Concept

Specified Complexity

Intro

How do you know, looking at something, whether it was designed or whether it happened by accident?

The mathematician William Dembski proposed a clean answer in The Design Inference (1998). A thing was designed when it has two features at once: it is highly improbable on its own, and it matches a pattern you could describe ahead of time independently.

Take poker. Any specific hand of five cards you deal is wildly improbable. There are over two and a half million possible five-card hands. But the particular hand you happened to deal yesterday is no more suspicious than any other; nothing told you in advance which to expect. Now deal a royal flush in spades, four times in a row. That is not just improbable; it matches a named, independently-specifiable pattern ("royal flush in spades"). The combination of improbability and pattern-match is what makes you reach for the deck and check for shuffling.

Dembski called the combination specified complexity. Complexity is the improbability part: the event has very low odds under chance alone. Specification is the pattern part: the event matches a target you could describe without referring to the event itself.

Apply this to biology. A protein is a chain of amino acids, usually a few hundred long, with 20 possible amino acids at each position. The number of possible sequences is astronomical (20 to the 300th power for a typical protein). Most of those random sequences fold into useless junk. The protein that catalyzes a specific biochemical reaction does so because it has a functional sequence matching a chemical task. The task is the independently-specifiable target. The sequence is enormously improbable. Both features are present. Specified complexity is present.

Dembski set a universal probability bound of 1 in 10 to the 150th, derived from the maximum number of physical events the observable universe could have produced over all of cosmic history. Anything more improbable than that is, on his account, beyond the reach of chance even if every atom in the universe were a separate roll of the dice.

The argument is controversial. Mainstream evolutionary biology contests that natural selection plus mutation can supply the relevant functional information over long timescales, so chance alone is not the only naturalistic option. The page below walks the formal framework, the application to biological systems (especially proteins and biological information), the criticism from mainstream biology, and the defenses and refinements offered by the Intelligent Design movement.

◎ See this page's graph and discover linked topics →

In full

A formal framework, developed primarily by William Dembski (The Design Inference, 1998; No Free Lunch, 2002), for distinguishing patterns that warrant a design inference from those that don't. The core idea: a pattern warrants the inference to design when it is both highly improbable under chance and matches an independently-specifiable functional pattern.

The two components

Complexity (high improbability)

A pattern is complex if its occurrence under the relevant chance hypothesis is highly improbable. Quantitatively, Dembski set a "universal probability bound" of 1 in 10^150, a threshold derived from the maximum number of physical events (atoms × seconds × Planck-time-resolved interactions) the observable universe can have produced over its history. Any pattern less probable than 10^-150 is, on Dembski's account, beyond the reach of chance even if every atom in the universe were a separate experimenter.

Specification (independent pattern-match)

A pattern is specified if it matches a target that can be described independently of the event itself. The classic illustration: a poker hand of 52 cards is enormously improbable on any specific deal, but no individual deal is specified, it doesn't match an independently-given pattern. By contrast, a deal that matches a named hand (e.g., "ace through ten of spades, in order") is specified, and combined with the improbability, warrants the suspicion that the deal was rigged rather than random.

A protein that folds and performs a particular catalytic function is specified in this sense: the description "catalyzes reaction R with selectivity S" is given independently of any particular sequence; sequences that satisfy the description are functionally specified.

When the two combine

When both complexity and specification are present, the chance hypothesis is overwhelmed and design becomes the inference to the best explanation. Dembski formalizes this as a one-tail probabilistic test ("rejecting" chance when complexity-times-specification crosses the universal bound).

Application to biology

The most-developed application is to functional protein sequences. A modest protein of 150 amino acids has 20^150 ≈ 10^195 possible sequences. Doug Axe's experimental work on protein-folding sequence-space (Undeniable, 2016; earlier Journal of Molecular Biology papers) suggests that the fraction of possible 150-aa sequences that fold into stable, functional proteins is on the order of 1 in 10^77 to 1 in 10^64, i.e., functional sequences are astronomically rare in protein-sequence space. Combined with the specification (the protein folds and performs a particular function), the combination crosses Dembski's universal bound.

The same framework extends to DNA sequences encoding functional proteins and to integrated cellular machinery (where multiple specified-complex parts must co-exist for the system to function).

Relation to the Information Argument

The Information Argument for Design uses specified complexity as its formal backbone. Where the information argument is the apologetic-grade inference ("life contains coded information; therefore mind"), specified complexity is the technical-grade framework ("here is the precise probabilistic-and-pattern-matching criterion that warrants the inference"). The two operate at different levels of formality but argue for the same conclusion.

Critical reception

Specified complexity has been heavily contested in the philosophy-of-biology and information-theory literature:

Mathematical / information-theoretic objections

• Wesley Elsberry & Jeffrey Shallit have argued that Dembski's "complex specified information" does not correspond to any standard information-theoretic measure (Shannon, Kolmogorov, algorithmic) and is not well-defined as a quantity.
• Olle Häggström and others have argued that the "no free lunch" arguments Dembski uses (in No Free Lunch) misapply optimization-theory results to biological evolution.

Philosophical / methodological objections

• Critics object that "specification" is too easy to confer post hoc, any pattern can be described after the fact in a way that makes it look specified.
• Defenders respond that genuine biological function is independently specified by the organism's survival requirements, not assigned post hoc.

From inside the design-inference camp

• Some ID-friendly thinkers (e.g., Michael Behe with irreducible complexity) prefer different formal frameworks; the design-inference movement is not monolithic.

Why the framework matters apologetically

Even if the formal Dembski apparatus is contested, the underlying intuition, that both improbability and specification are required to warrant the design inference, is sound. It blocks two common bad inferences:

1. From mere improbability to design. Any specific snowflake is astronomically improbable, but no snowflake is specified, so we don't infer design.
2. From mere specification to design. Any matching pattern after the fact looks specified, but if it's not improbable, it doesn't warrant the design inference.

The combination is what does the work. Biology, coded, functional, sequence-specific information in cells, meets both criteria, in the design-inference camp's view, and warrants the inference.

See also

• DNA, search-landing page; DNA is the headline exhibit of specified information
• Information Argument for Design, the apologetic-grade version
• Information Argument, the formal syllogism
• Abiogenesis, the origin-of-life problem the framework addresses
• RNA World, primary scientific scenario engaged
• Teleological Arguments, parent argument family
• Hubert Yockey, methodological forerunner (Shannon information applied to OOL)
• Fred Hoyle, pre-Dembski heuristic version (the "747" analogy)
• Naturalism, the worldview the framework is deployed against