Three kinds of self-models
Keep these three kinds of self-models in mind as you study the odd states of self-consciousness.
- Physical body model. This is a model of the physical body, with joints, bones, meat, weight, velocity, shape, etc. It is put into a model of the physical world, and it cannot occupy the same volume as other solids. So for example, the arm of the body cannot sink into the desk.
- Astral body model. This model has the same shape as the physical body model. However, it has no weight, is not affected by gravity, has no organs, and though it has a volume, it cannot exclude anything else from the volume (it's like turning off collision in a game). It is basically a ghost in the shell of the physical body model. It is usually a bit smaller than the physical body model.
- Viewpoint model. A point in space augmented with a Cartesian frame, to represent the orientation of your viewpoint. It represents your "true location". It is perched between the eyes of the astral body model, facing straight forward.
It is used to model where the sound you are hearing is coming from. So for example, if you hear a noise on the left, that would be modeled as a sound source to the left of the viewpoint.
It is also used to model where what you are seeing is. So for example, if you see a ball straight ahead, that is modelled as a ball straight ahead of the viewpoint.
Note that even in a blind-deaf person, there is still a viewpoint model. It is simply not used to assist in inference of audio and visual input. Even a blind-deaf person must feel like they are somewhere in spacetime. That "somewhere" is the location of the viewpoint model within the world model.
These three self-models are meant to synthesize different kinds of sensory data. Here, I am speculating since I can't find papers about this.
- Physical body model synthesizes sensory input on the body: location of localized pain, heat, coldness, touch, pressure.
- Astral body model synthesizes proprioceptive input, vestibular input, and active motor output.
- Viewpoint model synthesizes location of visual and auditory sources. Basically, if you hear something from the left, and sees something
Your viewpoint model is updated continuously by sensory input. There is a constant inference loop between your viewpoint model and the visual sources and audio sources in the world model, as your brain tries to integrate all visual, auditory, vestibular, and other senses into one consistent model of the world with a viewpoint. - If the ear senses an acceleration upwards, that would tend to move your viewpoint model up.
- If the eye sees most of the scenery moving to the right, that would tend to move your viewpoint model to the left, or turn it to the left.
- If the eye sees most of the scenery staying still, but a small blob moving to the right, you would keep your viewpoint model stable, but move the model of the blob.
The art of ventriloquism is all about confusing the viewpoint model, so that the sound source is incorrectly modelled.