1.1.2 Hypotheses

To develop a new theory of language comprehension, using a thought-brain-body information-processing model that is compatible with evolution and that deals with the integration of perception and action. More specifically, we are looking at the evolution of information processing in vertebrates. The investigation focuses on function rather than on structure.

To develop a computational model that can use 2D input (as from a video camera) to control the actions of a simple skeleton (primarily it's arms and legs at present).

• Hypothesis 1: All vertebrates have an 'inner language' to control action, and to support the utilization of information from perception
• Perception and action interact in terms of this 'inner language'
• All such vertebrates use the �inner language� to imitate perceived action
• All such vertebrates use the �inner language� for utilizing past experience (memory) for controlling current action
• All such vertebrates use the �inner language� for predicting and planning
• Hypothesis 2: Information processing requirements and constraints can help to explore the evolution of mechanisms to accomplish the information processing
• Hypothesis 3: The mechanism for information processing is inherited, but not the programming and data, so that the young of each species have to use various forms of learning to acquire the data and programming.

To develop a model that optimizes utilization of limited brain capacity

• Hypothesis 4: The �inner language� evolved to handle information compression and more complexity
• Only approximately 1011 neurons are available, which, at 1 bit per neuron, is not a lot of storage for perception and action. At 1 neuron per transistor, it makes for an impressive processor, but presumably quite a bit is used for buffering (cache), short-term memory, and long-term memory.

To develop theories of 'outer language', which will be explored in future projects:

• Hypothesis 5: All vertebrates have external communication with other members of the same species, based on action patterns such as gestures, postures, and sound production. The external communication is based on action, and thus is driven by the 'inner language'.
• Many vertebrates have external communication involving gestures and postures.
• Many vertebrates have external communication through sound production.
• The external cummunication is structured, repeatable, and similar between individuals of the same species.
• The external communication is recognized and acted on consistently by other members of the same species.
• Hypothesis 6: The �outer language� evolved to handle social interaction, cooperation, and finally human language communication
• Mating rituals and dominance displays are likely candidates

Since the mechanisms for language processing in the brain are not large persistent structural elements such as bones, the method of investigating fossil records is not pertinent. The data and programming we hypothesize as utilizing this mechanism is even more ephemeral. We are also not looking at biochemical structural elements such as DNA. Rather, we are looking at functional elements that can account for observed and conjectured behaviour. In other words, we are investigating information processing requirements for types of behaviour such as imitation, chasing prey, and remembering and then repeating behaviour sequences.

To investigate information processing requirements, we utilize simulation and computational competence. We assume that the actual mechanism in the members of a given species is equivalent and is working on similar data and programming to the computational model - to produce similar action sequences.

• We also analyze negative results. We can show that certain methods of information processing are inherently inadequate to generate the behaviour under consideration.

Behaviours under investigation

We have chosen types of behaviour that involve whole-body physical actions. We conjecture that the essential characteristics of these types of behaviours are common throughout the evolution of the vertebrates. (The list, and the behaviour specifications, are open to change.) Those behaviours that are present at various stages of evolution must be based on meeting the information processing requirements, and thus show that the information processing functionality must have been available. Examples of these behaviours are:

• simple learning and adaptation
• learning based on imitation
• prediction and planning required for picking up food or chasing prey
• generating novel action sequences for searching, playing, and experimentation