This is the third of three postings on symbolic thought.
Symbolic thought requires a community for its maintenance. It is
associated with the following:
1. long-range planning,
2. spatial organization,
3. advanced language (tense and aspect),
6. long-range movements,
7. status differentials,
8. complex social organizations,
9. religion, and
10. sexual selection,
11. extended childhood,
12. survival of post-reproductives, and
It is not associated with:
1. Basic technology,
2. Basic language (nouns, adjectives, imperfective verbs),
3. Social organization,
4. Burial practices,
5. Care for the aged,
6. Pair-bonding, and
7. Avoidance of change.
It does appear to involve a genetic change, since archaic populations in
Europe were replaced, despite attempts to acquire the new technologies
brought by the moderns. There is evidence that archaic and modern
populations did _not_ interbreed in Europe.
So what is the difference between symbolic and non-symbolic thought? I
doubt it is consciousness versus lack of consciousness. I also doubt it
was the evolution of an ability to model other people in social
situations. While that ability is associated with the frontal cortex,
damage to it results in autism, not non-symbolic thought. Hence there were
chaotic dynamics available to the cortex prior to the evolution of
symbolic thought. (These chaotic dynamics underlie social interaction, as
well as the pattern-matching mechanisms used by the cortex to evaluate
situations.) Also, symbolic thought is seen in bower birds, and they are
_not_ very bright.
It appears to me that symbolic thought is indefinitely self-referential.
By this I mean that non-symbolic thought, while being potentially
extendable in time and space and potentially self-referential, is not so
in normal practice. Instead, a non-symbolic stream of thought peters out
fairly quickly. There is no mechanism used to organize the stream for
retrieval other than direct rehearsal. Symbolic thought appears to support
a tree structure so that primary components in the stream can consist of
multiple patterns and so on indefinitely. Additionally, generalization
implies that these trees can be compared and common features identified.
This allows teaching by paradigm.
Note that this connects to Paul Werbos' dynamic heuristic programming. The
brain function linked to symbolic thought short-circuits the potentially
infinite regression seen in dynamic programming models.
What, then, grounds the tree of self-reference? It may not ground. It may,
like economic systems, be self-sustaining. The fact that communications
appear to be the most important component of human communities is
Internet: erwin at trwacs.fp.trw.com
Working on Freeman nets....