Strictly speaking, your dog sees phenomena which he can't pass through, we call it a "WALL" or a "STEP". If the object can be moved by a certain motion and then pushed or pulled, we call that a DOOR.
The best I can guess, is that a dog puts together an association of ideas: something like PRESS, PULL and the the idea of an OPEN AREA: the garden or the street, etc.
They lack linguistic concepts, so I have to assume that whatever goes in inside the skull, is an extremely watered down version of what we do.
Yes, I think most animals have an idea of orientation, which is why many newborn animals don't jump off a nest or off a table as soon as they're born, or why sea turtles known exactly where to go as soon as they are born.
So, It seems clear to me that our differentiation of objects cannot be arbitrary or entirely dependent on us. — Janus
I agree that it is not arbitrary: far from it. The "entirely dependent" part is very tough. It depends on what we assume the object must have, absent us.
What I would stress is, regardless of the world, what matters is how the creature reacts to the stimulation, more so than the actual world: moths flying to lamps (instead of the moon), dogs mistaking toys for food, tigers mistaking mirrors for other tigers, etc.
Granted, I am giving examples of deviation from the norm, but what I think this shows is animals react to stimulations,
regardless of if the trigger is the one the animals thinks it is.
You can reply that the tiger is reacting to a property of the mirror and the moth to a property of the lamp. And in a sense it is true, yet we would not want to say that a mirror is another tiger, nor a lamp the literal moon. So we may say certain properties of light, are mind-independent.
That' the difficult area for me.