Argument

Convergent Evolution as Design Signal Argument

Intro

A bat flies through a dark cave, snapping clicks off the walls and reading the echoes to navigate. A dolphin slips through black water and does the same. Same trick: echolocation. Sophisticated, biologically expensive, and (the textbooks say) evolved twice, independently, in two creatures whose closest common ancestor was a small land mammal that did neither.

A vertebrate's eye, with cornea, lens, retina, optic nerve, and image-forming optics, is a camera. An octopus's eye, made of the same parts in a similar arrangement, is also a camera. The two are built independently; cephalopods and vertebrates have not shared a sighted ancestor. The same engineering problem (build a camera-eye) got solved twice from scratch.

A platypus detects electric fields with specialized receptors in its bill; electric fish (sharks, knifefish, mormyrids) detect them with their own receptor systems. Independent invention again. The list goes on: echolocation in cave swiftlets, antifreeze proteins in arctic and antarctic fish, the bombardier beetle's chemical defense, the trap-jaw mechanism in ants and termites, parallel mimicry in butterflies.

Mainstream biology calls this convergent evolution. The standard story is that the same selection pressure (need to navigate in the dark, need to see, need to survive cold) produces the same engineering solution in separate lineages, and that this shows the power of natural selection. The trick was so good that nature found it twice.

But there is a problem with that explanation. The solutions are not simple. A camera-eye has many coordinated parts; it cannot half-function. Echolocation requires both the click-emission hardware and the neural circuitry to interpret returning echoes, and both have to be present from the start. The probability of natural selection plus random mutation hitting any one such solution, starting from scratch with the right combination of mutations, is extraordinarily low. The probability of hitting the same solution twice independently is the product, vastly lower.

A designer faces no such problem. An architect with a parts library will reuse optimal solutions across different projects. The same camera-eye in the cephalopod and the vertebrate is what one expects from one engineer, not from two independent random walks. Simon Conway Morris (a mainstream paleobiologist, not an ID proponent) argues in Life's Solution (Cambridge, 2003) that convergence is so pervasive that evolution is effectively predictable, that the same optimal solutions get hit again and again. Conway Morris is a theist; he frames this as evidence of a built-in directionality in the cosmos. The ID-side reading is closer to home: pervasive convergence on optimal designs across unrelated lineages is what one would expect if there is one designer reusing solutions.

In full

The Convergent Evolution as Design Signal Argument is an abductive argument that the pervasive independent appearance of highly complex features (camera-eye in cephalopods and vertebrates, echolocation in bats and dolphins, electroreception in platypus and electric fish, antifreeze proteins in arctic and antarctic fish) across lineages with no common ancestor possessing the feature is better explained by a single designer reusing optimal solutions across taxa than by unguided random-mutation-plus-selection independently producing the same vanishingly improbable combinations. The unguided-evolution rescue (convergent selection from similar pressures) requires the same low-probability combination of mutations to be hit independently in each lineage, a multiplicative probability collapse. The design rescue is predictive: an architect will reuse an optimal solution wherever the same engineering problem arises. Simon Conway Morris's mainstream defense ("convergence is predictable, evolution is constrained") concedes the directionality but locates the constraint in physical-chemical laws plus selection pressure; the ID-side reading takes the directionality as evidence of 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

Abductive (inference to the best explanation). P1 establishes the empirical phenomenon (independent appearance of complex features). P2 quantifies the cost of the unguided explanation (multiplicative probability collapse over independent occurrences). P3 shows that the design hypothesis predicts the phenomenon as a positive expectation rather than as an anomaly to be accommodated. On standard inference-to-best-explanation criteria (explanatory power, scope, prediction-vs-accommodation, parsimony), the design hypothesis wins or, at minimum, ties with unguided evolution, which is enough to undercut the inference from convergence to unguided power-of-selection. The argument does not depend on refuting evolution; it depends on showing that the pattern is at least as well explained by design.

P1, Highly complex features appear independently in lineages with no common ancestor possessing the feature

Affirmative case (second-order arguments)

1. The camera-eye in cephalopods and vertebrates. Octopus and squid eyes have cornea, iris, lens, retina, and image-forming optics, the same fundamental architecture as the vertebrate eye. The cephalopod-vertebrate common ancestor (a worm-like organism ~600 million years ago) had no such eye. Two independent evolutions of the same camera architecture. Detailed comparison: Sean Carroll (Endless Forms Most Beautiful, 2005) and Simon Conway Morris (Life's Solution, 2003).

2. Echolocation in bats and dolphins (and cave swiftlets). High-frequency sonar with neural decoding requires coordinated emission hardware (larynx modifications, nasal-passage modifications), neural circuitry for processing returning echoes (specialized auditory cortex), and behavioral integration. Bats and dolphins independently evolved the same system; their common ancestor was a small terrestrial mammal that did neither. Genetic evidence: Prestin (a motor protein in inner ear hair cells) shows convergent amino-acid substitutions in both lineages (Liu et al., Current Biology 2010).

3. Electroreception in platypus and electric fish. The platypus's bill contains electroreceptors that detect prey by their bioelectric fields. Many fish (sharks, knifefish, mormyrids) have independent electroreceptor systems. The platypus-fish common ancestor lacked any such system. Independent origin.

4. Antifreeze proteins in arctic and antarctic fish. Arctic fish (cod) and Antarctic fish (notothenioids) have antifreeze glycoproteins that prevent ice-crystal formation in body fluids. The two groups evolved their antifreeze proteins independently from different ancestral genes. Same engineering solution, two independent origins.

5. Mimicry in distantly-related taxa. The classic mimicry literature (Müllerian and Batesian mimicry in butterflies, moths, snakes) shows the same warning-pattern (color, shape, behavior) evolving independently in distantly-related species. The pattern often involves coordinated multiple traits.

6. The list keeps growing. Conway Morris's Life's Solution catalogues hundreds of convergence cases across phyla. Modern molecular biology has also revealed convergence at the genetic level: independent amino-acid substitutions producing the same protein function (Prestin in bats/dolphins; Na+/K+ ATPase resistance to plant toxins in monarch butterflies and unrelated insects; opsin-tuning amino-acid changes in independent vision evolution).

Anticipated objections

1. "Convergence is not as widespread as you say; many supposed convergences are actually parallelisms (similar genetic basis) reflecting shared developmental constraints."
2. "Some convergences (e.g., wings in birds, bats, pterosaurs, insects) are anatomically different at the detailed level; calling them 'the same' overstates the case."

Rebuttals

1. Even allowing for parallelism, the convergence literature is large and robust. Conway Morris's Life's Solution (2003) and Denton's Evolution: Still a Theory in Crisis (2016) catalogue hundreds of cases. Distinguishing parallelism from convergence often turns on whether shared developmental constraints "count" as common ancestry; this is itself a contested point. The core phenomenon (same engineering solution appearing in lineages without a sighted, echolocating, or electroreceptive common ancestor) is well-attested. Failure mode of the objection: chipping at terminology while leaving the empirical pattern intact.

2. Detailed anatomical differences do not undercut the convergence at the functional level. Yes, the bat wing and the bird wing are anatomically different in detail (bone composition, muscle attachment); both nevertheless solve the same engineering problem (flight) with the same functional outcome (powered flight using a membrane stretched over modified forelimb structures). The convergence is at the level of engineering solution to a problem, not pixel-by-pixel identity. The vertebrate and cephalopod eyes have the same functional architecture (cornea, lens, retina, image-forming optics) even though the cephalopod eye has its retina "the right way around" and the vertebrate eye has it inverted. Convergence at the design-pattern level is what the argument requires. Failure mode of the objection: demanding identity where similarity-at-the-engineering-level is the relevant criterion.

Live-cite kit

• Scripture: Psalm 104:24 ("in wisdom you made them all"); Genesis 1:20-25 (diversity of creatures within wise design).
• Scholarly: Simon Conway Morris (Life's Solution: Inevitable Humans in a Lonely Universe, Cambridge, 2003); Michael Denton (Evolution: Still a Theory in Crisis, Discovery, 2016); Sean B. Carroll (Endless Forms Most Beautiful, Norton, 2005, non-ID); Stephen C. Meyer (Darwin's Doubt, HarperOne, 2013); Liu et al., "Convergent sequence evolution between echolocating bats and dolphins", Current Biology 20 (2010): R53-R54.
• Aphorism: "Two lineages, no shared ancestor for the trick, same trick. That is a design fingerprint."

Tactical notes

• Lead with the camera-eye in cephalopods. It is the iconic example, easily explained, and visually intuitive.
• Use the Prestin echolocation paper for technical opponents. Current Biology 2010, mainstream venue. Convergent amino-acid substitutions in bats and dolphins are concrete molecular-level convergence.
• Don't claim no evolutionary biologist accepts the convergence problem. Conway Morris is mainstream. His framing differs; the data is shared.

P2, On the unguided hypothesis, the same vanishingly improbable combination must be hit independently in each lineage

Affirmative case (second-order arguments)

1. Complex features require multiple coordinated mutations to function. A camera-eye requires lens proteins (crystallins), photoreceptor proteins (opsins), retinal architecture, neural processing, transparent cornea, focusing mechanism, all to be present and integrated for vision to function. Each component is itself complex (opsins are sophisticated G-protein-coupled receptors). The probability of hitting a functional combination of any one of these by random mutation is low (Doug Axe's protein-folding work gives ~10^-77 for a 150-amino-acid functional fold). The probability of hitting all of them in a coordinated way is the product.

2. Independent occurrence multiplies the probability cost. If the probability of independently evolving a camera-eye in one lineage is P, the probability of independently evolving it twice is P². For echolocation, P². For electroreception, P². The unguided-evolution hypothesis predicts each independent occurrence requires its own low-probability path. Multiple independent occurrences across many features (camera-eye + echolocation + electroreception + antifreeze + mimicry + many more) compound the probability cost.

3. The "convergent selection makes it inevitable" rescue requires the selection landscape to be unusually narrow. Selection cannot create options it does not have access to; it can only choose among mutations that occur. For convergent selection to drive independent lineages to the same solution, the solution must be unusually narrow (only one or a few engineering choices work for the given problem), and the mutation rate must be high enough to reach the narrow optimum in each lineage's available time. Both conditions are contestable. Many problems have multiple workable solutions (think of the variety of bird and insect wing designs), so convergent selection should produce diverse solutions, not the same one.

4. The Axe-Behe edge-of-evolution math constrains the rescue. Behe (The Edge of Evolution, 2007) shows that random-mutation-plus-selection can produce at most 2-3 coordinated mutations on realistic mammalian timescales (extrapolating from malaria, HIV, and E. coli observed mutation rates). Camera-eye, echolocation, and electroreception each require far more than 2-3 coordinated mutations. The rate-of-evolution constraint applies to convergent and divergent evolution alike. See Edge of Evolution Argument.

Anticipated objections

1. "Selection guides each step; you don't need to hit the whole combination at once. Each intermediate is selected for its partial function.", gradualist defense.
2. "There are many examples of intermediate stages (e.g., the eye has evolved many times in many forms, from light-spots to camera-eyes); the gradualist path is well-documented.", intermediate-stages defense.
3. "Probability arguments against evolution are old and discredited; Dawkins's Climbing Mount Improbable shows how selection scales them."

Rebuttals

1. The gradualist defense assumes functional intermediates exist at every step. For irreducibly-complex systems (the camera-eye's coordinated parts; the echolocation's emission-plus-decoding integration), the intermediates may not be functional in the relevant way. Behe (Darwin's Black Box, 1996) and Meyer (Darwin's Doubt, 2013) develop this critique. The selection-guides-each-step defense works when each step has independent selective advantage; for coordinated functional systems, this is often what is in dispute. Failure mode of the objection: assuming the conclusion (functional intermediates exist) when the existence of intermediates is precisely the contested point.

2. Light-spot-to-camera-eye sequences are speculative reconstructions, not documented histories. The Dan-Eric Nilsson and Susanne Pelger (1994) eye-evolution simulation models a hypothetical sequence; it does not document a real evolutionary path. The simulation also assumes functional intermediates at every step, which is exactly what is in dispute. For the convergence problem, even granting Nilsson-Pelger fully, the independent traversal of the path in cephalopods and vertebrates remains unexplained: the same hypothetical path must be walked twice, hitting the same functional combinations. The intermediates-defense does not solve the multiplicative-probability problem. Failure mode of the objection: conflating possibility-in-principle with documented-in-history.

3. Dawkins's Climbing Mount Improbable (1996) is a rhetorical defense; the math is contested. Dembski (The Design Inference, 1998; No Free Lunch, 2002) provides the formal information-theoretic critique. Axe's experimental work gives the protein-folding probabilities. Behe's Edge of Evolution gives the realistic mutation-rate ceilings. Dawkins's mountain-metaphor assumes a smooth landscape from foothills to summit; the actual fitness landscape for complex functions is closer to discrete islands separated by deep valleys of non-function. The metaphor is rhetorically attractive but technically contested. Failure mode of the objection: citing a metaphor as if it settled a technical dispute.

Live-cite kit

• Scholarly: Michael Behe (Darwin's Black Box, Free Press, 1996; The Edge of Evolution, Free Press, 2007); William Dembski (The Design Inference, Cambridge, 1998; No Free Lunch, Rowman & Littlefield, 2002); Douglas Axe (Undeniable, HarperOne, 2016; "Estimating the Prevalence of Protein Sequences Adopting Functional Enzyme Folds", Journal of Molecular Biology 341, 2004); Stephen C. Meyer (Darwin's Doubt, HarperOne, 2013).
• Aphorism: "If the probability of evolution producing a camera-eye is one in a trillion, the probability of producing two independently is one in a trillion trillion."

Tactical notes

• Hit the multiplicative point hard. The cost of independent occurrence is multiplicative, not additive. This is the core P2 move.
• Be ready for the Nilsson-Pelger eye-evolution simulation. It is the most-cited gradualist defense. Note that it is a simulation, not history; it assumes functional intermediates; and it does not address the independent-occurrence problem.
• Pair with Edge of Evolution Argument. The Behe rate-constraint underwrites the probability case across all convergence examples.

P3, On the design hypothesis, convergence is unsurprising: a single designer reuses optimal solutions across taxa

Affirmative case (second-order arguments)

1. Engineering practice reuses optimal solutions. Automotive engineers reuse wheel-and-axle architecture across cars, trucks, motorcycles, and aircraft landing gear, because it works. Software engineers reuse algorithms (sort, search, hash) across applications. The same architect facing the same engineering problem in a new context will reach for the proven solution. Convergence is the signature of this practice.

2. Common design predicts convergence as a positive expectation. Wherever the engineering problem recurs (need to navigate in the dark, need to see, need to detect prey at distance, need to survive cold), the designer reuses the optimal solution. The design hypothesis predicts the convergence pattern in advance, not as an anomaly to be accommodated. This is a positive prediction, not a post hoc rescue. Stephen Meyer, Cornelius Hunter, Casey Luskin develop this point.

3. Simon Conway Morris's mainstream argument concedes the directionality. Conway Morris's Life's Solution (Cambridge, 2003) argues that convergence is so pervasive that evolution is effectively predictable, that intelligent humanoids are an "inevitable" outcome. Conway Morris is a theist and frames this as built-in directionality in the cosmos (compatible with theistic evolution). The ID-side reading takes the same data and infers a designer who builds in the optimal solutions directly. Conway Morris's argument concedes the structure of the phenomenon; the dispute is about cause.

4. The independence problem disappears under design. Where unguided evolution has a multiplicative probability cost for independent occurrence, design has no such cost. The same designer reusing the same solution across taxa is a single explanatory move, not multiple independent improbable events. On parsimony grounds, design is the simpler explanation.

Anticipated objections

1. "Conway Morris's argument is for inevitability-from-natural-laws, not design; you can't claim him for the ID side."
2. "The 'designer reuses solutions' move is unfalsifiable; any pattern can be retrofitted to it."
3. "Convergent evolution is evidence for the power of natural selection, not against it.", the textbook framing.

Rebuttals

1. Conway Morris's argument structurally concedes what the ID argument needs. He grants that convergence is pervasive and directional. He locates the directionality in physical-chemical constraints plus selection pressure. The ID side disputes the cause (natural law alone vs designed-by-an-architect-built-into-the-system), not the structure (directional evolution converging on optimal solutions). The disagreement is about cause; the phenomenon is shared. Citing Conway Morris is honest acknowledgment that mainstream paleobiology accepts the directionality. Failure mode of the objection: conflating an author's framing with the data the author documents.

2. The design-reuses-solutions move has positive content. It predicts (a) convergence on optimal solutions for recurring problems, (b) modular components reused across multiple lineages, (c) functional optimization at the engineering level (not random pattern), (d) discontinuity at the level of baraminic kinds. Each is testable; each has empirical content. Failure mode of the objection: stipulating unfalsifiability without examining the predictions.

3. The textbook framing presupposes the conclusion. Saying convergence shows the power of natural selection assumes selection is the mechanism, which is what is contested. The argument is not that selection cannot produce any convergence (low-complexity convergences are unproblematic for the gradualist account); the argument is that highly complex convergences (camera-eye, echolocation, electroreception) push the probability cost beyond what selection plus realistic mutation rates can deliver. The textbook framing works for simple cases and fails for the hard ones. Failure mode of the objection: using easy cases to license the hard ones.

Live-cite kit

• Scripture: Psalm 104:24 ("in wisdom you made them all"); Romans 1:20 (invisible attributes of God seen in what is made); Colossians 1:16-17 (Christ as architect).
• Scholarly: Simon Conway Morris (Life's Solution, Cambridge, 2003; The Crucible of Creation, Oxford, 1998); Michael Denton (Evolution: Still a Theory in Crisis, Discovery, 2016); Günter Bechly (The Fossil Record essays); Stephen C. Meyer (Darwin's Doubt, 2013; Return of the God Hypothesis, 2021); Cornelius Hunter (Darwin's God, 2001).
• Aphorism: "When two engineers build the same solution from scratch, you wonder. When the same engineer builds it twice, you don't."

Tactical notes

• Use Conway Morris carefully. He is a theist (an Anglican) and a mainstream paleobiologist; he is not on the ID-movement team. But he documents the convergence data in mainstream venues. Cite the data; acknowledge his framing; explain the structural agreement.
• Press the parsimony point. One designer reusing solutions is structurally simpler than many independent low-probability events. On standard inference-to-best-explanation criteria, parsimony favors design.

Conclusion

Convergent evolution is a design signal, or at minimum a serious anomaly for unguided neo-Darwinism that the design hypothesis explains naturally. Highly complex features (camera-eye, echolocation, electroreception, antifreeze proteins, mimicry) appear independently in lineages with no common ancestor possessing the feature. On unguided evolution, each independent occurrence requires hitting a low-probability combination of mutations, with the cost of independent occurrence multiplicative. On design, the pattern is what one expects: a single architect reuses optimal solutions across taxa whenever the same engineering problem arises. Simon Conway Morris's mainstream defense ("convergence is predictable, evolution is constrained") concedes the directionality without naming the architect; the ID-side reading takes the directionality as evidence of design. On standard inference-to-best-explanation criteria (explanatory power, prediction-vs-accommodation, parsimony), the design hypothesis wins or ties. Convergence is a design fingerprint pressed into the biological record.

Master objections to the argument as a whole

1. "Convergent evolution is one of the strongest evidences for natural selection's power, not against it.", Reply: this assumes selection is the mechanism; it presupposes what is contested. For simple convergences, gradualist accounts work; for highly complex convergences, the probability cost compounds and the gradualist accounts strain. The argument targets the hard cases.

2. "Conway Morris is not an ID proponent; misappropriating him.", Reply: the argument cites him for the data and the structure of the phenomenon, not for the cause attribution. Conway Morris accepts directionality; the ID-side argument takes directionality as design evidence. The framing is different; the empirical agreement is real.

3. "This is unfalsifiable: design fits any pattern.", Reply: design has positive predictions (convergence on optimal solutions, modular reuse, engineering optimization within constraints). Each can be tested. The unfalsifiability charge is dialectical.

4. "Even if convergence is anomalous, descent with modification is supported by other evidence.", Reply: conceded for some forms of descent; the argument operates within the cumulative case (phylogenetic incongruence, orphan genes, Cambrian all push the same direction). Convergence is one strand of the cumulative case, not the whole case.

5. "You're conflating engineering similarity with biological similarity; the analogy is misleading.", Reply: the analogy works because cells are engineered systems at the molecular level. The genetic code is a digital coded language; molecular machines are literal machines; the analogy is structural, not metaphorical. The engineering analogy is more apt for biological systems than for any other physical objects.

Tactical opening / closing

Opening line: "A bat and a dolphin both echolocate. An octopus and a human both have camera-eyes. A platypus and an electric fish both detect electric fields. Their common ancestors did none of these things. The same vanishingly improbable engineering trick has been hit again and again in lineages with no shared ancestry for the trick. Mainstream biology calls this convergent evolution and says it shows the power of natural selection. I am going to show you that the better reading, the one Simon Conway Morris approaches from one direction and the design community approaches from another, is that convergence is the signature of a single architect reusing optimal solutions."

Closing landing strip: "The Convergent Evolution as Design Signal Argument does not refute evolution; it relocates the inference. Where unguided evolution must hit the same low-probability combination twice independently (multiplying the cost), the designer hits it once and reuses it across taxa (no extra cost). Simon Conway Morris, a mainstream paleobiologist who is not on the ID team, calls convergence so pervasive that intelligent life is inevitable; he locates the directionality in built-in constraints. The ID reading names the source of the constraint: an architect. On parsimony, on prediction, on engineering-pattern-recognition, the design reading is at least as good and on the cumulative data may be better."

Connection to Scripture

• Genesis 1:20-25, creatures "after their kind"; the same designer giving different lineages the same engineering solutions.
• Psalm 104:24, "How many are Your works, O Lord! In wisdom You have made them all"; wisdom-of-design as origin of biological diversity.
• Job 38, God's creation challenge to Job; the architecture of life as God's signature.
• Romans 1:20, invisible attributes of God seen through what is made; design as natural-theology evidence.
• Colossians 1:16-17, "in Him all things were created… in Him all things hold together"; Christ as the architect of reused biological solutions.
• John 1:1-3, "all things came into being through Him"; the Logos as universal designer.

Patristic / scholarly note

Classical / patristic / medieval:

• Basil of Caesarea (Hexaemeron, c. 378), the wisdom of God displayed in the structure of creation.
• Augustine (De Genesi ad Litteram, c. 415), the divine wisdom as ground of creaturely order and pattern.
• Thomas Aquinas (Summa Theologica, I, Q. 47), the multiplicity of creatures expressing the goodness of God; design as the explanation for the patterned diversity.

Modern:

• Simon Conway Morris (The Crucible of Creation, Oxford, 1998; Life's Solution: Inevitable Humans in a Lonely Universe, Cambridge, 2003), mainstream paleobiologist; framing convergence as predictable / inevitable from constraints.
• Sean B. Carroll (Endless Forms Most Beautiful, Norton, 2005), non-ID; deep conservation of developmental toolkits.
• Liu Y. et al., "Convergent sequence evolution between echolocating bats and dolphins", Current Biology 20 (2010): R53-R54, molecular-level convergence in Prestin.
• Richard Dawkins (The Greatest Show on Earth, Free Press, 2009; Climbing Mount Improbable, Norton, 1996), strongest contemporary critic; convergence-shows-selection-power framing.
• Michael Denton (Evolution: A Theory in Crisis, Adler & Adler, 1985; Evolution: Still a Theory in Crisis, Discovery, 2016), structuralist-design framing; convergence as evidence for built-in archetypes.
• Günter Bechly (paleontologist, Discovery Institute fellow), convergence in fossil record.
• Stephen Meyer (Darwin's Doubt, HarperOne, 2013; Return of the God Hypothesis, HarperOne, 2021), ID-side synthesis of convergence evidence.
• Cornelius Hunter (Darwin's God, Brazos, 2001), philosophical analysis of the design-vs-descent inference.
• Casey Luskin (Discovery Institute essays on convergence as design signal), accessible summary papers.

See also

• Common Design vs Common Descent Argument, sister argument on the design-or-descent underdetermination.
• Phylogenetic Incongruence Argument, parallel critique of universal common descent.
• Orphan Genes Argument, parallel anomaly at the genetic level.
• Cambrian Explosion Argument, top-down body-plan origination.
• Edge of Evolution Argument, rate-of-evolution constraint underlying the probability case.
• Irreducible Complexity Argument, molecular-machine design evidence.
• Specified Complexity Argument, the information-theoretic design filter.
• Common Descent Critique, concept-side parent.
• Intelligent Design, framework that takes design hypothesis seriously.
• Stephen Meyer, primary contemporary defender.
• Origins and Science, concept-side master.
• Origins, master hub.
• Arguments, top-level master index.