Emergency Audit: Is Curiosity Trespassing?

A professor at the University of Chicago, Nicholas Polson, produced an extraordinary number of academic working papers with the assistance of artificial intelligence. By August 26, another economist had counted 258 papers in 2026. The output attracted substantial scrutiny, and SSRN subsequently removed 257 papers and suspended the relevant accounts while it assessed unusual submission activity.

Some of those papers may contain useful work. Others may be incoherent. Andrew Gelman examined an ambitious 80-page synthesis of human relationships and found little reason to take it seriously.

Research requires evidence, attribution, competence, and enough scrutiny to distinguish a contribution from an impressive arrangement of sentences.

Then, in his August 22 essay Intellectual Squatting, Makowsky begins speculating about why somebody would publish so much.

Perhaps the author is reserving intellectual territory. Publishing undeveloped ideas today so that somebody doing serious work tomorrow must acknowledge his priority.

Perhaps the goal is glory.

Makowsky is careful enough to acknowledge that some of the work might have real value, and that his account of the author's motives is speculative.

Modal Path Ethics would like to investigate the speculation.

Because the further the argument proceeds, the stranger its object becomes.

Scholarship becomes years of labor required to convince one's peers.

Originality becomes a contest over recognition.

Premature publication becomes a threat to the market for scholarly contributions.

And intellectual curiosity?

Well, apparently somebody will get around to that after the property dispute.

There is a distinction missing here, somewhere.

Fortunately for us, Modal Path Ethics has just spent several months repeatedly embarrassing itself while trying to understand this exact distinction.

Unfortunately for everyone, Modal Path Ethics has a Zenodo account.

And the Canadians recently rejected one of those papers.

This should be fun.

The Years Are Doing Something Suspicious.

Makowsky explains that scholarship is generally more perspiration than inspiration. The initial idea is exciting, but establishing a contribution can require years of work.

Of course it can.

Some problems require decades of investigation. Others require a clever experiment, a short proof, a new measurement, or noticing that two familiar things are connected in a way nobody previously considered.

Often, the difficult work is discovering why the original idea was wrong.

But there is an important transition in the argument.

The labor traditionally required to convince other scholars has begun to function as a defining feature of scholarship itself.

Why?

Years of labor can produce excellent evidence. They can produce expertise, understanding, failed approaches, and valuable intellectual discipline.

They can also produce years of labor.

A discovery does not become less correct because somebody found it too quickly. An argument does not become more coherent because its author suffered for several semesters before submitting it.

And recognition by peers, however important, remains an institutional event.

That tells us something about what a specific community has assessed, accepted, and decided to incorporate.

It does not tell us when intellectual inquiry first became legitimate.

A person has an idea.

That person might develop it for ten years. They might share it with another researcher tomorrow. They might construct a little experiment over the weekend and discover it fails. They might use artificial intelligence to explore the consequences, identify relevant literature, and locate the first obvious contradiction.

They might even publish the question.

None of these paths requires the person to have first earned the right to be curious.

The Property Right That Isn't.

Makowsky imagines a researcher using artificial intelligence to generate papers from undeveloped ideas, then demanding credit when somebody else eventually performs the substantial work.

That behavior would be extremely obnoxious.

It also would not automatically accomplish what that hypothetical researcher wants.

The United States Copyright Office explicitly distinguishes copyright in an author's expression from ownership of the underlying ideas. A sufficiently specific technical disclosure may affect subsequent patent claims under the applicable law. That does not itself give the person who disclosed it an exclusive patent right.

And your date stamp does not force another scholar to provide coauthorship, an eponym, or acknowledgment for a contribution that never materially entered the later work.

Priority, authorship, legal ownership, scientific validity, and citation are different relations.

Somebody can state a hypothesis first and be completely wrong about it.

Somebody can state a hypothesis first and lack the resources to test it.

Somebody else can independently discover the same relationship, establish it rigorously, and deserve substantial credit for doing so.

The first statement may still belong somewhere in the intellectual history.

That history is not a deed to the territory.

The institutional record is especially interesting here.

SSRN's own explanation of preprints explicitly identifies early dissemination, priority registration, and feedback as purposes of the service.

The National Institutes of Health likewise recognizes preprints as a way to accelerate dissemination, establish priority, and obtain feedback while maintaining clear requirements that the work be identified as unreviewed.

These instruments were built to let research enter public inspection before validation and publication have finished.

Now, artificial intelligence has made it dramatically cheaper to produce things resembling academic papers.

That changes the verification burden.

It can create enormous problems for the poor services carrying these papers.

It still does not make early disclosure a newly invented form of intellectual trespassing.

The College Newspapers Have Already Won This Case.

Modal Path Ethics has actually been through another, stupider version of this argument.

On August 25, I published a semi-coherent attack on Philosophy & Public Affairs for its artificial-intelligence writing policy.

My argument was that philosophical research and the professional craft of academic philosophical writing were fully distinguishable activities. Artificial intelligence had made that distinction increasingly difficult to ignore.

The journal could help preserve and credential one while philosophical inquiry continued elsewhere.

This was a reasonable distinction.

I then spent several thousand words treating the journal's decision to defend its particular professional craft as evidence that it had retired from philosophy, mostly for the title joke.

The resulting article was very funny. It also had a jurisdiction problem.

On September 8, I published Modal Path Ethics Apologizes to the College Newspapers.

The original insult had been to demote Philosophy & Public Affairs to a college newspaper. Unfortunately, college newspapers were always good.

Young journalists need somewhere to learn reporting, editing, deadlines, corrections, and all the complicated responsibilities of publishing material other people may rely upon.

They need an institution in which to develop a professional craft.

The college newspaper does not owe the world Reuters.

And a journal does not owe the world the entire field of human philosophy. It may quite legitimately maintain demanding standards for its particular publications, preserve specialist literature, train professional researchers, carry out adversarial review, and provide credentials that have material consequences for people's careers.

That is real work.

My corrective public service announcement was therefore quite simple.

• If your career requires journal publications,
  • please publish in the journals.
• If your work truly benefits from specialist peer review,
  • please use the journals.
• If you are an independent project with access to online publishing infrastructure and no tenure committee,
  • maybe stop running around demanding that every journal restructure itself around your activities.

Those people are actually using their institutions.

I have the internet. I should have noticed that sooner.

The same distinction applies here.

A scholarly venue can quite legitimately reject papers that impose unreasonable verification burdens. It can require adequate methods, credible citations, transparent artificial-intelligence disclosures, and evidence that the submitting author has actually inspected the work.

It can reasonably decide that hundreds of speculative papers are simply too much for its particular service to process responsibly.

There is no general obligation to publish or review every manuscript somebody generates.

But the boundaries of a journal, archive, profession, or credential are still boundaries of those instruments.

The intellectual field remains larger.

And this is where the story becomes considerably more interesting.

I Never Became Qichao Wang.

On September 9, Modal Path Ethics had a problem.

I had been working with ChatGPT on causal abstraction: how to distinguish causal organization carried by a system from computation supplied by the apparatus used to interpret, intervene upon, or explain it.

The research had produced some results. Unfortunately, the central novelty test had failed.

The trained-neural experiment had exposed an apparatus problem, but existing causal-abstraction theory already possessed the relevant resources for identifying it. So the candidate new distinction had reduced to prior work yet again.

I had explicitly decided that I would not keep a research program alive by dressing existing mathematics up in Modal Path Ethics terminology.

If the framework was going to actually contribute something technically distinctive, it needed to be able to change what the investigation could see.

So I went looking for another problem.

This is how I ended up in synchronizing automata.

I was not trying to become an official mathematician. I was specifically looking for a difficult technical object against which to test a structural intuition.

A finite automaton is a machine with a set of states and rules for moving between them. In the problem that interested me, every possible starting state receives the same sequence of commands. Different starting histories can eventually merge.

There was a peculiar question in Marek Szykuła's 2026 survey of open problems, attributed to Michalina Dżyga's earlier work.

Conjecture 8 concerned how long one distinguished state could avoid merging with any other starting state, even while the machine's overall collection of possible states was being compressed.

This is what caught my attention. Modal Path Ethics has spent considerable time arguing that aggregate change can conceal where contraction actually lands.

Here was a machine in which distinctions could disappear globally while one particular distinction remained stubbornly intact.

What made that possible?

Could an account of local contraction reveal something that global measures were missing?

This was the Modal Path Ethical angle I chose to pursue.

Solving an open mathematical conjecture was not an expected outcome here.

The first attempt failed.

We had tried to track the distinguished state through what we called tagged coalescence, developing a contraction ledger from global rank and defect.

This sounded pretty promising.

Then exact enumeration supplied counterexamples.

Machines with the same aggregate rank behavior could have different collision thresholds.

So the proposed invariant was too coarse.

The structure determining the distinguished state's survival was not captured by counting how many states the machine had merged overall.

It depended on which other histories remained available to merge with it, and how those histories could move.

Modal Path Ethics loves relational thinking. So the investigation moved into pair space.

Instead of following individual states and measuring global compression, we now followed pairs of states through the machine's transitions. A collision became a path to a diagonal where both members occupied the same state.

This representation was ordinary automata mathematics.

The route into it had come from a question about local contraction, trying to turn this machine into a Modal-Path-Ethics-shaped problem.

Just to be clear: Modal Path Ethics helped select the question and change the representation. It was never an assumption required by the mathematical result.

Then the computations began behaving strangely.

Across the Dżyga–Szykuła family, exact pair distances repeatedly exposed the proposed threshold of four-thirds the number of states.

The surrounding geometry was unexpectedly rigid. ChatGPT remarked that things had "escalated dramatically."

The investigation had started with a failed hunch about contraction.

Now it had encountered a structure that appeared to support a general lower bound.

And suddenly there was a much more demanding question:

Could we prove that every shorter route failed?

Over September 10–11, the work changed character.

Pair-space searches produced exact finite results. A Bellman-style potential supplied a candidate symbolic certificate against shorter collisions. Composing the machine's two commands exposed a cycle-and-fold mechanism relevant to synchronization.

We checked the source transitions, compared independent implementations, attacked endpoint and parity cases, tested altered machines, and asked symbolic solvers to find counterexamples to the proposed general inequalities.

The resulting candidate proof survived those internal attacks.

There was also a problem with the word printed in the source we used: its length did not agree with the proposed threshold. We investigated the inconsistency, recovered the corrected witness, and I even obtained confirmation from Szykuła himself about the intended expression and attribution. The correction was also part of independent 2026 work I obviously have no priority over.

• I had gone looking for a structural explanation.
  • I now possessed a computer-assisted candidate resolution of a mathematical conjecture.

That was very surprising to me.

And I had no intention of pretending that surprise constituted external verification.

So I prepared the evidence, diagrams, certificates, code, and theorem note, and sent the candidate off for specialist inspection, noting I understood if no one had time for this shit.

And so the claim remained a candidate.

On September 12, I published Revenge of the Theorem Scoreboard, which was already supposed to be about artificial intelligence outrunning the mathematical community's capacity to understand and digest new proofs.

Unfortunately, I now had one of my own.

The article acknowledged the candidate result while explicitly preserving its unverified status.

And it identified the problem I had realized I just created.

I had a large collection of mathematical evidence.

I could reconstruct how the result had been obtained.

But that did not establish that I could explain the important mathematics in any kind of form another mathematician would recognize, reuse, and carry forward.

I had just sent somebody a proof package.

I had also sent them an enormous amount of work.

This became the reason for Proof-to-Theory.

The project began the following period as an attempt to recover useful mathematical understanding from difficult proof artifacts: separate the successful construction from the obstruction that made it optimal, identify what carried the proof, find which explanations survived intervention, and produce a handoff that another person could inspect without reconstructing the entire investigation.

On September 30, I published I Gave a Mathematician Homework, including the larger Conjecture 8 explanation and reproducibility materials.

The status remained candidate proof, internally checked, independently unverified.

The article was an attempt to pay some of the explanatory debt I realized I had created.

And then Qichao Wang appeared.

His paper, The compress-with-another threshold of Szykuła's Figure 3 family, is dated October 3 and was posted to arXiv on October 5.

It independently establishes the exact threshold for the corrected automaton family and supplies a separate synchronization construction.

It also cites two Modal Path Ethics articles, Revenge of the Theorem Scoreboard and I Gave a Mathematician Homework, alongside other independent work.

Wang explicitly declines to make a priority claim.

So do I, of course. The mathematical difference is more interesting than the chronology of who reached the result first.

My candidate proof used exact pair-space computation and a symbolic progress certificate. I had spent substantial effort trying to make the obstruction readable.

Wang produced a self-contained argument using an excluded region and two entrance potentials.

He explained why the apparent shortcuts could not work.

He showed how the lower bound follows from the structure of the pair automaton, handled the small exceptional case, and separately constructed a reset word to establish synchronization.

The paper is eight pages long.

• I had approximately the mathematical equivalent of a police evidence locker.
  • Wang had an explanation.

On October 16, I published Qichao Wang Explains the Part I Couldn't.

It was an acknowledgment of a serious intellectual difference.

I had worked on this one machine intensely enough to investigate its conjecture, develop a candidate proof, construct checks, and recognize why the mathematical problem mattered.

At no point did that make me Qichao Wang.

Understanding one difficult automaton does not automatically supply expertise in the entire field of synchronizing automata.

I did not acquire his command of the surrounding literature, the standard methods, the mathematical taste involved in choosing a compact explanation, or the capacity to place the result in its proper disciplinary context.

• I could learn the object.
• I could contribute to the investigation, I guess.
• I could also recognize the quality of Wang's explanation when it arrived.
  • I was never going to write Wang's paper.

And that is fine.

His citation does not make me a coauthor. It does not verify my original proof package at all or establish that his argument depended on mine in any way. The mathematical result and its intellectual history remain available for independent inspection.

The point is that our contributions can be very different in both kind and value without making either illegitimate.

Which is precisely the possibility Makowsky's hypothetical squatter obscures.

An outsider may enter a field for reasons the specialists do not share.

They may ask an unfamiliar question, construct a new experiment, develop an unexpected result, and discover the limits of their own understanding while trying to explain it.

A specialist may subsequently establish the result more cleanly, identify its place in existing mathematics, or discover that much of the apparent novelty was already fully understood.

• The earlier investigation may still have been worthwhile.
  • The specialist's contribution remains his.

Nobody has to become the other.

And there is no requirement that intellectual curiosity first produce a marketable claim to priority.

Here I went looking for something I could learn about causal organization.

I found a mathematical machine that behaved in a way I did not expect.

I tried to understand it.

Later, I found somebody who understood it much better.

I would very much like academic scholarship to remain a place where that is considered a successful sequence of events.

Veto Mazes and You.

Unfortunately, we're not done, because Modal Path Ethics also wrote Bounded Last Words: Final Authority Without Sovereignty.

This was a much more conventional attempt to produce a scholarly manuscript.

It concerned a distinction I still consider useful.

An institution may need authority to make a final operational decision. That does not automatically give it authority over the conditions under which its decision can later be challenged, revised, or its office removed.

The argument developed through operational finality, meta-jurisdiction, constitutional pluralism, emergency authority, and the material architecture of correction.

It had a central claim I still think I can defend.

It also had approximately ten thousand words.

Many of them were doing unreasonable work.

Among the paper's more memorable expressions was veto mazes.

I still think that one is pretty good.

That paper went onto Zenodo.

I submitted it first to Global Constitutionalism, which declined it in August right away. Then to the Canadian Journal of Law and Jurisprudence.

Then I recognized it was pretty bad and assumed it just went in the trash again.

But on October 22, the Canadians rejected it following a substantive referee review.

And here is where Makowsky's distinction between having an idea and completing scholarly work becomes extremely important.

Because this manuscript was not good enough.

The central argument had become difficult to discern beneath its specialized vocabulary. Its use of the East Palestine train disaster lacked a sufficiently clear connection to the surrounding theory. Its encounters with existing scholarship too often summarized positions instead of doing the sustained argumentative work needed to confront them.

These are very real weaknesses. A reviewer should not have to reconstruct the argument the author intended to write. The reader should not have to guess how the historical case establishes the philosophical claim. And naming Bodin, Hobbes, and a collection of contemporary constitutional theorists is not the same thing as meeting their strongest arguments.

The manuscript may have contained useful distinctions. Some of its terminology deserved defense.

The manuscript itself still required substantial revision, and probably never should have been submitted in this condition.

I thanked the editor.

This still appears to have been a successful use of peer review.

The paper entered a correction process and encountered objections that should hopefully improve it.

Notice what did not follow.

• The rejection did not establish that I should never have asked the question.
• It did not make every concept in the manuscript worthless.
• Nor did the existence of its Zenodo DOI establish that that original submission had been adequate at all.

A persistent identifier had preserved an unfinished intellectual object.

The journal assessed whether that object met its requirements.

It did not.

Both instruments worked perfectly.

The Zenodo record did not need to disappear to make the rejection meaningful.

And the rejection did not need to become an indictment of curiosity to be useful and correct.

The Zenodo Shelf Is Becoming Embarrassing.

So I should probably disclose the rest of the shelf.

My ORCID research record lists four preprints.

• There is Bounded Last Words, which just received the assessment described above.
• There is Harm as Contraction of Reachable Continuation (v1.3 in progress).
• There is When Explanations Write Back.
• And there is Material Bounded Finality, with a separately archived reproducibility package.

These records are not interchangeable achievements.

Harm as Contraction had substantially more philosophical development behind it than the earliest version of Bounded Last Words. It underwent many revisions, prior-art comparisons, and engagements with neighboring work, including serious overlap with capability theory and Sangmok Kim's Structural Ethics.

That did not prevent an editorial rejection from Ethical Theory and Moral Practice on the question of sufficient novelty. And the subsequent Zenodo version did not finish the matter.

I have since independently reopened the arguments and their supporting claims for further investigation. The fact that I remain persuaded by the central insight is no substitute for completing that.

The original preprint may be useful. It is still clearly premature to treat this .pdf as the settled scholarly account of harm.

When Explanations Write Back has a different evidentiary structure.

It contains formal identification results, experimental protocols, controlled intervention data, reproducibility materials, and an explicit account of what succeeded and failed.

The Phi benchmark satisfied its individual primary criterion.

The stronger preregistered cross-architecture gate did not pass.

That failure remains in the paper. The followup is ongoing.

Material Bounded Finality is different again.

Its finite model exhaustively enumerated 62,328 reachable states under its declared assumptions. It constructed and checked the coexistence of several specified correction and succession conditions.

The paper expressly limits that result to a finite feasibility witness. It does not claim to have proven a general theorem of constitutional governance.

Some of the properties were protected by guards built into the model.

That is part of what the witness establishes and part of what it does not.

So these four records share an author.

They share an archive.

They share a general research program.

They do not share one evidentiary status, one level of maturity, or one degree of external scrutiny.

And none becomes peer-reviewed research simply because Zenodo has assigned it a persistent identifier.

Modal Path Ethics already investigated that problem in SLIME WATCH III: Ghostbusters.

The subject was a collection of synthetic academic identities and fabricated scholarly records that had acquired DOIs through Zenodo.

The identifiers worked. The papers could be permanently identified and cited.

That did not establish that their claimed authors existed or their research had occurred.

The ruling was:

A DOI is an address, not a witness.

The DOI had done the same thing for me.

It supplied a permanent address for research.

It could not tell anyone which claims were original, which arguments survived criticism, which computations had been independently replicated, or whether a journal would accept any of these manuscripts.

Some of those questions remained open.

Some have since received rather direct answers.

Thank you again, Canada.

The Slop Is Still Slop.

So this is why I cannot simply dismiss Makowsky's concern.

Cheap generation can overwhelm research.

An author who produces hundreds of plausible papers without checking their references, validating their mathematical claims, inspecting their evidence, or distinguishing speculation from results exports an enormous burden onto other people.

Readers must find the mistakes.

Reviewers must reconstruct the missing work.

Later researchers must distinguish real contributions from fabricated, duplicated, or unsupported ones.

Institutional credibility can cause this material to be taken more seriously than it deserves.

That is real research audit debt. And the existence of an actual idea somewhere inside the output does not discharge it.

Modal Path Ethics has to operate under the same rule.

If I upload an argument that has not earned its novelty claim, that claim remains unsupported.

If I submit a manuscript whose central contention is buried beneath its terminology, that is a problem with the manuscript.

If my attempt at mathematical exposition requires a specialist to reconstruct the central obstruction, then I cannot pretend that the exposition has already performed that work.

If a finite model contains assumptions and guards that make some failures unreachable, I cannot present its passing checks as universal certification.

This is where that distinction between an idea and a scholarly contribution earns its place.

A researcher may possess an interesting conjecture without possessing a proof.

They may possess a proof candidate without adequate verification.

They may possess a verified result without a good explanation.

They may possess an explanation without a defensible novelty claim.

They may possess an original insight without a publication that a specialist can responsibly rely upon.

These are all different states of being.

Artificial intelligence makes it easier to produce artifacts that imitate the final state before the intervening work has occurred.

The answer is to preserve those distinctions.

Preliminary work should be identified as preliminary.

Speculation should be identified as speculation.

Computational evidence should be distinguished from proof.

A proof sketch should not impersonate a theorem.

And an author should not treat a journal reader's labor as an unlimited resource simply because producing another fifty pages has become inexpensive.

But the reverse also matters.

An unfinished question is allowed to remain an unfinished question.

It can be recorded, circulated, tested, criticized, and eventually abandoned without becoming fraud or a territorial claim.

A research archive can preserve that history without certifying the result.

The right to investigate does not come with a right to everyone else's attention.

It also does not require institutional permission to begin.

The Marble Is Still There.

Makowsky supplies a memorable analogy.

A person puts a sketch on a block of marble and demands credit when Michelangelo subsequently produces a sculpture.

Right.

The sketch is not the sculpture.

But suppose the person discovered a technique that made a previously impossible sculpture achievable.

Or suppose the sketch identified a structural relation the sculptor had not recognized.

Then suppose the person simply said, "I think this might work, I haven't figured it out, does anybody know?"

Those are different contributions.

The first might be a technical discovery. The second might be an enabling idea. The third might become the beginning of somebody else's investigation.

Makowsky's hypothetical collapses them into one person demanding a share of another person's achievement.

This resulting person is very easy to dislike.

Unfortunately, that person cannot stand in for intellectual curiosity.

People have questions before they have answers.

Researchers form conjectures before they have proofs.

Inventors imagine devices before discovering which physical constraints will defeat them.

There is an enormous amount of intellectual activity between wondering about something and completing a publishable research article.

That interval seems to be where much discovery happens.

Why would making it easier to explore that interval be treated as a presumptive threat?

The Glory of Men.

Makowsky's imagined squatter is principally interested in receiving credit.

That is certainly possible.

Academic priority has been contested for centuries. Careers depend on acknowledgment. Professional recognition can support the income, access, and time required to continue research.

These incentives have real functions.

They can also distort the activity they are supposed to support.

Consider what happens when curiosity disappears from the account.

The researcher has an idea. But the institutional concern is whether publishing it now will interfere with the credit somebody else might receive later.

The idea becomes territory.

The publication becomes a claim.

The eventual discovery becomes a prize.

The community becomes the authority deciding which contestant deserves to have arrived at the tournament.

And the object that originally inspired the investigation?

Well, hopefully somebody will get around to that after the awards ceremony.

This is just an incredibly peculiar way to organize humanity's relationship with knowledge.

Competition can motivate discovery. Attribution preserves intellectual history. Peer review can prevent costly errors. Professional recognition can keep researchers working.

These are instruments.

The discovery does not exist to provide them with a market.

Curiosity is allowed to be very inconvenient to academic economics.

The Indulgence Market.

And then Makowsky closes with a truly remarkable historical comparison.

He suggests that flooding scholarship with artificial-intelligence-generated claims could disrupt existing intellectual-property and scholarly markets in something like the way the printing press disrupted the market for plenary indulgences.

I would like to revisit that comparison.

Specifically, I would like to know exactly who is selling the indulgences.

The printing press expanded the circulation of arguments, criticism, translations, religious controversy, and competing claims to authority.

It did not settle every theological dispute.

It did help make those disputes harder for established institutions to control through the previous channels of communication.

There were errors, lies, propaganda, conflicts, and enormous new opportunities for distortion.

Truth still required examination.

The authority to determine which arguments could enter public examination had changed.

That is the part of the analogy worth preserving.

A scholarly community must be able to assess research without drowning in worthless output.

It must also be willing to discover that some machinery developed to manage scarcity has mistaken scarcity for value.

If producing a serious research artifact becomes dramatically cheaper, the entire system of peer attention, recognition, and evaluation may need to change.

Some existing advantages will disappear.

Some professional functions will become more valuable.

Some institutions may discover that they have been charging an extraordinary amount of social effort for access to a process that could have been organized quite differently.

That deserves investigation.

The academy has valuable instruments to defend.

The market for academic glory is not the intellectual field.

Emergency Ruling.

Michael Makowsky has identified a real danger.

Artificial intelligence can make the production of apparently scholarly material cheap enough to overwhelm the institutions responsible for evaluating it.

The problem does deserve better instruments.

His speculative diagnosis of intellectual squatting does not establish the motives of the particular researcher who prompted it. Nor does the volume of that research establish its quality in either direction.

The individual claims still require inspection.

The general argument has reached the wrong object.

The value of inquiry cannot be measured by the length of time required to obtain recognition, the difficulty of entering a journal, the prestige of the researchers involved, or the market scarcity of the resulting publications.

Those measures may tell us something useful.

They can also tell us about the institutions producing the measures.

Modal Path Ethics now has several real examples of what happens when these things are distinguished.

A failed causal-abstraction investigation led me to deliberately test a structural heuristic against an unfamiliar mathematical problem. The investigation unexpectedly produced a candidate proof.

A mathematician subsequently produced an explanation I could not possibly have written.

The public record was preserved without either of us needing to annex the other.

A philosophical manuscript entered scholarly review and failed to meet the necessary standard. The reviewer had work left for me to do.

A separate philosophical preprint had stronger development behind it and nevertheless still required further scrutiny.

An empirical paper preserved its failed validation gate.

A formal paper preserved the limitations of its finite witness.

None of these events requires treating intellectual curiosity as a credential or peer review as an enemy.

The work must earn the claims made for it.

That is the standard.

Curiosity remains free to discover how much work is still missing.

I have no objection to demanding better evidence, rejecting bad papers, or refusing undeserved glory.

I would prefer not to make the right to investigate depend upon a market for glory in the first place.

The printing press did not owe the indulgence market a sustainable business model.

And curiosity does not need a landlord.