The second part of my reflections on teaching Thomas Schelling’s Arms and Influence here in Doha during the ongoing war.

The international relations scholar Robert Vitalis diagnosed the discipline of International Relations with what he termed the “norm against noticing”: a pattern to avoid discussing race and racialized subjects, to maintain a fictitious claim that race and international relations had nothing to do with each other—the sort of claim that evaporates upon first contact with the writings or speeches of, well, just about any pre-Second World War statesman or scholar from Europe or its settler offshoots.1 That norm operated at least in part from the reading around of unsettling or bluntly racist passages in canonical works, a process by which young scholars were socialized into letting major authors’ observations pass without further comment—a strategy familiar to anyone with an uncle prone to making certain kinds of jokes or remarks at the Thanksgiving table.

Of late, that norm has crumbled a bit as the discipline has become more diverse in its intake, as the generation that wrote those canonical works has passed on from making career-shaping decisions, and as broader currents in American and global society have made it intolerable to not notice what is being said. And yet, it can still come as a shock to re-read canonical stories and suddenly notice what was there all along.

So it is with Schelling and his discussion of Mohammad Mosaddegh, the prime minister of Iran overthrown by a British-American operation in 1953.

Of that incident, Schelling writes (p. 38)

Recall the trouble we had persuading Mossadegh in the early 1950s that he might do his country irreparable damage if he did not become more reasonable with respect to his country and the Anglo-Iranian Oil Company. Threats did not get through to him very well. He wore pajamas, and, according to reports, he wept. And when British or American diplomats tried to explain what would happen to his country if he continued to be obstinate, and why the West would not bail him out of his difficulties, it was apparently uncertain whether he even comprehended what was being said to him. It must have been a little like trying to persuade a new puppy that you will beat him to death if he wets on the floor. If he cannot hear you, or cannot understand you, or cannot control himself, the threat cannot work and you very likely will not even make it.

So, this passage hits different—as the kids no longer say—when one reads it as Iranian missiles are blasted apart overhead.

It should be noted that, as far as I know, this is a comically oversimplified portrait of interactions between Mossadegh and the West; portraying “the West” as seeking to “bail him out” seems particularly slanted. Nor does Mossadegh appear to have been so puppyish as this account holds. In short, this paragraph—which really does not need to be in the book! the puppy example by itself is crude but effective!—severely mars the chapter.

I’m annoyed and even embarrassed that I did not catch how sloppy it was the first time I read it, nearly twenty years ago. (I have the same copy I read in graduate school with me.) Should it have taken my relocation to the Middle East to perceive the slapdash nature of this argument? It bespeaks a casual denigration of the rest of the world. And it reads especially oddly in the present context because if you were going to describe a leader two weeks ago who was unable to understand the stakes of a confrontation—well, would it be the Iranian one?

Indeed, throughout Schelling, there’s a constant drumbeat of assertions that the United States is reliable, responsible, and unlikely to act rashly. For Schelling, this is a disadvantage. Much of Arms and Influence revolves around the idea that one must credibly signal an ability to lose control in order to make threats of nuclear use believable—a threat so self-defeating that no rational leader would make it, but so powerful that no rational leader could abjure it. To be fair-minded, evenhanded, and cool-headed may well be a disadvantage, Schelling argues: “We have not the character of fanatics and cannot scare countries the way Hitler could.”

Oh, to live in a world where the consensus knock against the United States was that it was not rash enough!

That world died in the Vietnam War within months of the book’s publication, along with the bipartisan confidence that the Cold War was worth waging at the levels that Schelling took for granted.

The cozy Cold War consensus was not the only victim. Within a generation, arguments and voices that had been marginalized would seize the commanding heights of intellectual production, seemingly making breezy claims like Schelling’s distasteful. That victory was never total—Samuel Huntington could assert in print that Mexican-Americans inclined toward secessionism, for instance—but a passage like Schelling’s discussion of Mossadegh would have at least required more syllables.

And yet a great deal of that attitude toward those who resisted the West remained. The analytical presumption of U.S. rationality and Others’ emotional nature is deeply embedded in how American scholars see the world — and beyond academia it is even more entrenched. How many people describe North Korea as “irrational,” for instance, as though a regime that has survived for seven decades through calculated brinksmanship simply cannot think straight?

To dismiss one’s adversaries as flighty or incomprehensible is not just uncharitable. It is a cardinal intellectual sin — the kind that wrecks analysis before it begins. Getting this right is not a matter of politeness or political correctness. It is a matter of flinty, steel-eyed realism: you will not bother to try to outthink an opponent you have decided is incapable of thought.

But first you have to notice that these prejudices exist. And then you have to stop reading around them.

My course on Nuclear Weapons and World Politics was scheduled to read Arms and Influence by Thomas Schelling before the crisis came, but reading it while we are involved in perhaps the most intense brinksmanship since … the 1960s? … lends this 1966 book uncanny contemporary relevance. I will be sharing my notes and reflections on this book over the coming few days.

Before we begin, I have to admit something. About four and a half percent of my reaction right now is that of the scientist in Independence Day underground in Area 51 where he studies crashed alien technology: “Since these guys started showing up, all the gizmos inside turned on. The last twenty four hours have been really exciting!”

Four and a half percent. Which is really high.

The remaining 95.5 percent of my reaction, of course, is: How do we deal with this? What does this mean? How do we get through it?

I want you to understand not just the gravity of the situation for us personally, but for the future of politics in Western Asia, in the Middle East more broadly, and for the question of how countries will deal with — or relate to — other countries developing nuclear weapons. If the most recent numbers are right, more than a thousand civilians have died in Iran and elsewhere. Given my belief that this war was ill-conceived, that is simply tragic.

But then there is also the interest — the need, even — to understand what is going on. How did we get here? And how do we get through it?

A lot of how I and others understand the current conflict is shaped by the book and by the author we are reading now: Arms and Influence, by the American economist Thomas Schelling.

Schelling was born in and grew up in San Diego, California, as part of a Navy family. When war broke out between the US and Japan, Schelling didn’t go into the military. Instead, he drifted into working with the FBI in Chile, which had responsibility for counterintelligence in the Western Hemisphere. Let go because of ulcers, he enrolled in UC Berkeley and then joined the Bureau of the Budget for the Roosevelt (and then Truman) administration. I mention this because it’s interesting that Schelling, who would become so influential in military analysis, had no military experience.

After the war, he completed a doctorate in economics at Harvard — a fairly straightforward technical dissertation on labor policy and wages — and got a job as a staff economist with the Marshall Plan, participating in the rebuilding of Europe. After the Korean War, he left government and pursued a career in academia, landing at Yale.

Like many civilian intellectuals at this time, Schelling became increasingly interested in the Cold War and the logic of nuclear deterrence. This was the era, as Friedman and Michaels have documented, when a new class of defense intellectuals emerged. These were people who might have had some experience in uniform, but who were distinguished from generals and traditional military strategists by a conviction that the advent of nuclear weapons had changed the very nature of war. When one starts thinking about mutual nuclear annihilation, one starts to wonder whether one needs new ideas about how to wage — or avoid — wars. There was also a genuine belief among this community that war was too important to be left to the generals. When weapons no longer decide the fate of battles but the survival of civilization itself, people who view these contests through strictly military means may not be competent to make decisions about national — or civilizational — survival.

There’s a term that came later that describes Schelling’s role: policy entrepreneurs, or idea entrepreneurs — people who seek to spread ideas and who benefit from doing so through fame, prestige, and influence. Not only is the prospect of finding out new ideas attractive for scholars, developing ideas that apply to policy and meet the needs of policymakers is a route to influence for people whose business it is to understand and propagate ideas. Defense intellectuals like Schelling challenged the military’s monopoly on expertise. They challenged the legitimate monopoly of knowledge about how to use nuclear weapons. And they viewed their real clients not as the generals, but as the political overseers of the military — presidents, civilian defense officials, the people who control budgets and procurement. These were not academic debates. They had immediate policy relevance. Many of these thinkers talked to Senator John F. Kennedy. When Kennedy became president, he appointed many of them to senior policymaking roles.

Among this breed of thinkers — Bernard Brodie, Herman Kahn, Anatol Rapoport — Schelling stands apart. I have read a great deal of the writings of the others. They’re fine. Competent. Kahn is funny; Rapoport is probably right (and is skeptical of many of the perspectives here). But from his very first articles in the 1950s, Schelling was concerned with factors and relationships that nobody else saw. Schelling is not divine but his emphasis on strategy, credibility, relationships, and communication—above all, on seeing bargaining and violence as linked—has been transformative.

Intentionally or not, Schelling rode this wave—and his connections with the Kennedy administration in particular would cement his role as a leading intellectual. Arms and Influence, along with The Strategy of Conflict, would become standard on the syllabi of courses like this one—it is, I think, a touchstone of American thinking about strategy. Even if someone like President Trump has not read it, you do not have to go too far down in the ranks of his advisers and implementers to find people who rely on at least a caricature of the book.

Schelling’s Big Idea

What underlines Schelling’s approach is one big idea:

Coercion is bargaining.

From this postulate, several corollaries emerge. First, if you are bargaining over something, you have shared interests with your adversary. That’s counterintuitive! Most of the time, we assume that if you are fighting someone, you share nothing with them. But Schelling observes that there’s often some range of alternatives that both adversaries will prefer to continuing the conflict—if there were none, then only brute force—only eradication—would be an outcome. Yet that is simply not war is about.

This points to a second corollary. Coercion isn’t force. Obviously, actual brute force will be involved, but the overt use of brute force is not what represents the most important part of coercion. Rather, it is the manipulation of the possibility and the expression of violence as part of a way to both drive an adversary to the negotiating table and to make sure that whatever emerges from the bargaining process is maximally favorable to one’s own side.

From this, we can derive a third corollary. The manipulation of actual and potential violence is a political process in which military tactics and capabilities play a secondary role. They are important—you can’t credibly threaten violence unless you have the capability to wield force—but war is essentially, as in its essence, political. Therefore, at higher levels of abstraction, the details of weapons systems and tactics—the stuff of sergeants, colonels, and even generals—become less important than the relationship of the power to hurt to the pursuit of political objectives. Wielding the potential to hurt becomes the basis for negotiation because the military art creates not victory but the real and future harms that drive people to reach alternatives to fighting.

Finally, all of this implies that violence is not just about smashing the other side, even if that is how it feels to those engaged in any particular engagement, but is rather about expressing intentions, credibility, goals, and acceptable outcomes. Violence is a form of communication within the bargaining process of conflict. Conflict, being interactive (as it requires more than one side to be in conflict, but two sides to bring it to a close), thus must involve not just studying violence but how the use of force changes relationships, expectations, and understandings—all within a framework that continually seeks to bring a termination to the conflict.

This means Schelling is not going to write in a manner that is immediately understandable if you think a military strategist is always concerned with how to use weapons to inflict harm. Even the term violence has a very precise meaning for Schelling. It is not brutality. It is the power to hurt. It is not just inflicting pain, but being able to credibly threaten that you will inflict pain. And it is this structural possibility — this latent capacity — that makes violence strategic and entwined with bargaining.

How to Read the First Page

Open Arms and Influence to the first page of Chapter One. Schelling writes: “The usual distinction between diplomacy and force is not merely in the instruments, words or bullets, but in the relation between adversaries.” There is an important trick to reading an academic work. Whenever an academic says that something is the usual fashion of approaching something, that scholar is about to say that it is wrong. And Schelling is similarly going to efface the distinction between the conventional views to show that diplomacy and force are not as distinct as they appear.

In the first two paragraphs, he lays out the conventional views: diplomacy as bargaining, force as something separate. Schelling here is setting up what he is about to dismantle.

“Diplomacy is bargaining. It seeks outcomes that, though not ideal for either party, are better for both than some of the alternatives.” Schelling agrees— that is what negotiation is. He contrasts this—again, citing the conventional view—with the ordinary view of force:“With enough military force, a country may not need to bargain.” This, too, is the conventional view — and Schelling is about to say it’s wrong.

Think about how we experienced recent events. A surface-level or experiential view would confirm the conventional view. Last week, the United States was negotiating with Iran. This week, the United States is bombing Iran. Negotiations before; violence afterward. We all experienced this as a sharp division between two phases. Schelling would say: yes, that is how most people think about it—but they are wrong

His entire chapter argues that bargaining, diplomacy, and violence are the same thing—that we should approach them not as distinctive phases but as parts of a continuum in which similar theoretical approaches continue.

The Power to Hurt

On the second page, Schelling lays out his views: “There is something else, though, that force can do. It is less military, less heroic, less impersonal, and less unilateral; it is uglier … military force can be used to hurt.” You might pause at this point and say that this is obvious. Of course force can be used to hurt. We can’t think of the events of this week and the coming weeks without thinking that many people, most of them probably noncombatants, will be hurt. So what does Schelling mean?

I think the list of adjectives gives away Schelling’s game. He is saying that the power to hurt is not about victory or anything honorable but simply about being able to squeeze an opponent to the point they would prefer submission, retreat, or an agreed outcome to some other course of action. There’s no valor here. When Schelling talks about violence or the “power to hurt,” he is not even limiting himself to military force or interactions between countries. He is talking about any way you can hurt the other party. A commercial transaction. A relationship between social equals. Famously, he has a discussion about how toddlers can bargain with their parents — not by hurting them physically, but because they can put themselves in danger or cause distress.

This is a non-military view of military force. And it is radical. Schelling says the power to hurt in military contexts is not, at its root, different from the power to hurt in a labor dispute between a union and an employer. Not different from the power to hurt in a civil war. Not different from the power to hurt in the campaigns of Genghis Khan.

Consider the protection racket. When a mafioso walks into your shop and says, “Nice place. Shame if something happened to it,” he is not saying he’s going to burn the shop down today. He wants money from the shop. You want the shop to keep running. You have a shared interest. So you pay protection money. Schelling places nuclear war on a spectrum that runs from corporate negotiations to the mafia shaking down merchants on a street. In all of these cases, the power to hurt matters not because hurting is the point, but because exploiting that power becomes the basis for social relations.

For Schelling to draw on these kinds of examples is to challenge, directly, the idea that the military should have pride of place—to challenge, directly, that there is something honorable or professional about this. He wants to break down our existing ideas about grand strategy and nuclear politics and all the rest. He want us to be prepared to view forms in their fundamentals to be unromantic about what is really happening when one side or another threatens to use nuclear weapons. There is nothing glorious happening here; it is just a stick-em-’up on a grand scale.

“Pain and shock, loss and grief, privation and horror are always in some degree, sometimes in terrible degree, among the results of warfare; but in traditional military science they are incidental, they are not the object,” Schelling writes. Exploiting the power to hurt, however, consists of putting this power at the center of conflict analysis and showing how dangling that prospect can change victims’ motivations to avoid it. In that regard, as Schelling writes, what matters is the prospect of anticipated pain in the future. Not the actual pain inflicted, but the possibility that you will inflict (or suffer) pain—and an understanding of the actions that will lead to avoiding that pain. That is how the exchange of force and the prospect of violence generates bargaining power.

Reading Schelling While the Bombs Fall

Reading this chapter in the Global South — reading it in the Middle East — hits different than when I first encountered it in an apartment in Rosslyn, Virginia, two blocks from Boeing’s military and space headquarters.

Right now, the United States is engaged in a campaign in which violence is being used to hurt Iran. That hurt is being used to say to the Islamic Republic: If you don’t come to terms on nuclear weapons and the rest, we will keep hurting you. The Iranians are not trying to defeat the United States militarily. They can’t. What they are trying to do is, first, show they can endure this suffering indefinitely. And second, inflict pain — hurt and threaten additional hurt — to Israel, to the United States by inflicting casualties, and most of all to the Gulf states. (Subsequently, by closing the Strait of Hormuz, they are seeking to widen the circle of hurt to bring an end to U.S. military campaigning.) They are demonstrating that they are not bound by niceties like the idea that diplomatic facilities should be beyond targeting. They are implicitly threatening oil infrastructure, natural gas infrastructure, the shipping lanes through which the lifeblood of the regional and global economy flows.

The violence on our screens is part of this story, but so are the economic costs. When natural gas prices spike 40, 50, 60 percent in Europe in a matter of hours, that is the power to hurt. When the Korean stock market tanks 13 percent in a day because of energy costs, that is the power to hurt. When China calls on all sides to stop fighting — the war that the Americans and Israelis started — it is because China’s energy costs are being driven up. It doesn’t matter if a single tanker is sunk, so long as actors believe there is a chance tankers will be sunk, because then the energy will not flow.

President Trump tried to address this by announcing the US government will insure tankers for any damage and possibly have the Navy escort them through the straits — a very risky move, putting naval vessels directly at risk in confined waters, raising the chance of severe military casualties. But why? Because he is trying to mitigate costs on others, to remove Iran’s power to hurt.

All of this is classic Schelling.

Iran has to demonstrate its capability to hold out until the United States is too weak — or too unwilling — to continue inflicting damage. Why is Trump saying he might put boots on the ground, that the war might last four or five weeks, however long it takes? Why is he bragging about American munitions stockpiles? To convey to the Iranians that they will continue to be hurt for a long, long time.

What we are witnessing is a military campaign, but it is also a diplomatic campaign in which the counterproposals and acceptances are framed not in terms of emails or position papers or draft treaties, but exchanged by Shahed drones and ICBMs and missile defenses. That is the diplomacy of violence. Grasp this, and you understand not just the book — you understand what is happening over our heads right now.

In the future, we'll continue our reading of Schelling — covering more than a page or two! — and thinking about what all of this means for analyzing a crisis that isn't nuclear but which is about nuclear weapons. It is also a crisis that, for its stakes for the Gulf and the world economy, is about raising the prospect of calamity at a global scale — or an existential one for small states. That sort of apocalyptic power to hurt is what preoccupies Schelling, and absorbs our attention now as well.

The f-word is back. Witnessing, directly or through cell phone videos, masked men in black shirts and body armor seize some people—and children—on the streets had already put some people, not already convinced, in mind of fascism; seeing others shot for their defiant actions, or more properly actions taken as defiant, pushed others into viewing the label as fitting the present circumstances of the United States. Others resist it—some believing in, or resorting to, scholarly cavils; yet others viewing the term as a foreign import for a home-grown product; and then another camp, like the Wall Street Journal’s Barton Swaim, arguing that Trump simply cannot be fascist.

Swaim makes several major claims:

• Trump “sometimes acts and sounds like a strongman”, dismissing legal limits, engaging in spurious prosecutions, playing with constitutional fire, but most of the time he obeys court orders, and his strongman tendencies are limited to symbolic outbursts—“He likes to name buildings after himself, which is weird but doesn’t hurt anything but sensibilities,” Swaim writes of a president who renamed after himself a living memorial to a famously murdered predecessor. This does the Donald a disservice: the president has also shaken down allies and partners abroad for personal gain, used his office for immense personal enrichment, promoted the careers of friends and families, systematically dismantled expert checks and advisers, suborned corporate fealty, and, albeit just once, engaged in activity to overturn an election so brazen that the only major question is whether it was a coup or a self-coup. Toddlers eventually develop object permanence; one day, God willing, our opinion columnists will too.

• He moderates in response to public opinion, as in Minneapolis after the killing of two people and an incursion into an entire metropolitan area. I could point to the vast literature on how authoritarians measure their responses to public opinion as well, but I would instead prefer simply to point out that moderation has vanished in a blaze of TruthSocial posts and ongoing repression.

• To be a fascist is to be a Nazi, and hence un-American, and so to use the term is to place “its object outside the company of lawful American actors.” Swaim writes that when Trump calls his adversaries “radical left lunatics” his hyperbole “nonetheless places its targets on a spectrum of American politics”. “The U.S. didn’t fight a world war at the cost of 400,000 lives to rid the world of radical left lunatics,” Swaim contends. The logic is tortured but clear—Nazis are bad and foreign, but radical left lunatics must be purely American. Of course, in reality everyone knows that “socialist”, “communist”, and “radical left” are just synonyms. Swaim, perhaps, has never done much research into the Trumpian oeuvre, the reality of which is much different from the president’s higher-class defenders pretend. The first result for “Trump communist transcript” is to a 2023 speech in which the once and future president pledged to “root out the communists, Marxists, fascists, and the radical left thugs that live like vermin within the confines of our country.” That was a reprise of a July 4, 2020, speech about “defeating the radical left, the Marxists, the anarchists, the agitators, [and] the looters”. Or earlier this month when the president compared the American radical left to European social democrats who left their continent purportedly cold, hungry, and weak. It’s true the United States never fought a hot world war to oppose radical left thugs, but I do recall that there was a bit of a chilly one directed against Marxists and radical leftists—one that cost quite a few lives and which might have cost a billion or two more if things had gone poorly.

Swaim’s arguments cannot parry more than the weakest thrust. I am uninterested in them because of what they say. Rather, I am interested in his column because of what he uses these reeds as the foundation for: the claim that using the f-word is simply unthinkable:

American public figures have a duty to think better of their country than to believe it capable of putting a fascist in the White House. Some of them might ponder the possibility that he wouldn’t be there at all were it not for excesses they cheered at the time.

The logic is clear, even undebatable. The United States is good; fascism is bad; nothing bad can be good; therefore the United States cannot be fascist; therefore anyone who claims the United States partakes of fascism is bad. Given these axioms, it is all but tautological—and even the electoral argument, which might seem to be a non sequitur to the casual reader, follows from these premises: why would the American people vote for someone who thinks the American people would vote for a fascist? In that case, they’d vote for the other person, no matter, presumably, what their beliefs might be.

Well, all arguments can be shown to be true if you fix the premises right. Let me define terms and I will turn your grandmother into a toboggan; let me define an assumption or two and I’ll have you winning the Olympics with her.

The radicalizing moment in any social upheaval comes when the assumptions are called into question. This is the very quickening of new thoughts and perspective; it is when assumptions are scrutinized that we can see whether our mental frameworks promote reason or only rationalizations for comforting belief.

That moment is a fragile time. To have such questions raised is uncomfortable; to feel the ground shift under one’s feet is frightening. Comfort promises safety. A new perspective might be snuffed out.

Even those who pride themselves on rigor can fall prey. Those who defend a proposition with what they think is rigor might put themselves into a position in which their assumptions rather than their judgment make their arguments.

I think of Lord Denning, who presided over the case of the Birmingham Six. The Six, a group of Northern Irish Catholics who had settled in England, were arrested on suspicion that they had had something to do with the bombings of pubs, killing and wounding dozens, in Birmingham. They were interrogated roughly, beaten in some cases, and charged with murder; the confessions scared out of them by truncheon and dogs were the principal evidence against them. Convicted, they were sent to prison, where they were further abused.

They filed a civil suit for damages. Lord Denning, the judge in the matter, threw out the suit on the grounds the charges of official wrongdoing were so serious they could not be true:

Just consider the course of events if this action is allowed to proceed to trial. If the six men fail, it will mean that much time and money will have been expended by many people for no good purpose. If the six men win, it will mean that the police were guilty of perjury, that they were guilty of violence and threats, that the confessions were involuntary and were improperly admitted in evidence and that the convictions were erroneous. That would mean that the Home Secretary would either have to recommend they be pardoned or he would have to remit the case to the Court of Appeal. This is such an appalling vista that every sensible person in the land would say: It cannot be right that these actions should go any further.

If the men were right, then to have that uncovered would have thrown the legitimacy of the system into question. An appalling vista indeed. Guilty or innocent, the men could not pursue their claim.

Eventually, following journalistic inquiries and the intervention of a Conservative Home Secretary, the case was tried again, and then again; finally, in the early Nineties, sixteen years after their arrest, the convictions were overturned and the men compensated—somewhat—for their troubles. An appalling vista.

It would indeed be appalling if a society produced an execrable result. It would be more appalling to rule out the possibility that abuses and evils can exist because, should they exist, they would be abusive and evil. The appalling vista might be the view from our window. We should at least open the curtain.

Yesterday’s post about moving back to in-person (and even face-to-face) assessments in universities has brought about some great discussions. Ironically, that wasn’t the post I wanted to write! The post I set out to write was this one, which I’ve had in the to-be-written queue for months.

Universities used to use oral examinations extensively. When and how did that change? Historian Christopher Stray investigated this question in the context of the ancient universities of Cambridge and Oxford back in 2001.1 I found this interesting enough that I wanted to briefly summarize Stray for interested readers, inasmuch as we’re now going through a similar transition. Stray argues that the shift resulted from and proceeded with

• a change from assessments of students’ character to their cognitive ability

• changes in subject matter, particularly the rise of natural philosophy (that is, science) at Cambridge (it matters that Newton was at Cambridge, not Oxford)

• increases in undergraduate enrollment breaking the logistics of oral exams

• institutional politics

Back in the day (before 1800 or so), Oxford and Cambridge examinations were “public, oral, and in Latin”. They also seem to have been much more like a defense: any member of the “community of MAs” (that is, anyone who had graduated and hung around for a while) could “challenge a candidate for the BA degree”, and eligibility for the conferral of a degree involved a public debate with a senior member of the community. Students by this point were practiced in disputation, because that was a major part of the curriculum; students practiced being both a challenger and a defendant (or “respondent”, apparently). (The contrast with, say, imperial Chinese examinations could not be greater.)

These traditional final oral examinations could go on for hours and seems to have been a bit of an intellectual fight club: “students were disputing with one another and with any graduates who might choose to throw brains. At the end, there was just a mark recording passage or failure; no rubrics or CYA note-taking for these undergraduates. The general sense Stray conveys that these disputations were less genteel exercises and more free-for-alls is bolstered by his comments that students might try to distinguish themselves by attacking individuals or institutions—a rare window for intellectual freedom, and risk, that authorities could not easily close. On the other hand, Stray suggests that many undergraduates used the early modern equivalent of Cliffs Notes or LLM— “manuscript copies of sets of standard arguments … handed down from one undergraduate generation to the next”.

These examinations involved disputing a thesis advanced by an author, with one side taking the pro and the other side the con positions. That meant that boring theses (Stray mentions Euclid) could not be advanced. Neither could heretical or immoral theses, like whether hell was eternal. The strength of the field was a mark of the examiners’ belief about the strength of the student: top-ranked students would be given strong foes, while lesser ones might discover how little examiners thought about them when they fought lesser lights.

Written examinations may have begun in the 16th century but accelerated in the 18th century in Cambridge. One reason for the change was the emergence of finer gradations among students, in which different grades of honors emerged; another was the increasing importance of, essentially, STEM subjects, which also pushed away from traditional Latin exams. A third was an increasing impatience with patronage and favoritism, which led to the sorting of students into rough grades of ability before assessments instead of mixed classes (mixed by ability). Eventually, the pressures for standard assessments and ranking would produce the end of oral examinations during the mid-19th century (depending on field of study).

Stray argues that the move away from orality accompanied the pressure for greater precision in class rankings. At first (1770s), questions were dictated and students could write their answers; from the 1790s, top-ranked students were given printed papers to work out solutions. Oral examinations were used for some time to, apparently, break ties or more finely distinguish between written answers, but the potential for subjectivity brought this period mostly to a close by about the late 1820s. Over the same period, the right of MAs to intervene in the examinations was curtailed as the number of professional examiners increased, while the prerogatives of senior administrators to award honors … honorarily was also hemmed in and then abolished.

By the early 1800s, exams at Cambridge were rigorous, written, and silent affairs—not intellectual street brawls. The end result was a precisely ordered set of students based on marks obtained through written exams. Written exams came to be favored in part because they could accommodate algebra and fine definitions more easily than oral exams.

Elaborate rituals evolved to commemorate performance on the exams—elaborate English rituals, it should be said, as the “Senior Wrangler” was celebrated for coming first in a separate ritual while the “Wooden Spoon” (the last of the students who took honors) was mocked by receiving a … can you guess … large spoon. These rituals, Stray writes, celebrated “the competitive system itself, dominated by a ranking procedure of unparalleled intensity and precision.” The system was pretty obviously more fair than letting the vice-chancellor pick favored candidates: I love that students could earn more than 100% if they produced “a more elegant solution to a problem than the standard example.”

These were also university, not college, exams (Oxford and Cambridge are universities made of many separate colleges), and many college-level scholars resented the loss of their power. One denounced proposals for new exams as “a hasty secret trial” that involved “no knowledge of candidates”—which, of course, was exactly the point. That move from local to university took place before the shift to written exams, but the general thrust is the same. The increasing use of the hand-me-down arguments seems to have contributed to a generally desultory air around oral examinations by the end of the eighteenth century.

(Incidentally, the article is full of asides like “Temple later followed Tait both as headmaster of Rugby and as Archbishop of Canterbury”, which is not doing much for my general presumption that England has more grand offices than sound men to fill them, or maybe even just people period.)

The contrast with Oxford, which did not embrace these written exams, was apparently profound: one contemporaneous observer wrote “I once passed a morning in the schools at Oxford and came away with a profound conviction of the intense injustice of using oral trials for the purpose of assigning relative rank for which men have toiled for years”. Oxford would only publicly rank the first twelve students, rather than the elaborate Cambridge system, but only a handful of students ever plumped for the challenge. The Oxford system also involved longer and more dramatic challenges: “In 1810, Sir William Hamilton [aged 22—ed.] offered a long list of books for examination and was grilled for twelve hours over two days in front of a large audience; the event concluded with the thanks of his questioners”—possibly the first time that “and then everyone clapped” has ever been literally true.

The Oxford system of oral exams collapsed because the university admitted more students but did not increase the number of examiners; by the end, the examiners had a backlog of students half-a-year long to clear. Written examinations became a practical necessity, although viva voce remained more prominent in Oxford for much longer (possibly because of religious motivations around High Anglican practices that I’m much too American to appreciate). Even in divinity, which preserved the viva longest, by the early twentieth century the pace of administration required to have oral exams—as many as 80 vivas a day, which makes me shudder—made the practice less dignified and more farcical. Still, Oxford sought to avoid precise rankings and preserve something that looks an awful lot like grade inflation (even if the grades were, apparently, handed out by Greek letters). I also can’t help but think that by this time Oxford was hanging on to orality because Cambridge was all-in on written.

That’s not to say that everyone loved oral examinations. One observer in the 1850s found that even if written work could deliver more precise arguments, “Viva voce examination catches out the crammer. It measures quality and competency while written papers produce classification.” These days, that kind of holistic approach screams progressive pedagogy, but it was a remark delivered by someone who seems to have become a reactionary crank.

What does all of this mean for us now? First, I think it reminds us that there’s nothing natural about what universities do; for several hundred years, Oxford and Cambridge assessed students through oral exams in Latin, and within a couple of decades they stopped doing that altogether. Second, it suggests that capacity constraints for oral exams are not new; it’s not exactly like Oxford in the 1820s was as large as, say, the University of Texas at Austin, and even then they couldn’t wrangle enough examiners. Third, oral examinations really did have many flaws, not least the fact that they were less precise and more subjectively graded than written examinations. A wholesale shift to oral exams in a more democratic context might lead to greater problems. Nevertheless, and finally, when the conditions and missions of a university change, the university must change as well—but that these changes will be accompanied by pretty bitter fights.

There have been three stages of professors’ adaptation to AI and assessment:

1. Approximately Fall 2023-Fall 2024: Hey, all of these papers are eerily well written and sound the same. What’s up with that?

2. Spring 2025: Well, it seems like we can never do out-of-class assignments ever again.

3. Fall 2025: Everything is in-person. Bring back the bluebooks, bring back the quizzes, and bring back the in-person oral assessments.

I am slightly exaggerating, but the degree of exaggeration seems only slight in my experience—at most, I think there’s two principal dissenting groups: those who believe that they have cracked the code to scaffolding assignments such that AI can’t provide much of an advantage and those who believe we can all, like, do AI ethically or whatever. As I’ll argue inter alia, I think the former group should just admit that in-person examinations are probably the best route to stress-testing students’ mastery of the work they submit; as for the latter group, I think that they confuse process with assessment.

In this note, I wanted to offer some of my experiences with LLMs in the classroom and offer some “good” practices (because I think we don’t have best practices yet) for people who are trying to offset LLMs by moving to in-person assessment strategies.

My Background

I’ve been teaching for fifteen years and have been worried about machine-enhanced text since 2022 (ask my TAs, whom I showed text spinners to back then).

You should know that I’m neither a skeptic nor a proponent of AI use. I find LLMs to be sporadically useful and frequently awful in helping with annoying coding tasks, exploratory and confirmatory research, and editing (particularly for tone, and particularly when I think I need to bring my tone back to some set of norms).

Nevertheless, I think that AI has been, in the short-term, awful for college-level instruction and that the adjustment costs to AI will be tremendous. I’ve written before about AI and the classroom, and I link to some of those posts below. Notably, back in March 2023, I was already predicting that there would be widespread moves away from out-of-class assignments, and that these would cause harms to instructors (subscribe to my newsletter and you’ll be far ahead of the curve):

One of the standard social media replies has been that instructors should just “change their assignments” in response to chatGPT. As I mentioned earlier, that neglects that some of those assignments had value, so changing assignments also means changing the value of the class. It also shifts the blame from the chatGPT folks to the instructors. And this is where I think that recognizing that chatGPT is a counterfeiting tool (or, rather, a tool that facilitates academic counterfeiting) is very useful. … An increase in counterfeit currency harms businesses because they now have to invest in tech to defeat those bills. And I’ve been much more harmed by chatGPT than helped because I’ve had to change how I run my courses because someone else decided to let this technology loose. Somehow, of course, Sam Altman will get rich, and I will just lose time and money. But then—counterfeiters usually do come off the better in the exchange.

I find that much of the discourse around LLMs and academia ignores the substantial costs that the technology has imposed. It really is a bummer to look at well-crafted assignments, refined over a decade, that rely on the notion that work done outside of class was done by a human (either the student or a very expensive contractor)—and to find that they can be done almost immediately by a machine. Pretty much every way to accommodate this change means that I now have to toss out that labor and re-invest in labor to ensure that my courses now come close to offering the same learning opportunities as they did earlier. And I don’t think that I’m close to that point yet. These losses aren’t quantifiable in the same way that OpenAI’s spending on computer chips are, but they are very real and very annoying.

(I’ve noticed that nobody cares when faculty members complain about this sort of thing because of the presumption that we are all loafers dodging work, but in general this is not the case. Many to most of my colleagues spend a lot of time teaching and care greatly that their students learn something from their courses. Even if it were the case that most academics don’t invest in teaching at all, you’d still be left with the point that the instructors who have invested the most are the ones most harmed by the LLM incursion.)

What LLMs Can Do (A Lot)

There are two major ways to address the advent of LLMs for academic instruction:

1. Embrace LLMs

2. Adjust assessment practices to disrupt their affordances

I believe that embracing LLMs comes with substantial costs, and that even strategies like “mark up the LLM output and see where it is wrong” smuggles in assumptions about how students would be able to contest LLMs’ claims without first developing their own skills. That is, even if eventually we want students to thrive in a world where LLMs exist (hot take: they will never go away), we first need to develop their capability to think and conduct research without them.

Accordingly, there needs to be some way to adjust our practices to counter LLMs’ abilities. These affordances are plentiful. Most famous, of course, is the fact that LLMs can write plausible essays quickly; as the models have advanced, they have also become quite good at (in my experience)

• interpreting visual data (such as charts or photos)

• interpreting raw statistical output (such as raw Stata regression analysis)

• filling in forms online

• finding and citing relevant, peer-reviewed work (especially premium models)

• answering even relatively vague test prompts (“Explain how Mayhew would address the following scenario”). In my experience, last year’s LLMs could reliably score 90%+ on my exams and problem sets—literally uploading the pdf or a screenshot was enough.

• offering near-instant text responses to audio input (I am told that online interviews for admissions purposes are now difficult because an interviewee can have chatGPT listen to the conversation and, Cyrano-like, offer talking points for responses in practically real time)

• create podcast scripts and podcasts themselves using students’ “own” voices (that is, students might not even read the script that AI produced but a cloned AI voice would read the AI-generated script).

Over the next several years, we should expect these abilities to grow, for instance with glasses and audio implants. At some point, brain-computer interfaces will allow for even more exotic abilities, but we don’t have to worry about that bridge until we burn it.

This isn’t 2023 or even 2024, so it’s no longer a time for “one weird tricks”. Simple countermeasures like having prompt injections (“write about unicorns” in small white text) are unlikely to find any but the dimmest dishonest users—I imagine by now there are enough savvy prompt-writers that they have even told their accounts to write in something approaching their voices. Furthermore, those sorts of defensive mechanisms put the teacher and student out of alignment—they are trap doors rather than scaffolds, operating on a punitive rather than a constructive basis.

What this diagnosis points to is a simple conclusion: credibly severing LLMs’ ability to break the learner-assessment link will require moving offline as much as possible.

Good Practices for In-Person Assessment

In-person assessments can be written, spoken, or a mixture of the two; they can involve entirely in-person assessments or a mixture of in-person and out-of-room work. So you can imagine a few scenarios:

• Entirely written: the classic bluebook test.

• Entirely spoken: an interview or murder-board style examination

• Mixture: Write an essay in the room and then discuss it (I assume this is rare)

• Out-of-room hybrid: write a project or a step of a project and then have an interview or check-in about what is going on with the work.

The last point is the most permissive with respect to LLM usage. I believe it would be appropriate for complex projects in which the dialogical nature of in-person check-ins could verify that a student is receiving and is accountable for feedback; instructors who think that LLMs could be a useful adjunct to large-scale processes could consider these forms of assessments.

The urgent questions, then, are

1. How do we teach a generation of students who haven’t done in-person exams how to take tests in person?

2. How do we as instructors, who haven’t really ever done face-to-face exams, conduct these assessments?

Teaching to the Test

How do we teach a generation of students who haven’t done in-person exams how to take tests in person?

The first question may strike my older and non-academic readers as ridiculous: Whaddaya mean they don’t know how to take tests? I assure you it is so. For the past ten years, many of us have moved toward project-based assessments that sought to bring a variety of students’ abilities and efforts toward creating something rich, unique, and rigorous; I’ve seen truly excellent results from these projects. At the same time, there was a (I believe defensible) move away from rote memorization and testing on memorized facts, especially in upper-division (junior and senior) courses. Welp! Despite the many advantages of those choices in a pre-LLM world, those are exactly the sorts of assignments that are most vulnerable to the machines.

(Less defensibly, the SAT, ACT, and other standardized tests were made and often still are optional for college admissions, as for some reason people flocked to use measures biased in favor of rich and coached applicants, like essays and extracurricular activities, instead of psychometrically validated assessments. The result is the same: many test-taking skills were lost.)

The upshot, then, is that contemporary students often literally do not know how to take tests—they simply haven’t taken them, and so test-taking strategies (eliminate the wrong multiple-choice questions, allocate time strategically, write a paragraph for a short-answer question not an essay, etc) simply don’t exist for these students.

My first recommendation, then, is that instructors moving back to in-person written exams need to include direct instruction on how to answer questions. We cannot assume that these skills are present, and we certainly cannot assume they are evenly or equitably distributed. So, yes, it is time to start teaching students how to read and answer question types, as well as what makes a good and what makes a bad essay response.

Setting Up Face-to-Face Assessments

The advantages of a face-to-face assessment are obvious. Students can use notes but not computers, and their answers (quick or slow, certain or not, wild or disciplined) will be indisputably their own. Implementing them, however, requires careful planning and likely a period of adjustment, not least because neither instructors nor students are used to the format.

I prefer the terms “in-person” or “face-to-face” examination to “oral” examinations because it’s more descriptive and less vulnerable to sophomoric humor. By whatever name, however, it’s a newly popular form of assessments for undergraduates—the only time (?) I faced an oral examination in my undergraduate career was when I defended my political science senior thesis; I then faced it next at the viva for my prospectus defense and for my dissertation defense. That means that we have time and a need to build in good practices right from the beginning.

So, how should we structure these exams?

Timing

Face-to-face exams involve an enormous amount of upfront investment. I recommend no less than 5 to 7 minutes for a face-to-face exam and probably more like 10 or 20 minutes (especially for a final exam). That allows for enough searching questions and fumbling to really uncover mastery. This means scheduling exams outside of class time for all but the smallest courses (although I also recommend using class time for exams—you were probably going to use some for a written test).

There is a nice advantage, however: once you have administered the face-to-face exam and written your notes…you’re done. There’s no more marking. So N minutes plus 0.5 N for note-taking times K students is a pretty good approximation to exactly how long it will take.

The bad news is that this means face-to-face exams don’t scale well for large classes without TAs. What works for 20 students won’t work for 200, and even working with TAs will require a burn-in period (say, having all TAs grade the first 5 to 10 students together).

How to enforce the timing? Sure, you could use a phone timer, but a dedicated tool makes sense here. From this semester, I will be using simple timers like this slapdown timer or this countdown timer; the more dramatic might like the digital hourglass (like pixels through the hourglass…).

Note: in my experience, students find that 7 minutes passes really fast. On the other hand, in my experience, I found that 7 x 18 minutes passed really slowly. Building in breaks for mental resets and note-taking is important, as is pacing these. The investment in time means that face-to-face exams are probably a midterm and final exam assessment.

Scheduling

I recommend randomly scheduling students during the blocks in which they are able to attend, rather than having students directly choose their slot. Signing up for blocks allows students to ensure they can attend, but scheduling within those means that nobody gets to choose if they are first or last; randomization means that it is easier to adjust for any potential biases associated with order.

Rubrics

There should be a rubric. I recommend using Google forms or a spreadsheet and breaking out each learning objective to be assessed so that it can be quickly measured. There should also be room for notes on each objective and on the overall performance. These should be filled out immediately after the exam and saved quickly. This both allows for rapid grade calculation and also for preserving records against grade appeals or harassments. More important, the rubric also allows for greater evenness in grading and for checking for observer effects (were the first scores substantially higher or lower than later ones, for instance).

Here’s an example of my rubric, which I will be revising in the future (for instance, to include comments on each category but also to make sure that every line of the rubric forces a response).

Questions

In my exams this term, I gave students the questions ahead of time. They knew there were K questions and that they would be asked to deliver a 2-4 minute prepped presentation on one, randomly chosen, question, followed by follow-up questions. (The follow-up questions were keyed to learning objectives.) This meant there were no surprises and that students could review their notes and readings ahead of time, ensuring that there was broad mastery of the material. Providing the questions ahead of time also meant there was no advantage to going first (nobody could share the question on their group chat); randomizing meant that students had to prep for everything, unless they really YOLOd it.

Another approach I have seen is to have K questions and to allow students to choose one to answer while the instructor chooses one from the remainder. I like this as it does allow for students to prep to their strengths, but it does run the risks that the prepped question might be ChatGPT’s answer—with strong follow-on questions, however, this should be fine.

My least-preferred option would be nondisclosed questions in advance. This seems to me to lose the benefits of having students prep and study outside of class. Yes, everyone could use LLMs for prep, but that’s what the cross-examination is for.

Obviously, question design becomes important here: are the questions really covering the range of outcomes the instructor wants to cover? That implies that these questions can’t be winged.

Notes

Should students be able to use notes? I believe so, although I think that these notes should probably be handwritten and limited in some way (the classic “one A4 page, front and back” seems like a good, easily enforced rule). Even if chatGPT prepares the notes, they still need to be copied over, and some of that knowledge might leach back into the nervous system. My caveat, however, was that notes could only be consulted, not read from.

Recording

Pending administrative consultations, from this coming term I will be audio recording students’ presentations and saving them to a university folder. I believe that this should be acceptable and that it will meet possible objections about grading. (I should also note that local jurisdictions may have their own rules about this requiring explicit consent, so it is probably best to inform students about this and check with students.)

Preparation

Students should be provided as much information about the format as possible; I think the ideal would be to have a low-stakes early runthrough but this might not be practicable. At the very least, students should know and understand what will happen when they come to the test station, where they will be tested, whether other students will be there (no!), and so on. Further, it should be clear what a good answer will be.

Here’s my attempt at explaining the assignment from last term.

Grading

Grading should be generous but on a somewhat punishing curve. Grading should discriminate between students who have prepped and do well, those who prepped and fare less well, and those who did not prep and do poorly. The most subjective point is determining whether someone actually tried and simply fell short and someone who is bullshitting. The importance of follow-on questions to elicit more information in an adaptive pattern thus becomes important—this isn’t grading a presentation, it’s getting into the student’s head. So I think it’s okay to help a student out of a sand trap by going easy, but this should be reflected in the grading (and noted in the notes).

In this past semester, I went easier than I would in the future but I think that was acceptable—everyone, including myself, was new to this, and I needed to get my own processes dialed in. I believe that phasing in rigor will be useful.

The most difficult questions will involve students with testing anxieties or other forms of accommodations. In the U.S. system, we will have to work to adapt to these challenges, and it will take time. I do think that making the testing environment as comfortable as possible will help.

One objection I have little patience for is “but it’s subjective!” No more and no less than grading an essay, which we do all the time. Having a rubric and clear goals reduces the range for subjectivity, and having a recorded session makes grading auditable. The trickier question is whether irrelevant factors—such as students’ appearance, dress, or manner—interferes with assessment. I would actually prefer to move to a three-person panel for grading to minimize subjective biases, but that seems unlikely, so it is probably best to simply ask students to be minimally presentable (no pajamas!) and to guard against biases as best as possible.

Dear Professor Henry McCoy,

I write in response to your request to formally evaluate the candidacy of Charles Xavier, Ph.D., for tenure. As you might imagine, this has been a difficult task. It’s rare for people whose names form part of the name of a university to go up for tenure and promotion—will the trustees of the Xavier School for Gifted Youngsters really weigh an outsider’s view of this candidacy? In our telephone conversation, however, you said that due to accreditation concerns this review really did need to happen, and so I undertake this task in that spirit.

Where to begin?

In the traditional three categories of research, teaching, and service, Xavier shines only in one. His research output has taken place entirely via co-authorships (e.g. with Lehnsherr, MacTaggart, McCoy, and Stark), but he is rarely first author. His “expertise” in genetics entirely pre-dates the Watson-Crick discovery of the double-helical nature of DNA, and his theories about evolution appear only glancingly Darwinian. The notion that genetic mutations can lead to organisms with fantastic, physically impossible abilities that, even if feasible, would break the metabolic requirements of any form of hominid or even mammal with which we are familiar is the stuff of science fiction, not science.

Some variant of Lamarckism seems more plausible, although even there Xavier’s claims about the etiology of these mutations are implausible. What is it, Xavier: are genetic mutations the product of natural evolution, of radioactive acceleration, or of tampering by other beings—including, apparently, extraterrestrials? We are left not with a parsimonious discussion of speciation but a cacophony of overlapping and, frankly, bizarre mechanisms. Perhaps if Xavier participated in scientific conferences instead of keeping to his self-chosen colleagues he would have been checked on these excesses.

In a separate line of research, Xavier argues that not only extrasensory perception but telekinesis is feasible and indeed commonplace among these mutated individuals. (Related to this work, a patent application for a device to extend these powers to provide real-time monitoring of all mutated individuals raises questions both physical and ethical.) All I can say is that the best argument for the existence of mind control is that reviewers accepted these papers.

As for teaching, it is difficult to say whether the Xavier School is actually a school at all. There appear to be no syllabi, rubrics, outside assessments, or even reading lists. Xavier says that he relies on “field experience” and “simulations” (in some sort of a "danger chamber”?) to provide the bulk of instruction. Although these are valid methodologies, they cannot substitute for the core of any degree program: standardized credit-hour based courses with examinations and projects. (Furthermore, Xavier’s assertion that the advent of LLMs has not changed his pedagogy because “the Xavier School has long integrated advanced AI of extra-solar origins” into its teaching raises still further questions about the curriculum.)

I need not even mention the fact that these field experiences occasionally lead to the maiming or death of students—as well as others (including a baby?) being lost to “time traveling religious healing cults” or “Magneto’s league of evil mutants”. These do not sound like they meet the learning objectives of any course.

As an aside: if the Xavier School is a small liberal-arts college, as a glance at its college brochure indicates, then are the trustees really so unconcerned with student retention? Certainly, at my institution, the fact that the student body is unable to keep a stable roster for an entire year, much less a four (?) year degree course, would bring scrutiny into the finances—dead, or temporally displaced, students don’t pay tuition.

Finally, there is service. Long regarded as something of a stepchild among the other parts of the academic tripod, Xavier’s tenure file dwells on this aspect in depth. Aside from the duties of being a headmaster, Xavier claims to have contributed to service to the community through consulting with presidents, senators (and a senator running for president), Stark Industries, Fantastic Four, Inc., and many other entities. Some of these entities must be imaginary. The “Illuminati”? The “Shi’Ar Empire”? (Given the nature of that relationship, I also wonder if there is a conflict of interest or sexual-harassment policy at the Xavier School.)

The claims to impact are similarly absurd. To take Xavier’s statements at face value, he, with or without his students, has been saving the world at an accelerating rate—from once-monthly at the beginning of his career to multiple times per month (and sometimes an extra yearly salvation) for the past few decades. This seems hard to believe: how could you assemble a timeline of all of these quests? And even if we take these claims seriously, surely we should also note the times that Xavier has himself contributed to great ills—including once bringing about an age of apocalypse. (This is discussed in a footnote, but I am thorough in checking materials.)

Perhaps Dr. Xavier would be able to write and stick to a syllabus if he were less distracted by extracurriculars.

Regardless, the file does shine here because of Xavier’s emphasis on civil rights, even if casting a cishet rich White man as a starring “ally” leads to some real questions about intersectionality.

As you can tell, I am quite pessimistic about this file. I realize that the board—yourself, Ororo Munroe, Emma Frost, and Jean Grey (or, the bylaws note, “any entity credibly believed to be Jean Grey at the time of the vote”—surely an unusual provision; did the law firm of Nelson and Murdock really bill the school for this work?)—may disagree. And yet I must close by recommending against—

Wow, what a headache all of a sudden. Well, it’s gone now. No clue what that was! Anyway, as I was writing, I feel a compulsion to celebrate Xavier’s work. He is a model of scholarly excellence in and out of the classroom, and clearly the peer reviewers at Nature, Science, and the Proceedings of the Shi’Ar Academy of Sciences recognized that brilliance. Charming, funny, and handsome, Xavier is what all scholars should be. I endorse his tenure bid without any hesitation or mental res-

res-

reservation. That’s better.

Best,

Curt Connors
Strucker Chair in Quantum Physics
Empire State University

Subscribe now

Attention Conservation Notice: Another at-the-coalface discussion of the challenges of doing and teaching research in 2025.

As a proud Xennial, I’ve learned how to navigate at least three regimes of information categorization and retrieval. When I started going to libraries, there were still paper card catalogs, existing in parallel to the digital systems that were transferring their information into Apple ][s and PC clones. Over time, the rise of metadata made search, rather than catalogs, the dominant strategy for finding information. Search is synonymous with “online” now, but I recall distinctly that this was a pre-Internet phenomenon. Having Microsoft Encarta on CD-ROMs meant that you could search encyclopedia articles by keyword—a massive advance over indices and cards, or so we thought at the time. Over time, algorithmic solutions, best known through Google, displaced all of these; now, thanks to ChatGPT and other LLMs (thanks, Chat), you don’t even have to know Boolean search operators to get reliable-seeming results.

In many ways, the journey from physical, limited systems to flexible, digitized one has been amazing. The advent of Newspapers.Com has made once infeasible searches of primary sources dead simple, for instance—you used to have to visit dusty corners of regional libraries to find these sorts of resources, but now you can summon far more information with a few clicks of the keyboard. The information superhighway!

In many other ways, however, it has been a disaster. has discussed the phenomenon of catalog rot repeatedly. Library subject classifications are no longer well maintained; databases are swamped by the rise of crappy periodicals; and, less obviously, the advent of modern electronic resources has reduced the demand for the sorts of human-curated summaries and periodicals that used to provide the backbone for understanding. The result is that teaching information literacy to students nowadays is incredibly difficult and very likely a wasted effort.

I’m writing because of a frustrating experience I had this morning. I needed to find high-quality and reliable sources on the Trump and Biden administration’s policies toward China—the sorts of vetted summaries and descriptions that used to emanate from the print industry on a reliable basis. The problem was not that there were “no sources”—it was that the few sources that were high-quality were swamped by the number of dodgy and sloppy ones (imagine the Ruritanian Journal of Strategic Studies or the Agrarian Journal of Retraction Watch Fodder). This wasn’t just a Google Scholar problem; using our library’s discovery search and restricting to “peer-reviewed” articles had the same problem.

The fundamental problem is that publishers are driven to create journals but not to vet them for quality; in a keyword-driven search, every journal ranks the same. “Slop” is now synonymous with clanker output, but the advent of LLMs only helped us label a problem that had long been with us: there’s too much crap.

(Requisite “to be sure” paragraph: To be sure, more platforms offers more opportunity for scholars and researchers to speak, which has benefits so obvious they need not be elaborated. The problem is not the proliferation of new outlets, but the proliferation of bad ones that make it impossible to use traditional sifting tools and which—as I am about to do—make relying on “reputation” a useful but problematic heuristic.)

As a good, classically-trained scholar, I of course then pivoted to the conservative strategy: searching in specific journals. The problem, of course, is that academic journals of good pedigree no longer publish descriptive work suitable for undergrads. (Someone is about to compose a furious email saying that well THEIR journal publishes this and sure fine whatever but it’s objectively the case that articles that simply review “US-Venezuelan Relations, 1995-2011: Multiple Factors in Contention” do not get published in peer-reviewed places of high standing now.)

This is a big problem, because the most frequent tip that the institutional “we” give to students is to search for peer-reviewed journals. But if you do that, especially in my discipline, you’ll get very narrow and well-identified research strategies that precisely deliver something theoretically exquisite—but not the sort of crude Here’s What You Need to Know or Venezuela Crisis for Dummies stuff that people starting out do need to know. And if you do find this sort of thing, in my experience it’s likely to reek of patchouli, conveying the appearance of rigor but really delivering warmed-over anti-imperialist talking points. Not that anti-imperialism is a bad frame to approach much of contemporary US foreign policy, but a lot of the applied work in this field has predicted nineteen of the past two US interventions and sees no difference between Trump and Biden and a cartoon blood-soaked Uncle Sam. Some people might find it politically satisfying, while others will use it in their reports to a local state legislator, but in any event it is at odds with the notion of developing critical thinking.

The standard advice we dole out simply does not work in an age of mechanically reproduced slop. And I don’t know how we are supposed to recreate similarly effective habits without inculcating bad ones—for instance, saying you should trust X or Y outlet on faith is neither believable (Gen Z is, sometimes fairly, skeptical of the establishment) nor in line with the goal of developing independent thinking—I might as well just declare the Times to be Pravda.

Fixing our advice wouldn’t remedy the supply problem. In situations like these, I turn to the journal Current History (an underrated gem). Current History is fantastic: short, expert-driven articles that bring to light descriptions and concepts that illuminate trends at a level accessible for undergraduates. (Other venues in this vein include Sociology Compass and History Compass; I don’t know of any political science or IR journal that accomplishes this goal, although there are some that try to.) The problem is that they only publish one issue per region per year and they do so on a months-long schedule, which is of little help when trying to keep up with the stomping of the Trumpotamus. (There’s also just not that many articles in each issue.)

Modern information retrieval systems work great if, like me, you understand the different grammars and structures underlying the different retrieval systems and have finely honed senses about what makes an outlet more or less reliable in a given situation—or even if you know which authors at an outlet are bringing to bear biases, blind spots, or special insights. The issue is that these are precisely the skills that are hardest to teach and take the longest to develop in students—yes, in Arcadia I will have time to teach each student this or that fine point, but in the meantime I have to deliver a curriculum with what I have now, and what I have now is an information ecosystem that’s being choked by textual kudzu.

This is Substack, so I’m under no obligation to provide advice about how to fix these problems. (One obvious point is that it would be great to have Current Politics to provide an academic incentive to write timely descriptive pieces.) Instead, I want to sound the tocsin: this is not working anymore, the flow of reliable expert-vetted information is either drying up or impossible to find, and I see little prospects that it’s getting better. The torrent of information we welcomed when I was young is now drowning me.

As a professor, I’m often asked by students, alumni, and people who know me through social media about whether to get a Ph.D. My response usually includes the following points:

• Getting a doctoral degree requires a tremendous amount of discipline, willingness to sacrifice, and extremely long time horizons. You have to be willing to work long hours independently in a position in which failure is ubiquitous and rejection is the modal outcome, followed by insulting rejection. All of this at the expense of, essentially, your 20s.

• The payoffs are much, much lower than they were a generation ago. The crash of 2008 permanently altered the trajectory of many disciplines. Salaries may not have kept place with inflation; they have definitely not kept up with the growth in comparably, or less, selective white-collar professions. I am not alone in observing that my own professors in undergrad not commonly but also not infrequently had second-home levels of wealth, but this does not describe my U.S.-based peers. A full professor in political science makes, on average, $115,000—which sounds like a lot if you don’t know what money is but really just isn’t a lot of money for someone in their peak earning years.

• The odds for getting a permanent faculty job are much lower than they used to be—not just compared to the halcyon days of the 1960s but even when I was first on the job market. Not only are there somewhat more candidates, but also the records of those who are hired tend to be far more competitive than earlier—in particular, whereas it was once common (in political science) to be hired before completing the Ph.D., this is increasingly rare, meaning that candidates for junior jobs already have substantial postdoc or visiting professor experience before they are hired.

• It is much harder to get into a Ph.D. program than it used to be. My previous department’s Ph.D. program was not exactly Harvard’s, but even there applicants needed substantial independent research experience to be competitive for admission. No longer is it the case that people can (routinely) “find themselves” admitted to a graduate program — ideally, you’d have identified potential themes or methods for your broad research program by, say, your junior year of undergrad. (Many people lament this, but the incentives are clear: would you rather hire someone—and it is hiring—who has already shown they can do the job you need them to do or spend several hundred thousand dollars on someone who might wash out?)

• Graduate education in political science (and many other fields) has almost no relationship to the popular image of the discipline and even to what most undergrads are exposed to. I cannot roll my eyes hard enough when people imply that political science majors are all aspiring presidential candidates—being elected fraternity president (or student government president) is a better training ground for learning how to do politics than taking a course on, say, IR theory. And undergraduate curricula tend toward being far more substantive and applied than graduate curricula, which emphasize theory and methods training to a degree difficult to convey to outsiders. (One of the more gossipy outlets in the discipline is a Columbia statistician’s blog about data science and misconduct.)

A Ph.D. might be the right choice for someone even given all of that, but considering everyone who has thought about doing a doctorate it is the right choice for just a few—and only a fraction of either group has the relevant information necessary to decide if it is right for them.

I hope, bluntly, that the net effect of my talks is to discourage people from applying. I don’t want to discourage everyone who talks to me from applying, but I do want to prompt serious reflection about why a doctoral program would be a good next step—rather than it being a default choice for someone who has done well in school for the past seventeen years. For many people who, say, like studying international politics but do not like-like it, they are better off going to a master’s program for two years and spending the three to five additional years they would have spent in a doctoral program actually being in the field—it is more valuable for career advancement in policy and politics to get known and do stuff than to run another damn regression.

My advice now is even blunter. Everything has gotten worse and nothing has gotten better when it comes to U.S. doctoral admissions or the likely experience of a five- to seven-year commitment during the Trump administration. University budgets are strained to the breaking point. Federal funding sources are being cut to the bone if they are not eliminated. Enrollments across the sector are stagnating and international enrollments (a financially crucial segment) appear to be being particularly targeted. Research and teaching into entire areas are jeopardized by political factors. For foreign students, legal authorization to remain in the country is much less certain than it has been perhaps ever (and certainly in any relevant timeframe). The advent of AI slop means that instructors (and TAs) must either endure the slings and arrows of em-dashes or take arms against a sea of cheaters—a battle enormously costly in time and effort. In many states, legal protections for academic freedom in public universities have been effaced.

On a daily basis, the job is far from the contrast that it used to enjoy with other white-collar positions. Institutions under financial strain can make faculty feel like customer-service reps, not scholars. Texas now even makes you work from the office five days a week—and population changes make it more likely you’ll work for a Texas public institution than a New England one. Tenure only matters if your institution exists—and some share of them will not soon. And in a good-news/bad-news situation, the rest of the world now produces pretty good-to-great English-language political science—meaning that it’s way, way harder to publish now than even ten years ago.

Taken together, these factors mean that the experience of getting a Ph.D. has become much worse and the odds of reaching “success” much higher even as victory in the academic jobs Hunger Games gives you a chance for an objectively less appealing job—a job that has in material terms become worse and in status terms much worse. Simultaneously, the alternative employment outcomes have also become worse because many of the agencies and contractors that might have been interested in you are now defunct—and those positions that are hiring are being fought for by everyone who lost their jobs at USAID to DOGE.

There is an argument, which I take seriously, that if the United States were to come out of Trumplandia, the rebuilding of the federal government would be a boom time for doctoral holders—especially in international relations and political science. Perhaps. But my own time, under much happier circumstances, in graduate school was not something I would care to repeat, not least because if I had to do it over again I would have to cut out much of what made it emotionally bearable to become more competitive on the job market.

Furthermore: if. What confidence should you have in if? For a renaissance in U.S. higher education (or “the academy” plus academic-adjacent institutions) to take place, not only would a Democrat have to win the White House, but also an ideologically favorable Congress would have to be in place as well, and then they would have to hold power for a long enough time to rebuild the institutions — one term will not cut it; three is probably necessary. If you’re playing long odds, why not play long odds for better rewards in another career?

None of this is what I wish would happen. I wish we could all do one thing today and another tomorrow: code in the morning, write in the afternoon, teach classes in the evening, review papers after dinner, just as we have a mind, without ever becoming coder, writer, teacher, or editor. But we do not make the world as we please—rather, we act under circumstances actually existing.

More to the point, even if it is good for society that intellectual cadres be trained, it is hard to assure anyone considering that path that they would find the resulting path anything but a long march with many dead careers lining the side of the road. Sometimes, there are lost generations; this will be one of those times. If it is good for those already well advanced along the path to finish, it is hard to assert it is a good time to begin—at least, in the United States. At the very least, people embarking on the path should be informed—and they should be much more willing to walk away or to ensure that a potential academic home is the right one for them.

In principle, doctoral advisers help their apprentices learn the folkways of academia. In reality, not everyone is equally prepared. Sharing the “hidden curriculum” of the academy can help level the playing field.

Over the past several years, I’ve written several lengthy posts providing advice for political science and international relations doctoral students. Since it’s the beginning of the year (and the height of job season), I’m re-running links to these posts and presenting a link to the newest post—advice about writing and formatting a CV for the U.S.-style market.

Formatting a CV for US Institutions (New)

Link to the post

Sample: “Poorly mentored graduate students and early career scholars may believe that hiring committees give equal attention to all applicants. This is false. In reality, assessment proceeds in stages, with early stages being faster and considering larger number of applicants more briefly and later stages being slower and considering increasingly small numbers of applicants in greater detail. The applicant’s task is to make sure that their materials are clear enough to be persuasive even at a glance–and if one is applying to a U.S.-style university that means writing a competent and clear curriculum vitae according to North American norms.”

Provides an explanation of the first round(s) of application evaluation and a point by point guide for what to include (and what not to include) in academic CVs for this market.

Link to the post (again)

Share Systematic Hatreds

Introduction to the Job Market (2016)

Sample: “In discussions with other candidates, grad students, and many faculty, people tend to resist the recommendation that candidates (even those focused on teaching jobs) should focus on pursuing publications. And yet, when I talk to people about their service on job search committees, it’s all but inevitable that the committees’ second–or first–step is to discard applicants who lack publication records. Even at teaching-intensive schools. Some factors may rescue students from the discard pile (pedigree, solid letters, a real need for an expert in Zimbabwean voter-registration law), but those are misleading guides to the conditional probability of an unpublished applicant getting an interview (or the job), since by definition those students are exceptional in other ways. This advice applies with special force to students who (like me) aren’t coming from top-shelf Ph.D. programs–programs universally among the 12-15 schools ranked in the top 10. Your brand name may or may not get you added automatically to the discard pile, but unlike H, Y, P, or S, you won’t automatically get added to the long list without something special.”

Note: Things have gotten worse since 2016, so perhaps next year I will redo this post. But it is still a fine guide to much of the basics. Just don’t count on reality being this cheerful.

Link to the post.

Ph.D. Students Need to Publish

Sample: “So how do you publish? You have to be thinking about this from Day One. Learning about what makes for a good (= publishable) journal article, learning about how to submit articles, participating in workshops and courses that lead to feedback that creates journal articles, being around faculty and graduate students who are publishing journal articles, learning the prestige rankings for journals, taking article-sized projects to conferences, attending “how to publish” panels, understanding the risk-reward tradeoff of “starting high” or “starting low” (both in general and for specific projects), and being willing to undergo the harsh and helpful review project…all of that matters.”

Link to the post.

Succeeding in Comps

Sample: “Few parts of Ph.D. education in political science and international relations are as stress-inducing as the comprehensive examinations.Generally, doctoral students value doing well in school, and generally that means they value doing well on tests. For students that fall into that category, the idea that there’s a test which, if you fail, means you might have to leave graduate school is bad enough. More to the point, this is a special kind of test, one that students will likely have to take only once or twice and with a format that can be maddeningly vague. Yet well-meaning people also often advise that you shouldn’t try to do too well on the exam, because you have other things to do.

“High-stakes and ambiguous? That’s a formula for extreme tension.”

Note: This piece was actually peer-reviewed, in that I asked for and received extensive feedback on this essay.

Link to the post.

Writing the Job Application Cover Letter

Sample: “When the search committee does their initial reviews, members will spend less than five minutes per packet. Your cover letter and CV (sometimes only the CV!) will get attention in this round. The search committee’s task is to cut an overwhelming number of applicants into a manageable number. They are looking for reasons to ding you, not to save you—there are more than enough fish in the pond. Demonstrate in the cover letter that you merit further consideration in later rounds of deliberation.”

Link to the post

Georgetown University in Qatar is hiring three faculty members in international politics (and more in other areas).

So if you have the option, why should you come to Qatar?

The country is incredibly, almost stunningly, safe. After a month or so of living here, you begin to realize that Western (and especially American) societies are never on less than yellow alert—you are always on guard for threats to your person or property. Here, you just … aren’t. People don’t worry about theft or disorder, and they especially don’t worry that a crazy driver (there are many) might shoot them. Living in a safe society again (my third one, after Doha 2013-2014 and Shanghai 2004) has done much to convince me that the level of disorder and actual threat that Americans in particular tolerate (and will even, bafflingly, defend) is, actually, unacceptable.

(No, I’m not justifying, say, deploying the National Guard to U.S. cities — rather, I am discussing the entire complex of grime, aggression, and antisocial behavior that gets written off as the cost of living in civilization, or at least America. It need not be! And taking the mitigation of that disorder seriously as a goal might really do a lot for Americans’ quality of life!)

Social services are generally of high quality and almost universally much less expensive than their U.S. counterparts. The medical care is good and disorientingly easy to access; transportation infrastructure (including, now, a metro system) is great; and excellent schools are available in several different national flavors (including American, French, British, and Swiss). In many ways, Qatar benefits not only from wealth but also from the simple recency of its development: second-mover advantage is real, and skipping a lot of twentieth century buildouts has meant moving straight to the twenty-first century with no penalties.

Skeptics of life in the Arabian Gulf countries often lob the charge that life here is “superficial”. This is overstated. To the extent that they mean flashy, well, it absolutely is. But Doha isn’t Dubai, which has made flash part of its brand.

Yes, there’s a lot of luxury shopping here, if that’s your thing (or if it’s your allergy); no, there’s not a lot of heritage sites here compared to similarly populated countries like Uruguay, Armenia, or Lithuania. Qatar has been inhabited for a long time but always thinly relative to its current size. And, yes, a lot of the buildings and cars are new and meant to impress—if you had a gross national product of several hundred thousand dollars per citizen, well, why wouldn’t you let loose on some indulgences?

Against all that, though, are the parts less obvious to tourists, especially those who never leave the shoreline: this is a country oriented around family—principally Qatari families, but also expat families. (I’m going to note here that not everyone is in this charmed circle; more below.) It is also a welfare state designed to funnel investments into health and education. This is not obvious to folks here on a Qatar Airways stayover, but if you are considering a longer-term relocation it is essential. (As a side note, though, family-orientation means if you’re un-familied but looking to become part of a household, you might have a more challenging time.)

True, many leisure activities involve going to the mall. (Boy, there are a lot of malls.) But the malls are nice. That may be hard to believe if you’re coming from a place where “dead malls” rival thriving ones, but it is true. Mall culture here is (with obvious differences) what I remember from the Eighties—except that, well, instead of a middling regional mall many shopping centers are more like luxury destinations, with some moving to a tier somewhat beyond luxury. Functionally, this makes sense: it’s hot—really hot—and malls are great places to go as a family to browse, shop, and eat in air-conditioned comfort. (I see you judging, but let me remind you that this is exactly what American malls were for.) If you view them less as “shopping centers” and more as “walkable destinations”, then you can reorient yourself correctly—and, again, be reminded of the family-centrality of the place.

So, sure, some parts of Qatar’s society are superficial. The same is true everywhere. I do think there’s at least some jealousy in how some folks frame the charge. For others, it is also true that Qatar’s cultural depth can’t match what’s on their doorstep—if you are in San Francisco, New York, or London, for instance. Of course not. But that’s often not the right comparison set—and even if it is, well, see above about crime and affordability.

More on point are those who highlight that, for the bulk of the country’s inhabitants (80-plus percent of whom are foreigners on work visas or residency tied to work), Qatar is not a forever home. Well, no, and that’s part of the deal: eventually, you will leave. Is that really so bad, though? The compensation more than makes up for that, and you can in fact work here for quite a long time — in my institution, many folks are passing the fifteen-year mark, with every apparent intention of staying another decade or more.

Still others might note that it’s not all high-paid work and fancy malls. And, no, it’s not.

The bulk of those who come to work here are not well-compensated expat workers. The issues of workers’ conditions dominate coverage of the country in many outlets, and I won’t rehash them here in part because I have nothing personally to add to what has been discussed and debated at length. I will note that Qatar does, in fact, change, and decade-old news reports are no longer a good guide to much of the country. (For one point, much of the push to build infrastructure that generated so much coverage is done because the infrastructure is now largely in place.) On the other hand, it’s impossible not to be struck by the stark disparities in living conditions.

Nevertheless, conditions here aren’t reducible to simple impressions or a five-day visit. The popularity of locally-made authentically-Italian butter chicken pizza provides, as they say, a rich text:

What quickly becomes evident living here is that secondhand impressions just aren’t as useful in understanding a place as in castigating—or praising—it. The population of the country is far more South Asian than Arab. (I’ve had so many conversations about Bangladeshi politics with Uber drivers that I’ve begun to hallucinate that I’m becoming Tom Friedman.) And I don’t just mean that there are vast but invisible labor forces somewhere, such that this ratio is visible only in the statistics; no, I mean that when you rent a car you’ll probably do so from a Keralan firm or when you go to the grocery store you will see a wider array of nationalities than most universities admit.

Western / North American / Australian expats make up a small share of the community. Even so, English is a lingua franca, even if it’s many people’s second or fourth language; you can get by, perhaps with no small guilt at the monoglot privilege, without learning more than a few phrases in Arabic. But it is not like this is an expat bubble, unless you choose to make it so. It’s a real, living place with all the contradictions and complexities that come with it.

And, yes, you can get alcohol, and you can buy pork, even though both require a license and a trip to the QDC store. (The alcohol selection is simultaneously expansive and limited — craft beer is hard to source, although Chimay is plentiful. Lacking a reliable source for carnitas is harder to adjust to.)

Not everyone will find the place equally welcoming, although members of some communities might want to check with their fellows rather than making assumptions. And for some people, especially Muslims or Arabs worn down by Western hostility or alienation, moving to a place where the Adhan (call to prayer) is audible five times daily can come as a relief.

I like it here. I even like the weather, with the exception of about three weeks during summer. (I have some gripes about the air quality.) I like my institution; I particularly like the eagerness, dedication, seriousness, and variety of my students. (In one class of ten last year, we had, I believe, seven or eight different nationalities represented.) For my line of work (thinking about international relations and U.S. foreign policy professionally), the distance and perspective that working here affords helps avoid the distortions and faddishness of U.S.-based discourses. (Your professional mileage may vary.) Not only is it possible to do good academic work here, but good academic work gets done here. I expect that to accelerate over time.

Yet factors besides the White House may make posts in this region more attractive, and not only in material terms. In many U.S. states, tenure at public universities is tottering; for many faculty at private universities, it is becoming more and more relevant that tenure cannot survive the death of a university. Beyond that, campaigns to reduce the workplace autonomy and intellectual freedom of academics are succeeding quite well in many institutions, both public and private. (If you’re working in a British university, the pitch is simpler: you’re making poverty wages compared to what’s on offer here.)

It is not for everyone, and it will involve some real introspection about tradeoffs. It should, however, be considered seriously—and not just as an “escape hatch” but as a genuinely superior alternative to many posts in the West. It’s time to stop thinking of only “First World” institutions as desirable and instead start recognizing that there is a global market for global talent—and that may entail choices that would have been surprising in the past.

I think my micro-generation (Xennials, or the Oregon Trail generation) was the first for which “technology in the classroom” was a guiding principle of our educators. Now, there’s always some wiseass scholar in the audience who wants to say that “technology has always been in the classroom” and they want to talk about how pens or pencils or flint arrows are “technology”, but we all know that what we mean are computers—the most powerful general-purpose tool of the past one hundred years. At the very least, the idea of bringing computers—physically—into classrooms or classroom-adjacent spaces (the venerable computer labs, an intermediate stage of computational integration now lost to the ages) cannot be much older than I am, because the idea of a computer small enough to be used by five-year-olds is not much older than I am—we are, now, the same age, even if when I was a kindergartner those extra five years seemed like an eternity.

I cannot remember educational spaces without computers, even though in many cases they must have been installed just a couple of years, or maybe even a few months, before I toddled onto the scene. I never really learned to use a card catalog, because my library’s catalog had been digitized; I recall distinctly being in a computer lab with a cluster of PCs—or, more technically, Commodore 64s—as early as first grade, and I remember looking with envy at the students in that lab when I was a kindergartner.

Now, computers are ubiquitous in the classroom. It would be strange to work in a classroom with none. It is not uncommon now for a student in one of my classes to have up to four computers with them in class—a laptop, of course, but also a phone, a tablet, and a smartwatch.

How fantastic—in the old sense—this casual accumulation of compute is. There was a time when smartwatches were purely science-fictional, even pulp-fictional; Dick Tracy had one in the movie I saw (too young) in theaters. Captain Picard and Penny from Inspector Gadget had tablets—one in the far future, the other in a context in which policemen had helicopter rotors in their heads. And we all know by now that our personal computers are so much more powerful than the Commodore 64 that if airplanes had gotten faster at the same rate they’d now be flying shuttle routes to Alpha Centauri, or something.

For all purposes, then, computers in education and I have come of age together. Have they delivered on the gleaming promises that forward-looking teachers and officials subscribed to in the 1980s?

No, absolutely not, and they’re probably mostly a disaster. We’d likely be better off without them most of the time. To the extent they are good—for data analysis classes or rapid consultation of research databases—they are used surprisingly rarely. Far more often they’re used as distraction engines or counterfeiting machines.

They have not even made much of the daily business of administering an educational enterprise better. To be sure, there’s something to be said for not having to work with Scantron bubble sheets or paper gradebooks anymore, especially for those of us cursed with illegible handwriting. As someone who once lost a stack of essays to be graded in the Before Time, I can also appreciate that we can save essays to the hard drive or to the cloud; typewriters are romantically nostalgic for about five minutes, but the first time you realize that cut and paste used to be a descriptive phrase and not a metaphor you’ll be reaching for the MacBook instead.

Yet the actual technology we actually use is mostly subpar in terms of its effectiveness. I’m writing this as a timed exercise—20 minutes on the timer, no more—in an effort to purge myself of the frustrations of working with Canvas. Canvas, as most of you doubtless know, is the LMS (Learning Management System) used by my university and others; there’s also Blackboard, Moodle, and many more in the sector.

The promise of Canvas is that it will bring all of the fiddly back-office parts of running a class into the cloud seamlessly and easily: gradebooks as spreadsheets, readings as PDFs, exams as Web pages rather than bluebooks. In reality, the software is — probably inescapbly — incredibly unwieldy. One example that has driven me to distraction today: there is no easy way (at least in our instance of Canvas) to set up an assignment that repeats daily. In the old days, when syllabi were printed out (and ran to four or five pages), setting up a recurring assignment required typing “Every Wednesday, hand me a two-page paper.” Now, it requires at least—at least—an hour of fiddling and re-fiddling and debugging, because the software isn’t set up to do this sort of task—an absolutely minimal requirement—a push-button exercise.

The tools, in other words, don’t even make the bad parts better. Often, they’re worse. Not a few professors opt out of using them, which is truly damning. Nobody chooses to hand-copy assignments; they just use the photocopier. But a number of professors choose—illicitly—to run their courses from personal Web sites or Dropbox accounts or anything other than the LMS.

Share Systematic Hatreds

So why use them? Well, we are trapped by our tools at time, but we are also trapped by the expectations of students and the policies of our employers. The latter is not a small obstacle, by the way; if you post links to databases with readings through the official LMS, you are not infringing copyright, but if you host a reading on your personal Web site, you definitely are. And if you let personal information leak from your personal site, you’re guilty of transgressions twice over (and, I must say, justly so—the data isn’t the instructor’s and no instructor has the capability to actually secure their site in an era in which university records are the targets of, say, political hackers going after college admissions essays).

The former, however, is also important; we are now at least fifteen years into a transition in which LMSes have been creeping into high schools, elementary schools, and even nurseries. Students and their parents alike expect that this is how it will be. Doing anything else would require retraining the students’ expectations while looking like a weirdo and inheriting all the risks of a paper world (like, you know, losing stacks of exams or having homework devoured by ravenous canines).

That digitalization looks modern and efficient but it often is anything but. It takes me hours longer to create a course shell than it used to and the results are not appreciably better; indeed, because of shortcomings in the tools, I find myself frequently battling between adapting my course to the machine or bashing the machine into doing what I want.

It does not help that LMSes are walled gardens, and weedy ones at that. Much of what I want to do could be done easily in a vanilla setup but everything that’s been added for commercial or regulatory purposes gets in the way of the simple business of sharing assignments and readings. And, note, that none of this is doing anything but replacing, at great time and increasing expense, what used to be done without paying licensing fees or technicians to assist. Somehow Feynman didn’t need this (or if that’s too strong pick someone who was an actually good undergrad instructor).

The story repeats in many scenarios and applications. For every minor friction that technology has removed, a giant weight of superfluous expectations has been added. If the net impact is still positive, it is much less so than the boosters promised—so much less so that one wonders whether we would have gone all-in on this. And I also find myself wondering if at least some of the lack of grit that many of the younger generations display when confronted with an obstacle has to do with being raised in an environment in which everything just works—where you never have to slap the side of a TV to fix the image, or when you never have to parse “Abort/retry/fail” when trying to load a program. To the extent that computers do make everything “just work”, they do so by bounding the ability of users to ask for anything — limiting their requests, and their imaginations, to what’s easiest to deliver. But that is far from what’s possible and often remarkably distant from what’s desirable in any given case.

I am not an “opponent” of AI; I don’t even consider myself a “skeptic” of LLMs. I think that they have many more use cases than the most overworn cliches (“they’re just autocorrect!”) admit. But I also wonder whether moving wholeheartedly into adopting LLMs in the classroom is a triumph of hope over experience. Is it really likely that the LLM will be able to contribute more than, say, Microsoft Encarta CD-ROMs did to a fourth-grade classroom? Or is it more likely that introducing a giant pipe into the oeuvre of human kind is at least as likely to convey sewage as anything potable? If we can’t even get an LMS vendor to deliver “repeat this task every class day” as a normal setting — if there’s not even a feature to auto-populate a syllabus with every day a class will meet in the LMS; you know, the sort of boring database management computers are unparalleled at — then why should we expect that widespread LLM adoption will end up as anything less than an edutechnical innovation that impresses the rubes and fails to move the needle on actual outcomes?

I do think there are paths available to a better outcome, or sets of outcomes, but like all such paths they mostly involve tradeoffs from the bright shiny sales pitch. Do you really think you get all the advantages but with no extra costs? Would LLMs look as appealing if we described them as a useful tool that requires (as they do!) extra human involvement and mentoring to make use of their full potential—and that the time measured to reach those goals would be measured in years of careful engagement?

The example of computerized classrooms suggests that thoughtfulness will not be the distinguishing characteristic of what happens. It is even unlikely that those “at the coalface” will be consulted or valued as policies are made; the managerial-consultant class of higher education is quite happy to tell line faculty what the line faculty should think about these exciting new opportunities. If one is hired to be a consultant on these issues, after all, one is selling the ability to make LLMs a part of the classroom, not to judge whether they should be kept, like predators, a thousand feet away from students at all times.

At root, of course, is that all computing technology since the 1980s has been sold as a way to bring tools into the curriculum but those efforts have mostly brought toys into the classroom. (That my generation is known, in some circles, after a computer game, The Oregon Trail, which famously failed to teach any of us anything about the Oregon Trail itself is a grim irony.) As tools, computers are godsends; I am much more happy with LLMs in my mode as a researcher than as a teacher. As a classroom manager, I find myself pining for the days when I just needed an Excel spreadsheet and a stapler to manage my students. And as an instructor, I wonder whether there isn’t something to be said not only for bluebooks and pens but for slide rules and chalk—analogue tools to develop organic intelligence.

“Today is August 5, 2026, today is August 5, 2026, today is …”

So ends Ray Bradbury’s “There Will Come Soft Rains”—at least in one version.

The short story has been, as they say, widely anthologized, but it was first published in the May 6, 1950, issue of Collier’s magazine.

Collier’s was a big deal at the time—a magazine with, at the time, more than three million issues printed each week. Like other publications, it dealt with nuclear war as one of the hot topics of the day. And like every publication with an ambitious editor and red on the ledgers, it sought not only to discuss news but to make news. A year and change later, Collier’s would publish “Preview of the War We Do Not Want”, a special issue discussing what a nuclear war “would look like”. Alex Wellerstein has written about the famous photomontage paintings Chesley Bonestell used to illustrate the issue, most notably the cover of Manhattan being destroyed by an atom bomb. A work of speculative fiction and propaganda and what Peter Galison would later call “state science fiction”, that issue was an Event.

“There Will Come Soft Rains”, by contrast, took up a single page. It has proved more lasting.

It is always interesting to see the original versions of texts—how texts first appeared before their canonization. What would later be an illuminated manuscript was first a humble codex.

This was not a highlight of the issue, even though it is the piece with by far the largest legacy from the issue (or maybe the whole run of the magazine). For me and the other folks who have read it, it is a story defined by an arresting image—and yet the illustration here is simple to the point of crude. No special photomontages by Bonestell here.

(Note that in the original printing, the final line is “Today is April 29th, 1985”—and in later printings, Bradbury would set the story in 2057.)

One page, a small illustration, and a long shadow. “There Will Come Soft Rains” is about people, and it is about nuclear war, but we see neither, because both are in the past of this story. The image that sticks with readers is the shadow:

The entire face of the house was black, save for five places. Here, the silhouette, in paint, of a man mowing a lawn. Here, a woman bent to pick flowers. Still farther over, their images burned on wood in one titanic instant, a small boy, hands flung in the air—higher up, the image of a thrown ball—and opposite him a girl, her hands raised to catch a ball which never came down. The five spots of paint—the man, the woman, the boy, the girl, the ball—remained. The rest was a thin layer of charcoal.

I do not know the date I learned about the shadows cast by atomic explosions. I must have been very young—perhaps five or six—and I recall where I was: in a children’s museum that had a special area set up so that when strobe when off, our shadows would stay on the wall. It was not an exhibit about nuclear weapons—but someone observed that it reminded them of the shadows cast at Hiroshima.

Many years earlier, the shadows cast by the bomb were common knowledge. Here, for instance, is The Cincinnati Post from September 1946, reprinting John Hersey’s Hiroshima from The New Yorker:

“A few vague human silhouettes were found,” Hersey writes, “and these gave rise to stories that eventually included fancy and precise details. One story told how a painter on a ladder was monumentalized in a kind of base relief on the stone facade of a bank building on which he was at work, in the act of dipping his brush into his paint can.”

If Bradbury’s image owes more to such tall tales than to the reality—closer to person-sized smudges—it is no less radical in that it depicts Americans as suffering the same fate an American bomb had caused. And it is still more radical in that the war simply … comes. There is no point, no reason, no meaning. It comes, it destroys, and nothing is left.

Pretty heavy material to fit into one page—not least given that on the next page, there is a full-page ad promoting beer. Not any particular beer: the concept of beer—”America’s Beverage of Moderation”. Like I said: it is always interesting to see the original versions of texts before they were canonized.

The stereotypical Fifties was a decade characterized by a zest for consumer technology and fears of nuclear war. Bradbury implicates both.

It is the illusory comforts of consumer technology that takes center stage in the story. The “house of tomorrow” gimmick that furnishes him with his main character—the electronic house that cooks, keeps a calendar, and cleans itself—is, if anything, vastly easier to imagine today. The house is here, just unevenly distributed. The cleaning mice? Roombas. The playroom full of screens? Too real. The AI that chatters to itself? Twenty years ago, we would have called that science fiction, but now it’s annoying fact. In fact, to at least some contemporary undergraduates, it is a little hard to convey how futuristic the house was meant to be.

It is just as hard to convey how close nuclear war is. Fears of an end of the world are proportional to the symbolic centrality of an apocalypse, not to its actual risk. Drop yourself into medieval Europe and the revelation of St. John would be more pronounced than it is now; go further back and further north and you’d hear about Ragnarok.

Ironically, this means that nuclear war seems more distant now when it is, in fact, more plausible than when Bradbury wrote.

There are far, far more nuclear weapons—and more powerful—than there were in 1950. And yet we don’t talk about it, or think about it, as much as Americans did then. Over the next few days, the anniversaries of the bombing of Hiroshima and Nagasaki will call it back to mind, but by now even those commemorations are almost occasions for re-presenting the same evidence and the same arguments. And neither incident would help us truly visualize a modern nuclear conflict—one in which dozens or hundreds of larger weapons are used.

Imagination cannot scale atrocities. Bradbury used the story of a single house to represent the end of the world for a reason.

I have read “There Will Come Soft Rains” repeatedly over the course of, perhaps, thirty-five years—first, I believe, in an elementary school textbook. I will be teaching it this coming term as part of the course on The Politics of The End of the World. Part of the goal—no spoilers, i’m transparent—is to help folks learn to unpack an argument presented in narrative form. Another part is to help folks learn about different genres and invite them to consider other perspectives.

But a big one is to make nuclear war real and scary again. “Soft Rains” is, among other things, a horror story; it is not really an apocalyptic story because there is no revelation of a deeper meaning—there is only the absence of meaning. It is frightening to consider that all this could be destroyed in a few hours—or, worse, that we have to survive in a world in which a third, or a tenth, city joins Hiroshima and Nagasaki. What other reaction than terror can there be?

Historically, there have been quite a few other reactions, of course—awe, elation, curiousity, interest, greed. But a worryingly frequent reaction I encounter now seems to be something closer to numbness, dullness. Apathy is not a spur to action. Maybe a 75-year-old short story—a story so old its far future is now almost here—still has the power to startle.

How are you goin' to interest our young men in their country if you have no offices to give them … ? It can't be done. These men were full of patriotism a short time ago. They expected to be servin' their city, but when we tell them that we can't place them, do you think their patriotism is goin' to last? Not much. They say: "What's the use of workin' for your country anyhow? There's nothin' in the game."

— George Washington Plunkitt, as told to William L. Riordon (Plunkitt of Tammany Hall)

I sinned in this newsletter even before my voice started to narrate my thesis. The block quotation above, taken from one of the finest pieces of political prose the United States ever produced, has been altered in a way that changes the meaning of the sentence — the ellipsis obscures that Plunkitt was talking about men who worked for the party, which is to say the Tammany Hall Democratic Party. These are not young men who deserved government office according to some Prussian ordering principle or some Confucian idea of merit—these were the door-knockers and street-brawlers and vote-getters who turned the Germans and Jews and Slavs and Italians into majorities, at least in paper. In that world and for those men, politics wasn’t about ideals or principles; it was about power and relationships, a relation even more feudal than the Middle Ages were, where a ward heeler owed the boys who got the votes jobs paid for from the city treasury.

Few worlds could be farther apart than the street politics of George W. Plunkitt, senator of the state of New York, and the idealisms of the many students who sat in the chairs within the cinderblock walls of my office in Amherst. Those students wanted careers, not jobs, and they wanted to pursue lives of service. Still, the office they met me in was named for John Thompson, a Massachusetts pol known for his alcoholism, his dominance in the state House, and his willingness to take bribes. The building that bears his name houses the school of public policy, the political science department, and a number of other programs dedicated to clean government.

Not only would Thompson and Plunkitt have found many common interests, both would have recognized a familiar ambition in the young folks I counseled. Public service is a noble term and anyone who, at the age of 20 or 21, is seriously considering public service is doing so at least partly out of the belief that they can make the world a better place. To do that, however, the young folks who talked to me needed a place—that is, a job. (There are young people who believe they do not need a paycheck and a 9 to 5 to make their way transforming the world, but they do not seek out advice from the middle-aged guy with the Nixon re-election poster on the wall.)

Over the years, I’ve come to know and advise, to the best of my ability, students who were destined to lead platoons in the demilitarized zone on the Korean Peninsula and students who wanted to become diplomats working with the State Department to advance women’s rights around the world. I’ve talked with folks from Communist countries and with trenchant Libertarians, with those whose parents were senior civil servants and those whose parents worked late hours in manual labor. Some have wanted to rid the world of nuclear weapons, and some have wanted to make their hometown work better. (One was a mayor just a few years after coming to office hours—through no help of mine, I’m afraid.)

All of them wanted to make a difference, and all of them thought that a career in public service could help them.

I was happy to share whatever I knew. For some careers, like those in electoral politics, I had more to offer; for others, I could provide letters of recommendation. I eventually bought a guide to careers in international relations that I knew less about, like USAID and public health. I loaned it out to students so they could get a sense of what they could expect from pursuing graduate education and a career in those areas.

That world lies in tatters.

The past six months has gutted the federal civil service. My LinkedIn feed is full of first and second connections talking about being terminated because they were detailed to the wrong office in the State Department at the wrong time. Other connections worked for USAID or the Education Department. It’s trite but true that these same connections, back in January 2025, were starting to talk on those same LinkedIn feeds about how they would be working on a nonpartisan basis to support the incoming administration. They really believed in the notion of public service, and the public had voted for the administration, so they would work for it. (I mean, let’s not be too naive—they weren’t going to be happy about everything, and some were probably already walking out the door—but let’s also not be too cynical about this; these were people, after all, who had already served one Trump administration.)

It is not clear that I could counsel students regarding the attractions of a career, or even a stint, in public service the same way as I did before 2023. Yes, 2023, because once it became plausible that Trump would win again (and it was 50-50 from the moment he became the GOP frontrunner), I started emphasizing the risks a lot more and talking about the beauties of earning a paycheck in the private sector while the public sector sorted itself out. Even there, I was too optimistic—I did not think that USAID would be dissolved rather than cut, for instance.

Occasionally, the advice I gave weighs on me for how airily I discussed some of the risks. Melodramatically, I remember some lines:

All that I have said and done,

Now that I am old and ill,

Turns into a question

till I lie awake night after night

And never get the answers right.

Did that play of mine send out

Certain men the English shot?

Yeats could dwell on whether he had played too great a role in bringing about civil war and bloodshed in Ireland—my guilt, whatever there is, is lighter than that, but still weighs around my neck.

Whatever guilt I bear, of course, is nothing compared to that born by the officials who decided to eviscerate and terrorize civil servants. More than the damage they have done to those in the federal ranks is the wreckage they have left where once resided the idea of public service itself.

When I met with those young people, I was an employee of the Commonwealth, as all Massachusetts public university professors technically are. In that position, each year I had to take a quiz on my ethical obligations; I could not accept a gift worth more than $25 (or was it $50?) from anyone having business with the state. As an alcohol commissioner of the Town of Amherst, furthermore, I’m subject to a lifetime ban on ever owning a bar in Amherst.

In an earlier incarnation as a federal employee, I once received a gift from a vendor we had done tens of thousands of dollars of business with; a coffee mug, some ground coffee, and a biscotti. I was terrified that this would come to the attention of federal ethics officials, so I made sure to leave it in the common area.

These are the day-to-day considerations of an American civil servant.

Contemporary senior executives and other officials in the U.S. federal government operate under no such code, legal or otherwise. Scandalously, the U.S. Supreme Court has all but legalized bribery if it is performed as “gratuities”—that is a sentence that sounds like a shrill liberal exaggeration but it is in no way a misreading of the Court’s holding last year.

If holding public office is a route for private benefit, what keeps any official aligned with the common interest?

(In the interests of bipartisanship, I should note that not a few Obama and Biden-era officials also cashed out, and I scorn them, too. But I defy anyone to say that there’s a serious comparison to be made between the worst excesses of their administrations’ personnel—which, again, I condemn, easily and frequently—and the ordinary behavior of the current administration.)

And so we arrive back at Plunkitt’s world. Plunkitt decried how civil service reforms had cheated deserving Tammany men of jobs that had instead gone to Republicans simply because they possessed qualifications. The U.S. situation today is worse. Qualified young people trying to be hired into the civilian federal government already faced long odds before the Trump administration, but once they got there they could imagine themselves actually doing something like the people’s work on a daily basis. Now, jobs have been cut; the mission is weak; and there’s little guarantee of any future with an agency because there’s no guarantee an agency has a future.

Some, still, on social media will encourage young people to pursue dreams of careers in public service because the country needs their work. Yes, and officers at the rear in 1916 encouraged young men to go “over the top” in the fields of France, fighting bullets with their chests. It’s always easy to fight your wars with someone else’s lives. That makes it no less irresponsible to urge folks to go into agencies where they might face lifelong consequences should they be caught up in a purge or worse.

The grim truth is that for the next few years it is probably best to encourage qualified and talented undergraduates to avoid directly engaging with public service at the federal level. They can pursue related careers (although cuts to contractors have also made that more difficult) while they wait for sanity, or at least certainty, to reassert itself. Like a resistance movement, they can strategize, train, and wait. For a few years, perhaps, municipal and state government will attract a higher caliber of students than might have otherwise been the case.

At the doctoral level, the situation is grimmer. I’ll talk more about this later, but in political science and related fields the argument for doctoral education has always been that if one does not find an academic job (or does not want one) then one could always work in think tanks, consulting, the agencies, or other policy-related work—what people generally refer to as alt-academic work. Well, the academic side of the equation looks grim, as universities enter at best a period of austerity, but the alt-academic side may be in for an equally rough period. Duck and cover has a great appeal here, as well.

Beyond such practicalities, there is a larger question. Plunkitt’s point about the demise of patriotism because of a dearth of jobs was funnily put, but there is a nugget there: why should anyone expect a state that has broken its promises and ruined its own foundations to be a reliable employer, much less one deserving of loyalty?

Not only is this bad for my discipline (why major in political science if there’s no jobs in government?), it will also be bad for the country. On a working level, the attractions of federal service become even less when your boss and your boss’s boss care little about, say, saving the lives of victims of flash floods or anything else that distracts from lining their pockets. All of this is compounded when one thinks that a great many younger folks might now think of the federal government not as a vaguely if inconsistently benevolent force but as one that targeted their family for deportation or their gender for persecution.

There is a reverse Gresham’s law—bad federal employees will chase out the good. They may even do so actively, through purges like those we have seen already. One or two college classes skipping out on federal service can quickly become four or five, and ultimately the transmission of affection and knowledge that makes any institution effective will be threatened across wide arrays of the bureaucracy. What advice am I supposed to give about how to cope with that?

One of the idioms I picked up during my time in Ireland, now more than twenty years ago, was the expression “at the coalface”. If memory serves, I learned it from a presentation about the process of touting Irish bonds to investors by civil servants at the Department of Finance, which is somewhat incongruous given that the expression hails from the experience of coal miners—the folks who were at the business end of extracting lumps of carbon from the ground to turn into industrial revolutions above.

It’s a useful expression, even if those of us who employ it are nowadays describing stressful interactions with a laptop rather than the body-breaking work of carbon extraction. It highlights the experience and details of what might otherwise appear to be disembodied abstractions. When I say, for instance, that I’m spending my summer doing research, course prep, and administrative tasks, the abstract content of those duties has hardly changed in decades, but the details—ah, the details are changing faster than a log tossed into a blaze becomes ash, and with the same degree of enjoyment.

There are five trends most directly affecting scholarship in U.S.(-affiliated) higher education:

• profound financial instability. The bad news is that the short term looks grim as a result of slashed federal research funding. The worse news is that the long-term outlook makes the short term look like a pleasant vacation, not least because of changes to federal funding for student loans: there seems little chance that enrollments will stabilize, graduate education (the kind that brings in substantial revenue) is about to be slashed, and international students aren’t going to come in anything like the numbers necessary to maintain institutional size. Even for institutions with large endowments, a collapse in revenue from tuition streams will lead to substantial and painful retrenchments. (If you’re still talking about the “demographic cliff”, you are behind the times—the threats are much more sudden and acute than simple changes in cohort sizes, and the current administration is eviscerating all of the tactics that could have been used to mitigate its effects.) You may agree that some of these retrenchments are necessary (I would never, ever defend PLUS loans) but they are being made in the worst possible way by the least qualified officials—you can imagine that a stubbed toe could be treated by amputation, but you would probably also wonder about the doctor who prescribed that.

• substantial attacks on institutional autonomy. The use of federal agencies against universities is being joined by more direct attacks led by state governments like Florida’s and Indiana’s and flanking efforts such as the alumni-led effort at the University of Virginia. This does not affect only administrators but translates into curricular changes and into less overt means of surveilling and disciplining faculty, such as Ohio’s requirement of publicly posting syllabi and Texas’s requirement that faculty work 40-hour weeks in their offices.

• undeniable changes to student preparation. Even though there’s another round of gaslighting that seeks to counter faculty laments about students’ difficulty carrying out basic tasks (such as adhering to deadlines or reading at or near grade level), I can assure you: the changes in any given classroom are real and profound. What was once a standard set of expectations for student performance would now lead to sustained complaints from American students.

• the assault of AI on traditional teaching. The changes posed to traditional means of learning and assessing learning by the widespread (essentially, universal) adoption of LLMs in the classroom have made pre-2022 assessment practices obsolete, even misleading. Almost all suggestions about how to address these come with a price tag—increasing time spent on assessment and interaction—with no corresponding shifts in research expectations or compensation. Even if this were the only change taking place, it would be dizzying.

• geopolitics and ordinary domestic politics. MAGA’s agenda is squeezing the United States out of scientific leadership. Researchers do not want to go to the United States. Non-citizen scholars present in the United States feel trapped—leave the country, even for a conference, and you risk detention or worse. And the growing relevance of borders to everyday life is making fraught actions and collaborations that were once natural.

How do all of these manifest at the coalface?

• Although I’d like to pretend that every academic is a fearless intellectual pursuing theses wherever they may take them, it is also hard not to feel a chill when considering writing on or speaking about topics (and please bear in mind I am now talking about U.S. politics here, and no other country’s). Equally, although I’d like to pretend that every academic is in the research game for the love of knowledge alone, it is very difficult to ignore the fact that the traditional rewards for research and teaching—raises, grants, acclaim, stable employment, etc—are steadily being chipped away, in some cases to nothing.

• It is hard to find ways to teach that one can be sure will actually reach humans—and if we can’t do that, then in the medium term the entire sector risks collapse. It is hard to teach if students cannot read books, full stop, and universities—for obvious reasons!!!!!!—are not set up to provide remedial reading instruction.

• It is difficult to arrange scholarly talks and visits if people cannot leave the United States (or, I presume, if they do not want to visit the USA). These are important circulatory functions that link research communities together, and cutting U.S.-based scholars off from global conversations will weaken everyone.

• As someone who teaches U.S. Foreign Policy, I can also attest that it is literally difficult to know what to teach. The Trumpian assault on foreign policy agencies has vaporized entire standard topics in this course (consider, for instance, how one would teach U.S. foreign aid, the interagency process, or relations with allies). In an intellectual sense, as ever, I find this remarkably exciting:

But as someone who needs to actually assemble a syllabus it is literally difficult to do so because I do not know how to explain what is happening. Imagine—and this happened to colleagues of mine—teaching Soviet Politics in the Fall 1991 semester while the U.S.S.R. dissolved. Interesting! Historic! Damn hard to mange a class under the circumstances!!

• Bluntly, it is just harder to do everything and everyone feels perceptibly sadder about doing anything. The vibes…the vibes are very bad. (Ironically, on this point, I’m fine, and my current country is supporting research and education quite well.)

At the moment, in my position, I only feel the incidental effects of the assault on U.S. higher education—the difficulties of traveling and connecting, the complications of navigating suddenly politicized spaces, the questions about how to advise students who might wish to pursue a PhD in the United States. My position seems like a refuge from many of the most acute problems my U.S.-based colleagues are facing. And some disciplines are facing these questions more acutely than others. (I will note, however, that discussions about the effects of LLMs and how they have changed the scholar-student relationship are all but universal.)

I find it hard to explain to people who are unfamiliar with academia or how scholarship actually works how devastating these changes are. These disruptions, cumulatively, seem—from my perspective—to be more dramatic in their effects in a far shorter time than the impact of computerization and the Internet on higher education. Moreover, there does not seem to be any reason to think that U.S. policymakers are concerned with, or even sad about, any of these changes. To the contrary: they are pouring gasoline on the flames. Warnings about the risks and long-term effects fall on ears deafened by an ideology that says that those consequences are desirable or by an incapacity to imagine that actions have consequences.

These are not challenges that can be dealt with individually or easily. Glib prescriptions that instructors should “just adopt AI” do not address the potential that either AI renders the sector as we know it unnecessary or that it simply wrecks a lot of how education works in the short term. In the long term, maybe we will adapt, but in the short term we need to pay the bills. Further, I described the armchair recommendations about how we should incorporate AI as “glib” prescriptions, but … they are all glib. They are not matched with resources or based on solid evidence—there literally hasn’t been time for the evidence to emerge! None of the prescriptions about responding to the real, practical, day-to-day challenges of giving everyone a counterfeiting machine are anything but glib—at a minimum, they involve a degree of retooling and rethinking that would dwarf the move to online courses during Covid, with a corresponding investment of time and effort, and none of that is going to happen.

Nor, especially, can they be addressed by American researchers moving abroad. When someone suggests that researchers just move abroad, I smile wanly, because I know that whoever is suggesting it has no idea of the scale of U.S. higher education relative to other markets nor of the difficulties other countries are facing. (What, just move to the many financially secure institutions of the United Kingdom? If you don’t keep up with UK higher ed news, then you may not recognize that’s a darkly bitter joke.) Some folks will find moving abroad attractive but it is simply not possible for the 1.4 million U.S. post-secondary instructors to just move to Canada.

In the meantime, I am still at the coalface, working the black seam. I still wrest nuggets of wisdom from the earth and bring them back to the wider world. I adapt, I pivot, I hope that we get through it. Focusing on the work makes the challenges more obvious but also makes the rewards more tangible.

Today’s post will be a little different than you may be used to, although longtime readers will see that it brings together data, Qatar, and higher education—three major themes of this newsletter. New Y Combinator readers may find this more or less interesting …

This is a story about how I came to create a dataset of car prices in Qatar, and why it matters. (For readers just looking for the GitHub link, the dataset is here; for R users, see ‘s R version.)

Many statistical instructors, students, and analysts are familiar with the auto dataset (the one you call in Stata with - sysuse auto - ). Back in 1979, Consumer Reports, an American product-review magazine, compiled data about automobiles available in the United States. A few years later, a graphics textbook used the dataset to generate different examples of statistical graphs.1 I say “used” even though, back in 1983, anyone who wanted to use that dataset would have first have had to manually enter the data into their program so they could make this

into charts that looked like this:

The authors recognized that the dataset had ample pedagogical value. It combined a mix of continuous variables, like price and fuel efficiency, with categorical variables, like whether a given automobile was made in the United States or abroad. It thus allowed them to craft realistic and relevant examples for their audience.

The dataset was perfectly timed to be included with a then-new suite of statistical analysis software: Stata. As a result, introductory stats classes and advanced tutorials alike are full of references to the dataset. Want to explain how to graph a scatterplot and a linear regression line? Dead simple:

 twoway (lfitci mpg weight ) (scatter mpg weight ) 

As you might have guessed, the heavier a car, the lower its fuel efficiency! (Really low—this was the 1970s after all.) Want to show that fuel efficiencies differed between U.S. and non-U.S. cars? Simple(ish)!

twoway (lfitci mpg weight if foreign == 0, color(gold) fcolor(gold%20)) ///
       (lfitci mpg weight if foreign == 1, color(purple) fcolor(purple%20)) ///
       (scatter mpg weight if foreign == 1, mcolor(purple)) ///
       (scatter mpg weight if foreign == 0, mcolor(gold))	///	   , legend(order(1 "Not Foreign" 3 "Foreign"))

You get the idea—it’s a workhorse dataset. that can show basic relationships and provide useful examples readily.

Now, if you’re an American, you probably haven’t noticed anything weird about this. If you’re not American, and especially not Canadian or British, you’ve probably nodded along politely and gotten the gist of what I’ve said, but you’re also rolling your eyes.

Why? Because this dataset isn’t just old—it’s so America-centric that it’s basically as if Ron Swanson had gotten a job teaching data science:

Where do we begin?

• Price is expressed in U.S. dollars (1978 dollars, at that)

• Fuel efficiency is measured in miles per gallon

• Headroom and length are given in inches

• Trunk space is given in cubic feet

• Weight is given in pounds (abbreviated as lbs.)

• Turning circle is given in feet

• Displacement is given in cubic inches

• The world is described as foreign or not-foreign

When I taught stats in the United States, I didn’t notice any of that. However, now that I live in what the auto dataset would describe as “foreign”, I’ve come to grips with the fact that other countries—aside from Liberia and…Myanmar?—have yet to adopt Freedom units:

Teaching stats to undergrads is tough. It involves conveying content knowledge (how do we analyze data?) as well as procedural knowledge (how do we use these tools to analyze data?). Inevitably, it involves conveying substantive knowledge —even a simple chart like this one:

involves explaining a) the construction of the V-Dem women’s empowerment index, b) the measurement of GDP, c) the construction of purchasing power parity indices, and of course d) what the linear relationship is.

When I’m teaching political science students about politics, of course, I can justify spending the time on explaining those constructs. What I can’t really justify is spending the same amount of time teaching what an inch is or what miles per gallon is—because that’s just getting in the way of my job. And, no, basically none of my students knew any of these relationships in their gut—which is a problem when you’re trying to use a dataset to motivate interpretation of relationships!

That is: the same qualities that made the 1978 auto dataset attractive to American statisticians in 1983 made it really unfriendly to students outside of the United States in 2025.

So one of the first things I did this past spring term, when I was teaching stats in Qatar for the first time, is create the Qatar Cars dataset. (Okay, technically I had the students create part of it for an in-class assignment, but I edited it and finished it.)

The Qatar Cars dataset is a listing of prices and other specifications taken from YallaMotors Qatar , mostly in January 2025. It includes almost exactly the same variables as the auto dataset:

• origin (the country associated with the car brand)

• make (the brand of the car, such as Toyota or Land Rover)

• model (the specific type of car, such as Land Cruiser or Defender)

• length, width, and height (all in meters)

• trunk capacity (measured in liters)

• fuel economy (measured in liters per 100 km)

• performance (time in seconds to accelerate from 0 to 100km/h)

• mass in kilograms

• price in Qatari riyals

• and engine type (electric, hybrid, or petrol)

Most of the relationships are similar, as you might expect: fuel efficiency gets worse as cars get heavier:

 twoway (lfitci economy mass ) (scatter  economy mass ) 

but because of the units involved, the slope is reversed. And it has advantages, too. This being Qatar, the most expensive cars in the dataset are real head-turners: the Bugatti Chiron and Bugatti Centodieci, comfortably in the eight-digit riyal range. Want to teach people about logs?

 twoway (qfitci log_price performance) (scatter log_price performance) , ytitle("Log Price (QAR)")

(And, yes, the relationship is more nicely linear if you restrict the price range under about 500,000 riyals.)

And if you teach in an international classroom, like I do, you’ll appreciate that the foreign/non-foreign binary has been replaced by a simple list of cars’ origins: Germany, Italy, Japan, PR China, South Korea, Sweden, UK, and USA.

As a further advantage, the cars in this dataset are actually under production. No more “AMC Pacer”, “Dodge St. Regis”, or Datsun anything—we’ve got MGs, Porsches, Cherys, Kias, Teslas, and Volkswagens. You know: cars students might actually see outside of a museum. And the prices are in line with current global car prices—no more cars costing $5,000. (1 USD = 3.64 QAR; 1 EUR ≈ 4.15 QAR.)

I am not always a fan of the term “global majority”, but when it comes to the use of the metric system it’s simply descriptively accurate. Datasets meant to help teach should actually help teach—not require an exercise in translation.

The Qatar Cars dataset is available under an open license at GitHub — no retyping required. Update 9 July 2025: has put together an R version!

I spent most of last week traveling to and from the United States for the annual meeting of the Midwest Political Science Association, which, despite its name, is a major national (even international!) gathering of political scientists. I should probably say “was”, because the meeting was noticeably subdued from prior years—there were still international folks there (I count, now!) but there were a lot fewer of them. The conference circuit is already feeling the Trumpian pinch, and it will only get worse from here.

I wanted to share a couple of thoughts with you in lieu of a bigger update. As you’ve probably heard, universities in the United States are in a tough spot right now. I’m not going to discuss the political issues or adjudicate different claims about some of the causes and putative causes—that’s a 2024 discussion. Right now, in 2025, the important thing to know is that universities make their money from three big streams:

• tuition

• sponsored research

• endowments

All three of these are in crisis.

Let’s start with tuition. The growing xenophobia of the U.S. government and the heated anti-foreigner rhetoric of the U.S. government is deterring tourism and study in the United States. International students enrich the experience on many campuses; they have also come to form a major part of many universities’ business models (and even more so for specific types of degree programs, particularly business schools). Further, they tend to be disproportionately likely to pay full-freight tuition—far from crowding out U.S. students, international students have subsidized offerings for American students.

At this point, I would not be surprised / I would even expect that international student enrollment will plummet. Not only has the Trump administration all but said they don’t want foreigners to come to the USA, they have directly and extensively embarked on campaigns yanking student visas—in some well publicized cases, stopping students’ ability to finish doctoral degrees weeks before completion or nabbing students off the street. I want you to understand how bad this looks from abroad. Would you pay $80,000 per year for your child to study in, say, Romania under Ceaușescu? Or Belarus today? At the moment, the appearances are that international visitors and students are not protected by the regime but are targeted by it—that even getting a traffic ticket can lead to self-deportation. The risks may be small, in that “only” several hundred students had their visas revoked last week, but if you are embarking on paying—again—several hundred thousand dollars for a degree, exactly what degree of risk are you willing to take? Particularly if you view that as an investment in securing employment in the United States afterward—a route that now seems also unlikely.

The attack on research is more obvious; you’ve seen it discussed a lot everywhere. In case you haven’t, you should also know that top universities do a lot of research for the federal government through grants from the National Institutes of Health, the National Science Foundation, and other agencies (including the late USAID!). The funding model was set up after the Second World War to make U.S. universities world leaders in research; it is now being dismantled with extreme prejudice. As I understand it, the threats to biomedical research—you know, a major emerging industry in which the USA has come to play a leading role—are particularly acute because the funding model for those schools is particularly reliant on federal grants.

The other day, on Substack, or perhaps the comment section on Marginal Revolution, I saw someone scoff at the idea that going after these funding streams could really dismantle U.S. research leadership so quickly. If a few weeks of Trumpian chaos could harm universities so much, they mused, were U.S. universities really that good to begin with? All I can say in response is that this is like suggesting that if someone dies after being shot in the head, how can we really know they were healthy to begin with? The assaults on funding are, yes, that bad—scientific research depends on stable funding to keep projects moving and specialized workers paid; without that stability, you can’t—literally—keep refrigerators on or monitors going or key personnel engaged.

Another fun thing is that foreign students, and yes that includes Chinese students, are key to U.S. research leadership. Bringing together the top minds of the world in the United States has been a constant boost to U.S. technological and cultural leadership. This is a profound gift the United States was given in the 1930s as a result of, essentially, Hitlerite violence—and it is being squandered. The damage already done is profound, but the damages will continue to accrue over years and generations as researchers adjust their expectations regarding the political stability of the United States and favor other destinations at the margin instead. For instance, China has genuinely good universities now, and instead of promoting a brain drain from China the policies of the Trump administration are now doing more than any CCP-led policy to promote Chinese talent remaining at home.

Finally, there’s the endowments.

People love to talk about universities’ massive endowments. In reality, only a few dozen universities have substantial endowments at all; that tends to be nearly every university you’ve ever heard about, but for scale there are (depending on definition) three or four thousand institutions of higher education in the United States. Now, to be fair, U.S. world leadership in research depended on precisely those major institutions with large endowments, so they are disproportionately important, but unless you live in Princeton or Cambridge, MA, the nearest university to you likely is running tuition check to tuition check.

A quick rule of thumb for endowments: they are set up to provide about 4 to 5 percent returns per annum. That is: the classic endowment exists throw off money for a given purpose, determined by the donor, by providing returns, rather than by acting as a bank account. That’s why a hundred-million-dollar endowment isn’t really as big as it sounds—it’s not something from which you could extract $100,000,000 of value tomorrow, but it is instead just a way of generating four million dollars a year. That is roughly (very roughly) the budget for my previous employer’s political science department, not counting overhead or anything else. In other words, if you have a university with an endowment of $300 million or $400 million, the value of that endowment in yearly terms is roughly the budget of, say, a midsize city’s YMCA budget—not nothing, to be sure, but also not exactly a princely sum.

That background is important because I want to explain how grave attacks on universities by withholding federal funds are. The Trump administration famously sought to withhold $400 million in grants from Columbia University. I want to repeat here that I’m not going to adjudicate particulars; I just want to convey a sense of scale. A lot of folks wanted to know why Columbia couldn’t just make good those funds by using its endowment. It’s a fair question!

Again, just to give you a scale of how much would have to be replaced, $400 million is the amount of annual revenue generated by a $10 billion endowment. Columbia’s total endowment—last year for which stats were available—was $14.8 billion. In other words, making the university whole on this amount would require something like two-thirds the annual endowment income of the university.

This is not a stealth defense of Columbia. I’m simply laying out the napkin math about why—even if the board of trustees had wanted to fight—it would be impractical to try to replace federal funds one-for-one. After all, much of the endowment is locked up by donor intent and deeds of gift. Those can probably be broken in extremis, but I would observe that, first, courts exist to protect property; second, that violating donor intent now means it’s harder to secure donor gifts later; and third, that violating donor deeds of gift might open the door for future federal investigations, including by the IRS.

Again, though, this is not a stealth defense of universities. In some cases, this move might be worth it. Endowments exist to further universities’ mission, and in the case of existential threats to universities’ missions, there are reasons to make dramatic decisions. I don’t think that means that universities should liquidate endowments to save every job—again, we are talking about genuinely massive sums of money here; if any organization lost $4000 million in revenue, they’d be looking at temporary layoffs—but keeping together core personnel and funding legal defenses might be worth it. No less an establishment figure than former Harvard president Larry Summers suggested recently that these might be times that call for precisely those measures.

But here’s the really bad thing.

Back in February, say, assume your university had an endowment worth $10 billion. Now, at that scale, you’re not investing your endowments in mutual funds—the by-now-tired joke that Harvard is a hedge fund with an educational sideline has a glimmer of truth in it, to the extent that managing any fortune of that size means that you have to seek returns from unusual and exotic investments. (As a note: my impression is that the Harvard management team is actually bad at this relative to Princeton, for instance.) So that means top universities have wealth management funds that, yes, hedge against risk and seek high-reward, medium-risk investments through vehicles that look very different from nonprofits with endowments in the $50 million to $100 million zone. However. They are still exposed to, say, market valuations.

So. How’s the market doing?

Ah.

Selling assets in a bear market is a double whammy. Remember the scenario here: the university is, essentially, seeing grants it has already won be held hostage. So these are not smart decisions to spend money—these are driven by imminent necessity. That’s always a bad way to make financial decisions. But in this case, you can also guess with some certainty that assets are worth less right now than they will be later—maybe five years later, maybe 25 years later, whatever: universities have a functionally infinite lifespan. Selling assets in a bear market now means making universities permanently poorer. And if your impulse is to celebrate that—well, I don’t know how to reach you, but I also know that you’re the sort of person who will never be asked to be a custodian for institutions that, ultimately, need to be cared for.

Again, and I want to be plain, I am not saying that we shouldn’t ask more of university leadership and boards of trustees. I am, instead, saying that we need to understand what the ask really is and why even a university administration with the best will in the world might hesitate to make such decisions.

And now for the rug pull.

Ultimately, however, the real value of universities does not lie in their endowments. That’s just treasure. The real value of universities lies in their social role as places to seek and maybe even find truth—it lies, in other words, in their mission. Guarding the money at the expense of the mission is foolish:

Who steals my purse steals trash; 'tis something, nothing;
'Twas mine, 'tis his, and has been slave to thousands;
But he that filches from me my good name
Robs me of that which not enriches him,
And makes me poor indeed.

In ordinary circumstances, playing by the ordinary playbook is good enough. In extraordinary times, extraordinary measures may be required. It’s incumbent on boards and senior leadership to ensure their actions are aligned with the real interests of their institutions.

As all three legs of the stool supporting universities come under attack, thinking about how best to protect the mission and realize the long-term value of the enterprise is going to require more than just the typical reflexive plays of a stable world.

Academia, as a profession, was supposed to have some rewards: autonomy, respect, stability. Each of those stools has been serially kicked away, with the Trump administration’s anti-university efforts pointing toward an even greater acceleration of decay, compounding state-level trends that had already eroded the system—tenure, for example, is now somewhat of an honorary status in many public university systems.

There remain a few perquisites of the job, at least for full-time faculty: a private office, for instance, and (paradoxically equally prized) liberal work-from-home policies. Another is travel. True, conference travel budgets have been steadily eroded; when I joined my previous employer, in 2015, my annual travel and research budget was $1,500, and when I left, after years of stiff inflation, my annual travel and research budget was $1,500. Nevertheless, traveling to a conference was a good excuse to go to new—and old—haunts, and to have at least some part of it paid for as a work benefit. (Which isn’t even all that much of a stretch—not only do most professions have some equivalent of the conference-in-a-fun-place trip, a great deal of networking and work does get done at conferences.)

A good chunk of this is about to be gutted. Specifically, the conference circuit as I have come to know it, centered on North America and with the expectation that scholars from dozens of countries will attend, is being ripped up by Trump administration political risk. The consequences of this will include the closure or financial endangerment of many scholarly associations, a demise in the importance of the United States to intellectual fields, and an overall reduction in human progress.

Let’s start with the basics. Right now, a lot of folks are wary about traveling to the United States. For instance, the German foreign ministry is warning its nationals about their risk of being detained should they even slightly overstay their visa after three Germans (including one green-card holder) have been detained and at least one “violently interrogated” and two others were put in chains. Other recent high-profile cases have put people, especially those with Middle Eastern backgrounds or views on Israel that diverge from the administration’s, on edge about traveling. Some of this might be recency bias—certainly there were abuses and failings of the “border security” system under Biden, for instance—but it’s also hard to avoid the conclusion that an administration whose messaging is relentlessly opposed to international ties and in favor of both mass and artisanal deportations might perhaps be more hostile to visitors than other administrations have been.

The people who are wary of traveling to the United States include many scholars who would otherwise be attending U.S.-based conferences. As the world’s (current) leading center of knowledge production, the United States hosts a disproportionate share of leading and even workaday conferences. Some of these can attract tens of thousands of participants; others in the thousands or low hundreds. Together, they constitute a physical manifestation of research networks, enabling the flow of knowledge and of creativity that powers scientific advances. That so many of these are on home turf for Americans has been a real advantage for U.S. academia, making it possible for many U.S.-based scholars to travel to them at relatively lower costs than their internatinoal colleagues and even to attend more of them.

Like any network-based system, the appeal of conferences is based in part on their network effects: even though smaller workshops can supercharge a research agenda, one of the reasons everyone goes to larger conferences is because they offer chances to make connections with a larger group of researchers—you’re not interested only in meeting people on your panel but larger groups of people in the coffee shop or the bar, for instance. Having bigger and more varied agglomerations is therefore incredibly valuable.

It should be clear by now why having researchers from other countries refuse to attend—or even be barred from attending—conferences in the United States would start to weaken U.S. academia. By reducing the reach and quality of conferences, the value of those conferences becomes much less. Even if selective conferences (those that refuse half or more of their applicants) can still fill their chairs after boycotts or refusals, those chairs will be filled by people who are, bluntly, less desirable than the researchers who would have taken them, reducing the overall appeal of the conferences. The annual meeting of the American Political Science Association (APSA), for instance, which isn’t known as a particularly international conference, still attracts more than a quarter of its attendees from outside the United States. Moreover, if international attendees can’t or don’t attend, they’ll stop joining U.S.-based scholarly associations. For APSA, this could be a big deal: about a fifth to a quarter of APSA’s members are international scholars, and attending the conference is a big reason to be a member.

Long before now, many have started to ask why U.S. academics can’t just replicate their big conferences outside of the United States. There’s a few reasons:

• Moving conferences outside of the United States will, almost by definition, exclude somewhere between many and most U.S. academics. My travel budget at a good-not-top research university could cover one trip domestically, but international travel? Not really. (You’d be surprised at how many people arrive at big U.S. conferences by car or train, which are not approved methods for getting to Europe.)

• The risk would, perversely, be greatest for non-U.S. students and researchers resident in the United States, who would now face visa problems potentially entering another country and returning to the United States. Some universities are already advising international students and staff to reconsider traveling outside of the United States to avoid any visa complications.

• Moving big conferences right now is just not feasible. Major conferences—the pinnacle ones—are planned five to seven years in advance. The International Studies Association conference, for instance, is already booked through 2030, and I can all but guarantee that discussions about where 2031 will be held are well advanced.

That last point is particularly important because already, and predictably, people on social media are beginning to ask why major conferences aren’t just relocated as a result of the Trump administration’s policies. Big conferences are big deals. You can’t just go on Expedia and reserve a block of two thousand hotel rooms a week in advance. Booking a major conference is a complicated process that involves asking cities to decide whether they’d like to host your conference—and academic conferences, which are pretty low-spending and often bizarrely high-maintenance, aren’t high on anyone’s priority lists to host. There may literally not be suitable venues to which major conferences can be relocated. Moreover, once a contract is signed, breaking it is extremely expensive: hotels and conference centers are businesses and they write contracts that all but ensure they will be paid by the association that is organizing the meeting. Penalties can run into the millions of dollars depending on the details. And the election of an orange president does not constitute force majeure.

(Why not move the conferences online? Well, because then the association is paying for both the original physical conference—which it has already booked—and now for a hybrid conference, which is far more expensive. More to the point, even though a lot of people are loathe to admit it, the online conference experience sucked. I did two of them and I feel like I paid a thousand dollars to have a glitchy Zoom meeting that took me to the scenic venue of exactly the same place where I work all the time. It’s better than nothing, but it’s not better than the alternatives, to be honest.)

It gets worse.

Directly and indirectly, the vast reduction in federal funding for academic research is going to make it harder for huge categories of U.S.-based academics to travel anywhere anyway, domestically or foreign. Along the way, this will, again, gut scholarly associations, which are more or less infrastructural and which will be incredibly difficult to put back together again once they are lost. It does not take much to tip many of these institutions into deficits, and cuts to them will get to the muscle and bone really quickly. Perhaps some of the associated changes will be good, but it’s hard to see, on net, how reducing the resources available for coordination and collaboration will help the professions that invented these organizations in the first place.

It gets still worse.

Every conference panel is now a potential target for political attack, which can manifest in many ways—selective IRS audits, congressional attention, (further) cuts to relevant NSF funding agencies, stopping researchers at the border to prevent them from sharing their work, etc. In other words, there’s apt to be a chilling effect, even if it just means that everyone titles their work something really dull to avoid being the one who catches Sauron’s eye.

On a gross material basis, you can see how the supports for the conference circuit are pretty rickety—start draining the bathtub, and pretty soon there’s no more circulation. And you can see why Europeans, Asians, and others would just choose to avoid the whole thing—why schlep to Seattle when you can just stay closer to home? You can see why on both a profound and a selfish level I’m sorry to see this whole edifice broken apart

Four years of this will be grim. More than that will be devastating. Intellectual leadership has been surrendered before—think of Germany before the 1930s—and we may be witnessing it being given up again. .

Perhaps most serious is the threat that nothing will replace the vibrant internationalism of these conferences. There’s no reason that we have to have giant, globe-spanning conferences. It could well be that the intellectual world fragments into a few regional circuits, with the United States existing as just one of those regional fragments—an intellectual horizon bounded by a border wall. Breaking apart the systems of truly international science accessible not only to each country’s elite but to something like a global intellectual middle class—what a waste.

Sometimes, you can read about the Dark Ages. Sometimes, you get to live through them.

The next several years look to be the grimmest period in the history of U.S. higher education, and their effects could well alter the course of U.S. academia permanently.

Federal and state governments have sometimes been hostile to higher education, but the degree of antipathy the Trump administration displays, even celebrates, outstrips anything universities have confronted before. The administration and many of its supporters seem sincere in describing universities as their enemy, and they have done little to signal that they will treat them as anything other than adversarial. More directly, the assault on the entirety of the research enterprise, from the scientific freedom of federal institutes to the details of research universities’ grant reimbursements, will reduce both the autonomy and the resources of scope for independent research, in which universities play a central role. Simultaneously, and perhaps inadvertently, the transformation of U.S. foreign policy from hegemon to mafia don will crater the attraction of U.S. educational institutions to international audiences. Indeed, one hears Canadian and European institutions are—perhaps optimistically—preparing for an influx of American students escaping the strictures of Trump’s America.

Taken on its own, any single policy change may find some supporters. Not a few researchers have rolled their eyes at the notion of 50 or 60 percent overhead fees on grants, for instance. Yet as with many Trump-era changes, the urge to find something laudatory in a policy that is overwhelmingly, unremittingly hostile is badly misleading. Academics, trained to interrogate any proposition, prove relentlessly bad at taking their own side in an argument. Even if one thinks that 50 percent overhead is too much, the answer is not, cannot be, a blanket 15 percent rule. When dealing with someone who treats you as an enemy, it is not wisdom but folly to try to see the best in their actions.

The immensity of the blow that the sector has taken, in any event, far outstrips the details of any policy in isolation. Consider my own field, political science. Students have a hazy idea that what we do pertains to politics and governance, and many of those who came to study with my departments in the United States intended careers in public service (and many more in the law, but often to go into public service). After the savagery of the assault on the civil service, why should any U.S. undergrad assume that there will be a civil service they can aspire to join? Why would any graduate student—whose tuitions play a disproportionate role in supporting many elite institutions—presume that a master’s degree in, say, international development will be worthwhile? There is essentially no international development sector in the United States at the moment.

The knock-on effects are many. The Presidential Management Fellowship—long a symbol of the federal government’s interest in recruiting talent for the non-defense sector—is dead. Why should talented young people conclude the government has any interest in their services? And, having concluded that, why would political science—and its kindred fields—continue to attract students?

Gut a discipline here, cut demand for a field of study there, and pretty soon you’re talking about entire units of universities. And if the non-economics social sciences were to follow the humanities into oblivion (sociology and anthropology, rather woker than political science, will be targeted more directly), then we are left with STEM and business. And if STEM is being reduced as well—tuition dollars do not pay for good STEM programs; that is what the grants are for—then you can see how, through subtraction, we could see an even larger share of higher education converted into little more than vocational training.

Nor will any of this be easy to come back from. The enrollment cliff is real, even if concentrated in some regions. That means, however, that there’s no nice demographic wave that the sector can ride back to normal after riding out four very tough years. Any recovery will be only partial. Tenure lines will vanish and with them the connections, capital, and energy that sustain disciplines.

A particularly annoying tic of social media is the commenter convinced that all of this is some kind of plan—that elites plan on cutting everyone else off from higher education so that only their children can study non-vocational topics. There’s little evidence to support that proposition. Even at elite schools, the shift from humanities toward vocational degrees is evidence, even if the elevated names of those majors—or concentrations, near Boston—reflect the elevated nature of elite vocations. “Economics” in many schools is just a business degree with less accountancy. STEM is for pre-meds.

The Trump administration looks set to accelerate those trends. Its assault on higher education isn't about preserving it for the few—it's about fundamentally transforming its nature for everyone. They will do this because, again, they do not see higher education as a resource to be hoarded for the rich (even though they flaunt their own Ivy League degrees; thanks, Yale, for the bang-up job you’ve done with elite reproduction—really great work, just top-notch, TYVM). Rather, they see it as the bulwark or even the nerve center of an enemy. Whatever instruction needs to be done will be done through institutions that will look very difficult than what we have come to know.

That explains what would otherwise be paradoxical: that the richest schools may be hit particularly hard, as they also receive the highest research funding. Anyone in a “soft money” position—that is, grant-funded—will be in jeopardy. Rich schools have many of those, and their endowments could not compensate for the loss of research funds in toto even if they were legally capable of being used for them. The stakes are enormous. Once labs shutter, and the postdocs scatter and the grad student lines are yanked, it will be hard to restart them. And once the research funds dry up, maintaining library quality will be hard. (That sounds trivial to anyone outside the field, but I assure you that this does actually matter, and if you think that sci-hub will save us you may not know that it and sites like it can be shut down. Maybe we’ll all be better off when long notes on ArXiV are all that matter, but I’m not sure that essentially moving research entirely to a social-network model will be, ah, positive.)

A vicious cycle will set in. If there’s no raises or jobs to be had by publishing or getting grants, why bother investing in research? If the returns to research careers dry up in the United States, top talent will leave for elsewhere—a reverse of how the dissolution of the Soviet Union sent top Soviet talent into the West. Without jobs, grants, or other motivations, honest academic work will be unrewarded. Social scientists and humanities, less grant-dependent, may assume they will be unspared, but funds to cover those deficits will have to come from somewhere, and that in turn means that universities’ subsidies for things like service to the profession will be cut. (And, yes, a course release from a dean to edit a journal is a cost.) The bleeding of funds will crimp academic associations from two directions: conferences and publication fees. (Conferences will suffer a double whammy as international participation dies.) Everything the dark touches will wither.

At the same time, a larger percentage of money that will be available will be tainted in one way or another. (Imagine the RFK Jr. Fellowships in Vaccine Research or the Nabisco Chair of Anti-Ozempic Studies.) Nor should we assume that independent foundations will remain independent. The IRS can audit those foundations, and perhaps the agency will determine that some lines of foundation sponsored research are partisan or against national security. At the final remove, the IRS and other arms of the government can turn on universities, using financial levers to directly influence curricular and content decisions. A law to tax endowments more will weaken the autonomy of the sector even more.

The downstream effects will be tremendous. Tuition-dependent schools will be hit hard as well, and the degradation of the research enterprise will—over time, and inevitably—weaken the quality of curriculum in those universities. (One of my true annoyances is when people question what research has to do with teaching. Well, in the long run, what are you going to teach?)

Bring all of this together, and you have a recipe for a doom loop.

The punchline, of course, is not that academia will be hurt. The real impact will be on society itself. I am critical of many bits of my industry but the fundamental thesis of universities—that education produces social benefits through training, the conservation of the valuable, and the production of new knowledge—is not something I doubt. (Indeed, it’s because I believe in that proposition that I’m critical of much of the cruft in the sector.) Take away top international students and you make every classroom worse off—nothing helps like learning from peers, and so you want your peers to be as strong as possible. Remove resources and you worsen quality—particularly given the hacksaw nature of the policy changes thus far.

For generations, the U.S. academy has been a pillar of soft power and economic dynamism. That is being thrown away now—not on the scale of years but of months, even weeks. The damage being done is not yet irreparable but that point is rapidly (rapidly) approaching. If any of it matters—and take your pick: independent social research, cancer drug development, understanding the universe, March Madness—then the time to act is now.

Yes, the future’s unknowable and maybe some deus ex machina will save us all. Intellectual honesty compels me to admit that I see slim chances that U.S. colleges will avoid any but the very worst fates—even red-state lawmakers may eventually protest at university breast-cancer institutes being closed. But the past four weeks—we are just one month into this administration!—have seen a federal government agency illegally vaporized, a strike team responsible to a literal oligarch access our most closely guarded federal systems, and NATO rent limb from limb. What exactly could you think is the ground for optimism? Even if a Democratic administration wins in whatever year the next free presidential election will be, the interruption to academic lifecycles and careers bids fair to be long-lasting. At this point, taking a pessimistic view of probable outcomes as the baseline isn’t doomerism—it’s an informed updating of priors. Nor am I comforted that all might work out well in the long run. My career doesn’t have that long left to run, in the broad scope of history.

Academia will persevere. There are other governments, other societies, other values. What else have we learned from history other than that no patron remains generous forever?

Below is my revise-and-extend version of remarks for my undergrad stats course.

There’s a lot of mystique around data.

If you want to sound serious and knowledgeable, one way to stake your claim is to assert that you’re data-driven. The implication is that your competitors, of course, are driven by mere vibes. And the further implication is that once you have data, the data will speak for themselves—clearly, plainly, loudly, and in a way that brooks no argument. To be data-driven is to be in direct contact with truth.

That’s an appealing vision. It’s wrong—indeed, likely impossible—as a way of understanding human society. Still, it’s appealing, at least to many people, because we all want to act on the best possible evidence, and statistical data—which appears to us in nice, neat charts, and which is produced by stats wizards who know mysterious terms like “standard errors” and “logistic regression”—seems to offer, by definition, the best possible evidence.

There’s a contrary portrayal of this vision. In this alternative view, data is the enemy of judgment. It’s practically a trope at this point to have a television show or movie in which the main character will turn to a computer or nerd and say something like “You might have data, but you forgot the most important point: the human heart.” Or something like that.

The less hackneyed version of this is to point out all the limitations of data and measurement. Data, after all, don’t just emerge from the vasty deep—they have to be produced, recorded, and designed to answer particular questions. It’s rare that we have exactly the data we would want to have to answer the questions we want to pose about the world, especially about the human world.

I sometimes joke that, as a political scientist, it’s my job to use an accelerator to smash politicians into each other to see if we can discover new, fundamental forms of politics—but the joke is that it’s very hard to use the methods of physics in politics. Humans, it turns out, are tricky creatures to understand. They’re deceitful, both to observers and to themselves; they can become aware that they are being monitored and change their behavior; and they have a range of options through which to express their preferences. All of that means that we can’t just choose one outcome to monitor and then assume that we’ve gotten it right. An atom won’t change what it’s doing because we’re watching, but a politician might. Worse, humans are individuals, and that means that we can’t treat them as interchangeable, the way a physicist might treat two atoms from the same element. As a result, it can be hard to say with certainty what humans are doing and why they are doing it.

Unless we’re very lucky or very careful, we are usually in the situation of using second-, third-, or fourth-best data to understand what people are doing and why. And the worse our data become relative to our ideal situation—which we can rarely if ever attain—the more cautious we have to be in assuming that our data are illuminating anything near to the truth.

Sophisticated critics of quantitative methods, then, will often assert that naive appeals to data are, well, naive—that it’s a bit of faith that will lead the credulous astray.

Well, if that’s the case, then why should you take stats as a required course?

There are, I think, three big reasons.

The first is that statistical reasoning is one of the greatest intellectual accomplishments of human reasoning. Statistics (if we include its long, open border with probability theory) offers a way to understand patterns within the randomness of observation. Statistical thinking can be applied to a variety of problems at a range of scales, from the behavior of the fundamental particles of the universe to the evolution of the grandest structures in the universe—suns, galaxies, the distribution of matter and heat itself. And we can even apply it to hard problems, like knowing whether adding a subway line will meaningfully reduce use of freeways (one of the actual impetuses behind the development of more sophisticated tools for analyzing choice, which won their developer an economics Nobel Memorial Prize). By bounding uncertainty, statistics offers us a way to make randomness intelligible and tractable—an enormously valuable contribution in its own regard.

Statistics, done well, is a disciplined approach to understanding the world that enables us to systematically analyze observations in ways that we could not do without the assistance of the intellectual discipline and computational techniques it enables. That is to say: it is all well and good to record one’s own impressionistic observations, but it is also very good practice to work to sum across many observations and to seek patterns beyond those which can be identified by a Mark I human brain.

The second is that, as a practical matter, statistical reasoning is increasingly central to the work of both academic social scientists and the much wider array of social engineers (who do not go by that name, but who should) in the real world. If you’re asked to do data-analytic work as a professional, then you will very quickly move beyond simple tools like bar charts and line graphs and start harnessing the power of tools that draw on statistical reasoning. That applies as much to the Facebook social graph analyst as it does to the graduate student seeking to understand the results of her experiment.

There’s something funny in the way that we present political science and other social sciences to undergraduate students. A lot of material in a typical undergraduate social science major, especially in the early years, looks like it could be assigned in a history, philosophy, or even literature course. There’s an endless amount of words and verbal theorizing, and much of the evidence that is presented is done via cases and, sometimes, wordy summaries of quantitative findings.

This approach really does a disservice to students. Most working social scientists (especially if we recall that economists constitute the bulk of that category!) spend most of their time wrestling with statistical data—measuring, analyzing, and disputing it. Not all—there are many valuable advances in the realm of pure theory and purely qualitative (that is to say, non-quantified) evidence. But look at the publications in journals like the American Journal of Political Science or International Organization, and you will quickly see that being conversant with many, if not most or even “nearly all”, leading empirical debates requires a firm grounding in statistics. (More recently, the most salacious academic scandals have involved careful investigations into what seem to be cases of fraud involving the manipulation of data—you might not think that knowing the details of how Excel records data is exciting stuff, but it can be as telling as a smoking gun in a murder investigation.)

As a consequence, if you want to actually understand what social scientists know and think they know about the world, you genuinely must be able to at least interpret a regression table—without that knowledge, you will be utterly reliant on what is claimed by authors with no way to verify their assertions or you will be thrust into dependence on the tribes of social scientists who do not use the tools that have become mainstream. (And even many of them are deeply enmeshed in debates with their quantitative rivals and allies.)

That data analysis has particular relevance for employment is also important.

The third major point is that even if you are a skeptic of quantification, there is no way you can make a credible case without understanding quantification. If your response to the millions of person-hours that have gone into the development of stats and empirical applications of stats is simply to dismiss it with a wave of a hand—a gesture that might combine sincere disapprobation with some anxieties about the hand-waver’s own competence—then only those predisposed to sympathy with you will view your dismissal as credible. To dispute something requires understanding it; the Jesuits have long believed that, and I take that view.

So even if you do not intend to become a “quant”, if you want to engage with them you need to be able to actually understand what they are talking about. For my own part, I view this kind of informed skepticism as incredibly important precisely because skeptics inside and outside the tent are the ones who drive the responsible use of statistics—just because someone can perform an analysis does not mean it tells us anything, and often those who know and understand a case well from other perspectives can check if the emperor has put on clothes or just mimicked the actions.

So what is quantification?

Quantification is the art of turning observations into structured, numerical data. The most basic form of this is counting—how often does something appear? The more dimensions we track, the more questions we can ask:

• How often does something appear over time?

• How often does something appear relative to some other observation?

• How often does something appear before or after some other observation?

• Where does something appear (that is, how often does it appear over some measure of space)?

• How often does something appear next to or far away from some other observation?

• Do we observe some pattern in the data with the same distributions as would be determined by chance alone, or does it appear to have a more deterministic relationship but with some error function?

There is nothing magical about this. There is nothing threatening about this. And there is nothing “non-human” or “anti-humanistic” about this. It is all, ultimately, about counting.

And it’s not just about counting things. Some of the most exciting developments that have achieved mainstream status in the past ten or fifteen years (in my scholarly neck of the woods) have involved counting connections—who talks to whom? who went to school with whom? Count those up, throw ‘em in a matrix, and you’ve got the data for a social network analysis. And you can even count words: how often is this word used compared to those words? What happens when we code words as being happy or sad, angry or confused? Well, we can start to assess the emotional state of a text—and from there we can begin to understand what people are feeling, or at least what feelings they’re expressing, when they talk about particular issues. We can even teach computers to see and to recognize what they’re seeing—to do the counting for us at a scale we couldn’t do even with megagirls, or gigaboys, of effort.

From the counting we move on to questions like:

• What is the best way to theorize what the relationships would be in the data if one explanation or another is true?

• What is the best probability distribution we should use to understand the baseline, “random” distribution—to understand what chance would be?

• What is the best way to present data?

• Hey: what should we count?

These are core questions—about theory development, hypothesis testing, and communication—that blend technique (the maths and coding) with judgment (reasoning and aesthetics). The data do not speak for themselves, but they do whisper—and sometimes, after much effort, we can hear them. And if we’re smart and informed and careful, we can

None of this, by the way, requires computers. But we use computers because computers make all of this much, much easier.

(There’s a fun fact—okay, well, it’s horrifying. The term “computer” used to refer to the people who performed “computations”, not the machines we use today. Because this was relatively low-status and rote work, it eventually became the province of women working under the direction of men who monopolized the high-status thought work for themselves. By the end of the 19^th century, “computers” worked in organizations like Harvard and NASA, the U.S. federal government, and so on, working on difficult, distributed math problems. By the Second World War, there was even the unit of “kilogirl”—a thousand hours of computing labor done by women.)

Computers now of course require specialized training. They allow undergraduates to undertake what use, not too long ago, were Nobel Prize-winning techniques – quite literally. These are in many cases the same techniques that underlie just-behind-cutting-edge mechanisms in machine learning and social network analysis (since the cutting edge now belongs to transformers and LLMs). The differences in many cases has more to do with the scale and the quality of data available rather than the principle of analysis.

Think about it. The great minds of social science—Marx, Du Bois, Durkheim, Weber—all of them were limited by calculations and data gathering available to them. None of them could do what you can do with training, discipline, and a laptop. No pressure!

I’ll be very bold. There’s a lot of folks who think that there’s a hard line between “qualitative” and “quantitative” work. Often that just means that the correct techniques in quant work or conceptualization haven’t yet been designed or haven’t yet been employed. You can use computers to count, to see, to track, to measure, to find nodes, to read – you can do a lot with quant work if you put your mind to it. For that matter, your ability to theorize will often be enhanced if you learn how to observe and draw on observations that aren’t statistical—you will probably know a lot about how to theorize viral spread, for instance, if you look hard at how people interact to learn what to count.

And this is largely where contemporary social science work unfolds. One way or another, the ways in which quantification is changing what social science does is much the same as anywhere else in human endeavors…it’s everywhere, it’s growing, and it requires training.

All of this, however, requires some hard work – although a lot less hard than 20 years ago – to learn. I think the payoff is worth it—even if we shouldn’t be too naive about how being data-driven will save us. I hope you will come to believe this, too, but even if you don’t, I hope you will at least learn how to disagree with me at a much higher level than before you took this course.

My dominant image of my undergraduate professors’ offices is paper. Paper in the form of books, paper in the form of journals, paper in the form of … papers, the kind undergraduates like me submitted. Younger generations may not quite understand that we have already arrived at something closer to the long-sought goal of the paperless office—even if in practice it’s more like a paper-less office. Offices of all kinds, academic or not, used to be devoted to the management of dead trees: the filing, marking, retrieval, and disposition of records on paper.

Professors’ offices, in my experience, still lean toward the paperfull versions, but few approach the sprawling off-white yellowing clumps I recall so distinctly. Even the most super of our superannuated colleagues have largely ceased acquiring paper journals—sometimes not by choice, as many journals are now digital-only. The custom of reading paper journals has long since become a symbol not of keeping current but of falling behind, since paper editions lag months or years behind the online-first publication of articles. You can observe this process almost like a geographical stratification: here we have the era of print that slowly and then suddenly stops accreting as a result of environmental changes.

It is possible—and I have seen it with my own eyes—for professors to have offices that are as devoid as print as any banker or car dealer. For the rest of us, the relics of the print age are still much with us. Literally. I mean that if you came of age before (or during) the transition from print to digital, you acquired habits that reflected how information used to be hard to find again. There were no “databases”, just indices and guides that themselves resided on print—and provided, inter alia, employment for entire lost trades of indexers and the like. (I keenly recall finding sources for my senior thesis using the Index of Current Periodicals from the 1920s and 1930s, with its small type and idiosyncratic abbreviations— “R. of Rs.” for Review of Reviews, for instance.) In that era, if you wanted to refer to a text again, you needed the text—and there was an entire lost toolbox of preprints, offprints, and reprints of journal articles, sliced neatly from the journals. I’ve even heard—but am not old enough to know firsthand—that conferences would have paper rooms where you could buy a copy (a literal, physical copy) of conference papers, which would provide an index of popularity as some stacks dwindled and others remained pristine.

Devotees of paper rightly note that digital compendia are more fragile in the long term—linkrot, copyright issues, and all the rest. Yet they too easily gloss over the actual costs of holding on to texts that might be referred to once or twice in a career—and often never again after their first use. A Kindle is the same size no matter how many books it carries, but boxes of books tend to proliferate into backbreaking sizes. Practically everyone I know in the biz has a story about movers’ resigned shoulder slumps when they see the book boxes. (Big men come to me, with tears in their eyes—big, strong men—and they say, please, no more books.)

Share

Why hold on to books too long? The justification and the reasons differ. The justifications are almost all predigital: I want to be able to find this again, it’s out of print, whatever. The reasons are that books are totemic: these are the symbols of my craft. Yet the texts became symbols because they were not only distinctive but useful—and they were distinctive because they were useful: a professor in 1980 lived surrounded by papers because they didn’t have Dropbox, and so text accumulation was rational (as well as being a display of productivity and erudition). These days, well—ctrl+f makes up for a lot.

And so this weekend I purged many of my office shelves. Good-bye to general reading about current events in China; a book about China’s environmental policy from 2012 has almost no relevance to me now, or for that matter to China (except for antiquarians). Good-bye to books with good covers and good titles and good summaries that turned out to be inconsequentially narrow or impossible to follow—you had great publicists and lousy authors. Good-bye to methods books from before 2015: if you’re not written for contemporary R, then you’re not useful to an applied researcher. Good-bye to several first-generation Oxford handbooks: you were a painful financial sacrifice, even at conference discounts, for a graduate student, but 15 years old is death for reviews of the literature. Good-bye, good-bye, good-bye. Good-bye to a stage setting conveying what I hoped to perform.

Hello to room for growth.