Common Descent

Common descent is the claim that all living things share ancestors, and that the differences between them accumulated along branching lineages.

The evidence is a pattern rather than any single observation: living things fall into groups within groups, with no overlap, and the same arrangement is recovered independently from anatomy and from molecular data that anatomists could not access.

The nested hierarchy#

Every organism belongs to exactly one group at each level of classification, and the groups nest cleanly.

A house cat is a felid, which is a carnivoran, which is a placental mammal, which is a vertebrate. There are no organisms that are half felid and half canid, and no trait distribution requiring an organism to sit in two branches at once.

Carl Linnaeus produced this classification a century before Darwin, while holding to special creation, because the pattern is in the specimens. Darwin's contribution was an explanation for why it exists: a branching process of descent produces nested groups automatically, since a lineage that acquires a trait passes it to all its descendants and to nothing else.

Designed objects do not form nested hierarchies. Wheels appear on cars, aircraft, wheelbarrows, and office chairs. Screens appear on phones, televisions, cameras, and refrigerators. A designer reuses whatever component suits the problem, which distributes features across categories freely.

Living things do not show that pattern. No mammal has feathers, no bird has mammary glands, no vertebrate has a chitinous exoskeleton, and no flowering plant has a notochord. The absence of these combinations is not required by function — bats fly without feathers, so feathers are not a requirement of flight.

Independent trees#

Separate bodies of evidence produce the same tree.

When molecular sequencing became available in the 1960s, it could have produced any tree at all. Cytochrome c is a protein a few hundred residues long, and nothing about its sequence is dictated by an organism's outward form. It reproduced the arrangement anatomists had drawn from bones and soft tissue.

The three-domain revision is the informative exception. Carl Woese's ribosomal RNA work in 1977 split what had been called bacteria into Bacteria and Archaea — a change to the tree that came from molecular data and was resisted before being accepted. The method was capable of overturning the existing classification, and in one case did.

Statistical strength#

The agreement is not approximate. Trees built from different genes converge on the same topology far more closely than chance permits, and the measure is quantifiable: the number of possible unrooted trees for even 20 taxa exceeds 10²⁰.

Independent datasets landing on the same tree out of that many possibilities is the core of the argument. Douglas Theobald set the calculation out in detail in the 29+ Evidences for Macroevolution FAQ.

Common objections#

"A common designer would produce similar designs"

The central alternative, and it accounts for functional similarity. Wings on birds and bats reflect the physics of flight, and shared solutions to shared problems are expected under either account.

It does not predict a nested hierarchy. A designer able to reuse any component would produce the pattern found in manufactured objects, where features cross categories freely. The systematic absence of combinations like feathered mammals requires explanation, and function does not supply it.

It also does not account for shared non-functional features — identical broken genes and identical viral insertion positions — which are covered in Pseudogenes and Endogenous Retroviruses.

"Molecular trees often conflict with each other"

They do, and the conflicts are real. Individual gene trees disagree at short internal branches, and some relationships remain unsettled — the position of turtles among reptiles was contested for years.

The disagreements are concentrated where the model predicts them: at rapid divergences, where lineages split faster than mutations accumulate to record the order, and where incomplete lineage sorting or horizontal transfer operates.

The comparison that matters is between the disagreements and the agreements. Trees conflict over the branching order within a group while agreeing that the group exists; no gene tree places whales among the reptiles or humans outside the primates.

"Horizontal gene transfer destroys the tree"

A substantive objection at the level of microbes. Bacteria and archaea exchange genes across lineages readily, and for those groups the metaphor of a tree is genuinely strained — a network is a better description.

W. Ford Doolittle has made this argument in the scientific literature, and it is taken seriously.

Its scope is limited. Horizontal transfer is rare in complex multicellular eukaryotes, where the germ line is sequestered, and the animal and plant trees are not affected in the way the microbial ones are.

It also does not support separate creation. A network of gene exchange is still a history of shared ancestry, and more of it rather than less.

"Similar organisms could just be similar without being related"

True of any single similarity taken alone, which is why the argument does not rest on one.

The pattern is what carries it. Similarity distributed at random, or according to habitat, or according to function, would not produce nested groups; it would produce overlapping ones. Marine mammals would group with fish, and burrowing animals with each other.

They do not. Dolphins group with mammals on every measure except body shape, and the exceptions of that kind are identifiable as convergence precisely because everything else disagrees with them.

What the evidence shows#

Living things fall into a nested hierarchy that was recognised before any evolutionary explanation existed and that manufactured objects do not exhibit.

The same hierarchy is recovered from anatomy, from protein sequences, from whole genomes, from viral insertion positions, and from the order in which groups appear in the fossil record. These are independent sources, and each was an opportunity for the pattern to fail.

Common design accounts for functional resemblance but does not predict the hierarchy or explain shared defects. Related evidence is collected in Special Creation, Pseudogenes, and Transitional Fossils.