Untitled Document

Robert Graves Diary Project : Description of Encoding (March 30, 2006)

I. INTRODUCTION

The primary aim of the mark-up segment of the project is to increase access to the diary of Robert Graves, which records his daily life and work from 1935 to 1939, while retaining the integrity of the original manuscript held in the University of Victoria Libraries Special Collections. The electronic edition will enhance and expand accessibility to students, scholars and the general public in the following ways:

- by creating an abstract for each month of the diary, which provides context for the reader at any point, as well as a summary of the diary’s contents.

- by linking the majority of names, places and titles mentioned in the diary to a reference file: the reader can access identification and bibliographical citations by a simple mouse over as and when needed.

- by introducing annotation by William Graves and Kenneth Gay (formerly Karl Goldschmidt). Kenneth Gay was Robert Graves’ personal secretary, who worked closely with Graves for most of the diary period, and whose annotated copy of the typescript was consulted. These notes identify many names, places and writings mentioned by Graves and provide further context for diary events, including occasional contemporary anecdotes and commentary. The diary transcription annotated by Kenneth Gay is at St. John’s College Oxford in the Robert Graves Trust Archives.

- by reproducing the Robert Graves manuscript as closely as possible, indicating substantive changes such as additions, deletions, and misspelled words as well as accidentals. When crossed out passages are replicated in the transcription, or when a phrase is marked “unclear” for example, readers are alerted, and can refer back to the manuscript image which is available for viewing on the same page as the diary entry. Readers are also informed when a drawing or other image occurs in the manuscript.

- by expanding abbreviations, and supplying missing letters or punctuation to provide clarity

- by including previously unpublished ancillary material such as enclosures (letters, clippings, cards, posters and photographs) and the logs of letters written (from 09-1937 on), which enhance the reading of the diary, as well as its research potential.

Digitization of the Graves diary manuscript will facilitate comparison of the original text with the transcribed version: readers can select a page and view it alongside the transcription online.

Searchability: Searches may be conducted on any word in the diary entries, logs, enclosures or abstracts. This includes names, places, or titles which may or may not be highlighted. Eventually, it is hoped that a mechanism will be in place to search names and places by code, so that abbreviated words and accented characters will be picked up in searches.

II. DESCRIPTION OF ENCODING: Diary, Logs, Enclosures and Abstracts

Note Display and the Reference File

The majority of names (people, organizations) and places mentioned in the diary are identified, either with a highlighted link to a reference file, or with footnotes in the text. Readers can access a note display for most names and places at any point in the diary, since they are highlighted whenever they occur, no matter how frequently.

Full bibliographical references are provided for most (published) titles mentioned in the diary. Works by Robert Graves, Laura Riding and associates are marked with a highlighted link to the reference file. Unpublished titles, and titles that changed before publication are identified. Citations for works by other authors will be located in footnotes.*

The Titles database is more complex than Names and Places because

1) Works by Robert Graves and Laura Riding take a variety of forms during this period, including poetry, novels, short stories, drama, essays and articles. They are also involved with editing and translation projects. When the genre is not clear in the bibliography, it may be indicated in square brackets after the title: No One Can Guess [poem]; A Trojan Ending [novel]; Politics and Poetry [essay].

2) Graves frequently mentions works by other writers who were part of the evolving Graves/Riding circle of friends and associates. Graves and Riding were often involved as editors and/or collaborators. Publications in this category are consistently tagged in the diary.

*Note: A third group of titles are not included in the Titles database, but are marked with a <note> in the text. Where they can be identified with certainty, a bibliographical reference is provided. Many of these contribute to a record of Graves’ reading:

- pleasure

- research

- critical evaluation

- news and current events

- literary reviews

a. FORMATTING THE TEXT:

<teiHeader>[TEI Header information]</teiHeader>
<text>
<body>
<div>
<header>[transcription of diary entry date]</header>
<p>
[transcription of diary entry]
</p>
</div>
</body>
</text>
</TEI.2>
</teiCorpus.2>

Note: This outline has not included a description of the Graves diary DTD (Document Type Definition), which is not administered by those doing textual markup, but is a vital component in the development of an electronic text. The DTD, which is placed at the head of the document, determines its parameters by telling the computer what kind of document it is, and by creating a subset of TEI that suits the purpose of that document. The TeiCorpus DTD was chosen because the diary’s contents are varied: from simple prose to verse to tables and images. In the process of validation, the document is parsed against the DTD.

<div> - marks sub-divisions within the text body. In the diary, each entry has been marked with: <div type=”diaryentry” n=”year-mo-day”>*

* The following attributes are used with the <div> element to describe and identify this section of the text:

type names the form of text to follow:

The value “diaryentry” is used for all diary text

n precedes the date value, which is used to identify the text.

<head> In the diary text, each entry (division) has a heading with the date of that entry, and any further information the author has added. The ordering of the date varies—sometimes with each entry—because it is transcribed from the original manuscript:

<head> Jan 1 1936 Wednesday</head>;
<head>Jan 3 < del hand=”RG”>Thursday</ del> Friday </head>

Graves occasionally makes corrections in his dating, and these are recorded also.

When it occurs in the ms, other information is added: e.g. “Laura’s birthday” or “All Saint’s Day” etc.

<p> - marks each paragraph (these approximate as closely as possible the author’s original paragraphing).

Since most diary entries are no more than one page, we have not normally needed to use the <pb\> (page break) element. Another encoding system is being devised for the few multi-page entries that occur – and for letter enclosures (please see enclosures below). This example of a multi-page entry is from July 23, 1936:

Aeroplane fired at. <rs type="person" key="JM1">G.</rs>
</p><pb/>


<p>returned in a hurry with only <rs type="person" key="Ju1">Juan</rs>, <foreign>Tio</foreign>
<note>Uncle eds.</note>
<rs type="person" key="Ti1">Pedro</rs>.

Additional divisions: In order to accommodate ancillary materials, additional <div>s are included in the diary text. There are two occasions when this expansion is necessary:

1. Logs: for most months between 1937-09 and 1939-05, Robert Graves included a list of correspondents and telephone calls at or near the end of his diary entries for that month. See sample of encoding below (with first and last entries only):

<div type="logentries" n="1937-09-30" id="gr01_0956_032">
<head>Log Entries for Letters from 29 August - 30 September 1937</head>
<!—the n attribute identifies the set of log entries by the date of their location in the diary, and the id attribute points to the ms image in the index of scans-->
<!—when the author has included a heading, it follows here -->
<p> <hi rend="underline">Letters</hi> and Phone-calls.</p>
<!—the individual log entries are embedded divisions: the n attribute identifies each by their date -->
<div type="logentry" n="1937-08-29"><head>Sunday. Aug 29<hi rend="sup">th</hi>.</head>
<p>To <bibl><rs type="cita" key="Tel"><hi rend="underline">Daily Telegraph</hi></rs></bibl> (letter)</p>
</div>
----------------------------------------------------------------------------------
<div type="logentry" n="1937-09-30">
<head>Thurs 30.</head>
<p>To <rs type="person" key="Du1">Dulan</rs> & Brough: ordering books. </p>
</div>
</div>

2. Enclosures: the Graves diary contains a variety of enclosures, usually adjacent to the relevant diary entry or date received. These include letters, cards, photographs, posters and news clippings. The markup for enclosures is outlined below by Martin Holmes:

Enclosures are marked up inside the diary where they are enclosed. They take this structure:

<div type="enclosure" n="1938-10-01_01">

<head>

Enclosure – Incomplete Letter from <rs type="person" key="DG1">David Graves</rs>
</head>

<div type="enclosure_page" id="gr01_1322_076_01" n="1">

</div>
</div>

b. FREQUENTLY-USED TAGS WITHIN THE TEXT BODY:

1) Names , Places: <rs> (referring string)

The majority of names mentioned in the diary are marked up. Exceptions include: names of passing acquaintances that do not reoccur (a <note> is provided instead if applicable), and minor figures who remain unidentified. Places are marked up unless they are widely familiar to a general audience (countries like France and Spain are not tagged, for example), or only occur once (such as restaurants that are not frequent haunts), are of peripheral interest only (a <note> is provided where possible), or have not been identified.

The element <rs> is combined with the following attributes in R.G. diary markup:

- type

- key

Examples: <rs type=”person” key=”AG1”> - “AG1”= Amy Graves

<rs type=”org” key=”Fa1”> - “org” = “organization”

- “Fa1”= Faber & Faber (publishers)

<rs type=”place” key=”Cas01”> - “Cas01”= Ca sa Siqui

Note about. codes: Code letters for names consist of two letters, always beginning with upper-case (the second letter can be either upper or lower) and a number. As a rule, these letters represent the initials of the person referred to (but there are some exceptions made for clarity, or when there are more than two initials).

The “org” value is treated the same as “person,” so that organizations are normally included in the Name database. Code letters for Places have three letters, and the number key is preceded by a “0.”

2) Titles: <bibl> (bibliography) and <rs> (referring string)

All titles mentioned in the diary—excluding those from Graves’ secondary reading, which are marked with a note—are tagged with the <bibl> element, with <rs> nested inside it. The referencing string was chosen instead of the <title> element because it would accommodate a link to the reference file.

Attributes used include: type (with the value “cita”), and key. The “cita” value was created to identify titles within <rs> because “cit” could not be used with the key attribute. The code contained in key=”__ “ is the link to our reference file and note display.

<bibl><rs type=”cita” key=”ICl”>I, Claudius</rs></bibl>

The code value for key is similar to those used for names and places, but there are minor differences. It consists of three letters representing the title, with no number. This creates a link with the Titles reference file so that note display will reveal the bibliographic reference for the work mentioned. Please note that the database contains a range of categories which accommodate different title levels (such as ‘m’ for monographic title, ‘a’ for poem or article, ‘j’ for journal,‘s’ for series, ‘u’ for unpublished) and types (such as ‘main,’ ‘abbreviated,’ ‘subordinate,’ ‘parallel’). There is more work to be done on bibliographies for some titles, for which the information was not readily available.

Note: Graves often underlines the titles he mentions, but not always! In keeping with the editorial guidelines, we underline only when Graves does. However, since the marked up titles are highlighted in a different colour online, it has an effect similar to italicizing, whether or not they are underlined.

3) Highlights and Quotations: <hi rend>, <emph> (emphasis) and <q>

The most common value attached to the attribute rend is “underline,” but “italic” and “sup” are also used.

An underlined title is indicated by placing the<hi rend=”underline”> tag around the title,within the <rs> string, as follows:

<bibl><rs type=”cita” key=”ICl”><hi rend=”underline”>I, Claudius</hi></rs></bibl>

Quotations: Regular quotations (“) are tagged with the element <q>; however, since this tag renders only as double quotes, we use the single quote on the keyboard when these occur in Graves’ text.

4) Emendations: < del> (delete), <add>, <unclear>,<gap>

a. Attributes for < del>: hand, type

Emendations indicated in the Graves diary markup will normally refer to changes originating with the author. When the ms shows words crossed out, they are transcribed if legible and marked with the < del> element as follows:

< del hand=”RG”>went yesterday</ del>

Since < del> is automatically displayed with a line through it, there is usually no need to describe the type of deletion, unless it is quite different from a cross-out.

b. If the crossed out portion is not legible, we use the following tag:

<unclear reason=”crossed out”/> (renders as a cross out: xxxxxxxxx), or

<unclear><gap reason=”indecipherable”></gap></unclear> (also renders as a cross out, with [indecipherable] in square brackets).

The <gap> element has also been used to describe an unaccountable space left in the ms.

The <unclear> tag marks an undeleted word or phrase that is not clear in the ms. Sometimes only a portion of a word is discernable:

<unclear reason=”incomplete”>what wh</unclear>

c. Attributes for <add>: hand, place

In most cases, <add> refers only to additions made, and indicated as such, by Robert Graves. For editorial intervention, the element <supplied> is used (see below). The place attribute takes the following values which apply to additions in the diary: “inline,” “supralinear,” “infralinear”(rare), “left margin,” “right margin” “top” and “bottom” – “supralinear” is the most common position used by Graves.

<add hand=”RG” place=”supralinear”>left on Wednesday</add>

NOTE: When an addition occurs immediately after a deletion, we omit the “hand” attribution for < del>, and allow the [RG] attribution for the <add> tag to stand for both. This approach is intended to minimize distraction for the reader. And within the diary text itself (excluding notes) the only possible “hand” is that of Robert Graves. Enclosures and abstracts include other “hands,” but they are in a separate division, shown distinct from the diary entries.

5) Editorial Interventions: <supplied>, <orig>, <abbr>, <foreign>,

<figure><figDesc>, <!—comment-->

Editorial changes are kept to a minimum in order to preserve the integrity of the original text. The element <supplied> is used for tagging missing words (or symbols, such as currency) inserted by the editor; and for supplying punctuation (only when needed for clarity). The supplied symbol or word is then displayed in [square brackets]. Note: square brackets are not used when marking up text unless they are placed inside another tag, such as <note>.

For misspelled words, we considered using <orig>, as we were advised that it would appear less intrusive in the text, however since we have not solved the difference in rendering, we have used <sic> <sic corr=” “> instead.

The element <sic> takes the attribute corr, which also allows insertion of a correction while retaining the original version: <sic corr=”received”>recived</sic>.

There are few misspellings in Graves’ text! We encounter them more often in the enclosures correspondence.

Graves frequently abbreviates names, and sometimes places. We now mark these as they are, since the note display will provide all the information needed. Other abbreviations are marked with <abbr> and its attribute expan, which displays the full word in square brackets.

The <foreign> tag marks all foreign words in the diary text, whether or not Graves has underlined them (when he has, the underline is included). Foreign words (rendered in italics) are then defined in a footnote. Proper names and places are a major exception to this rule, since they are generally tagged for note display, where they are identified.

When a sketch or other figure appears in the text, it is marked with <figure>, followed by <figDesc>, which contains a brief description of the figure. This notation (rendered in square brackets) will alert the reader to refer to the manuscript image if they wish.

6) Notes<note>

The <note> element is used without attributes in the Graves diary markup, as notes marked in this way are automatically numbered and placed at the foot of the text for that day’s entry. It is used primarily for explanatory notes and for bibliographical citations of Graves’ secondary reading.

The initials of the contributor are always placed at the end of the note: usually these would be KG (Karl Goldschmidt/Kenneth Gay), WG ( William Graves), or eds.

<note> Graves was still in hospital when the prize was announced. KG</note>

The other frequent use for this tag is to provide a translation for foreign words and expressions. It is placed immediately after the <foreign>tags, with the translator’s initials indicated.

7) Entities:* Accents, Punctuation, Currency

*PLEASE NOTE: The transition from entities to Unicode is still in process.

Accents are all presented in code: within the diary text, we use the Entity Name, not the Number.

Punctuation that requires entities is so far limited to ampersands (&) and dashes. While the shortest dash (i.e. a hyphen) can be supplied from the keyboard, longer dashes between phrases etc. are written as either – or —. The – is most frequently used, but the — is used occasionally to reflect noticeably longer lines in the ms.

Currency: is still marked with entities (primarily the pound sign) in much of the diary, but is being converted.

8) Verse: <lg>, <l>

The element <lg> identifies a group of verse lines that functions as a formal unit (such as a stanza). It will take the global attribute n to identify each stanza by number if required (i.e. if Graves has done this). The element <l> simply marks each line within the line group. The following example is an excerpt from an enclosure on May 6, 1939:

<p><hi rend="underline"><bibl><rs type="cita" key="MNi">The Moon Ends in Nightmare</rs></bibl></hi></p>
<lg type="stanza" n="I">
<l>I had once boasted my acquaintance</l>
<l>With the Moon's phases: <unclear reason="crossed out"/> I had seen her, even,</l>
<l>Endure and emerge from full eclipse.</l>
<l>Yet as she stood in the West, that summer night,</l>
<l>The fireflies dipping insanely about me,</l>
<l>So that the foggy air quivered and winked</l>
<l>And the sure eye was cheated,</l>
<l>In horror I cried aloud: for the same Moon</l>
<l>Whom I had held a living power, though changeless,</l>
<l>Split open in my sight, a bright egg shell,</l>
<l>And a double-headed Nothing grinned</l>
<l>All-wisely from < del hand="RG">the rift within</ del> the gap.</l>
</lg>

9) Lists<list>, <item>, <label> (Please see example)

The element list has an attribute (type) to further specify its arrangement, with values like

“ordered” when the list is numbered or lettered. Our example is of the simplest kind, as most of the lists in the diary are. The following example is from March 22, 1935:

<p>Planted seeds as follows:</p>
<list>
<item n="1">Mignonette</item>
<item n="2">Sweet William, edged</item>
<item n="3">Sweet William, plain</item>
<item n="4"> Ageratum</item>
<item n="5">Balsimina</item>
<item n="6"> Salpiglossis</item>
<item n="7"> African Marigolds</item>
<item n="8"> Zinnia</item>
<item n="9"> --------------</item>
<item n="10"> Wall</item>
<item n="12"> -------------- <foreign>A ver.</foreign>
<note>We shall see... KG</note>
</item></list>

10) Tables<table>, <row>, <cell> (Please see example)

<table> takes the attributes rows and cols (i.e. number of columns in the table)

<row>is elaborated with role (takes values such as “label” for descriptive information and “data” for “actual data values.”

<cell> has the attributes role, cols, and rows

Tables do not occur frequently in Graves’ diary, and when they do they tend to have a simple construction. The following example is from January 2, 1936:

<p>Corrected & sent off <bibl>

<rs type="cita" key="Foc"><hi rend="underline">Focus</hi></rs></bibl>:

<row role="data"><cell role="label"><rs type="person" key="MS1">Maisie</rs></cell><cell role="data">1</cell><cell role="label"><rs type="person" key="LL1">Len</rs> & <rs type="person" key="JL1">Jane</rs></cell><cell role="data">2</cell></row>

<row><cell role="label"><rs type="person" key="JR1">James</rs></cell><cell role="data">2</cell><cell role="label"><rs type="person" key="TM1">Tom</rs> & <rs type="person" key="JM2">Julie</rs></cell><cell role="data">4</cell></row>

<row><cell role="label"><rs type="person" key="HW1">Honor</rs></cell><cell role="data">2</cell><cell role="label"><rs type="person" key="IM1">Isabel</rs></cell><cell role="data">1</cell></row>

<row><cell role="label"><rs type="person" key="GG1">Gordon</rs></cell><cell role="data">2</cell><cell role="label"><rs type="person" key="JA1">John</rs> & <rs type="person" key="La2">Lucie</rs></cell><cell role="data">4</cell></row>

ABSTRACTS:

Abstracts for each month of the Graves Diary from 1935-1939 are now available online. These abstracts are not written by the author of the diary, but are provided to assist researchers and general readers by summarizing its contents. Readers accessing the diary at any point can become more familiar with the context of a given entry by scanning the monthly abstract.

The distinctive nature of the abstracts has resulted in some changes in application of the editorial guidelines. The two most significant changes are:

1. When a name or place is mentioned more than once within a single abstract, only the first mention of the name is tagged. Exceptions are made when the abstract is longer, or when omitting a tag may cause confusion for the reader. This approach contrasts with the diary text, where encoded names, places and titles are tagged whenever mentioned.

2. Titles of monographs are encoded with italics, while titles of poems and articles are placed inside quotation marks. This is done consistently, whereas the diary markup of titles varies depending on how they are treated in the manuscript. When Graves underlines a title, it is underlined in the transcription; if he places a poem in quotation marks, it is done likewise; if he omits any highlight, it is omitted in the transcription and markup.

The markup of the diary text, enclosures and abstracts aims to be comprehensive and informative, yet unobtrusive in its presentation. As it is still a work in progress, there remains work to do to make these aims a reality. With regard to presentation in particular, much can be done to minimize the distraction for the reader, and in the final phases of the project, issues like toning down the colour highlights on tagged items will be addressed.