So my linguistics textbook gives no credence to the connectionist model, which supposedly attempts to mimic the neurological reactions to specific input in the acquisition of language. I was just curious to see if anyone had any coginitely good ideas.
posted by:
|
Kansas
|
-
Re: Connectionist Model of Language Acquistion
Mon, December 5, 2005 - 7:12 AMAll sounds greek to me... for crying out loud girl, speak English!
LOL
-
Re: Connectionist Model of Language Acquistion
Mon, December 5, 2005 - 9:04 AMNo credence, or just no mention?
When I was leaving the discipline wayyy back in 1992, I attended a brief, brief talk on connectionism and harmony theory, neither of which I'd of before. Those frameworks made lots of sense to me...someone still in the field (theoretical, not applied) a few years ago told me they did take hold and gain popularity.
Does anyone know how widely they are taught, which schools expose the students to them?
Anyway, I'm sure you did this too, but I did a search on google for "connectionism language acquisition" and got quite a few reference leads, e.g. a couple bibliographies.
I'm interested in a general text on the ideas behind connectionism and linguistics, and a search on amazon got me a fair amount of hits, so if anyone has recommendations in that area, articles or books, I'd like to hear them. -
-
Re: Connectionist Model of Language Acquistion
Sat, December 24, 2005 - 6:16 PMWell, the connectionist model as far as i know it (heard about it first in 2001) is like a network. Every word (or better lemma) is set in your brain with connections to other words/lemmata. So when you hear the word cow, you also think of horse and other animals. There is a strong connection between these words semantically. Learning a (new) language is basically improving your network's strength. If the connections between the words are stronger, you should be a better speaker, because you can more easily come up with antonyms, synonyms and other related words, like rhymes.
-
-
Re: Connectionist Model of Language Acquistion
Thu, March 8, 2007 - 1:01 AMI'm defending my Ph.D. in computational linguistics tomorrow. For a class on neural networks, I wrote a survey of neural networks in computational linguistics. My first dissertation topic was first language acquisition, so I was very interesting in the neural network approach.
The center of neural–coginitive research in linguistics is the University of California at San Diego:
www.cogsci.ucsd.edu/index.php
The most famous proponents of neural networks for cognitive linguistics is there: Jeffrey Elman.
crl.ucsd.edu/~elman/
en.wikipedia.org/wiki/Jeff_Elman
Here's some background on neural networks. Isn't Wikipedia great?
en.wikipedia.org/wiki/Neural_network
If you want to delve into a more theoretical book about neural networks, grounded in statistics — my professor described neural networks as merely nice models to incorporate statistical decision making — read Chris Bishop's textbook.
research.microsoft.com/~cmbis...npr.htm
Hope these links answer your questions. If you like, I can post a pdf of my survey paper.
John
-
Re: Connectionist Model of Language Acquistion
Thu, March 8, 2007 - 3:18 PMSteven Pinker's "Words and Rules" does a surprisingly comprehensible trial of both the connectionist and Chomskyan paradigms on the specific example of English irregular verbs, and shows where both are lacking.
