The question is not whether design exists in nature. The question is whether we are allowed to say so. For fifty years, the rule has been: observe the pattern, measure the complexity, document the specification — and then conclude anything except the obvious. That rule is not scientific. It is philosophical. And it is breaking down.
The Explanatory Filter — William Dembski's formalisation of how we eliminate
chance and necessity before inferring design.
Every detection of design follows this path.
From SETI to forensics to archaeology — we use this logic already.
I. How We Detect Design — Already
When archaeologists find an arrowhead, they do not say: "Wind and erosion could have produced this — we cannot conclude design." They observe specific complexity — a shape that fits a purpose, with a pattern that cannot be explained by natural forces alone — and they infer a maker.
When SETI scientists scan radio signals from space, they have explicit criteria for what would count as a signal from an intelligent source. Not complexity alone — a crystal is complex. Not rarity alone — snowflakes are rare. They look for specified complexity: a pattern that is both improbable and matches an independently existing template. Like the prime numbers. Or a language.
DNA is a language. It has a four-symbol alphabet. It encodes information that is read, copied, edited, and error-corrected by molecular machines. The sequence is not determined by the chemistry — just as the letters on this page are not determined by the ink. Sequence specificity is the hallmark of information. Information is the hallmark of mind.
The pattern did not have to be this way. It could have been otherwise. It is not the inevitable product of natural law. DNA sequences are contingent — they are not chemically determined. The same nucleotides could spell a different message.
The probability of the pattern arising by chance is astronomically small. Not just unlikely — effectively zero. A functional protein requires specific amino acid sequences from a search space of 1077 possibilities. Eigen's Paradox: you need the code to build the machinery that reads the code.
The pattern matches an independent functional template. It is not just complex — it is complex in a way that works. A ribosome does not almost translate proteins. It translates them at 20 amino acids per second with an error rate of 1 in 10,000. The specification is precise.
II. The Butterfly Problem — Specified Complexity Made Visible
A butterfly wing assembled from random noise vs. from specified code.
The pattern that emerges from specification is not the product of chance.
It never is. You already knew that before you read this sentence.
Consider what it takes to produce a butterfly wing. Each wing is a surface covered in scales — each scale a modified cell, each cell reading the same DNA, expressing different genes depending on its precise position in the wing field. The colour pattern is not pigment poured on — it is a positional code. Cells at coordinates (x, y) receive a signal gradient and express a specific pigment gene. The result is a pattern that, in the case of the Heliconius genus, encodes species identity, mate recognition, and predator warning simultaneously.
That is not complexity. That is specified complexity. The same physical substrate — a surface of cells — produces meaningless noise if the code is scrambled, and a functional, beautiful, identity-encoding pattern when the code is intact. The difference between noise and signal is not chemistry. It is information.
III. The Thinkers Who Formalised It
Formalised the Explanatory Filter and the concept of Specified Complexity in The Design Inference (1998, Cambridge University Press). His framework does not invoke God — it describes the logical structure of how all rational agents infer design.
Author of Signature in the Cell (2009) and Darwin's Doubt (2013). Argued that the origin of biological information — not just complexity — is the central unsolved problem for materialist accounts of life's origin.
Developed Irreducible Complexity: systems where removing any component destroys function entirely. The bacterial flagellum requires all 40+ protein components to be present simultaneously. Natural selection cannot build it step-by-step because no intermediate confers a survival advantage.
Led academic atheism for 50 years. Converted to deism in 2004 after examining DNA complexity. Said explicitly: the argument from DNA's almost incredible complexity persuaded him intelligence must have been involved.
IV. The Standard Objections — and What They Actually Say
"Design requires a designer more complex than the designed object — who designed the designer?"
This objection proves too much. By the same logic, we could never infer that any human artifact was designed — because the human who made it is more complex than the artifact. We infer design from the properties of the effect, not from knowing the complete backstory of the cause. The Voyager probe was designed. We know this because of its specified complexity, not because we verified every detail about NASA.
"Design inference is not falsifiable — it is not science."
The design inference is falsifiable. It predicts that functional biological systems will exhibit specified complexity — and that removing components will destroy function rather than producing simpler functional intermediates. Behe's critics have, for thirty years, failed to demonstrate a step-by-step natural pathway for the bacterial flagellum. The prediction has held. Beyond that: SETI uses identical logic and no one calls it unscientific. Forensics uses identical logic. Archaeology uses it. The objection is not about method — it is about conclusion.
"Evolution explains biological complexity without design."
Evolution explains the distribution of biological forms across time. It does not explain the origin of the information required to build the first self-replicating system. The RNA World hypothesis — the leading materialist candidate — requires a functional RNA molecule to self-assemble before any replication machinery exists to preserve it. That is an information-before-replication problem that natural selection cannot address, because natural selection requires replication to operate. Meyer calls this the chicken-and-egg problem. It remains unsolved at the level of empirical chemistry.
V. Irreducible Complexity — The Mousetrap Made Real
The bacterial flagellum: a rotary motor with ~40 distinct protein components.
Remove any single component — the motor stops entirely.
There is no functional simpler version. Natural selection has nothing to select.
Behe's mousetrap is the cleanest illustration of the logic: a mousetrap requires a base, a hammer, a spring, a catch, and a holding bar — simultaneously. A four-component mousetrap catches no mice. A three-component mousetrap catches no mice. Only the complete, specified system functions.
The bacterial flagellum is a rotary motor. It has a base plate, a bushing, a hook, a filament, and a drive system — approximately 40 distinct protein types, each encoded by a different gene. It rotates at up to 100,000 RPM. It can reverse direction in a quarter turn. It is more sophisticated than any motor manufactured by human engineering at any scale.
The Type III Secretion System — the standard evolutionary precursor proposed — shares 10 of the 40 components. The other 30 have no known precursor. The injection system does not rotate, does not propel, and does not replicate the flagellum's function. Behe's critics have produced no step-by-step pathway that generates 30 novel protein types with simultaneous functionality. They have produced papers arguing it is theoretically possible. Theory and mechanism are not the same thing.
We already know how to recognise design.
We use that knowledge every day, in every court, in every dig site,
in every SETI listening station.
The same logic, applied to biology, leads to the same conclusion.
The only reason not to draw it is prior commitment to a different answer.
That is not science. That is philosophy.
And philosophy, at least, should be honest about its assumptions.