Retaining Traces of Composition in Digital Manuscript Collections: A Case for Institutional Proactivity  –  Millicent Weber


Abstract: Studies of digital manuscripts generally focus on the technical capabilities of collecting institutions, digital storage and preservation, recovery of corrupted or out-dated material, and provision of access. The potential content of future, digital or part-digital, collections, and their capacity to support sustained scholarly research, has been comparatively neglected by scholars and archival institutions alike. In response to this shortcoming, this paper presents a study into the potential content of future collections of poetry manuscripts and their capacity to support research into the process of composition. To predict this capacity, this paper compares a study of compositional process, using handwritten and typewritten manuscripts, with a small-scale survey of early-career poets’ compositional habits. The manuscript study used the draft manuscripts of three poems by the poet Alan Gould and three by the poet Chris Mansell to describe each poet’s compositional habits, while the survey component of the project obtained information about the drafting practices of 12 students of creative writing and poetry at the University of Canberra. This study identified five attributes of the manuscript collections necessary to support research into compositional process: the quantity of information; the completeness of information; the sequence of drafts; the preservation of material as it was created; and the availability of contextualising material such as letters. The survey showed that the majority of these attributes were only partially displayed by young writers’ draft manuscripts, but also indicated these writers’ interests in increasing and preserving the research value of their manuscripts. While the scale of this project is niche, the results are transferrable to the extent that they indicate the diversity of manuscript collections currently being created, and emphasise the importance of archival institutions adopting a more active advocacy role in encouraging writers to create and maintain comprehensive and well-organised collections of digital manuscripts.



‘Multiple, bastard, sequestered… uncouth script, black letter type’ (Nielson, 1993: 55): these are attributes of the handwritten literary manuscripts found in research libraries and other archival institutions. The relative complexity and, in some cases, inaccessibility of manuscripts to the casual reader is for the researcher an intrinsic part of these manuscripts’ value. The prominence of certain terms over others, the myriad revisions, the format of the work on the page and even the punctuation enable exploration of the creative and compositional process of writers. Such research is able to describe the genesis and the development of a particular work, and can even offer insights into the life, personality and psychology of literary figures. In the age of computers that are portable and affordable, there arises a concern that the attributes of these manuscripts that make them so useful for research may not be retained in future manuscript collections that are wholly or partially digital: digitally generated files may be corrupted or not retained, or may not contain the same kinds of information as handwritten manuscripts. This article explores how writers’ practices are changing, and the implications that this might have for researchers. Identifying connections between format and content, this research demarcates the responsibilities of both writers and collecting institutions in preserving valuable research material in a digital environment.

Background and Existing Research

Archival institutions are on the cusp of digital collecting, with institutions just now beginning to collect digital manuscripts and manuscript collections, and to provide research access to these collections. In recognition of this, scholars and archivists have begun exploring the practicalities and requirements involved in collecting this material, caring for it, and making it accessible to the public. Questions of technical capability of institutions, digital storage and preservation practices and standards, recovering corrupted material, and delivering content on- and off-site, are hugely important to continuing to uphold existing archival standards, and have accordingly garnered significant attention. They are a response to the everyday demands of this changing climate, and in many ways represent a transferral of traditional archival practices—acquisition, care, and access—over to a new, digital environment.

In his article on digital research material, David Zeidberg (1999) sketches a brief outline of some of the issues that libraries and other collecting agencies would need to address as ‘born-digital’ material—documents, websites, emails, ebooks, and other items created and accessed almost exclusively on computers—becomes an increasingly larger part of the material that society produces. Zeidberg’s insights are particularly relevant because he raises the issue of digital collection material with respect to the researchers’ needs. In advocating for greater attention to be focused on issues such as the authentication, preservation and accessibility of digital documents, there is nothing pioneering in his predictions: these, along with intellectual property, are central issues canvased in the majority of literature in the field of digital library material (Graham, 1998; Guercio, 2001; Hodge & Anderson, 2007; Zhou, 2010; Heidorn, 2011). This preoccupation with the technological issues associated with the management of digital collection material is not surprising, considering that these are practical issues that need to be addressed for appropriate infrastructure and policy to be developed to integrate this material into collections. Alongside these concerns, however, Zeidberg also raised the issue of the quality of the material that would be provided to the researcher, in terms of a comparison between handwritten manuscript drafts and their digitally created equivalents. Specifically, Zeidberg mentions concerns regarding preservation of versions and revisions and, through giving an example of a pivotal change of wording in a letter written by George III, observes: ‘Were George III to have had access to a word processor for composing his letters and documents, would we have ever seen this change or have had the opportunity to interpret the feelings behind the words? We probably would only see the final version, if even that were preserved.’ Unfortunately, he does not pursue this line of enquiry further. While there are other resources in the field that express the objective of exploring users’ research and their experiences, these are generally based on assessment of technological rather than content-based concerns (Hedstrom, Lee, Olson, & Lampe, 2006; Qayyum, 2008).

A complementary viewpoint to Zeidberg’s is that of Roy Rosenzweig (2003), which asserts the importance of the role that the creator of digital documents holds in the archival process. Rosenzweig of course reiterates the importance of authentication, preservation, and intellectual property issues, but also highlights the importance of encouraging individuals and organisations to take a more active approach in creating and maintaining their own collections of digital material as they are created. While Rosenzweig’s support for ‘grassroots’ style archival practices is advocated to mitigate against technical collection issues—mainly, the difficulty of access and preservation for a decades-old collection of digital documents, compared to a physical collection of the same age—the message also resonates with collection issues surrounding the quality of the research material produced. In the terms of Zeidberg’s hypothetical computer-literate George III, a well-regulated system of individual archival practice might have meant that both the original, and the amended, versions of the letter were saved to the hard drive and could be later transferred to an archive.

The British Library’s Digital Lives research project, led by Williams, Dean, Rowlands & John (2008), presents a stark contrast to the largely technologically-focused articles that make up the majority of those within this discipline. These researchers investigated the ways in which people create their own personal digital collections: their interactions with computers, the collection of documents that they create, how this collection is managed and used, and the way that this compares to the person’s storage and use of physical documents. By interviewing a range of people in different professions and age groups, this project collected valuable data about the material that might be entering institutions’ digital collections over the coming decades, and the ways that this material will already be organised. Some of the insights that this study provides were encouraging, particularly in relation to the quality concerns raised by Zeidberg: there was evidence of successive computer drafts being preserved, or even of drafts being printed for editing by hand.

The considerations canvased in each of these articles provide a general guide for a range of the issues that will need to be addressed as digital collections achieve more prominence within the archival structure. But this research is predominantly institutional in outlook: concerns centre on technological and infrastructural requirements for the storage, preservation and accessibility of digital items. When the researcher is referenced in these works, it is not primarily their use of the material for research which is addressed, as much as their access to that material. In approaching the considerations of digital collection management from the perspective of the researcher, this paper will therefore be able to provide better insight into the way that this material will be used – not just accessed – and the ways in which this use can best be facilitated by institutions and individuals alike.

Research Design and Data Collection

Recognising the strengths and limitations of the research conducted to date in this field, this article investigates the question: “How can we ensure that information about writers’ compositional process is preserved as writers migrate from physical to digital manuscript production?”

This question can be broken down into two distinct lines of enquiry. The first explores the insights that a researcher can gain into compositional process when comparing and analysing wholly physical collections of manuscript drafts of a literary work—the traditional manuscript collections incidentally created by writers as part of their practice, and subsequently collected by libraries and archives and made available for scholarly research. This initial line of enquiry relates to the theories and methods underpinning manuscript research as it is currently undertaken, asking how meaning is inscribed and discovered in manuscript collections. The second line of enquiry examines the partially- or wholly-digital manuscript collections that are currently being created by writers, and the extent to which these collections will offer similar insights into the writer’s compositional process.

Rather than positing an opposition between hardcopy and digital manuscript collections, this research seeks to identify the extent to which existing manuscript research—currently constructed with almost sole reference to physical manuscripts—maps onto a contemporary, hybrid situation, and to identify challenges that this may pose, and actions that can be taken, to ensure that researchers are supported. This article is framed by this focus on existing research collections, as it offers a productive and pragmatic means of investigating this particular aspect of digital manuscript collecting, but this does not diminish the significance and value of digital manuscript collections. Digital collections pose the challenges that this research identifies, but they also support research methodologies and provide important and unique material that paper-based collections cannot. The relationship between physical and digital collections is complex, collaborative and ongoing, and this article exploits one particular aspect of this relationship to explore the connection between format and content, and support the work of collection managers and researchers.

In order to explore these issues, the handwritten and typewritten manuscript drafts for a number of different poems were analysed and compared according to the methodology outlined by contemporary manuscript researchers such as Bushell (2003) and Davies (2008), and the process of composition of each poem was described. This initial research was then self-reflexively analysed to identify which of the various attributes of the collections being studied were necessary to form an understanding of the writers’ composition. Finally, a survey was distributed to a group of emerging writers to gauge the extent to which the attributes that informed the initial manuscript research are likely to be present in the collections of manuscripts that will enter archival institutions in the future. It also asked for respondents’ opinions on the possibility of creating research collections, and their receptivity to libraries’ encouragement of good archival practice. This survey emulated the work undertaken by Williams et al (2008) as part of the British Library’s Digital Lives project, but was re-designed to specifically address those attributes identified in the earlier stages of this research. The following is a brief outline of the specific methods used in this research:

Stage One: Initial Selection of Manuscripts

Following consultation with the manuscript librarians at the National Library of Australia, three poems each by Alan Gould and Chris Mansell were chosen for the initial manuscript analysis. These poets and poems were selected because they satisfied the following criteria:

  • The collections of drafts for each poem were considered to be complete, and were held in full at the National Library of Australia (in Gould’s and Mansell’s papers, MS 6635 and MS 7904).
  • The collections of drafts for each poem demonstrated a range of different compositional techniques.
  • The drafts were all created by hand or using a typewriter (no digital technologies were used).
  • Although writing at similar times, each poet has a distinctive and contrasting style.
  • Each poet was locally-based and available for interview to confirm the findings of the initial manuscript research.

Stage Two: Research into Compositional Process

The manuscripts for the poems chosen were then analysed to determine the process whereby each poem was composed. This was undertaken as follows:

  • Each draft was photographed and transcribed to provide better access in the analytical phases (see de Biasi, 2004; Bushell, 2003; and Davies, 2008)
  • A chronology of drafts was established (as per Bushell, 2003; Davies, 2008; mostly already completed by the National Library of Australia)
  • The changes within each manuscript draft and between subsequent drafts were listed and described. Changes that were identified and recorded included changes on the scale of individual words, larger changes such as the addition of a new theme or idea, and changes to the structure or arrangement of the poem. These changes correspond to the different levels of intention that Bushell (2003) describes.
  • These changes were then studied to determine patterns, building up a descriptive narrative of the composition of the poems. The compositional processes behind the different poems were compared with one another to describe each poet’s working habits and compositional style.
  • Finally, brief interviews were undertaken with each of the poets to confirm that these descriptions were consistent with their practice.

Stage Three: Identifying the Important Attributes

After this initial research was conducted, the important attributes of the draft manuscripts that contributed information about each poem’s composition were identified. This involved linking the statements about composition which were already made back to the concrete evidence that supported them, and relied heavily on the detailed descriptive work undertaken as part of the initial manuscript research.

Stage Four: Survey

The fourth stage of this research comprised the distribution of a survey to a group of emerging writers. This survey asked about writing practices and the storage of drafts, and was designed to determine to what extent those important attributes identified in the previous stages of research would still be present in the partially- or wholly-digital manuscript collections that will be available to researchers in the future. This survey was designed in recognition of the emphasis that other researchers (like Rosenzweig, 2003, and Williams et al, 2008) place on the role of the creator in ensuring digital manuscript collections are created and managed in a manner which ensures their future viability for scholarly research. The survey was administered to the creative writing cohort at the University of Canberra, and was similar to that conducted by Williams et al (2008), but with a focus on questions related to creative writing. The survey covered:

  • The extent to which composition is undertaken on the computer (or other devices).
  • The extent of computer skills and formal training.
  • The creation of different drafts of documents.
  • The filing, storage, and deleting of digitally-created and hardcopy drafts.
  • The backing up and archiving of digitally-created and hardcopy drafts.
  • Writers’ opinions about the possibility of creating research collections, and receptivity to libraries’ encouragement of good archival practice.

The results of the survey were then collated and compared with the data from previous stages of the research, and a series of recommendations were drafted that could potentially be distributed to writers, offering simple suggestions to promote the future research value of their future archives.

Findings and Recommendations

The research value of future manuscript collections will be determined by the presence or absence of those attributes of individual manuscripts and manuscript collections that were identified in the initial stages of this research project. These attributes can be collected together under five general traits that, when present, will assist detailed manuscript analysis to yield optimal results. These are: the quantity of information; the completeness of information; the sequence of drafts; the preservation of material as it was created; and the availability of contextualising material such as letters or creative stimuli. The following section of this paper connects each of these traits to the findings of this research project, and suggests some strategies that writers could use to ensure these traits are maintained in their collections of drafts. Practicing these strategies is the preserve of writers, but as this article demonstrates, ensuring writers’ practices incorporate these strategies is a challenge for collecting institutions to address. Proactive institutional intervention into writers’ storage becomes increasingly important as the permanence and fixity of draft material decline.


The first of these attributes is the quantity of information present: both the quantity of drafts, and the quantity of revisions contained within each draft. The two manuscript collections studied demonstrated this quantity of information in different ways. Chris Mansell made fewer drafts of her poems, but each draft contained myriad revisions and re-revisions. Alan Gould created less closely-annotated drafts, but the number of drafts he created—up to 20 drafts per poem—meant that these sets of drafts were each equally able to reveal the compositional process of their respective authors.

Half of the writers who responded to the survey indicated they do generally create a similar number or more drafts writing on the computer. It can therefore be hoped that this quantity of drafts would be as useful to researchers as copious handwritten revisions. Equally, all respondents indicated that some of their drafting is always done with pen and paper, particularly plans and early drafts. The problem here is therefore also one of keeping drafts of different formats together, and putting these into a sequence. Keeping this in mind, it is important that  when working on the computer writers frequently save new drafts, ideally whenever changes are made to their work, and/or use ‘Track Changes’ or a similar program or function which preserves any altered or deleted information. Writers should also be careful to retain any drafts that are printed or transferred to a tablet or another computer to be revised.


The second of these attributes is the completeness of information present: all drafts exist and are available for use, and all information is included in these drafts (nothing has been erased, destroyed or otherwise made illegible). This attribute of manuscript collections is connected to the researcher’s confidence in making conclusions about the progressive development of a poem from the manuscript drafts. With pen and paper writers might cross out, white-out or erase words, but these each leave a trace. With the computer, it is impossible to know what has been typed and then deleted. As the majority of survey respondents indicate they create some drafts by hand, this issue is perhaps not as severe as it could be, but on the flipside, half of writers admit they are at present unlikely to save a separate draft every time they make changes to their poems. To ensure collections are as complete as possible, writers should be encouraged to save separate drafts whenever changes are made and/or use track changes or a similar function, as per the previous recommendation.

It is equally important that no drafts are missing from the collection. As such, writers should make an effort to save all drafts of each poem in the same location, and to choose locations for saving drafts using some kind of system. Where any of these measures is not possible, or not taken, writers would ideally record any incompleteness in their collection (for example by keeping a log of their drafting practices, or writing a comment on a draft to indicate that it was preceded by a different draft in another location or format, or to indicate that parts of it were deleted).


The third of these attributes is that the sequence of drafts can be reconstructed: there is sufficient information within the drafts, or the drafts are numbered or ordered by the writer, so as to demonstrate the sequence in which they were created. Gould’s manuscript drafts were able to be ordered through a close examination of the phrasing and development of ideas in each draft. This process was possible because of the completeness of the collection of drafts, with many new elements introduced first as an amendment to the previous draft, and then incorporated into the body of the work in subsequent versions. Mansell ordered, numbered and stapled together her earlier drafts, which meant this careful sequencing was only necessary with her later versions.

While we can encourage writers to retain relatively complete information to assist researchers to sequence their drafts (as per the previous point) it is also important that writers include sufficient information with their drafts to assist researchers to determine their order of writing, whether in the title or content of the document, as numbers, dates, or other descriptions, and comprehending handwritten and tablet-created documents as well as drafts created on the computer.


The fourth of these attribute is that drafts are kept in the format in which they are created: This refers to the fact that a typed copy of a handwritten draft would be considered a separate and later draft to the original handwritten draft. The responses to the surveys indicated that a number of drafts are still handwritten, and that drafts are also written on tablets and other electronic devices, and are also often printed from the computer to be marked up by hand. It was an interesting and informative aspect of Gould’s and Mansell’s manuscripts that different drafts were written by hand, written with a typewriter, or included handwritten notes, revisions and additions on typewritten pages; different formats presented different content, but equally the choice of format hinted at the poet’s beliefs about the readiness of the poem to be typed up, or their difficulties with certain sections evident in a decision to return to handwriting.

Recognising this, it is important that handwritten and tablet-created drafts are retained in a format that preserves all contextualising information from the original draft. With handwritten drafts this is likely to be as simple as keeping the originals or scanning them to the computer; with drafts created on a tablet or other relatively short-lived electronic device, writers should ensure they use an application which allows documents to be exported from the device in a format readable on other devices (e.g. as an image or PDF document).

Supplementary Material

The fifth and final attribute is that supplementary material which inspired or influenced the work or otherwise provides context for the work is kept. It was important to the interpretation of the stimulus and development of Mansell’s poem ‘Goodbye Blue’ that the words to Leonard Cohen’s ‘Chelsea Hotel’ were included, as hastily written out by Mansell, in the collection of drafts. Other scholars also note the importance of including similar stimuli in manuscript collections; for example, Kidder’s (1980) analysis of EE Cummings’ poem ‘I am a little church’ was entirely dependent on his discovery of a photograph of the church which inspired Cummings’ reflections.

Other manuscript analysis has used correspondence or personal musings as contextualising information into a poet’s state of mind or inspiration, or to qualify their assumptions, in a similar manner to the interviews conducted with each of the poets as part of this research project. Encouraging writers to keep copies of relevant email correspondence or blog posts about their writing, as well as saving or noting any stimulus to their work alongside their drafts, would assist with this.

Survey Results: Further Discussion

The survey results indicated that poetic manuscript collections entering archival institutions in the future will be diverse and complex, often including both digital and physical manuscripts created on and accessed with a range of different devices. Every writer who responded said that they write both by hand and on the computer, and several of the writers also use their tablet or mobile phone for their writing. Not a single writer used just one format to create their drafts. We might infer that, as the range of technological devices increases, the diversity of different draft documents increases also. Hence, rather than addressing a situation where manuscripts are becoming bland and shallow, a far greater variety of different documents are being included in collections. As such, these collections have the potential to display equal or greater research value to their predecessors, but equally, gaps in collections caused by the ease of overwriting or deleting drafts and the potential for technological failures means information about the writers’ processes could be irretrievably lost.

Two thirds of writers surveyed indicated they would be encouraged by the prospect of future research potential to improve their archival practices, with multiple respondents commenting that it was something to look forward to in their careers as writers. Those who said this would not encourage them to create and keep more drafts gave two reasons. The first reason was that they did not believe their material was of a high enough standard: this implies that either they do not intend to become professional writers (and as such from a literary research perspective their manuscripts will be of limited use), or that when their writing does reach this standard they may then become open to advice in this area. The second reason was that they would not want to make draft documents available for researchers, as they would prefer the final version to stand alone as a single artistic unit. There is little that researchers can do to influence this kind of opinion, other than hope that, like in the case of Patrick White[1], threats to burn all manuscript material are hollow.


Writers’ practices are changing, and along with this the kinds of manuscripts that they are creating. Through combining an analysis of physical manuscripts with an exploration into contemporary writing practices, this research project identified changes in writing practices which are likely to present a problem, and detailed ways in which these might be mitigated.

The findings from this project indicate the receptivity of writers to improving the research quality of their drafts, and support more general arguments that other researchers (such as Rosenzweig, 2003) have made about creators needing to take more responsibility for the ongoing development and maintenance of their collections in an increasingly ‘ephemeral’ digital environment.

These findings indicate the necessity for archivists and librarians to advocate for creators using better practices to create and manage their drafts years before they are even considered for inclusion in manuscript archives, and the possibility for this advocacy to have a real impact of the research quality of future manuscript collections.

Finally, these findings indicate a broader need for all of us to take a proactive approach to ensuring that the incredible research value of our current manuscript collections is carried through to those collections entering archives in the future: to think about the ways in which format and content are inextricably linked, and to work productively alongside content creators to respond to the challenges and embrace the opportunities of a digital environment.


Works Cited

Bushell, Sally. “Intention Revisited: Towards an Anglo-American “Genetic Criticism”.” Text 17 (2003): 55-87. Print.

Davies, Alexandra. “Poetry in Process: The Compositional Practices of D.H. Lawrence, Dylan Thomas and Philip Larkin.” Diss. University of Hull, 2008. Print.

de Biasi, Pierre-Marc. “Toward a Science of Literature: Manuscript Analysis and the Genesis of a Work.” Genetic Criticism. Ed. Jed Deppman, Daniel Ferrer, and Michael Groden. Philadelphia: University of Pennsylvania Press. 2004. 36-68. Print.

Gould, Alan. Papers of Alan Gould, MS 6635. c.1972-2004. Canberra: National Library of Australia. MS.

Graham, Peter S. “New Roles for Special Collections on the Network.” College & Research Libraries 59.3 (1998): 232-240. Print.

Guercio, Maria. “Principles, Methods, and Instruments for the Creation, Preservation, and Use of Archival Records in the Digital Environment.” Trans. Kenneth Thibodeau. The American Archivist 64.2 (2001): 238-269. Print.

Hedstrom, Margaret L., Christopher A. Lee, Judith S. Olson and Clifford A. Lampe. ““The Old Version Flickers More”: Digital Preservation from the User’s Perspective.” The American Archivist 69.1 (2006): 159-187. Print.

Heidorn, P. Bryan. “The Emerging Role of Libraries in Data Curation and E-science.” Journal of Library Administration 51.7/8 (2011): 662-672. Print.

Hodge, Gail and Nikkia Anderson. “Formats for Digital Preservation: A Review of Alternatives and Issues.” Information Services & Use 27.1/2 (2007): 45-63. Print.

Kidder, Rushworth. “Picture into Poem: The Genesis of Cummings’ “I Am a Little Church”.” Contemporary Literature 21.3 (1980): 315-330. Print.

Mansell, Chris. Papers of Chris Mansell, MS 7904. c.1970-2004. Canberra: National Library of Australia. MS.

Nielson, James. “Reading between the Lines: Manuscript Personality and Gabriel Harvey’s Drafts.” Studies in English Literature, 1500-1900 33.1 (1993): 43-82. Print.

Qayyum, Asim. “Analysing Markings Made on E-Documents.” Canadian Journal of Information & Library Sciences 32.1/2 (2008): 35-53.  Print.

Rosenzweig, Roy. “Scarcity or Abundance? Preserving the Past in a Digital Era.” The American Historical Review 108.3 (2003): 735-762. Print.

Williams, Pete, Katrina Dean, Ian Rowlands and Jeremy Leighton John. “Digital Lives: Report of Interviews with the Creators of Personal Digital Collections.” Ariadne 55 (2008): n. pag. Web. 7 Mar 2016.

Zeidberg, David S. “The Archival View of Technology: Resources for the Scholar of the Future.” Library Trends 47.4 (1999): 796-806. Print.

Zhou, Yongli. “Are Your Digital Documents Web Friendly?: Making Scanned Documents Web Accessible.” Information Technology & Libraries 29.3 (2010): 151-160. Print.


[1] See Patrick White’s 1977 letter to the then Director General of the National Library of Australia Dr George Chandler, held in the NLA collection (number MS 8469) or available for download through the NLA ‘Media Zone’ website:



Millicent Weber is a PhD candidate in the Centre for the Book at Monash University, Melbourne. Her PhD research forms part of the Australian Research Council Discovery project ‘Performing Authorship in the Digital Literary Sphere’, and investigates audience experience at literary festivals. Millicent has worked with manuscript collections at the National Library of Australia and at the University of Melbourne Archives. Millicent’s honours project, completed at the University of Canberra, investigated the value of digital manuscript collections and the challenges that they pose for researchers and collection managers.