Okay, I have a couple of DMMs with diode check mode. I did as you said, and using that method the Q1338 shows a diode voltage of 0.773 from base to collector and the same (or nearly so) from base to emitter (red always connected to the base, black to either the collector or emitter), which means that this is definitely an NPN transistor, and the cheap component checker was wrong.
Okay, that clears up the core of my confusion. I'm not nearly as concerned with specific part numbers in different service manuals as I am with a completely different sort of transistor in the actual circuit than is specified in the schematic.
As for whether I got the pinout right: I had thought that the component checker didn't care and would try all combinations, but I just tried putting the transistor in with the pins offset by one position (the checker has a ZIF socket where the pins are labeled 1, 2, 3, 1, 1, 1, 1 so you can connect the part either as 1,2,3 or as 2,3,1, or as 3,2,1 if you flip it over), and now it gets correctly identified as an NPN transistor. This is what I get for $15: no manual, and a cargo-cult approach to electrical engineering.
For the record, the other three transistors that I used the component checker to identify as bad, I actually went back and checked them with the multi-meter, so I'm pretty sure that they were, in fact, bad: one was a dead short between all three terminals, the other two were complete open circuits between all three terminals. Clearly something fairly dire happened to the beam intensity amp which blew out two diodes and three transistors. I just hope I've fixed whatever the problem was.
-- Jeff Dutky