HTI Note: This introduction was based on use of the CD-ROM version of the English Poetry Database. Improvements in function and performance have been made for the HTI online version.
The word 'database' has an abstract ring to it. A library is a place as well as a collection of books, an august and public place, or a domestic and comfortable one; bodies are found in libraries (dead in murder mysteries, sleeping in real life). Nothing fleshes out a database; the appearance of this one is of five thin silver discs, almost without dimension or weight, precise in form, unprepossessing. The discs contain (but how? most of us couldn't say without making a specialist laugh or weep) that hard word data (a plural whose singular has given up the struggle for linguistic survival) and we know what to do with data: we process it, we take our places in the chain of manufacture and exchange. With chilling aptness the function of reading has passed to the machine: data may be machine readable, but not readable by us. A database does not have readers but users. It may be user-friendly: think of Captain Kirk being welcomed on board the alien vessel by a cool humanoid in a close-fitting, shimmering suit. Or perhaps not a presence, more a voice, guiding your search along the ordered pathways and circuits; not even a voice -- a noise, a click. Whatever the Muse of English poetry may be, the Muse of the English Poetry Full-Text Database has appeared to me in the form of a mouse.
The Bibliography of this project has therefore a special status. The books on this list are now both more and less than they were: more in being assimilated to each other in ways to marvel at and exploit; less in the loss of their singular bookishness, their materiality. In becoming texts, these volumes have given up their volume, their physical and historical density. Unbound, the texts, like saved souls, have soared into an electronic heaven where they rejoice in an infinity of potential combinations with other texts. Of the choice of the saved, and the limitations as well as the bliss of the electronic paradise, I shall say something further on. What matters now is to insist that the Bibliography does, in some sense, restore the loss of volume suffered by the books which have been incorporated into the database. At any rate the shadow or echo of the vast, the truly staggering manual labour that went in to the compilation of the thing can be traced. Like a pyramid or a cathedral, the database is a work of anonymous hands, the people who copied (or 'keyed') each poem in each book recorded here; and each poem was copied twice, then checked and proofed by others. From John Abbot's Devout Rhapsodies to the turgid canzons of Zepheria, each book on this list has been handled, each page turned, in a sustained collective, impersonal act of attention to English verse which has never been equalled and probably never will be. The pages of the Bibliography remind us of the sheer diversity of material practice represented by the publishing of poetry over twelve centuries (taking 'publishing' in its widest sense to mean the public circulation and distribution of writing, whether in manuscript or printed form). The database bears the marks of contemporary divisions over the ownership of texts, of particular versions in which economic investment has conferred value and title, and these divisions, too, can be followed through the Bibliography. Still, the scale of the project is the first fact. The Bibliography contains around 4,500 entries, each one of which records an individual volume with at least one item, one poem, which appears in the database. No single library has all these volumes, not the British Library or the Library of Congress. The Bibliography in that sense comprises the contents of an ideal or utopian library, which in fact exists nowhere except in the database, and then only under strange conditions. Walking around our ideal 'library' we would notice, for example, that many of the books have been mutilated: pages are missing, in some cases most of the volume has been tom out and only a few fragments of verse remain. A characteristic page offers volumes with 'Only verse included; Latin verse omitted ... Preliminaries and introductory matter omitted ... verse reproduced elsewhere in English poetry omitted'. It is not that we necessarily want these things, but that their omission makes a difference to the kinds of reading and understanding which are possible. But there are fascinating compensations, too. We tend to think of books of poetry as whole and self- defining: Lyrical Ballads, In Memoriam, Men and Women. The Bibliography vividly reminds us of the miscellaneous, eclectic, partial forms in which poetry has been cast. The work of an individual poet may be represented by a single monumental volume, or by items scattered in other people's books; poetry may be the substance of a volume, or an accidental feature of it; what counts as poetry for the purpose of the database may have been extracted from a surrounding context of discursive prose, of fiction, of drama, or of music.
The scale of the database means also that it usually includes only one version of each work. English Poetry is a vast collection of texts, but it is not a 'hypertext'; it does not offer multiple versions of texts in an interactive environment. Of course the database is not an ark; if English poetry were being embarked in it for good, then the sight of the animals going in one by one might well cause some dismay, but this is not the case. However grand in scale, the database is still only a collection of poems. Nevertheless its choice of texts cannot help having some influence on how it is used. Textual critics like Jerome McGann have argued in recent years against the concept of a single, stable, authoritative text, incarnating that over- worshipped god 'authorial intention'; but however much in sympathy with this view the compilers of the database might have been, it is obvious that its application to the project would have meant its immediate cancellation. The size of the database demanded a choice of primary texts, according to the best editorial principles that could be found to cover the enormous chronological, genetic, and bibliographical range of the material. Against the drawback that users of the database are presented with single versions of texts, where often multiple versions might be desirable, must be set the advantage of having so many single versions of different works.
The question of scale inevitably entails the question of value. My comparison of the database with a pyramid or a cathedral is open to the obvious objection that while each component of these buildings is necessary to the whole, the parts and the whole do not correspond in the same way in the database, and that the true analogy is with a vast rubbish dump in which heaps of broken images overlay the few items of real worth. The phrase 'English poetry' is most often used in the selective sense, and is thought of as a product of time's sorting and discarding. Occasionally, it is true, we become aware that time has done its job badly; some discarded poetry is retrieved, some we once prized is binned; we may worry, also, that some mediocre poets are preserved beyond their natural term by academic life-support systems, that critics, too, continue to draw the salaries of dead souls. But these arguments take place in a petty arena. Ask the average university lecturer who specializes in the poetry of any period to name fifty poets of that period and he or she would be hard pressed. I would guess that none could name a hundred from any period unless they had compiled an anthology of that period. I have just turned a dozen pages of this Bibliography and counted fifty-seven separate authors, of whom I recognized the names of six and was familiar with the work of two. We forgive ourselves our ignorance for both aesthetic and practical reasons. That the best poetry survives is comforting, and its small quantity makes it manageable. We rightly assume that the quantity will be small, that there will be only a handful of great poets in each generation, and of course it is their work we wish to preserve and study, their genes which we wish to keep in our breeding programme. (That the best poetry of the past is the breeding-ground for the best poetry of the present is a principle we rarely question.) In any case, the amount of work which would be needed to study poetry in any other way has hitherto made the task virtually impossible. A single book is one of the most sophisticated and flexible of all information systems, but a pile of books is one of the clumsiest. I am speaking here, I stress, about books as a means of study. The leaps of imagination and insight which you achieve in turning the pages of a book, in the pauses of reading, in the process of browsing backwards and forwards, in the establishing of a rhythm of absorption and reflection, cannot be replicated in any electronic form; they depend on physical touch. But comparative, thematic, and historical study are vexingly limited by the very physical properties which make the single book so versatile.
A tremendous amount of energy may go into the making of a list, but the result is often an inert mass. The Bibliography of the English Poetry Full-Text Database is, by itself, no more than a list of books. It is based on the New Cambridge Bibliography of English Literature (NCBEL) and constitutes an interpretation of that great flawed document in three modes: selection, material form, and arrangement. It contains the work of every author listed as a poet in NCBEL, but with some exceptions and additions; the particular books which form the work of each author have been chosen according to the editorial criteria of the project (needless to say these criteria have not been uniformly applied, in some cases because the preferred texts were unavailable, in others because one editorial principle overrode another: Byron's dramas are included, for example, though verse drama is generally excluded); and the arrangement in the Bibliography is alphabetical over the whole range, rather than within period or generic categories. The Bibliography on its own therefore represents an aspect of NCBEL, refined and specialized but not fundamentally different. NCBEL, too, is a list, gigantic and inert. But the Bibliography does not represent the database in the same way as it represents NCBEL. The records in the database are not inert, but dynamic, capable of being transformed, broken down, combined and re-combined.
Imagine that you are standing in a
library devoted solely to English poetry. On
the shelves are the separate volumes -- all 4,500
of them -- listed in the Bibliography. Now
imagine that you wish to get some idea of
how often the theme of Venus and Adonis has
been treated in English poetry, and specifically
whether it was popular in the nineteenth
century. Perhaps your attention has been
drawn to it by Wordsworth's (for him,
unusual) poem 'Love Lies Bleeding', or by
Shelley's fragment of Bion's Elegy, or by
Keats's Endymion, or by Elizabeth Barrett
Browning's lines in 'A Vision of Poets',
describing Keats himself as
the real
Adonis with the hymeneal
Fresh vernal buds half sunk between
His youthful curls, kissed straight and sheen
In his Rome-grave, by Venus queen.
You are less likely to have become interested in it by reading the Reverend Richard Cobbold's 'The Death of Adonis', a turgid and sentimental piece in Valentine Verses; or, Lines of Truth, Love, and Virtue, published in Ipswich 1827. Still, Cobbold's poem is one of forty- seven which English Poetry will turn up for you if you ask it to search for the words 'Venus' and 'Adonis' in poems by writers belonging to the three NCBEL periods of early, middle, and late 19th century.1 The full list is as follows:
This list results from a search of poems contained on discs 4 and 5 of the database. A search of the remaining three discs, which cover the period 600-1800, would turn up a further 263 poems. It would take you about ten minutes to produce this information in its rawest form, as a list of poems whose wide variation in length, subject matter, and tone would remain to be examined. Poems specifically about Venus and Adonis mix with poems which merely mention them in passing, originals with translations, sonnets and odes with narratives, closet drama and verse essays, serious with jocular (George Daniel's Democritus in London, for example, published with 'Notes Festivous' in 1852, in which a 'City Madam' appears as the Venus of Cheapside, a 'Venus very / Vain of her Adonis Jerry!'). The value of such findings might be very unclear at first (it does, however, look at first glance as though the theme declined in popularity in the 19th century, since it appears fewer times in proportion to a much larger number of poems) but one thing is certain: only the technology of the database allows such findings to be made in this form at all. In the physical library, with the actual books, it would take weeks, at the least, to produce the above list to the same degree of accuracy and completeness.
But the question of value is after all the crucial one: not what technology can do, but what you can do with it. 'We are in great haste to construct a magnetic telegraph from Maine to Texas,' Thoreau remarked in Walden a hundred and fifty years ago, 'but Maine and Texas, it may be, have nothing important to communicate.' And here I return to the question of scale. Suppose we think of the general descriptions on which our notions of cultural history and development are founded. 'Nineteenth-century poetry, we realize, was characteristically preoccupied with the creation of a dream-world.' That's E R. Leavis, writing in New Bearings in English Poetry in 1932. This is an extreme example, admittedly, and one not helped by its tone, but we in turn should realize how uncomfortably close it is to the way we habitually think about such matters. It may be objected that life is short, but the English Poetry database has greatly lessened one obstacle to the formation of valid general pictures of what the poetry of a given period is like. It may also be objected that the Leavis example is beside the point, because the staggering inaccuracy of his judgment is apparent without recourse to the big measuring-stick of English Poetry. But the point goes deeper than that. Leavis might reasonably defend his view by arguing that what he means by nineteenth-century poetry is, in fact, a truer definition of it than one given by the merely mechanical product of bibliographical and technological processes represented by the English Poetry database. 'The poetry of Shelley, Arnold, Tennyson, and Swinburne is nineteenth-century poetry for us' (I imagine Leavis insisting) 'because we are not concerned with quantity but with quality, and quality is measured by canonical presence and influence.' This really is an argument about what properly forms the material of literary study I confess to divided feelings here. I think some poets and poems better than others, I believe that there should be a canon and that its transmission is one of the functions of the academy. But I also believe that access to the mass of poetry is a precious supplement and corrective to the ways in which that canon is interpreted. In Browning's Sordello the hero finds, in the 'triple line of trees' surrounding a palace in Ferrara, the emblem of a political vision:
Real pines rose here and there; but, close among,
Thrust into and mixed up with pines, a throng
Of shrubs, he saw, -- a nameless common sort
O'erpast in dreams, left out of the report
And hurried into corners, or at best
Admitted to be fancied like the rest.
Reckon that morning's proper chiefs -- how few!
And yet the people grew, the people grew,
Grew ever ...
The 'real pines', the 'proper chiefs' of our poetry are 'mixed up', in the database, with the 'nameless common sort' (though here named, at least), who are certainly 'left out of the report' which we make to ourselves not only about what poetry is, but why it matters and should continue to matter. There is more bad art than good, but there is no good art without bad. The kinship of poets is as strong as what makes one of the family great and the others nonentities. The database offers at least the chance to explore these networks of kinship.
I have taken an exalted view of the 'common sort' who populate the database and I am conscious that this commonness is a matter of degree. The group of published poets is small in relation to the group of those who write poetry, especially in later periods, and the group selected for inclusion in NCBEL is smaller still. The publishers and editorial board of the database (of which I was a member) took a hard line about expanding the parameters of NCBEL, a policy which I would defend while regretting some of its consequences, in particular the perpetuation of NCBEL's male bias. This bias can be illustrated by two small examples. Richard Cobbold, whose mediocre poem on Venus and Adonis I mentioned above, appears in the Dictionary of National Biography not as a poet but as a novelist, and he is there described as the son of Elizabeth Cobbold; she too has an entry in the DNB, and in contrast to her son she figures as a poet who also wrote a novel (her collected poems were published in 1825). Although Richard was primarily a novelist who wrote poems he gets into the database, while his mother, primarily a poet who wrote fiction, does not.
The second example came to my attention because of my interest in the courtship correspondence of Robert Browning and Elizabeth Barrett.2 In the course of this correspondence the two poets discussed, with somewhat callous wit, the inferior poems of two admirers of theirs, who happened to have virtually identical names. William Cox Bennett (1820-95) wrote fan letters to Browning and poems in praise of him; Georgiana Bennet (1812-76) wrote in similar terms to Barrett Browning, who received her letter with sardonic relish:
Miss Georgiana Bennet -- did you ever hear of her? .. I never did before, but that was my base ignorance, -- for she is a most voluminous writer it appears .. & sent me five or six 'works' (observe), .. published under the 'high sanction' (and reiterated subscription) of ever so many Royal Highnesses & Right Reverends ... written in prose & verse, upon female education and the portrait of Harrison Ainsworth ("I gaze upon that noble face, & bright expressive eyes!") .. miscellaneous subjects of that sort! -- also, there is a poem of some length, called "The Poetess," which sets forth in detail how Miss Georgiana Bennet has found the laurel on her brow a mere nightshade, & the glories of fame no comfort in the world.
In another letter Barrett Browning mocks Bennet for her absurd ambition: 'As my Bennet says (Georgiana) when she catches vehemently at the laurel .. "I WILL NOT BE FORGOT." ... "I must die .. but I WILL NOT BE FORGOT" (in large capitals!)!' As for William Cox Bennett, Browning learned to recognize his handwriting on the envelopes, so that, as he drily told Barrett Browning, 'the contents give me no trouble'.
As it happens, William Cox Bennett may be found in both the DNB and English Poetry, where I have read his poems in praise of Browning (they are quite interesting, especially when read alongside poems which refer to Tennyson, Barrett Browning, Carlyle and other writers) but Georgiana Bennet has disappeared from view. Her determination not to be forgot has been answered only in the sense that she is a footnote in the story of people who really have achieved 'the glories of fame'. One would like to know whether 'The Poetess' (published with other poems in 1844) really is as bad as the line which Barrett Browning quotes, but one has to go to the British Library in the old-fashioned way to find out.3
Such limitations are mitigated by the possibility of supplements to the database. The electronic library, like other libraries and unlike single books, is not a sealed system and will very likely be expanded in the future to broaden its coverage. Making the database larger and more comprehensive, will enable it to give a better picture of the scope of its subject. But other aspects of this picture will not be affected by the addition of more material. On the contrary, they will be exacerbated, for better and worse, and in particular the aspect of juxtaposition.
The database powerfully emphasizes the textual identity of the poems it contains, especially when they are set alongside each other as the result of a keyword search. This point was made to me by Christopher Ringrose, of Nene College, in a conversation about a study I did recently of a group of mid- 19th century poems with the word 'mother' in the title or first line. I found (after some sifting) 213 poems of all shapes and sizes, which could be analyzed according to some striking patterns of resemblance, but which also exhibited great variation -- as great as that of the Venus and Adonis group -- in terms of genre, length, tone, and so on. My choice of keyword, Mr Ringrose pointed out, had created a context, an environment for these poems which overshadowed every other context, either in relation to their authors or to the books from which they had been magically lifted. The existence of poems as personal or social utterances or documents was suspended, so to speak, for the duration of an inquiry which considered them in a kind of abstract or virtual space. True, some of this missing context could be reinstated (starting with the bibliographical information about the source of the text provided by the database itself) but Mr Ringrose's point, I think, was that the psychology of the database discouraged such an effort. I am sure he is right, but I am also persuaded that this feature of the database may be considered as much a strength as a weakness. I recall Proust's expression of scepticism at the fashion for displaying works of art in 'period' surroundings, 'recreating' the environment in which they were originally produced and viewed, a fashion which coincided with an attack on museums as abstract, defamiliarizing spaces. This was precisely what Proust valued in them: the blank walls of a museum liberate a painting from its context, concentrate our attention on the essential quality which context obscures or devalues. The database is such a museum, or rather is a plurality of such museums, for each search creates a white-walled gallery in which texts hang like pictures, with only the briefest of descriptive captions. The value of this is not (or not necessarily) the aesthetic value with which Proust was preoccupied, the singularity of artistic genius whose encounter alone redeems our life; it is not so grand, but it is still of great importance. We too have a passion for context, for explaining and interpreting literature in terms of the conditions in which it was produced, whether those conditions relate to the life of the writer (still the most popular approach in literary biography) or to ideas and values at work in other discourses and other social practices, the approach we call 'new historicism'. English Poetry allows by contrast a critical approach which I would call 'new formalism'. It is new because its primary object of attention is not a work of art but a pattern made up of works of art (often of fragments), a pattern moreover which is not 'there' but which is created by critical speculation. Nevertheless the pattern created by a search of the database is predominantly formal. The screen itself, the medium of the search, with its shifts and dissolves, its associative leaps and couplings, gives the texts a formal energy and capacity to surprise which I have found both disconcerting and exhilarating.
The capacity to surprise is perhaps the most attractive feature of all. Recently a colleague asked me to use the database to find the phrase 'bright particular star'. I had just started the search when I remembered that it comes from All's Well That Ends Well (Helena says it of Bertram, whom she loves but who is out of her reach) and that it would therefore not appear in the search, since the database doesn't include Shakespeare's plays (they are included in Chadwyck-Healey's English Verse Drama). Nevertheless I finished the search, and gave the results to my colleague, who found herself interested in them for an unexpected reason. We were both surprised to find just a small number of occurrences that all came in the 19th century -- indeed, after 1830. This evocative Shakespearean phrase, unlike so many others, had journeyed across the 17th and 18th centuries without finding its way into a single poem, had left no trace in Romantic poetry, and even after that had only caught the attention of eight poets (among them Christina Rossetti and Swinburne, admittedly).4 Might the reason lie in the obscurity or unpopularity of the play? I suspect that there is a lot of work to be done in mapping the variations in Shakespeare's 'presence' in English poetry (or Chaucer's, or Milton's). I have already noticed, as one of the results of another search, that references to Hamlet far outweigh those to any other Shakespeare character or play in the mid-19th century.
We are at the threshold of the uses to which the database will be put. Some of these uses are already apparent: they are mainly the practical ones, such as the accessibility of texts for teaching purposes, and the opportunity to assemble customized or themed collections. It will be appreciated that the database functions at the most basic level as a vast concordance and dictionary of quotations, a function we should not be too embarrassed to ask it to perform, as though we were asking a genie to tie our shoelaces. The genie doesn't mind. But it can do more complicated and more complex things, some of which I can guess at and some of which I haven't the wit to imagine. The material in the database is divided into elements, arranged in a hierarchy: the poet's collected works, the individual works within that collection, and the structural features of each work: its division into parts, its stanzas or irregular groups, and its individual lines. Each element is tagged with a separate code and each interconnects with others in an intricate system that can be activated in a multiplicity of ways. The 'elements' of the database are the subject of a series of experiments whose principles are still in the process of being worked out. Some of these experiments will result in nothing more than a loud flash and bang. Others will have more lasting and productive outcomes. Literary criticism, like most intellectual disciplines, is none the worse for a touch of the unpredictable.
Daniel Karlin
University College London
[2] I shall refer to her subsequently as 'Barrett Browning'.