Review of Coins and Samian Ware

It is no secret that samian ware plays a key role in dating the archaeology of the western Roman Empire; the ubiquitous brown, orange and red-slipped sherds are distinctive and abundant enough for detailed study, which has in turn yielded valuable information concerning chronologies of production and deposition from the 1st  century AD onwards. Since dates inferred from samian deposits have obvious implications for the dating of other artefact classes and sites, it is clear that a critical reassessment of the foundations of samian chronology is a task of considerable importance for Roman archaeology more generally. Addressing this task is one of the main aims of this monograph, a much anticipated revised edition of Anthony King’s 1985 PhD study of the chronology of samian ware in the northwest Empire during the late 2nd  to mid 3rd  centuries AD.

It is no secret that samian ware plays a key role in dating the archaeology of the western Roman Empire; the ubiquitous brown, orange and red-slipped sherds are distinctive and abundant enough for detailed study, which has in turn yielded valuable information concerning chronologies of production and deposition from the 1 st century AD onwards. Since dates inferred from samian deposits have obvious implications for the dating of other artefact classes and sites, it is clear that a critical reassessment of the foundations of samian chronology is a task of considerable importance for Roman archaeology more generally. Addressing this task is one of the main aims of this monograph, a much anticipated revised edition of Anthony King's 1985 PhD study of the chronology of samian ware in the northwest Empire during the late 2 nd to mid 3 rd centuries AD.
The book follows a logical structure, divided in spirit into three sections. The first section, comprising chapters 1 and 2, provides a critical review of traditional approaches to samian chronology. The traditional chronology of the late Gaulish samian industry is introduced in chapter 1, where emphasis is placed on the reliance on independent historical dates and their subsequent relationship with archaeological and historical narratives of industrial decline. 1 Chapter 2 outlines the essence of this approach in more detail; a relative production chronology based on stylistic studies, tied to a relative depositional chronology based on seriated assemblages, in turn connected to an absolute depositional chronology based on the presence or absence of particular production types at sites dated by documentary references. The practical and theoretical limitations of the method are considered in detail, leading King to propose a system of absolute dating based on depositional associations between coins and samian sherds.
The second section of the book, chapters 3 to 5, provides a methodological framework for tying these deposits to absolute dates. The framework developed in these chapters provides an important attempt to extend the potential of coin dating beyond the straightforward termini post quem used so frequently in site and artefact chronologies (c.f. Davies, Hall and Milne 1992; Egan 2010). King focuses instead on likely dates of deposition, which might considerably postdate termini post quem for reasons well-known to applied numismatists (Lockyear 2012). Since this transition requires an assessment of the probable circulation-lives of different coin types, King turns to hoard evidence, marshalling an impressive dataset of several hundred hoards containing silver and bronze coins closing in the mid-2nd to mid-3rd centuries. In chapter 3 this evidence is used to assess coin circulation-lives in aggregate, regional-, pia Andrews: Review of Coins and Samian Ware Art. 9, page 2 of 3 period-and denomination-specific groups, resulting in a number of probability estimates for the dates at which particular coin types are most likely to have been 'lost' and entered archaeological deposits (pp. 23, 29). In chapter 4 a database of more than 200 coin-associated samian deposits from the western Empire is assembled, grouped by their probable extent of redeposition and rates of accumulation; in chapter 5 they are cross-referenced with the loss-date estimates to produce estimated deposition dates for distinct samian types, complemented by a cluster analysis of deposit associations between kiln centre and potter groups. The result is a tentative absolute chronology for samian deposition which, when combined with estimated life-expectancies for the pottery itself, produces a revised kiln-centre chronology (p 102). This bears some important differences with tradition -most notably a Central Gaulish industry active into the mid-3 rd century AD. The results of the extended chronology are drawn together in chapters 6 and 7 to propose a model for the decline of samian ware stemming from changes in the distribution network, purchasing power, and a shift in consumer preferences in the western Empire.

Review of Coins and Samian Ware
Although the discussion of the decline of the Gaulish industries in chapters 6 and 7 will likely meet much agreement from samian specialists, the controversy provoked by King's late dating of the Central Gaulish ware will no doubt provoke as much contention now as it has previously (c.f. Willis 2005, 5.8.5). It is therefore beneficial that considerable space has been devoted to data presentation; the book is lavishly illustrated with no fewer than 62 figures in 134 pages of text, while 144 pages of appendices provide an exhaustive quantity of numerical data and a most welcome summary gazetteer of the deposits analysed in chapters 4 and 5. While this may go some way towards answering the critics, it does little to address the fundamental question of how depositional dates are transformed into production dates, where the author treads on ground as difficult as that of the traditional chronology. King's new production chronology assumes that the main period of samian disposal took place some 20 or so years after production, and that reuse of moulds and the storage of older pots in warehouses and shops did not extend this life too much (p 101). Given the tendency for samian to occur as a curated or residual find in late deposits this assumption might be very optimistic indeed (Wallace 2006); it is possible that the purported longevity of the Central Gaulish industry might simply reflect the extended use-life of the wares, although King presents convincing independent evidence to bolster his argument in the introductory preface (p 1).
A few additional minor issues with the study should also be addressed. Throughout the thesis coins are assumed to enter deposits primarily through 'loss', although there is little discussion of what this actually means; consideration of the means by which coins enter archaeological deposits, drawing on recent object-biographical approaches in applied numismatics (Kemmers 2006; Kemmers and Myrberg 2011), would have been of considerable benefit to this revised volume. In addition, there is a frustrating tendency for graph axes to be ambiguously labelled or unlabelled (e.g. Figs 5.3 -5.17 and Figs 6.4 -6.5), although fortunately this does not detract significantly from the content of the work itself.
Although this volume -excluding appendices -is fairly slim, it is clear that the contents represent an important and robust study which, limitations notwithstanding, ought to stimulate much debate in Roman archaeology and artefact studies for years to come. The value of the work is extended by a potentially transferable methodology; it is easy to identify several avenues within Roman archaeology for exploring artefact chronology using King's methods, with amphorae and brooches standing out particularly prominently. Whether it could be adopted wholesale beyond the Roman Andrews: Review of Coins and Samian Ware Art. 9, page 3 of 3 period, however, is more doubtful; few artefact classes from medieval England, for instance, have been subdivided and classified with the same degree of precision as samian ware, and its coins are far less abundant as site finds than their Roman equivalents. Yet we cannot fault an innovative approach for difficulties in applying it to periods and finds it was never designed to address; instead the study should be seen as a welcome contribution which might inspire new detailed approaches to the dating of a diverse range of artefact classes.

Notes
1 At the time of original authorship in 1985 it was generally assumed that decline resulted from external historically-documented catastrophes rather than processes of economic or cultural change. This attitude has changed significantly in subsequent years.