< Working with Knowledge - Peter Horsman

back to conference papers page

Conference Papers Online

Keynote Address - as delivered 6/05/98

Peter Horsman

Netherlands Archives School 
International Council on Archives Committee on Information Technology, chair

Thank you very much for your kind words and for the invitation. My biggest fear is not the questions that I might get, not even the little fights but the technology. I donít trust technology at all. Just before conference it worked, so we take some time to get it working and I will use even two different types of technology as part of the sound.  

My dear colleagues, apart from the great honour there are a few advantages of being the first speaker of a conference and even more if one is supposed to deliver the keynote address. One of those advantages is that no-one can cut the grass under your feet by telling things that were so essential for your own paper. On the contrary, you are even expected to make statements for the whole conference and leaving the problems to answer to that to the speakers who follow. These advantages are so big that I even donít want to think of the eventual disadvantages of being the first speaker, such as forgetting issues or not having the opportunity to react upon previous speakers or not having smelled the atmosphere of the conference. What will be by paper about? It is about looking at archives and records from a distant position. It is about abstracting by modeling and founding archival description on abstract models and actually my paper will be more than about description. It is a little bit about the application of archival standards and eventually about seeking global archival concepts. Concepts that archivists share all over the world instead of stressing national peculiarities.  

Consequently my dear colleagues my paper shall be abstract. Indeed yesterday we were talking about fun, but Iím afraid will not be too much fun in this paper. I warn you in advance you can still leave.  

I will try to bring together some of the current concepts of archival description, including the concepts of contents, form, context, structure and metadata, and I talk about archival constructs as well. Most of these concepts and key words you will find in the conference program and no doubt the speakers whose name is related to one of these issues will explain them way better than I can. I consider my task to be to stay at a global level and to tell you truth that is where I feel most comfortable.  

I will build my paper upon generic data model where Gavan was referring to and how the ICA committee on information technology, how is called now, is working, is still working and is proceeding working. I wonít get into the details of this data model, but it will focus on the archival theoretical concepts and relaying it.  

I must tell you that these concepts are not without discussing and having debates within the bosom of the information technology committee. We have sometimes pretty tough fights about it, but usually after that we take a beer and have fun again. The model in itself was intended to support the international standard for archival description - ISAD(G) and basically the ISAD comply in database design and this links my paper to archival standards.  

There is one other thing I should tell you. I came in Monday night, so yesterday I had the opportunity to meet some people here in Australia, some other colleagues from abroad, and we talked a little bit about this conference and this urged me to re-write my paper. So the original paper is here, and the paper that I read is here, itís in handwriting, so sometimes when you see me hesitating it is just I cannot decipher my handwriting though I started my teaching career as a palaeographer, but that is long ago.  

I will get back to that probably later, but first letís focus a little bit on description. I even was considering to change a little bit my first slide, but I donít trust the technology too much, but I donít trust my skills with the technology not at all, so I decided to leave the title like this.  

So I will start with that, what is then archival description? Why do we describe for? Itís for control. The control of what? The control of records? The control of objects that contain knowledge? That are the sources of knowledge? What do we want to control intellectually and physically? So letís have a look, not to stay too abstracted for the whole paper, a look in the real world of records, and I should like to have now the video projector on. That is the other technology. It works. Thanks! I trust people.  

click on image to see full-scale version

Do we describe as archivists single sheets of paper or pages in a volume? Now I have to turn. This is really great! Or do we describe rather volumes as a whole, or series of volumes? I put you a little bit in the dark, itís just for a while, so donít fall asleep because I will see you. Do we describe the series of volumes or maybe the whole fonds? Or is it the context of the records, such as the record creating organisation, its functions or whatsoever? Or is it finally the context of the records creator or do we really describe it all? You can switch it off please. Thanks a lot.  

Letís go back to the kernel, to the problem. In 1993 David Bearman, who else, criticised, what else, the little progress archivists had made by that time in developing archival automated control systems. Indeed, he said, the history of archival automation has not been a story of great successes. Apparently he has never been in the Netherlands before that time. Or in Australia. Archivists, according to Bearman, were not able to adequately use the tools of information science. To tell you the truth, I am not too sure that since 1993 really much progress has been made, except in Australia of course, and the Netherlands. Sure, since then much work has been undertaken, and particularly in the field of developing archival standards to disseminate information, archival information all over the world. On the international level, the ISAD standard, the international standard for archival description, followed after a few years by the international standard for archival authority control for corporate bodies, persons and families - ISAAR(CPF) - or national standards such as the fabulous Canadian RAD - the rules for archival description, and all over the world archivists are breathless trying to adopt again new technologies for controlling records, inventing new techniques such as encoded archival description, but my dear colleagues, isnít it old wine in new barrels? Iím afraid it is.  

Letís put ourselves the question. Arenít we trying to catch the tiger by the tail instead of riding the beast? That is an expression that I borrowed from a Swedish colleague. It is like in 1993, looking at documentary standards rather than looking at record keeping systems. It is still basically the thinking of MARC standards rather than working on real archival description.  

Let me try to make my point clear, what itís all about. I think that the origins of the relative failure so far of archival automation is in the very fact that archivists take too much a documentary approach. The starting point for many archivists, except archivists who are sitting here, the starting point is the records material. Sometimes referred to with a metaphor of being the sediment of activities. At least in my country they do it, and they use for the word sediment even the same word that is used for rain. Something that falls down and thatís unpleasant. Sediment, is it like polluted mud in a harbour? The dirt of the bureaucracy good enough only for archivists in the basements of the administration? What business are we in, to use the words of David Gracy II. Are we in the business of being refuse collectors? Records my dear colleagues, are not the sediment of activities. They are the atoms and the evidence of activities, and very often the only few remains of the activities. They are as dynamic as the processes that created them, that used them. A sedimentary look pushes archivists towards a documentary approach. A static view instead of a dynamic view. The approach of reconstructing an old order, an original order, that never really existed because it was as dynamic as the organisation that created and used records and if you want to see an old order, an original order, please I invite you in my office to have a look at my desk.  

A documentary view, a static view, brings us basically documentary methodologies and techniques, and that is what we see happening all over the world where archives try to apply the ISAD standards as a methodology starting with the fonds as a dead collection of used papers and not with the records, record as the evidence of business transactions.  

The whole construct seems to me a library approach. Nothing wrong with library approach, but itís not an archival approach. Despite of the fact that many archivists are talking about archival principles such as the principle of provenance or respect des fonds and we use wrong terminology, but basically itís too often that library and documentary approach - this is not a critique, itís just an observation.  

What then is a record, what I am talking about? A fonds, or any kind of grouping of records is fundamentally different from a collection of documents, and thatís nothing new. Thatís the kernel of archival theory we have our principles for. But why do archivists neglect this so often in their description. what we really need is to recognise the dynamic nature of records. The very bond with the business processes that brought them into life and that use them. The primary bond is not a bond between the records, that we will read in the theories of Luciana Duranti for example. Itís not the bond between them, because by nature documents cannot have an interrelationship other than by a staple or a paper clip. Neither between the records and the record creator. It states the fonds concept pretty much, but the primary bond is the bond between the records and the process, or the business transaction to use the words of David Bearman again.  

The primary provenance is a functional provenance and not, letís say, a custodial provenance or an organisational provenance. A record is a record, not by its very nature, because itís just a document, but because of its function, its role in a business process. Thatís the basic idea of what we call in these days the functional archival science as Angelika Menne-Haritz worked it out in paper given at the ICA conference in Montreal and that she borrowed from Bruno Delmas from France.  

The business transaction that creates the record defines the contents and the form of the record. The contents is the representation of the act or the fact. The form is dictated by the procedures and the legislation about the transaction. A simple example if you buy real estate property you go probably to a lawyer or to a notary, I donít know how itís here in Australia, but this person is skilled in writing the right things on the right place in the right form so that it has legal value, so the act or the fact, thatís the contents of the whole document, and this person knows in what form to put it ,and itís the same with tax form that you fill in either by writing it down on paper or do it in electronic format.  

This simple sheet two, small boxes and a line, is the essence of archival science: 

click on image to see full-scale version

It is basically the essence, the very kernel of the data model where Gavan was referring to. Itís not the hierarchical diagram of the fonds and its component parts that underlies the ISAD thatís the kernel. The top-down model is too static, itís too rigid. It is an archival construct that I donít need. 

Yet both components, the transaction and the record are both part of a bigger whole. The transaction is part of a complex system of functions and organisations. The record is part of another complex system of files and series and sub series, and all that kind of archival constructs. Both worlds had their own characteristics, their own rule, their own dynamics and both worlds should be understood by the archivist of the next one.

click on image to see full-scale version

If we want to describe or to control archival objects, archival knowledge, to control it for whatever purpose we should look at both worlds separately, but we must understand the relationships at the very bottom. The relationships and what governs the behaviour of each of the both worlds. 

Record creation has to do with functions/tasks of an organisation. Record keeping, the other part take care of the orderly availability and reliability of the records by bringing them together into a record system. What Duranti terms setting aside. Thatís basically the two functions. 

Letís look first at the left hand part of the model, the world of the business transactions, the functions, and not surprisingly this part is underdeveloped in the work of archival standards. Some work in this field on standardisation has been done by the current ICA committee on archival standards and the ISAAR, though the underlying concept is borrowed from the library, but, and its rather as a sub alternative to ISAD, rather than as a separate component of an archival description system, or an archival control system.

click on image to see full-scale version

The fundamental entities in this part of an organisation are the business processes, the transactions, the functions that they carry out and the organisational structure that is responsible for the operations. At least for public organisations the relationships between organisations and their functions are defined by legal system, by law, and this relationship is called the competence. As one of the few archival institutions that I know of, the Federal Archives of Switzerland undertook a long lasting research into the competences of Swiss Federal Agencies since the half of the 19th century and implemented the results of this research into a database, the Kompetenz Kartei, and they started working on that about 10 - 15 years ago on paper and they are transferring it now into an automated system. When I saw that for the first time, I can tell you that the Swiss Federal Archives is the most well organised, clean archives all over the world. You wonít find any piece of paper just hanging around, its completely organised as the whole country is, except one room, the room with the Kompetenz Kartei. A classic archive. A more or less similar approach, a competence based approach, is undertaken by some other archival institutions including the fellow Archives of Canada and the National Archives of the Netherlands in the course of appraisal activities. I know that tomorrow Helen Morgan will say something about it, only nice words. 

In the current archival theory this part of the real world of records is referred to as the context. To be precise the context of record creating and of the record system. So I turn now contents, to form, context and archival constructs. So I am pretty far in my paper. 

Before going to the right hand side of the model, the part of the record system, I would like to stand still a moment at the momentum of the record creation.

click on image to see full-scale version

First the momentum itself. The relations between the process and the records are more complex than just creating a record. The relationships include also the use and deletion of records. Of course records can only be created and deleted once, but used an unlimited number of times by various business processes. Eventually business processes carried out by other organisations. 

It are these relationships, as David Bearman states, that they should be documented as well, documented by the record keeping system. It is impossible to do this afterwards, letís say, after disposition or disposal or whatever you will call it in your terminology. It should be done by the record system at the very moment of both creation, use and deletion. In fact any description afterwards is basically due to a defunct record keeping system, and unfortunately archivists have to deal with a lot of defunct record keeping systems. One could take the position, however, that archivists instead of designing description systems should put their energies into developing record keeping systems. This is particularly true and promising in our days, in which records become electronic and consequently record keeping systems need to be redesigned and should have the functionality of self documenting. 

Thatís really going into the forefront of record creation, and it would not surprise that in the next meeting of the Information Technology Committee we will talk again about the data model, but then not only use that data model for archival description, but also as the underlying data model for record keeping systems, and it would surprise you weíre going to cooperate, corroborate, with another ICA committee, the committee on electronic and other current records, who has it in its terms of reference. 

Such a record keeping system will then automatically capture the relevant contextual information and there has been done a lot of research already what really consist the kernel of the contextual information, its both the project on the University of British Colombia lead by Luciana Duranti and also the project at the University of Pittsburgh and currently also the University of Michigan and other universities are working on that. The record keeping system shall also automatically capture the data about the technology used at the momentum of creation and of use and the latter Iíd like to term metadata. This distinction between context information and metadata has been advocated by the ICA Committee on Electronic Records before it was called the committee on electronic and other current records in its guides for managing electronic records from an archival perspective. I know that indeed in the world of information science, what Gavan was referring to, the terms are used sometimes differently. I know, I am an information engineer, even a knowledge engineer, but who cares, the archival science is older so we have the first right.

click on image to see full-scale version

Then finally the record system, the whole of archival constructs that we call the structure of the fonds. ISAD, as I told before, defines a simple, rigid hierarchical structure with fonds at the top eventually with sub fonds underneath and then the series, sub series and so on and at the bottom line ISAD puts the item, the smallest physical unit to be produced in the reading room. 

I know that the ISAD Committee got a lot of critique particularly from Australia. I have seen Chris Hurley sitting in the room and we talked about it in one of the meetings in the Hague on that, and I agree with the critique on that. 

ISAD does not mention the word record, but eventually defines at that level as the smallest unit the item, letís say a document. I should like to interchange that for record. 

As a matter of fact I donít find the ISAD too much compliant with current archival thinking. They use the word fonds but what else do they mean than a record group? By no means ISAD seems to refer to the ideas and concept as has been laid down over the last years in publications by Michel Duchein and later by Terry Cook. 

Supporting particularly Terry Cook I would define fonds as a conceptual whole. Of all the records created by an entity, on whatever level, it is the concept with, I think, no direct practical meaning, although we have a principle on it. A database terminology, its just a view as a result of a very simple query such as, select all records created by, and with a proper database design and proper data input you can define any forms you want. 

The records might be put by the record keeping system into units, such as folders or volumes. A number of items we might call a series or a record group. Whatís the essential difference, whatís in a name. 

But my dear colleagues let me look a little bit in our electronic future. Do we need even the series? I know that what I am going to tell now is a little bit risky. 

A record keeping system, a proper record keeping system, should mirror the business functions like it does in my model. It should mirror the business functions and the transaction that created and used the records. A file, a case file for example, is the whole of records created and used in the course of one particular set of interrelated business transactions that you may call a file and it can give a similar definition for a personnel file. 

A series could be a set of files or records created and used in the course of a business function. Note the series records exist in that way because in both cases, file and series, the construct could be realised by simple database views and simple queries in asking for what function created it, and then we get all the records or what sort of functions used it. Do we then need the concept of series else than just a term. Isnít the series, just an archival construct developed for the paper world in order to establish administrative and intellectual control? 

When Peter Scott in 1966 wrote his famous article on abandoning the record group, didnít he put at the same time a time bomb under his own series concept? A time bomb that we see bursting in the world of electronic records? 

My dear colleagues, itís really about time not only to redesign archival description, but the whole of our archival methodologies and archival constructs. Hopefully, and I do expect this, this conference will contribute positively to this redesign, and if at the end of this conference it might come out that I was completely wrong, which I doubt, but might it come out Iíll be happy to destroy the paper here that I wrote here and to give this printed one for the proceedings of this conference and in this little disk is just the opposite of what I have said now. I thank you. Thank you very much. 

(Gavan McCarthy) Iíd just like to thank Peter very much, for I think its a very challenging and a very revealing presentation. Early on you mentioned the sediment. Sometime last year I read an article, probably in New Scientist, about a research project that was run in Melbourne, I think it is still running in Melbourne in Victoria where I come from, where they were indeed examining through core analysis the sediments of the Werribee sewage farm for heavy metals and other pollutants and things that were passing through people and using this as a way of reconstructing how life was in 19th century Melbourne. So there are roles for sediments I think in revealing our past! 

back to conference papers page

Published by: Australian Science Archives Project on ASAPWeb
Comments or questions to: ASAPWeb (asapweb@asap.unimelb.edu.au)
Prepared by: Helen Morgan and Sandra Peel
Graphics by Lisa Cianci
Date modified: 7 October 1999