GOLEM
Did gyre and gimbal...
While the TDE diagram (figure 4) explains external aspects of human language via its INTERsubjectivity at TDE level 3, it does not explain its internal aspects, called i-language by Chomsky. This problem is corrected in figure 5. Both TDE fractal architecture and GOLEM diagrams are needed to fully understand the mind's subjective and linguistic features.
GOLEM diagram
In figure 5(a), the thresholded input and output channels are depicted as being orthogonal (at right angles, independent ) to one another, such that the output channel is plotted horizontally and the input channel is plotted vertically. These channels are depicted side-by-side in figure 5(c) - this is called the GOLEM, or G(oal)-O(riented) L(inguistic) E(valuation) M(emory). The GOLEM diagram depicts the linguistic view of cognitive computing. While ALL cognitive computing is structurally linguistic, there is a central band in the GOLEM diagram where language proper is governed - see figure 5(c). This part of the model matches the duplex data path through Broca's/Wernicke's areas in the cerebrum, thereby providing more empirical support for TDE/GOLEM theory.
Constituency (Chomsky) vs Dependency (Montague)
How are these alternative viewpoints helpful? By means of the GOLEM diagram, the linguistic view of cognitive computing, we were able to finally understand the problem with Chomsky's so-called 'minimalist program'. As part of the 'language wars' a debate had arisen concerning a choice of grammars, constituency grammar as per Chomsky, and dependency grammar, as per Montague. By constructing the GOLEM diagram, one can see that this is a false dichotomy- rather BOTH formulations should be used, each depending on whether one's role is that of speaking/writing or listening/reading.
The producers problem- you know what the semantics (meaning) is, after all, that is your cybernetic-syntactic goal, to convey your intended meaning, by constructing an appropriate message, within which to encapsulate it. The constituency 'grammar' is the right choice.
The consumers problem- you know what the syntax (message) is, after all, that is your cybernetic-semantic goal, to extract the transmitted meaning, by deconstructing the message that carried it. The dependency 'grammar' is the right choice.
Semantics determines Syntax
To give Chomsky his due, he was the first to suggest the idea of i-language- that all cognitive computing was linguistic in its structure. TDE theory is based on a pan-semantic viewpoint which owes its origins to the i-language concept. Under this viewpoint, semantics determines syntax, and not vice-versa, as Chomsky believes. Cognition is driven by pan-semantic goals, at the knowledge level. When we use language, we do so either externally, to communicate meaning, or internally, to consolidate it. It is knowledge (a.k.a. meaning) which is the mind's lingua franca. The semantic form is based on predicate logic, equivalent to a potentially bottomless hierarchy of subjects and predicates, topics and facts. When we speak, or listen, write or read, we do so with only one goal in mind- to linguistically manipulate each others store of knowledge, either autobiographical in the LCH, or semantic, in the RCH. This is a cybernetic viewpoint just as much as it is a pan-semantic one. Sure, the brain is a computer- it is not, however, syntactic (procedurally coded, like a Turing Machine) but semantic (declaratively coded). In the following chapters, we will investigate implementation of a semantic computer according to TDE principles.