Here's the manuscript of my paper on fabric descriptions in ceramic analysis, presented at our meeting in Worcester in 2018 - hopefully it's of use and interest to many of you!
I came across a series of published fabric descriptions recently which did not seem quite right. They looked good and detailed the right markers we have recommended over the years in our three editions of The Study of (Later) Prehistoric Pottery: Guidelines for Analysis and Publication (1992; 1997; 2010); the sizes, shapes, sorting and frequency of identified inclusions according to Appendices 1-6. That was not the problem; one which lay instead in the interpretation of how these fabrics, as defined, had been created by their respective prehistoric potter(s).
Each fabric was labelled with the word “-tempered” attached to the principal inclusion identified as a name for that fabric; for example, “Limestone-tempered”. The actual definition of this particular fabric, however, stated that the limestone inclusions were subrounded in shape. And this is where the problem lies: subround inclusions in handmade, prehistoric pottery fabrics do not indicate that, in this case, the limestone had been crushed and added to clay in order to make the ‘fabric’ subsequently defined and labelled by the specialist as “-tempered”. In fact, it is the opposite that is most likely. These subround inclusions of limestone were probably naturally-occurring in the clay chosen by the potter to make their vessels. Had the potter gathered pieces of limestone and crushed them, the inclusions would have been very angular to angular in shape. Therefore, the presence of only subround limestone in that definition indicates that this is most likely to have been a fabric made from a clay with naturally-gritted limestone in it which had been selected by the potter. This indicates a type of expedient technology where a person, in this case the potter, finds something suitable that is readily available in the natural world to help solve a problem, i.e. to help make pottery vessels. In this case, no significant extra effort other than picking out the extra large pieces of limestone in order to ease the actual forming of the required vessel by hand would have been required. The major manufacturing stage of ‘creating the fabric’ by finding suitable nonplastic materials to crush, sort and add to a clay was not required and would have saved a considerable amount of time and effort.
I first came across this significant difference when working in the Caribbean on both handmade and bonfired prehistoric and colonial period Afro-Caribbean pottery in artefact assemblages recovered from archaeological excavations on the island of Nevis. Nevis is dominated by dacite, an intriguing type of intermediate igneous rock with a fine-grained texture bearing occasional phenocrysts and which is between andesite and rhyolite in composition; it is the volcanic equivalent of granodiorite. Dacite disaggregates readily into its component minerals through weathering (tropical sun, torrential rain) and degrades beautifully into the most practical of naturally gritted clays; really robust, red, quite rough clays with numerous pieces of plagioclase felspar and ferromagnesian minerals (hypersthene, hornblende, enstatite, augite), opaques (magnetite, haematite), rare quartz and quartzite, as well as fragments of dacite rock (Figure 1a-b). These inclusions are subangular and subround in shape typical of degraded rock outcrops which form residual clays nearby (Shepard 1954, 11). Large and small deposits of this gritted clay are easily accessible at ground level in several places on the island, which measures 9 x 12km. During the prehistoric (Wilson 1989, 2006) and colonial periods (Terrell 2005) as well as the post-emancipation and modern periods (Fay 2017), potters took advantage of this naturally-gritted clay to make cookpots, bowls and eventually leisure wares such as flower pots and hospitality pots for cooling water to give to visitors called monkey jars.
The important point of this presentation was that residual clays containing mainly disaggregated minerals with occasional pieces of a parent rock (whether igneous, metamorphic or limestone) which are suitable for the types of vessels that were being made are not the same as the more created or manufactured fabrics which were made with specific tempers which had been selected, altered and added by potters to clays for distinctive reasons, whether socially expected or practically required or both. Fabrics in Britain which contain mainly igneous and/or metamorphic rock that is angular in shape, such as the Malvernian Group A fabric (Peacock 1968, 415-21, plate XXXIII, a-b) and the Mountsorrel Complex variety (Knight 1992; Knight, Marsden and Carney 2003; Williams and Vince 1997), have few pieces of disaggregated components of that rock in the fabric, and those rock pieces which are present are angular in shape. These types of fabrics are most likely to have been created by potters crushing the rock, as potters did with calcined flint (heating the flint having been an additional stage of effort and planning prior to crushing) from the Neolithic through to the early Roman period, with potsherds for grog temper in the later Neolithic through later Bronze Age and with quartz/quartzite for large inclusions of same in Peterborough/ Impressed ware, for example. However, if the inclusions present whether igneous rock or limestone rock or shell-bearing rock and disaggregated pieces of shell from that rock are not angular in shape but rather subangular to subround or even round and have a majority of its inclusions being disaggregate components of a parent rock if igneous/metamorphic, then it is highly likely that these disaggregated inclusions were actually naturally-occurring in the clay which had been chosen by the potter to use.
Gibson and Wood (1990, 251-5, figs. 212-217) provide a really good reminder of what to look for with excellent examples if you haven’t thought about this topic in some time and need a refresher. The most important thing to remember is to be honest and open about what you know to be true and what you might not be sure about. Peacock (1968) declared that Group A Malvernian rock fabric was a tempered one. All signs point to this being absolutely correct. Knight, Marsden and Carney (2003) have also demonstrated that the granodiorite found in some Iron Age pottery made in Leicestershire was crushed and added as temper. However, we do not know about many of our sandy fabrics: are they naturally-gritted with quartz grains if they are subangular to subrounded or were they silty clays that were tempered with coarser grains from sedimentary deposits giving a bimodal pattern of fine versus coarse inclusions without any intermediary examples in the range of inclusion sizes within a fabric? Gibson and Wood give one example of each that they know to be true (1990, fig. 214/natural boulder clay with quartz sand and fig. 216/ Nottingham medieval pot fabric being a silty clay with added larger, nearly uniform sized quartz grains as temper), which is helpful.
And most importantly try to learn more about the nature of clays and sedimentary deposits which could have been exploited in the area where your assemblage was recovered. Is there a British Geological Survey geological memoir for that area available for you to examine, by any chance? These often rather elderly, small thin volumes are rich with information about what the solid and drift geology in an area is at a ‘walking-the-landscape’ scale of information including details about any different clays that have been noted. It is well-worth your effort to find out if one is available that relates to your site location. But above all be honest. Do not simply call something tempered if you cannot back this statement up with details about sizes and types of inclusions and an understanding of the likely geological deposit they derived from that justify using that term. If you are not sure, simply call a fabric ‘-gritted’ or ‘-bearing’, such as ‘quartz-gritted’ or ‘shell-bearing’.
Honesty is the best policy in fabric analysis and description. This does matter because what you write and publish tells us about the potters who made your vessels and the people and societies that used them.
Fay, P. J. 2017, Creole Clay; Heritage Ceramics in the Contemporary Caribbean. Gainsville: University Press of Florida
Gibson, A. and Wood, A. 1990, Prehistoric Pottery for the Archaeologist. Leicester: Leicester University Press
Knight, D. 1992, Excavations of an Iron Age settlement at Gamston, Nottinghamshire, Transactions of the Thoroton Society of Nottinghamshire 96, 16-90
Knight, D., Marsden, P. and Carney, J. 2003, Local or Non-Local? Prehistoric Granodiorite-Tempered Pottery in the East Midlands, in A. Gibson (ed.), Prehistoric Pottery; People, Pattern and Purpose, British Archaeological Report International Series 1156, 111-125. Oxford: Archaeopress
Peacock, D. P. S. 1968, A petrological study of certain Iron Age pottery from western England, Proceedings of the Prehistoric Society 34, 414-27
Shepard, A. O. 1954, Ceramics for the Archaeologist. Washington, D.C.: Carnegie Institution of Washington
Terrell, M. M. 2005, The Jewish Community of Early Colonial Nevis. Gainsville: University Press of Florida
Williams, D. F. and Vince, A. G. 1997, The characterisation and interpretation of early to middle Saxon granitic-tempered pottery in England, Medieval Archaeology 61, 214-20
Wilson, S. M. 1989, The Prehistoric Settlement Pattern of Nevis, West Indies, Journal of Field Archaeology 16(4): 427-450
Wilson, S. M. 2006, The Prehistory of Nevis, a Small Island in the Lesser Antilles. New Haven, Connecticut: Yale University Press