Author: Wesley R. Elsberry
Email:wre@cx33978-a.dt1.sdca.home.com
Date:1998/08/22
Forums: talk.origins
Message-ID: <199808221355.GAA02792@cx33978-a.dt1.sdca.home.com>
Adamski pointed out a web site featuring an article by William
Dembski. Dembski is one of the foremost members of the new
face of creationism, those who conjecture that Intelligent
Design is an inescapable conclusion from the evidence of
biological diversity.
Dembski should get credit for his attempt to place Intelligent
Design on an objective footing. Dembski knows that this is a
prerequisite before the scientific community is going to give
the concept any credence whatever.
[Quote]
The key step in formulating Intelligent Design as a scientific
theory is to delineate a method for detecting design. Such a
method exists, and in fact, we use it implicitly all the
time. The method takes the form of a three-stage Explanatory
Filter. Given something we think might be designed, we refer
it to the filter. If it successfully passes all three stages
of the filter, then we are warranted asserting it is
designed. 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?
[End Quote - W Dembski,
http://www.origins.org/real/ri9602/dembski.html]
The ordering that Dembski gives is important to the action of
his filter, and introduces error into the classifications made
by it. The limited selection of alternatives also negatively
influences the accuracy of Dembski's filter.
Dembski discusses various responses to small probabilities, and
concludes his essay with the following sentence.
[Quote]
From the perspective of the Explanatory Filter, all of these
moves have but one purpose: to block the conclusion that the
proper mode of explanation for life is design.
[End Quote - W Dembski,
http://www.origins.org/real/ri9602/dembski.html]
This assumes that the Explanatory Filter produced by Dembski
is the One True Explanatory Filter. Other filters are not
only possible, but preferable. Let me illustrate by
constructing a different Explanatory Filter and making a
case for its superiority.
Given a phenomenon to be explained, pass it through the
following Explanatory Filter:
1) If the phenomenon can be ascribed to random processes,
no further explanation is necessary. (The first condition
represents a null hypothesis. The random process hypothesis
serves admirably as a null hypothesis, and is to be preferred
as a null hypothesis to either law-like physical process or
intelligent causation.)
2) If the phenomenon can be traced to the action of known laws
regarding physical processes, no further explanation is
necessary. (Law-like physical processes posit fewer entities
than intelligent causation, and should be considered prior to
intelligent causation.)
3) If the phenomenon falls into the class of phenomena for
which we have past experience of having been produced by
the action of an intelligent agency, it is reasonable to
ascribe intelligent agency as a cause, and no further
explanation is necessary.
4) If no known random, law-like physical, or intelligent
causes can be identified, then the phenomenon can only be said
to be the result of a currently unknown process.
Note that my Explanatory Filter introduces a fourth
classification not present in Dembski's filter, that of a
phenomenon having an unknown cause.
In terms of a high-level programming language, Dembski's
Explanatory Filter is equivalent to
CASE phenomenon OF
physical_law_causation : cause := physical_law;
random_causation : cause := random;
specified_pattern : cause := intelligent_design;
ELSE : cause := random;
END;
whereas my Explanatory Filter would be equivalent to:
CASE phenomenon OF
random_causation : cause := random;
physical_law_causation : cause := physical_law;
intelligent_agent_causation : cause := intelligent_design;
ELSE : cause := unknown;
END;
The order in which alternatives are considered is very
important. Because Dembski utilizes a binary choice between
random causation and specified pattern at the end of his
filter, Dembski's filter is prone to false positive
classifications. This follows from the fact that Dembski
employs a "closed-world" assumption, where all possible
sources of causation are pigeonholed into random, law-like, or
intelligent agent categories. Because our knowledge of causes
is *incomplete*, there are actually two categories of law-like
physical processes: those we know about, and those that we do
not yet know. The ones that are unknown at the time of
classification imply the existence of a class of phenomena
which would *not* be filtered out at Dembski's first step, and
which could then be mistaken for specified non-random
phenomena, which are then erroneously given the status of
being intelligently caused.
My Explanatory Filter avoids the closed-world pitfall and the
false dichotomy pitfall that are apparent in Dembski's filter.
Because the null hypothesis of random process has to be rejected
before moving on to other alternatives, I avoid classifying
as random those phenomena which have characteristics of law-like
processes at work for which no known law-like process can be
identified.
Dembski provides examples of how we infer design in our
day-to-day activities as a prelude to his argument that we
can reliably infer design for which we do not have prior
experience.
[Quote]
The Explanatory Filter faithfully represents our ordinary
practice of sorting through things we alternately attribute to
law, chance, or design. In particular, the filter describes
1.how copyright and patent offices identify theft of intellectual property
2.how insurance companies prevent themselves from getting ripped off
3.how detectives employ circumstantial evidence to incriminate a guilty party
4.how forensic scientists are able reliably to place individuals at the scene of a crime
5.how skeptics debunk the claims of parapsychologists
6.how scientists identify cases of data falsification
7.how NASA's SETI program seeks to identify the presence of extra- terrestrial life, and
8.how statisticians and computer scientists distinguish random from non-random strings of digits.
Entire industries would be dead in the water without the
Explanatory Filter. Much is riding on it. Using the filter,
our courts have sent people to the electric chair. Let us now
see why the filter works.
[End Quote - W Dembski,
http://www.origins.org/real/ri9602/dembski.html]
In each case given above, design is able to be inferred
because we utilize our past experience with designed phenomena
to make that determination. All the examples given by Dembski
above will be identified as fitting the intelligent design
category in either his Explanatory Filter or the one which I
have provided. As I will show below, the assertion that
Dembski's Explanatory Filter reliably attributes design in the
*absence* of prior experience is false. My Explanatory Filter
works to cover the examples that Dembski cites without the
errors that use of Dembski's Explanatory Filter leads to. I
suggest that my version of an Explanatory Filter is closer to
what is actually employed in day-to-day identification of
causation.
Dembski discusses false negatives and false positives for
design due to operation of his Explanatory Filter. False
negatives are set aside as not harming the intent for which he
designed his filter, which is to detect design. Similarly, my
Explanatory Filter will assign designed phenomena that are
deceptive to one of the first two categories. Additionally,
my filter adds another class of false negative, which is where
design exists but cannot be recognized upon our basis of past
experience with similar designer/design relationships.
However, this is a harmless addition, since attribution to a
designer when we have no prior basis for that assumption is
erroneous. There will not be misclassification of phenomena
in that class as being due to law-like physical process since
all phenomena known to have been produced by law-like physical
process and no others are accepted at step 2 of my Explanatory
Filter. Unknown design and unknown law-like physical process
share classification as unknown and worthy of further inquiry.
I reject Dembski's assertion that his Explanatory Filter is
not subject to false positives. I have noted above and will
demonstrate below the existence of at least one class of false
positive classification to which Dembski's filter is
susceptible.
[Quote]
And this brings us to the problem of false positives. Even
though the Explanatory Filter is not a reliable criterion for
eliminating design, it is, I argue, a reliable criterion for
detecting design. The Explanatory Filter is a net. Things that
are designed will occasionally slip past the net. We would
prefer that the net catch more than it does, omitting nothing
due to design. But given the ability of design to mimic
unintelligent causes and the possibility of our own ignorance
passing over things that are designed, this problem cannot be
fixed. Nevertheless, we want to be very sure that whatever the
net does catch includes only what we intend it to catch, to
wit, things that are designed.
I argue that the explantory filter is a reliable criterion for
detecting design. Alternatively, I argue that the Explanatory
Filter successfully avoids false positives. Thus whenever the
Explanatory Filter attributes design, it does so correctly.
Let us now see why this is the case. I offer two
arguments. The first is a straightforward inductive argument:
in every instance where the Explanatory Filter attributes
design, and where the underlying causal story is known, it
turns out design actually is present; therefore, design
actually is present whenever the Explanatory Filter attributes
design.
My second argument for showing that the Explanatory Filter is
a reliable criterion for detecting design may now be
summarized as follows: the Explanatory Filter is a reliable
criterion for detecting design because it coincides with how
we recognize intelligent causation generally. In general, to
recognize intelligent causation we must observe a choice among
competing possibilities, note which possibilities were not
chosen, and then be able to specify the possibility that was
chosen.
[End Quote - W Dembski,
http://www.origins.org/real/ri9602/dembski.html]
It is a contradiction that Dembski's Explanatory Filter has
the property of not always accurately eliminating design and
also has the property of making no false positive
identifications of design. If the filter cannot accurately
eliminate all cases that have the appearance of design when
design is not present, then we necessarily have cases where
design can be inferred where that conclusion is not warranted.
Dembski's first argument quoted above is correct; when we have
knowledge of, and thus past experience of, a causal
relationship between a designer and a designed phenomenon, we
can reliably infer that subsequent recognition of those
phenomena imply the action of a designer. It is important to
realize that we utilize our experience of designed products
and the design process leading to those products in order to
specify and recognize critical attributes of the designer.
Those attributes are what we seek to identify when attempting
to utilize an explanation based upon causation by an
intelligent agency. This is nothing new, since David Hume
made the same point long ago in his rejection of the argument
from design.
Dembski's second argument is simply inadequate to the task of
leading to his conclusion. Our general capability to
recognize design is due exclusively to our past experience
with designed phenomena, where we know both the attributes of
the designer and the produced phenomena. We use the knowledge
of relationships due to design to identify examples of those
relationships. Using this capability to assert that we can
recognize previously unknown designed relationships is begging
the question. The extrapolation which Dembski assumes is
reliable is, in fact, unreliable. Dembski's Explanatory
Filter will permit false positives as follows: For phenomena
which are caused by law-like physical processes which are
currently unknown to humans, Dembski's filter may falsely
identify them as being due to the action of an intelligent
agency, since we will be able to specify a pattern in either
case. Phenomena in that class will pass Dembski's first
filter, because the law-like process which generates the
phenomena is not known and cannot be so classified. They will
pass the second filter, because the result of a law-like
process can demonstrate a pattern, and thus is distinguishable
from phenomena caused by random process. And they will pass
the third filter, because the pattern which a law-like
physical process can produce can be taken to form a
specification. In Dembski's filter, a false dichotomy at that
point exists which only allows classification as intelligent
design or random process. We need only examine history to
determine that application of Dembski's Explanatory Filter
would have falsely classified phenomena in the past, and by
inference that it will incorrectly classify phenomena in the
future. Let us take as an example the light emitted by
samples of certain ores before the discovery of radioactivity.
The composition of the ores comprised a specification. The
effect could not be attributed to law-like physical process,
since no process capable of producing the effect was known to
exist at the time. The effect was not random; light was
emitted from all ores of similar composition. By Dembski's
filter, this obviously was a case of an intelligent agent
causing the phenomenon, despite the fact that no prior
knowledge of a designer accomplishing such a phenomenon was
known. Under my Explanatory Filter, the phenomenon would have
been classed as due to an unknown process, and the
specification due to composition of materials in which the
phenomenon was observed could be used to direct further
inquiry to determine -- and comprehend -- a cause for the
phenomenon.
My Explanatory Filter retains all the power of the one which
Dembski provides, and fixes certain deficiencies present in
his filter, since phenomena for which we do not have an
understanding of the causative process are clearly identified
as such, where Dembski's filter can conflate phenomena which
obey physical laws that are as yet unknown to humans with the
action of an unspecified intelligent agency.
I will briefly examine Dembski's example, where an accusation
of ballot-fixing was leveled at Nicholas Caputo. In 40 of 41
ballots, Caputo fixed the Democratic candidate in the first
ballot position. Caputo was himself a Democrat. First ballot
position usually confers an advantage in voting results.
Dembski eliminates law-like physical process, then random
chance as an explanation for the results. The pattern that
leads us to conclude that Caputo cheated by design is that
favor to his own party was consistently shown. Dembski notes
that all who are exposed to this example infer design, that
Caputo cheated. But we have exposure to situations where
cheating of this sort is possible, and culturally are trained
to recognize this kind of behavior as being the result of
cheating. We conclude that Caputo cheated based upon our
prior experience with cheaters. We are able to identify a
motivation known by prior experience, a mechanism known by
prior experience, and impute those motivations to the actor
because of our prior experience with humans performing
political tasks. We do not find that Caputo cheated on the
basis of extrapolation; instead we look for and recognize the
attributes that we know in the past have been associated with
cheating in the political process.
The default case of my Explanatory Filter properly categorizes
phenomena as being subject to future comprehension with study
of the phenomena. The default case of random process in
Dembski's Explanatory Filter leads to the acceptance of false
positive identification of intelligent design because it
presents a false dichotomy in the filtering process. False
attribution of design will tend to inhibit the study of such a
phenomenon by making it appear to have an explanation when all
that has happened is that it has been *classified* as having
an explanation.
--Wesley R. Elsberry, Student in Wildlife & Fisheries Sciences, Tx A&M U.
Visit the Online Zoologists page (http://www.rtis.com/nat/user/elsberry)
Email to this account is dumped to /dev/null, whose Spam appetite is capacious.
I wonder what Mark Twain would say at a time like this?