Homo erectus

"morphological diversity in the African fossil Homo record around 1.8 Ma probably reflects variation between demes of a single evolving lineage, which is appropriately named H. erectus." — David Lordkipanidze and colleagues, on the fifth Dmanisi skull, 2013¹

"the predicted age of dental maturity (around 12.0–13.5 years) is more similar to chimpanzees and other great apes (11–13 years) than to modern humans (18–22 years)." — Christoph Zollikofer and colleagues, on the Dmanisi subadult, 2024²

"Any model of post-Babel distribution must account for the appearance and disappearance of these human species in stratigraphic order." — Todd Wood, 2010³

Summary

Homo erectus is the name given to fossils from Georgia, Africa, Java and China, spanning about 1.7 million years on the standard timescale.  The young-earth reading, held by Marvin Lubenow, Todd Wood and Peter Line, is that these were ordinary people:  a post-Babel population that spread early, lived hard, and was buried first.  Tested against the specimens, the dates, and the behaviour, that reading is half right.

Where the young-earth reading is strong.

1. The behaviour is human behaviour. Fire burned thirty metres inside Wonderwerk Cave, taking bone above 400 °C.⁴⁴ A rhinoceros was butchered on Luzon, across a sea barrier.⁵ A Trinil shell carries a geometric engraving.⁶ A Dmanisi man lived for years with one tooth,⁷ and the excavators infer that others fed him.²
2. The excavators of Dmanisi argue the creationist case about variation. Lordkipanidze's team say that the spread among their five skulls matches the spread inside a single chimpanzee population, and that the African fossils called habilis, rudolfensis and ergaster are probably one lineage.¹ That claim was challenged in Science within six months and is not settled.⁸
3. The body plan is human-like. The Skull 5 individual's stature, mass and limb proportions reach the low end of the modern human range.¹

Where it fails, and where no creationist source has answered.

4. The growth schedule is not human. A synchrotron count of daily growth lines in the Dmanisi subadult's teeth puts dental maturity at 12.0 to 13.5 years. Living humans reach it at 18 to 22. The molars erupted at about 3.8, 7.6 and 10.6 years, the ages seen in wild chimpanzees.² Christopher Dean found the same in 2001 from enamel growth rates,⁹ and Ronda Graves in 2010 from the skeleton.¹⁰ Peter Line's model for the small Dmanisi brains is cretinism, and his wider model for robust humans is slower development linked to long lifespans. The measurement runs the other way.
5. The brains are small, and the movement's own specialist says so. Line's survey of 48 crania gives 546 to 1,260 cc with a mean of 934 cc, about 70% of the modern mean.¹¹ He concedes that pathology cannot explain most of them and that the cause is unknown.¹¹ The smallest adult skull, Dmanisi Skull 5 at 546 cc, has an encephalization quotient inside the Australopithecus range.¹
6. Seven dating methods agree across three continents. Dmanisi sits between an argon-argon date on the lava beneath and a magnetic reversal above.¹² Zhoukoudian is dated by the ratio of two cosmogenic nuclides made together in quartz, so the date needs no assumption about how much daughter was present at the start.¹³ No creationist source found attempts the arithmetic for any erectus site.
7. The young-earth camp holds four incompatible readings of one skull. On Skull 5: normal human variation (Lubenow); a human with cretinism, with normal variation judged "less likely" (Line, 2013); a human whose small brain has no known cause (Line, 2020); and possibly an australopith wrongly filed as erectus (Rupe and Sanford). Wood says its status must "await further analysis."³
8. Two creationist claims fail against their own citations. Jeffrey Tomkins dates the Salkhit skullcap by a paper that placed its mitochondrial DNA in a modern human haplogroup, while keeping the archaic description that paper overturned.¹⁴ Rupe and Sanford print an erectus range of 700 to 1,400 cm³ in a book whose own endnote cites a comment on the paper that published the 546 cm³ skull.¹⁵
9. Wood names the hardest problem. Habilis lies lowest, African only; erectus next, across the Old World; sapiens highest, global. Wood writes that any post-Babel model must account for this order.³ None found does.

Verdict.  Homo erectus behaved like people and did not grow up like people.  The "fully human population" reading survives the tools, the fire, and the sea crossings, and fails on a measured developmental schedule and on the dates.  The "ape" reading, held by Jack Cuozzo, fails on everything in Part 3.  What is left for a young-earth model is Wood's position, humans of a different kind, with the stratigraphic order and the growth schedule unexplained.

Part 1: What the name covers

Eugène Dubois dug the type specimen, a skullcap, from the main bone layer at Trinil on Java in 1891.  Since then the name has been stretched over fossils from Kenya, Tanzania, Ethiopia, Eritrea, Georgia, Java, and China.  Susan Antón's 2003 review is the standard specimen-by-specimen survey.  She describes the vault as having "moderately sized cranial capacities, ranging from about 700 cc (in East Africa and Georgia) to over 1,200 cc (in China and Indonesia), with some gradual increase in average size through time."¹⁶  Ten years later the range got wider at the bottom.  Skull 5 from Dmanisi, a complete adult cranium joined to a mandible found five years earlier, measured 546 ± 5 cm³ by CT.¹

The table gives the specimens that carry the argument.  Cranial capacities are from Antón's Table 3 unless a later measurement replaced them.  Dates are from the papers named in the method column.

Antón's dates for the two Zhoukoudian skulls are the ones compiled in 2003.  The lower cultural layers 7 to 10 at the same site were later dated to 0.77 ± 0.08 Ma by cosmogenic burial dating,¹³ which is older than either;  neither paper on this article's file says which layer each calvaria came from, so the table keeps the compiled figures.

The span is 546 to 1,251 cc, a factor of 2.3.  Peter Line's own survey of 48 crania gives 546 to 1,260 cc with a mean of about 934 cc.¹¹  Philip Rightmire's mainstream compilation of 30 individuals gives a mean of 973 cm³.¹⁹  The creationist and mainstream compilations agree on the numbers.  The disagreement is about what a 2.3-fold spread means.

Two things about the spread are not noise.  First, the small values sit at the base:  the Dmanisi crania at 546 to 780 cc and the earliest Kenyan crania at 804 to 848 cc are all at 1.5 to 1.8 Ma, and the largest, at Ngandong and Zhoukoudian, are the youngest.  One skull now tests that order.  Yunxian 2 in central China, about 1,143 cc after digital reconstruction,²⁰ was redated in February 2026 from about one million years to about 1.77 million,²¹ which would put a brain near the top of the erectus range beside Dmanisi in time.  Antón's time-series analysis of the Asian material "supports a significant but gradual trend for increasing brain size through time in Asian H. erectus, at the rate of about 160 cc per million years."¹⁶  Second, the vault is not modern.  Antón finds that "The vault of African and Asian H. erectus is metrically differentiable from geographically diverse samples of H. sapiens (including the enigmatic Willandra Lakes 50 from Australia) in simple bivariate and principal components analyses."¹⁶  Antón also excludes Kabwe, Bodo, Ndutu, Ceprano, Dali and the Gran Dolina material from erectus, so any argument that erectus grades smoothly into modern humans has to say which specimens it is counting.

The body plan

The body is the genuinely human-like part, and it should be conceded plainly.  The Skull 5 individual's postcrania give a stature of 146 to 166 cm and a body mass of 47 to 50 kg.  Lordkipanidze's team write that Skull 5 "provides the first evidence that early Homo comprised adult individuals with small brains but body mass, stature, and limb proportions reaching the lower range limit of modern human variation."¹  Alan Walker's team, extending their method to five other East African individuals, "found that Homo erectus in Africa had an average stature of five feet seven inches (ranging from five feet two inches to six feet one inch) and an average weight of 128 pounds."²²

The brain does not match the body.  Lordkipanidze's team compute the Skull 5 encephalization quotient "as ~2.4, which is within the range of variation for Australopithecus."¹  Zollikofer's 2024 paper puts it the same way:  "the brain size of early Homo at Dmanisi (ECV range: 546–730 cm3) was only moderately larger than that of great apes."²

One species or several

The Dmanisi excavators argue that their five skulls come from one population, and that the variation among them is large enough to swallow the African species names.  "Evidence from skull 5 and the other four Dmanisi specimens indicates that cranial shape variation within early Homo paleodemes was similar in mode and range to that seen within modern Pan demes."¹  From that they conclude that "morphological diversity in the African fossil Homo record around 1.8 Ma probably reflects variation between demes of a single evolving lineage, which is appropriately named H. erectus."¹  They add that such variation "tends to be misinterpreted as species diversity, especially when single fossil specimens from different localities are compared with each other."¹  The same paper withdraws the species name Homo georgicus, given earlier to the Skull 5 mandible, and demotes it to a population name.

Creationist writers quote this as if the field had conceded the point.  It did not.  Jeffrey Schwartz, Ian Tattersall and Zhang Chi answered in Science in 2014 that "dental, mandibular, and cranial morphologies all suggest taxic diversity and, in particular, validate the previously named H. georgicus."⁸  Lordkipanidze's team replied that the critics' observations were "partly incorrect, and not calibrated against intraspecific variation in living taxa," and that after adjustment the data "are fully compatible with the hypothesis of a single paleodeme of early Homo at Dmanisi."⁸  Two things follow.  The lumping claim is live and defended, so a creationist may use it.  And the comparison Lordkipanidze makes is with chimpanzee populations, not human ones.  A creationist who accepts the lumping inherits a population whose largest brain, 730 cm³, sits below the 800 cm³ floor that Rupe and Sanford themselves give for normal humans.¹⁵

Part 2: How the sites are dated

A young-earth reading needs every erectus site to be a few thousand years old.  The sites are dated by different physical methods, and the methods agree.  The list below is the arithmetic a creationist model has to break.

Dmanisi.  The fossils lie in stratum B, on top of the Masavera Basalt.  The basalt is dated by ⁴⁰Ar/³⁹Ar to 1.85 Ma.  Palaeomagnetic sampling through six metres of conformable ash and soil finds the lower stratum normally magnetised and the upper reversed, which places the Olduvai/Matuyama reversal in the section.  That reversal is independently dated at 1.78 Ma.  Reid Ferring's team write:  "Here, the magnetostratigraphy and 40Ar/39Ar dating of the Masavera Basalt (4) constrain the age of stratum A deposits and the recently discovered artifact and faunal assemblages between 1.85 Ma and 1.78 Ma, the age of the Olduvai/Upper Matuyama reversal (14)."¹²  So the site has a radiometric floor, a magnetic ceiling, and a biostratigraphic cross-check, and they agree.  The section also records repeated visits:  "It is now clear that Dmanisi was occupied repeatedly over an interval of as much as 80 ka, strongly suggesting a sustained regional population."¹²  A post-Babel model has to fit 80,000 years of occupation, six metres of soil with carbonate horizons, and a full magnetic reversal into about four thousand years.  No published creationist source found for this article attempts that for Dmanisi.

Zhoukoudian.  The Peking Man site was dated in 2009 by cosmogenic ²⁶Al/¹⁰Be burial dating of quartz.  Guanjun Shen's team report that "The weighted mean of six meaningful age measurements, 0.77 ± 0.08 million years (Myr, mean ± s.e.m.), provides the best age estimate for lower cultural layers 7–10."¹³  It was "the first radioisotopic dating of any early hominin site in China beyond the range of mass spectrometric U-series dating."¹³  This method matters for a young-earth argument, but not in the way it first looks.  Burial dating is not immune to accelerated decay.  A uniform speed-up of every decay constant by a factor k shrinks a burial age by exactly the same 1/k as it shrinks a potassium-argon age, because the burial clock runs on the difference between two decay constants and that difference scales with k too.  What burial dating removes is a different objection.  The two nuclides are made together in the same quartz at a fixed ratio, so no assumption about starting daughter is needed, and the date is measured on the sediment holding the artefacts.  A young-earth model still owes an account of why the burial age, the magnetic stratigraphy and the loess correlation land on the same interval.  No creationist source in this article's file has taken it up.

Java.  Carl Swisher's 1994 argon-argon work on hornblende from pumice at two sites "yielded well-defined plateaus with weighted mean ages of 1.81 +/- 0.04 and 1.66 +/- 0.04 million years ago (Ma)."¹⁷  Those dates were challenged.  Langbroek and Roebroeks argued in 2000 that the whole Sangiran sequence is under 1.0 Ma.  Antón answers that the critique "relies on a single fission-track age on a tektite from Sangiran of uncertain provenience and on previous paleomagnetic data that suggested normal polarities ... but that did not apply thermal demagnitization techniques, and therefore likely measured modern (normal) overprints in the significantly weathered Sangiran sediments."¹⁶  The Trinil bone layer itself was re-dated in 2015 by argon-argon and luminescence on sediment packed inside fossil shells, "obtaining a maximum age of 0.54 ± 0.10 million years and a minimum age of 0.43 ± 0.05 million years."⁶

Ngandong.  This is the site a skeptic will raise, because its date has moved.  Carl Swisher's team published mean ages of 27 ± 2 to 53.3 ± 4 ka in 1996, on electron spin resonance and uranium-series dating of bovid teeth.²³  Antón, writing in 2003, allowed that "It remains possible that a systematic error in the U-series/ESR techniques on tooth and bone would prove these dates inaccurate, despite high reproducibility."¹⁶  It did.  Yan Rizal's team re-excavated the site and built a Bayesian model on 52 radiometric ages from four methods.  "Non-human fossils recovered during the re-excavation of Ngandong date to between 109 and 106 ka (uranium-series minimum) and 134 and 118 ka (US–ESR), with modelled ages of 117 to 108 thousand years (kyr) for the H. erectus bone bed, which accumulated during flood conditions."¹⁸  The abstract says the relationship between fossils, terrace and ages "have been heavily debated,"¹⁸ and that the new results "negate the extreme ages that have been proposed for the site and solidify Ngandong as the last known occurrence of this long-lived species."¹⁸  Creationists cited the 27 to 53 ka figure for decades as proof that erectus overlapped modern humans.  The field has abandoned it.

Wonderwerk, Luzon, and Flores.  The fire layer at Wonderwerk "shows a normal magnetic orientation and is bracketed between two cosmogenic burial ages of 1.27 ± 0.19 Ma and 0.98 ± 0.19 Ma."⁴  On Luzon the rhinoceros tooth was dated by combined ESR and uranium-series:  "The tooth yielded an age of 709 ± 68 ka (1 confidence interval), which is in agreement with the ESR results on quartz."⁵  On Flores, Adam Brumm's team "show using 40Ar/39Ar dating that an ignimbrite overlying the artefact layers at Wolo Sege was erupted 1.02 ± 0.02 Myr ago, providing a new minimum age for hominins on Flores."²⁴

Seven methods carry the dates above:  argon-argon, palaeomagnetic stratigraphy, ²⁶Al/¹⁰Be burial dating, luminescence, uranium-series, electron spin resonance, and faunal biostratigraphy.  Where two or more are applied to the same deposit they agree.

The scale of the disputes.  Every one of these dates has been argued over inside the field.  Ngandong moved from 27 ka to 117 ka.  Java's early end was disputed between under 1.0 Ma and 1.81 Ma.  Ferring notes that the earliest African erectus is "considered to be ca. 1.78 Ma (21) but probably is closer to 1.65 Ma (22)."¹²  The two ends of each fight are separated by factors of two to four, and even the youngest end, Ngandong's 27 ka, is about six times older than a post-Babel date of about 4,300 years;  the oldest, 1.81 Ma, is about 420 times older.  "Even the scientists disagree" is true and does not help.

Part 3: What erectus did

This is where the young-earth reading is strongest, and the mainstream's own excavators supply the evidence.

Tools.  Dmanisi's 73 in situ artefacts are all Mode I, simple flakes and cores with no handaxes.¹²  The Acheulean handaxe, shaped to a template on both faces, appears later in Africa.  Swisher's 1994 dates gave the standard explanation for why East Asia lacks them:  "If the ancestor of Homo erectus ventured out of Africa before 1.8 Ma, the dispersal would have predated the advent of the Acheulean culture at 1.4 Ma, possibly explaining the absence of these characteristic stone cleavers and hand axes in East Asia."¹⁷  The Gona crania published in 2020 pair a very small cranium of about 598 cc with Acheulean bifaces, and Peter Line argues that purposeful shaping to a design "has the hallmarks of human activity."¹¹  That argument is fair.

Fire.  Wonderwerk Cave in South Africa is the strongest early-fire claim because of its method.  Francesco Berna's team used sediment micromorphology on intact resin-impregnated blocks and infrared microspectroscopy on thin sections.  They report "unambiguous evidence—in the form of burned bone and ashed plant remains—that burning took place in the cave during the early Acheulean occupation, approximately 1.0 Ma."⁴  The infrared bands put the burned bone above 400 °C and below 700 °C, with the figured fragment between 400 and 550 °C, thirty metres inside the cave.⁴  Burning runs through the whole thickness of the layer, which rules out a passing wildfire.  The authors are careful about what they did not find:  "our data, although they do not show evidence of constructed combustion features, as listed by Roebroek and Villa as a criterion of controlled burning (3), demonstrate a very close association between hominin occupation and the presence of fire deep inside Wonderwerk Cave during the Early Acheulean."⁴  At Gesher Benot Ya'aqov in Israel, burned seeds, wood and flint at about 790 ka are "suggestive of the control of fire by humans,"²⁵ and the burned flint fragments cluster in spots "possibly indicating hearth locations."²⁵  Two sites, two continents, two methods.  Neither has a built hearth, and Wil Roebroeks and Paola Villa argue from the European record that fire became "a significant part of the hominin technological repertoire" only "from ∼300,000 to 400,000 y ago onward."²⁶  A fair reading is that fire was used at erectus grade and that how routinely is disputed.

Sea crossings.  Flores has been separated from Java by deep water throughout the Pleistocene, and tools lie on Flores at 1.02 Ma.²⁴  No fossil is associated with those tools, so the identification with erectus rests on who else was in the region.  Luzon is the better-documented case.  At Kalinga, Thomas Ingicco's team excavated "57 stone-tools associated with an almost complete disarticulated skeleton of Rhinoceros philippinensis showing clear signs of butchery."⁵  Thirteen bones carry cut marks, both humeri carry percussion marks, and the left humerus was smashed into five pieces.  The authors write:  "Although the identity of these archaic toolmakers remains unknown, it is likely that they dispersed over at least one sea barrier to reach Luzon Island."⁵  Note what they do not claim.  They do not name erectus, and they do not say boats.  Citing this as "erectus built boats" overstates the paper.

The engraved shell.  In Dubois's own Trinil collection at Leiden, Josephine Joordens's team "found evidence for freshwater shellfish consumption by hominins, one unambiguous shell tool, and a shell with a geometric engraving."⁶  They conclude that "engraving abstract patterns was in the realm of Asian Homo erectus cognition and neuromotor control,"⁶ while declining to say what the pattern meant.  Symbolic behaviour is the trait most often used to separate modern humans from everything else, so this is the most useful find for the creationist reading.  It is also the thinnest:  a single shell, found in the collection Dubois dug at Trinil in 1891 and recognised in the Naturalis museum in Leiden more than a century later, so its context cannot be re-excavated.

The toothless man.  Dmanisi skull D3444 with mandible D3900 belonged to an individual "who had lost all but one tooth several years before death."⁷  The 2005 paper calls it "the earliest case of severe masticatory impairment in the hominin fossil record to be discovered so far" and says it "raises questions about alternative subsistence strategies in early Homo."⁷  In 2024 the same team went further:  "The four mature individuals found at Dmanisi span an age range from early to late adulthood, including one senile individual who had survived several years without teeth.  The possible presence of post-reproductive individuals in the Dmanisi palaeopopulation can be viewed in the context of biocultural reproduction, which implies that intergenerational co-operation facilitated an extended period of growth sustained by alloparental provisioning."²  That is the excavators inferring that the group fed him.  The limits:  the 2005 paper offers soft foods without help as an alternative, and chimpanzees have survived major disability.  What D3444 shows for certain is a social species.

None of this is ape behaviour.

Part 4: How erectus grew up

This is where the young-earth reading fails, and it fails on a measurement rather than on an impression of a skull.

How the tall Turkana Boy was built.  The figure everyone repeats, that Turkana Boy would have stood six feet one inch, came from Alan Walker's team in the 1980s.  Walker describes the method in his own words:  "As a starting point we assumed the boy was age eleven or twelve, Holly's best guess based on human standards.  The estimate of adult height we arrived at was a staggering six feet one inch, with a possible height of up to six feet ten inches if he had yet to go into an adolescent growth spurt."²²  The growth curve came from living Dinka and related East Africans.²²  Two assumptions sit inside that number:  an age read off human dental standards, and a modern human adolescent spurt.  Lubenow reports the result faithfully:  "Designated KNM-WT 15000, it is thought to be that of a twelve- to thirteen-year-old boy.  Yet he was 5'4" to 5'6" tall.  Had he grown to maturity, it is estimated he could have been six feet tall."²⁷  Both assumptions have since been tested.

Enamel, 2001.  Christopher Dean and colleagues, Walker among them, counted daily growth lines in the enamel of thirteen fossil hominins, including Turkana Boy and Sangiran S7-37 from Java.  "Neither australopiths nor fossils currently attributed to early Homo shared the slow trajectory of enamel growth typical of modern humans;  rather, both resembled modern and fossil African apes."⁹  "It therefore seems likely that truly modern dental development emerged relatively late in human evolution."⁹

The skeleton, 2010.  Ronda Graves's team built growth curves between the human and chimpanzee patterns and asked which one fits the life-history events the boy had actually passed.  "Prior research on KNM-WT 15000 operates under the assumption that H. erectus experienced a modern human life history, including an adolescent growth spurt.  However, recent fossil discoveries, improvements in research methods, and new insights into modern human ontogeny suggest that this may not have been the case."¹⁰  On the best-fitting curve, growth was finished by 12.3 years, and "it is likely that KNM-WT 15000 would have only attained an adult stature of 163 cm (~5'4"), not 185 cm (~6'1") as previously reported."¹⁰  The six-foot Turkana Boy is withdrawn by the mainstream that produced him.

The synchrotron, 2024.  The decisive result is the newest.  Zollikofer's team imaged the teeth of the Dmanisi subadult, cranium D2700 with mandible D2735, at the European Synchrotron Radiation Facility, at voxel sizes down to 0.72 µm, and counted the daily lines in dentin and enamel.  "The individual died at the age of 11.4 ± 0.6 years, shortly before reaching dental maturity."²  They find that "the predicted age of dental maturity (around 12.0–13.5 years) is more similar to chimpanzees and other great apes (11–13 years) than to modern humans (18–22 years),"² with "the estimated eruption times of the molars (M1: around 3.8 years;  M2: around 7.6 years;  M3: around 10.6 years) being more similar to those observed in wild chimpanzees than in humans."²  Nariokotome plots with Dmanisi:  "the advanced dental and skeletal maturity of Dmanisi and Nariokotome at their respective ages of death are more similar to those of great apes than living humans."²

Read carefully what is human-like and what is not, because writers on both sides have quoted one half and dropped the other.  The pattern of maturation, which teeth mature in which order, falls inside the human range and outside the great-ape range.  The rate, how fast the whole dentition matures, is ape-like.  The paper's own summary is that "the first evolutionary steps towards an extended growth phase occurred in the genus Homo at least 1.77 million years ago, before any substantial increase in brain size."²

Three methods, three sets of specimens, twenty-three years, one answer.  Living humans reach dental maturity at 18 to 22 years.  This child was on course for 12.0 to 13.5, with molars at 3.8, 7.6 and 10.6.  Any account that calls erectus a human population has to explain a childhood five to nine years shorter than any population ever measured.  The one creationist mechanism on offer, cretinism, predicts the reverse:  congenital hypothyroidism delays dental development.  No creationist response to Zollikofer 2024 or to Dean 2001 was found on creation.com, icr.org, or in Answers Research Journal.  That is a real gap and it should be stated as "none was found," not "none exists."

Part 5: The young-earth readings

The creationist literature on erectus is not one position.  Todd Wood surveyed it in 2010:  "Homo erectus (sensu lato, including H. ergaster) is viewed as human by Hartwig-Scherer (1998), Lubenow (2004, chap. 12), and Wise (2005).  In contrast, Gish (1995, pp. 304–305) and Bowden (1981, pp. 208–210) view H. erectus as a mix of ape and human specimens, and Cuozzo (1998, p. 101) labeled H. erectus an ape."³  He reports that "among the most knowledgeable creationist scholars, there is a general consensus that the human holobaramin includes Homo sapiens, H. neanderthalensis, and H. erectus, that australopiths are apes, and that the status of H. habilis is uncertain."³  His own baraminic-distance analysis puts erectus inside a human group discontinuous from non-human primates;  the method and its critics are argued in Lucy and the Pre-Human Hominins.

Lubenow.  Bones of Contention is the treatment every later creationist leans on.  Erectus and Neandertal are one thing at different sizes:  "Homo erectus is on the lower end, with regard to size, of a continuum that includes Homo erectus, early Homo sapiens, and Neandertal."²⁷  Modern brain size spans a factor of three, so the fossil range is unremarkable:  "Modern humans have a cranial capacity range from about 700 cc all the way up to about 2200 cc."²⁷  Brain size has little to do with intelligence.  And the fossils are "postflood Ice Age people" spreading from the Near East.²⁷

Two of his numbers fail on his own sources.  He gives the erectus range as running "from about 780 cc (Gongwangling, China;  Lantian 2, China;  and Dmanisi D22880, Georgia, CIS) to about 1225 cc (the largest Peking Man skull, Zhoukoudian X)."²⁷  But the paper describing D2700, at about 600 cc, is in his own endnotes, and his own fossil chart lists the specimen.  The smallest Dmanisi brain then known is missing from the sentence that carries his argument.  Skull 5 at 546 cc came nine years after his book and cannot be held against him.  The 700 cc floor for modern humans rests on three textbook statements, and two of them disagree with it:  "A. H. Schultz cited a Melanesian with a cranial capacity of 790 cc as the lowest on record for a normal human.  The late anthropologist Marvin Harris (University of Florida) claimed that the variability of human cranial capacity starts at 850 cc.  However, Stephen Molnar writes, 'In fact, there are many persons with 700 to 800 cubic centimeters.'"²⁷  None is a measured series.  The Molnar sentence is real at the primary:  it is on p. 57 of Races, Types, and Ethnic Groups (1975), in a paragraph that gives no data and says "variation of plus or minus 400 cc about the mean is seen in most European populations."²⁸  Molnar then withdrew it.  The 1992 edition of the same textbook drops the 700-to-800 sentence and cuts the within-population spread by a factor of four:  "variation of ±100 cc about the mean is evident within most European populations."²⁹  The strongest of Lubenow's three citations is a sentence its own author deleted.  Which edition Lubenow used is unknown, because search-inside does not return his endnotes.  The Schultz original was not obtained.

Rupe and Sanford.  Contested Bones sides with the lumpers and argues most erectus remains should be reclassified as Homo sapiens.  Its distinctive move is on Dmanisi.  "This has led to taxonomic ambiguity and has opened the door for australopith (ape) skulls to be wrongly interpreted as Erectus.  There is at least one possible example of this in the Dmanisi site (skull D4500)."¹⁵  So the smallest erectus skull may not be erectus at all.  Their range sentence then reads:  "Normal human braincase volumes vary greatly, ranging between 800 and 2220 cm3, with an average around 1345 cm3.  Erectus cranial capacities are on average 940 cm3 and range anywhere between 700 and 1400 cm3."¹⁵  The 700 cm³ floor is wrong by the book's own citations.  Skull 5 was published at 546 cm³ in October 2013, and their own endnote 15 cites the Schwartz and Tattersall comment on that paper.  Line's 2020 survey gives 546 to 1,260.  Their counter-examples are Anatole France at 933 cm³ and "A man named Daniel Lyon had a cranial capacity of 660 cm3, yet he lived a normal life and worked for the Pennsylvania Railway Terminal for twenty years in the late 1800's."¹⁵  The Lyon case is a nineteenth-century newspaper report.  Their Dmanisi reporting is otherwise accurate, and they cite the Science comment rather than pretend the lumping claim went unchallenged.

Line, 2013.  Peter Line is the one creationist writer with a worked model for the small Dmanisi brains, and it deserves to be reported straight.  Robusticity across erectus, heidelbergensis and Neandertals comes from development linked to long lifespans:  "Changes in development, linked to longevity, most likely involving thyroid hormone secretion patterns, are proposed as the primary mechanism causing robusticity."³⁰  In a web article on the fifth Dmanisi skull, he explains the very small Dmanisi brains as pathology, cretinism from iodine deficiency:  "If iodine deficiency disorders like cretinism are still a problem in some parts of the world today, despite modern medicine and information about iodine deficiency at our disposal, how much more of a problem could it potentially have been for early post-Flood/post-Babel human populations migrating to uncharted regions of the earth ...?"³¹³¹  He calls the paper "speculative"³⁰ and ranks the alternatives:  "The above scenario is very speculative, and it may well be that the considerable morphological variation, as well as the extremely small brain sizes, evident in the Dmanisi Homo erectus population, were within normal limits of variation for robust humans, although I consider this option less likely."³¹  Note what that concedes.  Line judges Lubenow's whole argument, normal variation, less likely than mass pathology.

Three problems.  Charles Oxnard's cretinism argument, which Line borrows, was made about Homo floresiensis and is contested there.  Five individuals of five ages, all cretinous, is a strong claim, and it does not reach the single small crania in Java, China and Africa.  And the mechanism runs the wrong way.  Line's whole model is slower development:  he quotes Carl Wieland approvingly that "the robusticity may be the result of genetic longevity/delayed maturation."³⁰  The Dmanisi subadult's teeth matured faster than a modern child's, not slower.

Line, 2020.  Seven years later Line stated the problem more plainly than any critic has.  "However, while some small-brained Homo erectus specimens can possibly be explained as a result of pathology, there are too many specimens with small cranial capacity for pathology to be an explanation for all of them."¹¹  "The most logical conclusion appears to be that most Homo erectus specimens were not pathological and that there was a huge natural variation in cranial capacity of these humans ... For reasons unknown many members of this group had small brain size, and hence small cranial capacity."¹¹  On the idea that erectus and Neandertals arose by degeneration from post-Babel moderns:  "this would have to have happened independently, in a relatively short time, in many different parts of the world, which seems a little too coincidental to be realistic."¹¹  What he retains is that the tools are human tools and that the people were early post-Flood migrants.¹¹

Four readings of one skull.  On Dmanisi Skull 5, 546 cc, the young-earth literature now holds:  fully human normal variation (Lubenow);  a human with cretinism, normal variation "less likely" (Line, 2013);  a human whose small brain is unexplained (Line, 2020);  and possibly an australopith wrongly filed as erectus (Rupe and Sanford).  Wood says "the precise status of the Dmanisi hominids will have to await further analysis."³  These cannot all be right, and the disagreement is about one measured skull, not worldview.  The two most technically able writers also disagree on mechanism.  Wood writes that Neandertal DNA shows "that explanations of Neanderthal morphology as the result of disease (Acton 1978), environmental factors (for example, Lubenow 2004, pp. 264–265), or extreme longevity (Cuozzo 1998) are incorrect."³  Line's 2013 model is a longevity-plus-hormone explanation for the same set of forms.

Tomkins, 2019.  The current ICR article contains the clearest case in this file of a citation turned against itself.  Tomkins writes that the Salkhit skullcap from Mongolia shows "similarities with Neanderthals, Chinese Homo erectus, and West/Far East archaic Homo sapiens," and that "the most recent study of the Mongolian fossil now lists it at about 34,000 years old,"³² citing Devièse et al. 2019.  That is the study that removed the archaic reading.  Its abstract says the skull was "initially described as an individual with possible archaic affinities, but its ancestry has been debated since the discovery,"¹⁴ gives a compound-specific radiocarbon date of 34,950 to 33,900 cal BP, and reports a complete mitochondrial genome that "falls within a group of modern human mtDNAs (haplogroup N) that is widespread in Eurasia today."¹⁴  The 2019 paper dates a modern human.  Tomkins's second late-survival case, Kow Swamp in Australia, is handled with more care by Line, who reports Durband's finding that the crania "have strong shape similarities to known deformed individuals,"³³ Curnoe's contrary finding of "no evidence for such modification in KS 1,"³³ and that with the collection repatriated in 1990, "getting a definitive resolution to the argument may be out of reach."³³  Robust is not erectus, and no erectus-diagnostic trait is claimed for Kow Swamp.

The stratigraphic order.  Wood states the movement's hardest problem in the movement's own words:  "the dispersal of the human population from Babel would presumably have been led by H. habilis and H. rudolfensis, specimens of which appear stratigraphically lower than any other human species ... whereas the erectus-like species (H. erectus sensu stricto, H. ergaster, and the Dmanisi hominids) are the next stratigraphically and the first human species to disperse across the Old World.  Global distribution was achieved only by Homo sapiens, the stratigraphically-highest of the major human species ...  Any model of post-Babel distribution must account for the appearance and disappearance of these human species in stratigraphic order."³  That order is what a dispersal-with-descent model predicts, and within it brain size rises at about 160 cc per million years.  No published creationist model predicts the order from migration routes and burial probability.  Wood's sentence has stood unanswered since 2010.

Objections and responses

"Lordkipanidze says it is all one species, so the creationists were right about variation."  Half right.  The lumping claim is real and defended.  But the variation defended runs from 546 to 730 cm³ with an australopith-range EQ, and the comparison is with chimpanzee populations.  Accepting it makes five small-brained individuals one population.  It does not make them modern.

"Fire, boats and butchery prove they were human."  This is the strongest card and the article plays it.  Its limits are the excavators' own.  Berna found no built hearths.  Ingicco says the toolmakers' identity is unknown and does not say boats.  No fossil lies with the Flores tools.  The Trinil shell is one object.  The behaviour shows a species that planned, cooperated and crossed water.  It does not show a human growth schedule, and the teeth do.

"Brain size has nothing to do with intelligence.  Modern humans go down to 700 cc."  Lubenow's point, and partly fair.  But the issue is not whether a 546 cc person could be clever.  It is whether a population whose adults run 546 to 730 cc, whose EQ sits with the australopiths, and whose children reach dental maturity at twelve, is a modern population.  The 700 cc floor rests on three textbook statements that give 790, 850 and 700 to 800, none a measured series, and Molnar deleted his from the next edition of his own textbook;  Rupe and Sanford give 800;  Skull 5 is 546.  Chris Rupe has since struck cranial capacity from the creationist toolkit, listing "small cranial capacity" among traits that "should not be considered diagnostic."³⁴

"Turkana Boy would have been six feet tall."  That was the mainstream figure of the 1990s, faithfully reported by Lubenow.  It assumed a human dental age and a modern adolescent spurt.  Graves 2010 withdrew it:  163 cm, growth finished by 12.3 years.

"The Dmanisi teeth could be pathology."  Line's model is cretinism, inside a wider model of delayed maturation.  The Dmanisi child's teeth matured fast.  A model built on slow development cannot explain fast development.  Be careful with the clinical step, though.  A 2022 case-control study of 80 children with hypothyroidism against 80 matched controls found "no statistically significant difference in estimated dental age" between them.³⁵  That cuts both ways:  it removes the simple claim that hypothyroidism slows teeth, and it also removes cretinism as a way to produce a schedule five to nine years short of a modern one.  A creationist wanting to keep the pathology reading needs a different disease, and none has been proposed in print.

"The dates keep changing.  Ngandong moved by a factor of four."  It did, from 27 ka to 117 ka.  Every position in every one of those fights is still between about 6 and 420 times older than a post-Babel date of about 4,300 years.

"Australopithecus overlaps erectus in time, so one cannot be the other's ancestor."  Tomkins puts it this way:  "the supposed initial phase of human evolution is represented by the ape-like Australopithecus, which overlaps significantly in time with H. erectus.  If they coexisted, how could one have been the evolutionary predecessor of the other?"³²  The overlap is real.  It refutes a single-file ladder that nobody has defended since the 1970s.  A lineage can survive alongside its own descendants, so overlap alone settles nothing.  This article does not use the argument.

"Creationists have never engaged the primary literature."  False.  Line's 48-cranium survey matches Antón's table, Rupe and Sanford cite the Science comment against their own case, and Wood names the stratigraphic problem himself.  What is missing is an answer to the teeth and to the dates.

"Then erectus is a transitional form and humans evolved."  This is the objection to the article's own verdict, and it needs care.  What the evidence here establishes is narrower:  that erectus did not grow on a modern human schedule, that its adult brains were small, and that its sites are old by every method applied.  It does not by itself establish descent from an australopith;  the body plan, the tools, and the pattern of dental maturation are human-like.  Wood's position, human but of a different kind, is not refuted here.  What is refuted is the claim that these were ordinary people who look odd because life was hard.

Issues and unanswered questions

Computations that would settle what is still open.

1. How far outside the human range is a first molar at 3.8 years? Zollikofer's Figure 3a rests on population-level studies of M1 emergence. Report how many standard deviations 3.8 years sits from the human mean.
2. Regress endocranial volume on stratigraphic age across Antón's Table 3 plus Dmanisi and Gona, with a confidence interval. A model that reads the section as burial order must predict the same correlation from geography or ecology.
3. Check whether one decay-rate factor can do the work. A single factor applied to every decay constant compresses an argon-argon age and a ²⁶Al/¹⁰Be burial age by the same 1/k, so those two together do not discriminate. What would discriminate is a model that accelerates different decay modes by different amounts. Compute what such a model does to the beta-decaying pair ²⁶Al and ¹⁰Be set against an argon-argon age on the same section, and say what ratio of factors would be needed to bring both inside a post-Babel interval. Nobody has published this.
4. Pull the modern intracranial-volume distribution from a large MRI cohort. If healthy adults do not reach 700 cm³, the claim is dead; if they do, it is the strongest card the creationist side holds. Molnar's own retraction makes this the more urgent of the two open computations.
5. Count localities against stratigraphic position for habilis, erectus and sapiens. A sampling artefact should weaken as the sample grows.

Where the young-earth position has no published answer.  Dean 2001.  Zollikofer 2024.  The Dmanisi date structure of argon floor plus magnetic ceiling.  The agreement between the burial date, the magnetic stratigraphy and the loess correlation at Zhoukoudian.  Wood's stratigraphic order.

Sources not read in full.  Fourteen sources behind this article were obtained at abstract level only:  Swisher 1994, Swisher 1996, Dean 2001, Goren-Inbar 2004, Rightmire 2004, Shen 2009, Brumm 2010, Graves 2010, Roebroeks and Villa 2011, the two halves of the Schwartz and Lordkipanidze Science exchange of 2014, Joordens 2015, Rizal 2020 and Emeksiz 2022.  Every abstract was re-fetched as bytes from NCBI E-utilities or the OpenAlex Crossref deposit during verification, and nothing is attributed to any of them beyond its published abstract.  The Nariokotome monograph, Walker and Leakey 1993, is not on archive.org and was not obtained;  Walker's 1996 book, by the same first author, stands in for it.  Rightmire's 2005 Dmanisi description carries no abstract in any index reached and was not obtained;  the Dmanisi cranial capacities in the table come from Antón's Table 3 instead.

One creationist writer is quoted only through another.  Jack Cuozzo's position, that erectus is an ape, is taken from Wood's citation of him rather than from Cuozzo's own book, which was not obtained.  Nothing in the article turns on it beyond a single sentence in the Verdict.

Where this article's own side is weakest.  The Trinil engraving is one object from an old collection.  Wonderwerk has no built hearth, and Roebroeks and Villa's 300 to 400 ka threshold for habitual fire is a live position.  The "700 cc modern human" claim still has no measured series behind it in either direction.  Molnar dropping the sentence is evidence that he did not think it held, not a measurement that it is false.  And the decay-acceleration arithmetic in Part 2 is worked out here rather than taken from a published treatment.  An earlier draft of this article had it wrong, claiming a burial age resists a uniform rescaling when it does not.

Verdict

Homo erectus behaved like people and did not grow up like people.  The behaviour is the strongest evidence in the file and the excavators supply it:  fire thirty metres inside Wonderwerk, a rhinoceros butchered on Luzon across a sea barrier, an engraving at Trinil, and a Dmanisi elder fed for years by others.  The Dmanisi team argue in print that the variation creationists point to is real and within one population.

The "fully human population" reading fails on two measurements.  A synchrotron count of daily growth lines puts the Dmanisi child's dental maturity at 12.0 to 13.5 years against 18 to 22 in every living human population, with molars erupting at wild-chimpanzee ages.  Dean reached the same conclusion in 2001 by enamel growth rates, and Graves in 2010 from the skeleton.  The one creationist mechanism on offer, cretinism, sits inside a model of delayed maturation, and the teeth were fast.  And the sites are dated by seven independent methods across Africa, Asia and the Caucasus, in agreement.  No creationist source found answers the teeth or attempts the dating arithmetic.

What an honest young-earth position looks like.  Follow Wood and Line rather than Lubenow.  Say that erectus made human tools and human fires and crossed water, and that its people cared for their old.  Say that their brains averaged 70% of the modern mean, that pathology cannot explain most of them, that their children matured on a schedule no living population shares, and that the cause is unknown.  Say that the sites are old by every method applied and that a post-Babel model owes an account of Wood's stratigraphic order and has not yet paid it.  Then argue, if you wish, that a population can be human without being modern.  That is a position the evidence permits.  "They were ordinary people" is not.

Companion pages

• The literature — every source consulted, what it measured, and which way it cuts.
• Quarantine — the claims held back, and why.
• Lucy and the Pre-Human Hominins — the australopiths, the habilis rank question, and Wood's baraminology method.
• Neanderthals and Denisovans — the forms erectus is claimed to have given rise to.

Sources