I discovered that these two circuits behave in an identical way from a loose perspective (only the following schematics tested):
R1 and R3 both had 4.3v across them, and R2 and R4 both had 5v across them.
Similarly, when R1 and R3 were connected to +5v: R2 and R4 had zero volts across them.
Is this expected behaviour? From generic diagrams of the junctions of a PNP transistor, the junctions appear symmetrical, so from that perspective it makes sense.
What's frustrating: the BD438 datasheet gives absolutely no information on characteristics for reverse operation; Somewhere I read that reverse voltage for BJTs should generally be limited to 5v. Googling for me rarely uncovers anything about reverse operation of BJTs.
It would be great if someone could demystify reverse Vce behaviours and limitations, including interpolating from datasheets absolute ratings and typical characteristics for reverse operation given the forward operation information, (if it's possible).