Radar Perene / Archive / science

Multiple passes before publishing: internal review when there is no journal

◦ Index methodology v2.2 (working papers with DOI). See the methodology.

Science

Publishing a working paper requires nobody's permission. There is no editor to convince, no referee to answer, no response time to wait out — the repository accepts the deposit the day the author decides. That freedom is the format's greatest virtue and its most obvious risk: when nobody can say no, who says no? The house's answer is a simple principle, upheld without an audience: no study goes to the repository in the first version its author considered finished.

Internal review is the decision to put one's own text through more than one critical pass before publishing it, even when no external instance demands it. The number of rounds matters less than the principle: between "the author finished" and "the house publishes" there is a deliberate interval, and in that interval the text is treated as suspect, not as ready.

What the first version always hides

Whoever writes a study is the worst person in the world to review it on the day it is finished — not out of incompetence, but out of intimacy. The author reads what they meant, not what they said; mentally completes the step left implicit; accepts the number in the table because they remember calculating it carefully. All the vices this series describes — the adjustment that became a result, the cut that favors the thesis, the conclusion larger than the evidence — are invisible from the inside. The single pass does not fail out of laziness; it fails by construction.

Review in multiple passes breaks that intimacy in two ways: through time, because a text reread weeks later is read by a near-stranger; and through role, because each pass can interrogate the text with a different question — which is what separates rereading from merely reading again.

The house's principle, and what of it is public

Every working paper the house deposited went through internal review before Zenodo — more than one pass, always. How many rounds there are, what each verifies, in what order they happen and how they are recorded: that is internal process design, and it stays on the bench by deliberate decision. Not as a riddle — for consistency. The review process is itself a working instrument the house has been refining for cycles; describing it in detail would be publishing it, and publishing it would be freezing it.

What the reader can verify is the principle's external effect: the house's public series has dates, versions and history — and what got there got there after the deliberate interval. The public commitment is not "our process has N stages"; it is "nothing leaves in the first version considered finished". The difference between the two statements is exactly the difference between shop window and rule.

Internal review does not replace external review — and does not try to

One confusion needs to be undone head-on. Peer review — the rite we describe in What is peer review — has a property no internal process replicates: the reviewer owes the author nothing. Internal review, however rigorous, happens inside the same interest; it is a filter against error, not against institutional complacency. The house does not present one as the other: an internally reviewed working paper remains a working paper, on the rung the ladder of levels of research assigns it — no more, no less.

What internal review does is qualify the rung. Between two working papers — formally equivalent, both without external review —, the one that crossed critical passes before deposit carries less avoidable error. The reader cannot see the process, but sees its traces: versions that correct themselves openly, errata that did not have to be requested by third parties, a series that ages without scares.

The cost of having no customs office

It is worth imagining the alternative, because the market offers it in abundance: the text that goes public on the momentum of its own enthusiasm, the day it became "ready". Every error internal review would have caught becomes a public correction — or worse, does not: it stays in the text, cited, replicated, until someone outside finds it. The time saved by immediate publication is paid back with interest by the reputation of the entire series. An internal customs office costs days; its absence costs the currency with which a research house pays for everything else.

Frequently asked questions

What is internal review, in one sentence?

It is the practice of putting one's own study through more than one critical pass, with time and distinct roles, before publishing it — even when no external instance demands it.

Why doesn't the house describe the stages of its own process?

Because the process is a working instrument under continuous refinement, and the public commitment lies in the principle — nothing leaves in the first finished version —, not in the choreography. Principles hold; choreographies change.

Doesn't internal review become an echo chamber?

The risk exists and is the reason it does not replace external review. What the internal pass eliminates is avoidable error — what only the outside gaze eliminates is complacency. The two layers do different jobs.

Can someone who writes alone apply the principle?

Yes, with the tools of those who have no colleagues: the time interval before rereading, reading with a single question per pass, the printed version read as if written by an adversary. The essential part is not the committee; it is refusing the first finished version.

---

Continue the trail: The signal that only exists after waiting: the discipline of fixed windows →

House reading: the studies that crossed the customs office are at /pesquisa; what they produce daily, in the Diário.

A second, external review of material that already looks finished — the adversary's reading, commissioned — is a service the house provides to those who seek it.