Daniel sent us this one — he studied law back in his undergraduate days, and he remembers being struck by the sheer physical size of academic journals in print, these enormous tomes that would show up month after month. He gets why something like Dr. Schneiderman's work on the anatomy of the rhesus monkey jaw exists — an authoritative reference that becomes the ground truth for a discipline. But what he always found harder to grasp is how these journals fill up hundreds of pages every month, who's actually reading them, and whether the really obscure niche journals in the long tail have seen a contraction over the years. There's a lot to unpack here.
There really is. And before we dive in, quick note — today's episode is powered by DeepSeek V four Pro. So if anything sounds unusually coherent, that's why.
I was going to say, Herman's tangents are suspiciously well-organized today.
I resent that. But also, yes, possibly accurate. So let's start with the most basic question Daniel asked — who's reading these things? And the answer is genuinely surprising if you haven't looked at the data. A study from about fifteen years ago found that half of all academic papers are read by nobody except the author, the peer reviewers, and the journal editor.
That's a staggering number.
It's worse when you dig in. Research published in Scientometrics found that roughly twelve percent of medicine articles and about eighty-two percent of articles in the humanities go completely uncited. And citation isn't even readership — it's just somebody acknowledging the paper exists. Actual readership is almost certainly lower.
You've got these six hundred page journals arriving every month, and the median article might have an audience you can count on one hand.
Yet the system keeps expanding. Ulrich's Periodicals Directory — the closest thing we have to a comprehensive database of serial publications — currently tracks somewhere north of three hundred thousand active academic journals globally. That number has been growing at about five to six percent annually for the past two decades. There are roughly three million scholarly articles published every year.
Three million articles annually. Even if you wanted to keep up with a narrow subfield, you couldn't.
This gets at something Daniel noticed intuitively — the disconnect between the physical heft of these journals and their actual reach. In the print era, a law journal might arrive at a library, sit on a shelf, and maybe three people would touch it in a given year. The subscription model meant libraries were paying thousands of dollars for journals nobody was opening.
Which raises the question — if nobody's reading them, why do they exist?
This is where you have to understand that academic journals serve a function that has almost nothing to do with readership in the conventional sense. The primary purpose of a journal article in the modern university system is not to communicate findings to an audience. It's to establish priority, to secure tenure, to satisfy grant requirements, and to build a CV.
The audience is not the point. The publication is the point.
The act of publishing is the product. The readership is almost incidental. This is why you get these extraordinarily niche journals. There's the Journal of the Philosophy of Sport. The International Journal of Salt Lake Research. There used to be a journal called Bean Bag, a newsletter for scientists studying legume taxonomy. It ran for decades. These exist because for the people in those fields, having a dedicated venue to publish in is professionally essential. Whether anyone reads the articles is secondary. The journal exists as a credentialing mechanism.
The economics of this are completely inverted from normal publishing. In normal publishing, you write something, people buy it, you make money. In academic publishing, you write something, you often pay to publish it, the journal sells it back to university libraries at enormous markups, and the actual readership is negligible.
The people doing the peer review — the quality control — are unpaid volunteers. The authors are unpaid. The editors are often unpaid or receive a small stipend. The publishers — Elsevier, Springer Nature, Taylor and Francis, Wiley — they're the ones capturing the value. Elsevier's profit margins have historically run around thirty-five to forty percent. That's higher than Apple, higher than Google.
For a product where the content is created and vetted for free.
It's one of the best business models ever devised. And they sustain it through what's called the "big deal" — bundled subscription packages sold to university libraries. A library doesn't get to pick and choose individual journals. They buy a package of hundreds or thousands of titles, many of which nobody on campus will ever read, because the alternative is paying even more for the handful of journals their faculty actually need.
This is why the print volumes Daniel remembers were so enormous. The marginal cost of adding pages is low when you're already charging tens of thousands of dollars per subscription. If an editor has content, they'll run it.
The page budgets in some of these journals are remarkable. The Journal of Biological Chemistry, in its print heyday, could run to over a thousand pages per issue. Physical Chemistry Chemical Physics regularly hits six or seven hundred pages per issue. The physical objects were almost unmanageable.
I remember — and I say this as someone who has spent entirely too much time in medical libraries — you'd see these bound volumes and they were basically furniture. You could use them as step stools.
Yet there's a real question about whether the digital transition has changed any of this. Because Daniel asked specifically about the long tail and whether there's been a contraction. And the answer is complicated.
Walk me through it.
On one level, no — there hasn't been a contraction in the number of journals. The shift to digital publishing dramatically lowered the barrier to entry for starting a journal. You no longer need a printing press, a distribution network, physical storage. You need a website, some editorial board members willing to lend their names, and a peer review workflow. The result has been an explosion.
This is where the predatory journal phenomenon comes in.
There's this entire category of journals that exist primarily to collect article processing charges from authors desperate to publish. These are the emails academics get constantly — "Dear Dr. Poppleberry, we were impressed by your recent work and invite you to submit to the International Journal of Advanced Advanced Studies." They'll accept almost anything. There was a famous sting operation about a decade ago where a researcher submitted a paper consisting entirely of the phrase "get me off your fucking mailing list" repeated hundreds of times, and it was accepted by multiple journals.
I remember that.
The number of predatory journals has been estimated at somewhere between eight thousand and fifteen thousand globally. But the lines are blurry. There's a spectrum from outright fraud to legitimate but low-quality venues. And the growth has been fueled by pressure on researchers in countries like China, India, and Nigeria, where publication metrics are tied directly to career advancement and even cash bonuses.
The long tail isn't contracting. It's getting longer and weirder.
— and this is the important but — what has contracted is the print presence. Almost nothing in the long tail exists in physical form anymore. The journals Daniel remembers seeing on shelves, those enormous bound volumes, that's increasingly a legacy format maintained by a few hundred major titles. The vast majority of the three hundred thousand active journals are digital-only.
The physical object is disappearing, but the journal as a publishing entity keeps proliferating.
There's another layer of contraction happening that's less visible. It's in the consolidation of publishers. Over the past thirty years, the academic publishing industry has undergone massive mergers. A handful of companies now control something like fifty percent of all journal output. Elsevier alone publishes over two thousand seven hundred journals. Springer Nature has over three thousand. When these companies acquire smaller publishers, the titles persist because each one is a revenue stream. But the editorial independence shrinks.
The journal titles survive, but they're absorbed into these giant portfolios.
The pricing power concentrates. This is why the open access movement has gained so much momentum. The idea is that research funded by taxpayers should be freely available to the public, not locked behind paywalls where a single article download costs thirty-five dollars.
Which seems obviously correct as a principle, but the implementation has been messy.
It's been deeply messy. The dominant open access model has become "author pays" — the article processing charge, or APC. Researchers pay a fee, typically between two thousand and eleven thousand dollars, and the article is free for anyone to read. On the surface, this solves the access problem. In practice, it creates new ones.
Because now the journal's incentive is to accept more papers, not fewer.
If your revenue comes from authors paying to publish, you want to publish as many articles as possible. Quality control becomes a cost center rather than a value proposition. We've seen exactly this dynamic play out. MDPI, one of the largest open access publishers, has faced sustained criticism for rapid peer review timelines and high acceptance rates. They published over two hundred forty thousand articles in a single recent year.
Two hundred forty thousand articles from one publisher. How do you even find reviewers for that many papers?
You don't, or you do it badly. The peer review system was already strained before the open access explosion. The number of papers requiring review has grown far faster than the number of qualified reviewers willing to do unpaid labor. There's a phenomenon called "reviewer fatigue" — senior researchers get dozens of review requests per month and start either declining or doing cursory reads.
The whole system is under pressure from multiple directions. Too many journals, too many papers, not enough readers, not enough reviewers, and a business model that incentivizes volume over quality.
Yet, I want to push back on the cynical take a little bit. Because it's easy to look at all this and conclude that academic publishing is a giant waste of time and resources. But that's not quite right either.
Okay, make the case for the defense.
The case for the defense is that journals, even obscure ones, serve a genuine epistemic function. They create a permanent, citable record of research findings. That record might be consulted rarely, but when it's needed, it's there. The fact that only three people read a particular paper on the mating habits of a specific Central American frog doesn't mean that paper has no value. It means its value is latent — it exists for the moment when someone needs to know exactly that thing.
The long tail argument applied to knowledge.
In a world with only a few broad journals, that frog paper never gets published. The knowledge exists only in the researcher's field notes, and it's lost. The proliferation of niche journals means that highly specific knowledge gets preserved in the permanent record. Most of it will never matter. But some of it will.
The challenge is that we don't know in advance which obscure paper will turn out to be crucial.
This isn't hypothetical. The discovery of the CRISPR gene editing system — now one of the most important technologies in biology — began with a paper published in 1987 in the Journal of Bacteriology. It was a weird observation about repeating DNA sequences in bacteria. Nobody read it and thought "this will change the world." It was just another paper in a specialist journal, doing what specialist journals do — recording findings for the record.
That's a good example. And it highlights something Daniel was getting at — the tension between the apparent absurdity of these enormous journals that nobody reads and the genuine value of having a system that preserves niche knowledge.
The system is deeply inefficient, but the inefficiency might be partly unavoidable. If you tried to make academic publishing "efficient" by only publishing papers that lots of people want to read, you'd lose the long tail of specialized knowledge that occasionally produces breakthroughs.
Though I wonder how much of that long tail actually needs six hundred pages per month to accomplish its mission.
That's the right question. And the answer is almost certainly that it doesn't. The page budgets Daniel was struck by are largely a relic of the print subscription model. In the digital era, there's no reason a journal needs to be a certain length. You publish what's ready, when it's ready. The physical heft was always partly about justifying the subscription price.
The enormous print volumes were, in a sense, a marketing artifact. "Look at all this content you're getting for your money.
Libraries were complicit in this because their own prestige was tied to the size of their collections. A university library that could point to holding ten thousand journal titles was more impressive than one holding two thousand, even if most of those ten thousand were never opened. The metric was shelf feet, not usage.
That's a perfect encapsulation of the absurdity.
The digital transition has changed this metric, but it hasn't eliminated the underlying dynamic. Now it's about the number of titles in your database, the number of articles accessible through your institutional login. The collection size still matters as a prestige marker, but at least it doesn't take up physical space.
Let me ask you something about the readership question from a different angle. Daniel asked who's actually reading these things. We've established that for many articles, the answer is almost nobody. But for the articles that do get read — who are those readers?
The readership is almost entirely other academics. This distinguishes scholarly publishing from every other form of publishing. There's effectively no general audience for journal articles. Even articles in fields that seem broadly relevant — medicine, psychology, economics — are written in a technical register that makes them impenetrable to non-specialists.
That's by design, to some extent.
It's by design and it's by training. PhD programs teach you to write for your peers, not for the public. The incentives reinforce this — tenure committees don't care if you wrote a popular science book. They care if you published in top-tier journals in your field. So the audience is narrow by construction.
Which means the entire enterprise is a kind of internal conversation among specialists, funded largely by taxpayers and tuition payments, with the outputs locked behind paywalls that the same taxpayers can't access.
When you put it that way, it sounds almost deliberately perverse.
Is there any data on what the actual readership numbers look like for a typical journal?
There have been some attempts to measure this, and the numbers are sobering. A study of usage logs from a major publisher found that the median article was downloaded fewer than ten times in its first year. For journals in the humanities, the median was in the single digits. Even in biomedicine, the most heavily read area, only about twenty percent of articles got more than fifty downloads in their first year.
And these articles represent months or years of work, often funded by grants in the hundreds of thousands of dollars.
Those download numbers probably overstate actual readership. A download means someone clicked the PDF link. It doesn't mean they read the paper. It doesn't mean they got past the abstract. Some unknown fraction of those downloads are just people collecting PDFs they intend to read someday and never do.
The academic equivalent of a to-be-read pile that haunts you.
Every researcher has a folder on their computer with hundreds of PDFs they've downloaded and never opened. I'm guilty of this myself from my medical journal days.
We've established that the system is enormous, mostly unread, economically distorted, and yet not entirely without value. What about Daniel's specific question about contraction in the long tail? You said the number of journals keeps growing, but has anything actually disappeared?
Journals do die, but it's relatively rare and usually quiet. A journal might cease publication because its host institution cuts funding, or because the editor retires and nobody steps up, or because it gets absorbed into a larger title. But outright cancellations are uncommon because the marginal cost of keeping a digital journal alive is so low. Once you've built the website and established the ISSN, the ongoing costs are minimal.
Even journals with essentially zero readership persist indefinitely.
There's a phenomenon called "zombie journals" — titles that technically still exist but haven't published anything in years, or that publish so sporadically that they're effectively dormant. Nobody officially declares them dead because there's no mechanism for doing so. They just linger in databases, waiting for submissions that never come.
This is very different from normal publishing, where if nobody buys your magazine, you stop printing it.
Normal publishing has a market signal. Academic publishing, for most of its history, has been insulated from market signals because the people making purchasing decisions — librarians — are not the people consuming the content, and neither group is the one paying the bills in a direct sense. The ultimate funder is the taxpayer or the tuition-paying parent, and they're several steps removed from the transaction.
Has anything disrupted this? The open access movement tried, but you pointed out it created its own problems.
There have been disruptions at the margins. Preprint servers like arXiv and bioRxiv have become the de facto way that researchers in physics, computer science, and increasingly biology share findings. Papers appear on these servers before — or instead of — going through traditional journal peer review. In some fields, the preprint is the version that actually gets read, and the journal publication is just a formality for the CV.
The journal's function as a distribution mechanism is being bypassed.
In fast-moving fields, absolutely. If you're a machine learning researcher, you don't wait eighteen months for a journal to publish your paper. You put it on arXiv today, and by next week, hundreds of people in your field have read it and are building on it. The journal publication, when it eventually happens, is almost an afterthought.
The journal still matters for tenure.
It still matters for tenure. And this is the bind. The prestige economy of academia still runs through journal publication, even when the actual communication of ideas has moved elsewhere. Junior researchers can't afford to ignore journals, even if they know the real conversation is happening on preprint servers and social media.
The system persists because it's embedded in the incentive structure of the entire profession.
Changing that incentive structure is incredibly difficult because the people who control it — senior faculty, university administrators, grant committees — are the people who succeeded under the existing system. They have no reason to devalue the currency that built their own careers.
This is the self-perpetuating nature of prestige hierarchies.
It's exactly that. And it connects to something Daniel might appreciate from his legal background. The common law system works similarly — precedent accumulates, and the authority of past decisions constrains future ones, even when the reasoning behind those decisions has become outdated. Academic publishing has its own kind of precedent, its own accumulated authority that's very hard to dislodge.
That's an interesting parallel. Though in law, there's at least a mechanism for overturning precedent. In academic publishing, what would overturning precedent even look like?
It would look like a major university saying "we will no longer consider journal publications in tenure decisions. We will evaluate the actual quality and impact of research, regardless of where it appears." A few institutions have experimented with this — there was the Declaration on Research Assessment, known as DORA, which came out of a meeting in San Francisco about a decade ago. Thousands of organizations have signed it, pledging to move away from journal-based metrics.
Has it made a difference?
The intent is good, but the implementation is hard. Evaluating research quality directly requires actual reading and judgment. Counting journal publications is much easier. When you're a tenure committee with forty cases to review, the temptation to fall back on "how many papers in top-tier journals" is overwhelming.
The metric that's easiest to measure drives out the metric that actually matters.
Goodhart's law in action. When a measure becomes a target, it ceases to be a good measure. And journal publication counts have become the target for entire national research evaluation systems. China's system, for instance, has historically tied cash bonuses directly to publication in journals with high impact factors. The result has been a flood of papers, some excellent, many not, and a parallel industry of paper mills producing fraudulent manuscripts.
Organizations that sell authorship on fabricated or plagiarized papers. You pay a few thousand dollars, and they'll produce a paper with your name on it, complete with fake data, and shepherd it through submission to a real journal. It's a massive problem, particularly in biomedicine. Hundreds of papers are retracted every year because of paper mill activity, and that's almost certainly the tip of the iceberg.
The pressure to publish has created a black market in publications.
The journals themselves are often the victims here. Peer review wasn't designed to detect sophisticated fraud. If someone fabricates data convincingly, a reviewer reading a manuscript has no way to know. The system assumes good faith, and that assumption is being exploited at scale.
This is a long way from Daniel's memory of impressive bound volumes in the law library.
But I think the through-line is that the physical grandeur of those volumes was always a bit misleading. It suggested a solidity, a permanence, a quality control that the system never fully delivered. The journals looked authoritative because they were big and heavy and printed on nice paper. But the actual epistemic value has always been uneven.
Though I want to be fair — some of those volumes are authoritative. The Schneiderman monkey jaw anatomy book Daniel mentioned exists because someone spent decades building expertise and produced a reference work that advances the field.
And that's why I resist the fully cynical take. The system does produce important work. The problem is that it also produces an enormous amount of low-value material, and the signal-to-noise ratio keeps getting worse.
If you were redesigning the system from scratch, what would you change?
I think I'd start by decoupling the certification function from the dissemination function. Right now, journals do both — they certify quality through peer review, and they disseminate findings through publication. But in a digital world, those functions can be separated. Dissemination is basically free — that's what preprint servers do. Certification is the valuable service — "these findings have been vetted by qualified peers and are likely reliable.
You'd have a system where everything gets posted immediately, and then review happens separately?
Something like that. There are experiments along these lines — overlay journals that curate and review papers that already exist on preprint servers, post-publication peer review platforms like PubPeer where papers are evaluated after they appear. None of them have displaced the traditional model yet, but they're pointing toward something that might eventually work better.
The obstacle again being that the traditional model is deeply embedded in the incentive structure.
Let's not forget the money. The traditional model generates something like twenty-five to thirty billion dollars annually for publishers. The people benefiting from the current arrangement have every reason to resist change.
Daniel mentioned that he studied law as an undergraduate. Law reviews are an interesting case within this ecosystem, because they're student-edited. That's almost unique among academic disciplines.
It's completely unique. In virtually every other field, journals are edited by senior scholars. In law, the flagship journals at American law schools are edited by second and third year students. The Harvard Law Review, the Yale Law Journal — these are run by twenty-five-year-olds.
Which is wild when you think about it. You've got students making editorial decisions about articles written by professors and judges.
The selection process for those editorial positions is brutal. It's based partly on grades and partly on a writing competition that involves doing technical legal citations. The people who make law review are the academic elite of their class. But they're still students. They're not subject matter experts.
Does this create quality issues?
It creates a very different kind of quality issue. Student editors are meticulous about citation format — law review articles have famously exhaustive footnotes that can run for pages. But they may not have the substantive judgment to evaluate the actual legal arguments. There's a long-standing critique that law reviews publish articles that are technically polished but intellectually timid, because student editors are reluctant to challenge senior scholars on substance.
You get articles that are perfectly formatted and thoroughly footnoted, but maybe not as groundbreaking as they could be.
The articles are long. Law review articles routinely run to sixty, eighty, a hundred pages. There's a famous quip that a law review article is something that begins with a hundred pages of footnotes and ends with a tentative conclusion. The genre has its own conventions that don't necessarily serve readers well.
Though Daniel might push back on that characterization.
And to be fair, the law review system has produced important scholarship over the years. Some of the most influential legal ideas of the past century first appeared in student-edited journals. The system works in its own weird way.
We've been talking mostly about the supply side — who publishes, why, how many. Let's circle back to Daniel's core question about readership. Is there any trend toward more people reading academic work, or fewer?
The data suggests that total readership has probably increased in absolute terms because digital access makes it easier to find and download papers. But the number of papers has grown faster than the number of readers. So the readership per paper — the metric that matters for the question Daniel is asking — has almost certainly declined.
Each individual paper is being read by fewer people, even as the total amount of reading goes up.
That's the pattern you'd expect when supply grows faster than demand. And there's no reason to think demand for academic papers among the general public is growing significantly. Most people will never read a journal article in their lives.
Which is fine, in a sense. Not everything needs to be for everyone. But it does raise questions about the public funding model.
If the public is paying for the research but can't access the results, and wouldn't understand the results even if they could access them, and the results are mostly read by a few dozen other specialists — you can see why people get frustrated with the whole enterprise.
Though the alternative — defunding basic research that doesn't have immediate public uptake — would be catastrophic in the long run.
The challenge is designing a system that preserves the genuine value of specialized research while reducing the waste and improving access. Nobody has fully cracked that yet.
Now: Hilbert's daily fun fact.
A group of flamingos is called a flamboyance.
Where does this leave us practically? What should a listener take away from all this?
I think there are a few concrete things. First, if you're outside academia and you need to engage with scholarly literature, don't assume that journal publication equals quality. Look at whether the paper has been cited, whether it's been discussed on post-publication review platforms, whether the authors have a track record of reliable work. The journal name is a weak signal at best.
If you're inside academia?
If you're inside academia, you probably already know all this. But the actionable takeaway might be to support reform efforts where you can. Sign DORA if your institution hasn't. Do peer review when you're asked, and do it carefully. Push back against the use of journal impact factors as proxies for research quality.
The impact factor is particularly pernicious.
It's a terrible metric. The impact factor of a journal is the average number of citations received by articles published in that journal over a two-year window. It tells you nothing about any individual paper. A paper in a high-impact journal might have zero citations, while a paper in a low-impact journal might have hundreds. Using the journal's impact factor to evaluate the paper is statistical nonsense.
Yet it's used constantly in hiring and promotion decisions.
Because it's easy. That's the recurring theme. Metrics that are easy to calculate drive out metrics that are meaningful. Fighting that requires institutional will, and institutional will is scarce.
What about for someone like Daniel, who's just intellectually curious about this world? What's worth paying attention to?
I'd say pay attention to the preprint revolution. The most interesting developments in academic communication are happening outside traditional journals. arXiv, bioRxiv, SSRN for social sciences, Research Square — these platforms are where the real conversation is increasingly happening. If you want to see what cutting-edge research looks like before it gets polished for journal publication, that's where to look.
The traditional journals — are they going to disappear?
Not anytime soon. The prestige is too entrenched, and the money is too good. But their role is shifting from primary distribution to archival certification. They're becoming the permanent record rather than the daily newspaper. The daily conversation has moved elsewhere.
That's probably a healthier division of labor anyway.
The journal system was never designed to be a rapid communication medium. It evolved in an era when papers were physically mailed to subscribers and the review process took months by necessity. The fact that it's persisted so long in the internet era is more a testament to institutional inertia than to its fitness for purpose.
One last thing I want to touch on — Daniel mentioned the sheer size of these journals as physical objects. I think there's something almost poignant about that. These enormous bound volumes represented a kind of optimism about knowledge — the idea that scholarship was accumulating, that each issue added to the permanent edifice of human understanding.
The digital transition has made knowledge feel more ephemeral, even as it's become more accessible. A PDF on a server doesn't have the same weight as a bound volume on a shelf.
Though the bound volume was often a monument to unread pages.
The romance of print obscures the reality that most of those pages were never turned. The digital era has at least made the lack of readership more measurable.
Measurement is the first step toward reform. If you don't know how bad the problem is, you can't fix it.
Though as we've discussed, knowing the problem exists and fixing it are very different things.
One forward-looking thought — I wonder if the next generation of researchers, who've grown up with the internet and don't have the same attachment to traditional journal prestige, will finally tip the balance toward new models. The current system is maintained largely by people who built their careers in it. When those people retire, does the system go with them?
That's the optimistic scenario. The pessimistic one is that the next generation simply replicates the prestige hierarchies in new forms. Instead of publishing in Nature, they'll chase visibility on whatever platform becomes the new Nature. The human desire for status markers is pretty durable.
Something to watch. Thanks to our producer Hilbert Flumingtop for making this episode happen. This has been My Weird Prompts. Find us at myweirdprompts dot com or wherever you get your podcasts. I'm Corn.
I'm Herman Poppleberry. See you next time.
