Ideas and Theory - Environment - Representational Diversity Requires Dynamic Numbers

No matter what the opinion regarding details, there is one common similarity that all definitions of intelligence by all thinkers in this field share - all definitions embody:

o = f(i)

where o is output, i is input, and f is some function being performed on the input.  So, in it’s most basic level, intelligence is the result of some function being performed on some form of data.

It must be noted that for adequate input i, it should be determinable if the medium in question (f) (or rather, it’s operations) is intelligent or not by analyzing it’s output (o) because, as illustrated above, these 3 things are intimately related.

This means that any intelligent implementation, be it physical or even virtual, like our software intelligence, MUST receive some form of input.  The nature of the input will of course influence the very nature of the implementation, because, although the two may be seperate entities, they are really not seperate (mathematically, they are both part of a single equation).

So the lowest level of an intelligence is not photoreceptors or chemoceptors, for example, but the actual environment itself - the light and the airborne particles, in this example.

The environment is i in the above equation.  The environment is the source of stimuli, the data, which the animal/animat needs in order to perform the operation (f).  This means that any limitations of the dataset, will be limitations of the animat - if the dataset is not capable of somehow abstractly representing physical light, for example, than it follows that the animat will not be able to make use of an abstraction of biological vision.  The same applies for abstractions of sound, movement... and everything that can possibly be perceived in some way.  A capable animal requires a capable environment, and similarly, a capable animat requires a capable virtual environment.  And in this context, it is clear that ‘capable environment’ is defined by REPRESENTATIONAL DIVERSITY.  However, it is also defined by dynamic interactions between elements of the environment, because a static dataset is not of much use to an intelligence, because there will be no change, and therefore no relationships between changing elements - which will make it useless to an intelligent medium, because intelligence is in large part the mental embodyment of relationships between discrete elements in an environment (or in the dataset).

So my very first problem, is defining a system to represent diverse types of data, and interactions between elements of the dataset, in software.

The answer?  What can possibly be used to represent these two things with?

Numbers and equations.  Numbers represent values - and a value can represent ANY property assigned to that value - pitch, loadness, RGB color intensities, direction, speed, hardness, viscosity, inertia, thirst, and life, the universe, and everything.  And equations bind these numbers - it can represent a relationship between two or more numbers, and thereby give them meaning.  Equations make the universe dynamic - and as explained in the previous paragraph, dynamics is vital - a static universe is unintelligible.  This is simply perfect, because computers are very adept number crunchers.

Now we encounter the next problem; how should I represent values and equations in Farve?  After a little thought, it becomes clear that nothing could be simpler.  All we need is a basic element, which is simply a variable to contain a number, and a basic relator, which relates one such element to another by some mathematical function and rule of interaction.  How this is done will be discussed further in Section 2.