18

Pottery and Function

The problem of function is perhaps one of the more difficult faced by those study-
ing archaeological ceramics. The topic can be approached from three points of
view; first at the level of the function of the individual vessel, second, the func-
tional information that can be recovered from archaeological assemblages and third,
the overall orientation of a particular industry – the sector of ceramics usage at
which the principal products are aimed. To tackle all these aspects adequately it is
necessary to draw together information on form, nomenclature, fabric, technology,
trade, distribution and site-formation processes as well as historical, ethnographic
and literary references. It is perhaps not surprising that so much remains to be
done, for some of the necessary tools, such as the appropriate statistical techniques
for comparing between assemblages or the analytical techniques for identifying
organic residues, have only recently become widely available, although others, such
as experimental archaeology (Chapter 11) have been used for some time. It is also
not entirely clear what results may be expected from studies of vessel or assemblage
function or how such information is to be integrated into site reports or regional
surveys.

individual vessel function

Artefacts made of fired clay are ubiquitous, and their functions are diverse. Ceramic
bricks, tiles, pipes and other forms of building materials are very common, and tubes,
funnels and fittings for other industrial processes take advantage of the refractory
properties of fired clay. Moulded and fired-clay figurines have a long history, and
the use of clay as a medium for figurative art continues to this day. Pottery baths,
sarcophagi, portable ovens and similar exotica have been produced at some periods,
but perhaps the most important function of ceramics, both now and in the past, has

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 individual vessel function 247

been its use as containers, particularly for the storage, preparation, movement and
serving of food.

Functional Categories

The functions of pottery containers can be divided into three broad categories: stor-
age, processing (which includes various cooking methods) and transfer (including
serving and eating) (Rice 1987, 208–9), to which can be added funerary vessels, either
as containers for the remains of the deceased or as ancillary vessels, perhaps contain-
ing food or other accompaniments. Without too much difficulty we can imagine that
a vessel employed as a long-distance transport container for a liquid will tend to have
characteristics (durability, ease of handling and stacking, weight/volume ratio, low
porosity, sealable top and so on) that are rather different to those required by a vessel
whose primary function is the frying of eggs (thermal shock resistance, accessibility,
smooth or non-stick surface, perhaps even a nice flavour – Arnold 1985, 138–9).
An extension of this approach is to survey ethnographic and historical records for
correlations between aspects of form, technology or other characteristics and vessel
function (Hendrickson and McDonald 1983; Smith 1985). Rice has summarised the
‘predicted archaeological correlates’ for five broad functional categories: storage,
cooking (food preparation with heat), food preparation without heat, serving and
transport (Rice 1987, table 7.2). Such summaries may be a useful means of organising
the available information, but many of the decisions taken during the manufacturing
process require compromises between competing requirements, so in a particular
instance a predicted correlate may be masked by some other factor. Funerary vessels
may have been made especially for that purpose, but it is equally possible that vessels
produced for everyday use could have been used (Biddulph 2005).There will also be
vessels that have to fit into more than one category. Cooking vessels produced for an
export trade may acquire some of the characteristics of transport containers, such
as stackability and uniformity of size.
Size is sometimes related to function, but while there may be obvious contrasts,
e.g. between storage and serving vessels, there are many reasons for variation in size
(Mills 1999), and a straightforward correlation would be simplistic. More nuanced
studies of size can be revealing: Orton (1982b) distinguishes between pouring jugs (of
2–3 pints capacity) and drinking jugs (of about 1 pint capacity) in medieval pottery
from Cheam, Surrey, and elsewhere the same form is often seen across a spectrum
of sizes. Some potters may well have been capable of reproducing a form in a set of
distinct and well-defined sizes (Zapassky et al. 2006).

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 248 pottery and function

Sometimes lateral thinking can suggest radically different functions for apparently
well-known ceramic artefact types (e.g. van As et al. 2005).

Written Sources and Pictorial Representations

When we turn to the problems posed by particular vessel types there are a number
of potential sources of information on function. At the head of the list may be
placed those vessels which proclaim their function explicitly with inscriptions. Fine
beakers and jugs produced in the Mosel valley (western Germany) during the second
and third centuries AD occasionally included painted or barbotined inscriptions
amongst their decoration. These texts are dominated by such phrases as NOLITE
SITIRE (‘Thirst not’) and DA MIHI VINUM (‘give me wine’), and the association
between these vessels and the consumption of wine would seem to be reasonably
clear (Bös 1958). By extension, vessels of the same form from these factories, but
without the inscriptions, should also be drinking vessels (Fig. 18.1).
In literate societies there may be references to various aspects of ceramics and
particularly indications about the function of particular vessel types. The simplest
references may be merely names of pottery types, perhaps culled from lists or invent-
ories or scratched on the pots themselves, but with no other immediate indication
as to function. The post-firing incision of names on vessels is known, for example,
from the Roman period. They are commonly thought to refer to the owner of the
vessel, but Biddulph (2006) suggests that in the case of funerary vessels they may be
the names of mourners rather than of the deceased. At the other end of the spectrum
there may be complete printed catalogues of the products of particular industries
and descriptions of their use and manufacturing.
A further source of valuable information is the representation of pottery in figur-
ative art (Jacobs and Peremans 1976) which has the dual advantages that the vessels
are shown in use and can often be precisely dated (see Chapter 1; Fig. 1.4). By cross-
referencing all these sources it is possible to build up a reasonable picture of the
range of pottery types in circulation at some periods – broadly equivalent to the folk
taxonomies compiled during ethnographic observations (such as for Roman vessel
types, Hilgers 1969; White 1975). The information on pottery use that can be derived
from the examination of such pictorial sources extends to such items as the wicker
handles added to jars to turn them into water-carrying vessels (e.g. Faure-Boucharlat
1990, 92), or the basketwork used to protect vessels during transport (Laubenheimer
1990, 82, 85, 101). These of course would rarely survive in archaeological contexts. In
areas where traditional pottery is still in common usage it will often be the case that

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 individual vessel function 249

figure 18.1. Roman motto beaker of black colour-coated ware with a painted inscription
suggesting that the vessel was intended for drinking wine (by permission of the Museum of
London).

the best clues about the function of the pottery types recovered from archaeological
levels, as well as other aspects of the vessels, are to be had by examining their modern
counterparts.

Physical Properties

Just as there is a relationship between shape and function, so the physical character-
istics of the fired clay will be relevant to the use to which the finished product will
be put. Investigations have concentrated on three principal areas of interest: thermal
properties, particularly thermal stress resistance and heating efficiency, mechanical
strength and porosity. When ceramic materials are heated, either during firing or

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 250 pottery and function

during use, the constituents of the fabric will expand at differing rates – they have
different coefficients of thermal expansion. Combined with temperature gradients
through the vessel, stresses are set up which may lead eventually to cracking or
spalling. Although this much is certain there is rather more of a debate about the
significance of thermal stress as a factor in the choice of tempering or the shaping of
vessels intended for cooking (Tite, Kilikoglou and Vekinis 2001). Rye suggests that
the problems of thermal stress can be reduced by manipulating three factors: the
shape of the vessel, the porosity of the fabric and the mineral inclusions of the clay
(Rye 1976). Thermal stress should be minimised by the manufacture of round-based
globular pots with an even but thin wall rather than flat based or angular vessels,
where stresses will tend to concentrate along the angles. Fabrics with large pores may
inhibit the formation of large cracks as a developing small crack will be intercepted by
the pore and arrested. Some minerals, quartz in particular, have a thermal expansion
coefficient that is markedly higher than that of a typical clay, whereas others such as
feldspar and calcite expand at roughly the same rate (Rye 1976, fig. 3). Those in the
latter category should cause less stress to build up and thus might be preferred.
However, these factors are clearly not universally applied to the problem of redu-
cing thermal stress. Plog’s review of ethnographic data from the American southwest
highlights a number of examples of apparently contrary behaviour (Plog 1980), and
Woods points to the wide range of flat-based, quartz-tempered cooking pots in use
in Western Europe during the Roman and medieval periods (Woods 1986). In the
ethnographic literature it is often recorded that cooking pots have relatively thick
walls (Hendrickson and McDonald 1983, 632–4). It is clear that although some pot-
ters may be aware of, and take account of, the thermal-shock problem, what might
be considered to be the ‘appropriate’ solutions are not universally applied.
Tite et al. (2001) carried out experiments with clay test bars containing different
types of temper and fired to a range of temperatures, and found that producing
pottery with high strength requires high firing temperatures and low inclusion con-
centrations. Conversely, producing pottery with high toughness and thermal-shock
resistance requires low firing temperatures and high inclusion concentrations, with
platy or fibrous inclusions being most effective. They concluded that there is no
convincing evidence that strength and toughness requirements were a significant
factor in determining the technological choices (clay type, inclusion type and con-
centration and firing temperature) in the production of pottery used as containers
for transport and storage. In contrast, the routine use of high concentrations of
inclusions and low firing temperatures in the production of cooking pots suggests
that the requirement for high thermal-shock resistance was a factor that at least
influenced technological choice in this case.

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 individual vessel function 251

Porosity, in addition to being a potential factor in the reduction of thermal stress,

has a bearing on the problem of heating efficiency. A porous fabric may allow
liquids to seep through from one surface to another. For some vessel types this is
advantageous, indeed a basic requirement. In water jars employed in hot climates
the permeability of the fabric allows water to evaporate and hence cool the contents,
a process which is further encouraged by light-coloured surfaces. What may be an
advantage in these circumstances will be less beneficial in others – the long-term
storage or transportation of liquids for example. In a vessel employed for cooking
any seepage of liquid through the wall of the pot will reduce heating efficiency,
prolonging the heating process and wasting valuable fuel. Without reducing the
porosity of the fabric it is possible to reduce permeability by treating one or both
surfaces. Schiffer describes a series of experiments that demonstrate the relationships
between heating efficiency and permeability and the advantageous effects of different
surface treatments (Schiffer 1990). It is clear that the application of resins, slips
and even burnishing the surface are sufficient to raise heating efficiency, while still
retaining the potential benefits of a porous fabric. Even in areas and at periods where
glazes are in common usage and capable of providing a perfectly impermeable
surface, they are not frequently applied to cooking vessels. There are references in
the ethnographic record to the application of clay slips or other sealants to the
surfaces of pottery to reduce permeability, but these are not always to cooking vessels
but include vessels for storage as well. The benefits of corrugating the surface of a
cooking pot have been discussed: Young and Stone (1990) report that corrugated
vessels do not heat up any faster than plain ones, while Pierce (2005) suggests that
corrugation on different parts of the vessel may confer different benefits.
The mechanical strength of vessels can be considered under a number of headings:
there is resistance to sudden impacts, dropping the vessel for example, and there is
resistance to more gradual processes such as abrasion. If vessels are to be stored or
used in exposed environments then resistance to frost shattering may be an important
consideration. Strength is not only important in the finished product but also during
manufacture. It may be advantageous to produce thin-walled vessels for a particular
purpose, perhaps to improve the volume/weight ratio for transportation, but this
may require special procedures. It may be necessary to make the vessel in stages, or
in parts which are assembled only when they are partly dried, or the vessel may be
trimmed or beaten to produce a thinner wall.
It is usual to record the ‘hardness’ of a fabric as one of the characteristics con-
sidered during the standard process of pottery description (see Chapter 13, p. 158),
and a simple scheme such as reference to the Mohs scale is generally employed.
More sophisticated mechanical techniques for the assessment of impact strength

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 252 pottery and function

have been devised (e.g. Marby et al. 1988), and there have been experiments with
test briquettes to determine the relationship between the quantity and type of tem-
per, firing temperature and impact resistance. Bronitsky and Hamer (1986) sug-
gest that the incorporation of finely ground tempers made the finished product
significantly more durable. Schiffer and Skibo (1989, 606) record that tempered
briquettes were less resistant than those that were untempered, and that the dif-
ference in impact strength increased with firing temperature. Briquettes tempered
with organic materials were less durable than those tempered with sand. Abrasion
resistance has also been investigated experimentally (Skibo and Schiffer 1987), and
in this case it seems that a high percentage of coarse temper offers the greatest res-
istance to abrasion, particularly so when wet. Tite et al. (2001) review strength,
toughness and thermal shock resistance of ancient ceramics and conclude that
our current understanding of the factors affecting these properties is still far from
complete.
It is apparent that the various physical characteristics of fired clays outlined above
are not only interrelated, but the steps which might result in the optimum conditions
for one factor will in some cases have adverse effects on others. Investigations of the
precise effect of, for instance, particular types of tempering or surface treatment
increase our understanding of the behaviour of traditional ceramics, and as such are
valuable. For example, finite element analysis (FEM) has been used to study both the
mechanical behaviour of pottery (Kikiloglou and Vekinis 2002) and the relationship
between strength, shape and the method of stacking (Hein et al. 2008). They may
help to explain some the characteristics of a vessel of known function – a cooking
pot or water jug – but they do not by themselves provide immediate and direct
indication of the function of a ware or vessel.

Traces of Use and Wear

Many of the operations performed on ceramics will leave physical traces which can
give valuable clues about these activities. An individual observation may, by itself,
be of limited interest, but regular associations with an identifiable activity will lead
to functional interpretations which are of wider value.
It may be possible to identify the general function of particular forms, as cooking
pots, storage vessels and so on. In other cases a more specific association between
form, source and function may be uncovered, suggesting that a particular producer
was specialising in the manufacture of vessels which themselves had a specific and
specialised function. It is important to combine, as far as possible, aspects of shape,

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 individual vessel function 253

use-wear and residue analysis (see below) in attempting to elucidate the possible
functions of vessels (Vieugué et al. 2008).
Many pots retain traces of their role as cooking vessels. When a vessel is used
over an open fire traces of soot will often be deposited on the external surface, or
the colour of the surface will alter. In some cases fine cracks may develop. Hally
(1983) describes the variations in sooting patterns which develop under differing
conditions, in particular a vessel suspended or supported over a flame will tend to
develop sooting over the entire lower surface, whereas vessels set in a fire or amongst
hot ashes or embers will tend to develop sooting in a zone around the lower body
of the pot, but not directly on the base – while Skibo and Blinman (1999) show
the sooting patterns that result from different types of cooking. Unfortunately these
distinctions are of less value than they should be because it is probable that most
sooting is removed during washing and processing during post-excavation, leaving
only the barest traces surviving. Medieval drip-pans and pipkins are often sooted and
burnt on one side only, that opposite the handle, suggesting that they were placed
on the side of the fire rather than within or above it; contemporary manuscript
illustrations seem to confirm this (Moorhouse 1978, 7).
There is a potential source of confusion when one examines discolourations on
the surface of vessels due to heating or burning between those caused by cooking fires
and those resulting from the original firing procedure. Localised colour variations
known as ‘fire-clouds’ (Shepard 1956, 92) usually result from contact between the
vessel and fuel or hot gases during the firing cycle. They may occur in any type of
kiln, but are more common in open or pit firings where the fuel is arranged around
the vessels.
In addition to cooking or heating, many other food-preparation processes such
as scraping, cutting or stirring will leave traces on sherds which may be identified
(for example, see Biddulph 2008). Some Roman mortaria, bowls which often had
coarse grits embedded on the internal surface, were certainly used for grinding or
pounding, and most probably for the preparation of wheat and other cereals (Hartley
and Tomber 2006). The grits and the part of the fabric of the pot are often worn
away – in a few extreme cases this has resulted in a hole right through the vessel
(Fig. 18.2). The character of the product resulting from this mixture of cereal, grit and
pottery can be imagined. In general, abrasion observed on pottery can be ascribed
to various functions. For example, Reid and Young (2000) report evidence for grain
preparation, and Arthur (2002; 2003) discusses the effects of fermenting beer on the
interiors of the vessels used.
Repeated stress on a vessel in a particular area may eventually result in other types
of breakage, but may also be compensated for by thickening or strengthening. Weak

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 254 pottery and function

points may exist wherever separate elements were luted together, and in particular at
the attachment of handles to vessels. Extra strips of clay around part of a vessel may
have been added in reaction to a perceived fault, and it may be possible to observe
changes in the construction of vessels through time to strengthen such weak points.
The ways in which handles are made and attached may provide evidence for the
intended use of a vessel (Franken 1993/94). In some cases, repairs may be visible;
Nieuwenhuyse and Dooijes (2008) give an example where the nature of the repair
strongly suggests that the appearance of the vessel was not considered important.
Some classes of pot were made to be used once and then broken; indeed the
breakage was an important part of their function. The Roman writer Pliny describes
a type of bread or cake known as ‘Picenum bread’ which was baked in pots in an
oven. These were broken to get at the contents, which was then soaked in milk and
eaten (Pliny, Natural History, book XVIII, ch. XXVII; André 1961, 72).

Organic Contents, Deposits and Residues

Ceramics are used at most stages of food processing, and in many cases these
operations leave organic traces which may be identified. However the value of this
information is very varied. The identification of the contents of, for instance, a
transport container has a potential value that is quite different from the identification
of the organic contents of an individual cooking pot. In the former case there
will probably be implications far beyond the vessel in question; the immediate
implications of the latter lie, at least initially, in the context and site.
Occasionally a vessel will be recovered with the remains of contents which seem
to provide unequivocal evidence of its original function. Amphorae, the ubiquitous
storage and transport containers of the Graeco-Roman world, are commonly found
on wrecks or other underwater sites, and a very small number of these vessels are
recovered complete with their contents. There are amphorae containing olives stones
from a number of wrecks in the Mediterranean and, more surprisingly, one from
the Thames estuary (Sealey and Tyers 1989, 57). Amphorae containing fish bones
(from fish-based sauces) have been recorded (Sealey 1985, 83) – there are even a few
vessels which still contain wine (Formenti et al. 1978). In addition to their ceramic
connections, such large hoards of Roman foodstuffs are important archaeological
resources in their own right, with the potential to provide important insights into
agricultural or food-processing practices.
On a rather smaller scale are occasional vessels which contain remains of their
last contents recovered from destruction deposits and similar ‘primary’ contexts.

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 individual vessel function 255

figure 18.2. A third-century AD bowl in ‘Black-burnished ware’ from Ewell, Surrey,

showing wear on the inside of base (photo: UCL Institute of Archaeology).

Plates in Pompeian-red ware from the AD 79 Vesuvian destruction of Pompeii are

recorded as containing the remains of flat loaves – ‘somewhat overcooked’ (Loeschcke
in Albrecht 1942, 38; Greene 1979, 130).
It has long been noted that some pots recovered from archaeological contexts
contain traces of deposits or encrustations on their surfaces. Some of these derive
from the soils in which the vessels were discarded or buried, but others relate directly
to the function the vessel fulfilled during its life. The deposits may be burnt or
charred, either on the interior or exterior of the pot and probably resulting from
cooking, or they may be similar to the lime-scale created in modern vessels used
for boiling water for prolonged periods, such as kettles. Such deposits are derived
from natural products and can contain some biomolecular components from those
products even after the passage of much time.
In some cases, the surface deposits are deliberate coatings or sealings. The use
of pine resin, pitch and bitumen are known (see Reber and Hart 2008; Stern et al.
2008). There can be useful spin-offs from such evidence, for example in sourcing
pottery from the type of bitumen used (Stern et al. 2008), and in the observation that
wine can penetrate a pitch coating, but olive oil cannot (Romanus et al. 2009). On
the other hand, Skibo points out that resin coating, although resistant to abrasion,
would probably biodegrade in archaeological contexts (Skibo et al. 1997).

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 256 pottery and function

However, in addition to these visible traces of use, it has become clear that organic
compounds associated with human activity can be absorbed and retained by porous
ceramic materials but leave no visible trace on the vessel. Thus we cannot confine
the analysis of organic residues in pottery to that (possibly) small proportion of the
assemblage with visible deposits – rather we may have to consider the potential of
a very large group of material. The necessary procedures for the analysis of such
residues have become widely available only in recent years and include separation
(gas chromatographic) and identification (mass spectrometric) techniques (GC-MS)
to measure the concentrations of the saturated fatty acids. The identity and propor-
tions of the particular fatty acids present in lipids, for example, are significantly
different between animal and plant products, and also between species (Evershed
1999). In addition, much additional information on the environment of the natural
products can be revealed by stable isotope ratio analysis (bulk δ13 C, δ15 N and δ13 C
of individual fatty acids), with a gas chromatograph – combustion – isotope ratio
mass spectrometer (GC-C-IRMS). The concept of the archaeological biomarker has
been developed (op cit, 19), namely that the structure and even isotopic composi-
tion(s) of a given biomolecule or suite of biomolecules (the ‘chemical fingerprint’)
can be related to the compositions of natural products exploited by humans in the
past. Progress with the analysis of organic residues in archaeological material has
been reviewed most recently by Oudemans (2007), Evershed (1999; 2008) and Spiteri
et al. (2011), while Barnard and Eerkens (2007) have edited a conference volume with
an international range of contributors reviewing various techniques and practical
applications. Developments of instrumental methods in the future can be expec-
ted, as well as the greater use for archaeological applications of high performance
liquid chromatography (HPLC) coupled with soft ionization mass spectrometry and
MALDI techniques for high molecular weight substances.
Since its introduction, the application of stable isotope ratio analysis has been very
widely used to supplement the data on the proportions of saturated fatty acids in
organic residues. Spangenberg et al. (2006) however have shown from comparison
of ancient potsherd extracts and modern adipose and dairy fats from central Europe
with similar modern material from northern Europe that there is a small diversity
of carbon isotope ratios associated with the synthesis of C16:0 and C18:0 saturated
fatty acids in ruminant and non-ruminant animals. They stress the necessity of
comparing ancient residues to those of modern reference fats obtained from regions
with climatic conditions and vegetation that are similar to the catchment areas of
archaeological sites in antiquity. At the Late Neolithic (3384–70 BC) site of Arbon
Bleiche 3, Switzerland, the small variations of the δ13 C and δ15 N values in residues
on pottery were within the range expected for degraded animal and plant tissues,

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 individual vessel function 257

showing that the diet at the site is consistent with archaeological evidence of animals,
whose subsistence was mainly based on C-3 plants.
Two main points are worth consideration when planning a program of organic
residue analyses, or interpreting their results:

The results of the analysis in its ‘raw’ form are expressed in terms of various fatty acids
and glycerides – the building blocks of the original compounds. To translate these into
the original items of interest is not without difficulty as some compounds alter over
time, whereas others disappear. To identify the ‘original’ material, analysis of modern
samples for comparison is required, including where published descriptions do not
exist (Hill 1984).

A vessel used for cooking a range of substances, at the same time or separately, may
absorb organic residues from all of them. In addition, absorption from the post-
depositional environment is also possible (Heron et al. 1992). In a feature such as
a rubbish pit or midden, contact with organic compounds would seem inevitable.
Analysis of the surrounding soil may help to identify and eliminate sources of possible
contamination, but this would seem to rule out the use of material from old collections,
and even most material from recent excavations.

It is evident that the reconstruction of any ‘original contents’ from the extant or
altered parts of a mixture of compounds, complicated by possible contamination
and reuse of vessels demands great care, and any interpretation of these results
requires a full appreciation of the potential difficulties. It is also apparent that many
of the compounds retrieved from some classes of vessel (such as amphorae) relate
more to the tars, resins and other substances applied to seal the inner surface of
the pots rather than the commodities they carried (Heron and Pollard 1987; Pollard
and Heron 1996, 239–70). Similarly it is common practice to ‘prove’ earthenwares by
rubbing the inner surface with oil and baking them in a hot oven.
The survival of organic remains associated with pottery shows indications of
being region-specific: for example, lipids appear to be less preserved in the Middle
East than in Europe, possibly because of the greater antiquity of pottery, range of
variation in seasonal temperatures, calcareous soils and the particular extraction and
analysis protocols used to identify residues. However, Gregg et al. (2009) found direct
biochemical evidence of subsistence practices and pottery use at a late Neolithic site
at al-Basatı̂n, northern Jordan, including the use of vessels for cooking of meats or
extraction of marrow from both ruminant and non-ruminant animals.
An interesting example of the value of contents and residues is reported by
Bonnamour and Marinval (1985). They identify a group of early Roman jars from a

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 258 pottery and function

number of sites in the Saône valley (central France) with burnt deposits of millet. It
is suggested that the deposits result from the preparation of a gruel or beer. Many
of vessels are of a similar form (a jar with a rilled neck), and it may be that this was
‘chosen’ to match the function. It would be most interesting to know if the pots are
not merely of the same form, but from the same source. It could be that this function
required not only ‘a pot of that form’ but one from a particular source.
This leads us on to the identification of a sub-set of cooking wares which are not
‘general purpose’ but related to specific functions, perhaps even to the extent that
they are associated with the preparation of particular recipes. A modern example
will illustrate the point. The famous bean dish of south-west France, the cassoulet,
apparently takes its name from the original clay pot produced by the potters at Issel –
hence Cassol d’Issel – which was considered necessary for the dish (David 1959,
93). We can imagine that with growing popularity others potters would attempt
to make inroads into this market by producing their own versions of the cas-
sol, imitating features of the form, finish or refractory qualities of the original.
The later history of such a ‘type’ might be marked by the decline and loss of the
link with the original recipe and its integration into the range of general pur-
pose cooking vessels. In the archaeological record such a mechanism might appear
as the initial wide distribution of vessels from a single source, followed by ‘imit-
ations’ produced in secondary centres. This is only one instance of the general
problem of attempting to label a vessel type in a single functional category – a
glance at the contents of any kitchen, particularly during moments of stress, will
see all manner of vessels and containers which are not being used for their ‘proper’
function.
The final point to make about the identification of organic residues is the import-
ance of communication between the various specialists involved in writing up a
site. Those responsible for reporting on, for example, the fish bones from a site will
need to know that some of the amphorae from the same contexts originally carried
fish-based sauces (Partridge 1981, 243).

Reuse of Pottery

Vessels can have more than one use in their lifetime. For example, amphorae could
have a ‘second life’ as dry storage vessels, and even a ‘third life’ as urinals, as evidenced
by deposits of calcium and phosphates (van de Werff 2003). Even broken sherds can
have their uses, for example as tools in the manufacture of new vessels (López Varela
et al. 2002). The reuse of sherds as (amongst other things) spindle whorls, when

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 function, production and distribution 259

more suitable materials appear to be available, is thoughtfully discussed by Scham

(1998/99).
Some features, such as pierced holes, may be evidence of reuse, but they can also be
part of the original design, especially if they were made pre-firing. A range of possible
functions that could require such holes are discussed by Fulford and Timby (2001).
Beck et al. (2002) point out the importance, but also the difficulty, of distinguishing
between use-wear and wear caused after the life of a vessel.

function, production and distribution

Clearly related to the preceding is the overall emphasis of a production assemblage.

A producer of, for instance, transport containers for agricultural products will be
subject to economic influences and constraints which will have little, if any, effect
on a producer of lamps or pottery statuettes.
In addition to a consideration of the individual forms in the production
assemblage, the distribution of the products may be used to distinguish between
different functional categories. In general the ‘fall-off curve’ for high- and low-
value products will differ sharply – high-value products having a broader but more
even and lower level distribution, contrasting with the high concentrations but
more restricted distribution of low-value items. Certainly during the Roman period
the majority of the long-distance movement of pottery related to either its use
as bulk containers for agricultural products or as fine table wares, but there is
increasing evidence that long-distance movements in apparently utilitarian cooking
wares was possible if they were deemed to have particularly desirable characterist-
ics. The example of Black-burnished ware category 1 (BB1) in Britain is instructive
here (Peacock 1982, 85–7). These coarse hand-made cooking wares (jars and bowls)
were produced in south-west England throughout the Roman period, and indeed
the origin of the industry precedes the conquest. Before circa AD 120 they were largely
confined to their homeland, but after that date they are distributed widely, being
particularly common on military sites in the north of the country. Within a short
time many of the pre-existing industries in the south and east of the province begin
producing their own versions of the characteristic Black-burnished forms, but often
in wheel-thrown wares, and these eventually become the typical cooking pot forms
of the later Roman period. We have here an indication of the dramatic possibilities
when local forms are plucked from their source, promoted on a wider market and
then assimilated into the repertoire of competing industries. The question which
arises is whether the similarity of form can also be taken to indicate a similarity in

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 260 pottery and function

function. In the case of BB1 and the wheel-thrown versions the answer may be yes,
but it will not always be the case.

symbolic meaning

In addition to their functions as cooking pots, table-ware and so on, pots (indeed any
artefacts) may serve as transmitters of information about their producer, owner or
user (for example, see Harry and Frink 2009). Thus some classes of vessel may suggest
high status, while others indicate religious, social or tribal affiliations. There is a view
that artefacts are part of a ‘material culture language’, a means of communicating
information between individuals and groups, and more than this, a medium through
which social conflicts can be expressed and even resolved (Hodder 1986, 122–4). Some
of the flavour of this view of the symbolic functions of a pot has been summarised
in these words: a pot ‘may mean that I, as the ancient owner of this vessel, belong to
this group, and believe these things, that I have this level of wealth, and this much
status. I am also of a specific sex, and perform these labors defined by my sex, and
this vessel correlates with this sex and these labors’ (Strange 1989, 26).
Food preparation and consumption, and the myths and rituals that surround it,
are one of the central aspects of culture (Goody 1982). Eating and drinking beha-
viours are, on the one hand, subject to deeply held beliefs about what is ‘clean’
and ‘unclean’ (or good: bad, inside: outside and so forth), but on the other, an
area of culture open to outside influence in the form of new materials and tech-
niques and a means of expressing or promoting status and difference. Pottery,
the principal accessory to food preparation, storage and serving, will be inevit-
ably touched by many of the same taboos and become steeped in ritual and sym-
bolic meaning. Pottery has a demonstrable role in many cultures as a means of
distinguishing between groups, of dividing ‘them’ from ‘us’. The signals may be par-
ticular design elements, typological features, colours or manufacturing techniques.
In some instances, it is suggested, pottery becomes a medium for inter- and intra-
group power relations, a way of communicating information covertly that cannot be
expressed openly (Braithwaite 1982). The ‘obvious’ interpretation of the social func-
tion of a class of vessel should be treated with suspicion; for example, Loughton
(2009) challenges the traditional association of wine-drinking with a ‘warrior
class’ in Iron Age England, pointing out wider associations with various industrial
activities.
But how are we to apply these ideas to pottery in the archaeological record? It is
difficult enough through ethnographic observation, when vessels can be observed in

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press

 symbolic meaning 261

use and the individuals involved in the drama are at least on hand to be questioned
about their actions – or at least what they understand by their actions.

The Answer Lies in the Soil

The solution, if there is one, lies in the most powerful resource we have – the
structure of the archaeological record itself. We are concerned with the ‘multi-
dimensional’ location of pots in their complete context, their relations to other
pots and other classes of artefacts, and with archaeological features and layers. The
methodological tools needed if one is to pursue this approach are now available
through the quantification of entire assemblages and the stratigraphic relationships
between them.

https://doi.org/10.1017/CBO9780511920066.022 Published online by Cambridge University Press