0.10 Give us editors! re-inventing the edition and re-thinking the  (Page 11/18)

1) Our editions can reach a larger audience . First and most obviously, we can provide much better physical access to editions and commentaries, disseminating across space and preserving them over time. See Bodard 2008 for some of the new opportunities of electronic publication and digital classics, with a particular focus on epigraphy. We are now—and have for years been—able to deliver the results of our work to a global audience—reaching hundreds of millions rather than thousands of locations. We also have in our institutional repositories a mechanism to preserve these editions over long periods of time—certainly providing far longer access than the ephemeral print runs common in traditional publication. The guarantor is not the medium—paper vs. digital—but the reorganization of our library priorities. The need for digital repositories and services that can both preserve and provide sustainable access to the wide range of digital objects now becoming available and the challenge this provides to the traditional models of research libraries is a widely discussed topic; for some recent work see ARL 2009 and Sennyey et al. 2008. We have the resources already in our library collection budgets to pay for dissemination and preservation. The questions are political and social rather than financial or technical.

2) A larger audience can make use of our editions . As we can see with the Advanced Papyrological Information System (APIS) (External Link) . and with the Homer Multitext Project, (External Link)&bdc=12&mn=1169 . digital representations in 2d and 3d of manuscripts, papyri, inscriptions and other written artifacts can provide better visual data than any print facsimile edition could match. For more on the use of APIS and papyrological collections online see Hanson 2001, and for the Homer Multitext Project see Dué and Ebbott 2009. We can, as Peter Robinson has long demonstrated, encode the textual data in machine-actionable forms that allow us to analyze and visualize variants with greater precision. For one of Robinson’s most recent discussions of the creation of digital editions, see Robinson 2009. We can link Greek and Latin editions to modern language translations, either produced for the edition or already published elsewhere. We can add as much explanatory material as we have the time to produce and as we consider useful, including visual and textual explanations, static images, and dynamic visualizations. We can align our primary sources with the material record, not simply as a source for illustrations but to provide contrasting views of the lived world on which the textual and material record shed light.

3) Systematic annotation transforms the editorial process, redefines what readers can expect, and enables editions to interact directly with much larger collections . Our ability to annotate our primary sources has so far outstripped what we could do in print that annotation has evolved into something qualitatively different. We have long had indices of places, but efforts such as the Hestia Project have created machine-actionable databases with which to analyze the spatial relations implicit in our sources. For millennia, students of Greek and Latin have patiently answered such questions as: “What is the main verb?” “What is the subject?” and “What noun does this adjective modify?” When we systematically record the syntactic dependencies into a database and create treebanks of syntactically-analyzed sentences, we can convert impressionistic statements such as “common in tragedy” or “rare in late prose” into statements that are quantifiable and transparent, because we can call up the treebank and examine the decisions behind the numbers. The Perseus Project has been developing a Latin Dependency Treebank since 2006, and work on an Ancient Greek Dependency Treebank began in 2008. Both treebanks can be downloaded from (External Link) . We have always embedded our syntactic interpretations in our print editions—each comma and period reflects our interpretation of the language. Now we can make those assumptions explicit and then use them to support new questions and research. For example, the Latin Treebank has been utilized in various projects, including the development of a dynamic lexicon (Bamman and Crane 2008a) and the automatic detection of textual allusion (Bamman and Crane 2008b). And, at the same time, the treebanks that support large-scale linguistic analysis can help the reader understand a complex sentence in Plato and thus expand the role that sentence can play in intellectual life. Here, as so often, we find the automated system and human observer interacting symbiotically, with each driving the other.