CATALOGING RULES

by Martha M. Yee

June 15, 2008 draft, rev.

Contents:

Preface

0.1 Introduction

0.2 Objectives

0.3 Principles

0.4 Record structure

0.5 Importance of display

0.6 Importance of hierarchically structured human-readable identifiers for entities

0.7 Influence of access on scope of catalog

0.8 Generality of rules

0.9 Types of data

0.10 Change of name

0.11 Mandatory vs. optional

0.12 Relationship to FRBR

0.13 Title-manifestation

0.14 Serial title

0.15 Degression: begin with description of the work

0.16 Acknowledgements

1. Work

2. Expression

3. Title-manifestation description

4. Serial title description

5. Manifestation description

6. Item description

7. Relationships to other works

8. Creator identifiers

9. Subject identifiers

10. Discipline or perspective

11. Genre/form identifiers

12. Encoding of data

13. Indexing of data

14. Display of data

15. Examples

16. Works that influenced these rules and further reading

0 PREFACE

0.1 Introduction

These are rules to design a catalog. A catalog presupposes human intellectual effort, not raw computer indexing of an undifferentiated mass of keywords. Human intellectual effort is expensive, so it should be used as efficiently as possible to identify the entities of most interest to the users of collections represented in catalogs.

The bibliographic universe is complex. Catalogs should be designed to reduce this complexity for its users. Thus these rules are intended to support the labelling of bibliographic data by humans such that computers can use the labeled data to build complex indexes and displays that will appear simple to the users of the catalog. Like RDA, these rules could be used as a simple list of raw data elements that could be implemented and displayed many different ways in many different contexts. However, unlike RDA, these rules do not push most of what was formerly considered to be cataloging out of the rules and into implementation decisions and display decisions that are independent of the data elements. In addition to defining raw data elements, these rules include recommended implementation decisions and recommended display decisions, so as to try to ensure the greatest possible functionality and user service in the resultant catalog(s).

The approach recommended in these rules (to start with the work rather than the item) has the potential to allow us to create a complexly organized and extendable description for each work that can be exposed to commercial search engines such as Google. A Google searcher interested in that work can start from the work and end up with a particular manifestation of a particular expression of that work that is available at the searcher's local library, on the shelves, via interlibrary loan, or via licensed electronic delivery.

0.2 Objectives

The rules are designed to meet the following objectives:

The catalog should be an efficient instrument for ascertaining (in this priority order):

1. Which expressions of a particular work, specified by its creator and title, or, if the work does not have a single creator, its title alone, the institution contains or can provide access to

2. which works by a particular creator the institution contains or can provide access to

3. which works on a particular subject the institution contains or can provide access to

4. which works in a particular genre or form the institution contains or can provide access to

5. whether the institution contains or can provide access to a specific expression or manifestation of a work

To be an efficient instrument, the catalog must be designed to provide the user with efficient indexes and efficient displays that achieve the above objectives. In order to create efficient indexes and efficient displays in computer-based catalogs, the data must be adequately encoded to support such indexes and displays.

0.3 Principles

In order to meet these objectives, the rules are based on several basic principles.

Principle 1: First of all, these rules assume that the entities of most interest to users are the work, the creator (whether a person, a corporate body, or a conference), the subject, the genre, and the form. In order to build complex displays and indexes, it is important that each of these entities be given a language-based identifier so that the user can scan through thousands of entities matched on a search and recognize the one he or she is looking for. The language-based identifier must quickly and concisely identify the entity it represents to as many users as possible, including both users familiar with and users unfamiliar with the entity being identified. Otherwise, the catalog will not be an efficient instrument. This entity identifier should be the name by which the entity is commonly known in the language and script of the cultural community of the catalog. This is a departure from Anglo-American practice which hitherto has preferred the name commonly known in the country of origin, resulting in Russian works and corporate bodies being given the name by which they are commonly known in Russia even in American libraries whose users will know those works and corporate bodies only by their English names. With the current plans to create national authority files that link across national boundaries using computer-based identifiers rather than language-based identifiers, it should be possible to let each cultural region choose its own language-based identifiers. Indeed, it might be possible to allow individual users to choose a particular language and script/transliteration scheme for entity identification in any catalog, regardless of the cultural community of the catalog. This first principle is the principle of the name commonly known.

Principle 2: If the name commonly known is shared by more than one entity, it should be made unique by the addition of whatever information is necessary to differentiate each from the other within the cultural community of the catalog. Otherwise, the catalog will not be an efficient instrument; it will confound under one identifier the expressions of more than one work, the works of more than one author, or works on more than one subject. This second principle is the principle of the uniform identifier.

Principle 3: Use the language of the item before you in order to describe it. Begin with what the item being cataloged says and correct it only when it is known to be ambiguous or erroneous. Remember that not all users will know about error, so correction must be done in such a way that the item remains recognizable to the users unaware of the error. The identifying information on the item being cataloged provides a communication link between the catalog user and the cataloger. The fundamental assumption is that a majority of people who compose citations will do so using the identification information given in the item being cited. The user who comes to the catalog with a citation will be more likely to match a catalog record if the catalog record also uses the identification information given in the item as is. Additionally, it will not be possible in all cases for a cataloger to be able to resolve ambiguity in statements on the item being cataloged. In such cases, it is better to simply quote the item than to make assertions based on incomplete information that may eventually turn out to be false or misleading when all of the facts have been determined. Now that Unicode is being more widely adopted, it is time for us to try to use the identification information given in the item being cataloged even when the script that appears on the item being cataloged is not the script of the catalog. This third principle is the transcription principle.

0.4 Record structure

The focus in these rules is the design of displays and indexes in online catalogs. The focus here is not on the data structure, including the object of a record. It is intended that these rules could be used to design a catalog that consists of manifestation-based records (as in our current shared-cataloging environment), expression-based records, work-based records, or even a shared single virtual catalog that does not contain records at all in the conventional sense. Because these rules are data structure neutral, they can seem vague or ambiguous at times, especially when it comes to demonstrating relationships. Traditionally, we have demonstrated some relationships formally using citations and we have demonstrated other relationships informally using notes. An example of a formal demonstration of relationships: the fact that one work is based on another work is demonstrated by making a citation to the earlier work on all expression/manifestation records for the later work. Examples of informal demonstration of relationships: When one expression is based on another particular expression of the same work, a note is made to record the relationship between these two expressions of the same work. When one item is reproduced to create a new manifestation of the same expression of the same work, a reproduction note is made to record the details about which item was reproduced. It may be that we are moving into a world in which more universal machine-actionable identifers will be attached at the item, manifestation, and expression levels so that even these informal methods will eventually be transformed into formal linking using machine-actionable identifiers.

0.5 Importance of display

The creation of hierarchically structured displays is desirable such that a user looking for a particular work can see just the works that match his or her search, choose one, then see all of its expressions (as well as its related works, and the works about it, in separate links or lists), choose one expression, then see all of its manifestations. The user looking for a particular expression of a work should be able to see that expression in the foreground, but with the context of all the other expressions of that work readily available in the background. The user looking for particular types of manifestation (e.g. any online version) should be able to see the online versions in the foreground, but with the context of all the other manifestations and expressions of the work readily available in the background.

0.6 Importance of hierarchically structured human-readable identifiers for entities

The creation of hierarchically structured human-readable identifiers for entities of interest is desirable such that a user who has done a search that matches on a thousand entities can be given a quickly scannable display of all of the entities matched using identifiers that maximize the user's ability to recognize the entity sought. In order to accomplish this, the identifiers must consist of the name by which each entity is commonly known (see Principle 1 above), and, in cases in which two or more entities share the same name, conflicts must be broken with further data elements that are likely to increase recognition value for users and that are predictable and readily comprehensible, such that users can find the correct place in a sorted list of a thousand entities as quickly as possible (see Principle 2 above). These data elements used to break conflicts can be labelled in the RDF model as sort 1, sort 2, sort 3, etc., so that sorted displays of multiple entities can be maximized for user recognition value, and so that sorted displays can always begin with the name commonly known to the user (Principle 1), and subarrange under that only when additional sort elements are necessary to break conflicts (Principle 2). This will be an immense improvement over existing OPAC sorting software which mixes up entity names indiscriminately with elements supplied by catalogers to break conflicts in such a way that a user's search cannot be successful unless they know ahead of time about elements supplied by the cataloger (for an example, search in any OPAC for journals with the title Health or for the subject heading power as used in the discipline of political science rather than in the discipline of electrical engineering).

0.7 Influence of access on scope of catalog

A catalog should describe materials that actually exist and that can be obtained for the catalog user. These may be materials that physically exist on the premises of a particular institution, such as a library, archive or museum, or these may be materials that can be accessed via the Internet by the users of the catalog, including online materials that are licensed for access only to the members of that institution. If some users of the catalog are not members of the institution and not allowed to use licensed materials, these restrictions should be made very clear in the catalog.

0.8 Generality of rules

Undoubtedly specialist communities will create separate manuals to provide more detail, but , for the sake of creating universally compatible records, it is hoped that the basic structure in these rules would be adhered to.

0.9 Types of data

These rules call for the collection of several different types of data.

0.9.1 Transcribed and composed data

The first type of data is transcribed data, copied from the item being cataloged. (See Principle 3 above). In RDF terms, the datatype for this is literal. The second type of data is composed data. This is description composed by the cataloger. In RDF terms, the datatype for this is also literal. Both transcribed and composed data should be displayable and keyword indexable. Neither transcribed nor composed data is normalized. That is to say that care is not taken to ensure that we always use exactly the same string of characters to refer to the same entity. When we transcribe a publisher's name, we transcribe what we see (e.g., G.P. Putnam), even if we know that we have other books in our collection in which the publisher's name appears in other forms (e.g. Putnam).

0.9.2 Controlled (normalized) terms and composed data

There is, however, a type of data which is a combination of composed and normalized data. This type of data is selected from lists of controlled (normalized) terms such that normalized data is used when available (in RDF terms, datatype=non-literal), but when it is not, the cataloger is free to compose it (in RDF terms, datatype=literal). It is hoped that the lists of controlled terms will eventually take the form of embedded controlled vocabularies that exist and are defined and maintained outside the cataloging rules, allowing for timely update as terminology associated with new technologies and other types of new terminology appear and become widely adopted. This type of normalized/composed data should be displayable and keyword indexable. In addition, it should be available for limiting searches in an online catalog and for creating large lists or bibliographies for downloading, printing, etc. For example, a searcher might like to limit a particular search to DVDs only, or a library might like to download cataloging records for every title in their DVD collection so as to create a filmography.

Ideally, composed or composed/normalized data should be available in any language and should be switchable based on a users' stated preferences concerning language, script and transliteration.

Just a note about the lists of terms currently included in these rules. The current lists would need a lot of work to get to the most generally useful level of specificity, to define terms, and so forth. They were quickly cobbled together from existing lists in AACR2R, RDA, MARC 21, specialist cataloging manuals and the like just to give people an idea of what they might look like. It is hoped that criticism of these rules will not focus too closely on the lists which are, admittedly, flawed. If these rules were ever to be implemented, I'm certain that all of these lists would have to be redone by the cataloging communities using them.

0.9.3 Normalized human-readable identifiers or headings

The third type of data is normalized heading data or normalized human-readable identifiers. Normalized heading data is designed to be able to stand on its own in a display (or index) of many other normalized headings matched on a search, quickly and concisely identifying the entity it represents to as many users as possible, including both users familiar with and users unfamiliar with the entity being identified. "Stands on its own" means that the heading can identify the entity without the context of the entire work or expression description. (See Principle 1 above.)

There are four main functions for normalized heading data. First, the work and expression being described are given a normalized heading or human-readable identifier which identifies the work and expression to users; this used to be called the main entry. Secondly, normalized heading identifiers for works related in some way to the work being described are used to link the user back and forth between the work being described and the related work; these used to be called related work added entries. Thirdly, normalized heading identifiers are used to link back and forth between entities to which this work is related by virtue of membership in entity categories such as the works of a particular author, or works on a particular subject, or some other shared characteristic relationship; these used to be called author added entries, subject added entries, genre/form added entries, classification numbers, etc. Finally, variant normalized heading identifiers for all of these entities (works, expressions, creators, subjects, genre/forms, disciplines) are collected to enable the user to search under any variant for an entity and find the entity desired; these used to be called either title added entries (on bibliographic records) or cross references (on authority records).

There are four reasons why it is useful to create name-title human-readable work identifiers when applicable: 1) the name-title work identifier does double duty by collocating (i.e., bringing together in a display) not just the expressions of a work, but also the works of a creator; 2) the name-title work identifier collocates works on a subject or in a genre-form under creator name enabling catalog users to identify creators who have created extensively in a particular genre or written extensively on a particular subject; 3) the name-title work identifier allows linkage of the name part of the identifier to a record or other collection of data that represents that creator entity, thereby clustering all variant names for that particular entity; 4) this practice of creating name-title work identifiers conforms to the work identification practice implemented in millions of existing MARC 21 bibliographic and authority records. Name-title work identifiers constitute a continuum between our legacy cataloging data and the cataloging data of tomorrow.

Catalog use studies consistently show that the majority of catalog users look for known works and that they tend to know these works by their authors (creators) and titles in conjunction. Therefore, the name-title work identifier is so important in the building of effective catalogs that, even if systems can't solve the problem of linking the creator part of the identifier to the creator authority record to enable global updating (and most current global updating software can't do this anyway), the RDF data model behind these rules suggests that the creator name be used in conjunction with the title to identify the work, rather than kept separate so as to link up with the person entity. If computer software cannot be designed to be smart enough to provide maintenance when the creator name changes, human editing may be necessary.

In the RDF world, normalized data will correspond to entities that also have URI's. The entities could be referenced by means of the URI's (in RDF terms, datatype=non-literal), and the human-readable identifiers could reside in the location referenced by the URI, to be assembled on demand for display to catalog users. (It remains to be seen whether the internet can provide the speed necessary to make display assembly invisible to users! This approach was largely rejected in the mainframe and client-server environments because the necessary computer power was not yet available.)

Ideally, entities will have human-readable identifiers available in all languages, scripts and transliterations, and it will be possible to switch on demand to a preferred identifier in the users' preferred language, script and/or transliteration scheme.

Note that normalized human-readable identifiers or headings need to be designed not only to stand on their own without context in a heading display, as noted above, but also to sort effectively, so as to facilitate efficient scanning when a user’s search has matched hundreds of different entities. They also need to be designed to enable linking for the demonstration of relationships. In some ways, a normalized human-readable identifier should be conceived of as the name of a relationship or the explanation for a link; it should make it clear to a user what exactly is going to happen if they click on a particular hotlink.

0.10 Change of name

Unlike previous Anglo-American cataloging codes, this code considers that a change of name by itself does not constitute a change of identity (i.e., does not create a new entity). Change of identity should be determined by means other than tying it to a simple change of name. Change in identity should not occur lightly and should be done only when there is clear evidence that most users of libraries and archives around the world would consider that two entities exist after the change, rather than one entity which happens to have changed its name. Limiting change in identity in this way will cause our entity definitions to correspond more closely with users' entity definitions and will better support international cooperation, since no other cataloging rules internationally divide person entities into bibliographic identities, corporate body entities into new corporate entities with changed names, and serial works into new works with changed titles in the same fashion that AACR2R does. If considered useful, users should be provided access to works created under particular pseudonyms or earlier forms of corporate names by means of access to transcribed statements of responsibility. For earlier corporate names, date limiting can also be used. Certain situations do constitute changes in identity. For example, when two or more entities merge together into one, or when two or more separate entities result from the split of a previously existing entity, change in identity has occurred. For serials, the beginning of a new numbering scheme signals the creation of a new serial title rather than the creation of a new work.

0.11 Mandatory vs. optional

All data elements in these rules should be available for use whenever they are applicable. The assumption is that these rules would be applied by experienced and educated catalogers capable of good judgement. Such catalogers do not need to have data elements labeled "mandatory" or "optional." Labels such as these can only impede the application of good judgement to the infinite variety of situations routinely encountered by catalogers.

0.12 Relationship to FRBR

These rules refer to and adapt definitions of the entities work, expression, manifestation, and item as found in Functional Requirements for Bibliographic Records or FRBR, for short. FRBR is valuable because of its attempt to define concepts that have been implicit in Anglo-American cataloging practice for several hundred years at least. Unfortunately, the tables in the back of FRBR that attempt to map the elements of the bibliographic description to these four entities (work, expression, manifestation and item) are deeply flawed. This code suggests implicitly a better mapping of these data elements to the FRBR entities.

It should be noted that in this code Principle 3, the transcription principle (see above), is applied at the expression level as well as at the manifestation level. FRBR suggests that everything above the item level is "abstract" and some have argued that this means that only the item is concrete and only the item therefore can sit on the desk in front of the cataloger and be a source of transcription. (To be exact, FRBR defines a work as being "abstract" and the boundary between work and expression as being dependent on "the bibliographic conventions established by various cultures or national groups" (FRBR, p. 16). It also discusses manifestations as being "abstractions" describing the "shared characteristics of items" (FRBR, pp. 22, 78).) Perhaps it is unfortunate that FRBR chose to use the term "abstract," since it seems to have prevented people from noticing that cataloging has always been a process of generalizing about expressions and works from evidence found on specific items being cataloged. Most cataloging is carried out on works that have been published and/or distributed in multiple copies, and the assumption has always been that an accurate description of one copy would apply to all copies and that information about manifestation, expression and work could be deduced from information found on one copy. In this code, that assumption is continued. Transcribed data elements that pertain to expression are linked to expression and transcribed data elements that pertain to manifestation are linked to manifestation. It may be necessary at times to look at all the other expressions and manifestations of a work before you know whether, for example, a particular transcribed title pertains only to this manifestation (or, as we call it here, title-manifestation), whether it pertains to all manifestations of this expression (as when the expressions of the work have had different titles), or whether, indeed, it pertains to the work (as when all expressions of the work have had the same title). See also "the principle of transcription as applied to an expression" (section 2.2).

There will undoubtedly be objections raised about asking catalogers to make these decisions about whether particular data elements pertain to work, expression or manifestation. My cataloging experience would lead me to assert that a well-trained and well-educated cataloger can nearly always do this easily. There may be times when the same element pertains to two or three of the FRBR group 1 entities and will have to be coded as such, so this approach could lead to more repetition and redundancy than we have now, but I think it is do-able. If majority opinion is that catalogers are incapable of doing this, though, that position logically implies that it will not be possible for catalogers to code records in such a way that a computer can group all of the manifestations that make up a particular expression. In other words, it will not be possible for us to solve the multiple versions problem because our personnel is inadequate to the task. I have suggested elsewhere how we could use our current records created according to current practices to better FRBR-ize catalogs by simply creating software that pays attention to the work identifiers already present in our records and uses them to build indexes and displays (Yee, Martha M. FRBRization: a Method for Turning Online Public Finding Lists Into Online Public Catalogs. Information Technology and Libraries 2005;24:2:77-95. Also available on the Web at:

http://repositories.cdlib.org/postprints/715). The users would then have to look through all of the manifestation records representing various expressions of a particular work and make a decision about which expression is desired (and which manifestations seem to represent that expression). These new rules are an attempt to solve the so-called "multiple versions" problem by enlisting the aid of catalogers in determining which manifestations represent the same expression.

0.13 Title-manifestation

In addition to the FRBR entities work, expression, manifestation, and item, these rules define a fifth entity, the title-manifestation. A title-manifestation is a kind of manifestation (change only in distribution information or format) that occurs when significant marks of identification, such as the title, change without any change in underlying content.

0.14 Serial title

Title-manifestation, which by definition has no change in content, does not work for serials, since change in content occurs with every new issue as a matter of course. A sixth entity is therefore defined for serials, since they continue over time not by replacing themselves with revised editions, as monographs do, but rather by a continuous stream of issues each of which is a new compilation of other works (articles). Change in title has a profound effect on users' ability to find the serial work that contains the article work they are actually seeking. This new entity is called a serial title. In effect, it is used to summarize all manifestation changes (changes in publication information) that occur during the time that a serial's issues all bear the same title before the next major title change. Since a serial's content is continuously changing, it is possible for the subject matter of a serial to change below the expression level, and for subject headings to be serial title-change specific. This would not be possible for a monograph below the expression level, since any content change would create a new expression.

Serials that exist in several simultaneously released editions, such as different language editions or editions with different content for different markets, on the other hand, do seem to fit into the definition of expression that works for monographs as well.

According to the FRBR model, every item contains a manifestation, an expression and a work. Since title-manifestation is a kind of manifestation and since serial title plays somewhat the same role as title-manifestation for monographs, the following permutations are all possible according to these rules:

item--manifestation--expression--work (monographs and integrating resources)

item--manifestation--title-manifestation--expression--work (monographs)

item--manifestation--serial title--expression--work (serials)

0.15 Degression: begin with description of the work

In contrast with AACR2R and RDA, which begin with description of the manifestation, these rules begin with description of the work, proceed to description of the expression, then description of the manifestation, and finally description of the item. At each of these stages, the first step is to determine if this work is already described, if this expression is already described, or if this manifestation is already described. This procedure recognizes the actual process of cataloging, which does not take place in a vacuum, but is done against an existing set of cataloging records. As Ed Jones puts it:

"I think cataloging is actually two simultaneous activities: the cataloging of the object, which is primarily a manifestation-based activity, and the integration of the catalog record into the catalog, which is necessarily a work-based activity. Since one cannot begin to catalog an object until one has determined that it is new to the catalog, one must construct a provisional main entry for searching purposes before one can determine whether or not to catalog the object at all (or just add it as an additional copy to an existing record). [Sorry about all those "ones".] For the catalog search to be effective, one must determine the forms of the headings under which to search as well as the titles and variant forms. So one is tentatively cataloging the item (and performing much of the heavier intellectual work) before determining whether it is necessary to create a catalog record in the first place.

For some reason, all catalog codes are written with the assumption that cataloging is taking place in a vacuum, and the object being cataloged is the first to be entered in the catalog. In practice, the search of the catalog is the first step, and one must make a number of preliminary determinations before undertaking the catalog search:

(1) What is the title(s) under which I may find this (or something similar) in the catalog?

(2) What are the names under which I may find this (or something similar) in the catalog?

(3) What is the probable form of the headings that represent these names in the catalog?

A computer algorithm may already have searched the catalog by ISBN, etc., based on data in vendor records. If not, then a fourth determination is necessary:

(4) What are the standard identifiers under which I may find this (or something similar) in the catalog?

At this point, the cataloger searches the catalog--probably WorldCat--and tries to understand the results, which may be very complex, with dozens of candidate records for this object or related objects.

Only when a determination has been made that the object is new to the catalog, and only when its relationships with other objects in the catalog have been established, can one proceed to actually describe the object in some definitive way (though we have been examining the descriptive elements all along), identify it (perhaps within the framework of a family of objects), and provide appropriate access points.

I think the 'context of the catalog' would be very useful to include in a catalog code, not just for pragmatic purposes (because this is how we catalog) but because it tends to justify selecting the work as the Group 1 entity with which to begin the cataloging process."

Note that the approach to cataloging taken in these rules is such that data elements that apply to all expressions of a work are recorded at the work level and not repeated at the expression, title-manifestation, serial title, manifestation, or item level, and data elements that apply to all manifestations of an expression are recorded at the expression level and not repeated at the title-manifestation, manifestation, or item level, etc. This approach used to be known as degression. An important implication of this approach is that a particular data element may have to move down the work-expression-manifestation-item hierarchy over time as conditions change. For example, if all expressions of a work are illustrated in the first few years of the work's existence, the illustration statement may go at the work level, but if at some point in the future an expression of the work is published without illustrations, the illustration statement will have to move down to the expression level at that point in time. It is not really practical to practice degression in the current shared cataloging environment, but if the internet has the effect of moving us toward cataloging in a single virtual catalog, it might become more practical; it may even be possible to let computers do some of the data element movement, or at least signal to the cataloger when it might be necessary.

0.16 Acknowledgements

A number of very bright and expert people graciously agreed to read over this text and suggest improvements, among them Everett Allgood, Sherman Clarke, Renette Davis, Greta de Groat, Neil Hughes, Ed Jones, and Kelley C. McGrath. The text would not be nearly as readable and useful without their help. However, all flaws that remain are completely my responsibility. Jim Weinheimer has been kind enough to set up a public wiki site for the rules where anyone can edit the rules themselves, a brilliant way to experiment with creating a community of cataloging experts that is not geographically limited or cramped by lack of travel funds.

1. WORK

Follow Functional Requirements for Bibliographic Records (FRBR) in considering a work to be a distinct intellectual or artistic creation.

1.1 Work identification [type of data: normalized heading--work identifier; not repeatable]

Identify the work contained in the item being cataloged. If the work is already contained in the catalog, use the work identifier already established. If the identifier changes over time, it should be changed in all places in which it appears in the catalog.

1.1.1 Works identified using the name of a creator

If it is a work with a single personal writer, composer, painter, photographer, compiler or other type of creator, or a work for which one person or corporate body is represented as being principally responsible, identify the work using the name of that person or corporate body in conjunction with the title. If several creators are listed, consider the one named first as being represented as being principally responsible. Use the rules in Section 8 to determine the identifier for a creator used in a name-title work identifier. If it is not a work with a single creator represented as being principally responsible, identify the work using the title alone (see 1.1.2).

EXAMPLES:

Shakespeare, William, 1564-1616. Romeo and Juliet

Beethoven, Ludwig van, 1770-1827. Symphonies, no. 5, op. 67, C minor [the musical score]

Vermeer, Johannes, 1632-1675. Girl with a pearl earring.

Keating, P. J. Working-class stories of the 1890s [a compilation of stories by many authors, edited with an introduction by P.J. Keating]

Bishop, Henry R. (Henry Rowley), Sir, 1786-1855. Faustus [statement of responsibility reads: 'composed by T. Cooke, Charles E. Horn, and Henry R. Bishop;' Bishop's name is displayed more prominently than those of the others]

Dylan, Bob, 1941- The basement tapes [statement of responsibility reads: 'Bob Dylan and the Band']

National Center for Health Statistics (U.S.). Trends in infant mortality by cause of death and other characteristics, 1960-88.

Symposium on Tar Sands (1976 : Toronto, Ont.). Symposium on Tar Sands held as part of the Canadian Chemical Engineering Conference, Oct. 3-6, 1976, Toronto.

National Geographic Society (U.S.). Cartographic Division. South America.

1.1.1.1 Laws

Identify laws by treating the jurisdiction as the creator of the laws.

EXAMPLE:

United States. Constitution

1.1.1.2 Criminal proceedings

Identify criminal proceedings by treating the person or body prosecuted as the creator of the proceedings.

EXAMPLE:

Alley, Leavitt. Report of the trial of Leavitt Alley, indicted for the murder of Abijah Ellis, in the Supreme Judicial Court of Massachusetts

Identify civil and noncriminal proceedings by treating the person or body bringing the action as the creator of the proceedings.

EXAMPLE:

Goodwin Film and Camera Company. The Goodwin Film and Camera Company, complainant, vs. Eastman Kodak Company, defendant

1.1.1.3 Treaties

Identify treaties by treating the first named signatory jurisdiction as the creator of the treaty.

EXAMPLE:

Belgium. Convention between the governments of the United Kingdom, Belgium, and France regarding the supervision and preventive control of the African migratory locust

1.1.1.4 Archival collections, photograph albums and other collections

Identify an archival collection using the name of the person(s), family (families) or corporate body predominantly responsible for the creation, assembly, accumulation, and/or maintenance of the materials. Supply a title consisting of the name of the person(s), family (families) or corporate body predominantly responsible for the creation, assembly, accumulation, and/or maintenance of the materials, along with a term expressing the nature of the archival unit (such as collection, records, papers, manuscripts, photograph album, etc.), and a term expressing the topic of the archival unit, if applicable, in the form of 'on [topic].'

EXAMPLES:

Davenport, Mildred, 1900-1990. Mildred Davenport dance programs and dance school materials, 1914-1942

Bacot family. Bacot family papers, 1767-1887 (bulk 1845 1866)

Minnesota. Dept. of Health. Swine influenza immunization program records, 1975-1979

Boston War Camp Community Service. Bureau of Entertainment. Photograph album. Six views of United States soldiers and sailors entering, or waiting to enter, the Boston Theatre for Bureau of Entertainment-sponsored vaudeville shows or concerts, 1918-1919.

1.1.1.5 Works of mixed authorship

For works of mixed authorship in which different functions are carried out by different people, a judgment must be made as to whether one function and the person or corporate body which carried it out are so predominant as to render that person or body significant for identification of the work. For example, for purely musical performances, consider the composer of the music to be the principal author unless improvisation or adaptation is so significant as to create a new related musical work by the improviser or adapter. Consider a moving image recording of a purely musical performance using a static camera to be the work of the composer of the music. (Note, however, that most dramatico-musical works, such as operas, and most combinations of words and music, such as popular songs, are works of mixed authorship. These are not purely musical works. See 1.1.1.7. and 1.1.2.2)

EXAMPLE:

Beethoven, Ludwig van, 1770-1827. Symphonies, no. 5, op. 67, C minor [ a sound recording of a performance]

1.1.1.6 Compilations of musical performances

If desired, a sound recording compilation of works by many composers with a single performer can be identified using the name of the performer, as a kind of compiled work identified by its compiler (see rule 1.1.1 above).

EXAMPLES:

Harris, Emmylou. Pieces of the sky

Schumann, Elisabeth. Elisabeth Schumann

1.1.1.7 Two works intertwined

When two separate works are intertwined in a single publication, use judgment to determine whether one work is subsidiary to another, or the two works are conjoined in a third composite work. For example, when music and words have had independent existence as publications, treat the work as a compound work containing two works. Make separate records for each of the two works, or, optionally, create a record for one with the other work given a "work contained" added entry. When the words have existed for a long time, and the music is composed on just one occasion, treat the music as subsidiary to the words, creating a work identifier for the words consisting of name of author of words and title of words, and describing the presence of music as expression change. When the music has existed for a long time, and the words are written on just one occasion, treat the words as subsidiary to the music, creating a work identifier for the music consisting of name of composer of music and title of music, and describing the presence of words as expression change. For another example, consider the writer of a text with illustrations to be the principal author, but consider the artist to be the principal author of a pictorial work with commentary. There are times when it is best to consider a work consisting of two intertwined works to be a work of mixed authorship to be identified by title alone. For example, a popular song that has music by one person and words by another might then be performed in very improvisatory style by a number of different musicians. The best identification for such a work is the title alone.

1.1.1.8 Relationship of principal creator to work

For works identified using the name of a principal creator, include a relator term from the following list if possible to express the nature of the creator's relationship to the work; choose the most specific term, if possible; if none of these terms apply, compose a term to describe the relationship between the creator and the work being cataloged; if the relationship is ambiguous, do not include a relator term: