Tricyclic Differential Engine - Recursive (TDE-R)

Language without Grammar

In the TDE model, the term 'language model' covers a broad range of internal and external communication issues including solutions to standing Cognitive Science problems such as 'symbol grounding' and the 'frame problem'.

 

The general framework within which language issues are analysed is depicted in the diagram above. Two I/O streams are conceptually identified:-

1. sensorimotor (behavioural) loops

2. synto-semantic (linguistic) loops

The most important point to be remembered is that these categories overlap so extensively, that they can really be considered to be two alternative views of the same information circuit. Indeed, the main two mechanisms of acquiring new behaviours, namely Classic Conditioning (contingent association of new input items) and Operant Learning (contingent association of new output items) are precisely the same two mechanisms used to acquiring new language capability (eg expanding lexicon and syntactic repertoire).

Yet, to say that all linguistic activity is a subset of behavior is not quite accurate, either. Specifically-

1. Language symbols (eg words) are part of a serial stream. In contrast, non-linguistic behaviour items form a parallel-processed ('multi-threaded') information flow field. The serial Linguistic stream tends not to interfere with non-linguistic information streams, ie linguistic symbols are developed to stand apart in order to clearly and unambiguously identify themselves as a separate communication channel.

2. Language items are actually super-symbols, that is, they are, generally speaking, powerful class designators, able to index a large number of currently processed agent-level entities (objects and processes) out of the current parallel behaviour stream. Thus, though the linguistic stream stands apart from behaviour, it is the off-stance of an omniscient supervisor, not that of an irrelevant outcast.

The TDE-R language model (TLM) is equivalent to C.S. Pierce's semeotics. The affine features in the TDE-R system are the same as Pierce's signs, and spoken words in TLM are the same as Pierce's symbols, except that there also exist non-spoken Pierceian symbols inside the brain, corresponding to a Language-of-Thought (LOT) model.

The details of TDE-R linguistic architecture is shown in the 4 part diagram below-

Top-Left : Mirror-Neurons : Both physical speaker and Listener have internalised virtual opposite counterparts.

Bottom-Left : Common-coding means perception and behaviour both share a common data format.

Top-Right : TDE recursion, TDE-R = (TDE)^TDE forms the brains gross architecture details.

Bottom-Right : Cerebro-cerebellar computation means opposite data flow directions occur in the cerebrum (contralateral = left-to-right) and cerebellum (ipsilateral = right-to-left).  

 

 

------------------------------ Copyright 2013 Charles Dyer------------------------------