Two late Roman authors claimed that Hunnic parents cut the cheeks of newborn boys. According to Ammianus Marcellinus, writing in the late fourth century CE, the wounds left scars that interfered with beard growth. Jordanes, writing in the sixth century, gave the cutting a harsher meaning: he said the babies were made to endure wounds before they had even been fed milk. Later retellings have often treated these passages as descriptions of an established Hunnic custom. Both, however, were written for Roman or post-Roman readers by men who did not speak for the families whose behaviour they described. Establishing what the Huns themselves did is much harder.
The Huns became a major presence in Roman affairs in the 370s. Their advance north of the Black Sea disrupted Gothic communities and helped set large groups moving toward the Danube, the frontier of the Roman Empire. In the middle of the fifth century, Attila ruled a powerful confederation centred on the Carpathian Basin, the region around present-day Hungary. His empire encompassed people with different languages, ancestries and traditions. A practice found in one community would not by itself establish a custom shared by the entire Hunnic confederation. The ancient descriptions do not supply that information.
The difficulty begins with the nature of the surviving evidence. We have written claims about infant cheek cuts, but no identified child whose skin preserves the scars, no Hunnic account of the procedure, and no reliable measure of its prevalence. Archaeologists have found something different and much more securely documented: intentionally reshaped skulls in fifth-century central Europe. These remains show that some children in Hunnic-period communities underwent permanent changes to their appearance. They cannot tell us whether those same communities also cut cheeks. A reshaped skull can be studied centuries later; a healed cut to the skin usually cannot.
What Ammianus actually wrote about the Huns
Ammianus Marcellinus had served in the Roman army before composing his history of the empire, the Res Gestae. The surviving portion covers the later fourth century and concludes with Rome’s defeat at Adrianople in 378. As he explains the crisis brought on by the Huns’ advance and the resulting movements of Gothic groups, he pauses in Book 31 to describe the newcomers. Such a passage is called an ethnographic digression: the writer sets out what a foreign people supposedly look like, eat, wear and do in war or at home. It can tell us about Roman impressions, but it is written from outside the society being described.
In Book 31.2.2, Ammianus says that the cheeks of Hunnic children were deeply scored with iron from birth, and that scars prevented the usual growth of beards. He describes Hunnic men growing old beardless and compares their appearance unfavourably to eunuchs. His stated explanation concerns the marks on the face and their supposed effect on hair. There is no claim here that cutting served as soldier training, a test of bravery, a badge of rank or a religious observance. Those suggestions come from other accounts or from later attempts to make sense of his words. The supposed purpose of the wounds must therefore be distinguished from Ammianus’ account of their appearance.
Ammianus places the sentence about scars amid other descriptions designed to make the Huns seem strange and threatening. He portrays their appearance in almost inhuman terms, and he repeats stories that they cared little for cooked food or permanent dwellings. We have no reason to accept that entire portrait as a straightforward record of daily life. Ammianus knew Roman campaigning well, yet he also inherited literary conventions that portrayed peoples beyond the empire as the opposite of civilisation. He used physical appearance to suggest moral difference. The audience and purpose of the passage matter as much as the vividness of its details.
In his 2016 study, historian Diederik Burgersdijk examined Ammianus’ Huns and Alans digression. He argued that older ethnographic models and ideological concerns largely shaped Ammianus’ account. His analysis does not make the infant-cutting claim impossible. It does prevent us from treating the sentence as a verified firsthand observation. Ammianus never explains whether he personally saw adults with facial scars, heard a soldier or official describe them, or repeated a familiar story. His text establishes that the claim circulated in late Roman historical writing, without establishing how Hunnic parents treated their children or how many adults bore such marks.
There is a limited medical possibility behind the claim about beards. A sufficiently deep cut can damage hair follicles, and scar tissue may then remain without hair in the injured area. That mechanism would permit patches of reduced growth, depending on where the wound lay and how it healed. It would not by itself explain complete beardlessness, establish that cuts were routinely made, or identify why anyone intended them. Ammianus may have observed an appearance and supplied a cause, or he may have inherited the explanation along with the story. That biological possibility leaves the origins of Ammianus’ claim unresolved.
Why Jordanes said Hunnic babies learned to endure wounds
Jordanes wrote his Getica, a history of the Goths, around 551 CE. Attila had died in 453, and the imperial power associated with his name was already gone. Jordanes worked with earlier writings, among them the lost Gothic history of Cassiodorus, but scholars still debate what he selected from those materials and how he changed them. His account was produced in a different political and literary setting from that of Ammianus. It records what a later writer believed about Hunnic violence and Gothic experience, rather than observations made at a fourth-century Hunnic birth.
In Getica 127–128, Jordanes says that the Huns cut the cheeks of male infants with a sword before they had received milk. The child, in his telling, was thereby taught to bear a wound. Jordanes also repeats the claim that the resulting scars damaged beard growth. Those linked details, newborn wounds followed by beardless faces, closely resemble Ammianus. The repeated details may derive from the same written tradition. Jordanes may have read Ammianus himself, or received the same material indirectly. The route cannot be demonstrated, and his later account cannot simply be added to the earlier one as a separate confirmation of the custom.
Jordanes had a particular historical story to tell. In his history of the Goths, Hunnic expansion brings violence and upheaval to the rule of Gothic kings. The image of a baby being wounded before being fed makes cruelty appear to begin at birth and suggests that harshness defined the entire population. Readers also receive an explanation for the physical endurance associated with feared warriors. Yet Jordanes gives us the interpretation of a hostile historian, rather than a statement from a mother, father or healer who carried out the procedure. Even if cheek cutting occurred, the passage cannot recover a family’s intention. The infant’s suffering reinforces Jordanes’ portrait of a violent invader.
Between the two accounts, a claim about physical appearance also becomes a story about moral character. The earlier author connects scars and beards; the later author adds training through pain. Neither supplies the missing local testimony, and their overlap may reflect a common written tradition. Read together, the works show how Roman and post-Roman writers developed a recognisable image of the Huns for their audiences. The precise ceremony sometimes inferred from these passages, applied uniformly to children across Hunnic territories, has no comparable foundation in the surviving evidence.
What the scars might have looked like, and what we cannot know
No surviving description tells us exactly where or how the alleged cuts were made. Ammianus and Jordanes do not count the wounds, describe a healed pattern, or say whether the marks covered the cheeks or lay closer to the jaw. There is also no recorded Hunnic word for such a custom. Drawings with evenly spaced parallel scars, or faces crossed by dramatic slashes, necessarily fill in details the sources leave blank. They may help an artist imagine an account, but they are not visual evidence of what anyone in the fifth century looked like. Even the alleged pattern of wounds cannot be reconstructed.
Both ancient writers specifically mention boys. They tell us nothing dependable about whether girls or women were treated in a similar fashion, whether adults could receive such marks, or whether every boy underwent the supposed operation. Hunnic rule extended over diverse groups, including people formerly subject to Rome, and customs need not have been shared by the whole confederation. A practice restricted to an elite, a few families or one population might readily have become an outsider’s description of “the Huns.” The surviving accounts provide no basis for choosing among these possibilities. The evidence contains no means of determining the practice’s social or geographic extent.
Modern explanations sometimes call the marks an initiation ritual or a sign of belonging. Permanent bodily markings can indeed carry social meaning, but that general observation tells us nothing definite about this alleged Hunnic practice. No participant left an explanation, and no surviving description gives its Hunnic name or place in family life. Even the apparently practical idea that the cuts prevented beards comes from unsympathetic outsiders. One writer might have misunderstood the appearance of people he encountered; later authors might have made an established tale increasingly dramatic. Neither explanation can be established from the surviving evidence.
What archaeologists can prove about Hunnic-period head shaping
Intentional reshaping of a child’s skull, often termed artificial cranial deformation or cranial modification, is well documented archaeologically. Bindings placed around a young head could exert pressure as the bones grew, gradually producing a different shape that remained visible in adulthood. Unlike a surface scar, that change affects bone. It can therefore be measured on a skull recovered from a grave centuries later. The finds show that some families in Hunnic-period communities altered their children’s appearance. They offer a useful comparison with the literary cheek story, but the actual procedure and the type of surviving evidence are different. Skull binding cannot authenticate a report of skin cuts.
A 2017 morphological comparison by Peter Mayall, Varsha Pilbrow and Liana Bitadze analysed 21 modified skulls from Georgia with 16 from Hungary. The researchers found differences in the resulting shapes, which may reflect different methods of binding. The comparison cautioned against identifying every artificially elongated skull across Eurasia as Hunnic. Related practices occurred among several populations, and the Georgian material did not establish unequivocal Hunnic settlement there. A skull’s shape can document a modification performed during childhood; ethnicity and political identity require evidence from the burial and its wider setting. Appearance alone cannot label the person as a member of Attila’s people.
The cemetery of Mözs-Icsei dűlő in Hungary shows why burial context matters. It belongs to a time when Roman provincial communities and new post-Roman populations were interacting. In a 2020 study of 96 graves, Corina Knipper, István Koncz and their colleagues examined grave goods, burial practices and chemical isotopes in human remains. Isotopic signatures can suggest differences in where people spent their childhoods. The team observed changing funerary customs and individuals whose early environments differed from those of others in the cemetery. Fifty-one people there had artificially modified skulls. The combination of migration, shifting burial traditions and head shaping suggests a community where practices crossed social divisions rather than marking a single, uniform ethnic group.
At Mözs, modified skulls belonged to males, females and children. They also appeared with different forms of burial. Those findings are difficult to reconcile with the idea that head shaping was exclusively the initiation of male Hunnic fighters. Binding occurred early enough to alter a developing skull, and that change remained visible later in life; beyond that, the bones do not disclose what the family called itself, which language it spoke, or whom the person served as an adult. The evidence establishes the physical treatment of a growing head. The surviving bones cannot reveal what the altered shape meant to the people who produced it.
In 2014, Mónika Molnár and colleagues published another examination of nine modified crania from fifth- and sixth-century northeastern Hungary. The researchers considered how the heads had been altered and how the custom spread through communities on the steppe and into Europe. Their work adds evidence for the movement and adoption of head-shaping practices, while leaving their meaning for particular people uncertain. These graves document childhood head shaping in Hunnic-period settings. It does not provide independent confirmation for the different procedure described by Roman writers, in which infants’ cheeks were supposedly cut. Treating the two as interchangeable would turn a measurable change to bone into proof of an undocumented change to skin.
Why the cheek-cutting story leaves so little physical evidence
Most graves from the fifth-century Carpathian Basin contain bones and teeth along with objects durable enough to survive burial. Soft tissue, including the skin of the cheeks, usually disappears. A cut healed during infancy could leave a visible scar in life and no detectable mark on the skeleton. Even a serious wound to the skin need not damage the facial bones. An intact cheekbone found in a grave therefore cannot be used to declare that its owner had never been scarred. Equally, it offers no reason to declare that the owner had been scarred. The absence of visible scars on bones alone cannot settle whether cheek cutting occurred.
Archaeologists can measure a skull that changed shape during a child’s growth, and in some circumstances they can recognise a blade wound that damaged bone. They usually cannot recover healed marks that affected only the skin. Direct evidence for the cheek-cutting story would require something unusually well preserved, such as skin showing identifiable scars, or a trustworthy contemporary depiction or independent testimony precise enough to establish the custom. Ordinary graves provide none of those things. This is why the same burial material can settle questions about head binding while leaving the words of Ammianus and Jordanes beyond a direct physical test.
The absence of preserved cheek scars cannot be treated as evidence that cheek cutting never happened. There are good reasons why such traces would rarely survive. Yet cranial modification offers no shortcut to confirming the literary claim: a changed skull and a scarred cheek arise from different interventions, and only one can normally be recognised in ancient bone. That distinction limits what an archaeological comparison can establish. Archaeology has established a particular kind of intentional head shaping, while the specific infant cheek wounds reported by Roman writers remain unverified.
Priscus at Attila’s court: an eyewitness with different interests
Priscus provides a closer view of Attila’s world. This Greek-speaking Roman historian joined an embassy to the Hunnic ruler in the late 440s CE; reconstructions commonly date the visit to 448 or 449. Unlike Ammianus, he entered Attila’s court, although his account survives only in fragments and collections assembled by later writers. Priscus describes travelling across the Danube, speaking with interpreters and officials, and observing the strict etiquette of a banquet. He noticed that Attila drank from a wooden cup while other guests used expensive vessels. The banquet shows formal diplomacy and social distinctions missing from sweeping portrayals of Hunnic savagery.
The surviving account of the embassy and banquet records Roman envoys moving through a court that included Hunnic rulers, captives, craftspeople and speakers of several languages. Priscus paid attention to buildings and clothing, seating arrangements, gifts, meals and the treatment of visitors. Such details complicate Ammianus’ suggestion that the Huns had little use for houses or settled diplomacy. Yet Priscus was a diplomat describing an embassy, and his surviving work reflects that purpose. He did not attempt to catalogue births, childhood customs or every bodily mark he encountered. His eyewitness access gives the account unusual value without turning its silence on infant customs into evidence against their existence.
The extant account of Priscus’ visit contains no description of babies having their cheeks cut. The omission matters because a visitor to Attila’s court might have noticed facial scars on adults and offered a useful check on Ammianus. It cannot establish that there were none. Nor should descriptions preserved by other authors be credited to Priscus without evidence. The often-repeated detail that Attila had a thin beard streaked with grey occurs in Jordanes, Getica 182, rather than in Priscus’ surviving banquet passage. Jordanes uses Priscus elsewhere, but that fact alone does not establish this particular beard description as the preserved words of the eyewitness. The attribution is part of the evidence problem.
If Jordanes’ description of Attila is accurate, the best-known Hunnic ruler had at least a thin beard. That sits awkwardly beside a general claim that Hunnic men were rendered beardless by childhood wounds. It does not tell us that Attila was cut, or that facial hair grew around scars invisible to his biographer. Neither writer provides a close description of his cheeks. The two descriptions operate at different scales: an author can describe a whole people in absolute terms, while a particular person’s reported appearance may not fit the generalisation neatly. The passage should provoke caution, rather than an imagined reconstruction of Attila’s face.
The Huns were a political world, not a single archaeological type
The political reach of Attila’s confederation makes generalisation particularly risky. Roman writers often described “the Huns” as a single people with shared customs, yet Hunnic rule embraced Goths, Alans and other populations with histories along the Danube and across the steppe. Attila commanded through war, tribute, negotiation and alliances among groups with differing backgrounds. Burial objects cannot simply be assigned to one ethnicity either. Belts, bows, horse gear and other possessions moved through trade, gifts, imitation and military service. An object associated with Hunnic-period elites does not automatically identify the ancestry or childhood customs of the person buried with it.
A 2017 isotope study of five cemeteries in Hungary by Susanne Hakenbeck and colleagues investigated mobility and diet. Chemical signatures preserved in teeth and bones help reconstruct diets and childhood environments. The results revealed varied food habits and changes within individual lives, consistent with contact between pastoral and farming traditions. The researchers also cautioned against assuming that a set of ornaments or burial customs identifies a person unequivocally as a Hun. A modified skull alongside a weapon and jewellery gives us evidence about a burial and some elements of a life, but it cannot prove a fixed Hunnic ethnicity or the presence of a cheek-cutting ritual invisible in the grave.
A 2025 ancient DNA study led by Guido Alberto Gnecchi-Ruscone and colleagues added genetic evidence to the picture. The team compared genomes from the Carpathian Basin with individuals from central Asia and the earlier Xiongnu empire. Some people in the Hunnic-period region shared genetic connections with distant steppe elites. The analysis did not reveal a large, uniformly eastern-descended population throughout the region under Hunnic influence. Genetic ancestry cannot establish the customs an individual family observed. The study strengthens the case for examining Hunnic-period communities individually, rather than imagining everyone under Attila as a single biological group with the same practices.
Archaeologists also study large bronze cauldrons and other objects linked with Hunnic-period networks. Their forms and distribution offer clues to feasting, authority and exchanges between communities. Yet scholars still debate how particular artefacts were used and how widely they circulated. No fixed combination of grave goods reliably establishes the occupant’s ethnicity. Head shaping poses a similar challenge: its presence is physically demonstrable, while the political and cultural meanings attached to it may differ from one community to another. The surviving evidence points to connections and shifting loyalties rather than a population defined by one standard collection of bodily marks.
How Roman descriptions turned appearance into a weapon
Physical description had long been a way for Roman authors to characterise outsiders. Hair, clothing, food and bodily features were made to carry judgements about how people behaved. When Ammianus compared Hunnic men with eunuchs, he relied on assumptions about masculinity and status that his audience would understand. Jordanes intensified the image with children made to suffer before they could drink milk. Both portrayals belonged to a literary world in which an opponent could be feared for military ability and degraded as almost less than human. The descriptions reveal their authors’ prejudices and their audiences’ expectations, even where they preserve details of actual encounters.
Roman interest in the Huns was grounded in serious political events. Their armies had changed the balance of power in the fourth and fifth centuries, and their movements disrupted communities on the empire’s northern frontiers. Roman historians and officials had to explain events that inspired fear and strained established ideas about security. Accounts of alarming, unfamiliar bodies supplied a vivid language for that fear. They could also shift attention away from the complicated part played by Roman diplomacy, internal competition and imperial decisions in the resulting crises. The rhetoric reveals how Roman historians explained defeat and upheaval to their audiences; it cannot serve as a survey of Hunnic daily life.
The expansion of Hunnic power helped drive Gothic groups toward the Danube and contributed to the upheaval that culminated in Rome’s defeat at Adrianople in 378. During the fifth century, Hunnic campaigns became entangled with Roman subsidies, diplomacy and factional rivalries. Those relationships form part of the western Roman Empire’s transformation. Their political importance is well documented. Hostile descriptions of their bodies require separate scrutiny. Rome faced an opponent capable of negotiation and statecraft as well as warfare. A frightening portrait was useful to Roman narrators precisely because these military and political encounters demanded explanation.
The evidence comes from two written accounts and the archaeological record. Ammianus claimed that cutting infant cheeks limited beard growth. Jordanes repeated the physical detail and interpreted the cuts as early training in suffering. Independently, archaeologists have documented the modification of growing skulls in communities associated with the Hunnic period. The first two statements are testimony from literary sources that may share a tradition; the third rests on measurable human remains. There is no surviving evidence linking all three into a universal Hunnic initiation ceremony. Each tells us something about this period, but none can safely be made to prove a different claim.
So did the Huns scar their babies’ cheeks?
It remains possible that some Hunnic-period families cut children’s cheeks. The two ancient descriptions are specific enough to investigate, yet we cannot verify where the information originated. Perhaps Ammianus heard an accurate report; perhaps his explanation of facial hair was mistaken. Jordanes may have borrowed the image and added a moral interpretation. We have no secure way to identify which families, if any, observed such a practice, or whether their purpose matched either Roman writer’s account. There is no surviving Hunnic testimony to settle the issue. Both the existence of the cuts and their alleged purpose remain uncertain.
The claim about Hunnic infants survives because Roman authors made bodily appearance part of their account of a feared enemy. Excavations provide a more complicated picture: communities with different backgrounds adopted head shaping and other visible customs, while their political and family identities did not necessarily align. Modified skulls from fifth-century graves prove that childhood binding occurred. They cannot preserve scars that may once have marked soft skin. Until evidence closer to the supposed cheek-cutting practice emerges, its frequency, appearance and significance remain beyond secure reconstruction. The literary account of scarred cheeks and the archaeological record of reshaped skulls remain separate kinds of evidence.
Discussion
Comments