The Levantine Roots of the tophet Sanctuary

I Survey of votive sanctuaries outside Carthage

Phoenician and/or Carthaginian infant cremation sanctuaries, i. e., votive tophet precincts, spread throughout the central Mediterranean – in north-Africa, Sardinia and Sicily, and perhaps Malta – and remained active from about the eighth BCE to the second/third centuries CE.^[1] Urns and stelae serve as the characteristic artifact types from these open-air enclosures, with urns essential for containing the cremated remains of very young children (infants), ovicaprids (lambs/kids), or both. The tophet constitutes a unicum that cannot be confused with other cult places or even necropoleis in the Phoenician Mediterranean. In fact, despite the variety of installations and the different local choices, and in addition to a specific archaeological context – open-air precinct separated from other sanctuaries and necropoleis – tophet sanctuaries systematically present a common basic structure, primarily consisting of infant cremation burial in urns and, secondarily, votive stone markers, some with dedications to Ba‘l (and Tinnit).

Despite over a century of excavations, we have a puzzling information gap given that we have yet to find archaeological evidence of this basic structure in either the eastern Mediterranean (the Levant) or in the far western Mediterranean (e. g., Mauretania and Iberia). Two alleged Levantine sites have been labelled “tophet” because they share the primary structural components of urn burial in open-air precincts, although they lack stone markers. For one site, we await sufficient documentation, while the other must simply be excluded because it lacks necessary components. For the former case, the purported tophet at Amathus on Cyprus remains hypothetical, since evidence from that urnfield awaits full analysis and thorough publication. In any case, the precinct does not seem exclusive, with adult burials next to children.^[2] As for the second case, Achziv in Israel, the supposed precinct there must be excluded since the archaeologists, by their own admission, deliberately use the term tophet quite broadly (and quite misleadingly) as a “place of burning” in reference to an adult cremation necropolis.^[3]

The case of Malta presents a different problem because here we lack a secure archaeological context; nevertheless, the site provides important evidence. Poorly recorded amateur excavations from two centuries ago leave us guessing about the open-air precinct, although we have enough pertinent archaeological and epigraphic evidence to confirm the existence of an early tophet in the vicinity of Rabat. In particular, those amateurs extracted two stelae with typical votive inscriptions from the seventh century (one now lost) as well as significant ceramic and osteological finds (also lost), including 60 urns with unidentified bones.^[4]

To our research question: why do tophet precincts occur in the central Mediterranean, while we lack archaeological evidence for them in the east? Some tentative answers have been offered, but the complex problem has found only speculative responses that lie beyond the scope of this paper to resolve in all their implications. More modestly, we will review the available sources connecting tophet precincts with metropolitan Phoenician ones and evaluate their relevance.

At the outset, we should caution that one must very cautiously consider theories based on absence of evidence. In the Levant, this absence could simply be ascribed to the gap in our archaeological knowledge about the main Phoenician sites – particularly between the ninth and seventh centuries – essentially due to modern overbuilding that limits archaeological research. Further exploration may well bring new data to light that could change the documentary imbalance. Apart from this, however, both the nature and quantity of current data should be considered as purely coincidental and assume that the disproportion in our evidence may eventually find satisfactory explanation.

In addition, the possibility of a non-Levantine origin of the tophet is hardly sustainable. Despite such a marked imbalance in the documentation between east and west, this type of cult-place had Phoenician roots rather than a genesis exclusively in western contexts. Neither historical data nor cultural comparanda support invention ex nihilo of the tophet by early Phoenician migrants in new territories. Nevertheless, the conditions that settlers faced and the consequent strategies they pursued in their extraordinary “colonial” expansion would have played a formative role over time in the tophet phenomenon, even if they were not the Ursache of these sanctuaries. Particularly, interaction with the various indigenous peoples surely generated new ways of adaptation in one sense or another, and produced local variations. Indigenous populations may not have decisively influenced the function of the sanctuary, even though they intervened in various ways leaving traces of their participation in cultic activities.^[5] At any rate, we suppose that a shared social and religious ideology pre-existed among the settlers, a confessional identity that became even more salient during their so-called “diaspora,” whenever they came into contact with outside groups.

The chronology of settlement also provides a powerful argument in favour of a Phoenician origin for these special sacred places and their rites. All early tophet precincts begin their use phase at almost at the same time as their towns were settled, i. e., no later than the first/second generation of settlers. At Carthage (ca. 800 BCE), Sulci (ca. 750 BCE), Tharros (ca. 700 BCE), and Motya (ca. 750–725 BCE)^[6] on Sicily, the precincts date to the years nearly immediately following foundation. In each of these cases, the tophet appears as a fully mature and already paradigmatic institution. As a consequence, rather than create novel precincts, Phoenician settlers consolidated ancestral traditions brought from their homeland. We must presume a particular ideology, even if not necessarily one shared by all Phoenician migrants.

While a Levantine origin, therefore, seems assured, a whole series of new problems remain open concerning the most plausible general scenario for their transfer, starting with the identification of metropolitan origins for these migrants, and whether or not their daughter settlements share common social and “religious” customs.^[7] Does the translocation of these precincts reflect a planned action of settlement, or correspond to an “anarchist” diaspora, possibly conditioned by strategies adaptable to the specific places and to the ethnic and social contexts of the natives and their reactions?^[8] Certain theories assume diversified Mediterranean “circles” or “circuits,” while others attempt to elaborate compromise interpretations to reconcile the various hypotheses. The so-called “identity” issue involves the extent to which settlers consciously identified with tophet rituals, either as characteristic of the Phoenician “diaspora” as a whole, or as a marker of a particular sub-Phoenician identity, whether in the east or the west.^[9]

One hypothesis proposes a typological dichotomy of settlement, where, on one hand, the presence of a tophet indicates a large agricultural settlement of significant population with territorial ambitions; but where, on the other hand, the absence of a tophet would indicate a commercial enclave, chiefly interested in trading and extracting metals, and less in need of such a representative civic sanctuary.^[10] This approach has appeared too rigid and met less favour in recent times, and various proposals to blend it or modify it according to other criteria have been put forward. The tophet cannot be considered a generic identity institution of the whole Phoenician expansion in the west, but possibly a partial and circumscribed manifestation of it. We should not doubt the Levantine genesis of this unique sanctuary and its rituals.

As far as the precise center from which this ritual expansion radiated, we should look mainly, if not exclusively, at Tyre. The historiographic tradition marks this Phoenician metropolis as a central reference point for the migratory phenomenon, linked to Carthage as a sort of “New Tyre” from its very foundation.^[11] In addition to the evidence examined below, we look to Tyre and environs for the transfer of the cult of Ba‘l Ḥammon – not a minor deity of Tyrian pantheon – and rites of cremation, which has absolute continuity in the tophet, contrary to the variations found in the rituals at Phoenician necropoleis elsewhere, which oscillated between inhumation and cremation.^[12]

II The NṢB MLK at Nebi Yunis (RÉS 367)

From this presentation of general historical background we move to a different perspective, absent relevant archaeological evidence, and turn to various scattered textual clues that provide explicit and decisive support in favour of a Levantine origin of this cult-place and its related religious customs. With regard to the epigraphic sources in the Levant, Dewrell recently presented the status quaestionis (2017) and I generally agree with his conclusions. He rightly claims that, out of the available iconographic and epigraphic evidence from Syro-Palestinian and Jordanian excavations, and the evidence available from Cilicia and Cyprus, nothing provides unambiguous proof of a tophet precinct in the east. Consequently, we will not return here to the documents he has analysed – with one exception where we do not share his scepticism. The authenticity of this exceptional epigraphic artifact has been debated since its discovery, first published by Lagrange but now lost, namely the Phoenician inscription from Nebi Yunis, located between Jaffa and Ashdod (third-second century BCE).^[13] The importance of this inscription, engraved along the edge of an offering table, lies in its mention of a mlk-offering honouring the god Eshmun made by a large association of men:

Lidzbarski considered the Nebi Yunis inscription a forgery,^[14] identifying various grammatical anomalies as well as some personal names apparently derived from scholarly works that had appeared immediately prior to the inscription’s publication. The arguments of Lidzbarski were synthetically presented in his Handbuch (1898), then expanded in the first volume of his Ephemeris (1902). In the latter, responding to Lagrange’s counter-objections, he provided a good picture of the inscription and confirmed his judgment of forgery in more detail, but without making any substantial additions to the text itself. The authority of Lidzbarski has strongly conditioned scholarly opinion for many years but, eventually, other scholars have provided contrary arguments in favour of its legitimacy.

In their detailed study, Delavault and Lemaire have provided new photographs of the text and have certified its authenticity.^[15] In particular, the French scholars have shown how Lidzbarski directed his arguments against the form rather than the substance of Lagrange’s presentation, relativizing or even eliminating the suspicion of forgery. As regards grammar and syntax, Gianto has found no insurmountable obstacles, and he too concludes that the document must be authentic.^[16]

Recently Dewrell provided a new and careful review of this specific inscription. Although not completely convinced of its fraudulence, he does conclude that “both the circumstances of its discovery and the suspicious nature of its contents provide too much room for doubt to allow the Nebi Yunis inscription to serve as a secure basis for claiming evidence of mlk sacrifices in Phoenicia proper.”^[17] While generally sympathetic to Dewrell’s method, particularly his prudence regarding artifacts that derive from the antiquities market, his objections to authenticity remain weak and are essentially limited to the use of nṣb to designate an offering table rather than a stela. But even this objection can be overcome. At the start, we consider the restoration of [n]ṣb as correct (the first letter is illegible). We also concede that, in the vast majority of cases, nṣb does designate a stela or stone marker (in Phoenician) and that, in a few other cases, the term can refer to a statue (in the dialect of Sam’al), yet the root meaning generally signifies “something erected.”^[18] In our case, as Lidzbarski has already argued,^[19] that offering table would have had a base or pedestal for support, therefore its placement above would fit within the semantic range of “erecting.”^[20]

In terms of prosopography, Lidzbarski had deemed “suspect” the mention of personal names already attested in Phoenician, but this in itself does not provide decisive proof of fraud since it happens quite frequently (and fortunately) in epigraphy. Moreover, certain anthroponyms that occurred in the Nebi Yunis text, seemingly as hapax legomenona, have since been found in other inscriptions, names that a forger of Lidzbarski’s time could not have known or even imagined. As a striking example, the personal name ʿbdtwyn (line I.3) had occurred only here until it was found in an inscription from Ibiza over a century later.^[21] Finally, one powerful clue in favour of inscription’s authenticity derives from remarkably coherent and consistent palaeography, a factor Lidzbarski never took into consideration.

If our assessment is well grounded, then the Nebi Yunis inscription provides the first known mention of a mlk-offering in the Levant, in this case dedicated to the god Eshmun, a quite unexpected addressee for this kind of rite. Even though considered a healing deity, Eshmun’s personality must have been much more complex, given his role as Lord (b^cl) of Sidon in the Persian period.^[22] At the time of the Nebi Yunis monument (third-second century BCE), we must expect a very different cultural and religious climate from that of earlier epochs and from western Mediterranean contexts, for example, its collective character listing multiple dedicants.

Unfortunately, we lack an archaeological context and the inscription fails to provide cultic details, but, at the very least, the term mlk would have the general sense of sacrificial offering (etymologically, something “caused to go forth” to the divine recipient)^[23] without explicit attendant sacrifices. In any case, proceeding with the utmost caution, we see no reason for excluding this monument from our small dossier of evidence for Levantine tophet ritual. The Nebi Yunis inscription indicates that molk-sacrifices existed suo modo in Phoenicia as late as the Hellenistic period and alongside traditions that differed from those in western settlements.

III Tyrian Canaanite ritual in the Hebrew Bible

Biblical passages present a block of testimony on tophet, mlk, the rites of “passing through the fire,” its divine recipients, and its location in the valley of Ben Hinnom. The topic has been the subject of countless studies,^[24] and here we recount that no less than 25 passages in the Old Testament attest, more or less directly, that Israelites followed the Canaanites, sacrificing (and burning) their children (sons and daughters, possibly firstborn) in Tophet, near Jerusalem, according to the following sequence:

Israelites (“imitating” the Canaanites)

¯

slaughtered (šḥṭ) or sacrificed (zbḥ) in holocaust (ʿlh) or shed the blood (nqy dm)

¯

of their children (sons and daughters)

¯

which were passed through fire (śrp bʾš; ʿbr bʾš)

¯

as a mlk ritual (lmlk)

¯

in Tophet (at the gates of Jerusalem)

¯

to Ba‘l or to gods (unnamed).

In the Old Testament, we find the verb šḥṭ “to slaughter” (Isa 30.31–33; Ez 16.20–21 and 23, 36–39); the verb zbḥ “to (kill in) sacrifice” (Ez 20.28–29; Ps 106.34–39); the term “holocaust” (ʿwlh, Jer 9.3–16), exactly the same term used for Abraham and Isaac (Gen 22.1 ff.), Mesha and his son (2 Kings 3.26–27), and Jephtah’s daughter (Judg 11.29 ff.); the expression nqy / ytn dm “to shed / to offer the blood” (of children) (Ez 16.36; Ps 106.34–39), not to mention other allusions—thus the reality of child sacrifices in a “Canaanite” context is beyond doubt. As a consequence, given that Eissfeldt’s hypothesis linking Punic sacrifice mlk and the biblical lmlk still endures, biblical data can be considered as indirect evidence of the practice mutatis mutandis in the Iron Age Levant.^[25]

IV Tyrian Phoenicians in Hellenistic and Roman sources

In addition the general testimony provided by various Greek and Latin authors concerning Phoenician-Punic child sacrifice (mainly, but not exclusively, pertaining to Carthage and the western Mediterranean), a select few classical sources locate the origin of the rite directly in Phoenicia, with Quintus Curtius Rufus (likely following Clitarchus) explicitly naming Tyre as the place of origin when he recounted the siege by Alexander the Great.

According to Clitarchus of Alexandria (third century BCE), “the Phoenicians, but especially the Carthaginians who honor Cronus, when they want to get something very important, vow to sacrifice one of their children to the god, if they get what they asked for.”^[26] This marks a clear distinction between east and west, providing an eastern Phoenician origin for the rite that then continued “above all” (μάλιστα) in western Carthage, perhaps even as it faded in the eastern metropolis. In any case, we have here a confirmation of the exceptional value attributed to child sacrifice when deployed in extreme circumstances in order to gain divine aid. This “mechanism of the vow”^[27] proved quite common in the Ancient Near East in general, the Phoenician-Punic world in particular.

Concerning the siege of Tyre by Alexander (332 BCE), Curtius Rufus informs us that the Tyrians had an archaic custom: the sacrifice of a well-born child (ingenuus puer) to Saturnus (i. e., Ba‘l Ḥammon). The rite fell into disuse, but, when faced with the threatened of destruction of their city by the Macedonian king, someone suggested that they reinstate the custom, a suggestion that the council of elders opposed. He supposed that, at the moment of settlement, the founders (conditores) had transmitted the ritual from Phoenicia to Carthage, where it remained in force throughout the life of the city:

In addition, some gave the advice to reinstate (in Tyre) a ritual least pleasing to the gods, having been suspended for centuries: the sacrifice to Saturn of a child belonging to a free family. This ritual, sacrilege rather than sacred, was transmitted to Carthage by its founders where, it is said, it was practiced until the destruction of the city; and if the council of elders who held authority had not opposed it, a barbaric superstition would have triumphed over the sentiments of humanity.^[28]

It seems paradoxical to argue that this passage intentionally denigrates metropolitan Phoenicians, as has been suggested.^[29] In fact, the Tyrians refuse to restore the ancient and unpleasant custom, more superstitious than humane. Instead of denigrating the Tyrians, he posits a sort of ethical “progress” for them. Also note that the sacrifice of a child is not projected here onto the myth of Cronus (as in Philo, below), but onto an historical siege, drawing a parallel between the customs of Carthage and those of metropolitan Tyre.

A passage by Athanasius of Alexandria (fourth century CE) deserves only a brief mention, with its long list of cruelties committed by “pagan” peoples that includes Phoenicians (and Cretans) sacrificing children to Cronus. Particularly in patristic authors, we find generic polemical intent towards foreign peoples across the ancient world, with characteristic but stereotypical depravities tendentiously attributed to them.^[30]

Finally, in the mythical-historical fragments of Philo of Byblos (ca. 50–140 CE), cited in other authors like Eusebius of Caesarea, the Phoenician author preserves complex native traditions, including the foundational myth of child sacrifice. The standard tradition generally runs as follows—in situations of extreme danger (e. g., plague, famine, siege), leaders of Levantine cities, those from the most noble families, would offer “the dearest of their children” to appease the deity. In this foundation myth, El is rationalized as king of the country (later becoming the astral god Cronus), with an only-begotten son, Ieoud, whom he sacrificed.^[31] The details of this mythical ceremony suggest that the victims should be of noble rank and would, in the afterlife, enjoy the same prerogatives as the sovereign:

Among ancient peoples in critically dangerous situations it was customary for the rulers of a city or nation, rather than lose everyone, to provide the dearest of their children as a propitiatory sacrifice to the avenging deities. The children thus given up were slaughtered according to a secret ritual. Now Cronus, whom the Phoenicians call El, who was in their land and who was later divinized after his death as the star of Cronus, had an only son by a local bride named Anobret, and therefore they called him Ieoud. Even now among the Phoenicians the only son is given this name. When war’s gravest dangers gripped the land, Cronus dressed his son in royal attire, prepared an altar and sacrificed him.^[32]

We must distinguish here between the mythical and the historical (or pseudo-historical) in these testimonies. On the mythical level, Cronus-El (i. e., Ba‘l Ḥammon) had established the sacrificial rite of child immolation to save his (Phoenician) land from grave danger. On the historical level, however, our sources clearly confirm the Levantine origins of this Phoenician custom, a tradition transmitted to the Carthaginians in the west by their founding fathers. As we have seen, the rite had apparently fallen into disuse in the metropolitan homeland, while it continued in Carthage, albeit with softening and alternating intensities, as demonstrated by the active use of the tophet precinct until the destruction of the city.

To sum up, this somewhat heterogeneous dossier remains reliable and consistent in reference to the Phoenician homeland. Even if we may not have answered all research questions to our own satisfaction, we can confirm an eastern Phoenician origin for votive sacrifice in western Phoenician tophet precincts.