The Tophet at Carthage is an ancient open-air sanctuary in Carthage, in present-day Tunisia, where urns containing cremated remains of infants, young children, and animals were deposited alongside dedicatory monuments. Used for several centuries before Rome destroyed the city in 146 BCE, it was associated with the deities Baal Hammon and Tanit. The site is central to a longstanding debate in archaeology and ancient religious history: whether its human remains principally represent children killed as sacrificial offerings, children who died naturally and received special funerary treatment, or a combination of these practices.
Name and scope
“Tophet” is a modern archaeological designation borrowed from the Hebrew Bible. In biblical texts, Topheth was a place in the Valley of Hinnom near Jerusalem associated with the burning of children and condemned by the biblical writers. The word’s etymology is uncertain. Its application to Carthage reflects an interpretation of the archaeological discoveries, not a securely attested Punic name for the sanctuary.
Archaeologists also use tophet for comparable precincts at several western Mediterranean settlements associated with the Phoenicians and their Punic successors, including Motya in Sicily and Sulcis and Tharros in Sardinia. These sites share features such as cremation urns, young human or animal remains, and votive monuments. They nevertheless differ in chronology, assemblages, and local practices; evidence from one cannot automatically be applied to all the others.
The Carthaginian site is often called the Salammbo Tophet because of its location in the modern Salammbo district.
Discovery and development
The sanctuary was identified in 1921, and subsequent excavations revealed successive layers of urn deposits and stone markers. Its earliest use is generally placed in the eighth or seventh century BCE. Activity continued into the second century BCE, ending with the destruction of Punic Carthage in 146 BCE.
Repeated deposition gradually raised the ground level. New urns were buried above earlier ones, and monuments were erected or rearranged as the precinct developed. Excavators have distinguished phases through pottery, stratigraphy, and changes in the form and decoration of the monuments.
The excavated area represents only part of the ancient sanctuary. Estimates of its total number of burials depend on assumptions about its original extent, the density of deposits, and the duration of use. Such estimates are not equivalent to directly counted individuals, especially because some urns contain remains from more than one individual.
Deposits, monuments, and worship
Cremated remains
The sanctuary’s characteristic deposits are ceramic urns containing burned bone and ash. Human remains are predominantly those of very young individuals. Some deposits contain sheep or goat remains, either alone or together with human remains.
These assemblages establish that the precinct was a specialized place for the ritual deposition of cremated remains. They do not, by themselves, reveal how each individual died. Cremation can destroy evidence of injury, while later handling and burial can fragment or mix bones.
Animal deposits are important to interpretation. They may represent offerings in their own right or, in some circumstances, substitutes for human offerings. Their presence demonstrates that the precinct cannot be understood simply as an ordinary cemetery without considering its sacrificial and votive setting.
Stelae and inscriptions
Many urn deposits were associated with stone monuments, including small shrine-like structures and upright stelae. Their imagery includes divine symbols, human figures, animals, and the emblem conventionally called the “sign of Tanit.”
The inscriptions, studied through epigraphy, commonly dedicate an offering to Tanit and Baal Hammon. Votive formulas may state that a deity heard the dedicator’s voice or granted a blessing. They provide strong evidence that the precinct served a religious function involving vows and offerings.
Some inscriptions contain forms of the term conventionally transcribed mlk. Its precise meaning and the classification of offerings described by related expressions remain debated. It should not simply be treated as the name of the biblical figure or deity commonly rendered “Moloch.”
The child-sacrifice debate
Ancient literary testimony
Greek and Roman authors describe child sacrifice among the Carthaginians. A prominent account by Diodorus Siculus concerns a crisis during the invasion of Africa by Agathocles in 310 BCE. Other ancient writers also associate Carthaginian worship with the sacrifice of children.
These accounts are relevant but require critical treatment. They were written outside the Punic religious community, sometimes long after the events described, and may reflect hostile stereotypes or literary conventions. Greek and Roman authors also identified foreign gods with their own deities, obscuring distinctions within Carthaginian religion.
Their external perspective does not automatically invalidate their testimony, but neither does it justify reading every literary detail directly into the archaeological remains.
Sacrificial interpretation
Scholars supporting a sacrificial interpretation emphasize the convergence of several kinds of evidence: ancient testimony, votive inscriptions, animal offerings, the sanctuary’s distinct location and organization, and the restricted age range of the human remains.
On this interpretation, at least some children were deliberately killed as offerings in fulfillment of vows or in exceptional circumstances. The deposition of animal remains can be understood within the same sacrificial system, potentially including substitution.
Even among supporters of this interpretation, there is no necessary agreement about frequency, the circumstances of selection, or the proportion of deposits resulting from sacrifice. Evidence for the practice does not establish that every child buried in the sanctuary was killed.
Funerary and mixed interpretations
Other scholars argue that the sanctuary received infants who died naturally and were excluded from burial grounds used for older children and adults. High early-life mortality could have supplied substantial numbers of such deaths without deliberate killing.
This interpretation draws particular attention to age estimates suggesting that some individuals died around birth or possibly before birth. Those findings, however, depend on contested methods and assumptions.
Mixed interpretations allow for both natural deaths and sacrificial killings within a shared religious precinct. This possibility underscores the need to distinguish several questions: whether sacrifice occurred at all, how often it occurred, and whether it explains every human deposit.
Limits of the evidence
Age-at-death estimation is particularly difficult for cremated infants. Researchers examine tooth development, bone dimensions, and skeletal maturation, but burning can cause shrinkage, distortion, and fragmentation. Different analytical methods have therefore produced different age distributions.
Demographic comparisons introduce further uncertainty. An ordinary infant cemetery might be expected to contain many deaths close to birth, but its observed profile would also depend on burial rules, preservation, recovery methods, and access to the cemetery. A sanctuary’s burial population need not represent the mortality pattern of the wider city.
The surviving sources answer different questions. Inscriptions establish dedication and divine recipients more securely than cause of death. Bones reveal aspects of age and treatment more securely than the intentions behind the ritual. Literary accounts describe alleged practices but cannot identify the fate of particular excavated individuals.
The secure archaeological picture is therefore of a long-used Punic sacred precinct in which cremated young humans and animals were deposited through formal religious practices. The scale, organization, and precise relationship between sacrifice and funerary treatment remain subjects of scholarly disagreement.