documentation.md

title	subtitle	author	date
Technical Introduction	For the Provisional Transcription of the *Piers Plowman* Text in Takamiya MS 23	Ian Cornelius	October 7 2024

\newfontfamily\myfont{GFS Porson} \newcommand{\myalpha}{\begingroup\myfontα \endgroup} \newcommand{\mybeta}{\begingroup\myfontβ \endgroup} \newpage

Summary overview

The XML file transcription-takamiya-23.xml contains a provisional transcription of the text of Piers Plowman as transmitted in New Haven, Beinecke Library, Takamiya MS 23 (olim London, Sion College, MS Arc. L.40. 2/E), designated sigil 'S' in the textual scholarship. For description of the manuscript and its text, see @CorneliusTakamiyaMS232023, with references. This document serves as a technical introduction to the transcription file and policies that inform it.

The transcription agrees approximately with TEI P5 Guidelines for Electronic Text Encoding and Interchange and the conventions of published documentary editions of The Piers Plowman Electronic Archive (PPEA). See @TEIConsortiumTEIP5Guidelines2024 and @DugganPiersPlowmanElectronic2014a]. These authorities differ in detail; the present documentation does not attempt to reconcile them. Readers are referred to TEI P5 Guidelines for additional commentary on individual elements.

The transcription differs from published PPEA editions in presenting a bare diplomatic transcription, without a textual apparatus or editorial notes of any kind. It also differs from published editions in not being peer reviewed. For bibliographical and licensing details, see [Licensing and Citation].

Document history and responsibility

The initial transcription was produced by Matt Davis, J. Eric Ensley, Jim Knowles, and Timothy L. Stinson in 2014, based on archival-grade TIFF scans of the manuscript, captured in the Beinecke Library's digitization lab in that year. In 2017 responsibility for the transcription passed to Ensley and Ian Cornelius as co-editors, assisted by Paul A. Broyles as technical co-editor. Additional contributions were made by Jesse McDowell and Chelsea Skalak. Cornelius is responsible for documentation.

As co-editors Cornelius and Ensley have checked the transcription in full against the TIFF scans, spot-checked against the manuscript in cases of doubt, and aligned the transcription with articulated policy.

Document structure

The body element of the transcription file supplies a complete transcription of the Piers Plowman text. Divisions of the text into passus are recorded within div1 elements. For instance, the text of the Prologue is wrapped within the tags

<div1 type="passus" n="S.P" org="uniform" sample="complete">
...
</div1>

The sample attribute records that the transcription is complete, that is, comprises the entire text transmitted in the manuscript. This value makes no claim about the completeness of the manuscript.

Divisions of leaf and side are recorded by the element milestone,¹ for instance:

<milestone unit="fol." n="1r"/>

Bibliographical signatures, entered by the scribe to record the sequence of gatherings and leaves, are recorded within fw elements, for instance:

<fw type="sig" place="bottomRight">A2</fw>

Discourse lines (each usually construable as one complete metrical line of alliterative verse) are recorded within l elements and have the attributes xml:id, n, and n_Bx. These attributes have the following semantics:

• xml:id Consecutive numbering of discourse lines within passus, as transmitted by the manuscript. This attribute supplies a unique identifier for units of transcription.
• n Cross-reference to the corresponding line or lines in @KanePiersPlowmanBVersion1988.
• n_Bx Cross-reference to the corresponding line or lines in @BurrowPiersPlowmanBversion2018.²

Accordingly, the first line of Piers Plowman as transmitted by the Takamiya manuscript is encoded as

<l xml:id="S.P.1" n="KD.P.73" n_Bx="Bx.P.73">And cam vp kneling / to kisse
  his bulles</l>

Subsequent line-number references in this documentation are to the values of the attribute xml:id in l elements.

Marginalia are recorded within marginalia elements. The attribute id anchors each marginalia element to the nearest l element. Some marginalia elements supply a transcription of the marginal inscription (e.g., S.15.98.m.1); others only flag the existence of a marginal inscription (e.g., S.15.392.m.1).

Transcription

Character encoding

The transcription file employs UTF-8 character encoding. The scribe's Tironian logograph for English and and Latin et is transcribed by the XML character entity &. The transcription employs these non-ASCII characters:

• Small and capital thorn: þ (U+00fe) and Þ (U+00de)
• Small and capital yogh: ȝ (U+021d) and Ȝ (U+021c)
• Small o with tilde: õ (U+00f5)
• Small i with tilde: ĩ (U+0129)
• Small Greek alpha and beta: \myalpha (U+03b1) and \mybeta (U+03b2)

The characters õ and ĩ are employed in transcription of nomina sacra only: see [Nomina sacra]. The characters \myalpha and \mybeta are employed in cross-references to @KanePiersPlowmanBVersion1988.

The scribe writes an undifferentiated þ/y graph, transcribed as þ or y, depending on the phonetic value of the graph and the spelling conventions in Middle English manuscripts that distinguish these letters.³ The ȝ/z graph is similarly disambiguated.

Allographs are not differentiated in the transcription, except that we aim to record majuscules as capitals. For the letters B, D, H, K, L, V, and W, the scribe has a continuum of forms, running between unambiguous minuscule and unambiguous majuscule. Where these seven letters appear in word-initial position, the choice to transcribe as lower case or capital is often arbitrary.

Large two-lobed "buckled" a is transcribed as capital A in all instances.⁴ Initial double f is transcribed as F. Majuscule I/J is transcribed as I. We follow the scribe's disposition of u and v.

Line division

Where a discourse line runs across two or more text lines, line breaks are recorded with the tag <lb/>: for instance, S.1.32. Overflow onto an adjacent line at the right margin is not tagged. Instead, we record the punctuation-marks entered by the scribe around the textual overflow, usually a double virgula: for instance, S.5.291.

Punctuation

Marks of punctuation are transcribed with the nearest corresponding ASCII character:

• full stop (.) for punctus
• slash (/) for straight sloping virgula suspensiva
• comma (,) for smaller semicircular virgula suspensiva
• question mark (?) for question mark or punctus percontativus⁵

Scribal spacing around punctuation-marks is not retained: punctuation-marks within the line are transcribed with a space-character before and after. Line-ending punctuation-marks are transcribed with a single preceding space-character.

Tagged features

Text segments with distinctive features are enclosed within tags, as follows:

• add encloses text added by the scribe as correction. The location of the added text is recorded with the attribute location. Values are inline and supralinear.
• choice encloses certain scribal abbreviations and their expanded forms, with child elements abbrev and expan. See [Nomina sacra].
• damage encloses text affected by post-production damage to the manuscript, usually cropping of leaves.
• del encloses text deleted by the scribe. Methods of deletion are distinguished with the attribute rend. Values are blotted, linedThrough, overwritten, and razedOut.
• expan encloses alphabetic characters supplied in transcription as expansions of scribal abbreviations. See [Abbreviations].
• foreign encloses text not in English. Language identity is recorded with the attribute xml:lang. Values are LAT (Latin) and FRE (French).
• hi encloses text with graphic features different from the scribe's usual writing. The quality of the difference is recorded with the attribute rend. Values are ul (underlined text), boxed (boxed text), tx (engrossing script), and displayScript (engrossing script).⁶ We assume that the scribe's notional unit of underlining was the word: when, as often, an underline fails to coincide with word boundaries, the partially underlined word is transcribed as underlined in full or not at all, according to our interpretation of the scribe's intentions.
• sic encloses uncorrected scribal slips of the pen.
• unclear encloses unclear or illegible text. Illegible letters are recorded as full stops.

Examples of these tags may be found by searching the XML file.

Abbreviations

General policies on abbreviations

To the extent possible, scribal abbreviations are expanded in accordance with unabbreviated forms elsewhere in the manuscript. Latin text presents few difficulties, but otiose strokes and spelling variation in the English text generate difficulties typical of medieval and early modern forms of this language [see @ParkesEnglishCursiveBook1969, xxix--xxx].

For lexical items with variable spelling, the majority spelling determines the expansion of abbreviated forms. For lexical items always abbreviated by the scribe, expansion is guided by the usual value of a given mark of abbreviation in this manuscript and by common spelling conventions of sixteenth-century English. An example is SERVANT, a word always abbreviated in the scribe's writing but plausibly expanded as

s<expan>er</expan>u<expan>au</expan>nt

Treatment of tittle

Superscripted bar or tittle presents well-known difficulties [@DawsonElizabethanHandwriting150016501966, 19-20; @PettiEnglishLiteraryHands1977, 22]. We treat this mark as an abbreviation only where the resulting expanded form is confirmed in the scribe's unabbreviated spellings. Tittles judged non-abbreviatory are simply ignored in the transcription. This policy has the following general results:

• Tittle is ignored as non-abbreviatory in man, can, gan, son, and similar forms, for double -nn never appears in word-final position in the scribe's unabbreviated spellings.
• Tittle is usually expanded to n in the context of -Vne and -Vne-, where V is a vowel other than u. For instance, sone with tittle is transcribed as so<expan>n</expan>ne and manes with tittle is transcribed as ma<expan>n</expan>nes. Spellings with medial double -nn- regularly occur in this context.
• Tittle is ignored as non-abbreviatory in the context of -une, for such words are never spelled with double n. For instance, DOWN is always spelled doune, whether tittle is present or not.
• Expansion contexts for m differ from those to n. For instance, some with tittle is not expanded to somme, as the unabbreviated spelling somme does not occur.

These patterns of scribal spelling are obviously motivated to avoid minim confusion. That is, the scribe's unabbreviated spellings are themselves constrained, and this observation exposes a logical fault in the treatment of unabbreviated spellings as a neutral 'decoding key' for marks of abbreviation.

An alternative approach would transcribe tittle with a special glyph, as a functional component of the scribe's writing system, in certain cases possibly irreducible to any alphabetic character string. Support for such an approach comes from lexical items in which tittle may serve to disambiguate forms otherwise homographic. An example is supplied by the array of spellings for SOON, SON, and SUN (tittle is recorded by a trailing tilde [~]):

• SOON: sone (21x)
• SON: son~ (28x), sone~ (5x), son (1x), sonne (1x), sonne~ (1x)
• SUN: sonne~ (12x), sonne (8x)

In the five instances of sone~ SON, tittle has diacritic function, distinguishing the base spelling sone from both sone SOON and sonne SUN: that is, the form sone~ is unambiguous only so long as tittle is neither ignored nor expanded. On this evidence, tittle is a functional component of the scribe's orthography, but not a central component.⁷

Nomina sacra

Uniquely in transcription of nomina sacra, we employ a markup syntax that permits representation of both the scribe's brevigraph and a transliterated expansion. This is done with the element choice and its child elements abbrev and expan, for instance:

<choice><abbr>xpĩ</abbr><expan>cristi</expan></choice>

Brevigraphs are transcribed in Roman letters and expanded in agreement with unabbreviated spellings elsewhere in the manuscript. The transcription employs the following value-pairs for abbrev and expan:

• Ihu : Iesu
• Ihus : Iesus
• xp : crist
• xi : cristi
• xpĩ : cristi
• xpm : cristum
• xpõ : cristo
• xps : cristus

Word division

Words divided by the scribe

Contiguous morphs treated by the Oxford English Dictionary (OED) as constituents of a single word, but written with separating spaces by the scribe, are joined by a shadow hyphen in the transcription. For instance:

them<seg type="shadowHyphen">-</seg>self

Words joined by the scribe

Morphs treated by the OED as separate words, but run together by the scribe, are run together in the transcription, preserving scribal word-division. As this policy leaves affected forms unmarked in the transcription, their existence is recorded in this documentation, as follows.

• Negation of a noun or adjective phrase with no. The negative particle no may be run together with the following word: noking (1x), nolesse (1x), noman (31x), nomo (1x), nomore (13x), nowise (2x). In each case the following word is monosyllabic and the combined form is construable as a single accentual group.

• Negation of a verb phrase with not. The negative adverb not may be run together with a preceding verb. Most instances involve an auxiliary or form of be: benot (1x), couldnot (1x), darenot (2x), hadnot (1x), maynot (17x), shallnot (2x), shalnot (1x), shaltnot (1x), willnot (7x), wilnot (7x), woldnot (1x). More rarely, a lexical verb is involved: carenot (1x), lokenot (1x), workethnot (1x), wotnot (2x). The form workethnot is the only item involving a disyllable.

• Auxiliaries followed by be. Shall and will are run together with following be or become: shalbe (31x), wilbe (10x), wilbecome (1x). These combined forms have reduction of double -ll, seen also in shalnot, wilnot, alabout, and welnere. The two-word form shall be does not occur.

• Phrasal verbs. A phrasal verb may be run together with the following adverb: drawnfourth (1x), rapdoune (1x).

• Pronominal subjects. A pronominal subject may be run together with in following verb: theygo (1x), theybeleue (1x).

• Adjectives and quantifiers. An adjective or quantifier may be run together with the head noun: anyman (3x), allmankynde (1x), euilspeche (1x), lefthande (2x), lyfeholymen (1x), poremen (1x), poureman/-men (3x.).

• Prepositions. Prepositions may be run together with the following element: atones (4x), forthat (2x), ina (1x), incase (5x), infeith (1x), onhighe (2x). Several of these could be classified instead a compound adverbs. Forthat functions as a compound conjunction.

• Adverbs. A monosyllabic adverb may be run together with the following element: alaboute (1x), asmuch (3x), aswell (2x), howlong (2x), howoft (1x). somuche (6x), therbe (2x), thusmuche (1x), welnere (1x). The instances with how are conjunctive, introducing an indirect question.

Ambiguous cases

Where the scribe's spacing is inconclusive, permitting interpretation with or without word division, the transcription errs on the side of modern usage, as defined by the OED. Forms affected by this treatment are:

a while (S.18.168), hellewarde (S.18.117), forfare (S.15.139), forsworne (S.19.375), godhede (S.9.47), palfreyes mete (S.19.421), sometyme (S.16.46, S.16.47), tomorowe (S.2.45), westmynstre (S.20.282), without (S.5.633)

Elision and contraction

Scribal elisions and contractions are preserved. For instance: thaungell and tamende.

Licensing and citation

This documentation and the XML file transcription-takamiya-23.xml are released under a Creative Commons Attribution 4.0 International License.

An adequate citation is:

Cornelius, Ian, J. Eric Ensley, and Paul A. Broyles, eds. 2024. New Haven, Beinecke Library, Takamiya MS 23: A Provisional Machine-Readable Transcription of the "Piers Plowman" Text. Version 0 [2024-10-07]. https://github.com/icornelius/s-takamiya-23/.

Bibliography

Footnotes

1. The element milestone also redundantly records divisions of passus, in agreement with current practices of PPEA. See for instance the XML source of @JeffersonPiersPlowmanElectronic2014. ↩

2. An earlier version of @BurrowPiersPlowmanBversion2018 is available online at https://piers.chass.ncsu.edu/texts/Bx. The attribute n_Bx is a customized addition in this transcription of Takamiya 23, unsupported by PPEA validation schemas. ↩

3. This policy is supported by the scribe's care to differentiate forms otherwise homographic. For instance, though the scribe writes a single form for þ and y, ÞOU and YOU are precisely distinguished in the writing system: the scribe always writes YOU in full, without abbreviation, whereas ÞOU is written either with initial th-, as thou, or with the y/þ graph and superscripted u. ↩

4. For this graph see @PettiEnglishLiteraryHands1977 [plates 13, 17, and 24]; and @DawsonElizabethanHandwriting150016501966 [plates 3 and 7]. ↩

5. The scribe's form opens to the right. See @ParkesPauseEffectIntroduction1993 [53 and plates 34 and 35] and @PettiEnglishLiteraryHands1977 [plates 53 and 54]. ↩

6. To the extent that the scribe differentiated two grades of engrossing script, the one designated tx in the transcription is used for Latin, while the one designated displayScript is used for openings of page and passus. Yet the two grades shade into one another: a saner policy would recognize a single engrossing script, avoiding over-fine distinctions. ↩

7. The array of spellings for POOR and POWER supplies another example of a mark of abbreviation serving to distinguish lexical items otherwise homographic. The scribe writes POOR as poore (12x), pore (6x), poor (1x), or pou' with -er/-re abbreviation (64x), whereas POWER is written as poure, always without abbreviation (23x). The sole exceptions to this pattern are S.13.257 and S.14.236, where POWER is written with the -er/-re abbreviation; and S.15.338, where POOR is written as poure, without abbreviation. The usual form of each word is illustrated in S.3.169: pou<expan>re</expan> men maie haue no poure. ↩