17 August 1908
Repeated the comparison field after midnight. Seeing fair. No cloud. Plate 17 shows the same narrow quadrilateral among the fainter points.
Reduction gives parallax 0.023″. Recalculated twice. Distance absurd but not impossible.
Second reduction gives 1.18″. There is no arithmetic route from the first value to the second. Both emerge from the plate depending upon which comparison stars are admitted.
There is an error in the assumption, not the observation.
If the Earth is taken as fixed, the displacement becomes regular.
The stars move. The observer moves. This is not the same statement.
Parallax has always been treated as a measure of distance because we assume the baseline lies in space. But what if the baseline is the thing being displaced?
The stars are stationary. The observer is being displaced through something that is not space.
I believe Plate 17 is not a picture of an object. It is the shadow of a relation.