You still haven't spoken of models directly. The problem is probably at my end. I am not directing you towards the type of answer I think you should be able to provide given the ontological status of models suggested by your posts.

You learn basic Newtonian mechanics. You learn to apply some of the same equations to both a falling apple and the moon. You and I see unification in that. You feel a need, however, to supply a "wires and pulleys" explanation, where a model is the underlying mechanism (and not just something we ascribe to the visible scientific practice). You said:

A theory can be a form of description. You posit a mental version of this theory and attribute to it special unifying powers. That's what I wanted you to directly address. You need to show that you there is such an entity and how a such an entity might unify elements of the world.

I posted the quotes from Newton to show that we can usefully speak to what is visible and forego a "wires and pulleys" explanation. We see in our lives scientific theories unifying human practices in the world. To want to explain the unification by attributing what we see to an underlying mental model that has unifying powers is like wanting to explain a person's observed behavior by attributing it to an inner ghost of that person. That doesn't accomplish anything.

You suggest that by denying the role of models as I do, that I reduce people to simple trained pigeons. There's a quote from Gilbert Ryle in his book Concept of Mind that is relevant to the subject of models as well as ghosts: "Man need not be degraded to a machine by being denied to be a ghost in a machine. He might after all, be a sort of animal, namely a higher mammal. There has yet to be ventured the hazardous leap to the hypothesis that perhaps he is a man." We don't want to overlook how sophisticatedly we conduct our lives and the fact that we can and do recognize this sophistication--while omitting details of underlying mechanism.

I gather that a Cartesianist or Idealist view such as that is attractive because it appears to explain why we are sometimes in error or why we don't understand something fully. We want to posit a model or some sort of inferior representation of the real thing that then becomes the direct object of our cognitive activities. That has the odd implication, however, that we fully understand with complete certainty our inferior model. (Philosophers sometimes posit a special, error-free form of awareness for supposed inner entities called "immediate awareness" or somesuch. For example, they point out that you cannot be wrong about being in pain. I suppose, then, that you cannot be wrong about your model, otherwise that would involve the creation of a (wrong) model of the model, and the cycle would repeat.)

There is no need to talk like that. Understanding is an achievement verb. We understand toys. We just understand them more or less well. There can be many reasons why we might be wrong about toys. Understanding is constituted by proficiency, by how well we do. If you allow someone to assess their understanding purely introspectively, you are prone to get something like this, quoting an actual conspiracy theorist verbatim: "Im not a scientist but have done alot of intrest in it. It doesnt take a scientest to be educated! Check out youtube for good and informative vids on it."

The word "toy" and its many uses are not replacements for toys.

Look at this new toy train set I bought.
Go pick up your toys!
Don't toy around with these dials. You might break it.
You have quite a collection of classic toys here.
I had that toy as a kid.
Hand me that toy on the table.

Language, generally, is far too sophisticated to be replacements or pictures of the world. Rather, you will find that there are many and varied "forms of life," to borrow a phrase, surrounding the use of a word. That's why we have to learn language primarily from the context of use. Words are more like tools that we use to achieve particular ends. A hammer is not a replacement for a nail, of course, but something used to build a house.

Yes, there is more than one way to skin a ca--err, model something. You talk about a model changing here, but what we would actually see would be something like: I trip over a toy and the next time I am in such a room, I carefully look around and step over the toys on the floor. How I interact with the world has changed. We can still say that my model has changed, but here the purpose of the term "model" is not to refer to a model under my care, but to liken the change in my behavior to a change in a model (in the sense of the plastic solar system model).

If we feel a pressing need to explain why I became more careful around toys, my tripping over a toy as well as the results we have achieved by being careful would explain it. The world and the way it responds to me does shape my behavior.

You talk about "unified conceptual aspects" and "drawing a connection," but what we would see might be a teacher pointing out planets in the night sky and then configuring the model so that the students can get overhead and other perspectives on the orbs. What the students achieve is the ability to apply methods used on the model to the planets in sky. It's the system involving the model, the planets, and what the students say and do that the teacher is continually fine tuning until particular results are achieved.

The simplifications are due the solar system, the plastic model, the students, and the students' needs, and the way it all interacts. The plastic model presents some of the same visual characteristics the solar system would if viewed from a deep space perspective. The students learn to take advantage of that. To simplify is to treat the world simply, and to get away with it.

If a plastic model does not produce simplicity, then why should we think a mental model does? These sorts of problems vanish when you consider that simplicity becomes apparent only in the larger context involving the environment, the students, and how they perform in it.