UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Function-based records classification systems : an exploratory study of records management practices… Foscarini, Fiorella 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2009_fall_foscarini_fiorella.pdf [ 1.34MB ]
Metadata
JSON: 24-1.0067282.json
JSON-LD: 24-1.0067282-ld.json
RDF/XML (Pretty): 24-1.0067282-rdf.xml
RDF/JSON: 24-1.0067282-rdf.json
Turtle: 24-1.0067282-turtle.txt
N-Triples: 24-1.0067282-rdf-ntriples.txt
Original Record: 24-1.0067282-source.json
Full Text
24-1.0067282-fulltext.txt
Citation
24-1.0067282.ris

Full Text

FUNCTION-BASED RECORDS CLASSIFICATION SYSTEMS. AN EXPLORATORY STUDY OF RECORDS MANAGEMENT PRACTICES IN CENTRAL BANKS  by FIORELLA FOSCARINI  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE STUDIES (Library, Archival and Information Studies)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  June 2009  © Fiorella Foscarini, 2009  Abstract Records management and archival theory recommends that records classification, as a means to identify and organize the records made or received in the course of business, should be based on an analysis of the records creators’ functions and activities and reflect them. However, the purpose of classification, the meaning of the term function, and the methodology for conducting a business analysis are not clearly explained in the relevant literature. Additionally, no studies of actual applications of the functional approach to records classification in real organizational settings exist. This dissertation addresses the question of how the concept of function and the functional approach to records classification are understood by those who are responsible for the development and implementation of records classification systems as well as by the users of such systems. In order to contribute insights that can enrich the theory and methodology of records classification, an empirical, interpretivist research design, based on an initial survey of potential study subjects and a multiple-case study research, was conducted in four selected central banks in Europe and North America. One of the selection criteria was that the organizational cultures of the case study sites had to be as heterogeneous as possible. Findings showed that the meanings of function, functional approach, and even classification are subject to various interpretations, that classification developers find functional methodologies confusing, and that users do not usually appreciate the outcomes of their efforts. Furthermore, because the approach to classification was not always consistent with the nature of the records, some of the classification systems examined did not adequately serve either records management or business-related ii  purposes. The research also provided an explanation of the relationship between organizational culture and the understanding of both records management and business processes.  iii  Table of Contents Abstract ............................................................................................................................... ii Table of Contents ............................................................................................................... iv List of Tables .................................................................................................................... vii Acknowledgements.......................................................................................................... viii Dedication ........................................................................................................................... x 1. INTRODUCTION ........................................................................................................ 1 1.1 Overview................................................................................................................. 1 1.2 Identification of the Research Problem .................................................................. 1 1.3 Research Purpose and Approach Taken ................................................................. 5 1.4 Research Hypotheses or Propositions..................................................................... 7 1.5 Research Questions ............................................................................................... 10 1.6 Theoretical and Philosophical Framework ........................................................... 12 1.7 Research Methodology Overview ........................................................................ 16 1.8 Summary and Dissertation Structure .................................................................... 20 2. LITERATURE REVIEW ........................................................................................... 22 2.1 Overview............................................................................................................... 22 2.2 Review of Archival Literature .............................................................................. 23 2.2.1 Early Classification Methods: From the Origins to the German System ....... 23 2.2.2 Classification under Napoleon and Analysis of the Italian System ................ 26 2.2.3 Function and Structure in the Time of Modern Bureaucracy ......................... 30 2.2.4 Schellenberg and the Functional Analysis ...................................................... 33 2.2.5 Theory and Practice of Functional Classification in Canada.......................... 37 2.2.6 The Australian Approach to Recordkeeping: from Maclean to Records Management Standards ................................................................................... 43 2.2.7 Records Classification in the Most Recent Archival Literature. The Role of Classification in an Electronic Environment .................................................. 53 2.2.8 The Functional Approach in the Literature on Appraisal, Arrangement and Description, and Access.................................................................................. 60 2.3 Review of Organization Studies and Library Literature....................................... 76 2.3.1 Organizational Culture and Different Views of Organization........................ 77 2.3.2 Systems Approaches to the Study of Organizations ....................................... 84 2.3.3 Understanding Technology and Organizational Change through the Theory of Structuration ............................................................................................... 89 2.3.4 Organizational Behaviour and Related Issues ................................................ 95 2.3.5 Classification in Library and Information Science ......................................... 99 2.4 Summary ............................................................................................................. 102 3. RESEARCH DESIGN .............................................................................................. 105 3.1 Overview............................................................................................................. 105 3.2 Selecting Suitable Case Study Sites ................................................................... 105 3.2.1 Selection of Study Population ...................................................................... 106 iv  3.2.2 Survey Design and Administration............................................................... 108 3.3 Case Study Design and Implementation ............................................................. 111 3.3.1 Data Collection Methods .............................................................................. 113 3.3.2 Data Analysis Methods ................................................................................. 116 3.4 Ethical Issues ...................................................................................................... 120 3.5 Summary ............................................................................................................. 122 4. STUDY SETTING CHARACTERIZATION .......................................................... 124 4.1 Overview............................................................................................................. 124 4.2 Goals and Functions of a Central Bank .............................................................. 124 4.2.1 The Principles of Modern Central Banking .................................................. 125 4.2.2 Overview of Central Banks’ History ............................................................ 128 4.3 The European System of Central Banks (ESCB) ............................................... 130 4.3.1 The Relationship between the ECB and the National Central Banks ........... 133 4.3.3 Accountability and Transparency of the ‘Independent Central Bank’ ......... 135 4.4 Summary ............................................................................................................. 139 5. SURVEY RESULTS AND CASE STUDY PRELIMINARIES ............................. 141 5.1 Overview............................................................................................................. 141 5.2 Analysis of Survey Results ................................................................................. 141 5.2.1 Records Management-Related Questions ..................................................... 142 5.2.2 Organizational Culture-Related Questions ................................................... 145 5.2.3 Looking for the ‘Ideal Combination’............................................................ 147 5.3 Initial Reactions of Selected Organizations and their Consequences ................. 149 5.4 Summary ............................................................................................................. 154 6. COMPARATIVE ANALYSIS OF CASE STUDY FINDINGS ............................. 156 6.1 Overview............................................................................................................. 156 6.2 Records Management and Archival Framework ................................................ 157 6.2.1 The Role of Recordkeeping within the Organization. A Case-By-Case Analysis ........................................................................................................ 158 i. Organization A........................................................................................... 159 ii. Organization B ........................................................................................... 173 iii. Organization C ........................................................................................... 185 iv. Organization D........................................................................................... 196 6.2.2 Records Management Processes. A Cross-Case Analysis ............................ 208 6.2.3 Summary and Analysis of Common Themes ............................................... 224 6.3 Classification and Functions ............................................................................... 229 6.3.1 Purpose of Classification .............................................................................. 230 6.3.2 Descriptions of Functional Classification Systems ...................................... 241 6.3.3 Functional Approaches to Classification ...................................................... 258 6.3.4 Summary and Analysis of Common Themes ............................................... 276 7. CONCLUSION ........................................................................................................ 283 7.1 Overview............................................................................................................. 283 7.2 Discussion of Research Hypotheses and Questions ........................................... 283 7.3 Strengths and Limitations of Research Design ................................................... 292 v  7.4 Future Research .................................................................................................. 294 7.5 Final Thoughts .................................................................................................... 295 BIBLIOGRAPHY .......................................................................................................... 297 APPENDICES ................................................................................................................ 312 Appendix 1: Invitation Letter and Questionnaire ..................................................... 312 Appendix 2: Survey Results ..................................................................................... 316 Appendix 3: Case Study Invitation Letter ................................................................ 319 Appendix 4: Consent Form ...................................................................................... 320 Appendix 5: Interview Guide ................................................................................... 324 Appendix 6: Observation Guide ............................................................................... 333 Appendix 7: UBC Behavioural Research Ethics Board – Certificate of Approval (Survey) ............................................................................................... 334 Appendix 8: UBC Behavioural Research Ethics Board – Certificate of Approval (Case Study)......................................................................................... 336 Appendix 9: UBC Behavioural Research Ethics Board – Certificate of Approval (Case Study - Amendment) ................................................................. 338  vi  List of Tables Table 1: Questionnaire outcome sorted out by geo-cultural/political areas …………...148  vii  Acknowledgements Joining the Ph.D. program of the School of Library, Archival and Information Studies (SLAIS) at UBC would not have been possible without the loving support of my family and especially Michael, who decided to marry me, despite being aware of what living with someone who studies all the time means. I should also thank my employer, the European Central Bank, for authorizing me to fulfill my wish to go back to academia. I am particularly grateful to my then Head of Division Roman Schremser, who did not hesitate to recognize the importance of my personal and professional development and always supported any of my study-related requests for leave over the last four years. It was a fortunate event that, when I was in the midst of my study program, Stuart Orr became the Head of the ECB’s Archives and Records Management Section, as his strong interest in theoretical and methodological issues allowed a sort of continuity between study and work time. My student life at UBC could not have been more fruitful. I had the privilege to be guided throughout all the stages of the doctoral program by Luciana Duranti, a worldwide renowned scientist whose contributions to the archival discipline have been a permanent source of rich insights for me since I was a novice in the field. By working with her, I found out that Luciana, besides being an excellent supervisor and an inspiring mentor, is a very caring and ever-available person. I could have never overcome the moments of discouragements I had during the intense months spent over my dissertation if her reassuring words and thoughtful advice did not come to me exactly when I needed them.  viii  It has been a privilege to work with the other members of my examination committee as well. I had the luck to be at SLAIS when Terry Eastwood was still teaching there and I could benefit from his invaluable suggestions and heartily encouragements until the very last reviews of my dissertation. Ron Cenfetelli from UBC-Sauder School of Business introduced me to topics and methodologies derived from various sociological and organizational theories, which broadened my disciplinary background and allowed me to take different perspectives in my investigations of the records creators’ reality. The comments made following my dissertation defence by Angela Reddish from the Faculty of Economics at UBC, Victoria Lemieux from SLAIS, and my external examiner Barbara Craig from the Faculty of Information at the University of Toronto were essential to clarify a few aspects of my dissertation. I wish to include in my acknowledgements Heather MacNeil, who has always been available and generous in providing me with her precious advice and new ideas for my research, Francesca Marini, whose encouragement never let me down, as well as Ingrid Germani from the State Archives of Bologna, a unique point of reference for my profession and my life. Among the fellow students who supported me as true friends, I would like to mention Yvonne Loiselle, Genevieve Shepherd, and Sherry Xie. I wish I had the possibility to spend more time with them and the other Ph.D. students during these years. I would finally like to thank sincerely all the people who agreed to participate in my research as case study subjects. I wish I could mention their names because each of them has been so kind, helpful and full of sympathy for my hard work that they would all deserve to appear in alphabetical order on this page.  ix  Dedication  To the silent presence always next to me  x  1.  INTRODUCTION  1.1  Overview  This introductory chapter frames the issues at the core of this study, outlines justifications for the approach taken, and presents the hypotheses (or propositions) and research questions that have guided this author throughout her research. It offers an overview of the theoretical framework in which the study is situated, as well as the overall methodology employed. The last paragraph describes the structure of this dissertation by providing a summary of its chapters.  1.2  Identification of the Research Problem  The nature of a record and the relationships among records and between them and the activities from which they result are subjects of continuing debate. 1 The fundamental assumption on which this dissertation is based is that records, as instruments and byproducts (or residue) of practical activities, accumulate naturally and necessarily in a specific fashion that is determined by the ways in which the activities originating them are being carried out. 2 What distinguishes an archives (in the sense of a plurality of 1  See, among the latest discussions on this topic, Geoffrey Yeo, “Concepts of Record (1): Evidence, Information, and Persistent Representations,” The American Archivist 70 (Fall/Winter 2007): 315-43; Id., “Concepts of Record (2): Prototypes and Boundary Objects,” The American Archivist 71 (Spring/Summer 2008): 118-43. 2 According to the principles of archival theory, what characterizes the nature of records and archives, and qualifies archival science as an autonomous discipline, is this fundamental idea of a record as a ‘byproduct,’ in the sense of an unintentional outcome of a practical activity rather than a purposeful product of it. See the definition of record in Luciana Duranti, Terry Eastwood, and Heather MacNeil, Preservation of the Integrity of Electronic Records (Dordrecht: Kluwer Academic Publishers, 2003), 11: “A record is any document created by a physical or juridical person in the course of practical activity as an instrument and a by-product of it.” Similarly, Italian archivist Valenti defines archives as “the residue of practical activities.” See Filippo Valenti, “Riflessioni sulla natura e sulla struttura degli archivi,” Rassegna degli Archivi di Stato 15 (1981), 22. The relationship between record and activity (or action) is the ground on which American archivist  1  records) as an organic whole, or universitas rerum 3, that is, an entity structured according to the contingent circumstances of its creation, from a mere collection or sum of single items, artificially brought together for accomplishing any external purposes, is exactly this original, necessary, and incremental link (known as “archival bond” 4) existing among all records that belong together because they are originated during, and by virtue of, the same activity or business process. These tenets of archival science yield some important consequences for the ways in which active records (i.e., the records that are being used to carry out ongoing activities) are, or should be, interrelated and arranged in the records creators’ offices (or ‘living,’ current archives). First, the process of identifying and organizing the records that accumulate in the course of business, for instance, by means of classification, must be determined by the circumstances of records creation. A records classification scheme whose content and structure reflect the specific functions and work processes of any individual records creator allows records to be physically and/or logically aggregated in units (e.g., files, or dossiers, or series) that will be capable of revealing the meaning of the relevant records and the actions carried out through them to whoever will be looking at those units. 5  Schellenberg bases his description of how records aggregate: “Records are the by-product of action, and they naturally fall into groups that relate to action.” See Theodore R. Schellenberg, Modern Archives: Principles and Techniques (Chicago: University of Chicago Press, 1956), 53. 3 See Giorgio Cencetti, “Sull’archivio come ‘universitas rerum’,” Archivi IV (1937): 7-13. Reprint. Scritti archivistici (Rome, 1970), 47-55. 4 See Luciana Duranti, “The Archival Bond,” Archives and Museum Informatics 11 (1997): 213-18. 5 The next chapter, dedicated to a review of relevant literature, will describe the origins of classification as an administrative and archival tool. The idea of aggregating records in functional units, so that the original context of records creation can be captured and preserved, was described by British archivist Jenkinson in 1922 as follows: “The golden rule for the Administrator, so far as concern his papers, must be to have them always in such a state of completeness and order that, supposing himself and his staff to be by some accident obliterated, a successor totally ignorant of the work of the  2  Second, every records classification scheme is necessarily different from any another, although we can expect some similarities among those records creators that are entrusted with the same functions. In any case, one shall not superimpose any artificial or pre-established classification scheme to an existing accumulation of records, even where such a scheme might improve access to the records, because, by so doing, the records’ relationships would inevitably be altered or obscured. This does not mean that retrieval is not relevant to records classification. It is relevant, as a collateral benefit, to the extent that it does not contradict the primary purpose of records classification, which is, “to place individual records into the aggregates to which they belong, based on the creator’s mandate and functions.” 6 According to archival science, the practice of classifying records comes from the need to make explicit that ‘archival bond’ that exists among all the records participating in the same activity since the moment of their creation, as well as the broader documentary, procedural, and provenancial contexts characterizing and thus uniquely identifying each record. Through the act of classification, the network of relationships inherent in the nature of any record not only is brought to light, but it is also established and perpetuated. In this way, the meaning of each record in relation to all the others as well as the structure of the whole of records (i.e., the archival fonds) can be understood and transmitted over time. From what has been said, it emerges that records classification is an important method of procedural control over records creation (thus contributing to the records reliability), as well as a critical means for the identification of records in context over office would be able to take it up and carry it on with the least possible inconvenience and delay simply on the strength of a study of the Office Files.” See Hilary Jenkinson, A Manual of Archive Administration, 2nd ed. (1937), reprint (London: Percy Lund, Humphries & Co., 1965), 153. 6 Duranti et al., Preservation of the Integrity, 43.  3  time and space (thus contributing to establishing and maintaining the records authenticity). 7 While in the paper world the archival bond could manifest itself through the physical arrangement of the records and, thanks to the numerous signs inscribed on the paper (e.g., annotations, signature, etc.) and other elements of form, a diplomatic analysis could help reveal the context of records creation even after subsequent rearrangements of the archives had occurred, in today’s electronic environment the “physicality” 8 of traditional records and the implicit information conveyed by it do not exist any longer, and this makes records classification an even more essential tool than it ever was in the past. Additionally, by deducing from the primary purpose attributed to classification what a classification system should look like and how it should behave, it appears that a study of the functions and activities of a records creator is a prerequisite for the design of any records classification system. In other words, a functional approach to records classification development is justified by the nature of the records. However, there has been little theory building on the topic of functional classification 9 and classification practice demonstrates that the principles that should guide the design and implementation of records classification systems are generally not well understood by those entrusted with such a task, whether they are archivists or records managers. This statement is based on evidence provided by the variety of outcomes of uneven quality that diverse and inconsistent classification methods have produced both in Europe, where the fundamental ideas of records classification and filing  7  Ibid. David Bearman, “Item Level Control and Electronic Recordkeeping,” Archives and Museum Informatics 10, 3 (1996): 220. 9 For the purposes of this dissertation, functional classification may be defined as: “The process of devising and applying schemes based on the business activities which generate records, whereby they are categorized in systematic and consistent ways ....” See Australian recordkeeping Standard AS 4390 (1996). 8  4  have been first formulated, and in North America, where the debate around classification approaches is more recent. The literature review presented in the next chapter will show that the role and characteristics of records classification are often misinterpreted not only by practitioners but also by those who write about those matters for the sake of the latter. It will also reveal that the meaning of function, activity, business process, and the like lack a thorough elaboration and the methodology for analyzing them in organizational contexts is not well described either.  1.3  Research Purpose and Approach Taken  The primary goal of this research was to enhance our understanding of the concept of function and the functional approach as a methodology for the development and implementation of records classification systems. Because any method or means has its justification in the purpose or end that one wishes to achieve through it, an investigation of the purpose(s) of classification in general was also undertaken as a major component of this research. The study of both meanings – the one of function and the other of classification – involved an in-depth review of the literature concerning the functional approach as a records management and archival methodology relevant not only to records classification but also to appraisal and selection, description and arrangement, and access to information. As the functions and activities that interest this research are mainly those carried out in business environments, this author felt the need to expand the literature review delving into the territory of other disciplines, such as theory of organization, sociology, social-psychology, management science, and theory of administration. Likewise, because library and information science have also explored the topic of  5  classification, though from a perspective different from the archival one, it appeared relevant to include some ideas on the purpose of classification developed in that area of knowledge as well. The hypotheses and research questions formulated at the beginning of this study, and which are introduced in the next sections, mostly come from the analysis of the literature mentioned above. In part, they were also inspired by this author’s observations of existing functional (or claimed to be as such) classification systems and by her own experiences as a records classification developer. 10 The literature review findings also provide a justification for the approach taken by this researcher to answer the research questions so expressed. The complete absence of empirical studies on the design and application of function-based classification systems in use in real-world organizations convinced her that an inductive, interpretivist approach to the issues at stake would be the most appropriate to try to get new insights. Thus, through an exploration of the adoption and enactment of records management  10  This author, in her professional capacity as an archivist (a profession that in most countries of continental Europe involves records management responsibilities as well), was entrusted with the task of developing a classification system for the records of both the institutions she has been working for, namely the Province of Bologna in Italy (from 1995 to 2000) and the European Central Bank in Frankfurt am Main, Germany (from 2000 to present). Interestingly, in both cases, her mandate was the same, i.e., to design a function-based classification system; however, due to differences in the juridical and administrative framework, organizational structure, corporate culture, and expectations of the two entities, each experience was unique and very dissimilar from the other. The outcomes of her efforts were indeed substantially diverse not just content-wise (as a consequence of the distinct mandates and functions of either institution), but also structurally. The implementation phase also involved unequal challenges in both organizations, in virtue of their rather dissimilar administrative cultures as well as the different records management skills and attitudes of the respective system users. From these experiences, this author realized how important it is to agree with all the parties involved (e.g., area managers, users, IT experts, and colleagues within the records management and archives department) on the objectives that records classification tools are intended to achieve. The results of any analysis of an organization’s functions, activities, and transactions for purposes of classification rest on this understanding. Additionally, she became convinced that the type of organizational settings and cultures (e.g., hierarchical vs. flat; writingbased vs. meeting-based; working according to standardized workflows, routine processes and sequential procedures vs. working according to unstructured or semi-structured procedures and creative processes; service-oriented vs. knowledge-oriented; etc.) has an important impact on the way function and functional analysis are interpreted and applied.  6  concepts and practices in some selected organizational settings, this author attempted to shed some light on how people interpret and use their functional tools, with the conviction that from such ‘grounded knowledge,’ some new theoretical and methodological understandings of classification and function might emerge. Given the fact that this research is situated in the domain of an applied science, it was expected that, from its exploratory and explanatory aims, some practical outcomes would derive as well, for instance, in the form of recommendations for records professionals on how to design, implement, and use records classification systems that would eventually be able to meet their purposes.  1.4  Research Hypotheses or Propositions  As the section on research methodology included in this chapter will further explain, an interpretivist paradigm, such as the one framing this study, does not usually concern itself with the testing of hypotheses, and new hypotheses, or “working propositions” are expected to be generated from the analysis of the data collected during field work and observations. 11 However, as Williamson writes, “For some interpretivist studies, researchers develop propositions which are similar to research hypotheses. They do not require such precise wording, nor the rigorous testing associated with operational hypotheses. They, nevertheless, can help to provide a similar kind of clarification as hypotheses give to a quantitative study.” 12 The following hypotheses have to be read against this background, thus considering them as guiding devices that should primarily assist the development of this research, especially in its initial stages. 11  See Kirsty Williamson, ed., Research Methods for Students, Academics and Professionals. Information Management and Systems, 2nd ed. (Wagga Wagga, New South Wales: Centre for Information Studies Charles Sturt University, 2002), 26-32. 12 Ibid., 57.  7  Hypothesis 1: The way in which most of today’s archival literature interprets, describes, and prescribes the ‘functional approach’ does not help practitioners design and implement records classification systems that work. This criticism might be extended to the records management and archival formal education that one obtains by attending relevant courses and which would not be effective in teaching how to understand the essential features of real-world organizations and how to conduct a business analysis. Hypothesis 2: A records classification system does not have to be exclusively based on an analysis of an organization’s functions, activity, and transactions; other, ‘non-functional factors’ that might affect records creation must as well be taken into account. The message that the archival literature seems to transmit is that the functional criterion is to be applied as an exclusive, absolute principle. On the basis of her experiences and her readings outside the archival domain, this researcher came to the conclusion that the needs of the users, certain organizational structures, laws, regulations and other constraints might influence the ways records accumulate, and this usually is, and should be, reflected in the classification system. It is expected that this empirical study will reveal other, more specific non-functional factors that should be regarded as relevant by classification developers. Hypothesis 3: Treating every classification issue as a functional one is not appropriate, in that there are realms of human activity that cannot be categorized through functional lenses. Through her analysis of the literature relevant to the cultures existing in work places, organizational behaviours, and administrative processes in particular, this researcher 8  supposed that one cannot assume that the way in which work is carried out is, in every instance, so rational and recursively structured that the records professional, like an engineer, would be able to draw a function-based tree where every entry matches perfectly with an actual process or a phase of a process. The social reality would be much more complex and articulated, and human activities much more unpredictable and creative than the one represented in the logically structured hierarchy of business functions and processes described in the archival literature. Hypothesis 4: The organizational culture characterizing every work setting influences the way in which both the purpose of classification and the concept of function are understood. Such diverse interpretations are reflected in the structure and substance of existing records classification systems as well as in the ways those systems are enacted by the users in every different organizational context. This hypothesis derives from the observation that, due to the fact that they are insufficiently described, function and classification are both ambiguous concepts and, as such, they would be prone to be interpreted in different ways according to the organizational cultures under examination. In order to ‘test’ this hypothesis, this researcher chose specific organizational settings where to conduct her field work, so that each of them would potentially display a different organizational culture. Hypothesis 5: The implementation phase of a new, or revised, records classification system (including users’ involvement in testing activities, training, etc.) will have a crucial impact on the system’s acceptance and the ways in which the system will be interpreted and used within the organization.  9  The importance of users’ involvement during system deployment and the issue of training, of learning how and why to use the system in a given way, are emphasized by a certain literature exploring, from a structurational viewpoint, the relationship between technology and organization, and are central to this research.  1.5  Research Questions  The study here introduced mainly refers to qualitative methods of inquiry, thus the research questions formulated on the basis of the research purpose and the hypothesis mentioned above will necessarily be broader and more flexible than those that are normally used in quantitative research designs. Marshall and Rossman suggest that, in a qualitative study, “research questions should be general enough to permit exploration but focused enough to delimit the study. Not an easy task.” 13 Considering the primary goal of this research, a major operational research question was articulated as follows: Major Question: How do people in organizations understand the concept of function and the functional approach as a methodology for the design and implementation of records classification systems? Two broad categories of subjects potentially sharing different views on the topics under examination were identified, namely, the system ‘developers’ (i.e., archivists, records managers, members of project teams, or anyone else entrusted with the task of designing, maintaining, and/or implementing the records classification system in use in the organization) and the system ‘users’ (i.e., both ordinary and specialized users of records  13  Catherine Marshall and Gretchen B. Rossman, Designing Qualitative Research, 2nd ed. (Thousand Oaks: Sage Publications, 1995), 26.  10  classification systems, including, among the first group, area managers, experts, secretaries, etc., and among the second group, system administrators, records managers not involved in the development of the system, etc.). Taking this distinction into account, and with the purpose of breaking down the major question into more manageable units, the following sub-questions, or specific research questions, were elaborated. Sub-Question 1: What knowledge do developers and users respectively have of records classification theory and methods, and what expertise do they have in the practice of classifying? Sub-Question 2: Do non-functional factors influence the design of function-based records classification systems? If they do, how is such an influence exercised and why? Are developers aware of it and, in case they are, what are their opinions about it? Sub-Question 3: How is/was business analysis carried out in the organizations under investigation? Can/Could all activities be described in terms of structured business processes, or are there activities that just can/could not fit in? Sub-Question 4: How are business processes and functions perceived in relation to the characteristics of each organizational setting and culture? Sub-Question 5: How do users appropriate (i.e., adopt and adapt) the records classification system existing in their organization? How is/was their participation in the design and implementation phases of the system? How is/was user training provided? Both the major research question and the sub-questions have an operational nature, in the sense that they are asked in order to gain an understanding of the issues under 11  consideration, to start ‘mapping the unexplored territory’ of this research. Above them, on a more general level, this researcher identified two ‘ultimate questions’ that the answers to the previous questions would ideally allow her to respond to. Ultimate Question a: Why does the functional approach appear so difficult to apply? Ultimate Question b: How can records classification possibly be improved in the organizations under investigation, and in general?  1.6  Theoretical and Philosophical Framework  One of the components of the theoretical framework of this study involves the concepts, principles, and methods of archival science, with particular regard to those concerning records classification. These, including the ideas relevant to functional analysis for purposes of classification, which records keepers have been dwelling upon especially in the last few decades, will be thoroughly examined in the context of the literature review of the next chapter. As to the relationship between theory and practice in the archival (and records management) work, Trevor Livelton suggests: “It is reasonable to argue that archivists’ ideas inevitably underlie their practice, whether they are aware of it or not. Archivists may not necessarily employ fully developed concepts in their work (although they may); they may not necessarily be aware of those ideas (although they may); and the ideas they do employ … may have more of a methodological than a theoretical cast. This is not to say that theory, in the full meaning of the term, always underlies practice, but that ideas always do – and ideas are the stuff that theory is made of.” 14 Archival science is an applied discipline and, as such, its “methods, that is, orderly, logical, and systematic modes of procedure, act as a bridge between theory and  14  Trevor Livelton, Archival Theory, Records, and the Public (Lanham, Md. and London: The Society of American Archivists and The Scarecrow Press, 1996), 34-35.  12  practice.” 15 Although an underlying theory of classification may not have been fully articulated by the archival scholars and practitioners whose writings will later be reviewed, as Livelton observed, methods to organize records in the active phase of their life cycle have nevertheless been developed and used. In particular, the functional approach as a classification method seems to be considered suitable to the nature of records and the purposes of classification by those who have been devising such means. By analyzing the few theoretical works existing on the subject, but also the “discourse about methods,” 16 one can expect to find some enduring ideas about the purpose of classification, the meaning of function, as well as the nature and characteristics of the material being classified. Nevertheless, the novelty of this study lies in its inductive approach. As already emphasized, the practice of records classification has never been subject to any in-depth examination aiming at exploring how classification is actually understood by those using it in the ordinary course of business, or why a functional approach should be preferred to any concurrent methods (e.g., subject-based, organizational structure-based, record typebased) on the basis of the study of concrete outcomes of its application. By investigating current practices in real-world settings, including the needs and expectations of those developing and using classification systems which are, or at least are claimed to be, function-based, as well as successful and unsuccessful stories of system adaptations and adoptions, new ideas will ideally unfold. Such ideas may contribute to building a new  15  Terry Eastwood, “The Theory and Practice of Description in the Digital Era” (paper delivered at the Second Meeting on Archival Information Databases, sponsored by the Brazilian Society of Archivists, Rio de Janeiro, March 16, 2007), 2. 16 Ibid., 3.  13  theory, 17 or will at least enrich the landscape of the object under investigation by “piec[ing] together observed data, elements drawn from different frameworks … in order to gain access to the domain to be charted” 18 – in the words of Wolfgang Iser, describing the purpose of “soft theory” (i.e., the humanities and social sciences) in contrast with that of “hard-core theory” (i.e., the physical sciences). “… the humanities are not a problem-solving undertaking. Instead, their prime concern is to achieve understanding, to assess context-relatedness, to investigate meaning and function, to evaluate [the object of attention (i.e., in this case, records classification)], to address the question of why we need [it].” 19 This study shares the interpretivist approach to reality that goes back to the intellectual traditions of phenomenology and hermeneutics, where the prime datum is not the world external to the observer, but the observer’s mental process. Human beings in social processes are constantly creating their world in interaction with others. Thus, there is no “pre-given universe of objects,” but one which is “produced by the active doing of subjects.” 20 Such philosophical stance characterizes what Peter Checkland calls “soft systems thinking,” 21 a holistic way of looking at, and trying to make sense of, the “organized complexity” of social reality, that, since the mid-1950s, has been developed as an alternative to any form of “hard systems thinking” (e.g., systems engineering and systems analysis) with the aim of complementing the reductionism of the positivist approach. “Soft” systems concepts have informed this study, especially with regard to the interpretation of the research findings. 17  See Kathleen M. Eisenhardt, “Building Theories from Case Study Research,” The Academy of Management Review 14, 4 (October 1989): 532-50. 18 Wolfgang Iser, How to Do Theory (Malden: Blackwell Publishing, 2006), 5. 19 Ibid., 7. 20 Anthony Giddens, New Rules of Sociological Method (London: Hutchinson, 1976), cited by Peter Checkland, Systems Thinking, Systems Practice (Chichester, West Sussex, UK: John Wiley & Sons Ltd., 1981), 277. 21 See Peter Checkland, Systems Thinking, Systems Practice.  14  Another theoretical point of reference that should be mentioned here is Giddens’ theory of structuration and, in particular, Adaptive Structuration Theory (AST), 22 which draws on the concepts of structuration to examine the interplay existing between human action, social structures, and advanced information technologies. For the purposes of this study, records classification systems and the electronic document and/or records management systems (EDMS or EDRMS) in which the former are usually embedded in today’s organizational settings can both be considered advanced information technology in a structurational sense. According to Wanda Orlikowski, technology is one of the possible “instantiations of some of the structural properties of organizations” 23 and, as such, it is both structural and socially constructed. The concept of “duality of technology,” that she deduces from Giddens’ “duality of structure,” allows us to see technology as created and changed by human action (i.e., a product) and, at the same time, as a structure that both facilitates and constraints human action (i.e., a medium). 24 In this context, the relationship between technology and organizational change acquires a new, dynamic, and anti-deterministic meaning. Far from discussing the ‘impact’ of any given technology, AST focuses on the mutual influence of technology and social processes, on the always different outcomes that emerge when technology is enacted by human agents in any specific context. It is this author’s conviction that the recursive notion of technology and the emphasis of AST on the role played by the users may enrich our understanding of ‘archival technology.’ Indeed, as it will emerge 22  See Gerardine DeSanctis and Marshall S. Poole, “Capturing the Complexity in Advanced Technology Use: Adaptive Structuration Theory,” Organization Science 5, 2 (May 1994): 121-47; Wanda J. Orlikowski, “The Duality of Technology: Rethinking the Concept of Technology in Organizations,” Organization Science 3, 3 (August 1992): 398-427. 23 Orlikowski, “The Duality of Technology”, 405. 24 See Ibid.; JoAnne Yates and Wanda J. Orlikowski, “Genres of Organizational Communication: A Structuractional Approach to Studying Communication and Media,” Academy of Management Review 17 (1992): 299-326.  15  throughout the research report presented in Chapter 6, the explanatory power of AST has contributed new insights into the process of ‘appropriation’ with reference to the way in which users influence, and are in turn influenced by, the technology (whether the classification system or the ERDMS) they are using.  1.7  Research Methodology Overview  The methodology employed in the inductive, empirical research here introduced is mainly, but not exclusively, qualitative, and involves fieldwork (i.e., “the study of the phenomena under consideration in their natural setting” 25) as a major component. Besides a survey which was launched at the earliest stage of the project in order to select suitable cases from the chosen population and which consisted of factual and closed questions, 26 the main body of this research is based on a multiple-case study design. Because this study is concerned with ‘meaning,’ with the ways people ‘make sense’ of their world, from the very beginning, an interpretivist case study approach seemed to be the most appropriate means to answer the ‘how’ and ‘why’ questions listed above. Thanks to the “immersion in the milieu of the subjects under examination” 27 that it allows, ethnography might have proved to be a more effective strategy for the purpose of collecting data on the unstated practices shared among those subjects (i.e., their ‘tacit knowledge’) and eventually elaborating a “thick description” 28 of their cultures (e.g., in the present case, their recordkeeping cultures). However, because of constraints provided  25  Williamson, Research Methods, 31. See A.N. Oppenheim, Questionnaire Design, Interviewing and Attitude Measurement, new edition (London, New York: Continuum, 1992). 27 G. E. Gorman and P. Clayton, Qualitative Research for the Information Professional: A Practical Handbook (London: Library Association Publishing, 1997), 23. 28 Clifford Geertz, “Thick Description: Toward an Interpretive Theory of Culture,” in The Interpretation of Cultures. Selected Essays (New York: Basic Books, 1973): 3-30. 26  16  by the type of organization under examination (i.e., central banks, which, as it is wellknown, have very restrictive access regulations) a full ethnographic approach was not possible to apply. Nevertheless, various opportunities to exploit typical ethnographic techniques, such as, in-depth, unstructured interviewing and participant observation, occurred throughout this study and have thus been integrated in the case study design. 29 Case study research is indeed a flexible, though at the same time rigorous, approach that is particularly recommended “where the experiences of individuals and the contexts of actions are critical … or where terminology and a common language and set of definitions are not yet clear or widely accepted.” 30 As will be evident from the literature review, this is exactly the case with reference to the language of function or the definition of the role of records classification. A detailed description of each qualitative and quantitative method employed in this study and of the comparative approach used for analyzing the data collected is provided in Chapter 3 “Research Design,” being the purpose of this introductory chapter limited to expounding nature, advantages, and limitations of the research approach chosen. As anticipated earlier in this chapter, within an interpretivist paradigm, any initial hypotheses, research questions, as well as any other ideas derived from the review of relevant literature, have a guiding role in that they help to frame the whole project and to plan how to collect the data. The interpretivist researcher seeks, at the same time, to be totally open to the situations and subjects encountered during the enquiry, trying not to impose any pre-existing expectations. This approach would allow new themes to emerge 29  Examples of the use of ethnographic methods in archival research include: Elizabeth Yakel, “The Way Things Work: Procedures, Processes, Institutional Records,” The American Archivist 59 (1996): 454-464; Karen F. Gracy, “Documenting Communities of Practice: Making the Case for Archival Ethnography,” Archival Science 4 (2004): 335-65; Kalpana Shankar, “Recordkeeping in the Production of Scientific Knowledge: An Ethnographic Study,” Archival Science 4 (2004): 367-82. 30 Williamson, Research Methods, 113.  17  from patterns observed in the data collected in situ. To this end, in this project, data analysis has been undertaken throughout the research and not just in its concluding stage, as recommended by several authors. 31 The emphasis of this research is largely exploratory and explanatory. It is exploratory as it focuses on “phenomena that are thus far little-known or understood” 32 (e.g., how the concept of function is interpreted and used in a real work environment; how people in organizations perceive the role of records classification); and it is explanatory by virtue of the attempt made to identify “plausible causal networks shaping the phenomena being studied” 33 (with reference, for instance, to the relationship that seems to exist between specific organizational settings and cultures on the one hand and record-making and -keeping practices on the other). Because generalization of the kind that only nomothetic, positivist research designs allow (i.e., statistical generalization) cannot be achieved from the observation of a limited number of phenomena, one should not expect predictive explanations to proceed from this research – and actually, accomplishing external validity is not necessary within an interpretivist paradigm. However, several measures have been taken to ensure that the results of this study have the highest possible degree of reliability and external validity. First of all, in order to check the consistency of findings, different data collection methods (e.g., interviews, observations, analysis of documentation) have been used (methods triangulation). The reliability of findings is further enhanced by having cross-  31  See Robert K. Yin, Case Study Research. Design and Methods, 3rd ed. (Thousand Oaks: Sage Publications, 2003); Williamson, Research Methods. 32 Marshall and Rossman, Designing Qualitative Research, 39. 33 Ibid., 41.  18  checked for consistency the same information with different people within each participant organization (sources triangulation). 34 “… cases are not ‘sampling units’ and should not be chosen for this reason. Rather, individual case studies are to be selected as a laboratory investigator selects the topic of a new experiment.” 35 As will be shown in detail in Chapter 3, the selection of suitable case study sites for conducting this research has required careful consideration of the nature of the research questions, the unit of analysis, and expected outcomes. In the literature, this sampling technique, typical of qualitative research, is referred to as ‘theoretical sampling.’ 36 Under such circumstances, the kind of generalization one may obtain is ‘analytic generalization.’ The hypotheses developed prior to the conduct of data collection are used as a template with which the empirical results of the case studies can be compared. According to Yin, “if two or more cases are shown to support the same [hypotheses], replication can be claimed.” 37 This is one of the advantages that multiple-case designs offer in terms of external validity of the research findings. Walsham makes the point that ‘theoretical or analytic generalization’ has the potential to generate the following outcomes: “development of concepts; generation of theory; drawing of specific implications; and contribution of rich insight.” 38 Achieving at least the latter is one of the goals of this author. The case study approach has often been criticized on account of the researcher’s inability to be a ‘neutral observer’ of the reality under investigation. This would “limit  34  See Williamson, Research Methods, 36. Yin, Case Study Research, 32. 36 See Williamson, Research Methods for Students, Academics and Professionals, 32; Eisenhardt, “Building Theories from Case Study Research,” 537. 37 Yin, Case Study Research, 32. 38 Geoffrey Walsham, “Interpretive Case Studies in IS Research: Nature and Method,” European Journal of Information Systems 4 (1995): 74-81. 35  19  the validity of the research findings,” 39 although one cannot deny that bias may be inherent in any research strategy. As Sutton notes, “… one can understand something observed only through the tinted lens of one’s own experience.” 40 Interpretivists have turned this apparent weakness into the strength of their approach by demonstrating how the researcher’s point of view can actually become a source of understanding, as long as there is an awareness of it. In other words, “the concern with researcher objectivity is replaced by a focus on the impact of subjectivity on the research process.” 41 Throughout her study, this author has made the effort to take into considerations her personal characteristics, interests, and background as a potential source of bias, and has continuously tried to use them to interpret her findings in an insightful way, so as to “chang[e] analysed data into a contribution to knowledge and debate.” 42  1.8  Summary and Dissertation Structure  The present chapter has provided an overview of the problem being studied and approach taken to deal with it, in relation this researcher’s knowledge of archival theory, her experience as a records classification developer, and the overall goals and objectives of this research. The research hypotheses, or propositions, and research questions identified have been situated within the theoretical and methodological framework of the research. Chapter two contextualizes this research within relevant literature, which is in turn the source of the hypotheses and lines of enquiry guiding this study.  39  Williamson, Research Methods, 113. Brett Sutton, “The Rationale for Qualitative Research: A Review of Principles and Theoretical Foundations,” Library Quarterly 63, 4 (1993): 425. 41 Corinne Glesne and Alan Peshkin, Becoming Qualitative Researchers: An Introduction (White Plains, NY: Longman: 1992), 6. 42 Williamson, Research Methods, 300. 40  20  Chapter three presents an account of the implementation of the two major components of this research design, i.e., a survey and a multiple-case study research. Chapter four describes the main characteristics of the population chosen for this study, i.e., central banks. Chapter five reports on the selection of case study sites by means of the survey and provides an analysis of the initial stages of the case study research. Chapter six involves a detailed, cross-case report of the findings of the multiplecase research, including discussions of the most relevant findings. Chapter seven evaluates the outcomes of the empirical research against the objectives of this study, identifies its main contributions to the records management and archival discipline, discusses strengths and limitations of the research design, and outlines future work.  21  2.  LITERATURE REVIEW  2.1  Overview  The purpose of this chapter is to provide an “orientating framework [aiming at] contextualizing the problem that has led to the need for the research.” 43 To this end, it will start by situating records classification and, in particular, the functional approach to it, in the context of the relevant archival literature. 44 ‘Archival literature’ is here used as an encompassing expression which includes what in the Anglo-Saxon tradition is known as ‘records management literature.’ 45 Because in most recent decades the functional approach has become so pervasive that it is used as a pillar of archival methodology throughout the life cycle of the records, reference to archival functions other than records classification (namely, appraisal, arrangement and description, and access to archives) will be made insofar as they may shed light on the meaning of function in archival science. Observations related to this author’s examination of existing accessible classification systems will also be reported in this chapter as an additional argument for  43  John W. Creswell, Research Design. Qualitative, Quantitative, and Mixed Methods Approaches, 2nd ed. (Thousand Oaks: Sage Publications, 2003), 30. 44 This study looks in particular at the archival literature in English and in Italian, as the former is representative of the Anglo-Saxon and common-law approach to the issues in question, while the latter is representative of the Latin and civil law approach. 45 Due to the global influence of the Anglo-Saxon tradition, the distinction between archival science and records management has also gained ground in continental Europe, where, in the context of each national language, it is not uncommon to come across the Anglicism ‘records management’ or ‘records manager.’ In Italy, for instance, the traditional three stages of a record’s life cycle are termed archivio corrente, archivio intermedio (or archivio di deposito), and archivio storico. The concept underlying this terminology is that, independently of its status (whether active, semi-active, or inactive), any accumulation of records has archival nature. See Ernst M. Posner, Archives in the Ancient World (Harvard University Press, 1972). Reprint (Chicago, IL: Society of American Archivists, 2003). Consistently with this view, the term ‘document’ (in Italian, documento) in Latin countries encompasses the Anglo-Saxon terms ‘document,’ ‘record,’ and ‘archival record,’ because the quality of being ‘archival’ is regarded as inherent in any documentary by-product of the activity of a person or organization (in Italian, produttore d’archivio). See Giorgio Cencetti, “Sull’archivio come ‘universitas rerum’.”  22  adopting an empirical research approach. To provide further support to this methodological choice, literature of other disciplines such as organizational theory, sociology, theory of administration, as well as library and information science, will be presented in relation to some of the lines of inquiry followed by this project.  2.2  Review of Archival Literature  In order to give a sense of the developments throughout the centuries of archival thinking on the topic of records classification, as well as to highlight the various contributions provided by different countries and archival traditions to these developments, the following sections (except for the last one, 2.2.8) have, to the extent possible, been organized chronologically and according to examined traditions. Section 2.2.8 stands out as its topic is not records classification but the functional approach in the literature related to appraisal, arrangement and description, and access to archives.  2.2.1 Early Classification Methods: From the Origins to the German System In the ancient and medieval world, records used to be either spontaneously accumulated as they were sent or received (thus originating so-called “sedimentary archives”) or deliberately selected, always for practical and operational reasons, to make up series consisting principally of legal titles (so-called “treasury archives”) 46. Both systems coexisted in all European chanceries of the modern era where, however, in order to cope with the growing number of administrative activities, a subdivision based on the records state of transmission had to be introduced with reference to the former type of records arrangement. The categories were “records sent,” “records received,” “internal records,” 46  For the definition of “treasury” and “sedimentary” archives, see Valenti, “Riflessioni sulla natura e sulla struttura degli archivi”: 20-24.  23  and “miscellanea.” Such a system soon proved to be inadequate to control the mass of records produced by a bureaucratic machine that was increasingly becoming more complex and articulated. New types of record aggregations started to appear throughout Europe, such as, for instance, series based on the legal nature of the transaction originating the records (e.g., contracts, deliberations) or on the form of the records (e.g., circular letters, invoices). Finally, in the course of the 17th and 18th century, the Prussian state, which was renowned for its administrative efficiency, developed a revolutionary method to organize the records made and received by the government. All records related to the same subject, and secondarily to a given business transaction, activity, or procedure, independently of their status of transmission (i.e., degree of perfection, that is, draft, original, or copy), form, or value, would be incrementally put together in discrete physical and logical units, called dossiers or files, which would then be in turn aggregated organically according to various homogeneous criteria (e.g., names of persons or corporate bodies, geographic units, subject-matters, dates). This system – also known as Registratursysteme 47 – is the first example of a systematic method of classifying records following a comprehensive, subject- and function-based Aktenplan (i.e., file plan). The effectiveness of such a ‘mixed model’ of classification was related to its being a natural way of carrying out administrative work that was itself very rational, linear, and rigorous. Writers discussing the Italian archival tradition have pointed out that a major flaw in the German system was the fact that it was applied a posteriori, ex-post. That means  47  See Thea Miller, “The German Registry: The Evolution of a Recordkeeping Model,” Archival Science 3, 1 (2003): 43-63; Id., “The German Registratur” (Master’s thesis – University of British Columbia, 1997).  24  that explicit links among the interrelated records of a business transaction, for example, were established when the activity the records referred to was concluded and the relevant file was transferred to the central registry, rather than concurrently with the creation of the records and the development of the activity generating them. As a consequence – those critics suggest – the ‘original order’ of the records was somehow artificially created for the sake of administrative control. 48 In fact, the timing of records classification (i.e., throughout the active life of a file, or when the file gets closed) and the responsibility for it (i.e., records creators or staff dedicated to recordkeeping activities) are quite controversial issues. Different trends still exist today, some going back to established archival traditions 49, others just related to work circumstances. Theorists have investigated the conceptual consequences on the nature of the records aggregations that may derive from those two diverse approaches (including the issue of original order mentioned above). 50 However, no inductive research has ever been carried out to analyze the practical consequences and actual effectiveness of either method, or to explore why people in organizations adopt one option instead of the other.  48  See Elio Lodolini, Archivistica. Principi e problemi, 6th ed. (Milan: Franco Angeli, 1992), 76-80; Luciana Duranti, I documenti archivistici. La gestione dell’archivio da parte dell’ente produttore (Rome: Pubblicazione degli Archivi di Stato, Quaderni della Rassegna degli Archivi di Stato no. 82. Ministero per i Beni Culturali e Ambientali, Ufficio Centrale per i Beni Archivistici, 1997), 55-58; Maria Guercio, “Principles, Methods, and Instruments for the Creation, Preservation, and Use of Archival Records in the Digital Environment,” The American Archivist 64 (Fall-Winter 2001): 238-69. 49 For an overview of pre- and post-classification procedures in different archival traditions (namely, those of Germany, England, Australia, and the United States), see Schellenberg, Modern Archives. Part II: “Record Management,” 33-110. 50 With reference to the consequences on the establishment of the archival bond (which will be touched upon later on in this chapter), see Raffaele De Felice, “In margine ad alcune questioni di archivistica,” Rassegna degli Archivi di Stato XXXI, 1 (January-April 1971): 123-42.  25  2.2.2 Classification under Napoleon and Analysis of the Italian System At the beginning of the 19th century, the German system had spread through most of continental Europe thanks to the conquests made by Napoleon, whose administrative apparatus adopted and improved it by combining classification and registration capabilities in one single tool. The systematic identification and pre-organization of all incoming and outgoing correspondence was subsequently extended to cover internal records as well. With these adjustments, classification became the heart of the Napoleonic administrative system, which put great emphasis on recordkeeping. 51 In Italy, the twofold system developed by the Napoleonic administration, called sistema protocollo/titolario (protocol register/classification system), is still regarded as the core component of those recordkeeping systems which qualify as ‘trusted.’ For this reason, its use is mandatory for all public bodies. Entering a record into the system means to certify the exact moment in which the record is issued or received, to identify it in a unique manner, and to place it, by means of classification, within its procedural and documentary context. Within the Italian juridical system, the described register is considered a ‘public act,’ that is, a record itself, to be preserved indefinitely in virtue of its value as ‘the highest evidence’ before the court. 52 The legal value attributed to the act of classifying and registering records (the former always precedes the latter in the Italian system) brings this discussion to the writings of Raffaele De Felice, the Italian archivist who has most extensively discussed 51  See Lodolini, Archivistica. Principi e problemi, 85-90. The higher evidentiary value of the Italian ‘protocol register’ (inclusive of classification system) has been established by the Cassazione Penale (i.e., criminal court of appeal) sentence, sect. V, 6 October 1987. The Italian register should be considered in the context of a civil law system which requires that a specific set of formal elements apt to guarantee the certainty of law be provided for any document to be admitted as evidence before the court. See Antonio Romiti, Le principali sentenze sul protocollo delle pubbliche amministrazioni (Viareggio, 1995). 52  26  the topic of records classification from both a conceptual and a methodological perspective, and whose early works date back to the beginning of the 1960s. 53 De Felice’s writings on what he used to call “systematic classification by competence” 54 represent a first attempt to lend theoretical support and provide systematic rigour to one of the functions most neglected by the archival literature. In his view, the design of a records classification system should follow an “organic, logic, and coherent method,” based on the nature of the competences attributed by law to any given public authority. By classifying according to the competence criterion, “the purpose and means of each office, or each set of activities, become evident.” 55 De Felice does not define ‘competence’ which, in his writings, is used interchangeably with activity, function, and office. The definitions provided by the Concise Oxford Dictionary 56 do not help clarify the relationship between this and other functional terms as well as the diplomatic definition: “Competence [is] the authority and capacity of carrying out a determined sphere of activities within one function, attributed to a given office or individual.” 57 It is therefore evident that both functional elements (“sphere of activity within one function”) and structural elements (“office or individual”) coexist in the administrative concept of competence as ‘functional responsibility’ delegated to a juridical person. 53  See Raffaele De Felice, “La classificazione degli atti negli archivi moderni,” Rassegna degli Archivi di Stato XXIV (1964): 215-42. 54 See Raffaele De Felice, L’archivio contemporaneo. Titolario e classificazione sistematica di competenza nei moderni archivi correnti pubblici e privati (Rome: La Nuova Italia Scientifica, 1988). 55 Id., “Per la formazione dei titolari di archivio,” Rassegna degli Archivi di Stato XXVII, 1 (January-April 1967): 74. Translated by this author. 56 In The Concise Oxford Dictionary of Current English, 9th ed. (Oxford, NY: Clarendon Press, 1998), 271, “competence (also competency)” involves the following applicable definitions: “1 a ability; the state of being competent. b an area in which a person is competent; a skill. … 3 Law the legal capacity (of a court, a magistrate, etc.) to deal with a matter.” 57 Luciana Duranti, Diplomatics. New Uses for an Old Science (Lanham and London: The Society of American Archivists and Association of Canadian Archivists in association with The Scarecrow Press, Inc., 1998), 90.  27  Consequently, a classification system structured “by competence” will be ‘functionbased’ to the extent that the organizational structure corresponds to the hierarchy of functions and activities attributed to any given entity. This type of classification will necessarily show a low tolerance towards administrative change, because its flexibility is constrained by the organizational setting that actually informs its structure. As a matter of fact, the classification schemes currently used by some Italian public authorities, which were developed, whether explicitly or implicitly, according to a competence-based approach, all reveal a structure closely resembling the creator’s organizational chart. 58 A closer look at the classification model suggested by De Felice – which until recently, has been regarded as a standard (at least with reference to certain categories of public bodies), although it has rarely been applied literally59 – reveals a structure where the highest level of the scheme is fixed and is made of three main headings or “titles” (in Italian, titoli): “[titolo I] relevant to the activities concerning the organization and functioning of the offices; [titolo II] relevant to the activities aiming at guiding the acts of the administration in general terms; and [titolo III] relevant to the specific activities carried out by each office in carrying out any of the assigned competences.” 60  58  The Italian classification schemes examined by this author are the ones of the Province of Bologna (that she contributed to develop in the years 1995-98), State Archives of Bologna, City of Bologna, and Province of Bari. 59 De Felice’s model was consistent with the criteria set by Royal Decree no. 35 of 25 January 1900, the first Italian ‘records management’ regulation concerning all central administrative bodies of the State (in Italian, regolamento per la gestione degli archivi correnti delle amministrazioni centrali dello Stato). This regulation prescribed, inter alia, specific rules for the design of classification schemes. For the impact of this legislation on the understanding of the records management function and the role of the records manager in Italy, see Elena Aga-Rossi and Maria Guercio. La metodologia per la definizione di piani di classificazione in ambiente digitale (Rome: Scuola Superiore della Pubblica Amministrazione, 2005); Guido Melis, “Il deposito della memoria. L'evoluzione degli archivi amministrativi nella storia italiana,” Rassegna degli Archivi di Stato LXI (2001): 208-25, also available in English translation: “The Profile of the Archivist: Promotion of Awareness,” Archivum XLV (2000): 81-96. 60 De Felice, “Per la formazione dei titolari di archivio,” 64. Translated by this author.  28  Thanks to such a tripartition, which was meant to be common to all records creators, De Felice’s model would facilitate interoperability but also the identification of the creator’s most important policy records (mainly referring to titolo II) for purposes of preservation. It should however be noted that – in contrast with the more rational arrangement devised by Schellenberg, which will be examined later – this model involves some unnecessary redundancy, considering that the activities identified under titolo II will once again be repeated under titolo III, where the individual case files are supposed to be created. Nevertheless, a part from these technical flaws – which prevented his model from becoming popular – De Felice has the merit of having highlighted important points that make clear that records classification is more than a mere retrieval tool. “Classification … allows reducing the multiplicity of the affairs attended to a finite number of hierarchically arranged categories, so that the daily accrual of the archives will result in a logical accumulation [of records] that will faithfully reflect the growth and evolution of any given activity.” 61 The “cognitive act of classification,” De Felice adds, must necessarily be performed at the very moment of records creation “in order to guarantee the correct formation of the series through the rational categorization of the competences of the office.” 62 By “establishing the archival bond,” thereby determining the internal structure of an archival fonds, classification becomes, in his view, “the only means to accomplish the formation of archives” 63. Some commentators on De Felice’s writings have criticized what, in their view, appears to be an overestimation of the role of records classification, because  61  Raffaele De Felice, “In margine ad alcune questioni di archivistica,” Rassegna degli Archivi di Stato XXXI, 1 (January-April 1971): 135. Translated by this author. 62 De Felice, “Per la formazione dei titolari di archivio,” 67. Translated by this author. 63 Ibid., 68. Translated by this author.  29  “the archival bond exists independently of any administrative, cognitive, or cultural operation; archives which are generated without any classification linked to the records have nevertheless the archival bond.” 64 This is indeed theoretically correct; however, one should not underestimate the risks that may arise from not making the archival bond explicit through classifying the records. Where there is no evident, stable, expressed relationship among the records (which only classification can provide), their natural, original, and necessary order may get altered at any time, thus making the archival bond impossible to recognize and eventually reconstruct.  2.2.3 Function and Structure in the Time of Modern Bureaucracy The concept of competence previously examined has already demonstrated how business function and organizational structure may be intertwined to the point that they may coincide, or appear to coincide, both in the reality and in records classification as a representation of that reality. Indeed, this is not only an Italian phenomenon. Most of the classification schemes existing in early bureaucracies, as well as the descriptions of their methodological underpinnings, although function-based in theory, at a closer look, show that they actually reflect the current internal structure of the records creating organizations with their hierarchies of departments, divisions, offices, and so on. Function, defined in diplomatics as “the whole of the activities aimed to one purpose, considered abstractly,” 65 is an abstraction and, as such, it needs a structure made  64  Donato Tamble’, La teoria archivistica italiana contemporanea. Profilo storico-critico (1950-1990) (Rome: La Nuova Italia Scientifica, 1993), 109. Translated by this author. See also Duranti, “The Archival Bond.” 65 Duranti, Diplomatics, 90. Once again, one of the rare definitions of functional terms in the archival literature comes from diplomatics.  30  of “rules and resources” 66 to materialize. Where each function is carried out “without involving more than one organizational unit or department at a time” 67, the boundaries of either concept (i.e., function and structure) may be so blurred that making a distinction for the purpose of describing only the function, ‘abstractly,’ will almost be impossible. Although one can come across this type of organizational configuration (known as “machine bureaucracy” 68 or “full bureaucracy” 69) in any place and at any time, there actually was a time in history when organizations in the western world used to be primarily shaped that way. From the beginning of the industrial age (end of the 1700) until at least World War II, societal structures used to enjoy a relative stability. Both public and private bodies were characterized by rather simple and rigid hierarchical organizations, rational division of labour, and fixed sets of responsibilities assigned to each office or functional area in accordance with written rules and regulations. 70 Univocal, downward  66  Orlikowski, “The Duality of Technology,” 405. Structuration theory enables an understanding of structure that takes both its social and physical components into account. As a social construction, structure involves “rules” (e.g., hierarchy, delegation of authority, career paths, decision-making processes, etc.), while as a physical one, it is made of “resources” (e.g., people, office spaces, financial assets, etc.). See Anthony Giddens, The Constitution of Society. Outline of the Theory of Structuration (Berkeley: University of California Press, 1984); JoAnne Yates and Wanda J. Orlikowski, “Genres of Organizational Communication: A Structuractional Approach to Studying Communication and Media,” Academy of Management Review 17 (1992): 299-326. 67 Michael Lutzker, “Max Weber and the Analysis of Modern Bureaucratic Organizations,” The American Archivist 45 (Spring 1982), 124. 68 Henry Mintzberg, Structure in Fives: Designing Effective Organizations (Englewood Cliffs: Prentice Hall, 1983). See also Gareth Morgan, Images of Organization (Beverly Hills: Sage, 1986), 22-25. 69 Geert Hofstede, Culture’s Consequences. Comparing Values, Behaviors, Institutions, and Organizations across Nations (Thousand Oaks: Sage Publications, 2001), 377. In Hofstede’s analysis of national and organizational cultures, this type of organizational configuration is also called “pyramid model” and is mainly associated with Latin and Mediterranean countries. Although the value of Hofstede’s generalizations may raise some criticism, for the purposes of this discussion, it is interesting to note that the full form of bureaucracy, which has its roots in the time of the industrial revolution, appeared to be still massively present in France, Italy, Spain and other countries around that area throughout the 1990s. 70 This mechanistic type of organization has been fully described by Max Weber and Frederik Taylor in their classic writings of organizational theory and sociology. See also Morgan, Images of Organization; and JoAnne Yates, “Internal Communication Systems in American Business Structures: A Framework to Aid Appraisal,” The American Archivist 48, 2 (Spring 1985): 141-58.  31  communication flows and linear decision-making processes implied a minimum overlapping of tasks and no need for sharing them. In such self-contained, “monohierarchical structures, … decisions were made at one level and implemented at the next” 71 and the relevant records, “preserved in their original or draught form,” 72 provided a mechanism for monitoring an individual’s performance and set precedents for future actions. In the writings of British archival theorist Sir Hilary Jenkinson, one may find evidence not only of the alignment of function and structure typical of early bureaucratic organizations but also of the equally typical reliance on recordkeeping procedures as a basis for good administration. 73 It is in this context that statements like the following must be read so that they do not sound as inconsistent as they may appear to be: “Archive series must always refer to some Administrative Function, because without it they themselves would never have come into existence.” 74 “… a Class [(i.e., the highest level of a classification system) shall correspond to] … the division of office work which produced it.” 75 Similarly, Margaret Cross Norton, Illinois State Archivist between 1922 and 1957, in a paper presented in 1940, writes: “Archival classification is based upon departmental organization,” 76 a structure-related sentence that shortly precedes her most quoted functional statement: “It is a rule in government that records follow functions.” 77 As a  71  David Bearman and Richard H. Lytle, “The Power of the Principle of Provenance,” Archivaria 21 (Winter 1985-86), 16. 72 Cited in Lutzker, “Max Weber and the Analysis of Modern Bureaucratic Organizations,” 124. 73 See Jenkinson’s “golden rule for the Administrator,” already mentioned in the Introduction to this dissertation, in Jenkinson, A Manual of Archive Administration, 153. 74 Ibid., 111. 75 Hilary Jenkinson, “The Classification and Survey of English Archives,” (1943) in H. Jenkinson, The selected writings of Sir Hilary Jenkinson (Gloucester: Alan Sutton, 1980), 201. 76 Thornton W. Mitchell, ed., Norton on Archives. The Writings of Margaret Cross Norton on Archival and Records Management (Carbondale and Edwardsville: Southern Illinois University Press, 1975), 106. 77 Ibid., 110.  32  corollary, Norton adds that, in order to facilitate the transfer of records from an extinct agency to the one inheriting its functions as a consequence of re-organization, records should be classified on the basis of “the present administrative organization,” 78 thus demonstrating that, in her view, agencies were organized along functional lines. Both Jenkinson and Norton acknowledged the relationship to function as a fundamental characteristic of the nature of a record. Nevertheless, the type of administrative reality they were facing was probably not complex enough to make them appreciate the different effects of either approach (i.e., function- versus organizationbased) on the management of the records.  2.2.4 Schellenberg and the Functional Analysis More than ten years later, Theodor Schellenberg, the United States (US) National Archivist from 1950 to 1961, was still writing that “organization frequently corresponds to function” 79. However, of the three criteria for classifying records he identifies in his manual (i.e., action, organization, and subject matter), the first one, action, is indicated as the one records managers have definitely to privilege, in that, “Most public records are the by-products of action, and they naturally fall into groups that relate to action.” 80 Indeed, an important point that Schellenberg raised in his 1956 manual is that the practice of classifying records by subject matter had to be seen as an “exception,” 81 rather than a rule, as his contemporaries seemed for the most part to believe.  78  Ibid., 111. Theodore R. Schellenberg, Modern Archives: Principles and Techniques (Chicago: University of Chicago Press, 1956), 55. 80 Ibid., 53. 81 Ibid., 60. 79  33  The situation of recordkeeping in the United States must have been quite dramatic if, already in the 1940s, archivist Philip Brooks, not without concern, had noted: “Filing agents and methods are kaleidoscopic in their variety.” 82 Due to the exponential growth of organizations in terms of size, complexity, and volume of records handled, and to the simultaneous absence of adequate methods to keep the records under control, post World War II American society was indeed facing an emergency in terms of the organizational and public accountability83 that records should provide. Not by chance, the first part of Shellenberg’s manual is dedicated to the description of corrective measures that organizations are recommended to implement if they are to improve their efficiency and effectiveness. Mainly, these measures aim at the twofold objective of simplifying business functions and rationalizing recordkeeping practices, and consist of standardization of procedures, control of forms, reduction of duplicates and, last but not least, adoption of records classification. 84 As to classification methods, Schellenberg, who was influenced by the ideas of German archivist Brenneke, besides rejecting the subject-based approach that would only be suitable for reference and information files, emphasizes the importance of starting from the analysis of an agency’s functions, activities, and transactions.  82  Philip Brooks, “The Selection of Records for Preservation,” The American Archivist 3 (October 1940): 224. 83 See Terry Eastwood, “Should Creating Agencies Keep Electronic Records Indefinitely?” Archives and Manuscripts 24, 2 (1997): 264. “Organizational accountability” considers the records in their current environment and refers to the ability of “officers … to render an account of how they have fulfilled their obligations …;” while “public accountability” refers to the fact that “records account to the public for the discharge of the duties of its agent …”. 84 See Schellenberg, Modern Archives, 44-46. Interestingly, Schellenberg did not believe that a program to control records creation could by itself positively affect an organization whose functions and activities were not conducted efficiently. In other words, he was convinced that rationality of recordkeeping systems does not necessarily entail rationality of decision-making processes.  34  “Records, as a rule, should be classified according to function. They are the result of function; they are used in relation to function; they should therefore be classified according to function.” 85 As an alternative, he acknowledges the possibility of referring to the structure of the organization as a criterion for classification, given the usual correspondence between organization and function. However, he admonishes, “such a division into organizational classes is possible and advisable only in governments whose organization is stable and whose functions and administrative processes are well-defined.” 86 Schellenberg is the first author who elaborated a set of principles for classifying records and who highlighted the importance of functional analysis, an approach that was unfamiliar to his contemporaries. His classification development rules and, in particular, the hierarchy of functions, activities, and transactions he identified as the basic structure of his functional classification model (also known as “F-A-T model”), became a point of reference for the archival community, and not only in the US. Still today, his concepts and definitions are drawn on as a useful framework for the analysis of contemporary organizations. 87 Schellenberg’s method consists in an initial division of the whole of an agency’s functions (defined as “all the responsibilities assigned to an agency to accomplish the broad purposes for which it was established” 88) into two main groups of activities respectively called “substantive” (i.e., “activities relating to the technical and professional work of the agency”) and “facilitative” (i.e., “activities relating to the  85  Ibid., 62-63. Ibid., 59. 87 See, in particular, Elizabeth Shepherd and Geoffrey Yeo, Managing Records. A Handbook of Principles and Practice (London: Facet Publishing, 2003). 88 Schellenberg, Modern Archives, 53. 86  35  internal management of the agency, such as housekeeping activities”) 89. By splitting the structure of the classification into two parts, the first one specific to each agency and the second one potentially shareable among agencies, being made of activities that are common to all, Schellenberg on the one hand, avoided redundancy, and on the other, established a principle of uniformity that anticipated the need for interoperability among agencies that would be emerging in the future. In his model, all transactions deriving from the breaking down of both categories of activities are in turn subdivided into “policy” and “operational transactions.” According to Schellenberg, a classification system where all record aggregates belong to either type of transactions facilitates records appraisal and selection on the basis of the higher evidential and informational value of the records supporting policy decisions in comparison to those relevant to the specific individual transactions that follow, or precede, those decisions. Thus, an important methodological suggestion underlying his model is that, when designing records classification schemes, any predictable retention and preservation need should also be taken into account. Schellenberg also has the merit of having clarified that classification and filing are two distinct activities. While for the former a functional approach is recommendable, the way records should be grouped into files depends on the nature of the transaction. “All transactions – he explains – relate either to persons, or corporate bodies, or places, or topics.” 90 This is an insight that tends to be neglected by today’s practitioners who wish to develop so-called ‘business classification schemes’ and believe that everything (both classes and files) ought to be functional.  89 90  Ibid., 54. Ibid.  36  As to the purpose of classification, Schellenberg makes reference neither to the nature of the records nor to any other theoretical considerations (such as, the need for establishing and perpetuating the original context of records creation), but pragmatically writes: “Records must be put away in an orderly and accessible manner to be quickly retrieved when they are wanted.” 91 On the same vein, he concludes that, because “the purpose of classification is to facilitate the location of records when they are needed … records should not be overclassified.” 92 A good suggestion which, however, may clash with the complexity of affairs and does not take into consideration that different people may have different views of the same issue. This implies that classification may not be such a neat and straightforward matter like Schellenberg’s rules seem to posit.  2.2.5 Theory and Practice of Functional Classification in Canada Until the 1980s, in Canada like in the United States, subject-based systems were the rule. The first attempts to a functional approach to classification can be found in the systems developed by the Provinces of British Columbia and Nova Scotia, respectively called ARCS (Administrative Records Classification System) and ORCS (Operational Records Classification System), and STAR (Standard for Administrative Records) and STOR (Standard for Operational Records) respectively. 93 Thanks to the physical divide between  91  Ibid., 47. Ibid., 63. 93 Government of British Columbia. ARCS Online: Administrative Records Classification System (2003 ed.), available online at http://www.bcarchives.gov.bc.ca/arcs/index.htm (accessed on 05/05/2007); Id., ORCS: Operational Records Classification Systems; Government of Nova Scotia. STAR: Standard for Administrative Records (2006 ed.), available online at http://www.gov.ns.ca/nsarm/star/ (accessed on 05/05/2007); Id., STOR: Standard for Operational Records. The Canadian system is also known as “block numeric system” as it is “based on the assignment of blocks of numbers to represent the main groups, primaries and secondaries”. See Duranti et al., Preservation of the Integrity of Electronic Records, 44. 92  37  records resulting from common administrative activities (corresponding to Schellenberg’s facilitative activities), which are included in one system shared across all government agencies, and records resulting from the distinct operational functions of each agency (i.e., Schellenberg’s substantive activities), the overall system is very flexible and allows for interoperability. ORCS and STOR, which are unique for each agency, also share a common structure, thus providing a basis for further standardization across the country. The major advantage of the above mentioned systems derives, in our view, from the fact that the classification is fully integrated with a preservation plan, by associating each lower level of the classification with relevant retention rules expressing how long each identified record series is supposed to be retained first in the creator’s office (active stage) and then in a records center (semi-active stage). Retention information also specifies when the series are supposed eventually to be either disposed of or transferred to an archival repository for purposes of further preservation. 94 Thus, the retention mechanism facilitates the management of the life cycle of the records by guiding them through an established chain of responsibilities and by working as a ‘filter’ which takes into account the operational, legal, and potential long-term values attributed to the records. In other words, it is a sort of pre-appraisal. The challenge of integrating classification and retention considerations lies in balancing the prospective, analytic function of classification with the retrospective, evaluative view which is implicit in any form of records appraisal. This involves looking at the functions and related records not only in terms of actual business and records  94  See Duranti’s analysis of the BC and NS classification systems in Duranti, I documenti archivistici, 7376.  38  management needs, but also in terms of the significance of the records for purposes that might be different from the business ones (e.g., research or cultural purposes). From a methodological viewpoint, letting retention – which is not a primary objective of classification – influence the classification design might result in a scheme that both sacrifices current to potential future needs and imposes an order on the records that does not correspond to the one dictated by the requirements of business. An in-depth analysis of the Canadian classification schemes reveals that the declared functional approach is, in some instances, subordinated to retention requirements. Headings like “Policies” or “Contracts,” for example, which are not related to any specific function, seem to be meant to create typologically homogeneous series for purposes of preservation. Overall, one may say that the British Columbia systems and their Nova Scotia counterparts are definitely effective thanks to the number of functionalities they provide within one integrated tool. However, neither of them really constitutes a good example of the outcome of functional analysis in the proper sense. In fact, several classification criteria may be recognized in these schemes: record type (e.g., “Contracts”), structure (e.g., “Committee”), subject-matter (e.g., “Equipment and Supplies,” which in turn nestles “Clothing,” “Fuel,” etc.) and function (e.g., “Issuing of Permits”) are all mixed up at any classification level. Apart from this content-related issue, a structural issue also needs to be addressed at this point. The headings of the lower levels of the classification schemes under examination are listed in alphabetical order, instead of being arranged according to the sequence of activities, or stages, that usually occur when carrying out a function. Obviously, only the latter order allows displaying correctly the way in which given affairs or matters grow and develop, i.e., the workflow. The alphabetical arrangement of 39  subclasses is fundamentally inconsistent with a functional approach, and may be attributed to a lack of conceptualization with reference to the meaning of functional components. A Canadian archivist who has reflected on records classification methods and has analyzed existing ‘functional’ models criticizing their foundation is Paul Sabourin. At the end of the 1990s, as an archivist of the National Archives of Canada (NA – today Library and Archives Canada – LAC), Sabourin participated in a project for the review of the NA’s records classification system known as Subject Classification Guide. After the adoption of new disposition authorities called MIDAs (Multi-Institutional Disposition Authorities), which were based on a fully functional appraisal methodology known as macro-appraisal, 95 the old subject-based classification system had become an obstacle to the effective application of the functional categories identified in the MIDAs. That was the strong impulse for a change in the classification system as well. The first issue faced by Sabourin and colleagues was that of defining what a function is. After several years of structural-functional analysis in the context of the macro-appraisal approach, the NA agreed on the following “working definition”: “A function is (1) any high level purpose, responsibility, task, or activity which is assigned to the accountability agenda of an institution by legislation, policy, or mandate; (2) typically common administrative or operational functions of policy development and program and/or delivery of goods and services; (3) a set or series of activities (broadly speaking, a business process) which, when carried out according to a prescribed sequence, will result in an institution or individual producing the expected results in goods or services that it is mandated or delegated to provide.” 96  95  See Terry Cook, “Mind Over Matter: Towards a New Theory of Archival Appraisal,” in The Archival Imagination: Essays in Honour of Hugh A. Taylor, ed. Barbara L. Craig (Ottawa: Association of Canadian Archivists, 1992): 38-70. The macro-appraisal model will be discussed later on. 96 Paul Sabourin, “Constructing a Function-Based Classification System: Business Activity Structure Classification System,” Archivaria 51 (Spring 2001): 144.  40  It is understood – Sabourin explains – that the term function may be used with all three characterizations in mind or only one, according to the purpose for which it is used (i.e., description, appraisal, or classification). The one that better fits with the Business Activity Structure Classification System (BASCS) 97 that came out from the NA’s project appears to be the third part of the definition, which describes a function as a business process and each process as a cyclical, sequential series of fixed steps. The methodology for designing a functional classification system like BASCS rests on the assumption that the sequence of procedural steps as it is described in, and often prescribed by, legislation or other regulatory instruments potentially makes up the structure of a given kind of activity. With the expression Activity Structure included in the acronym BASCS, its developers meant exactly the decomposition of functions and activities according to both a hierarchical and a sequential order (not an alphabetical one!), down to the elementary units that correspond to the steps, or transactions, generating the actual files. According to this approach, the latter would thus reflect the natural (either prescribed or logical) development of each activity carried out by an agency. There is no doubt that this methodology is indeed very logical and purely functional. At first glance, one may even think that, with the implementation of such activity-based system, one could achieve the full integration of business processes and documentary procedures, which is considered by archival theory one of the fundamental methods to ensure records reliability in a trusted recordkeeping system. 98 However, if  97  BASCS structure and guidelines for its implementation are available on the LAC website at http://www.collectionscanada.gc.ca/government/products-services/007002-2089-e.html (accessed on 05/01/2009). 98 See Duranti et al., Preservation of the Integrity, 42-43.  41  applied systematically and exclusively, the approach described by Sabourin may result in some sort of abstract and self-referential system, a classification that would mirror the functions of the organization, yet would be totally unable to capture the actual ways of carrying out work in a real office. This shows just one of the limitations of a strict interpretation of the functional approach to classification. A second shortcoming refers to the fact that not every single activity behaves as a structured, repetitive process. There are areas of human endeavour (e.g., academic research, teaching, or artistic performance) that, on the contrary, have the characteristic of being creative and unpredictable, so that the relevant activities do by no means follow any pre-established, linear, or cyclical sequence of steps. Such a ‘freedom of action’ – which, following the “soft systems thinking approach,” is related to the nature itself of human beings, who “can always decide to act otherwise” 99 – may as well enter in any work procedures, even the most bureaucratic ones. This would imply that a fully functional approach to classification is destined to fail due to its own abstractness and ‘perfection.’ A last, record-related consideration inspired by the Canadian model is that, in a real work environment, not every step that makes up a process is bound to generate a distinct transaction file, as assumed by BASCS. Some offices may find it more convenient for the purposes of their business, for instance, to keep all of the records originated by an entire process, or even an entire function, together in one single file. In such a case, the higher activity level, not the transaction level, should be tagged as the entry point for file creation. Where the main driver for classification design is the  99  Peter Checkland and Jim Scholes, Soft Systems Methodology in Action (Chichester, West Sussex, UK: John Wiley & Sons Ltd., 1999), 2.  42  workflow, rather than being the records flow or the user needs, the lower levels of the scheme tend to become too detailed and as such, they may cause excessive fragmentation of files. As a consequence, users may find the classification complicated to apply and record keepers may have difficulty in maintaining it up-to-date. Actual results of the adoption of the BASCS model and relevant appropriation modes have never been analyzed critically. This looks like a gap in the literature that this author addresses at the conclusion of her review of the Canadian classification approach.  2.2.6 The Australian Approach to Recordkeeping: from Maclean to Records Management Standards The Australian archival tradition is an outgrowth of the British one. As Schellenberg explains, in the 1950s and 1960s, incoming and outgoing correspondence used to be registered in logbooks where links among the records belonging to the same files used to be annotated as well. 100 At that time, Ian Maclean was leading the Australian public records administration. His ideas have been very influential for the developments of recordkeeping in the country. In particular, he had the intuition of the continuity existing between records management and archives. Therefore, it was natural for him to pay particular attention to records classification, which he described as “the foundation of the study of modern records administration,” 101 but also to filing, as an important activity whose principles and rules have been rarely addressed in the archival literature.  100  Schellenberg, Modern Archives, 72. Specifically, he writes: “Under the present-day registry system, generally used both in Commonwealth and State governments, inward and outward documents are brought together into files just as in England.” 101 Ian Maclean, “Trends in Organizing Modern Public Records, with Special Reference to Classification methods,” Archives and Manuscripts 3 (1956): 17.  43  According to Maclean, files are usually created for two purposes, that is, either to carry out business (i.e., “transaction files,” that he defines as “file[s] that contain the sequence of papers deriving from a particular piece of business”) or to support the “action records” relevant to that business (i.e., “subject files,” containing “background information records”). 102 Taking this distinction into account, he formulated the following “rules of efficient recordkeeping”: “first … draw a clear line of demarcation between files established for the two different purposes; … and second … [strictly observe] the principle of respect for the sequence of administrative action.” 103 Maclean implicitly draws attention to the fact that not every records creating action is to be treated as a business transaction. There are in fact records that are put together just to support a given function or activity, and may therefore be arranged according to subjects or any other criteria useful to the records creator. However, is the difference between subject and function clear? Maclean admits that the way the words are used may be confusing: “Sometimes … [subject] means function or activity, sometimes the transaction that is the subject of a file, sometimes the event about which the department is taking action, sometimes the abstract subject that is the subject of documentation…” 104 In order to find out which criterion is most suitable to file a record, Maclean suggests that record keepers should analyze the specific purpose of the activity generating that record. Additionally, in line with Jenkinson, he stresses the importance of making the original order of the single acts comprising an administrative action or transaction evident to  102  Ian Maclean, “Australian Experience in Records and Archives Management,” The American Archivist 22, 4 (October 1959): 393. 103 Ibid., 395. 104 Ibid., 408.  44  anyone who looks at the relevant file. He does not specify though, in the article here analyzed, that it is through the classification that such sequence can be fixed and secured. Almost forty years after Maclean’s words, the National Archives of Australia (NAA) published the first edition of a recordkeeping manual known as DIRKS (Designing and Implementing Record-Keeping Systems). 105 The DIRKS Manual provides a rigorous and structured eight-step methodology designed to ensure that “records and information management is firmly based on the business needs of the organization.” The business-driven approach of the DIRKS methodology is already shown in its program foreword. As a further evidence of this approach, it will be enough to mention that, among the factors that concur to determine the recordkeeping requirements that are appropriate to each organizational context, DIRKS identifies the “accountability agenda of the organization,” which is based on a systematic analysis of its legal and regulatory obligations, business requirements, and broader community expectations, together with an assessment of the exposure of the organization to risk if those requirements are not addressed. The link between accountability and recordkeeping is a key issue in the administrative culture of Australia. Since the 1980s, when a series of accountability crises occurred in numerous government bodies, Australian archivists have been reflecting on their responsibilities towards their institutions and the society at large, as well as on the interdependence of administrative and recordkeeping practices. 106 It is against this background that one has to look not just at the DIRKS Manual but also at the 105  National Archives of Australia, Designing and Implementing Record-Keeping Systems – The DIRKS Manual: A Strategic Approach to Managing Business Information (Commonwealth of Australia, September 2001, revised July 2003), available online at http://www.naa.gov.au/recordsmanagement/systems/DIRKS/index.aspx (accessed on 21/11/2008). 106 See Sue McKemmish and Frank Upward, eds., Archival Documents. Providing Accountability through Recordkeeping (Melbourne: Ancora Press, 1993).  45  comprehensive recordkeeping framework, made of several standards, policies, and regulations, that the Australian archival community has been able to build in the last few decades. 107 The first two steps of the DIRKS methodology are those leading to the design of a Business Classification Scheme (BCS), that is, “a conceptual model showing an organization’s functions, activities, and transactions in a hierarchical relationship.” Organizations that intend to establish a new recordkeeping system, or to improve an existing one, should first engage in a “preliminary investigation of the business, social, and legal contexts” in which they operate through collecting relevant information by means of documentary sources and interviews with internal and external stakeholders on what the organization actually does. The second step, i.e., “the analysis of the organization’s business activities and processes,” involves identifying the “largest units of business activity” and then breaking them down into a set of logical sub-parts by means of a top-down functional analysis. To know how organizations carry out their business in detail, the manual suggests that “process analysis,” which presupposes a bottom-up examination of all steps involved in each activity or transaction, be conducted. It should first be noted that the DIRKS Manual uses various functional terms (e.g., function, business activity, process) but never defines them, thus confusing the reader. Additionally, in the same way as the Canadian BASCS, the Australian BCS fails to acknowledge the existence of any other type of functions besides those that are structured, sequential, and routinized. Furthermore, the DIRKS Manual does not include  107  See the web site of the National Archives of Australia at http://www.naa.gov.au/recordkeeping/default.html, which aims at providing “detailed, practical information to help Australian Government agencies improve their recordkeeping, following best practice approaches developed by the National Archives.”  46  in the discussion of classification design any examination of existing records, files, or recordkeeping procedures until later in the methodology. One may therefore argue that the Manual underestimates the amount of functional knowledge that may be gathered from a bottom-up analysis of record-related issues. In other words, the Australian BCS model seems to be the outcome of a mere, though very articulated, functional analysis process, and as such, it is again very close to the Canadian model. The only difference may be its wider focus, as the design of a BCS includes an investigation of the broader social context, or “ambient function,” 108 which an organization’s goals and strategies ultimately depend on. However, there is no indication in the DIRKS Manual of how to conduct such complex analysis of the wider environment of recordkeeping. In another section of the NAA’s recordkeeping strategy, the BCS is described as the “logical model” that archivists, or whoever in the organization is responsible for the relevant function, draw on to design “classification tools for records management.” 109 It seems therefore that the BCS is not itself a records classification tool, although the difference between that and a records classification, or a thesaurus (which is considered equivalent by NAA), would only emerge at the transaction level, i.e., at the point where records may happen to be created. In practice, the methodology recommends to translate the functional terms of the BCS into “topics and/or subtopics” to serve the “purpose of records classification … [that is,] to title the record for searching and retrieval.” Like Schellenberg, this approach reasons that the primary purpose of classification is records retrieval. It ignores the more substantial need to make explicit and fix the relationships among records in series and files. 108  Chris Hurley, “Ambient Functions – Abandoned Children to Zoos,” Archivaria 40 (Fall 1995): 21-39. See National Archives of Australia, Overview of Classification Tools for Records Management, available online at http://www.naa.gov.au/recordkeeping/control/tools.html (accessed on 05/05/2007). 109  47  Besides highlighting the above mentioned terminological issue, the manual does not elaborate on how to adapt the conceptual representation of business processes typical of a BCS into a workable records classification tool responsive to the requirements of the records creators. It only suggests that contents can be attributed to topics at the lower level of the classification. These contents could actually be anything: record type, subject-matter, transaction, the output of a small group of tasks, and so on. Thus, the records classification appears to be a hybrid tool in comparison to the purely functional BCS. These arguments bring us to discuss the idea of ‘terminological control’ (as a hierarchical and logical expression of predictable relationships) as opposed to the idea of ‘contextual control’ (as a non-hierarchical, contingent description of observed, unpredictable relationships), both of which have been formulated by Australian archivist Chris Hurley in his analysis of ‘contextual’ or ‘recordkeeping metadata.’ According to Hurley, because “records are time-bound … [in that] they evidence an event locked in time,” 110 the metadata relevant to circumstances that are contemporary to the making of the records, and are captured in record-keeping systems (by means, for instance, of records classification), require “external validation” 111 once the facts the records refer to have become ‘historical.’ Contextual control is what provides ‘ambience,’ i.e., the broader context that is needed to give meaning to any given body of records. Such highlevel knowledge, also known as ‘ambient function,’ is inherent in any recordkeeping situation; however, there, it does not need to be articulated. The place where contextual knowledge must be made explicit is archival description. At the moment of records  110 111  Hurley, “Ambient Functions,” 22. Ibid., 24.  48  creation and classification, the focus is on ‘business function,’ which requires terminological rather than contextual control. All this seems to be consistent with the attention paid by NAA to issues of terminology like those noted earlier and, in particular, to the thesaurus as a means of classification. As anticipated, records classification design and thesaurus building are, in the DIRKS Manual, treated in the same way. Organizations, it is expressly said, are free to choose, according to their retrieval preferences and needs, between the hierarchical structure of a “Records Classification Scheme” and the flat, alphabetical structure of a “Functional Thesaurus.” The thesaurus acts in a sense as an index to the BCS – which is the source of both records classification and thesaurus – and, thanks to the flexible approach it allows, offers more powerful retrieval capabilities. This method seems to confuse the purposes of records classification, on the one hand, with thesaurus or controlled vocabulary construction, on the other. Nevertheless, the issue cannot be dismissed without exploring the terms of a discussion that, since the end of the 1990s, has inflamed the Australian community of information professionals. A remarkable controversy started with a polemical paper by Maggie Exon criticizing Keyword AAA, a functional thesaurus released in 1995 by the Archives Authority of New South Wales and later on extensively implemented in most Commonwealth Government agencies. 112 Keyword AAA is a thesaurus of administrative terms widely used in Australia designed for the purpose of classifying, titling, and indexing any kind of information, and in particular records, by virtue of the 112  See Maggie Exon, “Contemporary Recordkeeping: The Records Management Thesaurus,” Informaa Quarterly 13, 4 (November 1997): 14-22. The response to Exon’s paper by Catherine Robinson and Janet Knight from the State Records NSW is available online at http://www.records.nsw.gov.au/recordkeeping/contemporary_recordkeeping_the_records_management_th esaurus_response_10470.asp (accessed on 05/05/2007).  49  acknowledged special relationship existing between records and functions. 113 Exon’s criticism was mainly related to the fact that Keyword AAA is not a thesaurus in the generally understood use of that term. Besides mere technical issues (e.g., how consistent functional thesauri and business classification schemes ought to be, being the latter the source of the former), the main controversial point was the actual ability of Keyword AAA to assist records management. However, in our view, the discussion appears to be primarily on information retrieval capabilities, which is not really at the core of the archival discourse. The real issue, as pointed out by Stephen Bedford, is that only a hierarchical classification can serve properly the purposes of record management. 114 As Barbara Reed states, “records ... are agents of action, active participants in business activity that can only be described through a series of parallel and iterative processes.” 115 One of these processes is classification, which identifies the business activity while it is being carried out and in the context of the higher function of which that activity is a component. To achieve the objective of revealing the functional context of the records, a records classification scheme must only display meaningful hierarchical trees of functional terms. Functional thesauri, on the contrary, thanks to their alphabetical arrangements and multiple entry points, allow connecting broader and narrower terms in  113  State Records Authority of New South Wales, Keyword AAA, 1st ed. (1995), available online at http://www.records.nsw.gov.au/recordkeeping/keyword_aaa_424.asp (accessed on 05/05/2007). “[Keyword AAA] covers terminology common to business functions and activities in most organizations and is normally used in conjunction with a thesaurus of functional terms relating to the organization's specific or core business functions, to provide comprehensive controlled vocabulary coverage.” 114 See Stephen Bedford, “The Thesaurus is Dead,” Informaa Quarterly 19, 2 (May 2003): 12-15. 115 Barbara Reed, “Metadata: Core Record or Core Business?” Archives and Manuscript 25, 2 (November 1997): 218-241.  50  combinations that may make no sense. Reed has named this the “loss of context” problem. 116 The discussion on the aims of records classification seems to be endlessly open. 117 Retrieval, the most obvious of these aims – yet just a secondary one, according to the traditional archival viewpoint – appears to be the only purpose recognized by a stream of literature that does not really qualify as archival in the proper sense. An example of this hybrid literature, which may be placed between library/information science and records management/archives, is a recent qualitative study of information seeking behaviour of electronic records management systems (ERMS) users conducted in Australia. One of its findings is that “none of the four organizations interviewed promoted or trained users to use their respective classification schemes to seek information in the ERMS.” 118 As a consequence, “users are not using the classification scheme to conduct their information seeking in the ERMS” 119 and prefer to refer to other metadata elements instead. The researchers conclude their article by making recommendations to improve the implementation of classification schemes as retrieval tools. While the utility of classification for information retrieval should definitely be promoted, it is not the only benefit of which records users should be made aware. 116  Cited in Stephen Bedford, “The Thesaurus is Dead”, 13. The problem identified by Barbara Reed is the following. In alphabetic thesauruses of three or more levels, broader and narrower term relationships are only drawn to one level. In this way, a term at the second level of the hierarchy can have many broader terms and many narrower but not all these broader terms make sense with all the narrower terms. On the contrary, in records classification schemes, relationships span more than one level, i.e., terms at a third level depend on the broader function-activity pair. Such schemes spell out the hierarchy of terms in their allowable combinations only; therefore, they preserve the context of the terms by presenting the entire classification string from broadest to narrowest term. 117 One of the loci of this discussion is the Australian Standard T21-09 Subcommittee that was established in 2002 to examine the issue of functional classification. Its ultimate objective is to produce a technical report on the construction of records classification tools compliant with the international standard ISO 15489 on records management. 118 Pauline Singh, Jane E. Klobas and Karen Anderson, “Information Seeking Behaviour of ERMS Users. Implications for Records Management Practices,” Human IT 9, 1 (2007), 173. 119 Ibid., 172.  51  The international standard for records management ISO 15489-2001 120 recognizes that classification establishes “linkages between individual records which accumulate to provide a continuous record of activity.” 121 Although it does not mention the qualities of such ‘linkages,’ this sentence can be regarded as consistent with the tenets of archival science. The Australian origin of the standard emerges though in the emphasis it puts on the classification of business activities “as a powerful tool to assist the conduct of business and in many of the processes involved in the management of records.” 122 Overall, the way of tackling the issue of classification shows that the standard is rather business-oriented, instead of focusing on the records. Additionally, although it embraces a functional approach, the standard does not suggest that the only way to classify records is by function. In its definition of classification, one reads: “Classification is the systematic identification and arrangement of business activities and/or records into categories according to logical structured conventions, methods and procedural rules represented in a classification system.” 123 However, in the section describing the design and implementation methodology for record systems, the whole discussion of classification refers to business classification schemes, which ought to be based on a hierarchical arrangement of business functions, activities, and transactions. What is remarkable is that ISO 15489 clearly distinguishes classification from other kinds of retrieval tools. Vocabulary controls, indexes, and thesauri are mentioned in  120  International Standard Organization, ISO 15489: Information and documentation - Records management. Part 1: General; Part 2: Guidelines, 1st ed. (September 2001). ISO 15489 derives from the Australian standard AS 4390-1996 and is the basis on which the DIRKS Manual is built. 121 Ibid., 13. 122 Ibid., 13. 123 Ibid., 4.  52  the standard, but just as additional aids that may be implemented in complex organizations to support their business/records classification schemes. Nevertheless, classification is not considered a mandatory aspect of recordkeeping systems.  2.2.7 Records Classification in the Most Recent Archival Literature. The Role of Classification in an Electronic Environment If one understands classification as a mere retrieval tool, then its role in the context of electronic records systems may seem somehow outdated or even superfluous, given the highly sophisticated search engines which are usually embedded in those systems. Thus, it is even more important today than it was in the past for archivists to stress that classification has other ends and values and that the intellectual control it exercises over the records is necessary and irreplaceable. In Italy, for instance, Maria Guercio, who has been writing plentifully on this topic 124, has emphasized that classification is a unique means to enable the systematic, logical, and functional organization of all kinds of records, whatever their medium. Thus, far from being an old-fashioned archival tool, classification – according to Guercio – can become “an essential instrument for the qualified management of meaningful contents on the web … against the risk of losing the notion of archives, structures, relationships in favour of an indistinct and unqualified ‘information’ dimension.” 125  124  See Aga-Rossi and Guercio, La metodologia per la definizione di piani di classificazione in ambiente digital; Maria Guercio, “Modelli efficienti di gestione documentaria nella societa’ dell’informazione. Il ruolo della classificazione d’archivio,” Archivi & Computer 2 (2005): 3-12. Guercio has chaired a project aiming at devising integrated models of functional records classification for a digital environment with the purpose of facilitating interoperability among specific types of organizations in the public sector (e.g., Universities, Regions, and Provinces). 125 Maria Guercio, “Records Classification and Content Management: Old Functions and New Requirements in the Legislation and Standards for Electronic Record-Keeping Systems,” Proceedings of the DLM-Forum (Barcelona 2002): 438.  53  What makes of classification a crucial tool in the electronic environment – more fundamental than it was in the paper world – is primarily the fact that it provides essential information about the contexts of records creation and use, information that would otherwise be unattainable. The fact that, through the classification, it is also possible to manage records retention, assign access privileges, protect records confidentiality, retrieve records in context, manage the work flow, etc. – as records management standards like ISO 15489 stress and EDRMS software vendors like to repeat – simply adds further benefits to the primary one, that is, by using Hans Hofman’s words, to express the “logical boundary” 126 (or archival bond) that determines the structure of an archives. In other terms, the process of assigning the same classification code and file number to all records participating in a given activity achieves the purpose that the elements of form and the physical arrangement of records used to fulfill in a traditional paper environment, that is, to link each individual record to the activity originating it and to the other records resulting from it. Classification, at the same time, makes the archival bond (i.e., the necessary link existing among all records belonging to the same file or series) evident and stable. MoReq2, 127 the revised and upgraded version of the “Model Requirements for the Management of Electronic Records” (MoReq) first issued by the European Commission in 2001, without explicitly mentioning those concepts, draws on them when it states that “the classification scheme lies at the heart of any Electronic Records Management  126  Hans Hofman, “Dealing with Electronic Records: Intellectual Control of Records in the Digital Age,” Janus 1998, 1: 155. “A record is no longer a physical entity, but physically fragmented, kept only together by a logical boundary.” 127 European Commission, Model Requirements for the Management of Electronic Records. MoReq2 Specification (Brussels, March 2008) is available online at http://www.moreq2.eu (accessed on 01/02/2009).  54  System (ERMS).” 128 This European de facto standard, despite its rather technical, antitheoretical approach, builds on a significant body of archival knowledge shared not only at a European level. The first requirement that any classification scheme must fulfill, in the digital as well as in the paper world, is “to reflect in its internal organization the hierarchical structure of business functions.” 129 MoReq2 does not go further in the elucidation of this statement, nor does it dwell upon methods of functional analysis. It is also not within its scope to explain the meaning of business function. One of its main concerns is that of ensuring that the internal integrity of an ERMS is guaranteed and maintained at all times. To this end, a number of control mechanisms and user access restrictions are identified and described in detail. By technically limiting the authority to make changes to any critical metadata – including classification code – to the system administrator, MoReq2 requirements ensure that content, structure, and contextual relationships of records and files are kept unaltered, thus providing a framework to establish the authenticity of an ERMS’s contents. Since the end of the 1990s, The National Archives of the United Kingdom (TNA – formerly, the Public Record Office), has promoted the use of function-based classification for the management of electronic records. 130 The main advantage associated with a functional approach would be that of “mak[ing] the relevant records easier to identify and relocate during times of administrative change.” 131 More recently, however, TNA has slightly modified its view and recommends a “hybrid approach,” 128  Ibid., 6. Ibid., 18. 130 See United Kingdom, Public Record Office, Management, Appraisal and Preservation of Electronic Records, Vol. 2: Procedures, 2nd ed. (Kew: Public Record Office, 1999). 131 Malcolm Todd, Business Classification Scheme Design (Kew: The National Archives, 2003), 22. 129  55  where only the higher levels of the classification are function-based, while sub-classes are subject-based. A hybrid approach is considered “more achievable” than a “purist functional” one, which anyway can never be applied as such because “a degree of compromise” is always necessarily brought in. The main difficulties refer to the fact that, first, “users do not understand and dislike [function-based classification schemes] because they are hard to use;” and second, “a strict functional approach will not support case files well.” 132 One may assume that the first weakness attributed to the functional approach (i.e., being not user-friendly) comes from feedback that TNA got from its government clientbase. In fact, the absence of empirical studies about user acceptance is one of the conclusions of this review of the literature. As to the case file issue, the arguments supporting the claim that ‘particular instance papers’ (as case files are traditionally called in the UK) are not suitable to be classified by means of a function-based system do not sound very convincing. The fact that case file contents are often cross-functional and refer to individuals or other subjects should not be a problem, especially in an electronic environment. Actually, Elizabeth Shepherd and Geoffrey Yeo’s textbook (referring back to Schellenberg’s F-A-T model) offers several examples of how to classify “instances of a process” by function. 133 Shepard and Yeo strongly support function-based classification as the most appropriate means of classifying records. They also present a rather detailed examination of the methods that can be drawn on to develop functional classification schemes, i.e.,  132 133  Ibid., 3. Shepherd and Yeo, Managing Records, 53.  56  “top-down analysis,” and “system (or process, or business) analysis.” 134 Both methods can be used together, as the DIRKS methodology had already explained. As to the bottom-up approach, Shepherd and Yeo recommend records managers to employ “system or process modeling,” a technique developed by systems analysts, to represent individual business transactions and the relationships among them. One may wonder, though, whether records professionals actually possess such knowledge, considering that, in general, current study curricula for records managers and archivists hardly contemplate any notions of business analysis, systems engineering, and the like. This British textbook also describes how to exploit the functionality of computers to enhance classification capabilities. Because in electronic systems storage is random, the use of folders “imitating” the records physical arrangement would no longer be essential. Instead of “translating the logical model of functions and processes into a hierarchy of folders and sub-folders,” contextual metadata from an authority file listing the various functional levels would be added to the records to allow a more flexible, virtual, faceted classification. 135 “Any aggregated record of a particular process or activity can be assembled on demand in response to a user’s search. The record series become virtual, as it is derived purely from metadata applied at item level.” 136 The “multidimensional approach to contextual metadata” 137 advocated by Shepherd and Yeo, together with their ‘atomistic’ interpretation of records management, may however expose one of the fundamental characteristics of the records (i.e., the necessary and  134  Ibid., 58-64. Ibid., 95. Similarly, Bearman wrote: “because electronic records do not have the physicality associated with … paper records, aggregation is unnecessary.” See Bearman, “Item Level Control and Electronic Recordkeeping,” 220. 136 Ibid., 96. 137 Ibid. 135  57  determined nature of their relationships) to serious risk. This author believes that the creation of virtual files “on demand” should not replace the ‘fixed’ arrangement that provides evidence of the way records had originally accumulated in the course of business. Another point that should be raised is that, while discussing the meaning of functional terms (which is something uncommon in the archival literature), this textbook may also generate confusion, as the specific use made of some of those terms does not correspond to the traditional one. For instance, ‘process’ is seen as ‘activity’ in the abstract, and what distinguishes activity from function is scope and hierarchical interdependency, as well as the fact that actions are time-limited, while functions are not. A study that would produce a taxonomy of functions has already been called for by many authors in the past, 138 however only very few examples exist and they are not really satisfactory as they have not been planned to serve archival purposes. 139 From their analysis of functional terminology, Shepherd and Yeo draw the conclusion that “most organizational activities are of a broadly repetitive nature: they are instances of a process that will recur many times.” 140 The idea that emerges is that the instrumental rationality described by Max Weber, typical of bureaucratic systems that – as Mary Douglas would say – “try to reduce uncertainty by means of abstraction and routinization,” 141 seems still to be a characteristic of our society. Furthermore, the authors claim that, despite their relative unpredictability, even “creative activities are 138  See Chris Hurley, “What, if anything, is a Function?” Archives and Manuscripts 21, 2 (November 1993): 208-18; and Bearman and Lytle, “The Power of the Principle of Provenance,” 14-27. 139 See Getty (The), Art and Architecture Thesaurus On Line, The Paul Getty Trust (2000). Available online at http://www.getty.edu/vow/AATServlet?find=activities&english=N&logic=AND&page=1&note=facet (accessed on 05/05/2007). 140 Shepherd and Yeo, Managing Records, 53. 141 Mary Douglas, How Institutions Think (Syracuse, NY: Syracuse University Press, 1986), 93.  58  mostly instances of types of activity that can be expected to recur.” 142 This observation bears interesting consequences with reference to functional classification development; however, its validity has not yet been proved empirically. The only inductive study investigating functional classification this author is aware of is Stuart Orr’s master’s thesis. 143 The main objective of Orr’s study was that of understanding whether the functional approach is indeed “a practicable method of classifying records.” 144 To gather such understanding, he employed the Delphy method, 145 which allowed him to obtain the views of experts, and a questionnaire survey of records management practitioners in Australia, Canada, and the UK. The main findings of Orr’s research are that, although both academic experts and practitioners seem to agree that function-based classification is a valuable way of classifying records, there exists no common model saying how functional schemes should look like, and records managers find them difficult to understand and to apply. Thus, Orr confirms that there are clearly usability issues around functional classification. From his study it also appears that the experts are more convinced of its claimed benefits than the practitioners. The Delphy method seems to be an appropriate instrument to bring knowledge and insight to complex issues, thanks to the ‘guided dialogue’ among experts that it elicits. It is therefore a good learning tool. However, being primarily based on opinion  142  Shepherd and Yeo, Managing Records, 55. Stuart A. Orr, “Functions-Based Classification of Records: Is it Functional?” (Master’s thesis – Northumbria University, 2005). Available online at: http://public.archiefschool.net/C8/Publicaties%20door%20derden/Document%20Library/Orr_Functional% 20Classification.pdf (accessed on 02/03/2007). 144 Ibid., 15. 145 Linstone and Turoff describe the Delphy method as: “a method for structuring a group communication process so that the process is effective in allowing a group of individuals, as a whole, to deal with a complex problem.” Harold A. Linstone and Murray Turoff, eds., The Delphy Method: Techniques and Applications, (Reading, Ma: Addison, 1875), 3. Cited in Orr, “Functions-Based Classification of Records.” 143  59  questions measured by a Likert scale, Orr’s study could not take advantage of more indepth data like those that unstructured interviews or direct observations allow to collect. Additionally, misunderstandings deriving from the ambiguity of some of the concepts under investigation could not be avoided due to the limitations inherent in the written form of communication. The latter problem also affected to an even higher degree the part of the study involving practitioners. Due to the rather low response rate to the questionnaire, this part may be said to have only partially fulfilled expected outcomes. Field work in this area appears therefore necessary, and this research aims at being the first attempt to fill in such a gap identified in the literature.  2.2.8 The Functional Approach in the Literature on Appraisal, Arrangement and Description, and Access Appraising records for the purpose of selecting those to be preserved is a ‘necessary evil’ that archivists have to face. It is not only a question of resources; rather, appraisal is fundamental to leave a good record of the activities undertaken by each entity in society for any kind of secondary use, including evidence, and research purposes. It is a tenet of archival science that records cannot be appraised at an item level, and not only because of their number, but mainly because that would break the necessary link existing between them. This is the reason why functional classification schemes, with their classes identifying groups of files that are instances of the same activity, are particularly suitable to be integrated with retention information (which should be seen as a form of ex-ante appraisal), as showed by many of the classification models examined earlier.  60  Additionally, by anticipating the judgement of the records’ value to the moment of their creation, archivists are closer to that context which lights up the meaning of the record. 146 As opposed to a “pertinence-based” (a.k.a., content-based) approach to appraisal, which focuses on the information contained in the records, a “provenance-based” (a.k.a. function-based) approach assesses “the purposes the record served and whether those purposes were ephemeral or lasting.” 147 However, until a few decades ago, the latter approach still involved a bottom-up analysis, in the sense that archivists used to study directly the meaning of the actual accumulations of records to understand their value. The idea of shifting drastically the focus of appraisal from the actual records to the functions generating them originated in North America, in the 1980s and ‘90s, when several appraisal and acquisition models based on a functional principle have been developed and have more or less successfully been implemented. Following American archivist Gerald Ham’s cry for a new role for archivists as “active documenters” 148 of their society, Helen Samuels, leader of an appraisal project at the Massachusetts Institute of Technology (MIT), devised an acquisition method called “documentation strategy.” 149 Samuels was inspired by the functional investigations  146  See Brooks, “The Selection of Records for Preservation.” Terry Eastwood, “Towards a Social Theory of Appraisal,” in Barbara L. Craig, ed., The Archival Imagination: Essays in Honour of Hugh A. Taylor (Ottawa: Association of Canadian Archivists, 1992), 82. See also Angelika Menne-Haritz, “Appraisal or Documentation: Can We Appraise Archives by Selecting Content?” The American Archivist 57 (Summer 1994): 528-42. 148 Gerald F. Ham, “The Archival Edge,” The American Archivist 38, 1 (January 1975): 13. Ham addressed the need for a drastic change in appraisal methods and archivists’ attitudes in general, if the latter “are to provide the future with a representative record of human experience in our time.” 149 See Joan K. Haas, Helen W. Samuels, and Barbara Trippel Simmons, Appraising the Records of Modern Science and Technology: A Guide (Cambridge, Mass: Massachusetts Institute of Technology, 1985); Id., “The MIT Appraisal Project and Its Broader Applications,” The American Archivist 49, 3 (Summer 1986): 310-14. 147  61  carried out by the Joint Committee on Archives of Science and Technology, whose final report, also known as JCAST report, was published in 1983. 150 The JCAST and MIT studies on the records of modern science and technology, as well as Samuels’ later work on colleges and universities’ records, 151 all lack practical details with reference to their respective methodologies, and show an inconsistent use of the concept of function. In particular, the editor of the JCAST report, Clark Elliott, writes that, in order to make informed appraisal decisions, “archivists have to know how and for which purposes records were produced.” Accordingly, he gives the following definition of “functional values:” “the functions of records while they were used by the creator.” 152 This definition is perfectly in line with the meaning of provenance-based approach provided above; however, later in the report, the term function is used in a different sense, so that in the end, it becomes an undifferentiated subject term. 153 In short, none of the early attempts of ‘functional analysis’ (including the one proposed by Samuels in Varsity Letters, which will later be analyzed in detail for its novelty) explains how to proceed in such analysis, that is, how to identify core and supporting functions of an institution (or type of institutions), how to derive from each identified function its component activities, and so on. As a matter of fact, Samuels, who claims to refer to the “definition of functional analysis adopted by anthropologists, 150  Clark A. Elliott, ed., Understanding Progress as Process: Documentation of the History of Post-War Science and Technology in the United States: Final Report of the Joint Committee on Archives of Science and Technology (Chicago: Distributed by the Society of American Archivists, 1983). 151 See Helen W. Samuels, ed., Varsity Letters. Documenting Modern Colleges and Universities (Metuchen, NJ and London: Society of American Archivists and the Scarecrow Press, Inc., 1992). 152 Elliott, Understanding Progress as Process, 12 and 18. 153 Other interesting conclusions drawn by Elliott, but unfortunately not followed up in the report, may be summarized as follows: 1) “having an understanding of the process of scientific research and technology innovation (i.e., knowledge of the organizational mission and functions, but also of the actual practices) is important in order to understand what is useful to preserve;” 2) “preserving the ‘top-of-the-iceberg’ documentation [like in the Schellenbergian approach] may not be enough; it is therefore necessary to analyze all levels’ functions and programmes;” 3) “different settings have different awareness of preservation needs.” Ibid., 24.  62  sociologists, and business managers” (i.e., “a descriptive technique to facilitate the examination of patterns across structures and cultures” 154), merely presents a list of seven functions which are supposed to be common to all higher education institutions in the United States, without really elaborating on her methodology. 155 Interestingly, in 1999, British archivist Elizabeth Parker published a function-based classification model relevant to the same type of institutions (though in the UK), and her top-levels differ in several aspects from those identified by Samuels. 156 According to Samuels, the way her contemporaries were used to approaching the analysis of organizational functions was “synonymous with structural analysis,” because the question archivists were usually interested in was “what is the function of a given office?” And she adds: “The traditional focus on administrative structures may be increasingly obsolete in light of the changing nature of modern institutions … Today, [the traditional pyramid] has been replaced by organizations that are differentiated not vertically, according to ranking and role, but flexibly and functionally.” 157  154  Ibid., 8. Ibid., 18. Samuels actually claims to have derived the “minimum set of functions” she identified from an “examination of the literature on higher education and particularly the vocabularies the academic community uses to describe itself,” as well as from “categories and concepts familiar to the archivists responsible for these records.” In other words, she appears to have conducted nothing more than a literature review and to have drawn on her own experience. 156 Elizabeth Parker, Study of the Records Life Cycle: Report by Elizabeth Parker, (Emerson Consulting Ltd., for TFPL Ltd for JISC, 1999). Available online at: http://www.kcl.ac.uk/projects/srch/reports/function.pdf (accessed on 5/12/2008). This observation confirms one of the points made in Orr’s thesis, i.e., that “the subjectivity in creating classifications creates problems.” Orr, “Functions-Based Classification of Records,” 71. 157 Samuels, Varsity Letters, 20. Italic added for emphasis by this author. David Bearman, in an article published in the same year as Samuels’ book, had similarly pointed out that a hierarchical analysis is not applicable where power and decision making cut through organizations rather than being concentrated at the top. In his view, post-WWII society has dramatically changed our perception of the internal dynamics of bureaucracy, due to the introduction of irrational and conflicting elements that “have transformed a stable framework into an inchoate type of organizational structure.” See David Bearman, “Diplomatics, Weberian Bureaucracy, and the Management of Electronic Records in Europe and America,” The American Archivist 55 (1992): 168-81. 155  63  As a consequence, Samuels claims that “appraisal methods must include the analysis of the functions of an institution no matter where they occur.” 158 The advantage of this purely functional method, absolutely independent of any structural constraint, would possibly be that of achieving an ‘integrated approach’ to documentation. In the context of documentation strategy, functional analysis becomes the tool that supports archivists in their search for the “activities that are to be documented” (i.e., what should exist, in whatever form or medium, and wherever, about a given “issue, function, or geographic area”). 159 It is however evident that, in this way, the conceptual boundary between function on the one hand, and subject, topic, or theme on the other, gets completely blurred. 160 Giving her insightful appreciation of the nature of modern bureaucracies, one may have expected Samuels, for instance, to suggest that, because existing analytical methods seem to have become inadequate to capture the complexity of today’s organizations, archivists should develop (or refer to) more sophisticated techniques of analysis. Instead, she addresses a method, called “institutional functional analysis,” which indirectly recalls the notion of ‘functional provenance’ as it was elaborated by David Bearman and Richard Lytle with reference to provenance-based access points to archives. In 1985-86, they had written:  158  Samuels, Varsity Letters, 24. Id. See also Helen W. Samuels, “Improving Our Disposition: Documentation Strategy,” Archivaria 33 (Winter 1991-92): 125-40. 160 Italian archivist Elio Lodolini pointed out that it is not unlikely to run across the phenomenon of ‘decontextualization’ when function and structure are split and the only criterion of arrangement of an archival fonds is the function. Although it may look as an attractive solution, in that functions are definitely more stable than organizational structure, in reality an arrangement by administrative function may lead to a subject-based arrangement, where “the subject is the function.” Lodolini refers to the massive rearrangements operated by Peroni in Milan at the end of ‘700-beginning of ‘800 as an example of such deleterious subject/function-based operations. See Lodolini, Archivistica: Principi e problemi, 56. 159  64  “It is probably more important to relate the records to a particular function than it is to relate them to an organizational component because there may be no relationship between the organization and the function. Functions are independent of organizational structures, more closely related to the significance of documentation than organizational structures, and both finite in number and linguistically simple.” 161 It seems to this author that splitting function and structure, and giving the primacy to function over structure, not only annihilate any difference between function and subject as seen before, but also contradicts the inseparable unity of the two concepts in the intellectual construct of ‘provenance,’ which is in fact to be seen as a representation of both an organization’s functions and its authority relations (i.e., structure). The latter are inherent in the organization design and manifest through its processes and roles (in structurational terms, they are part of the ‘structural properties of organizations’). The delegation of authority that establishes ‘functional responsibilities’ or ‘competences’ (in the meaning explained earlier in this chapter) is what brings conceptually together structure and function in any existing organization. These notions, which should be very clear to archivists not only when dealing with appraisal issues but also when arranging and describing archival holdings, are unfortunately often misinterpreted in the theory as well as in the practice. The weaknesses of Samuels’ argument did not pass unnoticed to Canadian archivist Terry Cook who, at the same time, recognized the validity of documentation strategy as a supplementary step after the appraisal of records according to their provenance, i.e., with reference to one single institution at a time. Cook’s method, known as ‘macro-appraisal,’ focuses on 161  Bearman and Lytle, “The Power of the Principle of Provenance,” 22. Actually, not the term but just the concept of “functional provenance” does appear in the Bearman-Lytle’s article; the term as such has been used for the first time by Angelika Menne-Haritz in 1993. See Angelika Menne-Haritz, ed., Symposium on the Impact of Information Technologies on Information Handling in Offices and Archives (New York, 1993).  65  “certain characteristics of the records creators and the record-creating process likely to produce records of high value before the resulting records themselves are actually appraised using more traditional criteria.” 162 The top-down approach of the first part of the macro-appraisal model aims at assessing the structural-functional circumstances which led to records creation with the purpose of identifying the agencies whose functions or programmes are the most relevant to society. The idea that “the interaction of structure and function together articulates the corporate mind (or programme) of the records creator” 163 is an outcome of Cook’s readings of Giddens’ structuration theory. This theory offers indeed a number of interesting insights with reference to the dialectic process of production and reproduction of social structures (including recordkeeping system) in social life. The structurational analysis of organizational dynamics could greatly assist archivists in their investigations of the functions and structures of their institutions. However, this author thinks that Cook’s conceptual elaborations on the nature of agents and acts and on the interrelationships between them do not yield operational outcomes that adequately correspond to the theory. The ranking of agencies and programs that the macro-appraisal approach would facilitate does not actually seem to be influenced by any structurational concepts. The same goes for the second step in the methodology, that is, the identification of the “points of sharpest interaction of the structure, function, and client,” 164 the “hot spots” where key functions, and thus key records, are likely to be found. How the theory of structuration is going to assist this process does not appear to be clearly addressed in Cook’s writings. 162  Terry Cook, “Documentation Strategy,” Archivaria 34 (Summer 1992): 185. On the macro-appraisal approach see also Catherine Bailey, “From the Top Down: The Practice of Macro-Appraisal,” Archivaria 43 (Spring 1997): 89-128. 163 Terry Cook, “Mind Over Matter,” 46. 164 Ibid., 50.  66  The macro-appraisal model has a precedent in the appraisal strategy elaborated by archivists Sante and Rohr in Germany after World War II and which was grounded on nearly the same principles. Decisions on the offices to be targeted were based on the “significance of an agency’s function,” which, at that time, meant: the higher the hierarchical position of an agency, the most relevant the records. Unavoidably, such a criterion for selection was soon interpreted as “whatever comes from the State possesses inherent value.” 165 In a similar way, the “theory of societal image formation” that underlies macro-appraisal also reflects a particular Weltanschauung, i.e., the one of Cook and the Canadian society of his time. 166 Therefore, despite the body of research that, according to Cook, must precede appraisal (which includes studies of administrative history, organizational structure and functions, decision-making process, records creating procedures, etc.), it appears that there is a high likelihood to fall into ideological traps when it comes to establishing criteria for value standards that are external to the records, like in the case of function-based appraisal. ‘Playing with functions’ has inspired a number of variations to the macroappraisal à la Terry Cook, all basically deriving from the consideration that because “there are altogether too many records ‘at the bottom’ for archivists to appraise,” 167 the traditional empirical approach would be unfeasible. Victoria Lemieux, for instance, through an analysis of Mintzberg’s theories, 168 suggests a method that would allow  165  See Hans Booms, “Society and the Formation of a Documentary Heritage: Issues in the Appraisal of Archival Sources,” Archivaria 24 (Summer 1987): 99-102. 166 According to Cook, archivists must be especially interested in circumstances where citizens consciously interact with the agency and have room for intervention and influence on the decisions made, and where therefore there is evidence of changes or distortions between an agency’s original goals and the actual results of a given programme. Also, where marginalized groups find a voice is a signal that should raise archivists’ attention. See Cook, “Mind Over Matter,” 50. 167 Cook, “Mind Over Matter,” 42. 168 See Mintzberg, Structure in Fives.  67  archivists to move away from evaluating record creators’ business functions towards evaluating record creators’ ‘functionalities.’ On the basis of the organizational configurations identified by Mintzberg, one could determine which functions are “organizationally significant” without the need to analyze the actual functions, and that would lead to the identification of the “sites of archivally significant records.” 169 No practical implementations of Lemieux’ approach has been attempted yet. The PIVOT project, 170 launched in the Netherlands at the beginning of the 1990s, is another example, this time a concrete one, of the application of macro-appraisal ideas with the purpose of achieving a “mass reduction” (one of the project’s keywords) of the overwhelming quantity of records accumulated by the Dutch administrations since World War II and not yet processed. Unlike the Canadian model, which provides for the testing of any hypotheses made on the basis of macro-appraisal criteria against the actual records by means of traditional appraisal techniques (also known as ‘micro-appraisal’), the Dutch project completely eliminates any bottom-up analyses of the records. The project was based on two major assumptions, one being that “the vast majority of modern records are of a transactional nature and do not provide any knowledge of government policy and activity.” 171 Only those records that derive from the main programmes put in place by government agencies, with particular reference to “those functions that show the highest contribution  169  Victoria Lemieux, “Applying Mintzberg’s Theories on Organizational Configuration to Archival Appraisal,” Archivaria 46 (Fall 1998): 32-85. 170 Project Implementation Reduction Transfer Period (PIVOT in Dutch). The project was launched by the Dutch National Archives as a sort of ‘emergency plan’ following the entering into force of a new archival law that would accelerate the public access to archives from 50 to 20 years after records creation. See Peter Horsman, “Appraisal on Wooden Shoes. The Netherlands PIVOT Project,” Janus 2 (1997): 35-41; Roelof C. Hol, “PIVOT’s Appraisal of Modern Records: A ‘Floody’ Tale from the Dutch Experience,” South African Archives Journal 38 (1996): 5-15. 171 Horsman, “Appraisal on Wooden Shoes,” 37.  68  to the attainment of some important output in society,” 172 are thus worth being preserved, as they would be capable of transmitting a representative image of society. The second underlying assumption was that “the evidential value of the records derives from the value of the function.” As a consequence, an appraisal of functions and actors (i.e., structure, agencies) should be sufficient to decide on the fate of the records “… without their having been given a glance.” 173 The PIVOT project was indeed quite effective and efficient as to its stated objective of reducing the records by nearly 95% in ten years. However, scholars in general and historians in particular, afraid of the blind destruction of important sources, soon pointed out its drawbacks. It was certainly a limitation of the PIVOT methodology that of believing that bureaucracy is, in any instance, a rational system governed by laws, regulations, and clear procedures. Modern theories of organization show that the reality of bureaucracy is, on the contrary, that of a social system with unwritten rules and selfset goals, where a permanent tension “between interest, conflict, and power … [is] resolved through political means.” 174 However, the above mentioned criticism can possibly be extended to all current attempts to apply a function-based approach to any archival activity, not just to appraisal. It is this author’s conviction that all archivists share an oversimplified, naïve idea of bureaucracy, being the actual business processes characterizing today’s “polyhierarchical, flattened, matrix, networking organizations” 175 mostly unknown to those who are in charge of managing the records (often unknown also to those who create  172  Ibid., 38. Ibid., 40. 174 Morgan, Images of Organization, 148. 175 Bearman, “Diplomatics, Weberian Bureaucracy,” 173. 173  69  those records). As far as appraisal is concerned, the location of the ‘significant functions’ and relevant record series identified by means of macro-appraisal is anything but a straightforward exercise, once one realizes that a plurality of offices and records systems may have been involved in the same decision making process. Additionally, we have no guarantee that the records supporting an important function will actually contain valuable information. In particular, e-business transactions, virtual team work, and all sorts of interactive and dynamic ways of conducting business enabled by present information and communication technologies make reality extremely complex and difficult to analyze, also because of the rudimentary analytic techniques that currently are at records professionals’ disposal. From this analysis of the relevant literature, one may conclude that, of all archival functions, records appraisal is probably the one that has appropriated the most the functional language and the idea that everything has to be done through a top-down analysis, without a deep examination of what such analysis would involve. The relationship between structure and function is a factor that complicates the management of records not only during the earlier stages of their life cycle but also at the point of their arrangement and description. Actually, the issue there becomes even more problematic in that the passing of time triggers a dynamism that affects both authority relations and functional relations, but in different ways. 176 Because of their different lifespan, records creators on the one side, and records systems, together with the functions they relate to, on the other, may present an inextricable puzzle that the simple application of the principle of provenance may be unable to solve. While records creators (whether juridical or physical persons) usually display a rather unstable nature due to continuous  176  See Terry Eastwood, “General Introduction,” in Terry Eastwood, ed., The Archival Fonds: From Theory to Practice (Ottawa: Bureau of Canadian Archivists, 1992): 1-14.  70  changes in their internal structure and authority relations, high level functions tend to persist unaltered. Based on the principle according to which “records follow functions,” 177 the sets of activities those functions are made of, together with the records supporting them, get allocated to new agencies or agents subsequently to the transfer of relevant functional responsibilities. Besides - or better, inside 178 - the history of the institution, one can also see another history, that is, the “history of custody, control and use of the records,” 179 which may further complicate the identification of an archival fonds. 180 Not to mention the remarkable adaptability that recordkeeping systems show: even in the presence of changes to functions, they “simply continue, shedding old functions and absorbing new ones with surprising flexibility.” 181 The first reconsideration of the principle of provenance from which a new method of archival arrangement and description derived took place in Australia at the end of the 1960s-beginning of the 1970s. The country was undergoing a series of tumultuous administrative changes, which had of course an impact on the archives of affected government agencies. The challenge involved in dealing with ‘multi-provenance series’ suggested to Peter Scott the idea of abandoning the concept of fonds as a principle of 177  Margaret Cross Norton’s famous sentence, canonically defined as principle of “functional sovereignty” “lends – in the words of MacNeil – a measure of continuity and stability to administrative activity and the records generated from them.” Heather MacNeil, “The Context is All: Describing a Fonds and Its Parts in Accordance with the Rules for Archival Description,” in Terry Eastwood, ed., The Archival Fonds: 207. 178 Whether the history of the institution and the “history of the purely archival vicissitudes” that any fonds is subject to in the course of the centuries or decades are two different histories or the second is part of the first one is an issue that raised quite some debate within the Italian archival community in the 1970s. In particular, the clash involved Elio Lodolini and Filippo Valenti. The latter was convinced of the autonomy of the “history of the fonds” suggested by German archivist Brenneke, a hypothesis that Lodolini would later firmly reject. See Lodolini, Archivistica. Principi e problemi, 160. 179 Eastwood, “General Introduction,” 7. 180 An archival fonds may be defined as “the whole of the records of a given body, including the whole of the relationships among its parts.” See Cencetti, “Sull’archivio come ‘universitas rerum’,” 9. Translated by this author. 181 Peter J. Scott, C.D. Smith, and G. Finlay, “Archives and Administrative Change: Some Methods and Approaches (Part 4),” The Journal of the Australian Society of Archivists 8, 2 (December 1980): 527 (reprint Archives and Manuscripts). See also ibid., “Archives and Administrative Change: Some Methods and Approaches. (Part 2).” The Journal of the Australian Society of Archivists 7, 4 (April 1979).  71  physical arrangement. The integrity of the records’ original order would then be maintained at the level of the series, being the latter an “independent element not bound to the administrative context,” 182 while the integrity of the whole fonds would be preserved at the intellectual level only by means of various descriptive sheets and inventories of series for each relevant creating agency. The Commonwealth Record Series (CRS) system, built on Scott’s ideas, represents the Australian solution to the issues of archival control. 183 Initially, the system emphasised a provenance-based method of information retrieval, thus priority was given to the identification and description of agencies, persons, and series. However, it soon became clear that functions may usefully be drawn on to enhance retrieval. The “functions concept” prompted the development of “function indexing terms that index agencies by terms describing major responsibilities and functions” and a “Functions Thesaurus … with authorized terms allocated at agency level.” 184 As will be mentioned later in this section, access to archives is the area of archival studies where the functionbased approach seems to have so far produced the most useful results. What should be added here is that, through the CRS system, archival arrangement and description became the natural extension of the recordkeeping processes (starting  182  Peter J. Scott, “The Record Group Concept: A Case for Abandonment,” The American Archivist 29, 4 (October 1966): 497. Scott’s definition of series is: “a group of records that are recorded by the same agency (or agencies) and that are in the same numerical, alphabetical, chronological or other identifiable sequence; or result from the same accumulation or filing process and are of similar function, format or informational content.” Ibid., 505. By separating description from physical arrangement and by linking all contextual information to the series level, Scott failed to appreciate the value of description as a top-down process (i.e., from the fonds, to the series, to the single items). 183 See Chris Hurley, “The Australian (‘Series’) System: An Exposition,” in Sue McKemmish and Michael Piggott, eds., The Records Continuum. Ian Maclean and Australian Archives First Fifty Years (Melbourne: Ancora Press in association with Australian Archives, 1994): 150-72. 184 Mark Wagland and Russell Kelly, “The Series System. A Revolution in Archival Control,” in Sue McKemmish and Michael Piggott, eds., The Records Continuum: 144.  72  from the accumulation of records in series) taking place in the creating agencies. This concept, that was echoed in the United States by Bearman’s considerations on recordkeeping systems as “the locus of functional provenance,” 185 later developed into the Australian ‘records continuum’ theory. 186 In opposition to the traditional stages identified by the life cycle model, the idea of the continuum is that of an uninterrupted, dynamic, and multi-layered process that provides the record with continuing contextual links to all the dimensions it participates in, and that makes any intervention on it (whether classification or appraisal or description) happen “at a number of points, at various times, and to different levels of aggregation.” 187 In this way, “records are in a constant stage of becoming” 188 and recordkeeping and archives become multidimensional. From this short outline of continuum concepts, it clearly emerges the influence of post-modernist and structurational ideas. Against this background, the role of ‘contextual’ or ‘recordkeeping metadata’ (i.e., the metadata connecting the record with information describing the actions surrounding its creation and use, also known as ‘process metadata’) acquires particular significance. As examined earlier in this chapter with reference to Australian classification, Chris Hurley elaborated the idea of ‘ambient function’ (as opposed to  185  David Bearman, “Record-Keeping Systems,” Archivaria 36 (Autumn 1993): 22. In the Australian interpretation, the idea of the ‘records continuum’ moved away from its original meaning and intents. In the ‘80s, Canadian archivist Jay Atherton had in fact contrasted the life cycle model (implying a clear-cut separation of responsibilities between records managers and archivists) with a new model based on a continuum of caretaking activities, thus fostering the integration of records management and archives. See Jay Atherton, “From Life Cycle to Continuum: Some Thoughts on the Records Management-Archives Relationship,” Archivaria 21 (Winter 1985-86): 43-51. 187 Barbara Reed, “Records,” Chapter 5, in Sue McKemmish et al., eds., Archives: Recordkeeping in Society (Wagga Wagga NSW: Charles Sturt University. Centre for Information Studies, 2005): 107. 188 See Frank Upward, “Structuring the Records Continuum. Part One: Post-Custodial Principles and Properties,” Archives and Manuscripts 24, 2 (1996): 268-85. 186  73  ‘business function’) to characterize the socio-historical context which represents the “context of provenance” 189 of the record. In another article where he investigates the meaning of function in the context of archival description, Hurley, in line with Scott’s observation of the different life-span of records systems and agencies, suggests the separation of the description of functions (which leads directly to the records) from that of the records creators. This would avoid the repetition of functional information for each agency entrusted with the same functional responsibilities. It would also allow concentrating all data about records and agencies that are linked to a given function in one single point, that he calls “functional unit of description.” 190 The International Standard for Describing Functions (ISDF), 191 recently published by the International Council on Archives (ICA) with the purpose of providing “guidance for preparing descriptions of functions of corporate bodies associated with the creation and maintenance of archives,” 192 is based on the same understanding. Despite their retrospective approach, functional descriptions provided by findings aids that standards like ICA-ISDF help to develop may be also beneficial to prospective approaches like that of classification. The language of functions has only recently attracted archivists’ full attention. Such an interest may be related, on the one hand, to the new capabilities offered by automated retrieval systems, and on the other, to the increasing public nature of archival description that fosters easier access to archival resources. Traditional archival finding 189  Hurley, “Ambient Function,” 25. Hurley, “What, if anything, is a Function?,” 214. 191 See International Council on Archives. Committee on Best Practices and Standards, International Standard for Describing Functions. First Edition (Paris: International Council on Archives, May 2007). Available online at: http://www.ica.org/sites/default/files/ISDF%20ENG.pdf (accessed on 15/01/08). ISDF aims at complementing ISAD(G) and ISAAR(CPF), which are well-accepted standards for the description of records and the preparation of authority records respectively. 192 Ibid., 7. 190  74  aids and provenance-based inventories require indeed a knowledge of archival methods that is not common within the general public. On the other hand, content-based indexing is not suitable to archival material. Functional access, which involves “function terms providing access to why records were created,” 193 has certainly the potential to become the most powerful access point in archives, as it would assist not only retrieval, but also classification, appraisal, and description. A controlled vocabulary of functional terms like the one provided by the Art and Architecture Thesaurus – Functions Hierarchy, 194 which is the outcome of several years of research in this area conducted by librarians and other information professionals in the United States, may be usefully drawn on also by the archival community. However, the basic issue here is again that our knowledge of functions, not in abstract terms but with reference to the way functions manifest and are enacted in the real world, is still not sufficiently developed to properly assist any standardization efforts. The fundamental question we should ask ourselves is, by borrowing Hurley’s words: “Is it our task, by observation, to discover and delineate what is there or to artificially construct an orderliness which is not real?”195 By examining the reality out there one may realize that, for instance, although functions are likely to be more stable than organizational structures, they “do evolve and change.” 196  193  Alden N. Monroe and Kathleen D. Roe, “What’s the Purpose? Functional Access to Archival Records,” in Toni Peterson and Pat Moholt, eds., Beyond the Book: Extending MARC for Subject Access (Boston: G.K. Hall and Co., 1990), 157. Italic added for emphasis by this author. See also Monica Scott and Flavia Fonseca, “Methodology for Functional Appraisal of Records and Creation of a Functional Thesaurus,” in H. J. Williamson and M. Hudson, eds., Classification Research for Knowledge Representation and Organization. New York, Elsevier, 1992, 127-34. 194 Getty (The), Art and Architecture Thesaurus On Line. 195 Hurley, “What, if anything, is a Function?,” 211. 196 Ibid.  75  In this respect, reference should be made to another attempt to standardization – though this time focused on business processes and documentary forms – that was different in its reasoning from previously examined taxonomies. In 1985, a project called “Commentaries on Sources” was launched in the Netherlands with the objective of identifying and describing the procedures and record types used by Dutch government bodies during the 19th century to carry out their mandates. 197 Although the outcomes of the project are limited by the specificity of the sample selected, and thus hardly generalizable, the approach taken is quite interesting. Researchers focused on trying to identify recognizable patterns of actions starting from a diplomatic analysis of the records, rather than relying on a logical breakdown of functions and activities identified through a study of the organizations’ mandates without examining any actual aggregations of records. This author is convinced that a bottom-up, empirical approach like the one here briefly presented, supported by the use of contemporary diplomatics, might generate new insights that could enhance not just descriptive standards but also the theory and practice of classification.  2.3  Review of Organization Studies and Library Literature  The following four sections provide an account of theories, concepts, and methods that have been developed in the context of other disciplines such as theory of organization, sociology, social-psychology, management science, and theory of administration, which the present study draws on in its attempt to understand organizational or business functions and how people interact with them in real-world situations. The fifth and last section is dedicated to some notions derived from the library and information science 197  See Peter Sigmond, “Form, Function and Archival Value,” Archivaria 33 (Winter 1991-92): 141-47.  76  literature on classification. Some of the ideas here presented have already been touched upon in previous sections where they have been used to discuss the theory of archival science in a new light.  2.3.1 Organizational Culture and Different Views of Organization ‘Culture’ is defined by sociologist Geert Hofstede as: “the collective programming of the mind that distinguishes the members of one group or category of people from another.” 198 Thus, culture is a ‘mental program,’ a ‘construct,’ which, as a collective phenomenon, may refer to any kinds of aggregation of human beings, from the most basic ones (e.g., families) to the highest complicated systems (e.g., societies). In his book, which reports the findings of an empirical study conducted in IBM firms in more than 50 countries during a period of time of four years, Hofstede is particularly interested in manifestations of ‘organizational’ and ‘national’ cultures and in the relationships among them. According to his interpretation of research findings, “cultural differentiations among countries have consequences for the functioning of and theorizing about organizations.” 199 Cultures are extremely stable over time. National cultures are characterized by deeply rooted ‘value systems’ that the people belonging to the same country introject since their childhood days. Such shared values become evident in individual and collective behaviours, as well as in the symbols, heroes, rituals, and other practices that, together, build the culture of a nation or – where people act as members of an organization – a specific organizational culture. 198  Hofstede, Culture’s Consequences, 9. Ibid., 10. As to “theorizing,” Hofstede discusses the cultural relativity of management theories in: Geert Hofstede, “Motivation, Leadership, and Organization: Do American Theories Apply Abroad?” Organizational Dynamics 9, 1 (Summer 1980): 42-63. 199  77  Hofstede identifies five main ‘dimensions’ (i.e., power distance, uncertainty avoidance, individualism vs. collectivism, masculinity vs. femininity, long-term vs. short-term orientation) along which value systems can be ordered and which determine the ‘character’ of any human groups. As far as organizations are concerned, ‘power distance’ (PD) and ‘uncertainty avoidance’ (UA) are the crucial dimensions. PD is that which answers the question of “who decides what.” 200 In other words, it defines hierarchical relationships in organizations: the larger the PD, the higher the concentration of authority. UA is related to the question of “how one can assure that what should be done will be done.” 201 In organizations, UA manifests in technology, rules, and rituals, which are all means potentially capable of reducing internal uncertainty caused by people’s behaviours, keeping people together, and exercising control on the future. 202 In  200  Power distance is defined as: “the extent to which a society accepts the fact that power in institutions and organizations is distributed unequally. … [PD measures the] degree of inequality underlying each human society.” See Hofstede, “Motivation, Leadership, and Organization,” 45. 201 Uncertainty avoidance is defined as: “the extent to which a society feels threatened by uncertain and ambiguous situations and tries to avoid these situations by providing greater career stability, establishing more formal rules, not tolerating deviant ideas and behaviours, and believing in absolute truths and the attainment of expertise. … [UA measures the] degree to which a society tries to control the uncontrollable.” (Ibid.) 202 Among the rituals, Hofstede mentions “business meetings, management training programs, writing and filing of reports and memos, accounting, planning and control systems.” Hofstede, Culture’s Consequences, 382. [Italic added for emphasis by this author]. Classification systems and any other records management and archival tools may therefore be regarded as both rational and symbolic means used to enhance UA in organizations. One may draw interesting parallels with what Hofstede writes about accounting: “Accounting is the handling of symbols that have meaning to the initiated in the business only … Objectivity in accounting is a myth … Accountants are the priests of business … The less an activity is determined by technical necessity, the more it is ruled by values and thus influenced by cultural differences. Accounting is a field in which the technical imperatives are weak. … So it is logical for the rules of accounting and the ways they are used to vary along national cultural lines. … In strong UA countries, accounting systems will contain more detailed rules as to how to handle different situations; in less strong UA societies, more will be left to the discretion of the organization or even of the accountant.” (Ibid., 382-83.)  78  short, while PD measures the authority of people, UA measures the authority of rules; PD relates to centralization, UA to formalization. Both dimensions have been used by Hofstede to categorize organizations in four basic types which are most likely to be associated with different countries. 203 These types are: 1. Personnel bureaucracy, or family model (characterized by large PD and weak UA, and typical of China, Hong Kong, Singapore, and other Asian countries); 2. Full bureaucracy, or pyramid model (characterized by large PD and strong UA, and typical of, inter alia, Latin and Mediterranean countries); 3. Workflow bureaucracy, or well-oiled machine model (characterized by small PD and strong UA, and especially present in, inter alia, German-speaking countries and Finland); and 4. Implicitly structured, or market model (characterized by small PD and weak UA, and most likely found in Anglo-Saxon countries, Scandinavia and the Netherlands). Hofstede’s categorization is consistent with Mintzberg’s fivefold classification of organizational structures, 204 although the methods of analysis employed by either sociologist as well as the conclusions they respectively reach differ in many aspects. Mintzberg sees organizations containing up to five parts (i.e., operating core, strategic The latter considerations have been taken into account when selecting the case study sites for this research. It was expected to find higher interest in records management and better developed records-related rules and tools in countries with a high UA than in countries with a low UA. 203 Ibid., 375-77. 204 See Henry Mintzberg, Structure in Fives. Mintzberg identifies the following five configurations: 1. Simple structure (corresponding to Hofstede’s personnel bureaucracy); 2. Machine bureaucracy (corresponding to Hofstede’s full bureaucracy); 3. Professional bureaucracy (corresponding to Hofstede’s workflow bureaucracy); 4. Divisionalized form (which is a mix of all four Hofstede’s types); and 5. Adhocracy (corresponding to Hofstede’s implicitly structured model).  79  apex, middle line, technostructure, and support staff). Additionally, organizations coordinate activities in one or more of five mechanisms (i.e., mutual adjustment, direct supervision, standardizing of work processes, standardizing of outputs, and standardizing of skills). Most organizations show one of the five configurations he identified because the part that, at each given time, is ‘key part’ in the organization is usually characterized by a work style that corresponds to one specific coordination mechanism (e.g., the operating core favours standardization of skills; the strategic apex achieves control and coordination through direct supervision; etc.). However, the key part may change at any time within one organization, and specific circumstances may require the adoption of different coordination mechanisms. An effective organization depends on developing a cohesive set of relationships between the internal factors (e.g., structural design, age, size, technology of the organization) and the external conditions in which it operates. Thus, Mintzberg’s model is more dynamic than Hofstede’s, and allows more configurations to be contemporaneously present in the same place. The necessity and predictability that Hofstede ascribes to the way in which “dominant value systems affect human thinking, feeling, and acting, as well as organizations” 205 has been criticized by various authors who have questioned the validity of his generalizations. 206 In particular, Hofstede’s tendency to identify cultures with nations shows an over-simplified understanding of both concepts, especially considering the global character of most of today’s nations and organizations. Nevertheless, his equation has been usefully drawn on by a number of studies. For instance, information management implications of the different organizational types identified have been 205  Ibid., 12. See, among the harshest critics, R. F. Baskerville, “Hofstede Never Studied Culture,” Accounting Organizations and Society 28, 1 (2003): 1-14. 206  80  discussed by Bearman with reference to recordkeeping issues, Davidson and Jordan with reference to the implementation of information systems, and Oliver with reference to the interaction of organizational culture with information culture. 207 This author’s study as well has referred to Hofstede’s categorizations in order to frame its scope and to establish a basis for comparison. The questionnaire employed for the selection of case study sites did include two questions explicitly referring to the PD and UA indicators. However, as will be discussed in a later chapter, it was not among the objectives of this research that of verifying, or falsifying, Hofstede’s conclusions. By examining the concept of organizational culture in more detail, one realizes that, in Hofstede’s view, national cultures and organizational cultures are in fact phenomena of a different order. Because the learning of organizational culture occurs in adulthood and people usually do not live in ‘total institutions,’ what an organization “is,” or “has,” 208 does not reach the depth and richness of the socially shared understanding typical of the cultures studied by anthropologists. Also, “at the organizational level, – Hofstede claims – cultural differences reside mostly in practices and less in values.” 209 Thus, “shared perceptions of daily practices should be considered the core of an organization’s culture.” 210  207  See David Bearman, “Diplomatics, Weberian Bureaucracy”; R. Davidson and E. Jordan, “Cultural Factors in the Adoption and Use of GSS,” City University of Hong Kong Working Paper (1996); Gillian Oliver, “Investigating Information Culture: A Comparative Case Study Research Design and Methods,” Archival Science 4 (2004): 287-314. 208 According to Hofstede, there is no consensus about the definition of organizational culture. Some authors treat it as something an organization is (which he calls “synthetic approach”), others as something an organization has (according to an “analytic approach”). Hofstede, on the basis of his definition of culture, provides the following definition of organizational culture: “Organizational or corporate culture is the collective programming of the mind that distinguishes the members of one organization from another.” Hofstede, Culture’s Consequences, 391. 209 Ibid., 394. 210 Ibid.  81  One may derive from this insight that recordkeeping, as a practice, is a factor that contributes to shape the culture of the organization. This attaches a certain importance to the role of records managers and archivists in organizations, an importance though that gets rarely recognized. “Organizations are symbolic entities: they function according to implicit models in the minds of their members.” 211 Hofstede seems to value particularly the minds of top managers, as they are the ones who may change an organizational culture to adapt it to a new strategic vision. They can do so, for instance, by changing the structure or the control system of the organization. However, as culture is hard to change not only because it exists in the “collective mind” of people, but also because it is “crystallized in the institutions these people have built together,” 212 any attempts to change it take time. Hofstede’s deterministic approach to organizational culture does not contemplate the possibility that unexpected outcomes may emerge from managers’ actions, or that external or other uncontrollable factors may shape the organization in unplanned ways. Gareth Morgan, who examined the image of “organizations as cultures” as one of the possible metaphors that may be used to describe organizations, has a different opinion with regard to managers’ ability to mould an organization’s culture. He writes: “Our understanding of culture is usually much more fragmented and superficial than the reality. … Like organizational structure, culture is often viewed as a set of distinct variables, such as beliefs, stories, norms, and rituals, that somehow form a cultural whole. Such a view is unduly mechanistic, giving rise to the idea that culture can be manipulated in an instrumental way. … Managers can influence the evolution of culture …  211 212  Ibid., 383. Hofstede, “Motivation, Leadership, and Organization,” 43.  82  but they can never control culture in the sense that many management writers advocate.” 213 Morgan’s view of organization is that of a phenomenon which is “generally complex, ambiguous, and paradoxical.” 214 Most people (organizational theorists, managers, but also archivists, as seen before) try to override such a complexity by assuming that organizations are ultimately rational phenomena to be understood with reference to their goals or objectives. This, according to Morgan, is not the right approach if one truly wishes to understand an organization. His method of analysis is multi-perspective and relies on metaphors as a powerful means to ‘read’ a situation, in that they allow being “creative and disciplined at the same time.” 215 Morgan’s analysis involves the following metaphors: organizations as machines, organisms, brains, cultures, political systems, psychic prisons, flux and transformation, instruments of domination. Each image is characterized by specific relationships among the internal components of the organization (e.g., structure, functions, people, and technology) and between the organization and its external environment. Each has its strengths and limitations, and has been implicitly or explicitly drawn on by various theorists of organization, sociologists, and philosophers. Morgan presents all these different views not to support any of them but rather to say that “organizations can be many things at one and the same time.” 216 Insights provided by the metaphors can also be used prescriptively:  213  Morgan, Images of Organization, 139. Ibid., 17. 215 Ibid. 216 Ibid., 321. 214  83  “As we understand an organization through the lens provided by a particular metaphor, we are shown a way of managing and designing the organization in accordance to a particular image.” 217 Effective managers and professionals are those who have the ability to read any given situation from different angles and to choose the most appropriate action suggested by those different views. This may be valid not only for top managers or the units in charge of organizational planning. Records professionals have also to make decisions that require a great deal of organizational understanding (e.g., how to design a classification system). Being supported by a wide and varied range of viewpoints about the social reality is certainly an asset to them. The problem is that both records managers and archivists may not know any other ways of reading the situation around them besides the ‘classic’ Weberian one, which they learn in the course of their formal education and which permeates most of their literature. 218  2.3.2 Systems Approaches to the Study of Organizations A system may be defined as “a set of elements connected together which form a whole.” 219 Each element in a system shows properties that are properties of the whole, rather than properties of its component parts. Types of ‘wholeness’ exist both in the natural sciences and in the social sciences; however, the “organized complexity” (which is another way of describing a system) that characterizes the latter is not subject to any of the principles developed by scientists to cope with it, i.e., reductionism, repeatability, and  217  Ibid., 331. Michael Lutzker had already in the ‘80s pointed out that “archivists keep on referring to a traditional, Weberian image of bureaucracy” which would prevent them from grasping how contemporary organizations actually work. See Lutzker, “Max Weber and the Analysis of Modern Bureaucratic Organizations,” 125. 219 Checkland, Systems Thinking, Systems Practice, 6. 218  84  refutation. The “‘messy’ nature of social phenomena,” 220 which does not allow generalizations or predictions on the pattern of the natural sciences, is related to the fact that one of the components of social systems involves individual human beings, who have self-consciousness and thus freedom of choice. This implies that, in any social facts or acts, there is always a mix of intended and unintended effects. Of course, social scientists have always been trying to bring the same kind of order and predictability of the natural sciences to their areas of interest. But evidence shows that they have not been very successful. Management science is one of those areas where both scholars and practitioners have the tendency to apply ‘hard’ systems methodologies to ‘solve’ problems existing in organizations. Systems engineering and systems analysis are just two of the most popular methods. Both proceed according to a similar problem-solving, goal-oriented approach: the problem and the desired outcome are given, and alternative ways of achieving that outcome are studied in order to select the best one to meet the identified need. 221 They both build models as simplified representations of the reality under examination and focus on optimizing such models. At the end of the process, the solution that is regarded as being the most effective, and possibly also the cheapest, is transferred into the real world. So much of current approaches to electronic document and/or records management systems (EDMS or EDRMS) development and implementation reminds of the engineering methodology! It is then not by chance that statistics and reports published in the last few years address the issue of the failure of several EDRMS  220 221  Ibid., 68. See ibid., 128-37.  85  projects. Some cannot be completed because of difficulties encountered in the implementation phase; others are ‘boycotted’ by the end-users who either refuse to use or misuse the new system. The same may apply to records classification design and implementation; the only difference is that the results of these projects usually do not get published. At a closer look, the standards lately developed to support the records management function (e.g., ISO 15489, MoReq, DIRKS), though they might be ‘optimal models’ of the reality, do not seem to offer a viable ‘solution.’ Indeed, even the records management program best compliant with a given standard may be ineffective in practice, because it may not align sufficiently with the needs of business, may not solve a perceived problem, or for some other reasons that are not system related. By trying to ‘engineer’ records management, EDRMS project managers as well as classification systems and standards developers implicitly apply ‘hard’ systems methodologies to “problem situations” that, being fundamentally unstructured, would require a different approach. In “human activity systems” (i.e., systems that “feature human beings in social roles trying to take purposeful actions”), 222 the primary uncertainty relates to the definition of the ‘problem’ and the precise objectives to be met. 223 To this purpose, no mechanistic and goal-oriented method can be effective. ‘Soft’ systems methodology (SSM) presents itself as an alternative, holistic way of dealing with ill-structured problems, not with the goal of ‘solving’ them, but rather of gaining understanding and, 222  Checkland and Scholes, Soft Systems Methodology in Action, 24. While structured problems are problems that can be explicitly stated and for which a solution exist (in the world of ‘hard’ systems thinking), unstructured problems manifest in a feeling of unease but cannot be explicitly stated without this appearing to oversimplify the situation. In other terms, “A problem relating to real-world manifestations of human activity systems is a condition characterized by a sense of mismatch, which eludes precise definitions, between what is perceived to be actuality and what is perceived might become actuality.” Checkland, Systems Thinking, Systems Practice, 155. 223  86  ideally, improving the ‘problem situation’ concerned. Peter Checkland describes the basic idea of SSM in these terms: “to formulate some models which hopefully will be relevant to the realworld situation, and to use them by setting them against perceptions of the real world in a process of comparison.” 224 So SSM is also interested in model building, yet in a way that is different from that of ‘hard’ systems thinking. First of all, we are not talking of one model only, but of as many models as the viewpoints that exist of the situation to be improved. The conceptual models of SSM are not ‘models of’ any given reality; rather they are ‘models relevant to a debate’ about a situation that is perceived as being problematic. They are constructed in order to explore perceptions of the real world, perceptions which are collected throughout the SSM exercise. Basically, what the researcher who embraces SSM is supposed to do is to try to find out as much as possible about the situation under examination, to “build up the richest possible picture,” 225 without imposing any particular structure on it. Thus, SSM, as a means to provide an “ordering framework for problem-solving,” 226 involves a process of continuous learning. Its objective is that of generating debate about possible changes together with the participants in the problem situation. SSM has been briefly introduced here not because the present study is designed according to its features. Actually, it is in action research that SSM finds its natural place as a research methodology. The reason for considering SSM, as part of the ‘soft,’ interpretive strand of thinking in the information systems world, relevant to this study is  224  Checkland and Scholes, Soft Systems Methodology in Action, 177. Ibid. 165. 226 Ibid. 61. 225  87  that its emphasis on the “irreducible complexity of real-world situations” 227 may help us obtain a richer picture of the concept of an organization than the conventional one. Instead of experimenting with metaphors like in Morgan’s example, with SSM we learn to ‘lend our ears’ to all the different voices that, in an organization, have a viewpoint on a given problem situation. Through this learning process, we may elaborate explanations of why our ‘models of’ the organizational reality, once they are translated into practice (whether as an EDRMS or classification system or records management standard), do not have expected outcomes. Also from the perspective of system development, SSM may provide a conceptual framework to make the shift in focus necessary to improve our records-related tools. While traditional engineering or requirement approaches focus on the “system that serves” (i.e., any information or records system as a system that serves and supports decision making in organizations), SSM prioritizes the “system served,” the real world. 228 Rather than perfect solutions that only work on paper, we may then achieve less perfect but practicable solutions that people can ‘appreciate’ (i.e., take into their “appreciative systems” 229) and use.  227  Ibid. 90. Peter Checkland and Sue Holwell, Information, Systems and Information Systems. Making Sense of the Field (Chichester, West Sussex, UK: John Wiley & Sons Ltd., 1998), 109-18. 229 The notion of “appreciation” is here used in the specific sense in which sociologist Vickers employs it. According to Vickers, our previous experiences create for us certain standards or norms which lead to the readiness to see (or appreciate) only certain features of the reality. These features or aspects of the reality are organized into “appreciative systems” which create for all of us, individually and socially, our “appreciated world.” The appreciative settings condition new experiences but are also modified by the new experiences in a recursive way that is similar to that of the theory of structuration. See G. Vickers, The Art of Judgement (London: Chapman and Hall, 1965). 228  88  2.3.3 Understanding Technology and Organizational Change through the Theory of Structuration As mentioned in Chapter 1, structuration theory is part of the theoretical framework this study is built on. 230 Within the so-called post-modernist trend in the archival studies, 231 Giddens’ ideas have often been drawn on as they provide a more dynamic view of the records and the interactions between them and the records creators. The present study has particularly been inspired by one of the outgrowths of such a theory, known as Adaptive Structuration Theory (AST), 232 which focuses on technology as a specific subset of social structures. Arguably, any records management or archival tool can be considered ‘technology,’ with reference not only to electronic records management tools but also, for instance, to classification as a τέχνη (in Latin characters, techné, i.e., craftsmanship or art, as a rational method which implies knowledge of principles and is oriented to practical outcomes) that helps us to do certain things in a given way and that embeds the ideas we have about its features and function. The theory of structuration offers an alternative way of conceiving the social reality that reconciles the long-standing opposition between objective and subjective  230  See Anthony Giddens, Central Problems in Social Theory: Action, Structure and Contradiction in Social Analysis (Berkeley, CA: University of California Press, 1979); Id., The Constitution of Society. 231 See Terry Cook, “Electronic Records, Paper Minds: The Revolution in Information Management and Archives in the Post-Custodial and Post-Modernist Era,” Archives and Manuscripts 22 (November 1994): 300-28; Id., “Fashionable Nonsense or Professional Rebirth: Postmodernism and the Practice of Archives,” Archivaria 51 (Spring 2001): 14-35; Frank Upward, “Structuring the Records Continuum. Part One”; Id., “Structuring the Records Continuum. Part Two: Structuration Theory and Recordkeeping,” Archives and Manuscripts 25, 1 (1997): 10-35; Sue McKemmish et al., eds., Archives: Recordkeeping in Society; Mark A. Greene, “The Power of Meaning: The Archival Mission in the Postmodern Age,” The American Archivist 65 (Spring-Summer 2002): 42-55; Heather MacNeil, “Trusting Records in a Postmodern World,” Archivaria 51 (Spring 2001): 36-47. 232 See Orlikowski, “The Duality of Technology”; DeSanctis and Poole, “Capturing the Complexity in Advanced Technology Use”; Marshall S. Poole and Gerardine DeSanctis, “Use of Group Decision Support Systems as Appropriation Process,” Proceedings of the Hawaii International Conference on Information Systems (1989): 149-57; Ibid., “Understanding the Use of Group Decision Support Systems: The Theory of Adaptive Structuration,” in J. Fulk and C. W. Steinfield, eds., Organizations and Communication Technology (Newbury Park, CA: Sage, 1990), 173-93.  89  understandings of it. Instead of considering social structures (i.e., the organization, in Giddens, or the technology, in AST) as either an external force that shapes and determines human actions (as the ‘decision-making school’ believes) or a product, a construction of human agents (as the ‘institutional school’ has it), structuration theory incorporates both views by recognizing that human actions are enabled and constrained by structures, yet these structures are the result of previous actions. 233 Structure must be understood as the “structural properties of social systems” 234 consisting of the rules and resources human agents draw on in their everyday interactions. These rules and resources mediate human action while, at the same time, they are reaffirmed through being used by human actors. This concept, also known as “duality of structure,” becomes in AST “duality of technology,” to signify that “technology is created and changed by human action, yet it is also used by humans to accomplish some action.” 235 The ‘structurational model of technology’ developed by AST acknowledges the mutual interaction existing between human actors and technology, thus considering the latter as both structural and socially constructed. In other terms, there would be a dialectic interplay between the structures that are embedded in a given technology and the structures that are brought about every time people use that technology. Thus, all technologies are always potentially modifiable, and there is nothing deterministic in any organizational change related to the introduction of a new technology. It is this author’s conviction that, not differently from when they refer to outdated concepts of bureaucracy or apply ‘hard’ systems approaches, when it comes to describing  233  DeSanctis and Poole, “Capturing the Complexity,” 121-25. Orlikowski, “The Duality of Technology,” 404. 235 Ibid., 405. 234  90  the reality of organizational change, records professionals still tend to see any external force as something that is beyond their control and that exerts a one-way, necessary impact on their world. The “technological imperative model” 236 of the decision-making school is indeed very much present in most of the archival literature discussing the ‘impact’ of the new technologies on archives, 237 but one may also find it in the workplace. However, as a justification for this tendency, one should see that the recursive relationship between technology and action may be difficult to recognize especially where technology design is separated in time and space from technology use. In such cases, it is normal for users to perceive technology as a ‘black box.’ Orlikowski calls this phenomenon “interpretive flexibility of technology,” 238 which means that different degrees of interaction are possible according to the characteristics of the material artifact (e.g., its age), the characteristics of the human agents (e.g., their reflexivity), and the characteristics of the institutional context (e.g., social and historical circumstances). Following this explanation, one may assume that a recently developed classification scheme has a greater potential to be challenged by its users who will try to modify it as much as existing circumstances allow. “Structuration is the process by which social structures are brought into action … [that is,] they are produced and reproduced in social life.” 239 Through the regular use of a technology, patterns of interactions become established as standardized practices in organizations. Over time, habitual use of such practices eventually gets institutionalized, forming the structural properties of organizations. These  236  Ibid., 400. See, for instance, Charles Dollar, Archival Theory and Information Technologies: the Impact of Information Technologies on Archival Principles and Methods, in Oddo Bucci, ed. (Macerata: University of Macerata, 1992), 45-49; and Menne-Haritz, ed. Symposium on the Impact of Information Technologies. 238 Orlikowski, “The Duality of Technology,” 407-09. 239 DeSanctis and Poole, “Capturing the Complexity,” 128. 237  91  are drawn on by humans and such use reinforces the institutionalized properties. However, ‘reproduction’ does not necessarily mean ‘replication.’ The recognition that human actors are “knowledgeable and reflexive” 240 is a central premise of structuration. These qualities imply that when enacting a technology, users discuss or at least think within themselves about how to use certain features, and they may, intentionally or unintentionally, change those features as they are using them. Thus, when the social structures of a technology are brought into action, they may take on new forms. “Appropriation” is the process by which “a group [of users] makes judgements about whether to use or not use certain [technology] structures, directly uses (reproduces) [them], relates or blends [them] with another structure, or interprets the operation and meaning of [the technology structures].” 241 The insight that users actively choose how to use the structures of a technology may explain why the results of the implementation of the same artifact (e.g., an EDRMS) may differ from organization to organization, and more generally, why it is impossible to predict how the implementation of a new technology is going to change an organization. Desired outcomes are not guaranteed, as human beings can always choose ‘to act otherwise.’ “Unintended consequences of technology” 242 is an expression that well captures this ‘getting away’ of the technology from its official, promoted use. The way people adopt and adapt any given technology depends on a series of factors, some related to group attitudes, some to the organizational environment, some to specific “appropriation moves.” 243 Among these, AST notes that technologies can be either ‘faithfully’ or ‘unfaithfully’ appropriated. Faithful appropriations are consistent 240  Orlikowski, “The Duality of Technology,” 406. DeSanctis and Poole, “Capturing the Complexity,” 129. 242 Poole and DeSanctis, “Use of Group Decision Support Systems as Appropriation Process,” 152. 243 Ibid., 153-54. 241  92  with both the spirit and the structural feature design, whereas unfaithful appropriations are not. Unfaithful appropriations are not ‘bad’ or ‘improper’ but simply not in line with the spirit of the technology. The latter is described as “the general intent with regard to values and goals underlying a given set of structural features.” 244 An interesting finding of AST research is that users system trainings in organizations mostly address structural features (i.e., the capabilities offered by the system) while the spirit of the system is hardly communicated to the users. Examining training material as well as how users get involved in the implementation phase of a new system appears therefore to be crucial to a study aiming at getting an understanding of the actual uses – or non uses – of classification systems in organizations. Structurational studies of the role of information technology (IT in a broad sense, from typewriters to cell phones) in ‘organizational transformation’ (i.e., “a shift in the way that work is done within a chartered collective” 245) reveal that the perceived causal relationship between the introduction of a new IT and radical changes in the organization of work is just “a widely held societal myth.” 246 The reality of ‘organizational transformation’ is that of a social process that unfolds gradually, over time, sometimes showing unexpected or inconsistent outcomes, and under the influence of nontechnological factors (such as, social, political, economic, and cultural forces) as well. The book on this topic edited by Yates and Van Maanen focuses not only on work practices, but also on the social structures supporting those practices, and the “ideologies  244  DeSanctis and Poole, “Capturing the Complexity,” 126. JoAnne Yates and John Van Maanen, eds. Information Technology and Organizational Transformation. History, Rhetoric, and Practice (Thousand Oaks, CA: Sage, 2001), xii. 246 Susan J. Winter and S. Lynne Taylor, “The Role of Information Technology in the Transformation of Work. A Comparison of Post-Industrial, Industrial, and Proto-Industrial Organizations,” in Yates and Van Maanen, eds., Information Technology and Organizational Transformation, 8. 245  93  and meaning systems” 247 that more or less legitimate them. The first part of the book collects studies that examine how IT gets adopted in organizations, including its intended and unintended consequences, from a historical perspective; the second part deals with the rhetoric of IT and organizational transformation; the third and last part concerns the practices that emerge when a new IT is made available to organizational members. The overall purpose of the book is that of “[raising awareness of] the multidimensional and ambiguous character of organizational change as well as the numerous uses to which IT can be put.” 248 Researchers must develop a special sensitivity to be able to appreciate what is below the surface of what may look like ‘technological progress,’ or ‘a story of success.’ The present research design does neither involve a series of longitudinal studies that would allow observing the interweaving of technology and human use throughout different socio-historical circumstances, nor does it employ ethnography, a methodology that would be ideal to make unstated meanings and invisible patterns emerge. Within the limitations of a case study research approach, this author will aim nevertheless to take advantage of the methodological suggestions derived from the examples collected by Yates and Van Maanen. In particular, a historical perspective may be obtained through asking informants and collecting material about previous arrangements and the history of the project that brought to the implementation of the system under investigation. Insights related to the rhetoric of the system may be elicited by engaging in discussions where the informants’ views are confronted with those of the researcher, so that networks of interpretations may emerge and reveal beliefs, disappointments, and other perceptions  247 248  Yates and Van Maanen, eds. Information Technology and Organizational Transformation, xii. Ibid., xvi.  94  about the system. As to studying shared practices and the interpersonal relationships formed around them, this researcher may try to ‘immerse’ herself in the daily work of the office examined within a quasi-ethnographic approach.  2.3.4 Organizational Behaviour and Related Issues In his discussion on the process of decision-making from the point of view of social psychology, Herbert Simon criticizes previous works in the area of theory of administration because of their separation of the world of ‘deciding’ from that of ‘doing’ and exclusive focus on policy-making. On the contrary, Simon states, “the task of ‘deciding’ pervades the entire administrative organization quite a much as the task of ‘doing’.” 249 By saying so, he invites us to analyze “purposive behaviours” 250 (i.e., behaviours oriented towards goals and objectives) at all levels in an organization. Although there is a hierarchy of decisions as there is a pyramid of goals, these are not perfectly integrated in any actual behaviour. Diverse and sometimes conflicting objectives make decisions hard to take, especially because not all possible alternatives are available under any given circumstance. Therefore, decision is always a matter of compromise, and organizations can never be perfectly rational due to the limited information-processing abilities of their members. In contrast to the assumptions made in economics about the optimizing behaviour of individuals, Simon concludes that individuals and organizations settle for a “bounded rationality” of “good enough” decisions based on simple rules of thumb and limited search and information. 251  249  Herbert A. Simon, Administrative Behaviour. A Study of Decision-Making Processes in Administrative Organizations, 4th ed. (New York: The Free Press, 1997), 14. 250 Ibid., 15. 251 Ibid., 23.  95  Morgan places Simon’s view of organization under the ‘brain metaphor,’ that is the idea that organizations are information-processing brains. 252 Everything in organizations, from departmental and job divisions, to hierarchies of authority, policies, programs, rules, standard operating procedures, exists for the sake of simplifying organizational reality in order to make it manageable to individuals. Among the accepted “principles of administration” elaborated by the theory of administration with the aim of enhancing organizational efficiency (principles that Simon demolishes one after the other) 253, one is particularly suitable to the topic of this research. The principle reads: “Administrative efficiency is increased by grouping the workers, for purposes of control, according to a) purpose, b) process, c) clientele, or d) place.” 254 The analogy with the terminology referring to the transactional level of classification schemes according to an archival tradition that goes back to Schellenberg is evident. Simon’s remark refers indeed to the ambiguity of key terms like ‘purpose’ and ‘process.’ To start with, he provides the following definitions: “‘Purpose’ may be roughly defined as the objective or end for which an activity is carried on; ‘process’, as a means of accomplishing a purpose.” 255 Thus, the same activity may actually be described as purpose or as process: it depends on the point in the hierarchy of purposes you look at the issue. Purposes form a hierarchy,  252  See Morgan, Images of Organization, 81-84. Simon, Administrative Behaviour, 29-30. Simon highlights the inherent ambiguity and insufficiency of each principle. His conclusion is that “Mutually incompatible advantages must be balanced against each other in the design of administrative organizations, as over-all efficiency must be the guiding criterion.” (45). One may say that the same works with reference to the design of records classification systems, although the efficiency criterion may not be the only relevant one. 254 Ibid., 36. 255 Ibid., 38. 253  96  each sub-purpose contributing to some more final and comprehensive end. In other words, there is no essential difference between purpose and process, but only a distinction of degree. The same can be said with regard to the distinction between function and activity levels in a classification scheme. Here is how Simon reformulates the two concepts: “A ‘process’ is an activity whose immediate purpose is at a low level in the hierarchy of means and ends, while a ‘purpose’ is a collection of activities whose orienting value or aim is at a higher level in the meansend hierarchy.” 256 Administrative gurus recommend arranging organizations by ‘major purpose’ so that all those who are dedicated to render a particular service can work together in a single large department. However, Simon asks, what is a particular service? His conclusion is that “there is no such thing as a purpose, or a unifunctional (single-purpose) [department]. What is to be considered as a single function depends entirely on language and techniques.” 257 Simon’s insight confirms the difficulties that are inherent in the language of functions and that make any classification work arduous. The means-end relationship mentioned above is used by Simon as a criterion to judge the correctness of administrative decisions: “An administrative decision is correct if it selects appropriate means to reach desired ends.” 258 “Rationality – he adds – has to do with the construction of means-ends chains.” However, these chains are seldom completely integrated and connected. Often the link between organizational activities (means) and ultimate objectives (ends) is obscure, or these ultimate objectives are either incompletely formulated, or there are contradictions 256  Ibid., 39. Ibid., 38. 258 Ibid., 56. 257  97  between the ultimate objectives and the means selected to obtain them. Once again, Simon stresses the fact that the “bounded rationality” of human beings is a limitation to rational decision-making and, in actual situations, it is usually impossible to separate completely means from ends. Simon provides a categorization of ‘rational behaviour,’ where rational is defined as something that serves a useful purpose. In particular, he states that “Rationality is concerned with the selection of preferred behaviour alternatives in terms of some system of values whereby the consequences of behaviour can be evaluated.” 259 Another interesting point he makes refers to the analysis of the role played by habits and routines in organizations. Habits and routines are also the outcome of decisions (‘once and for all’ decisions). Besides serving their purposes effectively, they help to “conserve scarce and costly decision-making time and attention.” 260 For this reason, a very large part of an organization’s activities is likely to proceed according to established rules and routines. Simon recognizes that, in today’s information society, “the critical scarce factor in decision-making is not information but time, attention.” 261 The quality of decisionmaking can be enhanced by “searching systematically, but selectively,” among potential information sources to find those that might be most useful. Thus, although he does not explicitly mention the role played by information or records systems in supporting decision-making, Simon seems to hint that, for the sake of administrative efficiency, good retrieval capabilities are the most valuable functionality those systems should provide. 259  Ibid., 86. Ibid., 102. 261 Ibid., 123. 260  98  2.3.5 Classification in Library and Information Science In the studies of Bowker and Star, classifications are seen as “powerful technologies” 262 which are embedded in all aspects of life, sometimes in invisible ways. “Classification schemes … literally saturate the worlds we live in.” 263 The stealthy presence of classification, once brought into light, may be used as a device for understanding the ethics, politics, hidden motivations, in one word, the cultures of any given societies. Anthropologists, who are used to studying classifications in these terms, are also well aware that “classifications arise from systems of activity and, as such, are situated historically and temporally.” 264 Classification, “the sleeping beauty of library and information science,” 265 not only is shaped by the culture of the social reality it represents, but is in turn responsible for shaping that culture. Bowker and Star provide the following definition of classification system: “A classification system is a set of boxes, metaphorical or not, into which things can be put in order to then do some kind of work – bureaucratic or knowledge production.” 266 Classification, in the sense of grouping things systematically, is therefore a basic human activity. Aristotle is claimed to be the first who stated the characteristics of systematic classification. In his “classical theory of categories,” the categories forming the classification scheme were arranged hierarchically, from the general to the specific, and a  262  Geoffrey C. Bowker and Susan Leigh Star, “Invisible Mediators of Action: Classification and the Ubiquity of Standards,” Mind, Culture, and Activity 7, 1-2 (2000): 147. 263 Ibid., 157. 264 Ibid., 149. 265 Susan Leigh Star and Geoffrey C. Bowker, eds., “How Classifications Work: Problems and Challenges in an Electronic Age,” Library Trends 47, 2 (Fall 1998), 185. 266 Geoffrey C. Bowker and Susan Leigh Star, Sorting Things Out: Classification and Its Consequences (Cambridge, Mass.: The MIT Press, 1999), 10.  99  category was described as “an abstract container with things either inside or outside the container.” What defined the category were “the properties the things inside the container had in common.” 267 In more recent years, a non-hierarchical type of classification, called ‘faceted classification,’ has been developed in order to provide more flexibility to the scheme. The suitability of the faceted method in records management and archives is questionable however for the same reasons presented earlier with reference to virtual files and multidimensional approaches to contextual metadata. 268 The properties of an ideal classification scheme have been described by Bowker and Star as follows: “- Each system should be based on a single classificatory principle; - The classes should be mutually exclusive; - The system [should provide] complete coverage of the world it describes.” 269 However, the authors admit that they have never seen a system that fully meets this ideal. Although not explicitly stated, the typical classificatory principle one may come across in libraries is the subject-based one. As to the reasons for classifying recorded information, librarians and information scientists seem to agree on the following main purposes of classification: “- To allow items to be arranged logically on shelves in order to: - Help users identify and locate items; - Group related items together so that users benefit from related items being co-located; - To allow a link to be created between items on shelves with entries in a catalogue or index.” 270  267  A.G. Taylor, The Organization of Information (Englewood: Libraries Unlimited, 1999), 174. See Shepherd and Yeo, Managing Records. 269 Bowker and Star, Sorting Things Out, 10. 270 See A. Maltby, Sayer’s Manual of Classification for Librarians, 5th ed. (London: Andre Deutsch, 1975); L.M. Chan, Cataloguing and Classification: An Introduction, 2nd ed. (New York: McGraw Hill, 1994); and J. Rowley and J. Farrow, Organizing Knowledge: An Introduction to Managing Access to Information, 3rd ed. (Aldershot: Ashgate Publishing, 2000). Cited in Orr, “Functions-Based Classification of Records,” 32. 268  100  Thus, classification in librarianship qualifies as a practical, retrieval-oriented technique that has nothing of the necessary-ness, or determined nature that characterizes archivists’ concept of classification. This makes the classification design definitely more straightforward in a library than it can ever be in an archives. The same may be said with regard to standardization efforts, which are strongly encouraged in the library community, much less in the archival one. American archivist Margaret Cross Norton expressed the different complexity that library and archives classification respectively entail in these terms: “The librarian has only to fit the books he is classifying into a preconceived scheme, but the archivist has to construct his classification scheme anew to fit the different types of records kept by each department.” 271 However, library classification has its difficulties and shortcomings too. In particular, with regard to its effectiveness, several authors, mostly quoting Spärck Jones who was writing in 1970, have pointed out that “for information retrieval (IR), theories of classification are inadequate and have not been sufficiently considered. … a substantive theory of classification is needed but does not exist.”272 IR experts seem to oscillate between two schools of thought, a “pragmatic” one, which sees classification as “a response to an objective,” and a “positivist” one, according to which classification is “an abstract process.” 273 The pragmatic method is based on the consideration that a classification is always required for a purpose; however, “whether  271  Mitchell, ed., Norton on Archives, 91. Birger Hjørland and Kartsen Nissen Pedersen, “A Substantive Theory of Classification for Information Retrieval,” Journal of Documentation 61, 5 (2005), 582. See also K. Spärck Jones, “Some Thoughts on Classification for Retrieval,” Journal of Documentation 26, 2 (1970): 89-101; Rick Szostak, “Classification, Interdisciplinarity, and the Study of Science,” Journal of Documentation 64, 3 (2008): 319-332; Birger Hjørland, “Core Classification Theory: A Reply to Szostak,” Journal of Documentation 64, 3 (2008): 333-42. 273 Hjørland and Nissen Pedersen, “A Substantive Theory of Classification for Information Retrieval,” 584. 272  101  that purpose can be stated formally is quite another question.” 274 Authors who follow this school emphasize the “investigation of goals, purposes, interests, and values.” On the opposite side, positivists tend to “keep to pure observations, logical deductions, and formal models,” while ignoring issues related to “interpretation and meaning as well as goals, purposes, and values.” In practice, according to the ‘positivist method,’ “classification of any kind of objects [should be] based on the properties of those objects.” 275 The interpretivist argument that the supporters of the other school invoke against the positivist method is that the properties of any objects are not ‘facts;’ rather, they “are only available to us on the basis of some descriptions,” and those descriptions may vary from one observer to another, or they may be theory-dependent. 276 These considerations are indeed quite appropriate to frame the archivists’ problem with classification as well. On the one hand, it seems that the purpose of records classification is so confused that a pragmatic approach becomes inapplicable. On the other hand, archivists in the post-positivist time may find it difficult to state dogmatically that classification should be based on the properties of the records, assuming that everybody shares the same understanding of those properties and that any other factor external to them is irrelevant.  2.4  Summary  The account of the literature of archival science and other disciplines provided in the previous pages demonstrates that the areas of functional and records classification knowledge deserve an in-depth re-examination, which should start ‘from the bottom,’  274  Ibid., 585. Ibid. 276 Ibid, 586. 275  102  that is, from an analysis of actual instantiations of function-based classification systems in real-world organizations. In particular, the review of the archival literature has highlighted that, although a functional approach not only to records classification but also to any other archival endeavour is greatly promoted as the only or principal means of managing records and archives, the concept of function does not seem to be thoroughly understood. Nor are the nature and purpose of classification clearly and consistently stated throughout the literature. Drawing, implementing, and maintaining a function-based classification scheme appears to be more an art than an established methodology. Such a lack of clear guidance confuses the practitioners, as an examination of actual outcomes of their efforts generally shows. Additionally, the absence of empirical, situated studies does not allow drawing any conclusions as to the relationship between specific organizational settings and cultures, and the representations of those organizations’ functions and activities. The literature concerning organizational studies confirms that there are many factors influencing the ways in which people in organizations carry out their activities and, in particular, interact with the tools and technologies that are meant to facilitate their work (including classification systems). An approach that takes those social and cultural factors into consideration, together with a multi-perspective understanding of the phenomenon organization, seems to be missing in the landscape of archival studies. On a theoretical level, the reasons for this research are supported by the gap identified between the overall limited, oversimplified archival understanding of the functioning of organizations and the complex, dynamic, multi-dimensional, and even incoherent view of social reality that emerges from the various human activity systems studies examined. By bridging that gap through the application of inductive reasoning 103  based on empirical evidence, rather than through the deductive approach usually employed in archival research, this author hopes to contribute rich insights to the archival theory. On the practical level, this work is justified by the need for some clarification on the design of functional classification systems and, more generally, on the role these systems play, or should play, in organizations. Again, a direct analysis of the actual ways in which ‘things get done’ in real-world settings will reveal why records classification is such a difficult issue and will ideally provide some suggestions on how to do it better.  104  3.  RESEARCH DESIGN  3.1  Overview  This chapter provides a detailed account of the implementation of the interpretivist, inductive methodology adopted for this research. The main features of the methodology, including its theoretical underpinnings, advantages, and limitations, have already been outlined in the introductory chapter to this dissertation. The first section of the present chapter is dedicated to the survey-based approach used for the selection of the sites where the case studies that are at the core of this research strategy would be conducted. This section will also provide an initial characterization of the specific population chosen. It will be followed by a section describing the methods employed to carry out the multiple-case research design, including data analysis and reporting techniques. The third and final section will deal with the ethical issues concerning this research and the way they have been resolved.  3.2  Selecting Suitable Case Study Sites  Selection of suitable sites for conducting case study research requires careful consideration of the objectives the researcher aims to achieve. Thus, the sampling technique employed did not follow the logic of random sampling, but it was a “nonprobability sampling” technique of the type known as “purposive” or “theoretical sampling.” 277 Yin suggests that each case involved in a multiple-case study should be  277  Williamson, Research Methods, 231; Yin, Case Study Research, 31-33; Eisenhardt, “Building Theories from Case Study Research,” 537.  105  considered like an “experiment.” 278 This insight implies that the researcher will approach the selection of cases according to “replication” logic. “Each case must be carefully selected so that it either (a) predicts similar results (a literal replication) or (b) predicts contrasting results but for predictable reasons (a theoretical replication).” 279 Considering the hypotheses and research questions formulated by this author with particular regard to those aiming at exploring the relationship between different organizational cultures and recordkeeping approaches, the logic that applies to this study is that of theoretical replication. As mentioned in the methodology overview included in the Introduction, through this approach, “analytic generalization” (as opposed to “statistical generalization”) of the research findings may eventually be claimed. 280  3.2.1 Selection of Study Population Gillian Oliver (whose research on information cultures shares with this study the reference to Hofstede’s categorization of organizations as a basis for comparison among cases) recommends “select[ing] an organization type that is represented, and which will have similar functions, in each subject country.” 281 Similar functions is a criterion that particularly suits a study like this one, which addresses issues of functional analysis and is interested in exploring how, keeping the variable of function stable, different environments conceive and construct their function-based classification systems. The researcher’s familiarity with the central bank environment oriented her choice towards this class of organizations, which has an established presence in every country of the European Union and in North America (i.e., the two targeted areas) and has more or less 278  Yin, Case Study Research, 47. Ibid. 280 Ibid., 32. 281 Oliver, “Investigating Information Culture,” 299. 279  106  the same functions, though national differences in the types of activities involved do exist. Goals, functions, and legal framework characterizing central banks as a very special type of financial institutions, including relevant recordkeeping and archival issues, are described in the next chapter. Here, it will be sufficient to mention that another reason for choosing them as study population is that, unlike commercial banks, investment banks and other financial intermediaries, central banks have a research component that may qualify them as ‘think-tanks.’ Like universities, research centers and other institutions sharing a similar mission, central banks perform research functions in the field of economics and finance, an intellectual work that is not necessarily oriented towards practical applications. This influences organizational behaviour, modes of accomplishing certain activities, work relations with internal and external customers, and types of services offered to the society in a specific way. 282 Central banks are therefore supposed to be a good ‘laboratory’ for observing activities that are unstructured or unique, or at least, do not share the characteristics of linearity and repetitiveness which are typical of most business processes. Oliver notes that “conducting case studies of one’s own organization will rarely be appropriate because of potential problems with reliability and objectivity.” 283 Besides that, where the researcher is affiliated with the organization to be investigated, almost all interactions will be based on prior acquaintance and shared existing knowledge, thus the exploratory aims of the research might be frustrated. Thus, this author refrained from including her own work place among potential study subjects. Nevertheless, she took  282 283  See the analysis of “organizations as brains” in Morgan, Images of Organization, 77-109. Oliver, “Investigating Information Culture,” 300.  107  advantage of her familiarity with the functions and the records management needs of the organization type under examination. Such ‘insider knowledge’ allowed her to take an informed approach to her interviews, observations, and analyses of findings, and this has considerably reduced the time she had planned to spend in each case study site. Additionally, being already part of the central banks’ business environment undoubtedly facilitated the initial phases of the project and expedited the researcher’s admittance in the banks, which, in general, are notoriously quite restrictive when it comes to access to their premises and files.  3.2.2 Survey Design and Administration The sampling of cases from the chosen population involved the design of a web-based questionnaire. 284 The main advantage of a self-administered questionnaire is that it allows the researcher to reach a large number of subjects who are widely distributed geographically, and in a relatively short time. 285 A link to the survey was sent via email to thirty recipients identified as the person, or one of the persons, responsible for the records management and/or archival function in each of the thirty central banks selected for the study, i.e., the central banks located in North America and those belonging to the so-called European System of Central Banks (ESCB), except the European Central Bank (ECB) for the reasons mentioned above. The identification of survey recipients was greatly facilitated by the researcher’s ‘insider’ knowledge. The online questionnaire was administered for the first time at the end of September 2007. About two weeks later,  284  The researcher used the online survey tool Zoomerang, kindly offered by Dr Cenfetelli from UBC Sauder School of Business. A static copy of the questionnaire (and attached invitation letter) is included in Appendix 1. 285 Williamson, Research Methods, 236.  108  reminders were sent to those who had not replied yet. The survey was considered closed at the end of October 2007. The questionnaire was developed in English as a pre-requisite that potential research subjects had to meet to be selectable. It should be noted that all central banks which are part of the ESCB use English as a working language to communicate with each other, especially in the context of inter-institutional committees, working groups, and task forces. Of course, this does not mean that everybody in the banks speaks English. In order to make sure that the language factor would not be a hindrance to the conduct of fieldwork, the Invitation Letter enclosed with the online survey made explicit that “participants in interviews must be able to communicate in English or Italian with the investigator.” 286 A second criterion put before potential subjects for purposes of pre-selection was expressed, again in the invitation letter, in the following terms: “organization[s] must be using, or be in the process of designing, implementing or reviewing, a corporate records classification system as a means to identify and to organize the records made or received in the course of business.” 287 The survey included structured questions, mainly of the types known as “factual” and “closed” questions. The former type is suitable to obtain straightforward answers (e.g., yes/no); while the latter, by providing “frames of reference” that guide respondents’ replies, is meant to help clarify the concepts used. 288 Indeed, the ambiguity of some of those concepts (e.g., classification based on function, rather than subject, organizational structure, or record type; power distance and uncertainty avoidance) made it necessary to 286  See Appendix 1, 1. Ibid. 288 Williamson, Research Methods, 237-38; and Oppenheim, Questionnaire Design. 287  109  provide some explanation of the terms used in order to achieve a base for common understanding. Two questions aiming at eliciting opinions about types of authority relations and degrees of bureaucratization in the organization (both measured by means of Hofstede’s dimensions ‘power distance’ and ‘uncertainty avoidance’) 289 were constructed by using a Likert scale. 290 Questions were developed on the basis of specific criteria considered appropriate for the envisaged theoretical sampling. The review of the literature and the hypotheses and research questions thereby identified are to be regarded as the sources for the following primary criteria: •  Selected organizations must use records classification systems that they perceive as being function-based;  •  Selected organizations must display characteristics (in terms of degrees of ‘power distance’ and ‘uncertainty avoidance’) from which one may assume that each of them belongs to a different type of bureaucracy according to Hofstede’s categorization. 291  The latter criterion enables the researcher to compare diverse organizational cultures and work styles, including relevant recordkeeping practices, and eventually to reveal the characteristics of the function-record relationship in each different setting. 292 Other criteria considered less determining though still important for