5.5.1 Information processing stages for vision to description

The preceding diagram illustrates typical imitation. The setting is a physical fitness class where the members follow the instructor. The instructor is on the left, and the imitator is on the right.

The green and blue arrows show the stages in information processing that are required to allow the imitator to copy the actions of the instructor.

The green arrows on the left show the stages that are required to convert the visual image to a description in an 'inner language'.

1. The first stage is the two dimensional image generated by the eye
• This image does not differentiate between foreground and background
• This images does not allow easy identification of the limb position to be copied
2. We conjecture that the next stage is somewhat like vector graphics, with outlines and shapes
• These line drawings allow us to infer three dimensional shapes
• This conjecture is not critical to this investigation, but it is supported by quite a bit of evidence in neuroscience
• Another component is the detection of relative motion and flow
3. We hypothesize that the next stage allows us to match up and identify the 3D shapes with known body parts
• Once we have identified some of the shapes we can infer other shapes and orient them relative to one another and relative to the environment
• Once we have identified the body and an arm, we know where to look for the head and legs
• Once we have identified the shapes and their orientation in space, we can label them
4. We hypothesize that the next stage allows us to describe the body parts and their orientation in space, using language
• This description does not have to be in any known language such as English.
• Groups of humans have been discovered even fairly recently that do not have a system for written communication.
• We hypothesize that such a description would exist even without outer communication through speech or writing.
• The "inner" description can be in an idiosyncratic language that differs from one individual to the next even within the same species.
• We furthermore hypothesize that many species have such "inner" descriptions that are not externalized in the form of "outer" language.
• The description does not have to be exact. Precision can vary from individual to individual, (which is well known by authors and by policemen.)