Cognitive Architecture

 

Mind (thought) and language involve mental operations which are at the same, high level of computational abstraction. Indeed, a persistent belief of cognitive science is that an I-language (I for 'internal', first attributed to Noam Chomsky) powers all higher thought processes.

 

One of the main reasons for the success of the TDE project was the way the TDE-R cognitive architecture fitted the known epistemological overview of the human mind. We know that the mind manipulates knowledge, indeed, that is the way we distinguish mind at the higher level from brain, the lower level which supports it. The commonly quoted statements that brain=hardware and mind=software may be approximations of the truth, but they are not very far from it.  From lesion studies and other experiments, we know that declarative knowledge (defined as 'know-THAT') is managed in the forebrain or cerebrum (containing the cerebral cortex), while procedural knowledge (defined as 'know-HOW') is managed in the hindbrain or diencephalon, the older part of the CNS which contains the brainstem, cerebellum and basal ganglia. Declarative knowledge, in turn, has two sub-types, episodic knowledge ('know-WHEN', that is, events), and semantic knowledge ('know-WHAT', that is, facts).

 

These three types of knowledge also emerge from the design of the TDE-R, quite independently of external evidence. Rather, they emerge from consideration of the internal information storage and flows of the TDE-R computational model itself. To see exactly how and why, the properties of the TDE must first be discussed, because the TDE-R (R stands for recursive) is the result of the global application of the TDE bio-computational template to the neuroanatomy of the human CNS. The TDE consists of four main 'lobes', each one inheriting the name of one of the lobes in each of the two cerebral hemispheres. They are classified along three dichotomous axes according to the kind of computational memory storage task that they perform. The first option is VOLATILITY, the choices being either short term buffers (matching the 'eraseable-squares' on a Turing Machine's tape) or long-term archives (matching its 'fixed-squares') .  The four TDE lobes are further sub-divided into two buffer-archive pairs, giving the second dimension, MEASURE, one option which handles spatial (position and length-based) data, and the other which performs the same functions for temporal (time and period-based) data. The final ABSTRACTION dimension decides whether the TDE template applies at the local, cybernetic level or the global, cognitive level. These paired dichotomous relationships are tabulated below, as well as being depicted on the cognitive architecture isometric perspective.