CHAPTER FOUR
This chapter, which deals with some fundamental concepts in the bibliographic universe, is prerequisite to the understanding of some of the major cataloguing principles to be discussed in chapters 5 and 6. An understanding, in abstract terms, of the components that make up the bibliographic universe and the relationships that exist between them will help to grasp the nature of bibliographic entities and their place in the hierarchy of the bibliographic universe.
The major aim of the present chapter is to identify what constitutes the bibliographic universe and also the value and role of attributes of entities viewed as data elements in bibliographic records. Another aim is to see in what ways the online environment may influence the treatment of bibliographic entities and their attributes as well as the relationships that exist between entities, between entities and attributes, and among attributes in bibliographic databases. Since this chapter deals with the conceptual analysis of the bibliographic universe, it aims also to see to what extent cataloguing concepts are inherent in the bibliographic universe and whether they are independent of the environment in which bibliographic entities are described, i.e., from the technology of catalogue construction.
A conceptual approach will be used to explore the bibliographic universe. The conceptual models employed in this study, as will be discussed in section 0.3, represent those elements of real world information (here, bibliographic entities, their attributes and the types of association that exist among entities). The advantage of such conceptual models is that they describe the inherent properties of data, which are independent of the hardware or software used; i.e., conceptual modelling is not affected by the constraints of a particular database system. Regarding conceptualisation as a way of interpretation, the conceptual approach in this chapter also provides a framework for the discussions in Chapter 5 concerning the functional analysis of the bibliographic record: which functions are carried out by which data elements and how data elements represent bibliographic entities and their relationships.
It is appropriate to define briefly a few basic concepts which are important to the introductory section of this chapter; these definitions are treated in detail in their corresponding sections.
Bibliographic entity. A general term for a work or any manifestation or part of a work which contains the intellectual or artistic creation and which is an object of interest to catalogue users and an object of description in bibliographic databases. In this thesis the term "entity" has been repeatedly used to express works at any hierarchical level. Entity also encompasses persons, corporate bodies and subjects.
Bibliographic universe. The totality of bibliographic entities and their relationships. In a sense, the bibliographic universe consists of all types of intellectual or physical objects in any format which contain works of imagination as well as information.
Attribute. A property or characteristic that is common to some or all of the instances of an entity. It is piece of information associated with an entity or a relationship. For example, 'title', 'language' and 'readership level' are some of the attributes of the entity 'work'. Entities are described and identified through their attributes.
Bibliographic relationship. An association between or among bibliographic entities that have common characteristics, such as the relationship between a person and a work or between a work and an item. The concept also applies to relationships between entities and their attributes and between attributes.
Over the past three decades of library automation, librarians have had the chance to become familiar with the basic elements of database management systems and their logic. Most database management systems are conceptually based on the entity-relationship model. The E-R model allows us to capture and preserve some of the important aspects of the semantics of the real world. Like most models in database management systems, this model is based on the concepts of 'entity', 'relationships' and 'attributes'.
Applicable also to bibliographic databases, the E-R approach is an attractive means for conceptual modelling within the bibliographic universe and has been applied by a number of librarians. No other conceptual models have been used to illustrate bibliographic entities. Tillett (1989a) applied the E-R model in her study of bibliographic structures and relationships. The IFLA Study Group on Functional Requirements of Bibliographic Records has used the same approach to delineate the different functions performed by the bibliographic record (drafts 1993, 1994, 1995, 1996). Emphasising the primacy of the object, or entity, over its relationships, Heaney (1995) uses this approach (but under the term 'object-oriented modelling') for studying the possible re-structuring of cataloguing rules and standards, namely AACR2 and MARC, in relation to computerised catalogues. In this chapter the E-R approach will serve as the foundation for later discussions about the identification of entities, their attributes and the relationships that exist among them.
The question of whether or not to generalise a given model often confronts anyone trying to use E-R models as an analysis tool. This is partly due to the fact that the bibliographic universe itself is a complex environment encompassing various entities at different aggregate levels. Furthermore, each category of library materials (i.e., bibliographic items) is different in terms of its nature and attributes. For example, monographs differ from serials in that certain types of relationship (e.g., equivalence, derivative, and whole-part) are more common to them than to serials (see Tillett, 1992c: 169). Since the nature of each material type is different from the other, the types of attributes and relationships in each category of library material are different and, therefore, demand different treatments.
Another problem is the wide variation in definitions and inconsistent application of terminology which will be discussed in the following section.
A clear identification of entities is a first step in the conceptual analysis of the bibliographic universe. If we are going to identify various attributes of, and relationships between, bibliographic entities at different stages and create bibliographic records in a way that could be functional in a multitude of environments -- i.e., from the creator to the publisher to the printer to the distributor to the library and in a local, national, or global catalogue -- the first step is to define different constituents of the bibliographic universe such as 'works', 'expressions', 'manifestations', 'items', 'persons', 'corporate bodies', etc. It will then be possible to design a conceptual framework for bibliographic entities to be described in any environment.
There is no definite terminology for entities at different bibliographic levels. Furthermore, the bibliographic vocabulary used by each community (e.g., library, book trade and A&I services) is somewhat different from that of other communities. In the library context terms have been, and are being, used interchangeably. Various definitions have been provided by different people, for example, by O'Neill and Vizine-Goetz (1989), Svenonius (1992), and by the IFLA Study Group on Functional Requirements of Bibliographic Records (drafts 1993, 1994, 1995 and 1996). Even in the context of book-like materials there is a dearth of consensus on the definitions of various entity types, i.e., different manifestations of works.
There are at least four justifications for arriving at a consensus on the definition of bibliographic entities:
1) Conceptual modelling of bibliographic entities would become easier and more understandable,
2) The increasing possibility that the same work may be produced in different manifestations and formats,
3) The proliferation of shared cataloguing systems and union databases that create and exchange bibliographic records, and
4) The trend toward cooperative creation and use of bibliographic records by publishers, library suppliers and libraries which demands definitions that are simple and understandable to all.
In this study the categories of bibliographic entities include 'work', 'expression', 'manifestation', 'item', 'copy', 'reproduction', 'person', and 'corporate body'. While referring to different definitions in the literature, this study uses the IFLA's (Draft 1996) definitions to set the framework for later discussions about attributes and bibliographic relationships. This would make the context of this research more consistent with IFLA's ongoing study on the Functional Requirements of Bibliographic Records.
Work. There is no universally accepted definition of 'work'. The term 'work' has frequently been used inconsistently in the literature of descriptive cataloguing and often without a clear definition. According to Yee (1994a: 9), a definition of work has never been included in the glossary of any Anglo-American code.
Ranganathan (1955: 26) defines 'work' as 'expressed thought'. Lubetzky (1979: 7) considers the work as an abstract entity which is the object of readers' interest but which is contained in the book. Lubetzky's definition is a book-oriented definition, of course. Tillett (1987: 22) defines a 'work' as the abstract intellectual content embodied in an item. O'Neill and Vizine-Goetz (1989: 167) clearly differentiate between a 'work' and a 'book': the former is the intellectual creation of an author and the latter is a material object or medium used to convey the intellectual work. They point out that: "Because the material book embodies and represents the intellectual work, the two have come to be confused, and the terms are synonymously used not only by the layman but also by the cataloger himself." Pointing out that there is no straightforward definition that applies to all sorts of work, Wilson (1989: 10) provides a definition which is very different from Lubetzky's: definition. To him, 'work' is an abstract entity in the form of a text consisting a string of ordered array of symbols. He claims that the text of a work is the work. His definition applies to any type of product, even the electronic documents which may be open-ended texts. The IFLA Study Group on Functional Requirements of Bibliographic Records (Draft 1996: 11) defines work as an abstract entity. In general, it is the abstract 'work' that is the basis for later expressions and manifestations.
Expression. Conceptually, 'expression' indicates the next level in the bibliographic hierarchy and provides a linkage between a 'work' and any 'manifestation' based on that work. It is a specific presentation of the abstract work in the form of text (i.e., sequence of symbols, words, sentences) and/or illustrations, etc. IFLA (Draft 1996: 12) defines expression as "the specific intellectual or artistic form that a work takes each time it is 'realized.'" Expressions include entities such as editions, adaptations, changes of genre, dramatisations, novelisations, and translations.
In the context of 'expressions', the term 'edition' should be defined clearly because it is a term very much used in dealing with the bibliographic universe. Defining what constitutes an 'edition' is one of the most important requirements for any conceptual model of the bibliographic universe and also in cataloguing practice (an issue which will be discussed in detail in Chapter 7); there is no general consensus on such a definition. Terms like 'edition', 'reprint', 'reissue', 'issue' and so on have always been problematic to cataloguers, especially in shared cataloguing systems. In practice, deciding on whether or not to create a new bibliographic record for the item being catalogued depends on the delineation of 'edition'. A new edition can be defined as any significant change in the intellectual or artistic content or form and/or layout of a work which makes it different from another edition of that work. Modifications, revisions and translations are therefore new editions of the same work.
Manifestation. Conceptually, 'manifestation' shows the transition from the abstract entity to the physical entity: any expression can potentially be embodied in different physical formats. Thus manifestation is a general term representing any kind of physical embodiment of an expression of a work. Yee (1994c: 227; 1994d: 356) defines 'manifestation' as a version or edition of a work with significant difference from another version of the same work. However, her definition does not identify physical form as a characteristic of manifestation. IFLA (Draft 1996: 14) defines 'manifestation' as the physical embodiment of the 'expression' of a work.
It should be noted here that the term 'version' has been used in the literature interchangeably for both the intellectual and physical forms in which an edition is produced and represented for use. For example, the same expression of a work can be represented in different physical formats, such as print and electronic formats (see figure 4.2), or the same work may be produced in several literary forms, e.g., a version of the 'Arabian Nights' for young readers. The problem of 'multiple versions' refers to the description of different physical formats of (the same) work in a catalogue. In this study, the term manifestation encompasses the term 'version'.
Item. The term 'item' is frequently used in the literature and in practice and has a more concrete meaning than the term 'work.' While 'work' is an abstract entity, 'item' represents the physical object. IFLA (Draft 1996: 15) defines 'item' as a single exemplar of a manifestation.
In the context of monographs, the terms 'book' and 'document' are often used interchangeably with the term 'item'. This is not a desirable approach in cataloguing because the term 'item' is broader than 'book' or 'document' and applies to different material types.
Copy. One physical unit identical to other units produced from a single type image, e.g., a single example of a book. Thus 'copy' is an instance of the item and has a content identical to other copies of the same item. However, a copy may be specific in that it contains new characteristics not common to other copies of the same edition. For example, a copy may include handwritten notes by author, or annotations added after production of the copy, or may have some parts missing, or may be in a special binding.
Reproduction. An imitation of a copy of an item in another physical form without any change to the text and content. For example, a photocopy, a facsimile or a microform reproduction has exactly the same content as the original item.
Person. An individual related to a work either as responsible for the intellectual/artistic creation of the work or as the subject of the work. IFLA (Draft 1996: 16) provides a similar definition: person is an individual. Persons are important entities in the bibliographic universe and in the conceptual models illustrating that universe.
Corporate body. A group of persons or organisations with corporate responsibility, action, etc. From a bibliographic perspective, corporate bodies act like persons; they are either responsible for a work or are the subject of a work. They may be of permanent or contemporary nature. AACR2R defines corporate body as an organisation or a group of persons that is identified by a particular name and that acts, or may act, as an entity. IFLA (Draft 1996: 16) defines 'corporate body' as an organization or group of individuals and/or organizations acting as a unit." Corporate bodies are important components of the bibliographic universe.
In addition to these levels, the following definition should be considered when dealing with the identification of bibliographic entities:
Superwork. Superwork is not an entity by itself. It is a concept which may be applied to the totality of a work, i.e., its different expressions and manifestations and other related works. Svenonius (1992: 6) defines superwork as: "The set of all manifestations of an original work and all manifestations derived from it." Superworks also encompass new works, such as reviews, criticisms, indexes and bibliographies, which are based on the same work. In this context, many bibliographies (for example, of writers, philosophers, etc.) have a superwork approach in bringing together all expressions and manifestations of a work and also other related works.
In general, what libraries actually collect and describe through records are entities at a lower level in the hierarchy of the bibliographic universe; that is copies of items that embody expressions of works. The catalogue or the bibliographic database should clearly demonstrate the hierarchy and the place of each entity in that hierarchy.
The following conceptual models (figures 4.1 to 4.4) represent a partial picture for some of the entity types in the bibliographic universe. In these models, square boxes represent entities, circles represent attributes and lines represent relationships. Figure 4.1 illustrates the concept of superwork as a set of related works and any expression, manifestation, and item derived from a work. IFLA (Draft 1993) represents a similar model for superworks. Figures 4.2, 4.3 and 4.4 show the place of each entity in the hierarchy of the bibliographic universe, i.e., in its relationship to higher-level and lower-level entities. In general, conceptual models illustrate four levels of 'work', 'expression', 'manifestation', and 'item' in the bibliographic hierarchy. But, the present models are extended to two more levels ('Copy and 'Reproduction') to show all possible entity types in the hierarchy. Figure 4.2 illustrates different manifestations derived from one particular expression of a work. Figure 4.3 represents a different approach, in that it shows different manifestations derived from different expressions of a work. Figure 4.4 illustrates the bi-directional relationships that exist between entities in the bibliographic hierarchy.
Figure 4.1 Bibliographic entities in a superwork context
(adapted from IFLA Draft Report, 1993)
Figure 4.2 An extended model for bibliographic entities
(different manifestations derived from one particular expression)
Figure 4.3 An extended model for bibliographic entities
(different manifestations derived from different expressions)
Figure 4.4 An extended model of entities with derivative relationship
Entities and relationships are qualified by attributes representing their descriptive properties. Each entity has a set of attributes. Attributes can be viewed as data elements associated with bibliographic entities to describe and identify them clearly during the processes of creation, publication, production and cataloguing. For example, the title of a work is an attribute that makes the work known to readers. To identify an entity uniquely, a combination of attributes such as the author's name, the title, the edition and the date of publication can be used. Even this combination might need further clarification, such as format (e.g., is it the paper or microfiche version?). A major concern of cataloguing codes has been, and still is, how to identify and record the least number of attributes that are necessary for adequate identification of, and access to, bibliographic entities in a catalogue. By 'adequate identification' is meant sufficient information, such as content, readership level, document type, publication information and imprint, that help users to decide on one particular item over another and confirm that the item described is the one for which they are looking in the catalogue. As will be discussed in the next chapter, bibliographic entities are identified, related, located and accessed through their attributes. Thus, each attribute has a specific function or set of functions that are essential to the catalogue.
Figure 4.5 shows possible attributes of bibliographic entities at different levels: 'work', 'expression', 'manifestation', and 'item'. As can be seen, new attributes are added to entities at the lower level of the hierarchy (i.e., at the manifestation and item levels). Attributes such as 'name of the publisher', 'place of publication', 'date of publication', 'series title', 'series number', 'standard numbers', 'physical format', and 'price' normally belong to manifestations. Nevertheless, attributes, such as 'title of the work', 'name of the author(s), 'content', 'genre', and 'language' which belong to entities at the work level are usually more useful for maintaining bibliographic relationships and for accessing entities.
Figure 4.5 Possible attributes of entities at different levels (based on IFLA Report, 1995)
Attributes can also be identified in terms of cardinalities (i.e., of primary importance) and optional/mandatory factors. In Matrix 4.1 optional attributes, which are not as essential to the major functions of the record as mandatory attributes, have been marked with an asterisk ('*').
It should be noted here that many of the following attributes have been used, at one time or another, as entry points in some manual catalogues. The concept of the machine-readable record, however, has made possible the inclusion of new attributes and/or new ways to make old attributes more easily retrievable. These attributes can be useful in an electronic environment and help to identify, retrieve, organise and display records according to users' needs. The language of the item, readership level (i.e., target audience), genre/form, document type or category of material, physical format, geographic area code, standard numbers and record number are attributes which, due to the constraints of the manual catalogue, were either not included in bibliographic records or were not usually search/retrieval elements.
Matrix 4.1 Common attributes for entities in the bibliographic universe
Attributes Work Expre-ss Manifes- Item Person Corporate
ion tation body
Name of person x
Dates associated with x
person x
Title of person*
Name of corporate body x
Other information
associated with
corporate body (e.g., x
date and place of
meeting) x
Title of work
Title of expression x
Other title associated x
with manifestation
Edition information x
Place(s) of x x
publication, etc.
Publisher(s)/distributor x x
(s)
Date(s) of publication x x
Date of work x
Date of expression x
Extent of work* x
Extent of expression* x
No. of pages/physical x
parts
No. and type of x x
illustrations*
Dimensions/size x
Terms of x
availability/Price*
Series statement x x
Content information
Subject description x
Summary* x
Language of work x
Language of x
expression
Accompanying x x x
material
Readership x x
level/audience
Content x x
format/genre
Codes (e.g., call no.,
accession no.,
bar code) x
ISBN/ISSN x
Ownership* x
Signs/signatures* x
Access restriction* x
*Optional.
Entities do not usually exist in isolation but are associated with one another through different types of relationships. In its simplest definition, bibliographic relationships are the associations between two or more entities in the hierarchy of the bibliographic universe. Any study of the bibliographic record must take into account relationships that exist between entities in the bibliographic universe and in the catalogue that is a partial representative of that universe. The end results of these relationships and the object of bibliographic records is access to adequate and precise information, to entities that satisfy information needs. As will be discussed in Chapter 6, section 1.1, the objectives and functions of the catalogue can be completely met through bibliographic relationships and linking devices. The study of bibliographic relationships is central to an understanding of the nature and structure of the bibliographic record, the catalogue, and ultimately, to the study of cataloguing principles.
In a broad sense, bibliographic relationships can be categorised into two types: those that hold between the entity and its attributes and among attributes (i.e., internal associations); and those that hold between or among entities (i.e., external relationships). Cataloguing principles and rules are concerned with both types of relationships.
In the first type of relationship, when a work is created by the creator (e.g., the author), relationships are established between the creator and the work. Associations in the data, such as 'an author has written a book', are called relationships. In this case, the relationship between the entity 'author' and the entity 'book' is 'has written'. This type of bibliographic relationship forms the structure of the record and is an integral part of the bibliographic record. No attributes can be included in a record without its having associations to the entity itself. This is a simple yet very important type of relationship in the bibliographic universe and one to which cataloguing codes devote considerable attention.
Some types of internal connections are themselves presented in the record in the way that data elements are displayed in relation to each other; for example, the relationship between the author and the title is directly maintained by displaying the author's name near to the title, either as the 'main entry heading' on top of the title or by repeating it with relating words, such as 'by', 'written by', 'editor:', etc., as the 'statement of responsibility' following the title. Similarly, on the title page or the chief source of information, the name of the author usually appears near to the title. Even if it does not, the cataloguer seeks to establish this relationship on the record.
More attributes are added when the abstract 'work' is embodied in an 'item' (see Matrix 4.1 and Figure 4.5). At this stage, attributes, such as the place(s) of publication, the name of the publisher(s), the date of publication, the extent and size of the item and the International Standard Book Number (ISBN) are introduced. The cataloguer's task is to identify various types of relationships and to create bibliographic records based on a certain structure for the display of the relationships between the entity and its attributes and among attributes, that is, data elements.
Another important type of relationship, which is established at the early stage but frequently not indicated clearly by the creator, is the relationship between entities. This is the type of relationship which the cataloguer attempts to distinguish and describe according to the rules provided for in cataloguing codes. The relationship between a newly created work and works by other authors that have been used as sources for the creation of the work is usually not defined except by indicating a general note such as: 'Includes bibliographical references'.
External relationships are considered very important and are a very complex element in the bibliographic universe (and hence in the catalogue). In order to identify this type of bibliographic relationship, different categorisations have been offered by different writers, for example, by Goossens and Mazur-Rzesos (1982), Hagler (1991) and Tillett (1987, 1992c). Goossens and Mazur-Rzesos (1982: 14) identify three general categories of bibliographic relationships: 1) the hierarchical relationships: the linking of the whole to its parts and of the parts to a whole, 2) the chronological relationships: the linking in time between the succeeding issues of an item, and 3) the horizontal relationships: the linking of versions of an item. Hagler (1991: 47) identifies five types of relationships between entities: 1) editions of the same work, 2) sequels, continuations, etc., 3) items both physically and bibliographically separate from one another but issued and intended to be used together, 4) important separately identifiable works contained within one publication and 5) items in the same series. Tillett (1987, 1992c) has developed one of the most comprehensive taxonomies in this area. She categorises bibliographic relationships into seven types: Equivalence, Derivative, Descriptive, Whole-part, Accompanying, Sequential, and Shared relationships. One more type of bibliographic relationship was added to Tillett's taxonomy by the IFLA Study Group On Functional Requirements of Bibliographic Records: 'work relationships', that hold between abstract works and their physical manifestations (IFLA Study Group on Functional Requirements of Bibliographic Records, Draft, August, 1993: 8, 10; Draft July 31, 1995). In the 1996 draft of the Functional Requirements of Bibliographic Records, the IFLA Study Group has modified Tillett's categories to 1) Work-to-work relationships, 2) Expression-to-expression relationships, 3) Expression-to-work relationships, 4) Manifestation-to-manifestation relationships, 5) Manifestation-to-item relationships, and 6) Item-to-item relationships. This study uses the IFLA's new categorisation as the framework for discussions about relationships.
??Table 4.1 Types of bibliographic relationships among bibliographic entities*
Relationships Definition and types of entities included
1. Work-to-work A basic premise of the work-towork
relationship relationship is that two different works
have been recognized to exist; that is, the
intellectual or artistic cintent of one work
has been judged sufficiently different from
the other to constitute a separate work.
Among the work-to-work relationships there
are two categories: referential works and
autonomous works. Three types of
relationships cut accross these two
categories: successor, supplement, and
complement.
2.
Expression-to-Expression
Relationships
3. Expression-toWork
Relationships
4.
Manifestation-to-Manifestat
ion Relationships
5. Manifestation-to-Item
Relationships
6. Item-to-Item
Relationships
*From the IFLA Study Group on Functional Requirements for Bibliographic Records, (Draft May 1996)
Bibliographic relationships, as well as being complex, may also be either uni-directional or bi-directional. This is an inherent characteristic of bibliographic entities and has implications in any form of catalogue. Although at the time of describing an entity it is difficult to know what publications or documents could be later published or added which may have some type of relationship with the entity being described, one should consider such a concept and allow for it in the model for the bibliographic relationships and also in the structure of the catalogue when applying that model.
From a general perspective, linking devices refer to devices to connect bibliographic entities. In other words, bibliographic relationships are maintained through some attributes working as linking devices. Tillett (1992b: 23) defines linking devices as "...those specific devices within the catalog that connect or link bibliographic records for related items."
A review of modern cataloguing codes shows that the devices for establishing bibliographic relationships have varied over the years and have been influenced by the form of the catalogue, that is, the technology to construct library catalogues (Tillett, 1987: 26, 196; 1989a: 151; 1992b: 23; 1992c: 163). The emergence of the concept of main entries, added entries, uniform titles and cross references, together with the abandonment of 'dash entries' are common examples of this dependency of linking devices on type of catalogue. In Matrix 4.3 various attributes working as linking devices are matched with the types of bibliographic relationship they support.
Matrix 4.3 Linking Devices and the types of relationships they support*
Equiv.R Deriv. Refer. Whole-par Augmen. Sequent. Work
Relationships el. Rel. Rel. t Rel. Rel. Rel. Rel.
Linking Devices
Main entry headings x x x x
Uniform titles x x x x x
Filing titles x x
Edition statements x
Physical description
accomp. matter x
statement
Notes x x x x x x x
Contents notes x
Multilevel x x x
description
Analytical entries x
Holdings statements x
for copies
Dash entries x x x x
Added entries x x
Name-title & title x
added entries
Series added entries x
Series statements x
Series content lists x
Subject entries x x
References x
*Based on Tillett (1987)
Bibliographic relationships are inherent in the bibliographic universe, and consequently in catalogues, and are not dependent on the medium by which and the environment in which records are created, manipulated and made accessible for searching. The new technology has made it possible for bibliographic relationships to be maintained and displayed more effectively and more flexibly in the online catalogue than was possible in the card catalogue. There has been a rebirth of interest in bibliographic relationships over the past decade and this may be as a result of the impact of the online environment on the structure of library catalogues, bibliographic files and indexes within the catalogue. To Tillett (1991a: 151), the demand for a new look at this concept has come from two factors: the development of online catalogues and the implementation of the second edition of AACR. There are at least two other factors: an increase in the types of media being catalogued and the growth of shared cataloguing systems and union databases.
When we move to a shared online environment such as shared cataloguing systems, union catalogues and national or international networks the need for expressing bibliographic relationships and the types of associations has become more important, since it is most likely that, with an increase in the number of participant libraries and library collections, the number of manifestations and editions of a work will increase. In such an environment it is important for the user to identify the relationship of entities to one another and to distinguish the place of entities in the bibliographic hierarchy. For example, a problem for the user appears when relating holdings to works or items which may be catalogued as collections or as separate parts (e.g., identical item catalogued as part of a series to which it belongs, or as a monograph in its own right).
In regard to the advent of computer catalogues, however, there has been little evolution in the concept of linking devices; many of the traditional devices have been embodied in online catalogues (Tillett, 1987: 196; 1992b: 23). As mentioned earlier, the provisions in current cataloguing codes for linking bibliographic entities are not adequate in terms of displaying all types of relationships explicitly and in a uniform approach. According to Attig (1989: 142):
The traditional technique used to link records is to rely on the data in the records to establish the relationships. In virtually all access points, the content of the heading field may be used to establish the relationship of the record to other records with identical heading. ... In current systems, data from linking fields are rarely used to establish a direct link between records. Instead, the searcher is expected to use the information about related items as the basis for further searching. However, the format does support more direct linkage, and the possibilities deserve exploration.
In an online environment each type of relationship (i.e., bibliographic, name, subject and access points) can be established in accordance with the structure, content and search retrieval capabilities of the computerised catalogue. Online catalogues have the potential to provide displays of related bibliographic records, clustered according to the type of bibliographic relationships. What has been retrieved can direct the user in further searching.
In terms of maintaining bibliographic relationships as well as the types and forms of linking devices, the question arises as to whether the rules in current cataloguing codes are relevant to an online environment: do they give instructions on how to provide the most appropriate devices for linking related records? Since the technology of catalogue construction determines the types of linking devices (either abandoning some, e.g., the dash entry, or introducing some, e.g., standard numbers), should cataloguing codes provide separate rules for bibliographic relationships and linking devices to take best advantage of the online environment?
Another question that seems relevant here is: which types of linking device to secure all types of bibliographic relationships are appropriate for an online environment? For example, the variety of linking devices for expressing the whole-part relationships may be a critical problem for the online catalogue. Is it possible and relevant to reduce this diversity in the treatment of relationships in an online environment? Is it possible to make one record for both the whole and its parts, or to make multiple records, i.e., each part to have a separate record citing the whole? This is an important question in a network environment since library policies may differ from each other in describing the parts of a whole. These are some of the considerations concerning the principles and rules for bibliographic relationships in an online environment.
The general approach for bibliographic records constructed under current cataloguing codes and MARC formats is not capable of properly maintaining and displaying different types of relationships. There is no clear and uniform approach towards the treatment of different types of relationships. In many cases, as noted earlier, different devices are used to demonstrate the same type of relationship. For example, the equivalence relationship can be displayed by different linking devices, such as dash entries; notes to acknowledge equivalent copy; notes to acknowledge the original; shared uniform titles and holdings statements for copies (see Matrix 4.5). This divergence in the treatment of bibliographic relationships through different linking devices, as also highlighted in Matrix 4.4, is not suitable to online shared environments, where uniformity in access to related entities is crucial to the exchange and understanding of bibliographic records. Regarding the treatment of bibliographic relationships and linking devices in cataloguing codes, Tillett (1991b: 398) states that:
With regard to bibliographic relationships, history has shown no rationale and little consistency in how we relate bibliographic entities. A review of cataloging rules since 1841 reveals differing methods and devices used over the years to show bibliographic relationships, but also reveals a lack of any theoretical rationale for the [linking] devices.
This researcher believes that online catalogues and future computerised systems may provide new and more effective linking devices to help the searcher to navigate a catalogue. The logic of database systems makes it possible to establish a more flexible linkage between data elements than exists in manual systems. It is possible to link two or more fields or subfields in machine-readable records in such a way as to better identify bibliographic relationships. For example, the computerised catalogue can match the value of the field 'uniform title' with the value of the subfield 'Language' or 'Part' or 'Date' in all records so as to retrieve and cluster those records that have these data elements in common. It is also possible to provide different structures for the retrieval and display of different types of relationships. This depends, to a large extent, on both the structure of machine-readable records and the software to design strings of fields and/or subfields so that the computer can retrieve and display all related records in a user-oriented manner.
The current approach in establishing bibliographic relationships in the online catalogue is through the incorporation of linking tags; in most MARC formats blocks 76X-79X in the USMARC formats (particularly formats for books and serials) are for linking entries. These linking entry fields are designed to display a note in the record in which the linking entry field appears and to provide machine linkage between two related records (USMARC Bibliographic Format, 1990: 76X-79X, p. 2). The linking fields (block 4XX) that are incorporated in UNIMARC for a limited number of types of bibliographic relationships could be expanded for other types, using new means for this purpose. More coded linking devices are needed in MARC formats. For example, devices for providing links to other expressions of the same work or other manifestations of the same expressions and so on. Another example is to assign 'edition' as a coded link to provide access from an edition to any other edition of the same work. This is a pre-coordinated approach, as can be seen in examples of super records in the Prototype Catalogue developed in Chapter 7. Nevertheless, in the MARC format all such devices need to be incorporated in a consistent way.
In online catalogues, there is the possibility of assigning more data elements as linking devices to establish relationships. Standard numbers, such as LCCN, ISBN, ISSN, OCLC record number and other similar numbers are one of the most useful elements for linking related records to one another. However, as Attig (1989: 143) has noted, these numbers do not always uniquely identify a single bibliographic entity and are not good candidates for the display of bibliographic relationships.
With regard to online catalogues, the question arises as to whether it is possible to have a special block in the Notes Area to include relationship information in a way in which related records can be retrieved and displayed. Tillett (1989a: 161) proposes that the types of relationships be tagged and reflected through notes. This will avoid the need for a redundant tracing in a machine-readable record. To achieve this, as Tillett (1989a: 161) points out, "We would need to slightly modify some of the MARC tags and indicators for notes which incorporate links to another bibliographic record."
Using hypertext techniques for linking related records is another promising method in bibliographic databases. This technique can be developed through pre-coordination, in that the cataloguer consciously creates links between related records in user-oriented ways in order to maintain different kinds of relationships. A prototype of such a catalogue has been developed, in section 2 of Chapter 7, for maintaining relationships between related works/items. It is also possible that a hypertext technique could enable linking which is not pre-coordinated.
A thorough understanding of bibliographic entities, their attributes and relationships is essential to capture the concept of the bibliographic universe and to construct useful catalogues, as tools to navigate that universe. Descriptive cataloguing is in fact an attempt to describe and provide access to bibliographic entities in the most effective way. From a conceptual analysis of the bibliographic universe we can arrive at the following conclusions:
-- The bibliographic universe encompasses a variety of entities at different levels with different types of relationships to one another. Entities may evolve into new entities and maintain complex relationships with their original entities. For an exhaustive description of bibliographic entities in catalogues and databases, we need to understand the nature of each entity type and the expressions in which entities at different levels are to be manifested and/or presented in databases.
-- A first step in the conceptual modelling of the bibliographic universe is to redefine its elements and to reach a consensus about the different entity types in the hierarchy of the bibliographic universe. This would help to describe entities more consistently and with respect to their relationships with other entities. It would also help the catalogue user to identify the right entity at the right level. However, due to complexities in the nature of bibliographic entities, their attributes and the types of relationships that exist at different levels of bibliographic hierarchy, it is difficult to reach a consensus on definitions for all entity types. A possible solution seems to lie in an international consensus on a set of definitions for each entity type such as 'work', 'expression', 'manifestation', and 'item'. A uniform approach to defining each entity type in each category of material is essential to cataloguing principles and cataloguing codes, especially in an international context.
--Attributes of entities have a most important role in the bibliographic universe. Entities are represented in bibliographic records through their attributes and can be identified, related, located and accessed through the same attributes.
-- As bibliographic entities in the bibliographic hierarchy move from the highest level (i.e., the 'work'), to lower levels (i.e., 'item' and 'copy'), more attributes are added to them. Attributes which are closer to the 'work' level are more useful in accessing and collocating entities.
--With an understanding of the significance of bibliographic relationships in fulfilling the catalogue's objectives and functions a new attention to the types of bibliographic relationships and to the linking devices which set out those relationships can enhance access to bibliographic information. For the online environment there is also a need to define explicitly the types of relationship that an item has with other items, not only in the same catalogue or collection but also with works or items in the bibliographic universe itself.
-- The concept of bibliographic entities and their relationships in the bibliographic universe is independent of the technology of catalogue construction. However, the electronic environment has the capability of creating links, maintaining and displaying relationships in ways that are not available in manual environments. New ways and techniques do exist that can link two or more records to each another in the electronic environment. A possible technique for linking two or more records to each another and to display bibliographic relationships more clearly could be to re-structure and enhance some of the MARC tags, or to use the hypertext facility.
In conclusion, a thorough understanding of bibliographic entities and their attributes and relationships is essential to any study of catalogue records, which are surrogates for bibliographic entities and which are carriers of attributes and relationships. This understanding is also necessary to the study of cataloguing principles upon which bibliographic records are created. These concepts are prerequisite to the discussions in chapters 5 and 6.
Back to: