Symbols
March 2, 2005
I'm taking an interesting class this semester on Self-Organizing Systems. So far, this has actually been a seminar on the evolution of language. In particular, how can we get computer agents to evolve their own languages for communicating about particular domains.
As part of this course, I read the book "The Symbolic Species" by Terence Deacon so that I could give a report on it. His thesis, as I understand it, is that language use requires a symbolic mental framework. He believes that symbol use evolved in humans from the use of ritual to communicate. From this point, languages evolved naturally and our brains with them. This explains why we don't encounter any simple languages in nature and why animals seem to be incapable of learning human language, despite their advanced problem solving abilities.
I found the book quite intriguing, especially as a counterweight to the innatist theories of language development. Deacon believes that there are some structures in the brain that help us understand language, however it is more important to recognize that languages evolve towards those structures which are easiest for children to understand. However, no amount of special hardware can make language possible without the capacity for symbolic thinking.
Of course, this raises the question of how symbols form, what they really mean, and what sort of selective advantage symbolic thinking might have for our species. That is, it is clear today that symbolic thinking has provided a huge benefit to our species, however it is unclear what sort of reproductive advantage it would have given to an individual or a group in the first 190,000 years of human history, before the development of civilization.
Finally, would it be possible to teach machines or animals to think symbolically? Can we use rituals as Deacon suggests to motivate this symbolic thinking? What would ritual even mean in the context of machines? The book provokes heavy thinking.