subject access, secondary subject matter, subject headings, research collections, subject matter, subject heading, inscriptions, Princeton Index, Iconclass, the history of art, historical scholarship, Erwin Panofsky, number, Photographic Archives, visual resources collections, Iconclass number, Index of Christian Art, the Department of Art and Archaeology, iconographical research collection, art historian, Princeton University, photographic reproductions, Christian art, Online Computer Library Centrr, Automatic Processingof Art History, Laura Corti, Rijksbureau voor kunsthistorische Documentatie, Yale Center for British Art, Regione Toscana, BFM, secondary subject, Iconclass numbers, Photographic Archive, Rijkbureau voor kunsthistorische Documentatie, Sarasan, Art of the Low Countries, Automatic Retrieval, Computerization, Marburger Index, Florence, Italy, Decimal Index, microfiche
Access to Iconographical Research Collections KARENMARKEY INTRODUCTION ICONOGRAPHICARLESEARCH COLLECTIONS are tools that support the iconographical approach to the history of art and offer access to secondary subject matter
and photographic reproductions of artworks. In this article, the four major approaches to art historical scholarship are defined. One of those four approaches-iconography-is targeted in this discussion because it supports iconographical research collections. A matrix of subject access to visual resources collections displays the many possibilities for providing subject access to these collections. Generally, iconographical research collections describe secondary subject matter through subject headings or a classification scheme; these collections employ phrase indexing of subject headings or classification codes. Access to three different iconographical research collections through secondary subject matter is described and their respective positions in the matrix are pinpointed. Typical users of iconographical research collections are scholars in Art History
, theology, and medieval studies who are trained in the iconographical approach to the history of art. Thus access to iconographical research collections through secondary subject matter is in accordance with the training and knowledge of users of these collections. Access to two different visual resources collections through primary subject matter is described. Users of these collections do not need special training in iconography because primary subject matter can be described by anyone with practical experience with everyday life. Subject access to these collections is contrasted with that offered by icono- Karen Markey, School of Information and Library Studies, University of Michigan, 580 Union Drive, Ann Arbor, MI 48109 LIBRARY TRENDS, Vol. 37, No. 2, Fall 1988, pp. 154-74 @ 1988 The Board of Trustees
, University of Illinois
MARKEYACONOGRAPHICAL RESEARCH COLLECTIONS 155 graphical research collections. The reasons for the limitation of the latter collections to secondary subject matter are presented along with future directions in subject access to iconographical research collections. APPROACHTEOSARTHISTORICASCLHOLARSHIP In general, art historical scholarship employs one of four approaches: 1. Form and style requires an analysis of formal conventions. It results in "the identification and definition of artistic styles such as Byzantine, Gothic, and Baroque, through artworks' morphological terms, as representing tendencies toward certain modes of composition, of color organization, proportion, treatment of space, etc." (Ackerman & Carpenter 1963, p. 9). 2. Connoisseurship is the "study of the artistic personality of an individual artist or of a group or School of Art
ists as revealed in their works" (Ackerman & Carpenter 1963, p. 203). 3. Iconography is "the study of subject matter or meaning in art" (Panofsky 1962, p. 3). 4. The Social History
of art "interprets a work of art in terms of the conditions that brought it into being" and "calls upon sources that art has in common with the other activities of society" (Ackerman & Carpenter 1963, p. 220). Scholars may span one or more approaches in their published work. In Painting in Florence and Siena after the Black Death, Millard Meiss brought "his background as a connoisseur and iconographer into partnership with a study of religious and literary history
to explain a dramatic shift in style following the plagues of the mid-fourteenth century in central Italy" (Ackerman 8c Carpenter 1963, p. 224). Underlying the great variety of approaches among art historians, there is a common need for certain basic identifying information that affects every historical operation. Regardless of the approach of art historians, nearly all ask of the work they study: ( 1 ) When was it executed (date)?(2) Where has it been since its execution (provenance)? (3) Who created it (responsibility)? (4)How was i t created (materials, tools, techniques)? (5) Why was it created (function)? Bibliographic tools and collections of visual resources have been established to support one or more of the four approaches to art historical scholarship. An art slide collection is a tool familiar to librarians and curators from academic institutions, museums, and art galleries. The art slide collection may be an administrative unit of the university library, museum, art gallery, art history department, or architecture department. However, the identifying information assigned to each slide in the collection answers the questions all scholars ask of a work of art, namely when the work was done, where, by whom, how, and why.
156 LIBRARY TRENDSIFALL 1988 Zconograp hy The prominent art historian Erwin Panofsky (1962)defined icono- graphy as "that branch of the history of art which concerns itself with the subject matter or meaning of works of art" (p.3). In the analysis of an artwork, the iconographer distinguishes three levels of subject matter:(1) primary subject matter, (2) secondary subject matter, and ( 3 ) iconological interpretation. Primary Subject Matter. Primary subject matter is "the identification of forms such as configurations of line and color, or peculiarly shaped lumps of bronze or stone as representations of natural objects such as human beings
, animals, plants, houses, tools and so forth" (Panofsky 1962, p. 5). Panofsky employs Leonard0 Da Vinci's Last Supper to explain how to identify primary subject matter. The Last Supper shows thirteen men sitting around a table eating dinner and engaged in conversation. This enumeration of objects, events, and expressional qualities is a @re-iconographical description of this artwork. The only knowledge needed to formulate a pre-iconographical description is practical experience with everyday life. Secondary Subject Matter. Secondary subject matter is "the identification of themes or concepts manifested in images, stories, and allegories'' (Panofsky 1962, p. 6). When viewers realize that thegathering of these thirteen men depicts Christ's last supper with his disciples, they have reached the second level of meaning called secondary subject matter. The identification of secondary subject matter is called an iconographical analysis. The iconographer relies upon his knowledge of literary sources, customs, and cultural traditions
peculiar to a certain civilization. Zconological Interpretation. Iconological interpretation is "the identification of underlying principles which reveal the basic attitude of a nation, period, religion, class, or philosophical persuasion" (Panofsky 1962, p. 7). For Panofsky, the culmination of iconographical analysis is iconological interpretation. "But when we try to understand it [Leonardo's Last Supper] as a document of Leonardo's personality, or of the civilization of the Italian High Renaissance, or of a peculiar religious attitude, we deal with the work of art as a symptom of something else which expresses itself in a countless variety of other symptoms" (Panofsky 1962, p. 8). The study of iconography (or subject matter in art) requires scholars engaged in iconographical analysis to pass through the first level of interpretation-i.e., pre-iconographical description. Thus, such scholars actually perform two interpretations of an artwork-i.e., preiconographical description and iconographical analysis. ICONOGRAPHICARLESEARCCHOLLECTIONS Iconographical research collections are tools that support the iconographical approach to the history of art and feature access to subject matter and photographic reproductions of artworks.
MARKEYACONOGRAPHICAL RESEARCH COLLECTIONS 157 In the late 1960s, discussions by scholars on the automation of visual art
s information revealed the availability of data for conversion to machine-readable form per level of subject matter. Art historian Kenneth C. Lindsay (1968)listed these three levels and added another level for empirical information-i.e., identifying information that tells when the work was done, where, by whom, how, and why (seeFigure 1) (p. 25). This figure shows that identifying information and secondary subject matter of artworks are presently described in iconographical research collections and thus are available for conversion to machine-readable form. Primary subject matter, however, is not described in iconographical research collections; consequently, primary subject matter is not available for conversion to machine-readable form. Lindsay considered the third level of interpretation-iconological interpretation-beyond the purview of automation. Since scholars must pass through the first level of interpretation to perform an iconographical analysis of a work of art, why has subject access to iconographical research collections been limited to secondary subject matter? Two explanations for this focus on the physical format and users of these collections: 1. The physical format of iconographical research collections is a card or microform format. These formats preclude the assignment of more than a few subject headings, descriptors, or classification codes per image. Secondary subject matter describes themes, stories, or allegories of an image and thus requires one or two entries per image. In contrast, primary subject matter requires the identification of many entities represented in an image-i.e., objects, events, and expressional qualities. Preparation and filing of multiple entries for primary subject matter for providing access points to a card- or microform-based collection would be prohibitively expensive in terms of the professional and clerical labor required. 2. Iconographical research collections have been established to support scholarly research; thus, secondary subject matter is in accordance with the training and knowledge of users of such collections and supports their scholarly pursuits. Matrix of Subject Access A three-dimensional matrix displays the many possibilities for providing subject access to visual resources collections (seeFigure 2). At the top of the matrix are the two types of subject matter-primary and secondary-suitable for conversion to machine-readable form. On the left side of the matrix are the three principal methods of subject access. At the bottom of the matrix, four indexing techniques for visual resources collections are given. The organization of this two-by-three-by-four-cellemdatrix seems to indicate that any of the four indexing techniques can be applied to any of the three methods of subject access and so on. In practice, this is
158 LIBRARY TRENDSIFALL 1988
Act of Interpretation A. Empirical (e.g., title, date, size,artist) I. Pre-iconographicaldescription II. lconographical analysis I II. lconological interpretation
Is data available for computer input? Yes No Yes No (but prc-ably not necessary)
Figure 1 . Data availability per level of interpretation for conversion into machinereadable form. Source: Lindsay, Kenneth C. "Computer Input Form for Art Works: Problems and Possibilities." In Computersand their Potential Applications in Museums; A Conferencesponsored b y the Metropolitan M u s e u m of Art, p. 25. New York: Arno Press, 1968.
Type of Subject Matte
Principal Methods of Subject Access: 1. Classificationschemes
2. Subject headings: a. Alphabeticallist b. Alphabetico-classedlist c. Classified list
3. Prose descriptions (or keywords)
Derived Phrase Keyword Keyword
search-key indexing in context indexing
Figure 2. Matrix of subject access to visual resources collections not true. This section features descriptions of three iconographical research collections and pinpoints each collection's position in the matrix. The three collections covered in this paper are characteristic of iconographical research collections, especially with regard to their respective position in the matrix.
MARKEY/ICONOGRAPHICAL RESEARCH COLLECTIONS 159 Index of Christian Art (Princeton Index). The Index of Christian Art (Princeton Index) was begun in 1917 by Charles Rufus Morey, an art historian at Princeton University
in the Department of art and archaeology
. The Princeton Zndex "grew out of iconographical research that was actively pursued in the Department of Art and Archaeology" (Woodruff 1942, p. vii). Today, the Index of Christian Art is still an administrative unit within that department. Its director, Nigel Morgan, is an art historian known for research in Gothic manuscripts. The mission of the Index since 1917 has been to catalogue by subject and "picture-type" all of the known (published) monuments of Christian art dated before the year 1400, to record briefly the history of the objects, to assemble the important bibliography relating toeach monument, and finally, when the literature of art history now available has been searched and exhausted, to maintain the catalogue by adding to i t yearly all of the newly published material and all of the pertinent bibliographical references." (Woodruff 1942, p. 2) In Figure 3, a portion of the matrix of subject access (Figure 2) is shown pinpointing the Index of Christian Art. In general, the Zndex is composed of two card files: (1)an alphabetical subject index, and (2)a monument file of an estimated 260,000 black-and-white photographic reproductions of artworks. A user of the Index of Christian Art who is interested in images of Lucifer first consults the alphabetical subject index under a subject heading. Figure 4 shows the alphabetical arrangement of subject headings under the main heading "Angel." Under the subdivided heading "Angel: Lucifer," the user first encounters a scope note or information card for this heading which bears other headings that the user can consult for this topic (see Figure 5 ) . Following this information card are main entry cards (see Figure 6) and reference cards (see Figure 7) that describe the secondary subject matter of the artworks and direct the user to the monument file for a photographic reproduction of the artwork described. In the monument file, photographic reproductions are filed by medium-e.g., metalwork, illuminated and illustrated manuscripts, sculpture, painting, mosaic-then alphabetically by city, then by institutional type, and institution. The user finds a photograph of the illuminated manuscript showing Lucifer by first consulting the monument file for illuminated and illustrated manuscripts, then under the city name Strasbourg, then under libraries in this city, and so on. Users of the Index of Christian Art are principally art historians and scholars in related disciplines. The photographic collection of the Frick Art Reference Library and Index of Jewish art
resemble the Princeton Index in their construction and access to secondary subject matter (Knox 1979; Narkiss & Sed-Rajni 1976). Iconclass. Iconclass is not an iconographical research collection; rather, i t is a hierarchical classification scheme used as a subject access tool in a
160 LIBRARY TRENDSIFALL 1988
II. Secondary subject matter B. Phrase indexing 2b.Alphabetico-classed list of subject headings
Figure 3. Pinpointing the Princeton index's position in the matrix
/ Anae1:Scene AngeLLucifer on Throne of God f
r I f
Angel:Fall Angel:Cherub T
Figure 4. Princeton index guidecards following "Angel"
MARKEYACONOGRAPHICAL RESEARCH COLLECTIONS 161 number of iconographical research collections. The scheme was initiated by the art historian Henri Van de Waal and completed by Leendert Couprie at the Department of Art History at the University of Leiden with financial assistance
from the Netherlands Organization for the Advancement of Pure Research and the Royal Netherlands Academy of Arts and Sciences (Van de Waal 1974-85; Couprie 1983). Iconclass distributes iconography into nine broad categories: (1) religion and magic; (2)nature; (3)human being, man in general; (4)society,civilization, culture; (5) abstract ideas and concepts; (6) history; (7) Bible; (8) literature; and (9) classical mythology, ancient history. Iconclass can be used by the keepers of iconographical research collections to effect a systematic arrangement of artworks by secondary subject matter. In Figure 8, a portion of the alphabetical Iconclass index under the index entry "courting" is shown. A user interested in "onesided courting" would access the photographic collection under the Iconclass number "33 C 31" to see depictions of one-sided courting or the systematic arrangement of Iconclass under this number to findother relevant IconClass Numbers
and topics of interest. In Figure 9, a portion of the systematic arrangement of Iconclass is shown beginning with the number "33 C 31 one-sided courting." The Iconclass user might find additional numbers by browsing this systematic portion such as "33 C 31 11 woman in flight-leaving something behind (clothing, shoe, etc.)." The user can then access the photographic collection at hand under relevant numbers or the systematic Iconclass bibliography to find published literature on this topic. Iconclass has been applied to organize iconographical research collections such as the Decimal Index to Art in the Low Countries (DIAL)and the Marburger Index. Iconclass can also be applied to provide a systematic arrangement of print materials. For example, three of the seventeen published volumes of Iconclass are a systematically arranged bibliography. DIAL (Decimal Index to Art in the L o w Countries).D I A L is published by the Department of Art History at the University of Leiden and the Netherlands Institute for the History of Art at The Hague (Rijkbureau voor Kunsthistorische Documentatie 1958). D I A L is a subscription service to over 15,000postcard-sized photographs of Dutch and Flemish art from the fifteenth to the seventeenth centuries
. D I A L subscribers are academic and museum libraries and archives around the world. In Figure 10, a portion of the matrix of subject access (see Figure 2) shows that D I A L describes secondary subject matter using a classification scheme. D I A L photocards bear one or more Iconclass numbers. Photocards also include identifying information such as artist, medium, size, title of artwork, owning museum, etc., but the detail of this identifying information varies from photocard to photocard. Subscribers usually file D I A L photocards by Iconclass number. D I A L users first consult the Iconclass alphabetical index for subjects of
162 LIBRARY TRENDS/FALL 1988
Angel: Lucifer See also Satan AngekFall Isaiah:Prophesy,fallof Lucifer
Figure 5 . Scope note and reference card for a subject heading
Strasbourg:Lib.,Bibl.de la Vilie.
Herradis of Landsberg, Hortus Delic.,fol.
Det. Lucifer nimbed, name inscribed,
#ant,flanked by groups 2 of nimbed angels,
two with pgarled nimbi holding inscribed scroll;
inscriptions: below, 2) &igel,Fall-haif figure
LVe of Archangel Michaelnimbednamein inscription,
1 3 8 flanked by half figures of angels 2,pearled nimbi,
in arc of heaven,all armed with tridents
piercing falling angels 5 including L;ucifer,name in
inscription,wearingdecoratedgarment,g!obe in L.
hand,floreated scepter in R.: inscriptions.'
AngekLucifer. LVe Green,R., Horfus Delic.(1979), p. 90(3,4);fig.5;~1.2. 138 For full bibliog. see Angel:Scene,foL3ro.
Figure 6. Main entry and bibliography cards under a main entry
MARKEYACONOGRAPHICAL RESEARCH COLLECTIONS 163
Strasbourg:Lib.,Bibl.de la Ville.
Herradisof Landsberg, Hortus Delic., fol.
Strasbourg:Lib.,Bibl.de la Ville.
Herradis of Landsberg, Hortus Delic., fol.
Figure 7. Reference cards directing user to main entry interest (seeFigure 8),find one or more relevant Iconclass numbers, then browse systematically-arranged D I A L photocards for artworks depicting secondary subject matter of interest. A D I A L user who has already identified Iconclass number "33 C 22 6 interrupted [lovers']meetingtaken by surprise" would find the photocard whose identifying information is shown in Figure 11 by browsing the photocard file under this iconclass number. Marburger Index. The Marburger Index also applies Iconclass numbers to photographic reproductions of artworks. Figure 12 pinpoints the location of this index in the matrix of subject access (see Figure 2). The Marburger Index describes secondary subject matter through a classification scheme. The Marburger Index was begun in the 1920s by Richard Hamann in a small university department called the Bildarchiv Foto Marburg (BFM) (1985) in the Art History Institute of Philipps University Marburg. With recent funding from the Volkswagen Foundation, BFM produced on microfiche an initial 500,000 photographs of German art from the Bildarchiv and over twenty other archives of German institutions. By 1987, the Marburger Index had approximately 800,000 photographic reproductions on microfiche.
164 LIBRARY TRENDSIFALL 1988 courtesy see also conduct courtesy, manners; etiquette 44 B 15 31 Humanity, Politeness; `Cortesia', `Humanita' (Ripa) 57 A 9' courtier(s) royal household; courtiers, retinue, train see 44 B 152 King Ahasuerus is furious with his wife Vashti and asks his courtiers' advice: Memucan(Muchaeus)suggest to depose the queen 71 Q 21 3 courting see also love unrequited see also proposal see also seducing see also suitor(s) animals courting and mating KEY (t43) to 25 F relations with friendly, non-aggressivecharacter (expressiveconnotations) see KEY ( +91) to 31 A etc. lovers; courting, flirting see 33 C 2 one-sidedcourting 33 C 31 rest during harvest (eating, dancing, making love, etc.) 47 114 9 Figure 8. Alphabetical iconclass index The Marburger Index consists of a monument file of photographic reproductions and multiple indexes to the identifying information and Iconclass numbers assigned to artworks. BFM provides online searching of computerized indexes to approximately 30,000of the total 800,000 images through the STAIRSretrieval system. This online system allows keyword access to Iconclass numbers and captions from the systematic arrangement of Iconclass. Manual searches of the Marburger Index can be performed by consulting the computer-produced microfiche indexes accompanying the monument file. BFM is constantly working to increase the number of images accessible through its online and microfiche indexes. The microfiche indexes provide access to much more information than could be provided through card or book indexes. Figure 13shows a portion of the primary iconography microfiche index under Iconclass number "33 C 21 courting." Marburger Index users must first identify this Iconclass number by consulting the alphabetical index and/or systematic arrangement of Iconclass. In this primary iconography index, brief empirical information is provided for photographic reproductions. Once users find relevant empirical information, they can consult the monument file to see images of interest. In the monument file, images are organized on microfiche alphabetically by city, city view, and building type, and each image is accompanied by brief identifying information (see Figure 14). Besides the primary iconography index, the Marburger Index provides subject indexes on microfiche for: (1) secondary iconography (by Iconclass number), (2)alphabetical por-
MARKEYACONOGRAPHICAL RESEARCH COLLECTIONS 165
Iconclass Number Description
33 C 31 31 1 31 11 31 2 31 21 31 3 31 4 33 C 32 32 1 32 11 32 2 32 21 32 3 32 31 33 c 33 33 1 33 2 33 3 33 c 34 34 1
one-sided courting seduction33 C 21 9 woman in flight lover chased away 33 C 22 9 leavingsomething behind(clothing, shoe, etc.) lover deserting woman deserted woman man accused by woman he has cast off man in flight pursuitof a woman, prowling-CC-pursuit of a man
rivalry of men for the love of women; trying to get the skirt; `Rivalita' (Ripa)
rivalryof women for the love of men; fight for the hose; `andouille' shaking lovers out of a tree, with cudgel shaking lovers out of a tree, by throwing things Love's labours lost (a third person interfering successfully) the succeedingthird
shifting out the candidatesby men -CC-by women
blindfoldedchoice choice betweenwealth and youth
marriage lottery computerized dating
Figure 9. Systematic arrangement (schedules)of Iconclass. Source: Van de Waal, Henri. Iconclass: A n Iconographic classification system
. Amsterdam: North-Holland, 19741985. trait catalog, (3) special catalog for the iconography of art and science, and (4) special catalog for Biblical iconography. Common Characteristics of Iconographical Research Collections. In Figure 15, relevant portions of the matrix of subject access are highlighted for iconographical research collections. These collections share the following- characteristics: They were begun by and are currently administered by art history departments and art historians, respectively. They serve art historians and scholars from related academic disciplines
. They describe secondary subject matter through subject headings or classification schemes
. Automation has been applied to few iconographical research collections.
166 LIBRARY TRENDSIFALL 1988
Figure 20.Pinpointing DIAL'Sposition in the matrix
92 B 3 ( + 13 Venus): 92 B 4 : 33 C 22.6
H.Goltzius engraving Hollstein V111,137~
Leyden, Printroorn photo NetherLArt 1nst.L nr.23917
Mars and Venus surprised
Figure 1 2 . Descriptive information on DIAL photocards 5. Automation of iconographical research collections has introduced keyword indexing of secondary subject matter. 6. Primary subject matter has not yet been described.
MARKEY/ICONOGRAPHICAL RESEARCH COLLECTIONS 167 Primary Subject Matter Users of two visual resources collections, the Repository of Stolen Art (RoSA) and The Historic New Orleans Collections (THNOC),can access artworks for their primary subject matter. These collections contain images that would be found in iconographical research collections. However, subject access approaches in THNOC and RoSA have been developed with the knowledge that users will not necessarily have special training in Christian iconography, theology, or art history. Repository of Stolen Art (RoSA). The Repository of Stolen Art is curently under development by the Royal Canadian Mounted Police (RCMP),Interpol Branch, Ottawa, Ontario. The objective of RoSA is to promote electronic exchange of information on stolen cultural property in support of RCMP's commitment to suppress theft, receiving, and trafficking in this property. This property includes objects from soils or waters, ethnographic art, military objects, decorative art, fine art, scientific and technical objects, books, records, photographs, and sound recordings
(Vance 1984, pp. 377-78). Figure 16 highlights relevant portions of the matrix of subject access for RoSA. Generally, phrase and keyword indexing is applied to keywords and phrases which represent the primary subject matter of artworks. RCMP collects and enters data into the Canadian Heritage Information Network (CHIN), the government-supported and centralized online catalog
ing service of the National Museum
s of Canada. RoSA is searched by detectives, police, and other local, provincial, national, and international law enforcement
agencies for identification and recovery of works. RoSA is also searched for graphing networks and patterns of receiving and disposing of works. Online interactive searching is provided by the keyword, Boolean-based PARIS (Pictorial Artifact Retrieval Information System). Individual RoSA records are built by answers to a yes-no questionnaire about: (1) empirical information-e.g., size and shape of lost object, function, materials, inscriptions, patterns; and (2) primary subject matter. An example of questions from the RoSA questionnaire is given in Figure 17. The Historic New Orleans Collection (THNOC). The Historic New Orleans Collection documents the history of New Orleans through approximately 150,000photographs, prints, and paintings. The goal of a project to automate access to THNOC is "to provide subject access to a broad spectrum of users with diverse interests including specialists, library scientists, and iconographers as well as students, the general public, and scholars in other fields" (Sarasan 1984a, pp. 317-18).To accomplish this goal, subject access to primary and secondary subject matter will be provided through pre-iconographical descriptions and
168 LIBRARY TRENDUFA1.L 1988
1. Classification Scheme
/ B. Phrasi indexing (Tanual and computer files) ,` 1. Classification Scheme
Figure ZZ. Pinpointing the Marburger index in the matrix iconographical analyses, respectively. A thesaurus to preiconographical descriptions will be developed, and Iconclass will serve as the controlled vocabulary for iconographical analyses. Other subject access approaches to THNOC will mix primary and secondary subject matter in data fields for proper names, places, and events depicted in images, and in controlled vocabulary terms added to the database by THNOC users. Figure 18shows relevant portions of the matrix of subject access for THNOC. MINISIS is a relational database management system that will be installed on a Hewlett-Packard minicomputer to offer THNOC users state-of-the-art retrieval capabilities such as free-text Booleanbased searching, range searching, truncation, and access to a multilingual online thesaurus (Sarasan 1984b, pp. 387-406). BARRIERTSO ACCESSOF ICONOGRAPHICAL RESEARCCHOLLECTIONS Iconographical research collections pose certain barriers to users who need to access the artworks in these collections. Since secondary subject matter is described in iconographical research collections, users must be sufficiently experienced with iconographical analysis to perform a successful search of the collection at hand. "Providing access to collections by primary subject matter can ease the task of users since it requires only one level of interpretation, practical experience, and knowledge of the history of style" (Markey 1986,p. 7). The recent design and development of ROSA and T H N O C may be indicative of a shift from secondary to primary subject matter to serve users who are not experienced with iconographical analysis and to take advantage of the increased number of subject access points that can be easily accommodated by computer technology. Subject access to iconographical research collections varies from one collection to the next. Differences between systems with regard to
33 c 21
10183 Bild: Blumenmadchen und soldat
Backer?,Jacob Adriaensz (1623-1 651 Niederlande)(Maler)
5' Y I. 8 3
Malerei, Tafel-Malerei Leinwand, 117 x 85
Standort . . . . : Pommersfelden Verwaltung . . : (halb-offentliche) Sammlung, Schloss Weissenstein,
Iconclass. . . . : 46 BB 22 1 (FLOWER-GIRL)& 33 C 21
Y 5' R
Abbildungen . : Marburger Index: MA1,02677,A,14 Witt Library WITI10936,E,6
20401 Bild: Die Liebe auf dem Lande
Watteau, Antoine (1698-1 721 Frankreich) (Maler)
Malerei, Tafel-Malerei 01,Leinwand, 56,7 x 80,5 Standort . . . . : Berlin (West)
Verwaltung . . : (offentliche) Sammlung, Staatliche Schlosser und Garten, Verwaltung, Schloss Charlottenburg, Inventar-Nr.GK 5337 Iconclass. . . . : 33 C 21 & 43AA 2
8 E bn
Abbildungen . : Marburger Index: MA1,00166,F,05-00166,F,O8
170 LIBRARY TRENDSIFALL 1988
X 192202 Berlin,Schijsseru. Garten
Jean Antoine Watteau, 1684-1721. Die Leibeauf dem Lande(L'amourpaisible) 0 1 a.Lw. 56,7x80,5 cm. Gen.Kat.5337.Seidel148.
Potsdam, Neues Paiais, ruletzt Charlottenburg. (Aufn. 1945/46).
Figure 14. Descriptive information on microfiche images in the Marburger index Type of
Subject Principal Methods of Subject Access: 1. Classificationschemes 2. Subject headings: a. Alphabetical list b. Alphabeticoclassed list c. Classified list
Phrase Keyword Keyword
indexing in context indexing
Figure 15. Pinpointing iconographical research collections in the matrix 2. The secondary subject matter of existing iconographical research collections is somewhat limited to the style(s), historical period(s),
MARKEY/ICONOGRAPHICAL RESEARCH COLLECTIONS 171 and/or geographical area(s) of the original collection. For example, Iconclass is closely associated with Dutch art and the Index of Christian Art is associated with Western Christian art from 400 to 1400A.D. 3. Subject heading lists have proliferated since the mid-1970s. Examples of lists published in this time period are the subject headings list used at the Centre National de la Recherche Scientifique
, the classified subject headings used at the Yale Center for British Art, the classified subject heading list of the National Museum of American Art, and the subject headings for Canadian iconography used at the Public Archives of Canada (Garnier 1984; Yale Center for British Art 1979; National Museum of American Art 1983; Castonguay 1981, pp. 269-80). Identifying information about artworks varies from one iconographical research collection to the next. Every collection provides identifying information but there the similarity ends. Differences between systems with regard to such information will require users to search for a known artwork under variant forms of artists' names, institution names, dates, etc. The publication of cataloging rules for describing graphic materials, Graphic Materials: Rules for Describing Original Items and Historical Collections (Betz 1982),and the availability of a machine-readable cataloging (MARC) format for representing the identifying information of graphic materials may foster consistency among collections (Online Computer Library Center 1986).The Art & Architecture Thesaurus (AAT)will eventually make a controlled vocabulary available for expressing style, geography, materials, and techniques connected with artworks (Barnett 1986, p. 135). Efforts to establish a centralized cataloging system for the visual arts have failed or been disbanded-e.g., Museum computer network
, Museum Prototype Project. Two results of failed efforts are the creation of separate databases within a single centralized system and the proliferation of separate local databases. Also, these databases vary from other databases with regard to the identifying information and subject cataloging assigned to each entry. FUTUREDIRECTIONS Separate collections and specialized databases will proliferate as long as there is no single organization responsible for establishing and maintaining rules, guidelines, and formats for identifying information and subject cataloging. As long as financial support and governance comes from art history departments, programs, and institutions, access to collections will be provided for scholars through secondary subject matter. If a centralized system is successfully established, the system will have to take advantage of data and functions that can be shared among institutions supporting iconographical research collections and other artwork collections. Identifying information for artworks and authority
172 LIBRARY TRENDSIFALL 1988
6.F h k e indexing
,#' , ,,
Figure 16. Pinpointing ROSA in the matrix Check all descriptions that seem true, even if they are not ronsistent. A. Shape B. Surface C. Human faces and figures
C1 Entire human figure C2 Portrait head or bust, partial figure C3 Torso shown without head C4 Other incomplete figure (e.g., sketch of eye, hand) e. D. Individual and types D1 Christ as child D2 Christ as adult D3 Christ with cross, crucifixion eee D19 Military person, warrior, soldier, etc. D20 Musician, singer, dancer, actor D21 Wearing plume or feathers e. E. Activities E l Procession, parade, caravan E2 Eating, drinking E3 Smoking em. F. Animals G. Setting H. Detail Figure 17. ROSA Questionnaire data for iconography, artist and school names, styles and periods, geography, building names, and institution names are data that can be shared among iconographical research collections, museums, art slide and picture collections, and academic institutions. The model of sharing data among institutions will be the reverse of the library automation model-i.e., institutions will add entries first to their own local systems and then add these entries to the centralized system. As long as there is variation among collections with regard to identifying information and subject cataloging, the burden of access is
MARKEY/ICONOGRAPHICAL RESEARCH COLLECTIONS 173 indexing
V Figure 18. Pinpointing THNOC in the matrix placed on users of iconographical research collections. At present, users must become proficient in the use of iconographical research collections that employ different subject vocabularies. It can be expected that when iconographical research collections are transformed from manual, card-based files to online databases, users will have to become familiar with many different online retrieval system interfaces and access techniques, especially in view of the proliferation of separate, specialized databases. References Ackerman
, James S., and Carpenter, Rhys. 1963. Art and Archeology. Englewood Cliffs, NJ: Prentice-Hall. Barnett, Pat. 1986. "AAT Ilpdate." Art Documentation 5 (Fall), pp. 134-35. Betz, Elisabeth W. 1982. Graphic Materials: Rules for Describing Original Items and Historical Collections. Washington, DC
: Library of Congress. Bildarchiv Foto Marburg, ed. 1985.Marburgerlndex: Inventory of Art i n Germany: User's Manual, 2nd ed. Munich: K.G. Saur. Castonguay, Denis. 1981. "Approche Sujet en Iconographie Canadienne a l'aide d'un Lexique de Termes Contr6li.s." In Computerized Inventory Standards for Works of Art: Proceedings: Collection of Papers, edited by Raymond Vkzina, pp. 269-80. Montreal, Canada: Fides. Couprie, Leendert D. 1983. "Iconclass: An Iconographic Classification System." Art Libraries Journal 8 (Summer), pp. 32-49. Gamier, FranGois. 1984. Thesaurus Iconographique: Systkme descriptif des representations. Paris: Le Lkopard d'Or. Knox, Katharine McCook. 1979. T h e Story of the Frick Art Reference Library: T h e Early Years. New York: The Frirk Art Reference Library. Lindsay, Kenneth C. 1968. "Computer Input Form for Art Works: Problemsand Possibilities." In Computers and Their Potential Applications in Museums (A Conference Sponsored by the Metropolitan Museum of Art
), pp. 19-35. New York: Arno Press. Markey, Karen. 1986.Subject Access to VisualResources Collections: A Model for Computer Construction of Thematic Catalogs. New York:Greenwood Press. Narkiss, Bezalel, and Sed-Rajni, Gabrielle (1976).Zndex of Jewish Art: Iconographical Index of Hebrew Illuminated Manuscripts. Jerusalem: Israel Academy of Science and Humanities. National Museum of American Art (NMAA), Office of Research Support. 1983. Subject Term Guide. Washington, DC: NMAA, Office of Research Support.
174 LIBRARY TRENDSIFALL 1988 Online Computer Library Centrr. 1986.Audiovisual Media Format, Znded. Dublin, O H : OCLC. Panofsky, Erwin. 1962. Studies in the Visual Arts, reprint. New York: Harper Fc Row. Rijkbureau voor kunsthistorische Documentatie. 1958. Decimal Index to Art of the Low Countries (DIAL). The Hague: Rijksbureau voor kunsthistorische Documentatie. Sarasan, Lenore. 1984a. "An Integrated, Automated System to Provide Subject Access to Large Image Archives Across a Broad User Base." In Census: Computerization in the History of Art, vol. 1, edited by Laura Corti, pp. 317-18. Florence, Italy: Regione Toscana. Sarasan, Lenore. 198413. "Visual Content Access:An Approach to the Automatic Retrieval of Visual Information." In Automatic Processingof Art History Data and Documents: Papers, vol. 1, edited by Laura Cmrti, pp. 387-406. Florence, Italy: Regione Toscana. Vance, David. 1984. "Identification of Objects." In Census: Computerization in the History of Art, vol. 2, edited by Laura Corti, pp. 337-51. Florence, Italy: Regione Toscana. Van de Waal, Henri. 1974-1985. Iconclass: A n Iconographic Classification System, 17 vols. Amsterdam: North-Holland. Woodruff, Helen. 1942. T h e Index ofChrzstian Art at Princeton University:A Handbook. Princeton, NJ
: Princeton University Press. Yale Center for British Art, Photographic Archive. 1979. Subject Authority Used by the Photographic Archives. New Haven, C T : Yale Center for British Art.