When three reports describe the same document and reach three different impressions of what it says, the most useful question is not which report is right — it is which sentences from the underlying document each report quoted, and which it paraphrased. The exercise below is unglamorous on purpose: it sets the reported language alongside the source language and notes the distance between them.
We selected the three highest-traffic accounts of the subject filing published in the two weeks following its release. Each is a reputable outlet. Each cites the same forty-one-page document. The texts of all three reports, and the document itself, are linked under Sources below. What follows is not a critique of any reporter or institution. It is a structural reading of how condensation happens, and what kind of context tends to drop out when an extended record is summarized in eight hundred words.
The "concession" that was not a concession
The first claim, repeated across all three reports with slight variation, was that the filing "concedes" a particular factual point. The word concedes appears in none of the three reports as a direct quote from the document — it is supplied by the reporters as a paraphrase. The paragraph it paraphrases reads differently when set beside the paraphrase.
As reported
"The filing concedes that the underlying meeting took place."
In the record
"The respondent does not contest that a meeting occurred on or about that date, but disputes the characterization of its purpose and the identity of one of the parties named."
The two formulations are not equivalent. "Concedes" implies a retreat from a position previously held. The filing's actual posture is narrower: it accepts a date, declines to accept a characterization, and disputes a name. None of the three reports quote the second half of that sentence. The first half, standing alone, carries a connotation the full sentence does not.1
The difference between "concedes" and "does not contest" is the difference between a retreat and a precision. The filing performs the second; the coverage describes the first.
From the analysis below.
Sequence and the appearance of consequence
The second pattern concerns order. In all three reports, the order in which two events are introduced creates the impression of a causal chain. The filings themselves describe the events as concurrent and unrelated; the chronology is laid out across pp. 18–24 of the document.
A reader who encounters the events in the order the reports present them, joined by the connector "in response to," will form a view of cause that the underlying document does not support. The events may, in fact, be related; the document does not say so, and the reports do not cite anything outside the document for the connection.
Silence read as position
The third pattern is the most common and the least visible: an absence of comment read as agreement. The subject declined to comment to two of the three outlets. All three reports report the declination accurately. One of the three then describes the declination as "consistent with" the position attributed to the subject; the other two leave the inference to the reader.
A statement on the record from the same subject — given to a fourth outlet earlier in the week — addresses the point directly. None of the three reports cite that statement. It is publicly available; we link to it under Sources.2
Timeline of the coverage
-
May 2 · Morning
Filing released to the public docket
Forty-one pages, including the paragraphs reproduced above. No press embargo; available immediately on the relevant docket.
-
May 2 · Afternoon
Subject issues a written statement to Outlet D
Statement addresses, by name, the characterization at issue in Section 1 above. Not subsequently cited by Outlets A, B, or C.
-
May 3 · Morning
Outlet A publishes first report
Uses the word "concedes." Cites pp. 11–13 of the filing in a hyperlink.
-
May 3 · Evening
Outlets B and C publish
Both adopt the framing introduced by Outlet A; one extends it with the "in response to" sequence discussed above.
What this analysis is and is not
This is a structural reading of three reports against a single document. It does not argue that any of the reports is dishonest, or that the underlying subject's position is correct. It argues something narrower and, we think, more useful: that there is a particular kind of context — the second half of a sentence, the absence of a connector, a statement made to a fourth outlet — that consistently drops out when an extended record is summarized under deadline. Readers who want to form their own view will find the document linked below.
Frequently asked
Why focus on framing rather than on the underlying facts?
Does this analysis claim the reports were inaccurate?
How was the underlying document selected?
Sources
Notes
- "Concedes" is, of course, a permissible paraphrase in many contexts. The point is not that the verb is forbidden, but that the verb supplied by the reporter carries connotations the verb of the source ("does not contest") does not — and the second clause that qualifies it ("but disputes the characterization") is omitted. ↩︎
- The statement to Outlet D was published five hours before the first of the three reports examined here. It was indexed by major search engines within the same window. We make no claim about why it was not cited; we note only that it was available. ↩︎