World Archaeology

ISSN: 0043-8243 (Print) 1470-1375 (Online) Journal homepage: https://www.tandfonline.com/loi/rwar20

Introduction to experimental archaeology

Alan K. Outram

To cite this article: Alan K. Outram (2008) Introduction to experimental archaeology, World
Archaeology, 40:1, 1-6, DOI: 10.1080/00438240801889456

To link to this article: https://doi.org/10.1080/00438240801889456

Published online: 01 May 2008.

Submit your article to this journal

Article views: 4060

View related articles

Citing articles: 20 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=rwar20
Introduction to experimental
archaeology

Alan K. Outram

What is experimental archaeology?

If one takes a scientific and ‘positivist’ view (Popper 1959), then experimentation is part of
a ‘hypothetico-deductive’ process. A hypothesis is formulated and then tested to see if it
can be ‘falsified’. If falsified then that hypothesis must be discarded and replaced with a
new, hopefully better one, which will, itself, then be tested. If a hypothesis resists
falsification, and is supported by experimentation, it can be regarded as valid. ‘Valid’, in
this sense, does not mean ‘true’, but merely that the principles behind the hypothesis can
continue to be used until falsified and replaced by a better set of principles. An
experimental, positivist approach can escape the shackles of simple historicism and
empiricism, because it allows one to move beyond the limited range of options made
available by records of the currently known world. It allows investigation of the counter-
intuitive and for the possibility of deductive leaps, rather than simply relying upon
probabilistic and inductive extrapolations of existing knowledge.
Positivism is still the underlying philosophy of modern science. While Kuhn (1962) very
clearly outlined his view of how science really works in the fallible and often prejudiced
world of human scientists, his critique was not so much a direct challenge to Popper’s
ideals, but more of a reality check. Experimentation remains a method that clearly sits
within the realms of science. The postmodernist attack on science and method (e.g.
Feyerabend 1975) presented instead a philosophy of ‘anything goes’, and gave no special
place to testing hypotheses though experimentation. This is not the place to debate, in
depth, the nature of the postmodern or post-processual challenge to science, but the reader
should note that this volume presents experimental archaeology as a scientific research
method. As such, while it is accepted that other theoretical viewpoints will interpret the
experiences of experimenters differently, this volume does not take an ‘anything goes’
approach to the topic, but it does investigate a range of styles and approaches to
experimental archaeology as science.
Having put forward a definition of experimentation, however, it is still not entirely clear
what ‘experimental archaeology’ exactly means. If experiment is the mainstay of modern
science, then, strictly speaking, is there really any difference between ‘experimental
archaeology’ and ‘archaeological science’? Readers of key works dating to when the term
‘experimental archaeology’ was first coming into common parlance (e.g. Coles 1973, 1979;

World Archaeology Vol. 40(1): 1–6 Experimental Archaeology

ª 2008 Taylor & Francis ISSN 0043-8243 print/1470-1375 online
DOI: 10.1080/00438240801889456
2 Alan K. Outram

Reynolds 1979) will clearly note that it relates only to certain types of activities. Coles
states that the aim of experimental archaeology is to ‘reproduce former conditions and
circumstances’ (1979: 1), and the same is echoed by Mathieu (2002: 1), who says it is
designed to ‘replicate past phenomena’. Indeed, most people, when thinking of
experimental archaeology, are conjuring with words like ‘reconstruction’, ‘re-enactment’,
‘reproduction’ and ‘replication’, because the activities of those who engage in the subject
usually seem geared around, in some way, re-creating activities, artefacts, structures and
processes that happened in the past. This concept provides a clear difference between
experimental archaeology and other forms of archaeological science, but it is perhaps a
defective definition. Reynolds (1999: 159) was very clear about his dislike of the ‘re-’ prefix.
After all, in the vast majority of cases, one does not actually know what the past was like,
so one cannot reconstruct it. Some aspects of an experiment must be hypothetical and
being tested, and, hence, not a reconstruction, otherwise there is little point in doing it.
This is the reason why Reynolds (1999: 159) referred to his hypothetical Iron Age houses
as being ‘constructs’. The European journal devoted to experimental archaeology nods its
head at this distinction in its title, while still acknowledging common parlance: EuroRAE:
(Re)construction and Experiment in Archaeology.
Perhaps a more useful term to use is ‘actualistic’. To present an example, one can
experimentally study the rendering of tar from birch bark, a material we know was used in
the past (Piotrowski 1999). One could test many factors about source materials, rendering
conditions and end products within the controlled conditions of a laboratory, using sterile
glassware and a gas powered heat source. Each experiment could involve holding all but
one variable constant at any one time. The rendering process could be quite well
understood, in physical terms, from such experiments and some conclusions might be
reached that were relevant to archaeological interpretation. However, a gulf is left between
such laboratory work and how such processes may have been achieved in the past, with a
limited range of materials, technologies and a lesser control upon the environment.
Experimental archaeology comes into its own at this point. What has been learned in the
lab can now be taken further; hypotheses can be tested with authentic materials and in a
range of environmental conditions that aim to reflect more accurately ‘real life’ or
‘actualistic’ scenarios. Such experiments investigate activities that might have happened in
the past using the methods and materials that would actually have been available. This is
not to say that all materials and methods need to be authentic in experimental
archaeology, but certainly those pertinent to the hypothesis.
The above two approaches to scientific investigation of past activity should not be seen
as in any way rivals. It is not an either/or situation. These two approaches naturally follow
on from, and complement, each other. Laboratory experiments provide a sound
understanding of scientific principles through the careful control of variables, while
‘actualistic’ experiments test out hypothetical scenarios using potentially authentic
materials and conditions. In the latter case, unpredictable phenomena are often given
more opportunity to act, thus enabling the refinement of hypotheses and archaeological
interpretation. Experimental archaeology should certainly be viewed as being in no way
separate from the rest of archaeological science. It simply represents a particular type of
experimentation that, in some cases, requires the application of skills or materials not
commonly available in our modern world. Actualistic experiments should be no less
 Introduction to experimental archaeology 3

rigorous than laboratory ones, but that rigour may be based upon a different set of
criteria. Maintaining rigour in actualistic scenarios is no mean feat.
Reynolds (1999: 158–62) defined five major classes of experiments that are the mainstay
of experimental archaeology. It is worth briefly summarizing them here:

1. Construct: 1:1 scale constructions that test a hypothetical design for a structure (e.g.
house) based upon archaeological evidence. It is a hypothesis that literally stands or
falls.
2. Processes and function experiments: investigations into how things were achieved in
the past. This includes investigations into what tools were for, how they were used
and how other technological processes (e.g. tar rendering or pit storage) were
achieved.
3. Simulation: experimental investigations into formation processes of the archae-
ological record and post-depositional taphonomy.
4. Eventuality trial: usually combining all three categories above, these are large-scale,
often longue dure´e, experiments that can investigate complex systems (such as
agriculture) and chart variations caused by unexpected or rare eventualities (e.g.
extreme weather).
5. Technological innovation: where archaeological techniques themselves are trialled in
realistic scenarios. A good example would be the testing of geophysical equipment
over a simulated, buried archaeological site.

What experimental archaeology is not

There is little doubt that the term ‘experimental archaeology’ is closely associated in many
people’s minds with re-enactment groups, outdoor education and public presentation
centres, and other demonstrations of past life and technology. The journal EuroREA
certainly covers all these aspects. Reynolds made his view, that these activities are not part
of experimental archaeology, bluntly and acerbically clear. The activities of re-enactors
dressed in period costume were, to him, ‘at best theatre, at worst the satisfaction of
character deficiencies’ (1999: 156). Indeed, such activities are clearly not, in philosophical
and research terms, experiments. It is perhaps unfortunate that the boundaries between
experimental archaeology (a research tool), experiences and demonstrations (educational
and presentational tools) and re-enactment activities (a recreational pursuit) have become
blurred in the minds of many. In some cases, one fears that this has coloured academic
perception of a valuable approach to research. Perhaps this is why Reynolds put forward
such a strong rejection of anything not truly experimental.
Of course, real-life demonstrations, three-dimensional (re)constructions and first-hand
experiences all have huge pedagogical benefits and are an excellent way to translate
archaeological research into a presentable form for the public. There is also no need to
deny re-enactors their fun. Furthermore, most true experimenters need to gain a degree of
competence through experience before conducting their experiments (e.g. those who
conduct flint-knapping experiments). However, from an academic point of view, it is
clearly beneficial to maintain a clear distinction between what is ‘experimental’ and what is
4 Alan K. Outram

‘experiential’. Experiential activities can be very valuable and can be easily associated with
an experiment to add a public or educational (translational) element, but that potentially
positive by-product should not be allowed to create confusion over experimental aims.
This volume deals only with experiments for the purpose of research, but this editor has no
wish to do down other valuable activities.

Publishing experimental archaeology: pitfalls and how to avoid them

Through editing this volume, reviewing for many others and teaching experimental
archaeology to graduate students, this editor has noted five major pitfalls that potential
authors of experimental publications regularly succumb to. Below, these are outlined
along with a discussion about how best to avoid such difficulties.

1. Lack of clear aims One of the most frequently encountered problems with
experimental articles is the lack of a well-thought-out hypothesis or a specific
archaeological question that is being addressed. It is likely that the authors are, in
fact, simply writing up a practical, experiential activity, after the fact. The content
may be interesting to like-minded specialists, but, in many cases, the decisions
relating to materials and recording methods would be very arbitrary because of the
lack of a clear aim. This may limit the usefulness of the activity somewhat. Such
experiential reports are not without value, however, and are frequently found
published in journals like EuroREA (in Europe) and the Bulletin of Primitive
Technology (in America). Such works are less likely to be accepted in more
academic publications.
2. Insufficient detail on materials and methods If a paper does not outline enough
information about materials and methods then it seriously limits its usefulness, as
other researchers will be unclear what exactly has been tested, and they will not be
able to replicate those experiments or build upon them. EuroREA, recognizing both
of the above common flaws, through feedback from its editorial board, has
attempted to encourage its contributors to move towards a more ‘scientific’ mode of
publication by printing a number of papers on producing good experimental
reports (Mathieu 2005; Outram 2005; Schmidt 2005).
3. Compromises over authentic materials Most experimenters will face difficult
decisions over when to compromise by using more readily available modern
materials (and methods) instead of using rare or totally unavailable authentic ones.
The key question is whether the compromise will materially affect the testing of the
hypothesis. To illustrate, it is unlikely to matter that one pours iron ore into a
smelting experiment from a plastic bucket, but the use of a modern high-grade ore,
not available in the past, could fundamentally defeat the purpose of the experiment.
An experiment that contains too many compromises risks losing any real value. A
compromised actualistic experiment neither has the control of the laboratory nor
does it test the authentic scenario it was designed to.
4. Inappropriate parameters This editor has seen a number of experiments that were
carried out apparently very well with authentic materials, but the basic parameters
 Introduction to experimental archaeology 5

were very badly set to answer the question being asked. Examples of this might be
that the timing of recording points was spaced too far apart or that the starting
temperature was too high. This is suggestive of a lack of experience that could be
addressed through collaboration with experienced craftspeople or technologists, or
through carrying out pilot studies first.
5. Lack of academic context Very able and experienced practitioners of ancient crafts
have much to offer experimental archaeology, but a problem that affects some
papers submitted by such individuals is a lack of academic context and appropriate
reference to the literature. This significantly weakens the work.

The last two points are perhaps best addressed through good collaborations between
craftspeople and academics.
Perhaps the most effective experiments are those that are totally integrated into a larger
scheme of academic research with the experimentation being just one of the methods being
employed in pursuit of a research goal. Where possible, there should be close collaboration
between different specialists and those with academic and practical skills. If all these
elements are present, then it seems far less likely that any of the above five mistakes will be
made.

In this volume

The papers in this volume address experiments relating to wide range of different
artefactual or ecofactual materials. These include phytoliths (Mithen et al.), bone (Seetah;
Domı́nguez-Rodrigo), stone (Aubrey et al.), pottery (Jeffra), metal (Molloy), organic
materials (Hurcombe) and residues (Evershed). Most of the papers deal with, on some
level, ‘process and function’ experiments. Two which fit very clearly in that category are
Jeffra’s study into the possible use of hair as an organic temper in pottery and Molloy’s
investigation of the functionality of Bronze Age weapons. However, these two have very
different approaches. Jeffra’s paper takes a very quantitative approach, whereas Molloy’s
work is qualitative. Molloy’s work is experimental, because he is addressing hypotheses
about the use of particular weapon designs, but his tests rely on qualitative observations
from an experienced martial artist (himself). This opens up an interesting line of debate.
Various quantitative measures about weapons and the damage they inflict could be
recorded, but this does not go very far in evaluating their actual use as a weapon. Rather
than putting these too extremes in opposition to each other, perhaps it is best to
acknowledge the value in both types of investigation, while understanding their respective
limitations.
Molloy’s paper is not the only one that relies heavily upon experienced practitioners.
Aubrey et al. make use of a large number of very experienced flint knappers in order to try
to interpret the spectacular lithic assemblage from Maitreaux, France. This assemblage
represents a Solutrean knapping floor (apparently in situ) which was recovered with
amazing spatial resolution. Much refitting work has been carried out, and the experiments
are designed to understand the chaıˆne ope´ratoire in the reduction sequences and potentially
identify the activities of individual knappers.
6 Alan K. Outram

Both Domı́nguez-Rodrigo and Seetah are concerned with bone surface modifications.
The former concentrates on the identification of cut marks, while the latter investigates
wider patterning created by butchery practices. Both investigate in detail the relationship
between experimental work and analogy. They are interesting companion pieces.
Hurcombe’s paper is a broad-ranging one that demonstrates how experiments,
combined with the analysis of inorganic finds, can shed light upon the use of perishable
organic materials. Organics almost certainly formed a huge proportion of the material
culture of prehistoric peoples, as glimpsed when preservation is exceptional, but,
otherwise, it is necessary to be inventive and develop new ways to enlarge our
understanding by proxy. Another way to understand organic products is through residue
analysis. Evershed shows how actualistic experiments have been used to help understand
the results from analysing lipid residues from ceramics. In this work there are elements of
both ‘process and function’ and ‘simulation’ experiments, but this is also an example of
Reynolds’ (1999) fifth class of work, ‘technological innovation’, where the experiments
support the development of new techniques of archaeological investigation. Mithen et al.
are also using experiments to help underpin new methodologies, in this case in the field of
phytolith analysis, but, as with any agricultural trial with multiple plots and a degree of
time depth, this also represents an ‘eventuality trial’.
It is hoped that this volume demonstrates the value of actualistic experimentation within
archaeological science, and how such experiments can be integrated into much broader
programmes of research.

Department of Archaeology, University of Exeter

References

Coles, J. 1973. Archaeology by Experiment. London: Hutchinson.

Coles, J. 1979. Experimental Archaeology. London: Academic Press.
Feyerabend, P. 1975. Against Method: Outline of an Anarchistic Theory of Knowledge. London:
NLB.
Kuhn, T. S. 1962. The Structure of Scientific Revolutions. Chicago, IL: University of Chicago Press.
Mathieu, J. R. 2002. Introduction. In Experimental Archaeology: Replicating Past Objects,
Behaviours and Processes (ed. J. R. Mathieu). Oxford: BAR International Series 1035, Archaeopress,
pp. 1–4.
Mathieu, J. R. 2005. For the reader’s sake: publishing experimental archaeology. EuroRAE, 2: 110.
Outram, A. K. 2005. Publishing archaeological experiments: a quick guide for the uninitiated.
EuroREA, 2: 107–9.
Piotrowski, W. 1999. Wood-tar and pitch experiments at Biskupin Museum. In Experiment and Design:
Archaeological Studies in Honour of John Coles (ed. A. F. Harding). Oxford: Oxbow, pp. 148–55.
Popper, K. 1959. The Logic of Scientific Discovery. London: Hutchinson.
Reynolds, P. J. 1979. Iron-Age Farm: The Butser Experiment. London: British Museum.
Reynolds, P. J. 1999. The nature of experiment in archaeology. In Experiment and Design:
Archaeological Studies in Honour of John Coles (ed. A. F. Harding). Oxford: Oxbow, pp. 156–62.
Schmidt, M. 2005. Remarks to the publication of archaeological experiments. EuroREA, 2: 112–13.