Concept

Common Descent vs Common Design

Intro

Humans and chimpanzees share a large majority of their DNA. Vertebrate forelimbs, a human arm, a bat wing, a whale flipper, a horse leg, are built on the same bone plan. The genetic code that translates DNA into protein is nearly universal across life. These are facts, and they are not in dispute.

The dispute is over what explains them. Two frameworks are on offer:

• Common descent says the similarity is inherited. Organisms resemble one another because they are modified descendants of shared ancestors, and the pattern of similarity is a family tree.
• Common design says the similarity reflects a shared engineering solution. A designer working across many organisms reuses what works, in the same way that a manufacturer reuses a proven chassis, a standard fastener, or a common operating system across an entire product line.

The point that decides how this argument goes is that similarity by itself does not distinguish between them. Both frameworks predict shared structures. Reuse is exactly what intelligent agents do, and it is the first thing anyone notices about designed systems: standard parts, common interfaces, modular components carried from one product to the next. So an argument that runs "these two organisms are similar, therefore they are related" has skipped the step that matters, because the same evidence sits equally well inside the other framework.

Two further points give the design side the edge on the specifics.

Reuse is a mark of good engineering, not of unimaginative design. A common genetic code across life is what a single coherent system looks like. If every organism ran incompatible code, the biosphere could not function as an interlocking whole: no shared nutrient cycles, no viral or bacterial exchange, no horizontal transfer, no digestion of one organism by another. Universality is a functional requirement, not a leftover.

The nested pattern is contested, not given. The strongest version of the descent argument is not bare similarity but the claim that similarity falls into a clean nested hierarchy of the kind branching descent would produce. That version has to contend with a large and growing body of data pulling the other way: convergent structures appearing on distant branches, and gene trees that conflict with one another and with the trees drawn from anatomy.

Quick reply line: "Shared parts do not settle it. Engineers reuse solutions constantly, and a shared genetic code is exactly what one coherent system needs in order to function as one. Similarity is compatible with both frameworks, so pointing at similarity is not an argument, it is the observation both sides are trying to explain."

◎ See this page's graph and discover linked topics →

In full

The dispute concerns the inference from homology, structural or genetic correspondence between organisms, to a specific causal history. Common descent treats homology as the signature of inheritance with modification from shared ancestry. Common design treats it as the signature of a common designer applying consistent solutions, with the correspondence reflecting shared function, shared constraints, and shared engineering logic.

The framing question is what the pattern of similarity, not its bare existence, is best explained by. See Intelligent Design and Evolution for the wider frameworks, and Irreducible Complexity for the complementary line of argument from system architecture.

Where the design inference is strongest

1. Convergence. The same sophisticated solution appears repeatedly on lineages that, on the descent account, cannot have inherited it from a common ancestor possessing it. The camera eye is the standard case, arising independently in vertebrates and in cephalopods, with the octopus version wired more sensibly than ours, no blind spot from nerves crossing in front of the retina. Echolocation appears in bats and in toothed whales, and the genetic work of Liu and colleagues (2010) found convergence in the Prestin hearing gene itself, not merely in the anatomy. Powered flight arose in insects, pterosaurs, birds and bats.

On the design framework this is expected and unremarkable: the same problem, solved the same way, because that is the solution that works. On the descent framework each case requires independent origination of a complex integrated system, invoked as many times as the data demands.

2. Conflicting gene trees. Trees built from different genes routinely disagree with one another and with morphology. This is not a fringe observation. It is discussed openly in the mainstream literature under the heading of incongruence, and it drove W. Ford Doolittle's proposal that the tree metaphor itself may not hold at the base. A pattern that requires this much correction is not delivering the clean signal the argument from nested hierarchy needs.

3. Function where the argument predicted junk. For decades, non-coding DNA was presented as wreckage from an undirected history and used as evidence against design. The ENCODE project (2012) reported biochemical activity across the large majority of the genome, and function has continued to be found in regions previously written off, including regulatory roles for repetitive elements. The design expectation, that apparent waste in a sophisticated system usually means unrecognised function, has been the better predictor. See Endogenous Retroviruses for the parallel case where sequences once treated as pure genomic litter turn out to carry essential functions, including in placental development.

4. The engineering logic of the code. The genetic code is not an arbitrary assignment. Its structure buffers against point mutations, since chemically similar amino acids occupy neighbouring codons, so many single-letter errors are silent or conservative. That is error-correction built into the mapping, which is what a designed information system looks like.

Where the descent argument is pressed hardest

Presented fairly and answered.

Shared errors. The strongest form of the descent case is not shared function but shared apparent mistakes: the same broken vitamin C gene in humans and other primates, retroviral insertions at matching genomic addresses, pseudogenes. Copying an error is hard to explain by independent design, and this is the argument to expect from a well-prepared interlocutor.

The reply turns on the premise that these are errors at all, and that premise keeps failing. Sequences classified as broken have repeatedly been found functional, which is the ENCODE pattern above and precisely what has happened with endogenous retroviral elements now known to be essential in placental development and in immune regulation. A shared feature that turns out to be functional is shared design, not shared damage. Further, insertion sites are not random: retroviral integration favours specific chromatin contexts and sequence motifs, so matching locations are expected on functional grounds rather than requiring inheritance. The argument depends on knowing that a sequence is useless, and the history of that judgment is a history of premature verdicts.

Nested hierarchy. That similarities sort into a branching pattern rather than an arbitrary one. Answered by point 2 above: the pattern is considerably messier than the argument requires, and where it is clean, a designer working from consistent principles and shared functional constraints produces grouped similarity as a matter of course, which is why human-designed artefacts also classify hierarchically.

Chromosome 2. The human chromosome bearing internal telomeric sequence and a degenerate second centromere, presented as a fused pair. The site does not match what an ancestral fusion should have left, being far shorter than expected and sitting inside a functional gene region with active transcription, which is a poor fit for a scar left by an accidental end-to-end join.

Why this matters apologetically

The argument is often conceded too early, on the assumption that shared DNA settles the question. It does not, and the reason is a point of logic rather than biology: a piece of evidence supports one hypothesis over another only if the two predict different things, and on similarity as such they do not.

This is also where the discussion should be kept. Similarity is the objector's opening move, and answering it well moves the conversation to the questions where the design inference is strongest: the origin of the information in the first place, the assembly of integrated systems that require all their parts, and the engineering of the code itself. See Irreducible Complexity and Intelligent Design.

Note that the two frameworks are not exhaustive of Christian positions. Some Christians accept common descent as God's method, which is treated at Theistic Evolution; the design inference engaged here is compatible with a range of views on the age of the earth.

See also

• Intelligent Design, the design framework
• Irreducible Complexity, the argument from integrated systems
• Evolution, the framework being compared
• Theistic Evolution, the mediating Christian position
• Endogenous Retroviruses, the shared-insertion evidence and the function findings
• Francis Collins, the principal advocate of the shared-error argument
• Michael Behe, the principal advocate of the design inference from system architecture
• Origins and Science, the domain hub