The differences in length and # of pins of the two sacrificial ICs were unimportant for this exercise: What matters is the different widths.
I unbent the top chip's right side pins, then soldered its left side pins to the bottom chip in the usual way, with one pin bent out that might be n.c. on a bottom chip, or an address bit a bottom chip might not have. Then I soldered the unbent pins to the bottom chip's pins with short lengths of stripped wire wrap, with a wire on one pin for a hypothetical /CE. If both chips would be 28 pins /WE would be at the same location for both, but Write for the top chip might have additional (or conditional) control, such as Write Protect. Hence the 2nd wire there.
This was as much a test of my soldering -- which has lately become noticeably unsteady
-- as of the feasibility of piggy-backing ICs of different widths.
But I didn't notice any examples of IC's of different widths being stacked.