In my opinion, in the example you wrote about, it can happen like this:
1) If you use nameMapping then TC searches for information about the object in NameMappingInfo and uses the data stored there to work with the actively called object in the script.
In this example, the behavior of TestComplete will also depend on the options set in this example:
Tools -> Options -> Engines -> NameMapping
2) If you don't use nameMapping (and I would recommend using it, though), you refer to objects directly via Sys.
In this example, in the "Object Browser" tab you have a list of actively accessible objects and t items related to them.
Personally, I don't know how Testcomplte reads objects from the operating system I work on. Maybe someone from other members of the forum has analyzed it before but personally I think it will be "stuck" in the code of the TestComplete application itself and you won't find it in the documentation.
If weaving, it will be answered by technical support.