Introducing the Online Catalogue of the Duke Papyrus Collection [1]

by Peter van Minnen

In this article I would like to present the nascent online catalogue of the papyrus collection at Duke University. Now that a first batch of records is already available online, it seems important to inform the papyrological public at large of the existence of the catalogue and to explain its rationale.

The Duke papyrus collection consists of about 1,200 discrete papyri, most of which are rather fragmentary. They were collected over a period of more than fifty years, from the first handful of texts bought in 1942 until the most recent gift from the private collection of emeritus Professor W.H. Willis, who has been instrumental in building up the bulk of the collection. Major purchases of papyri occurred from the late 60's (e.g. the largely unpublished Ammon archive from Panopolis)[2] through the 80's (e.g. the papyrus collection of the University of Mississippi).[3] The majority of the papyri are documentary, but there are some notable literary texts (e.g. the so-called "Comoedia Dukiana" acquired in 1984).[4] Not surprisingly, literary texts comprise less than 10% of the collection. Greek texts predominate but there are also a number of Coptic texts. These comprise about 20% of the collection and there are some notable texts among them (e.g. the unpublished martyrdom of Stephanos of Lenaeis). Then there are small pockets of Hieratic, Demotic, Latin and Arabic texts. Many Greek and Demotic papyri were extracted from mummy cartonnage. Less than 5% of the papyri in the collection have been published so far,[5] but one may expect that the work on the catalogue (among other things) will have positive effects on the rate of publication. In fact, the catalogue itself is a form of publication.

When I started cataloguing the collection in 1992, my task consisted of more than just drawing up records for the online catalogue. Most papyri were not yet properly conserved when I came to Duke, and I had to do that first.[6] I also made preliminary transcripts of most texts. Exact measurements and other physical properties of the papyri have been entered in a FilemakerPro database on a Macintosh along with information about acquisition and conservation. This database also records information about content and-if applicable-publication of the texts. The records in this database, which now contains information about approximately two-thirds of the collection, are then turned into online library catalogue records according to the US MARC (Machine-Readable Cataloging) format, the international standard for all automated library cataloging. More particularly, the Archival and Manuscript Control (AMC) format, a subset of the US MARC format, is used. While this procedure subjects ancient manuscripts to rules that were originally devised with more recent manuscripts in mind, the advantages far outweigh the disadvantages. The AMC format has been universally accepted for online cataloguing of archives and manuscripts. The format provides a clear structure for entering the data. What one puts in a record follows the widely-used Anglo-American Cataloguing Rules (AACR 2)[7] and, more particularly, Archives, Personal Papers and Manuscripts (APPM),[8] which customizes these rules for describing archives and manuscripts. This also dictates the use of standard reference tools such as the online Library of Congress Name Authority File (LCNAF), the Library of Congress Subject Headings (LCSH)[9] and the Art and Architecture Thesaurus (AAT),[10] and it therefore presents an opportunity to bring a measure of uniformity to cataloguing papyri.

The most promising aspect of using the AMC format is indeed that it provides a uniform and standardized approach for cataloguing papyrus collections. If the format is applied consistently, the records of one collection can go online along with the records of another collection without further ado. This opens up the possibility of an online union catalogue of all papyrus collections. Because both published and unpublished papyri will have to go in such a catalogue, the eventual product could be very useful as a searching tool for all those who are interested in papyri. Not only those who want to publish new texts, but also those who are looking for certain genres or topics in papyri would want to use this tool, which seems a natural complement to the Duke Data Bank of Documentary Papyri, at least as far as Greek and Latin documentary papyri are concerned. The DDBDP effectively replaces scores of indexes of Greek and Latin words and makes searching for such words, especially combinations of them, easy and attractive. If catalogues of papyrus collections are made according to the AMC format, an online union catalogue would provide a sophisticated subject index as well as a "General-Register" of papyri by genre. There are very few papyrus editions with subject indexes and the only comprehensive "General-Register" for documentary papyri in existence[11] does not include all that has been published since 1901.[12] Because papyrus collections usually contain literary and subliterary texts in addition to documentary texts and because languages other than Greek and Latin are represented as well, an online union catalogue could help break down the barriers between the various "papyrologies" in existence.

Another approach to online papyrus catalogues is also possible, but less attractive. An institution holding a considerable number of papyri can load a machine-readable version of its inventory or other finding aid (typically an ASCII file prepared with standard word processing software) onto a networked computer or server. Using appropriate software, scholars in other institutions can then locate that server and search the inventory using standard searching routines. At present only the inventory of the Yale papyrus collection is accessible this way. Now I assume that at some time in the future most libraries and institutions holding a good number of papyri will want to put their catalogues online somehow. If every institution follows Yale's example and devises its own format and therefore its own unique database, future papyrologists will have to browse through 50 or 100 idiosyncratic online catalogues to do a thorough job in searching the available records. If, however, the standard format currently applied at Duke is used a single search could suffice. The AMC format has been adopted by the technology committee of the American Society of Papyrologists as the preferred format for catalogue records to be included in the Advanced Papyrological Information System (APIS),[13] which is to provide an online union catalogue of papyri, both published and unpublished.

The Duke records will show passive "users" how the AMC format works for papyri. This should encourage them to become active "users" and to create similar records for the catalogues of the papyrus collections at their libraries and institutions. The "dirty" work of finding appropriate titles and subject headings for a vast range of Duke papyri is done from scratch. For other collections it will suffice to follow Duke's example for all those papyri that are similar to the ones at Duke. This will also facilitate the use of English by non-native cataloguers. In an international discipline like papyrology one cannot afford to be provincial.[14] Work from scratch will only be necessary when a certain type of text is not represented among the Duke papyri.

Because we do not want to wait until APIS is up and running, the Duke records are currently put online through the Online Computer Library Center (OCLC) and through the local Online Public Access Catalogue (OPAC) of the Duke University Libraries, which is accessible on the Internet. In the last-mentioned case the records will be merged with more than 4 million other records of books and manuscripts held exclusively by the Duke University Libraries. In the case of OCLC they will be merged with more than 24 million other records of books and other items held in libraries worldwide (the so-called Online Union Catalogue or OLUC; more than 10,000 libraries subscribe to it). In theory it works as follows for outside "users." They can access the records in OCLC using FirstSearch or PRISM, programs that allow a variety of searches. At present these programs do not allow random searches of all the information contained in the records, but one can search the note field for "P.Duk.inv." or "papyrus" in combination with the subject (e.g. land tenure) or work title (e.g. loan) one is searching for.[15] The catalogue of the Duke University Libraries does allow random searches in the keyword search mode. Outside "users" can access this catalogue either through gopher by selecting Duke Universities Libraries under Triangle Research Libraries in North Carolina or by entering "" or "" at the telnet prompt and "library" at the user prompt on the terminal they are using. The program running the catalogue of the Duke University Libraries was created by Data Research Associates (DRA) and is very user-friendly. In the keyword search mode one can search the papyrus records for just about anything in combination with "P.Duk.inv." In the regular search mode there are a few constraints, but the display of the results is more helpful than in the keyword search mode. In the keyword search mode one has to "set display long" to view the full records and not just summaries.For the time being searching through the catalogue of the Duke University Libraries will probably be more convenient than through OCLC. The search tools for APIS will be even more sophisticated.

Using LCSH in creating the catalogue records raises a problem. According to the rules of the game one is forced to use existing subject headings for the catalogue records. Rather than having the papyri determine what they are all about they have to fit an existing range of subjects. This might raise some eyebrows among papyrologists, but given the broad range of subjects available-LCSH contains over 5,000 pages of subjects-this should not be a problem at all. Rather choosing between two or more appropriate subject headings can sometimes lead to problems. After all one has to try to make choices consistently. Also the papyri are not really "about" the subjects assigned to them in the way secondary literature is. They only somehow relate to these topics. This minor inconsistency has to be taken for granted when one is dealing with manuscripts rather than printed books. Otherwise one could hardly come up with subject headings at all. And the rule is that the more precise subject headings one adds, the more useful the records become for "users" interested in particular subjects.

I will now give an example of a complete record[16] and explain the steps taken in producing it after each line of information. I show the version with the "tagging" (the numbers and letters identifying the various pieces of information) prescribed by the AMC format. With some of the programs used to call up the records the "tagging" will not appear on the screen. Instead more user-friendly captions (like "subjects" or "notes") are provided. In creating records for another collection "users" can always retrieve the full information (a) from the catalogue of the Duke University Libraries by typing "m" (for MARC record) when viewing a record in the regular search mode or (b) from OCLC by using PRISM instead of FirstSearch.

040 NDD $e appm $c NDD

The 040 field identifies the library holding the item (NDD=Duke University Libraries), the source of the AMC format (APPM=see n. 8), and the library cataloguing the item.

046 q $b 176 $d 165

The 046 field indicates a B.C. date, in this case a range that b(egins) in 176 and (en)d(s) in 165 B.C. The date is q(uestionable) because it can be either date. Note that e.g. the second century B.C. would be "q $b 199 $d 100" and the second century A.D. "q $b 100 $d 199."

049 NDDB $d [ $v = items ] $v 1

The 049 field identifies the Special Collections Library within the Duke University Libraries system as well as the number of items described in the record (here always=1). For papyrologists it seems most useful if all the items are described separately in some detail. One could conceivably put all the papyri in one big box and describe them in one very general record, but there is no point in doing that.

099 P.Duk.inv. 314

The 099 field contains the local call number, in the case of a papyrus collection the inventory number. Because this number will only show up in the catalogue of the Duke University Libraries, it is repeated in one of the 500 fields for the convenience of those who access the records in OCLC. If there are two unrelated texts on the same papyrus one has to make two records and distinguish the texts by adding (a), (b) or, if the texts are on recto and verso, R, V to the inventory number. Otherwise the information would get confused.

100 Theambesis $c (Daughter of Onnophris from Oxyrhyncha, Egypt), $d 2nd cent. B.C.

The 100 field identifies the author (=auctor intellectualis, not the scribe) of the text, in this case a woman named Theambesis. To create an "authority" for a woman from the ancient world not included in LCNAF is the documentary papyrologist's sweet revenge on those who know only of Sappho. It was not really necessary to add the c-subfield with additional information on Theambesis, because she is the only one in the database with that name. But it seemed helpful, especially in case another Theambesis shows up in another text. The d-subfield gives a broad indication of Theambesis' d(ate) rather than the exact date when she fl(ourished), either 176 or 165 B.C., which we cannot decide at present.

245 Petition, $f Year 6, Thoth 22 [176 B.C., October 27 or 165 B.C., October 24]

The 245 field gives the title of the "work" concerned (in this case a statement of the nature of the text) and an indication of the time at which it was written. The conversion of the date is given in [ ] because it is not unequivocal. For a provisional list of titles for papyrus documents see the list of titles.

300 1 $f item: $b papyrus, mounted in glass; $c 32 x 13 cm.

The 300 field gives standardized information about the physical properties of the item concerned. Dimensions are given as Height x Width. Width is left out if 0.5 Height < Width < Height.

500 Actual dimensions of item are 31.6 x 13.0 cm.

The 500 field has fewer restrictions in the way it is formulated and more detailed information about physical properties of the item can be given. Some of it could be helpful in identifying pieces of the same papyrus in other collections and it seemed best not to be too stingy with information of this sort.

500 44 lines.

Text on the verso is included in the overall line numbering of the piece.

500 Written along the fibres on the recto with the reply in a different hand; written along the fibres on the verso in different hands, each line inverse to the other with the address in large letters.
500 Upper margin of 3 cm.; lower margin of .5 cm.; left margin of 2 cm.; small right margin.
500 Extracted from mummy cartonnage with P.Duk.inv. 313 and P.Duk.inv. 315-323.
500 P.Duk.inv. 314 was formerly P.Duk.inv. MF84 2.

The temporary inventory numbers are being replaced with permanent ones.

The 510 field contains information about publication or citation of the item, but there is nothing to report in this case.

520 Papyrus petition from Theambesis, also known from P.Duk.inv. 316, wine seller and daughter of Onnophris from Oxyrhyncha in the Arsinoites (modern name: Fayyum), Egypt, to a strategos and hipparches, a high civil and army official in the region. Theambesis repeats her complaint that Heliodoros failed to repay a loan, and she asks the official again to tell Chairephanes, the epistates or local police officer, to get Heliodoros. In a postscript the official orders the police officer to get both Heliodoros and Theambesis. Dated to Thoth 22, year 6 (October 27, 176 B.C. or October 24, 165 B.C.).

The 520 field gives a brief description of the content of the text. Transliterations of ancient geographical and personal names stay as close to the original as possible. Thus one will find Ioannes, not John. Things Roman will clearly be marked as such (praefectus). In case of doubt alternatives will be entered. In the subject headings listed below it is not always possible to transliterate because of the existing rules. Arsinoites e.g. would appear there as Fayyum.

530 Item available as digitized image. Consult repository for details.

Digitized images of the papyri are made with a colour scanner. Presently 200 of these images are available online.

541 Purchased in 1984.

The 541 field indicates the source of acquisition and the date of entry in the collection.

545 Theambesis spelled her name as Thaambesis on another papyrus, P.Duk.inv. 316.

The 545 field contains biographical or historical information. Among other things it indicates how the name in the text relates to the "authority" established for it. In P.Duk.inv. 316 Theambesis' name is spelled Thaambesis. The 100 field, however, reads Theambesis as here, because that is the "authorized" form.

546 In Greek.

The 546 field indicates the language of the item.

555 Descriptive database available in repository.

The 555 field refers to the FilemakerPro database created by the undersigned with a more detailed description of the item.

600 Heliodoros $c (From Oxyrhyncha, Egypt), $d 2nd cent. B.C.
600 Chairephanes $c (Epistates from Oxyrhyncha, Egypt), $d 2nd cent. B.C.

The 600 field contains names of persons connected with the text but not its author.

650 Grievance arbitration $z Egypt $y 332-30 B.C.
650 Loans, Personal $z Egypt $z Oxyrhyncha (Extinct city) $y 332-30 B.C.
650 Loans $x Law and legislation $z Egypt $y 332-30 B.C.
650 Strategi, Greek $y 332-30 B.C.
650 Egypt $x Officials and employees $z Fayyum $y 332-30 B.C.
650 Women sales personnel $z Egypt $z Oxyrhyncha (Extinct city) $y 332-30 B.C.
650 Wine and wine making $z Egypt $z Oxyrhyncha (Extinct city) $y 332-30 B.C.
650 Police $z Egypt $z Oxyrhyncha (Extinct city) $y 332-30 B.C.
650 Complaints (Civil procedure) $z Egypt $y 332-30 B.C.

The 650 field contains the subject headings derived from the LCSH. They provide a variety of access points to the record. The addition of (Extinct city) to Oxyrhyncha is according to LCSH, which does not appear to know extinct villages. It saves us from calling Karanis a village, I guess. The broad time frame according to the y-subfield is the Ptolemaic period. The existing heading Strategi, Greek does not permit a geographical delimitation.

655 Documentary papyri $z Egypt $z Oxyrhyncha (Extinct city) $y 332-30 B.C. $2 aat.
655 Petitions $z Egypt $z Oxyrhyncha (Extinct city) $y 332-30 B.C. $2 aat.

The 655 field indicates the genre of the text. The 2-subfield indicates that the genre headings are taken from AAT (see n. 10). For a list see the list of genres. The genre headings are broader than the titles used for individual texts in the 245 field.

The 700 field contains an added entry for the (possible) author of a text. Here it has no use, but e.g. in the case of the so-called "Comoedia Dukiana" Archippus would find a place in this field.

This example will help potential "users" understand the value of having ready access to records like this, and especially of having ready access to records like this for all major collections of papyri. Some of the information in the "prototype" will need to be updated in a couple of years. This can best be done once the records are integrated in APIS.

If some of the acronyms and the electronic gibberish in this article appear daunting at first, readers can take comfort in the knowledge that it took me several months to understand what was going on. As always, I could not wait to tell them what I had learned myself.

(from Bulletin of the American Society of Papyrologists 31 (1994): 159-170; (c) American Society of Papyrologists, reprinted with permission)

Return to the papyrus home page

Return to Duke University Special Collections Library home page

Please see our page with contact information if you have any comments or questions about the Duke Papyrus Archive.

Last updated by Peter van Minnen and Suzanne Corr on 5/10/95