I don't know if there is a name, but this seems to be something you'd do to conserve memory.
First, a basic tile is very low resolution - just a few pixels across. But when rendered, they are magnified 2x, 3x, 4x, etc. and are much more "blocky" on the screen.
Next, older games will have a block of memory dedicated to screen display - what is in that memory is what's shown on screen. Keep in mind that memory is VERY tight in older platforms, so as a developer you want to save every little bit you can.
So rather than put your tiles into a separate memory space, you just make them part of your screen memory, thus saving a bit. And the artifact of this of course is that they show up on the screen.
From the perspective of memory access, the code doesn't care if the memory being accessed is being used for screen memory or variable storage. Basically, memory management in older games was very simple and there was just one block of memory available. There was no GPU, so no idea of separate processors and memory for programming versus screen display.
To give a sense of how tight memory could be programming earlier games, here is a quote from a 1983 article about programming the Atari and VIC computers, "Atari's super hi-resolution screen (and the 16-color GTIA modes) uses almost 8K of RAM." That seems like very little, but keep in mind the computer perhaps only had 48K of memory, a lot of which was taken up by other things besides the game.