Last fall, the chancellors of Vanderbilt and Washington University asked ten scholars to look at the state of scholarly standards in the humanities and the social sciences. Ashley Rubin, a sociologist at the University of Hawai’i, then sketched her own field. She spent six months on it and turned in a hundred and twenty-three pages.
At the end of her executive summary, before any of the findings, Rubin stops to say something odd for an audit though. The report evaluates sociology “according to criteria and assumptions about scholarship that are themselves contested within the discipline.” The commission had assumed, she explains, that careful systematic inquiry can produce reasonably trustworthy knowledge, and that scholarship should be judged on what it contributes to knowledge rather than on its social or political consequences.
Then the sentence: “Undoubtedly, scholars operating with different assumptions will come to different conclusions about the field.” Her word is “assumptions,” and the ones she has just finished listing are epistemic.
Interesting though, let’s imagine an accountant filing that. “Before we begin, please note that arithmetic is disputed by some members of this firm, and colleagues who reject arithmetic will assess these books differently.”
She isn’t being cute. She’s being accurate, and the accuracy is really the whole problem! She had a real assignment. She did the work. And she couldn’t state her findings without first conceding that the standard she used is one a portion of her discipline rejects.
Not fails to meet. “Rejects.”
That’s not a field that flunked an audit per se. That’s a field where a serious person can’t assume the audit means anything to the people being audited, and has to say so in writing.
On August 10 the same argument broke out in public, at a lower altitude and a bit higher volume, and it’s really the stimulus for putting together this post, because I found it interesting and a kicking-of-the-rusty-gears-before-classes-start.
Alice Sullivan, a quantitative sociologist at UCL, wrote that sociology “is a fractured discipline which contains communities with very different attitudes to evidence.” She was writing in the wake of the Jason Arday affair at Cambridge, where a prominent professor resigned on August 5 amid plagiarism allegations, which he denies while admitting to some mistakes, with a separate inquiry into his credentials still open. Her worry wasn’t really about him. It was that the whole discipline would eat the reputational bill for a hiring decision made in one corner of it.
Daniel Blight, who describes himself as a fellow sociologist, pushed back politely. He was surprised, he said, to see “anti-rigour” appended to a list of theoretical traditions as though it were a component of them. Don’t we need a more balanced view? There is rigor and a lack of rigor across strands in sociology.
Sullivan’s reply is the sentence this whole thing turns on. “A lack of rigour is possible in any strand of sociology. But some strands oppose rigour on principle. I believe we need to start being less polite about this.”
Both of them said something true. Only one of them was talking about the thing in dispute. Blight answered a question about design with an answer about distribution, and those aren’t the same question, and the difference between them is where everything interesting about these epistemological challenges lives.
Every field produces bad work. Sloppy papers, overstated findings, results that evaporate when somebody reruns them. I’ve written about the Science replication project and other numbers and they aren’t flattering to anybody, including my own discipline. But that kind of failure has a known fix, and the fix is machinery: data editors who check your code or measurements or identification strategies before publication, replication requirements, mandatory sharing of materials. Fields that installed the machinery got better. Fields that didn’t, didn’t. It’s an incentives story rather than a virtue story, which I find encouraging, because incentives are things you can change.
Now consider a claim whose defense is that your standard for evaluating it is merely “your standpoint.” The ultimate in untestable subjectivity, right?
No amount of machinery, or even epistemic care, touches that. You can’t assign a data editor to it. You can’t replicate it. You can’t even say what being wrong would look like, which means you can’t say what being right would look like either. Every tool in the kit assumes the thing being denied: a check outside the person making the claim, and the possibility that it comes back negative.
There’s no fixing that with better process. There’s nothing to fix.
So there are two different problems wearing the same word. One is a field that fails at rigor. The other is a commitment that rejects the criteria by which rigor could be assessed at all. The first is a quality-control problem. The second is closer to a design specification, and immunity from correction isn’t a side effect of it.
Immunity from criticism is the feature, and so criticizing it is not allowed on any grounds, moral, epistemic, or otherwise.
The usual move here is to reach for postmodernism and start swinging, which is why nobody who actually studied this stuff takes the criticism seriously.
As I’ve said before, the postmodernists have earned their due, so let’s give it to them. Their core insights are correct, even if you don’t want them to be.
Knowledge is contextual, and what counts as a question worth asking varies by time and place and who’s asking. Knowledge is constructed, assembled by people out of concepts those people made. And there’s no ahistorical, privileged vantage point, no view from nowhere, no observer standing outside the system taking neutral notes.
Contextualism, constructivism, aperspectivism. All three are present. Any essay that denies them deserves the laughter it gets, and gets it.
The error isn’t in the insight. The error is in what people did with it.
From knowledge is contextual, constructed, and perspectival, it simply does not follow that therefore nobody can check anybody, or discount their findings, or question their perspective. That’s a leap, and everything indefensible in the last forty years is still standing on it.
Alan Sokal and Jean Bricmont gave the distinction its name, and Helen Pluckrose has done a lot to draw it out. Specific skepticism, they call it, as against radical skepticism. Both start in exactly the same place: we can never be fully certain we’ve reached objective truth. Specific skepticism takes that seriously and builds methods to minimize being wrong, which is what falsification and replication and adversarial review are for. Radical skepticism takes the same premise and concludes, as Pluckrose puts it, that trying to ascertain the truth is futile and naïve, so we should sit in the uncertainty instead.
So, same premise. Opposite architecture. One of them produces a field that can be corrected and one of them produces a field that can’t, and the difference has nothing to do with how sophisticated the practitioners are.
So how did the second kind get a seat at the table?
Partly by invitation, and the invitation was generous, and I’d have signed it.
Ever hear the phrase “there are different ways of knowing”?
Epistemic pluralism is a real intellectual good. Fields innovate because nobody polices method from the top, because the ethnographer and the econometrician and the archivist all get to work, because the person with the weird approach is occasionally right, and the discipline that stamped them out lost something it can’t get back. I believe this. Most working scholars believe this, I think, though I question it some days. It’s one of the few things left that most of the professoriate agrees on.
Look hard at that phrase, though. It’s doing an enormous amount of work, and it does the work by meaning two entirely different things.
On the first reading, it’s plainly true and I’d defend it. An ethnographer, an econometrician, and an archivist know things in different ways, and each of those ways reaches something the other two can’t. Different instruments, different questions, different kinds of evidence. That reading is the reason the seat exists in the first place.
On the second reading, it says something else. Not different routes to the same check, but different and equally valid accounts of what counts as true, with no vantage point from which anyone could adjudicate between them. Not different tools. Different truths.
The first is a claim about methods. The second declines the “deal.”
And you cannot tell, from the phrase itself, which one you’re being asked to nod along to. That isn’t sloppiness. It’s the structure the philosopher Nicholas Shackel named the motte and bailey: hold the modest, defensible version whenever anyone challenges you, and operate on the expansive one the rest of the time.
The decent person in the room hears the first version and nods. Of course he does. The first version is correct.
But methodological and epistemic pluralism was always a deal, (wasn’t it?) and deals have terms.
I made a similar argument about tolerance and the word “liberal” a few weeks back, and the structure is identical. Tolerance was never an unconditional posture of accommodation. Locke wrote exclusions right into the founding text of tolerance, even. Popper’s famous footnote holds suppression in reserve for movements that have abandoned argument for fists. The tradition built a compact with terms and enforcement provisions, and the useful idiot in that story is the one who forgot the terms were ever there and called the forgetting a virtue.
Epistemic pluralism has terms too, and they’re not all that demanding even. You don’t have to use one method. You don’t have to be a positivist. The entry requirement is that your work can be wrong, and that somebody other than you gets to say so.
Nothing about that favors quantitative work, which is the objection I’d expect from people who suspect this is a mostly-numbers guy telling ethnographers their job isn’t real. Eboo Patel, who trained as a sociologist at Oxford, put the qualitative version better than I could this week. Qualitative sociologists, he points out, are curious about the same things everyone else is curious about, and they work with the same instruments everyone else has, namely watching people and talking to them. So the field’s claim to academic standing rests on one thing: “The only way that the field of qualitative sociology can justify academic standing is if our conclusions are more accurate than the conclusions of others. If we aren’t more accurate than others, we’re no better than the guy on the next bar stool spouting off opinions.”
No code to share. No dataset to post. No identification strategy to defend. Just a claim about accuracy, and a willingness to be proven wrong.
Patel found out. He went to Oxford in 1998 planning to show that white supremacy shaped the lives of middle-class South Asian Ismaili Muslims in London, spent twice a week for six months with his subjects, and kept hearing about religion instead. When he told his advisor he must be asking the wrong questions, the advisor suggested the other possibility: that his theory was wrong, and that he wanted racism to matter in their lives more than it did. He followed the data instead of the thesis, and it turned into his career.
That’s what the terms look like when somebody honors them. It cost him the argument he came in with, and the world got a little closer to actually understanding itself.
Call it pluralism’s deal if you like, but it’s older than that and it was never really ours to renegotiate. It’s science’s deal. Submit your claim to a check that could come back against you, and in exchange you get to call the result knowledge. Everything after that is administrative detail: peer review, replication, peer monitoring, the reason a credential means anything at all. Mill made the same bargain about speech in On Liberty, almost word for word. Kant put it in the preface to the first Critique: ours is the age of criticism, to which everything must submit, and reason grants respect only to what has withstood free and public examination. Same trade every time.
Earlier in my career, I assumed everyone at the table had accepted the deal, and that the problem was people cheating on it. The real problem is that a good number of them never signed on to that at all, and they said so, out loud, in printed rejections and catalogues of their epistemic commitments, decades ago.
But not all of them, and the distinction matters. Sandra Harding argued for a stronger objectivity, not the abandonment of it. Donna Haraway spent half of “Situated Knowledges” attacking relativism as “a way of being nowhere while claiming to be everywhere equally.” Those are people renegotiating terms, not walking out.
But a real strand made a different move. Lyotard’s whole argument was that the scientific metanarrative is one (privileged) narrative among others, and that where the criteria differ there’s no judging between them. The framework itself became the object of critique, an instrument of power, objectivity a claim staked by whoever already holds the position to stake it. That isn’t a hidden agenda. It’s a published one, and it has a bibliography.
It’s also sitting in the interview work Rubin collects. Sociologists telling researchers, on the record, that objectivity is either not possible or not desirable. Not that it’s difficult. Not that we keep falling short of it.
Not desirable.
So somewhere along the way, we stopped asking, in the voice of the most decent person in the room. “Who am I to say.” “Let a thousand flowers bloom.” “There’s rigor and a lack of rigor across all strands, isn’t a balanced view better?”
We heard the first version of the phrase and signed for the second one.
They told us which one they really meant. We set a place at the table anyway.
Rubin raises a much darker possibility, and she raises it as a question rather than a finding. Writing on her own Substack about what her data couldn’t settle, she wonders whether sociology is running a Keynesian beauty contest: people basing their behavior on what they believe other people are thinking. Is the majority, she asks, “just adjusting to fit an ideal very few people prefer”?
If she’s right, almost nobody in the room had to be persuaded of anything. They just had to spot the fashionable trend and grab on with both hands, even if it broke their inner compasses. They only had to believe that everyone else had already done it, which is a considerably worse problem than a discipline full of true believers, because a room full of people privately holding a standard they’ve stopped saying out loud has no way to discover that it agrees with itself.
Either way we, and what “we” “know,” end up in the same place, because the one commitment seated on pluralist grounds is the one commitment we can’t subsequently evaluate. Judging it requires the standard it denies. We handed out the seat and left the criteria at the door, and now when somebody like Sullivan says the obvious thing out loud, the room finds her impolite.
Shit’s broken if that’s the case. In my darker moments, I describe it as “epistemically fucked.”
The villain here isn’t the theorist with their own epistemic commitments that they’ve told us are their epistemic commitments. It’s the generous, conflict-averse colleague who thought refusing to discriminate between approaches was the “liberal” thing to do, and who has never once counted the cost of a vote he cast to end a meeting early and be left alone.
I’ve been that colleague. Most of us have, at one point or another. At least lunch was tasty.
Back to what Rubin actually found, because the findings are more careful than the fight around them.
Plenty of sociologists still do rigorous work and produce high-quality research, she says, and she says it repeatedly. What’s shifted is field-level norms. Writing later on her Substack about the commission’s findings across all six fields it examined, she put it this way: “There is clearly good research produced across all the fields examined. There’s also a clear commitment to politics and an explicit rejection of certain scholarly standards many folks take as bedrock.” Both can be true, she says, which is exactly why this is so hard to referee.
Note the word explicit. This isn’t inferred from bad papers. The interview and survey work she collects turns up contestation over whether objectivity is possible or even desirable, over whether scholars should separate advocacy from research, over whether certain questions may be asked at all. In one study of department chairs, a handful described their departments as rejecting positivism outright, though most described the opposite. People said these things on the record, to researchers. Rubin flags that the surveys are dated and wants them redone, which is the kind of caveat that tells you how she works.
Her second finding she calls monocropping. Inequality, power, and identity crowd the field’s attention while other questions go unasked: biology and evolution, religion beyond a narrow focus on Christian nationalism, socialization, cultural explanations for disparities, class as an alternative to racism, whole theoretical frameworks written off as conservative. That inventory is the empirical face of everything I’ve been describing. Immunity doesn’t announce itself in a bad paper. It shows up as questions nobody starts, because they aren’t allowed to be asked or have fallen out of fashion.
Then she does something that cuts against her own findings, and against me.
She observes that the strength of the evidence runs inversely to the importance of the metric. The most visible indicators of politicization, conference themes and presidential addresses and association statements, show the strongest evidence and tell you the least about what research actually gets produced. The substantive indicators, the quality of specific published work, hiring and tenure deliberations, are the hardest to measure and mostly unmeasured. And because the visible stuff is so visible, she warns, it can “convey a stronger impression of the field as overly politicized, hiding some of its important heterogeneity.”
I did that. In a piece a couple of weeks ago I leaned on an association election and a conference theme, which are exactly the cheap, loud, legible indicators Rubin is cautioning about. They’re real, and they’re weaker evidence than they feel like. I’d write that section differently now, methinks.
And here’s her hardest line, the one that constrains what anybody gets to claim from this report, including me. The evidence for politicized norms, she writes, is stronger than the evidence for their downstream effects on the rigor of the research produced, “which (to my knowledge) has not been systematically analyzed.”
So I’m not going to tell you the work is demonstrably worse. Nobody can tell you that yet. But look at what the narrower claim actually says, because I think it’s the more damning one: the commitment is avowed, it’s measurable in norms and in the questions nobody asks, and it removes the mechanism by which anyone could find out whether the output got worse. The unfalsifiability is the finding.
Everything I’ve described is a tool, and tools don’t stay with their inventors. Pluckrose again, and she puts it better than I can: the postmodern left didn’t create the post-truth right. What it did was weaken the defenses against it. Once “your standard is just your standpoint” counts as a legitimate move, the move belongs to everybody, and only the content changes.
Here’s the cheapest version, and it’s everywhere this summer. Carolyn Rouse, president of the American Anthropological Association, was asked by the Chronicle about the Vanderbilt report. Her answer opened with three words: they used AI. To be fair to her, she got to substance further down, and some of it lands. But she reached for that first, and reaching for it says something: a claim about how the document was made, ahead of anything in it. That’s a new kind of ad hom, even though there ain’t much hom in it.
I’ve been on the receiving end of the identical move from the other direction, which is how I know how good it feels. A writer ran one of my essays through an AI detector, got a high number, and posted it as a verdict on the argument. Not a word about the argument. Provenance in, content out.
All I do about that anymore is laugh. These are my ideas. Does AI help me execute them better? Yep, and I’ll say so in the disclosure. It’s production value, baby.
Alexander Kustov has written about this, and his conclusion is not the comfortable one. He used to argue that the detectors were bad and would stay bad, and he changed his mind a little: Pangram has third-party evidence behind it and, he writes, “claims a very low false-positive rate.” It seems to work ok, but it doesn’t distinguish between the different usages of the tool. That’s exactly why he thinks the ethics got harder, because “a working detector makes the ethics more urgent because it tempts people to treat provenance as a verdict.”
The line that matters for another point that I’ve been arguing: “A detector estimates the probability that a text is AI, never the probability that it is bad.”
Never the probability that it is bad. Kustov’s worry is that “we will police the em dashes while ignoring the hallucinated data underneath them,” which he calls a very academic way to lose the plot. His standard is the one I’d sign: if you put your name on the work, you own it. The facts, the claims, the errors, the taste, the structure, the judgment. Use the detector where provenance is actually part of the bargain, an exam with a no-AI rule or a prize for human craft. Everywhere else, treat the score as a prompt for judgment and never a substitute for it.
And he runs the test on himself in the same piece, which is more than most manage. He calls himself “that infamous AI professor who proudly writes with the help of AI,” then says the quiet part: an essay arguing that AI detection can become status policing is conveniently a defense of his own practice, so readers should ask whether he’s drawing the line in a way that flatters him.
He also names something: the accusation is a status move wearing a technical costume. The same instinct that once policed grammar and accent now polices em dashes and smooth transitions, and it “gives a scientific-looking license to dismiss work without reading it carefully.” It lands hardest on people who don’t write English well and can now reach an English-language audience. Too polished looks fake, too rough looks unserious, and the gatekeeper wins either way.
It’s a move that is cheap, it’s satisfying, and it works, because engaging a claim is expensive and this isn’t.
The structural version is worse than a dodge, and I’ve watched it run from every direction. Somebody sets out a demarcation criterion I’d happily sign. Real inquiry tests itself against reality; pretend-inquiry curates its evidence toward a conclusion it had already moralized. Good. Then the second half of the argument arrives, and the same person is attributing motives to whole classes of people, with nothing on offer that could ever come back negative. The criterion never once gets pointed at the hand holding it.
The irony here is that better work has become less trustworthy, while errant work is becoming more so. That is a screwed up set of incentives.
None of this should be a scoreboard. I’m not counting fouls on each side and calling it even. What these share is a structure: hold a standard, apply it outward with real force, never run it on yourself.
You can learn a lot from that, and it’s really the only test I trust anymore. Not whether somebody invokes rigor. Whether their sharpening ever points home.
Which puts the commission itself on the hook, eh?
The Vanderbilt–WashU report has taken a beating since June, and the criticisms sort into two piles that look similar but aren’t.
The first pile: the authors never consulted the scholars they cite. Who are they to judge. The report’s criteria are themselves contested. The whole thing is politicized research masquerading as an audit. Almost none of these is a claim that the report got a fact wrong. They’re claims about standing, process, and motive, and no possible version of the report satisfies them. A commission that had interviewed every cited scholar and named every research assistant would face the identical objections, which is how you know the objections aren’t really about the evidence.
The borderline case is the anthropologist Fernando Villanea, who says the report took his words out of context and flattened an argument about who anthropology should serve. That’s a real complaint, and it’s also the complaint every critic in history has made about every critique of them. I don’t know how to settle it from outside, and neither does anyone else, which is sort of the point.
The second pile has one thing in it that matters. Khiara Bridges, a law professor at Berkeley, pointed out that the report quotes her as endorsing a radical perspectivalism about truth when she was describing postmodernism in a primer, not endorsing it. She’s right, that’s a real error, and a report about scholarly standards can’t afford it. Paul Boghossian answered her the same day and said he was eager to correct any misrepresentation.
Then go look at the report. The sentence now reads that the view is central to many critical race theorists, “though it is not necessarily endorsed by her,” with a footnote explaining that it replaces the original wording at her request.
That’s the whole argument in one exchange. A checkable objection got a same-day reply, a change to the text, and a footnote documenting what changed. The unfalsifiable ones produced heat and nothing else.
Which is not a wash between two sides. It’s a test, and it came out the way you’d expect.
Which brings us back to where we started. Ten scholars wrote internal reports; the commission published its conclusions while most of that evidence stayed in a drawer, and that’s a fair hit, one even sympathetic readers have made. Two of the ten have since published in full. Only one of the two opens with a conflict-of-interest statement disclosing that she was paid, that her compensation wasn’t contingent on what she found, and that she came in with prior views. Only one keeps every caveat intact. Only one writes down, in her own executive summary, the sentence that gives her critics their best line of attack.
Rubin showed her work, including the parts that weaken her case, and she should be praised for that. It has nothing to do with anyone’s politics, and it’s the only difference that has ever mattered: she put it out there and made it falsifiable.
The point I’ve been trying to drive home is that when the people inside an institution won’t enforce its terms, enforcement doesn’t just go away. It moves. And where it lands is with people who have no stake in whether the institution survives, which is why what arrives isn’t correction at all. Nobody’s wage-gap paper improves because a legislature cut its funding.
Florida’s Board of Governors deleted introductory sociology from general education at all twelve public universities in March, on a surprise motion that wasn’t even on the meeting’s agenda.
And in June, the philosopher Matt Lutz reviewed the Vanderbilt report sympathetically and asked whether it would end up providing cover for administrators to dismantle humanities departments. Almost certainly, he answered himself. By late July he was writing that he’d been right about that. On August 3 the Secretary of Education sent a letter to university presidents and governing boards across the country, and cited the report by name: some academic disciplines and departments, it says, “have subordinated the scholarly pursuit to political and social goals, as documented in the recent Report on the State of Scholarship in the Humanities and the Humanistic Social Sciences.”
He predicted “administrators.” He got the federal government in eight weeks.
Sullivan’s fear was that the whole of sociology would eat the bill, and she’s right, and the people who’ll eat the most of it are the ones who were doing checkable work the entire time.
A permissive room and an epistemically fractured discipline are the same story told at two speeds. The first one produces the second, and the second one summons somebody with a bulldozer.
I don’t think you can put any of this toothpaste back in the tube. The postmodern turn was not a mistake to be reversed; the insights are here and they aren’t going anywhere, and anyone selling you a return to 1955 epistemology is selling you a fantasy that would be unrealizable even if it were available. It’s not a genie you get to put back in the bottle. How many metaphors can I torture in one paragraph, you ask? :)
But we can improve it from here, and improvement doesn’t require winning an argument about the nature of truth. Nobody is going to win that argument. It requires remembering the terms.
Those terms ought to be that you get a seat at the table if your work can be wrong and somebody other than you gets to say so. That’s the whole compact, and it’s science’s before it’s anybody’s. Simple enough that you’d think nobody would need it spelled out, and it’s the one thing pluralism can’t afford to be pluralist about, because it’s the condition that makes pluralism worth having in the first place.
Which does mean saying the uncomfortable part. If somebody has publicly declined the deal, they haven’t been excluded from anything. They declined. What they’re owed after that is a hearing, an argument, and the ordinary respect you extend a colleague. What they aren’t owed is the presumption that comes with having signed.
Which raises the fair question. Enforced by whom?
Not by argument. You can’t argue somebody out of a position that was built to survive argument, and it’s motte and bailey and bailey and bailey all the way down. Every philosopher who has tried has produced a beautiful paper and changed nobody’s mind. Not by the state either. Legislatures and education secretaries don’t enforce scholarly standards, they enforce whatever the last election said, which is the bulldozer, and I’ve spent enough words on why that’s worse.
The terms get enforced the way they always did, in the meetings. Somebody decides which article gets published, which candidate gets the line, which grant gets funded, which claim gets the deference that comes with expertise. Those decisions are made constantly, by people who have them to make, and every one of them is an act of discrimination in the old sense of the word. The question was never whether scholars would discriminate between kinds of work. The question is whether they’d admit that’s what they’re doing, and hold themselves to a criterion they’d say out loud.
Elinor Ostrom won a Nobel for working out how this goes when it goes well. Her subject was commons: fisheries, irrigation systems, grazing land, the shared resources that everyone draws on and nobody owns, which economists had long assumed would get wrecked unless a state or a landlord stepped in. Ostrom went and looked at communities that had managed theirs for centuries, and found a handful of arrangements that kept turning up. The people bound by the rules had a hand in making them. Monitoring happened as a byproduct of the work itself, since you notice things while you’re out there waiting your turn at the water. And violations drew what she called graduated sanctions, assessed “depending on the seriousness and context of the offense” by the other users or by officials answerable to them.
A discipline’s standards are a commons in exactly that sense. Nobody owns them. Everyone draws on them. Their value depends entirely on collective upkeep, and defecting is rational for any individual right up until the whole thing collapses. Peer review is Ostrom’s monitoring almost exactly: nobody hires an inspector, you just notice things while reading work you were going to read anyway.
So the question has a known answer. Not a legislature, and not a philosopher with a very good paper. The members, on each other, in proportion.
Graduated is the operative word, and it’s what separates this from the thing people picture when they hear it. Ostrom found the first sanction in durable commons is often almost nothing, and her explanation is disarmingly human: a small penalty may be enough to remind somebody that compliance matters, and besides, the monitor “might be in a similar situation in the future and would want some understanding at that time.”
Call that softness if you like, but the communities she studied kept their fines low on purpose. In the Spanish huertas the penalties were set so the guards “did not deeply antagonize the farmers” they had to keep working alongside. And she’s blunt about what happens when you skip the graduated part: “A large monetary fine imposed on a person facing an unusual problem may produce resentment and unwillingness to conform to the rules in the future.”
So nobody needs excommunicating. The response to a weak paper is a rejection letter with a reason in it. The response to a candidate who can’t say what would falsify their claim is a committee that declines to advance them and can explain why in a sentence.
Still uncomfortable, and I’m not going to sand it down for you. What I’m describing means sitting across a table and telling a colleague their work doesn’t qualify. Refusing to make that call buys you no neutrality at all. It just hands the call to whoever is loudest in the room.
Rubin’s report makes no recommendations at all, and says so twice: prescriptions were outside her charge, and in her view they shouldn’t rest on a single study anyway. But on her own Substack in June, months before her own report went public, she allowed herself exactly one: “we have to uphold (encourage, incentivize, teach) scholarly standards.”
Modest to the point of banality, I think. She guessed that “roughly half or 60% (at least)” of her field would reject it instantly, and she flagged her own numbers as outdated in the same breath, which is very much her way in reading her stuff. Sixty percent is her floor, not her ceiling.
Her care deserves the same in return. Rejecting a slogan is not the same as refusing to be corrected, and some of the people who’d wave that sentence away would do it because they think it’s empty, or coded, or a stalking horse for a report they distrust. Fair enough.
But hold it next to the paragraph she wrote at the top of her report, the one conceding that her criteria are contested. Put them together and you get a working scholar who can’t assume her colleagues would accept the standard, can’t assume they’d sign the mildest possible sentence about upholding it, and is careful enough to write both of those things down where anyone can check them.
She did the audit anyway, and that deserves more credit than it’ll get.
If I’ve gotten something wrong in here, the comments are open.
If you want these to keep coming, you can subscribe, and if you’d rather put a price on it you can actually see, you can always buy me a coffee at the coffee button!
All your support is appreciated.
Earlier in the stack:
An Election Can Defund a Field Like Sociology. It Can’t Replicate One., on what happens after a discipline stops correcting itself, and who shows up when it does.
Are “Liberals” Useful Idiots for the World to Come?, the parent argument: tolerance was a compact with terms, and forgetting the terms isn’t a virtue.
Half the Building Is Still Standing, the replication numbers, including the unflattering ones about my own field.
Reality Bats Last and The Fifty-Year War, where all of this comes from.
It seems that given the strong conflict between science and postmodernism that perhaps sociology needs to be split into two separate disciplines. In reality, this may already be happening since my check of the local Sociology Department website there was not a single appearance of the word "science". Assuming they are of the postmodernist persuasion.
Compelling read, thank you! Maybe I'm oversimplifying to the point of uselessness, but doesn't this just come down to paradigms (Kuhn and others)? Post-positivism vs. others, etc.?