The FRBR Family of Conceptual Models

Toward a Linked Bibliographic Future

Richard P. Smiraglia, Pat Riva, Maja Žumer

• Publisher: Taylor & Francis
• Year: 2014
• Print ISBN: 9780415641043
• eBook ISBN: 9781317850540
• Edition: 1
• Topic: Languages & Linguistics
• Subtopic: Library & Information Science
• Index: Languages & Linguistics

Since 1998 when FRBR (Functional Requirements for Bibliographic Records) was first published by IFLA, the effort to develop and apply FRBR has been extended in many innovative and experimental directions. Papers in this volume explain and expand upon the extended family of FRBR models including Functional Requirements for Authority Data (FRAD), Functional Requirements for Subject Authority Data (FRSAD), and the object-oriented version of FRBR known as FRBRoo. Readers will learn about dialogues between the FRBR Family and other modeling technologies, specific implementations and extensions of FRBR in retrieval systems, catalog codes employing FRBR, a wide variety of research that uses the FRBR model, and approaches to using FRBR for the Semantic Web.

Librarians of all stripes as well as library and information science students and researchers can use this volume to bring their knowledge of the FRBR model and its implementation up to date.

This book was published as a special issue of Cataloging & Classification Quarterly.

Be Careful What You Wish For: FRBR, Some Lacunae, A Review

RICHARD P. SMIRAGLIA

School of Information Studies, University of Wisconsin–Milwaukee, Milwaukee, Wisconsin, USA

The library catalog as a catalog of works was an infectious idea, which together with research led to reconceptualization in the form of the FRBR conceptual model. Two categories of lacunae emerge—the expression entity, and gaps in the model such as aggregates and dynamic documents. Evidence needed to extend the FRBR model is available in contemporary research on instantiation. The challenge for the bibliographic community is to begin to think of FRBR as a form of knowledge organization system, adding a final dimension to classification. The articles in the present special issue offer a compendium of the promise of the FRBR model.

If a man could have half his wishes, he would double his troubles.

—Benjamin Franklin¹

In 1989, two papers by Patrick Wilson appeared that seemed to address a long-forgotten subject in cataloging.² This was the “second objective” of Charles Cutter, from the “Objects” in the introduction to his Rules for a Printed Dictionary Catalog, and which are repeated in every textbook about cataloging.³ The first objective can be paraphrased as the ability to find everything by a given search term. But the second objective, and this was Wilson’s point, was the ability to observe a collocated set of the editions, translations, adaptations, and so on of a given work, and also to navigate among those various publications that all had as their common reference, a work. This, Wilson said, was really the primary objective of the catalog, and therefore should not have been rendered secondary to the nineteenth-century notion of an efficient inventory of books. Wilson’s papers appeared in The Library Quarterly and in the proceedings of a conference on the future of bibliographic control.

Interestingly enough, these two papers by Wilson appeared at an historical zenith in which empirical research, much of it inspired by Wilson’s earlier book Two Kinds of Power,⁴ had begun at last to create external validity for cataloging research,⁵ and to demonstrate the empirical reality that library catalogs as inventories of books were less than efficacious, because encyclopedic catalogs of works were what really was required for scholarship. Research by Tillett, Smiraglia, Yee, Vellucci, and Carlyle⁶ as well as work reported by O’Neill and Vizine-Goetz⁷ had demonstrated the advance of empirical knowledge about the role of the work in the library catalog. And major books by Svenonius and Smiraglia gave central attention to the importance of the work as an entity in its own right.⁸

These ideas were infectious. As the time came to revise or rewrite, yet again, the Anglo-American Cataloguing Rules, Second Edition (AACR2) it seemed appropriate for the authors of the code to take time to consider how the bibliographic apparatus might better be reconceived.⁹ This conceptual analysis took place at the international level, in the International Federation of Library Associations and Institutions (IFLA) Study Group on Functional Requirements for Bibliographic Records. The eventual answer was a document itself called Functional Requirements for Bibliographic Records, in which a conceptual structure was developed that incorporated bibliographic concepts into an entity-relationship model.¹⁰ Eventually christened by its acronym, FRBR, this model has dominated thought about the future of bibliographic control, in general, and the structure of future library catalogs, in particular, for the past 14 years. The model has been tested, examined, and applied in abundance. The present special issue is testimony to its influence.

There has been a lot of conversation about FRBR, and there is quite a lot of literature, much of it cited in this special issue. When I began this study in July of 2011 there were 266 citations in various search portals. To discover what authors thought were the lacunae in the FRBR model, I analyzed those 266 papers. In the end, I discovered only 21 papers in which empirically based lacunae in the model were mentioned. In this article I will review these 21 empirical approaches to problems, or lacunae, with the model, plus comments from one author in the present issue. Then I will offer some observations based on the differences between the FRBR family of models, and the empirical evidence that research has presented. The differences are subtle and not great, but they are important for the FRBR community to comprehend.

A catalog inventory of books must give way to an encyclopedic catalog of works. In this there is no dissent. But as various proverbial sources attest, we must be wary of getting that for which we have wished, lest we double our troubles. The question here becomes whether FRBR is the correct, or adequate, or real, model, of a bibliographic construct centered on “the work.” A second question is whether a bibliographic construct centered on “the work” is really the best direction for the future of the catalog. Science informs this discussion, as we shall see.

REPORTED LACUNAE IN THE FRBR MODEL

The Expression Entity

As we noted briefly above, there is a relatively small cluster of papers that report or analyze problems perceived in the basic FRBR model. Most of these problems arise from the definition of the entity work, or from expression, which is the entity that represents a work’s tangible intellectual form. In 2001 Patrick Le Boeuf, arguably the most prolific chronicler of FRBR’s evolution, wrote about the perhaps excessively abstract definition of a work, saying “in my opinion, this entity hovers at such an abstract level that no standard numeric identifier in the world could ever grasp it. Works are just thoughts that have not yet been materialized … .”¹¹ In that particular paper Le Boeuf sought to place FRBR’s development chronologically alongside the emergence of XML and the spread of the entity-relationship data model into bibliographic consciousness. Later in the same paper Le Boeuf reports on a thorough analysis of FRBR by the Italian Gruppo di Studio di Catalogazione, from which perceived problems with the expression entity emerge. Specifically, Gruppo di Studio suggested there actually were four medium-dependent levels conflated under the heading of expression, which ought explicitly to be arrayed hierarchically. As Le Boeuf explains the problem arising from musical works, the essence of the expression is changed according to what could be described as distance from the original work (my terminology), such that two performances in the same medium with variant notation (as might be the result of different improvisations), or transposition, or absolute change of medium (arrangement for a different medium, for example) are in essence different kinds of expressions of the work.¹² Interestingly, the Gruppo did not name the four levels, nor was their analysis supported by empirical research.

O’Neill, using again the example of Tobias Smollett’s work, The Expedition of Humphrey Clinker, for which cataloging of many editions was present in OCLC’s WorldCat, generated an early empirical analysis of the application of FRBR’s entity-relationship model. This was one of the first “FRBRization” experiments. O’Neill found the model useful for sorting editions and translations at a meta-level, but reported much ambiguity about the expression entity. He wrote that “identifying expressions was problematic and raised the question of whether they are valid entities.”¹³ This was echoed by Taniguchi, who reported that “the expression entity is not prioritized,” and Jonsson, who, agreeing with the Italian Gruppo, found the expression entity too ambiguous in definition to be identifiable.¹⁴ Chen and Chen tested using the FRBR framework for generating metadata to describe artworks; they reported that the expression entity was more difficult to apply to artworks than to bibliographic works.¹⁵ Eadie, working with images, and Albertsen and Van Nuys, working with digital documents agreed, reporting that automatic identification of expressions was impossible.¹⁶

More studies appeared throughout 2005, these by Ayres, Bowen, Jones, Jonsson, and Kilner—all reported difficulties modeling expressions.¹⁷ But, Le Boeuf, also in 2005, notably opined that the expression entity had become notoriously problematic, but that he thought that the real problem was not the model but cataloging practice; in other words, the problem was not the definition of expression so much as the lack of logic in approaching resource description.¹⁸ Subsequent research reported the same problem—Fee in 2008, and Chaudri in 2009, both told us that it was at the expression level that the model failed, because the partition of work and expression was artificial.¹⁹ Finally, in the present issue, Petrucciani writes that “the step-by-step progress from work to expression to manifestation … may possibly reflect the case of works created before the Gutenberg revolution. But it is less adequate to the more complex and interrelated phenomena of contemporary cultural production and publishing.”²⁰

Gaps in FRBR

Still other authors reported gaps they found in the FRBR model, including Bennett, Lavoie, and O’Neill who pointed out the model needed to expand to include aggregates (such as anthologies, or works that include other works), and Chen and Chen (working with ancient Chinese artifacts) who hoped to see the model incorporate all of the components of Functional Requirements for Authority Data (FRAD) and Functional Requirements for Subject Authority Data (FRSAD)—“person, time, space, concept, and event, as well as their relationships”—as a means of capturing inheritance.²¹ Albertsen and Van Nuys agreed the model required extensions to incorporate mechanisms for modeling aggregates, implying simple interpretation of manifestation and item entities could provide mechanisms for incorporating dynamic Internet documents.²² In fact, dynamicity is needed in metadata models, and is missed in FRBR’s interpretation of bibliographic models. This is echoed by Yin and Salaba, and by Lee and Jacob, who remind us that FRBR’s strict hierarchical structure prescribes rigidly explicit relationships, which do not necessarily mirror the reality of information objects.²³

As it happens, FRBR does not properly model even traditional bibliographic realities. Several authors—Jonsson, Kilner and Allgood—all report problems interpreting the normative bibliographic concepts of edition and state in the FRBR model, because these conditions of altered representation correspond neither to expression or manifestation.²⁴ Finally, it becomes clear that there is much confusion about the interpretation of the levels of Group 1 FRBR entities. And this is good, because it brings together the two sets of lacunae. This merges the interpretive problems with the expression entity with the expansion problems of the manifestation entity. Adams, Santamauro, and Blythe bring this to the fore by suggesting the reduction of the Group 1 entities to three levels by creating what they call a “superworkspression” for serials, which are works, and aggregates, and continuing resources.²⁵ Allinson agrees, with a very telling narration of the lacunae in the FRBR basic model:

In natural language, … A ScholarlyWork is expressed as zero or more Expressions. Each Expression is manifested as zero or more Manifestations. Each Manifestation is made available as zero or more Copies. Each ScholarlyWork has zero or more creators, funders, and supervisors. Each Expression has zero or more editors. Each Manifestation has zero or more publishers.²⁶

The upshot is, as Allinson and Petrucciani point out, FRBR demands a work have an expression, an expression have ...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• Citation information
• Foreword
• 1. Be Careful What You Wish For: FRBR, Some Lacunae, A Review
• Implementations
• Extensions
• FRBR and Cataloging Rules
• Research using FRBR
• FRBR and the Semantic Web
• Index