Biological Design

Taken together, the six articles find no test yet that design passes and evolution fails. The design inference, irreducible complexity and redundancy each wait on a number neither side has produced, and simple design reuse fails. The flaws critics cite mostly argue common descent, not against design. Parasites raise a moral question that well-made machinery sharpens.

By AI (Claude Opus 5.5, thinking level unavailable) Draft:  October 2026

"Clearly, designers who have the ability to make better designs do not necessarily do so." — Michael Behe, Darwin's Black Box, 20061

"Malaria was intentionally designed." — Michael Behe, The Edge of Evolution, 20072

Summary

This section asks two questions about living things.  Do they carry marks of engineering that point to a designer?  And do the flaws critics point to count against one?  Six articles take up the arguments one at a time.  This overview sets their verdicts side by side and asks what they show together.

The findings:

  1. That organisms look designed is common ground. Leslie Orgel wrote in 1973 that "living organisms are distinguished by their specified complexity."3 Richard Dawkins wrote in 1986 that they have "some quality, specifiable in advance, that is highly unlikely to have been acquired by random chance alone."4 The dispute is over whether any method can turn that appearance into an inference.
  2. The Design Inference: valid logic, no biological number. William Dembski's filter removes only the chance hypotheses fed to it. Its one worked biological figure prices a random draw of proteins, and in 2008 Dembski wrote that the example "doesn't fit the bill."5
  3. Irreducible Complexity: the narrow claim survives as an absence of evidence. Michael Behe's 1996 definition rules out direct routes only. Nobody has priced an indirect route for any system he named, and nobody has rebuilt one with each step's advantage measured.
  4. Convergent Evolution and Design Reuse: simple reuse fails. Where the same feature arose twice, either the parts differ, or a few working sites match while the silent DNA follows the ordinary family tree. The careful design readings fit the data, but only by predicting what selection predicts.
  5. Genetic Redundancy: a real test that nobody has run. Life carries more backup than a lean design needs, but most measured backup is partial, and selection keeps partial backups. A perfect backup is what evolutionary theory expects to decay, and nobody has counted them.
  6. Arguments from Bad Design: neither imperfection nor usefulness decides. A defect claim needs a stated task, a feasible alternative and a priced trade-off. The case articles under it find that most complaints are arguments for common descent, and that a few measured costs have no design answer yet.
  7. Predation, Parasites, and Pathogens: a different complaint. Corruption models explain how some partnerships turned harmful. They do not explain equipment built to harm, and that equipment works well.
  8. The two halves of the question meet at one point. The design arguments rest on how well parts fit a purpose. The parasite argument rests on the same fit, aimed at harm, and Behe accepts its premise: "Malaria was intentionally designed."2 The stronger the case for engineering, the sharper the moral question for a good Creator.

Verdict.  Taken together, the six articles do not show marks of engineering that only a designer could leave.  Nor do the flaws critics cite count against design as such, because design does not promise the best possible build.  In every article the popular form of the argument fails and a careful form survives, either as a claim nobody has priced or as a prediction nobody has tested.  The critics are in the same position, since their routes are sketched and not measured.  What the flaws do count against is narrower.  Vestiges and shared layouts are evidence for common descent, which a young-earth model must meet wherever they cross created kinds.  Harmful equipment that works well is not explained by later damage, and that leaves the young-earth account of harm unproved.

Part 1: The design inference

The Design Inference tests William Dembski's rule for rejecting chance.  An event that matches an independent pattern, and is improbable enough under every relevant chance hypothesis, is assigned to design.  The logic is sound, but Dembski wrote that the user must first "know what probability distribution(s), if any, were operating to produce the event."6  His one worked biological number, about 1 in 10234 for the bacterial flagellum, is the chance of drawing 250 suitable proteins at random from 4,289.7  No biologist proposes that route.  Dembski wrote in 2005 that the probability under a hypothesis with selection in it "has yet to be done," and in 2008 that the example "doesn't fit the bill."85  The article's verdict is that specified complexity is an accurate description of living things, not yet a proof, and that the critics' preferred likelihood method cannot be run either, because it needs the probability of a flagellum given a designer.

Part 2: Irreducible complexity

Irreducible Complexity finds that Behe's 1996 claim was narrower than its popular form.  On the page after his definition he wrote that "one can not definitively rule out the possibility of an indirect, circuitous route," and that such a route becomes less likely as the parts multiply.1  That is a probability claim, and nobody on either side has calculated it for a named machine.  Hermann Muller's 1918 route, in which an added part becomes necessary once other parts depend on it, has been shown once in a laboratory, and the authors call its generality "unknown."9  No route to the cilium, the flagellum, the clotting cascade or gated transport has been rebuilt with each step's advantage measured.  The two child articles hold the same line.  The Bacterial Flagellum finds that the flagellum has not been shown to be unevolvable and its evolution has not been shown.  Blood Clotting finds that the part of the cascade Behe actually claimed is essential in every knockout and present in every vertebrate genome examined, and that no step in building it has a measured advantage.

Part 3: Convergent evolution and design reuse

Convergent Evolution and Design Reuse tests the claim that one designer reused designs across unrelated creatures.  Fazale Rana states it plainly:  "it's reasonable to expect the same designs to repeatedly appear throughout nature."10  In every case the article read, either the parts differ, as in the octopus eye and the fish antifreezes, or the same few amino acids changed at working sites while the rest of the gene followed the ordinary tree.  In the hearing gene prestin, "when only synonymous substitutions are used, dolphin and cow are correctly grouped with 100% bootstrap support."11  The article concludes that the simple reuse argument should be retired.  Walter ReMine's and Winston Ewert's versions survive, but they have not predicted anything here that selection does not.  On the other side, convergence in anatomy and fossils still has no independent test and is assigned after a trait conflicts with the tree.

Part 4: Genetic redundancy

Genetic Redundancy starts from a sentence by four biologists, John Maynard Smith among them:  redundancy "is sensible from an engineer's point of view."12  Most genes that can be knocked out without visible harm turn out to matter under some other condition, so most silent knockouts are not backups.  Where real backup has been measured, duplicate genes carry a large share, and partial backups are kept by selection.  What survives is narrow.  A truly redundant gene "would not be protected against the accumulation of deleterious mutations," as Martin Nowak's team wrote, and a conservation test cannot see one.12  If perfect backups are common, that favours design.  Peer Terborg, writing in the creationist Journal of Creation, claims that "redundant genes cannot have resided in the genome for millions of years."13  On a young-earth timeline that makes a testable prediction:  perfect backups should still be intact.  Neither side has run that test.

Part 5: Arguments from bad design

Arguments from Bad Design is about method.  A critic must name the task a structure fails at, show a feasible alternative and price the trade-off.  A defender must do the same for any claimed benefit, and cannot treat an unknown purpose as a permanent exemption.  Behe denies that design entails the best possible build:  "Clearly, designers who have the ability to make better designs do not necessarily do so."1  Darwin, for his part, wrote that "Natural selection will not produce absolute perfection."14  So imperfection alone cannot exclude a designer, and usefulness alone cannot prove one.

The case articles under it give a consistent picture.

  1. Several famous complaints overstate the anatomy. The Inverted Retina finds that retinal cells guide light, though that does not show the layout is best. The Appendix finds a working immune organ that belongs on neither side's list of vestiges.
  2. The useless-organ argument is dead, and the live argument is about ancestry. Vestigial Structures finds that the live argument defines a vestige as a reduced version of a part other species share, and a newly found function does not answer that. Its whale-pelvis, snake-spur and horse-splint cases hold up as vestiges. The Human Body Under Criticism finds the same: the whole-body complaint is mainly an argument for common descent, and its one whole-body redesign has never been costed.
  3. A few costs are real and unanswered. The Human Body Under Criticism also finds back pain and groin hernia to be large measured costs with no design answer in numbers. Bad Design in Biochemistry finds that Fazale Rana's design answer on uric acid fails, that the reading of urate oxidase as a broken gene shared with apes holds, and that no young-earth treatment of it was found.

Part 6: Predation, parasites, and pathogens

Predation, Parasites, and Pathogens states the complaint that bad design misses.  Its first distinction is that working well is different from doing good:  a structure can do its job efficiently while doing another creature great harm.  Darwin put the problem in 1860, naming "the Ichneumonidæ with the express intention of their feeding within the living bodies of caterpillars."15  Young-earth writers answer with alteration, displacement and front-loaded capacities.  Warren Shipton allows "gene loss, the expression of front-loaded genes now facilitating pathological life cycles, or other scenarios."16  The article finds that the observed breakdown of partnerships and the damaged genes of dependent organisms support corruption in particular cases.  Coordinated equipment for harming another organism is the strongest difficulty.  Losing functions can explain why an organism needs a host, but not by itself how it enters and exploits one.  Predation, Venom, and Carnivory reaches the same result for weapons:  loss explains dependence, not fangs and venom.  The verdict is mixed:  the evidence shows neither that every harmful organism was created in its present role nor that every one descended from a harmless form.

Part 7: How the verdicts bear on each other

Argument Strongest design-side form, as the article states it Strongest critic form, as the article states it The article's verdict What would change it
The design inference A stated pattern, a stated chance model that includes selection, and a computed probability The filter removes only the hypotheses fed to it, and the only biological figure prices random assembly Valid logic, not yet a method for biology A published probability for a named structure under a hypothesis with selection in it
Irreducible complexity Indirect routes grow less likely as parts multiply, and no measured route exists for Behe's systems Muller's route works, the parts have relatives, and one machine was shown to gain a part The universal negative fails; the narrow claim survives as an absence of evidence A rebuilt route with each step's advantage measured, or a priced indirect route; for design, a system whose parts have no relatives and no simpler working version
Convergence and reuse ReMine and Ewert: alike but not identical, and a module changes only what its job needs The parts differ, and the silent sites follow the ordinary tree Simple reuse fails; the careful readings predict nothing selection does not Matching silent sites or introns in unrelated lineages would favour copying; an independent test would settle anatomical cases
Genetic redundancy Perfect backups should decay under evolution, and a conservation test cannot see them Most silent knockouts are not backups, duplicates carry much of the measured backup, and a model builds redundancy without copying Backup is real and abundant; the narrow argument about perfect backups is untested A count of genes whose loss costs nothing in any condition while a partner is present, and whether they are intact across a kind
Bad design Design does not entail the best build, and several complaints overstate the anatomy A defect needs a known task and a measured shortfall, not a full redesign Neither imperfection nor usefulness decides; most cases argue common descent A priced trade-off for a disputed structure, or a stated design prediction tested against repeated avoidable failures
Predation and parasites Alteration, displacement and front-loading, with observed breakdown of partnerships Equipment that works well for harm, which loss does not explain Corruption is supported in particular cases; harmful equipment and a recent date remain unexplained A harmless ancestor shown together with the origin of the harmful equipment, or a demonstrated earlier use for venom

Bad design and irreducible complexity aim at different targets

Irreducible complexity asks whether unguided steps could build a machine.  It says nothing about common descent or the age of the earth, and Behe finds "the idea of common descent (that all organisms share a common ancestor) fairly convincing."1  Most bad-design cases, by their own articles' findings, are arguments for common descent.  So a winning bad-design case does not touch Behe's argument, and a winning irreducible-complexity case does not answer the ancestry evidence.

The two meet only on the designer's competence.  Behe's page 223 shields irreducible complexity from "a designer would have done better," because his inference rests on how the parts depend on each other, not on how well the whole performs.  The bad-design article sets the price of that shield:  a design advocate who cites good performance as evidence of intentions must also say how poor performance bears on the same inference.  For a young-earth reader the result is uncomfortable.  Even if irreducible complexity wins, it wins inside common descent, and it does nothing to answer the vestige cases.

Redundancy and irreducible complexity use the same test

Both arguments read knockouts.  Irreducible complexity says that removing one part stops the machine.  Redundancy says that most removals cost little.  Behe wrote in 2001 that "The current definition puts the focus on removing a part from an already-functioning system," and called the gap "an asymmetry between my current definition of irreducible complexity and the task facing natural selection."17  The same limit applies to redundancy:  a knockout shows what a part does now, not how it came to be there.

The two also bear on each other directly.  A backup makes a machine look less irreducible.  The Bacterial Flagellum finds six proteins that can be deleted with swimming kept in whole or in part, and Blood Clotting finds every documented loss in whales, birds and puffer fish before the fork Behe set aside.  So the irreducible core is what remains once the backups and spare parts are removed.  A designer who adds backups would leave a smaller irreducible core, and Muller's route, which makes parts come to depend on each other, would leave a larger one.  Knockout data alone cannot tell these readings apart, which is why both articles end at numbers neither side has produced.

Redundancy answers "a designer would not duplicate"

A common skeptic's line is that duplicated genes are the leftover mess of copying.  The redundancy article answers that leftovers should decay, while many duplicate pairs keep overlapping jobs and some switch on when their partner fails.  Ran Kafri, Melissa Levy and Yitzhak Pilpel concluded that "compensation for gene loss is merely a side effect of sophisticated design principles using functional redundancy," using "design" for engineering logic, not for a designer.18  Engineers duplicate on purpose.  For the Curiosity rover, NASA's Gerard Holzmann writes, "Running the same landing software on two CPUs in parallel offers little protection against software defects," so the backup computer ran a different, simplified program.19  Duplication is therefore no mark against design.

It is no mark for design either.  Andreas Wagner notes that for any system with redundant parts "one can design a system without redundancy but the same transfer function, i.e. the same behavior".20  Evolution explains partial backups, and only perfect ones discriminate.  A broken copy is a different case, and Shared Genetic Errors treats it.

Convergence and redundancy show how much the design readings can absorb

Holzmann's point is that a backup built differently protects against a shared flaw.  ReMine's designer works the same way:  "a designer must not indiscriminately use the same design repeatedly."21  So the design side accepts the same parts in unrelated creatures as reuse, and different parts doing one job as sound engineering.  A reading that accepts both outcomes predicts neither.  The convergence article reaches this point for ReMine and Ewert, and the redundancy article judges the engineering analogy suggestive but not decisive.  Each article names the test that would discriminate:  silent DNA sites for convergence, which the simple reuse reading failed, and the decay of perfect backups for redundancy, which nobody has run.

Bad design and parasites are opposite complaints

The bad-design complaint says that life is too poorly built for a designer.  The parasite complaint says that some of it is too well built for a good one.  The reply that designers need not build the best answers the first and says nothing to the second.  The second is a moral argument, and every success of the machinery arguments strengthens it, because the evidence for purposeful arrangement is the same evidence in a flagellum and in a malaria parasite.  Behe accepts this:  "Malaria was intentionally designed."2  Biology can sort which harms arose by loss and which by gain, and the parasite articles do that case by case.  Whether animal death preceded the Fall is argued in Death Before the Fall.

Every design argument here waits on a missing number

The design inference needs a probability under a hypothesis with selection in it.  Irreducible complexity needs a priced indirect route.  Design reuse needs a prediction that selection does not make.  Redundancy needs a count of perfect backups.  The critics lack numbers too:  no measured route for any of Behe's systems, no probability of a flagellum given a designer, and no independent test for convergence in fossils.  The design side has made the improbability claim, so it owes its number first.  The critics' reply that a route is conceivable is unfinished in the same way.

Objections and responses

  1. "Every article here ends with 'nobody has computed it.'  The design side makes the claim, so that is a defeat for design."  Partly right.  The design inference and irreducible complexity both claim an improbability, so they owe the number, and after almost thirty years it has not appeared.  But the critics claim that routes exist, and they have not measured one for any of Behe's systems.  The redundancy test is different again, because there evolutionary theory makes the prediction that a perfect backup decays, and nobody has checked it.  The fair reading is that design has not been proven and the questions are not closed.

  2. "'Designers do not necessarily build the best' makes design unfalsifiable."  The bare inference to some designer is hard to refute.  The design inference article finds the same defect in Dembski's claim of no false positives:  if the natural Oklo reactors passed his test, "At worst it would indicate that certain naturally occurring events or objects that we initially expected to involve no design actually do involve design."7  But specific design claims can fail, and one already has.  Simple reuse predicted the same parts in convergent creatures, and the parts differ.  The redundancy prediction and the test for parts with no relatives can fail too.  The bad-design article states the choice:  a modest design inference with modest predictions, or a specific one that accepts a greater burden of testing.

  3. "The design side counts both sameness and difference as evidence."  Correct, and Part 7 says so.  Same parts in unrelated creatures are read as reuse, and different parts doing one job as engineering.  A reading that absorbs both outcomes gains nothing from either.  Only the tests at silent DNA sites and in the decay of perfect backups can separate the readings.

  4. "Bad design may not refute design, but it refutes young-earth creation."  This is the skeptic's strongest objection, and the case articles grant most of its premise.  The vestiges that hold, such as whale pelves, and the broken urate oxidase gene shared by humans and apes are evidence of descent between creatures a young-earth model treats as separately created.  Finding a function does not answer them.  The answer, if there is one, lies in the common-descent articles, Phylogenetic Congruence and Shared Genetic Errors, not in this section.

  5. "Parasites show that the designer is cruel, so design is no help to Christianity."  Biology cannot settle a moral argument, and none of these articles tries.  What they show is narrower.  Some harm arose when partnerships broke down, which fits a corruption model.  Some harmful equipment, such as venom, is a gain that works well, and the young-earth proposals for its origin, such as genes switched on at the Curse, have no evidence for an earlier use.  The moral question is real and grows with the strength of the design case.  It is a question for theology, and the scriptural side of it is in Death Before the Fall.

  6. "The appearance of design is what selection produces, so it is no evidence."  That is the critics' claim, and where a route has been measured it may be right.  For the machines in Parts 1 and 2 nobody has measured a route, so the claim is an expectation, not a finding.  The appearance is common ground, as Organisms Appear Designed documents.  Its cause is disputed, and neither side has priced its explanation for the hard cases.

Issues and unanswered questions

Weaknesses on the design side:

  1. No probability has been published for any biological structure under a chance hypothesis that includes selection.
  2. Irreducible complexity is silent on the age of the earth, and its author accepts common descent. Using it for young-earth creation needs a separate argument.
  3. No young-earth author has published an analysis of any convergence case, and no baraminology paper on convergence between kinds was found.
  4. No young-earth treatment of the urate oxidase pseudogene was found.
  5. No earlier use has been shown for fangs and venom, and loss does not explain them.
  6. Back pain and groin hernia are large measured costs that the design side has not answered with numbers.

Weaknesses on the critics' side:

  1. No route to any of Behe's systems has been rebuilt with each step's advantage measured.
  2. The likelihood comparison the critics prefer needs the probability of a structure given a designer, which nobody can supply.
  3. Convergence in anatomy and fossils is still assigned after a trait conflicts with the tree.
  4. The model that builds redundancy without duplication assumes the step that gives a gene a second job.
  5. The one whole-body redesign offered by critics has never been costed.

Open for both:

  1. How many perfect backups exist, and are they intact across a created kind? Competition assays across many conditions with double knockouts would answer the first.
  2. Why are workable designs so few? The convergence article calls this the design argument's best ground there, but it is an argument about the laws that limit life, not about reuse.
  3. Can a rejection region, the set of outcomes that would count as a match, be defined for a molecular machine?

Verdict

The six articles agree on more than their separate verdicts suggest.  In each, the popular form of the argument fails.  "No gradual route exists" cannot be shown, convergent creatures do not share their parts, most redundancy is not hidden spares, imperfection does not exclude a designer, usefulness does not prove one, and corruption does not explain every harm.  In each, a careful form survives as a question whose test is known and has not been run.  The design side owes a probability with selection in it, a priced route, a count of perfect backups.  The critics owe measured routes and a test for convergence in bone.

Living things do show features an engineer would recognise:  parts that depend on each other, backups, repeated solutions and machinery fitted to a purpose.  None of these yet separates a designer from selection.  The flaws critics cite do not count against design, because design does not promise the best build.  They count against narrower claims:  separate creation of creatures that share vestiges and broken genes, and a young-earth account of harm that has not explained harmful equipment.  This section should present design as a live reading that has not been proven and has not been refuted.  It should not present design as a demonstration.

Sources

  1. 1.

    Behe, Michael J.  Darwin's Black Box: The Biochemical Challenge to Evolution.  Free Press.  Tenth-anniversary edition.  2006.  Behe's argument that some cell machinery is irreducibly complex and could not be built by direct Darwinian steps, first published in 1996 and reprinted with a new afterword that answers his critics.  The article uses it for the limit Behe set on his own definition, for his acceptance of common descent, and for his denial that a designer must build the best.

    1. 1a.

      Page 5:  “For the record, I have no reason to doubt that the universe is the billions of years old that physicists say it is. Further, I find the idea of common descent (that all organisms share a common ancestor) fairly convincing, and have no particular reason to doubt it.” ↩

    2. 1b.

      Page 40:  “Even if a system is irreducibly complex (and thus cannot have been produced directly), however, one can not definitively rule out the possibility of an indirect, circuitous route. As the complexity of an interacting system increases, though, the likelihood of such an indirect route drops precipitously.” ↩

    3. 1c.

      Chapter 10, Questions About Design, section A Hole in the Eye, page 223:  “Clearly, designers who have the ability to make better designs do not necessarily do so.” ↩ ↩

  2. 2.

    Behe, Michael J.  The Edge of Evolution: The Search for the Limits of Darwinism.  Free Press.  2007.  Behe's book arguing, largely from malaria data, that random mutation and selection reach only a few coordinated changes.  Its tenth chapter argues that the malaria parasite's machinery was designed.  The article uses that chapter because Behe there accepts the moral cost of a design inference.

    1. 2a.

      Chapter 10, All the World's a Stage, section When Bad Things Happen to Good People:  “Malaria was intentionally designed.” ↩ ↩ ↩

  3. 3.

    Orgel, Leslie E.  The Origins of Life: Molecules and Natural Selection.  John Wiley and Sons.  1973.  A book on the chemical origin of life by an origin-of-life chemist.  The article quotes it because it uses "specified complexity" to describe living things decades before the design movement took up the phrase.

    1. 3a.

      Page 189:  “living organisms are distinguished by their specified complexity.” ↩

  4. 4.

    Dawkins, Richard.  The Blind Watchmaker.  W. W. Norton.  1986.  Richard Dawkins's book arguing that natural selection explains the complex adaptations of living things without a designer.  The article quotes his account of what makes living things complicated, because it grants the appearance the design argument starts from.

    1. 4a.

      Page 9:  “complicated things have some quality, specifiable in advance, that is highly unlikely to have been acquired by random chance alone.” ↩

  5. 5.

    Dembski, William A.  “Comment on "Some Thanks for Professor Olofsson".”  Uncommon Descent.  Dec 3, 2008.  A comment by Dembski on the intelligent-design blog he founded.  Its second point says that the flagellum example in No Free Lunch does not meet the standard a biological calculation needs.  A week later Dembski restated his support for the explanatory filter, but he did not take back this point.

    1. 5a.

      Paragraph point 2:  “The challenge for determining whether a biological structure exhibits CSI is to find one that's simple enough on which the probability calculation can be convincingly performed but complex enough so that it does indeed exhibit CSI. The example in NFL ch. 5 doesn't fit the bill.” ↩ ↩

  6. 6.

    Dembski, William A.  The Design Inference: Eliminating Chance through Small Probabilities.  Cambridge University Press.  1998.  The book in which Dembski set out the explanatory filter, which assigns an event to design once regularity and chance have been ruled out.  The article uses his statement that the filter needs the operating chance hypotheses as input.

    1. 6a.

      Page 41:  “before we even begin to send E through the Explanatory Filter, we need to know what probability distribution(s), if any, were operating to produce the event.” ↩

  7. 7.

    Dembski, William A.  No Free Lunch: Why Specified Complexity Cannot Be Purchased without Intelligence.  Rowman and Littlefield.  2002.  Dembski's application of the filter and of specified complexity to biology.  Chapter 5 works a probability for the bacterial flagellum.  The article uses that figure, which prices a random draw of proteins, and his treatment of the Oklo natural reactors as a possible false positive.

    1. 7a.

      Page 293:  “Randomly picking 250 proteins and having them all fall among those 500 therefore has probability (500/4,289)^250, which has order of magnitude 10^-234 and falls considerably below the universal probability bound”.  Recomputed, 250 × log10(500/4,289) = −233.3, which is the order of magnitude Dembski gives. ↩

    2. 7b.

      Page 27:  “The precise probabilities for such conditions have yet to be ascertained.”  “But suppose the Oklo reactors ended up satisfying this criterion after all. Would this vitiate the complexity-specification criterion? Not at all. At worst it would indicate that certain naturally occurring events or objects that we initially expected to involve no design actually do involve design.” ↩

  8. 8.

    Dembski, William A.  Specification: The Pattern That Signifies Intelligence.  Aug 15, 2005.  Dembski's restatement of specification as a measurable quantity, in a manuscript later printed in Philosophia Christi.  The article uses his admission that the probability of a biological structure under an evolutionary chance hypothesis had not been calculated.

    1. 8a.

      Page 25:  “is an evolutionary chance hypothesis that takes into account Darwinian and other material mechanisms”.  “The precise calculation of P(T|H) has yet to be done.” ↩

  9. 9.

    Finnigan, Gregory C. et al.  “Evolution of Increased Complexity in a Molecular Machine.”  Nature 481:360–364.  Jan 19, 2012.  A laboratory reconstruction, by Joseph Thornton's group, of ancestral proteins of the V-ATPase proton pump.  It showed a two-part ring becoming a three-part ring in fungi after a gene duplication, with each copy losing an ability.  The article uses it as the one tested case of Muller's route, with the authors' own caution about how general it is.  Mirror:  PMC

    1. 9a.

      Section closing discussion:  “We are aware of no other mechanistic analyses of a molecular machine’s evolutionary trajectory, so the generality of our observations is unknown.” ↩

  10. 10.

    Rana, Fazale.  The Cell's Design: How Chemistry Reveals the Creator's Artistry.  Baker Books.  2008.  A book by the biochemist Fazale Rana of Reasons to Believe, an old-earth creationist ministry, arguing that the chemistry of the cell shows a Creator's hand.  The article quotes his case that repeated designs point to a Creator who reuses them.

    1. 10a.

      Page 204:  “If human craftsmen made in God’s image reuse the same techniques and technologies, it’s reasonable to infer that their Creator would do the same. So, if life stems from his hand, then it’s reasonable to expect the same designs to repeatedly appear throughout nature.” ↩

  11. 11.

    Li, Ying et al.  “The hearing gene Prestin unites echolocating bats and whales.”  Current Biology 20(2):R55–R56.  Jan 26, 2010.  A two-page comparison of the hearing gene prestin across echolocating bats, dolphins and other mammals.  Protein-changing sites group the dolphin with the bats, while silent sites give the ordinary tree.  The article uses it as the clearest case of convergence confined to working sites.  Mirror:  PMC

    1. 11a.

      Paragraph 4:  “Indeed, the same misplacement of dolphin is observed in the Prestin tree reconstructed with only nonsynonymous nucleotide substitutions (Fig. S1B); but, when only synonymous substitutions are used, dolphin and cow are correctly grouped with 100% bootstrap support (Fig. S1C).” ↩

  12. 12.

    Nowak, Martin A. et al.  “Evolution of genetic redundancy.”  Nature 388(6638):167–171.  1997.  A theoretical paper by Martin Nowak, John Maynard Smith and two colleagues on whether selection can keep redundant genes.  It states the problem that a perfectly redundant gene is not protected by selection, and models conditions under which partial redundancy is stable.  The article uses its framing, which grants the engineer's view.  Mirror:  Archive.org

    1. 12a.

      Page 167:  “Such a redundant genetic organization is sensible from an engineer's point of view: important functions require backup devices that can take over in case of failure. But can natural selection favour the emergence and stability of redundant genes?” ↩

    2. 12b.

      Page 167:  “Yet there is a problem: genes encoding functional proteins must be under selection pressure. If a gene was truly redundant then it would not be protected against the accumulation of deleterious mutations. A widespread view is therefore that such redundancy cannot be evolutionarily stable.” ↩

  13. 13.

    Terborg, Peer.  “Evidence for the design of life: part 1—Genetic redundancy.”  Journal of Creation 22(2):79–84.  Aug 2008.  A paper in a creationist journal arguing that redundant genes are an engineered feature of genomes that could not survive millions of years of mutation.  The article uses its timing claim because it yields a young-earth prediction that can be tested.

    1. 13a.

      Page 84:  “It is also important to realize that redundant genes cannot have resided in the genome for millions of years, because natural selection, a conservative force, cannot prevent their destruction due to debilitating mutations.” ↩

  14. 14.

    Darwin, Charles.  On the Origin of Species.  John Murray.  First edition.  1859.  The first edition of Darwin's book.  The article quotes his statement that natural selection does not produce absolute perfection, because it means his theory, like design, does not predict flawless organs.

    1. 14a.

      Page 202:  “Natural selection will not produce absolute perfection”. ↩

  15. 15.

    Darwin, Charles.  “Letter DCP-LETT-2814 to Asa Gray.”  Darwin Correspondence Project.  May 22, 1860.  A letter in which Darwin tells the botanist Asa Gray why he struggles to reconcile the living world with a beneficent designer.  In the same paragraph he says he did not intend to write atheistically.  The article quotes his example of parasitoid wasps, because it states the problem of harm done by biology that works well.

    1. 15a.

      Section Paragraph beginning With respect to the theological view of the question:  “the Ichneumonidæ with the express intention of their feeding within the living bodies of caterpillars”. ↩

  16. 16.

    Shipton, Warren A.  “Destructive parasites: expressions of God's creation?”  Journal of Creation 33(2):84–92.  2019.  A paper in a creationist journal proposing how destructive parasites arose from once harmless relationships after the Fall.  It allows gene loss, newly expressed front-loaded genes and other routes, and admits that the available detail is not enough for a conclusive account.  The article uses it as the young-earth model's own statement.

    1. 16a.

      Section Abstract, page 84:  “gene loss, the expression of front-loaded genes now facilitating pathological life cycles, or other scenarios.” ↩

  17. 17.

    Behe, Michael J.  “Reply to My Critics: A Response to Reviews of Darwin's Black Box: The Biochemical Challenge to Evolution.”  Biology and Philosophy 16(5):685–709.  2001.  Behe's reply in a philosophy of biology journal to the reviews of his 1996 book.  He revises his definition and admits that it tests the removal of parts, while evolution must assemble them.  The article uses that admission because the same limit applies to knockout evidence for redundancy.

    1. 17a.

      Page 695:  “The current definition puts the focus on removing a part from an already-functioning system.”  “Thus there is an asymmetry between my current definition of irreducible complexity and the task facing natural selection. I hope to repair this defect in future work.” ↩

  18. 18.

    Kafri, Ran et al.  “The regulatory utilization of genetic redundancy through responsive backup circuits.”  Proceedings of the National Academy of Sciences 103(31):11653–11658.  2006.  A study of yeast duplicate genes that are switched on when their partner is lost.  The authors argue that such backups are part of the cell's regulation, not leftovers of old duplications.  The article uses it against the claim that duplicates are debris.  Mirror:  PMC

    1. 18a.

      Section Abstract:  “many such backed-up genes were shown to be transcriptionally responsive to the intactness of their redundant partner and are up-regulated if the latter is mutationally inactivated. … We thus challenge the view that such redundancies are simply leftovers of ancient duplications and suggest they are an additional component to the sophisticated machinery of cellular regulation. In this respect, we suggest that compensation for gene loss is merely a side effect of sophisticated design principles using functional redundancy.” ↩

  19. 19.

    Holzmann, Gerard J.  “Mars Code.”  Communications of the ACM 57(2):64–73.  Feb 2014.  An account by Gerard Holzmann of NASA's Jet Propulsion Laboratory of how the Curiosity rover's software was written and checked.  The article uses his description of the landing computer's backup, which ran a different, simplified program.  Mirror:  Archive.org

    1. 19a.

      Page 73:  “Running the same landing software on two CPUs in parallel offers little protection against software defects. Two different versions of the entry-descent-and-landing code were therefore developed, with the version running on the backup CPU a simplified version of the primary version running on the main CPU.” ↩

  20. 20.

    Wagner, Andreas.  “Distributed robustness versus redundancy as causes of mutational robustness.”  BioEssays 27(2):176–188.  2005.  A review by Andreas Wagner comparing duplicate genes with network-wide robustness as reasons why mutations often do no harm.  The article uses his point that a redundant system and a non-redundant one can behave identically.  Mirror:  Archive.org

    1. 20a.

      Page 186:  “General theorems of control theory imply that for any system with redundant parts, one can design a system without redundancy but the same transfer function, i.e. the same behavior (ref. 95, Chapter 4). … Beyond such enumeration, however, it may be impossible to distinguish distributed robustness and redundancy, because the two have indistinguishable signatures on the variables necessary to understand a system's function and robustness.” ↩

  21. 21.

    ReMine, Walter James.  The Biotic Message: Evolution Versus Message Theory.  St. Paul Science.  1993.  Walter ReMine's book proposing that life was designed to look like the work of one designer and to resist every natural explanation.  The article quotes his claim that such a designer must sometimes use different designs for one function.

    1. 21a.

      Page 354:  “To successfully send the biotic message, a designer must not indiscriminately use the same design repeatedly. Therefore, the designer is forced to sometimes use different designs to perform the same function.” ↩