PROPOSAL NO.: 2000-04

DATE:Dec. 6, 1999

NAME:Anonymous Attribution Information in Personal Name Headings

SOURCE:Art Libraries Society of North America, Cataloging Advisory Committee

SUMMARY:Works which cannot be attributed with certainty to a known person are common in the art world. Art historians express gradations of certainty about the relationship between a work and a known artist or group of artists by coupling a known artist's name with a qualifier. Librarians and visual resources professionals who use the MARC 21 formats in cataloging art works or surrogates of art works need to be able to record this information in an appropriate place in a MARC record. This paper proposes handling anonymous attribution information associated with personal names in a separate subfield.

KEYWORDS: Field X00 (BD, AD, CD, CI); Personal Names (BD, AD, CD, CI); Anonymous attribution information (BD, AD, CD, CI); Subfield $g, in X00 fields (BD, AD, CD, CI); Subfield $j, in X00 fields (BD, AD, CD, CI)

RELATED:DP115 (June 1999)


12/6/99 - Forwarded to the MARC Advisory Committee for discussion at the January 2000 MARBI meetings.

1/15/00 - Results of MARC Advisory Committee discussion - Approved Option 2. Participants generally felt that this information should be separated from that in subfield $g, since the latter is a dumping ground for miscellaneous information. The format should clarify the qualifiers to be used and assure that they follow the name of a known artist. The description of subfield $g should clarify that it is a dumping ground for information with which the user may not know what to do.

2/11/00 - Results of LC/NLC review - Agreed with the MARBI decisions.

PROPOSAL NO. 2000-04:Anonymous attribution information


Situations where the responsibility for a work is unknown, uncertain, fictitious, or pseudonymous can be found for virtually all forms of material. Such situations come up often enough for there to be specific guidance for dealing with them in cataloging rules such as AACR2 and other description conventions. Assignment of responsibility is most difficult for works of art, which generally do not have anything like a title page to convey information. Signed works of art are the exception rather than the rule. Art historians attribute responsibility for a work on the basis of stylistic analysis, traditional ascriptions, inventories, and other evidence. Because attribution is so often subjective and depends on a chain of reasoning or on sources of varying degrees of reliability, scholarly literature has evolved elaborate, nuanced terminology for referring to the creators of works of art.

Data dictionaries for the description of art objects regularly include a data element for attribution information. Categories for the Description of Works of Art includes an element "creation-creator-qualifier," for information about the attribution of a work to a known artist or group, including any possible relationship of an unknown artist to a known artist or group in whose style the work has been created (e.g., attributed to, follower of, in the manner of.)" (Categories for the Description of Works of Art: Definitions," Visual Resources XI, 3-4 (1996), p. 401). The Drawings Documentation Project Procedure Manual, developed by the Art Institute of Chicago, defines an "artist qualification field," which is used to characterize the relationship of the work to a known artist. It suggests relationships such as artist, attributed to, assistant of, pupil of, studio or workshop of, circle of, follower of, school of, style of, after, and imitator of. (Suzanne Folds McCullagh, "Nuances of Art Information," Visual Resources XI, 3-4 (1996), p. 275).

The degree of certainty of attribution of responsibility for works of art ranges from absolute certitude to nescience. At one end of the continuum, the artist's name is known, and cited with varying degrees of confidence ("signed," "attributed to", "traditional," "questionable," etc.). At the other end, the artist is unknown and unidentifiable except as belonging to a given school or period ("Rhenish School," "Barbizon School"). Schools in the regional or stylistic sense do not fall within the scope of this proposal, which deals only with attributions attached to personal names. Somewhere in the middle are the artists, who, while remaining unknown, can be individualized on the basis of features of their style, and assigned a name which consists of a descriptive phrase: "Achilles Painter," "Master of the Saint Bartholomew Altarpiece," "Entrelac Binder." Names of this type are analogous to the characterizing phrases ("An Oxford M.A.") which sometimes appear on the title-page of printed books and which are treated as names.

Up to now, the MARC 21 formats have not addressed specifically the encoding of qualifying attribution information, which when combined with the name of a known artist or creator, expresses the relationship of a work to that known person. Examples of such cases in both direct and "catalog entry" form are given below:

      School of Andrea Mantegna   (or, Mantegna, Andrea, 1431-1506, school of)
      Circle of Jacquemart de Hesdin   (or, Hesdin, Jacquemart de, circle of)
      Follower of Hokusai   (or, Hokusai,Katsushika, 1760-1849, follower of)
      Copyist of Rodin    (or, Rodin, Auguste, 1840-1917, copyist of)
As librarians catalog more art objects and surrogates of art objects (slides, photographs, digitized images), they are increasingly encountering this type of shadowy authorship. The same kind of problem with attribution is found in map cataloging as well, where sometimes the responsibility for an item can only be inferred (e.g., "map in the style of Ptolemy"). Book cataloging offers precedents for dealing with some, but not all attribution issues. Some questionable attributions to a known person can be handled by the use of the relator term, for which there is a clearly appropriate MARC 21 subfield $e (Relator term) defined for the X00 (Personal Name) fields. For example, subfield $e often contains the term "attributed name," which is defined in the MARC Code List for Relators, Sources, Description Conventions" to relate an author, artist, etc. to a work for which there is or once was substantial authority for designating that person as author, creator, etc. of the work. Attributions such as those expressed according to the "School of X" formula remain problematic for MARC encoding.


The encoding of attribution information associated with the name of a person is of considerable importance to the art and museum community since anonymous artist relationships are very common in describing works of art. Various metadata standards (Categories of the Description of Works of Art, Visual Resources Association Core Categories, and AMICO) have data elements that may be used to record relationship information and to indicate attribution. As institutions try to fit data about visual resources or art works into MARC- based systems, a decision needs to be made on the most appropriate data element for this information.

Several options are possible using currently defined MARC 21 content designators. Discussion Paper No. 115 suggested the use of subfield $c (Titles and other words associated with a name). The definition of subfield $c is open enough to accommodate this kind of information, but its long established use for titles such as "Sir", "Dr.", and "Mrs." resulted in fairly strong opposition to using it for attribution information. Objection was also raised about the use of subfield $c since it implies authority control of the name heading that includes attribution information. Comments were split over whether this was a problem, however. In the end, there was consensus that the use of specific data elements in the MARC 21 bibliographic format in access fields do not necessarily imply the creation of a MARC 21 authority record with a heading containing the same elements. There is a variation in practice as to when a corresponding name authority is created. Regardless of the solution found for artist attribution information in MARC 21, a subsequent decision would have to be made (perhaps at an institution level) about whether to require separate authority headings for names including attribution information.

A more acceptable option for recording anonymous attribution information could be the use of subfield $g (Miscellaneous information), which is already defined in all the X00 (Personal Name) fields. The Smithsonian Institution has already been using subfield $g in the X00 (Personal Name) fields for anonymous artist attribution information for several years. In general, subfield $g has not been widely used in MARC records in the X00 fields. It was defined mostly for consistency with the X10 and X11 fields, where it has been used for special qualifying data (e.g., the qualifier "never held", used with the names of meetings that were planned but never took place). The definition of subfield $g could be revised to address its use for anonymous attribution information. Examples of the use of this subfield for artist attribution information could be added to the format, adding greater justification for the definition of this subfield in the first place. It is not clear that the data recorded in subfield $g is of the type that requires indexing or authority control. Again, this may not matter, since indexing and authority control of specific MARC data elements are often decided upon locally. The main point to be made is that the X00 fields include a second defined subfield, that could be used for artist attribution information. This would avoid the need to define a new subfield.

The only other option for artist attribution information is the definition of a new subfield in the X00 fields. A new, separate subfield would allow the unambiguous encoding of such information in the X00 fields. There are still two subfield codes available in the formats in personal name heading field. The same subfields are available in the Bibliographic and Authority formats, should authority control of headings with such data be required. This solution may be the least desirable since it requires using one of the last available alphabetic subfields in the X00 field group. A new subfield would have to be either subfield $i or $j. Since subfield $i has been established in many fields for instructional phrases, subfield $j was preferred during discussion of DP115. Subfield $j has not been defined in the 1XX or 6XX fields, or fields 700-740 (the primary access fields). The following are examples of headings making use of each of the possible subfields.

      100 (0 )  Rembrandt Harmenszoon van Rijn,$d1606-1669,$gSchool of
      100 (1 )  Reynolds, Joshua,$cSir,$d1723-1792,$gPupil of
      100 (0 )  E. S.,$cMeister,$d15th cent.,$jFollower of
There was clear consensus on a solution at the subfield level since use of a separate field tag (such as 720) for headings with artist attribution information would not support the collocation and interfiling of headings with and without the information. There was also concern that use of field 720, as was suggested, would make use of a field that was specifically exempt from authority control. It is not clear that authority control would or would not be needed for headings with this data, but use of 720 would have precluded it. Regardless of the subfield chosen, use of a separate subfield, appended to the end of a heading for a known artist/creator, has the advantage of keeping the headings for the related person close to the known person, while distinguishing clearly between the two. The subfield could be treated for indexing purposes as a filing element, so that the headings with attribution information file after the heading for the known person. The Cataloging Advisory Committee of ARLIS/NA is willing to supply a list of suggested terms for use in whichever subfield is deemed appropriate for attribution information if it is desired to have a controlled list. The list would draw upon the terms most often used in the literature, and would offer a degree of standardization.

It is unclear that consensus could be reached over whether there would be occasion to provide authority control for headings including anonymous attribution information. Since some cataloging agencies indicated that they would want to control such headings, at least when cross references are needed, it is advisable to extend any solution decided upon to the MARC 21 Authority format as well. Since unique classification numbers are assigned to artists, and information about exhibits and programs dealing with unknown artists might be part of MARC 21 Classification and Community Information records, applying any format changes to the CD and CI formats should be considered as well.


In the MARC 21 Bibliographic, Authority, and Classification, and Community Information formats:

Option 1:

Option 2:

Library of Congress
Library of Congress Help Desk (02/11/00)