; rather, they keep their shelves stocked by way of a very elaborate, data-intensive logistical operation that delivers enough product to keep items on the shelf, but no more.  Hence the name “just-in-time” capitalism.  Why would you prefer such a system?  Excess stock is inefficient: it just sits there taking up space when it could be sold elsewhere, and you had to pay people to make it, even though you’re not getting paid for selling the product that’s sitting in the back of the store.

The innovations to supply-chain that enable just-in-time capitalism go well beyond the idea that stores should keep minimal stock, however.  If neoliberal financialization has taken a lot of the theoretical attention since 2008, it’s important to remember that financialization hasn’t been the only point of late capitalism.  For example, a pair of McKinsey reports in the early 2000s looked back on productivity growth in the late 1990s, widely attributed at the time to developments in IT.  As one argues, IT was not the cause: within the sectors that grew, “the most important cause of the productivity acceleration after 1995 was fundamental changes in the way companies deliver products and services.”  In “The Wal-Mart Effect,” McKinsey’s Bradford Johnson argues that:

“More than half of the productivity acceleration in the retailing of general merchandise can be explained by only two syllables: Wal-Mart. In 1987, Wal-Mart had a market share of just 9 percent but was 40 percent more productive than its competitors as measured by real sales per employee (the measure used for all company-level analyses in this study). A variety of Wal-Mart innovations, both large and small, are now industry standards. Wal-Mart created the large-scale, or “big-box,” format; “everyday low prices”; electronic data interchange (EDI) with suppliers; and the strategy of expanding around central distribution centers. These innovations allowed the company to pass its savings on to customers. By 1995, it commanded a market share of 27 percent and had widened its productivity edge to 48 percent” (McKinsey Quarterly 2002:1, p. 41).

In other words, our current retail scene – the one existing before Amazon – is attributable to Wal-Mart.  The efficacy of this logistical innovation was evident in the immediate aftermath of Hurricane Katrina, in which Wal-Mart was able to deploy trucks of aid to the New Orleans area much faster the George W. Bush’s mismanaged FEMA.   Amazon, in a sense, represents the intensification of this trend: maintaining brick-and-mortar stores is inefficient, if the logistical operation can be made sufficiently nimble and efficient.

Following Spring Breakers

As I noted earlier, cellphone tracking data is proving an interesting tool in trying to understand the spread of COVID-19.  You'll recall that Florida beaches featured a whole bunch of tightly-packed WOO-HOO SPRING BREAKers.  And that they then presumably left the state.  Where did they go?  Here's a map based on cellphone data.

Thanks Gov. DeSantis, who I never even had a chance to vote against!

The epistemology of COVID-19 – indicators edition

By Gordon Hull

A recent piece in the Guardian points to the limitations of models in understanding the spread of COVID-19: principally, those models are often not based on reliable data specific to the disease, so they can be wildly off.  As I argued earlier, the bottom line is that we really have no idea what we’re doing, and that lack of knowledge is an important reason why we’re having to adopt non-risk-based strategies like maximin to deal with the virus spread.  The decision to move to those strategies needs to be understood as a (bio)political one (just as the general preference for risk is); that they are not risk-based is probably part of why right-wingers are deciding it’s time to go “back to work.” It’s not that they have any better idea what’s going on than anybody else, they’ve just adopted a different political epistemology, and have decided that their idea of economics is more important than everybody else’s view of human life.

In the middle of any efforts at modeling or analysis are indicators.  These have been the subject of a fair amount of recent interest, as for example the recent (2015) anthology, The World of Indicators.  As the editors note in the book’s introduction, “indicators are typically presented as taken-for-granted facts. Yet, indicators are not neutral representations of the world, but novel epistemic objects of regulation, domination, experimentation and critique” (5).  More specifically, there is a fundamentally socio-political process at work behind any indicator.  We can all see the point with indicators like “women’s social equality,” since that concept means different things to different people.  So too, the point is clear enough in psychological categories like “normal” or “deviant” or “personality disorder,” as Foucault made clear.

Less obviously, it applies in more narrowly medical situations as well.  Consider malaria.  Malaria is a thing in that it’s a disease caused by a parasite.  Everyone knows that – it’s not an essentially-contested concept like “privacy.” And yet, as Rene Gerrets indicates in his contribution to the anthology, similar problems apply to malaria tracking in sub-Saharan Africa. 

Why We Are Not Bare Life: What's wrong with Agamben’s Thoughts on Coronavirus

By Gordon Hull

In a follow-up to a controversial piece in which he argued (in late February) that the social distancing and quarantining in Italy presented the temptation to universalize the state of exception, Agamben says this:

“Fear is a bad counsellor, but it makes us see many things we pretended not to see. The first thing the wave of panic that’s paralysed the country has clearly shown is that our society no longer believes in anything but naked life. It is evident that Italians are prepared to sacrifice practically everything – normal living conditions, social relations, work, even friendships and religious or political beliefs – to avoid the danger of falling ill. The naked life, and the fear of losing it, is not something that brings men and women together, but something that blinds and separates them. Other human beings, like those in the plague described by Manzoni, are now seen only as potential contaminators to be avoided at all costs or at least to keep at a distance of at least one metre. The dead – our dead – have no right to a funeral and it’s not clear what happens to the corpses of our loved ones. Our fellow humans have been erased and it’s odd that the Churches remain silent on this point. What will human relations become in a country that will be accustomed to living in this way for who knows how long? And what is a society with no other value other than survival?”

To be sure, there are and will be bad actors.  William Barr’s DOJ has apparently seen the epidemic as a good time for a power grab (notably, one that has been sharply critiqued from both the right and the left).  Ohio seems to be using it as a pretext to stop abortions.  So – and this will be my point – we must always be vigilant about the expansion of emergency powers.

But when you have a leitmotif, you stick with it, and so Agamben reiterates his basic thesis about the combination of state of exception and bare life:

Thoughts on COVID-19

By Gordon hull

The coronavirus outbreak showcases a lot of what is wrong with the Trump administration and the stupid, preening narcissist at its head (this is a president who claimed not to have known that flu kills thousands of people every year – but whose own grandfather was a victim of the 1918 pandemic).  At the Washington Post, Jennifer Rubin thinks that coronavirus is likely to bring down the entire GOP house of cards in November, as it relentlessly exposes the corrupt ineptitude of Trump; as goes Trump, so go his sycophantic enablers in the Senate.  This would be karmic.  It would also be no comfort to the many, many people that Donald Trump is about to kill.  Indeed, not just the actions of Trump’s egomaniacal incompetence but our collective need to think about it (guilty!) is itself part of why we aren’t prepared.

That is, people will die because the failure of the government to do anything meaningful towards containment in the early stages of the epidemic, in no small part because of Trump’s narcissistic dithering and lying.  Even now, as we’ve moved from “containment” to “mitigation,” nowhere near enough people are being tested.  This makes it much harder to know who is at risk of contracting the disease and spreading it to others, which in turn makes it more likely that they do so.  More people will get sick.  Not testing enough people also makes it impossible to know how deadly the disease is, because it denies you a sensible denominator for predicting the case fatality rate.  When Trump said he had a “hunch” that the WHO’s widely-circulated 3.4% fatality rate was too high, he was probably right (with lots of provisos) – but his own stonewalling on testing make it impossible to substantiate the point.  Trump will also kill people through his prior stupid policy decisions, like forcing asylum-seekers to wait in Mexico until their number is someday called, in makeshift camps that lack even the running water necessary for hand-washing.

A Deleuzian Undercurrent to Foucault’s “What is an Author?” (part 1)

By Gordon Hull

Toward the end of “What is an Author,” Foucault distinguishes between the “founder” and “initiator [instaurateur]” of a discourse.  Galileo is the paradigmatic example of the former, and Marx of the latter.  This is a puzzling distinction, to say the least.  Let’s begin with the terminology: Although “founder [fondateur]” is common enough, as far as I know, Foucault doesn’t use “instaurateur” anywhere else.  At least, a computer search of the text of Les Mots et Les Choses, Archéologie du Savoir and the pre-1975 Dits et Écrits didn’t turn up anything.  Other things being equal, those seem like the most likely places to find it (if I’m missing uses of the term, I’d love to learn about them!).  In particular, Order is a likely bet, because in the French seminar version (the one in D&E – see my initial thoughts here and Stuart Elden’s discussion of the textual history here) of “Author,” Foucault frames the text as partly responding to some leftover business from Order, where he admits that he both refuses to organize texts by authors, but also uses authorial names.  The nominal “instauration” occurs a few times in these texts in a way that something more substantial than a blog post would need to investigate, but as far as I can tell, the term of art in “Author” – “instauration discursive,” naming somebody rather than an event – is specific to that lecture.  So something is going on here!

It seems to me that it helps to understand this distinction by putting Foucault in conversation with Deleuze.  Specifically, it seems to me that the instaurateur is an application of what Deleuze calls difference or repetition outside the order of representation.  I’ll make an initial, obviously sketchy, case for that thought over the next few posts, with the caveat that I am not a Deleuze scholar.

Otto’s Phototagging

By Gordon Hull

Last time, I suggested that a recent paper by Mala Chatterjee and Jeanne Fromer is very helpful in disentangling what is at stake in Facebook’s critique of Illinois’ Biometric Information Privacy Act (BIPA). Recall that BIPA requires consent before collecting biometric identifiers, and a group of folks sued FB over phototagging. Among FB’s defenses is the claim that its software doesn’t depend on human facial features; rather it “learns for itself what distinguishes different faces and then improves itself based on its successes and failures, using unknown criteria that have yielded successful outputs in the past.” (In re Facebook Biometric Info. Privacy Litig., 2018 U.S. Dist. LEXIS 810448, p. 8). Chatterjee and Fromer apply the phenomenal/function distinction from philosophy of mind to the question of how mental state requirements in law apply to AI, with an extended case study of liability for copyright infringement. Basically, there’s an ambiguity buried in the mental state requirements, and we need to decide – probably on a case-by-case basis – whether the law’s objective is better served by a phenomenal or functional account of the mental state in question.

In applying the distinction, I suggested that we assume for the sake of argument that the software does not do the same thing that an embodied human being does when they identify a face. In other words, I was suggested that we accept arguendo that the software in question does not achieve the same phenomenal state as one of us does when we recognize a face. I also said I think that assumption, while clearly correct in a literal sense, may not be able to do as much work as it needs to. Here’s why.

It should be fairly clear that the experience of recognizing Pierre in a café is not identical between different people, or probably even the same person in different times. For that to be true, the molecular structure and electrical activity in their respective brains would have to be identical, which isn’t going to be the case. It’s also not clear that we don’t “learn[] for [ourselves] what distinguishes different faces and then improve[] [ourselves] based on [our] successes and failures, using unknown criteria that have yielded successful outputs in the past,” just like FB. After all, if you ask me why I recognize somebody, I will produce some criteria – but if it’s somebody I know, it’s not like I consciously apply that criteria as a rule. Neither the FB system nor I are using the old-fashioned “AI” of an ELIZA program. It would therefore at least require some argument to say that I recognize the face by means of that criteria, rather than offering it as a post facto explanation. Indeed, recognition does not appear to be a “conscious” process in the relevant sense at all. So that can’t be the issue.

Facebook, Biometric Privacy and the Philosophy of Mind

By Gordon Hull

Facial recognition technology is an upcoming privacy mess.  An early example of why is photo-tagging on Facebook.  The privacy problem was noted a while ago by Woody Hartzog and Frederic Stutzman: “once a photo is tagged with an identifier, such as a name or link to a profile, it becomes searchable …making information visible to search significantly erodes the protection of obscurity, and, consequently, threatens a user’s privacy” (47; on obscurity, recall here and here).  A while ago, I noted litigation surrounding Illinois’ Biometric Information Privacy Act (BIPA).  BIPA basically establishes notice and consent rules for companies that collect and use biometric information.  For example, it stipulates that “no private entity may collect, capture, purchase, receive through trade, or otherwise obtain a person's or a customer's biometric identifier or biometric information, unless it first” informs the person in question of what’s happening and what the entity is doing with the data, and then obtains a written release (740 ILCS 14/15(b)).  As I suggested, this regime is subject to the obvious problems with notice and consent privacy, but companies like Facebook are resisting providing even that de minimis protection for their customers.  In a landmark ruling last year, the Illinois Supreme Court upheld the law’s statutory damages provision.

More generally and as parallel federal litigation underscores, BIPA presents a significant threat to FB.  The issue in question is precisely photo-tagging.  As the 9^th Circuit described the process:

“In 2010, Facebook launched a feature called Tag Suggestions. If Tag Suggestions is enabled, Facebook may use facial-recognition technology to analyze whether the user’s Facebook friends are in photos uploaded by that user. When a photo is uploaded, the technology scans the photo and detects whether it contains images of faces. If so, the technology extracts the various geometric data points that make a face unique, such as the distance between the eyes, nose, and ears, to create a face signature or map. The technology then compares the face signature to faces in Facebook’s database of user face templates (i.e., face signatures that have already been matched to the user’s profiles). If there is a match between the face signature and the face template, Facebook may suggest tagging the person in the photo” (Patel v. Facebook, 6).

A group of representative Illinois FB users sued the company and filed for class certification,  arguing that “Facebook violated sections 15(a) and 15(b) of BIPA by collecting, using, and storing biometric identifiers (a “scan” of “face geometry,” id. 14/10) from their photos without obtaining a written release and without establishing a compliant retention schedule” (Patel v. FB, 7).

The Surveillance Gap and Neoliberal Necropolitics

By Gordon Hull

Consider the following, too brief summary: following Foucault, one can say that biopolitics is about optimizing populations, or something to that effect.  This involves a lot of work on the part of the administrative state, which sets itself up to provide services, everything from sewers and other infrastructure to social safety nets.  Different places do this differently, but the goal is to provide for the general welfare.  At the same time, as Foucault noted from the get-go (see the last lecture in Society must be Defended), the biopolitical “make live or allow to die” generated a correlative “if you want to live, they must die” (SMD 255).  Those Others were excluded from the “population” that the state tried to optimize and so were either allowed to die, or (in cases such as the Nazis) actively killed, or (in the case of American Jim Crow) actively suppressed and marginalized and often killed.  These various forms of necropolitics are intimately related as both a matter of historical fact (the Nazis thought Jim Crow an excellent example of race management) and conceptual structure under the rubric of something like state racism.  So too, the biopolitics of optimization is historically tied to the necropolitics of state racism, as scholars like Dell McWhorter make clear in the American case, or Agamben does in the case of the German Versuchpersonen.

But biopolitics has more than one variant, or at least so I’ve tried to argue.  I think it’s useful to view biopolitics in an earlier phase – what I call “public biopolitics” and a more recent neoliberal version.  The two can be distinguished in part by how they conceptualize the members of the population they are trying to benefit, and how they think they might do so.  For example, intellectual property on the public model is about public welfare and the benefits to everyone of encouraging inventive activity.  On the neoliberal version, the emphasis is much more on individual creators and markets. But the neoliberal version very much embeds a view of public welfare – a society of individual entrepreneurs and consumers, whose well-being is measured by various indicia of consumer welfare.

But what about the necropolitics of neoliberal biopolitics?  One obvious avenue to pursue is that those who are excluded from markets are allowed to die.  The rhetoric of consumer choice is often then utilized to suggest that they deserve their fate because of a failure to opt-in.  Whatever the merits of that avenue, it seems clear enough that more can be said.  One component of the U.S. piece is going to be data and surveillance.  Biopolitics itself arose in the context of data governance mechanisms like the census, and the importance of data to contemporary capitalism is large and growing.  One way the surveillance state implements necropolitics is by what Margaret Hu calls “big data blacklisting:” classifying certain kinds of people as ineligible for many of the things to which other members of the population are entitled (such as employment or the ability to board a commercial aircraft), or rendering them subject to ceremony-free death by drone strike.

In a fascinating recent paper, Michelle Gilman and Rebecca Green offer some important clues to another aspect of the story.  Most of the literature about data privacy focuses on what happens when people don’t have enough privacy.  Gilman and Green focus on the inverse problem: what happens when people have too much privacy, i.e., when they become invisible in a society predicated on visibility?  Working primarily with examples of the undocumented, the homeless, day laborers and those with felony convictions, Gilman and Green tell the story of “populations that remain outside seemingly omnipresent surveillance systems” (257), those who live in the “surveillance gap.”  If too much surveillance is bad, so is too little:

Review Essay of Privacy Books

In the shameless self-promotion dept., I have a new paper out - actually a review essay in Ethics & International Affairs (SSRN link here) of two recent books on privacy, Ari Ezra Waldman's Privacy as Trust and Jennifer Rothman's Right of Publicity.  Both books are well-worth the read!  The essay also pushes my thesis about the difficulties of assessing privacy on purely economic grounds.  Here's the abstract:

Most current work on privacy understands it according to an economic model: individuals trade personal information for access to desired services and websites. This sounds good in theory. In practice, it has meant that online access to almost anything requires handing over vast amounts of personal information to the service provider with little control over what happens to it next. The two books considered in this essay both work against that economic model. In Privacy as Trust, Ari Ezra Waldman argues for a new model of privacy that starts not with putatively autonomous individuals but with an awareness that managing information flows is part of how people create and navigate social boundaries with one another. Jennifer Rothman’s Right of Publicity confronts the explosive growth of publicity rights — the rights of individuals to control and profit from commercial use of their name and public image — and, in so doing, she exposes the poverty of treating information disclosure merely as a matter of economic calculation. Both books emphasize practical and doctrinal solutions to the problems they identify. In this essay, I take a step back and draw out the extent to which they converge on a fundamentally important point: the blunt application of market logic with its tools of property and contracts fails to protect the interests that lead us to turn to privacy in the first place; the tendency to economize privacy is a significant part of why we inadequately protect it.

The Normative Valences of Obscurity (part 2)

By Gordon Hull

Last time, I offered some thoughts on Woody Hartzog’s (and co-authors’) development of “obscurity” as a partial replacement for privacy.  On Hartzog’s account, privacy is subject to a number of problems, not least of which is that we tend to think in terms of an unsustainable binary: things are either “private” or “public,” which means that any information you disclose even once is permanently out there.  This doesn’t track how people live their lives: we share information all the time, for all kinds of reasons; however, we reasonably expect it to remain within certain social contexts, and we expect that it will take effort for someone to wrest it our of those contexts.  The latter of these is more or less what “obscurity” indicates.  There is a lot of information out there, but many of those who nominally have access to it don’t actually know enough to do anything with it.  Hartzog analogizes the situation to talking in a restaurant.  People at adjacent tables can likely hear the words you say, but they lack the context for them to be meaningful.  In that sense, communication, even in public, often remains obscure.  This is true both online and off; one of the reasons we need to worry about privacy now is that various technologies make it a lot easier to fill in that context, especially online.  So we don’t become less or more private online, but we do become less obscure.

I concluded by promising a point about latent ambiguity in this context.  Recall that in Lessig’s Code, he suggests that a number of important legal concepts – “privacy” and “fair use” – embed an ambiguity in their meaning.  That is, we don’t know quite what they mean because the people who wrote them into law had never thought through situations analogous to a current one.  For fair use in copyright, for example, it used to be difficult to stop people from making personal copies of works or to meter how many times they used them. So those uses became “fair” and defensible.  If you were accused of violating copyright, you could offer fair use as a defense, and norms arose against pursuing those violations.  Now that technological developments make it easy to stop copying and meter use, we have to confront the question of whether we want fair use for normative reasons, or if we simply had it because of those disappearing inefficiencies. Should fair use protect only use that is inefficient to meter?  There is a clear analogy to obscurity: do we have obscurity because it was difficult to know enough metadata to figure out what the neighbors were gossiping about over the fence, or because we think it’s a bad idea to pry?

The Normative Valences of Obscurity (Part 1)

By Gordon Hull

In a series of articles (and a NYT op-ed; my $.02 on that is here), Woddy Hartzog and several co-authors have been developing the concept of “obscurity” as a partial replacement for “privacy.”  The gist of the argument, as explained by Hartzog and Evan Selinger in a recent anthology piece (“Obscurity and Privacy” (=OP, pagination to the ssrn version)), is that “obscurity is the idea that information is safe - at least to some degree - when it is hard to obtain or understand” (OP 2).  This is because “we should not underestimate how much of a deterrent effort can be” (OP 2), and information that is hard to understand imposes similar costs in terms of effort.  They argue that obscurity functions better as a concept than privacy, in part because it avoids the binarism associated with the public/private dichotomy:

“Because activities that promote obscurity can limit who monitors our disclosures without being subject to explicit promises of confidentiality, the tendency to classify information in binary terms as either ‘public’ or ‘private’ is inadequate. It lacks the nuance needed to describe a range of empirically observable communicative practices that exist along a continuum” (OP 4)

The public/private dichotomy has been the object of sustained criticism, in part because it does not track how people live their lives.  For example, U.S. privacy law tends to regard information that an individual has voluntarily disclosed once as no longer private, as if the context of disclosure doesn’t matter at all.  The obscurity argument is designed to start with this basic thought: we share information all the time, and do so with the expectation that others will manage it appropriately.  This is Helen Nissenbaum’s point about the “contextual integrity” of information; it is also how Ari Waldman starts with his recent reformulation of privacy as trust.  The general stability of these information and contextual flows is behind Lior Strahilevitz’s “social networks” account of privacy and its violation, as well as Dan Solove’s account of how suddenly viral spreads of information online occasions the need to rethink reputation.