<<

. 2
( 7)



>>

action.
A similar approach is suggested by Wylie (1993), who pays close at-
tention to the way in which interpretative statements about the past are
validated through the use of data. Wylie suggests that, rather than making
statements about the past through a series of inferential steps (a process
she describes as chaining), we make statements by a process of tacking
back and forth between theory and evidence. This allows us to follow a
series of strands of evidence to create something like a web of meaning.
This process involves drawing on interpretations made at a general level,
and following through the effect these interpretations have on the concep-
tualisations of the evidence at a more particular level. The advantage with
this interpretative process is that while chains of inference may be easily
broken down by the removal of a single link in the chain, the process of
tacking back and forth between theory and data involves developing many
more or less concrete links between theory and data. The validity of these
multiple links depends more on the number of their associations than on
their testability. In effect, we create networks in which our theories and
data are inextricably linked within a web of signi¬cance. Tilley (1993,
18) has likewise proposed a process of interpretation which stresses the
importance of the values of connection, heterogeneity and multiplicity as
aspects of an interpretative process which creates networks of signi¬cance
between concepts. This process of interpretation has overall similari-
ties with the interpretative activity described by anthropologists as ˜thick
description™ (see Geertz 1973).
Wylie™s approach to interpretation has been criticised for its non-
committal to one epistemological position or another, in which aspects of
both positivism and relativism are retained (Fotiadis 1994; Little 1994).
However, if we enter into this argument there is a danger of an in¬nite
regress into the epistemological de¬nition of precisely where the core of
rationalism might stop and relativism might begin. These are consider-
able problems if we wish to remain within the domain of philosophy.
Instead, I want to shift the argument slightly to what I hope is a more
26 Archaeological theory and scienti¬c practice

productive level by examining how it is that we create knowledge, through
exploring the practices of scienti¬c knowledge production.
To reiterate, we have seen that the interpretative positions of archae-
ologists broadly rest on the premises which characterise the natural and
social sciences; that is, that the world is divided into inanimate objects
and animate subjects. Each is studied in different ways, using differing
methodologies. But this interpretative fault-line is problematic, especially
if we are attempting to write an interpretative archaeology which incor-
porates knowledge produced using a scienti¬c methodology. We have
also seen that one way out of this impasse is to develop an interpreta-
tive strategy that involves moving back and forth between theory and
data, developing weaker and stronger connections between each strand
of knowledge. In the next section I will develop this approach to inter-
pretation by examining the way in which networks of interpretation may
be understood within scienti¬c practice.


Science and culture
In order to reconsider the position of science within the interpretative
practices of the social sciences we need to interrogate the theoretical
position of science in greater detail. The central problem with science
is that its entire practice has traditionally been treated as if it were in a
privileged position with regard to interpretation. While other forms of
knowledge are treated as paradigmatic, scienti¬c knowledge is treated
as distinct, being beyond the problem of paradigms. In short, scienti¬c
knowledge is true knowledge with a privileged access to nature ˜as it
really is™. As a correlate of this, scientists themselves are considered to
be neutral and unaffected by cultural presuppositions. Notably, Binford
(1983, 45“57) developed exactly this kind of argument in relation to
New Archaeology™s position with regard to the interpretative position of
Hodder and others. Similarly, although Kelley and Hanen (1988, 99“
165) pay attention to the social context of archaeology, they are also
interested in distinguishing between the ˜logical™ and ˜non-logical™ aspects
of scienti¬c endeavour, an endeavour aimed at retaining an objective core
for the discipline.
As Rorty (1991, 46“7) notes, much of the discussion related to this
problem has focused on the attempt to demarcate science as a partic-
ular form of activity, de¬ned by a special method or a special relation
to reality. For instance, he notes that Hempel (1965) solves the prob-
lem by constructing a logical method of con¬rmation without worrying
whether this was scienti¬c activity or not. Others, such as Quine (1966),
con¬‚ated the whole of knowledge with science by supposing that the
Science and culture: interpretative networks 27

vocabulary of the natural sciences provided the true understanding of
reality. It is precisely this position which brings with it a whole series
of problems since, as science explodes outwards to include all human
action and knowledge, much of that activity and knowledge must be
described as irrational since it will not ¬t with the vocabulary of the nat-
ural sciences. Quine™s proposition takes us no further since we are still
unable, from this de¬nition, to solve the problem of competing claims
to other forms of knowledge. If all knowledge is rational, then how do
we treat the knowledge claims that do not ¬t with this image of ratio-
nality, except by classifying other knowledge systems as pre-rational or
misguided.
Objections to the attempts of science to de¬ne itself as special or def-
initionally rational have come from a number of quarters, including re-
cent work in the anthropology of science. The aim of such studies has
been to oppose the sort of de¬nition proposed by Quine. Rather than
viewing science as a special form of activity divided off from society,
the emphasis has been on considering science as simply another form of
activity which is conditioned by culture (Elkana 1981; Franklin 1995).
For instance, Martin (1990) has demonstrated that the language used
to describe the science of immunology is shot through with metaphor-
ical statements of a gendered nature. Fahnestock (1999) develops this
argument through an analysis of the tropes “ or modes of description
and rhetoric “ used in the construction of scienti¬c facts. In a similar
vein Strathern (1992) has investigated the constitution of the notion of
˜natural facts™ concerning biological reproduction and shown them to be
cultural products bound up with our culturally speci¬c notions of kin-
ship and genealogy. Furthermore, Haraway (1989) has suggested that
science, rather than being a uni¬ed and rational procedure, contains
concepts that are structured by gender biases which actively affect the
construction of scienti¬c facts. Each of these works indicates that the
very construction of scienti¬c concepts cannot be seen as value neu-
tral and unaffected by culture (see Longino 1990). Thus, much of the
critique of science has involved bringing science closer to culture, and
bringing science and its practice into line with society as an object of
study.
These studies are important, in that their central area of study is the
way in which scientists represent knowledge (Bloor 1976). However, a
problem still remains. The problem stems from the fact that such stud-
ies simply use society as a means of understanding science. What these
studies are in effect saying is that although scientists think they are rep-
resenting nature accurately they are in fact representing nature through
the distorting lens of either culture or society (but see Haraway 1991,
28 Archaeological theory and scienti¬c practice

1997; Strathern 1991). This still leaves us with the problem of nature. In
each case nature is still perceived as having a prior existence. In one case
scientists view themselves as having a privileged knowledge of nature; in
the other case social scientists indicate that scientists™ representations of
nature are distorted. However, in this second case the notion of nature
as an immutable entity is still retained (see Latour 1993, 91“9). In order
to solve many of the problems of relativism and rationalism, and in order
to enable science to be integrated within the interpretative positions of
social theory, we need to dissolve the distinctions which are held to obtain
between nature and society, and between objects and subjects.
Latour (1993, 91) has observed that it is precisely the demarcated
nature of scienti¬c knowledge and its privileged access to nature which
causes so many of the problems surrounding notions of relativism and
rationalism. This is due to the fact that if we consider science to have a
privileged access to nature then the knowledge constructed through this
privileged position also allows us to view culture as demarcated, since
those who are able to ˜see™ nature in its true form are also culturally
exemplary or special. As Latour notes, one of the major achievements
of our rational and modern view of the world is our ability to demarcate
ourselves off from other cultures. Thus we achieve the view that, while
there are innumerable other cultures within the world, these cultures can
all be set up in opposition to Western culture, since only the West has
a knowledge system “ science “ which directly represents nature (1993,
100“4). It is this point which lies at the heart of rationalism.
If we wish to integrate interpretative and scienti¬c methodologies, we
need to re-conceptualise these relationships. What if there is not one
single nature, but multiple natures each conceived differently by multiple
cultures? We have already considered this possibility under the aegis of
cultural relativism. But rather than retaining the rationalist concept of
a priori nature “ as a means of judging cultural constructions against “
we must now entertain the possibility that nature is contingent rather than
constant. Thus, nature as well as culture and society might be considered
to be in a constant state of ¬‚ux. Such a theoretical position allows for the
possibility of studying both inanimate and animate objects and subjects,
or as Latour puts it object-subjects or nature-cultures.
But this position brings its own dif¬culties. If we consider not one
nature, but multiple natures, each placed in relation to multiple different
cultures, how do we consider science to actually operate? If we no longer
retain the notion of a single constant nature how do we explain science,
and what is more, how do we explain the rationalist dictum that science
works in multiple cultural contexts? In order to solve these problems we
need to study in detail exactly how science does work. In other words,
Science and culture: interpretative networks 29

we need to perform a more detailed enquiry into the anthropology of
science.


Science in action
By examining the approaches of the anthropology of science (Latour and
Woolgar 1986), I want to draw out two things. First, I wish to attempt
to elucidate some of the problems we may encounter if we consider both
societies and natures to be contingent, and second, I wish to examine how
we may use these theoretical viewpoints as a way of integrating scienti¬c
practice with the interpretative social sciences. I wish to consider here
three main issues: the construction of facts, the issue of translation and
the creation of networks.
We have already observed in the previous chapter that one of the most
important theoretical standpoints available to the natural sciences is the
notion that nature is constant. It is this standpoint that enables natural
scientists to create ˜facts™ through the observation of nature and then
utilise them in further study. Latour (1987, 1“17) describes this process
as the creation of ˜black boxes™. These ˜black boxes™ may be considered in
precisely the same way that scientists traditionally conceive of them “ they
are concepts or devices whose workings are too complex to explain; all
that is understood of them is the relationship between input and output.
One of the de¬ning features of the methodology of the natural sciences
is that once facts are created they are very rarely re-examined “ they are
effectively ˜black-boxed™. This is in contrast to the social sciences, where
previous ideas are constantly being reviewed and argued over. While sci-
enti¬c facts are seen to be the product of the direct observation of nature,
Latour shows that this process of observation is far more complex. In
opening up the ˜black boxes™ he reveals the symmetry between the in-
volvement of animate subjects, in other words scientists, and inanimate
objects, the instruments with which they view the world. Once facts are
created the role of the former is underplayed and the role of the latter
is promoted as supreme. In a similar vein, Callon, Law and Rip (1986)
suggest that rather than considering facts as static pieces of information,
they should be considered as active. In other words, facts gain their power
in scienti¬c practice through their social use by people (see also Haraway
1989, chapter 1).
In order to understand more fully the process that goes into the creation
of facts, we must examine the issues of translation. Facts are typically
presented in paper form, as published articles. Once they are created, they
become incontrovertible. While the arguments contained within these
texts are themselves open to textual analysis (see Latour 1987, 21“62), it
30 Archaeological theory and scienti¬c practice

is not possible to critique factual information itself “ this must be treated
as a given. In order to understand the production of factual information,
we need to progress from the texts in which facts are presented to the
laboratories in which these facts are created. Instead of facts being self-
evident and directly observable, Latour (1987, 64“70) notes that facts
are actually translated or transformed into factual information through
the use of scienti¬c instruments.
In order to clarify this point I will use an example from my own work
using Gas Chromatography (for detailed methodology see Evershed et al.
1990; Jones 1997). If I wish to demonstrate the existence of cattle milk
within a pottery vessel I am required to do a number of things. First, a
section of the vessel is drilled and an amount of pottery fabric is retained.
This is then re¬‚uxed in organic solvent for six hours and the resulting
mixture is then reduced. The mixture is then reduced over nitrogen
and dissolved in ethanol. It is then reduced to dryness and derivatised.
The derivatised sample is then reduced to dryness, redissolved in ethyl
acetate and then injected into a High Temperature Gas Chromatography
unit, which is linked to an integrator. After an hour eluting through the
column, the derivatised fatty acids present in the mixture will be indicated
as peaks of varying height and width which are printed out on paper by
the integrator (see Fig. 2.1).
For more detailed con¬rmation I may decide to inject a different sam-
ple of the same mixture into a Gas Chromatography unit linked to a
Mass Spectrometer. In this case I will be presented with peaks which
represent the fragmentation pattern of the fatty acids on a computer
screen. Finally, if I wish to be certain about the existence of cattle milk,
I will compare my results against those of a sample of cattle milk that
has undergone the same procedure. At no point am I directly observ-
ing fatty acids. Rather, the presence of fatty acids has been indicated
to me by a process of transformation afforded by a set of instruments
both chemical and electronic. These results will be presented in
a quite different format, as a series of histograms (see Fig. 2.2). The
˜raw data™ will have been ˜interpreted™ “ transformed and cleaned up
for publication. This is a critical element in the rhetorical methods we
employ to persuade our peers of the validity both of our methodolo-
gies and of our resultant conclusions. It is part of what Lenoir (1998, 4)
describes as the ˜materiality of communication™ “ what and how the ma-
terial means of communication represent when we communicate with
others.
At no point am I observing nature directly during this process. More
importantly, this process cannot be easily criticised by an outsider, es-
pecially one who is unaware of the procedures of transformation and
translation. Latour (1987, 21“62) makes the important point that in
Science and culture: interpretative networks 31




Figure 2.1 GC retention graph of sample SF 2,000 from a GC integrator
Figure 2.2 Histogram of sample SF 2,000 from GC plot
Science and culture: interpretative networks 33

order to criticise scienti¬c facts it is necessary to carry out the same pro-
cedures yourself. For most people this would be a costly and impractical
business, so facts are taken on trust and become unchallengeable. This
brings me to a further point, regarding the creation of facts and con-
cerning the earlier discussion of facts as ˜black boxes™. The production
of my facts concerning the presence of fatty acids which indicate cat-
tle milk within the pottery vessels I was studying involves the drawing
together and articulation of a whole series of other facts. These facts
are effectively ˜black boxes™ that have been previously generated and on
which I rely in order to generate further meaningful factual information.
If I were to re-examine these ˜black boxes™ from the point of view of fatty
acids, I would be forced to re-examine the chemistry of fatty acids, which
would ultimately lead me into a review of the chemistry of carbon, hydro-
gen and oxygen and ultimately to sub-atomic physics. If I want to enquire
about the validity of the Gas Chromatography unit or the integrator, this
would lead me into the elution characteristics of various gas phase par-
ticles on capillary columns, or for the integrator into the characteristics
of microchip technology. However, my interests are in the generation of
information concerning food use in a particular period of human his-
tory. Like most scientists concerned with the generation of more factual
information I have neither the time nor inclination to investigate these
other avenues of information. Instead, I take them as read; I ˜black-box™
them. We can observe from this example that prior factual information
is tied together to make a seamless whole. It is this process of tying in-
formation together which allows scientists to proceed with the creation
of more facts. It is this process of tying information together which we
may describe as a network, and it is to the working of such networks that
I will turn next.
It is important to realise that networks are not just made up of facts and
instruments that articulate together in order to create a seamless work-
ing whole. Networks are also composed of people (Hughes 1983). This
point is crucial, since by looking at the role of people in the creation of
facts and the ways these facts are further used in the creation of more
facts, the analysis is biased towards neither objects nor subjects. Instead,
the analysis shifts between objects and subjects and looks at the way
in which they operate together. Science and the technologies produced
by science are not then simply the result of inanimate objects articulat-
ing together; rather, they are a product of the forces of both animate
subjects (people) and inanimate objects (instruments). It is the articula-
tion of these processes together which creates a network, and networks
are composed of weaker and stronger associations between people and
things.
34 Archaeological theory and scienti¬c practice

However, due to the fact that science and technology are ˜black-boxed™,
the human factor in the process of technological production is removed
from the equation. Technology is then perceived to have a life of its
own; it has become an animate object (Latour 1987, 1993; Pfaffen-
berger 1992). Both the production and use of science and technology
involve the articulation of a whole series of factors, some of which may
be physical processes, others political, others economic. Yet all of these
are brought together as part of a network in order to create a fact or to
make things work. The key is to understand how the associations be-
tween these disparate social and physical factors create facts and tech-
nologies. Most important, due to the social nature of science and tech-
nology, and the accretion of facts as ˜black boxes™, we can begin to think
of the process of technological advancement as something like the pro-
cess described in the previous chapter as structuration (see Pfaffenberger
1992; Giddens 1984). Thus new technologies are structured by, and
built up from, the available factual and material resources that surround
them, rather than always being conceived in a scienti¬c and technological
vacuum.
In order to clarify the set of assertions made above we need to turn to an
example. Using the example of Diesel, the inventor of the diesel engine,
Latour (1987, 105“6) notes that the products of scienti¬c technologies do
not simply emerge fully formed. While Diesel had the initial idea of creat-
ing an engine that ran according to Carnot™s theory of thermodynamics,
this was not the end of the process of creation but the beginning. His ¬rst
ally was Lord Kelvin, a reviewer of his ideas who placed his considerable
scienti¬c weight behind the notion. However, in order to build the engine
he had to enlist the services of MAN, a German machines ¬rm. They
were involved because of their own wish to develop a more ef¬cient steam
engine. During the testing of the engine there were considerable prob-
lems with fuel combustion that required building an elaborate system of
pumps and cylinders, making the engine less ef¬cient and more bulky.
Therefore, Diesel™s task was to tie together Carnot™s thermodynamics,
Lord Kelvin, the German machine ¬rm and the fuel combustion prob-
lems. Only by doing this could he create a ˜black box™ “ a working object.
As Latour illustrates, the ¬nished machine had severe problems and its
use as a taken-for-granted, a fact or ˜black box™ did not take place until
some years later, years in which Diesel had a nervous breakdown and was
close to suicide. The object was only ˜black-boxed™ some years later, and
by this time it was so transformed that its relation to Diesel was scant;
it had become transformed through the operation of many other objects
and subjects.
Science and culture: interpretative networks 35

This example demonstrates that networks are not simply the result of
the operations of machines or inanimate objects; they are also made to op-
erate through the ideas, politics and economics of people. What is more,
networks are composed of stronger and weaker associations. For instance,
the association between the fuel combustion and the German machines
¬rm was weak, while the association between Diesel and Carnot™s ther-
modynamics was strong. In order for something to be made to work these
associations must articulate together, and as Latour notes, the network
is only as strong as its weakest association (1987, 121“2). Finally, we can
observe that science and technology are drawn on in human life as a series
of animate objects, whose practicality, usefulness and utility are already
predetermined through the earlier role of human beings in their creation.


Theoretical implications of an anthropology of science
What does all this mean in relation to the problems we have been dis-
cussing for the past two chapters? Well, the most important point to arise
from this elucidation of science is the notion of associations. Rather than
considering objects and subjects as distinct and immutable entities, in
order to understand how science works we need to examine the associa-
tions between them. By examining the set of associations set up between
animate subjects and inanimate objects and interpreting them through
the same analytical procedures, we are able to begin to understand how
science and technology work. The important point here is that we treat
objects as social in the same way we treat people as social. Objects are
always bound up in the social projects of people, and it is this that makes
them animate. This theoretical and methodological treatment also al-
lows us to solve a number of the problems that have been a motif of the
previous chapter, namely the distinction between objects and subjects,
rationality and relativism.
We can now observe that the distinction between objects and subjects
has become less rigid, and we are therefore in a better position to as-
sess the way in which scienti¬c practices interlock with the interpretative
practices of social science. We can side-step the problem of retaining a ra-
tionalist core of knowledge onto which cultural speci¬c understandings
are constructed. By doing this we are able to re-evaluate the positions
of rationalism and relativism. Rationalism can be seen to be a product
of the series of associations that go to make up scienti¬c practice, espe-
cially the process of transformation and the setting up of associations into
networks. Science works because our networks of scienti¬c practice are
extensive; these networks can be seen to extend at different scales. The
36 Archaeological theory and scienti¬c practice

political and economic networks of capitalism extend further and with
stronger associative links between them than the networks of associations
which characterise traditional and small-scale societies (see Latour 1993,
107“11).
In order to clarify this point we need another example, again drawn
from Latour™s work. If we consider, for example, the highly speci¬c cate-
gorisation systems which anthropologists have described as ethnobotany
or ethnozoology, and then we compare these with the categorisation sys-
tems we use within Western botany or zoology, we have a contradictory
set of categories. However, if we consider the scale of these categorisa-
tion systems then the problem becomes clearer. It is not that the native
system is wrong and that ours is right, rather that these two systems are
working at two different scales. The native system is using and com-
paring a restricted set of plants and animals as a means of categorising
the world. In contrast, our categorisation systems compare plants and
animals from around the world, since Western botanists and zoologists
regularly meet to compare, discuss and standardise their collections. Our
categories are then operating at a quite different scale, and because of this
our network of associations is more extensive. It is no surprise, then,
that our categorisation systems work throughout the world, for they are
drawn up using comparanda from varied environments across the whole
world. Furthermore, as Strathern (1992, 17) observes, ˜A world made
to Euro-American speci¬cations will already be connected up in deter-
mined ways.™ This is the illusion of rationalism “ everything in the world
works in the same way because everything in the world has been made to
work in the same way.
Our task in comparing these networks of associations as anthropolo-
gists or archaeologists studying other cultures is in comparing the rela-
tionships between associations, and in comparing the scale and strength
of these associations. Latour describes this as relationism. By consider-
ing both cultures and natures symmetrically and by examining the cre-
ation of networks of associations between the two, we will no longer be
studying either one side of the equation or the other. Instead we will
examine the nature of the connection between the two, and the mode
by which this connection is articulated in practice. The stability of the
terms nature and culture begins to erode, and we begin to understand
science as a set of practices which create associations between things and
people.
For example, Franklin (1997, 212“13) makes the point that Darwinian
notions of genealogy and kinship have now been rendered artefactual
within the biological sciences since the advent of IVF (in vitro fertilisa-
tion). This technique enables conception outside the human body and
Science and culture: interpretative networks 37

thereby detaches gestation and genealogy from the reference point of the
female body. The practices of contemporary biological science have al-
tered not only our understanding of cultural practices such as kinship,
but also such natural processes as fertilisation.


Implications for scienti¬c and interpretative
archaeology: putting science into practice
So far the anthropology of science has provided insights into some of the
practices of science. We now need to move back from the theoretical prob-
lems discussed in previous sections and consider how to operationalise
these ideas. I wish, therefore, to return to some of the problems presented
in the ¬rst section of this chapter and consider how we may reconcile
both sides of the theoretical divide. I noted that one way of reconciling
the divide between the physical and textual models of the archaeological
record might be to retain the physical model as a method of approaching
the mechanics of the archaeological record, while simultaneously using
the textual approach as a means of understanding the social practices that
are represented by the record. As indicated earlier, this approach was seen
to be problematic since it consisted essentially of constructing a cultural
shell around a realist core. Rather it was proposed we should consider an
approach that sees interpretation as a process of creative tension in which
interpretations arise from the movement back and forth between theory
and data.
One of the simplest theoretical positions available to interpretative ar-
chaeologists involves the rejection of positivist science as an explanatory
framework for human action in the past. Traditionally, the rejection of
positivism involves adopting a theoretical position that leans heavily on
some form of theory which views human action as culturally speci¬c.
However, there is a problem with doing this in archaeology since, as we
observed above, positivist or empiricist ideas form the primary frame-
work of much archaeological science. In order to carry out practical or
¬eld-based archaeology, scienti¬c procedures will be utilised. These pro-
cedures will be formulated using something like a positivist or empiricist
approach. The interpretations made under the suppositions of scienti¬c
archaeology will be effectively ˜black-boxed™. This ˜black-boxing™ has an
interesting effect, since it allows these parcels of positivist knowledge to
be safely digested by interpretative archaeology. As Wylie (1992) notes,
both new archaeologists and interpretative archaeologists subscribe to
the existence of a core of objective or rationalist beliefs.
On the other side of the theoretical divide, we observe the familiar
problem of scienti¬c archaeologists producing knowledge up to a certain
38 Archaeological theory and scienti¬c practice

level and then simply hitting an interpretative brick wall. It is at this point
that the evidence will be ˜black-boxed™. Once the evidence cannot be
described with certainty, or is not amenable to generalising laws, then
interpretation ceases and the evidence is handed over to the theoretical
archaeologist, who then constructs a series of interpretations derived from
a quite different theoretical framework. The utilisation of archaeological
science in these ways is jarring, since the conditions under which each
form of knowledge is constructed are positively antagonistic. As an alter-
native we need to employ a strategy which uses archaeological science in
quite a different way.
The critique of science explored above was crucial, since it highlighted
a number of important points concerning scienti¬c practice. Under the
aegis of the anthropology of science we can see that the interpretation of
scienti¬c data can be considered to operate in a manner akin to hermeneu-
tics. The most important points to emerge from this discussion are that
the creation of factual information and the creation of novel technologies
are not simply the result of the measurement of a prior and immutable
nature. In order to understand how facts or technologies are created it is
essential to understand the social conditions of production. The process
of understanding the links between the social and the material is sym-
metrical: that is, the same analysis was conducted on both sides of the
object“subject equation. What is more, as we move away from a con-
cern with an explicitly epistemological position to one concerned with
the process of interpretation, we need to worry far less about the prob-
lems of validating our paradigmatic positions through data manipulation.
Instead, we become more explicit about the links between our theoretical
suppositions and that data (Rorty 1980, 315“56). In the next chapter
I will examine some of the problems inherent to an objectivist approach
to archaeological practice and explore ways in which we may move to-
wards an approach that embraces an interpretative framework. In doing
this I want to examine the institutional position of archaeological science
within the wider ¬eld of archaeological practice.
3 Archaeology observed




In this chapter I will take a broad view of some of the problems associated
with archaeological interpretation by examining the relationship between
archaeological science, the archaeological specialist and the practice of
archaeology as a whole. My account will provide a situated perspective
of archaeological practice (see Harding 1991 and Haraway 1997 for a
discussion of situated knowledge). My situated knowledge is derived from
the experience of working within the ¬eld of post-excavation analysis
in Britian as both a materials specialist and an archaeological scientist.
While this knowledge is speci¬c to this context, more general points may
be extrapolated from my account which can inform our understanding
of wider archaeological practice.
Throughout, I want to examine the process by which we come to make
archaeological interpretations. In doing so I will consider a wide range
of questions: How are archaeological reports constructed? Who provides
the information that makes up the archaeological report? What are the
conditions under which this knowledge is constructed? Is there an inter-
pretative distance between those who have a primary engagement with
the site, and those who report that encounter? How is this knowledge
deployed in the construction of subsequent archaeological knowledge?
Simply put, I will consider how it is that we create accounts of past soci-
eties using the medium of material culture.
For the purpose of discussion, archaeological practice can be divided
into three broad enterprises: excavation, post-excavation and publication.
These are crude divisions, but they will suf¬ce for the present. Two of
these enterprises, excavation and publication, have come under consid-
erable critical scrutiny over recent years, while the other has remained
strangely absent from critique. During the course of this chapter I want
to arrive at an understanding of why this might be the case, while outlin-
ing some of the problems associated with the normative perceptions of
each of these enterprises. Traditionally, excavation has been perceived as
an activity that simply involves a process of objectively recording the na-
ture and extent of archaeological layers and deposits and the position and

39
40 Archaeological theory and scienti¬c practice

status of the material objects found within them. In the same way, post-
excavation is considered as a series of objective analytical procedures,
which allow us through observation and description to more-or-less ac-
curately assess both the site and its material culture. Similarly, publication
has been traditionally treated as a process that involves a presentation of
the detailed objective descriptions of the results of excavation and post-
excavation analysis. If a problem with publication was perceived, it was
simply related to where the results of excavation should be presented
(report or archive), the infrequency with which sites proceeded to publi-
cation, and the medium in which sites were published (e.g. Lavell 1981).
In this chapter I want to consider some of the consequences of this ob-
jective view of archaeological practice.


Exploding excavations
Adams and Brooke (1995) have recently noted that archaeological prac-
tice is an essentially linear process that moves in an orderly series of stages
from excavation to post-excavation to publication. Given this observation,
we should consider the practices involved in post-excavation to be cen-
tral to the construction of archaeological knowledge as a whole. If we
wish to investigate the disciplinary position of post-excavation we need
to examine the effects of the linear nature of archaeological research as
a whole. While the linearity of archaeological practice is the paramount
structuring principle affecting the production of archaeological knowl-
edge, the linear structure of archaeological practice also produces two
further structuring principles, which have an important effect on the way
in which archaeology is practiced and the way in which that practice
affects our interpretations as archaeologists. These structuring princi-
ples include fragmentation, which manifests itself in the manner in which
archaeological information is both created and presented, and hierarchy,
which manifests itself in the organisation, management and dissemination
of archaeological information. The hierarchical and fragmented nature
of archaeological practice operates in tandem, one serving to reinforce
the other.
Prior to excavation the archaeological site is a holistic entity. Layer
covers layer, deposit is strati¬ed against deposit. Material culture, ani-
mal bones, seeds and pollen lie within these layers and deposits. The
layers and deposits are contextually related to each other, as well as to
the material that lies within them. However, excavation is a destructive
and non-reproducible process. As it is generally practiced, excavation
involves the careful removal of layer after layer of both anthropogenic
and non-anthropogenic deposits (Drewett 1999, 107“18). The extent of
Archaeology observed 41

each deposit is recorded spatially through the use of site grids and tem-
porally through the use of surveying equipment such as theodolites or
levels or of standardised excavation units such as spits. The relationship
of each layer or deposit to each other is established by sections cut through
each layer or deposit in order to establish a two-dimensional picture
of temporal sequence (Drewett 1999, 115). During the excavation pro-
cess these layers and deposits undergo a transformation; they shift from
being represented physically to being represented in the abstract (Harris,
Brown and Brown 1993). Through the process of excavation, the layers
and deposits are physically disassociated. They now only have meaning
because of their representation in plans, sections or as a set of ¬gures
and measurements on paper (see Drewett 1999, 143“4). They have no
material existence beyond this. The material within these deposits or
layers, whether manufactured by past human activity or biologically or
physically produced, is again recorded spatially and temporally. Unlike
the layers and deposits, the archaeologically signi¬cant material obtained
from these features still retains some semblance of physicality after the ex-
cavation process; however, the material is disassociated from the features
and contexts in which it once lay. It only retains any relationship to these
layers and deposits through its depiction on maps, plans and sections
and through a set of three-dimensional coordinates. Occasionally an area
will be sampled for subsequent detailed examination by sieving either on-
site or within the con¬nes of the laboratory. Alternatively, sections of the
site will be sampled for subsequent laboratory analysis procedures such
as soil micromorphology or heavy residue analysis. Again the location
of the sieved or sampled zone will be recorded spatially and temporally.
In this instance, elements of the original deposit may be retained for
inspection. In the case of laboratory samples, once again this is an ab-
stracted quantity of material that has no real physical relationship with
the layers or deposits from which it was derived, and any anthropogenic
or non-anthropogenic material obtained from such a sample will be de-
contextualised.
Through the process of excavation the site has become fragmented.
Physically, the site now only exists in the form of the individually bagged
and labelled material remains removed from the site. The recorded ma-
terial created by the archaeologists who excavated the site provides in-
formation on where the material remains came from and on the spatial
and temporal layout of the excavated features. However these features,
and the relationship that once pertained between them, no longer exist.
What happens to this fragmented information once it has been obtained?
In order to understand this, we need to turn to examine the nature of the
post-excavation process.
42 Archaeological theory and scienti¬c practice

Post-excavation is characterised by a further act of decontextualisa-
tion. The site has been utterly fragmented; it no longer has spatial or
temporal integrity. The post-excavation interpretation of the site occurs
at a remove from the site itself. Due to this, the information derived from
excavation is now spread out and further fragmented as it is handed over
to archaeological specialists. The plant macrofossil remains are handed
to the palaeobotanist, the pollen samples to the palynologist, the animal
bones to the archaeozoologist, the snail shells to the mollusca special-
ist, the human bones to the human anatomist. Material culture is also
divided amongst specialists. The ¬‚int, chert or obsidian will be handed
over to the chipped stone tool specialist, the ground stone tools will be
handed to the ground stone tool specialist, metalwork will be given to the
metalwork specialist, while the ceramics will be handed to the ceramics
specialist. Often the material may be further dispersed by period, with
particular specialists with de¬ned period specialisms working on parti-
cular classes of object. Thus in Britain and Ireland ceramics may be
divided between a prehistoric and Medieval specialist, depending on the
expected classi¬cation of the material from the site.
What happens next? We ¬nd yet more processes of fragmentation at
work when individual elements of the material residue of the site are
given to the archaeological scientist for analysis. Stone tools or cera-
mics will be thin-sectioned and examined by a petrologist, ceramics may
be examined using Gas Chromatography in order to determine func-
tion, and this may be undertaken by a biochemist or organic chemist.
Metalwork may be examined using a variety of techniques such as mass
spectrometry, inductively coupled plasma emission spectrometry, x-ray
¬‚uorescence or neutron activation analysis; this will often be undertaken
by another archaeological scientist with a background in physics or in-
organic chemistry. The interpretation of the results is again distanciated
from both the object of analysis and the site it was derived from.
So there are a number of levels of fragmentation involved as we move
from excavation to post-excavation, with each layer of analysis involving a
further process of abstraction in which the site, the objects from the site,
and the scienti¬c analysis of those objects are gradually pulled further
apart. As this process continues, the connections between them become
more abstract and less well de¬ned. In effect the site has exploded “ the
constituent elements of the site and its artefacts have become disengaged
(see Fig. 3.1). This process of ˜explosion™ has serious implications when
the time comes to publish the site. As noted above, archaeological exca-
vation is often presented as an objective technical exercise. We simply ob-
serve what we excavate and then present the results of those observations
in our detailed publications “ the site is simply recreated on paper. But let
Archaeology observed 43




Excavation



Sample Y Artefact X



Macro-analysis Macro-analysis by
by environmental specialist X
specialist Y e.g. ceramics report
e.g. faunal report
Mediation by
post-excavation
analysis
Specialist X2
Specialist X1
e.g. petrology
e.g. GC/MS
Micro-analysis by
specialist Y1
e.g. bone isotope studies




Report Y Report X

Structural
Partial connections
between report and
Report
specialist reports

Report Y1
Report X1 Report X2

Excavation Report

Figure 3.1 Exploding excavations


us examine this claim in more detail. I have presented a familiar view of
the excavation and post-excavation process above. In each case features,
objects and their contexts are gradually prised apart in order to better
understand them. However in doing this we effect a process in which our
interpretations of the site become ever less grounded on the primary ob-
servation of the site itself. Rather than presenting the holistic site that we
began with, excavation reports present something quite different. The for-
mula for publication generally involves publishing the site and its features
along with a detailed ¬nds catalogue. Finds are published according to
44 Archaeological theory and scienti¬c practice

category: we have ceramic reports, chipped stone reports, ground stone
reports, metalwork reports, plant macrofossil reports, pollen reports, ani-
mal bone reports, human bone reports, etc. Onto each of these will be
tacked, as something of an afterthought, the scienti¬c analysis pertaining
to each material. Somewhere near the end of the report we generally ¬nd a
short synthesis placing the site into its regional or archaeological context.
The publication of various elements of the site, such as particular classes
of ¬nds in micro¬che or fascicule form, is a more extreme manifestation
of this process of compartmentalisation in which elements of the site are
physically discrete from the site even when in published form.
This form of publication is a representation of the site that was exca-
vated. However, this is not the site which was originally excavated; it
is a site which has been frozen on paper in a very speci¬c form. What is
represented is a frozen moment in the process of ˜explosion™. The archae-
ological features, along with the material and biological remains which
go to make up the site, are presented in decontextualised isolation. The
publication of site reports in this way embodies a mode of representation
viewed through the lens of objectivist or rationalist science. It is a repre-
sentation of the past lives of people through a body of material culture
that, I argue, would be unrecognisable to those that had inhabited the site.


Fragmentation and hierarchy in archaeological practice
Nevertheless the conceptual fragmentation of objects and the analysis of
those objects at a distance from their contextual relationship with spe-
ci¬c on-site features is part of the practice of contemporary archaeology.
What causes this process of fragmentation? How has this fragmented
knowledge been produced? In order to answer these questions, we need
to examine the organisation of the excavation itself. What we ¬nd is that
knowledge is generated in a hierarchical framework. Individual areas of
the site are excavated by groups of individual excavators, and these may
then be supervised by an area supervisor or site supervisor. This per-
son is in turn overseen by a site-director. Interestingly, we ¬nd that the
primary interpretative engagement with the site is the domain of the indi-
vidual excavator (Chadwick 1998; Challands et al. 1998). These people
may also undertake the recording of the features they have personally ex-
cavated; nevertheless the process of recording is often undertaken by
the person who is on the next rung of the ladder of hierarchy “ the
area supervisor. Despite the fact that this division of labour exists on all
archaeological excavations, it is traditionally the site-director, the person
most distanced from the primary engagement with the site itself, who
collates the interpreted information from the site for publication.
Archaeology observed 45

The hierarchical process of excavation consolidates the process of frag-
mentation. Excavation involves a process of double fragmentation: ¬rst
the site is disassembled through the process of excavation and then the
process of recording that process of fragmentation is divisive (Chadwick
1998). The most important point to note, however, is not only that this
process of fragmentation occurs, but rather that this process is crystallised
within publication. This is due to the fact that both the primary excavation
of the site by individual excavators, and the interpretation of the materials
excavated from the site by post-excavation specialists and archaeological
scientists, are divorced from the ¬nal process of interpretation involved
in publication.
The methodology I have characterised represents empirical and ob-
jective archaeological practice. Obviously not all archaeologists work in
this way and this approach to archaeological practice has been criticised
(Hodder 1989, 1999). Nevertheless, this is the procedural prototype from
which much archaeological practice is judged; canonically it is the ˜cor-
rect™ way to do things. Yet, as I have shown, there are obvious problems
with this methodology. Not only is this approach to the interpretation of
material culture and archaeological sites a barrier to the holistic and inte-
grative interpretation of the site itself, but also this form of interpretation
is a conceptual barrier to the interpretations of sites, forms of activity
and types of material within their wider archaeological context. These
points will be discussed later in this chapter. Here I simply wish to ask
the question, why is archaeological practice ordered like this?
To answer this question, we need to examine a number of problems
encompassing the disciplinary organisation of archaeology and the philo-
sophical underpinnings of that organisation. To begin with, excavation
has been traditionally treated as the prototypical, or de¬ning practice of
the discipline of archaeology (Tilley 1989b). The results of this highly
visible activity are traditionally disseminated to a wider archaeological
public through the medium of publication. Therefore the twin practices of
excavation and publication are closely bound up with the measurement of
academic prestige. Archaeologists have traditionally forged their careers
through the excavation of prestigious sites and through the publication of
those excavations. No such disciplinary honour is associated with post-
excavation, and it is for this reason that post-excavation has remained a
peripheral activity. It never has been perceived to be central to the con-
struction of archaeological knowledge; indeed, during the emergence of
archaeological science as a routine part of post-excavation analysis, the
task of analysis traditionally fell to those outside the boundaries of the
discipline “ zoologists, anatomists, botanist and chemists. Bluntly put,
it is due to the peripheral nature of scienti¬c specialists, as well as other
46 Archaeological theory and scienti¬c practice

materials specialists, that scienti¬c analysis and specialist reports are often
marginalised in contemporary forms of archaeological reportage.


What is post-excavation analysis?
At a disciplinary level we see a distinction between those who excavate
and publish sites and those who analyse the material recovered from
those sites. When post-excavation analysis is undertaken by archaeol-
ogists (ceramic, lithics and metalwork specialists for example) they are
often individuals who have not been involved in the primary excavation of
the site from which material was recovered. We have seen that if we take a
strictly linear view of archaeological practice, the task of post-excavation
analysis should be central. However, due to the fragmented nature of
archaeological practice, post-excavation analysis is in fact at the periph-
ery of the interpretative process.
Despite a critical awareness of the problems associated with excava-
tion and publication, there has been little comparable criticism of the
processes involved in post-excavation. Yet if we are to consider some of
the problems associated with integrating archaeological science with an
interpretative approach to the past, then post-excavation analysis requires
careful consideration. It is during post-excavation that the detailed anal-
ysis of artefacts occurs, and it is as a result of this process that many
interpretations become crystallised. As important as this process is, how-
ever, little disciplinary interest appears to exist in the processes by which
the results of excavation move from site to publication. Indeed we might
easily characterise post-excavation as a form of ˜black box™ (Latour 1987)
in which the material results of excavation results are simply produced
from excavated sites for post-excavation analysis at one end, while the
results of this post-excavation analysis are reproduced in publications at
the other (Adams and Brooke 1995). Post-excavation practices remain
untheorised and incoherent; post-excavation is simply a stage between
the tasks of excavation and publication.
Here I wish to consider a de¬nition of post-excavation analysis that
includes the analysis of archaeological materials, bones, seeds, pottery,
stone tools, etc., as well as materials science, which includes a range of
instrumental scienti¬c techniques. I want to consider a number of prob-
lems concerning post-excavation analysis. First, what is post-excavation
analysis and how does it operate? How does post-excavation analysis in-
tercede in the interpretation of the site as a whole? We have seen that in
general post-excavation analysis is distinct from the process of excavation.
This image of post-excavation is enduring and characteristic of almost all
forms of post-excavation analysis. The process of post-excavation is itself
Archaeology observed 47

surrounded by an aura of objectivity, and herein lie many of its problems.
The analysis of material excavated from an archaeological site is tradi-
tionally understood to be the domain of objective or empiricist scienti¬c
practice. Materials, whether biological or cultural, are handed over to in-
dividuals with a high level of expertise. These individuals are considered
to be in possession of wide-ranging knowledge concerning their object of
study. Most post-excavation analysis proceeds by a combination of ob-
servation and description. Through the process of empirical observation
an accurate description of the material being studied is made, and these
observations are then compared against other known objects within the
class of materials being studied. It is against this background that a series
of objective recording methodologies have been developed. Knowledge
concerning different classes of archaeological object is published in jour-
nals or monographs, or detailed knowledge is simply stored within the
memory of the individual post-excavation specialist. This model of post-
excavation analysis works just as well for the analysis of archaeological
materials both through macroscopic identi¬cation and through instru-
mental scienti¬c analysis. Both of these aspects of archaeological science
operate within a well-established, independent body of data concern-
ing the phenomena that they are attempting to observe and measure.
Faunal specialists and human anatomists employ collections of skeletons
as comparative material, while they also utilise other forms of informa-
tion including broad ecological and morphological data on animal species
and more detailed osteological data such as that for tooth eruption and
epiphysial fusing derived from zoological analysis (Rackham 1987, 51).
Palynologists and plant macrofossil specialists utilise collections of known
pollen or seed species. Materials scientists, on the other hand, draw on
a more abstract body of data in order to assess their observations. Such
data encompasses the known characteristics of sub-atomic particles, bio-
chemical molecules, etc. (Pollard and Heron 1996).
Often the post-excavation specialist is in possession of little or no in-
formation concerning the context or nature of the feature from which the
material they are analysing was obtained. Once an interpretation of the
object is made, based on comparison with other phenomena, this infor-
mation is then passed on to the site excavator who uses it to con¬rm or
refute their initial observations. This image of post-excavation conforms
very well with the inductive logic of objectivist science. Post-excavation is
then treated as a form of independent process against which the material
and the site itself are tested. It is according to this philosophical model that
artefacts are simply used as a means of dating sites, in precisely the same
way that the techniques of radiocarbon dating and dendrochronology are
used as chronological indicators (Blinkhorn and Cumberpatch 1997).
Artefacts in context on site Artefacts decontextualised Artefacts reordered as
from site ˜data™ in laboratory



Transformation in
Excavation
conceptualisation of
artefact

Figure 3.2 The transformation of artefacts as data from excavation to laboratory
Archaeology observed 49

There are obvious problems with this approach to post-excavation. The
most immediate of these is that this form of analysis presupposes a large
body of site-independent information against which to compare the ana-
lysis of materials or phenomena. With this approach, we are unable to
deal with materials or phenomena that have been barely studied or are
entirely unique. Even if this information exists, this methodology leads
to stulti¬cation in the nature and level of knowledge that we are able to
develop concerning certain classes of material.


Transforming the laboratory
One of the critical points arising from the twin processes of fragmentation
and hierarchy in archaeological practice is that the constituent elements
of the site are prised apart through excavation in order to uncover and
analyse them. Moreover, as I noted, a necessary part of the process of
analysis involves the transformation of the physical components of the
excavated site into representational data. Here I want to further con-
sider the process of transformation that occurs during excavation and its
problematic relationship to the notion of the archaeological laboratory.
The main point to note, when we wish to consider archaeological
practice as a science, is that scienti¬c veracity is traditionally dependent
on the existence of a reproducible set of results from a given body of
data. The ability to conduct experiments, and to witness and thereby
judge the results of repeated experimentation, lies at the heart of science
as a cultural endeavour. As a cultural practice science is characterised by
the testability of its knowledge. However archaeology does not conform
to this model of scienti¬c practice. There is no such entity as a site-based
archaeological experimental laboratory in which conditions can be con-
trolled in order to reproduce testable knowledge. Where we do observe
controlled experimentation, this occurs in the present and the results of
experiments are related back to past processes through analogy.
Archaeological sites are individually distinct. After excavation the site
no longer exists in physical form. The archaeological site does not then
conform to the de¬nition of a laboratory since through the practice of
excavation the laboratory has been destroyed. Due to this the results of
the experiment cannot be tested. But what does remain from the site? The
site has become transformed and exists only as a series of representations
as plans and section drawings. Despite this, some physical remains are
preserved in the form of artefacts and environmental samples.
With the transformation of the site, the archaeological site-as-labora-
tory has effectively disappeared. However in order to retain an attach-
ment to scienti¬c objectivity a further interesting transformation takes
place (see Fig. 3.2). Although the site-as-laboratory has disappeared
50 Archaeological theory and scienti¬c practice

through the practice of excavation, we witness the creation of a new kind
of archaeological laboratory out of the transformed (decontextualised)
physical remains derived from the now extinct site. These remains are
taken into archaeological laboratories and they are subjected to analyti-
cal techniques that allow for reproducibility and testability. Although the
site-based archaeological laboratory has disappeared, out of its physical
remains (artefacts and environmental samples) the laboratory has been
reconstituted.
But the problem still remains that the archaeological site is both unique
and non-existent. Therefore we cannot consider it to be a component of
objective discourse since it does not conform to the basic tenets of objec-
tivity “ it is non-testable and can no longer be empirically observed. The
observations made on site have become transformed, and those artefacts
and environmental samples that do remain have themselves undergone a
transformation since they are only related to the site through the anno-
tation of site plans and sections of three-dimensional coordinates. Fur-
thermore these too are unique; they are recovered from speci¬c locations
around the site and are often derived from distinct contexts within the
site. There is little room then for statistical comparison of comparable
contexts; rather most sites are characterised by zones of contextual dif-
ference. It would appear that the practice of archaeological science is a
transformative process in which artefacts and environmental samples are
changed from decontextualised objects and samples into meaningful ob-
jects and samples by virtue of their status as objects of analysis in the labo-
ratory. It is due to this process of transformation that we ¬nd artefacts,
faunal remains and botanic samples discussed in terms of the induction-
driven methodologies described earlier. It is precisely because artefacts
and environmental samples are divested of their context that scienti¬c
analysis is required to situate them within a wider framework of objec-
tive knowledge. I will examine ways in which we might re-contextualise
scienti¬c knowledge in more detail below.


How do we de¬ne objective analysis in archaeology?
As archaeologists we state that we are producing an objective account
of the past in our excavation reports. But how do we de¬ne objectivity,
and whose de¬nition of objectivity are we employing? One of the seri-
ous problems inherent in the concept of objective data-recording is that,
once we believe that we have an adequate de¬nition of objectivity, we
do not feel the need to re¬‚ect on it; instead we crystallise it, or ˜set it
in stone™. For instance, while we have retained a notion of standardised
or objective recording, many of the principles guiding what we deem
Archaeology observed 51

worthy of record and publication have themselves been retained from an
earlier vision of archaeological objectivity. One obvious example of this
concerns the nature of reportage required in British and Irish excavation
reports. In excavation reports in these countries it is not deemed to be
a necessary component of an objective archaeological report to conduct
a detailed spatial analysis of the site and its artefacts, while it is deemed
to be a requirement of objective analysis to list parallels for site types
and artefact types found on other, more distant, sites at great length and
in some detail. In other words, despite the latent empiricism of much
archaeological enquiry, it is not deemed to make good sense to compare
and contrast empirical observations made between objects at a directly
observable intra-site level, although it is deemed to make good empi-
rical sense to compare and contrast both sites and objects at the inter-site
level. This is in part due to our attempts to create an objective scienti¬c
discourse that is both generalising and testable as opposed to particularist
and contextual. While this understanding of objectivity underpins much
archaeological practice, we also observe that the notion of objectivity we
retain in archaeological discourse is historically contingent. It is a notion
of objectivity that has been retained from culture-history, in which it was
a requirement, under the aegis of a normative vision of culture, that indi-
viduals should catalogue and list parallels for sites and objects as a means
of de¬ning and dating cultural units. Despite the widespread criticism
of this view of culture both archaeologically and anthropologically, the
requirements of culture-historical analysis are fossilised within the objec-
tive discourse of adequate site reporting. Here we observe an earlier set
of disciplinary practices simply ˜standing in™ for objectivity. This is a form
of statutory objectivity; it does not characterise objectivity in its strictest
philosophical sense.
It is the issue of statutory objectivity that lies at the heart of our prob-
lematic use of the results both of specialist analysis and of scienti¬c analy-
sis within excavation reports. This is because the use of specialist reports
and scienti¬c reports often becomes disengaged from the interpretation
of the site itself. Reports on archaeological materials and their scienti¬c
analysis are often undertaken as a ˜matter of course™, with little re¬‚ection
on how this information will integrate with the broader interpretation of
the site. An example of this non-re¬‚ective use of science may be drawn
from some of the published excavation reports for the Orkney Neolithic.
Many of the excavation reports from Orkney include a mandatory petro-
logy report for stone tools and ceramics (Williams 1976, 1979, 1983).
These reports are insightful in themselves, however the results of these re-
ports, if integrated into the body of the excavation report, generally state
that the pottery or stone tools are simply ˜locally derived™ (Ritchie 1983).
52 Archaeological theory and scienti¬c practice

While this may be considered to be a legitimate result, when these results
are integrated with a wider series of interpretative questions petrological
analyses may be of considerable interpretative value themselves, as we
shall see in chapters 6 and 7. In the instance of these reports it would
appear that, for the archaeologist interpreting these results, petrology was
treated as a simple technical procedure: a procedure which ˜ought to be
done™, but whose value had been overlooked or simply forgotten at the
wider interpretative level.
A similar example arises when we come to look at the nature of ani-
mal bone reports incorporated within many excavation reports. Faunal
reports are an interesting form of reportage since they often appear to
embody a series of requirements that have little relationship to archaeo-
logy in the strictest sense (Legge 1978); that is, if we de¬ne archaeology
to be an investigation of human activity in the past through the medium of
material culture. Many animal bone reports include minute comparative
details describing the morphological differences between bone material
using metrical analyses. These details are often used to provide infor-
mation on the level of domestication, the habitat of animal species, etc.
While these reports provide some information on archaeological details
such as economy, they also appear to indicate a preoccupation with in-
terests of a more broadly zoological nature (Legge 1978). All too often,
few faunal reports do not pay suf¬cient attention either to the context
from which material was derived or the intra-site spatial organisation of
animal deposits (see Legge, Payne and Rowley-Conwy 1998 for a dis-
cussion of the necessity of intra-site contextual differences in the study
of past diet). Here it is important to reiterate that we cannot simply read
off economy from site assemblages, as if sites themselves had economies
(Barrett 1989). Sites are not uniform entities, which means generalising
statements that rely on conditions of uniformity are problematic in the
reconstruction of economic regimes. Instead we need to embrace the het-
erogeneity of contexts in an attempt to understand both the differences
and similarities between different sites in terms of consumption practices
(see chapter 5 for further discussion; see Meadows 1997 for an example
of this approach).
Each of the examples discussed above calls into question the nature of
the de¬nition of objectivity embodied in many specialist reports. My point
here is that criticism does not solely rest with the specialist or archaeo-
logical scientist, it resides as much with the ¬eld archaeologist. We have
to ask ourselves whom these reports are for, what questions they are re-
quired to answer, and how the results of such reports are deployed within
the site report as a whole. We must remember that there is no such thing
as an objective scienti¬c report, rather many reports purporting to be
Archaeology observed 53

objective are simply reproducing a previous set of disciplinary concerns.
Due to the contingent nature of these de¬nitions of objectivity there can
be no ˜correct™ and objective manner in which to publish our interpre-
tations (Barrett and Bradley 1991), except when publication becomes a
prescriptive exercise enshrined in rigid policy (Adams and Brooke 1995).
Science in its broadest sense may be employed as a means of under-
standing a very broad series of questions. There is not a set of statutory
questions that we should feel impelled to ask of archaeological science or
the archaeological specialist; rather the questions that we ask of science
constrain the nature of the answer that we receive (Bradley 1998). We
need to think very critically, then, about a number of issues. What ques-
tions do we wish to ask of the archaeological scientist or archaeological
specialist? Can the particular techniques used provide answers to these
questions? In short, our expectations are intimately bound up with our
observations. Moreover, these questions will alter depending on the his-
torical period we are studying (see Bayley 1998). We need to be aware of
this and consider what kinds of archaeological reports we actually want,
rather than simply persisting with reports that reproduce an outmoded
vision of archaeological objectivity.


The philosophical nature of archaeological practice
There are a series of problems, then, in the way in which archaeological
knowledge is constructed. If we are to comprehend the construction of
archaeological knowledge in more detail, we need to examine why it is
that archaeological practice is conducted in this way. I will begin by look-
ing at the distinction between excavation and post-excavation. While the
peripheral nature of post-excavation in relation to excavation and publi-
cation may be explained by a recurrent distinction between ¬eld archae-
ologist and archaeological specialist, we need to examine the ideas that
underpin and structure these distinctions. If we return to the arguments
outlined in the opening chapters, we observe a practical application of the
distinction between the social/human sciences and the natural sciences.
At the most basic level this involves a disjunction between those who do
¬eld archaeology “ the province of cultural analysis “ and those associated
with the natural sciences “ the province of scienti¬c analysis, effectively
archaeological scientists.
While a fundamental distinction remains between those who practice
archaeology in the ¬eld and those who practice archaeology within the
laboratory, when we come to examine the division of labour within the
˜archaeological laboratory™ at the post-excavation level we see further ¬ne-
grained distinctions (Sofaer-Derevenski forthcoming) between those who
54 Archaeological theory and scienti¬c practice

examine the cultural aspects of archaeological data (pottery, lithics, met-
alwork, glass, etc.) and those who study the biological aspects of archaeo-
logical data (seeds, pollen, human and animal bones, etc.). Finally, we
also see a division between those who study the cultural aspect of objects “
¬nds specialists “ and those who study the physical aspects of objects using
procedures derived from materials science, engineering, chemistry and
physics. Again we observe the fundamental distinction that pertains be-
tween the two domains of knowledge “ the social sciences and the natural
sciences.
At another level we see yet further distinctions between the individuals
who study different forms of archaeological material “ ¬nds specialists
tend to be categorised by the material they study. What we are observ-
ing here is a categorisation scheme that is based on the material nature
of the object of study. This form of categorisation is itself based on an
objective understanding of the world in which artefacts may be divided
according to their essential natures. According to this view objects are
classi¬ed in bounded sets according to the possession of a series of mate-
rial attributes (Lakoff 1987). Each of these levels of differentiation goes
some way towards understanding why we observe a fragmentation of the
archaeological site through archaeological analysis.
We might consider the possibility that our fragmentary mode of analy-
sis is due to our belief in a set of overarching disciplinary divisions. This
is interesting, but we also need to consider the whole process of frag-
mentation in and of itself. Fragmentation is a characteristic aspect of
objective analysis in which one breaks down, or compartmentalises, the
world in order to de¬ne it. Once this is done, as rationalists we build up
generalising laws concerning the way in which each isolated unit ope-
rates in relation to its neighbours. This kind of analysis comprises the
central tenets of systems analysis. A characteristic of systems analysis
is that each component of the system and each constituent element is
treated as a discrete unit in order to aid comprehension. This then al-
lows the analyst to understand the articulation of these elements together
as part of a working model. Indeed the matrix analysis of Harris et al.
(1993) may be considered to be the archetypal form of systems anal-
ysis applied to archaeological stratigraphy, in which each stratigraphic
unit is broken down and understood in relation to its nearest neighbours
on the site. However, although the aim of the systems approach is to
create a working model, in the realm of archaeological practice this fun-
damentally objectivist procedure rarely leads to detailed model-building.
Instead we might argue that many archaeologists proceed as if, after hav-
ing laid the components of the site open to the empirical gaze, this was
suf¬cient.
Archaeology observed 55

There are serious problems raised by our decontextualised mode of
representation within excavation reports. The most pressing of these is
that the objects and remains that we excavate are meaningful; they are
contextually related to the features, layers or deposits in which they were
originally found. More importantly, when these objects actively inter-
vened in the lives of past peoples they were bound up in networks of
meaningful connections. If we are to recover and interpret these aspects
of the past, we must realise that we need to begin to create interpreta-
tive connections between different kinds of material, rather than simply
studying material in isolation and then presenting these isolated studies
in publications as objective reality.


Problems in the reproduction of objective knowledge
An objective approach to archaeological practice produces a report that
fragments archaeological knowledge. However the problems that this
fragmentation of knowledge brings with it does not stop with the ex-
cavation report. This form of knowledge production can be shown to
be unproductive to the reworking of subsequent archaeological knowl-
edge. At the level of primary interpretation, the disjunction of materials
from their original context prevents their contextual re-association. It is
often dif¬cult to locate artefacts spatially and stratigraphically within ex-
cavation reports, which prevents the interpretation of the functional and
social analysis of different areas of the site (Allison 1997). Hill (1995), for
example, draws our attention to the dif¬culties involved in reintegrating
¬nds from different contexts in his analysis of structured deposition in the
British Iron Age. A greater concern is that the decontextualised nature
of our reporting also structures the way in which we set about writing
wider accounts of the past based on site report data. This form of report-
ing in fact creates a number of problems, since it forces us to retain a
vision of artefacts and sites in the abstract. While it allows us to dislocate
material from its context and compare that material across widespread
geographical regions (a major component of culture-historical analysis),
it simultaneously structures the nature of wider archaeological discourse.
It is due to this fragmented view that we have societies and symposia de-
voted to distinct specialisms: metalwork analysis; ceramic analysis; lithic
studies; environmental analysis; and the chemical analysis of archaeolog-
ical materials. It is also the reason we ¬nd volumes devoted to prehistoric
ceramics, medieval ceramics, prehistoric metalwork, stone axe studies,
¬‚int and chert analysis, etc. Clearly, the way in which the discipline is
structured at the specialist level is divisive (Bradley 1987) and we see
little cross-referencing between areas of specialisation.
56 Archaeological theory and scienti¬c practice

The point here is that many of the problems we perceive as archaeolo-
gists are created by the manner in which we structure our discipline. What
is more, this structure pervades the discipline at the most fundamental
level (in the organisation of archaeological practice according to the trip-
tych of excavation, post-excavation and publication) and at the specialist
level (symposia, societies and publications devoted to particular kinds of
archaeological data). As Evans (1998) notes, we practice archaeology in
relation to other practitioners: ˜There are not just lineages of discovery
but matrices of interpretation™ (ibid., 198). The highly speci¬c manner in
which we structure archaeological discourse affects the way in which we
write archaeology at the wider synthetic level; it allows some problems
to be examined while systematically closing off other areas of research.
Most syntheses tend to present the evidence from a series of different
and discrete strands of material. Rarely do we see synthesis examining
the systematic connections between different classes of material.


The creation of archaeological truth
The concept of ˜objective™ analysis throughout the various areas of ar-
chaeological practice has brought with it a series of problems. Rather
than creating an objective standpoint by providing a unifying interpreta-
tive perspective, as initially proposed in Britain by Clarke (1973), it has
enabled archaeological practice to fragment along a series of fracture lines
which produce a dislocated vision of the past. There are further aspects
of objective analysis that are worth considering here. As noted in chapter
1, objectivity presupposes the existence of the real material world beyond
the gaze of the subjective observer. Objective analysis requires that there
is a real world ˜out there™ to be examined. One of the peculiarities of
objective analysis is that any contradictions in the analysis are considered
the result of problematic relationships in the real world, not in the nature
of the expectations and presuppositions of the observer. The presentation
of the results of objective analysis does not tend to include the analysis of
contradictions or problems; instead, peripheral contradictory conclusions
are jettisoned, while a single uni¬ed conclusion or explanation is retained
(Clifford 1994). Without this uni¬ed conclusion, any objective analysis
could not truly be counted as objective, since objectivity presupposes the
notion of certainty. Due to this notion, objective archaeological reports
tend to present a single interpretation of the site. Due to the hierarchical
production of archaeological knowledge this interpretation is usually that
of the site-director/site-publisher. Since the aim is to create an objective
record of the site, any contradictions inherent in the individual analy-
sis of various specialists, archaeological scientists, etc., will be concealed
Archaeology observed 57

or removed entirely from the text of the report (Hodder 1989). It is
precisely because of this problem of ahistorical objectivity that Barrett
(1987b) suggests that we need to carefully consider the suppositions and
expectations of ¬eldworkers before re-analysing excavation reports. In
precisely the same way, while Evans™ analysis of Bersu™s Manx excavation
campaign offers an insight into the ethnological presuppositions behind
his interpretative excavation strategy, it is unable to supply us with a list
of the artefacts discarded by Bersu as worthless due to this interpretative
strategy (Evans 1998, 193“8). We cannot simply consider site reports as
statements of fact. It is the process of creating truth through publica-
tion that lies at the heart of the fragmentary and hierarchical procedures
of archaeological practice. I will interrogate this view of archaeological
practice by focusing on a series of issues discussed in chapter 2, includ-
ing the creation of ˜facts™, the problem of translation and the creation of
interpretative networks.
Facts are created when information derived from one particular form of
analysis is ˜black-boxed™ and then utilised in further analysis. It is precisely
this process that occurs throughout archaeological practice at a number
of levels: at the level of the individual on-site excavator who digs the site,
at the level of the site supervisor, at the level of the individual specialist,
and at the level of the archaeological scientist performing an analysis on
the archaeological object. Each of these individuals produces a series of
discrete and objective bodies of information or ˜facts™ concerning the site
and its artefacts. These ˜black-boxed™ facts are then employed by the site-
director to create the site report. The site report itself then becomes a
fact and is utilised in subsequent archaeological analyses and syntheses.
It is this process of bringing together the individual ˜black-boxed™ facts
that involves a form of transformation or translation of knowledge.
In chapter 2, I discussed the way in which, during scienti¬c analysis,
this translation process was undertaken by a series of analytical instru-
ments. This also occurs in archaeological analysis in which archaeo-
logical scientists employ a variety of instrumental techniques as a means
of analysing archaeological materials. It is not only scienti¬c instruments
that are involved in this process of translation. More subtle “ or less ob-
vious “ processes of translation may occur using simple graphic media.
Many graphic systems of representation are used in the presentation of
archaeological facts: the site plan, section drawings, plans and elevations
of buildings, drawings of artefacts, graphs, text, etc. These are devices
that allow us, as archaeologists, to translate the physical world of the
archaeological site into a graphical and textual form. Each ˜black-boxed™
or translated component of the site is drawn together in these represen-
tations as components of an objective record of the site and its artefacts.
58 Archaeological theory and scienti¬c practice

This form of ˜black-boxing™ is a necessary requirement of the form of
systemic approach that I suggest characterises the organisation of archae-
ological practice. Although the goal of systems analysis is to create a co-
herent model of the given ¬eld of study, the practice of ˜black-boxing™ each
unit of analysis appears to have had the opposite effect. This is because
the knowledge associated with each component of the archaeological site
and its artefacts is produced in isolation and then ˜black-boxed™ and used
elsewhere. It then becomes extremely dif¬cult to present a coherent pic-
ture of the site as a whole “ each piece of knowledge is produced under
a different regime of expectations and judgements. Without critically ex-
amining these expectations it is very dif¬cult to bring each fragment of
knowledge together. While the fragmentary process is embedded in the
structure of archaeological practice, it also becomes embedded in the
presentation of the ¬nal publication.


Writing material culture: archaeological practice
and subjectivity
It is obvious from the critique I have undertaken throughout this chapter
that we need to think differently about our methods of archaeological
practice. Rather than considering either excavation or publication as
objective and technical exercises, a number of authors have suggested that
instead we bring a whole series of disciplinary expectations and prejudices
to the exercise of excavating a site (Tilley 1989b; Richards 1995; Hodder
1999). What is more, if we treat publication as the presentation of the re-
sults of excavation, then we need to be clear under what conditions these
results are constructed. We cannot treat excavation reports as if they were
neutral and ahistorical evaluations of ˜what was found™ on excavated sites.
As we have seen, excavation reports, like other forms of archaeological en-
deavour, are historically contingent; they crystallise the expectations and
concerns of the paradigm in which they are produced (Barrett 1987b).
Within their pages excavation reports should embed some of the expec-
tations and uncertainties involved in excavation. In short, they should
involve an adequate presentation of, and re¬‚ection upon, the conditions
upon which the knowledge contained within the report is based. Yet while
much recent critique has focused on the twin practices of excavation and
publication, we must also be signally aware of the interpretative nature
of post-excavation practices, especially the relationship between science,
excavation, post-excavation and the process of interpretation.
The issues outlined above concerning the nature of excavation and
publication encompass a wider critique of the way in which we construct
and represent knowledge about other cultures. Much of this has focused
Archaeology observed 59

on the anthropological problems of providing an account of other cul-
tures (Clifford 1986). This discussion has focused upon the position of
the observer in relation to the observed, and the process by which ¬eld
observations are transformed into ethnographies; a process which carries
with it claims of objectivity and authority (Clifford 1994). The problems
associated with reporting the results of excavation are, in many ways,
parallel to this, but there are further problems in our engagement with
past cultures. We are not simply interpreting one set of cultural idioms
through encounter and observation and then translating these experi-
ences through the medium of language and written text (Asad 1986).
Rather we are interpreting the nature of past cultural activity, through
an encounter with physical objects, and it is this understanding which is
then translated into a textual form. We are then transforming the nature
of our encounter with past cultures from a set of experiences which re-
lies on the physical nature of the archaeological object, into a descriptive
account of these experiences through the medium of texts and graphic
representations. We cannot simply treat our encounter with past material
cultures as a process of direct translation (Barrett 1994; Hodder 1986)
because by necessity we bring our own prejudices and expectations to bear
when interpreting material culture. It is due to the expectations bound
up in our interpretations that we need to critically assess the nature of
the interpretative accounts of our encounters with the residues of past
cultures.
If we are to examine the way in which we create archaeological infor-
mation, we need to consider in more depth what it is we are doing when
we encounter the residues of past human activity. Why do we think we
are competent to write about past material culture? In considering this
question I want to develop the ideas of both Barrett (1995) and Richards
(1995) with regard to the problem of interpreting excavation data. Both
authors have considered the nature of the experience involved in excava-
tion. Colin Richards has characterised it as an experience which embodies

<<

. 2
( 7)



>>