Introduction

1A factor often unemphasised or unremarked is the fundamental point that the percentage of any population in any given culture who directly encountered any foreign material at all was always very small. Even those who actually travelled abroad would have had only limited exposure to the foreign cultures they visited. Others exposed to foreign goods and iconography but in a more indirect and even circumscribed way would be the direct receivers onshore, royalty and the élite, some priests and artisans, and those associated with them, their families, social circle, entourage and even their slaves (who may themselves have been imported from abroad). The vast majority of the indigenous population in any culture falls outside these groups. Whilst such social considerations would affect exposure, so does geography: those living nearer to access points such as ports and trade routes also are better placed to encounter imported goods and ideas than others farther away. The actual imports, from however near or far, that we recover beyond their homeland probably would have been the conceit of only a very few individuals who had some means of access or exposure to them. Everyone else, this vast majority, would not. These observations can be applied to cultures worldwide, but the actual situations and influences can differ markedly depending on the actual circumstances and interests of the individual cultures involved. I have previously explored this construct over several publications focusing on ancient Egypt and the Bronze Age Aegean, especially Minoan Crete.1 In the present paper, I will be using Aksumite Ethiopia/Eritrea as a case study (fig. 1‑2).2 As our most detailed ancient source for actual imported goods to Aksumite world, the Periplus Maris Erythraei of ca 40‑60 AD, the goods to be offered for sale at Adulis is a starting point for my specific discussion. In focusing on the Aksumite polity, I found it extremely difficult to specifically equate my previous observations of Minoan acceptance and dissemination here, but I have come to realise that, although the specific patterns in acceptance of imports beyond their homeland are directly affected by the embedded individual as well as general understandings and situations specific to each culture, an underlying commonality exists and my overall thesis remains valid. The extent to which the two regions have been investigated archaeologically and published, and what has been revealed in each, is a major factor. The more numerous Bronze Age Aegean sites have a wider range of purposes, social levels, situations and developments over a longer period of time, and are more comprehensively excavated and published. Aksumite sites are typologically and chronologically more circumscribed, with a concentration on large-scale centralised and élite habitation, cemeteries and churches. Very few demonstrably “ordinary” graves and habitation sites are excavated, and few excavations are published in sufficient detail for useful discussion here.3

Fig. 1 – Simplified chart of Aksumite chronology (adapted from Bard et al. 2014, p. 312, fig. 15; British Institute in Eastern Africa; see also Phillipson 2012, p. 72, fig. 20).

Fig. 2 – Map of the Aksum region, indicating major sites and others mentioned in the text (created by N. Finneran, using www.arcgis.com base map).

2Recognisable imports transported from abroad have been recovered at excavated sites within Aksum, some in quantity and others only once or twice. Some are fairly limited in typological range, when compared to what potentially was on offer from their homelands.4 We may well first ask several questions: why were these particular imports transported beyond their homeland? More tellingly, why were these particular imports accepted on arrival? How and why did they impact on the society that received them? Why did they want them, and just what did they do with them? And, even more importantly, why – at the same time – not others, the majority that also would have been both actually and potentially available? Any receiving culture, or at least limited individuals and groups within it, will have been exposed to far more potential foreign goods and influences than they actually accepted.5 Individual decisions – plural – were made, as each import would have been individually considered and its potential repeatedly evaluated, whether consciously or not, by individuals and groups within the society. And the “ripple effect”, permeating beyond them to others who were not directly exposed, still would have some degree of influence. So, the questions of my title: why this one? And why here?

3What are the underlying rationales – not the means but the reasons – behind actual transfer and acceptance of goods and ideas beyond their homeland? These can range from the esoteric to the practical, summarised within seven general and partially interrelated circumstantial labels that can be applied to any culture: “trade”, “gift/gift exchange”, “ideology”, “personal memorabilia”, “war booty”,6 “foreign occupation” and “perceived need”.7 They can encompass commercial, political, religious and personal motives, sometimes overlapping. At least one, often more than one, label can be applied in every instance although not all would be applicable in all cultures, while imports can arrive directly or – actually or potentially already differing from the original – via an intermediary culture. But these underlying and often unconscious rationales for acceptance of each individual import will directly affect its individual impact on any – here Aksumite – society to individually varying degrees.

4Indigenous acceptance of an import must also infer varying levels of influence or intrusion, from severely restricted to widespread dissemination and assimilation within the society as a whole. And this, on a sliding scale between overwhelming acceptance and total rejection, incorporates the subtle and sometimes overlapping variations of my abstract: direct copying (“imitation”), adoption, adaptation, incorporation, interpretation (“misinterpretation”), and even enforced imposition of an imported thing or idea on Aksumite or any other society. Even these indigenous artefacts would be accessible only to some of the vast majority of the population who may be unable to distinguish between an import and its indigenous hybrid, but who potentially could distinguish between an indigenous hybrid and an indigenous non-hybridised traditional type. Yet, at the same time, the embedded influences of some imports may not even be recognised as derivative by the general population, if sufficiently removed from the original model and sufficiently common within society.

5Indigenous “copies” or “imitations” are not misunderstandings of the full originals, although sufficiently different that we can usually recognise them as local products, and we cannot presume the receiving culture intended – but failed – to reproduce the original intention and presentation. Others are deliberately intended as hybrids suited to local taste, with no attempt to directly “copy” the original. Indeed, no matter what is being “imitated”, the derivative will always exhibit some local character – it will be similar, but different.8 Imports might be accepted in toto, or alternatively only certain details or characteristics would be extracted for “imitation”. All aspects of an import need not be necessary or even acceptable within a society; only those perceived as relevant, having some meaning or useful for its own purposes (including aesthetic), would be – and were – assimilated, and these to varying degrees of “authenticity” by the receiving culture, or certain groups within it. The indigenous result is overlain by local interpretation, using only aspects of the original model that they had, consciously or unconsciously, selected for their own purposes. It also presupposes indigenous artisans capable of creating indigenous versions, which means in turn artisans who would first need to see, or visualise a description, in order to interpret the imported model, or its nearest available interpretation, through their own cultural norm and individual perception, for this very act of creation. Transmission and assimilation of imported models (material, ideological or functional) must be considered individually to build a comprehensive picture of their impact(s), as different traits may (and often do) reflect entirely different underlying acceptances. Each takes on a life of its own, diverging exponentially as it becomes farther removed from the original.

The Periplus Maris Erythraei and its goods for trade

6The anonymous author of the Periplus knew from previous experience that he had a ready market for certain goods at his various ports-of-call. Those destined for the Aksumite port of Adulis consisted of:

Articles of clothing for the Barbaroi, unused, the kind produced in Egypt; wraps from Arsinoë [in Egypt]; coloured abollai [a type of cloak] of printed fabric; linens; double-fringed items; numerous types of glass stones and also of millefiori glass of the kind produced in Diospolis [= Thebes, in Egypt]; brass, which they use for ornaments as well as cutting up for coins; copper honey pans (?) for cooking and for cutting up into armlets and anklets for certain of the women; iron which is expended on spears for elephants and the other wild animals as well as for war. Likewise there is also a market for: axes, adzes, knives; large round copper drinking vessels; a little Roman money for the resident foreigners; wine of Laodicea [Syrian coast/southwestern Turkey] and Italy, limited quantity; olive oil, limited quantity. For the king, silverware and goldware fashioned in the local manner; in clothing, abollai and kaunakai [another type of cloak, heavy and likely of wool], with no adornment and modest in price. Likewise also, from the interior of Ariakê [in north-western India]: Indian iron and steel; cotton cloth of the broader make, the so‑called monaché and sagmatogênê; girdles; kaunakai; garments of molochinon [a high quality cloth]; garments of cotton in limited number; lac dye.9

7Undoubtedly other goods also were available, although not listed in the text. Few of the imports on this list survive in the archaeological record, and some that do are insufficiently recognisable as imports to us. We will never know the influence (if any) of the various imported cloths and garments on Aksumite fashion, although clearly a ready market existed at all levels of society since a variety of cloth qualities, embellishments, styles and prices are itemised in the list. The Barbaroi living near Adulis were offered unused clothing produced in Egypt of unspecified but presumably cotton cloth, whilst those at Avalitês instead received (preferred?) presumably used clothing “cleaned by fulling”.10 Likewise, the quantity of Indian lac dye and extent to which it was coveted or used is unknown, as is the range of goods dyed – perhaps leather and some of the “unadorned” abollai and woollen (?) kaunakai imported for the king.11

Use and reuse of imports within Aksum

8Clear instances of intended Aksumite use, reuse and reworking of specific imported materials and goods are indicated in the Periplus text. “Brass, which they use for ornaments as well as cutting up for coins” was imported as raw. Although copper-based metal items rarely have been analysed, Ann Feuerbach12 reports almost all those tested in the “Tomb of the Brick Arches” are brass (copper with zinc), as are a bangle and earring acquired in Aksum (British Museum cat. 1989, 1209.5 & .8), so Aksumite metalworkers did make brass ornaments. However, “[…] cutting up for coins” is problematic. Askum did not mint its own coinage for another two centuries, in copper or bronze (copper with tin), but not in brass. What, if any, brass “coins” were the Aksumites making in the mid 1st century AD? Surely not for the “resident foreigners”, for whom “a little Roman money”, at that time the international currency, also was imported. Could the Aksumites have been imitating the only two known brass (orichalcum) coins of the period (ca 23 BC‑late 3rd century AD), the Roman sestertius and dupondius, perhaps only at Adulis to supplement any payment imbalances with the visiting merchants?13 The merchants themselves clearly knew about the practice, if so. Roman 1st‑2nd century AD coins have been recovered within Aksum, but only in later contexts, only in gold, and mostly adapted for reuse as jewellery.14 The author of the Periplus also mentions importing “copper honey pans (?) for […] cutting up into armlets and anklets for certain of the women”. Indeed armlets/anklets in copper-based metal have been recovered in the Early Aksumite tomb at Da‘erika near Aksum and elsewhere.15 Who are these “certain” women, what was their status or even profession within Aksumite society, and what would this particular jewellery visibly signify? Why not other women, what did they wear instead, and what – if anything – did the distinction infer in their society? Some sort of hierarchy or compartmentalisation elusive to us seems to be involved here. In addition to coinage, occasional evidence for cutting up, melting and casting copper-based metal(s), silver and gold for reuse also has been recognised archaeologically within 5th‑6th century AD Aksum.16 The over-life-size statue implied by the now-lost stone base at Aksum and the five statues mentioned in one Ezana inscription17 themselves represent a considerable outlay of these metals for deliberately impressive visual display.

9The raw materials and certain imports – brass, iron, steel,18 and cloths – clearly were intended to be fashioned into goods, and lac dye intended for use. Numerous iron tools and weapons have been recovered, and future analyses might enable us to distinguish between Roman and Indian iron that the Periplus says the Aksumites used to make spears for warfare and hunting, or the imported utilitarian adzes, axes and knives. All presumably would have a wide social distribution, by sheer population statistics the bulk likely for agricultural workers and soldiers although also the numerically fewer upper classes. David Phillipson’s thesis that iron also must have been obtained locally is now beginning to become apparent; an ironsmithing site located at Maryam Anza/Maryam Taqot is now under investigation, and smelting sites are known at 3rd‑4th century AD Aksum and reportedly “ancient” sites near Enticho.19

10Iron here was not solely for utilitarian use. Iron jewellery, especially bangles but also finger-rings, earrings and perhaps anklets, was worn by the wealthy although probably non-élite individuals interred in the Maryam Anza graves, and also known at Adulis, Yeha and Aksum.20 Pre-Aksumite/Early Aksumite copper-based bangles of different sizes, thicknesses and weights also are known.21 No iron jewellery, only tools and weapons, are listed in the Periplus, and indeed only a few iron weapons have been recovered at “Proto-Aksumite” Bieta Giyorgis and the 3rd century AD GT II tomb at Aksum.22 Iron jewellery may have become fashionable only in the 4th century AD, although apparently known earlier, and may signify some elevated degree of wealth or status. Locally-made or possibly imported Nubian iron applicators probably for use with kohl also were recovered at Maryam Anza, perhaps modelled on the visually similar imported Mediterranean copper-based applicators also found in the same graves and at Adulis.23

11Understandably, very few imported (as well as Aksumite) copper-based vessels have survived,24 although no “large round copper drinking vessels”. Even more understandably, neither have any silver or gold vessels but, if any “silverware and goldware fashioned in the local manner” brought “for the king” were recovered, how would we recognise them as imports? Interestingly, the Periplus text reverses the usual order of “gold and silver” to “silverware and goldware” only in the Adulis list, although the implications of this are unclear. Perhaps silver was considered more valuable than gold within the Aksumite world, or at least to Zoskales. Gold was more accessible as sources apparently existed in and around the polity and gold seems to have been processed at Aksum, but no silver extraction sites are known although it may have been separated from gold during processing,25 and Aksumite silver coinage consistently tends to be smaller in size and both quickly became – if indigenous – debased. However, bowls of undoubted classical Aksumite form and decoration in “Egyptian faience” would present a similar conundrum as otherwise we would have had no hint of indigenous faience production, much less of faience vessels, within the polity. Found only at Aksum itself, these then might be our only potential imports to be specifically “fashioned in the local manner” to survive.26

12Other non-surviving commodities are represented by associated goods. The “limited quantity” of both olive oil and “wine of Laodicea and Italy” in the Periplus suggests their distribution would have been highly circumscribed to the few who could afford them in lieu of any local oils and (possibly) wines available to the majority, or perhaps intended only for specific purposes.27 The few imported Early Roman amphorae fragments in Bieta Giyorgis OAZ Tomb II (ca 150 BC‑150 AD, so generally contemporary with the Periplus), are from Spain and southern France, presumably – at least originally – containing regional wines,28 the vessels themselves being incidental as imports. The more commonly recovered Late Roman ribbed amphora types, also representing products including regional wines, oil(s), date products and possibly garum attest to greater importation of at least some of these commodities in the late 4th‑7th centuries AD.29 Proportionately few of them originate in the western Mediterranean, but the most common is the late 4th‑7th century ribbed “Ayla amphora”. Both this and a less common costrel form were produced at the Nabatean port of Ayla and its predecessors.30 At least some are directly associated with wine transportation from abroad by the blackened resin interior coating,31 those from Ayla containing wine likely made from local dates. Although not mentioned in the Periplus, Red Sea fish garum recently has been identified at Early Roman Ayla and may also have been trans-shipped in later centuries.32 These and all other wheel-made clay vessels can only be imports into the Aksumite world as – for unknown reasons, despite the strong trading and political connections emphasised in the Periplus and other texts, and a highly sophisticated indigenous technological knowledge – Aksum entirely rejected wheel technology and Aksumite vessels are exclusively handmade. We can only speculate as to why, as the technology could easily have been adopted. It simply was not, and is an outstanding example of cultural rejection.

13Following their arrival and decanting, some amphorae were re-used as containers, as neonatal coffins (with neck and handles removed, and added stone lid), and stacked as a spout for filling a baptistry with water,33 so someone had realised they could be adapted for these purposes. Imported table amphorae and filter-mouthed amphorae and jugs were recovered at Matara, Adulis, Aksum, Fekada (near Adigrat) and the Black Assarca shipwreck near Adulis, some certainly and others possibly are reworked.34 Someone at Aksum, very likely unaware of its original purpose or even origin in the Nile Valley, carefully and deliberately removed both handles of an Aswani table amphora that perhaps were redundant for its new intended function. Someone else drilled and abraded an amphora sherd to fashion a “spindle whorl”, even abrading the ribbed surface that must have hindered its intended function – why else remove it?35 “Spindle whorls” of reused local potsherds are regular finds in domestic contexts although rarely published,36 and ribbed amphora sherds regularly found in excavation, so very likely whoever made this one merely chose it without knowing or even caring it was an import – it was that far removed from its origin.

14It has long been noted that glass beads are one of the “missing” items in the Adulis list of the Periplus. “Indo-Pacific” glass beads apparently were not imported to Adulis in the mid 1st century AD37 but are ubiquitous finds at least from the later 4th century.38 Nonetheless, five Mediterranean mosaic glass beads are published from “Proto-Aksumite” OAZ Tomb VI at Bieta Giyorgis, and one each at Adulis and in later 4th century AD or later contexts at Beta Samati and Wakarida.39 Technologically similar to mosaic beads, imported millefiori vessels are recovered at many sites, including Aksum, Bieta Giyorgis, Matara and Adulis.40 Millefiori vessels – what the Periplus describes as “the kind produced at Diospolis” – were produced within the 2nd century BC-early 2nd century AD Mediterranean, and some also are known in Late Roman Egypt.41 Although already imported during the 1st century AD and found in generally contemporary Bieta Giyorgis tombs, millefiori and other Early Roman glassware also is recovered in Late Aksumite “heirloom” contexts, suggesting latent importation even though no longer produced in the Roman world.42 Given the extensive range of glass colours represented at Aksum, and direct evidence for glassworking – but not glassmaking – at Beta Giyorgis, Aksum and Beta Samati,43 fragments must have been reused within Aksum, being remelted to produce indigenous vessels, beads, inlays and other goods, some in unusual colours and colour combinations.44 Some vessels follow existing Aksumite ceramic typology, whilst others are or may be unique indigenous developments of foreign types but conspicuously lacking direct comparanda abroad.45 Most glass beads and vessels are monochrome or monochrome combinations.

Influences and patterns of dissemination

15Actual imports are one thing, but their influence is quite another, and implies variable acceptance of a range of original models. What are the rationales for such indigenisation, however closely related to – or far removed from – the original? Is this, in the first instance, because some people unable to access or afford the original import and its associated prestige accepted an indigenous substitute instead, still implying almost the same sense of prestige? This prestige has always been important to certain people in any society, usually its rulers and élite, but even more so its “not-quite-élite” or “wannabes” not quite wealthy or socially important enough to acquire imports so attempt to emulate their “superiors” as best they can. Would the “limited number” of Indian “garments of cotton” have a similarly limited distribution? If so, why? One example of perceived upwardly mobile “wannabe” status pertinent here is the “coloured abollai of printed fabric”;46 coloured and presumably multicoloured cloth is emphasised here. The less expensive “abollai and kaunakai […] with no adornment and modest in price” imported specifically “for the king”, perhaps also had been purchased to order and possibly intended for distribution to his servants, soldiers or less important vassals farther inland,47 for whom a gift from the king himself outweighed its modest quality. The two different sources of kaunakai, these from the Mediterranean world and the other specifically from the Ariakê interior of India, might also imply differences in quality and perhaps price. While little cotton cloth has survived within Aksum,48 secondary evidence for Aksumite textile spinning or weaving includes the artefacts we call “spindle whorls” common on habitation sites.49 Surely textiles were produced, if only for the “ordinary” population for whom even “modest in price” imported cloth lay beyond their means or access. Or were they, or at least some of them, limited to wearing only locally-obtainable leather? If so, the wearing of cloth would have been a social, financial or hierarchal statement in itself. At the highest end of the social spectrum, the Aksumite emperor (likely Kaleb) was wearing “gold-threaded linen clothing […] [and] turban” when he met the Byzantine ambassador Julian in ca 530/531 AD.50 Was this cloth Aksumite or imported? And did he wear it specifically as a statement to Julian and his entourage, for whom wearing “cloth-of-gold” was very much a statement of rank?

16Another royal statement may be implied in a unique cast brass rondel depicting a rayed head in portrait view recovered in the “Tomb of the Brick Arches” at Aksum. It follows a decidedly classical model, but the Ge‘ez text around its edge confirms its Aksumite origin and later 3rd century AD date (fig. 3).51 What was its inspiration here? Possible sources include several classical and Roman portrait images surrounded by splaying wavy hair or radiating lines, including Medusa, whose head adorned aegides of Athena and Jupiter, and Roman military phalerae and medallions,52 and an anthropomorphic sun deity such as Helios who was also appropriated in Meroitic Nubia.53 This failed casting likely was intended to be placed at the top of a stela, but whether another one was produced or actually positioned is unknown.54 If so, flashing in the sun to be visible to all onlookers for a considerable distance, it would have made a strong political or empirical statement directed specifically towards the entire onlooking population. Whether or not they associated the portrait image with the Mediterranean world, it was recognisably different from the usual profile image and at minimum would be associated with their king and/or possibly an indigenous deity, not necessarily a sun god.

Fig. 3 – Disc (“rondel”) with face in frontal view, brass, dated by palaeography to the late 3rd century AD, in late 4th century context, dia. 38 cm (Phillipson 2000, p. 100, fig. 80; British Institute in Eastern Africa).

17King Ezana’s conversion to Christianity ca 350‑360 AD also is a royal statement that, in this case, was deliberately reflected and widely disseminated on his coinage, undoubtedly on his command. The Christian cross simply replaced the pagan “crescent-and-dot” symbol embedded around the coin edge above the king’s portrait and the text of his name, title or epithet on his gold coins (fig. 4), and thus continued to proclaim royal and state religious affiliation following the existing coinage design. The Aksumite cross design itself, from its earliest representations on Ezana’s coinage, is equal-armed with blunted triangular (cross pattée) terminals that is very uncommon on Roman coinage, perhaps then a deliberate choice on his part. Neither presentation nor positioning reflects 4th century Roman coinage design;55 he and his successors adopted only the presence and direct statement of the cross from the Roman world. Christian crosses otherwise do not regularly appear on “ordinary” objects, mainly pottery, until sometime in the 5th century, both pre-fired stamped or moulded designs – but also incised post-firing and presumably often an owner’s choice at any point after the vessel was made. The latter consist mostly of widely varying creatively individual designs, many far removed from the formal cross pattée ultimately based on the equal-armed cross with terminals of Ezana’s coinage. Found over much Aksumite territory, they would indicate a gradually increasing and possibly piecemeal acceptance of Christianity. Christianity apparently was not fully integrated into common Aksumite society for well over a century after Ezana himself converted, although evidence of Christianity is apparent.56 No Christian cross appeared either in the élite “Tomb of the Brick Arches” at Ezana’s capital in about his or his successors’ lifetimes, or the Maryam Anza cemetery farther east for well over half a century after his conversion. The origins and meaning of this symbol, at least, was well-known to the population, even if they did not follow the religion it symbolised.

Fig. 4 – Aksumite coins illustrating different metals and religious affiliations; all images are to scale; a: gold coin of Endybis, r. ca 295‑310 AD, dia. ca 15.5 mm; b: selectively gilded silver coin of Aphilas, r. ca 310-early 320s AD, dia. ca 12 mm; c: gold coin of Ezana as a Christian, ca 360‑380 (?) AD, dia. ca 16.5 mm; d: selectively gilded bronze coin of Armah, r. ca late 530s‑550s AD, dia. ca 19.5 mm, shown to scale; note the “crescent and dot” symbol of the pagan deity Almuqeh above Endybis and Aphilas, and the crosses around Ezana and on Armah’s staff and reverse; regnal dates cited are those in Hahn, West 2016, pp. 10‑15, 17 tab., but variant dates are cited elsewhere by other scholars and in earlier publications (Hahn, West 2016, cat. 6, 40, 130, 558; Ashmolean Museum, Oxford).

18The Periplus, of course, does not mention the indigenous permutations inspired by the goods brought to Adulis by the merchants for whom it was written. Andrea Manzo’s excellent investigation of Mediterranean influences on Aksumite ceramic vessel forms, potentially “skeuomorphic” adaptations of Mediterranean metal and glass vessels, highlights the acceptance of both antique forms and individual elements or details and a preference for tablewares in the Aksumite world, from as early as the Early Aksumite phase within which the Periplus is dated.57 Whilst I am not entirely convinced of all his correlations, this paper highlights the extent to which Aksum – or at least some Aksumites – adopted and adapted certain Mediterranean vessel types, some disseminating more intensively than others, and not all for the same purposes as the original. What is interesting is that very few of the original forms actually have been recovered in the Aksumite world, even as fragments. As noted above, metal vessels are few and many would not have survived to be recovered. But glass vessels, or at least identifiable fragments of glass vessels, have been found in considerably greater numbers even if others would have been melted for reuse. The only identifiable imports in the “Tomb of the Brick Arches”, for example, are various glass bead types and some vessels/fragments. The latter are not represented in the Aksumite ceramic repertoire, and some glass vessels might be Aksumite products.

19How were Aksumite potters exposed to these foreign vessels, to enable them to choose to accept these particular features whilst rejecting others…? It is not implausible that the greater Aksumite majority who made, used and possibly admired these hybridised clay vessels would not have been exposed, even remotely, to the original imports or known anything of their foreign inspirations. It is equally plausible that at least some may not have been able to differentiate between an import and a local adaptation, but could differentiate between a local adaptation and an unadapted local product. Such widespread adoption of any individual feature would also indicate a general need or appreciation for it. The reasons for choosing and combining foreign elements at least tell us that each choice was individual, filling an existing need, ambition or even fashion within Aksumite society to greater or lesser degrees – while other (rejected) elements did not or were not. This deliberately limited form/function choice appears repeatedly in the Aksumite repertoire of both imports and indigenous comparanda,58 suggesting a common perception of intended functionality and highlighting the extent to which some permutations may reach. I would like to focus on further aspects of one such transmission. Roman glass bowls and dishes having a pronounced sinuous concavity underfoot and two ribbed “grip” handle appliqués tapering to either end59 are 1st‑2nd century AD in date and mostly found in the Eastern Mediterranean (fig. 5). This distinctive appliqué appears on a very limited range of later 4th‑5th century AD indigenous bowl and basin forms, as a new decorative variant within an already existing repertoire (fig. 6).60 Fragments recovered at Berenike and Jalame61 are residual in Late Roman contexts, that at Berenike suggesting possible importation to Aksum at that time although no import has been found here.62 Why do these hybridised vessels only appear over two centuries after the Roman types no longer were produced? Why not earlier? More significantly, in addition to generally direct “imitation” of the original appliqué element (fig. 6), the tapering ribbed appliqués on these same forms then develop individual permutations – one, two and even four elements can appear on the same vessel – and also are combined with indigenous painted and other grooved, appliqué and modelled designs.63 Their wide decorative variety within even individual sites suggests down-the-line description rather than direct or even closely indirect exposure within Aksum. This in turn suggests individualistic “tweaking” either by individual potters or their individual customers. Many, if not most, potters and other artisans would not have come close to seeing the original element on a glass vessel, and may not have recognised it if they had. Widespread acceptance and the variety of permutations for this one appliqué element highlights its deep infiltration within Aksumite society from the later 4th century AD onwards.

Fig. 5 – Eastern Mediterranean 1st‑2nd century AD blown glass vessels with tapering ribbed “grip” handle appliqués; all images are to scale; a, d: “plate”, greenish-bluish, height 2.5 cm, dia. 14.0 cm (Israel Museum, 77.12.98; Israeli 2003, p. 156; cat. 154, collection and photo of the Israel Museum, Jerusalem, by D. Harris, P. Lanyi, Z. Radovan, N. Slapak, M. Zalzberger); b: “bowl with grip handles”, transparent pale green, height 5.7 cm, dia. 14.6 cm (Whitehouse 1997, p. 332, fig. 93; 79.1.242, collection of the Corning Museum of Glass, Corning, NY; bequest of J. Strauss, https://www.cmog.org/collection/search); c: “dish-like vessel”, clear colourless, dia. ca 12 cm (Nicholson in Sidebotham, Wendrich 2000, p. 209, fig. 8‑10).

Fig. 6 – Aksumite clay vessels with variations of tapering ribbed “grip” handle appliqués and associated indigenous embellishments; bowl g most closely follows the original appliqué element presentation; all images are to scale; a, b, c, d, e: rim fragments of large bowls and basins (Munro‑Hay 1989, fig. 16.126, 120, 136, 124, 255); f, g, h: round-bottomed shallow bowls and large basin (Phillips in Phillipson 2000, fig. 44.c‑d, j); i: pedestal bowl (Munro‑Hay 1989, fig. 16.358; British Institute in Eastern Africa).

20Iron weaponry is a particularly intriguing import/non-import of the people, only very minimally adopted and some even rejected (fig. 7‑8). Both the single-barbed arrowhead and stone archer’s loose64 found at Aksum may be “war booty” as both are standard Nubian types but only one of each has been recovered within Aksum; neither were accepted into the Aksumite repertoire. The few other published arrowheads recovered within Aksum,65 some possibly misidentifications, likely also are individual imports as we have no documentary evidence for archery in Aksumite warfare66 – another potential but rejected technology. The extreme thinness of the only indigenous elongated, ribbed and waisted lancehead (68.8 cm in length) recovered within Aksum (again, at Aksum) renders it functionally useless, but it is comparable to functional Nubian lanceheads in generally contemporary élite/royal cemeteries at Ballana and El‑Hobagi.67 Their similarity strongly suggests this was a ceremonial symbol of authority adopted from Nubia, a statement of something by its owner to his audience, but not disseminated farther. Manzo68 has proposed a Nubian origin or inspiration for functional ribbed non-waisted lanceheads that are well attested in both cultures. If so, and in contrast to the waisted form, it seems Aksumites embraced, used and may even have adapted this far more common type as an effective weapon in warfare.

Fig. 7 – Waisted and ribbed iron lanceheads (redrawn to scale by J. Phillips from cited sources); a: Tumulus III, El‑Hobagi, Sudan, late 4th century AD, present length ca 52 cm, weight ca 12.7 g, scale is 10 cm (Lenoble et al. 1994, p. 84, pl. 16); b: “Tomb of the Brick Arches”, Aksum, late 4th century AD, present length 68.8 cm, weight 13.2 g, scale is 10 cm (Phillipson 2000, p. 113, fig. 93).

Fig. 8 – Iron arrowheads at Aksum and Adulis (redrawn to scale by J. Phillips from cited sources); a, b, c, d: four “bay-leaf” blades from Adulis, length 4.8, ca 11.5, ca 10.4, ca 6.9 cm, scale is 5 cm (Zazzaro 2013, p. 82, fig. 14.16.1, lower row, 3 from left); e: from Aksum, triangular bronze blade with two barbs and iron haft, length 5.6 cm, one barb broken off, context 7th century AD, scale is 5 cm (Phillipson 2000, p. 406, fig. 356.b); f: single-barbed blade, likely Post-Meroitic import, length 7 cm, barb broken off, scale is 5 cm (Zahn in Krencker 1913, p. 221, fig. 437.113); g: barbless “laurel-leaf” blade, length 6.5 cm, tang missing, context 5th century AD or possibly later, scale is 5 cm (Munro‑Hay 1989, p. 222, fig. 15.127).

21There are many potential reasons for the acceptance of imports – goods, ideas, elements – into an indigenous society. People, even in ancient times, are individuals within their own culture, each having a different range of perspectives within the limitations of their own personal circumstances and boundaries, and occasionally beyond. We often fail to remember, or at least fail to consider, this point in the midst of modern research that can only provide a general blanket perspective. Not everyone was exposed to foreign influence first-hand, the percentage of any given population in any given ancient culture exposed at second- or even third-hand to imports was variably but extremely limited. Whether or not fully understood, each import or an element of it assimilated must have filled, in however small a way, a previously existing need within Aksumite (or any other) society – or at least some of it initially – for permeation even to begin. It may simply be someone liked it, and then someone else, and so on – or not, depending on how far it permeated into the society after its initial acceptance, before, and if, it went no further. By considering specifics individually, the extent and variety of their popularity, different contextual circumstances, and relationships with material already embedded into existing cultural traditions, it is possible to extrapolate variable degrees of influence in the acceptance and use of imports, the different scenaria within which they were used, the variety of roles they played in that society, and how individuals perceived themselves and their role(s) within it. In many senses, we must speculate based on existing knowledge but, if we don’t, we will not progress to further understanding the cultures we study, and their multidimensional relationships with others around them.