|
Author
|
Topic: Calculation of specificational resources
|
Poul Willy Eriksen
Member
Member # 1976
|
posted 18. May 2006 09:31
Hello!
![[Confused]](confused.gif) I have become involved in a disccussion about how to calculate the specificational resources and would appreciate some help.
William Dembski writes in his Specification paper p.18-19 about this; but it's unclear how to generalize Dembski's calculations.
To be more specific, let's assume the event "Four Aces and Ten of Diamonds" in a poker game. The target pattern would be "Four Aces", since the Ten of Diamonds is of no particular interest. The semiotic cost of "Four Aces" is two. A "Royal Flush" has the same semiotic cost and therefore the same specificational resources. So far, so good. However, to calculate the specified complexity I cannot use relative specificational resources, I need the absolute amount. I estimate this to be slightly below 100 (number of poker hands specifiable with 1-2 words = individual cards + pair + triplets + quadruplets + named hands = 97 in my calculation).
Can anyone confirm my calculation or give details of how to calculate the specificational resources? Thanks in advance for any replies ![[Smile]](smile.gif)
cheers
- Poul Willy Eriksen
IP: Logged
|
|
2ndclass
Member
Member # 1979
|
posted 24. May 2006 15:55
Hi Poul. I don't think that Dembski intended the word-counting method to be taken as a reliable heuristic. It's premised on the idea that very specific descriptions entail a lot of words, but this is often not the case. In natural languages, some words are more information-rich than others. For example, the description "rabbit" is more specific than the description "furry hopping animal."
One way to avoid this problem is to use a non-natural language, like a universal turing machine. This gives us a non-arbitrary semiotic cost based on the Kolmogorov complexity of the event. Using this method, we can calculate the semiotic cost of various card hands in a way that's independent of our card-playing linguistic background.
IP: Logged
|
|
|
Printer-friendly view of this topic