The easiest way is to just specify that an observation is an interaction. Any kind of interaction. Even a photon can be an “observer” and carry out an “observation” by interacting with something else. You run into no logical contradictions if you do this, although what you do find is that the variable properties of systems become variant.

Two observers may disagree as to when a particle took on a particular property. In the famous Wigner’s friend thought experiment, Wigner and his friend come into disagreement as to when the ψ wave “collapsed” and the particle being studied actually takes on a particular definite value. Although, they do not disagree as to what the value is, only when it acquired being.

Something that is variant is more like velocity. Velocity makes no sense if you do not specify a reference frame, i.e. a coordinate system. We can talk about the velocity of a train relative to a rock, relative to a car, but there is just no velocity of the train unto itself. Something that is invariant is identical in all possible reference frames. Two observers are not going to disagree as to whether or not the train is indeed a train, one’s not going to see it as a frog while the other sees it as a train. The train itself is invariant, but its velocity is variant.

If you were to state that all interactions qualify as observations, then you inevitably run into the conclusion that the variable properties of particles all depend upon a reference frame specified by ψ. Although, the specific values of particles do not depend upon reference frame, but when those values come into being depend upon reference frame.

You then do not need to posit any “intelligent creatures” or even “measuring devices.” You just have to take into account that it is meaningless to speak of the properties a system has without specifying some coordinate system given by ψ. The reason you “collapse” ψ after a measurement also then becomes rather obvious, because you change your relation to the system being observed so you have to update your coordinate system to account for it, kind of like taring a scale, and not because you “collapsed” some sort of physically existing ψ wave.

You have to do this because part of the coordinate system you can choose yourself, in the same way you can choose which reference frame you view the train from: either on the ground, in your car, or in the train itself, and in each you will measure a different value for the train’s velocity. However, what ultimately makes quantum mechanics unique is that part of the coordinate system you actually cannot choose yourself and will be randomly determined after an interaction, forcing you to have to constantly tare your scale, so to speak.

According to our interpretation, at the moment of measurement there is not a splitting of the world or consciousness, but a transition to this or that context in which a certain quantum correlation is already predetermined. Outside the context, a certain correlation is not predetermined, only the correlation itself is predetermined. Moving into one context or another corresponds to the choice of coordinate system (point of view); it is not a physical process. In that sense, the word “transition” isn’t exactly good. An observer simply discovers that he or she is in a certain context, within a certain point of view (in this case, unlike in classical physics, he or she cannot choose his or her context and cannot return to the original position). If the “coordinate system” is fixed, the correlated value of the physical quantity is fixed. So the quantum correlation is “coordinate”. It is coordinate both in the sense of the initial choice of the “coordinate system” and in the sense of the coordinate dependence of correlated physical quantities at a fixed choice of the initial coordinate system.

— Francois-Igor Pris, “Contextual Realism and Quantum Mechanics”