Special Creation
Separately created kinds cannot account for the genetic, fossil, and observed evidence for common ancestry.
6 min readUpdated
Special creation holds that God created organisms directly as separate kinds, which have varied within limits since but have not shared ancestry across those limits.
The claim is testable, because common ancestry and separate creation predict different patterns in anatomy, genetics, and geographic distribution.
What the text says
| Citation | Verse |
|---|---|
| Genesis 1:11-12 | Plants bearing seed "according to their own kinds" |
| Genesis 1:21 | Sea creatures and birds "according to their kinds" |
| Genesis 1:24-25 | Land animals "according to their kinds" |
| Genesis 2:7 | Man formed from dust, separately |
| Genesis 2:21-22 | Woman formed from the man's rib |
The Hebrew word is mîn, usually translated "kind." It is not defined in the text and does not correspond to any taxonomic rank.
The problem of defining a kind
Baraminology is the creationist discipline attempting to supply the definition, developed by Frank Marsh and continued by Kurt Wise, Todd Wood, and Jean Lightner.
| Difficulty | Detail |
|---|---|
| No consistent rank | Practitioners place kinds at family, genus, or order depending on the group |
| Boundaries move | Discovery of a hybrid merges two previously separate kinds |
| Ark capacity | Fewer kinds are needed to fit the ark, which pushes the boundary upward |
| Rapid diversification | Fewer kinds require faster post-flood speciation to produce current diversity |
The last two pull against each other. Answers in Genesis now proposes roughly 1,400 kinds on the ark, which requires that a single canid kind diversified into wolves, foxes, jackals, and domestic dogs within a few thousand years — a speciation rate substantially faster than any evolutionary biologist proposes.
Todd Wood, a young-earth creationist with a doctorate in biochemistry, has written that the evidence for evolution is substantial and that creationists should not claim otherwise. His position is that he rejects it on scriptural grounds rather than evidential ones, which is a clearer statement of the disagreement than most.
What the two models predict
| Observation | Common ancestry predicts | Separate creation predicts |
|---|---|---|
| Anatomical similarity | A nested hierarchy — groups within groups, with no crossing | Similarity by function; unrelated organisms may share good solutions |
| Molecular trees | The same nested hierarchy, from data anatomy could not access | No particular relation to the anatomical pattern |
| Shared broken genes | Present in exactly the descendants of the organism where the break occurred | No reason for a designer to distribute identical defects |
| Fossil order | Intermediates in specific strata, none outside their intervals | No required ordering |
| Geographic distribution | Species resemble their neighbours, not organisms in similar habitats elsewhere | Organisms suited to their environments wherever those occur |
Where the evidence is examined
Each prediction in the table above is tested in its own article:
- Common Descent— the nested hierarchy, and why anatomy and molecular data produce the same tree
- Pseudogenes— shared broken genes, including vitamin C and egg yolk protein
- Chromosome 2— fusion remnants at the predicted position
- Endogenous Retroviruses— shared viral insertion sites
- Transitional Fossils— intermediates in the predicted strata
- Biogeography— distribution tracking ancestry rather than habitat
- Vestigial Structures— reduced and repurposed inherited structures
- Speciation— whether a barrier to variation exists
Common objections
"Similar designs indicate a common designer, not common ancestry"
The central argument, made by Answers in Genesis and in most creationist textbooks. It has real force for functional similarity: wings on birds and bats reflect the physics of flight, and shared solutions to shared problems are expected under either model.
It does not extend to non-functional similarity. A broken vitamin C gene disabled by the same mutations at the same positions in humans and chimpanzees is not a design solution, and neither is a fused chromosome or a dead viral insertion.
The argument also does not predict a nested hierarchy. A designer free to reuse components would produce the pattern seen in manufactured objects, where features cross categories. Life does not show that pattern, and the absence requires explanation.
"Convergent evolution shows the nested hierarchy is not reliable"
Convergence is real and well documented — the eyes of vertebrates and octopuses, echolocation in bats and dolphins, and the body forms of dolphins and ichthyosaurs.
These cases are identified as convergent precisely because they conflict with the hierarchy established by everything else. Octopus and vertebrate eyes are built differently: the octopus retina has its nerves behind the photoreceptors, the vertebrate retina in front, producing a blind spot.
Convergence in a few traits against agreement across thousands is the situation the model predicts, since similar selection pressures recur. It would be a problem only if the conflicts were as common as the agreements.
"Pseudogenes have functions, so they are not broken"
Some sequences classified as pseudogenes do have regulatory functions, and this has been demonstrated in several cases.
For GULO specifically the function is known and lost: primates cannot synthesize vitamin C, which is why scurvy is a human disease and not a canine one. The gene's product is absent, and the specific mutations disabling it are shared.
Function would not resolve the pattern in any case. The evidence is not that the sequences are useless but that the identical damage occurs at identical positions in exactly the species that would have inherited it.
"Humans and animals are categorically different — we have language, reason, and morality"
The differences are real and substantial, and no biological account denies that human cognition is unusual.
Common ancestry is a claim about descent rather than about equivalence. Two species can share an ancestor and differ greatly, which is what a branching process produces.
Anatomically and genetically the pattern is not ambiguous. Humans fall inside the primate hierarchy on every measure — skeletal, molecular, developmental — and the shared errors above place us specifically with the African apes.
"The Cambrian explosion shows kinds appearing suddenly"
Most animal body plans do appear over roughly 20 to 25 million years in the early Cambrian, which is rapid in geological terms. Stephen Meyer's Darwin's Doubt (2013) builds the design case on this.
Twenty million years is not sudden in the sense the argument requires — it is roughly four times the entire history of the genus Homo.
The event is also not the appearance of everything at once. Precambrian Ediacaran fossils precede it, molecular clocks place the divergences earlier than the fossils, and the sequence within the Cambrian is ordered: sponges and cnidarians before bilaterians, simple arthropods before complex ones. Plants, vertebrates on land, and flowering plants appear hundreds of millions of years later.
What the evidence shows
The evidence forms a nested hierarchy, recovered independently from anatomy and from molecular data, and reinforced by shared non-functional errors that appear in exactly the species descent predicts.
Common design accounts for functional similarity. It does not account for identical broken genes, identical viral insertion sites, or a hierarchical pattern that manufactured objects do not display.
"Kind" has no consistent definition, and the versions that fit an ark require rates of diversification faster than mainstream biology proposes — which concedes the mechanism while disputing the timescale.