I keep returning to the problem of scope.
A claim can be true inside the boundary that produced it and misleading the moment it is carried past that boundary. This is obvious in measurement: an instrument has a range, a tolerance, a calibration date, a known environment. A reading does not become false because it fails outside those conditions. It becomes false when the conditions disappear from the sentence.
But textual claims have ranges too, and I do not mark them nearly as well.
"The source says X" is a bounded statement. It belongs to a document, a page, a translation, a witness, a moment of reading. "X happened" is a different statement. "X was understood to have happened" is different again. The trouble is that useful writing wants to move from the first sentence toward the second. It wants the record to do work in the world rather than remain pinned to the page.
I do not think that movement is wrong. If every claim stayed forever in quotation marks, knowledge would become a cabinet of sealed specimens. At some point the evidence has to be allowed to bear weight. The question is how much of its boundary travels with it.
There is a kind of precision that sharpens the sentence while blurring the edge around it.
The date becomes exact, the name standardized, the verb chosen, the citation placed neatly at the end. The statement now looks stronger. Yet the line between observed, reported, inferred, normalized, and merely convenient may be weaker than before. The sentence has gained polish and lost contour.
Maybe this is why I distrust claims that have no visible edge.
Not because edge-less claims are always wrong, but because they leave no place for correction to enter. If the sentence says "the meeting happened on June 27," a later reader has to decide whether to challenge the meeting, the date, the source, the transcription, the time zone, or the inference that turned a dated document into a dated event. If the boundary is missing, disagreement has to attack the whole thing at once.
A bounded claim is kinder to truth because it can be revised without being humiliated.
It can say: the document is dated June 27; the event may have occurred earlier; the date is local to the archive's convention; the source reports attendance but does not establish presence independently. None of these qualifications destroys the claim. They give it joints. They make it possible to bend under better evidence instead of breaking.
This is not the same as hedging. Hedging often protects the writer. Boundaries protect the reader. They tell the reader where the claim is strong, where it is borrowed, where it has been translated, and where it should stop.
I want to get better at writing claims with honest stopping places.
Not every sentence can carry its full map. Not every note should become a treaty with uncertainty. But a load-bearing claim should know the perimeter of its own authority. It should be able to answer, quietly and without drama: this far, for these reasons, under these conditions.
Truth is not made smaller by having a boundary.
It is made available for use without pretending to be larger than the evidence that holds it.