Further to @Plutos comments, and ref: the Coventry Canal and others, there simply wasn't a standard way of doing things, or even a standard performance spec. For example, locks on the Neath Canal only had one ground paddle at the top, and never had top gate paddles, speed was presumably not of the essence - other canals may have regarded speed of operation more highly from the outset.
Pluto has done far more research than any of us (probably more than the rest of us put together!) but my suspicion would be that (1) most earlier locks the design was based on what they already knew from nearby navigations (2) once ground paddles were understood gate paddles were not used at the top (with one possible caveat - see below) and (3) two ground paddles at the top was as much about redundancy as it was about speed, and even then may have largely been "other locks have two paddles" - as a general rule if something worked then one didn't fiddle with it.
Caveat - some locks on the main system only ever had one top ground paddle, notably the T&M east of Stone, and Marple Locks - Pluto and I have walked up Marple Locks looking for evidence or otherwise for top gate paddles when new, but there is none. On the T&M locks it would seem odd to have some locks with two top paddles and some with only one, so my suspicion is that those locks with only one top ground did have top gate paddles, but is is only a supposition. At some point the T&M standardised on two top ground paddles as the newer locks at Meaford have this whilst the top lock has only one.