Intelligence?
 

Ron Cottam, Willy Ranson & Roger Vounckx

Abstract
 

In a very specific way, intelligence and information-processing are synonymous.

Information-processing takes place between differently-scaled models of a relevant context. The way in which these models relate to each other, and therefore the result of an inter-scalar processing, depends on the constraints which are imposed on inter-scalar computation, and on the manner in which those constraints are applied.

Mono-scalar ‘systems’ can always be deconstructed more or less precisely to a set of rules: a single non-fragmented scale corresponds to a single formalization. Any consequent ‘systemic’ complexity is the result of formal incompleteness or of our lack of understanding.

Multi-scalar ‘systems’ constitute artificially- or naturally-constrained hierarchies, where the style of processing depends on whether the inter-scalar constraints are externally imposed or internally recursively generated by the information-processing itself. These two types of hierarchy, artificial and natural, have very different properties, and exhibit completely different styles of ‘intelligence’.

The unification of a processing assembly into a ‘system’ is always through our intervention, whether at a single scale or across multiple scales of operation. If an artificial hierarchy is not to be scale-fragmented, it must possess some kind of cross-scalar coherence, imposed through our manipulation of the inter-scalar constraints. A natural hierarchy generates this cross-scalar coherence itself, through an autonomy-negotiation between its various scales, creating a hyperscalar system.

This appears to be the ‘meaning’ of intelligence: it enables a multi-scalar system to operate as if it were simultaneously multi-scalar and mono-scalar. Different individual scales of operation retain a context-dependent degree of autonomy, but the entire assembly is unified at a hyperscalar level.

A ‘system’ is always hyperscalar, whether through artificially- or naturally-imposed constraints It is questionable whether it will be possible to generate sufficient cross-scalar correlation in an artificial information-processing assembly to exhibit an interesting or useful degree of independent ‘intelligence’.

______________________________________________
______________________________________________