I've addressed elsewhere whether deterministic systems, such as Neoplatonism, are consistent with ID. However, in that article, the assumption is the "necessity" that is eliminated from our causal explanations is unqualified. I've been assuming so far that the CSI metric is meant to eliminate all forms of necessity, whether they be natural causes, aliens, deities, etc.
However, a reader made the astute point that Dr. Dembski does not actually characterize necessity in such a broad manner, and that I am interpreting his work in a more general manner than it may be intended. In general, when Dr. Dembski talks of eliminating necessity, he is referring to natural causes, such as Darwinian evolution.
So, let's examine how well the concept of CSI works if the term "necessity" is restricted to refer to only certain forms of necessity, i.e. natural causes, and other forms of necessity are still fair game. Specifically, let's examine what happens if intelligent agents are necessary causes that necessarily cause CSI. P(CSI | Intelligent agent) ~ 1, or in other words the probability of an intelligent agent creating CSI is very close to 1.
First, notice that such a qualification is not necessary for CSI to be an indication of intelligent activity. It may be that P(CSI | Intelligent agent) ~ 0, and intelligent agents (IA) only create CSI in very, very rare circumstances. If intelligent agents are the only beings capable of creating CSI, then the detection of CSI would still indicate the activity of an intelligent agent. Therefore, even if P(CSI | IA) ~ 0, it is still the case that CSI functions as a design detector, and P(CSI | IA) ~ 1 is not a necessary condition for design detection to work.
Second, consider the conditions that CSI will be used for design detection. The premise behind using CSI is we do not know whether an intelligent cause has been at work in our given scenario. Consequently, we do not know whether a particular cause under consideration is a natural or intelligent cause. Thus, we must take all the causes into account when calculating the CSI for a particular event.
Now, let's say the event does have CSI, and it was created by an intelligent cause. Furthermore, the intelligent cause has the condition such that P(CSI | IA) ~ 1. P(CSI | IA) ~ 1 means that the probability of a CSI event occurring is 1 / specification resources. Therefore, the probability of the event under question occurring is 1 / specification resources.
The CSI formula is -log2 (P(E) * I(E) * PR). P(E) is the probability of the event occurring. I(E) is the specification resources available for specifying the event. PR is the probabilistic resources. We know from previous considerations that P(E) = 1 / I(E). This makes the formula now look like this: -log2 (1 * PR) = -log2 (PR). Since PR >= 1, and -log2 of any positive integer is <= 0, then CSI will always be <= 0. Consequently, if P(CSI | IA) ~ 1 and an intelligent agent is responsible for the event, the CSI calculation will never register positive, and can never detect design. Therefore, the condition P(CSI | IA) ~ 1 renders CSI an ineffective metric for detecting design.
The result from these considerations is that claiming intelligent design is compatible with determinism does not bode well for being able to actually detect intelligent design. It is in the interest of intelligent design to rely on non-deterministic metaphysics in order to remain logically coherent. One such non-deterministic metaphysic is libertarian free will, which attempts to be both non-necessary and non-random. Such a metaphysic is quite compatible with Intelligent Design:
http://appliedintelligentdesign.blogspot.com/2012/08/how-id-sheds-light-on-classic-free-will.html
However, a reader made the astute point that Dr. Dembski does not actually characterize necessity in such a broad manner, and that I am interpreting his work in a more general manner than it may be intended. In general, when Dr. Dembski talks of eliminating necessity, he is referring to natural causes, such as Darwinian evolution.
So, let's examine how well the concept of CSI works if the term "necessity" is restricted to refer to only certain forms of necessity, i.e. natural causes, and other forms of necessity are still fair game. Specifically, let's examine what happens if intelligent agents are necessary causes that necessarily cause CSI. P(CSI | Intelligent agent) ~ 1, or in other words the probability of an intelligent agent creating CSI is very close to 1.
First, notice that such a qualification is not necessary for CSI to be an indication of intelligent activity. It may be that P(CSI | Intelligent agent) ~ 0, and intelligent agents (IA) only create CSI in very, very rare circumstances. If intelligent agents are the only beings capable of creating CSI, then the detection of CSI would still indicate the activity of an intelligent agent. Therefore, even if P(CSI | IA) ~ 0, it is still the case that CSI functions as a design detector, and P(CSI | IA) ~ 1 is not a necessary condition for design detection to work.
Second, consider the conditions that CSI will be used for design detection. The premise behind using CSI is we do not know whether an intelligent cause has been at work in our given scenario. Consequently, we do not know whether a particular cause under consideration is a natural or intelligent cause. Thus, we must take all the causes into account when calculating the CSI for a particular event.
Now, let's say the event does have CSI, and it was created by an intelligent cause. Furthermore, the intelligent cause has the condition such that P(CSI | IA) ~ 1. P(CSI | IA) ~ 1 means that the probability of a CSI event occurring is 1 / specification resources. Therefore, the probability of the event under question occurring is 1 / specification resources.
The CSI formula is -log2 (P(E) * I(E) * PR). P(E) is the probability of the event occurring. I(E) is the specification resources available for specifying the event. PR is the probabilistic resources. We know from previous considerations that P(E) = 1 / I(E). This makes the formula now look like this: -log2 (1 * PR) = -log2 (PR). Since PR >= 1, and -log2 of any positive integer is <= 0, then CSI will always be <= 0. Consequently, if P(CSI | IA) ~ 1 and an intelligent agent is responsible for the event, the CSI calculation will never register positive, and can never detect design. Therefore, the condition P(CSI | IA) ~ 1 renders CSI an ineffective metric for detecting design.
The result from these considerations is that claiming intelligent design is compatible with determinism does not bode well for being able to actually detect intelligent design. It is in the interest of intelligent design to rely on non-deterministic metaphysics in order to remain logically coherent. One such non-deterministic metaphysic is libertarian free will, which attempts to be both non-necessary and non-random. Such a metaphysic is quite compatible with Intelligent Design:
http://appliedintelligentdesign.blogspot.com/2012/08/how-id-sheds-light-on-classic-free-will.html
