This is what I was referencing in my mind with that statement:
Under this scenario, theoretically you could have that turned off to achieve exactly that. 
Key words in context being “if only the grid and the axes were treated separately”, and each axis given the users the options to show them or not, etc.
The main thing I like when I imagine a scenario of this matter (bug free) is a ‘transparent’ option where I can see the grid and the axes, but only up to a particular ‘magnitude’ of ‘apparency’. 
I like to see the grid, when I want to see the grid, but only I’d like to see through the axes and the grid. Cause I like to see whats behind it and also even what lies exactly on it, say like if a planar surface resides exactly in the same space.
I like similar things like on a surface when I have the control points visible, sometimes I prefer to see through the surface if I’m editing points of that surface that may reside behind it or coincident ot it.
But when there’s bugs or confusing nomenclature workarounds, etc. then I just turn it off and dream of a day when the GUI is improved. 
In the meantime maybe this is a more clear iteration:
Maybe I forgot ‘visibility’ at the time