TESTING TIMES: AN EVALUATION OF THE RADIOCARBON CHRONOLOGY FOR EARLY CERAMIC VESSEL PRODUCTION AT UST’-KARENGA

This paper presents the initial results of a new phase of absolute dating at Ust'-Karenga. Three Optically-Stimulated Luminescence (OSL) dates were obtained on quartz grains extracted from Ust'-Karenga type ceramic sherds from Layers 4, 6 and 7 at Ust'-Karenga XII. These dates are used to test the reliability of the existing radiocarbon sequence and evaluate counter claims that sought to reject early dates for ceramics in the Transbaikal on the basis of a putative carbon cycle anomaly in the Transbaikal region. Our results strongly uphold the excavator's original interpretation of the site and independently confirm both the Late Pleistocene age and the long duration of the Ust'-Karenga pottery phase. The paper demonstrates the value of using independent absolute dating methods to test contested radiocarbon chronologies.

In many cases, early dates for ceramics, once published, are accepted or rejected by scholars with no clear rationale on either side. As a result, across Eurasia, key assemblages from this early phase of hunter-gather pottery production hang suspended in space without a generally accepted chronological context. The confusion that results from this continuous wrangling over dates makes it difficult to consider any broader patterns and, at a local level, effectively stifles discussions about the character of early pottery and its place in the lives of the people who made and used it. For a scientific discipline like archaeology, this is not a position that can endure.
Opposition to existing chronologies usually crystallises around the security of association between radiocarbon dates and the ceramic material they purport to date, and in many cases, this is a justifiable concern. Researchers have attempted to bypass the problem by directly radiocarbon dating surface residues or organic 'temper' from within the ceramics themselves, these approaches come with their own challenges. More importantlyat least for the purposes of this article -the testing of existing radiocarbon data with new radiocarbon dates fails to escape another common criticism from researchers who attribute archaeologically unacceptable dates at a regional scale to 'anomalies' in local carbon cycles and systematic errors in our estimates of age. Clearly, alternative approaches are required.
In cases where the difference between radiocarbon dates and the 'acceptable' age of the material is small the problem can be difficult to resolve without further excavation, well contextualized dating evidence, and a clear understanding of local carbon circulation patterns, reservoir effects and so forth. Fortunately, the impact of such cases is relatively minor, especially in earlier periods where chronological boundaries are imprecise. The impact of these problems is far more keenly felt where the discrepancy in date is in the order of millennia. In these cases, while the implications for archaeological interpretation are immense, the solution is potentially more ISSN 2415-8739 Reports of the Laboratory of Ancient Technologies Vol. 13 No. 1 2017 33 straightforward. When chronological interpretations are sufficiently divergent it becomes possible to verify existing chronologies with non-radiometric dating techniques, such as Thermo-Luminescence (TL) and Optically Stimulated Luminescence (OSL). Although they are often significantly less precise, these techniques can provide us with reliable chronological data that are entirely independent of the carbon cycle, enabling us to test both the position of existing sequences and the general coherence of the stratigraphic context.

A Contested Chronology: The Ust'-Karenga
Complex This paper applies this alternative approach to one of the most hotly disputed early ceramic finds in East Asia, the site complex of Ust'-Karenga, where pottery fragments have been 14 C dated, on the basis of both associated charcoal and organic 'temper' within the ceramics, to c. 12,200-10,500 calBC (Kuzmin 2006;Kuzmin and Vetrov 2007).
At the time of writing, the site of Ust'-Karenga in the Upper Vitim Basin (Fig. 1) is the most thoroughly dated early pottery site in Eastern Siberia, and even the briefest survey of the literature would be sufficient to see that chronology has been the focus of almost every paper published about the site in the last twenty years. With new radiocarbon data, ac-cessible information about earlier pottery finds in China and Japan (e.g. Keally et al. 2004;Kuzmin 2006;Wu et al. 2012;Zhao and Wu 2000) and a widening range of comparably dated sites in surrounding regions (e.g. Derevianko et al. 2004;Shewkomud and Yanshina 2012;Zhushchikovskaya 2005), it would be reasonable to assume that the chronology of Ust'-Karenga and other putative Late Pleistocene ceramic assemblages in the Transbaikal would have become increasingly secure. However, this has not been the case, and the dating of these sites and their ceramics continues to be regularly challenged.
Currently, the most complete discussion of the chronology of the Ust'-Karenga complex can be found in the proceedings of a regional conference held at Ulan-Ude (Vetrov, 2010). This paper was written as an indirect reply to two publications by a well-respected Palaeolithic archaeologist M.V. Konstantinov (2009a;2009b) in which it was suggested that the proposed phenomenon of early pottery in the Transbaikal was 'unsupportable'. These papers, which represent the tip of an iceberg of contention, very rarely expressed in publication, target their criticism at what their author deems the naïve and uncritical reliance on radiocarbon data among archaeologists (Konstantinov 2009b). He goes on to argue, quite rightly, that radiocarbon dates must be under- stood with reference to the typological and stratigraphic context of the finds they purport to date.
With specific reference to the position of the early ceramic finds within the sequence at Ust '-Karenga, Konstantinov (2009b: 190) suggests that the sediments described in Vetrov's papers are more in keeping with deposits from the latter half of the Holocene Climatic Optimum, while the ceramics themselves are typologically consistent with Middle Neolithic Bel'kachinsk culture finds in Yakutia (c. 4000-2600 calBC 1 ). Unfortunately, attempts to assess the validity of these claims are thwarted by the fact that no new results or specific evidence have been presented to explicitly support or refute these counter claims.

Stratigraphic situation
Usually taken as the 'type' profile for the Ust'-Karenga complex as a whole ( Fig. 2a and b), the stratigraphic sequence at the adjacent sites of Ust'-Karenga XII, XIV and XVI is located in the sediments of a 20-25m terrace at the mouth of the Karenga. The sequence can be split into two distinct geomorphological phases: subaqueous and subaerial. This stratigraphic sequence outlined below was remarkably consistent across the body of the terrace body, although the preservation of the lower cultural layers in different locations was affected by the topography of the underlying bedrock and proximity to the ancient river channels.
The subaqueous phase, which accounts for the larger part of the sequence extending from the bedrock to around 50cm (or less) below the modern surface, is composed primarily of finely laminated alluvial sands with lenses of silt and bluish grey clay (Ineshin 1979;Vetrov 1992). This phase contains four cultural layers which have revealed the earliest evidence of human occupation at Ust'-Karenga (cultural Layers 8a, 8 and 7a) and the earliest ceramic vessels in the region (Layer 7) -encountered at Ust'-Karenga XII, XIV and XVI (Fig. 3). These cultural layers are clearly visible as darker strata, thicker than the laminated sediments and sterile sands that surround them

Fig. 2: a -Map showing the location of the main sub-sites of the Ust'-Karenga complex; b -Generalised stratigraphic section derived from Ust'-Karenga XII (after Vetrov 2005) Рис. 2: a -карта, показывающая основные пункты усть-каренгского комплекса; b -сводная стратиграфическая колонка с местонахождения Усть-Каренга XII
(по: Ветров, 2005) ISSN 2415-8739 Reports of the Laboratory of Ancient Technologies Vol. 13 No. 1 2017 35 (Fig. 2b). At the upper interface with the subaerial sediments, the subaqueous sequence appears to be truncated and is certainly scarred by extensive polygonal frost-wedge formations, which locally deform the well-defined stratigraphy below. Another significant phase of cryogenic activity is also apparent in the alluvial phase below the cultural layers (Ineshin 1979;Vetrov 1992). There is some disagreement in the literature about which of these cultural layers constitute true paleosols and which do not. However, nomenclature aside, it seems to be generally accepted that the darker layers (which are presumed to be more humic in composition) are correlated with for periods of stability when the climate was comparatively warmer. There is also general agreement that the cryogenic features provide natural chronological brackets which can be used to constrain the dating of the cultural layers (Ineshin 1979;Konstantinov 2009a;Vetrov 1992). Dispute arises because, in the absence of other chronological evidence (or distrust in its validity), a number of equally plausible interpretations of this sequence can be made. This position is hardly unique in the archaeology of Eastern Siberia, but unlike many other early pottery sites where the association of absolute dates, material culture, and stratigraphy has been legitimately questioned on the basis of various forms of postdepositional disturbance (see McKenzie 2009), the cultural layers of the subaqueous sequence at Ust'-Karenga, including the earliest 'ceramic-bearing' layer (Layer 7) are conveniently delimited from the upper layers by substantial accumulations (0.4-1.0 m) of archaeologically sterile sediments (Kuzmin and Vetrov 2007;Vetrov 1992). Though the impact of more recent cryogenic disturbances is significant, it remains spatially discrete, leaving large areas of the lower levels of the site effectively in situ. In short, it is extremely unlikely that significant mixing of the upper (1-6) and lower cultural layers (7-8a) could have occurred, and the stratigraphy at Ust'-Karenga XII, as a whole, appears to be a promising context in which to explore the absolute dating of this period of Siberian prehistory. The sediments of the subaerial phase appear more homogeneous and are likely to have been formed by the drifting of unconsolidated sediments as much as by the action of periodic flooding. Within these deposits, clearly defined soil horizons (ancient and modern) are discernible and six cultural layers (Layers 6-1) have been distinguished on the basis of colour, texture, and associated material culture (Kuzmin and Vetrov 2007;Vetrov 2010). Ust'-Karenga pottery is also found in these subae- rial sediments -predominantly in Layers 6-4 (with occasional sherds in layers the upper layers), but whereas in Layer 7 it occurs in isolation, in these subaerial layers it is found alongside other ceramic types. In part, this blurring of cultural layers is a result contamination between these upper layers. This makes it even more important obtain direct dates for ancient events.

An Alternative Approach
Recent developments in OSL measurement technologies have widened the scope of this technique (Huntley et al. 1985) allowing it to be applied to smaller sample sizes and a wider range of materials, including pottery (Hood and Schwenninger 2015). This, therefore, provides the possibility of obtaining an absolute date on the production of the vessel. This technique is not usually sufficiently precise for this purpose, but as the aim of this study was to independently test the validity of a radiocarbon sequence, this was not a significant concern.
Ideally, for OSL analysis to have the greatest possible precision, we would rely on freshly excavated material, with directly associated sediments (which could also be dated), in situ dose rate measurements, and precise information about depth below surface, water content of the associated sediment, and postexcavation storage conditions. In spite of this, it was decided that this study would focus on existing collections of material.
There were several reasons behind this decision, including the practical difficulties of conducting expeditionary research in this remote region. However, the main reason was to allow us to evaluate a realistic research model that could be applied to similar chronological disputes in other regions. Many of these potential study locations are also remote from major cities, and though some are still a focus of research activity, many have already been excavated or otherwise destroyed. For this technique to be a viable way of testing existing dating sequences it must be able to incorporate curated material. Equally importantly, given the comparative rarity of this early ceramic material, the technique would need to be minimally destructive.

Sampling strategy and OSL Analysis
For this study five ceramic samples (Fig. 4) were selected from defined archaeological contexts. The samples were all typologically attributable to the Ust'-Karenga culture and chosen from petrographically defined groups with coarse inclusions dominated by quartz and quartz-rich rock fragments (primarily granitic in origin) (see Hommel et al. in press). Two samples were taken from the earliest ceramic-bearing layer (Layer 7) and two further samples were taken from the boundary of the overlying sterile alluvium and the low- est subaerial layer (Layer 6). A final sample was taken from Layer 4 which represents the uppermost stratigraphic layer in which material attributable to the Ust'-Karenga culture is routinely recovered in secure context. On the basis of optically stimulated luminescence measurements (OSL) of sand-sized quartz (125-180μm) extracted from the sherds, a series of three age etsimates was obtained. The extraction of quartz grains was carried out using standard preparation techniques including dry sieving, HCl (10 %) treatment to remove carbonates, HF treatment (48 %) to dissolve feldspathic minerals, heavy mineral separation with sodium polytungstate and final re-sieving of the treated mineral fraction. Measurements were performed in an automated Risø luminescence reader (Bøtter-Jensen, 1988; using a SAR post-IR blue OSL measurement protocol (Murray and Wintle 2000;Banerjee et al. 2001;Wintle and Murray 2006). Dose rate determinations are based on the concentration of radioactive elements (potassium, thorium and uranium) within the sherds (internal beta dose rate) as well as a representative sediment sample from Layer 7 at Ust'-Karenga XII in order to assess the external gamma dose rate. It was not practical as part of this study to undertake infield measurements at the site, so a large systematic error of 10 % was attached to the latter in order to account for any uncertainty. The dosimetric analyses were derived from elemental analysis of the samples (ceramic and sediments) by ICP-MS/AES using a fusion sample preparation technique. The final OSL age estimates include an additional 4 % systematic error to account for uncertainties in source calibration and measurement reproducibility. Dose rate calculations are based on Aitken (1985). These incorporated beta attenuation factors (Mejdahl 1979), dose rate conversion factors (Adamiec and Aitken 1998) and an absorption coefficient for the water content (Zimmerman 1971) based on a mean moisture content of 5 to 13 %. The contribution of cosmic radiation to the total dose rate was calculated as a function of latitude, altitude, burial depth and average over-burden density based on data by Prescott and Hutton (1994). The high palaeodose values for the samples reflect the antiquity of the prehistoric sherds but are mainly due to the high environmental dose rates, ranging from 3.7 to 8.5 Gy/ka. Whereas the sediment contains concentrations of radionuclides (K=3.4 %; Th=4.5 ppm and U=1.3 ppm) which may be regarded as normal, the same is not true for the clay fabrics which were all found to contain elevated concentrations of potassium (2.6-4.0 %), thorium (8.9 to 192.0 ppm) as well as uranium (3.2 to 32.0 ppm). A priori, there is no reason to question these values, nor the veracity of the calculated age estimates, but it is worth noting that these are unusually high levels of activity.

Discussion
Although it was necessary to introduce substantial systematic errors into our calculations -due to small sample size and the impracticality of conducting in-field measurements of environmental dose rate and sediment moisture content -the OSL analysis provided broad probability distributions for the production date of three ceramic fragments (Table 2). These results are consistent with the stratigraphic position of the ceramic samples analysed and span the expected range of the Ust'-Karenga culture (as estimated from calibrated ranges of existing radiocarbon analysis) (Table 1; Fig. 5). Of course, the correlation is far from perfect, but it was never expected that the results of these analyses would allow us to refine the existing chronology. Instead, the aim was to consider the general trend of dates obtained directly on ceramic material from across the stratigraphic section and to test the general position of the radiocarbon series using a fully independent dating technique. Critically, the aim was to use these results to evaluate two discordant interpretations outlined in the literature (Vetrov 2011). If we plot the OSL dates together with the ranges expected for both of these interpretations then it becomes immediately clear which is the more probable (Fig. 6). Future work on the dating of the site will allow us to further to confirm these results, ideally based on both new OSL dates on ceramics as   Vetrov 1995Vetrov Ветров 1995 No dates available Данных нет well as sediments, and new radiocarbon series (in clear stratigraphic relationship). Alongside the dating itself, it is vital to consider the environmental, climatic and cultural context in more detail. One of the principal criticisms levelled by Konstantinov (2009a) at the current stratigraphic interpretation, is that if the sequence were indeed attributable to the late glacial period, it would shows a series of four discrete phases of warming/stability (Layers 8a, 8, 7a, and 7). He considers this to be difficult to explain. Yet many interpretations of the pattern of late glacial climatic change suggest that this kind of multi-phase process should be expected in well resolved alluvial sequences such as this (see Ellis et al. 2004;Yu and Eicher 2001). If the sediments at Ust'-Karenga represent such a sequence, then the Upper Vitim presents an ideal opportunity to study human

Fig. 6. Showing the OSL results for Ust'-Karenga ceramic fragments from Layers 7, 6 and 4 plotted against (A) date range (calBC) of radiocarbon results from Ust'-Karenga Layer 7 and (B) the 'accepted' age of the Ust'-Karenga ceramic phase in Konstantinov 2009a and 2009b. Рис. 6. Демонстрация ОСЛ результатов анализа керамических фрагментов из культурных горизонтов 7, 6 и 4, наложенных на (А) диапазон радиоуглеродных дат (калиброванный возраст, лет назад), полученных по 7 культурному горизонту Усть-Каренги и (В) «допущенный» возраст
усть-каренгской керамической фазы (Константинов, 2009а, 2009б) adaptation to catastrophic environmental change. This would certainly require further fieldwork, perhaps at a significant scale. If we are to take archaeological context into account, as Konstantinov (2009b: 190) rightly requires, we need look at the specifics of the assemblage as a whole within a wider regional context. With this in mind, it is worth noting that the lithic assemblage associated with the early ceramics at Ust'-Karengawhich is based around multi-purpose bifaces and microblade production -is entirely consistent with the lithic industries found at other late glacial/early post-glacial sites in Eastern Eurasia (Vetrov 1995b;Ineshin and Tetenkin 2017). Perhaps more obviously significant is the fact that several early ceramic traditions in the Amur Basin have also produced secure Final Pleistocene/Early Holocene dates (e.g. Derevianko and Dorj 1992;Derevianko et al. 2004;Kuzmin and Jull 1997;Kuzmin and Orlova 2000;Shewkomud 2005;Zhushchikovskaya 2005). Some of the material shows clear technological relationships with the early ceramics of the Transbaikal, in general, and Ust'-Karenga in particular (Hommel in press;Shewkomud 2005). This, too, offers considerable strength to the excavator's interpretation of the site. Further support for early ceramic sites in the Transbaikal is found in the consistent results from a recent re-dating of a problematic sequence at Studenoye and new research at the site of Krasnaya Gorka (Razgildeeva et al. 2012;Tsydenova et al. 2017). Our research suggests that there is no systematic reason to expect these radiocarbon dates to be problematic, though it would be interesting to extend our evaluation to some of these contexts as well.

Conclusion
While it is important to maintain a critical stance in the face of scientific data, and while it is always essential to consider all available archaeological, environmental and stratigraphic contexts. It is vital that new data is presented to support or challenge existing interpretations.
In the case of Ust'-Karenga, errors vocally attributed to perceived problems with radiocarbon dates at a regional scale have been refuted in this paper by applying an absolute dating technique based on independent physical phenomena. While the broad probability ranges calculated for the dates leaves plenty of room for further research and discussion, the consistent correlation between luminescence dates and radiocarbon results strongly supports Известия Лаборатории древних технологий Том 13 № 1 2017 Reports of the Laboratory of Ancient Technologies Vol. 13 No. 1 2017 ISSN 2415-8739 42 the latter's validity. Although there is no doubt that the chronology of early Neolithic sites needs to be further refined. It is hoped that future discussions will rest on scientific data and that other forms of investigation into the character of life in the past will become an equally important focus. chaeology and the History of Art at Oxford for its institutional support.
Finally, we would like to acknowledge the research contribution to the field of early pottery studies made by our co-author Dr V.M. Vetrov, whose death in November 2015 came just a few months after the decision to publish this paper was made. Though there remain many questions to address, the new dates provide support for a position that he maintained for much of his career, often against vocal opposition. While sentimentality is somewhat out of place in a scientific journal such as this, we consider this result to be a fitting tribute to his memory. Инешин Е.М. Геохронология и материальная культура неолита Усть-Каренгского археологического комплекса // Неопубликованные тезисы / Иркутский государственный университет, 1979. Jordan, P. and M. Zvelebil, M., 2009. Ex oriente lux: the prehistory of hunter-gatherer