I had previously thought that ISRs are not materially used, but this turns out to not be strictly true.  It is true that most of the ISRs are stubs, but while chasing down all the 8000/C000 references, I found a reference in the maskable interrupt service routine:

E521             isrIRQ_E521:
E521 86 FF           ldaa    #$FF
E523 B7 80 00        staa    $8000  ; set all kbd (high) cols high (inactive)
E526 B7 C0 00        staa    $C000  ; set all kbd (low) cols high (inactive)
E529 3B              rti

Knowing (well, believing at any rate) that 8000/C000 are the keyboard column latches, and in the context of this code:

E406             sub_E406:
E406 14 21 02        bset    TCTL2_21 2      ; TCL2; Edge3 falling edge
E409 14 23 01        bset    TFLG1_23 1      ; TFLG1; IC3 enable
E40C 07              tpa
E40D 84 7F           anda    #$7F            ; clear the 'stop disable' flag
E40F 06              tap
E410 7F 80 00        clr     $8000           ; drive all kbd columns low
E413 7F C0 00        clr     $C000           ; drive all kbd columns low
E416 0E              cli                     ; enable interrupts
E417 CF              stop                    ; place system in low-power mode; waiting for interrupt
E418 01              nop                     ; (can't fiddle with mask bit yet)
E419 0F              sei                     ; disable interrupts for processing
E41A 39              rts                     ; get to it!

It seems that the design is to drive all columns low (normally only one is low at any time during scanning), enable interrupts, then issue a STOP to halt the processor and put it in low power mode.  An interrupt will rouse it from sleep, service the interrupt (and raising all the keyboard columns) and resume processing.  Resuming means disabling interrupts and carrying on until such time as there is nothing left to do, then returning to this code to park the processor.

Not exactly interrupt-driven keyboard support, but interrupt assisted.  OK, so what is providing the interrupt?  My educated guess is the 7430 8-input NAND gate.  (You can see it in the top board picture just above the keyboard connector with the row (alternatively column!) connector:

Recall from a prior post that the keyboard seems to be a garden-variety matrix design, with 16 columns driven by a pair of 74574 octal D, and 8 rows read by a 74573 with pullups.  Also these 8 rows feed into a 7430 8-in NAND.  So, if /any/ key is down in the selected column ('selected' means a 0 bit on the column line), then one of the row lines will be low, and the NAND will go high.  So the 7430 serves as 'there is some key pressed'.  And in the context of the sub_E406 setting 8000 and C000 all low (activating all columns) would make any key whatsoever cause a 'keydown detect'.  Such a notional 'any keydown detect' causing an interrupt would wake the system, service the interrupt, raising the column lines high, and proceeding with the main loop, which would then perform a proper keyboard scan, dispatching handling as appropriate.  So I think we've figured out the main system architecture.

If you examine that closeup, you'll see that pin 8 of the 7430 -- the output pin -- is on the component layer, and you can trace it around in [Eric]'s@EricHertz photo until it tantalizingly disappears between pins 21 and 22 of the CPU, never to be seen again.

*sigh*.  Well, there's other flies in this ointment of a theory, anyway, because the /IRQ lines is active low (or falling edge).  The logic of the 7430 as keydown detector is effectively active high.  So there would need to be some inversion somewhere.  But in the absence of physics, we have philosophy, so I'm holding on to this hypothesis.  There's a 7404 inverter nearby (though out-of-the-way for that trace), and a curious transistor looking thing right next to the /IRQ line.  If it were a transistor, it could invert.  However there are no resistors nearby as would be expected, and moreover the silkscreen indicates 'LV1'.  'Q' is typically used for transistors (and indeed there is a transistor on the other side of the CPU so marked, replete with resistors).  So I think this is some other kind of device.  It's also next to the /RESET line, so this line of inquiry could be a red herring.

Welp, for now we'll file this mystery away with all the others and motor on.