Extracted on 25/11/2008 from: https://www.ideals.uiuc.edu/handle/2142/392

THE ARTICLE:

Center for Electronic Texts in the Humanities

Rutgers and Princeton Universities

New Brunswick, New Jersey

Electronic Texts in the Humanities:

ABSTRACT

Electronic texts have been used for research and teaching in the

humanities ever since the end of the 1940s. This paper charts the

development of various applications in literary computing including

concordances, text retrieval, stylistic studies, scholarly editing, and

in the form of the Text Encoding Initiative's implementation of the

software packages currently available is posing significant problems

of support for libraries as well as delivering only partial solutions to

towards more advanced network-based delivery mechanisms.

INTRODUCTION

begun, and a body of knowledge and expertise has gradually evolved.

of electronic text usage in the humanities. Until recently, almost all

transcriptions, which are referred to as text, but the implications for

images will be noted briefly.

humanities. This can be literary text, which is prose, verse, or drama,

form of letters, memoranda, charters, transcripts of speeches, papyri,

inscriptions, newspapers, and the like. Other texts are studied for

linguistic purposes, notably collections of text forming language corpora

languages and/or writing systems. Others have variant spellings, critical

apparatus with variant readings, marginalia, editorial emendations, and

annotations, as well as complex and sometimes parallel canonical

needed for scholarly analysis.

is accompanied by a reference indicating the location of the occurrences

in alphabetical order, but they could also be alphabetized or sorted

hapax legomena (once-occurring words) can easily be seen. Specialized

speaker within a play) or sorted according to the words before or after

that the production of concordances was typically a mechanical batch

process that could generate vast amounts of printout.

Early on, attention was also paid to defining the alphabetical order

for sorting words in a variety of languages, for example, transcriptions

of Greek and Russian as well as Spanish where ch, II, and rr are separate

accented characters, editorial emendations, and the like were soon devised,

of two of the most widely used concordance and retrieval programs

be found in Howard-Hill (1979), Hockey (1980), and Sinclair (1991).

a vast collection of literary works since the time of the revolution, began

of Swedish (Gothenburg) (Alln 1970), Classical Latin texts at Liege

direction of Antonio Zampolli (1973). At that time, the only means

Burton (1981a, 1981b, 1981c, 1982) describes these projects and others

which makes interesting reading.

precompiled index or concordance of the text. Besides their obvious

application as a reference tool, concordance and text retrieval programs

can be used for a variety of scholarly applications, one of the earliest

of which was the study of style and the investigation of disputed

long time. Articles by T. C. Mendenhall at the end of the last century

describe his investigations into the style of Shakespeare, Bacon, Marlowe,

word-counting machine. Mendenhall (1901, 101-2) notes

the excellent and entirely satisfactory manner in which the heavy

it. ... The operation of counting was greatly facilitated by the

construction of a simple counting machine by which the registration

of a word of any given number of letters was made by touching

a button marked with that number.

Mendenhall's findings were not without interest, since he discovered

that Shakespeare has more words of four letters than any other length,

or function words. These are independent of content, and authors often

considered to be a classic authorship study, since the twelve disputed

Madison almost always uses "whilst." Other words favored by one or

the other of them included "enough" and "upon."

particles and prepositions, that were derived from his reading of printed

novels and to characterize their idiolects. These and similar studies

in the text and to distinguish between homographic forms.

Concordances can also be a valuable tool for the historical

lexicographer, and several large textbases were originally compiled for

of all this corpus and select citations from the concordances for the

occurs about 15,000 times, and it was just possible for a lexicographer

to read all the concordance entries for it. This is obviously not feasible

compiled using a twenty-million-word corpus of English (Sinclair 1987).

Other electronic texts have been created for the analysis of meter

and rhyme schemes. In the 1960s, scansion programs existed for Greek

and Latin hexameter verse (Ott 1973). Metrical dictionaries were compiled

for authors as diverse as Hopkins (Dilligan and Bender 1973) and

1987).

of some electronic texts. In simple terms, this process has consisted of

collating the manuscripts, establishing the textual tradition, compiling

an authoritative text, compiling the critical apparatus, and then printing

text. In the 1960s, computer programs to collate manuscripts began

to appear, and it was soon realized that collation could not be treated

as a completely automatic process and that, because of the lineation,

verse was easier to deal with than prose. Robinson's (forthcoming)

has a graphical user interface and is by far the most flexible collation

program.

constrained within the structures of database programs, particularly

relational databases where the information is stored as a set of rectangular

tables and is viewed as such by the user. Very little humanities-oriented

information fits this format without some restructuring, which, more

Hypertext has provided a solution to data modeling for the

interrelated information, which can be annotated by the user if desired.

and secondary material together. Images, sound, and video can be incorporated

edition can be represented very effectively as a hypertext, but hypertext

without privileging any particular one of them (Bornstein 1993). Other

experiments have used hypertext to model the narrative structure of

literature with a view to helping students understand it better (Sutherland

forthcoming).

as a by-product of research projects such as those described above. Large

collections of text have been assembled by a few research institutes,

of language and its relation to the cultural heritage. Most other texts

reflect the interests of those research groups or individuals, and it is

scholarly purposes. These texts are ASCII files, not files that have been

to 95 percent of texts fall into this category. For a variety of reasons,

However, it was soon realized that considerable time and effort is

keyboarded, and this is still the normal means of input. Optical character

systems tend to have difficulty with material printed before the end

of the last century, newspapers, or anything else where the paper causes

the ink to bleed, as well as material containing footnotes and marginalia,

nonstandard characters and words in italic, or small capitals. Those

and the collection of papers assembled by the Netherlands Historical

for most kinds of processing and analysis. Most large data entry projects

be inserted at that time.

Recognizing the need to preserve electronic texts, the Oxford Text

Archive (OTA) was established in 1976 to "offer scholars long term

storage and maintenance of their electronic archives free of charge."

of a British Library sponsored project to review the Oxford Text Archive.

^ There have been a few systematic attempts to create or collect and

for a specific language is the Thesaurus Linguae Graecae (TLG), which

a databank of almost seventy million words of Classical Greek (Brunner

using BRS-Search and can be accessed via Telnet to lib.dartmouth.edu

then, at the prompt, type "connect dante."

di Linguistica Computazionale in Pisa has a large collection of literary

collection. Material in Welsh and other Celtic languages has been built

and groups in various English-speaking countries are compiling corpora

completion of a hundred-million-word corpus of written and spoken

Center for Text and Technology maintains a catalog of projects and

institutes that hold electronic texts but not the texts themselves. This

Lancashire (1991) is the most comprehensive source of information in

print about humanities computing projects in general.

is the only attempt to catalog existing electronic texts using standard

by the Center for Electronic Texts in the Humanities (CETH). It contains

monographic electronic text files using Anglo-American Cataloguing

available by publishers or software vendors. These are the texts that

and are usually packaged with specific retrieval software. Examples

Patrologia Latina published by Chadwyck-Healey also fall into this

use with other software. Oxford University Press also publishes

electronic texts, which are ASCII files. Their texts are particularly well

program, which they also publish.

are mostly designed for scholarly applications on what are complex

the cost of supporting them will not be insignificant. Librarians are

Several of these products are designed for the individual scholar to use

or modified existing ones in between. Another issue is just what words

as good as the words that have been indexed. In some cases, this seems

understanding of the nature of the material and the purposes for which

it might be used. Other institutions have chosen to acquire texts in

the intellectual content of the material.

Creating an electronic text is a time-consuming and expensive

it. Texts that are created specifically for one software program often

to another and because they can be used for different purposes and

applications.

of a text (chapter, stanza, act, scene, title) and enables specific areas

or subsets of text to be searched and text that has been retrieved to

used by the Thesaurus Linguae Graecae, and the three-level referencing

which contain formatting instructions.

Following a planning meeting in 1987, a major international effort

to create guidelines for encoding electronic texts was launched by the

Association for Computers and the Humanities, the Association for

Computational Linguistics, and the Association for Literary and Linguistic

(TEI), brought together groups of scholars, librarians, and computer

a comprehensive list of the features within those texts.

is a metalanguage within which encoding schemes can be defined. It

is descriptive rather than prescriptive and thus can form the basis of

the reusable text. It permits multiple and possibly conflicting views

encodings can be added to a text without detriment to what is already

It provides base tag sets for prose, verse, drama, dictionaries, transcripts

textual criticism, transcription of primary sources, language corpora,

formulae and tables, graphics, hypermedia, analytical tools, and names

has very recently been published (Sperberg-McQueen and Burnard 1994).

associated with them. Often, it is difficult to establish what the text

that he or she did not find it necessary to record any documentation

any model to follow for documenting the text and thus recorded only

about electronic texts extremely difficult.

electronic text and the source from which it was taken, information

about the principles that governed the encoding of the text, any

classificatory material, and a revision history that records the changes

DIGITAL IMAGING

texts can also be applied to digital imaging of manuscripts and textual

the archival role is obviously very important. Most of the cost in digital

imaging is in taking the object to and from the camera, and so it makes

sense to digitize at the highest resolution possible. Storing the image

all the same problems that have been experienced with text stored in

a proprietary indexing program. It will not be possible to guarantee

that the image will be accessible in the future or that it can be used

for other purposes. Documentation and provenance information are just

to be used for images and offers a route to using both text

from the text to the image and can form the basis of a system that

operates on the transcription of the text but displays the image to the user.

Experience of working with electronic literary texts has highlighted

avoids the omission of some homographic forms; for example, the

want to search on that. In other languages, it provides a simple key

to the examination of the ends of sentences, for example, clausulae in

search on the ends of words is also useful, particularly for verse and

provide an index by endings. For others, this kind of search can take

within certain structural boundaries or allowing them to extend beyond

a boundary. For example, finding "tree" within ten words of "flower"

occurs at the beginning of the next chapter.

specifying strings of characters, possibly linked by Boolean operators,

whereas most users are interested in concepts, themes, and the like.

String searches cannot effectively disambiguate homographic forms, for

or the verb "bank" (used of an airplane), or Latin "canes" as "dogs" or

"you will sing."

syntactic analysis, and parsing have been used experimentally

for some time, but our understanding of these is still only

95 percent. Morphological analysis has been done reasonably well for

some languages, for example, Ancient Greek, but there are no widely

available general-purpose programs that are suitable for literature.

Father Busa recognized the need to lemmatize his concordance to St.

In Busa 1992, he reflects on the lack of intellectual progress and on

other types of text, and there have been relatively few attempts to apply

linguistics are turning to the compilation of large-scale lexical resources

and knowledge bases for use by natural language understanding systems.

dictionary and restructure that within the computer as a lexical database

that contains morphological analyses, lemmas, frequent collocations,

and other information that would help to disambiguate homographic

in language corpora, which are large bodies of text from which

initial methodologies used by computational linguists to analyze large

corpora are concordance based and are very similar to those that have

REFERENCES

Allen, Sture. 1970. Vocabulary Data Processing. In Proceedings of the International

Conference on Nordic and General Linguistics, ed. Hreinn Benediktsson, 235-61.

Reykjavik: Visindafelag Islendiga.

Bornstein, George. 1993. What is the Text of a Poem by Yeats? In Palimpsest: Editorial

Theory in the Humanities, ed. George Bornstein and Ralph G. Williams, 167-93.

Ann Arbor: University of Michigan Press.

Brandwood, Leonard. 1956. Analysing Plato's Style with an Electronic Computer. Bulletin

of the Institute of Classical Studies 3: 45-54.

Brunner, Theodore F. 1991. The Thesaurus Linguae Graecae: Classics and the Computer.

Library Hi Tech 9(1): 61-67.

Burrows, J. F. 1987. Computation into Criticism: A Study of Jane Austen's Novels and

an Experiment in Method. Oxford: Clarendon Press.

Burton, Dolores M. 1981 a. Automated Concordances and Word Indexes: The Fifties.

Computers and the Humanities 75(1): 1-14.

Burton, Dolores M. 1981b. Automated Concordances and Word Indexes: The Early Sixties

and the Early Centers. Computers and the Humanities /5(2): 83-100.

Burton, Dolores M. 1981c. Automated Concordances and Word Indexes: The Process,

the Programs, and the Products. Computers and the Humanities /5(3): 139-54.

Burton, Dolores M. 1982. Automated Concordances and Word-Indexes: Machine Decisions

and Editorial Revisions. Computers and the Humanities /6(4): 195-218.

Busa, Roberto, S. J. 1974-. Index Thomisticus. Stuttgart-Bad, Connstatt: Fromann-

Holzboog.

Busa, Roberto, S. J. 1992. Half a Century of Literary Computing: Towards a 'New'

Philology. (Reports from Colloquia at Tubingen). Literary b Linguistic Computing

7(1): 69-73.

2(1): 34-40.

Anciennes a 1'Universite de Liege. L'Antiquite Classique 30: 429-44.

Investigation of English Prosody. In The Computer and Literary Studies, ed. A. J.

Aitken, Richard W. Bailey, and N. Hamilton-Smith, 239-52. Edinburgh: Edinburgh

University Press.

Healey, Antoinette diPaolo. 1989. The Corpus of the Dictionary of Old English: Its

Delimination, Compilation and Application. Paper presented at the Fifth Annual

College, Oxford, England.

Hockey, Susan. 1980. A Guide to Computer Applications in the Humanities. London:

Duckworth; Baltimore: Johns Hopkins University Press.

Hockey, Susan. 1986. OCR: The Kurzweil Data Entry Machine. Literary b Linguistic

Computing 1(2): 63-67.

Hoogcarspel, Annelies. 1994. The Rutgers Inventory of Machine-Readable Texts in the

Humanities. Information Technology and Libraries 13( 1 ): 27-34.

Howard-Hill, T. H. 1969. The Oxford Old-Spelling Shakespeare Concordances. Studies

in Bibliography 22: 143-64.

Howard-Hill, T. H. 1979. Literary Concordances: A Guide to the Preparation of Manual

and Computer Concordances. Oxford: Pergamon Press.

Kenny, Anthony. 1978. The Aristotelian Ethics: A Study of the Relationship between

the Eudemian and Nicomachean Ethics of Aristotle. Oxford: Oxford University Press.

Kenny, Anthony. 1982. The Computation of Style. Oxford: Pergamon Press.

Lancashire, Ian., ed. 1991. The Humanities Computing Yearbook 1989-90: A

Comprehensive Guide to Software and Other Resources. Oxford: Clarendon Press.

Monthly 60 (December): 97-105.

Mosteller, Frederick, and David L. Wallace. 1964. Inference and Disputed Authorship:

The Federalist. Reading, Mass.: Addison-Wesley.

Netherlands Historical Data Archive. 1993. Optical Character Recognition in the

Historical Discipline: Proceedings of an International Workgroup. Netherlands

Historical Data Archive, Nijmegen Institute for Cognition and Information.

Ott, Wilhelm. 1973. Metrical Analysis of Latin Hexameter: The Automation of a

Philological Research Project. In Linguistica Matematica e Calcolatori (Atti del

Convegno e della Prima Scuola Internazionale, Pisa 1970), ed. Antonio Zampolli,

379-90. Florence: Leo S. Olschki.

Packard, David W. 1974. Sound-Patterns in Homer. Transactions of the American

Philological Association 104: 239-60.

Philippides, Dia Mary L. 1981. The Iambic Trimeter of Euripides. New York: Arno Press.

Proud, Judith K. 1989. The Oxford Text Archive. British Library Research and

Development Report. London: British Library.

Quemada, Bernard. 1959. La Mecanisation dans les Recherches Lexicologiques. Cahiers

de Lexicologie 1: 7-46.

Robey, David. 1987. Sound and Sense in the Divine Comedy. Literary b Linguistic

Computing 2(2): 108-15.

Robinson, Peter M. W. Forthcoming. COLLATE: A Program for Interactive Collation

of Large Manuscript Traditions. In Research in Humanities Computing, vol. 3, ed.

Susan Hockey and Nancy Ide, 32-45. Oxford: Oxford University Press.

Computing. London: Collins.

Sinclair, John M. 1991. Corpus, Concordance, Collocation. Oxford: Oxford University

Press.

Sperberg-McQueen, C. M., and Lou Burnard, eds. 1994. Guidelines for the Encoding

and Interchange of Electronic Texts. Chicago and Oxford: Association for Computers

and the Humanities, Association for Computational Linguistics, Association for

Literary and Linguistic Computing.

Sutherland, Kathryn. Forthcoming. Waiting for Connections: Hypertexts, Multiplots,

and the Engaged Reader. In Research in Humanities Computing, vol. 3, ed. Susan

Hockey and Nancy Ide, 46-58. Oxford: Oxford University Press.

Tombeur, P. 1973. Research Carried Out at the Centre de Traitement Electronique des

Documents of the Catholic University of Louvain. In The Computer and Literary

Studies, ed. A. J. Aitken, Richard W. Bailey, and N. Hamilton-Smith, 335-40.

Edinburgh: Edinburgh University Press.

Zampolli, Antonio. 1973. La Section Linguistique du CNUCE. In Linguistica Matematica

e Calcolatori (Atti del Convegno e della Prima Scuola Internazionale, Pisa 1970),

ed. Antonio Zampolli, 133-99. Florence: Leo S. Olschki.