It is the essence of our approach that appearance and reality are quite distinct and the child is designed to discover information about reality via appearance. See [McC99], and solve the problem it presents of finding the reality behind and appearance. We take a rather brute force logical attitude by making their relations explicit.
In our formalism, both appearances of objects and physical objects will be represented as logical objects, i.e. as the values of variables and terms. Thus the ontology includes both appearances and objects.
Our examples of appearances will mainly be visual appearances, because we understand them better than auditory, tactile or olfactory appearances. However, we would like a language that applies to combinations of all kinds of appearances--whatever happens to be available.
Natural language is better at describing objects than appearances. When it has to describe appearances, it often uses objects to describe them--as in ``a cloud shaped like a lion's head.'' This is for two reasons. First, appearances are represented in thought by something like pointers to the appearance itself and thus not readily communicated. Second, appearances are fleeting and can't be fully re-examined. Our robot's language of thought could use pointers to pictures, e.g. gifs. These would be communicable.
Show a hungry child a picture of a hamburger and ask ``What's that?''
Answer: ``A hamburger''.
``So eat it.''
``Don't be silly. It's not a real hamburger.''
The most obvious predicate in our logical language relates an
appearance to an object. Thus we may have a sentence
We need logical formulas for expressing these kinds of facts. It is more straightforward to do when the appearance is visual than when the appearance is tactile. How do we describe the appearance of an object to a blind person who has not yet felt it with his hands? We share with the blind Euclidean geometry extended to what we may call Euclidean physics.