Very nice. The current rendering of this would be to formulate with relay/resistors the 2 gate NOR as universal gate giving arbirary boolean functions and the 2 NOR crossed feedback setup as RS-flip/flop or elementary memory bit. All this can be done with hobyist materials.

Relay logic was certainly used prior to this. For example elevator call logic goes back to the 20s:

http://elevation.wikia.com/wiki/Elevator_control_system

Clearly the engineers involved had some way to think about it, but I suspect were not scholarly enough to write a paper about it. (not that I've researched this.. perhaps there were earlier papers).

Edit: Well here is a patent: https://patentimages.storage.googleapis.com/82/ae/57/d431cf0...

David Lindquist was Otis's lead engineer..

Interestingly, up until the early 1950s it was still controversial if Boolean algebra was useful for computer design. The argument was that buildable circuits were simple enough to analyze without Boolean algebra so it was pointless mathematical formalism.

Awesome. Was looking for this after reading that other awesome blog post about Boole. Once again HN to the rescue, saving me from a Google search.

Is this where ladder logic comes from?

This is great