What was it about slot 1 on the //e that it could do this? How and was it ever implemented? Was it only the //e or did the earlier ]['s or perhaps later IIgs support this?
From Sather's Understanding the Apple IIe on page 7-19, at the top of column 2:
Slot 1 has two signals connected which are not available at the other slots, ENKBD' and CLKEN'. These are present primarily for diagnostic and testing purposes, but innovative peripheral card designs could make good use of them. When ENKBD' is high, the keyboard ROM is inhibited and the other devices can place data on MD0-6 of the data bus when read access is made to $C000—$C01F. When the CLKEN' is high, the motherboard 14M signal is disabled, and a card in the auxiliary slot can inject an alternate master clockpulse reference onto the 14M line. If, as normally is the case, ENKBD' and CLKEN' are not connected on the Slot 1 card, motherboard pull-down resistors pull these lines low so the affected circuits can function normally.
The advantage of keyboards interfaced via ENKBD in slot 1 is that they would be compatible with software that directly accessed the soft-switches instead of using the firmware routines.
Many 3rd-party keyboards existed, but I can't confirm if any used ENKBD. Here's a long list of disability keyboards with around 30 supporting the Apple II family, but none exclusive to the //e.
The name "Concept Keyboard" was familiar to me, and there are some references to it using a large interface card, but it's listed as being also compatible with the Apple //c and IIgs, even though according to the Hardware Reference manual the IIgs didn't retain ENKBD in slot 1. The Concept A3 model appeared to use a much smaller card that doesn't use pin 35 which carried ENKBD on the //e.