MadSci Network: Neuroscience
Query:

Re: Do connectionist approaches imply that there is no central executive unit?

Date: Wed Feb 21 07:12:31 2001
Posted By: Benoit A. Bacon, Post-doc/Fellow, Psychology, University of Glasgow
Area of science: Neuroscience
ID: 981982247.Ns
Message:

Dear Katja,
 Connectionism is one of the fastest growing approaches in 
psychology. In the last 15 years, it has taken its rightful place as a 
prominent perspective in our quest for the understanding of  mental 
phenomena.
 As you already know, connectionism is an attempt to explore the 
computational properties of brain-like mechanisms. Some 
connectionists regard their models as real networks of real neurons, 
whereas others see their investigations as an attempt to directly 
understand the computational structure of thought. It must also be 
remembered that the term "connectionist networks" actually include 
more than 50 categories of networks currently under investigations 
(for a review, see Ellis and Humphreys, 1999).
 This wide variety in networks makes it difficult to answer a question 
like yours. The difficulty is increased tenfold when one considers that 
most models tackle only a restricted part of mental processes (e.g. 
memory, language, vision, attention) with the consequence that any 
"central executive unit" postulated commands only a part of the 
whole.
 You mention the "central executive system" as the postulated 
controlling unit of working memory (Baddeley, 1986) and you link it to 
the notion of "self". The actual role of  Baddeley’s central executive is 
still a matter of much debate more than 15 years later, and it is 
refereed to by its detractors as a "black box", meaning a component 
whose inner workings are undefinable. Not unlike the mysterious 
"self", is it? Still, this central executive reigns only over working 
memory and was not meant as an expression of the cognitive, 
metacognitive and moral processes that define the self. The all 
encompassing connectionist model has not been elaborated yet, as 
even models of individual processes are still "in development".
 Later you wonder if "our brain makes decisions based on previously 
internalized knowledge" and whether "we come to realize these 
decisions only afterwards". This is a question that has bothered 
humanity since the ancient Greeks, and is has become one of the 
most famous debates in philosophy and psychology: "determinism 
versus free will". Are indeed all our decisions the arithmetical result 
of our past experiences (determinism) or do we have the possibility to
express our own, personal, original conclusion (free will)?
 If I understand you correctly, then, the "sensible link" that you 
propose focuses on the concept of "self" in decision making, and 
taps into connectionism for insight about the existence of free will. I 
think this is impressive and that it shows interest for the larger 
questions, which is often lacking in empirical scientists. I’m not so 
surprised, however, that you were told to "scale it down" for your 
dissertation. 
 It seems that you are more interested in high level processes (e.g. 
decision making) than in more basic systems. Many connectionists 
are tackling high-level cognition, thinking and reasoning and their 
developing models are more likely to include elements relating to 
self or consciousness issues (for a review, see Ellis and 
Humphreys, 1999). I should however point out that the study of more 
simple sensory systems, especially vision, has been known to give 
rise, from time to time, to great insights into the workings of human 
consciousness (e.g. Kolb and Braun, 1995, Crick 1996). 
 I hope that this will help you in finding a suitable subject for your 
dissertation. Remember that scaling down your topic is best 
achieved not by narrowing your views to small ideas, but rather by 
focusing on a specific concept within the larger framework.
 Good Luck!
  Benoit  
  
References:
Baddeley (1986) Working Memory. Oxford: Oxford University Press.
Crick (1995) Nature, 379, 485-486.
Ellis and Humphreys (1999) Connectionist Psychology. East Sussex: 
 Psychology Press.
Kolb and Braun (1995) Nature, 377, 336-338.




Current Queue | Current Queue for Neuroscience | Neuroscience archives

Try the links in the MadSci Library for more information on Neuroscience.



MadSci Home | Information | Search | Random Knowledge Generator | MadSci Archives | Mad Library | MAD Labs | MAD FAQs | Ask a ? | Join Us! | Help Support MadSci


MadSci Network, webadmin@www.madsci.org
© 1995-2001. All rights reserved.