Bryn Mawr Classical Review 04.04.05


ANNOUNCEMENT: An Electronic Database of Greek Lexicography: LSJ 9 and friends on-line


In early June, a proposal for placing LSJ on-line was posted to a variety of sources. Many people responded with their ideas and reactions. The dominant message that we received seemed to be that LSJ by itself was not quite enough. Even the most stinging critics of LSJ and its weaknesses seemed in the end to agree that LSJ needed to be the starting point, but that we needed to push beyond the existing model.

The following is a revised summary that attempts to address these suggestions. The final proposal will go in in three weeks (Aug 31), but we still have time to deal with more suggestions, great and small.

In any event, thanks to all who responded the first time around!

August 9, 1993

NOTE: Permission is included to anyone who wishes to repost this notice to any other source.


Students of classical Greek still rely upon the 9th edition of the Liddell Scott Jones Greek-English Lexicon (LSJ 9), published in 1941. With its 150,000 dictionary entries and 750,000 citations to source texts, LSJ 9 remains a monumental achievement, representing a century of continuous lexicographic work. Nevertheless, the form has overwhelmed the content. LSJ 9 has remained untouched for more than fifty years, with new material published in supplements (one in 1968, another forthcoming). Even these supplements, however, ambitious they may be, cannot fully reflect the advances in our knowledge of Greek.

We propose to create a lexicographical database that will allow students of Greek to gain better access to information about language and that will make it possible to publish, in a central place and in a timely fashion, new ideas about the meanings and usage of words.

Placing LSJ 9 in electronic form is the obvious starting point for such an effort, and placing LSJ on-line will be our primary task, but LSJ is only one step in a larger process. After discussions with classicists, we have decided to include in this project additional lexicographic tools so as more fully to demonstrate the capabilities of the electronic medium and to provide a more complete foundation on which others can build. These will include specialized lexica (e.g., Slater on Pindar, Ebeling on Homer) and classic commentaries (Jebb on Sophocles and Adams on Plato), as well as an index of Greek terms discussed by the scholia and lexicographic authors already on-line in the TLG. All of these sources put together are far smaller than LSJ 9 (40 megabytes), but they make the final database far more diverse and a better model for future work than LSJ 9 by itself.

The Electronic Oxford Electronic Dictionary, which has been available for years, allows readers not only to call up specific definitions but to ask new questions not feasible with the printed version: e.g., search for definitions (which definitions contain the word "wealth"?) and locate citations (which definitions cite King Lear?). Nevertheless, the OED provides only a partial model of how such a database for Greek might work. More than 90% of the citations in LSJ 9, for example, are already on-line in the Thesaurus Linguae Graecae (TLG). Lexica, commentaries and source texts can interact dynamically in ways that are not feasible for any one document by itself.

Roughly 700,000 the citations in LSJ 9 and all of the citations in the specialized lexica would become "active links." The LSJ article on KATADIASPLEKO/W, for example, cites Sch. Ar. Pl. 1082. The reader could point to this reference and call up the text in a separate window. Conversely, we will provide back references so that anyone looking at Sch. Ar. Pl. 1082 will see that the LSJ article on KATADIASPLEKO/W cites this passage.

Over the past ten years, we have developed a morphological analyzer for Greek. The work proposed here will allow us to include the morphological information of LSJ 9, increasing our database of words from 40,000 to 150,000. We will be able to support morphologically intelligent text searches for any word in the database. Users will be able to locate disparate forms such as OI)/SW and H)/NEGKON by asking specifically for FE/RW. Likewise, we will systematically run this morphological searcher over the TLG for each dictionary entry, creating a precompiled database indicating where each word is used and how often.

Simply by collating the passages discussed by ancient lexicographic sources will allow modern scholars to see more fully how the Greeks themselves viewed their language. The scholar asking for information on TAFW/N at Iliad 9.193, for example, would encounter "see TE/QHPA in LSJ, also Schol. on Od. 16.806." The Odyssey Scholion preserves an ancient folk etymology not included in LSJ but useful as an indication of how the Odyssey was read.

We also seek to create a reference tool that will grow over time. Where a traditional print lexicon begins to drift out of date as soon as it has been published, the lexicographic database that we propose will be designed to become increasingly comprehensive and up to date as time progresses. Placing LSJ 9 in electronic form paves the way for a comprehensive new LSJ that integrates the forthcoming LSJ Supplement and that can, for the first time in a century, be regularly updated. Specialized lexica -- a crucial tool for scholarship -- have all but died out in classical scholarship. An electronic Ebeling would, however, be a step towards a new, badly needed lexicon of Homeric Greek. Jebb's commentaries on Sophocles, by contrast, remain classics in themselves. By publishing Jebb, we would also provide a ready framework that new commentaries could follow. We thus hope to encourage commentaries that appear both in electronic and in printed form.

Virtually all of the material that we propose to enter is in the public domain and we will thus be able to make the resulting database available inexpensively, as a TEI conformant text file for which others may write software, as a GOPHER server on the internet, as part of an electronic LSJ incorporating LSJ 9 and the new supplement, and as part of a preexisting system (e.g., a future version of the Perseus database).