Hi Michael,
thanks for your question. I am very glad to answer, since this is a main point in my essay and maybe I was not explicit enough here. Especially because the view I want to push forward is very unorthodox.
First of all: there is only one kind of interaction. The distinguishing feature is in the kind of object or system, that is interacting with the primary object, which makes it a measurement instrument or not. I did not want to enter the discussion of what makes objects a good measurement apparatus. Symmetry of the measurement apparatus certainly plays a role on whether an apparatus is capable to correlate to some quantity that one wants to measure.
It is enough, that the interaction is such, that it can correlate quantities of one object to an other objects. In footnote 3 of page 2 I even claim, that interaction might be uniquely derivable from invariance regarding the symmetries, and the ability to correlate the variant quantities contingent properties of an object (relative position, velocity and spin for instance). But this is still under investigation.
One point is regarding Wigner's friend type situations (but maybe in general): Since every observation is always indirect, we must rely on the validity of some laws of nature (our theory) in order to be able to rely on the result of the measurement. If my lab is shaking because of a train is passing by then I cannot rely on having on the results that have been measured. In quantum measurement, I need to rely on symmetry and unitarity in order to be able to make valuable propositions on the measure state.
Another point is a bold ontological and epistemological claim I make that is central to my essay. The idea comes from Poincaré. There is a hierarchical dependency of definitions.
Only if the free object is defined (by his normal behaviour - the free equations) then external forces can be defined (by causing a a change of the normal behaviour - change of momentum). Only if all this is set empirical claims can be made. But in order for these empirical observation to manifest at all, the system must be separable from the environment.
The bold ontological claim now is, that under different environmental conditions different symmetries and hence objects and laws manifest. Objects are contingent and emergent parts of reality.
I hope this becomes clearer. I will discuss it again in your blog.
Luca