ACHAEMENID ROYAL INSCRIPTIONS FROM PERSEPOLIS IN ELECTRONIC FORM
By Matthew W. Stolper, Professor of Assyriology in the Department of Near Eastern Languages and Civilizations, and
Gene Gragg, Professor of Near Eastern Languages and Director of the Oriental Institute
The University of Chicago
(This article originally appeared in The Oriental Institute News and Notes, No. 157, Spring 1998, and is made available electronically with the permission of the editor.)
From 550 BC on, Cyrus the Great and his successors, the Persian kings of the Achaemenid dynasty, conquered and held an empire on a scale that was without precedent in earlier Near Eastern history, and without parallel until the formation of the Roman Empire. At its greatest extent, its corners were in Libya and Ethiopia, Thrace and Macedonia, Afghanistan and Central Asia, and the Punjab. It incorporated ancient literate societies in Elam, Mesopotamia, Egypt, and elsewhere. It engaged the emerging Greek states in a long confrontation that had profound effects on Greek and later European historical consciousness. It lasted without substantial loss of control until it was conquered by Alexander the Great, and then dismantled by his successors after 330 BC.
Unlike the Assyrian conquerors before them, the Achaemenid kings devoted little space in their royal inscriptions to the struggles that forged the empire. It was not for want of literary ability or for want of an understanding of the force that royal propaganda could have. In fact, the longest Achaemenid royal inscription, the apologia of Darius the Great (DB), carved in three languages to accompany a relief on the cliff face at Bisitun (near Kermanshah, in western Iran), is an elegantly constructed narrative of Darius's triumph over his competitors for control of the empire, and it was not only addressed to posterity, but also translated and disseminated to the conquered lands, and an Aramaic version was copied out in Egypt about a century after the text was composed. But the other inscriptions, for the most part, present the empire not as an accomplishment, the result of royal efforts, but as a divinely sanctioned order.
The characteristic that best represented this order was its diversity, which was in turn an expression of the empire's size. In a world where few people had seen maps, where area and distance could not be expressed in geometric figures, the most impressive way of putting them in words was to name the many people who served the king, and so some of the inscriptions give lists of subject lands, twenty, or twenty-three, or twenty-eight items long. This characteristic was condensed in the Old Iranian words vispazana, "(with) all kinds (of people)," or paruzana, "(with) many kinds (of people)," which the Achaemenid kings' translators rendered with Akkadian phrases indicating, literally, "of many languages" or "of all languages."
This diversity is a characteristic embodied in the Achaemenid royal inscriptions themselves. They are mostly cast in three unrelated languages: Old Persian, the language of the rulers; Elamite, the pre-Iranian language of the part of Iran that the Persians had made their homeland; and Akkadian, the language of the Assyrians and Babylonians who had ruled much of western Asia and parts of western Iran before the Persian conquests. Monolingual inscriptions are sometimes coupled with related inscriptions in the other languages. And when trilingual versions are incorporated into architectural elements they also represent the political order. The Old Persian version always has pride of place, on the top in a vertical arrangement or in the middle of a bilateral arrangement.
Many of the Achaemenid inscriptions were incorporated into the foundations or the architectural ornaments of palaces built for rulers. The court moved often from one royal complex to another. In the classical and biblical accounts of the Achaemenid court, most episodes take place in the palaces of Susa, and many of the Achaemenid inscriptions were found there by some of the earliest systematic excavations in the Near East. But long before that, European travelers had brought back descriptions of the extensive standing remains at Persepolis, and accurate copies of the Achaemenid multilingual inscriptions from Persepolis were the basis for the first steps in deciphering the cuneiform scripts. Georg Grotefend's early successes with the Old Persian script relied on a comparison between the texts of Darius I and Xerxes from Persepolis that are now called DPa and XPe. The near completion of the Old Persian decipherment and the decisive work on the Babylonian script done by Edward Hincks and Henry Rawlinson relied above all on the great inscription of Darius I at Bisitun (DB).
The decipherment opened the way to exploring pre-Achaemenid cultures and societies of ancient western Asia that had been entirely unknown or only dimly understood, and explorers of these areas soon left the Achaemenid texts behind as the domain of subspecialists in Iranian or Elamite philology, or in the history of the encounters between Persians and Greeks. These texts, terse as their contents often seem, are the very words of the Great Kings, but the fact that the Great Kings themselves produced them in more than one language is often ignored or merely acknowledged in passing. The inscriptions are often cited, for example, from handy editions of the Old Persian version without reference to the other versions, and at other times continuous translations of one version are interrupted by remarks on additions or omissions or divergences in the other versions.
Close study and accurate use of these texts, and even simple appreciation of their intended effect, calls for a synoptic presentation of the versions. No handy synoptic edition has replaced F. H. Weissbach's magisterial Keilinschriften der Achämeniden of 1911, in part, at least, because the development and divergence of scholarship on Old Persian and Old Iranian, Elamite, and Akkadian have made replacing it with an equally compendious and authoritative printed edition a forbidding task. Even Weissbach's edition presented only the verbal components of the texts: parallel transliterations and translations of the versions, with textual apparatus. Readers who wanted to see what the inscriptions looked like in their original setting still had to search for drawings and photographs in publications by early visitors to the sites. When modern excavation reports - above all, the Oriental Institute's final publication of the excavations of 1931-1939 at Persepolis and nearby Naqsh-i Rustam, in three volumes produced between 1953 and 1970 - made good photographs and facsimiles of many of the inscriptions and their contexts available, looking at the textual, architectural, and archaeological aspects of these documents at the same time still required a lot of table space in a very good library.
With the development of computers and then of the Internet, the possibilities of on-line presentation of linked texts and pictures offer an opportunity for a different sort of synoptic presentation of texts, including these inscriptions. The Achaemenid Royal Inscriptions project (ARI) is an exploration of these possibilities, with the creation of an electronic study edition of Achaemenid inscriptions in all their versions, accompanied by translations, glossaries, basic text-critical and reference apparatus, pictures of the original inscribed buildings and objects, and site plans to indicate their original locations, beginning with the inscriptions from Persepolis and Naqå-i Rustam.
The project began with a grant from the Provost's Program for Academic Technology Innovation in 1996. Matthew W. Stolper and Michael Kozuh (graduate student in NELC, specializing in Assyriology and Achaemenid History) prepared most of the data: transliterations and text-critical apparatuses in uniform formats, working translations of each version of each text, glossary entries, and so on. Gene Gragg wrote most of the programs that index, search, link and display the data; in this task the on-line Bible Browser of Richard Goerwitz (Ph.D. in NELC, 1993) was a model vehicle for the display of groups of texts in several simultaneous versions. Charles Jones (Oriental Institute Research Archivist) and John Sanders (Head of the Oriental Institute Computer Laboratory) have contributed to all tasks.
Most of the project's work has been done on Hewlett-Packard X-Terminals donated to the Oriental Institute by the LaSalle Banks. The current state of the work can be viewed at http://oi.uchicago.edu/research/projects/ari/ (the URL is case sensitive; see figure 1).
The ARI project aims to make it possible for the user to move quickly among various ways of examining the inscriptions, from looking at whole texts of individual versions to section-by-section comparison among the versions in different languages; or from a glossary entry in one language to the context of each use of a particular word in all inscriptions in that language; or (eventually) from a display of the versions as a whole or in sections to pictures that show the visual relationships among the versions and plans that show where a person moving among the buildings at Persepolis would have encountered the inscription - or, to put it more accurately, how the Oriental Institute's archaeologists saw and photographed the inscription and its setting in the 1930s.
The phrase "current state of the work" is used advisedly. In addition to the flexibility that this form of presentation gives to the user, it offers a similar flexibility to the editors, allowing additions, improvements, or changes of a sort that would be impossible in printed form: corrections of typographical and editorial errors, additions to the bibliography or notes to the texts, reconsiderations of decisions about translation, incorporation of critical responses, additions of new texts or images, or whole new classes of information. Where printed presentations are fixed, on-line presentations are protean, since it is also possible to revise the layout in which the information is displayed on the screen, or the organization of the connections among parts. This project is an experiment with solving problems of design and construction that will also appear in the presentation of other bodies of texts. Like other such projects, this one is under constant change, maintenance, and progress, so what can be described is its state at a particular moment and ambitions for further developments. As this is written, the project is very much under ongoing design and construction. But the core elements are in reasonable working order. Here is what they do. The basic data are 28 royal inscriptions on building elements, tombs, and movable objects from Persepolis, amounting to a total of 63 Old Persian, Elamite, and Akkadian versions. The long texts of Darius I from his tomb at Naqsh-i Rustam are still to be added. The individual inscriptions are identified with a system of sigla that is usual in Achaemenid scholarship, as DPa, DPb ... DPi (= Darius I, Persepolis, inscriptions "a" through "i"), XPa, XPb ... XPo (= Xerxes I, Persepolis, inscriptions "a" through "o"), etc. Sections of individual texts (numbered 01, 02, 03, etc.) are identified in a way that is also usual, as paragraphs marked off either by the beginning of the inscription, or by either of two formulaic phrases in Old Persian, Elamite, or Akkadian, one that means "I, so-and-so, king," the other that means "declares so-and-so, king."
The user can follow links from the home page or the introduction to "Browse Whole Text" to see the transliteration of the entire text of a single version of a particular inscription (see figure 2). For example, the user who responds to the prompt "Select Text" by choosing the text "XPa" and responds to the prompt "Select Version" by selecting "Old Persian" and then selects "Submit Display Options," will receive a complete transliteration of the best-preserved of the four copies of Xerxes' inscription on the Gate of All Lands that faced the top of the monumental stairway, the main entrance to the Persepolis platform.
Whether he chooses the Old Persian, Elamite, or Akkadian version, he may be dismayed to find an unfamiliar system of transliteration, in which the various special characters, super- and subscripts that are used in conventional printed transliterations have been replaced by combinations of characters from the ordinary ASCII character set (for example, s$ replaces sh, c% replaces ç, h# replaces h, and so on; the transliteration conventions are explained in the introduction). It is hoped that in a later stage of the project this feature (which makes it easy for the programs to search and index the transliterations, and makes it simple for users to type Old Persian, Elamite, or Akkadian words on U.S. keyboards) can be amended, so that the texts appear in more familiar form.
The user can also follow links to "Browse Versions by Section" to see a section-by-section parallel display of the versions, with or without their English translations (see figure 3). For example, the user who selects the text XPa, selects section "02," selects all three of the boxes marked "Old Persian," "Elamite," and "Akkadian," and selects "Include Translations," will receive transliterations and translations of the three versions of the second paragraph of the inscription on the Gate of All Lands, the paragraph in which Xerxes identifies himself as the son of Darius, a member of the Achaemenid dynasty, and king of "countries containing many kinds (of men)" (in Old Persian and Elamite) or of "lands of the totality of tongues" (in Akkadian). And if this difference in the versions interests him, he will also see that "containing many kinds (of men)" is represented in the Old Persian by a phrase paruv zananam, in Elamite by a transcription of the Iranian phrase, parruzananam (not by an Elamite translation), but in Akkadian by a paraphrase, sha naphar lishanati, "of the totality of tongues."
That is not quite accurate - he will actually see sign-for-sign transliterations p-ru-u-v : z-n-a-n-a-m rendering Old Persian paruv zananam, par2-ru-za-na-na-um rendering Elamite parruzananam, and s$a2 nap-h#a-ar li-s$a2-nu.MES$ rendering Akkadian sha naphar lishanshati. If he knows the orthography and philology of the languages, he can recognize the forms transcribed this way in their ancient spellings. Otherwise, he can move to "Browse Lexicon, see an alphabetical list of all the Old Persian or Elamite or Akkadian words in all the inscriptions. For example, if he selects "Old Persian" and scrolls through the list to "paru-: much, many," and clicks on it, he will receive a lexical entry that shows all spellings of all forms of the word and the passages in which each form occurs. If he selects the citation "xpa02" of the spelling p-ru-u-v, he will receive a display of the Old Persian version of section 2 of the Old Persian version of XPa, including the phrase paruv zananam. If he returns to the list of Old Persian words, he will see another entry, "paruzana-: of many kinds" and recognize that the phrase was treated as a single word in two other inscriptions of Xerxes and one of Artaxerxes I, and if he follows links to those passages, he can see the Elamite and Akkadian equivalents in each case. A future version of the lexical entry will indicate the equivalents of the glossed word in the other languages (with links to the corresponding entries in the glossaries of the other languages), and a future version of the response to a request for a particular context will offer the option of seeing the parallel contexts in other versions.
The user who begins not with texts, but with language, can start at "Text and Lexical Search," submit a search for any string of characters that are part of a word in one of the glossaries, select the language of the glossary to be searched, and receive a list of glossary entries (see figure 4). For example, if he submits the string "paru-," selects the option "Words Beginning with String," and selects the language "Old Persian" and the scope of the search "Search Lexical Entries Only," he receives a response that identifies the glossary entry "paru-: much, many," from which he can move to the full lexical entry, containing forms, spellings, and citations, as above. If he submits the string "paru" (without final dash), he receives a list of three glossary entries: "paru-: much, many," "paruva-: former, previous," and "paruzanah: of many kinds," from each of which he can move to complete lexical entries and then to the contexts of each citation. If he submits the string "zana" and selects the option "Lexical Entries Containing String," he receives a list that includes the compound form "paruzana-: of many kinds," and the independent form "zana-: human being, kind," each with links to full lexical entries that lead in turn to citations in context.
The user who begins not with lexicon but with epigraphy can also start at "Text and Lexical Search," submit a string of characters that are part of the transliteration of any of the texts, select the appropriate language, select a response that includes occurrences in transliterated texts, occurrences in the lexical entries, or both, and receive a list of lexical entries and/or contexts which can then be examined in detail. For example, if he submits the transliteration string p-ru-u selects the language Old Persian, selects the option "Search Lexicon and Text," and selects the search type "Words Beginning with String," he receives a list of six spellings of forms of the three words paru-, paruva, and paruzana-, with links to those entries, and a second list of the same spellings with links to the passages of Old Persian texts in which each spelling appears.
A look at the lexical entries shows some of the information that is still to be supplied. The form contains headings marked "See also:" (for cross-references to compounds and cognates, with links to the respective lexical entries), "Equivalent to:" (for notation of the lexical equivalents in the other languages, with links to the respective lexical entries), and "Remarks:" (for notation of epigraphic, philological, or lexical problems, with links to appropriate bibliography).
Two of the main components of the project are still at an early stage. One is the catalogue, with the identifying information for each version of each inscription: location among the buildings of Persepolis (and excavation numbers of inscribed movable objects), and primary publications of facsimiles, mostly photographs. If the user follows links to the sample entry for XPa Old Persian, for example, the display indicates that there are four exemplars, describes the position of two of them on the door jambs of the Gate of All Lands, and indicates where photographs of each are published, in an abbreviated form of bibliographical citation with links to the running bibliography where complete publication information is supplied. The graphics are the other component at an early stage of work. Sample images reproduce published photographs of the two exemplars of XPa, with the three versions displayed side by side at the openings of the Gate of All Nations.
The user who follows links to "Bibliography" can inspect a running list of all publications cited anywhere among the pages of the site, in alphabetical order by author and date. If he selects an entry in this list, he can see the complete bibliographic information for the reference. This information is also linked to each of the occurrences of that reference in the site.
The flexibility that this form of presentation affords to both the user and the presenter answers the question of what should be done next with an embarrassingly large array of possibilities. Beyond the additions and corrections already mentioned, three main tasks have priority.
One immediate desideratum is the addition of plans, excavation photographs, and perhaps other illustrations, linked to the various forms of text display, to fulfill the intent to display synoptic versions of the inscriptions not only as linguistic items, but also as archaeological artifacts and architectural elements. This task also offers an opportunity to make an accessible digital archive of many of the photographs from the Oriental Institute's excavations, unpublished as well as published; and an opportunity to amend some of the published plans to accommodate the results of later excavation and restoration.
A second desideratum, perhaps somewhat paradoxical, is to accommodate users who dislike computers. Ideally, the transliterations can be displayed in a familiar form, without the need to indicate diacritics with arbitrarily substituted ASCII characters. The user should be able to transfer every form of display to paper at his own printer easily.
Finally, and most importantly: one of the main purposes of this project has been to use this small corpus of well-known texts as an arena for gaining experience with problems of preparing, displaying, and analyzing texts in a variety of unrelated languages, and also to begin defining editorial and design standards for the electronic publications under the Oriental Institute's aegis.
We anticipate that electronic/web publication will be an important medium for the distribution of Ancient Near Eastern texts and data, and we expect the Achaemenid site, as it evolves, to become an important test arena for standards and tools of this medium. As a first step, represented by the Achaemenid site, is the preparation in electronic form of a corpus texts containing an explicit representation of all the information needed for a scholarly edition. We intend to provide at the site periodically updated information on our emerging format, its relation to developing public standards such as XML, and the possible extension of the format to other bodies of texts. In one such extension, for example, the programs and markup standards being developed here, are being used by Jones and Stolper to prepare on-line publication of the Achaemenid Elamite and Aramaic administrative texts recovered by the Oriental Institute's excavations at Persepolis, beginning with about 2,000 Persepolis Fortification tablets that the late Richard T. Hallock transliterated, an unpublished sequel to his Persepolis Fortification Tablets (OIP 92 ), and with about 100 Persepolis Fortification tablets that the late George G. Cameron studied in the Tehran Museum but was unable to publish, and with about 400 Aramaic tablets, also from the Fortification archive, treated in a manuscript left by the late Raymond Bowman.
At the same time as we are preparing texts, we are also in the process of assembling a set of tools for the preparation, indexing, and displaying of these texts, and their associated glossaries, bibliographies, catalogues, and critical apparatus. For the moment we are using for this purpose home brew scripts written in the Perl programming language. These prototype programs have in fact turned out to be more robust and extensible than initially anticipated. They have proved useful, for instance, in setting up an Afroasiatic etymological database, in the preparation of an electronic study edition of the corpus of Aksumite inscriptions from early first-millennium ad Ethiopia, and in designing a prototype for the Middle Egyptian on-line instructional program described elsewhere in this issue of News & Notes (see page 13). However, we are following closely the current flurry of activity in the development of tools for electronic text, and we expect that we will be gradually replacing provisional tools with more polished and professional versions as they become available.
Matthew W. Stolper has been on the faculty of the Oriental Institute since 1980. His research has concentrated on Achaemenid Babylonian texts and history and on Elamite texts and history. Gene Gragg has been on the faculty of the Oriental Institute since 1969, and became Director in 1997. In addition to research and teaching in the peripheral languages of the ancient Near East, Gene Gragg has long been occupied with the Semitic and Cushitic languages of Ethiopia. He did lexical research in Ethiopia and has published a dictionary of the Cushitic language Oromo.
Revised: July 30, 2007