When a human being is born, he learns body coordination just like language only through the presence of his body,

Yes. For example I remember reading about a simulated kitten.

The artificial neural network running the kitten program was the size of a few large refrigerators. The kitten body was a small robot. The "brain" of the kitten had no way of detecting its "true self". All sense detection was exclusively through the kitten body (and attached sensors).

One of the points I found interesting was the idea that even if you had a "human" brain, or an artificial neural network capable of (roughly) human capability, IFF you could only "see" through kitten eyes and hear through kitten ears and speak with kitten vocal chords, you would be indistinguishable from a kitten.