Argument

Phylogenetic Incongruence Argument

Intro

Charles Darwin closed On the Origin of Species with one diagram, a branching tree. Every living thing, he proposed, traces back to a single trunk through a network of forking branches. The picture was simple, powerful, and testable. If everything alive descends from one common ancestor, then every gene in every creature should tell roughly the same family story. The tree built from one gene should match the tree built from another, and both should match the tree built from anatomy.

For most of the 20th century the tree picture was confirmed by morphology alone, bones, organs, embryos. Then came genome sequencing, and the data got messier. Trees built from one gene contradicted trees built from another. Mitochondrial DNA (inherited only from the mother) told different ancestry stories than nuclear DNA. Trees built from molecules contradicted trees built from fossils. Even within a single organism, different genes pointed to different ancestors.

Mainstream biology responded with rescues. Horizontal gene transfer (genes jumping between unrelated organisms, especially in bacteria). Incomplete lineage sorting (genes carrying their own histories that don't match the species history). Convergent molecular evolution (the same DNA sequence evolving independently in unrelated lineages). Each rescue is real and documented. Each also concedes that the original prediction, one gene one tree, has failed.

The argument here is not that no descent ever happened. It is that the universal singular tree-of-life thesis, the claim that every living thing traces to one ancestor through one branching pattern, has been undercut by the very genomic data that was supposed to confirm it. The honest reading is that common design (a single architect reusing components across taxa, plus limited descent at lower taxonomic levels) fits the data at least as well as universal common descent.

W. Ford Doolittle, a mainstream biologist, put it bluntly in Science (1999): "Maybe the tree of life is not a reality that can be discovered, only an artifact imposed by the methods used." Eugene Koonin (a Russian-American computational biologist, not an ID proponent) reached similar conclusions in The Logic of Chance (2011). When the field's own senior practitioners say the universal tree is in trouble, the design alternative gets a hearing.

In full

The Phylogenetic Incongruence Argument is an abductive (and partially reductio) critique of universal common descent that proceeds from a structural mismatch between the singular-tree prediction of the Darwinian thesis and the actually-observed pattern of incongruent gene trees across the genomic record. Universal common descent predicts that ancestry should be a single nested hierarchy testable by sampling any independently-inherited locus; the actual data shows widespread discordance between morphological and molecular trees, between mitochondrial and nuclear trees, between different nuclear genes, between different domains of life, and between fossil-based and sequence-based ancestral relationships. The mainstream response invokes horizontal gene transfer, incomplete lineage sorting, hybridization, and convergent molecular evolution; each rescue is empirically real but each shifts the explanans rather than vindicating the original prediction. The conclusion is that common design (a single designer reusing optimal components across taxa, combined with limited descent within baramin / kind boundaries) is at least as good an explanation as universal common descent, and on parsimony grounds may be better. 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, with a partial reductio component. P1 states the universal common descent prediction (singular tree). P2 states the observation that falsifies the simple version of that prediction. P3 examines the rescues and shows they preserve the framework only by ad hoc additions that concede the original prediction has failed. The argument does not claim universal common descent is logically refuted; it claims the inference from genomic data to universal common descent is no longer compelling, and competing inferences (common design, baraminology, limited descent within kinds) should be on the table. On standard inference-to-best-explanation criteria (explanatory power, scope, parsimony, ad-hoc-ness), the design alternative fares at least as well.

P1, Universal common descent predicts a singular nested-hierarchy tree of life

Affirmative case (second-order arguments)

1. Darwin's own framing made the singular-tree prediction explicit. Origin of Species (1859), final paragraph, "from so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved." The single tree with a single trunk is the iconic image of the book; it is the only diagram Darwin included. Universal common descent is the thesis that there is one root and one branching pattern, testable by independent traits.

2. The mainstream philosophical literature has identified nested hierarchy as the core evidence for common descent. Elliott Sober (Evidence and Evolution, Cambridge, 2008) and Sahotra Sarkar treat the convergent nested-hierarchy structure across independent character sets as the strongest probabilistic argument for descent over design. The argument's force depends on the convergence of independent trees on the same hierarchy.

3. Independent loci should converge if descent is true. Theobald's "29+ Evidences for Macroevolution" (TalkOrigins) explicitly grounds the case for common descent in the prediction that morphology and molecules, mitochondria and nucleus, conserved genes and selectively neutral sites, should all give the same tree. The prediction is testable; failure to converge counts against the thesis.

4. The singular-tree prediction has historic apologetic weight on the naturalist side. Francis Collins, Kenneth Miller, and most of BioLogos build their human-chimp common-ancestor case on shared genetic patterns including ERVs, pseudogenes, and synteny; these arguments presuppose that the molecular data confirms a singular tree. Removing the singular-tree prediction removes the foundation of the standard theistic-evolutionist argument as well.

Anticipated objections

1. "Universal common descent never predicted a strict singular tree; it predicted a web of life including horizontal exchange.", the Doolittle-Koonin "web" reframe.
2. "The nested hierarchy is statistically supported even with noise.", Theobald-style probabilistic defense.
3. "You're attacking a straw man, no contemporary biologist holds the strict singular-tree view."

Rebuttals

1. The "web of life" reframe is itself a concession. If the prediction was originally singular-tree and is now web-with-frequent-crossings, the original prediction has been weakened, not vindicated. The mainstream literature freely acknowledges this: Doolittle's 1999 Science paper is titled "Phylogenetic Classification and the Universal Tree" and explicitly asks whether the universal tree concept should be abandoned. The reframe is a retreat, not a defense. Failure mode: goalpost-shifting masquerading as theoretical refinement.

2. Statistical convergence still requires the underlying signal to dominate the noise. Theobald-style arguments assume the genuine phylogenetic signal is robust enough to recover the tree from noisy data. When the actual data shows different genes giving statistically significant conflicting trees (not just noise around a consensus), the statistical-convergence defense collapses. Casey Luskin (Discovery Institute) catalogs cases where the same dataset gives different trees under different methods, ML vs Bayesian vs parsimony. Failure mode: assuming what is to be proven (signal-dominance) when the data is what is contested.

3. Contemporary biologists explicitly defend the singular-tree thesis when it suits their argument. Francis Collins's case for human-chimp common ancestry depends on the universal-tree prediction at the human-ape branch; he does not invoke web-of-life reframings there. Pop-science presentations of evolution universally use the tree diagram. The "no one holds it strictly" defense is rhetorical; the moment a theist questions human-ape ancestry, the strict tree returns. Failure mode: heads-I-win-tails-you-lose framing of the prediction.

Live-cite kit

• Scripture: Genesis 1:20-25 (creatures "after their kind"); Genesis 1:11-12 (plants "after their kind"); Psalm 104:24 ("in wisdom you made them all").
• Scholarly: Charles Darwin (On the Origin of Species, 1859); Elliott Sober (Evidence and Evolution, Cambridge, 2008); Douglas Theobald ("29+ Evidences for Macroevolution"); Stephen C. Meyer (Darwin's Doubt, 2013, ch. 6); Casey Luskin (Discovery Institute essay series).
• Aphorism: "Darwin drew one tree. Modern genomics gives us a thicket."

Tactical notes

• Lead with Darwin's own diagram. It is the only illustration in Origin; the universal-tree prediction is not a strawman.
• Force the opponent to commit to either singular-tree or web-of-life. If singular-tree, P2 hits hard. If web-of-life, the original prediction is conceded as weakened.
• Don't get drawn into method-of-tree-building technicalities. The point is the prediction, not the methodology. The prediction is failing across methodologies.

P2, Actual genomic data shows widespread incongruence

Affirmative case (second-order arguments)

1. Morphological vs molecular tree disagreements are common and well-documented. The classical example: the placement of whales. Morphology placed them with mesonychids (extinct ungulates); molecular data placed them with hippos. Both can't be right; the molecular data won, but the morphology lost despite being the historical basis of the tree. Similar conflicts: bats vs primates in some early molecular analyses; the relationship of mammalian orders. Stephen C. Meyer (Darwin's Doubt, ch. 6) catalogs the genre.

2. Mitochondrial vs nuclear DNA give conflicting trees in many lineages. Mitochondrial DNA is inherited only through the maternal line; nuclear DNA reflects both parents. In sheep, cattle, and various rodents, mtDNA-based phylogenies disagree with nuclear-DNA-based phylogenies at significant branches. The mainstream explanation is hybridization or mitochondrial capture, but each invocation concedes that the simple inheritance picture fails.

3. Different nuclear genes give different trees, the "gene tree vs species tree" problem. Eugene Koonin (The Logic of Chance, Princeton, 2011) and W. Ford Doolittle (Science 284, 1999) document that across prokaryotes and even within eukaryotes, individual gene trees frequently disagree with the species tree assembled from the consensus. The disagreement is not noise; it is statistically significant and pervasive.

4. The base of the tree (LUCA and the three domains) is especially problematic. The relationship of bacteria, archaea, and eukaryotes has resisted resolution. The Woese three-domain tree was revised by the Eocyte hypothesis, then by various two-domain models. Doolittle's 1999 Science paper concludes the deep tree may not be recoverable in principle because horizontal gene transfer dominated early life. The root of the tree, the very claim of universal common descent, is the most contested branch.

5. Convergent molecular evolution produces the same sequences in unrelated lineages. Echolocation-related genes (Prestin) in bats and dolphins show convergent amino-acid changes; the convergence creates a fake "ancestry signal" between bats and dolphins on those genes specifically, even though no one believes bats and dolphins share a recent ancestor. The implication: molecular similarity does not always trace to shared ancestry.

Anticipated objections

1. "Incongruence is just noise; the consensus tree is robust.", statistical-defense move.
2. "Horizontal gene transfer fully explains the incongruence in prokaryotes; the eukaryote tree is solid.", restricted-rescue move.
3. "Cherry-picked examples; most genes give the same tree.", frequency-defense move.

Rebuttals

1. The literature acknowledges that incongruence is often statistically significant, not noise. Casey Luskin's review papers cite multiple cases where competing trees are statistically supported by their respective datasets. The "consensus tree is robust" defense works when noise dominates; when conflicting genes give significantly supported conflicting trees, the consensus is itself an averaging artifact, not a discovered truth. Doolittle's framing ("an artifact imposed by the methods") gets at this. Failure mode: assuming noise when the data shows signal-conflict.

2. The HGT-restricted-to-prokaryotes defense underestimates eukaryotic HGT and other incongruence sources. Recent literature documents HGT into eukaryotes (e.g., bdelloid rotifers, plant-to-fungus transfers). Even setting HGT aside, eukaryote phylogenies suffer from incomplete lineage sorting and hybridization at significant frequencies. The "eukaryote tree is solid" claim has weakened as more genomes have been sequenced. Failure mode: defending the tree by carving off the troubled parts.

3. The frequency defense undercounts the structural problem. Even if 80% of genes give "the" tree, the 20% that don't are not random noise; they often involve key transitions (the base of mammals, the placement of major groups). A theory that predicts singular tree and gets 80% there is in worse shape than a theory that predicts variability and gets 80% concordance. The prediction was universal; the data shows predominant but contested. Failure mode: converting a strict prediction into a statistical tendency post hoc.

Live-cite kit

• Scripture: Genesis 1:20-25 (kinds); Job 38 (God's creation challenge to Job).
• Scholarly: W. Ford Doolittle ("Phylogenetic Classification and the Universal Tree", Science 284, 1999); Eugene Koonin (The Logic of Chance, Princeton, 2011); Michael Syvanen ("Evolutionary Implications of Horizontal Gene Transfer", Annual Review of Genetics 46, 2012); Stephen C. Meyer (Darwin's Doubt, 2013, ch. 6); Casey Luskin (Discovery Institute summaries).
• Aphorism: "The molecules don't tell the same story as the bones, and different molecules don't tell the same story as each other."

Tactical notes

• Lead with the whale example. It is famous, mainstream-accepted, and concrete. Morphology and molecules disagreed; one was wrong; the field reorganized.
• Use the Doolittle quote. "An artifact imposed by the methods" is a senior mainstream biologist saying the tree may not be a real-world structure.
• Don't claim no genes agree. Many do. The argument is about the failure of universal agreement, not the absence of any agreement.

P3, Mainstream rescues explain incongruence ad hoc rather than resolving the predicted-vs-observed gap

Affirmative case (second-order arguments)

1. Horizontal gene transfer (HGT) was not predicted by the original Darwinian framework; it was added when the singular-tree picture failed. Carl Woese's discovery of HGT pervasiveness in prokaryotes forced the framework to accommodate it. The accommodation works (HGT is real), but it concedes that genes do not always track organismal ancestry. Once HGT is allowed, the singular-tree prediction becomes "the tree as recovered by the method that filters HGT out", which is circular: the tree is whatever the method recovers, and the method is calibrated to recover a tree. Failure mode: rescue-by-method-tuning.

2. Incomplete lineage sorting (ILS) is invoked when gene trees disagree with species trees. ILS says ancestral polymorphisms can persist through speciation events, so genes can sort independently of speciation. Real phenomenon, mathematically tractable. But the more ILS is invoked, the less the gene trees are evidence for the species tree; the species tree becomes an inference from multiple discordant gene trees rather than a tree confirmed by independent loci.

3. Convergent molecular evolution lets the same sequence arise twice. Documented (Prestin in bats and dolphins; antifreeze proteins in unrelated fish lineages). Each documented case is also a case where molecular similarity does not trace to common ancestry, undercutting the general inference from molecular similarity to descent.

4. Each rescue is real; the cumulative effect is to insulate the tree thesis from data. Common design proponents make this point: the tree is now defended by a battery of auxiliary hypotheses (HGT, ILS, convergence, hybridization, gene loss) each of which is invoked when the tree fails. The framework becomes effectively unfalsifiable. Casey Luskin, Stephen Meyer, and Cornelius Hunter develop this critique.

Anticipated objections

1. "HGT and ILS and convergence are real biological phenomena; invoking them is not ad hoc, it is science."
2. "Common design is unfalsifiable; you are critiquing rescues while having no testable alternative."
3. "The framework's accommodations are themselves predictions that have been confirmed."

Rebuttals

1. The phenomena are real; the use as rescues is the issue. A rescue is ad hoc when it is invoked specifically to save a failing prediction, not when it is independently motivated and predicted in advance. HGT was discovered (by Woese), not predicted by the singular-tree framework. ILS quantification arose to explain failure cases. The phenomena are real; the role they play as universal-tree rescues is what the argument identifies. Failure mode of the objection: conflating the reality of the phenomenon with the legitimacy of its rescue use.

2. Common design is testable in principle and in practice. Design predicts reuse of optimal components across unrelated lineages (which is what convergence looks like), modularity, hierarchical organization compatible with multiple top-down hierarchies, and discontinuity at higher taxonomic levels. Baraminology (creation biology) makes specific predictions about within-baramin descent and between-baramin discontinuity. Todd Wood (Understanding the Pattern of Life, 2006) develops this. The objection that design is unfalsifiable is dialectical, not technical. Failure mode: stipulating unfalsifiability rather than demonstrating it.

3. Predicted accommodations would be a strong defense if the accommodations had been predicted before the failures. They were not. HGT, ILS, convergence at the molecular level, mitochondrial capture, all were discovered as anomalies and then accommodated. The framework's flexibility is post hoc, not predictive. Failure mode: rewriting the historical sequence to make accommodations look predictive.

Live-cite kit

• Scripture: Romans 1:20 (the invisible attributes of God are clearly seen, being understood through what is made); Colossians 1:16-17 (all things created through Him and for Him, in Him all things hold together).
• Scholarly: Carl Woese (HGT discovery papers, PNAS 1990s); W. Ford Doolittle ("Phylogenetic Classification and the Universal Tree", Science 284, 1999); Eugene Koonin (The Logic of Chance, 2011); Michael Syvanen (Annual Review of Genetics 46, 2012); Todd Wood (Understanding the Pattern of Life, 2006); Cornelius Hunter (Darwin's God, 2001); Stephen C. Meyer (Darwin's Doubt, 2013).
• Aphorism: "When every contradiction gets a rescue, the theory has stopped being tested."

Tactical notes

• Use the historical sequence. Each rescue (HGT, ILS, convergence) was discovered as an anomaly, then accommodated. The accommodation does not retroactively make the framework predictive.
• Press for a testable failure condition. Ask the opponent: "What pattern of data would count against universal common descent? If every incongruence is rescued, what would falsify the thesis?"
• Don't dismiss the rescues as fake. They are real phenomena. The critique is their use as universal-tree insulation, not their existence.

Conclusion

Universal common descent is not well-supported by the genomic record; common design (with limited descent at lower taxonomic levels) better fits the data. Darwin predicted a singular tree of life testable by any independent locus. Genomic data has shown widespread incongruence between morphological and molecular trees, between mitochondrial and nuclear DNA, between different nuclear genes, and especially at the deep base of the tree. The mainstream rescues (horizontal gene transfer, incomplete lineage sorting, convergent molecular evolution, hybridization) are real but each shifts the explanans rather than vindicating the prediction. The cumulative effect is to insulate the tree thesis from data, leaving the framework effectively unfalsifiable. On standard inference-to-best-explanation criteria, common design (a single architect reusing components across taxa, with descent constrained to within-baramin variation) is at least as good an explanation and, on parsimony grounds, may be better.

Master objections to the argument as a whole

1. "Universal common descent is overwhelmingly supported by multiple independent lines (anatomy, fossils, biogeography, genetics).", Reply: most of these lines presuppose the singular-tree framework rather than independently confirm it. Where independent lines have been examined critically, conflicts have emerged. The cumulative case is weaker than its presentation in textbooks.

2. "You're misreading the experts; Doolittle and Koonin still affirm common descent.", Reply: they affirm a much-weakened form. Doolittle's web-of-life and Koonin's discussions of HGT-dominated early life are substantially weaker than Darwin's singular-tree prediction. The argument concedes that descent at some level is real; what it disputes is universal singular-tree common descent.

3. "Common design is religiously motivated and not science.", Reply: this is methodological-naturalism gatekeeping. (See Methodological Naturalism Critique.) The inference structure (best explanation of pattern in independent loci) is the same as the inference used to support common descent; switching to design changes the conclusion, not the structure.

4. "Even if the universal tree is wrong, common descent within mammals (or within vertebrates) is rock-solid.", Reply: the argument is most-effective against universal common descent; the limited-descent-within-kinds picture is compatible with the data and with the argument. ID-side defenders like Behe accept limited common descent; only YEC rejects it more broadly. The argument does not require rejecting all descent.

5. "This is just argument-from-controversy; science is full of unresolved questions.", Reply: the argument is not that controversy exists; it is that the controversy is at the predicted-vs-observed level, and the rescues are ad hoc. Unresolved questions are normal; failed predictions defended by ad hoc rescue are the issue.

Tactical opening / closing

Opening line: "Darwin drew one tree of life. Modern genome sequencing was supposed to confirm it gene by gene. Instead, different genes give different trees, mitochondrial DNA disagrees with nuclear DNA, and the molecules disagree with the bones. The senior mainstream biologist W. Ford Doolittle published in Science asking whether the universal tree was real or just an artifact of method. Let me walk you through why this matters for common-descent claims."

Closing landing strip: "The Phylogenetic Incongruence Argument does not say no descent ever happened. It says the universal singular-tree prediction has failed on the genomic data, and the rescues that defend it (horizontal gene transfer, incomplete lineage sorting, convergent molecular evolution) are ad hoc accommodations rather than predictive successes. The honest reading is that common design, a single architect reusing optimal components, is at least as good an explanation. The case for universal common descent is weaker than the textbook presentation admits."

Connection to Scripture

• Genesis 1:11-12, plants "after their kind"; vegetation reproduces within kind boundaries.
• Genesis 1:20-25, animals "after their kind"; sea creatures, birds, land animals each created within kind.
• Psalm 104:24, "How many are Your works, O Lord! In wisdom You have made them all"; diversity attributed to wisdom of the designer.
• Romans 1:20, the invisible attributes of God are clearly seen, being understood through what is made; design as evidence-source.
• Colossians 1:16-17, "by Him all things were created… in Him all things hold together"; Christ as designer-sustainer of all biological order.
• Job 38, God's challenge to Job; the diversity and architecture of creation as God's witness.

Patristic / scholarly note

Classical / patristic / medieval:

• Basil of Caesarea (Hexaemeron, c. 378), each kind reproduces according to its kind; against Greek-philosophical eternalism and unlimited transmutation.
• Augustine (De Genesi ad Litteram, c. 415), develops rationes seminales (seminal reasons); kinds were created in seed-form by God to unfold over time. Often misread as proto-evolutionary; the kinds themselves are God-given.
• Thomas Aquinas (Summa Theologica, I, Q. 73), the distinction of species as part of God's creative work in the six days.

Modern:

• Charles Darwin (On the Origin of Species, 1859), originator of the singular-tree thesis.
• Carl Woese (1928-2012, PNAS papers on HGT, 1990s), discovered the pervasiveness of horizontal gene transfer; reframed prokaryote phylogeny.
• W. Ford Doolittle ("Phylogenetic Classification and the Universal Tree", Science 284, 1999), the senior-mainstream-biologist tree-skeptic; "an artifact imposed by the methods."
• Eugene Koonin (The Logic of Chance, Princeton, 2011), Russian-American computational biologist; documents HGT and incongruence at the base of the tree.
• Michael Syvanen ("Evolutionary Implications of Horizontal Gene Transfer", Annual Review of Genetics 46, 2012), HGT critique of universal tree.
• Stephen Meyer (Darwin's Doubt, HarperOne, 2013, ch. 6), ID-side synthesis of the incongruence literature.
• Casey Luskin (Discovery Institute summary papers), accessible reviews of mainstream incongruence literature.
• Todd Wood (Understanding the Pattern of Life, 2006), creation-biology baraminology; develops the within-kind descent picture.
• Cornelius Hunter (Darwin's God, 2001), critiques the philosophical assumptions underlying common-descent inference.

Critics:

• Douglas Theobald ("29+ Evidences for Macroevolution"), defends the singular-tree probabilistic case.
• Elliott Sober (Evidence and Evolution, Cambridge, 2008), philosophical defense of nested hierarchy as evidence for descent.
• Jonathan Marks (What It Means to Be 98% Chimpanzee, 2002), anthropological-genetics nuance; cautions against over-reading molecular trees but defends common descent overall.

See also

• Common Design vs Common Descent Argument, sister argument on the same underdetermination.
• Convergent Evolution as Design Signal Argument, parallel critique using independent appearance of complex features.
• Orphan Genes Argument, related anomaly at the genetic level.
• Common Descent Critique, concept-side parent.
• Cambrian Explosion Argument, parallel anomaly at the paleontological level.
• Edge of Evolution Argument, rate-of-evolution complement.
• Methodological Naturalism Critique, why mainstream rejection is not decisive.
• Intelligent Design, framework that takes design hypothesis seriously.
• LUCA, the universal-common-ancestor concept this argument complicates.
• Endogenous Retroviruses, a separate locus where shared-ancestry inference is debated.
• Human Chromosome 2 Fusion, adjacent contested case in human-origins genetics.
• Stephen Meyer, primary modern defender of the design alternative.
• Origins, master hub.
• Arguments, top-level master index.