Mapping Scientific Memory:
understanding the role of record-keeping in scientific practice
Gavan McCarthy and Tim Sherratt
This article appeared in Archives and Manuscripts, vol. 24, no. 1, May 1996.
About the Authors
Gavan McCarthy, MA Monash, BA LaTrobe, has
been the archivist in charge of the Australian Science Archives
Project since its inception in 1985. Prior to that he worked on
science, medicine and technology collections at the University
of Melbourne Archives. He is a professional member of the Australian
Society of Archivists Inc and a member of its federal council.
His research interests include the history and archives of science,
technology and medicine in Australia and more particularly the
development of resources and tools for continuing research in
these areas. His major publication is Guide to the Archives
of Science in Australia: Records of Individuals, D.W. Thorpe,
Port Melbourne, 1991. In recent years he has seen ASAP develop
its work to include major consultancy projects for the electricity
and pharmaceutical industries and become an important information
provider on the World Wide Web.
Tim Sherratt, BA (Hons) Melbourne, established ASAP's
Canberra Office in 1993 and, as Head of Information Services,
has been primarily responsible for the development of ASAP's Internet
resources. These include Bright Sparcs - with information on archival
holdings relating to over a thousand Australian scientists, the
WWW Virtual Library for the History of Science, Technology and
Medicine, and the WWW version of the ASA's Directory of Archives
in Australia. He is a historian by training with an active
interest in the history of Australian science. In his spare time,
Tim enjoys writing funding applications and backing up computers.
How do scientists document their research? As electronic means
of communication become the norm, this question has taken on special
urgency. If we don't understand the process of record-keeping
within the sciences, we are in danger of losing our scientific
memory - with severe legal, financial and cultural consequences.
This article introduces the connection between scientific practice
and the records-keeping process, indicating how little we know
of the technological, administrative and cultural dimensions of
this relationship and how it has changed over time. Archival research
that analyses this connection will enable the development of strategies
to deal with current and future problems. But how can we fund
In recent years a dramatic change has occurred in most areas of
scientific practice - the work that scientists do is no longer
being systematically documented. In the past, activities like
the keeping of notebooks and the writing of correspondence were
an unremarkable, but integral, part of being a scientist. Records
such as these provided a documentary safety-net, ensuring that
crucial information was not easily lost. This is no longer the
case. The rapid uptake of electronic means of communication, data
collection and reporting has brought a crisis in scientific record-keeping
- the safety net has gone.
This crisis, although not of equal severity in all scientific
disciplines or technology-based industries, has shown itself dramatically
in a number of environments. In the late 1980s a large organisation
in Victoria introduced an organisation-wide electronic document
creation and transmission system. Prior to its introduction a
structured paper-based record-keeping system was in place and
although use was not uniform across all work sites it provided
a workable framework for meeting the record needs of the organisation.
When the new system was introduced a disclaimer was made stating
that the system was not a record-keeping system. However, the
nature of the system led users (who had little knowledge of record-keeping
systems) to abandon their old practices which now seemed unnecessary.
Some seven or eight years later the organisation decided to abandon
the service-provider in favour of new PC-based systems. However,
many hundreds of thousands of documents which should have migrated
to the new systems could not be salvaged. The cost to the organisation
is in the millions of dollars and the costs will only increase
with time as the need for the lost records increases. As this
example indicates, scientific organisations and individuals in
Australia, and all over the world, are in danger of suffering
from an irrecoverable memory loss.
How can we deal with this crisis when we know so little about
the nature of the scientific memory -about the documentation processes
that occur within science? There is little existing research,
and most of this tends to concentrate on the salvage of
records within particular disciplines, or the salvage of personal
records of particular scientists. The archivist will always be
called upon to perform this sort of work, but what we need are
the knowledge and the tools to prevent situations developing where
there is nothing to salvage. This is the threat being posed by
the indiscriminate use of electronic documents and transmission
To embark on such a research program, we have to look beyond the
records themselves to that point within scientific practice where
records are created. We have to understand the value that scientists
ascribe to their records. We have to know how administrative and
technological change constrain their decisions. We have to map
the archival pathways that constitute the scientific memory. Only
then will we be able to develop strategies that meld with the
ongoing practice of science, instead of just cleaning up its messes.
A solid foundation
Much archival research has tended to focus on appraisal
- assisting archivists to make judgements about the significance
of existing bodies of records.
But waiting around for the leftovers is hardly a satisfactory
method of documenting a culture. Consequently there has been a
call for archivists to intervene in the process of records creation.
'Functional analysis' or 'documentation research' has been suggested,
by Helen Samuels and others,
as a way of moving beyond this after-the-fact archiving. Such
research examines the functions of the institution within which
the records are generated and seeks to identify what records would
be required to adequately document these functions. The application
of these principles by Joan Warnow-Blewett
to high-energy physics has shown that such research techniques
are extremely useful in seeking to document modern science. However,
the process by which records are created still remains somewhat
mysterious. Functional analysis can help us identify key points
in the process, but what goes on at those points?
The principles involved in the functional assessment of record-keeping
practices are also being addressed in the Pittsburgh Project,
led by David Bearman and Richard Cox.
This work provides a solid theoretical basis for future study
and practice, but it is only a beginning. Once enumerated these
principles need to be tested, implemented and assessed. They provide
a new avenue for research, not a substitute for analysis and reflection.
The real-world environment of modern scientific practice will
provide a challenging test-bed.
In the meantime, researchers in the social studies of science
have been invading scientific laboratories with their notebooks,
tape recorders and video cameras. Like the archivists, they too
have been keen to move beyond the documents, but their aim is
an understanding of the way science is actually practised. Latour
and Woolgar's pioneering
study of the Salk Institute portrays the laboratory as a factory
for the production of papers, or rather 'inscriptions'. Only when
data is inscribed in the form of graphs or tables can it be used
in the process of persuasion that results in the construction
of a scientific fact. Such inscriptions are thus highly valued.
Latour comments in one article on the "extraordinary obsession
of scientists with papers, prints, diagrams, archives, abstracts
and curves on graph paper".
Archives? Few of us who deal with the archives of science could
say that they had noticed such an obsession. Locked away in the
inscription process is the creation (or non-creation) of records,
but the picture is rather more complex. Inscriptions have value,
but they are cleaned-up, modified and finally incorporated into
published papers. What value is ascribed to the products of these
various stages? And what of documentation not relating to the
data itself - administrative material, correspondence, funding
applications? Anthropological or ethnomethodological studies of
scientific institutions open up the practice of science for inspection,
but the focus is on the creation of facts, not records. The tools
they provide are useful, but the apparatus of inscription needs
to be unpacked to reveal the many moments of record creation (or
non-creation) that constitute it.
Functional analysis helps us to identify sites for investigation,
and anthropological studies of science provide us with some of
the tools to conduct the investigation, but neither offers much
of an historical dimension to the study. We do, after all, need
to understand the way that record-keeping processes in science
change over time. To do this we need to find a way of exploring
past practices. We can't walk into a working laboratory of the
1920s, but we can analyse the documentation that survives from
the laboratory. The toolkit we put together from archival practice
and sociology of science will need to be applied across time,
in order for us to understand the nature and significance of the
current situation in the sciences. The various methods of historical
investigation will be called into play - biographical research,
institutional history - as will a general understanding of the
development of the scientific disciplines both in Australia and
overseas. Together these approaches will draw from the existing
records the story of their own creation.
Existing research and techniques will provide us with valuable
landmarks in our mapping of the scientific memory, but how do
we begin to fill in the outlines? We suggest a three-stage research
program that begins by investigating changes in practices over
time through the analysis of surviving records. This would then
be compared against the current situation by studying record-keeping
practices within specific scientific institutions. The knowledge
gained by these studies could then be used to design and test
appropriate tools and strategies for dealing with current and
The process of change
Science is undertaken within a matrix of social and
technological forces. The tools that scientists use, the way they
communicate, the means in which their projects are funded and
administered have all changed over time, and they continue to
change. The process of records creation is equally dependent on
such forces, but how are the two linked?
This question can only be addressed by mapping changes over time
- by looking at what records have been preserved and locating
these records within contemporary scientific practice. A combination
of archival expertise and historical understanding are required
to draw out the connection between the science and the records.
An excellent starting point for this research would be the analysis
of information on existing archival holdings relating to Australian
science already collected within RASA (the Register of the Archives
of Science in Australia) by the Australian Science Archives Project.
By looking for patterns within the remaining records and correlating
these against developments in Australian science, it would be
possible to provide a baseline for further studies that looked
at the contemporary scientific environment. Once the process of
change is better understood, current and future problems will
be able to be placed in a meaningful context.
A snapshot of current practices - who is responsible?
Rather than being overwhelmed by the vast amount
of electronic information currently being generated in the sciences,
we need to examine in detail the processes according to which
this information is being generated, and assess its significance
in providing a record of scientific activity. We need to explore
the various levels of responsibility at which the creation of
records can or should occur. Such responsibilities start with
an obligation that scientists have to themselves, to their career
and to their discipline in general. However, responsibility extends
beyond the individual to workgroups, project teams, supervisors
of research students or staff, departmental heads and, in science,
extends into much more complicated arrangements that include multi-institutional multi-national collaborative research projects.
Detailed case-studies within scientific institutions are required.
Through interviews with scientists coupled with an examination
of their administrative environments and communication practices,
detailed data could be gathered that would provide a snapshot
of current practices, needs and problems. But as well as identifying
the holes in the record-keeping process, such studies would suggest
targets for intervention - critical points within the practice
of science that could be directly addressed by improved documentation
New tools and strategies
We need to comprehend the problems facing record-keeping
in the sciences more clearly than ever before. Once understood,
these problems can be dealt with. The aim should be to develop
a range of intervention strategies that address the technological,
administrative and cultural dimensions of the current situation.
New software tools that aid the documentation process may be one
aspect, but equally there will be a need to work with scientists
to find ways of incorporating appropriate record-keeping processes
within scientific practice. Any such strategies will need to be
rigorously tested in real-life situations.
A new significance
This research is urgently required and of international
significance. It is no longer acceptable for archivists to trundle
along in science's wake, picking up the records that just happen
to be left behind. Electronic
records will not afford us such a luxury. Whereas paper documents
can only be consciously disposed of, electronic data can
be easily lost by changes in technology, inadequate backup procedures
and poor documentation. And what of concerns about fraud and misconduct,
intellectual property, legal and financial accountability? The
demands being made on our record-keeping systems are growing,
but how do we make them work better, when we know so little of
how they work? These are not new problems - they are all too familiar,
multiplied by inaction, exacerbated by lack of communication.
In recent times, some work has been carried out within organisations
or by private consortia, however, the results of this work has
not been made public. General outcomes need to be published to
reveal both the results of the studies and the research techniques
employed. These techniques are likely to be of international interest
and of use in a wide range of research-based environments.
But as we turn to finally face the horror, we must be careful
not to be merely reactive. As we pursue this long-overdue research,
we have the opportunity to develop a deep understanding of the
processes involved. We need to extend the archival research agenda
beyond the analysis of institutions and projects to the moment
of creation - the point where the actions of a particular scientist
determine whether a record is created or not. That moment where
scientific practice and the records-keeping process connect. This
is new territory, drawing together archival, historical and sociological
techniques. There is both urgency and opportunity.
Its time for funded archival research in Australia
For too long archival research in Australia has been
conducted in the margins and with very poor funding. We have relied
heavily on a few individuals who have been given the time by their
institutions to carry out this work and on the Australian Society
of Archivists to lead and cajole members to produce policy and
position papers that help further the understanding of our profession.
The Australian Council of Archives has also attempted to further
archival knowledge through working groups and meetings. However,
despite having a number of Universities teaching archives and
records management at the post-graduate level, the Australian
Research Council still has not created research categories for
archives or records management.
The need for archival research in the practice of contemporary
science is well demonstrated to the extent that some scientists
and scientific administrators are funding small preliminary projects
to deal with their immediate concerns. However, we need to progress
beyond this piecemeal effort and find funding for larger-scale
research resulting in the publication of substantive contributions
to our professional knowledge. This will empower us to meet the
demands being placed before us. A positive first step would be
for the Australian Society of Archivists and the Australian Council
of Archives to lobby the Australian Research Council for the inclusion
of archives and records management as categories of research.
We need to make the fund-givers aware that there is important
research that must be conducted. Such research will reap significant
benefits not only for the scientific community but for government,
business and industry.