Here is how personal computer visual navigation usually seems to work:
A computer, with a mouse.
Icons and pieces of information on a desktop:
The mouse chases after fixed objects (icons and pieces of information) on a 2D plane.
I like mice within certain types of programs. Mice arrows have the ability to shoot across a screen in response to a relatively small hand nudge from the physical mouse. This makes them good for word based programs when the need to click on visual symbols is very infrequent, and also good for repetitive process modelling programs where new conceptual thinking is not required on a continuous basis because the focus is on producing a set of drawings to a certain consistent standard for example.
But mice do distance the user from closer engagement with the visual information on a screen that would be allowed for by using hands directly onto a touch screen.
In the case of the desktop, the use of a mouse for navigation results in limited options to create a meta-organisation of objects and creates a proliferation of disconnected objects.
This is how personal computer visual navigation could work for organising desktop based objects with more direct engagement with the visual information (and associated concepts):
Ability to drag objects into variable size groupings.
Ability to pinch whole groupings smaller if don’t want to focus on them at that moment.
Ability to draw links between groupings to make mindmaps. For groupings that don’t have a perceived significance to the rest of the mind map, you can leave these to the side.
I think that this ability to cluster and then link objects together through lines of connection would be a closer reflection of actual mental space and therefore a more intuitive way to organise information on a desktop, particularly for the purpose of growing really big ideas or projects from a large variety of different pieces of information and tools for organising information (apps).
Some further desktop object organising ideas: