Consilience is the unity of knowledge. I get a rush (feel awed, am inspired) when I see deep principles that tie together apparently disparate areas of knowledge. WeiQi and Exercise insights brought me awe this morning (OK, the morning tea helped too).
WeiQi is a very tough game, probably one the most difficult strategy games known. Players who reach high ranks in this game tend to be very bright. But it is not just brightness that takes you to the top, spending hours learning the specific implications of WeiQi’s rules is huge. Though challenging, this game can be very fun for a beginner who is playing an established player because (unlike Chess) it is played with handicaps that give players of very different skill levels a 50-50 chance of victory over each other.
Bob Hearn , an AI scholar, commenting in Mile Gu’s article on AI & WeiQi, tells how he took his 3-dimensional Go set to US Go Congress where he talked a 6-dan pro into playing his 3-dimensional game against a 6-kyu amateur. If this were regular 2-dimensional Go, the amateur should get a 12 stone handicap but they played without handicaps. The 3D rules were exactly the same as 2D WeiQi, only the particular matrix they played on changed (it was a diamond lattice). And guessed who won. Wow, you guessed it, the 6k won !
OK, this is only an n=1, but the lesson may be that a pro’s Go knowledge is very highly specialized, and it was useless in this [3D] game.
Reading this brought Consilience Awe when I saw the tie between WeiQi specialization and exercise specialization.
The Paleo blogs of late have debated skill training vs strength training. In Conditioning Research we learn that
The major outcomes [of research on skill vs strength training] suggests that the goal should be to gain strength in a safe, efficient, and effective way and then learn how to use that strength in a given sport.
Conclusion:
The Consilience Tie: General vs Specialized
In WeiQi circles: General intelligence is largely inherited and can be lost without use. Though WeiQi keeps the mind active and may preserve general intelligence, WeiQi intelligence is a specialized sort of intelligenct which does not necessarily build general intelligence [except perhaps for young minds]. You need to practice music, art, relationships and other areas of mind to keep a versitile mind. Likewise, just because you are bright, don’t be disappointed if you do not excel in WeiQi without putting time into it !
In Paleo circles: Paleo exercise is great for general strength and health, but if you want to be good at a particular sport, practice that sport !
Sabio's reply:
Fascinating ! Fun to hear confirmation in other fields.
Here is a link for Herb’s On-line stuff: at CMU ! Thank you !
Posts
The 3D Go anecdote surprises me none at all. Expertise is gained primarily through pattern recognition, since the variability of actual processing capacity isn’t nearly as great as the expert-novice gap.
I learned this from Herb Simon (the most brilliant professor I ever had btw), and even wrote an experiment back in grad school that supported his theory. He, as an AI pioneer, used chess in many experiments and examples. He noted, among other things, that chess masters could very quickly memorize the position of all the pieces on a board when a novice could not. However, if the pieces were positioned in such a way to be non-sensical to the rules and gameplay of chess, the expert and novice were equals again (reverting back to their 7 +/-2 STM capacity).
I tested it for random sequences of non-verbal sounds (sort of like a modified game of Simon; the pun wasn’t lost on me). I interviewed subjects and captured their level of musical training, and then tested them. I found statistically significant correlations between longer remembered sequences by the musicians than otherwise. However, when the sounds were not musical instruments, that expertise gap vanished.
Recognizing patterns allows people to chunk information together, effectively increasing their overall memory capacity. In effect, given nearly constant nominal memory capacity for all people, those who can put more than one item into a single available ’slot’ will produce a net gain in recall. These chunks must be easily defined such that the contents may be derived from a simpler definition, the latter of which is actually ‘remembered’.
If you haven’t had the pleasure, read some of Herb Simon’s many writings. They’re meaty, but fascinating.