Archives & Museum Informatics








last updated:
November 30, 2017 12:56 PM


Archives and Museum Informatics Technical Report #9 (Pittsburgh, Archives and Museum Informatics, 1989)

by David Bearman, Archives & Museum Informatics

Table of Contents

Introduction: The Archival Endeavor
Chapter I: Selection and Appraisal
Chapter II: Retention and Preservation
Chapter III: Arrangement and Description
Chapter IV: Access and Use
Chapter V: Intelligent Artifices: Structures for Intellectual Control
Chapter VI: Recorded Memory and Cultural Continuity



These essays were originally drafted in the late spring and summer of 1986 to be given as talks at professional society meetings, and were refined during the 1986 Research Seminar on Modern Historical Documentation at the Bentley Historical Library, University of Michigan. Being awarded a fellowship to attend the 1986 seminar, which was jointly funded by the Andrew W. Mellon Foundation and the National Endowment for the Humanities, enabled me to take the time to read deeply in the literature of the profession, where I discovered that the failings I had been seeing in American archival institutions were not primarily a consequence of bad management, but of the fact that the best methods of the profession were inadequate to the task at hand.

Since 1986, I have not pursued this research directly and the work has remained dormant. I have spoken about it publicly from time to time, but have not forced others or myself to deal with its conclusions. Increasingly I have felt that this is irresponsible of me, and am glad to be bringing it out for public scrutiny at this time. In the process, the first four chapters have been substantially rewritten, but the underlying structure of the argument, and its conclusions, have not been affected. Although references throughout the text were updated to reflect recent literature, the modest amendments and refinements of archival methods and practices proposed during the past three years have not required me fundamentally to revise my premises.

The first four essays address the four fundamental activities involved in the management of the physical record: selecting an appropriate record from the great volume of evidence, preserving that record against time, describing the record that has been retained, and providing for access and use. Of each of these activities I ask whether our present methods are adequate and if not, how they can be adjusted within the practical limitations which cultural repositories face. Throughout, my audience is the professional, but I do not mean simply those who work in repositories called archives. I include here museum curators, librarians and others whose efforts are devoted to preserving cultural memory and making it available to future generations, and who frequently are also responsible for archival materials.

Chapter five explores an idea relevant to the changes in methodology proposed in the first four chapters. Intelligent artifices are presented as a tool for expanding intellectual control, and consequently, access and use. Chapter six returns to the themes in the introduction to consider the role of the archivist and the nature of the archival endeavor, and to propose new views of both based on the ideas in the intervening chapters.

More important even than the generous financial support I received from the Mellon Foundation and NEH is the intellectual support of Francis Bluin and Bill Wallach of the Bentley Library, and of my colleagues in the 1986 Seminar - Chris Baer, Greg Bradsher, Judy Endelman, Avra Michelson and Peter Sigmond - who read and discussed drafts of these papers with me that summer and stimulated me by their own research. In addition, I would like to thank Helen Samuels and Ed Bridges who read and critiqued drafts in 1986, and Richard Cox who read drafts in 1989. Finally, these essays have been greatly improved by the careful editing of Lynn Cox, Managing Editor of Archives and Museum Informatics Technical Report, without whose assistance they would not have been published.

Introduction: The Archival Endeavor


It is a hallmark of human societies that they seek to preserve a memory of the past, and have always done so. Indeed, keeping and using the past is central to our concept of human cultures and civilization.

In the oral culture of non-literate societies, in the myths and fables refined by centuries of retelling, we are delighted to discover a source of fundamental truths, of the themes which make us essentially human. This record of apocryphal events, reflecting historical facts but not bound utterly by them, is a distillation of a once living testimony. The thematic integrity and size of the oral record, made it possible for a single human mind to comprehend the whole, and to relate it to the next generation in appropriate contexts.

The advent of literacy posed a new problem for human societies, for in freezing the memory of a culture in material formats, in creating a written or recorded memory which resists transformation, man provided himself with an objectified past which stands in opposition to the present in a way which the oral tradition, so bending and resilient, never did. Because societies were enabled to extend themselves further across space and time through writing, and subsequently through the recording of sounds and images, than through the oral record, the "technology of preserved communication" has allowed for the development of a more variegated cultural repertoire1. In so doing, it has created a world so complex that no individual can possibly master it all. Recently, the record of human endeavor has grown so large that no individual can even master where to find it all. In such a world, special roles have been created to assure the interpretation, retention and transmission of culture and special institutions have been established to support these functions.

Archives, historical societies, libraries and museums are the institutions created by our society to play the role of selecting, conserving, and providing access to the record of our culture. To an extent far greater than the society admits, these professional keepers of cultural evidence are not simply custodians of our recorded past, but shapers of the cultural memory. This is true in the first place because archives and museums are in no position to acquire a complete record of our culture (though libraries might, in the aggregate, acquire a nearly complete set of the published record of our society). But it is true for a more basic reason as well: all ordering, all conceptual schemes, all means by which man comes to name his reality, are, reflections of culture. Consequently, like Adam naming the creatures of the earth, they frame our conceptual universe and constitute at the same time both our cultural heritage and the record of that heritage. The first, and greatest, challenge of archives is to select the archival record, and shape that heritage.

The cultural record is continuously being deposited and eroded. It is deposited at different rates by different groups within the society and at different times in the life of the culture (reflecting the socio-economic and cultural-political requirements of the society). The cultural record is also subjected to differential rates of erosion due to the properties of the physical medium upon which it is recorded, the care taken of it by the sub-culture in which it is created, and the need which the culture has for the particular information it holds. Archivists live with the certainty both that all activities leave some recorded memory in our society, and that each recorded memory will disappear in time without intervention by some preserving agent. No technologies currently exist that will permanently arrest those natural processes of decay that erode even the tiny portion of the overall record which is retained in archives. Thus the second challenge of the archival profession is to preserve a record, by necessity a portion of the total record, and thus further to shape the memory of our society.

Ultimately, archivists are expected to provide access to whatever is kept for however long it can be used. The memory of the society is not a dead record but a tool in the continuing reinvention of the culture. The record is the seed of a developing culture, it is the source of identification that individuals feel to their state and their society. The record also details the obligations of individuals to the state and to society, and the accountability of social institutions to individuals. In the immediate need that a culture has for its own memory lies the third major challenge to the archival professions - to assure the use of cultural evidence in the continuing construction of the culture.

Finally, just as the culture has grown beyond the memory of the individual, so have its individual repositories grown too large for individuals to serve as guides to the materials therein. Already our repositories threaten to become mere warehouses, mocking us as monuments to our propitious ability to generate records, because we are unable to direct potential users to the information which they seek. Furthermore, information contained in the card catalogs and finding aids designed to assist researchers in each individual repository must ultimately also be merged into reference sources for counties, states, countries and the entire world. Means of identifying like materials must be developed which do not require more resources than the society is willing to expend. Solutions to the problems of intellectual access, access to meaning, are far more complex than those to the technology of access, but will have to be addressed if the cultural archives we are acquiring are to serve their larger purposes.

The Sisyphean task of archivists is to overcome an almost logarithmic (though presumably not inevitable) growth in the volume of historical evidence and reverse the effects of physical decay so as to be able to focus an unbiased view of the past through any given facet of the crystal of cultural memory. These essays examine the task of archivists in the United States. They accept as fact the assessment of the extent of archival challenges documented in the literature of the profession, and they analyze the best methods for managing records proposed by the profession. Although the problem in the U.S. is probably no different from that of other developed countries, these essays were limited by the need to focus on a discrete setting so that data estimating the size of a specific problem could be used to assess the weaknesses of a specific body of practice.

Questioning received wisdom is a modern means of advancing methods, one which anthropologists assure us is itself a product of literate society, of a society which makes records, and which cannot therefore "discard, absorb or transmute the past" through the "homeostatic process of forgetting or transforming those parts of the tradition that cease to be necessary or relevant." It is appropriate for archivists to reflect, as they question their own received views, that the product of their professional activity is extremely subversive to the society in which they live, not simply for the obvious reasons of its conscious manipulability, explored in George Orwell's 1984 , but because the very nature of recorded knowledge is culturally disturbing. As Jack Goody and Ian Watt put it in their extraordinary examination of the "Consequences of Literacy," members of literate societies "are faced with permanently recorded versions of past beliefs, and because the past is thus set apart from the present, historical inquiry becomes possible. This in turn encourages skepticism; and skepticism not only about the legendary past, but about received ideas about the universe as a whole."2

These essays, like all scholarly and professional discourse, are a manifestation of just such skepticism. Why else would archivists feel compelled to question their roles as culture shapers, keepers and transmitters? What, fundamentally, is our goal? Is what we are trying to do desirable? And is it possible? If we are to select evidence from the universe of documentation deposited by our society in its every activity, what criteria shall we use to determine what to keep? What methods shall we use to identify and measure the recorded evidence of our culture by these criteria? What benefits will be realized by the society if we succeed? What risks if we fail? Are our expectations for preserving cultural evidences against the ravages of time realistic? Are the means at our disposal adequate to make the effort credible? Are they even sufficiently adequate for us to feel that we are being professionally responsible in pursuing them? And finally, if we succeed in selecting a cultural record and retaining it, to whom will we be able to provide access? How will they be able to interpret and use the memory of our society? And how can we most effectively convey to future users the nature of the record which has been kept?

The questions are, of course, not new. One SAA President treated his audience to many of them a decade ago.3 But I believe the perspective I want to bring to answering them has not been applied rigorously to the problems of archives. Simply stated, I will ask two basic questions about each major goal of the archival challenge and about our current assumptions and methods:

These essays attempt to quantify, wherever possible, both the size of the archival task in modern America and the capabilities of the American archival profession given its current methods and resources.

Because of the nature of the problems which these essays treat, comparison of the magnitude of the tasks and the magnitude of our capabilities often reveal substantial discrepancies. In each chapter, I have found that the shortfall between documented needs and proven methods is greater than one order of magnitude (a factor of ten). The concept of an order of magnitude is central to these essays, so readers deserve a brief reminder here of its meaning. An order of magnitude is represented as an increase of one power of ten; 10 squared (100) is an order of magnitude greater than 10 to the first power (10). While we may meet people during our lifetimes who are widely different in height and weight, they will all be between 0.1 and 9.9 times 10 inches tall, indeed they will be between 3 and 8 times 10 inches tall. If we were to hear of someone an order of magnitude larger or smaller, we would say that they were not a person, but a different species altogether. Similarly, while we can strive to improve any given human method by a substantial percentage through careful systems study and the adjustment of its application, refinements of our methods rarely yield 100% improvements (a mere doubling of output). Order of magnitude improvements of human methods (1000% for each order of magnitude), are unheard of without the implementation of fundamentally new tactics, technologies or goals.

Therefore, when these essays discover time and again that archivists have themselves documented order of magnitude and greater discrepancies between our approaches and our aims, they call for a redefinition of the problems, the objectives, the methods or the technologies appropriate to the archival endeavor. In this respect, these essays differ from most official studies and reports to the profession, which uncover such discrepancies, but too often simply call for greater resources. I hope that by examining our methods critically, and seeking solutions in changes to archival premises, and techniques, these essays can make a contribution to the profession.

Chapter I. Selection and Appraisal


The Problem:

The first challenge confronting archivists is to select the archival record of our society. The centrality of this function, called appraisal in the jargon of the field, is properly recognized by the profession. In the SAA Basic Manual Series, Maynard Brichford calls it 'the most significant archival function and quotes Marcel Baudot, who characterized it as the "sine qua non of all sound archival practice."1 The identification and retention of records of enduring value was the first of three goals articulated by the 1986 Report of the Society of American Archivists, Task Force on Goals and Priorities.2

The properties of the cultural record that pose challenges to our methods are numerous. These methods must somehow accommodate the facts that records generating activity is dispersed while records assessment is localized, and that records generation is under the control of autonomous individuals while records appraisal is institutional, and usually centralized. Our methods must be sensitive to the fact that cultural activity takes diverse forms, yet its record is restricted by the technologies of the day. And they must acknowledge that while cultural activity is purposive, the records of this activity are mere by-products. As much as record creation and cultural activity are time bound, records are retained in order to overcome the bonds of time. Finally, our methods must enable us to cope with an almost unimaginable volume of recorded evidence, of which only a tiny fraction can reasonably be kept in archives.

We have meager resources with which to address these challenges. A recent survey found archives have a median annual budget of $82,000, of which over 75% is devoted to a staff of 3.5 professionals and non-professionals.3 The total annual expenditure of all the archives in the study did not exceed the $100 million appropriated to the National Archives. Extrapolating from the survey results to the entire archival community is risky, but if we project that the budget of all the archives in the U.S. is less than $500 million, we are certain to be estimating high. The budget of all the State archives is less than $20 million.4

If we examine the records archives manage, we discover that the vast majority of existing archival repositories are only equipped to deal with records on paper or film, whether text or image. The holdings of all the archives in the country probably do not exceed 10 million cubic feet of text on paper, and a trivial quantity of television and radio, images and computer readable records. Yet the record of modern society is vast. It is created as a consequence of virtually every human activity and resides in every institution and with every individual.

There are 4000 archives reported in the most recent census of archives.5 Even if we assume that more than half of all repositories have somehow failed to be reported, and allow that there might be as many as 10,000 archives in the United States, we must be struck by the disparity between this figure and the over 3 million corporations (over 500,000 of which report incomes over $1 million per annum) and 83,000 governmental entities in our nation today.6 The organizations which could have archives and which archivists would normally wish to document include over 50,000 banks, 20,000 radio, television and cable stations, 6800 hospitals and 3300 universities and colleges. There are more than 25,000 associations listed in the 1988 edition of the Directory of Associations.7 Obviously, records are being created in many more institutions than have archival programs or are appraising records, and the principal focus of all archival programs is the records of their own institutions.

The record of our society is found in every medium and every format. Based on the 108th edition of the Statistical Abstract of the United States, we know that Americans consume 4,500,000 tons of writing paper annually, which translates into roughly 4.5 billion cubic feet of paper records, based on 2000 sheets to a foot and one foot to 20 pounds. More than one hundred thousand gigabytes (each one billion bytes) of data can be recorded on the 425 million blank floppy disks sold every year to say nothing of the greater quantities stored on magnetic tapes and large disk drives. In addition, television programming fills dozens of channels twenty-four hours a day, 365 days a year. One hundred ninety two million cassettes of unrecorded video tape, representing over 500 million hours of recording, are sold annually in the United States along with 268 million blank audio cassettes. Photographic equipment, supplies and services account for $10.8 billion of economic activity annually. This translates into hundreds of millions of photographs, x-ray and other developed images that are added to hundreds of thousands of images drawn, painted, and otherwise produced, creating a vast archive of images that contributes to the material evidence of our society.

Over 240 million telephone conversations are initiated every day in the United States! (As recently as 1970, there were only 109 million a day.) Our documentation of these conversations is limited to the transient contents of 10 million telephone answering machines and telephone message pads and memoranda of conversations, but this archive is about to be augmented by a potential digitized record of almost unimaginable dimensions representing billions of telephone conversations conducted over the digital communications network of the 1990s. Will we want to capture any part of this documentary record that dwarfs the 76.5 million pieces of first class mail delivered by the Post Office annually?

Because this vast recorded archive of interactions is generated by every person, in the context of every activity, its retention for minutes or hours, to say nothing of years and centuries, is under the control of each of these agents. Records are created and stored in homes and schools and offices, in publishing houses, studios and advertising agencies. They are made on film, tape, and paper, in sound, images and text. And they are stored in boxes and filing cabinets, bound volumes, scrap books and tape libraries.

Unlike libraries, which have a finite and knowable amount of published information to collect (there are a mere 3,000 periodicals publishers and 2,000 book publishers in the United States, for instance), and which can determine what part of that literature each needs to acquire based on collecting goals that serve local needs and national programs established to share resources, archives are faced with a body of evidence of human activity of unknowable dimensions and significance. How adequate are our methods for shaping this memory? How well do they serve local and society wide needs? How effective are they at assuring the best use of our resources so that our attentions are directed to important records? If carried out perfectly with existing resources, how well would they satisfy the needs of the future?

The size of the annual accumulation of recorded evidence, and the variety of its sources of creation, prohibit archivists, even collectively, from ever coming into contact with all but a tiny percentage of it. Indeed, reliable statistics of the volume of records offered to or appraised by archives in a year are not available, but there can be little doubt that hundreds of times the volume of twentieth century records ever offered to archives are stowed in closets, basements and storage rooms. If somehow archivists could review it all, how much manpower would be required to appraise it, using present approaches?

Archival Methods:

Records may be consciously retained by the organization which creates them because the law requires it, or because the organization needs the information they contain for ongoing operation, but when these tests are not met, and the organization chooses to keep records, they are applying archival criteria. The concept that some records have continuing value as historical evidence, and that this value is a product of their separately assessed, evidential value and informational value, is basic to archival appraisal.8 Evidential value is the property of records to document the organization, functions, policies, decisions, procedures, operations and other activities or events that are of importance to an organization. For traditional archives, whose mission it is to guarantee organizational accountability, evidential value is the primary reason to keep records beyond the period of their administrative usefulness or "active" life. Informational value is the value of the contents of the record for future analysis. While informational value is a secondary concern to institutional archives, manuscript libraries and special collections libraries, look first for informational, or research value in what they refer to as "archives". Records of informational value, such as the census, may be kept by traditional governmental archives, but they would consider it remiss not to have the evidential record, in this case the records of how the Census was organized and run and how the data was compiled, in addition to the data itself.

The concept of values as criteria to guide archivists in selecting which records of their organization merit archival retention is generally put into practice as a sort of catechism or flowchart of questions that ought to be asked of a potential accession before acquiring it. Despite the extensive listing of qualities that contribute to each criteria found in some exhaustive catechisms,9 the theory of values has serious shortcomings as a tool for making appraisal decisions within an institution, and is fatally flawed in helping to make broader appraisal decisions.

Within institutions, these failures derive from conflicting premises about "value," exacerbated by the fact that archivists call the process of selecting records for retention, appraisal, thereby emphasizing the cost-benefit analysis that is implicit. In theory, the institutional costs of retaining records are weighed against the social benefit of having them preserved. Early discussions of appraisal theory in the United States made the cost accounting aspect of appraisal decisions quite explicit10, but in subsequent models, costs have been present as one, somewhat minor ingredient of institutional policies if at all.11 In any case, archivists are at a disadvantage arguing within an institution for benefits presumed to be achieved for the whole society, even if their position was not fatally flawed.

Cost-benefit methodology cannot be employed when one side of the equation cannot be calculated and the other side is infinite, as it is when appraising records for permanent retention. The cost of retaining anything permanently is infinite. The value of any continued retention is based on hypothetical, counter-historical arguments. In such a situation no cost-benefit justification can long stand the scrutiny of serious managers, and if anything is to be retained at all, the archivist finds himself qualifying the costs side of the equation until eventually almost any hypothetical future benefit provides an adequate justification for retention.

To earn the respect of their fellow managers, and to begin to construct a viable practice of appraisal, archivists need to substitute the language and methods of risk management for cost-benefit analysis. Instead of asking what benefits would derive from retaining records, they should insist on an answer to the probability of incurring unacceptable risks as a consequence of disposing of records. This will very likely dramatically reduce the volume of records that are judged essential to retain. And it suggests an approach to solving the second dilemma of our current appraisal methods: their focus on records rather than the activity they document.

The practical constraint on appraisal as it is now practiced is one of limited manpower. Although we lack data to document how much manpower appraisal requires under normal circumstances in an individual repository, we do have some reliable measures derived from four carefully monitored studies of massive appraisal efforts: two in the Federal government, one of a state court system and one in a cooperative project of private repositories appraising the records of the now dissolved Pennsylvania Railroad Corporation.12

Each of these projects employed a team of appraisers using sampling methodologies to assure that an adequate record would be retained. The results were surprisingly consistent in ratios of staff assigned to the tasks and in the volume of both records appraised and records retained. One full-time appraiser was responsible for less than 10,000 ft. of records appraised per annum. Each staff person accounted for an average of 1600 ft. of records retained as a consequence of appraisal. Such numbers indicate that the overall requirement for appraising the nation's recorded heritage far exceeds the resources available.

At ratios of 1 person year to 10,000 cubic feet of records appraised, it would require 450,000 man years to review the 4.5 billion feet of paper records created annually in the United States, to say nothing of the machine readable data, images, sound recordings, video tape and other media. At 1:1600 feet retained, the archival profession in the United States, which numbers at most the full-time equivalent of 1000 professionals involved in appraisal, would decide to keep 1,600,000 feet of the 10 million feet it was able to review per year. In other words, archivists using sophisticated sampling based methods of appraisal, could review .2% of the annual record, and decide to keep .035%. The input they could review would fall short of what must be appraised by three orders of magnitude, while their output would still be an order of magnitude larger than any we have ever kept to date!

The appraisal efforts mentioned above assembled teams of archivists whose experience exceeds that which most archives could marshal, and equipped them with tools, especially the statistical tools, that are much more elaborate than those most archivists could employ. For this reason, the extrapolation of the results of these systematic efforts to the country as a whole is even more discouraging than the raw data suggest. These projects addressed themselves to culling record series measured in tens of thousands of feet so that each appraisal decision made affected huge volumes of material. Perhaps the highest efficiency archivists could hope as a profession to achieve without nationwide mobilization would be to review 2-4 million feet of records per year and decide to keep 200400,000 cubic feet. This would appear, from the external evidence, to be approximately the current situation.

Not only will our usual methods not permit us to review a plausible quantity of the overall documentation of our society, but if we used methods approaching the sophistication of those employed in these huge appraisal projects which could exploit statistical sampling and expert panels to assist in decision-making, we would apparently decide to keep between 15% and 25% of the records appraised. To appreciate what this would mean, archivists need only consider their widely held belief that they select only 1-3% of the record for retention, a belief based on calculations such as those by Greg Bradsher, which compare the volume of holdings (in this case 1.5 million feet of records in the National Archives) to the universe of records created by the institution (in this case the 25 million feet of records created by the Federal government during the 140 years prior to 1940 and the many millions created since).13

Is 15-25% the right amount to keep? This question brings us to yet another fundamental problem with the concept of values in appraisal. The theory of differential t1values" in appraisal has equipped the archival professional with a rationale for sculpting the past, but it has done nothing to clarify the goal of the effort. Not only does this make values a blunt tool, it leaves us in no position to evaluate success. Appraisal theory, until very recently, has failed to ask, "What kind of record do we want to preserve19? The profession has not critically examined the effects of applying these criteria, possibly because the cultural goals of archival appraisal are unstated.

It would seem that archivists have rarely asked the teleological question because we believe that the answer, within the context of modern democratic self-evident: we have assumed that the purpose of appraisal is to select a representative record.14 The concept of representativeness certainly arises whenever such goals are discussed, but the profession does not agree whether this record is intended to be "representative" of all of recorded memory, or "representative" of the activities of members of the society, or "representative" of those aspects of social activity perceived by members of the society at the time as important to the understanding of the culture. Most archivists apply appraisal criteria to records, not to activities or social policy processes, and therefore assume that the goal is not to skew the record as received.

Among the practical problems posed by any appraisal process which takes representativeness as its goal is the reality that appraisal takes place within an institutional context, isolated from either a meaningful knowledge of the "universe of documentation" or from the appraisal activity of other repositories. Recent writing on appraisal has addressed both these issues, but the impressive efforts to define various "universes" of documentation15 and to make appraisal decisions known to the professional community16 have yet to be translated into new methods. Therefore archivists still face the challenge to either develop a method of selection for archival retention that discriminates with sufficient efficiency to increase the productivity of appraisal by about one order of magnitude and reduce the retained record by an order of magnitude, or to redefine their objectives. Better methods must permit the average appraisal archivist to select archival materials from at least one million (as opposed to the 10,000 possible now) cubic feet of records per year without loss of effectiveness. The alternatives of increasing the size of the profession by an equal magnitude (to 500,000 professionals) or accepting a 1% relevance of retained records, are both equally unrealistic.

Recently, those suggesting that appraisal has failed to identify an appropriate universe of records to appraise have become a powerful voice in U.S. archival circles.17 They have forcefully argued the inadequacy of appraising whatever records come to hand, and coined the concept of a concerted and cooperative documentation strategy in place of less systematic, institutionally based appraisal of records. But the documentation strategies approach is itself flawed by an absence of methodologies by which to define what constitutes an appropriate subject of a documentation strategy, and by the excessive manpower requirements of conducting such coordinated archival strategies. The only real case study to date recounts a 25 year effort of several American Institute of Physics (AlP) professionals to define and implement a documentation strategy focussed on the activity of modern physics in conjunction with archivists from universities and Federal research institutions throughout the United States.18 While most are prepared to credit the AlP effort with tremendous success, we should also be ready to confess that a vast number of activities are as important to contemporary society as physics. How many such documentation efforts can the United States support, and what would be the consequences? In discussions with Helen Samuels over the documentation strategy she proposed for "Route 128"19 the post-industrial society's industrialization phenomenon in suburban Boston, this author asked rhetorically whether the decline of the cod-fishing industry, the rise of tourism, the remaking of Boston harbor, the arts community of the Berkshires and dozens of other Massachusetts phenomena didn't warrant documentation strategies as well? While all the archivists in Massachussetts are planning these documentation strategies, who will engage in national and international strategies of equal significance?

The intellectual attraction of the documentation strategies approach should be that it focuses on appraisal of activities and functions rather than of records.

Helen Samuels has always framed the task in terms of appraising functions and activities, rather than records.20 I also contend that we will only be able effectively to appraise larger volumes of records if we focus our appraisal methods on selecting what should be documented rather than what documentation should be kept, and develop tactics for requiring offices to keep adequate documentation, rather than trying to review what they have kept to locate an adequate record.21 While a focus on selecting functions raises difficult questions as to how archivists should assist in documenting activities which do not produce a recorded memory (questions debated under the guise of discussions about the "value'9 of oral history and the concept of "adequacy of documentation"), it has the advantage of permitting the archivist to be proactive, making appraisal judgments before records are created, and to provide instructions that are explicit and targeted. In place of vague General Records Schedules based on media characteristics and types of records series, the archivist can require that a record of a stated activity or event be kept that is adequate to reconstruct what happened and why. In place of sampling records to determine what individual items or cases can be discarded while remaining representative, the archivist can require records creators to be accountable for avoiding institutional risks by documenting specified activities.

Finally, we began by stating that archivists must recognize that any collecting activity, unless utterly comprehensive, is a selecting activity, and that selection always involves shaping. In order to avoid shaping the record too much, archivists have articulated new approaches based on documentation strategies and sampling procedures. But when we examine these strategies in depth, they must fail for lack of resources. Ironically, if we discard the preferred methods of trying to engineer "representativeness", we can nonetheless accept it as a objective because our methods for selection are so utterly inadequate for the size of the available documentation, that random and accidental processes outside of the domain of conscious culture preservation dwarf the impact of our activity.

The law of averages as it pertains to the survival of records assures the retention of vast classes of materials without further attention by archives. Consider, for example the cumulative impact of archivists in the United States on the size and character of the record of the 1 980s that will remain in the year 2080. By our estimates, the country has created close to 100 million cubic feet of records annually during this decade in addition to 2 million hours of TV broadcasts, 9 million hours of radio broadcasts, 50,000 terabytes of data, and 1 billion still photographs. Archivists have appraised less that 1% of this output, and decided to keep 25% of the paper, and less than .1% of the rest. If we assumed that all the materials in archival custody survived until the year 2080, and that the paper outside of custody has a loss and decay half life of 25 years, there will be still be 25 times as much record of our decade remaining as a result of random accidental non-destruction in 100 years as there is in archives. If we assume that magnetic media and photographs have a half life of 10 years, ten times the volume of these media would remain outside of archives as inside. And this of course assumes that no further deposit of materials in archives would take place during the end of the intervening century.

I would argue that our selection methods must do more than acknowledge the impact of actions by non-archivists, of retention by other cultural repositories (though this is a valid point of the documentation strategists) and of conscious disposal and known physical destruction. Archival appraisal should, first and foremost, take into account the effects that random retention will have in preserving records one hundred years from now. Archives should only strive for representativeness of the informational record when there is reason to be1~Leve that random effects will not adequately assure the survival of an historically valid sample.


Archival appraisal approaches based on the concept of values inherent in records fail as rational management policy because they create a false impression of being based in cost-benefit analysis. This aspect of appraisal decisions should be replaced by risk management strategies.

Archival appraisal approaches based on the assessment, or sampling of records, fail as practical means of appraising the volumes of records with which we as archivists are confronted. These approaches, especially the much studied sampling approaches, also retain an excessively large proportion of the record, approximately 15-25%, which is not a practicable quantity to extend across the society. They should be replaced by approaches focussed on appraisal of activities that generate records, and assignment of retention responsibilities to those conducting the activities.

Archival appraisal approaches based on trying to build a representative record of human culture through actively shaping the archival record fail to make a dent in the record as it will be preserved a hundred years from now. At the same time, they distract attention from a legitimate teleological debate that would force archives to define their role in society.

Chapter II. Retention and Preservation


The Problem:

When archivists determine that records possess evidential or informational value sufficient to warrant their archival retention, they designate them as "permanent" or of "enduring value" and accession them. They tell themselves, their institutions and their donors, that they intend to keep such records forever. Nevertheless, they must recognize that recorded memory is a material manifestation of the past, and as a material manifestation it will need to be stored and it will undergo material decay. Indeed, these two physical requirements, storage and conservation, are closely related.

The media on which mankind has been recording its history since the advent of literacy have become increasingly fragile and susceptible to decay with each succeeding technological development. The shift from stone to clay tablets, from clay to papyrus, from cloth paper to wood pulp paper, from paper to photographic media and now to magnetic recordings has produced ever shorter format lifetimes. Along the way, we have introduced media and recording techniques which are not "eye-readable". To listen to a wax cylinder or analyze a data tape we also need the hardware or software which was designed to use them. The generations between such technological changes are shorter now than ever before - thousands of years elapsed between stone and clay, hundreds between clay and papyrus and cloth paper, wood pulp paper is less than three hundred years old, photography and sound recording only one hundred-fifty, magnetic tapes only fifty, and optical storage media are less than ten years old.

Against such rapid change stands a concept of permanence, dwarfing any of the scales which apply to the realm of recorded memory or the preservation of its material manifestation.1 Recorded history is measured in tens of thousands of years. The immediate ancestors of modern man have populated the earth for about one million years. Man has been evolving recognizably for the past ten million years. Mammals emerged a little over one hundred million years ago. Permanent is a very long time indeed! In this context, it is odd and somewhat frightening that traditional conservation efforts still refine methods to extend media by factors as little as two or three times.2 Usually the costs of these approaches are such that one cannot extend them to more than a fraction of the archival record, even then without any particularly sound reason for believing that the ravages of time will thereby be suspended.3 If the intent of such intervention is to save the cultural record even for one thousand years, that is, for a noticeable period in the scale of recorded history, these interventions will have to take on heroic and untested proportions.4

The importance of preservation of the record is such that both the Society of American Archivists and American Association of Museums reports on the future of these professions identify it as central to the purposes of the administration of cultural repositories, and major preservation programs have been launched in the past several years by the National Endowment for the Humanities and other national organizations to secure the preservation of important cultural documentation.5 The breadth of the cultural commitment to preservation, and the scale of its potential technical and financial requirements, are illustrated in the 1984 report of a panel convened at the request of the U.S. Senate Subcommittee on Science, Technology and Space to recommend policies for the archiving of weather and earth satellite data in the event that the satellites were sold to a private corporation. The panel found that "it is in the public interest to maintain an archive of land remote sensing satellite data for historical, scientific and technical purposes. The data in question are a national resource worthy of preservation...and while the cost of archiving these data is not insignificant, it is extremely small relative to the investment in the space segments of the satellite remote-sensing systems...." At the same time, the challenge of preservation is underscored in the body of the report where the Earth Resources Observation Data Center proudly asserts that "with proper environmental conditions and handling procedures the data and black and white film) are expected to be good for 20 years -- color film for ten years."6

A recent study by the National Association of Government Archives and Records Administrators (NAGARA)1 estimated that the costs of preserving the one million cubic feet of records (2.5 billion items) currently held by state and local government archives at almost half a billion dollars over the next decade ($135 million from non-state sources to match an additional state expenditure of $350 million in addition to the $17 million that would be spent based on 1986 spending levels).7 This expenditure would go toward the passive maintenance of materials, improved storage conditions, and the transfer of information on media that are seriously deteriorated to more stable media, not toward active conservation treatment. The report does not attempt to estimate the length of time by which such an investment in preservation would extend the life of the records. In any case, the costs of conserving the relatively small volume of records already held by state government archives exceeds the current social expenditure for the maintenance of these archival records by a factor of ten. To fund the preservation of existing records, suspending all new acquisitions or public services, would require more than a tripling of current budgets. Such findings led NAGARA to call for increases in expenditures sufficient to offset the expense of the estimated conservation effort; they lead me to ask whether there isn't something fundamentally wrong with our goals.

The recent emphasis within libraries, archives and museums on collections management has fostered a revision of the traditional focus of conservation programs on the repair and restoration of individual items. Increased emphasis is being placed on collection wide storage and handling practices which slow deterioration or stabilize the current situation. While commendably more realistic, both information oriented "preservation" and collections management fail ultimately to prevent decay, and both represent financial drains of unprecedented magnitude. Archivists must adopt a considerably more radical stance.

Archival Methods:

If we are really going to store and preserve records forever, of course, preservation measures would need to be taken forever and space for records storage would have to be maintained forever. The costs of doing anything forever are infinite, by definition. Because the very concept of permanent retention is preposterous and flies in the face of the laws of physics and economy, archivists have permitted themselves to ignore the consequences of acts of preservation that fall short of permanent retention. They have overlooked the obvious fact that conservation of the original records of contemporary society, comprised as they are of materials that nearly defy preservation, is impractical in the extreme.

The problem is, of course, of our own making in that archivists have declared that certain records are to be retained "permanently". The life of 19th and 20th century records on paper is estimated at less than fifty years.8 With exacting storage conditions and treatment to deacidify the paper and stabilize the inks used to record information, this can be extended to several hundred ears. Extending the life of such materials to one thousand years, would require regular intervention such as on-going deacidification and heroic measures of storage, such as retention in cryogenic vaults.9 On the other hand, copying onto microform combined with reasonable storage measures and low use may be able to assure an information life approaching one thousand years at acceptable cost.

Even with the best care, the magnetic records of the twentieth century, including sound tape, videotape and data tapes, will lose their signal after less that twenty-five years without recopying. The signal recorded on optical recording media may well last two or more times as long, but this is more likely to impede future access than to improve it.10 The problem with preservation of electronic media is that the media can easily be preserved longer than the capability of reading the signals recorded on them. Magnetic and optical media for recording of sound, image and data are all subject to market forces and technological change which are occurring at a rate that requires us to continuously recopy media to newer physical and logical formats in order to preserve access. Archival guidance concerning sound recordings in various formats recommends that older types of recordings be transferred to reel-to-reel tape in order to avoid having to maintain a museum of outdated sound playback equipment,11 but this assumes that archivists have the playback equipment now and that reel-to-reel tape won't soon suffer the same obsolescence. In a recent study for the United Nations, this author concluded that such recopying of records will almost certainly have to take place within offices with each migration they make to new systems, media or software.12 The recommendation is based on the presumption, which is urged as a matter of policy, that there will be a continuing requirement to migrate the information, certainly not less often than once a decade.

A strategy which can instantly reduce the dimensions of the problem to manageable size and free archival energies for other tasks is to simply declare the problem away. Rather than setting our sights on posterity, we need to replace the concept of "permanent retention" with the more realistic concept of "retention for period of continuing value" and adopt policies based on the premise that no preservation measures should be taken to extend the "format life" of the materials. Format life is the length of time which, given reasonable care, the information contained in the records will remain usable in its original format. Materials should be reformatted when format life has expired, if the information in them has continuing value. These redefinitions are not simply a sleight of hand means for instantly reducing the size of the threatened volume of "permanent" materials in custody, or a bureaucratic ploy to provide a more convenient definition of reality (both of which they are to some extent); they are offered as a way of redefining archival purposes.13

In their definition of archival, the authors of the official glossary of the Society of American Archivists defined archives as records kept for their "continuing value."14 But archivists have not taken the concept of continuing value seriously as a policy framework. If they decided to keep materials for their "continuing value" only, archivists would need to reassess this continuing value on a regular basis, freeing archives to deaccession materials found to no longer deserve retention. I suggest this is the only responsible management strategy. But my proposal is not equivalent to Leonard Rapport's 1981 proposal to deaccession existing collections.15 while I am sympathetic to Rapport's proposal, and it would save us some resources for storage of records over the longer run, it does not contribute significantly to the resolution of our preservation problems because it employs the same appraisal methods that have already been demonstrated to be inadequate. Further, it intimates that the need for reappraisal arises from previously impaired decision making, rather than that it should be a regularized function that is a desirable consequence of recognizing changing constituency needs and cultural values. what is proposed here is that we rethink the entire framework of accessioning materials and make all decisions about their continued preservation based on realistic assumptions about the length of time that they will be retained.

Minimal preservation measures, called "holdings maintenance," do little to extend the life of the medium, they simply make it possible for the medium to be usable for its full physical life. Yet even extending such minimal measures to all their holdings is clearly beyond the capabilities of most archives. After several years experience, it is dear that the substantial investment that the U.S. National Archives has made in "holdings maintenance," toward its 20 year plan to preserve the records of the Federal government, will prove inadequate.16 In 1988 NARA was able to provide holdings maintenance attention" for 121,700 cubic feet of records at a time when the holdings of the archives were increasing by over 100,000 cubic feet annually.

Even if we were able to extend holdings maintenance treatment to all our holdings, the aggregate cost of storing records in specially constructed facilities, and environmentally controlled, space approaches $25 per square foot per year. The costs of storing the 1 million cubic feet that the state archives wish to preserve at a cost of $500 million will be an additional $250 million per decade. The costs of storing the 100,000 cubic feet accessioned annually by the National Archives, is $2.5 million, a sum greater than the average annual increase to the NARA budget over the past few years. Clearly, even holdings maintenance is unaffordable if permanent retention remains our goal.

Once we shake the commitment to retain materials permanently, we are free to determine which materials in our custody ought to be retained as cultural artifacts, as well as information sources, and which might require preservation only as information sources, without concern for their physicality. The National Archives itself estimates that only .4% of its holdings possess intrinsic value. Thus 99.6% of its holdings should be subject to media transformation, and an equal percentage of new intake should be transformed upon accessioning.17

Recognizing these realities, some pragmatists within the archival community have begun to express the view that since we cannot prevent the eventual disappearance of most of this record, only halt the relative speed of its displacement and decay, we should focus less attention of the conservation of the physical medium, and more on the preservation of its information content. Advocates of this position promote the concept of "preservation microfilming", in which originals are relegated to the ravages of time, but microfilm copies are retained, presumably forever.18 However, the basic text for American archivists on conservation, Mary Lynn Ritzenthaler's SAA Manual, mentions microfilming only once, and only as an alternative to photocopying materials that might be damaged by the kind of handling required to photocopy them.19

Traditionally, conservators have found preservation microfilming alarming because it violates the "rule of reversibility," which states that we should not take actions (finding and destroying originals) that cannot be reversed, and because it is arguably not the "appropriate treatment" in that it takes an action that is not necessary in order to forestall future conservation requirements. I would suggest that both doctrines lead to bad management and should be reconsidered. Preservation microfilming is a contradiction in terms only if we accept the dogmatic definition of preservation offered by the National Conservation Advisory Committee that "preservation is action taken to retard or prevent deterioration or damage in cultural properties by control of their environment and/or treatment of their structure in order to maintain them as nearly as possible in an unchanging state".20

We must begin by accepting the information life of specific recording formats as a fact of physics. While we can influence the production of new media and formats and encourage current information recorders to use formats with longer lives, the "format life" of any given format is the outside boundary beyond which we cannot rationally plan to retain the information without transforming the medium. If the expectation is to keep the information beyond that time, then plans to reformat it should be made when it is acquired, and executed as soon as it is cost-effective to do so, usually immediately. If the expectation is to retain a record for less than its format life, a probable retention period should be estimated and the benefits of reformatting should be calculated. Media should be reformatted if they meet economic tests to do so. Retaining records in the formats in which they were created, even employing the most minimal preservation measures, simply to achieve format life is almost never the most economical solution.

If we were to accept the NAGARA premise that 2.5 billion documents in state archives should be preserved at a cost of $500 million, the job could be done more cheaply, last longer, cost less to maintain, and have greater potential impact on more possible uses, if we microfilmed the entirety and made it available on interconnected Computer-Aided-Retrieval systems nationwide. By William Saffady's calculation, at 580,000 pages per cubic foot, microfilm storage of the contents of all the state archives would require 4310 cubic feet!21 According to the latest price lists from the Mid-Atlantic Preservation Service (MAPS), the costs of copying documents is now $0.15 per page or $375 million for 2.5 billion documents. The state archives are, thus, looking to preserve their existing holdings largely in the form of paper at a cost greater than the cost of microfilming the entire lot, and have convinced themselves to retain holdings in future decades at costs thousands of times greater than the cost of storing preservation microfilm.22

In facing the requirement for preservation of information content, archivists will find several new challenges. One of these challenges will be to consider the nature of the use to which records will be put, and make to decisions about appropriate media transformation based on those findings. Microform is not the only media transformation we should consider. For example, records now of value only for genealogical research might well become important sources for social history research if converted to machine-readable databases.23

Another emerging challenge is that even when we decide to retain records that require machine reading for their format life, whether computer data, images or sound recordings, we will have to cope with the rapid transformation of our modes of producing, recording and storing information. The changing requirements of "playback" threaten even the most immediate aims of the archival profession. The challenge we face in managing the technological change of even as short a time as one hundred years, far exceeds that which confronts, and largely confounds, businessmen and government planners in our age.24 The challenge is to manage a revolution in the nature of information acquisition, storage and transmission more radical than any since the invention of writing. With ongoing maintenance of current electronic information systems and software costing 15% of the acquisition costs of equipment per year, and equipment and software becoming obsolete after 5-7 years, the costs of retaining operating information systems for obsolete data can easily run between 25% and 50% of the cost of the system annually, as long as it is even possible to maintain. Therefore, as a practical matter, no electronic information can, or should, be retained in its original systems environment after that system is migrated to new equipment and software. When the National Research Council studied the question of preserving electronic information five years ago, it recommended that this data should be preserved in Computer-Output-Microfilm because all other media were inherently too unstable and would require such massive investments over time.25 The suggestion provoked an outcry from data archivists who correctly recognized that the greater benefit of manipulability possessed by electronic records would be sacrificed to the demands of preservation.26 The dilemma presents a classic case of trading the preservation of the data for the loss of the ability to use it conveniently (and the loss of its evidential context). If the technologies of computer-input-microfilm could be improved to an extent that it was possible to recover the electronic databases in a reasonable amount of time for manipulation as test sets and destruction when the researcher was through with them, it would warrant further consideration As it is now, the preservation of electronic information systems leaves archivists in the proverbial position between a rock and a hard place.

In addition to the obvious challenges of coping with the new physical formats of information in the electronic age, two more fundamental challenges confront us. First, we will have to confront what may be a culturally intrinsic threat posed by communication across time: we might not, for the very reasons which make our archive so important to us, so defining of our culture and our modes of thought, be able to communicate to persons in the future, or they to understand us. Secondly, we will need to confront the preservation of the evidential context of information creation, a context we have long claimed is essential to understanding the meaning of records, and which is now threatened by changes in our modes of communication.

Because evidential data has meaning only in the context of its use, and because that context is not self-evident for machine-readable data in the way that it is for paper records, which leave behind their "original order" and the evidence of how they were exploited in their active setting, archivists will need to concern themselves in the electronic era with the preservation of system functionality. Indeed, archivists will need to migrate both archival data and functionality along with current records, and make provisions to continue to do so for as long as the data is to be retained. This part of the problem of preservation will be with us even if we decide not to keep anything longer than its natural period of use.

Conclusions: An assessment of the demands of preserving records permanently forces us to discard the notion of permanent retention in favor of a more responsible management formula based on "continuing value".

When we accept continuing value as the basis for retention, we are forced to address whether to retain records in their original formats. In place of the presumption that records should be kept in such formats, we should substitute a serious assessment of format life of different recording media and take actions, such as microfilming and media migration, required to maintain the information content of records in a usable form, for the period and purposes, for which it is required.

In keeping the information content of records, we must become more sensitive than we have been to the preservation of evidential context, and to the preservation of the functionality associated with the use of records. In addition we will need to consider (and here we can wait until we address other issues associated with use) erosion of accessibility to records caused by changes in cultural frameworks and modes of expression over time.

Chapter III. Arrangement and Description


The Problem:

Archivists describe their holdings in order to provide potential users access to them. They also record information about the processing, conservation and use of their holdings to "control" (i.e., manage) them, but traditionally have not considered this administrative information part of the "description" process. Here we will examine both documentation activities as conducted within the U.S archival community.

The sparse archival literature on description backlogs, combined with a disciplinary research literature replete with evidence that existing finding tools and archival information systems are incomplete and inadequate, suggest that the majority of all materials now held by archives in the United States have not been described, or at least not described in such a way as to provide sufficient access to potential users. One hint about the magnitude of the problem is that for thirty years the National Union Catalog of Manuscript Collections (NUCMC) has reported on manuscript holdings of the nation's archives and manuscript repositories, yet has cumulatively reported only about 75,000 collections. Informal surveys of archival repositories reporting to NUCMC suggest that less than 5% of the holdings of those institutions are being reported. Fewer than 10% of all repositories known to the NHPRC Directory report to NUCMC. Within the last five years, less than one hundred members of the Research Libraries Group have entered into the RLIN database more than three times the number of records reported to NUCMC in its lifetime, and they have not yet comprehensively reported the holdings of their institutions.1 Clearly most description remains to be done. But our methods prevent its completion.

The traditional process of describing archives, or subjecting them to "intellectual control", as archivists now often call it, involves two complementary activities: arrangement and description, which take place as steps in the life cycle of records following their accessioning by the archives and preceding their being made available for use.2 The archivist processes the materials to discern their original order (a reflection of the way in which they were created and used), or to impose upon them an order where none exists. The archivist constructs an inventory of the records themselves and then describes the history of the person or body that created the records.

The nature of the problems experienced by archivists in describing their holdings is suggested by the proposals made regularly to the National Historical Publications and Records Commission (NHPRC) to describe new accessions. In these, a repository will often request support for a full-time archivist for a year or two to describe 100 cubic feet of accessioned records. Most archives have substantial backlogs which exist because the description process is so manpower intensive. Estimates made by SAA committees in 1976 were that processing modern collections required 25 manhours per cubic foot, which works out to 18 months for a 100 cubic foot collection.3 One recent study actually found an average rate of 25.2 hrs., with manuscript collections averaging about twice as long per cubic foot as archives.4 An extensive study at the University of Illinois archives over the course of an entire year revealed that personal papers required nearly 10 hours per cubic foot, while corporate records could be processed at a rate of approximately one cubic foot for every 3.5 hrs, and institutional publications fell in between.5 A subsequent study at another university of records considered more corporate than personal, confirmed a measure of about 3.8 hrs. per cubic foot.6

While comparison of processing methodologies is difficult based on the descriptions provided in published reports, we can conclude that an archivist can process between 120 and 500 cubic feet of organizational records per year and between 80 and 200 cubic feet of personal papers. Some part of the manpower requirement derives from the need occasionally to impose a systematic internal arrangement on collections, or record series that lack such a structure. In the University of Illinois study, 58% of the time of professional archivists was spent in arrangement of materials above or to the series level. It is evident that the description process itself consists both of constructing an inventory of the records as arranged, and conducting the research required to augment the inventory with a description of the scope and contents of the collection and a biography or agency history, setting the records into the context of their creation. All this information is then packaged into an inventory/register of the collection, and may be indexed by systems of cards or automated controls.

Breakdowns of these processing costs, in terms of the amount of effort expended in tasks required for subsequent description, are not available in comparable terms, although the Illinois study estimated 1.7 hours was the "time required for producing one page of finding aid or one control card." The expenses associated with such processing times have risen along with salaries since these studies were conducted, but methods of processing appear not to have been modified, so the cost of processing one foot of archival materials may now be estimated as between $60 and $1000.

These costs are not incurred because the materials are being described in too great a detail, because we are too exacting about description, or because we are providing such fine tuned access. On the contrary, traditional archival description emphasizes description of very large collectivities of materials over the content description of the items they contain. Originally such descriptions reflected the "record group," or all the records of a given agency. Although archivists are increasingly focussing on the record series, largely because practice has emphasized the use of internal physical arrangement of the series to provide access to specific items, the practices of archivists still do not provide access to the contents of items, or even to case files or volumes within the larger aggregation.7 So the source of the problem is not overly detailed description; indeed most information scientists and librarians still find archival description too imprecise.

Nor are we sacrificing detailed intellectual control for superb administrative control systems. Most archival holdings are not adequately controlled for administrative purposes, whether for administration of the repository or management of the archival community, by comparison of figures on any activity across repositories. Efforts to compare archival processing costs, to gather information on the extent of archival holdings, or to estimate the nature of access barriers, all founder on the same shoal. Ironically, the extent of our lack of knowledge about archival holdings is hidden by the very inadequacy of our documentation of archival holdings. Efforts over the past decade to rectify this dearth of information by establishing standards for reporting on archival actions, have been largely unsuccessful.8 How can we estimate the extent of the informational inadequacy?

If we assumed that the archival holdings of the nation were under perfect intellectual control today, but that accessions of new materials into archives continued at the present pace, bringing approximately 400,000 cubic feet into custody a year, 4,000 full-time archivists in the U.S. would need to be employed doing description to keep up with current rates of accessioning. We know, of course, both that most material already in custody is inadequately described or not described at all, and that 400,000 cubic feet of annual accessions is not providing the range of documentation we desire. In addition, we know that the less routine the records being described, the greater the effort involved in their description, and that we are particularly short of adequate descriptions of records reflecting policy making at high levels in organizations, which is precisely what the archivist is currently most committed to documenting. And, we know that 4,000 archivists are not available for this job, to say nothing of the 10,000 that might be needed.

Archival Methods:

If describing archives requires more time and investment than archivists can afford, then they must speed up the process of describing holdings by factors exceeding what can reasonably be achieved by improving current methods. Identifying new methods will require that we be clearer about what information we require in archival information systems, why it is needed, and therefore how it should be structured. The proposed archival description methods would involve a strategy for description of records which rests on a single axiom accompanied by three theoretical premises, grounded by a unifying tactic. The axiom is that archivists describe the context out of which records were created, rather than describing the content of the records themselves. It should be understood however that this axiom is valid because the description of that context is the most powerful proxy for the content description of records, and that should it fail to achieve the envisioned goal, the axiom would need to be revised. The premises are that the most practical means of achieving intellectual control are top-down definition of holdings, description of provenance, and exploitation of the lifecycle of records in the description process. These premises should be realized in practical description approaches in inverse order from that stated above: lifecycle records systems control should drive provenance-based description and link to top-down definitions of holdings. The tactic is that archivists should find, not make, the information in their descriptive systems.

These theoretical premises are not new, but they have been poorly realized in practice. Top-down description, which requires that the record series as a whole should be described before its component parts, including specific accessions to it or items within it, has often resulted in rudimentary accession reports serving for many years as the sole source of information about what an institution has acquired. The records of public institutions are most often the subject of what are called "preliminary inventories," but which are, for all practical purposes, the only controls that are likely to be produced in most cases. Nevertheless, the principle of top-down description does by and large govern practice, and defines a solution to description which is, by definition, affordable and practicable. In effect, it states that a brief description of the whole collection is mandatory, but any further definition of its contents will be undertakingy insofar as the repository can afford it. The consequences are, of course, that a very superficial level of information about holdings of archives is available.

Top-down description has often been taken for a necessary evil, imposed by lack of resources, rather than as a positive strategy. As a result, Systems to exploit archival information often treat all access points to records as a single file whether the system is a card catalog with access points filed alphabetically on cards describing thousands of cubic feet, interfiled with the same vocabulary linked to cards describing individual items, or a single automated database with the same failing. As such, systems have failed to take advantage of the inheritance of characteristics of the parts from the whole and of the whole from the parts. The first failure results in redundancy of each lower level record, which must carry information about the collection of which it is a part. The second failure results in dated descriptions at the higher level, because discovery of detail at the lower levels, such as the dates of items within the collection, or additional forms of material, do not "ripple up" the higher level description. Because the description of records themselves is ~e description of a physical nclusion relationship, such inheritance of physical properties can and should be built into archival information systems.

The second principle of archival description, that the context out of which records arose is the actual object of the descriptive record, because provenance is the heart of archival description, is observed in ways that are so formalized as to be nearly meaningless. Provenance has been equated in American archival practice to "agency of origin," and assignment of records to the proper record group has been a proxy for description of the context out of which they arose. The sole point of access provided by provenance when it focuses on agency of origin is a corporate name. Only recently has any further attention been paid to the function that generated the records, and no satisfactory structure for archival information systems yet supports separate databases for organizations and their activities.9 Unless we can come to see that description of information creating contexts is utterly independent of the description of physical records, that such contexts exist whether or not we acquire their records, and that information about such contexts, including how and why they create records, does not derive from the records themselves, we will not design appropriate ways of acquiring or reflecting information about provenance.

Archivists have traditionally focussed their descriptive activity on the records, creating finding tools about groups of records. Though the professional archivist may actually devote much of his energies to writing administrative history and provenance statements, and devote less time to listing the contents of the holdings, the holdings remain the organizing core of the information system and the inventory comprises the bulk of the description. While traditional archivists recognize that it is a great advantage to obtain an index to records together with the records themselves as part of an archival accession, even aggressive records managers do not set up systems that produce such indexes or finding tools as a matter of course, and no concrete guidance is provided in archival manuals about how to integrate such externally produced finding tools into archival descriptions. Even when indexes are obtained with records, archivists tend to make their own finding tools, cognizant of the fact that the context in which information is found is crucial to its understanding.10

The third principle, life-cycle management, is paid considerable lip service in the profession, but the absence of integrated systems serves as testimony to its impotence. Although archivists have been having a flirtation with life-cycle information systems, the limitation of their present vision is best demonstrated by the history of the National Archives (NARA) in the past decade.11 An institution which once controlled records management, but then gave it up, has now embraced life-cycle management, but only of its own holdings. By failing to arrange for the systematic acquisition of information about information systems, manual or automated, which generate the archival record, the agency must generate such information in order to describe the records it acquires. The task is greater than can be achieved given the resources of the organization, so the records acquired will, by and large, never receive this level of description. NARA has issued standards for all its information systems to assure that data collected anywhere in the life-cycle will be consistent, but they are proceeding apace with the design of new incompatible control systems, such as the GAPS and R/ACE databases.12

Information about physical features ought also not be captured in an isolated process called archival description, but rather through the operation of an on-going information collection system based on actions.13 The actions taken by archivists and curators to their holdings, unlike those taken by libraries, have an impact upon the meaning of the records as evidence. Weeding collections, publishing about them or from them, exhibiting items from a collection in conjunction (or juxtaposition) to other materials, all effect our uture understanding and interpretation of the materials. Therefore we need to keep records of archival actions not just for the administrative purposes of tracking work and improving our methods (though these should be an adequate justification), but for the reason that the intellectual content of records is affected by processing.

The vocabulary of actions consists of a quite simple dialogue - who, when, why, under what authority, where, and with what result. These statements, repeated for every action and step of every action, constitute a disaggregated transaction history of the collection. Based on analysis of these transactions, we can establish work patterns and costs, missing authorizations, needs for procedures, and many other management requirements. Based on viewing the latest status (with what result) values, we can report for users on the ownership, location, condition, and other salient characteristics of holdings. After all, "owned" (borrowed, on approval, etc.) or stacks" (circulating, in conservation, etc.) are statuses of prior actions; subsequent actions may result in new statuses - "sold" or "on exhibit". Ideally, the user of an archival information system would see the latest status reports on many actions, but would not be privy to the detail in each transaction record. Researchers might see the entire history of certain kinds of actions (such as publish, exhibit, transfer ownership) which have bearing on the interpretation or significance of the materials.

Yet the fundamental transformation of archival description practices will come only if we employ new tactics in addition to strategy and principles. Specifically, we should embrace tactics that use description and data created by others, rather than requiring archivists to create the descriptions of archival holdings. That is, archival description and control systems must be designed from the first to capture information from external sources either before materials are acquired or when they are accessioned, rather than to depend on information provided by archivists in the years following acquisition, through laborious analysis of the records in hand. Archivists need to plan now to take advantage of the rich information sources that already exist outside archives, and which contain the data archivists need to describe all aspects of their holdings, except for the most trivial physical characteristics such as the actual volume of materials accessioned and their condition.

Several years ago I proposed to the National Archives of the United States that it should be capturing information about documentable activity from its own publication, the Federal Register, which is a daily accounting of all the changes in organizational structure, all the publicly announced meetings and hearings, and all the rules and regulations of the Federal government. The Federal Register contains all the information required to describe every program of the Federal government, to document its authorities, and to identify its functions, activities and public interface. In short, it is an index to every activity about which the Federal government is obliged to inform its citizens today. As such it is the record of organizational structure and function, and the record of activity. Obviously it serves as a potential index to all future records of the government as well. Incredibly, though this record is constructed by the National Archives, and though its archival description staff uses the Federal Register as a primary source of agency history information when, forty years after the fact they try to describe records accessioned into the archives, NARA produces the document as an unstructured electronic text file and does not retain it as an index to future Federal records.14

Other organizations have similar sources of external information. The AT&T archives and records program has established relations with the corporate public relations function to automatically bring full text of issuance's from that office into its system as a finding tool as well as an archival holding. Personnel systems with organizational structure information, central directives systems documenting policies, and legislative information systems documenting laws and statutes establishing new offices and assigning functions, can all serve a similar purpose.

One of the major challenges facing archivists in the next decade will be to control the information in office automation systems so as to properly dispose of transient information and identify records for archival retention. Previous studies have shown that the only viable framework for making appraisal decisions on electronic records is the thorough documentation of systems functions; this same information is no less essential to assure meaningful access to the records that are retained.15 Without information reflecting the way in which the data was created, accessed and used in the operating context, one randomly organized bit-stream, or even sequentially organized data file, looks very much like another.

Higher level standards are beginning to make possible the interoperability of varieties of office equipment and even office systems.16 These standards depend on data being transmitted from one system or peripheral carrying self-defining information to the target system or peripheral. For example, Standard Generalized Markup Language permits any printing system to print the same document image and to change instructions for the type face or size of any given document component in a single command, because the printer markups in the document, identifying the components of text are logical rather than physical. Similarly, the x.400 electronic mail envelope carries information about who created a document, where it is intended to be sent, and how it is to be acted upon when it is received, in a header that is independent of the document itself. As a result, the same document can be transmitted to a variety of stations on different electronic mail networks controlled by different software, each with different action instructions that will be properly interpreted by the receiving system.

These self-referential data packet descriptions, and higher level self-referential descriptions of database environments and systems functionality, as envisioned by the ISO 8660 standard, are ready made for importation into archival reference Systems. Systems that can employ such external descriptors will be able to avoid having to describe archival holdings after they are accessioned, and will be able to provide much more detailed, indeed item level, access to electronic information. More importantly perhaps, failure to capture and use this information will render the task of appraisal impossible.

Archivists will face the same challenges that have confronted the data managers and configuration managers of automated information systems when they attempt to control information in electronic form. In the process, they will do well to turn to tools used by those disciplines, particularly to the Information Resource Directory System and data dictionaries, to document systems. What they will find may prove a surprise: archival description systems have been meta-data systems, systems of information describing information systems. And when they have worked well, archival meta-data systems, like those of information resource managers, have described record systems and the context of the activities those systems support.

To keep the focus on the question of methods, rather than data structures, requires that we draw the proper conclusions from the preceding. Just as archivists need to look outside the repository for information that already exists about the context of records and utilize the self-referential descriptions of electronic records, they need to look inside their own organizations to build information systems that execute and document procedures taken on holdings, so that description of such actions is automatically appended to the description of holdings over time. Such systems ideally would also incorporate opportunities for users of records to report on what they find in holdings, and by describing what they find to provide future users with access points not previously available. Whenever staff have an opportunity to examine a body of records, whether to answer a reference request or take a conservation action, they too would be enabled to add information to the description.

The methodological innovation proposed here is to develop systems that enable descriptions of records to grow dynamically from their entire history of creation and use. Until archivists adopt such an approach, the backlogs of materials awaiting archival description will grow, making a mockery of the mission of the archives to provide access to the records of the past.


Archivists cannot adequately describe what they currently hold or will acquire if they continue to employ current methods based on examination of their holdings, even if they rely on only the highest level of top down description. To increase the effectiveness of description and control by the necessary order of magnitude, archivists will need to identify that information which can be obtained from outside, and import it into their systems automatically. This information will provide access by provenance, based on the nature of the activities documented, and by the structure of information systems. They will also need to design systems that capture administrative data throughout the life-cycle of records, both before and after they come into archival custody, and use that information in the on-going control and management of the records.

Finally, they will need to design means to capture what their users learn about records, and the knowledge they bring to the repository about the context of records creation, so that this information can enrich the description of holdings for subsequent users.

Chapter IV: Access and Use


The Problem:

Most potential users of archives don't. Even though the influential report of the GAP Task Force, Planning for the Archival Profession, acknowledged that use is the ultimate purpose of archival work, Paul Conway was forced to note in his discussion of the 1985 Census of Archival Repositories, that archivists know very little about their users, and seem not to care.1 In 1989, the National Coordinating Committee for the Promotion of History, the major lobbying group that has supported the National Archives, felt compelled to issue a report. calling upon NARA to recognize that "in the final analysis it is the use of records that makes archives valuable" and to point out that NARA has not studied its users at all since 1976, and now doesn't know "who the users are, why they visit the archives, how they approach the records, what research strategies they employ, how successful they are in retrieving desired information, and how they use this material."2

What little we do know tells us that on a per user basis, it is expensive to run an archives and that we are reaching only a small portion of the population. The median number of users visiting archival repositories responding to a recent SAA survey was 325 per year, or barely one a day. Large institutions, of course, had larger user populations. The National Archives was visited by over 200,000 users in 1988; an additional 500,000 letters and phone calls about the holdings were answered by staff. At a median expense to the National Archives of $145 per research visit, fewer than one in four citizens will visit, write or phone the National Archives in their lifetimes.

Those who do use archives are not the users we prefer. As Elsie Freeman puts it, "we have what can most kindly be called an adversarial relationship with genealogists, one of our largest clienteles, and with other avocationists."3 This relationship is reflected even in the SAA manual Reference & Access, which states that "public repositories will want to be as generous as possible in their access policies, welcoming the general public (including genealogists and local history buffs, undergraduate students, and bottle collectors) as well as the academic researcher." It is hard to miss the derogative connotation of juxtaposing buffs and bottle collectors, and the suggestion (however unintentional) of the opposite of these literal words of welcome.4

We would expect government archivists to consider the citizen to be the ultimate beneficiary of record keeping, since an accountable government is one that can assure the citizen of his rights. The 15 million filers of tax returns, tens of millions of social security beneficiaries, veterans, and other individuals who annually have contacts with government agencies, are all potential users. Yet when in the spring of 1988, the National Archives began receiving 1000 requests a week from Japanese Americans seeking to verify their internment during World War II, in order to receive the compensation authorized by Congress, it turned the process over to the Justice Department rather than assume the burden of answering citizens requests. Instead of reinforcing in their minds the value of the Archives as a repository of citizens rights information, NARA chose not to serve its primary users.5 Perhaps like William Joyce, the Archives felt that it was more important to get to know their holdings in order to serve users better, rather than studying their users in order to serve them better.6

Our preferred user, the scholar, does not come in great enough numbers to justify archives,7 and does not use archival resources very heavily when he comes. Frederic Miller, looking at social historians, found that on an average they used about 2.5 record series per article.8 And Miller's average was more than Jacqueline Goggin or Clark Elliott found in similar studies of women's history and the history of science.9 Even academic users find out about our holdings by word of mouth and other informal means, rather than through published guides and national bibliographies, even after two decades of reporting to the National Union Catalog of Manuscript Collections.10 One of the major questions put to the participants in the recent grant made by the National Historical Publication & Records Commission to the Research Libraries Group was to demonstrate whether research use could justify reporting to the national bibliographic networks.

Perhaps it is not surprising that we have so few users. As a matter of principle, archives do not acquire materials because our users might want them. The criteria we use, even when considering whether to keep records once accessioned and long unused, are disdainful of use.11 Archivists will not usually accession records that are still in active or semi-active use, even when they have been scheduled for archival retention and conditions of care in the archives would be better than those in the office of origin. Records recalled too frequently from records centers are seen as evidence of poor management decisions by records managers rather than as opportunities to raise the profile of the archives.

When users do come to the repository, we have strict rules to keep them from seeing too much at a time or using it in a place that might be more convenient for them than the archives reading room. Most of the basic manual on reference and access consists of guidelines for rules: rules about security, use of restricted materials, reproduction, loan, copying by researchers, taping by researchers, copyrights. We almost never lend materials, even though few of our holdings have "intrinsic value." Although we have long talked of exit interviews, we still conduct only the initial interview, and none of the excellent considerations in Paul Conway's article "Facts and Frameworks: An Approach to Studying the Users of Archives," which focuses on how archivists can learn from their users, not only who the users are, but what the archives holds, how it serves its users, and how it can continue to serve them after their research is completed, are reflected in practice.12

Ten years after Richard Lytle noted that we could not design sensible national information systems (to say nothing of local information systems) without such studies, we still have none.13 Archivists consider it necessary and reasonable that "the archivist is necessary, even indispensable, for subject retrieval."14 We expect our users to come to us fully prepared by research in all secondary sources, to be able to associate names of persons with subjects of interest15 and to understand the nature of provenance.16 Yet we must confess that our finding aids are inadequate even for us to retrieve what the user wants.

Discouraged by finding that simply preparing better finding aids does not attract academic researchers, archivists have not yet succeeded in developing public programs to attract more users, as have most other cultural institutions. When the New York City Department of Cultural Affairs recently reviewed the outreach programs of city funded cultural institutions, it included zoos, theaters, museums and botanical gardens, but not archives.17 Could it be that no New York City archival repositories are sponsoring any outreach programs innovative enough to be mentioned? Whether or not it is true, the impression is dangerous, but it is little wonder when we can search in vain through the past two decades of the American Archivist for a description of a single outreach program in an archives anywhere in the nation.

If the average citizen does not use archives from one decade to the next, it is no wonder they are not better appreciated. If we extrapolate the findings of the SAA survey to 5000 repositories (recognizing of course what an exaggeration of numbers this produces, since institutions which did not respond are likely to be smaller and less visited), archives might experience seven million visits per year. A given citizen might visit an archive once each generation; to achieve a minimum outreach of one use per citizen per year, the archives in this country would have to increase their current level of use by well over an order of magnitude.

Archival Methods:

Archives are caught in a dilemma. They are, on the one hand, clearly a housekeeping function of the organization of which they are a part, and are beholden to that organization for efficient management of an information resource, and on the other hand they have pretensions of serving a broader cultural master, one who rarely pays the bills and is stingy in its praise. Archivists have adopted a strategy of urging their importance within their own organizations based on the value to the greater society of their cultural mission. They have succeeded in convincing resource allocates that they are well meaning, largely harmless, and largely irrelevant functions that should be permitted to continue to exist as long as they do not demand too great a resource.18 With few resources however, heavy use becomes a burden. Larger numbers of external users would not justify a greater resource allocation, so the preference of the archivist is for a "quality" user, perhaps a "serious scholar", and especially users who do not demand too much reference assistance, too quick a response, or to specific a piece of information, all of which require greater investment by the archivist.

In 1983, I labeled this pathological strategy and its resultant condition "the poverty of archives."19 I continue to believe it is an attitude that accepts the tapeworm as inevitable. In its place we need to develop methods that make archives indispensable to the institutions in which they are located, thereby creating a vocal internal constituency requiring our services, and which make use the measure of success and the ultimate (though not necessarily short-term) criteria for record retention, thereby providing archivists with powerful incentives to find users and sell services.

The ingredients for achieving the first result, making archives indispensable to the institutions they serve, are found throughout these essays. They involve using enforceable institutional policy to require program managers to retain documentation of functions identified by archivists as critical, and managing the establishment of manual and automated information systems to assure the capacity to report archival records as they are created, and to assure the retention or destruction of records as appropriate.20 Creating an internal constituency necessitates archivists maintaining, as information systems and ndexes to potential archival holdings, knowledge of the institution's history, its functions and activities and their assignments, from the present day back to the distant past, through the construction of independent reference databases serving as archival authorities.21 Fundamentally, a new relationship to internal constituencies depends on archivists taking control of records of archival value and of information about the context of creation of such records, from the time of their creation or earlier. The control that archivists exercise over such records need not be physical, it is only critical that the administrative authority of the archives over such records be recognized and that the creators of such records accept responsibility for their archival retention and for reporting on them to the archives.

When the archivist focuses on control of all archival information in the organization from the moment of its creation, and uses the importation (of data from elsewhere in the organization to build reference databases on the fly in which the organization functions, its sources of authority, and its vision of history, he will be in a position to serve as information broker for the organization. From this position of power, the archivist should focus ~ vesting himself of responsibility for the day to day care of the physical record. 'his divestiture is increasingly important in the electronic age, where records will be uneconomical to manage except in the distributed settings from which they arose, until the implementation of true interoperability standards for all electronic information systems applications.

Internally, archives should be the interface to all corporate information. The AT&T archives mentioned earlier is one of a handful of menu choices available to staff of AT&T worldwide when they log into their terminals daily. Corporate news (press releases), organization charts, and policy directives, are available through the archives menu choice. These are obviously records of archival value, needed on a daily basis by users throughout the organization. The archives has recognized that it can supply information, and that if it does, its importance to the organization will be apparent. In some ways this is the flip side of the earlier discussion about the Federal Register and the National Archives. Not only might NARA recognize the value of the Federal Register as a potential index to functions and activities that must be documented (and thus an index to the future archives), but it could use the Federal Register to make itself a presence in every agency in the Federal Government, by providing the best online access service to information from the Register. Instead it falls to the private sector to provide the Register online to government offices!

The ingredients for instituting use as the measure of programmatic success and the criteria for record retention are also found throughout these essays. Continuing value looks to use for justification of retention. It will result in considering such highly used series of records as birth, death and marriage certificates as archival, thus assuring heavier use of archival records by the public. Appraisal based on activity looks at functions that had a direct effect on potential users, especially on their rights as citizens in a governmental archives. Decisions based on appraisal of records by functions with substantial potential impact on constituents will result in saving and servicing records that are particularly needed. Description that incorporates user input over time will enrich future access, especially with the elusive perspectives that develop from changes in cultural perceptions of the value and meaning of records. Such description methods will empower users over time to shape the cultural record through their use of it, establishing a dynamic relationship between the user and the repository.

There are direct methodological implications of a focus on building and maintaining a constituency as a strategy for becoming and remaining a service organization in a competitive culture. For example, archives are unique among cultural institutions for their scant membership and development activities. They should actively establish and nurture a community of supporters, providing them the usual newsletters and special events, in return for the equally typical financial and political support. Archives should target this constituency to serve in expressing support for the archives, whether to state legislators or corporate headquarters.

Making users more aware of archives is not simply a desideratum, but a necessity. If the electronic record of tomorrow is going to survive long beyond its creation, users will need to become more conscious of archives and what it means to make an appraisal decision. Managers will need to become as conscious of their responsibility for accountability for actions, as of their responsibility for the actions themselves. Archivists need to focus as a profession on mechanisms for identifying the archival record that have an impact on the entire society, that will fundamentally effect our cultural repertoire of documentation. Just as we once taught students to write business letters with "Dear Sirs" as a salutation, we will need to educate an entire society to address electronic mail envelopes with header information that will be critical both to transmission and to ultimate retention.

And we need to learn to be more responsive to users, treating their requests for information, especially in the public sphere, as the exercise of rights, and filling them promptly and accurately. We need not equate this with taking on another burden. The costs involved in filling citizens requests for information required by another agency should be charged to that agency by the archives. The archives needs to adopt the position that it is indeed the ultimate arbiter of what must be retained, but it does not for that reason become responsible for upkeep any more than the auditor becomes responsible for misappropriation of fluids by identifying where it has taken place.

We will face increasing challenges from users to exploit the information that archives hold, rather than to retrieve the records that contain the information. While the federal government of the United States and the states have not yet resolved the question of whether they must create a record that previously did not exist in order to satisfy a Freedom of Information Act request for data from an electronic data base,22 the Canadian government pointed the way with its decision that agencies are obliged to provide information in the form requested if they can create it.23 We will be faced with these requests both for the physical form of records24 and for their intellectual content.

For both internal and external users, archives must begin to make serious study of user presentation language. Although we acknowledge that we do not know how users ask for materials when they do come to the archives, I believe that I conducted the only study yet of user presentation language in the spring of 1989. While the results are not yet tabulated, it seems likely that they will change the assumptions on which we have based archival finding aid systems. The strategies for access that are required for the next generation of archival records and the next generation of archival users require that we develop "intelligent artifices," or better understandings of the structures of language used for intellectual access to archives. The first step in the design of such systems is to understand fully the mental images our users bring to archives and the ways in which they articulate their requirements. Success in this undertaking does not assume knowledge of artificial intelligence, a desire to design semantic parsing routines or a capability for implementing gateways to the world's knowledge bases; it only requires that we listen carefully to what our users actually say they want, and build access languages based on what we hear, so that users can begin to access our holdings without archivists as intermediaries. In this way the archives becomes free to deliver its resources to remote users, to extend its "service catchment area" as health and welfare bureaucrats might say, and thereby to enlarge its potential clientele.

Information, it has often been noted, is the only resource that is increased by use. Archivists also need to capture the ways in which information is used, much as a hypertext system captures new pathways defined by users, so that connections can be made, illuminated, and reinforced between sources of information relevant to a wide variety of user queries. Developing methods to de-brief users, of which the most promising is to provide users with access to computers and software facilities that enable them to take notes, then manage and analyze the notes, and use the information thereby contributed as part of the description of records available to subsequent users, should be a research and development objective for the profession.

The aim is to make the archives without holdings become also an archives without walls. Freed of the primary custodial responsibility (though still often the custodian of last resort for records whose creating office can no longer manage them), the archivist adds value to the information acquired about records, from information sources regarding activities and functions and from knowledge of user presentation language, to make the information available to anyone, anywhere. The archivist who has reformatted what holdings he does have, taking into account the formats that best support use, can make holdings available by loan to the user wherever he is. The resultant expanded constituency may be exploited both to draw in further users through word of mouth, and to support the archives financially, politically, and intellectually.


Archives do not receive sufficient use to justify their expense, and archivists have not succeeded in convincing resource allocators to increase the funding for their programs in order to support an abstract cultural benefit. Because archivists do not acquire materials that are in heavy demand on an ongoing basis, and do not acquire any materials just because they might be used, the lack of use is not surprising. Archives don't advertise their services, rarely engage in systematic efforts to build and nurture a constituency, and do not aggressively support their primary clienteles. The result is that few Americans visit archives or know what they are. Most cultural institutions perform these services for their clienteles, and archives could as well. Archives could and should actively seek control over records that demand heavy use or are critical to potential users, in order to increase their clientele. They should seek to become centers for information about records they do not hold, and to provide users the opportunity to borrow reformatted materials as needed. And archives should seek to be recognized in organizational policy as the source of information about what functions and activities are currently being performed by which offices, and about what information systems exist in the organization and their contents.

To play this role archivists will have to acquire control over records earlier in their life cycle, collect all available information about functions and activities of their parent organization, and promote their institutions as sources of information rather than of records. When they do, they will face the difficulties of organizing archival knowledge so that external sources can be used in its construction (Chapter V), conveying meaning over time (Chapter VI), and dealing with an increasingly electronic future.

Chapter V: Intelligent Artifices: Structures for Intellectual Control


The new methods suggested thus far in this study as substitutes for our current approaches will not, by and large, require any fundamental new theory or new technologies.1 In fact, they are essentially conservative, often literal, about what archivists have long claimed makes them distinctive. The glossary definition of archives published by the SAA in 1974 declared that archives kept records of "continuing value." We have since simply neglected the corollary that archives dispose of records which cease to be of continuing value. The notion of permanence has fewer adherents daily, though we have yet to take seriously enough the limitations of potential future resources. The need for archivists to build descriptions from lifecycle management data, and to look to self-defining information packet standards, are both recognized, and although no concrete steps have been taken on a large scale to import information from external sources for authority reference files of agency histories and biographies, these approaches are being discussed. Following through will require tremendous effort on the part of the profession, and is likely to require specialized archival education far beyond what has been offered in the United States to date, but it does not involve fundamental research.

Other challenges faced by the profession will require more study. Minding the past tomorrow will be much more difficult that it is today. In the relatively near future the greater part of the archive of our society will consist of vast, machine readable, databases consisting of randomly stored and indiscriminately collated, primary and secondary, published and unpublished materials. How these archives will be stored and how we will retrieve from them is, as yet, unclear, but it is evident that wherever in organizations we hold them and whatever the technology of their storage, the immense size of the potential virtual database, its diverse authorial sources, and the range of its potential uses will present tremendous challenges to those who seek to provide access to it.

The problems of access to such materials cannot be solved simply by throwing huge amounts of raw computing power at them although some have suggested this, knowing that it will be extremely cheap by today's standards.2 Nor is it sufficient to assume that the problem will be resolved by the emerging implementation of expert systems and other applications of artificial intelligence. These technical developments will be employed in the solution we devise but they are not, in themselves, the answer. We still understand too little about the nature of archives and the fundamental intellectual constructs of which their description consists, and we are too ignorant of the dimensions of search strategies which would be relevant to retrieval from such a storehouse of information to instruct an artificially intelligent system in becoming an expert.

Therefore, we need to construct information systems that describe activities and other information systems, and which can be constructed of data that is accessible from other sources. To do so will require that we analyze the way in which we represent reality in language, which is the basis for both description and access.

Knowledge, Language and the Structure of Archival Information Systems:

Persons (or other subjects) execute functions (organized activities) with respect to objects, in a particular place and time and in a manner to which we attribute a form and a content. Or so our culture inclines us to believe. In our culture the fundamental dimensions of reality are space, time, subject, action, object, form and function.

The dominance of one or another of these dimensions in any particular construct of the world is a reflection of perspective of the user. User queries are intended to retrieve those representations of realities which have something in common along one or more selected dimensions. Thus, a biologist may be interested in an activity of an object in a place - reproduction behavior of gulls in Terra Del Fuego. The activities in which he is interested, biological behaviors, are controlled by a language special to his discipline. The objects in which he is interested, species, are likewise governed by a complex, hierarchical language, called taxonomy. And the place in which he is interested is a geo-morphological construct. An art historian interested in the dimensions of time, agents 'and form, say the inlay work characteristic of 15th century silversmiths, could formulate a query which used none of the dimensions of interest to the biologist. The category of agents employed in this query is a social construct, occupation, which is by no means the only social categorization available. Consider, for example, a social scientist, who wishes to approach evidence by exploring the relations between function and agents, political martyrs or teenage parenting. Instead of classifying agents by occupation, this researcher sees agents as belonging to different social constructs, in this case an age cohort and a political affiliation.

It is important that we be aware that the principal intellectual approaches currently supported by cultural repositories to their holdings are structured around object themselves. When we examine user queries however, we find that the object is usually what the user wants to find, not what the user already knows about and asks for. The object in the archive, museum or library is not significant to the researcher in itself, but only as evidence for something else. This something else is the actual subject of research. Our users query objects directly only when, because of prior experience and training on their part in the limitations of the access approaches which we support, they resort to using the object as a proxy for a particular time and place or a specific function and agent.

Rather than enumerating the weaknesses of our existing systems of access and control, we must focus on the potential utility of understanding that the number of dimensions of reality are limited, but that the number of views of each dimension are myriad. This is not simply a reflection of the nuances of meanings of concepts but rather of the discrete vocabularies, and views, required by particular disciplines. Practically speaking, this means that if we design information systems which are given, or can identify, the relevant dimensions being used by a search, we can exploit this information as a navigation device to identify the information resources, such as disciplinary vocabularies and indexes, suggested by the search itself.

Therefore, we must explore how to identify each of these seven dimensions (space, time, subject, action, object, form, function) in a query, so as to use that knowledge to map the relationships between concepts identified by different disciplinary views within each dimension. This strategy would, in principle, make it possible to cross between disciplinary perspectives within a single dimension and retrieve information which was originally recorded by persons with intellectual perspectives different from those of the immediate researcher. For the purposes of this chapter, because I have discussed it elsewhere, I will not dwell here on debates about data values properly associated with particular elements, but only on the structural properties of human thought and language.3

Over the past several years, the proliferation of on-line databases and machine readable information sources has made information scientists painfully aware that the problem of intellectual transportation across disciplinary perspectives is not resolved by making data available on-line or in full-text. Indeed it may be exacerbated, since in manual retrieval systems the human mind makes leaps across categories which are not supported by existing mechanisms in automated systems. As a practical matter, if we are to integrate a variety of externally developed databases into archival information systems in order to provide for retrieval without much in-house description of records, we need to determine how we can best make large machine readable data stores, consisting of a variety of sources, each collated for particular purposes and audiences, accessible through a single user interface.4 In this case the user interface is not a mere technical presentation facility, but a meat-language (a language about languages or vocabularies) by which the appropriate intellectual contents of discrete information resources are aligned.

It is theoretically, if not technically possible, if two databases support the same access approach, to bridge a query into each at the same time in order to make a combined retrieval. If these two databases also use different data elements beyond those which support the common approach, they will permit iterative hopping from the results of one combination, to other databases with commonalities along these different dimensions. For instance, if we find that two databases support a search by time, and retrieve records linked on this dimension, but the second database also supports corporate actors which are missing in the first, we can use corporate actors as a bridge into a third database which does not allow for selections on the basis of chronology. Naturally since these databases will use different vocabularies within the same approach, retrieval recall may be extremely, and perhaps unacceptably, poor. But in principle the way in which such a search could be conducted is clear, and it suggests a strategy we should pursue.

First, we will need to analyze what approaches each separate database which might be incorporated as an information resource supports in its native implementation. Some databases may be accessed by chronological terms (dates or periods), others by personal names and yet others by geographical terms (political units or morphological entities). Second, we need to identify the vocabulary employed in each facet within the specific implementation. In each case the particular vocabularies supported for each dimension may be the same, may overlap, or may be discrete. If discrete, they may be related to each other in definable ways or not. Mapping each disciplinary language to a internal usage of the archival system will prove both difficult and limiting, since relations between languages will not be defined. Instead, we need to explore the domain of each approach to determine if some neutral, non-verbal, ground for conceptual mapping can be identified.

Let us consider two databases which provide access along the same approach but use different vocabularies entirely. Here the mapping of different vocabularies along a common dimension is conceivable. It is theoretically possible, for instance, to map all time along a linear dimension. Mapping a variety of religious-calendric time, with politico-calendric time, for example is relatively simple. The year 515 B.C. in our Christian calendar, and the sixth year of the reign of Darius, correspond to the year of the completion of the reconstruction of the temple in Jerusalem (used as a starting point of some Old Testament timekeeping) and the 36th Zoroastrian year. Mapping chrono-political, but non-calendric time is not much more difficult, since the time of any given reign or war can be expressed in calendric time, as can some chrono-ecological time concepts, such as the time of a particular plague. But many chronological concepts, especially chrono-cultural terms such as "the Renaissance," are not sufficiently definitive to be mapped, not simply because the idea is fuzzy, or because particular scholars would disagree about dates, but because they apply to a variety of human endeavors taking place throughout Europe, each of which and each location of which, has its own criteria and its own dating. Other vocabularies, such as chrono-morphological terms (Pleistocene epoch of the Quaternary period of the Cenozoic era) are on such a different scale from human time that their boundaries alone are wider than all of recorded history, making it impossible to map them against cultural time.

We confront a similar problem with geo-terminology, in which we can state that in principle one can map all of space along four dimensions, only to find that many geo-vocabularies defy explicit mapping, or are of a scale which so dwarfs other geo-concepts as to make their integration unwieldy. Geo-forestation or ecological regions are clearly more relevant to the location of biological species than are geo-political concepts. Geo-morphological vocabularies are likewise more useful to understand the origin of minerals. The fourth dimension, time, is always present, whether explicit or implicit, in geo- concepts because the face of the earth and the structure of the universe are being continually transformed. Still, even when both time and space are used to locate a particular empire or nation (a geo-political construct) or the land occupied by a tribe or migration (as in a geo-ethnological vocabulary) features which are named in county gazetteers rest uncomfortably beside the name of the planet earth, and the tundra is less clearly delimited that the boundaries of Ann Arbor, Michigan.

The possibilities of mapping time and space to neutral dimensions beyond language constraints should be exciting for archivists because they suggest new ways of presenting archival resources to users. Unfortunately, the difficulties of creating such common maps for persons and organizations have in the past constrained archival retrieval by creation context. The need for a more universal access vocabulary is nowhere more evident than in a comparison of the successful use of names as an access point into the published literature (where the authors and the famous alone are named), to the failure of names to provide easy access to the public records of the twentieth century, where every citizen is reflected in at least some public record at nearly every level of government. The proliferation of personal (and organizational) names is so complete in contemporary archival records that the databases cannot be initially explored using this means of access. Only "prepared" researchers, knowledgeable about the specific local institutional landscape and potential actors, can now access archival records by creator. However, if other databases containing names could be accessed online, researchers might identify logical groups, by demographic and similar characteristics, and use the results of their searches as means of access into the archives, rather than having to conduct external research for names of individuals and look them up in archival finding aids, hoping for a lucky convergence.

It is equally fruitless to attempt to identify potential organizations as agents of actions and to use their names as access points to archival holdings, since the number of "legal persons in our modern corporate society is not much less overwhelming than the number of individuals. when we add to the number of corporate entities the virtually limitless number of sub-divisions and departments, working groups and committees which generate records within such organizations, names alone become a very imprecise approach. As a consequence of the limited usefulness of names and the difficulties of mapping disciplinary specific vocabularies for place and time, I would argue that archivists must explore the potential value of form and function as means of accessing the archival record of our society. Since the culture of which we are a part, indeed any culture, creates only a relatively limited number of cultural forms of expression, and recognizes a relatively small number of functions, the undertaking seems possible.

Cultures are in effect, distinctive modes of behavior supported by a shared language. The ways in which organizations may act are constrained by the culture, which assigns social functions to particular agents at various times. Marriages will be performed by the state and religious authorities, and even ship captains in our society, but will take a distinctive form regardless of agent. The licensing function establishes criteria for eligibility, evaluates applicants and grants and records licenses in our society, whether licensing nurses or exterminators, restaurants or drivers. The analysis cannot depend on verbal clues alone however, because we speak of the licensing of motor vehicles and of dogs, which are actually the function of taxing.

Despite these sorts of hazards, function has many advantages as a tool for archival access. As a society becomes more complex, functions once assigned to somewhat undifferentiated agents are assigned to agents who specialize in only the individual function. At one time in the history of the western United States, the sheriff performed virtually every function of government. Subsequently some of these were assigned to a county clerk. Now each office within a typical township or municipality will perform only a small set of unctions, which may in larger cities be even more finely divided. Also functions do not remain with the same offices of government forever; archivists are aware that the archival function in the Federal Government was once independent, then under the authority of the General Services Administration, and has recently become independent again. Because functions are independent of structure, they can provide an access tool across structures and without reference to the peculiar historicity of any given organization. Because it is the activity that archives are intended to document, functions are also more closely tied to the reasons records are retained than are structures. Functions can thus serve as avenues for retrieval of the subjects of activities where names cannot serve.

Similarly, form can serve as an avenue to access the content of evidence where content analysis is impossible or uneconomical. Every culture has its own distinctive signature, by which members of that culture know each other and through the use of which they communicate implicit understandings. The signature is composed of a variety of signs. Forms of material culture, including record systems and records themselves, are such signs. In the realm of recorded memory, the medium is indeed the message; the society generates specific, relatively fixed information packages for specific cultural purposes. Declarations are not used to order provisions, bills are not used to woo suitors, and memoranda are not used to propose marriage.

Forms of material are culturally and chronologically specific information packages which are used to convey relatively consistent types of information. These packages change as cultural conditions change. Thus the registers of penitentiaries and hospitals changed into case files toward the end of the nineteenth century when treatment records became more important, and new social technologies of control were introduced to regularize practice. Journals became segregated into financial journals and daybooks with the development of bookkeeping and the segregation of personal and business lives in the modern office. While this process is on-going, the norm at any given time in the history of a culture is continuity in forms of material.

Occasionally a technological revolution introduces new forms of communication, and with them new cultural definitions of the information content they convey. Archivists today are living in such a time and witnessing the emergence of a vast new cultural repertoire of forms of electronic records.

Designing Intelligent Front-Ends:

There are two purposes in considering these issues about the structure of knowledge and the implications they might have for archives. One is to determine how to create data constructs that will permit the importation of data from external sources to populate archival information systems, as proposed earlier. Another is to use this analysis to construct systems in which users can approach a variety of databases through a single user language, positing a single question and receiving a unified answer. How can we design systems that will enable users to put their query to the widest range of applicable databases and retrieve from these with the highest possible precision or recall?

The solution to this problem is variously called an intelligent gateway or a scholars' workstation. The outlines of what is required for intelligent gateway systems are now fairly clear. Quite a number of approximations are commercially available, a few near things are on the drawing boards, and the standards required for common user query languages are under development. The functional definition of an intelligent gateway recently published by Martha Williams in the journal of ASIS describes most, though not all, of the modules which have been envisioned. But the outlines all leave a couple of gaping chasms across which we must leap.5

If we accept that databases will continue to be made by a large number of different communities, principally to satisfy their own requirements, and that it is desirable, if possible, not to have others reinvent the same databases simply in order to fit them better into the design frameworks of their own systems, then we will need to concern ourselves with two purely intellectual questions - one of structure and one of content. Structurally, I have proposed that we explore seven dimensions of reality as meta-categories which are sufficiently general to be used to bridge views in cross database retrieval. The content issue is how to identify like values in these fields, and how to identify other relationships, like parent-child or broader-narrower. We have already noted that pure synonymy is not always achievable, even in translations to neutral frames of reference (as in the case of the dimension of time), and other relationships are more problematic yet.

There are two tactics we could adopt to overcome the lack of fit between perspectives. The first is to construct a meta-language, a database of databases, an index to indexes, or a super-thesaurus, in which the relationships between all concepts are worked out. The second approach is to use semantic dues to dentify terms belonging to a common dimension, and then to exploit the knowledge of the user, and the learning ability of systems to effect the translation within a dimension between unlike views, and therefore unlike vocabularies. For practical and theoretical reasons, I prefer the latter approach, which I have labeled the construction of intelligent artifices. The practical reasons are self-evident; the construction of a meta database it extremely costly and time consuming. Its utility depends on the constancy of the structure of the databases to which it provides access, and we know that not only is the universe of on-line databases growing in absolute terms very rapidly, but that databases are disappearing, merging and becoming other than what they were on a regular basis. Finally, constructing a meta-file which serves as an index to such a primordial data-sea may be logically impossible, or at least the rules would appear to be too complex and idiosyncratic for use in design.

On the other hand, reasons for designing intelligent artifices are extremely attractive. The construction of relationships between unlike vocabularies by users has the benefit of being incremental and constructivist. As such, it has the practical attraction of being individualized, so that while any given user may (depending on the implementation of the intelligent front-end processing system) be able to take advantage of linkages deemed relevant by other users, they are also enabled to make connections which they personally find acceptable, thus in effect generating a meta-vocabulary for their particular view. The theoretical attraction of supporting the legitimacy of conflicting interpretations of reality reflects the heuristic problem which was posed; different views of the world incorporate data from outside their own perspective selectively. Not all values of chronological terms which can be meaningfully mapped to a given point on that dimension are relevant to the view the user is exploring.

Thus, if I am an art historian, concerned for the moment about the trade in ivory carvings of the Bapende sculptors during the Bakongo kingdom, I need to consider the cultural time of the Bakongo style which represents a limit on the production of such artifacts, the term for the political time corresponding to the 17th-19th centuries in other countries which were trading in such artifacts, and the term for the ecological period that includes the time during which these ivories were acquired. I need not consider numerous other kinds of time which can be mapped to the same place in my chronological dimension, such as sidereal states or economic(technological stages of development. I may decide also, that for my purposes, the only political dimensions of time which need interest me are those concerning the abolition of the slave trade. Thus each nation is to be considered only in terms of the presence or absence of this trade, which provided the alternative commodity for the market in art forms. while this choice may be perfectly reasonable in terms of my construction of the issue of my research, it would not be a generalizable feature of the divisions of the political time spectrum in an abstractly developed meta-index. Here the conceptual potential of the model for the pursuit of particular research endeavors becomes most evident, and compelling.

Access to the Archive of Digital Images, Sounds and Texts of the Future:

If, as seems increasingly likely, the future archive of man will be comprised of images, sounds and symbols (including but not limited to text), stored randomly in digital stores with access to these provided by an intermediary system which provides a large degree of intelligent understanding of the use of language, with a fair knowledge of the modes of thought associated with the particular user, what implications does this have for archivists and information managers today?

First, and most immediately, it compels us to recognize that the forms of material generally held by archival repositories today are only segregated from those held by libraries and museums by accidents of the informational format in which they are recorded.6 Archivists can contribute to the retrieval of archival and bibliographic information by focusing their description practices on form of materials and function of the context of creation.

Second, we must reconsider the business which archives and libraries have been in for the past century: the production of document surrogate databases to provide access to their holdings. Catalogers, abstracters and indexers may be destined for the same future which faced copyists prior to the invention of movable print and commercial correspondents in the age before the inauguration of the international news service bureaus. If we can learn to take advantage of databases describing the contexts out of which records were created, we can avoid the costly and inefficient steps of analyzing the records themselves in order to find clues to context, and then entering this data into our own descriptive systems by keyboard.

Third, archivists need to take very seriously the threat of the disappearance of explicit provenance. If we do not know where a piece of information arose from, or what kinds of activities it supported during its active life in a database environment used by many different persons, we will lose the link between the record and the actions it documents, depriving the archives of their special value as a source of "evidential" documentation.

Recognizing that archival information retrieval is uniquely dependent on exploiting intelligent artifices, we can see that the age of electronic information sources is both a boon and a threat to future access. It permits us to invent means of accessing archival records that do not depend upon content analysis of documents, or upon uncertain retrieval by proper names. At the same time, electronic information systems are the source of a new kind of records, that lack the forms with which we have been acquainted, and which lack many of the provenance defining attributes of traditional documentation. One of the challenges for archival methods, as we pursue strategies for harnessing the structures of our knowledge of social actions for purposes of description and access, is to adjust to such changes in the character of documentation.

Chapter VI: Recorded Memory and Cultural Continuity


The wealthy societies of human history, those we call civilized, established specialized secular roles for preservers of culture.1 Originally, these secular culture curators, the storytellers and chroniclers, were not passive custodians of the heritage but active interpreters and teachers, responsible for weaving the knowledge of the past into the pattern of the age in which they lived. With the advent of widespread literacy and printing in the modern world, a bifurcation between the roles of culture conservers and culture creators has been gaining acceptance. Librarians, archivists and museum curators have been cast into the presumably custodial role of culture preservers, while the creative function of the culture recorders or culture makers has been delegated to archaeologists, filmmakers and writers.

As curators, we tended to accept a passive definition of our function, one which permits us to act safely neutral in a society which believes in objective truth, one which is professional because it is technocratic. It is telling that none of the twelve priority goals established by the recent SAA Task Force on Goals and Priorities addresses the interpretation of archival holdings, public programs to make users aware of archival materials, or reference services for users of archives, though four goals are detailed under the broad heading of "Priorities for the Availability and Use of Records of Enduring Value". We tell ourselves that society values our efforts for their balanced contribution to posterity, but I will argue that our role is more closely akin to that of the storyteller remaking the past in a fashion relevant to our time. It is not simply that we are not the passive conduits of recorded knowledge; we cannot and should not be. Our value rests with the contribution we make to the continuity of culture, by connecting the present with the recent past, not by passively conserving the evidence of a distant past for the unmeasurable benefit of some equally remote future.

Nevertheless, we must ask whether the future will know more about us due to our efforts to preserve the cultural record, to retain recorded knowledge. I have required an answer to this question to justify the monumental efforts of the Smithsonian to inventory its holdings of over 125 million objects, to evaluate grant proposals for projects that plan to preserve thousands of thousands of cubic feet of records and millions of millions of bytes of data at substantial cost in human effort and dollars, and to assess the methods of the archives and museum professions. In each case the argument advanced has been some presumed future benefit. Is the task of compiling a comprehensive record of human history for the use of posterity possible? Is it even desirable? On purely theoretical grounds, is it a worthy intellectual enterprise? Whether or not the compilation, storage and preservation of evidence for the future is a viable project, we must ask whether it is a powerful rationale. Does it succeed in shaping our programs and securing their support?

The society has defined the cultural purposes of the institutions we serve sufficiently broadly that if we find our efforts on behalf of posterity are not practical, or useful, or desirable, we need not abolish libraries and archives, but simply dedicate them to serve the needs of the present for its immediate past, rather than the needs of the future for the comprehensive and representative record of human history. This will, of course, have significant implications for how libraries, archives and museums expend their resources, direct their programs and envision their highest purposes, but it need not diminish them as institutions.

At the 1985 meeting of the Society of American Archivists, Dr. Kenneth Foote of the Department of Geography at University of Texas at Austin, presented a paper subtitled "Artifacts and Memory in Communication and Culture.2" His outsider perspective shattered the comfortably accepted assumptions of the archival community by viewing recorded knowledge, and the written record in particular, as just another modest, transitory, and fragile source of evidence of the past, rather than as the principal source to which material evidence might contribute.

Accustomed as we are to viewing the record as physically fragile and in need of preservation, archivists were unprepared for Foote's two case studies suggesting the intellectual impermanence of recorded memory. The first case reviewed the efforts of an unusual interdisciplinary deliberative body, the Human Interference Task Force of the U.S. Department of Energy, which was charged with developing means to inform persons living 10,000 years from now of the presence of radioactive materials buried by our society. This group of semioticians, linguists, historians and scientists considered every possible way in which to notify the future of the simple fact of the danger of a radioactive site, and while it recommended a combination of markers and written records for such communications, it raised serious doubts about the value of either.3 The undamental reason we cannot design a means to assure communications with the future is that human history, human languages, human cultures are too tentative to support communications across such distances of time. Like the electro-mechanical technologies that limit transportability of data in our own era, social technologies and constructs are likely to turn our best constructed messages to noise.

Dr. Foote's second argument was based on case studies of how societies purposefully forget, how they manufacture stigma and efface material evidence in order to erase the past. He examined instances in which the psychological health of a community depended upon forgetting, as in Salem following the witchcraft trials. Foote noted that Germany has eradicated all Nazi party sites because, he suggested, they are far more dangerous to the present society than concentration camps since they reaffirm that the Nazis rose to power through democratic political mechanisms, however skewed and manipulated. One effect of Dr. Foote's work is to bring into crisp focus the shortness of civilized time in the scale of human history, just as paleobiologists so crudely alert us to the brevity of human time in the scale of life.

M.T. Clancy's From Memory to Written Record: England 1066-1307 demonstrates not only how recent our reliance upon written records as evidence is, and how slowly the culture accepted the replacement of the "sworn word of men true" by their signature, but also how the introduction of the written record as evidence was itself a political legerdemain, benefiting certain groups at the expense of others. Clancy's study not only shows how the world we take for granted, in which written evidence is considered to stand for the event, is less than 800 years old and limited until very recently to an extremely small population, but it also provides us one way of imagining the demise of the written record - a political environment in which it benefits the ruling classes to reverse the process of the slow acceptance of written evidence. Written memory could then be but a dim recollection, a myth, a background thread in the fabric of oral memory.4

This may seem farfetched, but the technological ingredients for doubting the written record are already in place even though our cultural attitude has not yet been affected by them. In the past few years we have witnessed the emergence of digital copying machines which make each copy different from the original. The electronic age has brought us "publication from databases which are constantly undergoing change, so that like the manuscripts of medieval monks, no one "copy" is precisely like another.5 There is as widespread a cultural illiteracy surrounding the ability to "read" an electronic mark as authentically that of the individual vouching for the truth of the evidence as there ever was in the middle ages before signatures replaced seals and witnesses. All these technological ingredients undermine the legitimacy of the printed and written word, erasing its claim to facticity and its immutable tie to a concrete time, place and person, which permitted it to substitute for the evidence of the sworn man. My eleven year old pays as little attention to incredible feats on television as to depictions of murders and mayhem, illustrating that the visual record has no more credibility as a document of reality than the written word. Why is it difficult to imagine that additional technological developments will further reduce the credibility of recorded evidence?

Three hundred years ago, halfway between the present and the birth of the concept that the written record was evidence, Sir Thomas Browne mused on the discovery of ancient graves that the iniquity of oblivion blindly scattered her poppy, and deals with the memory of men without distinction to merit of perpetuity."6 This essential randomness of surviving historical evidence is, of course, one of the driving forces behind the efforts of librarians, archivists and museum curators to preserve a comprehensive document, but it should make us wonder if we are erecting sand dunes to protect the coast against the ocean.

Sir Thomas did not stop with this challenge to building evidential dikes, but went on to note that cultural memory might not simply be random, it might be systematically unjust; as he put it - "Herostratus lives that burnt the Temple of Diana - he is almost lost that built it." Librarians, archivists, and museum curators were all confronted by the truth of this second observation as they scrambled over the past decade to fill their repositories with evidence of women and minorities. What we should come to understand from this experience is not that we have become more representative, but that cultural blindness and sudden cultural insights are both equally culture bound. The truth is not an absolute, but a contextual reality. The past is not a given, but a mutable creation. There is no law in history to select the forces of causality - causality is an intellectual tool we use to make sense of too much evidence. The pasts we construct are all discussion with the present, axes in today's intellectual battles, not monuments to some fixed reality. And this is good for us, because if it were not so, scholarship would cover every territory but once and never need to return, and we as the holders of cultural evidence would be impoverished by having only one monograph for each age level representing each Dewey class on our shelves.

The final argument against the comprehensive record lies in our extraordinary ability as a culture to do without it. We can construct most satisfactory versions of the past from the fragments we possess. The recent Annales School accounts of daily life in medieval France, of the working, eating and sexual practices of peasants and workers, constructed through the exegesis of randomly preserved texts rather than from the records of surveys made by suitable bureaucracies (which, of course, did not exist), and from the analysis of randomly surviving market counts and production statistics rather than from complete censuses, are, nevertheless, rich and convincing.7 Archaeological reconstructions of the thinking of pre-historic man from the remnants of his material life and the semiotics of his cave paintings, while not fully answering all the questions we might have about these early societies, stand as impressive testimony to our ability to make sense of other cultures from clues. As we survey the information pollution which clogs our cultural arteries today, can we really say that the survival of a random 1% of the documentary record would give the future less of an understanding of our dizzying society than would a total record, or a selected record which reflects the biases and limitations of the cultural curator, us, captured in the limits of his or her own moment?

The science fiction novel Canticle for Liebowitz depicts a time in the future following the third world war in which a dedicated community of monks, having preserved fragments of the recorded knowledge of the twentieth century, is able to convince the authorities to canonize their patron, E. Liebowitz.8 As proof of his historical actuality they display recorded documentation: a shopping list complete with pastrami, which they discover in a fall out shelter. Several hundred years later, with additional scraps of information, they are able to make an arc lamp, with an electric generator driven by monk power, to illuminate the library in which they study these precious artifacts of the past, and concoct most fanciful histories based on their fragmentary knowledge. The lessons I draw from this fictionalized thought experiment are two. First, the preservation of recorded memory for the future, whether systematic or not, is a perilous endeavor, most likely to fail. Second, the record of the past contributes to our understanding of the past by its continuity with our own culture. We are able to reconstruct the lives of medieval peasants because they share fundamental aspects of our culture (or rather we share theirs since we inherited it from them), while the Leibowitzian monks are unable to use the information they have to reconstruct the twentieth century because their tie has so utterly been broken. The Department of Energy study suggests the same - if we maintain cultural continuity with the future, we will be able to communicate with a high degree of certainty, but a gap in the continuity makes it as likely that our descendants will worship the radioactive sites as that they will avoid them.

If, as I have been suggesting, we cannot be certain that recorded memory will communicate with the future, because the written record is transient and is always a biased reflection of the present rather than of an objective, absolute past, and if accident provides for reasonably reliable statistical sampling, and if historians do very well without plenitude, what can be said for the role of archivists, librarians and museum curators as custodians of a comprehensive record of the past? What should be said of this role, as a matter of professional strategy?

The storytellers and the chroniclers made the accounts. They passed them to their contemporaries, those already born passing along what those near to the end of their lives retold. These culture conservers kept the past through its transmission, through its remaking, for the present, not through preservation of the thing in itself In this context authentication meant testimony that possessed authenticity to listeners, not some abstract validation as actual historical fact. Could it be that librarians and archivists really inherited this role, a role which finds its meaning in the contemporary society, not in the future? Have we made a tactical error in emphasizing the survival of information for posterity over continuity between past and future?

Concerned about the image of archivists in society and by the problems archives have in securing funding in competition with claims of others, the Society of American Archivists several years ago year contracted for a study of how archives were perceived by resource allocators.9 What they discovered raises serious doubts about the value of resting the claims of culture custodians upon the benefits of their efforts for the future and should be a lesson to the administrators of all cultural institutions. Resource allocators professed great respect for the role of archives in preserving the past for the benefit of the future; they understood quite well what archives did and held the professionalism of their staffs in high esteem. But they did not allocate adequate resources to archives because the other organizations competing for resources had more immediate claims, and because they viewed the task of the archives as intrinsically less resource demanding. The image of a well organized closet or attic, fought and feared by librarians, archivists and museum curators, is alive and well. I contend that it is being fostered and nurtured by our own actions. The sooner we admit to the futility of efforts to accumulate a comprehensive and unbiased record for some future generation, the easier it will be to argue our benefit to the present and to compete for resources with other essential services. If the lessons of Canticle for Liebowitz and the DOE study are to be taken seriously, the resource allocators are right - they should not invest in the systematic preservation of cultural knowledge, if only because it cannot succeed.

I conclude from this that we should restate our ends. In the process we will need to remake the means we have been advancing for the past decade. think it is possible, with only a subtle change in our rationale, to make a stronger case.

For example, the Committee on the Records of Government, a blue-ribbon panel sponsored by the American Council of Learned Societies, the Social Science Research Council and the Council on Library Resources, and funded by the Mellon, Rockefeller and Sloan Foundations, recently issued a report which opened with the claim that: "The United States is in danger of losing its memory."10 Does anyone care? Within the report there are other threats which are taken more seriously by policy makers, such as potential losses of investments and productivity, serious erosions of rights of citizens and potentially critical failings to retrieve information relevant to national security decision-making. Are we so captive to our definition of ourselves as preservers that we cannot use the weapons at our disposal to attract more support?

A recent evaluation at the Smithsonian of the priorities of the entire Institution in light of substantial budget cuts focussed on the need to preserve the collections, and proposed restricting access and reducing research efforts if necessary to assure the protection of the heritage. The Library of Congress recently made similar decisions in response to the 1986 budget cuts required by the Gramm-Rudman-Hollings act, and when these sparked what should have been a welcome demonstration of support, the Library appeared to assert the importance of the future over the present. I believe it is critical for librarians, archivists and museum curators to demonstrate their importance in today's society, to point to the need for their services now, not in some indefinite future, and to emphasize the importance of their function in terms of everyday, real, economic needs, not identification with the future and our heritage.

Records are reflections of human activity. As such they document activity. Their form makes cultural connections, their content makes intellectual connections and their function supports political and social connections relating to accountability. The perspectives which have separated librarians and museum curators are that the former focus almost exclusively on the content of their documentation while the latter are more concerned with the form. A scientific journal to a librarian will be retrieved by what it contains for persons interested in high-temperature alloy stress, while for a museum curator the scientific journal is an artifact reflecting the conventions and networks of scientific discussions in the twentieth century. Archivists occupy a middle ground between subject and object oriented views, emphasizing function over either form or content.

In his annual report for 1969, Smithsonian Secretary Dillon Ripley stated that "the Smithsonian's uniqueness and value depend on our success in being a different kind of marshaling center, where recorded knowledge gives wide access to pertinent inquiry."11 This vision of a "marshaling center," articulated in the years before Ripley's initiatives resulted in the creation of the Smithsonian Associates, Smithsonian Magazine, Smithsonian Radio and TV, and the Smithsonian Traveling Exhibits Service, has inspired me for the past decade. It points to a proximate role for the cultural repository in the life of the society because it recognizes no boundaries between the library, archive and museum in the quest for evidence pertinent to inquiry, and it recognizes that there are different sources of pertinent inquiry, what I have come to call different user views in the design of databases.

Society generates a record of its activity to serve distinct social ends: as tools for the design of the next building or machine records are applicable knowledge; to assure the rights of its members or as simple excretions of the activity itself the record serves accountability; to train youth, the communication of the record is contained in its cultural form. As intermediaries in the process of delivering these message-bearing remains, we are the sub-contractors of engineers and planners, lawyers and social workers, teachers and journalists. To claim a social role, to demand our share of resources, we point not to the needs of the indeterminate future and the nostalgia of the unappreciated past, but to the immediate requirements of today. These are the requirements for accountability, for applicable knowledge, and for cultural connectivity. Our strategy should be to organize and rationalize the world of cultural repositories so as to create an information implosion - a focussing of evidence, a marshaling to use Ripley's term - so that sources from all cultural repositories focus like a laser on the needs of today, falling into place in a puzzle, harnessing the energy of their differences on a point.

If the survival and interpretability of the vast amounts of documentation generated by our society are dubious, so that the cultural endeavor cannot be justified on the grounds of its benefits to the future, and if what we learn from the past is only the lessons which we reconstruct to suit our present, still it is possible to find meaning, purpose and direction in the calling of curation. Our excitement is real when we learn of a library kiosk in the shopping mall which provides special consumer advisory information, or of a museum program which uses secondary school children and their teachers to reconstruct historic sites based on evidence from digs and from records in the local archives. We are truly inspired by the connections made when the Austin History Center opened an exhibit featuring a festival which was historically significant for firemen in the past, with a festival for the firemen of today and their families, at which the present day fire department graduated its cadets and honored its retirees. Such programs reflect the calling of the culture custodian. The challenge is to make sense of the documentation - not to keep it. To deliver it where it is needed - not to store it. Let us stress making comprehendible connections over acquiring comprehensive collections, or we risk amassing a rotting storehouse of knowledge.

To begin this project, we need to better understand the nature of the cultural document itself and its connections to social action. We need to explore its internal structure and its use. We need to design systems for its retrieval which provide access by a variety of perspectives and allow users to move between views in the conduct of research. But the promise of the effort is that each of us can contribute, at the local level, in our own settings, in unique ways and with rich results. We need to develop mechanisms for outreach which go beyond accepting users to actively demanding their involvement.

In redirecting ourselves to this effort, we need to revisit our rhetoric, removing the unconvincing references to our role in preserving evidence for posterity, and replacing them with our role in focussing and connecting the past and the present. Instead of envisioning ourselves as victims of an information explosion, we need to emphasize a vision of archives, libraries and museums joining to bring about an information implosion.


Informatics: The interdisciplinary study of information content, representation, technology, and applications,
and the methods and strategies by which information is used in organizations, networks, cultures, and societies.