CLASSIFICATION PROBLEMS OF EASTERN MEDITERRANEAN AMPHORAE : REVISIONS ON AMPHORA TYPOLOGIES

Associate Professor, Dr Ahmet Kaan Senol

Typological studies of amphorae and their contents came into prominence after it was accepted that such studies could act as indicators of the ancient economy. Although many scholars tried to construct new typologies after the publishing of E. Dressel’s long-lived table, many of these efforts resulted in divergent typologies, incomplete tables and an excess of terminologies. As the preparation of these studies depends on the accessibility of find contexts, museum collections, regional and/or chronological limits have caused plenty of problems in recent recording methods, and so we should question the utility of the habitual classifications.

Present situation :
Today, most of the identified classes have more than ten different names generally attributed by publications, authors or excavation sites. On the other hand, unidentified forms are still considered as absent in dispersion diagrams because they are not compiled in catalogues until now. Therefore, while we have lots of names for well-known types, many of the amphora forms are nameless because we do not know their production centres. As this artificial absence of unidentified classes causes lots of troubles in data logging, researchers utilise the catalogue numbers of similar published forms with the names of the excavation sites or the authors in order to distinguish the new forms. In some cases, this application system is very beneficial for the statistical methods as in accepted international publications such as Athenian Agora or Pompeii finds. But how should the researchers integrate the new groups into the data banks? Will they continue to put new numbers by using their names? We will discuss these problems later on.

Among the documentation problems, standardisation has a vital importance. Classification problems have been solved partially for Roman Imperial amphorae and some African groups, but the classification tables of the amphorae produced during the Archaic, Classical and the Hellenistic periods are still deficient. At the same time, the classification of Byzantine amphorae is still in a draft state. Therefore, we are still far away from presenting complete amphora dispersion tables. Now we have to ask if the recent statistical methods reflect the distribution patterns? Undoubtedly, different separation methods of amphora shards affect the statistics due to their error rates. I will not discuss the results of diverse statistical methods but I will focus on the labelling problems of amphora groups in this paper.

Amphorologists have accepted the legitimacy of different classifications and have used their labelling systems in their data-bases and amphora dispersion tables. The major disadvantage of this attitude is the misinterpretation of economic movements in the case of the same amphora type that was produced in different centres. Eventually, the dispersion tables only reflect the contents of the assemblages without determining the production centres, which is essential for the construction of ancient economic patterns. Although wrongly identified and unidentified shards affect the validity of statistics, we should examine the rate of amphora imitation on dispersion tables.

The most reliable data for amphora classification are obtained from production centres, especially from the deposits or kiln sites. The main problem of labelling the Archaic and Classical Greek amphorae is the scarcity of discovered production centres. That is why scholars are obliged to use the amphorae found in the assemblages of the consumption centres, necropoleis or shipwrecks, which are far from representing the whole series or their origins, in order to construct the classification tables. Therefore, publishing many different classifications prepared from different contexts causes a multiplication of labelling which makes the data unserviceable. Moreover, we should not forget the existence of considerable amounts of imitations in these tables. Besides, much evidence of small-scale amphora production has not been included in data systems as it has not yet been denominated. The results of this deficiency of data can be seen especially in the diagrams of the Hellenistic contexts. Although scholars are able to separate the amphorae of lesser known groups, they have not been integrated into dispersion diagrams as their denominations have not been updated till now. Therefore, the effects of small-scale production on Mediterranean commerce cannot be confirmed sufficiently. Eventually, wrongly classified or non-classified amphorae and imitations constitute remarkable error rates in dispersion diagrams as we observe similar misinterpretations on the stamped amphora tables because of the uncertainty of the percentages between stamped and unstamped amphorae.

Although the typological studies of occidental amphorae made progress after the publication of Dressel tables, identified sub-groups or the production of new centres have scarcely been integrated into the dispersion tables. On the other hand, as many contexts and ensembles of oriental amphorae from excavations have been published, the presentation of the oriental amphorae in the dispersion tables still causes misinterpretations of the eastern Mediterranean commercial relations. The main problem of the classification of Oriental amphorae in data bases is the adequacy of the conventional typologies. Even the most utilized classifications do not answer the scholars’ need. Although the scientists are adding new sub-groups into the main classification tables, these new groups are not presented in the dispersion tables or data banks as their denominations are not standardised. Today, we should convert the LR typology to a new classification system covering the newly discovered workshops and regions. If we continue to denominate similar types without concern as to their production centres, we will continue providing erroneous data to studies of ancient commercial relations. For example, it is hard to distinguish the role of Cyprus, Syria and Cilicia in Mediterranean commerce, due to the dispersion diagrams prepared by the amphora ensembles as only one name covers these production sites. To avoid these confusions, scholars should use the regional names of the production centres in addition to their catalogue numbers, the surname of the first classifier or at least add their regions such as Cilician LR 1a, b, c, Syrian LR 1a, b, Cypriot LR 1a, b, c, Rhodian LR 1a, b, etc. At the same time, we should also re-interpret the presence of LR 5-6 shards in the dispersion diagrams as to whether they indicate Palestinian or Egyptian wine.

In order to avoid this misuse, we should construct a complete database accepting a new terminology using the manufacture sites and the sub-groups covering all known types and unidentified forms with their temporal denominations. If we take into consideration the artificial complexity of the classification of oriental amphorae depending on the usage of different typologies, this project should start with eastern Mediterranean amphorae. As hundreds of different groups should need re-arrangement, teams of amphorologists should study under chronological divisions in order to re-issue the permanent catalogues without neglecting any period. Thus, this new typology should start from the earliest amphorae to the contemporary vessels.

Although, Levantine and Egyptian amphorae are the earliest evidence of large-scale maritime commerce, neither of the production centres have been discovered, nor have their different classifications been unified in publications. Thus, the denominations of Levantine and Canaanite amphorae and the confusions of the typologies of their sub-groups cause troubles in data processing. Unfortunately, we cannot obtain the type, period or production region of even well distributed forms from the dispersion diagrams, as they only indicate the names of general groups like Canaanite amphorae, Levantine amphorae, torpedo or Syrian/Palestinian group. Scholars meet a similar problematic when using the data of Mycenaean amphorae as the denominations of their subgroups have not been used in dispersion diagrams yet.

After the early years of the first millennium, a distinctive influence of Aegean goods started to be felt in eastern Mediterranean commerce. Archaic amphorae of pioneer Aegean centres found in the excavations of emporions or consumption centres are the major evidence of this commerce. We have two main problems in drawing a conclusion from the data provided from diagnostic Archaic amphora shards; first is the classification problems of the amphorae, caused by the abundant production of trial versions during the fifth century BC before the city-states decided on their special forms. Today, we have hundreds of different variations of Archaic amphorae from excavations mostly along the Black Sea coasts and we are still far from categorizing their production centres as many of them are still consider as unidentified forms. The second problem is the different denominations of well-known Archaic amphorae and the adequate forms in the published catalogues.

Similar problems occur in the data processing of the amphorae produced from the fifth to the late fourth century BC. After the establishment of new Hellenistic consumption centres, South Aegean city-states converted their production and exportation systems to answer the great demand, due to the extension of Mediterranean commerce. During the boom of wine production, they all produced similar amphorae without hesitating to use the form characteristics of their neighbours. Thus, similar amphora forms produced in different production centres started to be used extensively in commerce. Among the amphora forms that cause confusions in classification, different typologies of the amphorae with mushroom rims and hollow feet need unification. We propose to classify these amphorae according to their production sites instead of denominating them as amphorae with mushroom rims. We may call them Knidian, Rhodian, Coan, Type 1.a,b,c etc. As we have not determined all of the production sites yet, we should label the sub-groups of mushroom rims temporally in order to supply valid data for statistical diagrams.

Another reason for adding the production centre to typological classification is the affect of the increased imitation trends in the Hellenistic Period. Today, we should consider how the shards of imitated amphorae cause misleading data in dispersion diagrams. Specialists should integrate the imitated forms into general classification tables with diagnostic denominations to avoid confusion and wrong interpretations. A general conformity about the classification of imitated amphorae should urgently be agreed upon.

The standardisation process of amphora forms in the Mediterranean Basin under the Pax Romana causes identification problems of similar amphorae produced in different centres. Although the problems of the production centres of DR 2-4 amphorae have been solved partially by using the name of the production centres before their labels, many amphora forms such as Zemer 41, Late Rhodian amphorae etc. need re-denomination. If we insist on using their habitual denominations in dispersion diagrams and databases, we will continue to repeat the problem as mentioned above.

Today, amphorologists need to integrate the new classifications and typologies of the Late Roman and Proto-Byzantine amphorae into general catalogues as a result of the new discoveries of amphora workshops and deposits which have been widely found in the Eastern Mediterranean zone. As we are able to distinguish the production centres of the shards of LR 1 to 3 amphorae more precisely than 20 years ago, why do we not use these important data in dispersion diagrams?

In addition, Proto-Byzantine and Medieval amphorae are scarcely used in statistical analysis in detail because of their denomination problems. Catalogued items in publications need confirmed classifications or at least unifications of different typologies as unidentified Medieval/Byzantine amphorae. If we consider the age-long usage of ceramic vessels in the commercial life of Eastern Mediterranean centres, a project of labelling the identified forms and constructing a basic classification with denominations of unidentified Oriental amphorae should be undertaken urgently.

In order to have access to valid interpretations about commercial patterns in the Mediterranean Basin, scholars should convert and integrate their data to optimise Geographic Information Systems. Moreover, the typological classifications of commercial amphorae should be updated and published periodically so as to provide accurate data for statistical analysis.