Argument

Interdependency Argument

Intro

Think about a mousetrap. It has five parts: the base, the hammer, the spring, the catch, and the holding bar. Remove any one of them, and the trap does not catch fewer mice; it catches none at all. It needs all five to do anything.

Michael Behe noticed that living cells are full of systems like this, only much more complex. The bacterial flagellum, the little whip-tail motor that lets bacteria swim, has about 40 protein parts arranged like a precision engine. Take out one, the motor does not run slower; it does not run at all. The blood-clotting cascade has about a dozen factors that have to fire in sequence; missing one piece does not give you slow clotting, it gives you hemophilia. The immune system, the cellular cilium, even the basic machinery of life itself, all have this all-or-nothing shape.

At the deepest level, life needs three things at once: a way to copy its instructions (DNA replication), a way to make energy (metabolism), and a wall around it (membrane). Each one needs the other two to exist. Replication needs proteins, proteins need translation machinery, translation machinery needs replication. Metabolism needs enzymes, enzymes need replication. Membranes need proteins to stay alive, proteins need replication. You cannot get there one piece at a time; the first cell required all three already coordinated.

This is the Interdependency Argument. Step by step evolution by natural selection works only if each step gives the organism some survival advantage. But for these systems, no partial version works. There is nothing for natural selection to keep around because there is no functional half-built version to select. So either the parts all came together at once by extreme luck (astronomically unlikely), or they came together because something coordinated them. Coordinated outcomes, in every case we have ever traced, come from minds.

The page lays this out as debate prep, with the strongest counter-arguments treated honestly. The big objection is co-option: the parts could have evolved doing other jobs and then been repurposed later. The page handles that one carefully. It is not a knockout proof, but it sits cleanly inside the same kind of inference scientists use elsewhere: when you see coordinated machinery, look for a coordinator.

In full

Living systems are built from mutually dependent parts that cannot function in isolation. The bacterial flagellum's 40+ proteins, the blood-clotting cascade's coordinated factors, and, at the cellular-origin level, the joint requirement of replication, metabolism, and membrane all exhibit irreducible interdependency: removing any single component disables the whole. Random stepwise selection cannot easily produce systems where the partial precursors are non-functional and therefore non-selectable. The honest inference is to a coordinated origin, i.e., to design. This page is structured as debate prep, each premise carries a second-order positive case, anticipated objections, rebuttals, a live-cite kit, and tactical notes.

Argument structure

Form

Deductive in form, with an empirical premise (P1) about cellular and molecular interdependence and a probabilistic-modal premise (P2) about what stepwise processes can and cannot produce. Closely related to Michael Behe's irreducible complexity framework (Darwin's Black Box, 1996), now applied across scales, both at Behe's original molecular-machine level (bacterial flagellum, blood-clotting cascade, immune-system cascade) and at the OOL level (replication / metabolism / membrane mutual dependence). The argument is structurally an inference-to-best-explanation that exploits the unselectable-precursor problem.

P1, Some biological systems exhibit irreducible interdependency

Affirmative case (second-order arguments)

1. Behe's molecular-machine cases are well-established cell biology. The bacterial flagellum (~40 protein components organized as a rotary motor with stator, rotor, drive shaft, hook, and filament, non-functional in isolation); the blood-clotting cascade (~12 coordinated factors that must activate in correct sequence; partial cascades produce hemophilia, not partial clotting); the adaptive immune system's V(D)J recombination (multiple coordinated components, RAG enzymes, scaffolding, signal sequences); the eukaryotic cilium (~250 protein parts in a coordinated structure). Each is documented in standard cell-biology textbooks; the interdependence is uncontested.
2. The OOL-level interdependency is even more striking. Life's three core functions, genetic replication (DNA / RNA), metabolism (energy capture and use), and membrane (selective enclosure), are mutually interdependent. Replication needs proteins (which need translation, which needs ribosomes, which need replication); metabolism needs enzymes (which need replication); membrane needs membrane-protein machinery to maintain gradients (which needs translation). The 2024 LUCA reconstruction (Moody et al., Nature Ecology & Evolution) confirms all three systems were already integrated at the earliest reconstructed life ~4.2 Bya.
3. Mousetrap-style structural analysis. Behe's mousetrap analogy, a five-part mousetrap (base, hammer, spring, catch, holding-bar) cannot work with any subset; remove any part and the mousetrap doesn't catch fewer mice, it catches no mice. Biological irreducibly-complex systems exhibit the same all-or-nothing structure.
4. Mike Gene's "rugged-fitness-landscape" reframing strengthens the case: even granting evolutionary smoothing, irreducible interdependency creates fitness valleys between functional configurations that natural selection cannot cross.

Anticipated objections

1. "Co-option / exaptation: components originally serving other functions are repurposed into the irreducibly-complex system." The classic Kenneth Miller / Nicholas Matzke move on the bacterial flagellum (Type-III secretion system as ancestor).
2. "Scaffolding: temporary intermediate structures hold the assembly until the final form is in place."
3. "The 'irreducibly complex' systems are not actually irreducibly complex, they have functional substructures." Knockout experiments showing partial function.
4. "The flagellum / blood-clotting examples have been refuted." The post-Dover claim.

Rebuttals

1. Co-option is itself an unsolved evolutionary puzzle, not an answer. The Type-III secretion system (TTSS) hypothesis for flagellar origin requires that ~10 of the flagellum's ~40 components were ancestrally TTSS, and the remaining ~30 were assembled separately. Even granting the TTSS ancestry, the integration of TTSS components with the new components in functional configuration is the very problem the argument identifies. Co-option names a process; it does not demonstrate the stepwise pathway with selectable intermediates. Failure mode: labeling the puzzle is not solving the puzzle. (See Stephen Meyer's response in Signature in the Cell, ch. 17.)
2. Scaffolding accounts work for modifications of already-living systems, they presuppose life. They do not address the OOL-stage interdependency (replication / metabolism / membrane), where there is nothing yet alive to scaffold from. At the molecular-machine level, scaffolding is empirically claimed but rarely demonstrated; the burden is on the proposer to show the actual scaffold and the actual transitions. Failure mode: just-so storytelling without empirical demonstration.
3. The "knockouts show partial function" argument concedes Behe's point. Behe's claim is that the original system is irreducibly complex; that disabled cells with knock-outs sometimes survive on alternate pathways does not show the original system's components arose stepwise, it shows redundancy in the modern system. The historical-origin question is independent of current knockability. Failure mode: conflating extant redundancy with evolutionary stepwise origin.
4. The "Behe was refuted at Dover" claim is overstated. Judge Jones's 2005 ruling addressed the constitutional question of whether ID-as-curriculum violated the Establishment Clause; it did not adjudicate the scientific dispute. Behe's specific arguments have been engaged in the literature, with continuing technical debate (see Behe's The Edge of Evolution, 2007, and Lehigh University's continuing publications). The post-Dover narrative is sociology, not science. Failure mode: conflating legal rulings with scientific refutation.

Live-cite kit

• Scripture: Genesis 1.1 (coordinated creative acts); Genesis 1:11-25 (life "after its kind"); Colossians 1.16-17 ("in Him all things hold together"); Psalm 139:13-16 (intricate fashioning); Job 38-41
• Scholarly: Michael Behe (Darwin's Black Box, 1996; The Edge of Evolution, 2007); Stephen Meyer (Signature in the Cell, 2009, ch. 17; Darwin's Doubt, 2013); William Dembski (No Free Lunch, 2002); Jonathan Wells (Icons of Evolution, 2000); Scott Minnich (Lehigh; flagellum work); Moody et al. ("The nature of the LUCA", Nature Ecology & Evolution, 2024), for the OOL-level interdependency
• Aphorism: "A mousetrap with four parts catches no mice, and a half-flagellum doesn't half-spin."

Tactical notes

• Lead with the bacterial flagellum, it is the single most rhetorically vivid case, and most opponents have heard of it (one way or another).
• For OOL audiences: pivot to replication / metabolism / membrane. The 2024 LUCA paper is a fresh data point that catches mainstream-science opponents off-guard.
• Don't defend Behe's specific mathematical formalism in The Edge of Evolution if the opponent is technically sophisticated, the edge (two coordinated mutations as the practical limit) is contested; the interdependence of the systems themselves is much harder to deflect.
• Be ready for the "Behe was refuted" trope, name it as legal-conflation, redirect to the actual molecular-biological literature.
• Force-commit: "Name the published, demonstrated stepwise pathway with selectable intermediates for the bacterial flagellum." There is none, the opponent will retreat to "co-option" without specifics.

P2, Random stepwise processes cannot easily produce such systems

Affirmative case (second-order arguments)

1. Joint probability collapses multiplicatively. If each of the required components is independently improbable under chance, the joint probability of all of them arising together (in spatial proximity, in functional configuration) drops multiplicatively. For 40 components each with even 10^-3 individual probability (extraordinarily generous), joint probability is 10^-120, beyond Dembski's universal probability bound.
2. Natural selection requires functional intermediates that are selectable. Random stepwise assembly works only when partial products are themselves viable, but a partial pre-cellular system missing replication, or missing metabolism, or missing membrane is not living and is therefore not subject to selection. There is no smooth gradient of "partial life" up which selection could climb. Selection is impotent for the OOL.
3. For molecular machines: fitness valleys. Even within already-living systems, irreducibly-complex machines require traversing fitness valleys (where partial assemblies provide no advantage or net cost). Behe's Edge of Evolution (2007) argues two coordinated mutations is the practical limit for natural-selection traversal in vertebrate timescales; the malaria-parasite chloroquine-resistance data (~10^20 parasite-generations to get two specific mutations) anchors the empirical ceiling.
4. The "scaffolding then removal" requirement is itself complex. If a temporary scaffold held an assembly and was then removed, the scaffold is itself a coordinated structure requiring its own explanation. Each scaffolding move adds complexity rather than reducing it.

Anticipated objections

1. "Natural selection is a creative force, given enough time, it can climb any landscape."
2. "The probability calculations assume independent variation, but biological evolution is constrained and channeled."
3. "Computer-simulated evolution (Avida, etc.) shows irreducibly-complex systems can arise stepwise."

Rebuttals

1. Selection's creative power presupposes selectable intermediates. If the intermediates are non-functional, there is nothing to select for. The objection re-asserts what is in dispute. Selection is real, but it works on variation that is itself selectable; for irreducibly-interdependent systems, the stepwise variation is unselectable. Failure mode: petitio principii, assuming what must be proven (that selectable intermediates exist).
2. "Constrained variation" actually strengthens the design inference. If variation is non-random in the direction of functional configurations, that channeling is itself information-bearing, i.e., the channels themselves require explanation. Stuart Kauffman's hoped-for self-organization principles (Origins of Order, 1993; At Home in the Universe, 1995) have not delivered the goods; the channels remain undemonstrated. Failure mode: shifting the explanandum to a constraining process that is itself unexplained.
3. Computer simulations smuggle in design. Avida-style simulations build in fitness functions, replication mechanisms, mutation rates, and selection criteria, i.e., they presuppose the very informational structure (replication + variation + selection) that OOL needs to explain. The simulations show what already-functioning evolutionary algorithms can do; they do not demonstrate that such algorithms arose without intelligent setup. Dembski-Marks's "active information" critique applies. Failure mode: embedded design hidden in initial conditions.

Live-cite kit

• Scholarly: Behe (The Edge of Evolution, 2007, malaria data); Dembski (The Design Inference, 1998; universal probability bound); Dembski & Marks (The Search for a Search, 2010); Doug Axe (Undeniable, 2016, protein-folding combinatorics); Eugene Koonin (The Logic of Chance, 2011)
• Aphorism: "Selection works on what's already alive. The first cell can't be selected for, there's nothing yet alive to do the selecting."

Tactical notes

• The "selection presupposes life" point is the strongest move at OOL level. Most opponents reflexively appeal to natural selection without realizing it begs the question for OOL.
• Behe's malaria data (chloroquine resistance requiring ~10^20 parasite-generations to get two coordinated mutations) is empirically grounded and difficult to deflect. Use it as the concrete anchor for the "edge of evolution" claim.
• Don't get drawn into specific simulation debates, Avida defenders have years of practice. Stay at the meta-level: simulations show what designed algorithms can do, not what blind chemistry can do.

P3, Coordinated origins fit irreducible interdependency

Affirmative case (second-order arguments)

1. The inference structure is standard inference-to-best-explanation. Where stepwise random-plus-selection fails (P2), and the system requires multiple coordinated components (P1), the natural inference is to a process capable of coordination, i.e., agency. This is the same inference structure used when archaeologists infer human intent from coordinated artifact-arrangements.
2. Coordinated origins are positively known to produce irreducibly-interdependent systems. Engineers design machines with mutually-dependent parts every day, automobile engines, computer processors, software systems with interdependent modules. The known cause-type for irreducibly-interdependent systems is intelligence. Applying this empirical generalization to biology is consistent epistemic practice.
3. The alternative, coordinated origins by undirected processes, is undemonstrated and prima facie implausible. "Self-organization" has been hoped for by Kauffman et al. for decades; the hoped-for principles have not materialized in producing coded informational systems with irreducibly-interdependent components.

Anticipated objections

1. "Designer-of-the-designer regress: if life requires a coordinated designer, who coordinated the designer?"
2. "This is just Paley's watch updated, Darwin refuted Paley."
3. "Why infer God? Aliens-as-designers (panspermia) is equally consistent with the data."

Rebuttals

1. The inference is to a necessary, eternal, non-physical designer (God), not to a contingent designer. The contingent-things-need-explanation principle does not apply to a necessary being. The Dawkins regress objection assumes the designer must be of the same metaphysical type as designed contingent things, but classical theism explicitly denies this. (Aseity, Contingency Argument.) Failure mode: category mistake, applying the explanatory rule for contingent beings to the necessary being.
2. Darwin's refutation of Paley targeted biological design at the species level, not OOL. The OOL question precedes natural selection; selection presupposes a self-replicating population, which presupposes the very informational machinery in question. Darwin himself acknowledged that the origin of life was beyond his theory. The Information / Interdependency arguments are post-Darwin updates that target a different explanatory question. Failure mode: misidentifying the target, assuming OOL falls under Darwinian selection.
3. Panspermia just defers the question. If aliens designed life on Earth, who designed the aliens? Panspermia (Crick, Hoyle-Wickramasinghe) is sometimes invoked precisely because the OOL problem is so severe, but panspermia does not solve the problem, it pushes it. Eventually the regress terminates either in a necessary being (theism) or in a brute fact (which violates PSR). Failure mode: deferring rather than answering.

Live-cite kit

• Scholarly: Stephen Meyer (Signature in the Cell, 2009; Return of the God Hypothesis, 2021); William Dembski (The Design Inference, 1998); Behe (Darwin's Black Box, 1996); William Paley (Natural Theology, 1802, for the watchmaker analogy in updated form)
• Aphorism: "Engineers know how to design irreducibly-complex systems. Chemistry doesn't."

Tactical notes

• The Dawkins regress is the most-common deflection. Have the necessary-being response ready; name it as a category mistake.
• Be willing to concede the inference is to design without immediately claiming it is to God. The Christian-God conclusion comes from the cumulative case, not from this argument alone.
• Don't waste time on panspermia, most opponents do not seriously hold it; they raise it as a deflection. Note the deferral and redirect.

Conclusion

Irreducibly interdependent biological systems are best explained as the product of a designing intelligence capable of organizing the multiple components together. The bacterial flagellum, the blood-clotting cascade, the OOL-level replication / metabolism / membrane integration, each exhibits structure that random stepwise selection cannot produce because the partial precursors are non-functional and unselectable. The known cause-type for irreducibly-interdependent systems is intelligence. The honest inference is to design.

Master objections to the argument as a whole

1. "Irreducible complexity has been refuted in the scientific literature.", Reply: contested, not refuted. The literature has technical debates, but Behe's continuing publications (and the 2024 LUCA paper at OOL level) keep the argument live. The "refuted" claim is sociological framing.
2. "This is God-of-the-gaps reasoning.", Reply: the inference is from positive evidence (intelligence is the known cause-type for irreducibly-interdependent systems), not from ignorance. Same structure as archaeology / SETI / forensics.
3. "You haven't shown the designer is the Christian God.", Reply: conceded; this is part of a cumulative case. (See Christian God is the Only True God.)
4. "ID is not science by definition (methodological naturalism).", Reply: this is gatekeeping by definitional fiat, not an epistemic argument. The inference structure is identical to accepted scientific inferences in archaeology, SETI, forensics. (See Methodological Naturalism.)

Tactical opening / closing

Opening line: "Imagine a mousetrap with four of its five parts. It doesn't catch fewer mice, it catches no mice. The bacterial flagellum has 40 parts, all required. Let me explain why blind chemistry has no plausible pathway to such systems, and why the honest inference is to design."

Closing landing strip: "The Interdependency Argument doesn't claim we've proven design beyond all doubt. It claims that random stepwise processes can't produce systems where the intermediates are non-functional and therefore non-selectable, and that intelligence is the only cause-type we've ever observed produce irreducibly-interdependent systems. That's a positive inference, not a gap-fill."

Connection to Scripture

The interdependency argument resonates with the Genesis 1 pattern of coordinated creative acts. God does not produce ecology piecemeal but in coordinated days (light, separation of waters, dry land, vegetation, luminaries, sea creatures, land creatures, humans), each act presupposing prior acts and integrating into the whole. The cellular case scales this down: the first cell is described in the design-inference literature as analogously requiring a coordinated rather than piecemeal origin.

• Genesis 1.1, coordinated creation
• Genesis 1:11-25, life "after its kind"; coordinated emergence
• Colossians 1.16-17, "in Him all things hold together", cosmic-stability anchor
• Psalm 139:13-16, "fearfully and wonderfully made"
• Job 12:7-10, "in His hand is the life of every living thing"
• Acts 17:25, God "gives to all life and breath and all things"
• John 1:3, "all things came into being through Him"

Patristic / scholarly note

Classical / patristic:

• Basil the Great (Hexaemeron, c. 378), Genesis-account engagement with Greek natural philosophy
• Augustine (De Genesi ad Litteram), rationes seminales (seminal reasons): God created the seeds-of-things to unfold over time
• William Paley (Natural Theology, 1802), the watchmaker analogy; argument from biological contrivance

Modern intelligent-design movement:

• Michael Behe (Darwin's Black Box, 1996; The Edge of Evolution, 2007; Darwin Devolves, 2019), irreducible complexity; foundational cases
• William Dembski (The Design Inference, 1998; No Free Lunch, 2002), formal framework
• Stephen Meyer (Signature in the Cell, 2009, esp. ch. 17 on flagellum response; Darwin's Doubt, 2013)
• Scott Minnich (Lehigh University), bacterial-flagellum experimental work
• Douglas Axe (Undeniable, 2016), protein-folding combinatorics
• Jonathan Wells (Icons of Evolution, 2000)

Critics (for steelmanning):

• Kenneth Miller (Finding Darwin's God, 1999), TTSS-as-flagellar-ancestor argument
• Nicholas Matzke, flagellum stepwise pathway proposals
• Russell Doolittle (blood-clotting cascade evolution)
• Robert Pennock (Tower of Babel, 1999), methodological-naturalism critique of ID

OOL-level interdependency (newer literature):

• Moody et al. ("The nature of the LUCA", Nature Ecology & Evolution, 2024), confirms integrated complexity at earliest reconstructed life
• Eugene Koonin (The Logic of Chance, 2011), major OOL researcher
• James Tour (Rice University), public lectures on the OOL gap
• Stuart Kauffman (Origins of Order, 1993; At Home in the Universe, 1995), self-organization hopes (mostly unfulfilled)

See also

• Biogenesis Argument, sibling syllogism on Pasteur's empirical landmark
• Information Argument, sibling syllogism on coded information
• Argument from Origin of Life, broader OOL deployment
• Information Argument for Design, apologetic-grade synthesis hub
• Specified Complexity, formal framework
• Fine-Tuning Argument, sister design argument at cosmic scale
• Argument from Mathematical Truth, sister: math's "unreasonable effectiveness"
• Intelligent Design, broader ID program
• Irreducible Complexity, concept hub
• Abiogenesis, what the argument targets
• RNA World, immediate scientific scenario the argument engages
• LUCA, empirical evidence of integrated complexity
• Naturalism, the worldview the argument argues against
• Methodological Naturalism, the gatekeeping move ID confronts
• Genesis 1.1, biblical anchor
• Christian God is the Only True God, cumulative-case home
• Arguments, master index