. 3
( 7)


all the interest, excitement and intellectual challenges of an ethnographic
encounter with another culture. John Barrett has discussed the fascina-
tion embodied in the process of discovery, a fascination heightened by
a ˜feeling of otherness or difference™ (1995, 6). For both these authors
excavation is felt to be primary to the process of interpretation, as it is
during excavation that our ¬rst interpretations about the past are formu-
lated. However, as Barrett notes (1995, 6), this initial engagement with
the past, through archaeological excavation, begins to stale as the remains
of the past are ¬ltered through the processes of post-excavation and pub-
lication. The excitement and interest engendered in the primary task of
excavation is rarely communicated in excavation reports. As we have seen
60 Archaeological theory and scienti¬c practice

this is due, in part, to the processes of fragmentation, hierarchisation and
transformation. As the remains of the past are ¬ltered through the various
stages of archaeological investigation, they are gradually dislocated from
the site itself and the knowledge concerning them is ordered according
to a hierarchical interpretative convention. Interpretation at this stage is
quite simply divested of its interest “ the past has become entirely abstract
and objecti¬ed. To paraphrase Barrett (1995, 6), it is only when material
is placed in a frame of reference other than our own that it gains its his-
torical signi¬cance. It is precisely this frame of reference that is lacking
when we shear material from the site and subject it to a series of objective
I want to begin to reconsider these problems through recourse to an
analysis of the parallel problems encountered by anthropologists. As I
have noted, one of the principal problems which anthropology has con-
fronted in the late twentieth century is precisely how, as observers of other
cultures, we are able to provide accounts of our experiences and encoun-
ters with other cultures. That is, in essence, what we attempt to do as
archaeologists. In the opening chapters of this book I noted that, follow-
ing Tambiah (1990), one of the views available to anthropologists was
to employ what he described as a ˜double subjectivity™. This viewpoint
involves attempting to understand the intentions of people inhabiting an-
other cultural setting, while simultaneously taking up a neutral stance
removed from this cultural setting. This ability to occupy a position both
within and without a given culture enables the anthropologist to translate
or write the experience of that culture within the terms of our own cul-
ture. As both Asad (1986) and Tambiah (1990) indicate, it is through this
process of translation, through an attempt at understanding the idioms
of another system of thought, that we come to understand ourselves. The
interesting point about this form of enquiry, as it applies to anthropology,
is that it presupposes that at each stage of interpretation, from the pri-
mary encounter to the publication of a monograph, the anthropologist
moves back and forth between native cultural understandings and the
understandings of the anthropologist™s own conceptual system. It is this
movement back and forth which is the essence of translation.
I noted above that one of the effects of fragmentation and hierarchisa-
tion through excavation was the shift in emphasis in the status of what
we might describe as the ˜archaeological laboratory™. The excavation site
as the object of analysis disappears, and we are left with the tangible
physical remains of the site in the form of artefacts and environmen-
tal samples. At this point the remains of excavation are transformed; as
they enter the laboratory in which archaeological scientists work, they
become the objects of scienti¬c analysis. They are then resituated within
Archaeology observed 61

the wider discourse of archaeological science, and their nature is deter-
mined through recourse to extensive bodies of independent scienti¬c
data. If we are to adopt the mode of enquiry suggested above, we need
to be aware that artefacts and environmental samples are physically situ-
ated within the realm of objective scienti¬c analysis, while conceptually
they are simultaneously situated within the contextual parameters of the
archaeological site. Moreover, the artefacts and environmental samples
were once contextually situated in, and intervened in, the lives of people.
Hodder correctly notes that both object and context are mutually con-
stituting “ the meaning of the context is provided by the objects found
within that context and vice versa (1999, 84“6). It is critical, then, that
the physicality of the object to be studied is taken into consideration as
well as its physical, historical and cultural context.
How does this viewpoint affect the way in which we practice our ana-
lyses in the laboratory? While the transformation of artefacts and envi-
ronmental samples occurs from the site to the laboratory, archaeological
scientists need to consider re-contextualising artefacts and environmen-
tal samples both during and after analysis in the laboratory. Both the
object of analysis and the interpreter (the archaeological scientist) need
to be at once situated “ like the anthropologist “ within Western concep-
tual systems and non-Western conceptual systems. By doing this during
the process of scienti¬c analysis, material will be placed in a very differ-
ent frame of reference, and will thereby regain its historical and cultural
One way of accomplishing this change is to alter our view of archae-
ological practice. Rather than simply viewing excavation to be the site
of primary interpretative encounters with the physical remains of past
cultures and post-excavation analysis in the laboratory to be the domain
of objective analysis, we need to cultivate each element of archaeolog-
ical practice as part of a single interpretative process. At each stage of
the interpretative process we need to consider how it is that we formu-
late interpretations as we move back and forth between our observations
of material and our theoretical assumptions about that material (Barrett
1990; Wylie 1993).
However the crucial moment in the process of reworking our relation-
ship with past material culture arrives with publication. We need to be
especially aware of the kinds of narratives that we employ in writing about
archaeology (Pluciennik 1999). I have argued that it is through excavation
publications that we not only reproduce particular kinds of knowledge but
also particular modes of archaeological discourse. The narrative forms
that we deploy in excavation reports are therefore crucial. As Pluciennik
notes (1999, 655), narratives are constituted of characters, events and
62 Archaeological theory and scienti¬c practice

plots. Importantly each element of a narrative is linked together by dif-
ferent levels of scale and context that ultimately lead to different kinds of
explanation. I will take this point up in the next chapter; here it is suf¬-
cient to point out that in excavation reports we need to experiment with
the inter-relationships between these elements of narrative. The process
of excavation lends itself to a variety of plot devices. We need not simply
write reports as the presentation of data recovered by a linear scienti¬c
methodology. There are other ways of telling that include the usage of
narratives about the excavation process (Bender, Hamilton and Tilley
1997). Other modes of writing might employ artefacts and environmen-
tal samples as components of the story, used to describe signi¬cant kinds
of on-site events, such as processes of building or rebuilding, or other fun-
damental alterations in the site. Alternatively artefacts, architecture and
environmental evidence may be used to explore speci¬c kinds of social
practice, such as food production and consumption, the social construc-
tion of the body, hunting and agriculture, defence and warfare, etc.
Earlier in this chapter I employed the metaphor of ˜explosion™ to des-
cribe the effects and after-effects of archaeological excavation. Here I
am advocating that writing excavation reports should involve the oppo-
site of ˜explosive™ linear narratives; they should encompass a process of
re-connection that involves the creation of interpretative networks. This
process of reconnection should be twofold. First, it is essential that we
recreate the connections which de¬ned the relationships between site and
artefact. By doing this we will go some way towards providing a clearer
picture of the site that was excavated. Rather than presenting a picture of
the site with its compartmentalised artefacts and scienti¬c analyses, we
need to create an image of the site with its components in context, either
in the order and sequence in which they were excavated or according to
an order and sequence that makes sense within the overall interpretative
framework of the report. By doing this we will also be creating interpre-
tative connections by bringing together the various analyses undertaken
at both the excavation and post-excavation level in an inclusive form of
interpretation. Rather than prioritising one element of the site, or one
interpretation, over the other, we need to be aware of the dynamic that
exists between them.
4 Materials science and material culture:
practice, scale and narrative

In the account I have provided so far, science has been represented as a
somewhat monolithic entity. However, if we consider scienti¬c practice to
have a critical effect on the kinds of knowledge generated, then it follows
that different sciences engender quite distinct cultural practices and pro-
duce distinct forms of knowledge (Knorr-Certina 1999). This means that
we cannot assume the existence of a uni¬ed laboratory-based practice
known as ˜archaeological science™ that unites archaeobotanists, zooar-
chaeologists, soil micromorphologists, ceramic petrologists, etc. Within
the next two chapters, I will set aside other areas of archaeological sci-
ence in order to focus on the practice of materials science “ the study
of archaeological materials using techniques derived from engineering,
chemistry and physics (Kingery 1996). In taking this step, I do not wish
to present a further divisive view of archaeological practice; rather, I want
to examine how we might re-orientate materials science analysis in terms
of the wider goals of interpretative archaeology. In order to undertake this
task I will focus on the interface between these branches of archaeology
(see Renfrew 1982).
Hand in hand with the creation of the archaeological laboratory we
observe the conceptual transformation of the physical traces of the site
(both artefactual and environmental) as they become the focus of objec-
tive scienti¬c analysis. In order to re-contextualise artefactual and envi-
ronmental samples within a historically and culturally meaningful frame-
work, I advocated a mode of enquiry that takes account of both context
and content, both the physical dimensions of artefacts and environmental
samples and their cultural and historical dimensions.
This methodology can be described in terms of a process of translation
in which alliances are created between observations and interpretations
on a step-by-step basis. In other words, an interpretative analytical process
for archaeological science that requires simultaneous attention both to
the rigour of scienti¬c analysis and to the rigour of cultural and historical
understanding. Such a framework of analysis attends to the traditional
explanatory mode of analysis of science, in which the veracity of results

64 Archaeological theory and scienti¬c practice

are judged by empirical analysis and reproducibility, while at the same
time attending to the interpretative mode of analysis promoted in the
social and human sciences, in which analysis is judged according to the
kind of framework of understanding in which it is situated and depends
much on the clarity and persuasiveness of the discussion (Toulmin 1990).
If we are to examine this mode of enquiry, it is necessary ¬rst to concern
ourselves with the practices of materials scientists. First, we need to focus
on the nature of the problems that materials scientists study, and second
on how they go about solving those problems at the practical level. In
previous chapters I noted that epistemologically there were a number of
alliances made between science and certain kinds of archaeological the-
ory. For example palaeobotanists and zooarchaeologists often ally them-
selves with ecological or Darwinian approaches, and materials scientists
often ally themselves with broad-based studies of exchange networks or
studies of ancient technology. However I also noted that there appeared
to be an interpretative ˜brick wall™ created between the kinds of analysis
found in much recent interpretative archaeology and that traditionally
allowed for in scienti¬c analysis. Certain theoretical approaches appear
to articulate better, or more solidly, with the results of certain kinds
of scienti¬c analysis than others. If the practices of materials scientists
are to articulate differently, we need to understand how the connections
are made in practice between different kinds of scienti¬c and theoretical
knowledge. We need to consider what it is that we might want to know
at the theoretical level and examine how our practices as materials sci-
entists might be modi¬ed in order to create clearer articulations between
one body of knowledge and another.
In this chapter I want to develop this methodology by examining three
main points which arise from this proposal. First, I want to examine the
nature of the relationship between materials science and material culture
as a means of highlighting some of the possibilities open to archaeological
scientists in the integration of analytical approaches from both the human
sciences and the natural sciences. Following on from this, I want to look
at the important issues of scale and resolution. These issues are critical
if we are to fully realise an alliance between material culture studies and
materials science analyses. Finally, using a short case study, I will explore
what such a scienti¬c analytical process might look like.

What is material culture?
The ¬eld of material culture studies has grown considerably over the past
twenty years or so. Due to this we tend to take the term ˜material culture™
for granted (Prown 1996), but I believe we need to scrutinise the term
Practice, scale and narrative 65

in more detail. An objective view of the world would see a logical dis-
tinction between inanimate objects, which are part of the material world,
and animate subjects, that are part of the cultural world. The concept of
˜material culture™ fuses these distinct elements of the world together in
a single term. From an objective viewpoint it appears logically inconsis-
tent. If we are to employ the term, then we may be required to embrace
an element of subjectivity if we are to proceed with our enquiries. We
are required to take on board the view that the material world is appre-
hended through a series of cultural devices, including perception and
language, and that the way in which we engage with this world is heavily
laden with social and cultural signi¬cance. More maturely, from a per-
spective informed by an understanding of the philosophical tenets of phe-
nomenology, we might view the material and cultural aspects of the world
to be mutually constituted and mutually specifying. As Christina Toren
(1999, 5) puts it: our ideas, the cultural aspects of life, are constituted in
material relations with each other, and we communicate with each other
in and through the materiality of the world. Or, to put it another way,
people and things are mutually constructed; things are enmeshed within
human affairs (Latour 1999, 174“215; Strathern 1998). In short, in order
to propagate social relations between each other, people use things.
However, this proposition poses a considerable puzzle. In many ways
we might think of socially constructed things as mediators of social re-
lations. But if things are socially constructed and therefore composed of
social relations, why do we need social relations (in the form of things)
to mediate with another set of social relations? One way of reorienting
this conundrum is to consider the possibility that we might view the con-
struction of things as one technique by which humans make society more
durable (Latour 1991); since things retain, in material form, the ideas
and notions that make societies operable and reproducible. If we take
this idea on board, we then need to consider how it is that things mediate
for humans.
This exploration of both how and why alliances are constructed be-
tween people and things is what makes the ¬eld of material culture studies
so dynamic. Importantly this theoretical perspective places archaeology
at centre stage, since it is one of the few disciplines dedicated to exploring
the relationship between people and things. Furthermore, the develop-
ment of this outlook within archaeology has led to increased discussion
at a theoretical level with those working in other disciplines, such as an-
thropology, history of science and technology, human geography, design
studies, art history and museum studies (see Miller 1987; Gosden 1999,
152“79). Moreover, since the discipline of archaeology encompasses an
interest in the social and physical dimensions of the material world, it
66 Archaeological theory and scienti¬c practice

is critical that we engage with these ideas not only at a theoretical level
but also at a practical level. We are required to consider simultaneously
how it is that artefacts are socially and culturally constructed, while also
taking into account the physical and mechanical construction of arte-
facts. We do not need to study these two aspects of artefacts as separate
and distinct entities. Materials science and material culture studies are
therefore engaged in the same project of enquiry, as we will see in this

Sampling, scale and abstraction
How are we to relate these points to the concerns and expertise of the
materials scientist? First, materials science provides a series of techniques
for characterising the physical and material basis of material culture;
these range from image intensi¬cation techniques, such as low power
light microscopy, to high intensi¬cation techniques, such as scanning
electron microscopy (SEM). A series of techniques are also available
to determine the material composition of artefacts (for overviews see
Kingery 1996; Pollard and Heron 1996; Wachtman 1993), ranging from
thin-section petrology to more complex analytical procedures such as
neutron activation analysis (NAA), inductively coupled plasma spec-
troscopy (ICPS) or x-ray ¬‚uorescence spectroscopy (XRF). Each of
these techniques is aimed at providing precise data on the major, mi-
nor and trace elements from which an object is constituted. Second,
through techniques based on the principles of engineering, the phys-
ical properties of material culture may also be determined (Cotterell
and Kamminga 1990). Fundamentally aspects of cultural choice deter-
mine both the material composition and the mechanical properties of
material culture. If we are to understand these aspects of material cul-
ture, then materials science has a critical role to play (Kingery 1996,
If we are to gain an understanding of the manner in which cultural
choices have been made in relation to the material construction of arte-
facts, we must be especially aware of the scale of analysis at which we
operate. In general we can distinguish between techniques designed to
examine the elemental basis of artefacts, the microstructure of artefacts,
and those used to determine the macrostructure. Different scales of anal-
ysis require different instrumental techniques in order to describe the de-
tails of artefact composition (see Kingery 1996; Killick 1996). Our task
remains to move between our interpretation of objects as determined by
material science techniques and what these interpretations might tell us
Practice, scale and narrative 67

Scale Analysis of artefact

Macroscale analysis
Microscale analysis


empirical instrumental
perceptual enhancement
Figure 4.1 The changes of perception allied to changes in analytical

in terms of the intentional and meaningful aspects of material culture.
Herein lies a problem.
As a broad rule of thumb, we appear to observe an interesting relation-
ship between scales of analysis and scales of resolution (see Fig. 4.1). As
we deal with the elemental and microstructural properties of artefacts,
the ¬ner-grained and more abstract the scale of resolution (e.g. trace
element analysis, stable isotope analysis or scanning electron microscopy)
the less inclined we feel to interpret our results in terms of past social
practices. What is more, the more abstract the results, the greater the
scale of the interpretative framework we feel we need to operate at in
order to make sense of the results. Conversely, as we turn to deal with
the macrostructural properties of artefacts, the less abstract our object of
analysis (e.g. microwear analysis on stone tools) and more coarse-grained
our resolution, the more we feel inclined to interpret our results in terms
of past social practices.
For example, when we employ elemental analyses of pottery we tend
to situate our results in very broad schemes of interaction and ex-
change (Tite 1996). We tend to say less about individual acts of pottery
68 Archaeological theory and scienti¬c practice

production; rather, we feel more secure discussing production in terms of
very broad frameworks. On the other hand, when we employ much more
coarse-grained techniques such as petrology we feel more con¬dent in
assessing strategies of pottery production in a localised framework.
This problem is demonstrated by the recent work of Arnold et al.
(1999, 2000). Their work involves the use of instrumental neutron acti-
vation analysis (INAA) as a means of determining discrete compositional
groups of pottery collected during ethnoarchaeological ¬eldwork from
Guatemala and Yucatan. Their work was aimed at testing the ability of
INAA as a means of resolving groups of pottery at a number of lev-
els of interaction. The results of this analysis indicated a high degree of
resolution in distinguishing between communities of potters working in
different regional centres of production. The results were also able to dis-
tinguish between communities that used at least one resource in common.
However, the analysis was unable to distinguish between different pro-
duction households within a single production community. This example
indicates that while studies that utilise elemental analysis are extremely
useful in resolving differences between the characteristics of materials at
the macroscale level over a large geographical scale, the same technique
is unable to resolve differences in the composition of objects at the mi-
croscale or local level. I use this example not to indicate that either the
technique or its application should be viewed as problematic, but rather
to illustrate how closely we need to be aware of how our techniques relate
to our scales of analysis.
There is a further rule that applies to the issue of abstraction. Often
there is a tendency to feel that the more complicated our techniques, the
better our science. Simple techniques are associated with crude levels
of analysis; complex techniques are associated with sophisticated analy-
sis. This is obviously a generalisation, but we need to critically examine
our tendency to relate complexity with sophistication, since on critical
examination there is no obvious correlation between complex analyti-
cal practice and sophisticated analytical interpretation. If we are to give
this point serious consideration, we need to be aware that there are two
crucial issues we must tackle. First, we need to consider what it is that
we want to say with materials science in terms of past social practices.
Second, we need to be aware that, as indicated in the previous chapter,
materials science techniques are situated within the broader process of
interpretation. If we take this on board, we might judge the sophistica-
tion of our analytical techniques not just by the complexity of our results
but by the sophistication of the resultant interpretation. In other words,
we need to consider not just the technique, but how we interpret the re-
sults of that technique. This symmetrical mode of analysis therefore pays
Practice, scale and narrative 69

attention to the mode by which we tie the analytical results together with
our interpretative frameworks.

Narratives: scienti¬c facts and interpretative ¬ctions
The historian of science Donna Haraway (1989) points out that in the
modern West we have tended to rigidly divide fact from ¬ction. She ar-
gues that, in order to maintain a belief in the veracity of science, we have
created a distinction between the results of scienti¬c practice and the
results of imagination created by novelists. We have maintained a rigid
divisive line between the non-constructed and the constructed. Facts are
hard-edged, real and non-constructed, while ¬ction is soft, related to fan-
tasy and entirely constructed. Throughout this book I have suggested that
scienti¬c facts are in some ways socially constructed. The intention here
is not to pour scorn on the truth or falsity of scienti¬c knowledge, but to
draw attention to the way in which we practice the creation of scienti¬c
facts. One of the most important aspects of this argument is that once
we realise that we are constructing or creating facts we become aware
that we may, within reason, construct them in multiple ways. In other
words, we need to consider what it is we want to know before we begin
the practices of fact construction. If we take this argument on board,
we need to realise that just as ¬ctions are created within certain narra-
tive frameworks, so facts are situated in certain narrative frameworks.
As Knorr-Certina (1981, 88“91) has it, all scienti¬c facts are situated
within wider frameworks associated with broad-based research goals and
problems that transcend the individual laboratory or ¬eldworker.
The proposition that we might treat facts as constructed should not be
perceived as a degradation of the veracity of facts; rather it provokes an
awareness of the relationship between our analyses and wider scienti¬c
questions. Equally we may note, with Ingold (1993), that ¬ction need not
be perceived as entire fabrication, a process of covering up or concealing
the raw facts. Rather ¬ction may be considered as a device for opening up
or providing an entry into an argument. We need to embrace the realisa-
tion that both fact and ¬ction are equally constructed, although they are
constructed in different ways and for different audiences (Knorr-Certina
1981, 94“133). Our task then is to critically reconsider the relationship
between different types of construction and the kinds of statements we
might make using knowledge of different qualities.
A consideration of narrative leads us to think about some of the issues
related to the nature of abstraction, and the narrative and analytical scales
at which we operate. The issue of narrative forces us to realise that we
need to be mindful of what it is that we are trying to say about the past;
70 Archaeological theory and scienti¬c practice

we have to concern ourselves with the kind of narrative scales at which
we work. Pluciennik (1999) notes that different types of narrative relate
to different scales of analysis. We might usefully distinguish between two
broad narrative scales, the macroscale and microscale. Macroscalar nar-
ratives include those in which events are abstracted and linked to common
processes. We can recognise this form of narrative in the generalising ap-
proach of rationalism in which human activity may be linked together by
a series of cause-effect laws. This is the narrative scale at which much
culture-history operates; cultures are used to stand in for characters on
the world stage. Similarly this narrative scale is often deployed in pro-
cessual archaeology, since the aim is to work towards building a series of
generalisable laws of human behaviour. Such laws of behaviour are often
set within broader narrative schemes, such as those of social evolution
(Pluciennik 1999, 662).
On the other hand, microscalar narratives emphasise the individual or
historically speci¬c, events are linked by very speci¬c reason-action expla-
nations and there is little claim to extrapolate to wider causal explanations.
Microscale narratives are context dependent, and we can see an obvious
alliance between this scale of analysis and relativism. Such approaches
have been employed most readily in post-processual and interpretative
accounts of the past in which human action is viewed as situated in, and
structured by, local cultural and historical contexts.
But are we required to operate with just two polarised scales of analy-
sis, or are the different narrative scales interlocking? These questions have
had a considerable currency within both history (Braudel 1980; Ginzburg
1982) and archaeology (see Bintliff 1995; Knapp 1992 for archaeological
overviews). But they have recently been re-appraised by a number of au-
thors (Dobres and Hoffman 1994; Gosden 1994; Hodder 1999). Hodder
(1999, 130“1) suggests that the two narrative scales are incommensurable
since localised events are underdetermined by large-scale structures “ we
can never determine all the factors that lead to events, and they cannot be
explained by large-scale structures alone. Despite this pessimism Hodder
goes on to describe, using the example of Otzi the ice-man, how individ-
ual actions are related through practice to long-term structures. Dobres
and Hoffman (1994, 213), discussing the scales of analysis required in
technological studies, suggest that while microscale and macroscale anal-
yses are different, it is possible to work from the local, microscale towards
the macroscale. They suggest that ˜dynamic social processes operating at
the microscale may have impacted upon or substantially contributed to
more macroscalar processes™ (ibid. 213). Similarly Gosden (1994, 15“17)
notes that each action is linked in space and time to other future or past
actions. These actions form chains which together form networks, and
Practice, scale and narrative 71

these networks form what Gosden describes as ˜frames of reference™. It is
these frames of reference that together make up large-scale structures of
action. Our conceptualisation of scales of analysis must therefore encom-
pass the point that, both historically and geographically, the scales within
which individuals act are constructed by societies, and these prescribed
scales of action are both monitored and transformed by the individuals
that act within them. The scales within which we act are multidimen-
sional “ for example, the same individual may operate simultaneously at
the level of personal interaction, as a member of a localised social group,
and as a member of a larger social group, such as a nation-state. At each
point the ability to work within and transform these scales depends on
the action taken with regard to these various scales of analysis.
We need to keep in mind the point made by Barrett (2000, 63) that
˜it is possible to write narratives which mark the passing of time but
without reference to agency. Such narratives work at a level of abstrac-
tion “ economic processes operate without labour, ideologies arise with-
out the struggle to maintain belief “ and the reasons for choosing such
abstractions must be explicitly understood.™ Therefore we need to remind
ourselves that, although macroscale processes may provide a basic level
of explanation, our analytical framework should commence by bringing
into focus the scale of human endeavour and interaction “ at the mi-
croscale level “ that provide the structural conditions that inform these
wider processes.
How do these issues relate to the analyses performed by materials scien-
tists and why are they important? I have already argued that the material
world is not socially and culturally neutral: it is imbued with meaning.
However the ways in which we can ˜get at™ some aspects of the mean-
ingful nature of the material world using materials science techniques
depend very much on the type of technique that we use and the an-
alytical scale at which that technique operates. This problem is again
illustrated by the work of Arnold et al. (1999). The samples collected for
analysis were microscale in nature “ they were derived from individual
potters in each area “ but the technique employed (INAA) to analyse the
samples, although successful at distinguishing compositional difference
at the macroscale level, was unable to distinguish between samples at
the microscale level at which they were sampled. The technique there-
fore restricted subsequent analysis to the discussion of broad macroscale
The main point I am making here is that if we undertake our anal-
yses within a narrative framework that only attends to the macroscale,
our interpretations will tend to be restricted to this scale of analysis. If
our analyses attend to the microscale, then our analysis may also be able
72 Archaeological theory and scienti¬c practice

to inform us about processes that occur at greater scales. I would
argue, then, that the microscale represents the most appropriate scale
with which to initiate our analyses. This is because attention to ¬ne-
grained microscale analysis will not only allow us to consider ¬ne-grained
localised differences in material culture patterning, it will also enable us
to ˜scale up™ our discussions to consider localised differences within a
wider macroscale framework. This attention to scale has important im-
plications with regard to our sampling procedures and the subsequent
analysis of results.

Scale, structure and narrative
Curiously, we tend to see macrostructural analyses of artefacts associated
with microscale narratives, and elemental or microstructural analysis of arte-
facts linked to macroscale narratives. Why is this? Because as we move
away from the tangible (and observable) analysis of material culture to-
wards more abstract modes of analysis, we begin to suffer a crisis of
con¬dence. We begin to question our ability to assess the signi¬cance of
our results within the framework of human observation. This is due to the
fact that when we analyse the concrete macrostructure of materials such
as potsherds or ¬‚int ¬‚akes our results are amenable to empirical observa-
tion. This observation may be with the naked eye, or it may be enhanced
(or transformed) by optical instruments. Furthermore, the practice of
analysis does not require the destruction of the artefact in order to pro-
duce results. Therefore our interpretation of these results can be re-
assessed at any time hence.
However, when we consider abstract elemental and microstructural
analytical procedures, such as stable isotope analysis or trace element
analysis or petrology, designed to examine the microstructure of materi-
als, the destructive nature of our analytical techniques requires that we
sample. Signi¬cantly, each distinct sample is derived from a different lo-
cation and therefore no two samples are coterminous. Each sample is
unique. This means that not only is each sample non-observable in the
empirical sense, but also we cannot return to the same sample location
on the artefact and reassess it at a later stage.
We employ a number of devices in order to circumvent this problem.
When we present the results of our analyses, we ¬lter them through a
series of inscription procedures. The analysis of concrete observable arte-
facts, such as potsherds and ¬‚int ¬‚akes, is fairly straightforward. We sim-
ply observe, measure and sketch the artefacts. The relationship between
our observations and our interpretations is viewed as reasonably unprob-
lematic. We present a text-based interpretation of our analysis alongside
a quantitative analysis, employing simple histograms related to minimum
Practice, scale and narrative 73

number of vessels, or numbers of different types of ¬‚int instrument. We
also use a further mode of representation “ artefact drawings “ which are
presented as accurate and conventionalised representations of artefacts
˜as they really are™. These drawings provide an illusion of concrete famil-
iarity. Because we are trained to recognise and interpret these drawings,
we therefore ˜see™ these artefacts in their natural state much as if we were
observing them in reality. For us they are real and empirically observable.
When we come to examine more abstract modes of destructive anal-
ysis, our observations are ¬ltered through far more complex inscription
devices. First, as noted in chapter 2, instrumental techniques such as
XRF, ICPS, NAA and GC/MS encode information. They transform the
raw material analysed into data. We observe further layers of encryption as
data are transformed statistically. Finally these data are presented visually
in the form of three-dimensional graphs or histograms. The material
world has undergone a series of transformations as it changes from raw
material to numerical data to graphic data. My point is that due to the
abstract nature of this kind of information, and the destructive, non-
repeatable nature of analytical practice, we employ a series of inscription
devices, statistics and graphs, as a means of persuading ourselves of the
veracity of our data. This is not to say that these results are not ˜real™, or
do not adequately represent archaeological facts. Rather, I simply wish to
point out that certain analytical procedures and the presentation of the
results of these procedures are constructed. Not only this “ they are also
subject to multiple layers of construction.
To return to the ¬rst point made above, we have problems with em-
ploying abstract materials science analyses as a means of discussing past
social practices. But this is precisely because of the nature of our analyti-
cal practices. We create a distance between past social practices precisely
because we disassociate material from the artefact and subject it to a
series of instrumental, mathematical and graphical transformations. A
far greater number of stages of analysis and presentation (each associ-
ated with a fresh stage of transformation) lie between, say, NAA anal-
ysis of pottery and macroscopic analysis of pottery. Therefore we ¬nd
it problematic to retrace our steps and relate our analysis to some form
of past social practice. There are simply too many layers of transfor-
mation between speci¬c social practices and the results of instrumental
We ¬nd it easier to start from a position in which we examine the mi-
croscale contextual differences between the observable macrostructural
properties of artefacts. We then ¬nd it simpler to transform our under-
standing of this observable data to the next stage of interpretation, in
which we try to say something meaningful about past social practices in
relation to our empirical observations.
74 Archaeological theory and scienti¬c practice

However if we start from an initial position in which we have already ab-
stracted material from an artefact in order to examine the microstructural
differences in artefact composition, we are then required to make a far
greater series of inferential steps back towards the empirically observable
artefact in order to say anything meaningful about past social practices in
terms of our microstructural observations. Rather than torturously retrac-
ing our steps, we instead opt to ¬t our observations into wider generalised
schemes of knowledge in order to make sense of them “ we therefore rely
on comprehending the results of our analysis in a macroscale narrative
I argued above that attention to microscale contextual differences al-
lows us to eventually ¬t our observations within wider macroscale nar-
ratives. While this is desirable, I have also outlined some of the prob-
lems we face when we attempt microstructural analyses of artefacts. We
face problems because, due to our analytical practices, our results are
abstracted from the artefact and we therefore ¬nd them dif¬cult to in-
terpret in terms of microscale contextual differences. The problems of
transformation and subsequent interpretation arise in part because of
the archaeological practice of shearing the object from its original ar-
chaeological context. As indicated in the previous chapter, not only is
the material sheared from its context, but that context very rarely trav-
els with the object through subsequent stages of analysis. This relates
to my point that often materials scientists have little knowledge of the
historical or cultural context of the material on which they base their
analyses. The decontextualised object actively creates the ˜brick wall™ of
interpretation that divides materials science from interpretative archaeol-
ogists. Again, our practices structure the mode by which we construct our
If we are to get around this problem, and work towards knitting to-
gether tighter and more plausible alliances between materials science and
interpretative archaeology, then we need to work back along the chain of
transformations described above. At each stage the materials scientist
must keep in mind two aspects of data:

1 While data are transformed through presentation, the artefact remains
the same. The data are derived from an archaeological artefact of some
form, and our interpretations must be plausible in terms of the material
and physical nature of that artefact.
2 The artefact is grounded in a context. Just as the artefact remains the
same, despite the subsequent alteration of the data derived from the
artefact, so too the ¬xed nature of the context of the artefact must also
be kept in consideration.
Practice, scale and narrative 75


artefact artefact
Scientific Interpretative
and and
analysis archaeology
context context

Figure 4.2 Artefacts and their contexts as boundary objects

Both of these features of analysis, the artefact and its context, provide
what historians of science describe as ˜boundary objects™: an object that
keeps its constancy, although the data and alliances surrounding that ob-
ject change (Star and Griesemer 1989). Such boundary objects, in this
case the artefact and its context, allow distinct groups of people, in our
case materials scientists and interpretative archaeologists, to ¬nd some
form of bridging device that provides some constancy of purpose and
interpretation in what are otherwise distinct and contradictory projects
(Fig. 4.2). In the case of archaeological analysis, the property that de¬nes
and maintains the constancy of the boundary object is the documentation
that surrounds the artefact “ it is this that gives it contextual de¬nition.
It is the methodological constancy “ the retention of documents con-
cerning context “ at each stage of the analysis that allows us to begin to
bridge the gap between the two approaches to the archaeological artefact.
This constancy enables the results from the preliminary stages of labora-
tory analysis to have an effect on the subsequent stages of interpretative
analysis, and vice versa.
There is, of course, a further important problem, and this concerns
resolution. If the results of microstructural analyses are very coarse grained,
then we feel little con¬dence in differentiating one set of results against
another at the microscale contextual level (see Tite 1996, 241“2). This
problem with resolution often affects the way in which we sample arte-
facts. As Tite (1999, 197) observes, the chemical analysis of artefacts
76 Archaeological theory and scienti¬c practice

using minor and trace element analysis may provide a compositional ˜¬n-
gerprint™ for characterising objects made of the same raw materials. How-
ever in some cases, as with pottery and metals, we may have problems
due to the variability in chemical composition of source materials. In
an attempt to surmount this problem there is a tendency to sample large
numbers of objects in order to place them in broad compositional groups.
It is this practice of homogenising numerous samples that then prevents
us from discussing differences between artefacts except in terms of wide
macroscale geographical or chronological differences.
How are we to get round this problem? While problems of resolution
are best solved by the application of alternate methodologies, we may also
consider the possibility of multi-stage sampling procedures. These could
commence with techniques designed to securely observe the differences
between artefacts in terms of macroscale differences and then, armed
with the results of this analysis, proceed to a stage in which more ¬ne-
grained microstructural techniques are applied. As Kingery (1996, 178)
rightly notes, we need to be aware that no single technique is entirely
secure in its ability to provide conclusive results. It is therefore bene¬cent
to work with a number of complementary techniques.
A further problem also relates to the nature of the on-site sampling of
artefact assemblages. Here I am not especially interested in the nature of
sampling in terms of the recovery of a representative sample of the assem-
blage population (Cherry 1978). Obviously these issues are critical to our
understanding of any archaeological site; however, I am more interested
in how we move from on-site sampling to the sampling required of ma-
terials scientists in a laboratory setting. As the example above indicates, if
we are to understand the microscale differences in artefact assemblages
we need to consider sampling according to the contextual differences of
the assemblage, rather than at random. It is only by adopting a contextual
approach to sampling artefacts that we avoid the problems associated with
homogenising artefacts in subsequent analysis. If we are to meaningfully
categorise artefacts in terms of compositional groups and relate compo-
sitional differences to decisions made in selecting materials for artefact
manufacture, we must be careful in our use of this sampling procedure.
As materials scientists, if our analytical results are to have any bearing on
our understanding of issues of artefact categorisation and construction,
we are required to tie the results of our analytical methods back to the
speci¬c raw materials that constitute the artefact.
We should conclude that these issues are critical to scienti¬c analysis
since the issue of narrative scale lies at the heart of what we sample, how
we sample, and how we relate this analysis to a speci¬c framework of un-
derstanding. What I am suggesting here is that our attention to issues of
Practice, scale and narrative 77

scale as materials scientists needs to be manifold. We must relate our the-
oretical interests to a set of multi-layered practical considerations. First,
we need to consider what kind of narrative scale we wish to operate with.
Second, we need to decide what kind of techniques are most appropriate
to answering the kinds of questions we are interested in, and how we may
go about sampling a given artefact in order to produce the answers to
these questions.

Linearity and re¬‚exivity in practice
Having discussed some of the problems associated with the scales of
scienti¬c analysis in relation to material culture, I will now shift the dis-
cussion towards considering another scale of analysis “ how this overall
approach to archaeological science integrates with wider interpretative
research questions. Traditionally scienti¬c analysis has proceeded along
a familiar path from hypothesis, to experiment, to the refutation or con-
¬rmation of the primary hypothesis. At this point the hypothesis is refor-
mulated, abandoned or retained. This process of investigation is charac-
terised by its linearity. Furthermore, as we saw in the previous chapter,
a linear process of excavation, post-excavation and publication similarly
de¬nes archaeological practice. Here the analysis of objects and samples
retrieved from the site is traditionally undertaken off-site and after the ex-
cavation process has ended. I have already argued that this rigidly linear
structure promotes a series of problems in the subsequent interpretation
and publication of archaeological data.
As an alternative, Hodder (1999, 80“104) has proposed what he de-
scribes as a re¬‚exive methodology in which both description and inter-
pretation are simultaneous. One aspect of such a methodology involves
the classi¬cation and analysis of those objects and samples otherwise
analysed under the aegis of post-excavation at the stage of excavation “
in other words on-site “ as objects and samples are retrieved from the
ground. This process allows the interpretation of artefacts and environ-
mental samples to actively re¬‚ect back on the methodology of excavation,
allowing for an increased sensitivity to the exigencies of the site itself and
thereby enabling a closer ¬t between the interpretation of site and arte-
fact. This suggestion is commendable and certainly side-steps some of
the problems associated with the process of ˜explosion™ outlined in the
previous chapter. However we must be wary of two problems, the ¬rst
philosophical, the second practical.
First, we must be careful when utilising the on-site expertise of ar-
chaeological scientists that science does not simply stand for objectivity
and that the analyses of objects and samples do not remain as immutable
78 Archaeological theory and scienti¬c practice

statements of fact which are then deployed back on-site in a rigid and
prescriptive manner. Rather we must be aware that, as suggested above,
the formulation of scienti¬c analyses must be structured as much by con-
textual considerations as by recourse to a set of objective scienti¬c data.
Second, the practical aspects of retaining on-site specialists depend very
much on the contingencies of national laws concerning the exportation of
scienti¬c materials beyond national borders. More importantly, funding
considerations may also hamper such an approach “ since few funding
bodies are prepared to part with post-excavation monies until the signif-
icance of the site has been established through excavation.
Nevertheless the broad notion of re¬‚exivity may be harnessed in other
ways. For example, it is common for many ¬eld projects to be conducted
over a number of seasons, and it is entirely feasible that specialist scienti¬c
analysis may proceed off-site between excavation seasons. This process
would potentially allow information from such an analysis to be fed back
into the excavation process in subsequent seasons. Ultimately, while the
place and position of the archaeological scientist or specialist is important,
we also need to consider how interpretations are arrived at in practice and
how these interpretations cohere with the results of others within the exca-
vation report or within other mediums of reportage. The point here is that
close contact between specialists and other practitioners is required at all
stages of analysis in order to produce a coherent and satisfying account, an
account in which knowledge is not presented as disparate and fragmented,
but in which each area of knowledge mutually structures the other.

Case study: interpretation and materials science
In the previous section I discussed the importance of considering research
questions in terms of certain narrative scales. This process involves three
1 Framing a set of interpretative questions and setting out to answer them
by working with artefactual data which are ¬rmly related to context.
2 Examining this data using analytical procedures with enough resolution
to distinguish between the contextual differences in artefacts.
3 This procedure should also be re¬‚exive in its approach: working back
and forth between research questions, site and artefact.
In order to examine what this interpretative approach to materials sci-
ence might resemble, I will consider the work of Peter Schmidt (1996,
1997), since his studies encompass many of the approaches to archae-
ological science advocated above. Schmidt is interested in studying the
nature of iron technology in East Africa. His stated aim is to ˜recuperate
Practice, scale and narrative 79

Chimney: constructed
of slag and mud

Iron ore and


Iron ore

Ground level

Figure 4.3 A schematic view of the Haya furnace

the history of African iron technology™ (Schmidt 1997, 1). This is be-
cause Africa has long been viewed through the lens of primitivism. For
Europeans, Africans lacked history. Europe was the home of the techno-
logically advanced and literate, while Africa was inhabited by the tech-
nologically inferior and illiterate. Such views often manifest themselves
in depictions of the development of iron technology, which usually use
prehistoric Europe as their model. Due to this impoverished view of iron
technology Schmidt (1996) was interested in examining the technology of
iron smelting and investigating the use of a technological process known
as preheating. This process involved heating gases running through the
tuyeres into the furnace, which enabled a higher reducing temperature to
be achieved during smelting and produced a higher yield of pure metal
(Fig. 4.3). The very existence of this process was contested by a number
of people on the grounds that the process had not developed in Europe
until the Industrial Revolution. At the outset Schmidt™s aims were to
investigate this particular smelting technology as a means of contesting
the colonialist bias through which much African technology is viewed.
His approach to this problem include ethnoarchaeological documenta-
tion, thin-section analysis of slags and blooms, and an analysis of the
80 Archaeological theory and scienti¬c practice

performance characteristics of tuyeres using a thermocouple during use
and through visual examination after use.
Schmidt approached a number of retired Haya smelters and black-
smiths who inhabit the region around the western edge of Lake Victoria,
Tanzania to reproduce furnaces in the present that allowed the perfor-
mance characteristics of both furnace and tuyere to be determined. The
investigative procedure itself was not linear in its aims and outcomes.
The creation of furnaces and the production of iron blooms were fraught
with problems from a number of quarters: social, ritual and technolog-
ical. Schmidt™s account draws together each element of the procedure
and allowed him to establish the veracity of the preheating process, using
temperature measurement of the tuyeres and thin-section investigation of
the slags and iron blooms.
From this analysis the existence of the preheating technology was con-
¬rmed. The con¬rmation procedure involved a cohesive alliance between
the scienti¬c examination of the furnace components and the products
of the furnace with the experimental and ritual procedures and oral testi-
monies of the local iron smelters. This then led Schmidt to investigate the
time-depth of this technology through the excavation of ancient smelting
and forging sites. This exercise established that smelting technologies of
this kind potentially dated to the second century AD, con¬rming the tech-
nological process of preheating to be a process that predated European
expansion by at least a millennium.
This investigation of the antiquity of East African iron smelting pro-
cesses is admirable in itself since it provides a ¬rm platform from which
to challenge Eurocentric beliefs concerning the technological abilities of
non-Europeans (see Harding 1991, 191“218). But Schmidt goes beyond
this to examine how iron production relates to ideological beliefs and the
construction of local histories. His observations of smelting and forging
encompass both the ˜functional™ dimensions “ the shape of the furnace,
the position of the tuyeres, the temperature of the furnace and the phys-
ical and chemical characteristic of the iron blooms “ and the ˜ritual™ di-
mensions, such as the songs performed while smelting, the deposition of
smelting charms beneath the furnace, the construction of spirit houses,
the observation of taboos, etc.
As in many parts of Africa, the symbolism of iron production here is
associated with sex and fecundity. It is men who undertake iron produc-
tion and they must abstain from intercourse during the smelting pro-
cess. In Buhaya, the area studied by Schmidt, the technological process
of production bears many similarities to reproduction and birth. More
widespread practices also make this explicit connection. For example
Herbert™s (1993) study of iron smelting in West Africa notes how the kiln
Practice, scale and narrative 81

is modelled both conceptually and physically as a female ¬gure; the fur-
nace is fed with charcoal and iron ore, it excretes slag and gives birth to
an iron bloom. The sequence of activities relates the production process
to the biological process undergone in human birth.
The presence of the phallic-shaped tuyeres and the use of bellows to
force air into the furnace complete the sexual metaphors that pertain to
Haya iron production. Sex is considered to be ˜hot™ “ like iron production.
Therefore an important requisite for fertilisation is that the phallic tuyeres
should be hot; for this reason they are inserted deep within the furnace,
preheating air as it enters the furnace and enabling higher temperatures
to be reached. An understanding of the coherence between the symbolic
and physical elements of action clari¬es how it is that the preheating
process came into existence. Without an understanding of the symbolism
surrounding iron production, any elucidation of the preheating process
would have been technologically one-sided “ and would have taken little
account of the social.
Such an analysis allows us to gain a detailed understanding of how a
particular non-European technology was developed and used within a
distinct world of cultural meaning. Furthermore, it enabled Schmidt to
investigate the manner by which iron production was tied to both the
mythology and kingship ideology of the region. Forging was an impor-
tant element of the inauguration of kings since iron, like the power of the
king, was related to fecundity. What is more the association between a
mythological iron tower and the inauguration of a king of some histori-
cal importance was demonstrated to have a concrete relationship to an
important iron production centre. This analysis strengthens our under-
standing of the time-depth of oral histories in the area.
This discussion highlights the use of a gamut of materials science tech-
niques from the domains of engineering, chemistry and physics within
a research framework derived from social anthropology or interpretative
archaeology. The critical point I wish to draw out is that the examination
of East African iron technology began with a signi¬cant question “ can
we use a scienti¬c examination of technology to rewrite the conventional
colonialist histories of Africa? During the investigation, the data derived
from detailed microscale scienti¬c analysis were allied with an investiga-
tion of symbolism that not only explicated certain problems inherent in
our understanding of the nature of belief systems and history but also
clari¬ed our understanding of the technological process itself. The mode
of investigation moved from the microscale to examine the macroscale
historical processes and, having done this, returned to explicate the mi-
croscale. This form of investigation enables us to envisage how it is that we
may proceed with laboratory-based scienti¬c analysis while also retaining
82 Archaeological theory and scienti¬c practice

an eye on the historical and cultural context of the objects and samples
that we analyse. Such a process encompasses a shift between scales and
also employs a re¬‚exive approach, since it works back and forth between
European and non-European understandings in order to create a cohesive
and holistic interpretation.
Having examined a case study that emphasises the alliance between
materials science and material culture studies, I now wish to extend this
investigation in order to examine how we might use a speci¬c kind of
microscale analysis, the scale of the human life cycle, as a means of under-
standing the production, use and deposition of material culture. I will also
explore how, within this framework of analysis, we can employ materials
science techniques to open up our understanding of these processes.
5 Material culture and materials science:
a biography of things

Introducing a biographical perspective
One of the major points to emerge from our discussion of scales of anal-
ysis is that modes of analysis that only attend to large-scale structures
have little to tell us about how people lived and structured their lives on
a daily basis. In order to understand these issues, we have to consider
temporal and spatial scales of a more limited nature and duration, and
work from these to consider how activities performed at these smaller
scales transform larger-scale structures. What we are interested in, then,
is how material culture is used to create and maintain meaningful social
relations, relations that af¬rm the de¬nition of identity and belonging at
individual, local and wider scales.
One scale of analysis that provides a useful starting point is the human
life span (Gilchrist 2000, 325). The narrative structure of human life
cycles provides an extremely broad framework determined by biograph-
ically important events such as birth, life and death by which people
make sense of their lives. Much of the literature concerning the way in
which artefacts are invested with meaning focuses on their biographical
relationship to human beings.
The notion of biography has arisen out of our understanding of the per-
ception of objects in gift-based economies (see Mauss 1925 for his classic
delineation of the notion of ˜the gift™). Gifts, unlike commodities, are ex-
changed as a means of establishing relationships between people.Gifts are
not simply conceived of as objects distinct from people (a notion that is
critical to commodity-based exchange systems). Instead, things are often
considered to possess some of the qualities of people. If objects can be
thought of as having some of the qualities of people, then it seems reason-
able that objects have lives that conform to the same structure as those
of people: they are born, they live and they die. Just as we might think
of objects as intervening in human lives and becoming ˜animate™ in the
process (see chapter 2), likewise in many societies objects may be thought

84 Archaeological theory and scienti¬c practice

of as sharing aspects of human lives, and thereby becoming imbued with
biographical status.
A useful distinction between the concept of ˜biography™ and that of
˜use-life™ has recently been made by Gosden and Marshall (1999, 169).
Use-life is often framed from a materialist perspective: artefacts are cre-
ated, they have a ¬nite use-life, they become worn and are discarded.
The concept of biography, while embracing the insights that a use-life
perspective brings to artefact analysis, also encompasses the idea that
objects are used to construct and maintain social identities. This point
has been reiterated by Janet Hoskins (1998), who examines the way
in which objects are not only socially identi¬ed with particular peo-
ple, but also used to describe and relate their life histories. She is in-
terested in the way in which artefacts are used by people as a means
of making meaning of their lives. All in all, the concept of artefact bi-
ographies is a useful metaphor for thinking about the way in which
people and artefacts are mutually related over time. This concept is es-
pecially useful for thinking about how it is that social identities are ex-
pressed through the medium of artefacts over different stages of the their
How are we to examine this concept archaeologically? First and fore-
most we are aided in this project by the notion of ˜contextual archaeology™
(Barrett 1987; Hodder 1986; Shanks and Tilley 1987). This allows us to
consider the possibility that the meanings associated with an artefact are
not ¬xed; rather, they are transformed or altered according to the context
within which the artefact is situated. This means that we are able to exam-
ine how an artefact may change its meaning over the course of its life as
it shifts from one context to another. However, we also need to be aware
that the notion of artefact biographies is not to suggest that objects can be
read as ciphers for the human life cycle. Rather the notion of biography
encompasses the idea that objects are utilised to express differing modes
of identity at different points in their own lives, while various objects may
themselves be used to express varied kinds of identity over the course of
the lives of human beings (Sofaer-Derevenski 2000).
Thomas (1996) adopted this approach in his examination of the dif-
fering character of deposits from the British later Neolithic. For Thomas,
artefacts are used as active components in the process of constructing dif-
ferent identities, and identities are constructed through the juxtaposition
and manipulation of sets of artefacts. Tilley™s (1996) account of object
use and deposition in the early-middle Neolithic of southern Scandi-
navia lays more emphasis on the symbolic identities of artefacts at various
stages of their lives and examines the way artefacts, in particular axes and
A biography of things 85

pots, are perceived in relation to each other. This account re-emphasises
the point that different types of artefact are materially distinctive and
may be employed in the construction of different, but related, kinds of
cultural biography. Sofaer-Derevenski (2000) has clearly delineated the
mutual relationship between objects and identities, in her examination
of the signi¬cance of copper technologies (beads and arm“rings) to the
categorisation of gendered individuals at various points over the course
of their lives in the late Neolithic and Copper Age of the Carpathian
It is important to reiterate that if we are to examine the nature of the
relationship between people and things over the course of an artefact™s
life, we need to consider not only aspects such as exchange or deposi-
tion but also production and use. In other words, we need to empha-
sise the entire life of an artefact as the object of analysis. In doing this
we should also be aware that when we are considering whole classes of
artefacts, we operate at a scale of analysis which may transcend indivi-
dual biographies to construct instead an image of the ideal career of an
object (Tilley 1996, 248). While the concept of biography is a useful
way of understanding the use-life of a single artefact, we may also ex-
amine the nature of artefact biographies at much larger temporal and
spatial scales. Such an approach informed Bradley™s (1990) examination
of the nature of depositional practices from the Neolithic to Iron Age at
a pan-European scale. The biographical approach to artefacts informs
our understanding of the ˜cultural life of things™ at a number of scales of
I will now look at how social and cultural identities are expressed at
various stages of an artefact™s life, through the technology of production,
through use and consumption and through exchange and deposition. But
more importantly, I will examine how these issues might be clari¬ed and
given more resolution through the application of a series of techniques
from the ¬eld of materials science. Tite (1999) has usefully examined
some of these issues in relation to the ˜use-life™ of ceramics; here I wish to
broaden the discussion to include other types of archaeological material.
As Kingery (1996, 185) notes, the physical narrative of artefact produc-
tion, use and discard is enmeshed with the narrative of the utilitarian,
spiritual, emotional, creative and aesthetic life of artefacts. However, he
argues that much materials science begins with production and works
towards discard, while many interpretative accounts operate in the oppo-
site direction. In this chapter it is essential that we align our studies and
articulate the materials science perspective with a biographical approach
to artefacts.
86 Archaeological theory and scienti¬c practice

Making material culture: artefact production
and landscape
A number of studies of the procurement of materials have provided a use-
ful framework within which to examine the way that resource procure-
ment relates to environmental factors (Arnold 1985; Torrence 1986).
These broadly ecological approaches (see Matson 1965 for original for-
mulation of concept of artefacts and ecology) describe the parameters
within which people procure and use resources. Here the environment is
seen to have a crucial effect on the organisation and scale of production,
on distribution networks, etc. While these approaches are important in
delineating the possibilities and constraints available to people within dif-
fering kinds of environment, I believe that we need to consider not only the
environmental factors but also the cultural perceptions of that environ-
ment prior to the procurement of resources. Site-catchment approaches
such as ceramic ecology (Matson 1965) operate within a paradigm that
views space as neutral and homogenous. Moreover, the environment may
be treated with a degree of economic rationality, as an abstract framework
within which human communities act.
But there are other ways of conceptualising landscape and environ-
ment. We may consider the natural world to be appropriated socially
and culturally. As a component of the classi¬ed and lived human world
landscapes are heterogeneous, and various features of the human envi-
ronment will possess a multitude of different meanings. Such an approach
embraces an understanding of human experience as being intimately in-
volved in the process of actively interpreting, classifying, and being in,
the material world (Ashmore and Knapp 1999; Thomas 1996; Tilley
1994). Within this interpretative framework the initial appropriation of
natural products used in the technological processes of production is an
active part of this classi¬cation process. If we are to understand in more
detail how the natural world may be appropriated in the construction of
material culture, it is essential to look in more detail at the issue of place.
Places are an important component of the socially classi¬ed landscape.
Through the alteration and inhabitation of speci¬c places, people are im-
plicated in the land (Gow 1995), and as such different places are imbued
with meaning. What is more, through the history of inhabitation, places
also become identi¬ed with particular people (Weiner 1991). This un-
derstanding of place, which views places to be bound up with the history
and identity of people, also proposes that the inhabitation of place is char-
acterised by particular memories. For instance, Casey (1987) notes the
way in which memories are place-speci¬c. Activities conducted within a
particular place are an important part of the act of remembrance. This
A biography of things 87

perspective is particularly relevant to the way in which we view the pro-
curement and use of materials for the production of material culture.
Since landscapes are made up of different kinds of place, each associ-
ated with different memories and identities, the use and incorporation of
materials derived from different places is an important means by which,
through production, the identity of material culture may be initiated.
Indeed, the incorporation of materials from a series of different places
may be an important means of expressing particular types of identity
through production (Munn 1986; Battaglia 1991; Tilley 1999).
The use of speci¬c materials from culturally signi¬cant places is one
means by which the power of place is grounded (Chapman 2000a, 25).
This is because, by virtue of their material constituents, artefacts carry
the social and cultural signi¬cance of particular places beyond the con-
¬nes of a single zone of cultural signi¬cance. Although physically suitable
materials are necessary for the construction of artefacts, physical proper-
ties need not be the only thing to structure the use of materials in cultural
production (Tacon 1991). The production of material culture involves a
process of acquiring the physically and culturally appropriate materials
for the task, which may in itself involve a process of exchange between
social groups, possibly between boundaries de¬ned by age, gender or
kinship (Mackenzie 1991; Tacon 1991). These approaches to landscape
need not require the abandonment of the broad frameworks established
by an ecological approach to the procurement of resources. Rather, we
simply need to note that the selection of these resources will be deter-
mined by their cultural classi¬cation. I believe that this approach allows
us to consider the procurement process with a more ¬ne-grained degree
of subtlety.
Interestingly, Arnold has examined the nature of the categories in-
volved in the procurement and use of pottery temper, the raw materials
of pottery production, amongst the Ticul of Yucatan, from an ˜emic™
perspective (Arnold 1971). His ˜ethnomineralogical™ approach was con-
cerned with de¬ning and relating archaeologically observable data with
native categories. Importantly, while physical attributes such as colour
and hardness were seen as important, one of the key means of categoris-
ing material was by source (ibid., 27). Both the material (clay) and the
source were bound up in a single idea of place that de¬ned how the clay
was to be employed in the manufacture of pottery. Other archaeologists
have recognised the importance of place-speci¬c identities to the manu-
facture of material culture. For instance, the physical location of stone axe
quarries has been recognised to be of great signi¬cance to the subsequent
distribution of their products (Bradley and Edmonds 1993; Cooney and
Mandall 1998). Similarly, the location of copper mines has also been
88 Archaeological theory and scienti¬c practice

viewed to be critical to understanding their social role within the wider
landscape (O™Brien 1994, 1999).
A variety of analytical techniques may be applied to answer the ques-
tion of place-speci¬c artefact provenance, ranging from simple geological
provenance techniques such as petrology, used in the case of stone tools
and ceramics (e.g. Clough and Cummins 1979; Cooney and Mandall
1998; Peacock 1969; Rice 1987), to geochemical techniques such as
isotopic analysis, in the case of metals (Gale and Stos-Gale 1992). How-
ever we may also apply a number of chemical characterisation tech-
niques which range from optical emission spectroscopy (OES), used in
early obsidian characterisation studies (Renfrew, Cann and Dixon 1966;
Williams-Thorpe 1995), to techniques such as XRF, ICPS and NAA,
used in the analysis of materials such as ceramics (Neff 1992), metals
and stone.
When considering the issue of provenance, we must again be aware of
the issue of scale. Here we have to consider both the resolution of various
techniques, and the survey and sampling procedure prior to analysis. For
example, obsidian characterisation has been extremely successful using
fairly low-resolution procedures such as OES. Broad sampling proce-
dures coupled with OES have enabled researchers to establish exchange
patterns across the eastern Mediterranean. The use of NAA at a later
stage enabled increased resolution of speci¬c obsidian sources (Aspinall,
Feather and Renfrew 1972). Much of the subsequent analysis of the pat-
terns of exchange determined from these analyses has been ¬rmly placed
within a macroscale narrative of contact and supply zones for obsidian
procurement and exchange (Renfrew, Dixon and Cann 1968). Attempts
to understand the localised procurement of obsidian at places such as
Melos similarly subscribed to analysis within this broad framework
(Torrence 1986). If we wish to obtain a more detailed understanding
of obsidian procurement and distribution, then we are required to un-
dertake detailed ¬eld sampling methods (Hughes 1994) coupled with
high-resolution analytical techniques. This form of analysis linked with a
microscale examination of technology and production (see Dimitriadis and
Skourtopoulou forthcoming) should provide an increased understanding
of the social relations involved in more extensive exchange.
Similar problems of resolution have been encountered in the isotopic
analysis of bronze. Much work has been undertaken in re¬ning the res-
olution of the analysis of lead isotopes in bronze in order to provide
an increased resolution for the source of copper (Gale and Stos-Gale
1992), but the ability of this technique to provide precise information
on the sources of copper has been questioned (Budd et al. 1995). As ar-
gued previously, an integrated combination of techniques is required in
A biography of things 89

order to gain clear resolution. For example, the work of Joel, Taylor, Ixer
and Goodway (1995) combines the use of lead isotope analysis with ore
petrography to provide a distinct signature for the metal sources of the
copper mines of the Great Orme, Wales. Similarly, Mandal (1997) has
shown how the combination of petrology and XRF has distinguished the
products of two stone axe quarries in northeast Ireland.
Studies that utilise a combination of techniques coupled with ¬ne-
grained ¬eld-based sampling procedures should allow us to construct
detailed narratives concerning the place-speci¬c origin of artefacts. Such
analytical rigour enables us to tie together information concerning the
analysis of materials with a concomitant level of detail concerning the
provenance of materials. This detail then enables us to discuss the sig-
ni¬cance of the source of materials and the social relations involved in
their extraction with far more clarity (for such an attempt in relation to
stone axes see Cooney and Mandal 1998; Cooney 1999). Similarly, tech-
niques of sampling which pay attention to the exigencies of local geology
have also provided a detailed understanding of the source of clays used
in regional potting traditions (see Howard 1981; Cleal 1996).
If we are to answer questions regarding the microscale signi¬cance
of materials, then wherever possible we are also required to undertake
detailed regional surveys in order to obtain compositionally different ma-
terials from a variety of locations around the landscape. It is only by
attending to the twin problems of resolution of analysis and resolution
of sampling that we will be able to distinguish the distinct locales from
which materials are obtained. By attending to these problems of analysis,
we can consider the social signi¬cance of the locations from which these
materials are derived at a greater level of interpretative sophistication.

Shaping technology: production and materials science
As observed in chapter 1, the study of technology and production in ar-
chaeology has traditionally been a focus of study for materials scientists “
it is considered to be one of the primary subject areas in which science
has a clear and obvious role to play (Sillar and Tite 2000). Nevertheless,
studies of technology and production have fairly recently come under the
critical attention of anthropologists and interpretative archaeologists, as
well as historians of science and technology (Callon 1991; Bijker, Hughes
and Pinch 1987; Lemonnier 1992; Pfaffenberger 1988).
These studies have questioned the linear evolutionary narrative preva-
lent in many studies of technology. As we well know, the notion of the
progressive nature of technology is one of the prime structuring prin-
ciples of Western capitalism. As such, technology is often treated as an
90 Archaeological theory and scienti¬c practice

extra-societal force that acts independently of human beings. However
if we are to fully understand technology and its effects on society we
need to place human beings back into the picture (Pfaffenberger 1992;
Latour 1993; Dobres 2000). One of the simple points that arise from this
theoretical reorientation is the realisation that technologies are created
through intentional human action.
Technologies are created by techniques of the body. As Mauss (1935)
noted, these techniques are ¬rst and foremost cultural; moreover, they
are also culturally speci¬c. The creation of new technologies therefore
requires a cultural choice (Lemonnier 1993). There is never simply one
way to execute a technical action; there are many possible ways. What
is more, each choice in technical production is linked to the next by its
outcome. We are therefore able to distinguish a series of interlocking
technical actions, chaˆnes op´ratoires or operational sequences, that go to
± e
make up technologies and through which we are able to examine the
trajectories by which different cultural actions are executed on materi-
als. In short, different cultural choices ˜shape™ technologies in different
ways. Given this approach we need not view techniques as something to
which meaning is added on at a later stage, but rather see techniques as
complex phenomena that encapsulate symbolic considerations from their
inception (Dobres 2000).
Technologies, therefore, weave together the material, social and sym-
bolic dimensions of human life. We need not view culture as a determining
factor over the constraints of the material world. Rather we need to re-
member that, as with our analysis of science, our task is to examine the
creative alliances made between people and the material world; we should
think of technologies as the mode by which humans act through ma-
terial means. By studying technologies we simultaneously investigate how
it is that people make things work for them and how things act to struc-
ture subsequent human action. It is through this process of interaction
that technologies may be considered to create different categories of per-
son (Dobres 2000). At a basic level, personal identities are instantiated
through their association with particular kinds of activity; however, we
also need to realise that the structures by which technologies are organ-
ised also serve to structure the nature of interaction and the nature of a
person™s identity. Pfaffenberger (1999) discusses this point with regard to
the organisation of the production of canoes and yam-houses in two
cultural contexts in Melanesia and the organisation of tin mining in
nineteenth-century Britain. He notes that it is not so much the arte-
fact created, but the experience of making the artefacts that conditions
or disciplines people. I will take this point up further, later in this chapter.
A biography of things 91

There are a number of points we need to consider in relation to tech-
nology and materials science. First, materials science may be useful in
determining artefact composition, the sequence of artefact production
and the mechanical performance of artefacts. Determinations of the com-
position of artefacts may also reveal considerable details about the tech-
niques involved in production, and the mechanical abilities of material
culture. For example, elemental analysis of metals enables us to deter-
mine the precise smelting technology used in winning metal (Budd,
Gale, Pollard, Thomas and Williams 1992); similarly, techniques such as
petrology enable us to determine the composition of pottery and there-
fore the choices made in production. What is more, studies of ¬ring tem-
perature and atmosphere, whether through thermoluminescence studies
or simple observation of colour changes in native clays (Matson 1968),
enable us to discuss the parameters of choice and functionality in the
production. Techniques aimed at de¬ning the composition of parts of an
artefact, for example glazes on pottery (e.g. Mason and Tite 1997) or the
hafting adhesive for ¬‚int sickles (e.g. Endlicher and Tillman 1997), also
have an important function in determining the sequence of steps required
to achieve these effects.
We may also wish to employ materials science techniques in order to
reveal the sequences involved in the production of artefacts. For example,
xeroradiography is useful in ceramic analysis (Carr 1990) as a means
of understanding the techniques involved in ceramic production such
as coiling, slab building, etc. Similar analytical techniques may disclose
production techniques involved in the production of metalwork, such as
welding (Craddock 1995, 7; Killick 1996).
As Killick (1996) notes, the techniques involved in the production of
artefacts may be elucidated using image intensi¬cation methods, such as
light microscopy and scanning electron microscopy (SEM). Such tech-
niques of identi¬cation have been routinely used to examine the working
of metals and of other materials such as stone. Furthermore, techniques
such as SEM, coupled with elemental detectors such as energy disper-
sive analysis by x-ray (EDAX), have the added advantage of enabling the
composition of artefacts to be determined at the trace level.
Other materials science techniques aimed at determining the mechan-
ical properties of objects enable us to assess the relationship between the
material and non-material properties of artefacts. Measurements of ten-
sile strength developed in engineering may allow us to understand to what
extent mechanical properties were a consideration in the production of
stone axes (Bradley et al. 1992). Similar approaches to the assessment
of thermal shock resistance (Schiffer 1988; Skibo 1993) in pottery may
92 Archaeological theory and scienti¬c practice

also enable us to determine to what extent this was an important factor
during production.
In these enquiries we need to be especially aware of the mode by which
our materials analysis coheres with our theoretical framework. There is
a tendency for many materials scientists to adhere to an evolutionary
view of technological progress. For example, Craddock™s (1995) account
of early metal mining and production is situated very clearly within the
discourse of social evolution. Smelting and mining techniques are con-
sistently referred to as either primitive or advanced. In these kinds of
narratives we must take into account the view raised in the previous
discussion. Techniques and operational sequences are not simply pro-
gressive in order, but are determined by the cultural outlook of those
employing them. It is critically important that account is taken of the
coherence between the techniques employed and the symbolic system
within which they are situated. This is well illustrated in the example de-
scribed by Schmidt discussed in the previous chapter. In a similar vein


. 3
( 7)