Visualising the Multiplicity of Relationships: Changing Nature of Archival Description in a World of Digital Representations
Referat auf der Sektionskonferenz der Hochschularchive und Archive wissenschaftlicher Institutionen ICA-SUV, Barbados, 26.-29.6.2013, Karsten Kühnel
Dear Ladies and Gentlemen,
“By attempting to define, to categorise, pin down, and represent records and their contexts of creation, management, and use, descriptive standards and metadata schema can only ever represent a partial view of the dynamic, complex, and multi-dimensional nature of records, and their rich webs of contextual and documentary relationships.”
These “rich webs of contextual and documentary relationships” how Sue McKemmish has put it in this quotation, is exactly what I mean with that “Multiplicity of Relationships” which needs to be visualised. The necessity of a better visualisation is supported by Tom Nesmith assessment in the 2011’s Helen Willa Samuels-commemorative publication by Terry Cook, where he says:
“that the knowledge of records and archiving actions that a wide variety of archivists and others have thought necessary in metadata and descriptive systems and standards has become so voluminous that no formal practical system could convey it.”
The sheer amount of multi-layered descriptive information, with which the modern archival discipline of archival description operates and which results – or might result – from the consequent application of international descriptive standards, brings the danger with it that the archives user is simply overwhelmed and that the data seems to be unmanageable.
I think there might be a certain danger that archivists produce such a lot of metadata which is incomprehensible to the user. The more complicated the relationships get – which are actually supposed to be of help to the user in understanding the data-, the clearer the presentation has to be. The more the understanding of the networks of relationships needs the knowledge of the archival units’ structures, the more the archival units themselves have to be visualised together with the descriptive information on them. The key prerequisite to take concrete and effective steps is the digitisation of archival material. Let me consider some aspects in this context.
What happens if an archivist removes a couple of record volumes from different locations of the depot and presents the user with this material? What happens if the user takes a single sheet out of the record, turns it around and places it on top of the record’s folder?
The archival context is disturbed in both cases and the material creates a new context.
The user who takes a single sheet out of the record has altered the record’s composition and has isolated the sheet. Following that, he has put it back in its usual context.
The archivist has removed files from the record groups and has thereby altered the groups’ composition and has isolated the files. And – what is more – he has created a new composition by presenting it to the user. Decontextualisation and the subsequent altered recontextualisation, these are the processes to be observed here.
An archive is digitising its material. New images are created sheet per sheet. Then these images are reworked, the front and back are merged to a single digital file, multi-page image files are created, the first few pieces of metadata are included in the files. The process of recontextualisation has started.
The recontextualisation, the assembly of the initially isolated image files into an archival composition is conducted by the concept of representation. The newly created contexts are supposed to represent the original contexts. These digital representations are – without a doubt – new entities and to bring the new contexts in accordance with the original ones is just one possible recontextualisation out of many.
Digital recontextualisation is the consequence of a decision made by the archivist. He decides to make certain contexts visible which are then either newly described or described for the first time. Thus, digitisation leads to the process of creating documentary units out of archival units. The documentary units need to have one or several documents descriptions. This process creates a two visualisations: concerning the images in the form of documentary unit, concerning the description in the form of document’s description.
The image is taken from the metadata-model for the archives portal EHRI and is a snippet of the physical domain model (draft).
The assembly of isolated parts –even if they had belonged to a greater whole in the first place – involves the danger of a random recontextualisation which does not refer back to its original contexts. There is the danger that the principle of the provenance is carelessly abandoned and that archival material is interpreted rather subjectively.
The ICA has published standards for the description of the four main entities which shall give also the framework for the destination of recontextualisation of digitised material and of the recontextualising manner of archival description.
Recontextualisation signifies the pointing out of relationships which eventuate in the origination and use of the archival material. These relationships are exactly that context of that describing effects the process of archival description to become the process of defining fonds and collections . Heather MacNeil initiated an evolution in discussion and practice in 1992 with her essay “The context is all”, which led to the fact that the record group or fonds was first a construct and then a constructible entity. At first this concerned archival units on a fonds and collection level. This could be broken down to the file level. Physical units on the depot shelf were the smallest undividable units of constructs which were able to recontextualise relationships and make them visible in finding aids.
The digitisation offers us the opportunity also in the microcosm of units of description defined by archival reference codes, that is on the deepest descriptive level to evolve from a – how Terry Cook said – „static identification of records with a structure“ to a„dynamic relationship with a creating or authoring activity“ or at least to make that visible.
Relationships are the connecting dots in the abundance of contexts. Peter Horsman illustrated a Four Domain Model for relationships of records in his essay „Wrapping records in narratives“ in 2011.
Several of them show contexts of origin, causalities at Corporate Bodies, persons, families or in functions, activities and processes. They highlight the provenance of the archival material, the organisational provenance as well as the functional.
The relationships in documenting deserve special attention. They refer to the creation and use of documents. Those relationships mirror the de- and recontextualisation which in return define the life cycle of the material. The files could have been generated out of certain functions but could then be re-used for other functions.
The process of de- and recontextualisation can turn the hierarchical system of the descriptive levels upside down and collections can find themselves recontextualised as parts of records and files.
Lastly there is the area of relationships which is not reflected in the archives at all.
The visualisation of relationships by means of new assemblies of archival units in the form of digital documentary units is in itself a visualisation which is created by the definition of an affiliation of formerly isolated images.
The usage of the Metadata Encoding & Transmission Standard METS is best suited for a sort of clamp which is applied to the images in a defining manner. The METS-based compositions are determined by the descriptive attributes which are applied as document description of every documentary unit by the archivist. First of all he has to start his analysis of the relationships with the analogue material, that is the archival unit, to be able to define a documentary unit on the basis of his analysis results. Then he is able to generate the corresponding description. On a last step follows the act of attribution of analysed relationships as parts of the description of the virtual creation of fonds or collections. The “conceptual fonds” has been made visible and its borders have been documented and explained now.
This necessary sequence of the workflow leads to the fact that it is nearly impossible for an outsider to detect additional relationships based on the digital documentary units offered to him.
So as to not reduce the description as a process to the subjectively limited cognitive faculty of single archivists it is recommendable to reproduce the digital archival unit in an unmodified manner as one way of the digital representation of the original material.
“It does, however, mean that the provenance of the records must be clearly reflected in the description, that the description must enable retrieval by provenance, and that a descriptive system must be capable of representing together all the records of a single creator held by a single repository.”
I quoted the 2013 edition of the DACS (“Describing Archives: A Content Standard”).
Given that, it is possible to argue that the strict respect des fonds or to put in German the “Registraturprinzip” or in English the principle of original order as a principle of arrangement is the prerequisite for a comprehensive and transparent analysis of relationships.
The identifying of relationships by the user could be annexed to the finding aids by means of appropriate tools for the integration of user-generated content. If the user has to combine images of digital archival material, he has to be enabled to attach the added content in the form of a self-defined documentary unit with a set of appropriate metadata. If the usage of METS for the definition of documentary units is continued, tools would have to be provided which can mirror the process of generating image sequences in a METS-XML-file format in a flexible manner.
The encoding of the content-related information to such a new METS file would have to result in an additionally generated EAD file.
Concerning the EAD encoding of relationships which correlates to functions, activities, etc., there is now a problem similar to the missing EAC-F for the description of functions. Functional provenance is a type of provenance which is not taken into account as yet. A solution would be to define the attribute <origination> through a type more closely that heralds the indication of either the organisational or the functional provenance. An example could look like that:
In this example description on file level five provenances are illustrated. A set of a pre and a final provenance indicates the organisational provenance, that is corporate bodies, persons or families, who were responsible for the file’s creation. Furthermore, there are three functions depicted which have changed through the sequence of usage of the file. They are as well to be regarded as causative and are therefore listed as functional provenances under the attribute <origination> as well.
The links to the authority records for describing the agents and functions listed as provenances would have to be put in the paragraph of the history of the file’s origination or better known as “administrative or biographic history” which is encoded as <bioghist>. According to that example, the EAC-authority records have to operate analogous to the standards ISAAR-CPF and ISDF as EAC-CPF and EAC-F where the last one is missing as yet.
The display of functions deserves special attention in the visualisation of relationships. The standard ISDF defines functions as follows:
“any high level purpose, responsibility or task assigned to the accountability agenda of a corporate body by legislation, policy or mandate. Functions may be decomposed into sets of co-ordinated operations such as subfunctions, business processes, activities, tasks or transactions.”
Through describing, however, only those functions can be made visible that are not only mandated to the agent but also really performed by him. The result of the high level functional analysis which Helen Willa Samuels has worked out for university archives shows an approach which aims at the effect in the society as a whole as cause and intent in actions.
When I talk about “functional provenance” from there, I think of functions as a cause for actions and effects – the display of which matches the provenance principle in the definition of the ISDF standard.
Functional relationships or the web of relationships were already depicted by David Bearman und Richard Lytle 1986 (Archivaria 21) as an increasingly transinstitutional entanglement. The entanglement originates in the change of the structuring of administrative bodies from self-contained, virtually supreme entities into institutions with flexible, ever-changing power structures, competences, responsibilities and interinstitutional interactions. The diversity of manifestations and the de-hierarchisation of administrative bodies associated with sharing of functions are clearly described by Bearman und Lytle as follows:
“In the modern world of task forces and committees, staff roles and sub-contracting, this seemingly simple structural relationship is in reality immensely complex. On organisation charts this complexity is indicated by dotted lines, influence arrows and circles, two-way authority links, and other shorthands which represent a host of non-hierarchical relationships. Management by consensus, collegial relationships, professional boundaries and rights, job responsibilities limited by union contracts, independent ombudsmen, or central agency arbiters further complicate these relationships.”
The functions remain constant in their character while they are flexible with regard to their institutions. For the archival description this means, however, that interconnective description has grown into a necessity. The interinstitutional coherence of the accrual originating in a functional provenance requires finding aids which depict functions in a larger context than merely in the radius of operation of a single player.
The visualisation of relationships on the basis of digital representations is very simple from a technical point of view. The technical possibilities and the results of the digitisation give archivists as well as users the chance to illustrate complex describing systems with complex information on a descriptive level with the help of links and graphics and at the same time provide them with a direct access to the described objects and to present them in the respective structures which shall be highlighted with the respective description.
I thank you for your attention with this summarising link to a slide about functions of the new Europeana Data Model.
 Sue McKemmish, Placing Records Continuum Theory and Practice. In: Archival Science 2001, p. 354.
 Tom Nesmith, Documenting Appraisal as a Societal-Archival Process: Theory, Practice, and Ethics in the Walk of Helen Willa Samuels, In: Controlling the Past – Documenting Society and Institutions, Essays in Honor of Helen Willa Samuels, hg. v. Terry Cook, Chicago 2011, p. 31-50, in particular p. S.34.
 Heather MacNeil, “The Context is All: Describing a Fonds and its Component Parts According to the Rules for Archival Description”, in: Terry Eastwood (ed), The Concept of the Fonds: From Theory to Practice. (Ottawa, 1992) 198-224.
 Terry Cook, What is Past is Prologue: A History of Archival Ideas Since 1898, and the Future Paradigm Shift. In: Archivaria 43.1997 [http://journals.sfu.ca/archivar/index.php/archivaria/article/view/12175/13184].
 Peter Horsman, Wrapping Records in Narratives: http://www.its-arolsen.org/fileadmin/user_upload/Dateien/Archivtagung/Horsman_text.pdf
 Cf. Geoffrey Yeo, The Conceptual Fonds and the Physical Collction. In: Archivaria 73. 2012, pp. 43-80: In this essay he also presents a comprehensive overview about the discussion on the conceptual fonds starting with Terry Cook’s theories.
 DACS 2013, hg. v. SAA, S. xvi.
 Helen Willa Samuels, Varsity Letters: Documenting Modern Colleges and Universities, 1992, Reprint SAA, Lanham (Md.), 1998
 Bearman, Lytle, The Power of the Principle of Provenance. In: Archivaria 21 (1985/1986), S. 16-17.