Results for the 2023 RationalWiki Moderator Election have now been posted. Thank you for participating in this election, and congratulations to the winners!

# Explanatory Filter

The Explanatory Filter is a "protocol" devised by intelligent design creationist William Dembski which purports to detect whether or not any entity (object, living thing, etc.) was designed by an intelligent agency. If such a protocol did exist, it would be a valuable addition to the set of scientific tools; sadly, Dembski's Explanatory Filter does not live up to its press releases.

## Process

The Explanatory Filter operates by 'filtering out' everything which is not Design, and declaring whatever is left to be Design. As Dembski puts it:[1]

Roughly speaking the filter asks three questions and in the following order: (1) Does a law explain it? (2) Does chance explain it? (3) Does design explain it?

Dembski explains:[1]

At the first stage, the filter determines whether a law can explain the thing in question. Law thrives on replicability, yielding the same result whenever the same antecedent conditions are fulfilled. Clearly, if something can be explained by a law, it better not be attributed to design. Things explainable by a law are therefore eliminated at the first stage of the Explanatory Filter.

Suppose, however, that something we think might be designed cannot be explained by any law. We then proceed to the second stage of the filter. At this stage the filter determines whether the thing in question might not reasonably be expected to occur by chance. What we do is posit a probability distribution, and then find that our observations can reasonably be expected on the basis of that probability distribution. Accordingly, we are warranted attributing the thing in question to chance. And clearly, if something can be explained by reference to chance, it better not be attributed to design. Things explainable by chance are therefore eliminated at the second stage of the Explanatory Filter.

Suppose finally that no law is able to account for the thing in question, and that any plausible probability distribution that might account for it does not render it very likely. Indeed, suppose that any plausible probability distribution that might account for it renders it exceedingly unlikely. In this case we bypass the first two stages of the Explanatory Filter and arrive at the third and final stage.

Dembski notes that "exceeding improbability is by itself not enough to preclude something from happening by chance",[1] although other creationists disagree. To explain how, except via improbability, we can determine design, Dembski writes:

Invariably, what is needed to eliminate chance is that the event in question conform to a pattern. Not just any pattern will do, however. Some patterns can legitimately be employed to eliminate chance whereas others cannot.

A bit of terminology will prove helpful here. The "good" patterns will be called specifications. Specifications are the non-ad hoc patterns that can legitimately be used to eliminate chance and warrant a design inference. In contrast, the "bad" patterns may be called fabrications. Fabrications are the ad hoc patterns that cannot legitimately be used to eliminate chance.

Specifications would become Complex Specified Information in Dembski-speak.

## "Examples"

Dembski provides several examples through which design is supposedly found.

### Caputo

Dembski provides the example of Mochary v. Caputo (1985)[2] to illustrate these points:[1]

To see how the filter works in practice, consider the case of Nicholas Caputo. Back in 1985 Nicholas Caputo was brought before the New Jersey Supreme Court. The Republican party had filed suit against him, claiming Caputo had consistently rigged the ballot line in Essex County, New Jersey where he was county clerk. It is a known fact that first position on a ballot increases one's chances of winning an election. Since in every instance but one Caputo positioned the Democrats first on the ballot line, the Republicans argued that in selecting the order of ballots Caputo had intentionally favored his own Democratic party. In short, the Republicans claimed Caputo had cheated.

The question then before the New Jersey Supreme Court was, Did Caputo actually rig the order, or was it without malice and forethought on his part that the Democrats happened 40 out of 41 times to appear first on the ballot? Since Caputo denied wrongdoing, and since he conducted the drawing of ballots so that witnesses were unable to observe how he actually did draw the ballots, determining whether Caputo did in fact rig the order of ballots becomes a matter of evaluating the circumstantial evidence connected with this case. How then is this evidence to be evaluated?

In determining how to explain the remarkable coincidence of Nicholas Caputo selecting the Democrats 40 out of 41 times to head the ballot line, the court had three options to consider:

Law
Unbeknownst to Caputo, he was not employing a reliable random process to determine ballot order. Caputo was in the position of someone who thinks she [sic] is flipping a fair coin when in fact she [sic] is flipping a double-headed coin. Just as flipping a double-headed coin is going to yield a long string of heads, so Caputo, using his faulty method for ballot selection, generated a long string of Democrats coming out on top.

Chance
In selecting the order of political parties on the state ballot, Caputo employed a reliable random process that did not favor one political party over another. The fact that the Democrats came out on top 40 out of 41 times was simply a fluke. It occurred by chance.

Design
Caputo, knowing full well what he was doing and intending to aid his own political party, purposely rigged the ballot line selection process so that the Democrats would consistently come out on top. In short, Caputo cheated.

The first option-that Caputo chose poorly his procedure for selecting ballot lines, so that instead of genuinely randomizing the ballot order, it just kept putting the Democrats on top-was dismissed by the court because Caputo himself had claimed to use a randomization procedure in selecting ballot lines. And since there was no reason for the court to think that Caputo's randomization procedure was at fault, the key question therefore became whether Caputo actually put this procedure into practice when he made the ballot line selections, or whether he purposely circumvented this procedure in order for the Democrats consis- tently to come out on top. And since Caputo's actual drawing of the capsules was obscured to witnesses, it was this question that the court had to answer.

With the law explanation eliminated, the court next decided to dispense with the chance explanation. Having noted that the chances of picking the same political party 40 out of 41 times were less than 1 in 50 billion, the court concluded that "confronted with these odds, few persons of reason will accept the explanation of blind chance."

One of the examples is

how copyright and patent offices identify theft of intellectual property

However, only designed objects can be copyrighted or patented, so copyright and patent offices do not detect design, but non-design, that is, plagiarism. But for Dembski, plagiarism counts as "intelligent design".

### SETI

Another example is

how NASA's SETI program seeks to identify the presence of extra- terrestrial life

Except that it isn't, according to the SETI Institute.[3] As this article says, the SETI program does not actually search for anything "intelligent" (like a series of prime numbers, as Dembski, with inspiration from the movie "Contact", frequently claims), but for something as simple as a single sinusoidal wave that exhibits Doppler-shift - a possible indication that it might come from a planet rotating around itself or orbiting a star.

## Attempted formal definition

Given an event E, let H be a "non-design" hypothesis that can explain E, then

1. if P(E | H) is high, conclude regularity (law);
2. if P(E | H) is medium, conclude chance;
3. if P(E | H) is low, and there is a known simple specification for E, conclude design.

In the drop-out case, that is low probability and no known simple specification, the conclusion is chance.

Note that there is only mentioned one hypothesis (H) here, and Dembski usually only refers to one hypothesis, though occasionally he refers for "all relevant chance hypotheses", but then subsumes them into the one hypothesis H = {Hi}i.

## Problems

The problem with this is that we cannot filter out everything which is not Design!

### Argument from ignorance

See the main article on this topic: Argument from ignorance

Many "design inferences" are actually wrong, though they may be reasonable relative to the general knowledge of their own time. If used on an entity that is the result of processes not currently known, or of a currently-unknown concatenation of known processes, the Filter would tell us that that entity is Designed. As a practical matter, the Filter cannot actually be successfully applied by non-omniscients. Thus, it is an excellent example of the Sherlock Holmes Fallacy.

By way of demonstration, consider that historically, nearly everything has been considered the direct result of an intelligent agency before people understood it. Storms were caused by Thunder gods. How could water get up into the sky without divine help? How could it gather? What else could explain lightning? But it doesn't end there. How could the planets stay up without angels? What could explain gems or metal veins being embedded inside other rocks? What else but angry gods could cause plagues? The list is endless. And the list is, for every currently known phenomenon, a long line of errors. That is, Dembski's filter has been applied repeatedly throughout history and been wrong every single time. Historically, science has always begun when arguments from ignorance like Dembski's are rejected and stymied when they are proclaimed science.

Dembski, aware of these flaws, wrote:[1]

It needs to be stressed that this third and final stage does not automatically yield design-there is still some work to do. Vast improbability only purchases design if, in addition, the thing we are trying to explain is specified.

Given this, it seems premature to conclude anything based on the Filter, unless we accept that our ignorance is a valid reason to do anything. Rather than a design conclusion, we may be better off saying that we don't know yet - further research is needed. In a court case, new evidence can turn the tide.

### Lack of application

It is worth noting that in spite of IDists' claims that Dembski's Filter demonstrates that life was Designed, nobody has ever actually applied the Filter to any life-form, nor yet even to any identifiable aspect nor feature of any life form.

### One-pass?

In the Caputo case, there were two different chance hypotheses, one that Caputo was in good faith, but used an unreliable randomization process, and one that Caputo, as he claimed, used a reliable randomization process. To determine design, we would need to rerun the filter over and over again, until all chance hypotheses had been removed. Saying that they are part of the same chance hypothesis may be stretching the notion of a hypothesis a bit too much.

### Clarity of "design"

It's unclear what "design" really means. Dembski provides an example of electoral fraud, intellectual property fraud, and electronic transmissions as times in which the Explanatory Filter applied, but they certainly don't all use the same definition of "design" or "intelligent".