This week’s news cycle seems certain to be dominated by The Washington Post’s huge scoop: its acquisition of a confidential memo, written by General Stanley McChrystal and sent to, among others, Secretary Gates and President Obama. By obtaining the highly anticipated review of the nation’s Afghanistan policy, Bob Woodward has ensured that he will shape the narrative and progress of yet another administration. And by releasing a version of the document on the Post’s Web site, the paper has ensured, laudably, that the public and other reporters get a peek inside a complex policy decision.

But what exactly are they seeing?

The Post was forthright that the document they released was not the exact version that Woodward originally obtained. Here’s how Woodward’s article puts it:

Senior administration officials asked The Post over the weekend to withhold brief portions of the assessment that they said could compromise future operations. A declassified version of the document, with some deletions made at the government’s request, appears at washingtonpost.com.

While Woodward’s phrasing isn’t crystal clear, the most straightforward interpretation of his article’s disclaimer suggests that the administration asked that limited, extremely sensitive portions be withheld, and that the Post agreed to make those deletions from some declassified version.

But that isn’t what’s implied in a short note the paper used to introduce the searchable, text-only version of the document:

The Department of Defense on Sunday evening released a declassified version of Gen. Stanley A. McChrystal’s assessment of the war in Afghanistan. The Post agreed to publish this version, which includes minor deletions of material that officials said could compromise future operations, rather than a copy of the document marked “confidential.”

This accounting clearly says that the government released a copy of the document to the Post with the deletions already made, a scenario that would have seemed to have been ruled out by the phrase “with some deletions made at the government’s request.” (After all, a request to delete requires a party to request and a party to delete—how could that phrase explain a situation where the government did both?)

Kris Corrati, a Washington Post spokesperson, cleared up some of the confusion in an e-mail to CJR.

After being told Saturday by The Washington Post that it had a copy of Gen. McChrystal’s assessment, the White House asked The Post to delay publication of any story long enough for the Department of Defense to review the document with Post editors and present a case against the release of certain information in the report. The Post agreed to that request. After a meeting Sunday with senior Department of Defense officials, the Post agreed to redact certain material from the document that has now been posted at washingtonpost.com.

The Pentagon then produced a version of the document with the agreed redactions and released it back to us declassified. We posted that version online, after confirming its accuracy.

Corrati said the Post would have no comment about whether their staff disagreed with or rejected any of the Pentagon’s requested redactions.

But once they reached agreement, it’s clear the Post outsourced the final editorial step of redacting the twenty-nine agreed-upon portions to the Pentagon.

Despite repeated requests, Corrati refused to explain why the Post allowed the Pentagon to make the deletions—rather than hear out the Pentagon’s concerns and then, to the point they agreed necessary, themselves redact the information the paper agreed would be unwise to release.

At first glance, this decision creates an obvious and troubling opportunity for the Defense Department to release a document that was, in some way, modified beyond the Post/Pentagon agreement. Corrati says this didn’t happen, and that the paper checked the document to ensure it was complete and that it “included [only] the agreed upon redactions.” Indeed, Woodward presumably has the original, non-redacted copy that the Post is still free to release—which would make it a simple matter to thwart any possible attempt by the government to pull a fast one.

“In our view, it did not matter who made the redactions as long as they were consistent with what had been discussed,” Corrati maintains.

No matter the reason, it’s likely the Pentagon undertook an unprecedented course of action by making the deletions on its own, according to Steven Aftergood, one of the capital’s foremost experts on classified information and documents.

“They essentially produced a new document,” says Aftergood, who leads the Federation of American Scientists’ Project on Government Secrecy. “It is unusual. I can’t think of a similar case where an altogether new document was produced.”

While the novelty of the agreed-upon process is striking, the Pentagon’s method of redaction deprived the public of a limited, but often valuable, category of information—one that would be intact had the Post redacted the document as it’s typically done: by going over it with a black, felt-tip marker, or, more likely these days, a digital approximation thereof.

Instead, the Pentagon chose to replace the redacted material with the word “REDACTED,” sometimes bracketed, or followed by or introduced with ellipses.

As anyone who has seen traditionally redacted documents would attest, the length of the black mark—is it about the length of a word, a sentence, or a paragraph?—tells you a lot about the quantity of the information being withheld. While there’s no reason to doubt the Post’s assertion that the twenty-nine redactions were “minimal,” it’s hard to know what that word means in the context of a document that spans sixty-six pages. (And who knows whether it would be any longer without the redactions?)

Furthermore, by extrapolating from the length and context of any given redaction, it’s often not hard to tell whether what’s missing is, say, a first and last name, or a number. Sometimes you can plainly judge whether a redaction is appropriate or not. The bottom line is that these simple black boxes are evidence of what’s missing—and a check against over-redaction.

But all that’s missing from the PDF of the Pentagon’s version.

It’s rare that we hear about situations where the government makes a request that a journalist withhold some details learned in reporting that could truly endanger lives. And while it’s tricky territory, there are certainly cases where it is justifiable to trim what, exactly, readers are told in support of some other interest. Though in this case there’s no reason to doubt the decisions, the fact is that we simply don’t have the information required to judge the propriety of these twenty-nine particular agreements.

But put that, and the missing black boxes, aside. The Post’s decision to let the Pentagon make its own redactions—even though the paper agreed with them all—seems to turn the order of things on its head.

Once obtaining the report, the Post is free to publish any portion of it that it pleases. The Pentagon is also free to make its case that pieces of information shouldn’t come to light.

But on each point, the final decision should be made by journalists. And without a good explanation, there’s no reason they shouldn’t execute those decisions themselves. To do otherwise is to farm out a portion of the editorial process—to, no less, the subject of the coverage.

Clint Hendler is the managing editor of Mother Jones, and a former deputy editor of CJR.